Engineered Niches to Analyze Mechanisms of Metastasis and Guide Precision Medicine

Abstract

Cancer metastasis poses a challenging problem both clinically and scientifically, as the stochastic nature of metastatic lesion formation introduces complexity for both early detection and the study of metastasis in preclinical models. Engineered metastatic niches represent an emerging approach to address this stochasticity by creating bioengineered sites where cancer can preferentially metastasize. As the engineered niche captures the earliest metastatic cells at a non-vital location, both non-invasive and biopsy-based monitoring of these sites can be performed routinely to detect metastasis early and monitor alterations in the forming metastatic niche. The engineered metastatic niche also provides a new platform technology that serves as a tunable site to molecularly dissect metastatic disease mechanisms. Ultimately, linking the engineered niches with advances in sensor development and synthetic biology can provide enabling tools for preclinical cancer models and fosters the potential to impact the future of clinical cancer care.

Introduction

Currently, cancer often goes undetected until identified by radiological studies, or the patient becomes symptomatic. While improvements in cancer screening, diagnostics, and treatments have decreased overall cancer mortality by approximately 30% since 1991 (1), over 90% of cancer deaths are due to the development of metastatic disease (2). Consequently, a critical need exists to improve metastasis-specific diagnostics and therapeutics. However, such research is limited by the stochastic nature of the initially microscopic metastatic lesions. A gap exists in current technologies for the early identification and analysis of metastasis in both preclinical models and clinical practice.

Synthetic metastatic niches represent an emerging technology for the mechanistic interrogation of the metastatic microenvironment. This tool can also have implications for the development of cancer diagnostics. The synthetic niches provide a pre-defined location for metastatic events, enabling mechanistic studies of dormancy and disease progression and, when implanted subcutaneously, act as readily accessible sites for clinical monitoring. Bioengineered metastatic niches support a platform for non-invasive imaging, cell collection, histological analysis, and biomarker identification that could be a rich source of information in clinical and preclinical settings (Figure 1). Importantly, synthetic metastatic niches reflect many aspects of metastatic disease observed in the native metastatic niche, e.g. the recruitment of an aggressive population of tumor cells similar to those in lung metastases (3).

Figure 1:

Engineered diagnostic sites contribute unique information to current clinical paradigm and emerging technologies.

In this review, we provide an analysis of engineered niches for mechanistic investigations of metastasis. Furthermore, we describe the potential of these niches to integrate within current clinical practice. The intersection between cancer biology and engineering is becoming ever more intertwined, and new engineered technologies will enable transformative insights and provide a means to screen novel therapeutics.

Bioengineered Niches for Metastatic Insights

Engineered metastatic niches represent a technological opportunity to study extravasation and colonization of metastatic sites (4). A century ago, Paget postulated the “Seed and Soil” hypothesis, in which he recognized that the distribution of metastatic cells to specific secondary organs was initiated by a favorable environment in those organs (5). Decades of work have helped to elucidate that the “soil” of these environments was altered by signaling from the primary tumor to create a pre-metastatic niche (PMN), and that mobilization and extravasation of bone-marrow derived cells played a pivotal role in this process (6). Secreted factors and exosomes from primary tumors enabled this mobilization and facilitated the formation of the PMN (7–11). The stochastic nature of pre-metastatic and metastatic lesions challenges the ability to analyze the dynamics and molecular mechanisms of localized sites at early time points in disease progression. Most studies isolate entire organs, or sections of organs with metastatic lesions, resulting in confounding data associated with the adjacent healthy tissues. Synthetic niches, to which tumor cells metastasize, provide tunability to enable enrichment or depletion of particular factors within a PMN, to examine the role of specific cell types in the microenvironment, tumor cell recruitment, and phenotypes of invading tumor, immune, and stromal cell populations. In vitro models have provided insights into the initial alterations at metastatic sites, with mechanisms subsequently validated in vivo (12–16). However, in vitro models cannot fully recapitulate the in vivo environment, and thus the synthetic niche may provide an alternative or complementary approach for studying mechanisms of metastasis.

The biological component of the engineered niche is comprised of infiltrating cells that contribute to the formation of connective tissue, a vascular network, and immune cells that are recruited as part of the foreign body response to an implanted material (17). The presence of disease alters the immune system, and these alterations are reflected within the niche. The niche is formed by conditioning of immune cells from the vasculature, which ultimately attracts aggressive metastatic tumor cells (3,18–20). In a murine breast cancer model, tumor inoculation induced an influx of Gr1^hiCD11b⁺Ly6C⁻ myeloid derived suppressor cells, accompanied by decreases in CD11c⁺ dendritic cells (DCs) and F4/80⁺CD11b⁺ macrophages (21). Single-cell RNA-seq performed on the immune populations in synthetic niches showed gene expression by multiple cell types that were associated with immunosuppressive, pro-tumor phenotypes. This pro-tumor environment was illustrated through the up-regulation of S100a8 and S100a9, facilitating T cell suppression and colonization of metastatic cells in vital organs (22–24). In this context, the niche was formed by the progression of immune cell recruitment to a synthetic implant. This non-specific attraction of immune cells from the circulation to the engineered niche led to a suppressive environment formed within the niche when disease was present, which reflects the systemic disease dynamics (22). Furthermore, although engineered niches enable studies of metastasis, these engineered sites may not identically recapitulate metastatic sites. The synthetic niches contain cells and other factors not found in native metastatic lesions (e.g. connective tissue fibroblasts vs lung fibroblasts) that interact with cells recruited from the vasculature. Simultaneously, the tunability of these systems and ease of access positions the engineered metastatic niche as a provocative tool for studying cancer metastasis.

Analysis of metastatic tumor cells captured at a synthetic niche indicated that these cells possessed a migratory and metastatic phenotype, and a transcriptomic profile similar to the metastatic cells that homed to the lung (3). Furthermore, engineered metastatic niches redirect systemic immune cells, altering the primary tumor microenvironment (25). This attraction of tumor-conditioned immune cells and metastatic cells indicated that the engineered niche largely phenocopies a natural metastatic niche (natural MN). As described below, engineered metastatic niches have taken many forms based on the specific question being addressed. These engineered niches range from entirely synthetic implants, to cell-laden sites that require ex vivo culture prior to implantation. The relatively nascent stage of these systems implies that the possible design criteria for specific applications has significant potential, and this potential for plasticity is the focus of a recent review (26). Collectively, the engineered niche thus provides a unique platform to study the interactions between disseminated tumor cells, stroma, and immune cells (27).

Niche-driven insights into factors driving the metastatic cascade

Exosomes shed from the primary tumor are known to modify distal sites including lung tissue and bone marrow, to promote an environment conducive to metastasis (8). Building on this knowledge, incorporating exosomes into engineered MNs increased accumulation of metastatic cells, demonstrating that exosome-derived signals can influence MN function (28).

Extracellular matrix (ECM) proteins also contribute to the capture of metastatic cells, which has been confirmed using synthetic MNs. Polycaprolactone (PCL) scaffolds coated in collagen IV and fibronectin recruit more metastatic cells than uncoated scaffolds (29). Alternatively, cancer-associated fibroblasts encapsulated in retrievable alginate beads produce ECM that enables capture of metastasizing peritoneal cells (30). These platforms identified the functional role of specific ECM components in cancer metastasis, with the potential for use as an engineered cell-capture therapy.

Secreted factors play complex roles in MN formation and are involved in tumor cell attraction, immunomodulation, and ECM deposition. Delivery of factors from the synthetic niches enabled investigation of the role of these proteins in metastatic progression. For example, erythropoietin (EPO) and CXCL12 loaded scaffolds demonstrated that the release of EPO enhanced melanoma cell recruitment, while CXCL12 release had no impact on cell recruitment (31). A separate report showed that CXCL12 delivery increased the metastasis of CXCR4+ melanoma cells to hyaluronan-based engineered niches. These disparate findings suggest that engineered niches can be harnessed to examine unique aspects of MN formation across cancer models (32). Another study implementing synthetic MNs reported that the presence of haptoglobin increased metastatic cell colonization (33). Localized lentiviral gene delivery also induced over-expression of specific cytokines at the implant. Lentiviral overexpression of CCL22, CCL2, CXCL12, and IL10 was performed from the engineered MNs. Gene delivery has been performed both prior to and following tumor cell inoculation. In both cases, delivered factors altered the immune cell composition of the implant. CCL22 increased metastatic tumor cell recruitment, while IL-10 reduced tumor cell recruitment (20,34). Importantly, these studies highlight the mechanistic insights associated with manipulating the synthetic niche. Changing one component, such as CCL2 for IL10, can lead to alterations within the environment that can obscure the contribution that a factor makes to a specific cellular response. For example, lentiviral expression of CXCL12 and CCL2 had no effect on tumor cell recruitment, but CXCL12 delivery increased immune cell recruitment (34). The ability to bias the abundance and phenotype of localized immune cells to screen tumor cell attractors illustrated the diverse capabilities of the engineered niche to dissect the metastatic cascade in vivo.

Hypoxia plays a large role in both primary tumors and metastatic sites as this environment promotes immune evasion, epithelial-mesenchymal transition, and the formation of the MN (35). Hypoxia stabilizes hypoxia inducible transcription factors (HIFs) that modulate many downstream pathways. Importantly, HIF signaling increases expression of lysyl oxidase (LOX) and similar proteins, which lead to alterations in ECM crosslinking and assembly. Elevated LOX expression and the concomitant changes in ECM are important steps in the formation of the PMN (35–38). While the role of hypoxia in cancer has largely focused on primary tumors, the impact of hypoxia on MN function has been analyzed. This study was performed by loading CoCl₂ into synthetic engineered niches, which served to stabilize HIF-1α, and to simulate some aspects of hypoxia. Scaffolds loaded with CoCl₂ mimicked many hypoxic responses for HUVECs in vitro. In vivo, the hypoxic niches exhibited increased vascularity compared to controls, yet did not influence tumor cell recruitment (39). The extent of HIF and LOX induced remodeling of ECM in the synthetic niche has yet to be determined in vivo, but engineered 3D microenvironments have shown cancer cell-directed ECM remodeling in vitro (40). Biomaterials have been developed with aligned paths for metastatic cell migration, indicating niche architecture and ECM-mediated alterations may significantly impact metastasis (41). Collectively, engineering the synthetic niches with specific factors or architectures allows for the function of novel factors to be identified or investigated using several disease models.

Organ-Specific Synthetic Niches

Conditioning of the PMN has been proposed to account for specific cancer types metastasizing to specific organs (e.g. lung, brain, liver, and bone for breast cancer) (42). MN development in the lung is associated with distinctive tumor-conditioned stromal and immune cells (38,43). The ability to engineer the MN allows for the localized presentation of organ-specific ECM proteins, chemokines, and organ-derived cells to analyze their impact on metastatic events and cancer cell recruitment.

In the late 1990s, human bone was implanted into humanized SCID mice to study the metastasis of prostate cancer, identifying that human prostate cancer cells metastasized to human bone but not human lung and intestine fragments (44). This study motivated multiple biomaterial-based models to study metastasis and the leukemic niche within bone marrow (45–52). Biomaterial scaffolds have been developed to mimic bone microstructure, introduce osteogenic cues, and encapsulate osteoblastic cells (48,53–55). These approaches established humanized and tunable bone metastasis models (56). Furthermore, expression of human cytokines reduced tumor burden in hematochimeric mice compared to non-humanized mice (57). A more comprehensive review of tissue engineered bone for modeling malignancies can be found in Mcgovern et al (58).

Lung and liver-mimetic synthetic niches were developed by harnessing ECM derived from metastases of the respective tissues to coat microporous scaffolds (29). ECM from the organs of tumor bearing animals (i.e., diseased ECM) enhanced tumor cell colonization to the synthetic niche, which was further modified to verify myeloperoxidase as a key regulator of tumor cell colonization. Interestingly, the ECM from tumor-free animals did not promote tumor cell recruitment, consistent with the role of conditioning by the primary tumor in the function of natural and engineered niches. Lung tropic tumor cells proliferated at a higher rate on diseased lung ECM, demonstrating the organ and disease-specific influence of the MN. In particular, with further tissue engineering advances, the ability to create an array of organ-specific synthetic niches may be possible. The ability to create a bone, brain, lung, and liver mimic that could all be easily accessed will facilitate greater understanding of the homing of cells to these distinct microenvironments. We expect with improved understanding of the processes driving organotropism, therapies could be designed to interfere with these interactions.

Engineered mimics of metastatic sites are novel tools to enable the cellular and molecular mechanisms of metastatic cancer to be investigated. Studies examining factors driving the metastatic cascade generally have been conducted in breast, melanoma, ovarian, and colon cancer models. Studies of engineered niches and organ-specific metastasis have not typically spanned disease models, and thus the robustness of a scaffold in multiple models of metastatic disease is an area of active study. Furthermore, animal models of cancer, while widely used, are not fully representative of all aspects of human disease. Human cancer cells or patient derived xenografts may better recapitulate some aspects of human disease, yet are typically performed in immunocompromised mice, removing or minimizing the role of adaptive immunity, an important facet of cancer metastasis. Several of the studies reported in this review have used both immunocompetent and immunocompromised mice, supporting that concept that the engineered niches are capturing key components of the metastatic cascade. Nevertheless, pairing of the engineered niches and metastatic models with humanized mice may help to accelerate discoveries most relevant to human metastasis.

Clinical Potential of Engineered Diagnostic Sites

Current Clinical Paradigm

The diagnosis of cancer involves a combination of analyses, including imaging, blood tests, and biopsies to identify tumor presence, blood-borne biomarkers, and tumor phenotype (Table 1). Imaging often provides the first indication of a cancer diagnosis and, once the diagnosis is confirmed, imaging is routinely used to measure changes in tumor size as the disease progresses. However, most imaging technologies can only detect tumors that are approximately 1 cm in size (59). The diagnostic and prognostic utility of such scans is limited, as tumor size is not a direct correlate for aggressiveness, heterogeneity, or response to certain therapies (59). Additionally, while small tumor size is often correlated with Stage I disease, a tumor of 1 cm may have up to 100 million cells (60). By this point, cancer cell heterogeneity can be extensive, with new mutations driving aggressive disease biology and treatment resistance. To this end, modalities that provide an earlier window of opportunity to detect and diagnose disease would provide a means to treat a patient when cancer cell heterogeneity is low, potentially optimizing treatment response. Blood tests and biomarker quantification allow for relatively easy and repeated analyses, yet many blood tests are not malignancy-specific and cannot be used in isolation (61). As an example, CA125 is used for monitoring ovarian cancer and has been effective for determining recurrence, though monitoring has not conclusively led to improved outcomes (61,62) . Markers for breast cancer recurrence (CA-27.29) are not widely applied, as their sensitivity and specificity have not been well established (63). Biopsy of primary tumors and subsequent histological analysis can be valuable for prognosis, defining stage and grade of disease (64). Histopathology of a tumor can identify disease subtype (e.g., receptor status in breast cancer) and aid in initial treatment selection, but this tissue resource cannot be monitored after tumor resection. Biopsy of distal sites can be used to confirm the presence of metastatic disease, and the corresponding disease subtype, which can then be compared to the primary tumor to track disease evolution. Thus, analysis of the MN could be used to inform and guide treatment decisions.

Table 1:

Limitations of the technologies used in the current clinical paradigm

Method	Examples	Limitation	Clinical Need
Imaging	MRI, CT, SPECT, PET and ultrasound	Cannot reliably detect tumors < 1cm3	-Earlier detection -Move beyond volumetric assessments
Blood tests	CEA (colon cancer) CA-125 (ovarian cancer) CA-27.29 (breast cancer) Complete blood count	Not always cancer specific	-Multivariate measurements more specific to malignancy and progression
Histopathology	Cytokeratin (carcinomas) Ki67 (proliferation) ER, PR, HER2 (breast cancer)	Snapshot measurement	-Location to biopsy after resection -Longitudinal, real-time measurements

As next-generation sequencing and computational power increase, the reach of precision medicine into the field of oncology is expanding. Clinical decisions made with current tools are based on population-level results from large clinical trials, providing benefit for many. However, embedded in this paradigm is an empiric philosophy where a treatment is applied, the treatment response measured, and, if indicated, the patient is transitioned to next line therapy. This course can be long and arduous, and individual patient care can suffer at the expense of this treatment process. A critical need exists to move beyond this paradigm, as patients could benefit from a personalized medicine approach capable of detecting early systemic events, improving prognostic stratification, and monitoring response to therapy in real-time. Furthermore, therapy selection is typically guided by information obtained from the primary tumor, which provides perspective on the initial disease biology, yet may not reflect metastatic sites. We propose that engineered niches provide the ideal platform to expand upon current diagnostic capability.

Biopsy for metastatic diagnostics

Metastatic tumor cell detection and analysis:

Research into early metastatic diagnostics has focused largely on liquid biopsy as an alternative to clinical diagnostics and therapeutics based on the primary tumor. Most of this effort stems from the analysis of circulating tumor DNA (ctDNA), exosomes, or circulating tumor cells (CTCs) (65). Such technologies have been reviewed extensively (66–68). Elevated concentrations of tumor markers such as ctDNA and exosomes are found in the blood of cancer patients (68,69). ctDNA concentrations have been studied as surrogates for tumor burden in patients with metastatic disease, although some reports have found a lack of correlation (70,71). While, in theory, detecting ctDNA is an attractive target for diagnostic purposes, concerns associated with sampling have been identified. For example, breast cancers with a 1 cm tumor diameter are detectable by mammography (93%), but do not typically release enough ctDNA to be detected in a blood sample (72). The presence of ctDNA is even lower for a difficult to detect 0.5 cm diameter tumor. Cancer-specific exosome markers have shown prognostic value, including transforming growth factor beta 1 (TGF-β1) as a marker for ovarian cancer and epidermal growth factor receptor (EGFR) for glioblastoma (73). These technologies have also shown promise for monitoring the response to immunotherapies, as the primary tumor may initially gain volume due to the immune infiltrate (74). Of note, exosomes and mutated ctDNA can also be found in healthy individuals, decreasing the sensitivity and specificity in interpreting these findings (68,75). Conceivably an engineered diagnostic site could be leveraged for these diagnostic tests, and may also provide a sample enriched in factors of interest.

The direct analysis of metastatic cells has the potential to examine active disease and its evolution, along with the capacity to identify targets based on the biology of the cancer cell. At present, only one CTC device, CellSearch, is approved by the FDA (76). CellSearch implements antibodies for EpCAM and cytokeratin to isolate and identify CTCs, which are generally associated with poor prognosis (77). However, CellSearch and CTC isolation cannot distinguish cells with metastatic potential (typically 0.01% of CTCs), from those cells that will ultimately be biologically inert, thereby limiting the clinical relevance of CTC identification (78). Engineered diagnostic sites represent an alternative to bulk analysis of CTCs, because these sites have the capacity to accumulate an aggressive population of cancer cells and also decrease the systemic metastatic burden (20,21,28,79). Engineered niches capture cells prior to colonization of solid organs (20). Importantly, early detection with scaffold implantation, combined with surgical resection of the primary tumor, provided a survival advantage in pre-clinical models (22). Information collected from these niche-derived metastatic cells, which are phenotypically dissimilar from circulating CTCs, and cancer cells that make up the primary tumor (32), may be analyzed to direct specific therapies most relevant to the treatment of metastasis.

Immune cell detection and analysis:

Elevations in specific immune cell populations in the blood have shown prognostic value in cancer management. Accordingly, the platelet-to-lymphocyte ratio (PLR) and neutrophil-to-lymphocyte ratio (NLR) were associated with poor prognosis, but little consensus on clinically meaningful values is available (80,81). In addition, high concentrations of regulatory T cells (Tregs) in the blood were identified as biomarkers for cancer, but the prognostic value of these T cells was dependent on the type of cancer analyzed. Tregs indicated poor prognosis for hepatocellular carcinoma and breast cancer (82,83), yet suggested a favorable prognosis for colorectal cancer and specific lymphomas (84,85). Such variability suggests that identifying changes in blood cell ratios and correlating these findings to cellular events in the MN may augment the function of the MN as a diagnostic.

Peripheral blood gene expression signatures have been developed to monitor cancer progression, particularly in lung cancer and breast cancer models. For lung cancer, studies with varying gene signatures have reported relatively high diagnostic accuracy (AUC = 0.81–0.98) (86–88). In the context of breast cancer, two separate studies developed gene signatures that distinguished breast cancer patients from healthy controls with accuracies of 80% and 77% (89,90). More recently, such signatures have been applied to predict disease relapse. One study established a signature that predicted post-surgical recurrence with an AUC of 0.88, where clinical parameters predict recurrence with an inferior AUC of 0.66 (91). Although these methods are promising for the early detection and management of cancer, they are limited by many of the same challenges as CTC analysis, namely that many of the changes observed do not necessarily correlate to metastatic burden.

Gene signatures derived from engineered metastatic sites monitor the immune dynamics of the local microenvironment to reflect disease progression and response to therapy. Profiling gene expression at the engineered niche, primarily in immune cells, was employed to generate a multivariate gene signature as a method to predict the likelihood of metastatic progression (22). Implementation of the signature separated healthy and early stage cancerous mice from animals with moderate or late stage disease with an accuracy of 92.3%, using supervised and unsupervised algorithms. As metastatic sites are conditioned by the immune system, measuring immune alterations at the engineered niche identified metastatic progression prior to the arrival of tumor cells. This strategy also demonstrated the capacity to monitor responsiveness to tumor resection based on analyzing immune cells at the niche (22), which may suggest the potential to monitor response to immunotherapies. Engineered diagnostic sites could be harnessed to molecularly stage metastatic disease, as an alternative to traditional staging methods based on imaging and histopathology of primary tumors. In a translational setting, a patient treated with the standard of therapy, yet at high-risk for distant recurrence, would undergo implantation of the scaffold at an easily accessible site (abdominal wall). This site would then be sampled and the tissue molecularly profiled and results stratified akin to OncotypeDX or MammaPrint, which could then inform clinician decisions regarding treatments at early time points in the metastatic setting. Molecular staging may impact the understanding of disease subtypes and serve as a powerful tool for precision medicine that can be harnessed prior to substantial distant disease burden (92).

Biomarker measurement and sensors

The utility of an engineered niche as a diagnostic will be enhanced with the development of non-invasive analytical technologies that identify disease-relevant features. Relative to native metastatic sites, the engineered niche represents a pre-defined location for analysis, which supports the use of advanced imaging technologies and the integration of sensors. Importantly, non-invasive monitoring supports longitudinal analysis of the niche, which can be coupled with discrete analysis by biopsy for more detailed molecular information. Finally, sensors can be designed to analyze the microenvironment of the niche, including cancer cells, immune cells, and other tissue-associated factors, reflective of the MN, which cannot be assessed by liquid biopsy.

Multiple imaging approaches have demonstrated the feasibility of detecting distinct structural features of the niche that reflect disease progression. Inverse spectroscopic optical coherence tomography (ISOCT) measures the micro and nanostructure of tissues, and when applied to engineered diagnostic sites detected structural changes sufficient to distinguish tumor bearing from tumor free animals at a pre-metastatic stage of disease (20,79). More recently, spectral ultrasound imaging (SUSI) was able to detect microenvironmental alterations in a synthetic niche during metastasis (18). The changes detected by ultrasound were determined to be cellular in nature, and may represent immune cell changes at the scaffold or the increased presence of endothelial cells in tumor bearing animals (18). Collectively, these studies highlight the opportunity for non-invasive disease detection that may be eventually translated to simple clinical or at-home monitoring.

Engineered diagnostic sites can also be designed to measure local concentrations of cancer-associated biomarkers that may not be detectable through liquid biopsy. In a recent study, a polymeric chamber with a permeable membrane was implanted in the resection bed of ovarian and testicular tumors (93). The chamber contained two populations of superparamagnetic nanoparticles tagged with antibodies for different components of the beta subunit of human chorionic gonadotrophin (hCG-β). The nanoparticles aggregated as hCG-β diffused into the device, which could be monitored with MRI through changes in the transverse relaxation time. These sensors are designed to serve as integrators of biomarker concentration. Measurement of integrated local expression of biomarkers over time may be useful for early detection of recurrence by MRI (93). This approach demonstrates the capacity for non-invasive in vivo monitoring of soluble cancer biomarkers and may enable common clinical imaging techniques to measure recurrence long before a detectable tumor forms.

Sensors that have been employed for monitoring primary tumors may ultimately be adapted to the engineered MN. An implantable NMR sensor and wireless reader were employed to monitor pH continuously over several days (94). Changes in relaxation time associated with decreased intratumoral and peritumoral pH were identified when compared to the pH from the contralateral control side of the study animals. Additionally, ongoing work on the Implantable Microsystems for Personalized Anti-Cancer Therapy (IMPACT) project at University of Edinburgh seeks to measure signals, such as pH and oxygen, using tumor implanted wireless sensors (95). As these technologies continue to develop, they may provide unique insights into MN formation by providing measurements of dynamic changes in the microenvironment during niche development.

Responsive implants

Synthetic biology offers the opportunity to create diagnostic systems that respond to dynamic molecular signals in vivo (96,97). Macrophage cell lines have been engineered to express luciferase, for non-invasive visualization under the arginase-1 promoter, which was activated in the tumor microenvironment (96). Another approach harnessed a synthetic circuit in human embryonic kidney cells (98). Cells were encapsulated in alginate-poly(L-Lysine)-alginate beads to protect against xenogeneic rejection and were injected subcutaneously in mice to create a melanin biomedical tattoo that darkened upon the development of hypercalcemia. This innovative strategy utilized endogenous human proteins to create a persistent, in vivo synthetic circuit enabling the early detection of cancer progression. However, this detection strategy is nonspecific as many cancers do not induce hypercalcemia, while at the same time, other benign medical conditions are associated with this anomaly. Note that cell-based approaches may be challenging to translate, due to the regulatory hurdles, technical challenges, and expense associated with a cellular engineering and maintaining cell survival.

A number of technologies are emerging that may ultimately be translated toward the non-invasive or minimally-invasive monitoring of engineered niches. Analyte-responsive smart tattoos (99) for glucose sensing could be modified for tumor monitoring. Traditional sensors for glucose, pH, or other analytes could be adapted to the niches, and in the circumstance where a long-term implant is undesirable, degradable electronics could be used (100–102). Microneedle sampling is being developed for vaccine delivery, which penetrates only superficial layers of skin to sample the skin microenvironment and may prove to be a useful tool for minimally invasive monitoring of engineered MNs (103–105). Although these technologies have yet to be harnessed for oncology, future work will likely couple engineered MNs with these minimally invasive technologies to enable deeper analysis of alterations associated with metastasis and ultimately serve as precision tools in the diagnosis and prognosis of metastatic disease.

The pre-clinical data combined with non-invasive detection suggests that engineered niches have promise for use as a diagnostic for early detection, and for monitoring responses to therapy in the clinic. Multiple platforms have been able to capture metastatic cells with direct utility for monitoring disease progression, for which clinical studies have only recently been initiated (e.g., MTrap, NCT03085238). Engineered niches are fabricated from FDA-approved materials, and studies have been designed to establish that the synthetic niches do not pose risk to patients as they capture and retain tumor cells. Similar technologies have begun testing as a component of cancer vaccines in human clinical trials, and will be the first to examine the safety of these platforms (NCT01753089). The synthetic metastatic niches have been effective for monitoring disease recurrence and the response to therapy in pre-clinical models, yet the course of cancer progression in humans will require rigorous scrutiny.

Outlook

Although synthetic MNs are nascent in their development, this technology holds promise for preclinical studies of cancer metastasis and clinical translation for early detection, metastatic staging, and monitoring response to therapy (Figure 1). Synthetic niches reflect many aspects of native MNs, and these implants can be engineered to tune the metastatic microenvironment. Tunability enables controlled presentation of ECM or soluble proteins to molecularly dissect niche composition and function in vivo. Clinically, engineered diagnostic sites can provide unique, dynamic information about a metastatic niche that is not captured by the current clinical paradigm. Clinical technologies for the early detection and monitoring of metastatic cancer are a major area of interest, but many emerging technologies are limited by sampling issues. Synthetic niches enable an alternative and potentially enriched source that can feed into many existing analysis pipelines. For example, gene expression signatures can be used to monitor changes in synthetic niches similarly to their use in examining primary tumors for prognosis, and repeated analysis of the synthetic niche provides the opportunity for dynamic monitoring (22).

Development of engineered niches is occurring at a critical time in cancer therapy, with a recent expansion of treatment options for patients with metastatic disease. Implementation of these niches enables diagnosis when disease and mutational burden are low, presenting unprecedented opportunities for therapeutic intervention. Using engineered diagnostic sites as devices for the identification of early metastatic events, may improve therapeutic ability to forestall distant disease progression. Additionally, engineered niches also exploit the possibility for identification of novel markers, and the discovery of new targeted treatments. For example, vaccines could be directed towards neoantigens identified from metastatic cells found in the niche, or the molecular analysis of immunologic markers could help to direct the appropriate immunotherapy. Cells from the niche can also be converted to an experimental platform, akin to cell lines or organoids, to evaluate the effectiveness of emerging interventions. This system could be particularly useful for examining the metastatic microenvironment and factors associated with dormancy, stemness, and disease progression.

While additional work remains to fully realize the potential of these synthetic niches, tissue engineering and precision medicine are converging, and the resulting tools may have transformational benefits for science, medicine, and patient care. These niches are, in part, capturing systemic immunological changes, which are present in cancer, yet are also observed in autoimmune diseases, transplant rejection, and other conditions with prolonged, asymptomatic prodromes. In addition to enabling diagnosis and implementation of precision therapies, translation of this technology to humans could enable mechanistic insights into the early stages of human disease that have previously been elusive.