The progression of cancer and development of metastasis is closely associated with the interrelation between cancer cells and their microenvironment, including immune responses – termed the tumor immune microenvironment (TIME). Innate and adaptive immunity are both involved with immune surveillance to eliminate pre-cancerous and cancerous cells, yet, transformed cells can evade immune control and generate immunosuppressive signals capable of modulating the activity of the immune system towards a pro-tumoral phenotype (1). Lymphocytes - including CD8⁺ cytotoxic and CD4⁺ helper T cells, natural killer (NK) and NKT cells – together with dendritic cells (DC) and pro-inflammatory macrophages (M1-like) have the power to induce an anti-tumoral response leading to tumor cell death. However, as tumor growth progresses, pro-tumoral populations of cells are recruited and generated at the tumor site, such as anti-inflammatory macrophages (M2-like), myeloid-derived suppressor cells [MDSCs, comprised of immunosuppressive monocytes, macrophages, and granulocytes (2)] and regulatory T (Treg) cells, all which act to counteract the anti-tumor response by suppressing the activity of effector cells and hampering the activation of newly recruited lymphocytes. Tumor cells also evade immunity by downregulating or losing expression of tumor antigens, releasing immunosuppressive molecules, and shedding soluble major histocompatibility complex (MHC)-I, among other mechanisms (1).

Collectively, the density and the diversity of the tumor-infiltrating cells affects prognosis and treatment response. For example, the levels of pro-inflammatory cytokines produced, and the degree of T cell infiltration are used to categorize tumors as either “cold” or “hot” (3). “Cold” tumors are non-T cell inflamed, characterized by T cell absence or exclusion, while “hot” tumors are T cell inflamed, characterized by T cell infiltration and molecular signatures of immune activation. This means “hot” tumors have higher response rates to immunotherapy (4, 5). Triple-negative breast cancer (TNBC) is, among all breast cancer subtypes, the most immunogenic (6–8), presenting with higher T cell infiltration, mutational burden, and PD-L1 expression (9, 10); however, when compared to other tumor subtypes, TNBC is considered a “cold” tumor (11–13). Despite recent advances in therapies, patients with TNBC remain at high risk for disease recurrence and early metastasis (14–17), with a combined 5-year relative survival rate of 77%; for patients with metastasized disease, that value lowers down to 12% (18). Therefore, innovative therapeutic strategies that leverage the immunogenicity of TNBC and enhance molecular signatures of immune activation provide an opportunity to improve patient outcomes.

One promising approach is cryoablation, a method that kills tumor cells via rapid freeze/thaw cycles while leaving the dead tumor cells in vivo. The exposure of the tumor to subzero temperatures (≤ -40°C) results in tumor necrosis, an inflammatory type of cell death that promotes immune recruitment while preserving tumor-associated antigens (TAAs) (19–21). Since the ablated tumor (and consequently, its TAAs) remains in vivo, this introduces the possibility of augmenting the anti-tumor immune response, leading to an improved systemic immune activity, also known as the abscopal effect ( Figure 1A ). This contrasts with surgical resection, in which tumors and their TAAs are removed (21). Currently, cryoablation is clinically approved for the treatment of breast fibroadenomas and has shown promising results in low-risk breast cancer (22–26). There is great interest in applying this technique to treat TNBC, which is more immunogenic than the above-mentioned subtypes and, therefore, could benefit from the augmentation of the availability of TAAs and inflammatory signals.

Clinical studies have shown the long-term abscopal effects of cryoablation in hormone receptor positive breast cancer patients with lymph node and/or liver metastasis, which regressed by 5-6 months after the treatment (27, 28), however clinical data for TNBC remains limited. In a recent long-term pre-clinical pilot study from our group, we showed local treatment of TNBC by cryoablation increased systemic immune responses compared to surgical resection where, six weeks after the procedures, 40% of the mice in the resected group showed local recurrence and lung metastasis, while no recurrence or metastasis occurred in mice post-cryoablation (29). Other groups observed similar superior immune activation with cryoablation compared to resection in breast cancer (30) and colon cancer models (31). However, the early drivers of the cryoablation-generated abscopal effect in TNBC remain unknown. Knowledge of such information could help guide decisions regarding follow-up therapies for TNBC patients and support the development of combinational therapies for cryoablation, by supplying the field with clinically actionable data, such as key immune populations and gene pathways directly affected by cryoablation. This information could be used for identifying suitable immune modulators, such as ICIs, and choosing which is the most appropriate for combination with cryoablation, ultimately providing patients with more efficient and safer therapies that can enhance tumor elimination and immune surveillance, improving survival.

Herein, we used the bilateral-tumor murine breast cancer model, flow cytometry, immunofluorescence and RNA-sequencing (RNA-seq), to identify differences in immune cell frequencies and immune gene signatures after cryoablation of primary tumors. We observed global T cell immune responses in distant tumors post cryoablation, which may help guide follow-up and combinational therapy decisions. Furthermore, we were able to use RNA-seq deconvolution from paraffinized tissue to determine the frequency of immune populations in the primary resected tumor, which may aid in determining optimal biomarkers to predict patients’ response to cryoablation in a clinical setting.

Surgically resecting aggressive tumors while maintaining cancer-free margins remains the mainstay of local therapy, however, the eventual prognosis largely dictated by tumor biology remains poor (43); though chemotherapeutic agents and radiation may aid in that scenario (44). Despite recent treatment advancements, particularly with the use of immunotherapeutic agents, TNBC patients remain at high risk for recurrence and metastasis (14–17) with addition of toxicities from these newer agents. Thus, strategies that elicit an improved anti-tumor immune response and overcome resistance are necessary for advancing the current standard of care. Cryoablation is a promising alternative to resection, since it kills the tumor cells and can provide a vast repertoire of preserved tumor cell associated antigens to boost the activation of immune cells (21). The activated anti-tumor T cells can then recognize both local and potential distant metastatic tumor cells. While cryoablation is already in clinical use to treat small/low-risk breast tumors, its clinical efficacy for TNBC remains unknown. Our prior study using animal models suggests cryoablation is a suitable substitute for resection, possessing the same ability to kill the tumor mass, while generating a superior immune response, characterized by increased infiltration of abscopal tumors by lymphocytes, and decreased metastasis rate at six weeks after the treatment (29). However, the early markers and the mechanisms that lead to the improved response are still unknown and knowledge of those will aid in the development of combinational therapies for cryoablation and pave the way for its use for TNBC.

Here, we show for the first time in a mouse model of TNBC that systemic changes are observable as early as one-week post-cryoablation. One such effect was reflected in tumor weights, with abscopal tumors from mice with cryoablated primary tumors presenting a significantly lower weight on day of sacrifice, than abscopal tumors from resection. The ability of cryoablation to induce superior control of tumor volume distantly at this early time-point suggests that the approach might result in improved surveillance for metastasis, which could significantly extend survival of TNBC patients, since currently only about 12% of patients with distant disease survive 5 years after diagnosis (18). Immune profiling by flow cytometry supports a model that early changes in immune cell frequencies contribute to the greater abscopal tumor elimination in mice from the cryoablation group. For example, we observed a significant increase in the frequency of NK cells in the abscopal tumors from cryoablation compared to abscopal tumors from resection, and trends for having more CD4⁺ and CD8⁺ effector/effector memory T cells.

NK cells are cytotoxic innate lymphocytes that can recognize and destroy tumor cells independently of major histocompatibility complex class I (MHC-I) molecules (45, 46). Tumor cells often downregulate the expression of MHC-I to avoid recognition by CD8⁺ T lymphocytes. Elegantly, NK cells activate when engaged with MHC-I-null cells resulting in direct tumor cell killing (46). Besides monitoring for the expression of MHC molecules, NK cells also detect stress-triggered self-ligands, which makes them able to recognize tumor cells even when they have retained MHC-I expression (47, 48). Therefore, having NK and T cells present in the tumor microenvironment is a powerful combination that allows the elimination of the tumor cells, where one cell can suffice this function when the other is not able to. Sabel et al. observed that NK cells obtained from the spleen of mice one-week after breast tumor cryoablation presented increased cytotoxic activity in vitro (30). In combination with our in vivo data, this suggests that cryoablation may influence the frequency and activity of NK cells. Future studies should evaluate the cytotoxic tumor killing ability of NK cells obtained from the abscopal tumors to completely elucidate this mechanism.

The molecules CD62L and CD44 have traditionally been used in flow cytometry to classify T cells into different subsets. Recently, a group defined the rarely studied double-negative (CD62L^-, CD44^-) CD4⁺ and CD8⁺ T cells as “pre-effector” cells, which have evidence to originate from naïve cells and can give rise to effector/effector memory T cells (33). In our results, we observed a significant decrease in the frequency of pre-effector CD4⁺ and CD8⁺ T cells in abscopal tumors from cryoablation, which was accompanied by a trend for having more effector/effector memory CD4⁺ and CD8⁺ T cells, compared to abscopal tumors from resection. These data suggest that cryoablation of the primary tumor might accelerate the development of effector/effector memory T cells in the abscopal tumors compared to resection.

The superior activation of CD4⁺ and CD8⁺ T cells could be induced by a higher frequency of migratory cDC1s (XCR1⁺ CD103⁺), which possess the unique ability to transport tumor antigens to lymphoid structures and excel at cross-presenting the TAAs to CD8⁺ T cells (49). At one-week post-cryoablation, we did not find statistical increases for those populations in the tumors. Nevertheless, we observed an overall trend for having more migratory DCs in the abscopal tumors and TdLNs in mice treated with cryoablation, and a significant increase in the frequency of that population in the spleens. Considering that one DC has the ability to scan over 500 T cells per hour (50), minor increases in their frequency may improve important T cell responses. Our results suggest the increased frequencies of migratory cDC1s may be one of the mechanisms resulting in heightened frequencies of effector/effector memory T cells. This model agrees with previous studies supporting the hypothesis that cryoablation improves anti-tumor immunity by recovering the function of DCs present in TdLNs (51) and enhancing the migration of DCs from the tumor to the TdLN (52). Further evidence that cDC1s play a role in the cryoablation-generated abscopal effect is supported by previous studies with other tumor models in which cryoablation in combination with dendritic cell therapy increased the anti-tumor immune response (53, 54). Indeed, the profile of immune cells in the axillary TdLNs is known to provide prognostic value for breast cancer patients, with increased CD4⁺ T cells and CD1a⁺ DCs correlating with higher rates of disease-free survival (55).

Another mechanism that may contribute to the improved anti-tumor response is the reduction of immunosuppressive signals in the abscopal tumors from cryoablation, partly provided by myeloid cells. We observed a significant decrease in the frequency of M1-like macrophages, and an overall trend for having fewer Ly6G^int neutrophils, immature macrophages, and inflammatory monocytes. M1-like macrophages are classically defined as anti-tumoral, while M2-like macrophages are pro-tumoral; however, these are only two states along a spectrum of different macrophage phenotypes (56, 57). For breast cancer, previous studies showed that having higher macrophage density (regardless of phenotype) correlated with poor prognosis (58, 59); therefore, having fewer M1-like macrophages at the abscopal tumors could be beneficial. Combined with the decreased immature macrophages and inflammatory monocytes signature, which are also considered pro-tumoral (60–63), the results suggest cryoablation alleviates immunosuppression at the abscopal tumors compared to those from resection. Importantly, Ly6G^int Siglec-F^hi neutrophils were established as immunosuppressive in lung cancer and acute spleen infections (34–36), and the trend for having decreased frequencies of those cells at the abscopal tumors from cryoablation corroborates the above-mentioned hypothesis, but further studies are needed to confirm the immunosuppressive activity of these cells in mouse breast cancer models. Additionally, it will be important to measure immunosuppressive activity of all populations mentioned here, to understand whether cryoablation is affecting functionality as well, or exclusively recruitment.

Besides the significant increase in the frequency of migratory cDC1s in the spleen of cryoablated mice, other immune cell frequencies were similar in abscopal tumors from cryoablation and resection in the spleen and blood. Our findings agree with results from Sabel et al., which showed that tumor-specific T cell responses after cryoablation were present locally (TdLN) with no major changes in the periphery (spleen), with the caveat that this study examined only cytokine secretion and cytotoxic activity, and not immune cell frequencies (30). Interestingly, we observed a trend for reduced systemic central memory CD4⁺ T cells. Central memory T cells are important to provide a fast and augmented immune response upon secondary exposure to previously encountered antigens and, therefore, are desirable in the tumor context. However, since our analysis was carried out to evaluate early events, it was expected that there would not be increased frequencies of central memory CD4⁺ T cell generation with cryoablation and it would be more appropriate to analyze this population at later time points in future studies.

We used RNA-seq analysis to identify potential mechanisms of the early immune response leading to an enhanced abscopal effect induced by cryoablation compared to resection. Results confirm that it is possible to observe the abscopal effects of cryoablation as early as one-week after the procedure, with multiple pathways being affected by the treatment, such as leukocyte chemotaxis, lymphocyte activation and NK cell activation, which aligns with the flow cytometry data. Some of the upregulated genes at abscopal tumors from cryoablation could be underlying the differences found at the protein level by flow cytometry; the increased expression of Il12b and Il2ra might be one of the mechanisms leading to higher NK cell frequencies at the abscopal tumors by inducing their generation, recruitment, and activation (64, 65). The increased expression of Pfr1 at abscopal tumors from cryoablation further suggests that NK/NKT and T cells at the TIME are potentially secreting more perforin, which could be one of the means leading to the decreased abscopal tumor weight, but additional functional experiments are needed to confirm this hypothesis. We observed that some of the DEGs in the abscopal tumors from cryoablation compared to those of resection suggest negative modulation of the anti-tumor activity, however, considering the decreased abscopal tumor size, increased infiltration by anti-tumor immune cells and increased anti-tumor gene signature, we believe that this could be part of a mechanism to prevent an exacerbated response that might cause damage to non-cancerous tissue.

Remarkably, GSEA analysis revealed upregulation of the IL-2/STAT5 and IL-6/STAT3 pathways in abscopal tumors from cryoablation compared to abscopal tumors from resection. Various studies showed that expression of STAT5 by tumor cells could act as either a tumor suppressor or oncogene in breast cancer under different circumstances (66), while STAT3 expression in tumor cells has mainly been associated with a pro-tumoral microenvironment (67). As discussed by Halim et al. (66), STAT5 signaling is thought to have oncogenic roles during tumor development, and tumor suppressor functions during tumor progression. Moreover, evidence suggests STAT5 has dominance over STAT3 (68) and that tumor suppressor activities of STAT5 can counteract the oncologic effects of STAT3 (69). Therefore, the results suggest that while also inducing higher STAT3 signaling, cryoablation induces STAT5 signaling, which has the potential to inhibit STAT3 oncogenic signaling and lead to reduced tumor cell proliferation, higher sensitivity to chemotherapy (69), and decreased angiogenesis, tumor cell migration, and metastasis (70, 71). This could be very important for breast cancer, since STAT3 is known to promote breast cancer malignancy and clinical trials have yet to show successful inhibition of its pro-tumoral effects; in that sense, cryoablation could make TNBC more amenable to STAT3-inhibitor treatment, however, confirmational analyses are needed to support these findings.

Under current standard of care practices, it is near impossible to do routine flow cytometry analysis for tumor immune cell infiltrate characterization and so H&E TIL counts are currently used for that purpose with biopsies (72). However, H&E TIL counts does not allow analysis of lymphocyte subpopulations. To identify immune populations without flow cytometry, we used an approach that would be technically achievable based on clinical protocols. We used tissue sections from histology slides (representative of tumor biopsy) for the RNA-seq analysis and deconvolution of the immune cells using Cibersort analysis, representing the most pragmatic clinical approach. We found that this technique allowed identification of multiple and different immune populations for each individual mouse. The goal of this approach was to develop a theoretical approach that can provide insight into the TIME that could be translated to clinical application for identifying best-fit TNBC candidates for cryoablation and to develop appropriate combinational therapies, such as use of immune checkpoint inhibition (ICI).

Lumpectomy with surgical axillary followed by adjuvant treatments is recommended for early-stage breast cancer by the National Comprehensive Cancer Network (NCCN) 2020 guidelines (73). Cryoablation is currently in use for small, low-risk tumors for patients who are not candidates for standard treatment or woman who refuse surgery, and is advantageous over other ablative techniques, such as laser thermotherapy and radiofrequency ablation (RFA), since it causes minimal discomfort and can be performed in the office with local anesthesia, while the others have to be performed in the radiology department, and require anesthesiology-administered sedation for pain control (74). ICE3, a recently completed prospective, multicenter trial found that hormone receptor (HR)-positive and human epidermal growth factor receptor 2 (HER2)-negative invasive ductal carcinomas (IDC) treated with cryoablation had a decreased 5-year overall recurrence compared to surgical resection. Subset analyses on patients additionally treated with adjuvant endocrine therapy and combination adjuvant endocrine and radiation therapy found decreased recurrence (75). There are clinical trials investigating cryoablation combined with ICI therapy, as well. With such a promising response in less immunogenic breast cancer, cryoablation could be the key to improving patient outcomes with TNBC. As of August 2024, there were 3 active trials investigating cryoablation in combination with immunotherapy in TNBC on ClinicalTrials.gov, 2 recruiting and 1 not yet recruiting, all utilizing pembrolizumab and one additionally utilizing ipilimumab and nivolumab. Patient selection is rooted in TNBC status and while these results will be interesting to see, screening patients further utilizing molecular profiling could help predict which patients may best benefit from cryoablation and checkpoint inhibition combination therapy, since triple negative tumors encompass both basal-like and non-basal cancers on genomic profiling (76). Further investigation into this would allow individualized treatment with the greatest impact in a poor-prognosis disease.

Although promising results were observed in the animal experiments reported here, some caveats are worth highlighting. Since mice are small, the cryoablation technique had to be adapted. In humans, the procedure is minimally invasive, where the probe is inserted into the breast tumor through a small incision in the skin with guidance from imaging techniques, such as ultrasound; a saline hydro dissection procedure is used to prevent skin frostbite (77). Meanwhile, in mice it is necessary to surgically retract the skin to expose the tumor and, instead of trying to insert the probe into these small tumors, it is laid directly on top of the tumors for freezing ( Figure 2A ). We acknowledge these modifications make it possible for the surgery to play a role in our results; however, based on the resection controls, we still see significant changes in the immune response to cryoablation compared to resection. Wu et al. showed similar changes in the abscopal tumors as observed in our study, using an analogous cryoablation technique for a model of colon cancer (31).

Besides its application for breast cancer, cryoablation has also been used to treat benign and malignant liver, colon, prostate, skin, kidney, lung and bone tumors, as well as non-tumor heart diseases (78, 79), with advantages and shortcomings involved. Similarly to results observed for breast cancer, cryoablation was shown to be just as effective as resection for small tumors (22, 79–82); however, for larger or multifocal tumors, cryoablation showed decreased efficiency, with higher rates of tumor progression and adverse events (83–86). Therefore, it seems that combination of cryoablation with other therapies would also be ideal for other tumor types besides TNBC, however, additional research is needed to elucidate the immune effects of the technique in those different microenvironments.

Overall, we observed an improved anti-tumor response due to primary tumor cryoablation. Our gene expression results were reflective of those changes and may provide physicians with clinically actionable information, such as knowledge of genes and pathways affected by cryoablation, which could be implemented to guide decisions for best therapy strategies. In that scenario, a promising follow-up therapy is the combination with immune checkpoint inhibitors, since those will allow sustained activation of effector cells (21, 87). The greater availability of tumor antigens and boosted inflammatory signals offered by cryoablation, combined with the relief of immunosuppressive signals provided by the immune checkpoint inhibitors, may result in a strong enough response to safely treat high-risk breast cancer.

Cryoablation induces a systemic abscopal effect, leading to increased frequencies of cytotoxic immune cells at distant tumor sites, which is accompanied by higher frequencies of migratory dendritic cells in the spleen, detectable as early as one-week after the procedure. Gene expression analysis supports those findings and reveals genes and pathways affected by cryoablation, providing the field with new information that could assist in developing combinational therapeutic approaches with cryoablation and potentially help guide therapy decisions.