Recent advances in CAR-MSCs: the new engine of cellular immunotherapy evolution

Abstract

In recent years, the development of chimeric antigen receptor (CAR) technology has greatly promoted the progress of cellular immunotherapy. Among them, CAR-T cell therapy has shown remarkable clinical effects in the treatment of hematological malignancies. However, this therapy still faces a series of challenges, including immunogenicity, toxic side effects, and insufficient maintenance of long-term efficacy. The latest research progress has extended CAR technology to mesenchymal stem cells (MSCs), and the resulting CAR-MSCs combine the precise targeting ability of CAR molecules with the inherent immunomodulatory, tissue homing, and regenerative repair properties of MSCs, providing a new therapeutic strategy for cancer and immune-related diseases. This review examines the engineering design, biological characteristics, and applications of CAR-MSCs in oncology and immune-related disorder therapy. Preclinical studies have shown their effectiveness against glioblastoma, Ewing sarcoma, acute myeloid leukemia, and lung cancer, as well as graft-versus-host disease, through TRAIL secretion, bispecific antibody production, and Treg induction. Despite promising results, significant hurdles persist in CAR-MSC manufacturing scalability, cell persistence, heterogeneous MSC tissue sourcing, and undefined application protocols, all of which are critical for clinical translation. We investigated corresponding strategies, including nonviral gene delivery, metabolic engineering, senescence-resistant MSC clones, and microenvironment-specific activation. Standardized production workflows incorporating rigorous quality control are essential for future applications. CAR-MSCs represent a paradigm shift in precision immunotherapy by providing dual therapeutic modalities for cancer and immune disorders. Fully unlocking their therapeutic potential will require interdisciplinary efforts to overcome biological and technical barriers while advancing combination therapies.

Introduction

The progress in chimeric antigen receptor (CAR) design has considerably accelerated the development of cell-based immunotherapies, playing a central role in biomedical studies. Cells engineered with CARs, such as CAR-T and CAR-natural killer (NK) cells, have exhibited impressive outcomes in addressing a range of blood-based and solid cancers [1–5]. The technology itself has advanced through several generations—from first to fifth—each aimed at improving treatment effectiveness and reducing risks [6, 7]. Nevertheless, issues like immune reactions, adverse effects, and short-lasting responses remain substantial barriers [7]. These limitations have encouraged the exploration of new cell types to augment the capabilities of CAR-based treatments.

Mesenchymal stem cells (MSCs), constituting a type of multipotent stromal cell, can be sourced from diverse tissues such as bone marrow (BM), adipose tissue, and the umbilical cord (UC), followed by extensive in vitro expansion [8]. These cells are characterized by their low immunogenicity, capacity to facilitate tissue repair, and potent immunomodulatory functions within the microenvironment [9]. Their broad therapeutic applicability and an excellent safety profile are exemplified in over 1,300 clinical trials conducted globally, notably for conditions like graft-versus-host disease (GvHD) and osteoarthritis [10, 11].

By integrating CAR technology with MSCs, researchers have developed CAR-MSCs, which combine the targeting specificity of CARs with the inherent advantages of MSCs. This fusion positions CAR-MSCs as a novel frontier in cell-based therapies [12–14]. CAR-MSCs can be engineered to target tumor-associated antigens while simultaneously modulating the tumor microenvironment (TME), potentially mitigating the systemic toxicity. Moreover, the integration of CAR technology with MSCs expands the spectrum of potential targets for cancer treatment and opens new avenues for applying CAR technology in the treatment of nonmalignant conditions, such as immune-mediated disorders, neurological diseases, and tissue repair and regeneration [15–17].

The development of CAR-MSCs marks a significant step forward in cellular immunotherapy, offering more effective and safer treatment strategies. This review systematically summarizes the evolution of CAR engineering in conjunction with the biological characteristics and therapeutic applications of MSCs. For the first time, we comprehensively discuss the engineering design, biological properties, and clinical applications of CAR-MSCs in the treatment of tumors and immune-related diseases. Furthermore, we also discussed the potential challenges and corresponding strategies that the future clinical application of CAR-MSCs might encounter. Through a critical evaluation of the literature, our aim is to highlight the significant potential of CAR-MSCs in the field of immunotherapy.

Current status of CAR engineering

The CAR is a complex synthetic receptor with a history spanning nearly three decades [18]. As a specific receptor, CAR enables immune cells to recognize antigens and elicit antigen-specific cytotoxic responses. The significant progress in immune cell therapies offers robust evidence supporting the strategic decision made decades ago to integrate CAR technology into T cells [19].

The CAR structure includes an extracellular domain composed of a hinge region and a single-chain variable fragment (scFv). The scFv, derived from antibody variable regions, binds tumor cell surface antigens independently of major histocompatibility complex (MHC). The hinge region connects the scFv to the transmembrane domain. The intracellular part of CAR contains signaling domains from T-cell receptors and costimulatory molecules. These domains are typically derived from CD28 family receptors or TNF receptor family members (e.g., 4-1BB). Together, they boost T-cell expansion and cytokine production by activating costimulatory pathways, enhancing the antitumor activity of T cells [20–22]. Signal transduction domains usually include the TCR/CD3ζ or FcεRIγ chain, both containing immunoreceptor tyrosine-based activation motif essential for T-cell signaling. CARs are capable of autonomously recognizing tumor-associated antigens (TAAs) without restriction by the MHC. These engineered receptors are designed to increase the ability of the immune system to recognize and eliminate cancer cells [23, 24].

Since the 1980 s, CAR technology has undergone multiple iterations of refinement, significantly increasing the breadth of cellular immunotherapy (Fig. 1). The advancements in cellular immunotherapy centered on CARs exemplify the translation of fundamental research into clinical applications. First-generation CAR-T cells contain only a single intracellular signaling domain, such as CD3ζ or FεRIγ, without any costimulatory domain. These CAR-T cells resemble endogenous TCRs; however, their limited ability to generate interleukin-2 (IL-2) results in a diminished response during clinical trials. Consequently, supplementation with exogenous IL-2 is often necessary for first-generation CAR-T cells to ensure an efficacious response [25].

Fig. 1.

Architecture of CARs and their successive iterations. First-generation CARs contain a single activation domain, such as CD3ζ, which triggers cell activation and scFv-specific cytotoxicity. Second- and third-generation CARs include one or two costimulatory signals, enhancing both costimulatory molecule activation and intracellular signaling. Fourth-generation CARs, or TRUCKs, combine previous features to target and eliminate tumor cells while simultaneously secreting cytokines like IL-12 through cytokine gene cassettes, boosting the local immune response. Abbreviations: CAR, chimeric antigen receptor; TNF, tumor necrosis factor; TRUCK, T cell redirected for universal cytokine-mediated killing

The limitations of first-generation CARs were addressed in second-generation CAR-T cells by incorporating costimulatory signaling domains (e.g., CD28 or 4-1BB). Studies have demonstrated that engineered costimulatory domains (e.g., ICOS and CD27) exhibit significant antitumor effects [21, 22]. These enhancements improve T-cell proliferation, cytokine production, cytotoxicity, and in vivo persistence. The effects of different costimulatory domains vary to some extent. Compared with CD28, the 4-1BB costimulatory domain exhibits longer persistence and reduced cell exhaustion [26]. These advancements have substantially improved the therapeutic efficacy of CAR-T cells. Currently, most FDA-approved CAR-T-cell therapies utilize the second-generation CAR design.

Third-generation CARs incorporate two costimulatory signaling domains in conjunction with the CD3ζ chain [27]. This design aims to optimize T-cell activation, proliferation, persistence, and antitumor functionality via dual costimulatory signaling. However, preclinical studies of third-generation CAR-T cells have yielded inconsistent results; some investigations have demonstrated superior performance compared with that of second-generation CAR-T cells, whereas others have reported no significant difference in efficacy between the two [28, 29]. The reasons for the disparate outcomes observed in third-generation CAR-T-cell therapies remain unclear and may be attributed to factors such as limited sample sizes, potential instability in product manufacturing, or variations in the immune microenvironment. Third-generation CAR-T-cell therapy targeting CD19 has been reported to be an effective and safe treatment for systemic IgM light chain amyloidosis (AL) associated with underlying marginal zone lymphoma [30]. However, the enhanced efficacy from dual costimulatory signals may also increase the risk of adverse events, especially severe cytokine release syndrome (CRS). Therefore, the selection and optimization of costimulatory molecules are crucial for preventing excessive signaling, which may lead to immune cell exhaustion. More clinical trials are needed to evaluate and monitor the efficacy and safety of these products, helping to unlock the full potential of third-generation CARs.

Fourth-generation CARs, or TRUCKs (T cells redirected for universal cytokine killing), recognize tumor cells and induce cytotoxicity through CARs while also releasing engineered immunomodulators (e.g., IL-12, IL-15, and IL-2) to modulate the TME. For example, to combat T-cell dysfunction, Zhao Y et al. engineered CAR-T cells to secrete IL-10 [31]. This modification improved CAR-T cell proliferation and effector functions, leading to complete regression of multiple cancers, such as colon cancer, breast cancer, melanoma, and pancreatic cancer, in mouse models. Moreover, TRUCKs can enhance CAR-T-cell functionality by expressing chemokine receptors or chemokines (e.g., CCL19 and CCL21), enabling directional migration to the TME [32]. These strategies have been validated in preclinical studies and shown promising efficacy and safety in clinical trials for various malignant tumors [33, 34]. The ability to elicit a secondary antitumor immune response distinctly endows fourth-generation CARs with significant therapeutic potential for treating solid tumors, particularly in targeting antigen-negative cancer cells within the TME [35].

Furthermore, similar strategies involve engineering CAR-T cells with gene-edited noncytokine proteins, which can be categorized into secretory and nonsecretory types. Secretory proteins (such as Helicobacter pylori neutrophil-activating protein, drug-activating enzymes, scFv-based PD-1 blockers, and heparinase) typically target the immunosuppressive TME [36–38]. Nonsecretory proteins (such as c-Jun, PD-L1-specific chimeric conversion receptors, anti-B7-H3 chimeric cytokine receptors, anti-CD38 chimeric cytokine receptors, and TGFβR 4-1BB hybrid receptors) primarily enhance the functionality of CAR-T cells [39–42]. Both approaches have demonstrated potential in preclinical studies for treating hematological malignancies and solid tumors. Consequently, fourth-generation CAR-T-cell therapy represents a highly promising therapeutic strategy. In the future, it will be essential to further investigate and optimize the structure of TRUCKs and evaluate their efficacy through additional preclinical and clinical trials to achieve more efficient tumor cell clearance.

Research on fifth-generation CARs remains in the exploratory phase, with mainstream investigations and clinical applications focused primarily on the first four generations of CARs. Currently, fourth-generation CARs are undergoing active research and early clinical trials. The term ‘fifth generation’ may refer to further enhancements and innovative therapies derived from the fourth generation. Some scholars have reported the development of fifth-generation CARs by modifying second-generation CARs with the addition of the IL-2Rβ cytokine receptor in the intracellular domain. This change enables coordinated activation of the JAK-STAT signaling pathway between costimulatory signals (e.g., 4-1BB or CD28) and CD3ζ, promoting improved cell proliferation and persistence [43]. To enhance CAR-T cell efficiency, Yuti P et al. developed fifth-generation anti-BCMA CAR-T cells containing three costimulatory domains and capable of secreting anti-PD-L1 scFv blockade molecules. These cells demonstrated superior antitumor efficiency, enhanced proliferative capacity, and reduced T-cell exhaustion in response to multiple myeloma (MM) cells [44]. However, there is no widely accepted definition of the terminology for the fifth generation within the academic community. In summary, the available data suggest that the overall efficacy and safety profile of CAR-T-cell therapy is highly encouraging [45, 46].

However, the comprehensive advancement of CAR-T-cell therapy has been significantly impeded by several challenges, including limited tissue penetration, antigenic evasion, tumor heterogeneity, an immunosuppressive TME, increased treatment expenses, immune effector cell-associated neurotoxicity syndrome (ICANS), and CRS [47, 48]. Consequently, the industry has acknowledged that the full potential of CAR technology should not be confined solely to T-cell modification; it is equally imperative to leverage the antitumor properties of other immune cell types, including NK cells, macrophages, invariant natural killer T (iNKT) cells, γδT cells, regulatory T cells (Tregs) and MSCs [15, 49–51]. In addition, evidence from both clinical and preclinical studies has built upon the success of CAR-T-cell therapy in oncology to extend its applications to various noncancerous diseases, including autoimmune disorders (e.g., systemic lupus erythematosus (SLE) and inflammatory myopathy), fibrotic conditions (e.g., cardiac disease, liver disease, chronic kidney disease, pulmonary disorders, and skeletal muscle diseases), and senescence-associated pathologies (e.g., liver fibrosis, solid tumors, atherosclerosis, and natural aging). Infectious pathologies such as HIV, hepatitis B/C, cytomegalovirus, and tuberculosis—initial targets of CAR therapy—are being re-evaluated in light of a deeper understanding of immune responses and the advent of enhanced therapeutic tools [52]. Compared with tumors, noncancerous diseases typically present a significantly smaller target burden, where partial clearance of diseased cells can yield therapeutic benefits and present a lower mutational load. In addition, the immunosuppressive environment characteristic of the TME is largely absent in most chronic diseases, allowing physiological targeting of diseased tissue by the immune system [52].

Biological properties of MSCs

MSCs represent a heterogeneous population of multipotent adult stem cells with the ability to self-renew and differentiate [53]. Initially identified as colony-forming unit fibroblasts in the 1970 s, MSCs were discovered within the BM and are morphologically defined as spindle-shaped, adherent nonhematopoietic stem cells (non-HSCs) [54, 55]. Besides BM, these cells can be obtained from various tissues and organs such as adipose tissue, UC tissue, Wharton’s jelly, amniotic fluid, placenta, liver, heart, lung, adrenal glands, pancreas, and dental tissues. MSCs can also be derived from embryonic stem cells or induced pluripotent stem cells (iPSCs) (Fig. 2) [8, 56–61]. To better define human MSCs, the Mesenchymal and Tissue Stem Cell Committee of the International Society for Cell Therapy (ISCT) has proposed that MSCs should express CD73, CD90, and CD105 in over 95% of cases, while lacking CD45, CD34, CD14 or CD11b, CD79a or CD19, and human leukocyte antigen DR (HLA-DR) surface molecules in less than 5% [62]. Under specific conditions, MSCs can differentiate into key cell lineages, including osteoblasts, chondrocytes, myocytes, and adipocytes. The differentiation direction, differentiation potential, and heterogeneity of MSCs vary according to their tissue origin. For example, MSCs derived from the BM, synovium, and periosteum present the highest alizarin red positivity rate for osteogenesis; amniotic MSCs demonstrate superior chondrogenic potential [63]; the number of BM-MSCs decreases sharply with age. Moreover, fetal MSCs possess greater proliferation capacity [64].

Fig. 2.

The origins and therapeutic biological properties of MSCs. Initially discovered in BM, MSCs have subsequently been obtained from a range of other biological sources such as adipose tissue, UC, dental pulp, amniotic fluid, and the placenta. These cells release a wide variety of biologically active substances—including growth factors, cytokines, and extracellular vesicles—that contribute significantly to processes like immune modulation, tissue regeneration, blood vessel formation, and neural protection. Abbreviations: MSC, mesenchymal stem cell; BM, bone marrow; UC, umbilical cord; TGF, transforming growth factor; IL, interleukin; CCL, chemokine (C-C motif) ligand; miR, microRNA; HGF, hepatocyte growth factor; VEGF, vascular endothelial growth factor; KGF, keratinocyte growth factor; PDGF, platelet-derived growth factor; FGF, fibroblast growth factor; IGF, insulin-like growth factor; OPG, osteoprotegerin; BDNF, brain-derived neurotrophic factor; SDF, stromal cell-derived factor

Emerging evidence has expanded the surface marker repertoire of MSCs, complementing well-established canonical markers with novel biomarkers, including SERPINF1, S100A9, LRRC75A, CMKLR1 (CCRL2), CD142, CD271 (LNGFR), TNFAIP6 (TSG6), CSPG4, CD106 (VCAM-1), CD317, and CD146 [65]. These refined surface marker signatures enable the stratification of functionally distinct MSC subpopulations, particularly those orchestrating niche-specific functions in HSC maintenance, neoangiogenesis, neurogenic niche modulation, and immune regulation, as evidenced by recent single-cell RNA sequencing and secretome profiling studies (Table 1). In a myocardial infarction model, CD73⁺ adipose-derived (AD) MSCs demonstrated superior cardiac recovery through enhanced angiogenesis. Compared with their CD73-negative counterparts, these cells showed significantly increased secretion of vascular endothelial growth factor (VEGF), stromal cell-derived factor-1α (SDF-1α), and hepatocyte growth factor (HGF) [66]. CD271⁺ AD-MSCs exhibited elevated angiogenic gene expression with concurrent downregulation of inflammatory genes, maintaining characteristic MSC surface markers while demonstrating enhanced neoangiogenic potential [67]. TNFAIP6⁺ murine MSCs displayed increased immunosuppressive activity and therapeutic efficacy in acute inflammation models [68]. Neuroglial antigen 2 (NG2)-expressing BM-MSCs presented increased proliferative capacity, preserved stemness, and improved functional recovery in hepatic fibrosis models [69]. CD106⁺ MSCs from UC sources demonstrated superior adhesive properties, angiogenic capacity, wound-healing potential, and immunomodulatory functions [70]. CD317⁺ MSCs exhibited increased differentiation potential alongside increased immunosuppressive activity [71]. CD146⁺ cell populations have greater cartilage-protective effects and Th17 cell suppression capabilities [72].

Table 1.

Functional properties of biomarker-defined MSC subpopulations in preclinical models

Surface Marker	MSC Source	Disease Model	Functional Outcomes	Mechanistic Insights	Ref.
CD73	Mice AD	MI in rats	Superior cardiac recovery via enhanced angiogenesis	Upregulated VEGF, SDF-1α, and HGF secretion	[66]
CD271	Human AD	Tissue regeneration	Improved neoangiogenesis with an anti-inflammatory profile	Elevated angiogenic gene expression with downregulated inflammatory genes; retains typical stem cell surface marker complement	[67]
TNFAIP6	Mice BM, placenta, and AD	Acute inflammation in mice	Augmented immunosuppressive activity	Accelerated anti-inflammatory cytokine TNFAIP6 release	[68]
NG2	Mice BM	Liver fibrotic/cirrhotic injury in mice	Marked motion and proliferation; stronger inhibition of inflammatory infiltration and fibrosis; better balance of immune response	Greater stemness potential with increasing Ki-67/SSEA-3 expression	[69]
CD106	Human UC	Full-thickness wound mouse model	Superior adhesive properties, angiogenic capacity, wound healing, and immunomodulation	/	[70]
CD317	Human AD	Acute inflammation in mice	Enhanced differentiation potential and immunosuppressive activity	Higher levels of TNFAIP secretion	[71]
CD146	Human UC	Collagen-induced arthritis in mice	Greater cartilage-protective effects	More conspicuous Th17 cell inhibition resulting in a lower level of IL-17 A	[72]

MSCs, mesenchymal stem cells; AD, adipose-derived; BM, bone marrow; UC, umbilical cord; MI, myocardial infarction; VEGF, vascular endothelial growth factor; SDF-1α, stromal cell-derived factor-1α; HGF, hepatocyte growth factor; NG2, neuroglial antigen 2; SSEA-3: stage-specific embryonic antigen-3; Th17, helper 17; IL-17 A, interleukin-17 A; TNFAIP6, tumor necrosis factor alpha-induced protein 6; Ref., reference

Building on our previous research [73–76], we employed single-cell sequencing to delineate MSC heterogeneity. A CDCP1-expressing MSC subgroup showed enhanced migration and superior acute GvHD suppression compared to its CDCP1-negative counterpart. CDCP1, designated CD318, functions as a transmembrane glycoprotein predominantly expressed in tissue-specific progenitor populations, including HSCs, neural progenitor cells (NPCs), and MSCs [77]. Mechanistically, we hypothesized that CDCP1 enhances the homing of MSCs to GvHD-affected tissues via RhoA pathway activation, amplifying their immunomodulatory effects. Current animal studies are validating these properties to optimize the selection of clinical-grade MSCs for acute GvHD prophylaxis. In conclusion, the functional heterogeneity of MSCs is closely associated with their surface marker profiles. Moving forward, leveraging single-cell transcriptomics and proteomics will be key to precisely identifying, functionally enhancing through genetic engineering, and advanced sorting of MSC subpopulations, ultimately enabling the isolation of disease-specific MSCs. On this basis, CAR-engineered MSCs hold promise as potent tools for personalized medicine.

Bidirectional interactions between MSCs and CAR-engineered cells

Emerging evidence has revealed bidirectional interactions between MSCs and CAR-engineered immune effectors, particularly CAR-T cells (Table 2). The current literature reveals three distinct paradigms: MSC-mediated suppression of CAR-T-cell functionality, enhancement of CAR-T-cell activity, and neutral effects on cellular potency.

Table 2.

Bidirectional interactions between MSCs and CAR-engineered cells: Outcomes, mechanisms, and therapeutic implications

Interaction Type	Key Findings	Mechanism	Model System	Potential Clinical Application	Ref.
Suppression of CAR-T-Cell Activity	BM-MSCs inhibit AML-targeted CAR-T-cell expansion via IDO-1 upregulation; reduce IFNγ/IL-2 secretion	IDO-1, COX-2, PD-L1, PD-L2, and ICAM-1 mediated immunosuppression	Cell coculture	IDO-1 inhibitors enhance CAR-T-cell expansion	[78]
	hUC blood-MSC-derived STC1 suppresses the efficacy of CAR-T cells in killing lymphoma cells	STC1-induced TME modulation	Cell coculture; xenograft mice	STC1 blockade to enhance CAR-T efficacy	[79]
	BM-MSCs protect MM cells against low-affinity, moderate-lytic CAR-T via cell-contact dependency	CAR affinity-dependent resistance	Cell coculture	Optimizing CAR affinity/design for MM	[80]
	BM-MSCs induce resistance to NK cells in MM; CD38-CAR engineering reverses protection	CAR-dependent evasion of MSC-mediated resistance	Cell coculture	CAR engineering to bypass MSC-mediated resistance	[81]
Enhancement of CAR-T-Cell Function	UC-MSC mitochondrial transfer reduces CAR-T-cell apoptosis and enhances cytotoxicity	Increased mitochondrial activity of anti-apoptotic Bcl-2/BARD1 pathways	Cell coculture	MSC mitochondrial transfer potentiates the therapeutic effects of virus-free generated CAR-T cells	[82]
	BM-MSCs with OAd/IL-12/PD-L1 inhibitors enhance HER.2 CAR-T-cell-mediated tumor killing in lung tumors	Combination therapy boosted T-cell effector cytokines (IFN-γ, granzyme B, perforin)	3D tumor spheroids; NSCLC xenograft mice	MSC-vectored combination therapies	[83]
	IL-7/IL-12-secreting BM-MSCs boost CAR-T-cell expansion and Th1 polarization	IL12 improved CAR-T cell proliferation, reduced AICD, and enhanced target cell lysis	Colorectal cancer xenograft mice	Genetically engineered MSCs for CAR-T adjuvants	[84]
Neutral Effects on CAR-T Cells	BM-MSCs derived from B-ALL pediatric patients suppress conventional T cells but spare CD19-CAR-T-cell cytotoxicity	CAR-T-cell insensitivity to B-ALL BM-MSC-mediated immunosuppression	Cell coculture; acute colitis mice and B-ALL xenograft mice	Decoding BMM-driven CAR-T-cell resistance mechanisms	[85]
	MSCs promote hematopoietic recovery without impairing CAR-T-cell antitumor efficacy	CXCL12/NF-E2-mediated BM reconstitution	B-cell lymphoma mice	MSCs for post-CAR-T-cell hematopoietic support	[86]

BM-MSCs, bone marrow-derived mesenchymal stromal cells; hUC blood-MSCs, human umbilical cord blood-derived mesenchymal stromal cells; UC-MSCs, umbilical cord-derived mesenchymal stromal cells; TME, tumor microenvironment; MM, multiple myeloma; OAd, oncolytic adenovirus; CAR-T, chimeric antigen receptor T cell; AML, acute myeloid leukemia; IDO-1, indoleamine 2,3-dioxygenase 1; COX-2, cyclooxygenase 2; PD-L1, programmed death-ligand 1; ICAM-1, intercellular adhesion molecule 1; IFNγ, interferon-gamma; IL, interleukin; STC1, stanniocalcin-1; NK, natural killer; CD, cluster of differentiation; Bcl-2, B-cell lymphoma 2; BARD1, BRCA1-associated RING domain protein 1; NSCLC, non-small cell lung cancer; AICD, activation induced cell death; B-ALL, B-cell acute lymphoblastic leukemia; BMM, bone marrow microenvironment; BCMA, B-cell maturation antigen; Th1, T helper 1 cell; CXCL12, C-X-C motif chemokine ligand 12; NF-E2, nuclear factor erythroid-derived 2; B-ALL, B-cell acute lymphoblastic leukemia; Ref., reference

MSC-mediated suppression of CAR-T-cell efficacy

Towers et al. demonstrated that BM-MSCs suppress acute myeloid leukemia (AML)-redirected switchable CAR-T-cell expansion through IDO-1-mediated mechanisms, concurrently attenuating IFNγ and IL-2 secretion [78]. Zhang et al. identified stanniocalcin-1 (STC1) as a critical MSC-secreted factor that modulates the TME to impair CAR-T-cell cytotoxicity against lymphoma in murine xenograft models [79]. Holthof and colleagues employed a diverse set of CAR-T cells exhibiting a range of antigen specificities—including BCMA, CD38, and CD138—along with varying affinities for target cells [80]. The results showed that MM patient-derived BM-MSCs significantly protected MM cells against lower-affinity, moderately lytic CAR-T cells. Similarly, MSCs protected MM cells against KHYG-1 NK cells by inducing resistance mechanisms. However, the immune resistance mediated by MSCs can be reduced by genetically modifying KHYG-1 cells to express the CD38 CAR [81]. These findings suggest that this immune escape can be addressed by inhibiting antiapoptotic molecules or enhancing immune killer cells.

MSC-mediated enhancement of CAR-T-cell function

Recently, Court et al. demonstrated that mitochondrial transfer from MSCs reduced CAR-T-cell apoptosis postelectroporation while enhancing cytotoxic activity, suggesting metabolic reprogramming as a mechanism for improved stress resistance [82]. McKenna et al. engineered MSCs to deliver oncolytic adenoviruses (OAds) coexpressing IL-12 and PD-L1 blockade, ​which enhanced​ CAR-T-cell antitumor activity against solid tumors by combining localized viral oncolysis with immune checkpoint inhibition [83]. Hombach et al. pioneered the secretion of IL-7/IL-12 by transgenic MSCs, which potentiated CAR-T-cell expansion in colorectal cancer models, achieving superior tumor control compared with unmodified MSC coadministration [84].

Effects of neutral MSCs on CAR-T-cell activity

Certain studies have demonstrated that MSCs mediate immunomodulation without direct interference with CAR-T-cell functionality. Zanetti et al. characterized BM-MSCs from pediatric B-ALL patients and confirmed their immunosuppressive effects on conventional T cells while preserving CD19-CAR-T-cell cytotoxicity [85]. Xia et al. identified key mechanisms by which MSCs support hematopoietic recovery following CAR-T-cell therapy. Notably, the murine model demonstrated that coadministration of MSCs does not impair CAR-T-cell antitumor efficacy [86].

Applications of MSCs

MSCs possess significant potential for clinical applications because of their unique therapeutic properties, which include low immunogenicity, inherent tumor tropism, ease of isolation, regenerative capabilities, and multidirectional differentiation potential [9]. Through direct cell contact and the release of cytokines, chemokines, and growth factors, MSCs help sustain immune balance across most tissues by influencing inflammation, blood vessel formation, and tissue fibrosis (Fig. 2) [87]. The distinctive characteristics and mechanisms offer exceptional therapeutic prospects, which has prompted the use of MSCs as a cell therapy technology in most research worldwide [88]. As of February 2025, ClinicalTrials.gov hosts over 1300 projects focused on MSCs.

Immune disorders

MSCs block the T-cell cycle, promote Th2/Treg polarization, inhibit B-cell proliferation and antibody secretion, suppress NK-cell proliferation, and drive macrophage polarization from the M1 phenotype to the M2 phenotype, thereby reducing inflammation, prolonging neutrophil survival, and inhibiting DC differentiation [89–91].

Hematopoietic stem cell transplantation (HSCT) holds an irreplaceable position in treating malignant hematological diseases [92, 93], as it offers patients the possibility of a cure by eliminating malignant cells and reconstructing the hematopoietic and immune systems. However, GvHD is a major complication after HSCT that seriously threatens patient survival and quality of life [94]. MSCs show strong effectiveness in treating or preventing GvHD, with clinical studies confirming their ability to improve patient survival and quality of life [10, 95–98]. Moreover, across multiple trials targeting immune-related conditions, most also demonstrated both safety and efficacy [99–102]. Notably, a recent study demonstrated that CAR-modified MSCs can significantly enhance the immunosuppressive effects at inflammatory sites of colon GvHD in murine models, mitigate the severity of GvHD, prolong survival duration, and concurrently preserve both the stem cell phenotype and safety profile of MSCs [103]. This discovery signifies a transformative advancement in the application of cell therapy for immune-related disorders, offering robust conceptual validation for designs and enhancements of CAR-MSCs.

Hematological disorders

Researchers have demonstrated that the combination of MSCs and HSCs can increase transplantation success rates [104–106]. Mechanistically, MSCs may promote the expansion of HSCs, thereby enhancing hematopoietic recovery after CAR-T-cell therapy through the upregulation of C-X-C motif chemokine ligand 12 and nuclear factor erythroid derived 2 [86]. The AKT signaling pathway—a key modulator of MSC biological activity—has also been shown to govern HSC proliferation and differentiation. MSC-derived microvesicles, which carry hematopoietic niche-associated mRNAs that enhance hematopoietic reconstitution through paracrine reprogramming, have been identified as critical mediators of HSC engraftment. Furthermore, MSCs suppress donor lymphocyte-mediated cytotoxicity and regulate post-transplantation inflammatory responses, thereby enhancing engraftment efficiency in allogeneic transplantation models [104, 107]. Some studies have also preliminarily investigated the potential therapeutic benefits of MSCs in aplastic anemia and immune thrombocytopenia [108, 109].

Tissue regeneration and repair

MSCs possess robust tissue repair capabilities and are utilized to address various types of tissue injuries, including those affecting bone [110] and cartilage [111], tendinopathy [112], and those affecting the liver [113] and the spinal cord [114]. Current trials have shown that MSCs are safe and effective for alleviating pain and enhancing cartilage thickness in osteoarthritis patients [11]. However, they have not been shown to significantly improve quality of life [115]. Additionally, MSCs show promise in cardiac regeneration by promoting cardiomyocyte repair and improving heart function [116]. Intracoronary infusion of MSCs after acute myocardial infarction has been proven to improve the left ventricular ejection fraction [117]. Although direct intramyocardial injection of MSCs is safe for treating chronic heart failure, it is ineffective in restoring cardiac function or alleviating symptoms [118]. The application of MSCs for skin wounds, limb ischemia, and tissue regeneration is expanding, with increasing evidence supporting their role in promoting angiogenesis and facilitating cell repair [119, 120].

Neurological diseases

MSCs have shown potential for enhancing neurological repair functions, particularly in conditions like Parkinson’s and Alzheimer’s diseases [121, 122]. They support neural regeneration by differentiating into neural cells and secreting neurotrophic factors that enhance nervous system function [123]. Moreover, MSCs have demonstrated beneficial effects in treating stroke (including ischemic stroke and acute cerebral hemorrhage), chronic motor deficits, and optic neuropathy, broadening their therapeutic potential [124–128]. Overall, clinical trials have confirmed the efficacy, feasibility, and safety of MSCs, laying the groundwork for future research to benefit more patients.

Cancer

The mechanisms by which MSCs exert their effects in cancer therapy are intricate and multifaceted [9]. The key mechanisms include the following: [1] Immunomodulatory capacity: MSCs modulate the immune components of the TME through direct cell–cell interactions and paracrine signaling via soluble factors like TNF-α and IFN-γ, inducing cancer cell apoptosis and reducing invasiveness [129, 130] [2]. Induction of cancer cell apoptosis: Studies show that coculturing MSCs with specific cancer cell types induces apoptosis and inhibits proliferation. MSCs induce apoptosis in breast cancer cells in a dose-dependent manner [131], and BM-MSCs significantly trigger both early and late apoptosis in CD34⁺ leukemic stem cells [132] [3]. Homing effect and tumor tropism: MSCs exhibit homing and tumor tropism, migrating to primary tumors, metastatic sites, and areas of tissue damage or inflammation. This property highlights their potential as a targeted delivery system for anticancer agents. Engineered MSCs can produce higher levels of anticancer proteins at tumor sites, effectively inhibiting tumor growth and metastasis [133] [4]. Increased effectiveness of radiation therapy and chemotherapy: BM-MSCs, in conjunction with low-dose radiation, can significantly increase the proliferation of peritumoral astrocytes and exhibit notable antiangiogenic properties [134]. BM-derived MSCs pretreated with doxorubicin exhibit significant cytotoxic effects on breast cancer cells and gastric cancer cells [135, 136] [5]. Paracrine mechanisms: MSCs can be engineered to secrete bioactive molecules that regulate tumor-related processes such as proliferation, survival, angiogenesis, and immunosuppression, thereby inhibiting tumor growth and exerting antitumor effects [137, 138].

Fewer clinical trials have explored MSCs for cancer treatment compared to other diseases, with only a limited number completed and most still in recruitment or ongoing stages. Orae-Yazdani S et al. first assessed the safety of herpes simplex virus-thymidine kinase gene therapy using MSCs in recurrent glioblastoma (GBM) patients, confirming its safety [139]. In the future, multiarm clinical trials are needed to compare the efficacy of this MSC-based regimen with that of standard treatment, not only in GBM but also across other tumor types.

Other diseases

MSCs have also been utilized in the treatment of various other conditions, such as type 2 diabetes and its related complications [140–142], COVID-19/ARDS [143, 144], idiopathic nephrotic syndrome [145], chronic low back pain [146], and radiation-induced hyposalivation [147], among others. In conclusion, almost all trials have shown that the treatment methods for MSCs are at least safe. As knowledge of MSC mechanisms advances, future studies will focus on their potential use in treating additional diseases.

MSCs as CAR carriers

MSCs are recognized for their capacity to regulate the immune system, combat inflammation, and promote immune tolerance. They continuously produce or overexpress various proteins and exosomes, enabling direct delivery of therapeutic substances or mediating gene transfer. Thus, MSCs hold promise for use in gene delivery and regenerative medicine [148, 149]. Although MSCs have been extensively studied for the treatment of immune-related diseases and are generally considered safe, their therapeutic efficacy remains unsatisfactory. This may be attributed to inadequate targeting and tracking accuracy, as well as limited immunosuppressive capacity [150]. We propose addressing these challenges by optimizing targeting strategies and enhancing immunosuppressive mechanisms.

The CAR is a genetically engineered protein widely utilized in cell therapy. Its primary functions include (1) targeting and binding to specific antigens on the surface of target cells; (2) providing a stable binding interface to ensure sustained interaction between the CAR and the target; and (3) activating signal transduction pathways to trigger robust immune responses. CAR technology has achieved remarkable success in treating hematological malignancies, leading to FDA approval of CAR-T-cell therapy [47]. The intrinsic advantages of MSCs make them among the most promising candidates for developing CAR carriers. MSCs can be engineered to express CARs as molecular tools for identifying disease-associated markers.

In this way, compared with traditional CAR-T and CAR-NK cell therapies, CAR-MSCs demonstrate unique advantages in targeting precision, safety profiles, and application versatility. First, CAR-MSCs not only preserve the inherent immunomodulatory properties of native MSCs but also enhance the secretion of anti-inflammatory factors through engineered CAR signaling, thereby achieving a dual therapeutic effect of target localization and tissue repair [32, 148]. In contrast, CAR-T therapy may trigger CRS due to excessive T-cell activation. Allogeneic CAR-T cells carry the risk of GvHD, as T cells may recognize host antigens [48]. Although CAR-NK cells are associated with a lower risk of CRS, they may still release inflammatory cytokines [151]. Notably, MSCs themselves possess immunosuppressive properties that reduce the likelihood of immune rejection, eliminating the need for HLA matching and significantly lowering the risk of GvHD. Second, CAR-T cell therapy involves patient-specific customization, making the process time-intensive and expensive. While CAR-NK cells can be used in an allogeneic setting, their large-scale expansion remains technically challenging. MSCs, on the other hand, are readily available from diverse sources, can be efficiently expanded ex vivo, and maintain stable functionality after cryopreservation—making them highly suitable for mass production as off-the-shelf therapeutic products. Furthermore, CAR-T and CAR-NK cells are predominantly applied in hematological malignancies. However, T cells and NK cells often exhibit poor infiltration and are readily suppressed by the immunosuppressive TME in solid tumors [24, 151]. In contrast, MSCs naturally home to inflammatory and injured tissues and demonstrate superior tumor penetration. These core attributes render CAR-MSCs particularly well-suited for the treatment of autoimmune diseases and the stromal microenvironment of solid tumors. Current evidence suggests that CAR-MSCs hold the potential to become a transformative platform in the field of cell therapy.

Preclinical studies on CAR-MSCs

In GvHD

Recently, researchers at the Mayo Clinic have developed an innovative immunotherapy strategy that may lay the foundation for treating various immune diseases [12, 103]. This novel approach involves integrating CARs with MSCs, leveraging synthetic biology to engineer MSCs into specialized cell types known as CAR-MSCs. The development of CAR-MSCs enhances their ability to home to target tissues, which is mediated by a tissue-specific antigen-binding domain and further supported by the immunosuppressive properties conferred through the incorporation of an immunosuppressive intracellular signaling domain. The bioengineering process of CAR-MSCs has been validated as both safe and effective in maintaining stem cell characteristics, achieving a transduction efficiency exceeding 80% for the CAR construct in human AD-MSCs.

Researchers have utilized CD28ζ technology to genetically modify CAR-MSCs, enabling them to target E-cadherin (Ecad) for the treatment of GvHD. Ecad serves as a critical target for activated T cells in immune-related diseases. The modified CAR-MSCs demonstrated specific immunosuppressive effects in both in vitro and in vivo experiments, outperforming traditional MSC treatments. CAR-MSC intervention significantly alleviated symptoms in mice, suppressed T-cell activity, and increased survival rates. These findings indicate that the design of specific CAR constructs can generate CAR-MSCs with enhanced immunosuppressive capabilities while maintaining nontoxicity.

In tumor intervention

CAR-MSCs combine the targeting specificity of CARs with the immunomodulatory and tissue repair capabilities of MSCs, aiming to enhance therapeutic efficacy while minimizing side effects. This innovative strategy offers promising new avenues for the treatment of malignant tumors, particularly solid tumors. Currently, several studies have explored the application of CAR-MSCs in various tumor models, including GBM, Ewing’s sarcoma (ES), AML, and breast cancer, yielding encouraging results.

Golinelli et al. developed bifunctional MSCs that simultaneously express TRAIL and anti-GD2 CAR (GD2 tCAR), which can selectively target and eliminate GD2-positive GBM cells [13]. TRAIL is a ligand capable of inducing apoptosis in malignant cells but is relatively safe for normal cells [152]. Tumors of neuroectodermal origin, such as GBM, sarcoma, and neuroblastoma, frequently overexpress GD2 [153]. In the ES model, bifunctional MSCs can effectively target and eliminate ES cells, particularly demonstrating significant efficacy in the lung metastasis model [14]. In the treatment of leukemia, Aliperta et al. modified the single-cell-picked clone 1 (SCP-1) MSC line to continuously secrete CD33/CD3 bispecific antibodies (bsAbs), thereby redirecting T-cell toxicity to CD33⁺ leukemia cells and enhancing tumor-specific killing effects [154]. The findings of these studies collectively demonstrate that the MSC-based platform can function as a modular biomanufacturing system for precision immunotherapy. Unlike the aforementioned research preparation techniques, Tang S et al. employed a non-viral method to engineer iPSCs with stable overexpression of NKG2D-CAR, which were subsequently differentiated into MSCs. In a murine model of lung cancer, these NKG2D-CAR-expressing iPSC-derived MSCs exhibited enhanced tumor tropism and robust proliferative capacity [155]. Despite the limited number of studies on CAR-MSC treatment, their significant potential in breast cancer therapy has been highlighted in the literature. CAR-MSCs represent a promising therapeutic option for breast cancer and merit further exploration [156].

The findings from these studies suggest that genetically engineered MSCs expressing CARs exhibit robust anticancer and immunosuppressive properties (Table 3), mediated by targeting precision and immunomodulatory capabilities via gene transfer mechanisms and cytokine secretion (Fig. 3).

Table 3.

CAR-MSCs in preclinical trials

MSC Origin	CAR Structure	Disease	Optimization	Ref.
Human adipose tissue	EcCAR containing CD28ζ	GvHD	Immunosuppression; tissue regeneration; trafficking	[103]
Human adipose tissue	truncated anti-GD2 CAR with TRAIL	GBM	Targeting potential; anticancer activity	[13]
Human adipose tissue	truncated anti-GD2 CAR with sTRAIL	ES	Tumor targeting; persistence	[14]
SCP-1 cell line from human bone marrow	anti-CD33-anti-CD3 bsAb with CD137L	AML	T-cell redirection; facilitation of T-cell-mediated elimination of autologous leukemic cells; immune response	[154]
iPSCs	NKG2D-CAR	NSCLC	Migration and adhesion; tumor-homing ability	[155]

MSCs, mesenchymal stem cells; EcCAR, E-cadherin chimeric antigen receptor; GvHD, graft-versus-host disease; sTRAIL, soluble TNF-related apoptosis-inducing ligand; GBM, glioblastoma; ES, Ewing sarcoma; SCP-1, single-cell-picked clone 1; AML, acute myeloid leukemia; iPSCs, induced pluripotent stem cells; NSCLC, non-small cell lung cancer; Ref., reference

Fig. 3.

Schematic of CAR-MSC architecture and potential mechanisms of action in disease contexts. MSCs express CARs via gene transduction, thereby enhancing their tissue-targeting and immunoregulatory capabilities. Different ScFvs in the extracellular domain of CARs enable the targeting of distinct tissues, while varying intracellular signaling domains can result in differential functional enhancements, including immunomodulation, sustained proliferation, enhanced secretion, or cytotoxicity. Abbreviations: MSCs, mesenchymal stem cells; CAR, chimeric antigen receptor

Clinical studies on CAR-MSCs

Despite the promising therapeutic potential of CAR-MSCs, research on their antitumor efficacy and immunomodulatory properties remains in its nascent stages with a limited scope. Leveraging their promising results, Mayo Clinic’s Cellular Engineering Laboratory—the pioneer in CAR-MSC development—is designing a phase I trial to evaluate Ecad-CAR-MSCs with CD28ζ signaling in the treatment of steroid-refractory acute GvHD, with plans for subsequent multicenter studies in collaboration with industry partners [103]. However, detailed information regarding the survey (including inclusion/exclusion criteria, primary and secondary endpoints, and study duration) currently lacks supporting data.

To explore the use of MSCs in the treatment of GvHD, a series of randomized clinical trials led by our center (ChiCTR-IOR-15006330; ChiCTR-IIR-16007806; ChiCTR1900022292) has been conducted over a decade to produce three major achievements. In the first trial, we established that repeated UC-MSC infusions initiated at the 100-day mark post-transplantation markedly lowered the incidence and severity of cGvHD (27.4% vs. 49.0%; P = 0.02) [74]. The subsequent investigation demonstrated protective effects with early-phase infusion repetition from day + 45 + 81 posttransplant, resulting in a remarkable reduction in the cGvHD rate (5.4% vs. 17.4%; P = 0.03) [95]. Most recently, the third clinical trial extended this paradigm by evaluating sequential MSC infusions initiated within the first 3 months post-HSCT, representing the earliest intervention window among these studies [10]. Collectively, these phased clinical investigations substantiate that prophylactic MSC administration effectively reduces the incidence of cGVHD, with an excellent safety profile. CARs can be applied in future work on the basis of previous clinical trials by leveraging their antigen specificity and immune activation. Specifically, we aim to explore and compare the effectiveness of CAR-modified MSCs in further reducing GvHD at different time points.

CAR-MSCs may serve as an affordable “off-the-shelf” cell therapy platform applicable for the treatment of not only GvHD but also other immune-mediated diseases, such as Crohn’s disease and ulcerative colitis. As of August 2025, no clinical studies on CAR-MSC therapy have been registered on international clinical trial platforms, including ClinicalTrials.gov, WHO ICTRP, EudraCT, and ChiCTR. This gap highlights the urgent need for more clinical research to facilitate the translational application of CAR-MSCs.

Advances in CAR-MSC manufacturing technologies

The technology used to manufacture CAR-MSCs is critical for their clinical application and involves five main aspects: [1] optimization of the CAR structure; [2] stringent standards for MSC isolation; [3] precise gene editing; [4] GMP-compliant amplification; and [5] comprehensive quality assessment (including efficacy, identity testing, and sterility). While the production of CAR-MSCs can leverage existing technologies for CAR-T-cell therapy, parameter optimization must still be tailored to their unique characteristics.

First, compared with that of CAR-T cells, the design of the CAR structure for CAR-MSCs must consider the unique biological characteristics of MSCs. For example, MSCs exhibit low immunogenicity and broad regulatory functions; however, their antigen-specific recognition and cytotoxic capabilities are relatively limited. Consequently, molecular modules that increase migration and homing abilities are frequently incorporated into CAR designs.

Second, MSCs can be isolated from various tissues. However, owing to the limitations associated with BM collection, AD- or UC-MSCs may become more prevalent in the future. Collected samples are subjected to isolation and adherent culture in specialized MSC expansion medium under controlled conditions (5% CO₂ and 37 °C) to obtain a high-purity cell population.

Third, the core of CAR-MSC manufacturing involves introducing the CAR gene into MSCs. Common gene modification techniques include viral vector-mediated transduction and nonviral vector-mediated transfection. Viral vectors, such as lentiviruses, are widely used because of their high integration efficiency and stable expression. Lentiviral vectors have been successfully applied multiple times in CAR-MSC preparations [13, 14, 103, 154], enabling efficient CAR gene transduction while preserving MSC proliferation and differentiation capabilities. However, it is crucial to consider the impact of the viral titer and cell condition on transduction efficiency, as well as the potential carcinogenic risks and high costs associated with this approach. Researchers are developing new delivery systems, such as adeno-associated virus (AAV) and integrase-deficient lentiviral vectors (IDLVs), to improve viral vector design. These systems provide greater transduction efficiency, a reduced immune response, lower production costs, and safe genomic integration [157, 158]. Recently, Ye et al. developed an innovative gene delivery platform that combines Sleeping Beauty (SB) transposase mRNA with AAV-encoded SB transposons containing therapeutic transgenes [159]. The platform enables efficient CAR delivery to T cells while also achieving the successful transduction of NK cells, myeloid lineage cells, and iPSCs. Comparative analyses have revealed advantages over lentiviral vectors and electroporation-based delivery, including sustained transgene expression, enhanced activity, increased cell yields, and increased viability. This system is designed to address current limitations in cell engineering, particularly for MSC therapies that require persistent CAR expression. Although nonviral vectors (such as electroporation, nanoparticle, and transposon systems) have not been extensively reported for CAR-MSC preparation, they offer advantages such as low toxicity, high biocompatibility, and ease of scale-up production, warranting further exploration. Gene editing technologies, such as CRISPR, offer expanded opportunities for enhancing immunomodulatory functions or improving specific antigen recognition through the precise modification of the MSC genome [158, 160].

Fourth, genetically modified MSCs need to be expanded in vitro to obtain a sufficient quantity. Although MSCs have a strong ability to expand, multiple passages may lead to phenotypic and functional changes. Studies suggest that MSCs can be cultured for up to 20 passages [103], with genetic modifications using MSCs between P8 and P13 being viable [14]. As reported, human MSCs have been immortalized through lentiviral transduction via the gene encoding human telomerase reverse transcriptase (hTERT) [154]. In addition, cell proliferation and functional integrity can be maintained by optimizing culture conditions via the addition of growth factors or cytokines.

Fifth, stringent quality control protocols are indispensable in CAR-MSC manufacturing, encompassing comprehensive assessments of CAR surface expression uniformity (via flow cytometry or immunofluorescence), cellular purity, functional potency (proliferation kinetics, apoptosis resistance, targeted migration efficiency), and microbiological safety (sterility testing, mycoplasma screening). The implementation of these analytical benchmarks guarantees that therapeutic-grade cell products meet clinical translation requirements.

Despite notable advancements in CAR-MSC manufacturing technology, critical challenges persist in achieving clinical-grade cell products. Overcoming technical barriers, including suboptimal gene delivery efficacy, off-target risks, process variability across GMP facilities, and the implementation of phase-appropriate quality control platforms, remains imperative for successful translation from preclinical development to investigational new drug (IND)-enabling studies.

Potential challenges and strategies for CAR-MSC applications

Despite the enthusiasm elicited by preclinical studies reporting the potential of CAR-MSCs, the clinical translation of CAR-MSC applications has proven to be more challenging than expected. Owing to the limitations of cell therapies such as CAR-T cells and CAR-NK cells [161], several factors might pose potential difficulties for future researchers to focus on for improving the development of CAR-MSCs (Fig. 4). These include the following: [1] Ensuring the targeting ability of CAR-MSCs to the greatest extent possible. This might involve chemokine receptor expression, genetic manipulation of MSCs, and modification of MSC membranes. Additionally, the variance in the CAR structure is also a significant factor influencing the targeting force [2]. Further augmenting the antitumor and immunosuppressive effects of CAR-MSCs [3]. Prolonging the durability of CAR-MSCs. Although MSCs are highly desirable for reducing inflammation and promoting tissue regeneration, the harsh microenvironment in disease states can compromise cell survival, and the limited amplification capacity and sustainability of MSCs restrict their broader application [4]. Comprehensively optimizing the process of CAR-MSC preparation to attain the maximum therapeutic effect from the perspective of MSC source selection, culture protocol, condition, cryopreservation, and thawing; diverse donor characteristics, such as age and sex; type of CAR vector, e.g., viral or nonviral; and other factors [5]. Selecting the method of CAR-MSC administration for different diseases, as well as standardizing the treatment dose and frequency. In addition, the future clinical application of CAR-MSCs also calls for strict regulatory and approval procedures to ensure their safety and efficacy.

Fig. 4.

Potential challenges and strategic optimization for the clinical translation of CAR-MSCs. This multilayer pie chart provides a comprehensive overview of the potential challenges and anticipated strategies for addressing these challenges in the future development of CAR-MSCs. The middle layer highlights five major obstacles to CAR-MSC development and clinical translation, each denoted by a distinct color. The outermost layer, extending from each of the five segments, details specific solutions corresponding to these challenges. Abbreviations: MSCs, mesenchymal stem cells; CAR, chimeric antigen receptor; PCPs, positively charged patches

Increasing the targeted efficacy of CAR-MSCs

MSCs are valued for their tumor tropism. Studies have shown that MSCs respond to the selective recruitment of new cells induced by the TME during different stages of tumor growth, thereby promoting their migration to the tumor site after intravenous injection. This tumor tropism may involve three mechanisms, representing the main strategies for enhancing MSC tumor tropism identified in recent research. First, tumor cells secrete various chemokines and cytokines, such as monocyte chemotactic protein-1 (MCP-1), platelet-derived growth factor (PDGF), and VEGFα, into the TME. MSCs express receptors for these factors on their cell membranes, enabling them to be recruited to the tumor site [162, 163]. Second, the metabolic characteristics of the TME serve as a powerful driving force for MSC migration to the tumor site. Hypoxia can upregulate MCP-1 expression by increasing nitric oxide and HIF-1α production, which enhances MSC tumor tropism [164]. Third, MSCs express adhesion molecules on their membranes, such as vascular cell adhesion molecule-1, intercellular cell adhesion molecule-1, and activated leukocyte cell adhesion molecules, which are also expressed by leukocytes, monocytes, and DCs. These findings indicate that MSCs might share similar migration mechanisms with these immune cells [164].

In addition, studies have suggested that MSCs expanded in vitro for prolonged periods exhibit a weaker tumor-homing ability than freshly isolated MSCs do, with lower expression levels of CXCR4, possibly partially attributed to the influence of in vitro expansion conditions [165, 166]. Therefore, the migration and homing abilities of MSCs can be enhanced by pretreatment with growth factors (such as insulin-like growth factor 1 (IGF-1) and HGF) or cytokines (such as IL-3 and IL-6) to upregulate CXCR4 expression [167, 168]. Inserting the CXCR4-encoding gene via viral vectors is an alternative approach [169]. In summary, optimizing the tumor tropism of MSCs requires consideration of not only their intrinsic characteristics (internal factors) but also tumor features (external factors). The former encompasses factors such as heterogeneity within the cell population and molecular expression patterns involved in migration, whereas the latter includes aspects such as tumor type, tumor surface marker patterns, tumor burden, oxygenation status, and inflammatory state [170, 171].

Increasing the targeting ability of CAR-MSCs via CAR modification is crucial for obtaining functional CAR-MSCs. On the basis of strategies for optimizing CAR-T-cell targeting efficiency, the following potential approaches can be considered [1]. Dual (multiple) specific CARs: Designing CAR molecules to target two or more antigens simultaneously can mitigate antigen escape [172]. For example, the tandem CAR-T (tanCAR-T) strategy enhances targeting efficacy by linking recognition domains for antigens such as HER2 and IL13Rα2 [173] [2]. Boolean logic gates: Using AND, NOT, and OR logic gates improve CAR specificity [174, 175]. For example, the AND gate requires the coexistence of two distinct antigens to activate CAR-T cells, reducing off-target risk [176]. The synNotch receptor recognizes tumor-associated factors and induces the expression of a CAR that can recognize a second TAA, activating carrier cells [177] [3]. Exploiting the TME: Engineering CAR-MSCs to express chemokine receptors upregulated in target cells enhances MSC infiltration [178, 179] [4]. Masking CARs: Adding a masking peptide and protease-sensitive linker conceals the antigen-recognition domain until protease-mediated hydrolysis exposes it, leveraging high protease activity in the TME for improved targeting accuracy [180]. In addition, Volta et al. designed adaptor CAR-T cells (AdFITC-CAR) that specifically target fluorescein-labeled diabody adaptors binding to AML antigens. This system enhances targeting specificity and therapeutic efficacy against AML, demonstrating potential for antigen-versatile CAR-T-cell therapies [181]. Through continuous development of the CAR design, the improvement of signal transduction mechanisms, and the performance of rigorous clinical trials and validations, the targeting ability of CARs is expected to be significantly enhanced, offering a more promising and feasible outlook for the application of CAR-MSC therapy in the treatment of diseases.

Enhancing the anticancer or immunosuppressive efficacy of CAR-MSCs

While MSCs possess potential for anticancer therapy, their effects on tumors are dualistic. They can inhibit tumor growth through immune modulation and disruption of the TME. They might also facilitate tumor progression by potentially transforming into cancer-initiating cells and fostering a microenvironment that promotes proliferation and metastasis [9]. This underscores the complexity and challenges of using MSCs in cancer therapy. Future studies should therefore aim to clarify MSC mechanisms in cancer to establish safe and effective therapeutic strategies.

The following section outlines strategies to enhance the anticancer and immunomodulatory effects of CAR-MSCs [1]. Integration of domains or costimulatory molecules: The immunosuppressive capability of CAR-MSCs can be enhanced by incorporating the functional CD28ζ domain (which has been shown to outperform CD28 and CD3ζ alone) into the CAR [103]. This incorporation allows CAR-MSCs to transmit signals more efficiently and trigger an immune response upon recognition of target cells. Furthermore, in the design of CAR-MSCs, incorporating costimulatory molecules such as 4−1BBL can significantly enhance their cytotoxic effects on AML cells [154]. This strategy is linked to T-cell activation, thereby enhancing anticancer effects while minimizing damage to normal tissues. Clearly, the design of CARs and the optimization of corresponding CAR-MSCs warrant further investigation to achieve maximal efficacy across various diseases. Additionally, it remains unclear whether CAR-MSCs can interact with other immune cells to augment therapeutic outcomes [2]. Gene transmission and expression: Previous studies have shown that MSCs can be engineered to express TRAIL, thereby increasing their antitumor capacity [13, 14]. This has prompted researchers to consider CAR-MSCs as carriers for gene delivery, enabling the continuous or direct transfer of therapeutic genes into target cells to optimize treatment efficacy. Leveraging naturally occurring mutations represents a promising strategy. Garcia J et al. identified a gene fusion, CARD11-PIK3R3, in cutaneous T-cell lymphoma that markedly enhances the activity of the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and activating protein-1 (AP-1) signaling pathways, increases IL-2 production, and contributes to counteracting T-cell dysfunction. This fusion preserves phenotype memory and enhances antitumor efficacy [182]. This discovery offers a novel perspective on how insights gained from the evolution of malignant cells can improve a wide range of cell therapies [3]. MSCs can secrete exosomes that contain immunoregulatory molecules capable of modulating the behavior of immune cells [183–185]. CAR-MSCs can harness this mechanism to deliver specific immunoregulatory factors, thereby suppressing immune responses or enhancing antitumor immunity [4]. MSCs have the inherent ability to home to damaged tissues or tumors, allowing them to migrate directly to the TME [186]. This property can be harnessed to increase the accumulation of CAR-MSCs at the tumor site, thereby enhancing their anticancer efficacy [5]. As previously mentioned, by incorporating CAR technology, CAR-MSCs can recognize and bind to specific markers on the surface of diseased cells, such as E-cad, which serves as a target for overactive T cells in immune-related diseases [103]. This targeting capability enables CAR-MSCs to precisely identify and locate tumor cells or diseased cells.

There is undoubtedly significant potential for CAR-MSCs in both treating cancer efficacy and regulating the immune system. However, research in this field remains in progress, and further studies are needed to investigate the roles of differently CAR-MSC designs in various cancer types and diseases.

Reinforcing the proliferative capacity and persistence of CAR-MSCs

Proliferation refers to the ability of CAR-MSCs to divide in vitro or in vivo and is influenced by culture conditions, gene editing efficiency, MSC source, and cell state. Persistence refers to the duration for which CAR-MSCs survive and function in vivo. Proliferation and persistence are correlated; strong proliferation can enhance in vivo impact and persistence. However, persistence also depends on the microenvironment, antigen stimulation, and immune rejection.

Tonic signaling—the spontaneous activation of chimeric antigen receptors (CARs) in the absence of tumor antigen stimulation—plays a critical role in determining CAR-T-cell therapeutic efficacy. Chen et al. identified that positively charged patches (PCPs) on the CAR antigen-binding domain promote receptor clustering and initiate tonic signaling [187]. For CAR constructs exhibiting high basal signaling activity (such as GD2-CAR or CSPG4-CAR), strategies like reducing PCP numbers or elevating ionic strength during in vitro expansion can effectively suppress spontaneous activation and mitigate T-cell exhaustion. Conversely, adding PCPs to CARs with low tonic signaling (e.g., CD19 CAR) enhances in vivo persistence. These findings offer a strategy for improving CAR-MSC durability and adaptability through PCP modulation. Preventing CAR protein degradation and promoting CAR recycling may increase CAR-MSC persistence. Li W et al. reported that tumor antigen engagement triggers rapid CAR ubiquitination, leading to CAR downmodulation [188]. By mutating the lysine in the CAR cytoplasmic domain (CAR^KR) to inhibit ubiquitination, lysosomal degradation of the CAR protein is prevented. These findings suggest that CAR^KR-MSCs can reduce surface CAR loss, increasing long-term viability. The screening and identification of mature, safe candidate drugs that can be combined with CAR-MSCs to increase persistence and cytotoxicity represent promising strategies. A recent study demonstrated the antitumor efficacy of afatinib (AFA)-pretreated CAR-T cells in murine leukemia models [189]. Coculturing AFA with CAR-T cells during preparation revealed that AFA inhibits the TCR and PI3K/AKT/mTOR signaling pathways, reducing cell exhaustion and enhancing persistence and cytotoxicity. These findings provide valuable insights for future research on the development of pretreatment drugs to improve CAR-MSC antitumor cytotoxicity.

Unlike previous studies on CAR-T-cell regulation, Liu Y et al. combined a genetic code expansion system with CAR-T-cell therapy by introducing a stop codon into the CAR protein. The CAR protein expression and cytotoxicity of CAR-T/CAR-NK cells are controlled via an exogenous unnatural amino acid (BOCK). Specifically, BOCK activates CAR-mediated cytotoxicity; without BOCK, CAR activity is inhibited, keeping cells quiescent. Notably, BOCK deficiency reduces tonic signaling and CAR-T-cell exhaustion during in vitro preparation, increasing CAR-T-cell persistence and efficacy [190]. This reversible BOCK-modified CAR system offers an innovative approach for CAR-MSC development. This strategy improves both proliferation control and durability, optimizing treatment outcomes.

Researchers have explored the expansion of CAR-T cells via the use of engineered cytokines such as IL-15 and IL-7 and the transcription factor Forkhead box protein O1 (FOXO1) [191]. A recent study revealed that coexpressing IL-15 in GPC3-targeted CAR-T cells enhances their expansion, intratumoral survival, and antitumor activity in solid tumor patients [192]. Two recent studies identified FOXO1 as a factor that improves CAR-T-cell fitness and efficacy against solid tumors. Doan AE et al. demonstrated that FOXO1 overexpression induces a T-cell memory-associated gene expression program, conferring memory potential and metabolic fitness to CAR-T cells and enhancing in vivo persistence and tumor control [193]. Chan JD et al. revealed that FOXO1 overexpression promotes a stem-like phenotype in CAR-T cells, which is correlated with improved mitochondrial fitness, persistence, and therapeutic efficacy in vivo [194]. These findings suggest that genetically enhancing FOXO1 expression in CAR-MSCs is a promising strategy to improve proliferation, durability, and efficacy via favorable metabolic changes and high transformation potential. Furthermore, key transcription factors, such as TCF1, BACH2, MYB, and TOX, are critical in defining and maintaining the exhausted T-cell phenotype. Recent CAR-T studies have highlighted their importance in molecular interactions [195]. Specifically, bromodomain protein 4 (BRD4) inhibitors alleviate CAR-T-cell exhaustion and inhibit the terminal differentiation of exhausted T cells by downregulating BATF and upregulating EGR1 expression [196]. A deeper understanding of the molecular mechanisms underlying exhausted T cells could reveal innovative therapeutic strategies for CAR-MSCs.

Lee SS et al. recently demonstrated that MSCs briefly exposed to tropoelastin display an upregulation of proliferation-related genes and a concurrent downregulation of senescence-related genes. Importantly, the senoprotective benefits of tropoelastin persist throughout continuous and long-term MSC culture and significantly extend the replicative lifespan of MSCs [197]. The use of genetic engineering to modify the genome of MSCs is also a potent approach to increase the proliferation ability of CAR-MSCs or endow them with antiapoptotic characteristics. Jin HJ et al. reported that hUCB-MSCs with high expression of CD146 presented a greater growth rate and greater multidirectional differentiation, expression of stem cell markers, and telomerase activity [198], indicating that upregulation of the level of CD146 could contribute to promoting the proliferative capacity of CAR-MSCs. In addition, the overexpression of MIF rejuvenated aged MSCs through the activation of autophagy. The abrogation of miR-195 can activate telomere relengthening by increasing the expression levels of telomerase reverse transcriptase (Tert), thereby contributing to the rejuvenation of MSC senescence [199]. The overexpression of Erb-B2 receptor tyrosine kinase 4 (ERBB4) in senescent MSCs can reverse their aging phenotype and strengthen their resistance to oxidative stress [200]. The overexpression of Yes-associated protein (YAP) or forkhead box D1 (FOXD1) is also capable of rejuvenating hMSCs [201].

Strengthening the microenvironmental adaptability of CAR-MSCs enhances their durability. For example, low-oxygen culture inhibits MSC aging, preserves stem cell characteristics, and improves growth dynamics [202]. Studies have shown that supplementing media with fibroblast growth factor 2, PDGF-BB, ascorbic acid, and epidermal growth factor significantly increases MSC expansion in vitro [203]. Recent findings confirmed that the overexpression of sirtuins (SIRT1, SIRT3, SIRT6, and SIRT7) delays MSC functional decline and increases MSC differentiation capacity [204–207]. Improving the immunosuppressive environment, providing continuous antigenic stimulation, and combining therapies (e.g., anti-PD-1 blockade) may increase CAR-MSC persistence [151]. In conclusion, enhancing the proliferation capacity and durability of CAR-MSCs is a complex and challenging task that demands comprehensive and meticulous consideration from multiple perspectives.

Optimization of CAR-MSC production

Selection of MSC source

In CAR-MSC production, selection of the MSC source is a critical factor, with common sources including BM, UC tissue, adipose tissue, placental tissues, and others (Table 4). BM is one of the most frequently used MSC sources and is typically harvested from regions such as the iliac crest and sternum. However, BM-MSCs exhibit a limited proliferative capacity that decreases markedly with age, reducing both cell quantity and proliferative potential. The UC, an accessory tissue present at birth, is rich in MSCs. In terms of cellular abundance, the UC contains the highest concentration of MSCs among tissues. Moreover, UC-MSCs demonstrate a short doubling time, a strong colony-forming ability, and a multilineage differentiation potential that is retained even after 40 passages. These cells exhibit rapid proliferation and potent immunomodulatory capabilities. Adipose tissue, which is rich in both adipocytes and MSCs, can be harvested with minimal donor discomfort. AD-MSCs display robust in vitro expansion capabilities and low immunogenicity, making them highly suitable for CAR-MSC production. The amniotic membrane and placenta serve as alternative sources of MSCs, which exhibit superior immunomodulatory properties compared with those of BM- and AD-MSCs [208].

Table 4.

Characteristics of MSCs derived from diverse sources

Tissue Source	Proliferation	Differentiation	Immunomodulation	Availability and Ethicality
BM-MSCs	+	Advantages of three-line and vascular cell differentiation	Relatively weak immunomodulatory effects compared with AD-MSCs	Invasive, risk of patient discomfort, low cell content; ethically controversial
AD-MSCs	++	Adipogenic differentiation	Advantages in regulating DCs	Invasive, risk of patient discomfort, high cell content; ethically controversial
UC-MSCs	+++	More efficient osteogenic differentiation	Stronger immunomodulatory effects and lower immunogenicity	Noninvasive, painless, higher cell content; less ethically controversial
PL-MSCs	++	Better performance in adipogenic and vascular cell differentiation	Relatively weak inhibitory effect on T-cell proliferation	Noninvasive, painless, rich cell content; less ethically

BM-MSCs: bone marrow mesenchymal stem cells; AD-MSCs: adipose-derived mesenchymal stem cells; UC-MSCs: umbilical cord-derived mesenchymal stem cells; PL-MSCs: placenta-derived mesenchymal stem cells; DCs: dendritic cells; Ref., reference

Accessibility and safety are critical considerations that should not be ignored, as these properties vary among MSCs derived from different tissue sources. Obtaining BM-MSCs requires BM aspiration, which may cause harm. In contrast, AD- and UC-MSCs are easier to acquire with minimal donor damage. Certain sources, such as embryonic tissue, may raise ethical and legal concerns during selection. The phenotypic, morphological, and functional characteristics of MSCs can be modulated by various donor-specific variables, including age, gender, body mass index, and overall health condition [209]. Fetal MSCs exhibit superior properties compared to adult MSCs, including lower immunogenicity, stronger anti-inflammatory and proliferative properties, colony formation ability, osteogenic differentiation ability, higher osteogenic gene expression, and longer telomeres [210]. In brief, for CAR-MSC production, optimal sources should be evaluated on the basis of cellular traits, procurement feasibility and safety, and disease-specific needs. Stringent quality controls are essential to ensure consistency and therapeutic potential.

Freezing‒thawing of MSCs

MSCs are highly vulnerable to cryopreservation-induced apoptosis. During the freezing‒thawing process, a significant proportion of MSCs lose metabolic activity, and their immunosuppressive function and adhesion capacity are markedly impaired, posing a major technical challenge for obtaining sufficient active cells for therapeutic use [211, 212]. Thus, modulating biochemical pathways involved in postthaw degradation, particularly apoptosis, could increase the safety and efficacy of CAR-MSC applications, facilitating their development into economically viable and sustainable therapeutic options. Several studies have shown that Rho-associated kinase (ROCK) inhibitors improve the viability and physiological functions of cryopreserved stem cells and T cells after thawing [213, 214], suggesting valuable strategies for optimizing CAR-MSC engineering.

Several strategies have been proposed to reduce defects from the freeze‒thaw process in MSCs [215, 216]. These include using rate-controlled freezing devices, optimizing cryoprotectant formulations, and preconditioning MSCs with cytokines (e.g., IFN-γ) before cryopreservation, which enhances their “fast-start” performance after thawing. Cell culture rescue is another approach to improve engineered MSC products. This involves allowing thawed cells to recover under optimal conditions for a defined period to restore metabolic activity. The recovery time typically ranges from 24 to 72 h, depending on the study [215]. In summary, the optimization of the freeze‒thaw process enables CAR-MSCs to maintain pluripotency, differentiation potential, viability, and cytotoxicity after long-term storage and subsequent in vivo infusion, thereby improving therapeutic efficacy. This advancement is crucial for both cell biology research and clinical applications, providing a solid foundation for future developments.

Nonviral CAR-MSCs May hold promising prospects

In CAR product preparation, the viral vector is a critical raw material. Primary T cells are easy to transduce, but robust CAR transduction in MSCs requires significant optimization. Sirpilla et al. demonstrated that adding protamine sulfate led to high and stable CAR expression in MSCs, resulting in the first successful generation of CAR-MSCs [12]. The viral vector used for CAR transduction can be mass produced and stored stably at −80 °C for several years [217]. As final CAR-related products cannot be sterilized by filtration, open processing must be minimized, and quality control tests should ensure the sterility, purity, efficacy, and safety of the vector. Furthermore, selecting viral vectors, monitoring long-term safety, and transitioning to multisite production present challenges for manufacturing CAR-MSCs.

Concerns about the safety, high cost, extended production time, and scalability of viral vectors have led to research into nonviral alternatives for gene delivery [218]. Compared with traditional viral CAR therapies, nonviral CAR-T therapeutics do not use viral vectors to deliver CAR genes but instead modify T cells through genetic engineering techniques, such as mRNA delivery, programmable nucleases (ZFNs, TALENs, and CRISPR/Cas9) or transposon systems (SB, piggyBac, and Tc Buster), to reduce potential immunogenicity and safety issues [219]. Moreover, multiple nonviral transfection strategies—such as electroporation, lipid-based nanoparticles, or cell-penetrating peptides—can be employed to deliver CAR-encoding nucleic acids into MSCs, utilizing approaches analogous to those applied in NK cell engineering [220].

The conclusions of the above studies on nonviral methods for the production of CAR-T/NK cells reflect a highly consistent trend: virus-free CAR immune cell therapies are preferable to traditional viral vector-based therapies. For example, Bexte et al. reported that, compared with those modified by lentiviral vectors, SB-modified NK cells presented high and stable CAR expression and more frequent vector integration into safe genomic harbors. SB-generated CAR-NK cells demonstrated greater cytotoxicity than nontransfected NK cells did [221]. Additionally, novel hyperactive TcBuster transposase-engineered CAR-NK and CAR-T cells, as reported by Skeate JG et al., presented a low integrated vector copy number, a safe insertion site profile, robust in vitro function, and improved survival in a Burkitt lymphoma xenograft model [222]. These studies provide valuable insights for the future development of nonviral CAR-MSC engineering to fully realize the therapeutic potential of MSC-based immunotherapy.

Standardization of the CAR-MSC administration protocol

Before CAR-MSCs are used clinically, a standardized protocol must be established for selecting the optimal administration route, determining the dosage, and defining the frequency. This will ensure efficient delivery to target areas and maximize therapeutic potential. The administration route is critical for MSCs to reach tumor tissues. Recent studies have reported methods such as intravenous infusion, intra-arterial injection, intraperitoneal administration, intracranial injection, intranasal delivery, and direct intratumoral injection.

MSCs can be delivered to tumors and metastatic sites via intravenous injection through the vascular system. However, most MSCs (15–30 μm) become trapped in the pulmonary capillary network due to the “first-pass effect” [223, 224], limiting delivery efficiency. To overcome this, strategies include the preadministration of vasodilators to widen pulmonary capillaries, although this may reduce the blood supply by lowering blood pressure [225]. Alternatively, the cultivation of smaller MSCs or the use of MSC-derived exosomes or nanoparticles can enhance delivery [226]. Injection into specific veins improves organ targeting. For example, Li et al. reported greater homing selectivity with intraparenchymal hepatic vein injection than with intrahepatic or inferior vena cava injection in an acute liver injury model [227]. Intra-arterial injection of MSCs bypasses the first-pass effect, as most cells reach peripheral tissues before the lungs. This ensures high MSC bioavailability in peripheral organs and makes selective intra-arterial infusion a feasible strategy for targeting specific organs [228]. However, similar to intravenous administration, the large size of MSCs may cause arterial occlusion, highlighting treatment-related side effects due to their diameter. Owing to the extensive vascular network in the peritoneal cavity, MSCs administered intraperitoneally can access both the lymphatic and circulatory systems [229]. Kimura et al. reported that, compared with intravenous delivery, intraperitoneal administration of MSCs resulted in more focused localization within neuroblastoma tissue [230]. This may result from the lack of a first-pass effect in intraperitoneal administration. Furthermore, various routes of administration, such as intracranial (e.g., subarachnoid injection), intranasal, and intracavitary routes, have been explored, each with distinct characteristics. Intracranial administration offers an effective strategy for targeting intracranial tumors with MSCs by bypassing the blood‒brain barrier (BBB). Wang et al. demonstrated in a glioma mouse model that paclitaxel-loaded MSCs delivered via intracranial injection could migrate throughout the entire brain hemisphere [231]. Intranasal administration offers a noninvasive way to bypass the BBB, showing promise for brain targeting [232]. Studies have shown that treatment efficacy increases with increasing proximity of the injection site to the tumor. Repeated direct injection of MSCs into a tumor ensures better cell distribution in the TME and improves treatment outcomes [233].

Systematic regulations for MSC dosing and frequency remain insufficient. The typical infusion dose of MSCs ranges from 1 × 10⁶ to 2 × 10⁶ cells/kg, varying according to the study design. Infusion regimens may involve single or multiple doses based on the trial objectives. In summary, the administration route, dose, and frequency of CAR-MSCs are determined by the clinical trial design and therapeutic goals. Treatment plans should be tailored to individual patient conditions and study requirements. It is worthwhile to explore the genetic modification of CAR-MSCs, enabling them to express specific genes that increase their proliferation and persistence, potentially leading to a reduction in the required infusion dose in future applications.

Conclusion

The integration of CAR engineering and MSCs represents a novel impetus for cellular immunotherapy. Leveraging advancements in life sciences, the synergistic application of stem cell biology and immunotherapy strategies holds substantial promise for clinical translation. The use of CAR-MSC technology to enhance the endogenous stem cell reservoir and immune cell population may enable precise immune modulation while maintaining long-term systemic homeostasis. However, current CAR-MSC research is still in its infancy. The safety issue involves both the complexity of theoretical deduction and experimental verification, encompassing risks such as tumor formation, long-term survival and genetic stability, non-target tissue damage, and excessive immunosuppression. Breakthroughs in genetic engineering, biomaterial science, and single-cell analysis are expected to accelerate the development of CAR-MSC therapies. Through continued innovation and evidence-based optimization, this emerging therapeutic approach promises to address key challenges in precision medicine and open new avenues for the personalized treatment of malignancies and immune-mediated diseases.

Abbreviations

AAV Adeno

Associated virus

AD Adipose

derived

AFA 

Afatinib

AKT

Protein kinase B

AMI

Acute myocardial infarction

AML

Acute myeloid leukemia

AP-1

Activating protein-1

ARDS

Acute respiratory distress syndrome

BBB

Blood–brain barrier

BM

Bone marrow

Bax

Bcl-2-associated X protein

BCMA

B-cell maturation antigen

BRD4

Bromodomain protein 4

bsAbs

Bispecific antibodies

CAR

Chimeric antigen receptor

CAR-MSCs

Chimeric antigen receptor-engineered mesenchymal stem cells

CDCP1

CUB domain-containing protein 1

CLTI

Chronic limb-threatening ischemia

COVID-19

Coronavirus disease 2019

CRISPR

Clustered regularly interspaced short palindromic repeats

CRS

Cytokine release syndrome

CXCL12

C-X-C motif chemokine ligand 12

CXCR4

C-X-C chemokine receptor type 4

DCs

Dendritic cells

EASI

Eczema area and severity index

EGF

Epidermal growth factor

eGFR

Estimated glomerular filtration rate

Ecad

E-cadherin

ES

Ewing sarcoma

FDA

Food and drug administration

FGF

Fibroblast growth factor

FOXD1

Forkhead box D1

FOXO1

Forkhead box protein O1

GBM

Glioblastoma

GMP

Good Manufacturing Practice

GvHD

Graft-versus-host disease

HFrEF

Heart failure with reduced ejection fraction

HGF

Hepatocyte growth factor

HIF-1α

Hypoxia-inducible transcription factor-1α

HLA-DR

Human leukocyte antigen DR

HSCs

Hematopoietic stem cells

HSCT

Hematopoietic stem cell transplantation

hTERT

Human telomerase reverse transcriptase

ICAM

Intercellular cell adhesion molecule

IDLVs

Integrase-deficient lentiviral vectors

IDO

Indoleamine 2,3-dioxygenase

IFNγ

Interferon-γ

IGF-1

Insulin-like growth factor 1

IL

Interleukin

iNKT

Invariant natural killer T cells

iPSCs

Induced pluripotent stem cells

M1

Proinflammatory macrophage phenotype

M2

Anti-inflammatory macrophage phenotype

MCL-1

Myeloid cell leukemia sequence 1

MCP-1

Monocyte chemotactic protein-1

MHC

Major histocompatibility complex

MM

Multiple myeloma

mTOR

Mammalian target of rapamycin

MSCs

Mesenchymal stem cells

NF-κB

Nuclear factor kappa-light chain enhancer of activated B cells

NF-E2

Nuclear factor erythroid 2

NG2

Neuroglial antigen 2

NK

Natural killer

NPCs

Neural progenitor cells

OAd

Oncolytic adenovirus

OS

Overall survival

PAIS

Perinatal arterial ischemic stroke

PAX5

Paired box 5

PCPs

Positively charged patches

PDGF

Platelet-derived growth factor

PGE2

Prostaglandin E2

PI3K

Phosphoinositide 3-kinase

Ref

 Reference

ROCK

Rho-associated kinase

SB

Sleeping beauty

scFv

Single-chain variable fragment

SCP-1

Single-cell-picked clone 1

SDF-1α

Stromal cell-derived factor-1α

STC1

Stanniocalcin-1

SLE

Systemic lupus erythematosus

TAA

Tumor-associated antigen

TCR

T-cell receptor

TGF-β

Transforming growth factor-β

TME

Tumor microenvironment

TNF

Tumor necrosis factor

TRAIL

TNF-related apoptosis-inducing ligand

TRUCK

T cells redirected for universal cytokine-mediated killing

Treg

Regulatory T cell

UC

Umbilical cord

VCAM

Vascular cell adhesion molecule

VEGF

Vascular endothelial growth factor

YAP

Yes-associated protein

Author contributions

C.Z. and X.Z. defined the overarching research direction and provided critical guidance for this study; Y.C. drafted the manuscript; Y.C. and J.J.F. prepared the figures and tables; Y.C., J.L., Y.Y.M., Y.Y., and L.Y. revised the manuscript; all authors have read and approved the final version of the manuscript.

Funding

This study was supported by grants from the National Natural Science Foundation of China (82170212, 82370181, and 82341201), the National Key R&D Program of China (2023YFC2508905), the Chongqing Science and Health Collaborative Major Project on Medical Science and Technology Innovation (2024DBXM004), the Chongqing Medical Scientific Research project (Joint project of Chongqing Health Commission and Science and Technology Bureau) (2025CCXM002), the Clinical Research Special Project of the Second Affiliated Hospital, the Army Medical University (2024F010), and the Special Project for Talent Construction in Xinqiao Hospital of Army Medical University (2022XKRC001).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

All authors consent to publication.

Competing interests

The authors declare no competing interests.