Tumor-Infiltrating Lymphocytes (TILs) as a Biomarker of Abscopal Effect of Cryoablation in Breast Cancer: A Pilot Study

All mice successfully developed bilateral tumors and were subsequently divided into two experimental groups for local treatment: (i) Standard of care - Resection group, and (ii) Cryoablation group. The mean (SD) tumor volume for T_Baseline, Abs_Res, and Abs_Cryo was 279 (41) mm³, 276 (122) mm³, and 224 (56) mm³, respectively. There was no significant difference for tumor volume or weight between any of the three tumor groups. Cryoablated tumors remained in vivo and showed complete destruction by in vivo imaging (Fig. 2). For the cryoablation group, the average freeze/thaw/freeze cycles were 84/180/81 seconds. At 24-hour posttreatment assessment of the two local treatment procedures, no mice showed luciferase activity at the treatment site, indicating complete tumor ablation or tumor resection; however, the nontreated distant tumors still had a high fluorescent measurement (Fig. 2C).

One week after local treatment, the distant tumor was resected and labeled according to schema based on local therapy of primary tumors (Abs_Res or Abs_Cryo). All tumors were analyzed for TILs (Fig. 3). T_Baseline were remarkably consistent for overall TIL percentages. There was no correlation between TIL scores to tumor volume with all resected tumors having a stromal TIL score of 1% (Fig. 3B; Table 1). Both Abs_Res and Abs_Cryo had an overall increase in stromal and invasive margin TILs compared with T_Baseline. There was a significant increase in the mean (SD) stromal (2.8 [1.1]%; p = 0.015) and invasive margin (50 [12.2]%; p = 0.02) TILs for the Abs_Cryo group compared with T_Baseline of 31 (4.9%) (Fig. 3; Table 1). The Abs_Res did not show significant increase from T_Baseline.

TILs were further assessed for changes in CTL markers in the Abs_Res and Abs_Cryo compared with T_Baseline by immunofluorescent staining for CD3ε (T cells), CD8α (CTLs), and granzyme B (activated CTLs) (Fig. 4). The Abs_Cryo tumors had an overall increase in CD3ε normalized fluorescence staining compared with both the T_Baseline and Abs_Res tumors. There was a slight increase in CD8α staining in the Abs_Res group compared with T_Baseline, but virtually no change in CD3ε or granzyme B staining. It is noteworthy that in the Abs_Cryo tumors, there was a significant increase in mean (SD) normalized fluorescence for CD8α (15.7 [12.1]; p = 0.020] and the CD8 CTL activation marker granzyme B (4.8 [3.6]; p = 0.048) compared with T_Baseline (5.2 [4.7] and 2.4 [0.09], respectively), indicating increased activated CTLs in the tumor (Fig. 4).

An evaluation comparing the standard of care group to the cryoablation group showed that cryoablation was superior to resection in preventing tumor recurrence (Fig. 5). At 4 weeks post-cryoablation, all mice were tumor-free by palpation and IVIS analysis. At 6 weeks post-cryoablation, two mice in the resection group had pin-point tumors at the resection sites by palpation, which for one of the mice was easily detected by IVIS by 7 weeks (Fig. 5A). Two of five (40%) mice in the resection group versus none of the five mice in the cryoablation group developed local and systemic failure of disease. Both tumor-positive mice had to be sacrificed due to humane endpoints, whereas the cryoablation group all appeared healthy and showed no signs of disease at sacrifice time points (Fig. 5B). Moreover, the cryoablation procedure showed minimal scar tissue with healthy mammary tissue at time of sacrifice (Fig. 5C). Of the mice in the resection group that developed tumor recurrence, one was at the initial resection-treated tumor site and the other was at the excised Abs_Res site (Fig. 5D; Table 1). Neither tumor had substantial number of TILs with 1% central tumor and less than 6% invasive margin TILs. Of note, two of the five mice in the resection group that did not mount a central tumor TIL response were the same mice that failed long-term control of tumor growth. Proportion test showed significant association between increase in TIL% and prevention of tumor recurrence (p = 0.02). Overall, TIL scores for the recurrent primary tumors was much lower than any of the previous tumor TIL scores analysis. The stroma TILs remained at 1%, whereas the intratumor TILs decreased to 0% and the invasive margin TILs decreased to 3.5% and 5.5%. Mice that developed tumor recurrence had increased spleen and liver weights (Fig. 5E) compared with the other mice with both mice developing splenomegaly (Fig. 5E) most likely due to tumor-induced granulopoiesis. Additionally, mice that had tumor recurrence exhibited metastasis to the lung (Fig. 5F) and suspicious regions in the liver (data not shown). These results indicate that local cryoablation therapy generated a strong systematic antitumor immune response and long-term antitumor immunity that protected against tumor recurrence and metastases.

Among the hallmarks of cancer is a competition between the tumor attempting to escape immune surveillance and antitumor immunity versus the immune system recognizing and eliminating the tumor.²⁸ This phenomenon is being extensively investigated in the local and systemic environment created by the tumor and attempted therapies. Whereas local antitumor immunity is studied within the tumor microenvironment, the systemic response is a subject of investigation in terms of abscopal effect. The tumor microenvironment, comprised of immune cells, signaling molecules, extracellular matrix, adjacent blood vessels, and fibroblasts, is the battleground and has the potential to provide an advantage to one side, depending on its composition. Immune cells, in particular, play an essential role in this battle by providing the factors that comprise the tumor microenvironment—both adding to the antitumor defenses and blocking them. Even within lymphocytes, there are subtypes on either side. For example, tumor infiltration by CD8⁺ effector T cells is associated with significant antitumoral response and longer survival in breast carcinoma.¹³ Conversely, T_Reg cells, which normally suppress self-reactive T cells to fight autoimmune diseases, are believed to contribute to tumor tolerance by suppressing other immune cells.¹³ However, the overall percentage of lymphocytes in the tumor bed observed without subtyping are still predictive of better prognosis in TNBC and HER2⁺ breast cancer.^12,18‐20

Breast cancer TIL score predicts the effectiveness of breast cancer treatment and prognosis.^12,18‐20 Breast cancer patients with a higher TIL score on diagnostic biopsy specimens appear to do better with adjuvant and neoadjuvant systemic therapy and result in less tumor metastasis.^12,18‐20 The robust evidence supporting TIL scoring as an important prognostic marker led to the formation of the International Immuno-Oncology Biomarker Working Group on Breast Cancer that recommends regular TIL reporting on breast carcinomas following the guidelines provided in their 2017 report.²¹ However, TIL scores have mainly been reported within the context of primary tumor. To the best of our knowledge, no studies have looked at the role of TIL scores in distant tumors as a potential biomarker of abscopal effect resulting from local therapy of primary tumor.

It is well established the systemic immune response plays the predominant role in executing the abscopal effect.²⁹ Previous studies have shown that primary tumors treated by ablation exerts a direct cytotoxic effect on the tumor cells and reprograms the tumor microenvironment.³⁰ Tumor ablation initiates immunogenic cell death (ICD) releasing damage associated molecular pattern (DAMPs) molecules, cytokines, and chemokines, which recruit immune cells into the tumor and induces potent antitumor immune responses with increases in dendritic cell antigen presentation and activation of CD8⁺ T cells.³¹ These activated CD8⁺ CTLs then target the primary treated tumor and also can enter into the periphery as the prime driver of the abscopal effect targeting distant tumors and metastasis.^32‐34 Therefore, methodology to measure abscopal TILs theoretically could potentially serve as an innovative tool for monitoring and predicting cancer ablation treatment efficacy and long-term survival.

Cryoablation of breast cancer provides a viable alternative to traditional resection and thermal ablation where the tumor undergoes a series of freeze/thaw cycles but remains within the patient.³⁵ As the “killed” tumor and surrounding tissue architecture is preserved, there remains the potential for augmentation of an immune response to the preserved TAAs (potentially measurable by TIL scoring),²⁴ leading to an abscopal effect. This aligns with our findings of increased stromal and invasive margin TILs in the unmanipulated distant tumors 1 week after local therapy of primary tumor, with a significant increase for stromal and invasive margin TILs in the cryoablation group. Moreover, we demonstrated that the TIL score of distant tumor correlates with systemic tumor burden and survival, albeit in a small sample. Thus, the distant, unmanipulated tumors represent a target to study abscopal effect resulting from the treatment of the first treated tumor using TIL score as a proxy measure for systemic disease control. Interestingly, we also found a lesser increase in TILs after resection of the first treated tumor, albeit not statistically significant. This modest effect could have resulted from the natural inflammatory response secondary to local trauma of handling and resection. Notably, 40% (2/5) of the mice in the resected group developed both tumor recurrence and metastasis to the lung. The same mice were the only ones in the entire experiment (the resected and cryoablated groups combined) with no increase in stromal TILs compared with the baseline control tumors. This suggests that cryoablation provided a more robust antitumor response both locally and systemically, describing an abscopal effect. More importantly, because the relapse was associated with lack of TIL response, TIL counts can be used as a proxy measure for abscopal effect.

While local resection, which is the mainstay for local control in the present time, removes the primary tumor and provides a minimal local inflammatory response, micrometastasis may still exist. More importantly, by removing the TAAs, resection denies an opportunity for the immune system to mount a robust response needed to overcome the potential for targeting metastasis. Cryoablation is clinically used for the treatment of fibroadenomas and low-risk cancers,^35‐39 mainly because of convincing evidence supporting complete destruction of tumors up to 15 mm in size. However, the potential of augmented immune response by virtue of leaving TAAs in vivo, is not very well-leveraged in the present day practice. This raises the concept of expanding cryoablation to high-risk tumors (TNBC and HER2⁺) that are more immunogenic⁴⁰ but have a poor prognosis to conventional treatments. Cryoablation may provide the immunologic boost needed for improved prognosis in these already immunogenic tumors.

Additionally, cryoablation could be combined with immune modulators to create a more robust immune response. Immune modulators target and inhibit checkpoint molecules like cytotoxic T lymphocyte antigen-4 (CTLA-4) and programmed death-1/ligand-1 (PD-1/PD-L1). A 2016 pilot study found that the combination of ipilimumab, an immune modulator, with cryoablation in women with invasive breast cancer of any HER2⁺ and hormone receptor status led to peripheral elevations in Th1-type cytokines, activated and proliferating CD4⁺ and CD8⁺ T cells, and post-treatment proliferative T-effector cells relative to T-regulatory cells within the tumor. Study with a larger sample size focused on already immunogenic tumors could be performed to determine the potential benefit.⁴¹ As these studies on local therapies alone or in combination move forward in the clinical realm, it will be important to identify relevant and pragmatic biomarkers that could be utilized as outcome measures (indicative of impact on survival) to foster cost-effective clinical trials. This translational work is ideally suited for development of animal models that endorse biomarkers that could be successfully translated into clinical research and practice.

Animal models for translational research provide insight into cancers and therapeutic responses that cannot be performed through clinical research studies. Although animal models have successfully helped to define mechanisms of cancer initiation, development, and treatment, there are clearly differences between rodent models and human cancer development. For example, breast cancer in humans tends to spread lymphatically to lymph nodes, then metastatically to the bones, brain, and adrenal gland, and then the liver and lung, whereas in the mouse, metastasis is almost exclusively through the hematogenous route to the lung and liver.⁴² Therefore, when utilizing animal models for cancer research, these differences should be considered for both analysis and interpretation of the results. For our study, we focused our metastasis analysis on the lung and liver; however, the immune system in both humans and mice is systemic and functions similarly. Therefore, correlations for the abscopal effect in the mouse can be extrapolated to humans. We found that metastasis was only observed in the mice that had tumor recurrence. It is possible that local tumor recurrence is a marker of systemic disease rather than the cause of metastasis. While immune cells were present at the site of metastases, they were primarily comprised of neutrophils, while lymphocytes, although present, were sparse (Fig. 5F). This also suggests that the surgical resection did not induce as strong of an antitumor CTL response in these mice to control tumor recurrence and metastasis. Most importantly, these outcomes were predictable in our model via TIL scoring of the abscopal readout tumors. Intuitively, the difference in TIL score from baseline to postablation in humans could potentially predict the desired abscopal effect.

Most previous breast cancer mouse models use a single tumor model, perform cryoablation or resection, and analyze the results for the immune response.^33,34,43 Sabel et al. demonstrated that cryoablation of MT-901 tumors resulted in a tumor-specific immune response in the tumor draining lymph nodes, which correlated with rejections of tumors upon rechallenge.³³ In our approach, we designed an experiment using two tumors to evaluate TILs in the distant tumor as an immunological readout to measure the abscopal effect. We found that changes in TIL scores in the abscopal-tumor appear to reflect whether mice developed recurrence and emphasize the importance of TILs for predicting long-term outcomes. Although a larger study is needed to confirm these results, this preliminary study reflects the utility of using TILs for measurement of abscopal effect.

Overall, we found that both resection and cryoablation induce an antitumor response with cryoablation being superior. It is likely that tumor resection results in both detrimental and beneficial effects on the systemic immune system, which might explain why 40% of the mice in the resection group failed to control disease. Surgery can trigger healing programs that lead to induce immunosuppressive states primarily driven by myeloid subsets, which can promote protumorigenic niches and suppress T-cell responses. Conversely, the benefit of resection with sufficient recovery time is reduced primary tumor burden allowing for restored systemic adaptive immune responses.⁴⁴ We observed increased TILs in the distant tumors with a significant increase in stromal and invasive margin TILs for the cryoablation group. Furthermore, TIL phenotyping showed that the Abs_Cryo tumor TILs represent an increase in CTL markers by immunofluorescent staining. CD8⁺ T cells only express granzyme B following stimulation and differentiation into cytotoxic T cells.⁴⁵ Remarkably, cryoablation induced a stronger abscopal CTL response with a significantly higher increase in both CD8α and granzyme B staining when comparing the two treatment procedures to baseline tumors. We cannot attribute the granzyme B specifically to only CD8⁺ T cells, because activated CD4⁺ CTLs and NK cells also express granzyme B. However, the data show an increase in immune cytolytic activity in the Abs_Cryo tumors. Interestingly, only mice that did not have an increase in stromal TILs posttreatment in the resection group had recurrence and metastasis, suggesting that induced TIL response is important in predicting overall disease control. The TIL score and CTL analysis in combination with the lack of tumors posttreatment showed that cryoablation resulted in an overall greater abscopal effect in preventing tumor recurrence and metastasis. Because the tumor microenvironment influences the TIL response, it will be important to further characterize the TILs for both antitumor and suppressive phenotypes and quantitate the ratios of activated CTLs (CD8⁺ICOS⁺⁺) versus T_Regs (CD4⁺FOXP3⁺) posttreatment. These ratios could potentially be used as an additional TIL scoring method for predicted treatment efficacy for cryoablation or novel therapies. Future experiments will focus on using RNA-seq analysis to detect “immune gene signature” differences between resection and cryoablation using gene expression levels of immune-related genes to describe the composition and functional status of the immune cell infiltrates, which could be translated for clinical applications.

Questions remain about why some individuals immunologically respond to treatment better than others and how to enhance the response in poor responders. These could be helped by studying the baseline TIL responses mounted by individuals and the ability of local therapies to enhance this response.

Primary local treatment of triple-negative breast cancer by both resection and cryoablation increases systemic immune response, albeit the response is superior with cryoablation. Unmanipulated tumor TIL scoring provides a viable biomarker for studying the abscopal effect of local therapies.