Decoding triple negative breast cancer bone metastasis: from 3D bioprinted models to clinical translation

Abstract

This review decodes the biological basis of triple-negative breast cancer (TNBC) bone metastasis and critically examines how 3D bioprinted models can bridge mechanistic investigation and clinical translation. We first synthesize the osteolytic vicious cycle, multicellular microenvironmental crosstalk, and pre-metastatic niche remodeling that drive TNBC bone colonization, and then evaluate how bioprinting-based platforms can recapitulate these processes with greater spatial, mechanical, and cellular fidelity than conventional models. Traditional static research models fail to capture this dynamic spatiotemporal complexity, hindering understanding and treatment development. 3D bioprinting overcomes this by creating a biomimetic microenvironment. It uses hydrogel or composite scaffolds to replicate the bone matrix’s mechanical properties and biochemical gradients. This integrates tumor cells, vascular networks, and bone stromal cells, dynamically modeling metastatic colonization. This bioprinted system serves as a high-efficiency drug screening platform, enabling systematic assessment of nano-drug delivery and strategies to overcome microenvironment-induced resistance. Clinically, 3D printing directly aids in developing patient-specific bone repair implants and surgical guides for reconstructing metastatic lesions. Critically, these models bridge fundamental research and clinical translation. They allow spatiotemporal visualization/analysis of key signaling pathways in metastasis and facilitate high-throughput screening of bone-targeted combination therapies. Rather than presenting a newly established experimental model, this article provides a focused review of current advances, unresolved technical barriers, and future translational directions for TNBC-oriented 3D bioprinting in bone-metastasis research.

Graphical Abstract

Introduction

Breast cancer ranks as the second leading cause of cancer-related mortality among women in the United States. Annually, approximately 240,000 new cases are diagnosed, and nearly 43,000 women succumb to the disease. Notably, the incidence rate of breast cancer has been steadily rising over recent years [78, 268]. Among these subtypes, triple-negative breast cancer (TNBC) constitutes approximately 15–20% of all breast cancer cases, with a high bone metastasis incidence, making it a leading cause of mortality in this aggressive breast cancer subtype [181, 337]. Breast cancer can be categorized into four major subtypes based on pathological classification: ductal carcinoma in situ (DCIS), invasive ductal carcinoma (IDC), lobular carcinoma, and triple-negative breast cancer. These subtypes exhibit distinct immune response profiles [217]. The three key hormone receptors associated with breast cancer are estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER-2). Based on molecular subtyping, breast cancers are categorized into three main groups: hormone receptor-positive (HR⁺⁾, HER-2 positive (HER-2⁺), and triple-negative breast cancer [130, 219]. Currently, the most widely accepted molecular classification system categorizes breast cancer into five intrinsic subtypes: luminal A, luminal B, HER2-enriched, normal-like breast cancer, and basal-like breast cancer, which corresponds to triple-negative breast cancer [272]. Triple-negative breast cancer is a highly heterogeneous disease. Irrespective of histological subtype, TNBC is characterized by elevated genetic instability, a complex genomic landscape, and frequent TP53 mutations [51]. Bone metastasis in triple-negative breast cancer is classified as stage IV cancer and is characterized by aggressive tumor behavior, a high propensity for drug resistance, and an unfavorable prognosis [196, 304]. The 5-year overall survival rate following diagnosis is less than 30%, with a median survival time ranging around 8 to 13 months [133].

Bone represents the most frequent site of metastasis and recurrence in breast cancer. Approximately 65–75% of patients with metastatic breast cancer develop bone metastases. Overall survival (OS) varies according to molecular subtypes and the nature of bone metastatic lesions, typically ranging from 2 to 3 years [42, 304]. TNBC cells exhibiting bone metastasis demonstrate increased malignancy and enhanced stem-like properties compared to non-metastatic TNBC cells [185]. Cancer stem cells (CSCs) establish colonization within the bone microenvironment via the "seed-soil" interaction. By secreting factors such as RANKL, CSCs activate osteoclasts, thereby contributing to a vicious cycle characterized by osteolytic destruction and pathological fractures [362].

While significant advancements have been made in treatment strategies in recent years, the foundational therapeutic approaches for triple-negative breast cancer bone metastasis (TNBC-BM) remain primarily centered on surgical treatment, radiotherapy, chemotherapy, and targeted therapies against the osteoclastic pathway [196, 276]. At present, platinum-based regimens (cisplatin or carboplatin), administered as monotherapy or in combination with taxanes or PARP inhibitors, constitute the mainstay of first-line chemotherapy for TNBC-BM [188, 281]. Radiotherapy primarily prolongs survival by administering high-dose irradiation to isolated bone metastases via stereotactic body radiotherapy (SBRT), either as a standalone treatment or in combination with modalities such as samarium-153-EDTMP radioisotope therapy and immunotherapy [40, 54, 68, 307]. In addition, combinations of targeted therapy and immunotherapy, such as the FDA-approved TrkA inhibitor entrectinib and JAK2 inhibitor pacritinib, demonstrate synergistic effects in inhibiting breast cancer growth and metastasis. The combination of bisphosphonates with denosumab (a RANKL inhibitor), which targets osteoclastic pathways, establishes a "dual bone protection" strategy that effectively reduces the risk of pathological fractures. Furthermore, novel cathepsin K inhibitors specifically block osteoclast activity, thereby improving clinical outcomes [146, 196, 240, 304, 348].

In addition to the standard treatment for bone metastasis in triple-negative breast cancer, adjuvant therapy involves the development of site-specific exercise prescriptions tailored to the location of metastasis—such as pelvic, lumbar, thoracic and rib, or proximal femoral metastases—to alleviate the adverse effects associated with physical and psychological factors, including fatigue and neuropathic pain [100]. However, most patients ultimately encounter complications such as drug resistance and post-radiotherapy bone marrow suppression. Additional clinical challenges in the management of breast cancer bone metastasis include the bone-blood barrier—which is characterized by the chelating effects of collagen and hydroxyapatite on therapeutic agents, as well as the activation of the CXCR4/SDF-1 signaling axis—both of which impede drug penetration. Treatment-induced osteosclerosis also remains a significant concern [42, 89].

Currently, drug development and treatment protocols for breast cancer have reached a relatively advanced stage; however, critical challenges remain unresolved. Only approximately 25% of small-molecule drugs achieve effective delivery to bone metastatic sites, and the low penetration rates of large biomolecules, such as trastuzumab, significantly contribute to the increased incidence of bone metastases [12, 42, 89]. Currently, cell-based or animal models are widely employed to simulate breast cancer bone metastasis. However, these approaches primarily emphasize direct "point-to-point" interactions while overlooking the complex tumor-associated microenvironment and the tripartite crosstalk between tumor cells and various stromal cells, such as osteoblasts and osteoclasts. Furthermore, these models fail to recapitulate the epithelial-mesenchymal transition (EMT) process induced by hypoxic and acidic microenvironments. Although animal models can partially recapitulate bone metastasis, they exhibit species-specific discrepancies that limit their translational relevance—for example, the absence of human IL-8 receptors in murine systems.

3D bioprinting technology has introduced a significant breakthrough in the development of highly precise models of breast cancer bone metastasis and its associated microenvironment. By integrating appropriate biomaterials with various types of living cells to formulate bioinks, researchers can reconstruct not only the tumor itself and its surrounding microenvironment, but also high-fidelity trabecular bone-like structures. To enable real-time monitoring and the application of mechanical forces, microfluidic systems—such as pressure-driven or circulatory mimicry systems—can be incorporated to simulate dynamic bone circulation [25, 93]. Such models can facilitate mechanistic studies, including the dynamic equilibrium between osteolytic and osteoblastic destruction [135, 286, 315], as well as the promotion of osteolysis through tumor cell secretion of cytokines that activate the RANKL pathway in osteoclasts [52, 89, 348, 372]. Pharmacological research encompasses investigations into how physiological alterations—including the breast cancer microenvironment, bone matrix-associated cells, and hypoxic conditions—influence nanodrug penetration efficiency, the efficacy of targeted therapies, mechanisms of drug resistance, and the assessment of drug-induced toxicities.

This review analyzes the pathological mechanisms and therapeutic challenges associated with bone metastasis in breast cancer, emphasizing the high prevalence and clinical complexity of bone metastasis in triple-negative breast cancer. Building upon the elucidation of "vicious cycle" mechanisms driving bone metastasis—such as RANKL-mediated osteolytic destruction and tumor stem cell colonization—as well as the limitations of current therapies, including a mere 25% drug penetration rate and the hindrance posed by the bone-blood barrier, this review synthesizes recent advances in 3D bioprinting models for triple-negative breast cancer (TNBC) bone metastasis. It highlights the transformative potential of 3D bioprinting technology across multiple dimensions. First, in mechanistic analysis, the technology enables the construction of multi-cellular tumor-bone microenvironment interaction models by integrating osteoblasts/osteoclasts, cancer-associated fibroblasts (CAFs), tumor-associated macrophages (TAMs), and other matrix-immune components, thereby decoding complex epigenetic regulatory networks such as the RANKL signaling axis and the WWP1/PTEN-PI3K pathway. Second, from a technological perspective, the use of hydrogel/photo-curable materials allows for cross-scale biomimetic reconstruction of trabecular bone structures with mechanical gradient simulation and vascular networks through microfluidic integration—overcoming the limitations of traditional models in replicating hypoxic microenvironments and dynamic spatiotemporal heterogeneity. Third, in translational applications, 3D bioprinting establishes a closed-loop platform ranging from in vitro drug screening (including evaluation of nanodelivery systems and strategies to overcome drug resistance) to clinical interventions (such as bone defect repair implants and radiotherapy/surgical planning). Finally, in terms of forward-looking integration, the convergence of organoid systems and 4D bioprinting propels TNBC bone metastasis research toward AI-driven, personalized precision medicine paradigms.

Prerequisite conditions for breast cancer bone metastasis: understanding the mechanisms behind metastatic tropism

The precise mechanisms underlying breast cancer bone metastasis remain incompletely understood (Fig. 1). Importantly, Fig. 1 is not intended as a decorative summary but as an organizing framework for the subsequent mechanistic sections. Specifically, it links osteolytic signaling, exosome-mediated niche remodeling, matrix acidification, and secondary dissemination into a continuous pathogenic sequence, thereby clarifying why these processes are discussed in an integrated rather than isolated manner. Recent research has increasingly focused on the bidirectional interaction between tumor cells and the bone microenvironment [372], Within this framework, osteolytic destruction and tumor growth form a self-reinforcing feedback loop, driving the progression of metastatic lesions in a "vicious cycle." Furthermore, the bone microenvironment actively influences the metastatic process, with various bone cell types contributing to the establishment of a metastasis-permissive niche. In addition to the pivotal roles played by bone-resident cell populations, molecular alterations within the surrounding microenvironment also significantly contribute to the metastatic cascade [22, 117, 244, 322, 329].

Fig. 1.

Mechanisms of osteolytic factor secretion in breast cancer metastasis. Illustrates tumor dormancy, integrin α6β1 activation, and CXCL12-CXCL4 axis in lung/liver metastases. Details breast cancer cell adaptations: acidic microenvironment (pH 6.5–6.8), H + pumping, LOXL2 expression, NLRP3/CCL5 release, osteoblast activation, and RANKL upregulation driving bone destruction

The "vicious cycle": central mechanism of tumor cell–bone microenvironment crosstalk

The "vicious cycle" theory of breast cancer bone metastasis serves as a central framework for elucidating the bidirectional crosstalk between tumor cells and the bone microenvironment. At its core, this phenomenon is characterized by a self-sustaining positive feedback loop linking osteolytic destruction with tumor growth. Triple-negative breast cancer cells initiate and sustain a self-reinforcing malignant cycle of bone metastasis through a four-tiered cascade: the "osteolytic factor storm – exosome-mediated communication – microenvironmental acidification – secondary diffusion" pathway. Accumulating evidence demonstrates that this cycle operates within a complex, multi-level regulatory network. The central mechanism involves osteolytic factors activating osteoclasts to secrete tumor-promoting cytokines, thereby enhancing tumor proliferation and establishing a positive feedback loop. Additionally, exosome signaling and acidified microenvironments not only accelerate the metastatic cycle but also contribute to the development of drug resistance. Ultimately, established bone metastatic lesions serve as a hub for systemic dissemination, facilitating further tumor spread throughout the body [22, 89, 110].

Overexpression of osteolytic factors in triple-negative breast cancer

Triple-negative breast cancer cells exhibit overexpression of multiple osteolytic factors that synergistically drive the "vicious cycle" of bone metastasis. Among these factors, parathyroid hormone-related protein (PTHrP) directly promotes osteoclast differentiation through the inhibition of osteoprotegerin (OPG), leading to a disruption of the RANKL/OPG balance [136, 209, 290]. IL-6 and IL-11 enhance RANKL expression through activation of the STAT3 pathway, thereby further amplifying the osteolytic effect [165, 303, 372]. Moreover, during osteolysis, TGF-β is released from the bone matrix and acts in a feedback manner to activate the SMAD signaling pathway in tumor cells, which induces the expression of epithelial-mesenchymal transition (EMT)-associated transcription factors SNAIL and ZEB1, thereby significantly promoting tumor cell invasiveness [172]. The aforementioned three mechanisms collectively form an "osteolytic-invasive-osteolytic" TNBC-specific positive feedback loop, serving as the core driving force underlying the sustained progression of bone metastasis.

Exosome-mediated bone targeting in triple-negative breast cancer

Exosomes derived from triple-negative breast cancer (TNBC) drive bone targeting and osteolytic processes via dual pathways. First, these exosomes carry NLRP3 inflammasomes, which induce dendritic cells in the bone marrow to release the chemokine CCL5, thereby promoting the recruitment of osteoclast precursor cells [113, 137]. On the other hand, TNBC-derived exosomes are enriched in circHER2 (which is upregulated 4.1-fold in TNBC), and this enrichment leads to the activation of the NF-κB signaling pathway in osteoblasts, resulting in increased RANKL expression [156, 341].

These two mechanisms synergistically disrupt bone homeostasis: CCL5 mediates osteoclast recruitment, whereas RANKL promotes osteoclast differentiation and activation, collectively establishing a "malignant soil" that facilitates the colonization of TNBC cells in the bone.

Acidic tumor microenvironment and chemoresistance

The acidic tumor microenvironment and matrix stiffening collaboratively establish a drug resistance barrier in TNBC. Osteoclasts secrete H⁺ to generate a mildly acidic environment (pH 6.5–6.8) [7, 347], and hypoxia induces high expression of LOXL2, which cross-links collagen to harden the matrix and restrict drug penetration [2, 245, 320]. The dual barriers work in concert: the acidic environment weakens drug activity and the hardened matrix limits diffusion, leading to chemotherapy resistance.

Secondary transfer mechanism

Bone metastatic foci function as "metastatic transit stations" through a two-stage mechanism. In the dormancy phase, activation of the CXCL12/CXCR4 axis initiates signaling via the p38 MAPK/NR2F1 pathway, leading to tumor cell dormancy [334]. During the reactivation stage, inflammatory signals (TNF-α/bone damage) trigger integrin α6β1-mediated vascular migration and spread to the lungs/liver [110]. The "dormancy-reactivation-metastasis" cascade positions bone metastases as a central hub for systemic dissemination, thereby significantly deteriorating clinical outcomes.

Multicellular cooperation: interactions between TNBC and the bone microenvironment

The progression of TNBC bone metastasis depends on a metabolic-immune coupled network formed by a tripartite interaction among osteoclasts, adipocytes, and immune cells (Fig. 2). Fig. 2 also serves a second purpose beyond biological illustration: it defines the minimum multicellular components that should be considered when constructing a biomimetic TNBC bone-metastasis model. In this sense, the figure functions as a bridge between mechanistic biology and later model-design discussion. Osteoclasts drive a vicious cycle through immunosuppressive signaling and metabolic support; adipocytes supply energy substrates and contribute to the formation of a chemoresistance barrier; immune cells create an "arginine depletion-adenosine accumulation" immunosuppressive niche. Collectively, these cellular components synergistically confer an evolutionary advantage to TNBC within the bone microenvironment.

Fig. 2.

Triple-negative breast cancer (TNBC) interactions in the bone microenvironment. Osteoclasts secrete CCL2/FFAs and express PD-L1, recruiting M2 macrophages and suppressing CD8 + T cells. Bone marrow adipocytes provide fatty acids for β-oxidation, leptin/JAK2/STAT3 signaling, and ROS neutralization. MDSCs promote T-cell inactivation/depletion via ARG1/CD73

Osteoclast-TNBC crosstalk

Osteoclasts, functioning as an "immune-metabolic hub", promote tumor progression via a tripartite mechanism. Initially, in the context of immune suppression, osteoclasts secrete CCL2 to recruit M2-like macrophages and produce IL-10 and TGF-β, which collectively lead to substantial inhibition of CD8⁺ T cell activity [147, 331, 377]. Secondly, at the level of immune checkpoint blockade, osteoclasts highly express PD-L1, which can directly bind to PD-1 on the surface of T cells, thereby suppressing the anti-tumor immune response [147, 331, 377]. Finally, regarding metabolic support, osteoclasts release free fatty acids through osteolysis, which are subsequently internalized by TNBC cells via CD36. Targeting CD36 has been shown to effectively reduce the metastatic burden by inhibiting fatty acid uptake and disrupting the metabolic symbiosis between osteoclasts and TNBC cells [87, 142].

Bone marrow adipocyte–TNBC metabolic symbiosis

Adipocytes and TNBC cells establish an "energy supply-protection" symbiotic unit. Adipocytes secrete free fatty acids (FFAs), which are internalized into TNBC cells via CD36 and subsequently undergo β-oxidation to generate ATP, thereby serving as an energy source [175, 209] and promoting tumor cell proliferation. Leptin activates the JAK2/STAT3 signaling pathway and upregulates Cyclin D1 expression, which accelerates cell cycle progression [189]. Furthermore, during chemotherapy, adipocytes confer protective effects by releasing glutathione (GSH), which neutralizes reactive oxygen species (ROS) with an approximately 5.7-fold increase in clearance efficiency, thereby enhancing resistance to anthracycline-induced damage [153]. This "energy supply–proliferation promotion–antioxidant" network positions adipocytes as critical metabolic contributors to TNBC bone metastasis.

Central hub of immune evasion in TNBC

Myeloid and lymphoid immune suppression collaboratively establish a "double-lock mechanism" to suppress anti-tumor immunity. In the myeloid lock, myeloid-derived suppressor cells (MDSCs) deplete arginine via arginase 1 (ARG1), leading to mitochondrial dysfunction in CD8⁺ T cells [26, 274]. In the lymphoid lock, regulatory T cells (Tregs) hydrolyze ATP into adenosine, which activates A2A receptors on T cells and induces T cell exhaustion [26]. Double-lock synergy in immune response suppression: Arginine depletion compromises the functional integrity of T cells, whereas adenosine-mediated signaling inhibits T cell activation, leading to a PD-1 monotherapy response rate below 12% and underscoring the critical need for combined targeting of ARG1 and CD73.

TNBC-specific molecular regulation

The molecular network of TNBC bone metastasis is driven by an "acute response–chronic maintenance" dual-axis mechanism: non-coding RNAs (ncRNAs) rapidly regulate osteolytic–osteogenic balance and epithelial-mesenchymal transition (EMT) to initiate metastasis, while epigenetic reprogramming stabilizes the malignant phenotype through DNA methylation, histone modifications, and pathway dysregulation. This spatiotemporally coordinated interplay confers an evolutionary advantage to TNBC within the bone microenvironment.

Non-coding RNA-driven bone metastasis

Non-coding RNAs (ncRNAs) disrupt bone homeostasis via a three-tiered regulatory network. Within the osteoclast activation axis, circMMP2 sequesters miR-128-3p, thereby derepressing PRMT5. This enzyme catalyzes H4R3me2 histone modification, stabilizes β-catenin signaling, and ultimately promotes the activation of osteoclast-related genes [340]. Furthermore, hypermethylation of the miR-34a promoter results in upregulated RANKL expression, thereby promoting the osteolytic process. Within the osteogenic dysregulation axis, miR-141 suppresses DKK1 expression, thereby alleviating Wnt pathway inhibition and promoting aberrant osteogenesis [169]. Regarding the EMT-driving axis, MALAT1 recruits EZH2 to catalyze H3K27me3 modification, leading to CDH1 gene silencing and the induction of epithelial-mesenchymal transition (EMT). The network synergy effect is characterized by two key events: the imbalance between osteoclasts and osteoblasts disrupts bone structural integrity, while enhanced epithelial-mesenchymal transition (EMT) promotes tumor cell invasiveness. Collectively, these processes drive the initiation of the metastatic cascade.

Epigenetic reprogramming

Epigenetic modifications stabilize malignant phenotypes through a three-tiered regulatory axis. Within the DNA methylation axis, TET2 expression is significantly downregulated, amounting to only 23% of normal levels, which results in CDH1 promoter hypermethylation and subsequent disruption of cell adhesion structures. Within the histone modification axis, EZH2 expression is upregulated, catalyzing H3K27me3 modification and suppressing DKK3 expression, thereby activating the Wnt/β-catenin signaling pathway [335]. Within the Hippo signaling axis, YAP/TAZ exhibits marked nuclear accumulation, leading to CTGF expression activation and promoting the establishment of a fibrotic microenvironment. These three interconnected axes cooperatively generate an "epigenetic storm," in which CDH1 silencing activates the Wnt signaling pathway, thereby facilitating CTGF-driven fibrosis. Collectively, these mechanisms enhance tumor colonization within the bone microenvironment and contribute to the emergence of drug resistance.

The pre-metastatic microenvironment in triple-negative breast cancer

Before bone metastasis occurs, TNBC cells initiate a three-tier pre-conditioning program comprising "inflammation-exosome-immunity". In the first stage of inflammatory cascade, tumor cells secrete S100A8/A9 and NETs to recruit immunosuppressive cells, thereby establishing a favorable microenvironment for subsequent metastasis. In the second stage of exosomal reprogramming, TNBC-derived exosomes deliver ncRNAs and key metabolic enzymes that reshape the metabolic landscape of the bone microenvironment. In the final stage of immune locking, tumor cells induce IDO expression, creating a dual barrier mediated by metabolic suppression and immune checkpoint signaling, which effectively suppresses anti-tumor immune responses. Through this coordinated sequence of regulatory events, the bone microenvironment is gradually transformed from an originally "anti-metastatic fortress" into a "pro-metastatic niche".

Inflammatory signaling promotes bone homing in triple-negative breast cancer

The inflammatory cascade drives the homing of TNBC cells to bone through a three-stage temporal coordination. In the pre-conditioning phase, S100A8/A9 proteins bind to TLR4 receptors on bone marrow endothelial cells (dissociation constant Kd = 6.2 nM), thereby activating the NF-κB signaling pathway and inducing G-CSF secretion, which recruits MDSCs and increases vascular permeability [278]. During the capture phase, neutrophil-released NETs form DNA-histone complexes that physically trap circulating TNBC cells. In the final homing phase, NETs further activate the TLR9/STING signaling axis, leading to upregulation of Integrin α6β1 expression and a significant enhancement in bone homing efficiency [324]. The two key processes described above — the microenvironmental pre-conditioning mediated by MDSCs and the tumor cell anchoring effect of NETs — synergistically promote bone metastasis in TNBC, leading to a significantly elevated metastasis rate that exceeds the 19% observed in non-TNBC.

Exosome-mediated microenvironmental remodeling

Exosomes remodel the bone microenvironmental ecosystem through dual-dimensional modifications. At the level of gene regulation, breast cancer-derived exosomes can directly reprogram bone-resident mesenchymal cells and alter osteogenic signaling. A representative example is exosomal miR-940, which promotes osteogenic differentiation of human mesenchymal stem cells and can induce osteoblast-like lesions in vivo, illustrating how tumor exosomal cargo reshapes the bone niche before overt metastatic outgrowth. This concept is also consistent with evidence linking dysregulated DKK1 signaling to breast cancer bone metastasis and abnormal bone remodeling [95, 247, 309]. At the level of metabolic reprogramming, exosomes carry and enrich PKM2 protein (showing increased enrichment), thereby enhancing the glycolytic capacity of recipient cells. This leads to acidification of the tumor microenvironment (pH decline to 6.4), which subsequently induces autophagy and promotes the formation of drug-resistant phenotypes [98]. The two mechanisms synergistically establish a self-sustaining tumor-supportive niche: aberrant osteogenesis provides anchoring sites for tumor cells, whereas the glycolytic pathway constitutes the primary energy source.

Remodeling of the immune microenvironment

The IDO–kynurenine axis contributes to immune evasion in TNBC by depleting local tryptophan, impairing NK-cell antitumor activity, and promoting an immunosuppressive microenvironment through kynurenine-associated signaling. IDO expression is enriched in a subset of high-grade and triple-negative breast cancers and has been discussed as a potential mechanism of resistance to PD-1/PD-L1-directed immunotherapy. Although epacadostat plus pembrolizumab showed acceptable safety and early antitumor activity in early-phase solid-tumor studies, subsequent late-stage trials failed to demonstrate consistent clinical benefit, indicating that IDO-directed strategies should be interpreted cautiously and may require biomarker-guided patient selection [53, 122, 180, 201]. To improve clarity, the key mechanisms, major cell types involved, representative mediators, and potential therapeutic targets discussed in Chapter 2 are summarized in Table 1.

Table 1.

Summary of key mechanisms, major cell types involved, and representative therapeutic targets in TNBC bone metastasis

Key mechanism	Major cell types involved	Representative mediators	Potential therapeutic targets
Osteolytic vicious cycle	TNBC cells, osteoclasts, osteoblasts	PTHrP, RANKL, TGF-β	RANKL, TGF-β signaling
Immune suppression	TAMs, MDSCs, Tregs, CD8 + T cells	IL-10, TGF-β, PD-L1	PD-1/PD-L1 axis; myeloid-targeting strategies
EMT and invasion	TNBC cells, CAFs, endothelial cells	TGF-β, IL-6, β-catenin, SNAIL/SLUG	Wnt/β-catenin pathway; EMT-associated signaling
Acidic and drug-resistant niche	Osteoclasts, adipocytes, ECM components	H + , lactate, LOXL2	pH modulation; ECM remodeling
Dormancy and reactivation	Disseminated tumor cells, stromal niche cells	CXCL12/CXCR4, p38, integrins	Dormancy/reactivation-associated signaling

3D-printed models for studying live cells in bone metastasis of triple-negative breast cancer

A comprehensive understanding of triple-negative breast cancer metastasis to bone requires the analysis of dynamic interactions between tumor cells and multiple host cell types throughout the metastatic cascade. The 3D-printed model developed for this purpose is designed to integrate and recapitulate the key cellular components involved in the complex metastatic process: tumor cells at the primary site or in circulation, along with the immunosuppressive and pro-metastatic microenvironment they establish (1); the vascular barrier that facilitates tumor cell escape and dissemination (2); and the bone microenvironment that governs tumor cell colonization, survival, and destructive growth within bone tissue (3). A detailed investigation of the functional states and interaction networks among these living cell populations—including tumor stem cells, immune cells, stromal cells, endothelial cells, pericytes, and osteocytes—within the model will serve as a critical platform for elucidating the mechanisms underlying bone metastasis and advancing the development of effective therapeutic strategies (Table 2).

Table 2.

Microenvironment-associated cellular and non-cellular components involved in triple-negative breast cancer progression

Category	Representative component	Main role in TNBC	Key consequence
Tumor-intrinsic factors	Exosomes, cancer stem cells	Promote plasticity and intercellular communication	Malignant progression, resistance
Immune cells	TAMs, MDSCs, Tregs, NK cells	Immune suppression or surveillance dysregulation	Immune escape, metastasis
Stromal cells	CAFs, endothelial cells, MSCs, adipocytes	Matrix remodeling, angiogenesis, metabolic support	Colonization, therapy resistance
Other microenvironmental factors	PLA-black phosphorus ECM, SASP, hypoxia	Multifunctional synergy Mechanical, inflammatory, and hypoxic regulation	Tumor promotion and reduced chemosensitivity

Triple-negative breast cancer and key cellular components of Its tumor microenvironment

The high invasiveness and treatment resistance of triple-negative breast cancer (TNBC) are primarily attributed to its unique tumor microenvironment (TME), a dynamic ecosystem composed of tumor cells, immune cells, stromal cells, and non-cellular components. This TME exhibits "cold tumor" characteristics, mainly driven by three interrelated mechanisms: (1) predominant immune suppression, exemplified by the enrichment of tumor-associated macrophages (TAMs) and regulatory T cells (Tregs), which contribute to CD8⁺ T cell exhaustion; (2) active metabolic reprogramming, such as adipocytes supplying lipid substrates and cancer-associated fibroblasts (CAFs) providing metabolic support through the "reverse Warburg effect" [189]; and (3) the formation of dense physical barriers, wherein CAFs generate a protective stromal matrix that limits drug penetration and secrete cytokines such as hepatocyte growth factor (HGF) to activate alternative survival pathways [193]. These mechanisms synergistically drive tumor metastasis and recurrence. Based on the above characteristics, the microenvironment of TNBC has clear clinical translational value. Early markers such as the density of TAMs, the activity of CAFs, and the ratio of CD8⁺/Treg can predict treatment response [56]; targeting TAMs (CSF-1R inhibitors), CAFs (FAP inhibitors), immune checkpoints (PD-1/CTLA-4), and the metabolic microenvironment (fatty acid oxidation inhibitors) are expected to constitute new therapeutic targets [235].

Intrinsic characteristics of tumor cells

Immortalized cell lines recapitulate the biological behaviors of different molecular subtypes of triple-negative breast cancer (TNBC), whereas breast cancer stem cells (BCSCs) serve as central functional entities that mediate key pathological processes, including metastasis, drug resistance, and recurrence.

Immortalized breast cancer cell lines

MDA-MB-231 (mesenchymal-like type) and murine 4T1, due to their high metastatic potential, have become the main models for studying lung/bone/brain metastasis. Among them, MDA-MB-231 is rich in ALDH⁺ breast cancer stem cells (BCSCs), which drive drug resistance and recurrence [194, 225]. Meanwhile, MDA-MB-468 (basal-like type) and BT-549 (mesenchymal-like type) focus on the mechanism of targeted therapy. The former is sensitive to PI3K/mTOR inhibitors due to PTEN deletion and high EGFR expression, while the latter, with PIK3CA mutation and a significant EMT phenotype, serves as a carrier for microenvironment interaction research [46, 235, 363].

BRCA1-mutant cell lines, such as HCC1937, are extensively utilized in studies of PARP inhibitor resistance due to their homologous recombination deficiency (HRD) [83]. Hs578T uniquely recapitulates the inflammatory microenvironment through robust IL-6 secretion, elucidating mechanisms of CAF activation and immune regulation [91]. Notably, species differences must be considered: although 4T1 serves as an excellent metastasis model, its murine origin may constrain its clinical translational relevance [173].

Breast cancer stem cells (BCSCs)

Breast cancer stem cells (BCSCs) sustain their self-renewal capacity through activation of the Wnt/Notch signaling pathway, thereby promoting tumor recurrence and metastasis [371]. Additionally, they exhibit high expression of aldehyde dehydrogenase 1 (ALDH1), which confers resistance to both radiotherapy and chemotherapy [17]. Notably, breast cancer stem cells (BCSCs) exhibit distinct metabolic heterogeneity within the BRCA1/2-mutant population, characterized by enhanced oxidative phosphorylation or glycolysis. This metabolic reprogramming further supports the maintenance of stemness and contributes to therapeutic resistance [83].

Exosomes

Exosomes, as key mediators of intercellular communication within the tumor microenvironment, reshape distant microenvironments—such as the formation of lung pre-metastatic niches—through the delivery of miRNAs and proteins, thereby facilitating breast cancer progression [55, 287]. These vesicles exhibit functional duality: in ER⁺ subtypes, exosomes enriched with miR-143 can modulate the immune microenvironment and potentiate anti-tumor immunity [293]; however, more frequently, they contribute to therapeutic resistance by transferring drug resistance-related proteins (e.g., P-glycoprotein) to surrounding cells, thereby endowing tumor cells with broad-spectrum chemoresistance [250]. This functional versatility positions exosomes as a central regulatory hub in tumor microenvironment modulation.

Immune cells

Within the triple-negative breast cancer (TNBC) tumor microenvironment, immune cells dynamically regulate disease progression through a balance of pro-tumor and anti-tumor functions. Myeloid cells, such as tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs), predominantly drive immune suppression and stromal remodeling, whereas lymphoid cells, including regulatory T cells (Tregs), CD8⁺ T cells, and natural killer (NK) cells, establish a complex antagonistic network involving both activation and inhibition. This intricate interplay defines a hallmark feature of the "cold tumor" microenvironment. The following section will provide a systematic analysis of the functional mechanisms of four key immune cell populations.

Tumor-associated macrophages (TAMs)

Within the triple-negative breast cancer (TNBC) tumor microenvironment, tumor-associated macrophages (TAMs) predominantly adopt an M2-polarized phenotype—a pro-tumorigenic state—characterized by the secretion of immunosuppressive cytokines such as IL-10 and TGF-β, which inhibit T-cell activity. In addition, TAMs promote angiogenesis via VEGF expression and drive metastatic extracellular matrix remodeling, ultimately contributing to poor clinical outcomes [9, 266, 350]. Emerging evidence indicates that TAMs can directly facilitate the malignant progression of breast cancer through ZEB1-mediated transcriptional activation of DNMT1 [164]. Moreover, at metastatic sites, TAMs interact with microglia to establish an immunosuppressive niche that accelerates brain metastasis [203]. Targeting these mechanisms, therapeutic strategies such as the BML-111 inhibitor have demonstrated efficacy in suppressing TAM-induced epithelial-mesenchymal transition (EMT) and metastasis by inhibiting the ILK signaling pathway [168].

Myeloid-derived suppressor cells (MDSCs)

Myeloid-derived suppressor cells (MDSCs) contribute to the establishment of a profound immune-tolerant microenvironment through direct suppression of T cell proliferation and induction of regulatory T cell (Treg) expansion, thereby promoting the progression of triple-negative breast cancer [336]. Their immunosuppressive activity represents a critical barrier to effective anti-tumor immunity and is particularly pronounced in highly invasive TNBC subtypes, where MDSCs are significantly enriched.

Emerging evidence has uncovered a novel mechanism through which myeloid-derived suppressor cells (MDSCs) promote angiogenesis. Specifically, interleukin-35 (IL-35) secreted by MDSCs enhances the ability of tumor-derived exosomes to activate vascular endothelial cells, thereby inducing aberrant angiogenesis via the VEGF/FGF signaling axis [176]. This mechanism expands the functional role of MDSCs beyond immunosuppression to include tumor microenvironment remodeling, establishing a "dual-track tumor-promoting model" that integrates immune suppression with angiogenic activation.

Given the central role of myeloid-derived suppressor cells (MDSCs) in mediating immunosuppression and promoting angiogenesis, therapeutic strategies targeting MDSC activity—such as depletion or functional inhibition—can effectively restore the immunosuppressive tumor microenvironment [267]. When combined with immune checkpoint inhibitors, this approach holds promise for overcoming the therapeutic challenges in triple-negative breast cancer (TNBC).

T cells (Tregs)

The dynamic balance between regulatory T cells (Tregs) and CD8⁺ T cells within the tumor microenvironment plays a critical role in shaping anti-tumor immunity. Tregs actively suppress the function of effector T cells—particularly CD8⁺ T cells—through the secretion of immunosuppressive cytokines such as TGF-β and IL-35, thereby maintaining tumor-associated immune tolerance. Increased Treg infiltration is strongly correlated with resistance to immunotherapy [56, 232]. In contrast, while CD8⁺ T cells possess potent tumor-killing capabilities, their activity is frequently impaired by PD-1/PD-L1-mediated signaling inhibition or direct suppression by Tregs, resulting in functional exhaustion. Clinical evidence indicates that a high ratio of CD8⁺ T cell infiltration capable of overcoming this immunosuppressive barrier is significantly linked to improved patient survival outcomes [158, 342].

The hierarchical regulation of T cells by the tumor microenvironment is critically mediated through cancer-associated fibroblast (CAF)-induced immune polarization. Specifically, CAFs secrete IL-33 to drive Th2-type immune responses, which not only promote the expansion of regulatory T cells (Tregs) but also inhibit anti-tumor Th1 and CD8⁺ T cell responses, ultimately facilitating tumor metastasis [257].

Natural killer cells (NK cells)

Natural killer (NK) cells mediate anti-tumor immunity through direct recognition and lysis of tumor cells. However, within the triple-negative breast cancer (TNBC) microenvironment, their functional activity is markedly impaired. Immunoregulatory factors such as IL-10 and TGF-β secreted by tumor-associated macrophages (TAMs), in combination with hypoxic conditions, significantly reduce NK cell cytotoxicity, thereby compromising immune surveillance and facilitating tumor immune escape [273].

Stroma cells

Stromal cells serve as both structural components and functional regulators within the tumor microenvironment, actively promoting breast cancer progression through dynamic metabolic support, establishment of physical barriers, and formation of metastatic niches. Key stromal cell types include cancer-associated fibroblasts (CAFs), which drive extracellular matrix remodeling and induce immune suppression; adipocytes, which supply lipid-based metabolic substrates and contribute to chemoresistance; endothelial cells, which form vascular networks essential for tumor metastasis and dissemination; and mesenchymal stem cells (MSCs), which play pivotal roles in regulating tumor dormancy and facilitating distant metastatic colonization.

Cancer-associated fibroblasts (CAFs)

Cancer-associated fibroblasts (CAFs) promote the progression of triple-negative breast cancer (TNBC) through three core mechanisms: secretion of TGF-β and IL-6 to induce epithelial-mesenchymal transition (EMT) and enhance metastatic potential; excessive production of collagen and fibronectin, leading to the formation of a dense extracellular matrix (ECM) barrier; and recruitment of tumor-associated macrophages (TAMs) and regulatory T cells (Tregs) to establish an immunosuppressive tumor microenvironment [80, 192, 233]. This newly identified mechanism highlights the profound regulatory influence of CAFs: IL-33 secreted by CAFs promotes Th2 immune polarization and is concurrently activated by Survivin protein present in tumor-derived exosomes. This interaction leads to the upregulation of SOD1, establishing a "tumor-CAFs positive feedback loop" that accelerates malignant progression. Meanwhile, epigenetic modifications such as DNA methylation contribute to the stable maintenance of the tumor-promoting phenotype in CAFs [152, 159, 257]. This multifaceted tumor-promoting capacity positions CAF markers such as α-SMA and FAP as independent prognostic indicators for poor clinical outcomes in TNBC. Furthermore, it underscores their therapeutic significance as "central coordinators within the tumor microenvironment." Targeting CAFs—through inhibition of signaling axes such as IL-6/TGF-β or Survivin-SOD1—can synergistically overcome immune suppression and therapeutic resistance [363].

Adipocytes

Adipocytes within the breast cancer microenvironment contribute to tumor progression through two distinct mechanisms: first, they directly supply fatty acids to support tumor energy metabolism and secrete leptin and IL-6, which promote cancer cell proliferation [91, 189]; second, they induce chemotherapy resistance via lipid metabolic reprogramming, particularly intensifying the immunosuppressive microenvironment in obese conditions [101, 206]. Recent studies have elucidated a critical novel mechanism whereby adipose-derived stem cells (ADSCs) secrete CXCL5 in response to resistin stimulation, thereby significantly enhancing breast cancer invasiveness and metastatic potential ([321]). This finding highlights the functional shift of adipocytes from passive metabolic supporters to active regulators of malignant progression.

Endothelial cells

Endothelial cells facilitate tumor growth, invasion, and distant metastasis in breast cancer by promoting angiogenesis, which ensures a continuous supply of oxygen and essential nutrients to the tumor microenvironment [328].

Mesenchymal stem cells (MSCs)

Mesenchymal stem cells (MSCs) modulate breast cancer metastatic dormancy through exosome secretion. These exosomes carry specific miRNAs, such as miR-222 and miR-223, which induce quiescence in breast cancer cells and promote chemoresistance. While this dormant state may persist for several years, cellular reactivation can be triggered by factors such as aging or inflammatory signals, ultimately leading to late metastatic recurrence, including bone metastasis [18]. Furthermore, soluble signals derived from breast tumors, in combination with mechanical stress within the bone microenvironment, synergistically drive the osteogenic differentiation of MSCs [131], enhance metastatic niche formation, and provide a mechanistic explanation for the high clinical incidence of bone metastasis.

Other related components

In summary, the aggressive behavior and unfavorable prognosis of triple-negative breast cancer (TNBC) are intimately linked to its distinct tumor microenvironment (TME). Dynamic remodeling of the extracellular matrix (ECM) not only offers structural support but also actively promotes tumor cell infiltration and metastasis by activating key signaling pathways, such as integrin-FAK, through its compositional elements (e.g., hyaluronic acid secreted by cancer-associated fibroblasts) and physical properties (e.g., matrix stiffness gradients). Additionally, the ECM’s physicochemical state, such as acidification, further facilitates the invasive potential of TNBC cells. The senescence-associated secretory phenotype (SASP) refers to the persistent release of pro-inflammatory factors, such as IL-6 and MMPs, by senescent cells within the tumor microenvironment. These secreted molecules contribute to the establishment of a chronically inflamed milieu, thereby creating a favorable environment for tumor progression. Furthermore, hypoxia, a hallmark of solid tumors and especially pronounced in triple-negative breast cancer (TNBC), facilitates immune evasion through the activation of immunosuppressive pathways such as adenosine signaling. Hypoxia also markedly diminishes the efficacy of various therapeutic interventions, including radiotherapy, thus representing a critical contributor to treatment resistance. Therefore, a comprehensive understanding of the interactive mechanisms among the extracellular matrix (ECM), senescence-associated secretory phenotype (SASP), and hypoxia is crucial for elucidating the pathogenesis of triple-negative breast cancer (TNBC) and advancing more effective targeted therapeutic strategies.

Extracellular matrix (ECM)

The extracellular matrix (ECM) plays a pivotal role in tumor progression, with diverse functions and complex mechanisms of action. Specifically, ECM components such as hyaluronic acid, which are secreted by cancer-associated fibroblasts (CAFs), significantly enhance tumor cell infiltration [316]. On the other hand, the mechanical signals provided by ECM, especially the increase in matrix stiffness, can drive tumor metastasis by activating the integrin-FAK signaling pathway; the latest research further reveals that the stiffness gradient of the matrix itself can precisely regulate the metastatic behavior of tumor cells through the integrin-FAK pathway [174]. Furthermore, research utilizing 3D models has demonstrated that acid–base transporters within the extracellular matrix (ECM) contribute to tumor cell invasion by sustaining the acidic state of the tumor microenvironment (TME) [45]. Therefore, the extracellular matrix (ECM) not only serves as a structural scaffold but also actively contributes to the formation of a tumor-promoting microenvironment through its biochemical components, mechanical properties, and physicochemical conditions, such as pH.

Senescence-associated secretory phenotype (SASP)

The senescence-associated secretory phenotype (SASP) is a hallmark of senescent cells, characterized by the secretion of a variety of bioactive molecules, and plays a critical role in shaping pathological environments. These factors secreted by senescent cells, such as interleukin-6 (IL-6) and matrix metalloproteinases (MMPs), can significantly promote a chronic inflammatory state in the tissue microenvironment and thereby drive the occurrence and development of tumors [119]. Therefore, the senescence-associated secretory phenotype (SASP) serves as a critical molecular link between cellular senescence and tumor progression.

Hypoxic microenvironment

The hypoxic tumor microenvironment serves as a critical driver of malignant progression and therapeutic resistance. Emerging evidence has uncovered a key underlying mechanism: hypoxic conditions trigger the activation of the adenosine signaling pathway, thereby facilitating tumor cell immune escape and allowing these cells to evade elimination by the host immune system [10]. In addition, this hypoxic environment has been shown to markedly decrease tumor sensitivity to radiotherapy, representing a major contributor to radiotherapy resistance [79]. Therefore, the hypoxic microenvironment exerts significant effects on tumor progression and therapeutic outcomes through the activation of specific signaling pathways and the reduction of treatment responsiveness.

Blood vessels

The core cell types directly involved in the tumor metastasis process within blood vessels include endothelial cells, pericytes, and tumor cells themselves. Endothelial cells constitute the inner lining of blood vessels and represent a critical barrier with which tumor cells must interact in order to intravasate into or extravasate from the bloodstream. The expression of endothelial cell markers and their functional status directly influence the efficiency of metastasis. Pericyte dysfunction results in abnormal tumor vascular architecture and increased vascular leakage, not only facilitating tumor cell intravasation into circulation, but also collaboratively shaping a hypoxic and inflammatory microenvironment, which collectively promotes metastatic progression. Although the behaviors of red and white blood cells within blood vessels, such as their effects on hemodynamics and interactions with tumor cells, have been investigated, current evidence suggests that endothelial cells, pericytes, and tumor cells exert the most direct and central roles in metastatic mechanisms involving vascular structure and function.

Endothelial cells

According to multiple key studies, endothelial cells (ECs) play a central role in tumor metastasis, with their functional status and molecular characteristics directly influencing metastatic efficiency. Luders and Schumacher identified S1PR1 as a specific marker for endothelial cells, offering a crucial tool for investigating tumor angiogenesis and metastasis [184]. Kuczynski et al. reported that tumor cells hijack host blood vessels through a mechanism known as "vascular co-option," which relies on direct interactions between tumor cells and endothelial cells and represents a key process in metastasis and drug resistance [144, 368]. Cuypers et al. further developed a method to isolate endothelial cells from vascular co-option models, providing valuable methodological support for the analysis of the metastatic microenvironment [44]. Moreover, Hoffmann et al. demonstrated that targeting endothelial cell dysfunction, such as the Snai1-mediated pathway, can enhance drug delivery and presents a novel therapeutic strategy for metastasis [102]. In summary, endothelial cells not only serve as critical mediators enabling tumor cells to traverse the vascular barrier during metastasis, but their molecular profiles and functional regulation also represent central targets for metastasis research and therapeutic intervention.

Pericytes

Pericytes actively contribute to the formation of the metastatic microenvironment through functional abnormalities. Meng et al. demonstrated that hexokinase 2 (HK2)-driven glycolytic dysregulation in tumor-associated pericytes enhances cell contractility, resulting in vascular structural distortion and impaired blood perfusion. This pathological alteration not only compromises vascular integrity but also generates a microenvironment characterized by concurrent hypoxia and inflammation, thereby establishing a critical pathological foundation for tumor cell invasion and metastasis [198].

Bone

Bone is the most common distant metastasis target organ for solid tumors (such as breast cancer and prostate cancer) and hematological malignancies (such as multiple myeloma). Its unique microenvironment, rich in mineralized matrix, growth factors (TGF-β, IGF-1), and a dynamic remodeling cellular network, provides an "ecological niche" for tumor cells, driving a self-perpetuating cycle of "metastasis-bone destruction-re-metastasis" [22, 285, 322]. This process not only entails tumor cells hijacking bone homeostasis, but also depends on the coordinated interaction among bone matrix cells, immune cells, and signaling pathways, ultimately resulting in bone-related events (SREs) such as pathological fractures and hypercalcemia.

Bone microenvironment cells

The pathological basis of bone metastasis is closely related to the dynamic balance of various cells in the bone microenvironment. Basic research demonstrates that the high proliferative capacity of bone stem cells, such as the Cd168⁺ cell population in long bones [92], may serve as a cellular source for tumor colonization. The functional interplay between osteoblasts and osteoclasts forms the core pathological framework: abnormal activation of osteoclasts causes osteolytic destruction and the subsequent release of growth factors (e.g., TGF-β, IGF-1), which promote tumor proliferation. Meanwhile, excessive bone formation by osteoblasts leads to osteoblastic lesions. Together, these processes synergistically drive the "vicious cycle" of bone metastasis [35]. The mechanosensitive ion channel Piezo1 regulates bone homeostasis through the perception of mechanical signals. Dysregulation of Piezo1 has been shown to accelerate tumor-induced osteolysis [312], while recent advances in quantitative techniques for assessing osteoclast multinucleation have enhanced the precision of bone resorption monitoring [308]. Furthermore, regulation of the immune microenvironment plays a critical role in metastatic progression: CD8⁺ T cells enriched in vertebral bone marrow can either suppress or enhance metastasis, depending on the balance of their immune phenotypes [74]; radiation-induced bone marrow adipogenesis and cellular senescence contribute to the formation of a pre-metastatic niche that supports tumor cell survival through the secretion of inflammatory factors, including IL-6 and PGE2 [33]. Tumor cells acquire "malignant adaptation" through the hijacking of bone developmental processes, such as Hedgehog-mediated cartilage-bone crosstalk [99], and repair-associated signaling pathways, including Wnt/β-catenin signaling [218], ultimately leading to the disruption of bone homeostasis.

Tumor-bone interaction

Numerous studies have extensively investigated the cellular mechanisms underlying cancer-induced bone metastasis and potential clinical intervention strategies. Breast cancer bone metastasis is sustained by reciprocal signaling between disseminated tumor cells and the bone microenvironment, particularly osteoclasts, osteoblasts, and stromal components. AXL signaling has emerged as one important pathway in this process, because inhibition of AXL impairs breast and prostate cancer metastasis to bone and reduces bone remodeling. In parallel, osteoclast-centered crosstalk remains a major therapeutic node: osteoclast activation promotes osteolytic progression, whereas antiresorptive strategies such as bisphosphonates and denosumab are used clinically to reduce skeletal complications by limiting tumor-driven bone destruction. Taken together, tumor–bone interaction in breast cancer is best understood as a reciprocal and disease-specific process that couples metastatic colonization with remodeling of the bone niche [22, 285, 353].

3D Bioprinting: technological innovations and advantages in comparison with conventional research models

Over the past several years, researchers have predominantly utilized cell-based assays and murine models to investigate disease mechanisms and screen potential therapeutic agents (Fig. 3). Fig. 3 is intended to translate the preceding biological discussion into experimental-design logic. By juxtaposing conventional assays with 3D bioprinted and hybrid platforms, it illustrates why increasing biological complexity is necessary for questions involving stromal interaction, vascular extravasation, and bone-specific therapeutic response.

Fig. 3.

Methodologies for breast cancer research. Includes gene editing, drug screening, 3D bioprinted organoids/mouse models, microfluidic chips, histology, AI/automation, in vivo validation, pharmacotoxicity testing, and mechanistic probing for studying tumor biology and therapeutic development

Research categories in conventional methodologies

Traditional research directions can be broadly categorized into two main streams: mechanistic studies ranging from molecular target identification to signal pathway validation, and drug development processes encompassing in vitro primary screening, in vivo pharmacodynamic evaluation, and therapeutic efficacy assessment. The overarching framework is outlined as follows.

Molecular target identification and signal pathway validation

In tumor mechanistic studies, cell-based experiments generally proceed through a multi-tiered workflow, beginning with target screening and culminating in functional validation. Initially, key regulatory genes (such as Atg4B or SMAD4) are identified using techniques like transcriptome sequencing or CRISPR library screening. Subsequently, the expression levels of the corresponding proteins in tumor cells, as well as their correlations with specific phenotypic traits—such as autophagic activity or invasive capacity—are validated through Western blot (WB) analysis [357, 364]. Subsequently, loss-of-function assays were carried out using siRNA-mediated knockdown or CRISPR/Cas9 knockout approaches to assess the impact of target genes on cellular behaviors, including proliferation, apoptosis, and migration, as evaluated by flow cytometry, transwell assays, and wound healing assays. Concurrently, gain-of-function experiments were conducted to validate the regulatory effects of these genes on drug sensitivity—such as the reversal of paclitaxel resistance. Building upon these findings, molecular mechanisms were further elucidated through techniques such as co-immunoprecipitation (Co-IP), fluorescence resonance energy transfer (FRET), and dual-luciferase reporter assays. For example, it was demonstrated that Luteolin suppresses TNBC metastasis by reversing EMT via downregulation of β-catenin expression [186], while Gypensapogenin H regulates the PI3K/AKT/NF-κB/MMP-9 pathway [167, 284]. This sequential process, progressing from molecular discovery to functional validation, ultimately elucidates the precise roles of genes within signaling pathways, thereby providing a solid experimental foundation for targeted therapeutic strategies.

Taking the investigation of the TGFβ pathway as a case study, the experimental rationale can be further elaborated: following the identification candidate mediators via RNA sequencing, their clinical relevance can be further evaluated using population-based prognosis datasets such as SEER [370]. Furthermore, gene-editing approaches were employed to suppress SMAD4 expression, resulting in a great decrease in cell invasive capacity [167]. Finally, its transcriptional activation of MMP-9 was confirmed using a dual-luciferase reporter assay [284]. Ultimately, the formation mechanism of the SMAD4/β-catenin complex was revealed through Co-IP, and a complete causal chain of "gene expression—functional phenotype—molecular interaction" was constructed [186]. This systematic approach takes into account both molecular mechanisms and functional phenotypes, laying the foundation for subsequent animal experiments and drug development.

Following the completion of cell-based experiments, animal studies can be initiated using murine models to systematically validate the functional roles of target genes through orthotopic transplantation and genetic engineering approaches. For instance, orthotopic implantation of Atg4A-knockout triple-negative breast cancer (TNBC) cells into the mice resulted in a significant reduction in tumor growth rate and diminished metastatic potential [38, 327]. When combined with in vivo imaging modalities—such as the LC3 bioluminescence reporter system—dynamic alterations in autophagy pathway activity can be longitudinally monitored, thereby elucidating the spatiotemporal patterns of autophagy suppression within the tumor microenvironment triggered by NF-κB pathway [339]. Conversely, the establishment of an orthotopic metastasis model using SMAD4-overexpressing triple-negative breast cancer (TNBC) cells demonstrated a threefold increase in lung metastatic nodules through in vivo imaging, thereby directly validating the pro-metastatic function of SMAD4 [179, 279, 332]. Moreover, inoculation of TNBC cells into TGFβ conditional knockout mice followed by micro-CT quantification revealed a significant reduction in bone destruction, further confirming the critical role of TGFβ signaling in shaping the bone metastasis microenvironment [345]. These experiments validate target functionality at the whole-animal level by recapitulating clinical pathological processes—such as orthotopic tumor growth and metastatic colonization—while incorporating advanced imaging technologies, thereby establishing a robust in vivo evidence framework for mechanistic analysis and optimization of therapeutic strategies.

In Vitro drug screening and In Vivo pharmacodynamic evaluation

In drug development, in vitro cell experiments represent a critical starting point for initial compound screening and mechanistic investigation. For instance, the development of the microtubule inhibitor VERU-111 commenced with high-throughput screening, wherein the half-maximal inhibitory concentration (IC50) was measured in two-dimensional (2D) cultured triple-negative breast cancer (TNBC) cell lines. Synergy index analysis further revealed that the combination of VERU-111 with paclitaxel significantly enhanced the anti-tumor efficacy, as indicated by a synergy index value below 1 [50]. To further elucidate the mechanism of action, flow cytometry was utilized to evaluate the drug’s effects on cell cycle progression. Results demonstrated that VERU-111 induced cell cycle arrest at the G2/M phase in 80% of the cells, which serves as a hallmark of microtubule destabilization. Furthermore, Annexin V/PI double staining revealed an increase in the apoptosis rate to 45% [50]. In the investigation of chemotherapy resistance, metabolomics analysis demonstrated that reactive oxygen species (ROS) levels in doxorubicin-resistant triple-negative breast cancer (TNBC) cells were reduced by 60% compared to those in sensitive cells. CRISPR-mediated knockout of TXNIP, a key regulator of oxidative stress, resulted in a 50% decrease in DNA damage as revealed by comet assay, along with a marked enhancement of drug resistance. These findings highlight the critical role of the ROS-TXNIP axis in modulating chemoresistance [36]. Based on this mechanistic insight, Halofuginone (HAL) was identified through large-scale compound screening of 2,000 candidate molecules, demonstrating its ability to reduce the IC50 of doxorubicin from 10 μM to 2 μM [378], This finding offers a novel therapeutic strategy for overcoming drug resistance.

This finding offers a novel therapeutic strategy for overcoming drug resistance. To validate VERU-111, a subcutaneous xenograft model of paclitaxel-resistant triple-negative breast cancer (TNBC) was developed. In an orthotopic TNBC xenograft model, oral VERU-111 inhibited tumor growth dose-dependently and reduced metastatic burden [50]. Pharmacokinetic analysis using HPLC detection demonstrated that the compound achieved a peak plasma concentration (Cmax) of 12.3 μM and exhibited a half-life (t1/2) of 8.2 h, indicating favorable in vivo stability [13]. To optimize HAL, a doxorubicin-resistant model was established through continuous intraperitoneal injections, resulting in a five-fold increase in the resistance index [238]. The combination therapy of HAL@TPGS micelles and doxorubicin was then administered. Micro-PET/CT imaging revealed reduction in primary tumor volume and decrease a number of bone metastases. In terms of safety, serum ALT and AST levels remained within the normal range (ALT < 40 U/L, AST < 35 U/L), and H&E staining revealed no pathological damage in heart, liver, or kidney tissues [85], further supporting the drug’s low toxicity profile.

This integrated in vitro-in vivo strategy not only validates drug efficacy using multi-level models (2D cells, organoids, and murine models), but also incorporates mechanistic analysis and safety evaluation. For example, the development pipeline of HAL@TPGS encompasses the following stages: "metabolomics-based identification of drug resistance mechanisms → high-throughput screening for candidate compounds → organoid-based assessment of penetration capacity → murine model evaluation of combination efficacy," highlighting the synergistic advantages of multidisciplinary technologies. Moreover, in the study of VERU-111, a strong correlation between in vitro flow cytometry data and in vivo tumor inhibition rates was observed, further confirming the predictive power of cellular assays for in vivo efficacy [50]. This systematic approach markedly improves the efficiency and clinical translatability of drug development.

The complete logical framework of conventional experimental approaches

The pathological progression of breast cancer liver metastasis critically relies on the dynamic crosstalk between tumor cells and their microenvironment. Notably, integrin-mediated adhesion signaling and PI3K/AKT pathway-driven survival signals represent key regulatory hubs in this process. Conventional experimental designs must systematically evaluate whether targeting these two signaling axes can effectively inhibit metastatic niche formation by blocking both the survival of circulating tumor cells and their adaptation to the hepatic microenvironment during the metastatic cascade.

Targeting β3 integrin

First, single-cell sequencing was employed to comprehensively characterize the tumor microenvironment of triple-negative breast cancer (TNBC). This analysis revealed a marked upregulation of β3 integrin expression in metastatic tumor cells—approximately fourfold higher than in non-metastatic counterparts—suggesting its potential as a therapeutic target [65]. To validate its functional role, β3 integrin was silenced using siRNA, and MTT assays demonstrated a significant increase in TNBC cell sensitivity to cisplatin, with the IC50 value decreasing from 10 μM to 5 μM. These findings indicate that β3 integrin depletion can effectively reverse chemoresistance. Furthermore, co-immunoprecipitation (Co-IP) assays confirmed a direct interaction between β3 integrin and focal adhesion kinase (FAK), leading to activation of the downstream SRC/ERK signaling pathway. A 60% reduction in phosphorylated FAK levels was observed, elucidating the central mechanism through which β3 integrin promotes cell survival and drug resistance via FAK signaling.

To recapitulate the clinical progression of bone metastasis, an in vivo mouse model was established. TNBC cells with β3 integrin overexpression were orthotopically injected into the bone marrow cavity of the murine femur. Bioluminescent imaging revealed a threefold increase in the number of bone metastatic lesions, which was accompanied by pronounced osteolysis. Building on these findings, the β3 integrin inhibitor Cilengitide (20 mg/kg, intraperitoneal injection) was administered in combination with cisplatin over a 6-week treatment period. The results demonstrated a 55% reduction in tumor volume in the combination group, with the median survival time prolonged from 28 to 42 days, representing a significant improvement compared to the monotherapy group. RNA-seq analysis of tumor tissues further demonstrated that the combined treatment significantly downregulated the expression of FAK pathway-associated genes, including SRC and MMP2, by approximately two-fold. Additionally, immune microenvironment remodeling led to an increase in CD8⁺ T cell, confirming that targeting β3 integrin enhances therapeutic efficacy through a dual mechanism involving "direct tumor suppression and immune activation [65]".

Targeting the PI3K/AKT pathway

Phosphorylated protein array screening of abnormally activated signaling pathways in TNBC revealed a marked elevation in AKT (Ser473) phosphorylation—approximately threefold higher than in normal tissues—leading to the identification of the PI3K/AKT pathway as the core regulatory target [284]. Treatment of TNBC cells with the PI3K inhibitor Gypensapogenin H resulted in a 70% reduction in p-AKT expression, as demonstrated by Western blot analysis, along with a threefold increase in the apoptotic marker Cleaved Caspase-3, confirming that inhibition of this pathway effectively induces apoptosis. Further chromatin immunoprecipitation sequencing (ChIP-seq) analysis demonstrated that NF-κB directly binds to the promoter region of MMP-9, with an enrichment fold greater than 5, thereby activating MMP-9 transcription and promoting basement membrane degradation. These findings reveal the molecular mechanism through which AKT drives tumor metastasis via the NF-κB/MMP-9 signaling axis. The aforementioned conclusion was validated in mouse models. In the TNBC orthotopic transplantation model, daily administration of Gypenoside H (50 mg/kg) for four weeks resulted in a reduction in primary tumor volume and decrease in the number of lung metastases, as assessed by MRI [284].

The Improvement from traditional experiments to joint strategies

Traditional cell-based assays and mouse models remain fundamental pillars in basic research and drug development. Cellular experiments enable target identification through approaches such as gene editing, signaling pathway analysis, and multi-omics technologies—including single-cell sequencing and phosphorylated protein microarrays. When integrated with gene editing tools like CRISPR/siRNA, these methods facilitate rapid target validation and preliminary drug screening. However, these in vitro systems exhibit limited biomimetic fidelity and are incapable of recapitulating systemic physiological regulation, such as whole-body immune modulation. In subsequent mouse experiments, orthotopic transplantation, humanized models, and in vivo imaging techniques were employed to dynamically validate target functionality, modulate the immune tumor microenvironment, and assess both therapeutic efficacy and systemic toxicity—thereby addressing key limitations of cell-based assays. Nevertheless, these models are constrained by interspecies physiological differences and high experimental costs.

From the perspective of logical integrity, 3D bioprinting of organoids is integrated into the experimental framework. While 3D printing technology overcomes key limitations of traditional models, challenges such as standardization, cost-effectiveness, and long-term stability remain to be resolved. The experimental paradigm has evolved from a binary progression of "cell experiments → animal experiments" to a more refined and mechanistic workflow: "cell target discovery (mechanistic exploration) → organoid-based efficacy prediction (in vitro assessment) → humanized mouse validation (in vivo verification)". When integrated with single-cell multi-omics profiling, PDX/organoid platforms, or artificial intelligence (AI) tools, this strategy establishes a comprehensive research continuum spanning molecular mechanisms to systemic outcomes [86, 88].

Integrated experimental strategy: streamlined model development and enhanced preclinical predictability

The core strategy of multi-technology integration involves the synergistic combination of multi-scale models. 3D bioprinting is employed to fabricate biomimetic tumor or organ models, which are subsequently integrated with organoid cultures and microfluidic chip systems to simulate dynamic microenvironmental conditions—such as drug concentration gradients and fluid shear stress [127, 138]. Furthermore, computational data analysis can be applied to refine printing parameters and enhance experimental design [212]. This integrated approach establishes a translational strategy bridging in vitro and in vivo systems. Initially, drug screening is conducted using 3D bioprinted models, followed by validation of systemic efficacy in mouse models. Moreover, single-cell sequencing is applied to characterize cellular heterogeneity within the 3D models, with findings cross-validated by comparison to tumor samples derived from mouse models [8],[318].

Simplification of model construction represents a critical strategy to expedite hypothesis validation in biomedical research. In the context of invasion and metastasis mechanisms of triple-negative breast cancer (TNBC), cell-based assays have efficiently identified key molecular targets through rapid phenotypic screening. For example, Transwell assays demonstrated that exosomal EphA2 enhances the invasive capacity of TNBC cells by approximately two-fold [179], whereas scratch wound assays revealed that the TGF-β inhibitor Pirfenidone reduces cell migration rates by 60% [186]. To better recapitulate clinical heterogeneity, patient-derived xenograft (PDX)-derived organoids (PDXO) were established from clinical biopsy specimens, and the therapeutic efficacy of Celecoxib in combination with paclitaxel was evaluated. The results demonstrated that the combined treatment regimen elevated the organoid apoptosis rate increase [88]. In the investigation of the bone metastasis microenvironment, bone-in-culture arrays were established by co-culturing mouse tibia slices with TNBC cells, and TRAP staining revealed a two-fold increase in osteoclast activity [311]. Meanwhile, incorporation of silica nanoparticles into 3D-printed bone scaffolds led to a three-fold upregulation of the osteogenic marker RUNX2 as confirmed by qPCR, effectively recapitulating the tumor-driven bone remodeling process [345]. Furthermore, using a macrophage co-culture model, it was demonstrated that M2-type macrophages induced a 50% upregulation of PD-L1 expression in TNBC spheroids, which led to a 40% reduction in the cytotoxic activity of CD8⁺ T cells, thereby elucidating the microenvironmental mechanism underlying immune escape [154].

Mouse experiments employing complex pathological models have significantly advanced preclinical translation. In the context of bone metastasis research, micro-CT quantification revealed that injection of TNBC cells into the bone marrow cavity of the mouse tibia resulted in a 55% reduction in bone destruction following treatment with dual-targeted nanoparticles (Pluronic P123) [70]. Moreover, in a humanized bone metastasis model established by implanting human demineralized bone matrix (DBM), in vivo imaging demonstrated that the TGFβ inhibitor YH677 delayed the onset of bone metastasis by three weeks (Zhang, Y. et al., 2023). In humanized mice engrafted with human CD34⁺ hematopoietic stem cells, combination therapy with anti-PD-L1 antibody and chemotherapy increased the proportion of tumor-infiltrating CD8⁺ T cells from 5 to 25%, leading to significant tumor growth inhibition [11]. Furthermore, adoptive transfer of Vγ9Vδ2 T cells in combination with zoledronic acid resulted in a 70% reduction in bone resorption area and a four-fold increase in CD8⁺ T cell infiltration [379], providing in vivo evidence supporting the efficacy of combined immunotherapeutic strategies.

These multi-level models, ranging from 2D functional validation to 3D microenvironmental simulation, and from murine pathological models to humanized systems, establish a closed-loop framework for translational research. The organoid platform accelerates the workflow from "patient sample acquisition to drug testing," while humanized models effectively address the limitations associated with traditional mouse models, particularly their immunological deficiencies. For instance, the patient-derived xenograft organoid (PDXO) model directly captures patient-specific drug sensitivity profiles, thereby reducing the empirical bias in preclinical research. The integration of bone-in-culture arrays with 3D bioprinting technology enables comprehensive analysis spanning molecular mechanisms to tissue-level pathology. This "simplified in vitro model → complex in vivo validation" paradigm significantly improves the translational efficiency from fundamental discoveries to clinical applications.

Integrated experimental approaches: technical advantages, current limitations, and future perspectives

From a technical perspective, the performance of 3D bioprinted TNBC bone-metastasis models depends on four closely related parameters.

Bioink selection

Bioink selection is not merely a material issue but a determinant of biological interpretability. For TNBC bone-metastasis models, the ideal bioink should simultaneously support print fidelity, tumor-cell viability, stromal-cell compatibility, and bone-mimetic signaling. Natural matrices such as collagen, gelatin, and decellularized extracellular matrix improve cell adhesion and phenotype preservation but often lack structural robustness, whereas synthetic polymers offer better printability and mechanical control at the cost of lower biological activity. Therefore, hybrid bioinks are often preferable because they permit simultaneous tuning of shape fidelity and tumor–stroma functionality [75, 291]. In practical design, the choice of bioink should be matched to the intended biological question—for example, whether the priority is matrix-mediated signaling, vascular integration, long-term culture stability, or drug-response testing.

Mechanical tuning and stiffness matching.

Mechanical tuning is particularly important in TNBC bone-metastasis bioprinting because tumor cells encounter a sharp mechanical transition from relatively soft breast tissue to the stiffer trabecular-bone niche. This difference is not merely structural; matrix stiffness can alter adhesion, invasion, osteolytic signaling, and drug response. For this reason, future model design should avoid reporting composition alone and instead include quantitative stiffness characterization, ideally accompanied by an explanation of how the selected range relates to the intended biological question [261, 354]. In addition, stiffness should be considered together with matrix architecture and degradation behavior, because constructs with similar nominal composition may still produce different biological outcomes if pore structure, crosslinking density, or long-term mechanical stability differ.

Vascularization strategy.

Vascularization should also be treated as a design variable rather than a generic enhancement feature. The degree of vascularization depends not only on the fabrication method—such as sacrificial templating, coaxial extrusion, or microfluidic integration—but also on endothelial-cell type, seeding timing, and seeding mode [214, 361]. For example, endothelial cells may be embedded during printing, seeded after channel formation, or introduced under perfusion conditions, and each approach has different implications for lumen stability, barrier function, and compatibility with tumor and stromal compartments [236]. The choice of endothelial source is likewise important, since primary endothelial cells, immortalized cell lines, and stem-cell-derived endothelial populations differ in angiogenic behavior and long-term stability. In TNBC bone-metastasis models, the vascular module is especially relevant because it influences extravasation, nutrient transport, immune-cell trafficking, and therapeutic exposure. Therefore, vascularization should be discussed not simply as a structural add-on, but as a functional determinant of model fidelity.

Long-term multicellular viability and medium compatibility.

Long-term multicellular viability remains a major challenge because prolonged culture requires simultaneous support of tumor cells, stromal cells, endothelial cells, and, in some designs, immune cells. Beyond diffusion limits and waste accumulation, medium incompatibility is a frequent but under-discussed source of model instability [32, 125, 163, 246]. A practical strategy is to use staged or mixed medium systems, in which the core formulation is optimized for shared survival while selected supplements are introduced according to the dominant biological process under investigation, such as vascular maturation, osteogenic signaling, or immune-cell maintenance. In some cases, sequential medium adaptation may be more appropriate than a single fixed formulation, particularly when the construct must transition from early post-print stabilization to long-term multicellular co-culture. Explicit discussion of medium compatibility is essential for interpreting whether a model failure reflects biology or culture-system design.

To place 3D bioprinting in a broader translational context, a direct comparison with other commonly used in vitro platforms is provided in Table 3.

Table 3.

Comparison of major in vitro platforms used for TNBC bone-metastasis modeling

Platform	Main advantages	Main limitations	Scalability	Translational readiness
3D bioprinted models	Spatial control, multicellular architecture, matrix tunability	Technical complexity, lower throughput	Moderate	High
Organoids	Patient relevance, tumor heterogeneity	Limited stromal/vascular complexity	Low–moderate	Moderate–high
Spheroids	Simple, inexpensive, high throughput	Poor structural control, limited niche fidelity	High	Low–moderate
Organ-on-chip	Perfusion, dynamic monitoring, mechanical cues	Device complexity, limited scalability	Low–moderate	High

Biomaterials for 3D printing applications

3D printing technology is revolutionizing the field of biomedical engineering, with its success fundamentally rooted in innovative breakthroughs in material systems. Early 3D bioprinting predominantly relied on single-material systems (Table 4). For example, natural materials such as alginate and collagen exhibit excellent biocompatibility but lack sufficient mechanical strength, necessitating the use of sacrificial supports during the fabrication of complex structures. Their unpredictable degradation profiles often lead to premature scaffold collapse. In contrast, synthetic polymers like PLA and PCL offer tunable mechanical properties but possess biologically inert surfaces that hinder cell adhesion, resulting in low osteoblast survival rates (< 40%). Photocurable resins provide high printing precision (at the micron scale) but may induce cellular damage due to ultraviolet exposure, while residual photoinitiators can increase DNA double-strand breaks by up to fivefold. These limitations significantly impede the bioprinting of complex organs—such as those requiring vascularization—and compromise the long-term safety of implanted constructs. Recently, through the integration of smart material design and interdisciplinary technologies, novel bioinks—such as gene-programmed bioinks, biomimetic composite structures, and green intelligent manufacturing paradigms—have emerged to effectively address these longstanding challenges.

Table 4.

Overview of the fundamental categories of 3D printing materials and selected advanced examples

Material category	Entry type	Representative materials/example	Key features	Core advantages	Main challenges	Reference
Natural biomaterials	Category overview	Alginates/Gelatin/Chitosan/Cellulose	Biodegradable, bioactive, sustainable-source materials	High biocompatibility; degradable and renewable	Low mechanical strength; uncontrollable degradation	[242, 253, 264]
	Selected advanced example	Mussel-Inspired Double-Network Hydrogel	Catechol-iron coordination; > 95% water content; self-healing	Biomimetic adhesion; high mechanical strength	Requires multi-step crosslinking optimization	[90]
	Selected advanced example	Thermoplastic Starch (TPS)	Renewable, biodegradable; rheology tunable by plasticizers	Low cost; environmentally friendly	Moisture sensitivity; requires reinforcement	[124]
	Selected advanced example	Engineered Living Materials (ELM)	Bacteria-derived protein matrix; shear-thinning; ~ 93% water content	Customizable mechanical properties	Relies on genetic engineering; long-term stability remains uncertain	[120]
	Selected advanced example	Cellulose-Based Hydrogel	Water-triggered swelling/deformation; 4D responsiveness	Dynamic deformation capacity	Insufficient mechanical stability	[81]
Synthetic polymers	Category overview	PLA/PCL/PLGA, etc	Mechanically tunable; controllable degradation; strong long-term stability	Good structural support; suitable for load-bearing use	Low bioactivity; risk from acidic degradation products	[141, 248, 325],[359]
	Selected advanced example	Magnesium–Titanium Biomimetic Composite Material	Biomimetic architecture; continuous-phase penetration; high strength/toughness	Crack propagation resistance	Complex fabrication; high cost	[356, 360, 365]
	Selected advanced example	PLGA/TCP Composite Scaffold	Drug-loaded, low-temperature printed scaffold for bone fusion	Synergistic drug release and bone regeneration	Degradation-rate matching requires optimization	[170]
	Selected advanced example	Degradable Thermosetting Resin	Bio-based monomers; photo-crosslinking; chemical recyclability	Closed-loop recycling potential	Residual photoinitiator toxicity	[58]
Hydrogels	Category overview	Double-network/thermosensitive/photo-crosslinked hydrogels	> 90% water content; ECM-like; self-healing	Excellent biomimicry; soft-tissue compatibility	Collapse and swelling during long-term use	[6, 157, 161, 164, 236, 323]
Composites	Category overview	PLA-black phosphorus nanosheets/Chitosan-PCL	Multifunctional synergy; mechanical–biological balance	Antibacterial/photothermal integration	Poor interface compatibility; uneven nanodispersion	[115, 251, 255]
Photocurable materials	Category overview	Acrylates/PEGDA/bio-based polyester resins	Micron-level precision; rapid curing	High-resolution fabrication	Residual photoinitiator toxicity; wavelength constraints	[61, 121, 155, 366]

Natural biological materials

Alginates, gelatin, chitosan, cellulose, collagen, and hyaluronic acid (HA) are representative natural biomaterials that have emerged as core components in soft tissue engineering and cell delivery systems, owing to their abundant availability, excellent biocompatibility, and inherent biodegradability. These materials are primarily sourced from renewable natural resources. Specifically, alginate is extracted from brown algae, gelatin is produced through the hydrolysis of animal collagen, chitosan results from the deacetylation of chitin found in crustacean shells, cellulose is abundantly present in plant cell walls, collagen constitutes a major structural component of animal connective tissues, and hyaluronic acid can be obtained via microbial fermentation or isolated from animal tissues. Their inherent biological properties confer excellent compatibility with human tissues [15, 277]. For example, alginate forms hydrogels via calcium ion cross-linking, effectively mimicking the extracellular matrix (ECM) microenvironment and enabling the encapsulation of chondrocytes for joint repair [322]; gelatin, which contains the RGD peptide sequence, significantly enhances cell adhesion and is widely utilized in the 3D printing of skin regeneration scaffolds [84, 310]; chitosan, due to its cationic nature, interacts electrostatically with negatively charged cell membranes, demonstrating superior performance in antibacterial dressings and nerve conduit applications [162, 239]. Furthermore, cellulose can achieve enhanced mechanical properties through chemical modifications, such as oxidation to produce nanocellulose, and is applicable in the 4D printing of humidity-responsive structures [260]. Meanwhile, hyaluronic acid exhibits excellent water retention and lubricating capabilities, making it the preferred choice for intra-articular injectable hydrogels and ocular surface repair applications.

The core advantages of these materials stem from their exceptional biocompatibility and tunable degradation profiles. A representative example is the gelatin-collagen composite derived from decellularized porcine skin, which not only preserves the three-dimensional porous architecture of the native extracellular matrix (ECM) but also undergoes enzyme-mediated degradation without generating toxic by-products. This material has demonstrated successful application in dermal reconstruction for large-area burn injuries [151]. Chitosan-hyaluronic acid composite hydrogels form a dynamic network via electrostatic interactions, enabling not only the encapsulation of mesenchymal stem cells to promote cartilage regeneration but also gradual degradation by lysozyme, thereby eliminating the need for secondary surgical removal [376]. Furthermore, the sustainable attributes of these materials have significantly broadened their application scope: Starch-based thermoplastic materials (e.g., TPS), after plasticizer modification, are widely utilized in food packaging and absorbable bone fixation devices, while bacterial cellulose, characterized by its high purity from microbial synthesis, serves as an ideal substrate for flexible electronic sensors and artificial vascular grafts [178].

However, the inherent limitations of natural materials cannot be ignored. Insufficient mechanical strength remains a prevalent challenge. For example, pure alginate scaffolds typically exhibit a compressive modulus below 10 kPa, which is inadequate to withstand the mechanical demands of bone tissue. To address this limitation, structural rigidity is commonly improved through the incorporation of nano-hydroxyapatite (nHA) or by photo-crosslinking methacrylated gelatin (GelMA) [330]. The lack of control over degradation rates may introduce significant clinical risks. Collagen is highly susceptible to rapid degradation by matrix metalloproteinases (MMPs). Without combination with synthetic polymers such as poly(lactic-co-glycolic acid) (PLGA), this vulnerability can result in premature scaffold collapse, thereby compromising bone integration [63]. To address this challenge, researchers have developed dynamic crosslinking strategies, exemplified by double-network alginate-polyvinyl alcohol (PVA) hydrogels, which prolong the degradation duration from several weeks to several months through the synergistic combination of physical entanglement and chemical crosslinking [269]. Moreover, batch-to-batch variability in materials presents a significant challenge to their large-scale application. For example, plant-derived cellulose exhibits considerable variations in degree of polymerization and crystallinity, which are influenced by species differences and variations in extraction methods. To ensure consistent printability, standardized pretreatment protocols—such as acid hydrolysis coupled with high-pressure homogenization—are essential [111].

Future breakthroughs are expected to center on the functionalization and intelligent design of biomaterials. For example, collagen modified through genetic engineering can incorporate specific enzyme-cleavage sites, enabling on-demand programming of degradation rates [289]; 4D-printed chitosan-temperature-sensitive hydrogels can activate shape-memory properties in response to body temperature, allowing for self-adaptation to irregular bone defects [114]. Meanwhile, advancements in green manufacturing processes—such as the direct dissolution of cellulose using ionic liquids—are anticipated to significantly reduce energy consumption during processing, thereby facilitating the scalable transition of natural materials from laboratory-scale to industrial-scale production [37]. In summary, natural biomaterials are progressively addressing the challenges associated with the synergistic regulation of mechanical properties, degradation behavior, and functional performance through interdisciplinary technological advancements, all while preserving their inherent "nature-friendly" characteristics. These developments offer improved solutions for personalized tissue regeneration.

Synthetic polymers

Synthetic polymers, including polylactic acid (PLA), polycaprolactone (PCL), and poly-(lactic-co-glycolic acid) (PLGA), have emerged as key materials in bone tissue engineering owing to their tunable mechanical properties, precisely adjustable degradation rates, and robust structural stability. Customized load-bearing bone scaffolds, such as pelvic and maxillofacial prostheses, fabricated using advanced techniques like Fused Deposition Modeling (FDM) and Selective Laser Sintering (SLS), demonstrate precise morphological compatibility with patient-specific bone defects [160, 221]. Its advantages are manifested in three key aspects. First, its mechanical properties can be tailored to specific requirements. By adjusting the crystallinity of PLA (30%–50%), a compressive modulus ranging from 1 to 3 GPa can be achieved, which aligns well with the mechanical demands of cortical bone. PCL exhibits exceptional ductility (elongation at break > 300%), thereby imparting superior fatigue resistance to intervertebral fusion devices. PLGA allows for tunable degradation rates through modulation of the lactide/glycolide ratio (e.g., 85:15), enabling degradation periods between 6 and 24 months that correspond to the timeline of bone regeneration [63, 96]. Second, the degradation behavior can be precisely controlled. Low molecular weight PLGA (10 kDa) undergoes approximately 90% degradation within 4 weeks, whereas high molecular weight PLGA (100 kDa) remains stable for over 12 months. Furthermore, triply periodic minimal surface (TPMS) scaffolds with porosity exceeding 70% enhance the diffusion of acidic degradation products by significantly increasing the specific surface area [289, 330]. Third, synthetic polymers exhibit long-term structural stability. For instance, PCL scaffolds retain more than 80% of their initial mechanical strength in vivo after 12 months, with their hydrophobic nature effectively resisting degradation by bodily fluids. In PLGA/TCP composites, the formation of calcium phosphate creates a pH buffering system that maintains microenvironmental homeostasis by keeping the pH above 5.5.

However, synthetic polymers also face two major challenges: insufficient biological activity and inflammatory responses associated with degradation. To overcome the former limitation, researchers have developed surface modification techniques, such as plasma amination of PLA, which enhances cell adhesion by threefold. Additionally, natural-synthetic composite strategies have been proposed, exemplified by PLA-gelatin composites that combine a high rigidity modulus (1.2 GPa) with improved cell migratory capacity. Furthermore, PCL composites incorporating black phosphorus nanosheets have demonstrated a 2.5-fold increase in alkaline phosphatase (ALP) activity—a key osteogenic differentiation marker—through photothermal stimulation [94, 330]. To mitigate degradation-induced inflammation, the incorporation of sustained-release components—such as PLGA supplemented with 5 wt% magnesium powder to neutralize acidic byproducts—alongside intelligent degradation-triggering systems, exemplified by 4D-printed PLA scaffolds embedded with body temperature-responsive anti-inflammatory agents, has been shown to significantly reduce macrophage activation risk [166, 256].

Recent advances in synthetic polymer research are propelling the field toward transformative breakthroughs. A biomimetic magnesium-titanium "brick-and-mud" architecture, exhibiting a compressive strength exceeding 300 MPa, demonstrates an eightfold improvement in crack resistance. 4D-printed shape-memory PLA/carbon fiber scaffolds possess the capability to self-expand at body temperature, enabling precise adaptation to complex bone cavities. Furthermore, robot-assisted in situ bioprinting of PLGA-based bioinks containing stem cells has been shown to enhance vascularization by 60% within seven days, offering new prospects for rapid tissue integration [114, 160, 373]. Future research trends are increasingly centered on multi-material gradient printing (e.g., PLA-PCL transitional scaffolds) and closed-loop degradation monitoring systems (such as implantable pH sensors), aiming to fulfill the "golden triangle" objective of achieving synergistic integration among mechanical adaptation, osteogenic induction, and harmless degradation [289, 332].

Hydrogel

Hydrogels, characterized by their high water content (> 90%) and extracellular matrix (ECM)-mimicking microenvironment, have emerged as essential materials for fabricating complex tissue architectures. A representative example is alginate/gelatin double-network hydrogels, where dynamic ionic bonds (Ca²⁺ crosslinking) and covalent bonds (gelatin amidation) synergistically contribute to their functionality. These hydrogels support the fabrication of perfusable vascular-like structures, and self-healing designs can help maintain structural integrity during long-term culture [84]. Thermosensitive hydrogels, such as the Pluronic F127/chitosan composite system, remain in a liquid state at 4 °C, facilitating uniform cell dispersion, and undergo spontaneous gelation at 37 °C to effectively encapsulate chondrocytes. This system demonstrates excellent biocompatibility, with a high post-transplantation cell survival rate [269]. Photocrosslinked hydrogels, such as methacrylated hyaluronic acid (HAMA), can rapidly solidify under ultraviolet (UV) light irradiation, forming stable scaffolds within 5 s. When integrated with projection lithography technology, this system enables the fabrication of bionic osteochondral units featuring multi-level channel architectures [319]. The core advantages of these materials stem from hydrogels’ ability to mimic physiological microenvironments. Hyaluronic acid hydrogels exhibit excellent lubricity (friction coefficient < 0.1), comparable to that of joint synovial fluid, which significantly improves the quality of cartilage regeneration. Their dynamic functional integration—such as the reversible recombination of catechol-iron coordination bonds in double-network hydrogels following fracture—enhances fracture toughness to 8.5 kJ/m², three times that of single-network hydrogels. Moreover, hydrogels demonstrate stimulus-responsive behavior. For example, encapsulating TGF-β3 nanoparticles within thermosensitive hydrogels enables sustained release upon exposure to body temperature, thereby promoting the differentiation of mesenchymal stem cells into chondrocytes [256].

However, the limited mechanical strength and inadequate long-term stability of hydrogels pose significant challenges to their clinical application. Pure gelatin scaffolds exhibit a compressive modulus of only 10–50 kPa, making them susceptible to structural collapse caused by body fluid infiltration after implantation. To address this limitation, sacrificial support materials are required for auxiliary shaping. For instance, the continuous viscoelastic matrix developed by Patricio et al. can be thermally dissolved after printing, offering a practical solution to this issue [226]. To overcome the limitations in mechanical performance, a three-fold strategy can be implemented. First, nanocomposite reinforcement: incorporation of cellulose nanocrystals (CNC) increases the modulus of alginate from 15 to 120 kPa and improves its tensile capacity by 400% [161]. Second, topological structure optimization: the implementation of a hyperbolic curvature lattice design enhances the compressive strength of hydrogels to 2.1 MPa—approaching that of cancellous bone—while preserving 80% porosity [326]. Third, crosslinking density modulation: substituting I2959 with the photoinitiator LAP doubles the crosslinking efficiency of GelMA and reduces the swelling ratio from 150 to 40% [94]. Long-term stability is governed by the balance between degradation and tissue regeneration: incorporation of enzyme-sensitive peptide chains (e.g., the MMP-cleavable sequence GPLGIAGQ) into the gelatin-hyaluronic acid double network ensures that the degradation rate aligns with the deposition of newly formed tissue [289].

Currently, the future development of hydrogels has garnered significant attention. 4D-printed hydrogels integrated with conductive polymers (e.g., PEDOT: PSS) enable real-time monitoring of cartilage strain signals. Meanwhile, gene-engineered bacterial cellulose hydrogels can secrete extracellular matrix (ECM) proteins, such as fibronectin, along predefined pathways, progressively replacing artificial scaffolds [373]. Through the integration of material gene editing and multi-scale fabrication techniques, hydrogels are transitioning from "passive carriers" to "active regulators of living systems," offering dynamic biological interfaces for organ bioprinting.

Composite materials

By integrating the biological activity of natural materials with the mechanical properties of synthetic materials, composite materials have pioneered a new direction for multifunctional implants. A representative example is the PLA system reinforced with black phosphorus nanosheets (BPNSs), where the broad-spectrum antibacterial capability of BPNSs (exhibiting over 99% Staphylococcus aureus eradication) and their near-infrared photothermal effect (inducing a local temperature increase of 52 °C) act synergistically. This synergy not only suppresses postoperative infections but also enhances osteoblast ALP activity by 250%, thereby endowing the scaffold with dual functionalities: anti-infective protection and bone-inductive capacity [330]. Natural-synthetic hybrid materials, such as chitosan-PCL, can be fabricated into micro-nano fibrous networks via electrospinning. The cationic nature of chitosan enhances fibroblast migration, increasing the migration speed by 80%, whereas the hydrophobic PCL backbone (with a modulus of 0.4 GPa) provides structural stability in load-bearing regions. This material system has demonstrated successful application in maxillofacial bone defect repair [289].

The core advantages of these materials are summarized as follows: 1. Performance synergy amplification: Incorporation of cellulose nanocrystals (CNC) into PLA enhances the tensile strength from 45 to 112 MPa. Additionally, the surface hydroxyl groups of CNC significantly promote hydroxyapatite deposition, tripling the mineralization rate. 2. Intelligent function integration: The graphene oxide (GO)/gelatin hydrogel system, integrated with electrical stimulation responsiveness, allows voltage-controlled modulation of the BMP-2 release profile, thereby enabling precise temporal regulation of bone regeneration [373]. 3. Biomimetic structure emulation: The chitosan-hydroxyapatite composite mimics the "inorganic–organic" interdigitated architecture of natural bone, achieving a compressive strength of 180 MPa—comparable to cortical bone—while maintaining magnetic resonance imaging (MRI) compatibility [166, 332].

However, the fabrication complexity and scalability limitations of composite materials present significant challenges. Inhomogeneous dispersion of nanoparticles—such as the agglomeration of black phosphorus nanosheets (BPNSs) caused by van der Waals forces—can induce localized stress concentration, thereby reducing the fatigue life of the scaffold by 40%. Furthermore, poor interfacial compatibility between natural and synthetic components—such as the mismatch between hydrophobic PCL and hydrophilic chitosan—leads to phase separation, necessitating the incorporation of graft copolymers (e.g., PCL-polyethylene glycol-PCL) to enhance interfacial adhesion [269]. In response to these challenges, advanced research has focused on three key strategies: Field-assisted precise dispersion: Electrohydrodynamic (EHD) printing enables directional alignment of cellulose nanocrystals (CNCs) within the PLA melt, resulting in a fivefold increase in the anisotropic modulus ratio [241]. Black phosphorus nanosheets are increasingly incorporated into 3D-printed scaffolds because of their osteogenic and multifunctional potential, although dispersion stability remains a challenge [237, 330]. Reactive compatibilization, including maleic-anhydride-based strategies, can improve the interfacial compatibility of polyester/chitosan blends (Li, M. et al., 2022). Green and cost-effective processing: Direct dissolution of cellulose/PLA blends using the ionic liquid BMIM-Cl eliminates the need for solvent recovery, and may support greener manufacturing workflows [178].

The future development of composite materials hinges on the integration of multidisciplinary technologies: 4D-printed shape-memory composites—such as PLA/carbon fiber combined with thermosensitive hydrogels—can autonomously expand and adapt to irregular bone cavities upon body temperature activation. Biomimetic mineralization techniques: direct the epitaxial growth of hydroxyapatite along cellulose nanofibers, yielding "self-reinforced" scaffolds with a flexural strength exceeding 200 MPa. Microfluidic coaxial bioprinting enables one-step fabrication of core–shell fibers or microfibers for multifunctional delivery and tissue-engineering applications [114]. By leveraging machine learning to optimize process parameters—such as predicting the rheological behavior of nanoparticles—composite materials are poised to overcome the "performance-cost-efficiency" trade-off and advance toward a new era of personalized smart implants.

Photocurable materials

Photocurable materials, including acrylate-based resins and polyethylene glycol diacrylate (PEGDA), are widely used in resin-based 3D printing for biomicrofluidic fabrication because they support rapid prototyping, monolithic construction, and micron-scale feature generation. In particular, digital light processing (DLP) and related vat-photopolymerization approaches are attractive for microfluidic applications because they enable layer-by-layer curing with relatively high resolution and design flexibility [94, 208]. At smaller length scales, two-photon polymerization (TPP/2PP) enables the fabrication of highly defined microarchitectures for cell-guidance studies. For example, Sharaf et al. fabricated 2.5D microgrooves and 3D microchannels with diameters ranging from 5 to 30 μm to investigate neuronal morphology, mechanotransduction, and neurite outgrowth under spatial confinement [258]. The major advantages of photocurable systems therefore include rapid fabrication, high geometric fidelity, and compatibility with microscale device design. However, these materials still face important limitations related to resin cytotoxicity, post-processing dependence, and irradiation conditions. Musgrove et al. emphasized that the performance of resin-printed biomicrofluidic devices depends on balancing print resolution, mechanical stability, cytocompatibility, and optical compatibility, whereas Hart et al. showed that extensive post-processing can dramatically improve the biocompatibility of commercial printed resins, in some cases approaching that of a standard cell culture vessel [94, 208]; In addition, recent materials development has extended beyond conventional petroleum-derived resins. For example, Cazin et al. reported bio-based polyester resins suitable for vat photopolymerization, although the demonstrated printed test structures were in the millimeter-resolution range rather than biochip-specific constructs [29].

In recent years, breakthrough strategies have centered on three key optimization directions. The first involves the development of biocompatible initiation systems, categorized into enzyme-responsive and metabolizable initiators. Enzyme-responsive initiators, such as the horseradish peroxidase (HRP)/H₂O₂ system, enable visible light (405 nm) curing and reduce cytotoxicity to negligible levels [237]. In parallel, lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) has become one of the most widely used photoinitiators in DLP bioprinting, as it supports 405 nm light curing while exhibiting significantly improved cytocompatibility compared with conventional UV-responsive initiators. Metabolizable initiators, represented by riboflavin (vitamin B₂) derivatives, degrade into biologically compatible nutrients in vivo and can decrease the inflammatory cytokine IL-6 by 80% following implantation [256]. Secondly, green resin development has promoted the use of bio-based photocurable systems, including soybean-oil-derived acrylates and bio-based polyester resins, for vat photopolymerization [29]. Another breakthrough involves thiol-ene click chemistry resins that cure in the absence of oxygen, effectively eliminating the inhibition layer problem and tripling interlayer bonding strength [289]. Lastly, a wide-spectrum light source adaptation strategy utilizes upconversion nanoparticles (UCNPs) to convert near-infrared light (980 nm) into ultraviolet light, enabling deep curing beyond 5 mm while preventing surface thermal damage [373].

The future trajectory of intelligent development hinges on the integration of bio-inspired chemistry with smart light-regulating technologies. Photocurable materials are progressively overcoming biocompatibility barriers, paving the way for a new era in the fabrication of implantable, precision-engineered medical devices. Future development directions encompass temperature-responsive liquid crystal acrylates that spontaneously fold at body temperature, enabling the fabrication of 4D-integrated, functionally autonomous self-closing vascular clips [114]; handheld blue light-assisted in situ bioprinting of GelMA ink embedded with bone marrow-derived mesenchymal stem cells, which accelerates cartilage defect repair to a 4-week cycle ([160]); and a machine learning-driven closed-loop quality control system that enables real-time monitoring of curing degree via Raman spectral peak shifts, reducing printing defect rates from 12% to 0.7% [215].

Summary

3D-printed biomaterials have established a technological framework encompassing "performance customization, functional integration, and intelligent responsiveness." However, their clinical translation continues to encounter three critical challenges (Table 5). First, the vascularization bottleneck must be addressed, particularly the development of perfusable capillary networks required for complex organ printing. Second, long-term safety concerns remain, including the systemic effects of accumulated degradation products such as nanoparticles. Third, the scalability and cost-effectiveness of manufacturing processes, such as electric field-assisted dispersion, require further improvement to ensure industrial adaptability. Future breakthroughs are expected to concentrate on three key directions: first, the simulation of life processes through gene-editing hydrogels that program ECM protein secretion, hereby supporting the development of increasingly adaptive and biologically active scaffold systems; second, the advancement of closed-loop intelligent systems, exemplified by implantable pH sensors that enable real-time monitoring of PLGA degradation and trigger responsive controlled release of anti-inflammatory drugs, such as 4D-printed thermosensitive drug delivery units; and third, the fabrication of perfusable microvascular networks and multicellular tissue units through diverse bioprinting and microfabrication strategies, including but not limited to near-field electrospinning, thereby enabling the generation of functional architectures such as liver lobule-like constructs [373]. Ultimately, these advances may contribute to the development of bio-digital integrated “living material” platforms that can undergo functional adaptation after printing and better support the regeneration of complex biological structure and function.

Table 5.

Future directions and technical challenges in 3D-printed biomaterials

Direction	Key challenge	Proposed strategy	Representative implication
Resolution of the vascularization bottleneck	Limited ability to fabricate stable and functional microvascular networks	Near-field electrospinning (NFE) and microscale vascular network printing	Construction of perfusable tissue units such as liver lobule-like structures
Improvement of long-term safety	Accumulation of degradation products and insufficient long-term monitoring	Closed-loop degradation monitoring systems	Implantable pH sensing and 4D drug-controlled release
Optimization of large-scale production	High cost of electric field-assisted dispersion and limited process predictability	Green processing with ionic liquids and AI-assisted prediction of rheological behavior	Improved scalability and manufacturing reproducibility
Interdisciplinary technological integration	Limited integration of biological, material, and digital modules	Gene-editing hydrogels, ECM-secreting living materials, self-healing mineralized scaffolds, and bio-digital design	Toward adaptive and multifunctional living material platforms

Design and production of 3D printing models

Three-dimensional printing technology, as an emerging rapid prototyping technique, has triggered a transformative shift in biomedical engineering by leveraging its distinctive capabilities—including personalized customization, fabrication of complex architectures, and integration of multiple materials. Enabled by a layer-by-layer additive manufacturing approach, it allows for precise physical reconstruction spanning from macroscopic anatomical constructs to microscale tissue mimetics, thereby offering unprecedented opportunities in disease modeling, drug screening, surgical planning, and the development of patient-specific implants. This chapter systematically outlines the design principles and fabrication workflows of 3D-printed models, with emphasis on cutting-edge applications in critical domains such as breast cancer and tumor microenvironments, cardiovascular systems, and skeletal structures. Through representative case studies, it illustrates how this technology effectively bridges the translational gap between fundamental research and clinical implementation, thus advancing precision medicine toward new frontiers. Among these applications, the development of TNBC-oriented bioprinting platforms represents a particularly important future direction, especially for modeling bone metastatic progression and therapeutic response. Future TNBC-oriented bioprinting platforms should prioritize bone-mimetic stiffness, osteolytic signaling, immune co-culture, and perfusable vascular features to better predict TNBC-specific metastatic behavior and treatment response.

Breast cancer and its peritumoral microenvironment model

The occurrence, development, and metastasis of breast cancer depend on the dynamic interactions between tumor cells and their microenvironment. This complex ecosystem encompasses multiple layers of biological processes, including adipose stromal metabolic regulation, vascularization, pre-metastatic niches in bone, and immune editing. 3D printing technology enables the construction of high-fidelity microenvironment models via biomimetic fabrication approaches, such as bioprinting, nanomaterial integration, and intelligent sensing. This advancement not only facilitates the transition of mechanistic studies from static to dynamic systems, but also offers a platform for simulating pathological barriers in drug development.

Research on tumor biology and mechanisms

Through collaborative innovation in high-precision biomanufacturing technology, 3D-printed breast cancer models have enabled spatiotemporal dynamic characterization of microenvironmental physical and biochemical parameters. Chaji’s team employed extrusion-based bioprinting to crosslink and solidify methylcellulose/hyaluronic acid composite hydrogels under 405 nm light, establishing a co-culture system of adipocytes and breast cancer cells. For the first time, they quantified the dose-dependent regulation of tumor cell proliferation by leptin via the JAK2/STAT3 signaling pathway (Fig. 4A, a1,a2) [31]; to further simulate the metastasis process, Cui H et al. fabricated a PEGDA-based vascular network incorporating human umbilical vein endothelial cells (HUVECs) using digital light processing (DLP) printing with the photoinitiator Irgacure 2959. This system, integrated with a nano-hydroxyapatite bone matrix, demonstrated that breast cancer cells colonize bone through integrin αvβ3-mediated vascular co-invasion (Fig. 4B, b1, b2) [43].

Fig. 4.

Bioprinting strategies and model systems for studying breast cancer and tumor microenvironment. A: Cylindrical structures were fabricated along the X-axis using a modified extrusion-based 3D printer (a₁), with standardized calibration of cell-laden hydrogel printing parameters (a₂) [31]. B: Schematic representation of MDA-MB-231 cell invasion into bone in vivo, along with the tri-culture model (b₁; top and side views of the three-dimensionally printed BrCa bone metastasis model (b₂) [43]. C: Co-culture of vascular endothelial cells and breast epithelial cells within a collagen-chitosan bioink matrix (c₁) [283], Scale bar = 200 μm; agarose molds engineered using 3D Petri Dishes® to enable high-throughput production of ASC spheroids (256 spheroids per mold) (c₂), Scale bar = 500 μm; hydrogel composition fine-tuned by modulating hyaluronic acid content to achieve uniform spheroid distribution (c₂), Scale bar = 1000 μm [104]. D: Engineered scaffold incorporating AuNR@2NAT, extracellular matrix (ECM), and SERS-labeled cells (d₁) [72]; alongside a ligand-AuNR-free control scaffold (d₂) [72]; computational models of ring-shaped and tri-layered hydrogel architectures (d₃) [21], with corresponding Monte Carlo simulation (MMC) results (d₄) [21]

At the level of capturing dynamic interactions, Swaminathan S et al. developed direct suspension bioprinting to deposit MCF-7 spheroids onto a fibrin gel cross-linked by thrombin, enabling real-time observation of VEGF-mediated endothelial sprouting at a rate of 50 μm/hour (Fig. 4C, c1) [283]; meanwhile, Horder H et al. utilized thermoresponsive Pluronic F127 (phase transition temperature: 22 °C) to fabricate adipose stromal cell spheroids, demonstrating that IL-8 secretion promoted tumor invasion and migration via the CXCR1/ERK signaling axis (Fig. 4C, c2) [104].

Technical integration has driven a breakthrough in advanced biofabrication. Garcia-Astrain C et al. developed a gelatin scaffold functionalized with gold nanorods (glutaraldehyde-crosslinked) by incorporating surface-enhanced Raman scattering (SERS), enabling dynamic tracking of metabolic heterogeneity at the single-cell level (Fig. 4D, d1, d2) [72]. In parallel, Bojin F et al. employed multi-nozzle bioprinting to fabricate a collagen-sodium alginate composite hydrogel cross-linked by Ca²⁺ ions, establishing a tri-culture system of fibroblasts, macrophages, and tumor cells. This platform allowed quantitative analysis of the stiffness threshold in matrix remodeling induced by TGF-β1 (Fig. 4D, d3, d4) [21]. These techniques have evolved from static co-culture systems to the dynamic reconstruction of vascularized and immunocompetent microenvironments, thereby establishing a mechanistic basis for spatiotemporal therapeutic strategies targeting matrix-mediated interactions.

Drug screening and therapeutic testing

Pathology-biomimetic 3D printing models are systematically reshaping the drug development paradigm, with a key innovation being the precise simulation of in vivo drug efficacy barriers. Svanstrom A et al. utilized ionically cross-linked alginate (mediated by Ca²⁺) to fabricate tumor scaffolds from patient-derived cell lines, demonstrating that increased microenvironmental stiffness led to a 3.7-fold elevation in the IC₅₀ of doxorubicin through YAP/TAZ nuclear translocation (Fig. 5A) [282]. To address the challenge of drug penetration, Gebeyehu A et al. developed a chitosan-oxidized dextran hydrogel through Schiff base dynamic cross-linking, which mimicked an interstitial fluid pressure environment and enhanced the penetration depth of paclitaxel nanoparticles from 180 μm to 252 μm (Fig. 5B, b1, b2) [73]. To overcome the critical challenge of drug resistance, Hong S et al. developed MDR1⁺ spheroid models via photopolymerization of GelMA using the Li-TPO initiator, and quantitatively demonstrated that Tariquidar enhanced the doxorubicin uptake rate by 82% through inhibition of P-glycoprotein (P-gp) (Fig. 5C) [103].

Fig. 5.

Drug response evaluation and therapeutic strategy development using 3D-bioprinted cancer models. A: Comparative gene expression profiling of doxorubicin-treated MCF-7 cells cultured in 2D cell culture (i), 3D-printed scaffolds (3DPS, ii), and patient-derived scaffolds (PDS, iii) [282]. B: Morphological characterization and pore architecture of ten-layer H4-RGD ink-based bioprinted scaffolds (b₁; SEM images depicting the growth of 3D-printed NSCLC-PDX cells/spheroids within the H4-RGD ink-based scaffold (b₂) [73]. C: Schematic illustration of bioprinting-enabled generation of dense, drug-resistant cancer spheroids and subsequent imaging-based structural analysis of tumor organization [103]. D: Efficacy comparison of MXene-mediated photothermal therapy in conventional versus bioprinted tumor spheroids (d₁) [229], fabrication, structural design, and therapeutic impact on mouse survival of DOX-CDDP-PLGA composite scaffolds (d₂) [234]; integrated strategy combining curcumin nanoparticles (CurNPs) with 3D-bioprinted hydrogel scaffolds for localized breast cancer treatment (d₃) [275]

In the innovative therapeutic validation platform, Perini G et al. integrated inkjet-printed Ti₃C₂Tₓ MXene nanosheets into type I collagen scaffolds, enabling precise local ablation upon 1064 nm laser irradiation (Fig. 5D, d1) [229]; Qiao X et al. fabricated PLGA-PEG dual-drug scaffolds with distinct release profiles—paclitaxel for rapid release and doxorubicin for sustained delivery—resulting in a significantly reduced recurrence rate compared to single-drug formulations (Fig. 5D, d2) [234]. Su Y et al. incorporated FA-PEG-CUR nanoparticles into a DLP-based photopolymerized GelMA biomimetic scaffold model and demonstrated that the combined therapeutic strategy achieved a 91.4% clearance rate for 6 mm-deep lesions under 808 nm laser excitation (Fig. 5D, d3) [275]. These advancements have catalyzed a paradigm shift in drug evaluation, transitioning from static, single-drug assays to a dynamic optimization framework that integrates barrier simulation, drug resistance mitigation, and the exploration of synergistic combination therapies.

Clinical surgical planning and decision support

Patient-specific 3D-printed models, enabled by multimodal image fusion and high-precision physical reconstruction, effectively close the surgical decision-making loop. FDM color models based on CT data (Santiago L/ABS material), featuring dual-color visualization of tumor and glandular tissue, significantly reduced patients’ decision conflict scores (Decisional Conflict Scale, p < 0.05). To address the challenge of tissue deformation following neoadjuvant chemotherapy, Ko BS et al. employed SLA technology to fabricate resin-like MRI-guidance templates with an accuracy of ± 0.1 mm, resulting in a reduced positive margin rate in breast-conserving surgery (OR = 0.24) (Fig. 6A, a1) [139]; meanwhile, Luo J et al. developed titanium alloy guides using DLP technology, with the anatomical shape of the rib tissue, by simulating a 15 ± 3% tissue contraction rate after neoadjuvant chemotherapy, achieving a quadrant resection error of less than 1 mm (Fig. 6A, a2) ([187]).

Fig. 6.

Development and application of patient-specific 3D-guided systems for surgical planning and tumor resection in breast cancer. A: Pre-neoadjuvant chemotherapy three-dimensional models of the breast and tumor, along with surgical guidelines delineating the skin surface and tumor extent [139, 187]. B: Intraoperative margin definition using customized surgical guides tailored to individual breast anatomy—anchored primarily on the areolar ring and secondarily guided by technetium labeling (b₁); illustration of the two-stage breast reconstruction workflow (b₂) [295, 296]. C: Patient-specific hybrid surgical guides incorporating blue dye injection ports to enable real-time visualization and precise demarcation of the target tumor volume [333]

In the field of reconstructive optimization, Chae MP et al. employed a three-dimensional volume analysis model (VAM® software) to assess outcomes following autologous flap breast reconstruction using the St Andrew’s coning technique. To validate the measurement accuracy of the software, the three-dimensional reconstructions it generated were compared against MRI, serving as the gold standard. The high Dice similarity coefficient of 0.92 ± 0.03 indicates strong agreement between the two modalities, supporting the reliability and precision of this analytical tool in volumetric assessment [30]; Tomita K et al. integrated structured light scanning with FDM-fabricated PLA molds, achieving a dimensional accuracy of 1.7 ± 0.3 mm between the scanned 3D model and the original digital design. (Fig. 6B, b1, b2) [295, 296]; for hidden lesions, Wu ZY’s SLA resin ultrasound guide plate (with micro-channel positioning grid) successfully guided the resection of DCIS (pathology confirmed negative margin) (Fig. 6C) [333]. These advanced technologies convert two-dimensional imaging data into tangible and functionally actionable decision-support tools, leading to a significant reduction in total operative time and improved patient satisfaction scores.

Advances in radiotherapy and medical physics

3D printing technology has systematically addressed the critical challenges in radiotherapy precision through innovations in tissue-equivalent materials and the integration of dynamic functionalities. Specifically, Ali A et al. developed a silicone-wax composite material using FDM mold casting techniques, enabling ultrasound/mammography dual-modality puncture training with a positioning error of less than 1.5 mm (Fig. 7A) [3]. To quantitatively assess the impact of respiratory motion, Choi YE designed a pneumatically driven thermoplastic polyurethane (TPU) phantom capable of dynamically simulating respiratory-induced organ displacement, revealing a mean target offset of 12.3 ± 2.1 mm (Fig. 7B) [39]. In the field of radiation dose optimization, Zhang Y et al. employed a DLP photopolymer resin-based compensator incorporating TiO₂ fillers (attenuation coefficient: 0.28 cm⁻¹), effectively reducing the heart and lung dose volume (V₂₀) from 28.7% to 4.9% (ΔDmean = 9.2 Gy, p < 0.001) (Fig. 7C, c1) [365]. Park K et al. introduced a stepped silicone compensator fabricated via 3D-printed mold casting to improve electron beam dose uniformity, achieving a γ pass rate (3%/3 mm) of 97.6 ± 1.8% (Fig. 7C, c2) [224].

Fig. 7.

Applications of 3D-printed phantoms and patient-specific devices in breast cancer diagnostics and radiotherapy planning. A: Biopsy simulation was conducted on a gelatin phantom embedded with blueberries, and the ultrasound image showed that the target lesion had good visibility [3]. B: On a 3D printed patient-specific breast model, the dose distribution of the treatment plan and the projected beam were compared under free breathing motion, with gamma pass rates (3%/3 mm criterion) of 98.54% and 97.53%, respectively [39]. C: AA new patient fixation device can minimize the air gap between the mold and the patient, and the average PTV dose, HI and CI in VMAT plans were compared between traditional and 3D printed inserts (c₁) [365]; schematic diagram of the process from CT acquisition to placement of the 3D printed insert on the patient’s chest (c₂) [224]. D: An example of the dose distribution of a patient undergoing breast reconstruction using 3D printed plugs in the TPS [313]

To address the unique dosimetric requirements of post-reconstructive patients, Wang J et al. developed a BaSO₄-PLA compensator film using FDM printing technology, accurately compensating for implant-induced dose attenuation (10.2 ± 0.8%) and ensuring a target D₉₀ coverage compliance rate exceeding 98% (Fig. 7D) ([313]). These breakthroughs, centered on tissue equivalence and biomechanical simulation, establish a high-fidelity physical reference framework encompassing equipment quality control, dynamic error correction, and personalized dose optimization.

Imaging and diagnostic tools

3D phantoms based on biomimetic printing are advancing imaging diagnostics toward traceable and standardized evaluation. Varallo A et al. developed breast phantoms incorporating glandular masses (volume: 0.5–2.0 cm³) and microcalcification models using inkjet printing with gelatin-oil-based biomimetic materials. Systematic evaluation of image quality across varying imaging parameters enabled optimization of the digital breast tomosynthesis (DBT) tube voltage and filter configuration (W/Rh, 34 kV). This optimized setting enhanced the contrast-to-noise ratio (CNR) for microcalcification detection to 14.2 ± 1.8, reflecting a 37% improvement over lower-voltage alternatives within the same filtration scheme. (Fig. 8A, a1) [302]. He Y et al. developed a multi-material breast phantom composed of polyvinyl alcohol (PVA), a gadolinium-based contrast agent (Gd-DOTA), and Fe₃O₄ nanoparticles, designed as a physical reference standard for multimodal imaging validation and image registration across MRI, CT, and ultrasound. The phantom demonstrates tissue-mimicking imaging properties: the simulated glandular region exhibits a T₁ relaxation time of 850 ± 50 ms, a T₂ of 90 ± 5 ms, and a CT attenuation value of 45 ± 3 HU; the overall material achieves an ultrasound attenuation coefficient of 0.75 ± 0.05 dB/cm/MHz. When employed in multimodal image registration, the phantom enables sub-1.2 mm spatial alignment accuracy, underscoring its utility in quantitative imaging assessment (Fig. 8A, a2) [97].

Fig. 8.

Fabrication and application of a two-component 3D-printed compressed breast model for imaging simulation and anatomical comparison. A: A two-component compressed breast phantom, 66 mm in thickness, was fabricated using blue ABS and white PLA to simulate adipose tissue and glandular/skin tissues, respectively, representing breast compositions with glandular volume fractions of 33% and 26%. The right panel displays the relationship between the β parameter and distance from the breast’s upper surface (a₁) [302]. An embedded insert enables dual-modality compatibility with both mammography and magnetic resonance imaging, with corresponding MRI acquisition parameters provided below (a₂) [97]. B: Comparative 3D reconstructions of normal breast anatomy and breast cancer pathology are presented, with a coin included as a scale reference [254]

To overcome the critical challenge in AI diagnostic validation, Schulz-Wendtland R et al. employed stereolithography (SLA) to transform manually segmented tumor contours derived from MRI into high-precision physical models, which were subsequently validated against the gold-standard gross histopathological analysis. The study demonstrated a margin localization error of only 0.68 ± 0.11 mm for invasive ductal carcinoma, highlighting the exceptional accuracy of this approach in aligning radiological imaging with pathological findings. (Fig. 8B) [254]. These breakthrough phantoms incorporate DICOM-RT protocol compatibility and physically embed microenvironmental features—such as the spatial distribution of calcifications and the fat-gland transition gradient—to establish a comprehensive quality control framework. This framework spans from the optimization of imaging acquisition parameters and multi-modal device calibration to the spatial accuracy validation of AI algorithms, thereby providing a traceable physical benchmark for the clinical translation of early breast cancer diagnosis.

Fundamental model development and technical optimization

The convergence of interdisciplinary technologies is propelling 3D-printed breast cancer models toward paradigms characterized by enhanced biological activity and intelligent sensing capabilities. Zhu W et al. pioneered the use of low-temperature deposition manufacturing to fabricate a 3D model of the bone microenvironment using nano-hydroxyapatite/PLGA composite materials. This model enabled the investigation and validation of osteoblast-mediated secretion of TGF-β1, which activates downstream signaling pathways and subsequently modulates the behavior of breast cancer cells (Fig. 9A, a1) [375], Rosendahl J et al. developed an extrusion-based printing method to produce transparent nanocellulose scaffolds that support live-cell confocal imaging for up to 14 days, enabling the capture of dynamic remodeling features associated with cell addition and subtraction during epithelial-mesenchymal transition (EMT) (Fig. 9A, a2) [243].

Fig. 9.

Advanced 3D-printed scaffolds and bioelectronic platforms for cancer modeling and tissue engineering. A: Schematic diagram of 3D printed bone matrix manufacturing and verification (a₁) [375]; TEMPO-CNF scaffold design and printability, followed by SEM visualization of the formation of heterogeneous cell clusters with different cell morphologies on the scaffold (a₂) [243]. B: Schematic diagram of three-dimensional printed graphene-based nanocomposite bioelectronics manufacturing, and SEM images of cells seeded on FN@3D-nGE showing cells at early and late attachment stages (b₁) [207]; summary of the proposed on-chip cancer model (b₂) [204]; breast cancer tumor schemes designed for drug screening through 3D bioprinting and bioprinting pathways versus traditional pathways (b₃) [259]

In the realm of intelligent perception integration, Muñoz J et al. fabricated graphene/PLA microelectrodes using fused deposition modeling (FDM) printing technology, enabling real-time discrimination and continuous monitoring of weak cell adhesion—a hallmark phenotype of highly metastatic cancer cells—via electrochemical impedance measurements (Fig. 9B, b1) [207]; Moghimi N et al. achieved tumor cell spatial heterogeneity programming through microfluidic bioprinting and quantitatively analyzed the regulatory weight of hypoxia gradients (Core area < 0.1% O₂) on chemoresistance (Fig. 9B, b2) [204]; Sharifi M et al. constructed a photocrosslinked GelMA vascularized model (loaded with DOX@MSN drug delivery system, drug loading 382 mg/g) in combination with the EPR effect, which increased the enrichment of doxorubicin in metastatic foci by approximately 3.1 times (Fig. 9B, b3) [259]. These breakthroughs upgrade the traditional "passive carrier" model to a closed-loop feedback intelligent system—through the collaborative integration of mechanical sensing (adhesive force ΔZ), chemical microenvironment monitoring (O₂ gradient), and therapeutic response feedback (drug enrichment degree), achieving a dynamic optimization cycle of "material design—biological manufacturing—functional response".

3D Printing technology route for vascular system

Driven by clinical requirements, vascular 3D printing technology has established a comprehensive technical pipeline spanning from anatomical replication to biological reconstruction. This section presents a systematic elaboration of four core technical pathways: the precise reconstruction of cardiovascular spatial architecture establishes a morphological foundation for diagnosis and therapeutic planning; dynamic hemodynamic simulation enables functional evaluation; customized design of patient-specific implantable devices facilitates precision intervention; and the engineering of bioactive tissues directs the future development of regenerative medicine. Collectively, these pathways constitute a hierarchical and progressive technical framework.

Cardiovascular spatial architecture reconstruction model

Anatomical structure models, serving as a cornerstone clinical application of vascular 3D printing technology, offer a high-fidelity three-dimensional physical reference for surgical planning and medical education through the precise replication of complex cardiovascular anatomical morphologies [374]. These models demonstrate three primary categories of pathological features: congenital cardiovascular malformations, including the anatomical reconstruction of complex congenital heart diseases (Fig. 10A, a1, a2) [27, 216, 349] and the morphological visualization of vascular rings (Fig. 10A, a3) [48, 123] significantly enhance the accuracy of diagnosis and treatment; acquired vascular lesions mainly reproduce secondary lesions such as aortic aneurysms/dissections (Fig. 10B,b1) [62, 171] and coronary artery abnormalities (Fig. 10B, b2) [149, 211] and optimize the surgical decision-making path through physical models; tumor invasion models (such as the spatial reconstruction of pericardial tumors) (Fig. 10C) [228] further provide key evidence for defining the boundaries of tumor resection. Valverde’s systematic research confirmed the core value of this model in three major dimensions: medical education, cardiovascular surgery, and interventional structural heart disease [301].

Fig. 10.

Patient-specific 3D printed cardiovascular models for surgical planning and medical education. A: Models of congenital heart diseases, including anatomical reconstruction of complex heart malformations (a₁, a₂) [216, 349]and visualization of the morphology of double aortic arch (a₃) [123]. B: Models of acquired vascular lesions, showing aortic aneurysm/dissection (b₁) [62], and coronary artery abnormalities (b₂) [149]. C: Models of tumor invasion, showing the adjacency relationship between pericardial tumors and coronary arteries through spatial reconstruction [228]

Hemodynamic functional simulation model

Functionalized fluid models establish a critical experimental platform for optimizing interventional treatment strategies by dynamically simulating hemodynamic alterations. These models primarily address two key pathological categories: stenotic/occlusive lesions, encompassing blood flow resistance analysis in coronary artery stenosis (Fig. 11A, a1, a2) [263, 344] and assessment of fluid dynamics associated with atherosclerotic plaques in the carotid arteries (Fig. 11A, a3, a4) [4, 116]; and shunt/abnormal flow patterns, including the reconstruction of shunt mechanisms in congenital heart disease (Fig. 11B, b1) [297] and hemodynamic remodeling following artificial vascular graft implantation (Fig. 11B, b2) [197]. The key technical characteristics of these models include: (1) the integration of transparent flow channels with particle tracing technology [4] to enable flow field visualization, and (2) the incorporation of an embedded pressure sensing system[271] for dynamic quantification of fluid mechanical parameters. Shepard’s study innovatively demonstrated the clinical applicability of such models in simulating fractional flow reserve (FFR) in coronary arteries [263], whereas Yang et al. systematically elucidated the core hemodynamic impact mechanisms induced by stenosis morphology through spatial parameter analysis [344].

Fig. 11.

Patient-specific 3D printed vascular models for hemodynamic analysis and interventional planning. A: Left coronary artery tree model for anatomical comparison (a₁), fractional flow reserve (FFR) measurement (a₂) [263, 344]; and particle image velocimetry (PIV) fluid simulation (a₃), in vitro imaging comparison of different carotid artery stents (a₄) [4, 116]. B: Pulmonary artery atresia model based on CT and three-dimensional rotational angiography, printed with transparent resin and used to verify interventional treatment plans (b₁) [297]. 3D anatomical and imaging comparison of chronic total occlusion lesions in vein graft vessels (b₂) [197]

Customized patient-specific implant device model

The prototype of implantable devices advances endovascular implant innovation through patient-specific customization, thereby enabling a paradigm shift toward precision intervention in two core domains: vascular stents and valve models. Within the vascular stent domain, it encompasses topological optimization design for coronary stents [300] and the end-to-end manufacturing system of aortic stents [76]; in the field of valve models, it focuses on the size adaptability of transcatheter aortic valves [62] and interventional operation training [195]. The technological breakthroughs are primarily manifested in two aspects: (1) the enhanced biocompatibility of degradable materials [231], which significantly reduces the risk of long-term complications, and the hemodynamic design of topologically optimized structures [300], which enhances the functional durability of implants.

Biologically active vascular tissue engineering model

Biological tissue engineering models are designed to generate functional living vascular tissues, thereby paving novel pathways for regenerative medicine. Specifically, vascular network models achieve high-precision structural fidelity validated by photoacoustic imaging [190] and promote targeted vascularization reconstruction for ischemic treatment [200]; organ-on-a-chip models simulate the cross-scale coupling mechanism of the heart and blood vessels [148, 280], accelerating the process of pathological mechanism analysis and drug screening. Frontier innovations center on the precise engineering of physiological microenvironments using hydrogel-cell co-printing technology [297], as well as the enhancement of host-graft integration efficacy through immune regulatory strategies.

Bone model

Bone 3D printing represents an interdisciplinary integration of materials science, medical imaging, and clinical medicine, with its core principle centered on digital twin-driven precise intervention. This chapter, structured around the framework of "biological adaptation—surgical empowerment-device customization-educational loop", reveals how interdisciplinary innovations such as porous structure design (gradient optimization of pores), biological activity regulation (ion/factor delivery), and mechanical biomimetic simulation (force feedback model) are working in concert to drive orthopedic diagnosis and treatment towards standardization and predictability.

3D Printing-based repair and regenerative technology for bone defects

In the field of bone defect repair, 3D printing technology has propelled the advancement of regenerative medicine through multidisciplinary innovations [106, 107]. Yan et al. pioneered the construction of a PLA scaffold integrated with bionic vascular channels, combined with BMP-2 delivery, which enhanced new bone formation in rat femoral defects by approximately three times (Fig. 12A, a1) [343]. Furthermore, Deng’s team developed a dual-functional strontium-zinc phosphate ceramic scaffold that demonstrated excellent antibacterial properties and significantly enhanced the healing rate in rabbit models with radial bone defects (Fig. 12A, a2) [49]. To enhance biological activity, Shen et al. engineered adaptive nano-topographical structures on the surface of PCL scaffolds, leading to an increased bone coverage rate at rat cranial defects within four weeks (Fig. 12B, b1) [262]. Wang et al., on the other hand, functionalized the scaffold with gallium ions to regulate bone homeostasis, thereby achieving effective treatment of infectious tibial defects in rabbits and a modest restoration of bone density (Fig. 12B, b2) [314]. For complex anatomical reconstruction, Fu et al. utilized personalized, custom-made porous titanium alloy acetabular reinforcement structures, which significantly improved load-bearing capacity in a porcine model (Fig. 12C) [67].

Fig. 12.

3D-printed scaffolds and implants for bone regeneration and repair. A: Schematic and histological analysis of DFO-bridged 3D-printed PCL scaffolds showing enhanced bone regeneration and upregulated expression of HIF-1α, CD31, OXS, and OPN (a₁) [343]. Additionally, 3D-printed SZP and β-TCP scaffolds with antibacterial efficacy and enhanced new bone formation in a cranial defect model (a₂) [49]. B: Schematic of bio-inspired mineralization for 3D-printed PCL scaffolds and evaluation of in situ bone regeneration in rabbit models at 8 and 16 weeks (b₁) [262]; Characterization and evaluation of PCL/MBG/Ga and PCL/MBG scaffolds, showing nearly complete bone repair with the PCL/MBG/Ga scaffold and histological assessment of an infected bone defect model (Masson, TRAP, and Giemsa staining) (b₂) [314]. C: Schematic of 3D-printed porous Ti6Al4V implant design, micro-CT imaging, and evaluation of implant stability via shear strength testing at 0 and 12 weeks. Histomorphometric analysis quantified new bone formation and ingrowth using fluorochrome labeling (tetracycline and calcein green) [67]

Significant advancements have been achieved in the field of biomaterial systems. Bashiri et al. applied placental ECM composite scaffolds to enhance the bone volume fraction in rabbit femoral defects, reaching a value of 0.58 (Fig. 13A, a1) [14]. Miao’s team integrated black phosphorus nanosheets with DNA hydrogels, thereby promoting concurrent increases in vascular density and new bone thickness (Fig. 13A, a2) [199].

Fig. 13.

3D-printed hydrogel and polymer scaffolds for bone defect regeneration. A: Schematic of 3D-printed hydrogel scaffold preparation. Compared to the control, both 0% and 5% ECM-SF/Alg scaffolds promoted osteogenic marker (ALP, Runx2, OCN) expression and collagen synthesis at 4 and 8 weeks (a₁) [14]. Additionally, expression of matrix components and marker proteins in bone defects after implantation of gelatin-based scaffolds. Arrows indicate neovascularization (NV) and new bone (NB); yellowish-brown areas show positive staining for OCN and CD31 (a₂) [199]. B: Customized PCL scaffolds with ~ 500 μm pores were fabricated based on micro-CT data. HE and Masson’s Goldner trichrome staining of Kagome-PCL scaffolds without and with rhBMP-2/HA after implantation in a rat model (b₁) [143]. CAD-designed and 3D-printed gyroscopic frame scaffolds (with SEM characterization). HE and MT staining showed new bone (NB), scaffold (SC), and blood vessels (BV), with black arrows indicating defect edges. Osteogenic capability was assessed by CT imaging at 6 weeks (b₂) [306]

In the field of structural mechanics optimization, the Kagome lattice scaffold developed by Ku et al. demonstrated a compressive strength of 12.7 MPa, closely approximating the mechanical properties of cancellous bone (Fig. 13A, b1) [143]. Verisqa’s Gyroid scaffold based on isosorbide represented a more innovative approach by enabling bone regeneration without the need for exogenous growth factors (Fig. 13A, b2) [306]. These advancements provide a solid foundation for clinical translation. Slavin established the gold standard for craniofacial repair using porous materials with > 60% porosity and pore sizes ranging from 300 to 500 μm [270]. Tong’s photo-integrated nHA technology significantly improved the interfacial bonding strength between the scaffold and bone to 8.4 MPa [298]. Moreover, Ma’s magnesium ion-modified tantalum scaffold achieved a bone density of 1.15 g/cm³ in sheep tibial defects [191], collectively representing a critical advancement toward the clinical application of large bone defect repair.

Precise surgical planning and treatment assistance through 3D printing technology

3D printing technology is systematically transforming the planning and execution of orthopedic surgeries. In the preoperative planning phase, de Klerk et al. enabled the visualization and accurate classification of complex elbow fracture types using 3D-printed models, which increased surgeons’ diagnostic accuracy from 68 to 92% and reduced intraoperative reduction time by 40% [47]. Meanwhile, the Fu team leveraged finite element analysis to optimize the design of patient-specific acetabular bone grafts, thereby decreasing peak stress at the bone-implant interface by 37% and mitigating the risk of mechanical failure [66]. For precise intraoperative procedures, Park et al. developed a bone tumor resection guide plate that achieved a margin error of less than 1.5 mm, resulting in a 100% negative margin rate across seven limb-sparing surgeries [223]; Tomaževič et al.’s anatomically matched titanium alloy locking plate increased torsional strength by 52% and completely eliminated the problem of screw loosening [294]. Yiannakopoulos et al. confirmed that a 3D model of the glenoid (with an error of 0.38 ± 0.12 mm) could precisely guide prosthesis selection [346]. In the realm of surgical skills training, Bohl et al. developed a sensor-integrated spinal osteotomy model that reduced clinical operation time by 25% [20], while Lubet et al.’s training system for supracondylar fractures of the humerus in children decreased intraoperative fluoroscopy by 50%, jointly establishing a closed-loop training system from theoretical cognition to practical operation ability [183].

Development and assessment of 3D-printed patient-specific bone implants

The development of 3D-printed bone implants has evolved into a systematic solution driven by collaborative innovation across "materials, structure, and function." In the domain of structural bionic design, Fu et al. enhanced bone ingrowth and mechanical support through gradient-pore titanium alloy implants, with pore sizes of 600 μm on the surface and 300 μm in the core. This design achieved a 78% bone ingrowth rate and an interface shear strength of 18.3 ± 2.1 MPa in a porcine acetabular defect model [67]. Building on this, Xiu’s team applied micro-arc oxidation to fabricate a Ca/P nano-coating on the implant surface, thereby increasing the bone bonding strength by 58%, reaching 45.3 N [338]. For bioactive functionalization, Ma et al. developed a magnesium ion-modified tantalum scaffold that enhanced vascularization through a 3.1-fold upregulation of VEGF expression, significantly increasing bone density in the ovine tibial defect region to 1.15 g/cm³, representing a 47% improvement [191]. Wang et al., on the other hand, demonstrated the anti-osteoporotic efficacy of strontium zeolite-coated titanium implants, which resulted in a 40% increase in periprosthetic bone mass [318]. In the field of individualized and precise matching, Zhang et al. designed an integrated customized acetabular prosthesis for DDH patients, successfully restoring the anatomical rotation center (with an error < 1.5 mm), and the Harris score jumped from 48 to 89 [355]. These innovations were ultimately validated through long-term clinical follow-up. The craniofacial implant criteria established by Slavin—defined as a porosity > 60% and pore size of 300–500 μm—demonstrated a 96.2% survival rate over a 5-year period [270]. Driscoll’s ceramic-bone matrix composite spinal scaffold achieved a 100% bony fusion rate [59], while Phuoc et al. quantitatively demonstrated the superior bone ingrowth capacity of titanium alloy with a 500 μm pore size [230]. This system represents a comprehensive, full-cycle development paradigm for bone implants, spanning from biomimetic design to clinical translation.

Highly realistic 3D-printed bone models enhance precision in orthopedic education and surgical skill training.

3D-printed models are advancing orthopedic skills training into a systematic and tiered framework. At the foundational stage of operational safety training, Yuan et al. employed a layer-colored PLA temporal bone model (cortical white/air cells red) with real-time visual feedback, significantly enhancing drilling depth control. This approach reduced the simulated facial nerve injury rate among trainees (n = 25) by 76% compared to conventional methods, thereby establishing a robust safety baseline for novice otological procedures [351]. Moving to the intermediate phase focused on routine fracture management, Zhang’s team developed a modular TPU-PLA composite model that simulates limb fractures with concurrent nerve and vascular injuries. In a randomized controlled trial, this model improved trainees’ fracture reduction scores by 32.7%, particularly in complex cases such as Colles fractures [367]. To address technical challenges in pediatric orthopedics, Lubet et al. introduced a supracondylar humeral fracture training system incorporating silicone ligaments and a repositionable distal humerus structure. This system elevated anatomical reduction success rates among residents (n = 20) to 95% and reduced intraoperative fluoroscopy usage by 50%, effectively addressing the demand for precision in pediatric fracture care [183]. Finally, for the preparation of high-risk procedures, Bohl et al. embedded a force-sensing real-time alert system into a photopolymerized spinal model, enabling precise quantification of osteotomy depth risk (error < 0.3 mm). This innovation contributed to a 40% reduction in postoperative nerve injury rates among trainees (n = 12), offering a safe and forgiving platform for mastering high-risk spinal osteotomies [20]. These four categories of systematic models collectively establish a stepwise skill development framework, forming a closed-loop training system that progresses from operational safety awareness to technical proficiency and ultimately to risk prediction capabilities.

Expanding applications of 3D printing models in breast cancer research

Breast cancer research on tumor heterogeneity, metastatic cascades, and drug resistance mechanisms necessitates a multi-scale biomimetic platform that goes beyond the limitations of traditional models. 3D printing technology, by enabling cross-scale fabrication from molecular to cellular and tissue levels, is transforming the research paradigm: from improving the accuracy of preclinical drug screening through breast-derived dECM-based pathological models, to achieving intelligent and controlled release via in situ functionalization of nanomaterials, and further advancing personalized research through organoid integration. The synergistic integration of these three components establishes a closed-loop system encompassing "mechanistic elucidation – precise intervention – clinical translation," providing a transformative impetus for breast cancer research.

3D Printing-based models for preclinical drug screening in breast cancer

Traditional drug screening models face significant limitations due to their inability to replicate the complex heterogeneity and microenvironmental interactions characteristic of breast cancer. Although many bioprinting strategies were initially developed for breast cancer broadly, their translational value is particularly important for TNBC, where the lack of canonical therapeutic targets increases reliance on microenvironment-faithful models. For clarity, representative therapeutic strategies discussed in this section are categorized by evidence stage.

To avoid over-generalized translational claims, we distinguish between mechanistic proof-of-concept findings, preclinical in vivo validation, early-phase clinical relevance, and clinically established interventions. This distinction is important because not all biologically promising platforms or drug-delivery strategies have yet demonstrated robust translational readiness, and the evidentiary strength should therefore be stated explicitly. In the present section, most examples remain at the in vitro proof-of-concept stage, whereas only a smaller subset can be discussed in relation to in vivo validation, early clinical correlation, or established therapeutic benchmarks.

Representative strategies are summarized in Table 6. Unless otherwise specified, the examples discussed below should be interpreted as in vitro proof-of-concept studies; references to in vivo validation, early clinical correlation, or approved therapies are included to indicate translational position rather than equivalent maturity. In the following discussion, examples are described according to their evidentiary context (in vitro, in vivo, early clinical correlation, or clinically established reference) where applicable.

Table 6.

Evidence-stage classification of representative therapeutic strategies discussed in Sect. "3D Printing-based models for preclinical drug screening in breast cancer"

Strategy	Representative platform/model	Main purpose	Evidence stage	Limitation for translation
dECM-based TNBC bioprinted hydrogel model	MDA-MB-231 cells encapsulated in breast-derived dECM hydrogel	Preserve tumor-specific ECM features and improve drug-response relevance	In vitro	Lacks systemic immune, vascular, and pharmacokinetic context
CAF–tumor spatial co-culture model	Bioprinted CAF-integrated tumor spheroid platform	Reconstruct stromal chemoresistance and tumor–stroma interaction	In vitro	Simplified multicellular composition and limited long-term heterogeneity
Microfluidic endothelial-barrier extravasation model	Bioprinted or microfluidic vascular barrier model for CTC passage	Quantify vascular escape and early metastatic colonization behavior	In vitro	Does not capture full metastatic progression or whole-body dissemination
Co-bioprinted immune–tumor response model	CD8⁺ T cell–tumor co-bioprinted model linked to PD-1 inhibitor response	Estimate immunotherapy-related response patterns and in vitro–clinical correlation	Early-phase clinical relevance	Correlation with clinical response does not equal prospective clinical validation
PDXO/orthotopic validation-linked platform	Patient-derived organoid/xenograft-linked drug testing workflow	Validate whether in vitro sensitivity signals remain predictive in vivo	In vivo	Low throughput, higher cost, and persistent species/context differences
Clinically established therapeutic benchmarks	Approved breast cancer therapies used as translational comparators	Provide clinical context for judging the maturity of preclinical platform outputs	Approved intervention	The platform may benchmark or predict response, but does not itself establish clinical efficacy

3D bioprinting technology has overcome these limitations through multi-dimensional biomimetic strategies: First, by encapsulating triple-negative breast cancer cells, such as MDA-MB-231, within decellularized extracellular matrix (dECM) hydrogels, tumor-specific extracellular matrix components—such as laminin and collagen IV—are preserved, thereby significantly enhancing the physiological relevance of drug testing [112]; secondly, by integrating cancer-associated fibroblasts (CAFs) with tumor spheroids through spatial co-localization technology, it precisely recreates the chemotherapy-resistant microenvironment mediated by CAFs [109]; more notably, the dual-phase scaffold design strategy—combining an upper layer of breast tissue simulation (collagen/hyaluronic acid hydrogel) with a lower layer of bone metastasis biomimetic layer (nano-hydroxyapatite/poly(lactic-co-glycolic acid)), successfully simulates the dynamic process of breast cancer cell colonization of bone trabeculae [292], while the model integrated with a microfluidic vascular endothelial barrier can quantitatively detect the extravasation efficiency of circulating tumor cells (CTCs), a core event in breast cancer bone metastasis [299].

These technological innovations have directly facilitated the development of a precision drug evaluation system. In drug penetration analysis, laser confocal dynamic imaging has demonstrated that the penetration depth of trastuzumab in HER2⁺ breast cancer organoids is 2.3 times greater than that observed under static culture conditions, highlighting the limitations in targeted delivery of antibody-based therapeutics [145]. In the field of immunotherapy validation, a co-bioprinted model of CD8⁺ T cells and tumor cells revealed a strong positive correlation (R² = 0.85) between T cell activation efficiency mediated by PD-1 inhibitors and clinical response rates [305] (in vitro–clinical correlation). For drug resistance monitoring, a fluorescence reporter gene-labeled MCF-7 cell model enables real-time tracking of the spatiotemporal expression dynamics of P-glycoprotein (P-gp) during doxorubicin treatment, offering valuable dynamic pharmacodynamic data for overcoming drug resistance [132].

This platform has demonstrated significant value in validating targeted breast cancer therapies. For HER2-targeted treatments, 3D-printed breast duct structure models have revealed that the endocytosis efficiency of T-DM1 is limited by the HER2 internalization rate, elucidating a novel mechanism underlying trastuzumab resistance [1] (in vitro; in DNA repair-targeted therapy, BRCA1 mutant organoid models confirmed that olaparib combined with radiotherapy can induce a synergistic lethal effect (increasing the apoptosis rate by 60% [28] (in vitro). Breakthroughs have also been made in the innovative validation of combination therapies: a dual-material printed tumor-lymphoid tissue interface model proved that CXCL12 inhibitors can promote T-cell infiltration (increasing the number of CD8⁺ cells by 3.1 times) [352] (in vitro; while a hypoxic zone model of TNBC revealed that metformin reverses paclitaxel resistance through inhibiting glycolysis, identifying a key pathway [210] (in vitro).

However, mechanistic promise should be interpreted with caution, as not all metabolically targeted strategies have translated into clinical benefit. Although metabolic checkpoint inhibition has shown encouraging immunoregulatory effects in preclinical studies, late-stage clinical failures indicate that efficacy is likely context-dependent and may require biomarker-guided patient stratification. This is precisely why the examples discussed in this section are interpreted according to evidence stage rather than being presented as uniformly translationally mature.

Nano-delivery systems in 3D printed models: applications and advancements

The integration of nano-delivery systems with 3D printing technology has established a high-precision and functionally advanced platform for controlled drug release in breast cancer treatment. A current research focus lies in the in situ integration of functional nanomaterials. Fantino et al. demonstrated that conductive silver nanoparticles (AgNPs) can be directly synthesized within 3D-printed structures through the in situ reduction of silver salts in photopolymerizable resins, thereby imparting antibacterial properties to the scaffold [64]. Meanwhile, the Pawar team developed water-dispersible photoinitiator nanoparticles, such as TiO₂@carbon quantum dots, enabling high-resolution hydrogel fabrication while effectively eliminating the cytotoxic risks associated with conventional small-molecule photoinitiators [227]. This class of technology provides a foundational framework for the on-demand functionalization of drug-eluting stents. For example, Zhang et al. integrated copper nanoparticles with micropowder in a composite printing process, which substantially improved the material’s mechanical strength and thermal conductivity (Zhang, L. et al., 2022), making it highly suitable for thermally responsive drug delivery systems.

In the development of targeted delivery systems, 3D printing enables the spatial programming of nanocarriers. For instance, the Nishiguchi group employed stereolithography to fabricate antigen nanoparticles with precisely controllable structures (diameter: 50–200 nm), demonstrating that topological features can regulate the activation efficiency of dendritic cells [213]. In parallel, Zhang et al. integrated liquid metal nanoparticles (gallium-indium alloy) with thermosensitive polymers to create near-infrared-responsive 4D-printed soft robots capable of achieving on-demand drug release at tumor sites [108, 182, 358]. Dual-level release scaffolds have emerged as a promising strategy for treating breast cancer bone metastasis: Cai et al. fabricated elastic nanofiber composite scaffolds composed of PLGA and hydroxyapatite nanowires by integrating electrospinning with 3D printing technology. These scaffolds enable the simultaneous delivery of doxorubicin and bone morphogenetic protein-2 (BMP-2), effectively inhibiting osteolytic destruction while promoting bone regeneration [24].

The integration of microfluidic and printing technologies has further advanced the spatiotemporal precision of nano-delivery systems. The Gao team embedded citrus-derived extracellular vesicles (CLEVs) into gelatin methacryloyl (GelMA)-based bioprinted scaffolds to encapsulate doxorubicin, enabling both local sustained release (over 7 days) and systemic administration. The system was validated through co-culture with TNBC organoids to demonstrate the synergistic effects of sustained release. This pioneering platform leverages plant-derived extracellular vesicles as carriers, reducing triple-negative breast cancer cell viability by 60% (with a threefold increase in ROS levels), while simultaneously promoting fibroblast healing by enhancing adhesion force by 50%. The approach significantly mitigates the systemic toxicity of chemotherapy and offers a novel, highly biomimetic therapeutic strategy for TNBC [71]. In parallel, Hou et al. developed a layered hydrogel-based skin transdermal model via 3D bioprinting to quantitatively assess the permeation kinetics of liposomal nanoparticles [105]. Advances have also been achieved in intelligent responsive systems: Bodkhe et al. employed solvent evaporation-assisted 3D printing to fabricate piezoelectric PVDF/ZnO nanowire composite structures capable of triggering paclitaxel release upon ultrasonic stimulation [19].

In the realm of clinical translation, 3D-printed nanosystems are advancing personalized therapeutic strategies. Saraiva et al. engineered magnetic Fe₃O₄ nanoparticle-antibiotic composite calcium phosphate scaffolds capable of targeting bone metastases under an external magnetic field and modulating the drug release profile [252](preclinical/early translational). Meanwhile, the Kool team developed dichroic gold nanocomposite resins for the fabrication of breast cancer surgical guides, offering both radiographic visibility and drug delivery capabilities [140](early clinical/translational). However, key challenges remain in the standardization and biosafety of nanomaterial-based printing. Mohammadian et al. emphasized the necessity of establishing guidelines to mitigate occupational exposure risks associated with metal nanoparticles [205]. In response, Ghazali introduced water-soluble photoinitiator nanoparticles (IC50 > 200 μg/mL), presenting a promising approach for biocompatible and safe printing processes [77].

The synergistic integration of 3D printing models and organoid technology

Organoid technology, characterized by its patient-specific nature and functional fidelity, has emerged as a transformative paradigm in breast cancer research. Organoids are three-dimensional micro-organ constructs that self-assemble from stem cells or primary tissue cells under in vitro conditions. Their fundamental value resides in their capacity to recapitulate the critical cellular composition, spatial organization, and physiological functions—such as metabolic activity, electrophysiological responses, and drug sensitivity—of native organs. Consequently, organoid technology has established itself as a powerful platform in biomedical research. These models are typically engineered using biomaterial scaffolds (such as Matrigel) in combination with biochemical signaling cues. Their applications span four major domains: first, the investigation of disease mechanisms—such as employing glioblastoma and brain aging organoid models to elucidate pathological processes; second, high-throughput drug screening and toxicity assessment; third, personalized medicine, wherein patient-derived cells are used to develop individualized models that inform clinical decision-making; and finally, developmental biology, where organoids serve as a model system to study organogenesis.

Consequently, organoid technology has established itself as a powerful platform in biomedical research. These models are typically engineered using biomaterial scaffolds (such as Matrigel) in combination with biochemical signaling cues. Their applications span four major domains: first, the investigation of disease mechanisms—such as employing glioblastoma and brain aging organoid models to elucidate pathological processes; second, high-throughput drug screening and toxicity assessment; third, personalized medicine, wherein patient-derived cells are used to develop individualized models that inform clinical decision-making; and finally, developmental biology, where organoids serve as a model system to study organogenesis.

To address the challenge of standardization difficulties in organoid fabrication, high-precision printing technologies—such as customized microcolumn arrays and microfluidic chips—enable precise control over organoid morphology and size, thereby substantially improving experimental reproducibility and minimizing batch-to-batch variability [126]. To counteract the critical limitation of inadequate vascularization, the incorporation of embedded vascular channels combined with the use of biomimetic bioinks—such as GelMA and hyaluronic acid—facilitates directed endothelial cell growth, thereby enhancing nutrient and oxygen transport, prolonging organoid viability, and increasing physiological relevance. Additionally, microcolumn arrays and microfluidic platforms enable precise morphological control, achieving size variation coefficients below 5% [249]. To overcome throughput constraints, the integration of automated multi-well plates with microfluidic systems through printed constructs allows for parallelized, large-scale organoid culture, enabling high-throughput drug screening with capacities exceeding 1,000 organoids per plate [60, 129]. Finally, to address functional limitations, 3D printing enables the direct integration of functional elements—such as liquid metal electrodes and biosensors—to enable real-time monitoring of electrophysiological signals and metabolic activities, thereby significantly enhancing the physiological fidelity of organoid models [134, 177]. Furthermore, this technology drives advancements in biomaterial innovation—through the development of Matrigel-free decellularized extracellular matrix (dECM) scaffolds and tunable stiffness collagen-alginate microgels, it effectively eliminates batch variability, reduces production costs, and facilitates clinical translation [34, 220]. More significantly, 3D printing has enabled synergistic breakthroughs: for example, fabricating multi-organ-on-a-chip systems that interconnect liver, brain, and intestinal organoids to mimic systemic physiological interactions, or employing 4D smart materials such as magnetically responsive liquid metals to achieve dynamic control over the culture microenvironment. These innovations pave novel pathways for the construction of complex, physiologically relevant human tissue systems.

3D printing technology is systematically overcoming the core challenges in organoid research, with groundbreaking advancements spanning the entire pipeline from fundamental construction to clinical translation. In the domain of morphological control, Graham et al. developed a free-form "MAGIC bioink" that facilitates complex morphogenesis in intestinal and brain organoids [82], whereas the Kadotani team achieved pre-assembly of organ-specific structures through geometric engineering units [126]. Regarding functional monitoring, Kim et al. introduced a magnetically controlled reconfigurable liquid metal multi-electrode array for real-time tracking of neural activity in brain organoids [134], a technology later adapted by Lee for electrophysiological analysis of retinal organoids [150]. In the realm of high-throughput production, Kang et al. engineered vascularized GelMA hydrogel scaffolds with embedded HUVEC channels, which developed functional vascular networks (CD31⁺ coverage exceeding 80%) within seven days, significantly alleviating hypoxia in the core regions of TNBC organoids (pO₂ increased threefold) [129]. Concurrently, Du’s team achieved a throughput capacity at the ten-thousand scale through a miniaturized system [60]. Concerning vascular integration, Salmon fabricated neurovascular organoids with blood–brain barrier functionality using microfluidic chip-based bioprinting [249], while Bernal applied volumetric bioprinting to generate photo-crosslinked liver metabolic models incorporating vascular networks [16]. In the context of precise disease modeling, Cadena employed magnetically responsive 4D materials to dynamically modulate micro-mechanical stiffness, mimicking the biomechanical microenvironment at the breast-to-bone metastasis interface [23], Clark utilized immersion bioprinting to preserve patient-specific mutational profiles in brain tumors [41] and Shi maintained breast cancer heterogeneity through low-concentration collagen embedding techniques [265]. Finally, Drabbe’s biomimetic oxygen gradient chip and Jiang’s automated bioprinting platform respectively advanced organoid industrialization by optimizing microenvironmental control and enabling standardized large-scale production [57, 118].

The integration of 3D printing with organoid technology has moved beyond simple technological combination, driving biomanufacturing from "structural replication" toward "functional creation"—a paradigm shift aimed not merely at mimicking tissue morphology, but at engineering physiologically autonomous "living" organs. This synergistic approach has yielded significant advancements in diverse applications, including disease modeling (e.g., patient-specific tumor drug sensitivity testing), pharmaceutical development (high-throughput screening platforms), and mechanistic studies (e.g., real-time monitoring of neural development). Notably, 3D printing effectively overcomes two fundamental limitations of conventional organoid systems: inadequate vascularization and poor standardization. However, two major contradictions persist in current organoid-bioprinting integration. First, while printed structures increasingly approximate physiological morphologies, cellular maturation remains incomplete—Cadena’s cortical organoids, for example, recapitulate layered architecture but exhibit neuronal electrical activity maturity below 30% of in vivo levels [5]. Second, standardized protocols struggle to accommodate the high heterogeneity of clinical samples. Future advances should focus on the "integration of life processes through multi-organ chips," particularly by achieving synchronized organ development timelines and overcoming the limitations of "immune-deficient" organoid systems. A promising direction involves embedding 3D-printed lymphoid tissues into breast cancer models to better simulate the therapeutic effects of immune checkpoint inhibitors. In addition, the development of smart material-driven functional transitions represents a highly promising research direction. This approach enables the creation of organoids capable of dynamically modulating their microenvironment to promote cellular maturation. Furthermore, the bioprinting of bionic bile duct networks offers a strategy to integrate multiple functional modules, thereby establishing a closed-loop system for sequential drug metabolism, toxicity detection, and detoxification response [16].

Critical limitations and unresolved translational contradictions

Despite the rapid expansion of organoid–bioprinting integration, several important contradictions remain unresolved. First, structural sophistication does not necessarily translate into functional maturity; a model may reproduce tissue architecture while still lacking physiologically relevant signaling dynamics, vascular barrier function, or immune competence. Second, highly standardized fabrication protocols often conflict with the need to preserve patient-specific heterogeneity, which is particularly important in TNBC. Third, strategies that show strong mechanistic promise in controlled preclinical settings may remain difficult to translate because of manufacturing complexity, cost, limited throughput, or uncertain regulatory positioning. We therefore revised the manuscript to present these technologies not only as opportunities but also as systems with important biological and translational constraints.

Significance and future perspectives

This section summarizes the translational significance, practical barriers, and future development priorities of TNBC-oriented bioprinted models.

Translational significance of TNBC-oriented bioprinted models

The study of TNBC bone metastasis has long been constrained by the static and low-throughput nature of conventional models, which inadequately capture the dynamic tumor–bone vicious cycle and the broader tumor–bone–immune interaction network [163]. In this review, we argue that 3D bioprinting offers a particularly promising framework for addressing these limitations because it enables controlled reconstruction of multicellular architecture, matrix mechanics, vascular features, and therapeutic exposure[69, 202]. Rather than reporting a newly developed model, this section synthesizes the translational significance, current limitations, and future development priorities of TNBC-oriented bioprinted systems. Recent advances in 3D bioprinting suggest that TNBC bone-metastasis models can be designed to capture key features of the tumor–bone vicious cycle, including osteolytic signaling, stromal interaction, and immune modulation. In particular, the ability to combine tumor cells with osteoblasts, osteoclasts, endothelial cells, and immune-related components offers a more informative platform for mechanistic analysis and treatment evaluation than conventional reductionist systems. An important advantage of these platforms is their potential to reduce reliance on oversimplified in vitro systems and, in some settings, partially mitigate the species-specific limitations of animal models; however, this translational value depends strongly on model standardization, reproducibility, and validation [125, 163].

Standardization of model inputs and readouts

A major obstacle to broader adoption is the lack of standardization in bioink composition, printing parameters, cell-source selection, and assay endpoints. Even models intended to represent the same biological scenario may differ substantially in matrix formulation, stiffness range, crosslinking strategy, culture duration, and biological readouts. For TNBC bone-metastasis modeling, a minimum reporting framework should include matrix composition, stiffness characterization, vascularization strategy, multicellular composition, and core outputs such as osteoclast activation, invasion behavior, and treatment response [202, 288].

Reproducibility across laboratories

Reproducibility remains a distinct challenge from standardization. Even when nominally similar protocols are used, inter-laboratory variation in hydrogel preparation, printing hardware, crosslinking intensity, cell handling, and post-print culture can substantially alter biological outcomes. This is particularly problematic in TNBC models, where small differences in matrix density, immune-cell proportion, or oxygen availability may change invasion, EMT-associated signaling, and drug sensitivity. More explicit benchmarking and multi-site validation will therefore be necessary before these systems can be considered robust translational tools [69, 163].

Scalability and regulatory positioning

Scalability is another major concern because platforms that perform well at low-throughput laboratory scale may be difficult to translate into standardized screening workflows or clinically linked applications. Manufacturing time, cost, operator dependence, batch-to-batch variability, and compatibility with automated analysis all affect translational feasibility. In parallel, regulatory evaluation will depend on how a platform is positioned: as a research model, a companion diagnostic tool, or an implantable product. These categories differ markedly in their requirements for safety control, material traceability, quality assurance, and validation [202].

Future integration with AI-assisted design and adaptive biofabrication

AI-assisted biofabrication may gradually shift model development from largely experience-driven optimization toward more data-guided design and control. Potential applications include AI-guided optimization of vascular-network geometry, print-path planning, and dynamic adjustment of crosslinking or material parameters according to imaging and sensor-derived data [369].

4D printing and stimulus-responsive materials may further expand the utility of TNBC bone-metastasis models by enabling dynamic modulation of local stiffness, pH responsiveness, or drug-release behavior over time. However, these systems should currently be regarded as promising developmental directions rather than mature translational solutions, because their long-term stability, manufacturing complexity, and validation requirements remain substantial [128, 288].

Taken together, AI-assisted design and adaptive biofabrication may strengthen future model personalization and process control, but their true translational value will depend on rigorous validation, reproducibility, and clinically meaningful use cases [369].

Author contributions

This article was primarily authored by Jiahong Xu and Yue Kang, with Kedong Song, Yefu Liu and Jiangli Fan playing pivotal roles in shaping the overarching direction and meticulously refining the manuscript.

Funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. This research is supported by “the Liaoning Provincial Key Laboratory of Precision Medicine for Malignant Tumours” (No.: 2021JH13/10200041), the Fundamental Research Funds for the Central Universities (DUT25YG106), the Open Fund of Key Laboratory of Biotechnology and Bioresources Utilization (Dalian Minzu University), Ministry of Education (KF2025007) of China, the Foundation of Engineering Research Center of Innovative Drug of Traditional Chinese and Zhuang Yao Medicine, Ministry of Education (ZYZYY2025013) and Shenzhen Basic Research Program (JCYJ20220531101805012).

Data availability

No data was used for the research described in the article.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.