Immunophenotypical characterization of colorectal cancer liver metastasis of pre-treated and chemo-naive patients

Background There is extensive evidence for molecular heterogeneity in colorectal cancer (CRC) and CRC metastasis (mCRC). At the present, there are only few information regarding the immunological characterization of mCRC and about the effect of chemotherapy and targeted drugs on the interplay between tumour and the immune system of the host. It has been proposed an integrated classification molecular and of microenvironment of mCRC, considering that no data regarding the evolution over time of these features are available so far. New immunotherapeutic strategies are currently under development and will be further studied starting from refractory settings of heavily pre-treated mCRC patients. On this basis, a specific immunological characterization of mCRC will be relevant to direct future clinical and pharmacological research. We planned an exploratory, prospective, observational study for the immune phenotypical characterization of mCRC leukocyte infiltrate from pre-treated or chemo-naive patients undergone surgical metastasis resection. Material and Methods Patients will be divided in two groups according to previous administration of systemic chemotherapy (pre-treated or chemotherapy vs chemo-naive or not pre-treated group). Collected metastasis need to be at least 8 cm3 in volume to allow standard pathological and flow cytometry analysis. Flow cytometry samples were isolated in operating room selecting viable-tumor portion of metastasis, and then carried to the laboratory to be minced in small pieces and enzymatically digested. Tumor-infiltrating leukocytes have been characterized with three panels of different antibodies in terms of subset composition (CD45, CD3, CD20, CD56, CD68), maturation and activation grade (CD45, CD3, TCRγδ, CD4, CD25, CD127, CD8, CCR7, CD45-RO, PD-1, Tim-3), cytokine production, degranulation and cycling activity (IL-2, IL4, INFγ, TNFα, CD107a, Ki-67). Results Eleven liver metastases have been analyzed. Seven cases are in chemotherapy group and three patients were naïve. Oxaliplatin-based chemotherapy was administered in 62.5% of patients, Irinotecan-based in 25% and immunotherapy (Ipilimumab/Nivolumab) in 12.5% (one case). One case received anti-VEGF (Bevacizumab) and four cases received anti-EGFR (Cetuximab) targeted therapy. Primary tumor location was right colon in 18.2% (2 cases), left colon in 63.6% (7 cases) and rectum in 18.2% (2 cases). One B-RAF and one N-RAS mutations were mapped, the other 80% of patients had no mutations in analyzed genes. There were two patients with micro-satellite instability (MSI-H), whereas the remainder 8 were stable (MSS). Five patients (62.5%) underwent on surgery within 6 months from the end of systemic chemotherapy administration and three (37.5%) after 6 months. All patients are alive after a mean follow-up period of 12.19 +/- 3.3 months after operation. During follow-up, 60% of patients present a tumor recurrence with a mean progression-free survival (PFS) of 9.15 +/- 2.17 months. In particular, one case had a hepatic recurrence only, two cases pulmonary and lymph-nodal, one case peritoneal only and two cases peritoneal and pulmonary. CT-patients have a higher percentage of infiltrating leukocytes among viable cells compared to naïve patients (p=0,048). CT-patients show a polarized response towards CD4+ T-cells (p=0.0001), whereas naïve patients towards a CD8+ T-cells response (p=0.05). CT-patients display a very consistent maturation signature independently from CT-regimen. Together TEM and TEMRA represent more than 95% of all T-cells in both CD4+ and CD8+, but with a different ratio. TEM were more than 80% of all CD4+T-cells; within CD8+, 60% are TEM and more than 35% are terminally differentiated effector. IL-7 receptor (IL-7R) is higher both in CD4+ (p=0.003) and CD8+ T-cells (p=0.045) in CT-patients compared to naïve, as IL-7R expression intensity (p=0.043 in CD4+ and p=0.035 in CD8+). Chemotherapy regimen analysis is limited because of the small sample, but the highest percentages of infiltrating leukocytes among viable cells have been found in the patient treated with immunotherapy and in the Oxaliplatin-Irinotecan-treated one; Oxaliplatin-treated patients display variable CD45+ percentage ranging between 10 and 60%, and Irinotecan-treated patient reveals the lowest leukocyte percentages among total viable cells. To analyze time-dependent differences of immune response in liver metastases, we divided patients in two categories considering a cutoff of 6 months from the end of chemotherapy. Patients operated later than 6 months have higher percentage of infiltrating leukocytes among viable cells (p=0.0008). and in particular of B-cells (p=0.024). On the other hand, IL-7Rexpression among CD8+ was higher near the chemotherapy (p=0.043), even though the CD4+ degranulation activity in these patients was lower (p=0.033). Discussion and conclusion The leading aim of our exploratory study is to describe and characterize immune infiltrate in mCRC of naive and pre-treated patients. Results can drive further studies and possibly identify the most favorable subset of patients for receiving novel immune modulating therapies. In all samples, cytotoxic potential of infiltrating lymphocytes is a quite variable among CD4+, but high and uniform in CD8+, consistent with a T-cell effector pattern. Almost all infiltrating Tlymphocytes are CCR7-negative effector memory cells and express high levels of IL7-R that allows the maintenance of this memory pool in the metastases. Chemotherapy reduces tumor cells leaving a higher amount of immune infiltrating cells in liver metastases (p=0.048). Moreover, our results demonstrate that chemotherapy treatment polarizes immune response towards CD4+ T-cells (p=0.0001). This effect could be partially explained by tumor microenvironment modification induced by chemotherapy, in particular because the treatment induces tumor cell death, promoting antigen presentation and CD4+ T-cells expansion. In pre-treated patients T-cells display higher level of IL7-R+ and at higher expression intensity compared to naïve patients, both in CD4+ (p=0.003 and p=0.043, respectively) and in CD8+ T-cells (p=0.045 and p=0.035, respectively). Data display that chemotherapy increases the support to memory pool maintenance inside the mCRC. Analysis of immune infiltrate in liver metastases according to chemotherapy regimen is limited by the small sample size of patients that have been enrolled for each regimen. Nevertheless, we find some differences among regimens that can be use as working hypothesis in further works. Analysis of immune response variation according to time from the end of chemotherapy before surgery, shows that infiltrating leukocytes in general (p=0.0008) and B-lymphocytes in particular (p=0.02) are higher after longer period from CT suspension. These patients display also a higher degranulation activity (CD107a) especially in CD4+ (p=0.03). On the other hand, there is a timedependent affection of CD8+ memory pool maintenance through a lower IL7-R expression in CD8+ of late operated patients (p=0.043). This finding could open an interesting research field about the optimal moment for surgical resection. The main limitation of this study is the reduced sample size. We analyzed only three naïve patients (and in some tests only two), but there were objective difficulties in enrolling such patients. Further developments of the present study will necessary include an extension of sample size, aimed to enroll more naïve, Irinotecan-treated and Oxaliplatin plus Irinotecan-treated patients.