The Impact of PD-L1 Expression in Patients with Metastatic GEP-NETs.

Programmed death-ligand 1 (PD-L1), which is expressed on many cancer cells, interacts with PD1 expressed on the surface of T cells, inhibiting the T cells and blocking the antitumor immune response. Expression of PD-L1 in gastroenteropancreatic neuroendocrine tumors (GEP-NETs) has not been studied. We investigated the impact of PD-L1 expression in 32 patients with metastatic GEP-NET. The expression of PD-L1 was evaluated using an anti-PD-L1 immunohistochemistry (IHC) antibody optimized for staining of formalin-fixed paraffin-embedded (FFPE) tissue samples. The correlation between PD-L1 and clinicopathological data including survival and response to systemic treatments was analyzed. Primary sites were 24 foregut-derived GEP-NETs, including stomach (n=1), duodenum (n=2), biliary tract (n=7), and pancreas (n=14), and 8 hindgut-derived GEP-NETs of the distal colon and rectum. Among the 32 patients with metastatic GEP-NET analyzed in this study, 7 (21.9%) had expression of PD-L1 in tumor tissues. Expression of PD-L1 was significantly associated with high-grade WHO classification (grade 3) (p=0.008) but not with gender, primary site, and number of metastatic sites (p>0.05). The status of PD-L1 expression was statistically associated with progression-free survival (PFS) for first-line systemic treatment (p=0.047). Moreover, the status of PD-L1 expression could significantly predict overall survival (p=0.037). The expression of PD-L1 was associated with higher WHO tumor grade (grade 3) in metastatic GEP-NETs. PD-L1 expression had both predictive and prognostic value for survival of patients with metastatic GEP-NETs.

Introduction

The immune system influences the fate of developing cancers by functioning not only as a tumor promoter that facilitates cellular transformation, promotes tumor growth, and sculpts tumor cell immunogenicity 1-6, but also as an extrinsic tumor suppressor that either destroys developing tumors or restrains their expansion 2, 5, 7. Programmed death-ligand 1 (PD-L1), which is expressed on many cancer and immune cells, plays an important role in blocking the cancer immunity cycle by binding programmed death-1 (PD-1) and B7.1, both of which are negative regulators of T lymphocyte activation 8. Indeed, tumor cells from different locations express PD-L1 and thus can inhibit the immune response. Recently, clinical trials testing anti-PD-L1 or anti-PD1 agents for restoration of antitumor immunity have reported very promising outcomes, notably in melanoma and renal, lung, prostate, and bladder cancer. PD-L1 expression has been studied in different cancers including kidney, lung, pancreas, esophagus, ovary, colorectal, head and neck, breast, and skin (melanoma), with evidence of correlations with clinicopathological tumor features in several studies 9-15. Neuroendocrine tumors (NET) are a heterogeneous group of malignancies derived from neuroendocrine cell compartments with roles in both the endocrine and the nervous system. The majority of NETs are gastroenteropancreatic (GEP) in origin, arising in the foregut, midgut, or hindgut 16. Although NETs are known to be very rare 17, recent studies based on the Surveillance, Epidemiology and End Results (SEER) cancer registry and European studies demonstrated an increasing incidence of NETs 17. Moreover, a Korean study showed a remarkable increase in the incidence of GEP-NETs during the last decade 18. Although GEP-NETs are increasingly being diagnosed, a concomitant improvement in outcomes has not been noted. Recently, two molecularly targeted agents, sunitinib and everolimus, have been approved for advanced pancreatic NETs 19, 20 and might prolong the survival of patients with GEP-NET. However, most GEP-NET patients experience resistance to these treatments and additional effective treatment options are very limited 21. Cancer immunotherapy is presently one of the areas in which major medical breakthroughs are being witnessed with impressive results reported by several groups 8, 22. One key target of cancer immunotherapies is the PD1-PD-L1 pathway; however, this pathway has not been studied in metastatic GEP-NETs. Here, we investigated the impact of PD-L1 expression on survival in 32 patients with metastatic GEP-NET and searched for correlations between PD-L1 expression and clinicopathlogic data.

Material and methods

Patients

Among patients who were pathologically diagnosed with metastatic GEP-NETs in Samsung Medical Center between June 2011 and September 2014, 32 were analyzed for PD-L1 expression in this study. The following clinicopathological characteristics were collected for all 32 patients: age, gender, primary site, tumor grade according to the 2010 WHO classification, liver metastasis, number of metastatic sites, site of metastasis, and information on chemotherapy.

Immunohistochemistry (IHC) of PD-L1

Tumor sections were freshly cut to 4 mm and dried at 60℃ for 30 minute. PDL-1 immunohistochemistry (rabbit anti-human PDL-1 monoclonal, 1:25, clone SP142; Ventana, Tuscon, AZ) was performed on an automated immunostainer (Benchmark; Ventana). Antigen retrieval was performed for 92 min with CC1 and the antibody was incubated for 120 min un 37℃ in Ventana BencaMark XT. Signal visualization was achieved with the Optiview DAB IHC detection kit (Catalogue number 760-700) and Optiview Amplification kit (Catalogue number 860-099).

PD-L1 expression was evaluated on tumor cells. Tumor infiltrating immune cells are not identified in all cases. The proportion of PD-L1-positive cells was estimated as the percentage of total tumor cells; tumor cells typically showed membranous staining with a variably component of cytoplasmic staining. Specimens were categorized as IHC negative or positive if < 1% or ≥ 1% of cells were stained by PD-L1 mAb, respectively (Figure 1).

Figure 1

Representative image of PD-L1 IHC staining of tumors from patients with GEP-NETs

Statistical analyses

Correlation between the status of PD-L1 expression and clinicopathological variables was analyzed using the t-test or the Fisher's exact test as appropriate or one-way analysis of variance. Treatment outcomes were estimated as response rate (RR), disease control rate (DCR), progression-free survival (PFS), and overall survival (OS). PFS and OS were defined as the time from the first study treatment to the date of disease progression or death, respectively. Descriptive statistics were reported as proportions and medians. Kaplan-Meier estimates were used in the analysis of all time to event variables, and the 95% confidence interval (CI) for the median time to event was computed.

Results

Patient characteristics

Among patients who were pathologically diagnosed with metastatic GEP-NET in Samsung Medical Center between June 2011 and September 2014, 32 were analyzed for PD-L1 expression in this study. Baseline characteristics of these 32 patients are listed in Table 1. The median age of the patients was 60.0 years (range, 35-88) and the male to female ratio was 1.29. Primary sites included 24 foregut-derived GEP-NETs [stomach (n=1), duodenum (n=2), biliary tract (n=7), and pancreas (n=14)] and 8 hindgut-derived GEP-NETs of distal colon and rectum. According to WHO classification, 17 patients had grade 3 neuroendocrine carcinoma (NEC) and 15 had grade 2 NET. All 32 patients had liver metastases. Twenty-one patients (65.6%) had 2 or more metastatic sites. The majority (78.1%) of patients received the recommended systemic treatment after the diagnosis of metastatic GEP-NETs. However, 7 patients (21.9%) were managed by only best supportive care because of patient preference or poor performance status (PS).

Table 1

Baseline characteristics in 32 gastroenteropancreatic neuroendocrine tumors (GEP-NETs) patients.

Variables	n	%
Gender
Male	18	56.3
Female	14	43.7
Age, years
Median, range	60, (35-88)
Primary tumor site
Stomach	1	3.1
Duodenum	2	6.2
Gall Bladder	6	18.8
Biliary	1	3.1
Pancreas	14	43.8
Rectum	8	25.0
WHO classification
Grade II neuroendocrine tumor	15	46.9
Grade III neuroendocrine carcinoma	17	53.1
Liver metastasis
Yes	32	100.0
Number of metastatic sites
1	11	34.4
2 ≤	21	65.6
First line treatment
Only supportive care	7	21.9
Systemic treatment	25	78.1

The correlation between PD-L1 expression and clinicopathological variables

Among the 32 patients with metastatic GEP-NET analyzed for PD-L1 expression, 7 (21.9%) exhibited expression of PD-L1 in tumor tissues. The expression of PD-L1 was significantly associated with high-grade WHO classification (grade 3) (p=0.008) but not with gender, primary site, or number of metastatic sites (p>0.05) (Table 2). Among 21 patients receiving the recommended systemic therapies as first line, 4 patients had PD-L1-positive tumors and 17 patients had PD-L1-negative tumors. There was significant difference in the tumor response according to the status of PD-L1 (Table 3).

Table 2

The difference of PDL1 expression by immunohistochemistry (IHC) according to clinical characteristics.

(A) Gender (Male vs Female), p=1.000
PDL1 IHC
Gender	Positive	Negative	Total
Male	4	14	18
Female	3	11	14
Total	7	25	32
(B) Primary site (Foregut vs Midgut vs Hindgut), p=1.000
PDL1 IHC
Primary	Positive	Negative	Total
Foregut	5	19	24
Hindgut	2	6	8
Total	7	25	32
(C) WHO classification (Grade II vs Grade III), p=0.008
PDL1 IHC
WHO	Positive	Negative	Total
Grade II	0	15	15
Grade III	7	10	17
Total	7	25	32
(D) Number of metastatic sites (1 vs 2≤), p=0.719
PDL1 IHC
Metastatic sites	Positive	Negative	Total
1	2	9	11
2≤	5	16	21
Total	7	25	32

Table 3

The difference of response for 1st line treatment according to PD-L1 expression by immunohistochemistry (IHC) p=0.028, p=1.000.

	PD-L positive	PD-L1 negative
1st line chemotherapy	N=4	N=17
Partial response	3	2
Stable disease	0	9
Progressive disease	1	5
Response rate	75.0%	11.8%
Disease control rate	75.0%	64.7%

The impact of PD-L1 on patient survival

In 21 patients receiving systemic therapies as first line, the median progression-free survival (PFS) was 4.9 months (95% CI 4.2-5.5). There was a significant difference in PFS between patients who were positive and negative for PD-L1 (4.2 vs. 5.1 months, p=0.047) (Figure 2). The median overall survival (OS) was 17.3 months (95% CI 11.7-22.8) for all 32 patients. According to the status of PD-L1, median OS was 16.0 months (95% CI 7.7-24.1) in patients with PD-L1-positive tumors and 24.8 months (95% CI 2.6-46.9) in those with PD-L1-negative tumors (p=0.037) (Figure 3).

Figure 2

Kaplan-Meier curve of progression free survival (PFS) for systemic treatment of 1st line by PD-L1 IHC status (p=0.047).

Figure 3

Kaplan-Meier curve of overall survival (OS) by PD-L1 IHC status (p=0.037).

Discussion

Inhibition of PD-L1 is a promising novel target in cancer immunotherapy 23. This study analyzed a heterogeneous collection of metastatic GEP-NETs for expression of PD-L1. We aimed to investigate the role of PD-L1 expression as a predictive and prognostic marker for survival in 32 patients with metastatic GEP-NET. We also searched for correlations between PD-L1 expression and clinicopathlogic data. This analysis revealed that the expression of PD-L1 was significantly associated with high-grade WHO classification (grade 3) (p=0.008) but not with gender, primary site, and the number of metastatic sites (p>0.050). Moreover, PD-L1 might be a useful predictive and prognostic biomarker for survival in metastatic GEP-NETs. To our knowledge, this is the first report evaluating the characteristics and value of PD-L1 expression in patients with metastatic GEP-NET.

PD-L1 expression has been studied in many cancers 9-15. Studies of breast cancer showed that the expression of PD-L1 was associated with poor prognosis features: large tumor size, high grade, negative hormone receptor status, positive HER2 status, and high proliferative rate 24. The characteristics of indolent or aggressive GEP-NET were determined by tumor grade or differentiation 17, 25. In general, tumor grade or differentiation has been based on the World Health Organization (WHO) classification. The 2010 WHO classification divides NETs into two main subgroups, tumor (NET) and carcinoma (NEC), according to Ki67 value and mitotic count. NET is defined as <20% Ki67 and/or <20/high power field (HPF) mitotic count, and is further sub-classified into Grade 1 (G1) (Ki67 ≤2 and/or mitosis ≤5) and Grade 2 (Ki67 3-20% and/or mitosis 5-20/HPF) 26. All NETs with Ki67 >20% and/or mitosis >20/HPF are Grade 3 tumors (NECs), which includes small-cell or large-cell carcinoma. Thus, grade 3 GEP-NETs have more proliferative and aggressive clinicopathologic features. Our finding that expression of PD-L1 was related to a more proliferative and aggressive grade 3 type of GEP-NET was consistent with previous studies 24, 27, 28. This correlation between the expression of PD-L1 and more proliferative tumor type might be caused by the higher aberration rate of hyperproliferative tumor cells, which is responsible for higher immunogenicity due to the rapid appearance of neoantigens.

Anti-PD-L1 therapies are under the spotlight in oncology as next-generation anticancer agents. Herbst et al. showed that anti-PD-L1 therapy (MPDL3280A) was safe and active in several cancer types, including non-small cell lung cancer, melanoma, renal cell carcinoma, and other solid tumors 8. Powles et al. also reported that MPDL3280A was a new therapeutic option for patients with bladder cancer. In these two studies, tumor expression of PD-L1 was associated with a higher response to anti-PD-L1 therapy 22. In our study, 7 of 32 (21.9%) patients had expression of PD-L1 in tumor tissues. In particular, in the subgroup of grade 3 GEP-NETs 7 of 17 (41.2%) patients had tumor expression of PD-L1. There are limited therapeutic options for grade 3 GEP-NET (NEC). The high proportion of PD-L1 expression in grade 3 GEP-NETs suggests that anti-PD-L1 therapy may be a useful new therapeutic modality for this subgroup of GEP-NETs. However, although overexpression of PD-L1 might serve as a predictor of tumors that have successfully evaded an immune response, nivolumab, anti-PD1 antibody, could alleviate inhibitory T-cell signaling independently of tumoral PD-L1 expression 29. In view of similar observations of clinical activity of anti-PD1/PD-L1 therapeutic antibodies in PD-L1 negative patients 30-32, PD-L1 expression might not be necessary for achieving objective response with these agents.

Our study was a retrospective analysis with a small sample size. In addition, there was a lack of standardization of PD-L1 IHC, notably in terms of specificity and reproducibility of available antibodies and the definition of optimal positivity cut-off and interpretative subjectivity. Nevertheless, this analysis identified clinicopathologic characteristics related to expression of PD-L1 in metastatic GEP-NETs and evaluated the predictive and prognostic values of PD-L1 in GEP-NETs for survival. These data may provide useful information and background for future research for immunotherapy for GEP-NET.