Abstract
Aim: To study whether mismatch repair (MMR) status is related to the expression of programmed cell death-ligand 1 (PD-L1) and CD8 counts in a series of grade 3 endometrial carcinomas. Materials and Methods: The expression of MMR protein PD-L1 and CD8+ cell count were evaluated by immunohistochemistry and related to several clinicopathological parameters. Results: Among 105 endometrial carcinomas, 40% were of endometrioid and 60% of non-endometrioid histology. MMR deficiency was observed in 28.6% of cases and was related to endometrioid histology (p<0.001), positive PD-L1 expression (p=0.047) and high CD8+ cell count (p=0.022). When examined by histotype, endometrioid MMR-deficient tumors were related only to PD-L1 expression (p=0.032) but not to high CD8+ cell count (p=0.231), whereas non-endometrioid MMR-deficient carcinomas were not related to either of these markers. MMR deficiency was associated with PD-L1+/CD8high status (p=0.006), whilst MMR proficiency was associated with PD-L1−/CD8low status. In MMR-proficient tumors, high CD8+ cell infiltration alone and combined with PD-L1− status was associated with better progression-free survival (p=0.013 and p=0.04, respectively). Conclusion: MMR-deficient high-grade endometrioid tumors might be more likely to benefit from immunotherapy compared to other grade 3 endometrial carcinomas.
Endometrial cancer is the sixth most frequently encountered malignancy worldwide and growing incidence rates have been noted over the years. Overall, endometrial carcinomas are characterized by a favorable prognosis (1, 2). However, high-grade disease is an aggressive malignancy of significant metastatic and recurrent potential (3). The management of advanced endometrial carcinoma remains highly controversial concerning the extent of the optimal surgical approach and the appropriate adjuvant treatment (4).
Individualized targeted therapy is swiftly developing in gynecologic oncology (5). The whole genomic sequencing of endometrial cancer by the Cancer Genome Atlas has served remarkably in this direction. In 2013, the Cancer Genome Atlas research project classified endometrial cancer into four new categories based on molecular features: Polymerase-E ultramutated, microsatellite instability (MSI) hypermutated, copy-number low and copy-number high (6). The two former groups, polymerase-E ultramutated and MSI hypermutated, are characterized by a higher mutational burden and are associated with an increased neoantigen load. There have been studies, recently, supporting the notion that a high mutational burden in tumors represents a favorable feature for the implementation of immunotherapy (7-10).
The first results of immunotherapy as a treatment option in endometrial cancer seemed promising in a few clinical trials, such as the KEYNOTE-028, and in some sporadic cases (11, 12). Response to immune checkpoint inhibitors has been linked to DNA mismatch repair (MMR) deficiency, which leads to microsatellite instability (10, 13). MMR deficiency has been reported in approximately 27.1-47.9% of endometrial carcinomas, with most studies suggesting a positive effect on prognosis (6, 14-17). On the other hand, Backes et al. (16) and Nelson et al. (18, 19) noted that MMR deficiency was associated with a worse prognosis in high-risk endometrioid carcinomas, whereas other reports found no significant relationship between MMR status and survival (20, 21).
Apart from MMR deficiency, the efficacy of immune checkpoint inhibition has been linked with increased CD8+ T-cell infiltration and the expression of programmed cell death ligand 1 (PD-L1) (22-24). Recent studies have shown that MMR-deficient endometrial carcinomas are more likely to express PD-L1 and to have higher numbers of tumor-infiltrating lymphocytes than MMR-proficient carcinomas (25). CD8+-infiltrating lymphocytes can halt tumor progression by attacking cancer cells after recognizing neoantigens presented on them (26). However, cancer cells develop various immune escape mechanisms. Endometrial cancer cells express PD-L1, in up to ~80% of cases (27), and stimulate the programmed cell death 1/PD-L1 axis. This axis is an established target in the therapy of several types of cancer, and it is being considered as a potential promising alternative in the treatment of selected endometrial carcinoma cases.
Our goal, in the present study, was to evaluate by immunohistochemistry the expression of MMR proteins and relate it to the expression of PD-L1 and CD8+ tumor-infiltrating lymphocytes in a series of high-grade endometrial carcinomas of all histological subtypes in order to evaluate whether MMR-deficient tumors have an immunologically activated tumor microenvironment, thereby being more likely to benefit from immunotherapy.
Materials and Methods
Women treated for high-grade endometrial cancer between late 2001 and early 2018 at our center were evaluated retrospectively for inclusion in the study. Exclusion criteria involved administration of pre-operative chemotherapy, incomplete surgical staining and inadequate immunohistochemical staining for MMR proteins, PD-L1 or CD8. Complete surgical staging was defined as total hysterectomy, bilateral salpingo-oophorectomy, pelvic with/without para-aortic lymphadenectomy or even more radical surgery when clinically justified. Of the 182 patients with grade 3 endometrial cancer who were treated at our center during this period, the eligibility criteria were met by 105 women. More specifically, there were 42 endometrioid, 38 serous, four clear-cell, six unclassified, and seven endometrial carcinomas, as well as eight malignant mixed Mullerian tumors. Post-operatively, most of these patients (88.6%) received adjuvant treatment (chemotherapy, radiotherapy or a combination of chemotherapy and radiotherapy), whereas the remaining 11.4% received no adjuvant treatment at all.
Hematoxylin-eosin and all immunohistochemically stained tumor slides were reviewed by two pathologists specialized in gynecological pathology (EG, EK) who were unaware of the histological diagnosis. In cases of disagreement, the slides were re-evaluated by a third pathologist (KP). The most representative formalin-fixed paraffin-embedded tissue block was selected for the immunohistochemical evaluation.
The study was approved by the Institutional Review Board (1718025202/24-4-18).
Immunohistochemistry. From each representative archival paraffin-embedded tissue block, six 4 μm tissue sections were cut and assayed for mutL homolog 1 (MLH1), MLH2, mutS homolog 6 (MSH6), PMS1 homolog 2, mismatch repair system component (PMS2), PD-L1 and CD8 by immunohistochemistry.
The following antibodies were used as previously described (28, 29): FLEX Ready-To-Use rabbit and mouse monoclonal antibodies against the mismatch repair proteins (clones ES05, FE11, EP49 and EP51 for MLH1, MLH2, MSH6 and PMS2, respectively; DAKO, Carpinteria, CA, USA), PD-L1 assay 22C3 (dilution 1:40; DAKO) and mouse monoclonal anti-CD8 antibody clone C8/1448 (dilution 1:70; DAKO). For each antibody the appropriate positive and negative controls were used.
Tumors were considered MMR-deficient when they showed complete absence of immunoreaction for at least one of the four MMR proteins, otherwise they were considered MMR-proficient (Figure 1A and B). PD-L1 expression was deemed as positive for carcinomas with membranous staining of any intensity in ≥1% of the tumor cells, whilst expression of <1% was considered negative (Figure 1C and D).
The count of CD8+ intraepithelial cells was manually assessed in five random high-power-fields ×400 for each case. The median CD8+ cell count over the whole population was 19.0/high-power-field. Tumors with CD8+ counts of ≥19.0/high-power-field were regarded to have high infiltration (Figure 1E and F).
Statistical analysis. Chi-square and Fisher's exact test were employed to evaluate relationships between categorical variables. Parametric and non-parametric tests were used for continuous variables, as needed. Univariate survival analysis was performed with Kaplan–Meier method, and log-rank test. Multivariate analysis was not performed due to the limited number of events. p-Values of less than 0.05 were regarded as statistically significant. All analyses were conducted in SPSS version 23.0 (IBM, Armonk, NY, USA).
Results
Patient characteristics. A total of 105 patients with high-grade endometrial carcinoma were enrolled in the study. The mean age at diagnosis was 65 years (range=31-85 years), while the mean follow-up period was 64.91 months (range=18-190) months. Out of these carcinomas, 42 (40%) were of endometrioid and 63 (60%) of non-endometrioid histology.
MMR deficiency was observed in 30 cases (28.6%). Of the clinicopathological parameters under evaluation, it was statistically related only to endometrioid type of carcinoma and to deep myometrial invasion (Table I).
Analysis of PD-L1 and CD8+ cell infiltration in relation to MMR status. As shown in Table II, there was a statistically significant higher rate of tumors with PD-L1 expression in MMR-deficient (18/30, 60%) than in MMR-proficient cases (29/75, 38.7%) (p=0.047). Moreover, the majority of MMR-deficient cases revealed high CD8+ all infiltration (21/30, 70%) in contrast to MMR-proficient tumors (34/75, 45.3%) and this correlation was statistically significant (p=0.022). When analyzing the group of endometrioid tumors separately, MMR-deficiency was statistically related only to PD-L1 expression (p=0.032) and not to high CD8+ cell counts (p=0.231). In the group of non-endometrioid cancer, no statistically significant relation was found between MMR status and PD-L1 or CD8 (p>0.99 and p=0.450, respectively) (Table III).
Representative images of immunohistochemical staining of grade 3 endometrial carcinomas. A: DNA mismatch repair (MMR) deficiency (original magnification, ×200). B: MMR proficiency with positive nuclear staining (original magnification, ×100). C: Programmed cell death-ligand 1 (PD-L1)-positive membranous staining (original magnification, ×200). D: PD-L1 negativity with positive staining of the lymphocytic infiltrate (original magnification, ×40). E: High intraepithelial CD8+ cell infiltration (original magnification, ×40). F: Low intraepithelial CD8+ cell infiltration (original magnification, ×40).
Association between combined PD L1/CD8+ cell infiltration and MMR status. In order to further explore the relationships between MMR status and PD-L1 and CD8+ infiltrating cells, these were evaluated in relation to the combined PD-L1 (negative or positive) and CD8+ cell infiltration (low or high) status (Table IV). Four distinct groups were formed, according to tumor PD-L1 and CD8+ infiltration: PD-L1+/CD8+ high, PD-L1−/CD8+ high, PD-L1+/CD8+ low, and PD-L1−/CD8+ low. Evaluation of each group with MMR status revealed that PD-L1+/CD8+ high status, was statistically associated with MMR deficiency (p=0.006), whereas PD-L1−/CD8+ low was associated with MMR-proficient tumor (p=0.003). There was no statistical association with MMR status for the other groups.
Survival analysis. Survival analysis was performed in the MMR-deficient and MMR-proficient subgroups. Of note, only one death and three recurrences occurred in patients with MMR-deficient tumors during the follow-up period. Therefore, due to the limited number of cases, no statistics were computed on overall survival for these patients. Likewise, statistics on progression-free survival in MMR-deficient patients was only possible concerning CD8 counts, combined PD-L1+/CD8 high and combined PD-L1+/CD8 low expression. However, none of the aforementioned markers was shown to have a statistically significant relationship with progression-free survival (p=605, p=0.697 and p=0.154, respectively).
On the other hand, in MMR-proficient tumors, high CD8+ count and combined PD-L1−/CD8+ high status were significantly associated with a better progression-free survival (p=0.013 and p=0.04, respectively). PD-L1 positivity and all other combinations of PD-L1/CD8+ had no significant association with progression-free survival (PD-L1+, p=0.345; PD-L1+/CD8+ high, p=0.426; PD-L1−/CD8+ low, p=0.077; PD-L1+/CD8+ low, p=0.281). In MMR-proficient tumors, there was no statistically significant relationship between any of the markers under study and overall survival.
When studying by histotype, in those with endometrioid MMR-deficient tumors, all statistically measurable parameters (CD8+ count, combined PD-L1+/CD8+ high and PD-L1+/CD8+ low) were found not to be significantly related to progression-free survival (p=0.546, p=0.950, and p=0.241, respectively). On the other hand, patients with endometrioid MMR-proficient PD-L1+/CD8+ low tumors had significantly worse progression-free survival compared to those belonging to the other groups (p<0.001).
Association of DNA mismatch repair (MMR) status with the clinicopathological features under evaluation.
Association of DNA mismatch repair (MMR) status with programmed cell death-ligand 1 (PD-L1) and intraepithelial CD8+ cell infiltration.
Statistical analysis revealed no significant results for any of the above markers concerning their relation to progression-free or overall survival in those with MMR-deficient nor in those with MMR-proficient non-endometrioid tumors (data not shown).
Association of DNA mismatch repair (MMR) status with programmed cell death-ligand 1 (PD-L1) and intraepithelial CD8+ cell infiltration according to histological subgroup.
Evaluation of the association of DNA mismatch repair (MMR) status with combined programmed cell death-ligand 1 (PD-L1) and intraepithelial CD8+ cell infiltration.
Discussion
Immunotherapy is gaining significant ground in the treatment of gynecological malignancies. Even though more and more research is going to shed light on this promising field, there is still a lot to be learned. The Food and Drug Administration approval of PD1/PD-L1 axis inhibitors for the treatment of solid MSI tumors regardless of their origin might potentially serve as a valuable tool in the clinician's armamentarium for treatment of endometrial cancer (30). Addressing impaired MMR from an immunohistochemical point of view has been considered as a more clinically applicable and reliable alternative for the initial testing of MSI (22, 31, 32). However, there are limited reports focusing on the possible role of MMR status as a biomarker for immunotherapy in metastatic grade 3 endometrial carcinoma.
In the present study, the MMR status of a series of high-grade endometrioid and non-endometrioid endometrial carcinomas was evaluated, in relation to the expression of PD-L1 and presence of CD8+ infiltrating cells. All the above biomarkers are considered to be predictive for response to immune checkpoint inhibition (32). It should be noted that regarding CD8 expression, only intraepithelial CD8+ counts were evaluated, despite the fact that previous studies emphasized that both intraepithelial and stromal CD8+ lymphocytes should be counted, since their density in each compartment might be of different prognostic significance (33). Nevertheless, in a previous study by our group, comprising a more limited number of cases, the supremacy of intraepithelial over stromal CD8 expression in predicting a longer progression-free survival of endometrial carcinoma was outlined (28).
Our results indicate that MMR-deficiency was statistically associated with increased rates of PD-L1 expression and high CD8+ counts, which is in accordance with current literature (25, 34-37). Apparently, an increasing number of mutations and neoantigens that are related to deficiency in MMR mechanisms, steps up the innate immune response against the tumor and creates a favorable environment for the implementation of immunotherapy (7). Apart from high CD8+ counts, increased rates of PD-L1 expression have also been accredited to the higher mutational burden of MMR-deficient tumors (38). However, in our population, the above associations were observed only when considering the total cohort of carcinomas under evaluation. When the analysis was performed separately for each histological subgroup, endometrioid MMR-deficient tumors were significantly associated only with PD-L1 positivity, whereas there was no relation with CD8+ count. On the other hand, MMR-deficient tumors of non-endometrioid histology were not related to PD-L1 positivity nor to a high CD8+ count. Such differences at the molecular level might indicate that not all sub-types of high-grade endometrial carcinoma featuring MSI would benefit equally from the implementation of immunotherapy. However, in the recently published recommendations by the European Society for Medical Oncology on MSI testing for immunotherapy in cancer, a significant percentage (19.9%) of MSI-high endometrial carcinomas were found to have a low tumor mutational burden, a finding which is not common in other tumor types such as colorectal and esophagogastric adenocarcinomas (32). The results of our study are compatible with the above statement since 16.7% of our MMR-deficient tumors had negative PD-L1 expression and low CD8+ counts. The question that emerges is whether high PD-L1 and CD8 expression is a prerequisite for the enhancement of the response to immunotherapy in MMR-deficient tumors in endometrial cancer. Unfortunately, the complexity in the relationships between the different markers questions the use of their combinations for building treatment protocols. Furthermore, some tumors that lack MSI or PD-L1 expression are known to have a high response rate to immune checkpoint inhibitors (32). The majority of clinical trials on immunotherapy in MMR-deficient endometrial carcinomas had positive PD-L1 expression as a prerequisite (11, 12, 39, 40). Nevertheless, in a relatively recent study, a significant and lasting response to immune checkpoint inhibitors was presented in a patient with MMR-proficient and PD-L1− endometrial cancer (41). It is our belief, in accordance with other investigators, that other biomarkers, probably related to DNA-repair pathways or mutational signatures that contribute to a high tumor mutational burden independently of MSI status should be investigated in endometrial cancer as novel predictors for response to immunotherapy (32, 41).
In a recent study, Nagle et al. reported no significant effect of MMR status on the survival of their endometrial carcinoma population, whereas MMR deficiency was associated with worse prognosis when the analysis was focused solely in those with endometrioid tumors (21). In our cohort, when MMR status was evaluated in relation to PD-L1 and CD8+ count, no relation was found in progression-free survival for the total number of cases nor for endometrioid or non-endometrioid MMR-deficient carcinomas when studied separately. Only patients with endometrioid MMR-proficient tumors were found to have a worse prognosis when related to PD-L1+/CD8+ low expression. Of course, our survival analysis did not consider the stage of the disease nor the implementation of chemo- or radiotherapy and therefore could be considered only as indicative.
In summary, in high-grade endometrial carcinoma, MMR deficiency was significantly associated with PD-L1+ and high CD8+ count, features implying the existence of an immunologically activated tumor microenvironment. Endometrioid grade 3 MMR-deficient tumors were more likely to be PD-L1+ compared to MMR-proficient tumors, whereas this was not true for non-endometrioid carcinomas. Our data suggest that MMR-deficient endometrioid grade 3 carcinomas might be more likely to benefit from immunotherapy compared to other high-grade endometrial carcinomas. Whether MMR-deficient tumors featuring a PD-L1−/CD8 low phenotype would benefit equally from the implementation of immunotherapy is a question that remains to be answered.
Acknowledgements
This research work was supported by the Onassis Foundation – Scholarship ID: G ZO 001-1/2018-2019.
Footnotes
↵* These Authors contributed equally to this study.
Authors' Contributions
SV and TD: Study design, data collection and analysis, and drafting/final editing of the article. EG, EK, and HG: Evaluation of the hematoxylin-eosin and immunohistochemistry slides. CP: Critical revision of the article. KP: Supervision, critical analysis, drafting/final editing. All Authors approved the final version of the article.
Conflicts of Interest
The Authors declare that there are no conflicts of interest.
- Received January 21, 2020.
- Revision received February 3, 2020.
- Accepted February 4, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
