記住我
Abstract
Context: The programmed death-1 (PD-1) is an immune checkpoint molecule that suppresses T-cell response. The binding of PD-1 to PD-L1/PD-L2 results cytokine production, and T-cell proliferation are reduced. Tumors expressing PD-L1 and PD-L2 escape from cytotoxic T-cells and are exposed to tumor progression. For this reason, immunotherapy has become a new option in the treatment of cancer. Aims: In this study, we examined the PD-L1 and PD-L2 expression in colorectal carcinoma (CRC), and evaluated the relationship between clinicopathological parameters and CD8+ T cells. Methods and Material: We evaluated CD8 expression in tumor-infiltrating lymphocytes and surrounding tumor lymphocytes with PD-L1, PD-L2 staining in tumor cells and immune cells formalin-fixed paraffin embedded samples of 124 patient diagnosed with CRC. Statistical Analysis Used: Pearson Chi-Square, Fisher Exact Chi-Square, and Pearson Exact Chi-Square analyses were used in the analysis of the cross tables. Survival distributions predicted Kaplan--Meier method and it was evaluated using log-rank statistics. Results: In our study, a significant correlation was found between PD-L1 expression and female sex and tumors with medullary morphology. No expression of PD-L2 was observed in tumors containing medullary morphology, and a statistically inverse relationship was observed between PD-L2 and the medullary component. PD-L1 positive tumor-infiltrating lymphocytes were determined to be an important predictor for recurrence-free survival. Conclusions: We believe that the evaluation of these parameters may be useful in the selection of patients who will benefit from immunotherapy in CRC cases.
Keywords: Colorectal cancer, immunotherapy, PD-L1, PD-L2
How to cite this article: Introduction 
Colorectal carcinoma (CRC) is the third most common cancer in men and women worldwide.[1],[2] In CRC, surgery, chemotherapy, radiotherapy and immunotherapy, or combination therapies are preferred. However, due to the tumor microenvironment that occurs with immunosuppression, these treatments have different outcomes in different patients. This situation requires the development of individualized treatment methods in patients with CRC.[2],[3]
The programmed death receptor-1 (PD-1) is a member of the CD28/CTLA-4 co-stimulatory receptor family and is one of the T-cell regulators. It is expressed on T-cell, B-cell, and natural killer cell surfaces. The programmed death ligand-1 (PD-L1, B7-H1, and CD274) and the programmed death ligand-2 (PD-L2, B7-DC) are the ligands of this receptor, which are expressed on many cells such as tumor cells, T- and B-cells, macrophages and dendritic cells. The binding of PD-1 and PD-L1/L2 results in an inhibitory signal. As a result, cytokine production and T-cell proliferation are reduced.[4],[5] Positive results are obtained with anti-PD-1 and anti-PD-L1 in melanoma, non-small cell lung carcinoma, and renal cell carcinomas. PD-L1 expression in patients with gastrointestinal malignancy and its relationship with prognosis is unclear.[6],[7],[8] The current study aims to examine the PD-L1 and PD-L2 expression in CRC, and evaluate the relationship between clinicopathological parameters and CD8+ T-cells.
Subjects and Methods Patients and tumor characteristics
In this study, 124 patients diagnosed with CRC between 2010 and 2013 in pathology department of Medical Faculty were included. All cases were selected from patients with resection materials. Endoscopic biopsy materials, recurrent tumors, and patients receiving neoadjuvant therapy were not included in the study. Demographic characteristics, tumor localization, and tumor size were recorded. The pathologic preparations of the cases were obtained from the archive and re-evaluated in terms of histological type, differentiation, necrosis, small vessel invasion, large vessel invasion, presence of perineural invasion, depth of invasion, and number of metastatic lymph nodes. Evaluation was made from clinical files to obtain follow-up intervals, and time to recurrence.
Immunohistochemical staining and evaluation
In this study, paraffin blocks were selected that represent the tumor and the inflammatory cells around the tumor are evaluated and most intensively were selected.
Four-micron thick sections were prepared from blocks. PD-L1 (dilution 1:100, rabbit monoclonal, ab205921; Abcam, Cambridge, MA), PD-L2 (dilution 1:300, rabbit monoclonal, clone D7U8C, Cell Signaling Technology, Danvers, MA), and CD8 (SP16 clone, Neo marker) (in ready-to-apply form) without dilution were applied to tissues. Tissue sections using DAKO Omnis automated immunohistochemical staining device (DAKO Omnis, Denmark). Sections were washed with distilled water, two times with 96% alcohol, one time with 99% alcohol, and three times with xylene. The stained preparations were made ready for evaluation by covering with xylene-based slide sealer. Membranous staining with PD-L1 was considered positive. Staining of tumor cells and immune cells were evaluated. For both groups, PD-L1 expression of 5% and above was considered as positive and below 5% was considered negative.[6] Cytoplasmic and/or membranous staining of the tumor cells was considered positive in immunohistochemical analysis with PD-L2. The staining intensity of tumor cells was divided into three groups as mild, moderate, and severe. In moderate and severe intensities, 10% and above were considered as positive, and less than 10% were as negative. In the mild intensity, staining of 50% and above was considered as positive and staining below 50% as negative.[9],[10]
CD8 staining was evaluated in two different categories as tumor-infiltrating lymphocytes (TIL) and surrounding tumor lymphocytes. The presence of 50 and more than 50 CD8 positive lymphocytes in the ×200 amplification area within the tumor was considered as high, and the presence of less than 50 CD8 positive lymphocytes was considered as low. The presence of 200 CD8 positive lymphocytes in the ×200 amplification area within the tumor surrounding lymphocyte was considered as high, and under 200 lymphocytes were considered as low.[11] Two pathologist assessed the immunohistochemical staining independently without knowledge of the clinicopathological data.
Statistical analysis
Continuous data were given as mean ± standard deviation and median value (25--75%). Qualitative data were given as percentage (%). Shapiro--Wilk test was used to investigate the suitability of the data for normal distribution. Mann--Whitney U test was used for comparison of two groups with non-normal distribution. Pearson Chi-Square, Fisher Exact Chi-Square, and Pearson Exact Chi-Square analyses were used in the analysis of the cross tables. Survival distributions predicted Kaplan--Meier method and it was evaluated using log-rank statistics. Time to event was calculated from the date of colon resection to the event date. A Cox regression model, employing the backwards conditional method, was used to adjust for confounders. IBM SPSS Statistics 21.0 program was used in the analysis. For statistical significance, P < 0.05 was accepted as the criteria. Our study was approved by Eskişehir Osmangazi University Faculty of Medicine Non-Pharmaceutical Clinical Studies Ethics Committee (Date:14/11/2016, No:02).
Results Patient's characteristics
The study cohort (n = 124) had a mean age of 64 years (range: 26--90), and 39.5% were female. The majority of the tumors were located in the left colon (n = 100). Mean tumor diameter was 4.7 cm (range 1.2--13). Clinicopathological features of 124 patients are given in [Table 1].
Immunohistochemical findings
PD-L1 was positive in 15 (12.1%) cases and negative in 109 (87.9%) cases of CRC [Figure 1]a, [Figure 1]b. In 75 (60.5%) of the cases, PD-L1 was positive in immune cells and negative in 49 (39.5%). PD-L2 was positive in 52 (42%) and negative in 72 (58%) cases. When PD-L2 staining intensity was evaluated, 44 patients had mild staining, 22 patients had moderate staining, and 20 patients had severe staining [Figure 1]c.
Figure 1: (a-c): (a) PD-L1 staining in a tumor with medullary features (×200) (b) PD-L1 staining in moderately differentiated tumor (×200), (c) Severe PD-L2 staining in tumor cells (×200)When surrounding tumor tissue was evaluated with CD8, 53 (42.7%) cases had lymphocytes high level, and 71 (57.3%) cases had lymphocytes low level. CD8 with TIL was high level in 12 (9.7%) cases and low level in 112 (90.3%) cases.
PD-L1 expression in tumor cells was significantly higher in females (P < 0.05) and tumors with medullary morphology (P < 0.05). When the T stage was evaluated, PD-L1 expression was significantly higher in the immune cells surrounding the tumor in early-stage tumors (T1 and T2) (P < 0.05) [Figure 2]a. PD-L1 positivity rate was also higher in tumor cells in early T stage tumors. However, no statistically significant result was found (P > 0.05). PD-L1 expression was not observed in signet ring cell carcinoma and mucinous carcinoma. In patients with medullary morphology, PD-L1 expression was found higher in immune cells surrounding the tumor, but no statistically significant result was found. There was no statistically significant difference between PD-L1 expression in tumor and surrounding immune cells and age, tumor location, tumor size, histologic grade, T stage, N stage, perineural invasion, small and large vessel invasion and presence of necrosis (P > 0.05) [Table 2]. Since CD8-positive T-cells are thought to be associated with adaptive immune resistance in PD-L1 positive tumors, CD8 positive lymphocyte infiltration in the tumor stroma was evaluated in our study. PD-L1 positivity was found to be significantly higher in the immune cells surrounding the tumor in cases with CD8-positive peritumoral lymphocyte and TIL (P < 0.05) [Table 3]. There was no significant correlation between PD-L1 positive tumor cells and CD8-positive peritumoral and TIL [Figure 2]b.
Figure 2: (a and b): (a) Colorectal cancer with PD-L1 positive peritumoral and tumor-infiltrating lymphocytes (×200), (b) Colorectal cancer with CD8 positive peritumoral and tumor infiltrating lymphocytes (×200)
Table 2: Comparison between PD-L1 expression in tumor cells and clinicopathological parameters
Table 3: Relationship of PD-L1 expression by immune cell and CD8 infiltrationPD-L2 expression was not observed in tumors with medullary morphology. In statistical analysis, there was an inverse and significant relationship between PD-L2 staining and medullary morphology (P < 0.05). Three of the PD-L1 positive staining cases were strongly stained with PD-L2, one was moderate, 6 were weak, and 5 were negative. In 5 of 12 cases, CD8 high TIL staining were PD-L2 positive and in 25 of 53 cases with high level of surrounding tumor lymphocytes were PD-L2 positive. There was no statistically significant relationship between PD-L2 expression and CD8 staining and other clinicopathological parameters [Table 4].
Table 4: The relationship between PD-L2 expression and clinicopathological parametersIn 10 of 12 cases with CD8- high TIL staining, there were also high level lymphocytes around the tumor. There was a statistically significant relationship between these two groups (P = 0.007). Intense TIL staining by CD8 was significantly higher in right colon tumors and tumors with medullary morphology (P < 0.05).
Survival analysis
The follow-up time was determined as 56 ± 3.9 months (range, 49--64 months). The average recurrence free survivor of all patients was 32 ± 3.7 months (range, 24--39 months). Univariate analysis of the patients cohort identified the PD-L1 positive immune cells (P 0.025) as significant predictor for recurrence-free survival [Table 5]. On multivariate analysis, PD-L1 positive immune cells were shown to be independent factors associated with recurrence-free survival (hazard ratio 0.56, P 0.03) [Table 6]. In cases with PD-L1 positive and negative tumor cells, the median (months) values of recurrence-free survival and follow-up times were 7 (0--29) and 30 (5--79), 7 (0--30), and 13 (1--49), respectively (P = 0.471, P = 0.724). The median (months) recurrence-free survival and follow-up values in cases with PD-L1 positive and negative immune cells were 11 (1--26) and 19 (4--70), 16 (0--42), and 42 (17--72), respectively (P = 0.332, P = 0.165). In the statistical evaluation, no significance was found in either of them. In cases with PD-L2 positive and negative tumor cells, the median (months) values of recurrence-free survival and follow-up were 7 (0--24) and 40 (7--74), 6 (0--41), and 40 (7--84), respectively. The median (months) recurrence-free survival and follow-up values in cases with PD-L2 positive and negative immune cells were 8 (0--31) and 31 (8--72), 10 (0--56), and 21 (4--86), respectively. No significant difference was found when PD-L2 positivity in tumor cells (P = 0.543, P = 0.579) and immune cells (P = 0.863, P = 0.961) were compared with recurrence-free survival and follow-up times. There were no statistically significant relationship between peritumoral/intratumoral CD8 staining and recurrence-free survival (P = 0.672, P = 0.952) and follow-up times (P = 0.147, P = 0.551). A statistically significant difference was found between the cases with perineural invasion and the follow-up periods (P = 0.006). No statistically significant difference was found when recurrence-free survival and follow-up times were compared with other clinicopathological parameters.
Table 5: Relationship between clinicopathological characteristics and RFS
Table 6: Relationship between clinicopathological characteristics and RFS Discussion CRC accounts for approximately 9.7% of newly diagnosed cancer cases worldwide. It is the third most common type of cancer in men and women in the world.[1],[2] The distribution of CRCs varies according to geographical regions. It is common in northwestern Europe, North America, and Anglo-Saxon regions, and is less common in parts of Africa, Asia, and South America. Although it may differ in every ethnic group, it is more common in men than in women.[8],[12],[13] In our study, 60.5% of the cases were male and 39.5% were female. The incidence of CRCs increases with age. The mean age at presentation is 62 years. Cases under the age of 40 tend to have one of hereditary CRC syndromes.[14] In our study, the mean age was 64.9 years.
Approximately 50% of all CRCs are located in the rectum and sigmoid. However, the incidence of the right colon has increased in the last decade. In the literature, there are publications indicating that right colon tumors have increased from 22% to 38% in white males in the last 30 years.[15] In our study, 80.6% of the cases were located in the left colon, and 19.4% were in the right colon. Thirteen (54.2%) of the right colon tumors were female, and 11 (45.8%) were male.
Lymphocytic infiltration has prognostic significance in CRCs. Numerous studies have reported a better prognosis in tumors infiltrated with CD8-positive cytotoxic T-cells. This reflects the impact of T-cell-mediated anti-tumor immune response on prognosis.[16],[17] Lymphocytic infiltration is significantly more prominent in microsatellite instability-high (MSI-H) tumors.[18] Tumor cells express some molecules that are negative regulators of the immune system to escape the T-cell-mediated immune response.[19] PD-1 is one of the immune control point molecules that suppress the T-cell-mediated response. PD-1 has two ligands, PD-L1 and PD-L2. Binding of PD-1 with PD-L1 and PD-L2 suppresses the immune system by inhibiting the activation of T-cells. Tumors expressing PD-L1 and PD-L2 escape from cytotoxic T cells and prevent tumor progression.[20]
PD-1 is a member of the CD28 family and is expressed in T-cells. The expression of PD-Ls is variable. PD-L1 is more widely expressed than PD-L2. PD-L1 is expressed on B-cells, dendritic cells, macrophages, T-cells, as well as non-hematopoietic cells such as endothelial cells, pancreatic islet cells, astrocytes, keratinocytes. PD-L2 is expressed on dendritic cells, macrophages, and mast cells.[4],[20] Tumor cells can express both PD-L1 and PD-L2.[21]
PD-L1 expression of melanoma, non-small cell lung carcinoma, renal cell carcinoma, breast cancer, esophageal cancer, gastric cancer, and lymphomas have been reported and associated with the poor clinical course.[22],[23],[24],[25],[26] PD-L1 expression is associated with a poor prognosis in a wide variety of solid tumors and hematologic malignancies, making treatment on the immunomodulatory axis attractive.[27] In cancer immunotherapy, there are antibodies that specifically block PD-1 and PD-L1. However, the role and importance of PD-L2 in regulating the immune response is not yet clear. Therefore, there has not yet been a clear target for cancer immunotherapy.[28]
In CRC, Wang et al.[29] found PD-L1 expression by 8% in tumor cells and PD-L1 expression by 21% in tumor-infiltrating immune cells. They observed PD-L1 expression in both tumor cells and immune cells infiltrating the tumor in 4% of the cases. Staining of PD-L1 in tumor-infiltrating immune cells was associated with disease-free survival. No correlation was found between tumor staining of PD-L1 and clinicopathological data.
Gasser et al.[9] observed membranous and cytoplasmic staining with PD-L1 and PD-L2 in tumor cells in primary CRCs. They found PD-L1 expression in 42.2%, PD-L2 expression in 41.4%, and both PD-L1 and PD-L2 expression in 24.1% of the cases. PD-L1 and PD-L2 expression were significantly higher in tumor cells in advanced-stage disease. Regardless of the stage, the survival of tumors expressing PD-L1 and PD-L2 was significantly worse. They found PD-L1 and PD-L2 expression as independent prognostic factors in survival. Survival of PD-L1-expressing tumors was worse than PD-L2-positive tumors. In addition, survival in both PD-L1 and PD-L2-expressing tumors was significantly worse than in those not expressing both ligands.
In stage 3 CRC cases, Koganemaru et al.[11] observed PD-L1 expression in 8.1% of tumor cells and in 15.3% of mononuclear cells infiltrating the tumor. They found that disease-free survival was significantly shorter in PD-L1 expression in tumor cells. In contrast, disease-free survival was significantly longer in cases with PD-L1 expression in mononuclear cells infiltrating the tumor.
Lee et al.[30] found that PD-1 and PD-L1 expression were significantly higher in patients with MSI-H. In accordance, in this study, they also found that medullar features, high TIL and high peritumoral lymphocytic aggregation were also significantly associated with PD-L1 expression. Rosenbaum et al.[12] found that PD-L1 expression was significantly higher in female, medullary phenotype, intensive TIL, and MSI-H cases. Unlike other tumor types, CRC responds to PD-1 and PD-L1 blockade in a low rate.[6] Le et al.[31] in their study investigating the clinical effects of pembrolizumab in patients with metastatic CRC, reported that patients with MSI-H genotype responded significantly to this treatment. As a result of this study, they emphasized that it is important to determine MMR status in the selection of patients to be treated with pembrolizumab.
Masugi et al.[32] in their studies investigating the relationship between PD-L2 staining and lymphocytic reaction and clinicopathological parameters in CRCs found that PD-L2 expression was significantly higher in female, well and moderately differentiated tumors, and tumors showing high PD-L1 expression. They found PD-L2 expression to be significantly lower in patients with Crohn-like lymphoid reaction. They suggested that PD-L2 expression in tumor cells inhibited the development of lymphoid response.
In our study, PD-L1 expression was observed in 12.1% of tumor cells and 60.5% of immune cells surrounding the tumor. PD-L2 expression in tumor cells was detected in 42% of the cases. Similar to the literature, PD-L1 expression in tumor cells was significantly higher in females and in patients with medullary morphology (P < 0.05). PD-L2 expression was not observed in tumors with medullary morphology, and a statistically inverse relationship was found between PD-L2 and medullary morphology (P < 0.05). Three of the PD-L1 positive staining cases were strongly stained with PD-L2. In 5 of 12 cases with CD8- high TIL staining were PD-L2 positive and in 25 of 53 cases with high level of surrounding tumor lymphocytes were PD-L2 positive. In this study, a statistically significant relationship was found between PD-L1 expression in tumor-surrounding immune cells and in early T stage (T1 and T2) cases (P < 0.05). In addition, in cases with dense CD8 and TIL and lymphocytes around the tumor, PD-L1 expression was found significantly higher in the tumor-surrounding immune cells (P < 0.05). Unlike the literature, PD-L1 expression in tumor cells was observed more in early T stage tumors but no statistically significant difference was found. Similar to the literature, we have shown that PD-L1 positive tumor infiltrating lymphocytes can actually affect recurrence- free survival.
In this study, PD-L1 was expressed on both tumour and immune cells in CRC. High PD-L1 expression in patients with high CD8+ T cells supports the association with an adaptive immune resistance mechanism. Female gender, medullary morphology, intense lymphocytic infiltration in and around the tumor are common features in MSI-H tumors.[21] In our study, PD-L1 expression was found to be significantly higher in females, medullary morphology and TIL and tumor-rich lymphocytic infiltration. These findings support the relationship between PD-L1 expression and microsatellite instability in CRCs reported in the literature. Significant expression of PD-L1 in tumors with medullary morphology and in female gender suggests that it may be useful in patient selection.
Financial support and sponsorship
Our study was supported by the scientific research Project unit of Eskişehir Osmangazi University.
Conflicts of interest
There are no conflicts of interest.
References 
Correspondence Address:
Canaz Funda
Osmangazi University, Medical Faculty, Meşelik, Eskişehir
Turkey
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/ijpm.ijpm_814_21
留言 (0)