Introduction to NK Cells in Cancer Immunotherapy

Natural killer (NK) cells are cytotoxic lymphocytes of the innate immune system that exert antitumor activity through the recognition of stress ligands on transformed cells without prior sensitization. Unlike T cells, NK cells rely on a balance of activating and inhibitory receptors to discriminate between healthy and malignant cells. Once activated, NK cells release perforin, granzymes, and engage death receptors such as FasL and TRAIL to induce apoptosis in target cells13, 14. In breast cancer, particularly in immunologically “cold” subtypes such as luminal A, NK cells face suppression from the tumor microenvironment (TME), which is rich in immunosuppressive cytokines such as TGF-β, IL-10, and PGE2. These signals downregulate NK-activating receptors like NKG2D and impair their infiltration and cytotoxic function5.

Clinical Applications and Comparative Context

Multiple therapeutic strategies have been developed to harness NK cells in breast cancer, including adoptive transfer of expanded NK cells, chimeric antigen receptor-engineered NK cells (CAR-NK), and bispecific/trispecific NK cell engagers. Among these, CAR-NK cells are particularly promising, combining the tumor specificity of CAR-T cells with reduced risk of cytokine release syndrome and graft-versus-host disease (GVHD). In the Phase I/II trial (NCT04220676), HER2-targeted CAR-NK cells demonstrated early signs of clinical activity in patients with metastatic breast cancer, with a favorable safety profile. However, in comparison to CAR-T therapy, which has shown dramatic effects in hematologic malignancies, CAR-NK therapies for solid tumors face additional barriers, including limited persistence and challenges in trafficking to tumor sites.

Compared to mesenchymal stem cell (MSC)-based therapies, NK cell-based immunotherapy offers a more direct cytotoxic effect but lacks the regenerative potential and immunomodulatory capacity of MSCs. Nevertheless, NK cells are faster acting and do not require prior priming. The therapeutic advantages of each platform suggest a synergistic potential, particularly in combining MSCs engineered to modulate the TME or secrete NK-stimulatory cytokines.

Barriers and Challenges in Clinical Translation

Despite advances, several hurdles limit the clinical success of NK cell therapy. The immunosuppressive TME is a significant barrier, as factors like TGF-β and IL-10 inhibit NK cell recruitment, cytokine release, and receptor activation. Moreover, the hypoxic and fibrotic nature of many breast tumors creates a hostile environment that limits NK cell infiltration and survival. Unlike CAR-T cells, NK cells have short half-lives post-infusion and require cytokine support (., IL-15 or IL-2) to sustain activity. ALT-803, an IL-15 superagonist, is currently under clinical evaluation for enhancing NK persistence. CAR-NK cells are a promising cellular immunotherapy for cancer. Patient heterogeneity, such as differences in tumor antigen expression (., HER2, MUC1), HLA genotype, and immune checkpoint ligand density, can dramatically impact CAR-NK efficacy5, 15. These factors influence target recognition, persistence, and immune escape, contributing to variability in clinical outcomes.

Another translational barrier is manufacturing scalability. Unlike conventional pharmaceuticals, NK cell therapies must be produced under Good Manufacturing Practice (GMP) conditions, with strict batch-to-batch consistency. Autologous NK cells present challenges of variable quality and limited expansion, while allogeneic sources raise concerns about immune compatibility. The production cost of NK therapies remains high, with limited automation and regulatory complexity slowing commercialization.

Personalized Strategies, Biomarkers, and Artificial Intelligence

Personalizing NK therapy requires a better understanding of predictive biomarkers. Tumor HLA class I expression plays a pivotal role in regulating NK cell activity via inhibitory KIR receptors; tumors with low HLA-I expression are more susceptible to NK-mediated lysis. Additionally, the presence of activating ligands (., MICA/B, ULBPs) for receptors like NKG2D correlates with better outcomes. Recent studies also suggest that stromal expression of CXCL9/10, which recruits CXCR3 NK cells, could be used as a predictive indicator of NK cell infiltration.

AI is transforming NK cell therapy by enabling the identification of such biomarker signatures through high-throughput multi-omics data analysis. For example, AI-driven algorithms have been used to stratify triple-negative breast cancer (TNBC) patients based on NK gene expression profiles, improving patient selection and treatment matching12. Additionally, machine learning models are aiding in optimizing CAR designs and predicting potential off-target toxicities by simulating ligand–receptor interactions before preclinical testing. Recent work has applied convolutional neural networks (CNNs) to classify histopathological images and predict NK cell infiltration in breast tumors, improving immune landscape profiling16, 17. Support vector machines (SVMs) and random forest models have also been trained on transcriptomic and single-cell RNA-seq data to identify NK-sensitive tumor subtypes and stratify patients likely to benefit from CAR-NK therapies18, 19. These algorithms enhance patient selection and make preclinical testing more targeted and efficient.

Future Integration and Synergistic Approaches

Looking forward, the most promising strategies involve combining NK cells with complementary approaches that overcome the limitations of monotherapy. For instance, immune checkpoint inhibitors such as anti-TGF-β or anti-PD-L1 antibodies can be used to reverse TME-mediated suppression. Gene-editing tools like CRISPR/Cas9 are being employed to delete inhibitory receptors or insert cytokine-supportive genes into NK cells, thereby improving their survival and tumor-killing potential. Additionally, metabolic reprogramming targeting pathways such as mTOR or glycolysis may enhance NK function under nutrient-deprived conditions within the TME.

Critically, synergy between NK cells and MSCs is an emerging frontier. MSCs can be engineered to secrete IL-15 or other NK-activating cytokines locally within the TME, enhancing NK infiltration and cytotoxicity without systemic toxicity. Furthermore, MSCs may be used as delivery vehicles for NK-engaging molecules, bridging innate immunity and tissue repair. This dual-modality approach could transform post-surgical management of breast cancer by simultaneously addressing residual tumor burden and promoting tissue regeneration, a concept explored further in the next section.

Introduction to MSCs in Breast Cancer

Mesenchymal stem cells (MSCs) are multipotent stromal cells capable of differentiating into osteoblasts, chondrocytes, and adipocytes. Their tumor-homing ability, immunomodulatory functions, and low immunogenicity make them attractive for cancer therapy and regenerative medicine applications. In breast cancer, MSCs have been investigated for their dual utility: as therapeutic vehicles for targeted anti-tumor delivery and as facilitators of soft tissue regeneration following mastectomy20. These cells can migrate to sites of inflammation and tumor activity, enabling site-specific delivery of therapeutic agents with reduced systemic toxicity.

Genetic Modification for Enhanced Therapeutic Efficacy

Genetic engineering of MSCs is a key strategy to enhance their therapeutic specificity and efficacy. For example, Liu (2018) demonstrated that MSCs engineered to express interleukin-18 (IL-18) significantly suppressed tumor growth in a murine model of breast cancer by enhancing local immune activation and reducing angiogenesis21. Similarly, MSCs engineered to express the suicide gene cytosine deaminase (CD) can convert the prodrug 5-fluorocytosine into the chemotherapeutic agent 5-fluorouracil (5-FU), leading to localized cytotoxicity in the tumor microenvironment while minimizing systemic toxicity22. These genetic modifications allow MSCs to act as “smart” delivery platforms for targeted cancer therapy.

Safety Concerns and Mitigation Strategies

Despite their promise, the use of MSCs in oncology raises safety concerns, particularly regarding their potential to support tumor growth, metastasis, or immune evasion under certain conditions. Some studies have reported pro-tumorigenic effects of MSCs, including secretion of growth-promoting factors and immune-suppressive cytokines. To address these risks, several safety engineering strategies have been proposed. One approach involves integrating suicide gene systems into MSCs, such as inducible caspase-9 (iCasp9), which can be pharmacologically activated to eliminate MSCs if adverse effects occur23. Additionally, preconditioning MSCs with cytokines or modifying their culture conditions can promote an anti-tumorigenic phenotype. Another emerging strategy is to use extracellular vesicles (EVs) derived from engineered MSCs instead of whole cells to mitigate risks associated with cell proliferation and transformation.

Clinical Translation: Challenges and Early Trials

The clinical translation of genetically modified MSC therapy in breast cancer is still at an early stage. A Phase I clinical trial (NCT02807805) investigated the safety of MSCs expressing interferon-beta in patients with advanced solid tumors, including breast cancer. The trial demonstrated that the therapy was well-tolerated with no dose-limiting toxicities, although therapeutic efficacy was modest, highlighting the need for optimization in delivery, persistence, and tumor targeting (ClinicalTrials.gov, 2024). Translational challenges also include variability in MSC function across donors, difficulty in large-scale expansion under Good Manufacturing Practice (GMP) conditions, and complex regulatory oversight due to the use of genetically modified cells.

Integration with Artificial Intelligence and Omics Technologies

Artificial intelligence (AI) and omics technologies are increasingly being used to optimize MSC-based therapies. Transcriptomic and proteomic profiling of MSCs can identify subpopulations with enhanced tumor-homing or immunomodulatory potential. AI algorithms trained on secretome or surface marker data can predict therapeutic potency and assist in donor screening, bioprocess optimization, and batch quality control20. Integration of such tools into the MSC manufacturing pipeline offers opportunities to reduce variability, improve clinical predictability, and streamline regulatory compliance.

MSC–NK Cell Synergy: A New Therapeutic Paradigm

Recent studies suggest a promising synergy between MSCs and natural killer (NK) cells in the treatment of solid tumors. MSCs can be engineered to secrete cytokines such as IL-15 or IL-21, enhancing NK cell activation and infiltration into tumor sites. This combinatorial approach has shown superior tumor suppression in preclinical models compared to monotherapies20. Importantly, such an integrative strategy may provide dual therapeutic benefits, targeting residual cancer cells via NK cells and promoting post-surgical soft tissue repair through MSC-mediated regeneration.

Implant-Based vs. Autologous Approaches: Opportunities and Limitations

Implant-based reconstruction (IBR) remains the most commonly performed method due to shorter surgery time and hospital stay. Advances such as acellular dermal matrices (ADMs) have reduced capsular contracture and improved cosmetic outcomes. However, IBR is often contraindicated in patients receiving radiation therapy because of poor tissue quality and higher complication risks. Autologous reconstruction, using flaps like TRAM or DIEP, offers natural results and is better suited for irradiated patients. Yet, these procedures are resource-intensive, require microsurgical expertise, and are associated with donor-site morbidity. Additionally, volume loss and fat necrosis remain concerns even with successful flap transfer.

Fat Grafting and Stem Cell Integration

Fat grafting, commonly used as an adjunct to refine shape and contour, has gained attention for its regenerative potential. Adipose-derived stem cells (ADSCs) within lipoaspirates are thought to promote angiogenesis and improve fat graft survival. Clinical studies have shown that ADSC-enriched fat grafts can improve volume retention compared to standard fat grafting. However, concerns about ADSCs promoting tumor recurrence in the post-mastectomy setting persist. Trials such as RESTORE II (NCT01555654) and the ongoing SAFE-ASC trial are evaluating long-term oncologic safety. To mitigate risks, newer approaches are using ASC-derived EVs instead of live cells, showing reduced fibrosis and improved fat graft retention in irradiated models.

Scaffold-Based Tissue Engineering in Reconstruction

Tissue engineering has enabled the development of scaffolds that provide structural support and guide tissue regeneration. As summarized in Table 1 and Supplementary Table S1, materials such as ADM, dECM, and GelMA show high compatibility with MSCs and support vascularization, both critical for sustained graft integration. For instance, decellularized adipose tissue scaffolds seeded with ADSCs have shown enhanced adipogenesis and angiogenesis in animal models. Clinical trials are now assessing bioengineered constructs with MSCs or ASCs to improve reconstructive outcomes (, NCT04892537). However, translating these materials into routine clinical use demands resolving challenges such as GMP-compliant production, immunogenicity control, and cost-effectiveness in low- and middle-income settings.

Among the various scaffolds explored for breast tissue engineering, natural materials such as acellular dermal matrices (ADMs), collagen, and decellularized extracellular matrices (dECMs) offer superior biocompatibility and vascularization potential—crucial for long-term graft survival and integration. ADMs, in particular, are already used clinically in implant-based breast reconstruction. In contrast, synthetic polymers like PLGA and PEG offer tunable mechanical and degradation properties but often require a combination with growth factors or cells to enhance angiogenesis. Natural hydrogels such as fibrin and silk fibroin also show promise due to their pro-angiogenic and cell-supportive environments, although their mechanical strength may be a limitation. The choice of scaffold depends heavily on the intended application—be it structural support, delivery of MSCs, or long-term regeneration. As detailed in Supplementary Table S1, specific scaffold–cell pairings such as ADSCs with decellularized adipose tissue (DAT) and GelMA hydrogels have shown robust adipogenic differentiation and angiogenesis . These findings directly inform clinical scaffold selection, especially for patients undergoing reconstruction post-radiation, where enhanced vascularization is needed to counter fibrosis and improve graft retention30. Such preclinical evidence provides a translational bridge toward selecting biomaterials tailored for patient-specific reconstruction challenges.

Clinical Trials and Translational Progress

Several early-phase trials have demonstrated the safety of MSC-based therapies in solid tumors, including breast cancer. In trial NCT02807805, IFN-β–secreting MSCs were administered to patients with metastatic cancers and showed no dose-limiting toxicity, although clinical efficacy was modest. Similarly, NCT05109929 is currently exploring MSC–NK cell co-therapies for their potential to modulate the immune microenvironment post-surgery. To transition from feasibility to routine use, MSC therapies must overcome donor heterogeneity, inconsistent potency, and safety concerns. Standardization of MSC manufacturing (, cytokine release profiles and differentiation assays) is a critical next step. Additionally, regulatory agencies like the FDA and EMA now require functional potency assays as part of investigational new drug applications.

Personalized and Integrated Reconstruction Strategies

With the convergence of imaging, AI, and 3D bioprinting, breast reconstruction is becoming increasingly individualized. Patient-specific scaffolds designed via CAD modeling can match anatomical contours and tissue volume more precisely than traditional methods. Ongoing studies aim to evaluate whether AI-optimized scaffold architectures improve fat retention and vascularization outcomes compared to conventional flaps. A truly personalized strategy may involve integrating pre-vascularized scaffold systems to enhance graft survival, acellular EV therapies to reduce fibrosis, and patient-specific imaging and omics data to tailor scaffold–cell combinations.

Regulatory and Manufacturing Barriers

A major hurdle in clinical translation is the lack of standardized manufacturing protocols for MSCs. Variability in donor source, isolation techniques, and cell expansion introduces significant heterogeneity in therapeutic performance. Regulatory agencies such as the FDA and EMA now require validated potency assays, such as cytokine secretion profiles, immunosuppressive function (., IDO, TGF-β1), or angiogenic capacity, as part of MSC release criteria. Establishing GMP-compliant processes and quality control standards remains a priority before large-scale implementation.

Safety Concerns in Oncologic Settings

Concerns persist about the potential pro-tumorigenic effects of MSCs, particularly in the post-mastectomy bed, where dormant tumor cells may remain. While several studies report a lack of tumor-promoting behavior in immunocompetent models, data in irradiated or high-risk patients remain sparse. For instance, the RESTORE-2 trial (NCT01555654) evaluated ASC-enriched fat grafting in breast reconstruction and found no increase in local recurrence over three years. Similarly, Lee . (2023) demonstrated that ASC-derived exosomes reduced radiation-induced fibrosis without stimulating tumor markers in preclinical models31.

Clinical Trial Landscape and Current Status

Several early-phase clinical trials have evaluated MSC-based therapies in breast cancer patients. In NCT02807805, MSCs engineered to express interferon-β were administered systemically to patients with advanced solid tumors. While no dose-limiting toxicity was reported, antitumor efficacy was limited. The ongoing NCT05109929 trial is evaluating the co-administration of NK cells and MSCs in post-surgical breast cancer patients to assess safety and immunomodulatory outcomes. In the context of scaffold-assisted therapies, NCT04892537 is exploring the use of 3D-printed biodegradable scaffolds seeded with ADSCs to enhance volume retention and reduce fibrosis in autologous reconstruction. These trials highlight the clinical potential of MSC-based interventions but also the slow progression of such approaches due to regulatory, ethical, and technical complexities.

Future Trial Design Considerations

To move the field forward, future trials should incorporate standardized endpoints, such as long-term fat graft viability, fibrosis scoring, and oncologic surveillance over five or more years. Stratification of patients by prior radiation exposure, immune profile, and genetic risk may reveal responder subgroups. The use of acellular MSC-derived EVs may offer a regulatory and safety advantage over cell-based methods, particularly in high-risk populations. Multi-arm comparative trials—comparing standard autologous flap reconstruction, fat grafting, MSC-assisted scaffolds, and exosome-enriched therapies—are warranted to guide clinical best practices. As this translational pipeline matures, collaborative consortia among academic institutions, regulatory bodies, and industry partners will be essential to establish robust, safe, and scalable therapeutic platforms.

Enhancing the Immunoregenerative Interface

Future research must address the interaction between therapeutic cells and the immunosuppressive tumor microenvironment (TME). While NK cells and MSCs show promise individually, their combined application remains poorly understood. Investigating how MSC-derived cytokines (, IL-15, IL-21) modulate NK cell persistence and cytotoxicity could uncover new therapeutic synergies. Moreover, fine-tuning this balance to avoid MSC-induced immune dampening in cancerous contexts will be critical32.

Scaffold Optimization Beyond Mechanics

While materials science has focused on mechanical compliance and degradation kinetics, the next wave of scaffolds must incorporate biological intelligence: dynamic release of immunomodulators, spatially programmed angiogenesis, and responsiveness to inflammatory cues. Decellularized adipose tissue and ECM-mimetic hydrogels offer such potential, but comparative studies are needed to determine which materials best support long-term graft survival, especially in irradiated beds33.

Safety and Standardization in Cellular Therapies

The clinical translation of MSC- or ASC-based strategies remains constrained by donor variability, potency uncertainty, and long-term oncologic risk. While emerging acellular options like exosomes and EVs may mitigate these concerns, their mechanisms of action, dosing strategies, and persistence need a clearer definition. Moreover, global standardization of potency assays will be essential for regulatory progression20, 31.

Personalized and AI-Driven Reconstruction

Advances in imaging, 3D bioprinting, and AI-enabled design suggest a future in which breast reconstruction is personalized to each patient’s anatomy, biology, and treatment history. However, the integration of patient-specific immune profiles, radiation exposure, and tissue quality into scaffold or graft selection remains underexplored34.

Cross-Disciplinary Integration

Perhaps the greatest opportunity and challenge lies in bridging traditionally siloed disciplines such as oncology, biomaterials, computational biology, and surgery. Effective translation will require not only technological innovation but also crosstalk between immune modulation and regenerative design.