Correlation of PD-1/PD-L1 polymorphisms and expressions with clinicopathologic features and prognosis of ovarian cancer

Abstract

OBJECTIVE:

To explore the correlation of PD-1/PD-L1 polymorphisms and their expressions with clinicopathologic features and prognosis of ovarian cancer.

METHODS:

A total of 164 patients with ovarian cancer were enrolled as case group and 170 healthy women as control group. We conducted quantitative reverse transcription-PCR (qRT-PCR) to determine PD-1/PD-L1 expressions in peripheral blood mononuclear cells (PBMCs). Polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) and allele-specific amplification were used to detect PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G.

RESULTS:

PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G polymorphisms increased the risk for ovarian cancer. PD-1 rs2227982 C>T was associated with FIGO stage and differentiation grade, while PD-L1 rs4143815 C>G was correlated with histological type and differentiation grade. Besides, PD-1/PD-L1 expressions were positively correlated in PBMCs of patients with ovarian cancer to be associated with differentiation grade. Compared with wild homozygous patients, PD-1/PD-L1 expressions were significantly decreased in PBMCs of ovarian cancer patients carried with the mutant genotypes of rs2227982 C>T and rs4143815 C>G. The PFS and OS in ovarian cancer patients with wild homozygous genotype of rs2227982 C>T and rs4143815 C>G were significantly higher than those with mutant genotypes, which were significantly lower in patients with low expressions of PD-1/PD-L1 than those with high expressions. Univariate COX regression analysis identified FIGO staging, differentiation grade, rs2227982 C>T, rs4143815 C>G and expressions of PD-1/PD-L1 as the prognostic factors, and multivariate COX regression analysis demonstrated that high FIGO stage and low expressions of PD-1/PD-L1 were independent risk factors for the prognosis of ovarian cancer.

CONCLUSION:

PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G polymorphisms increased the risk of ovarian cancer, leading to a poor prognosis, associated with low expressions of PD-1 and PD-L1. While high PD-1 and PD-L1 expressions are indicators of a favorable prognosis in ovarian cancer.

Keywords

Ovarian cancer PD-1 PD-L1 polymorphism clinicopathologic features prognosis

1. Introduction

Ovarian cancer, as one of the three most common malignant tumors in gynecology, ranks top two in morbidity and top one in mortality respectively from gynecological malignancies [1]. Statistically, about 238,700 individuals were diagnosed with ovarian cancer, and 151,900 patients were died of ovarian cancer in 2012 [2]. Unfortunately, owing to the unspecific symptoms and a lack of early adequate screening method, ovarian cancer is usually diagnosed at advanced clinical stage [3]. Currently, a variety of factors are known to involve in ovarian cancer, with the continuous research and the development of technique, to influence its clinical prognosis, such as age at diagnosis, fertility, family history, mental, environmental, and dietary factors [3, 4], as well as the genetic susceptibility [5], which may possibly result in gene expression difference in individuals, and thereby playing an important in tumor survival [6], including in ovarian cancer [7].Programmed cell death-1, also named as PD-1 or CD279, a newly found gene related to immune escape, is located on chromosome 2q37.3 and encodes a 50 $\sim$ 55 kDa type I transmembrane glycoprotein belonging to CD28 immunoglobulin superfamily [8]. PD-1 contains two ligands, PD-L1 and PD-L2, that are cell surface glycoproteins as the members of the B7 family called B7-H1 and B7-DC, and PD-L1 is the major ligand located in chromosome 9p24 [9, 10, 11]. Moreover, PD-1/PD-L1 has a vital impact not only on the peripheral immune tolerance but also on the tumor development, however, its mechanism is quite differ from various types of cancers, for example, high expressions of PD-1/PD-L1 can lead to plenty of human cancers with poor clinical prognosis, such as in breast cancer [12] and lung cancer [13], but with the contradictory results in pancreatic cancer [14] and rental clear cell carcinoma [15]. Interestingly, there was a study stating that high PD-1/PD-L1 expressions were indicators of good prognosis for ovarian cancer [16], on the contrary, ovarian cancer patients with higher PD-L1 expression was reported by Hamanishi et al. to be an independent risk factor for overall survival (OS) and progression free survival (PFS), showing a poorer outcome [17]. In terms of genetic polymorphism, although there are some SNPs within PD-1, rs2227982, a missense mutation in exon 5, which could cause the amino acid substitution from valine to alanine and possibly changes the structure and function of PD-1, was interestingly found no statistical association with overall cancer risk by meta-analysis [18, 19]. However, previous studies demonstrated that the mutation of the polymorphism increased the risk of female patients with esophageal squamous cell carcinoma (ESCC) [20], decreased the risk of breast cancer [21], and had no influence on the susceptibility of gastric cancer [22], indicating there is controversy regarding the association of the rs2227982 polymorphism in the PD-1 gene with cancers. In addition, several lines of evidence suggest a relationship between a somatic mutation at a naturally occurring polymorphism locus rs4143815 in the 3’-UTR of PD-L1 gene [23], which has been reported to alter binding efficiency of miR-570 to the elevated PD-L1 transcripts and protein expression [24], and susceptibility to cancers, like gastric cancer [23], non-small-cell lung cancer (NSCLC) [25]. However, little research had explored the association of PD-1 rs2227982 and PD-L1 rs4143815 with the risk of ovarian cancer. Meanwhile, whether the polymorphisms affect the transcript levels of PD-1 and PD-L1 to have an impact on the occurrence, development and prognosis of ovarian cancer is unknown.

Table 1
Comparison of baseline characteristics between case group and control group

Characteristics	Cases group ( $n=$ 164)	Control group ( $n=$ 170)	$P$
Age(years)
$\leqslant$ 50	64	71
$>$ 50	100	99	0.656
Number of pregnancy
1	66	69
2–3	50	52
$\geqslant$ 4	48	49	0.996
Body mass index
$\leqslant$ 19	27	19
0–25	70	83
26–29	31	38
$\geqslant$ 30	36	30	0.654
Smoking status ${}^{\ast}$
Never	38	41
Ever	64	53
Moderate	24	29
Heavy	38	47	0.482
Alcohol intake ${}^{\#}$
Never/rare	47	52
Light	38	43
Moderate	25	31
Heavy	26	25
Ex-drinker	28	19	0.589
Family history of ovarian cancer ${}^{\&}$
Yes	56	51
No	108	119	0.482
Menarche
$\leqslant$ 12 years old	102	92
$>$ 12 years old	62	78	0.150

Note: ${}^{\ast}$ Non-smokers (never), subjects smoking 10 cigarette/day for 10 years (ever), those smoking 20 cigarette/day for 20 years (moderate), and those smoking 20 cigarette/day for 30 years (heavy). ${}^{\#}$ Light drinkers 1–8.9 U/week, moderate drinkers 9–17.9 U/week, or heavy drinkers 18 U/week, here 1 U $=$ 22 g ethanol. ${}^{\&}$ A positive family history of ovarian cancer was defined as reporting of ovarian cancer in one or more first degree relatives.

Therefore, this study enrolled 164 patients with OC and 170 healthy women by conducting quantitative real-time polymerase chain reaction (qRT-PCR) to determine PD-1/PD-L1 expressions in peripheral blood mononuclear cell (PBMC) as well as polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) and allele-specific amplification to detect PD-1 rs2227982 and PD-L1 rs4143815, to further analyze the associations of PD-1/PD-L1 expression and its genetic polymorphism with clinicopathological features and prognosis of ovarian cancer.

2. Materials and methods

2.1 Ethics statement

Following the Declaration of Helsinki [26], this study has been obtained informed consents in written form from all participants and approved by the ethics committee of the First People’s Hospital of Jingzhou City.

2.2 Study subjects

A total of 164 women with ovarian cancer aged 38 $\sim$ 78 (mean age of 56.15 $\pm$ 11.71 years) who were treated with surgery in department of obstetrics and gynecology from December 2008 to December 2012 were enrolled as the case group. According to histologic subtypes, there were 106 serious adenocarcinomas, 28 mucinous adenocarcinomas, 13 endometrioid adenocarcinomas, 11 clear cell adenocarcinomas and 6 other cases. Based on Federation of Gynecology and Obstetrics (FIGO) stage of the National Association of Gynecology and Obstetrics [27], 50 cases were in stage I, 21 cases in stage II, 88 cases in stage III, and 5 cases in stage IV. On basis of histological stage, 44 cases were in high differentiation, 47 cases in moderate differentiation and 73 cases in low differentiation. Inclusive criteria: ovarian cancer was the first tumor; all patients were not treated with radiotherapy or chemotherapy before surgery; all specimens were histologically confirmed with complete clinical data. Exclusive criteria: patients were combined with other malignant tumors or abnormal liver and kidney; patients with severe blood coagulation disorders; patients with primary cardiovascular and cerebrovascular diseases. Besides, 170 healthy women aged 37 $\sim$ 82 (mean age: 57.45 $\pm$ 13.77) were included as the control group. No significant differences were found in age, body mass index (BMI) and other data between case group and control group ( $P>$ 0.05, Table 1).

2.3 Quantitative reverse transcription-PCR (qRT-PCR)

The peripheral venous blood was collected from fasting patients using 3 ml ethylenediaminetetraacetic acid, and PBMCs were isolated by Ficoll gradient centrifugation. Following protocols of manufacturer, TRI (Life Technologies, Invitrogen, Thessaloniki, Greece) was used to isolate the total RNA. Then, with random 6-mer oligonucleotide primer (50 pmol/mL) (Roche, USA) and M-MLV reverse transcriptase (Invitrogen), total RNA of 1 mg was reversely transcribed into complementary DNA (cDNA), according to the manufacturer’s introduction. In the automated thermocycler RotorGene 6000 (Corbett Life Science, Sydney, Australia), SYBR-Green PCR Supermix (Invitrogen, UK) was used to determine PD-1/PDCD1 (encodes PD-1) and PD-L1/PDCD1LG1 (encodes PD-L1) at the mRNA level. The $B2M$ gene was regarded as an internal control for sample normalization, namely reference gene. We conducted duplicate qRT-PCR reactions using the cDNA reaction product (1/20 aliquot) and averaged all measurements. Primers for the amplification were purchased from Qiagen (Valencia, CA, USA) and designed by the Oligo 6.0 software (NBI, Plymouth, MN, USA). In Table 2, primers for the amplification and thermocycling conditions of the analyzed genes were shown. The relative expression level was analyzed using 2 ${}^{-\Delta\Delta CT}$ .

Table 2
Primers and PCR conditions for the amplification of the analyzed genes

Gene	Sequence	PCR conditions
PD1/PDCD1	Commercially obtained by Qiagen Cat No PPH13086E	95 ${}^{\circ}$ C for 10 min, followed by 40 cycles (95 ${}^{\circ}$ C for 15 s, 60 ${}^{\circ}$ C for 60 s)
PDL1/PDCD1LG1	Forward: GGTGGTGCCGACTACAA Reverse: TAGCCCTCAGCCTGACAT	95 ${}^{\circ}$ C for 2 min, followed by 40 cycles (95 ${}^{\circ}$ C for 10 s, 58 ${}^{\circ}$ C for 10 s, 72 ${}^{\circ}$ C for 20 s)
B2M	Commercially obtained by Qiagen Cat No PPH01094E	95 ${}^{\circ}$ C for 10 min, followed by 40 cycles (95 ${}^{\circ}$ C for 15 s, 60 ${}^{\circ}$ C for 60 s)

2.4 Genotyping

The whole blood was taken to extract genomic DNA using Universal Genomic DNA Extraction Kit Ver. 3.0 (TaKaRa, Japan). With amplification in a T-Gradient Thermoblock PCR System (Biometra, Germany), the genotypes of PD-1 rs2227982 polymorphism was detected by PCR-RFLP method. The total of 50 $\mu$ l PCR reaction solution consisted of 0.3 $\mu$ g genomic DNA, 1 $\times$ PCR buffer, 0.3 mM MgCl ${}_{2}$ , 0.2 mM dNTPs, 2.5 U Taq DNA Taq polymerase (Takara, Japan), and 0.1 $\mu$ mol primer (Invitrogen, USA). Here, primer sequences were 5’-GGGACATTGAGCCAGAGAG-3’ (Forward) and 5’-CAATGGTGGCATACTCCG-3’ (Reverse). The restriction enzyme BbvCI (NEB, UK) was used to digest PCR products with the guidance of manufacturer. Allele specific amplification was carried out to detect the PD-L1 rs4143815 C>G polymorphism, with Forward: 5’-CCACTCAATGCCTCAAT GTG-3’, Reverse: 5’-GCAGCAAGTTTAGTTTGGC G-3’. The PCR reaction system (50 $\mu$ l) contained 0.5 $\mu$ l of 5U/ $\mu$ l Taq DNA polymerase, 10 $\times$ Buffer of 5 $\mu$ l, 4 $\mu$ l of 2.5 m M/L d NTP Mixture, 2 $\mu$ l of 10 $\mu$ M/L Primers, 100 ng DNA template and 33.5 $\mu$ l of dd H ${}_{2}$ O. Next, PCR products were separated via agarose gel electrophoresis (AGE), stained with ethidium bromide and observed under vilber lourmat.

2.5 Follow-up

The follow-up started from discharge time after systemic treatment till December 2016. During follow-up, patients who died of other causes, or were lost to follow-up or were survivors till the end of research, were recorded as censoring, and patients lost to follow-up were treated according to the last statistical time. The follow-up was carried out via outpatient service, telephone or medical record. The survival time was calculated as month. The OS was the time starting from date of diagnose with ovarian cancer till the date of death caused by any reason. The PFS was the time beginning from the date of diagnose with ovarian cancer till the date of disease progression and death.

2.6 Statistical analysis

All data were analyzed using SPSS 21.0 software (SPSS Inc., Chicago, IL, USA). Represented as mean $\pm$ standard deviation, the measurement data complying with normal distribution were compared by $t$ -test, and the correlation analysis was conducted by Pearson correlation coefficient. Expressed as percentage and radio, enumeration data were analyzed using Chi-square test. The Hardy-Weinberg equilibrium test was carried out to analyze expected value and observed value of genotype frequencies. The odds ratio (OR) and 95% confidence interval (CI) were used to express the relative risk level using non-conditional logistic regression analysis, and Kaplan-Meier method was applied to conduct the survival analysis. Both univariate and multivariate analyses were performed using Cox proportional hazards regression model. A $P$ value less than 0.05 was regarded as statistical significance.

3. Results

3.1 Association of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G polymorphisms with the risk of ovarian cancer and its clinicopathologic features

Table 3
Association of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G polymorphisms with the risk of ovarian cancer

SNP	Cases group ( $n=$ 164)	Control group ( $n=$ 170)	$\chi^{2}$	$P$	OR	95%CI
rs2227982 C>T
CC	87 (53.05%)	111 (65.29%)				reference
CT	60 (36.59%)	48 (28.24%)	3.778	0.052	1.595	0.995 $\sim$ 2.557
TT	17 (10.37%)	11 (6.47%)	2.779	0.096	1.972	0.878 $\sim$ 4.427
CT $+$ TT	77 (46.95%)	59 (34.71%)	5.185	0.023	1.665	1.072 $\sim$ 2.586
C	234 (71.34%)	270 (79.41%)				reference
T	94 (28.66%)	70 (20.59%)	5.870	0.015	1.549	1.086 $\sim$ 2.211
rs4143815 C/G
CC	31 (18.90%)	54 (31.76%)				reference
CG	82 (50.00%)	78 (45.88%)	4.879	0.027	1.831	1.067 $\sim$ 3.142
GG	51 (31.10%)	38 (22.35%)	7.573	0.006	2.338	1.271 $\sim$ 4.301
CG $+$ GG	133 (81.10%)	116 (68.24%)	7.279	0.007	1.997	1.203 $\sim$ 3.316
C	144 (43.90%)	186 (54.71%)
G	184 (56.10%)	154 (45.29%)	7.795	0.005	1.543	1.137 $\sim$ 2.094

Table 4

Association of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G polymorphisms with clinicopathological features

Parameter	N	PD-1 rs2227982 C>T					PD-L1 rs4143815 C>G
		Genotype distribution			CT $+$ TT vs. CC		Genotype distribution			CG $+$ GG vs. CC
		CC	CT	TT	P	OR	CC	CG	GG	P	OR
		( $n=$ 87)	( $n=$ 60)	( $n=$ 17)		(95 %CI)	( $n=$ 31)	( $n=$ 82)	( $n=$ 51)		(95 %CI)
FIGO stage
I–II	71	48	15	8			18	33	20
III–IV	93	39	45	9	0.001	2.890	13	49	31	0.065	2.090
						(1.515 $\sim$ 5.511)					(0.945 $\sim$ 4.621)
Histological type
Serous	106	60	42	4			26	72	8
Non-Serous	58	27	18	13	0.218	1.498	5	10	43	0.013	3.445
						(0.787 $\sim$ 2.850)					(1.244 $\sim$ 9.538)
Differentiation
grade
G1 $+$ G2	91	68	19	4			26	45	20
G3	73	19	41	13	$<$ 0.001	8.403	5	37	31	$<$ 0.001	5.278
						(4.152 $\sim$ 17.010)					(1.912 $\sim$ 14.560)

The genotype frequencies of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G in the control group were in accordance with Hardy-Weinberg equilibrium ( $P=$ 0.075 and $P=$ 0.334). Non-conditional logistic regression analysis showed that the risk of ovarian cancer was significantly increased for carriers with CT $+$ TT genotype and T allele of rs2227982 C>T (CT $+$ TT vs. CC: OR $=$ 1.665, 95%CI $=$ 1.072 $\sim$ 2.586, $P=$ 0.023; T vs. C: OR $=$ 1.549, 95%CI $=$ 1.086 $\sim$ 2.211, $P=$ 0.015). Similarly, as shown in Table 3, patients carried with C/G, G/G and CG $+$ GG genotypes and G allele of rs4143815 C>G had an obvious elevation in the risk of ovarian cancer (CG vs. CC: OR $=$ 1.831, 95%CI $=$ 1.067 $\sim$ 3.142, $P=$ 0.027; GG vs. CC: OR $=$ 2.338, 95%CI $=$ 1.271 $\sim$ 4.301, $P=$ 0.006; CG $+$ GG vs. CC: OR $=$ 1.997, 95%CI $=$ 1.203 $\sim$ 3.316, $P=$ 0.007; T vs. C: OR $=$ 1.543, 95%CI $=$ 1.137 $\sim$ 2.094, $P=$ 0.005). In addition, the genotype of PD-1 rs2227982 C>T was strongly linked to FIGO staging and differentiation grade of patients with ovarian cancer (both $P<$ 0.05), but did not show any association with histological type ( $P>$ 0.05), Meanwhile, the genotype in PD-L1 rs4143815 C>G polymorphism had a close relationship with histological type and differentiation grade (both $P<$ 0.05), but was irrelevant to FIGO staging in patients with ovarian cancer ( $P>$ 0.05, Table 4).

3.2 Association of PD-1/PD-L1 expressions with its clinicopathologic features in PBMCs of patients with ovarian cancer

The qRT-PCR results demonstrated that the relative mRNA expression levels of PD-1 and PD-L1 were significantly increased in PBMCs of patients with ovarian cancer as compared with the control group (both $P<$ 0.05, Fig. 1A), which was positively correlated (Fig. 1B). Besides, both PD-1 and PD-L1 expressions were unrelated to histological type of patients with ovarian cancer (both $P>$ 0.05), but were clearly associated with differentiation grade (both $P<$ 0.05).By comparison with ovarian cancer patients in FIGO stage I, patients in FIGO stage III had a significantly decreased PD-1 expression, as well as a reduced PD-L1 expression in FIGO stage IV (both $P<$ 0.05, Table 5).

Table 5
Correlation of PD-1/PD-L1 expressions with clinicopathologic features in PBMCs of patients with ovarian cancer

Parameter	N	PD1 mRNA expression	$P$	PDL1 mRNA expression	$P$
FIGO stage
I	50	9.08 $\pm$ 1.14		13.09 $\pm$ 3.20
II	21	8.95 $\pm$ 1.40		12.32 $\pm$ 3.42
III	88	8.56 $\pm$ 1.24 ${}^{\ast}$		12.09 $\pm$ 2.92
IV	5	8.26 $\pm$ 0.39	0.069	9.06 $\pm$ 2.29 ${}^{\ast}$	0.027
Histological type
Serous	106	8.82 $\pm$ 1.22		12.56 $\pm$ 3.21
Mucinous	28	8.94 $\pm$ 1.08		11.69 $\pm$ 3.46
Endometrioid	13	8.57 $\pm$ 1.43		10.71 $\pm$ 3.42
Clear cell	11	8.56 $\pm$ 1.22		10.32 $\pm$ 2.85
Other	6	7.62 $\pm$ 1.35 ${}^{\#}$	0.157	12.33 $\pm$ 3.12	0.127
Differentiation grade
Well (G1)	44	9.16 $\pm$ 1.22		13.45 $\pm$ 2.77
Moderate (G2)	47	9.26 $\pm$ 1.11		13.52 $\pm$ 3.17
Poor (G3)	73	8.2 $\pm$ 1.08 ${}^{\&}$	$<$ 0.001	10.89 $\pm$ 2.69 ${}^{\&}$	$<$ 0.001

Notes: ${}^{\ast}P<$ 0.05, compared with ovarian cancer patients in FIGO stage I; ${}^{\#}P<$ 0.05, compared with serous and mucinous adenocarcinomas; ${}^{\&}P<$ 0.05, compared with Well (G1).

Figure 1.

Expressions of PD-1 and PD-L1 in PBMCs of patients with ovarian cancer. Notes: A, Expressions of PD-1 and PD-L1 in PBMCs in ovarian cancer patients and controls by qRT-PCR. * $P<$ 0.05, compared with the control group; B, Correlation analysis of PD-1 and PD-L1 in PBMCs of patients with ovarian cancer.

3.3 Correlation of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G polymorphisms with PD-1/PD-L1 expressions in patients with ovarian cancer

Figure 2.

Expressions of PD-1 and PD-L1 in PBMCs of ovarian cancer patients with different genotypes of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G Notes: A, Expressions of PD-1 and PD-L1 in PBMCs of ovarian cancer patients with different genotypes of PD-1 rs2227982 C>T; B, Expressions of PD-1 and PD-L1 in PBMCs of ovarian cancer patients with different genotypes of PD-L1 rs4143815 C>G; ${}^{*}P<$ 0.05, compared with wild homozygous; ${}^{\#}P<$ 0.05, compared with heterozygous.

Figure 3.

Effects of PD-1/PD-L1 polymorphisms and their expressions on the prognosis of patients with ovarian cancer. Notes: A-B, Effects of PD-1 rs2227982 C>T (A) and PD-L1 rs4143815 C>G (B) on PFS of patients with ovarian cancer; C-D, Effects of PD-1 rs2227982 C>T (C) and PD-L1 rs4143815 C>G (D) on OS of patients with ovarian cancer; E-F, Effects of PD-1 (E) and PD-L1 (F) expressions on PFS of patients with ovarian cancer; G-H , Effects of PD-1 (G) and PD-L1 (H) expressions on OS of patients with ovarian cancer.

As revealed in Fig. 2, when compared with wild homozygote carriers, both PD-1 and PD-L1 expressions were significantly decreased in PBMCs of ovarian cancer patients who were the mutant carriers of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G (all $P<$ 0.05). Additionally, the expression of PD-L1 in ovarian cancer patients with a homozygous mutation in the PD-1 rs2227982 C>T was significantly lower than those in heterozygous mutation carriers ( $P<$ 0.05).

3.4 Effects of PD-1/PD-L1 polymorphisms and their expressions on the prognosis of patients with ovarian cancer

The ovarian cancer patients were followed up for 5 years, and the loss-to-follow-up rate was 4.88% (8/164). Kaplan-Meier survival curve showed that significant differences in PFS and OS were found in patients with ovarian cancer in different genotypes of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G (all $P>$ 0.05), and PFS and OS of wild homozygous genotype carriers were significantly higher than those of mutant genotype carriers (all $P<$ 0.05, Fig. 3A-D). In addition, according to the median of PD-1 and PD-L1 expressions in PBMCs of patients with ovarian cancer were divided into high expression group and low expression group (Median ${}_{\textit{PD-1}}=$ 8.800, Median ${}_{\textit{PD-L1}}=$ 12.251), and patients with low expressed PD-1 and PD-L1 had significantly decreased PFS and OS than patients with high expressed PD-1 and PD-L1 (all $P<$ 0.05, Fig. 3). Univariate COX regression analysis showed that the risk factors for prognosis of patients with ovarian cancer were high FIGO stage, low differentiation grade, CT $+$ TT genotype of PD-1 rs2227982 C>T, CG $+$ GG genotype of PD-L1 rs4143815 C>G and low PD-1 and PD-L1 expressions (all $P<$ 0.05). While the multivariate COX regression analysis identified high FIGO stage and low PD-1 and PD-L1 expressions as the independent risk factors for prognosis of patients with ovarian cancer (all $P<$ 0.05, Table 6).

Table 6
Risk factors for prognosis of patients with ovarian cancer by Cox regression model

Parameter	Univariate regression analysis						Multivariate regression analysis
	B	SE	Wald	$P$	HR	95%CI	B	SE	Wald	$P$	HR	95%CI
Age	0.091	0.196	0.215	0.643	1.095	0.745 $\sim$ 1.610
Number of pregnancy	0.197	0.115	2.931	0.087	1.218	0.972 $\sim$ 1.527
Body mass index	0.062	0.099	0.384	0.535	1.064	0.875 $\sim$ 1.292
Smoking status			1.107	0.775
Ever vs. Never	$-$ 0.031	0.248	0.016	0.899	0.969	0.596 $\sim$ 1.576
Moderate vs. Never	$-$ 0.197	0.330	0.357	0.550	0.821	0.430 $\sim$ 1.567
Heavy vs. Never	0.137	0.270	0.259	0.610	1.147	0.676 $\sim$ 1.947
Alcohol intake			3.776	0.437
Light vs. Never/rare	$-$ 0.336	0.269	1.561	0.212	0.715	0.422 $\sim$ 1.210
Moderate vs. Never/rare	$-$ 0.032	0.299	0.012	0.914	0.968	0.539 $\sim$ 1.739
Heavy vs. Never/rare	$-$ 0.350	0.312	1.259	0.262	0.705	0.383 $\sim$ 1.298
Ex-drinker vs. Never/rare	0.131	0.280	0.220	0.639	1.140	0.659 $\sim$ 1.973
Family history of ovarian	$-$ 0.092	0.203	0.204	0.651	0.912	0.613 $\sim$ 1.359
cancer (No vs. Yes)
Menarche	0.204	0.193	1.108	0.293	1.226	0.839 $\sim$ 1.791
FIGO stage			10.847	0.013					7.963	0.047
II vs. I	$-$ 0.002	0.321	0.000	0.994	0.998	0.532 $\sim$ 1.871	$-$ 0.209	0.325	0.414	0.520	0.811	0.430 $\sim$ 1.533
III vs. I	0.306	0.219	1.944	0.163	1.358	0.883 $\sim$ 2.087	0.191	0.226	0.712	0.399	1.210	0.777 $\sim$ 1.887
IV vs. I	1.674	0.536	9.739	0.002	5.332	1.864 $\sim$ 15.257	1.364	0.548	6.189	0.013	3.910	1.336 $\sim$ 11.447
Histological type			1.898	0.755
Mucinous vs. Serous	0.159	0.249	0.410	0.522	1.173	0.720 $\sim$ 1.910
Endometrioid vs. Serous	0.356	0.355	1.010	0.315	1.428	0.713 $\sim$ 2.862
Clear cell vs. Serous	0.204	0.397	0.264	0.608	1.226	0.563 $\sim$ 2.667
Other vs. Serous	$-$ 0.352	0.590	0.356	0.551	0.703	0.221 $\sim$ 2.235
Differentiation grade			15.657	$<$ 0.001					2.688	0.261
G2 vs. G1	0.028	0.278	0.010	0.920	1.028	0.597 $\sim$ 1.772	$-$ 0.060	0.280	0.046	0.831	0.942	0.544 $\sim$ 1.631
G3 vs. G1	0.774	0.240	10.389	0.001	2.168	1.354 $\sim$ 3.470	0.324	0.263	1.516	0.218	1.382	0.826 $\sim$ 2.313
rs2227982 C>T	0.968	0.194	24.810	$<$ 0.001	2.632	1.798 $\sim$ 3.852	0.507	0.295	2.947	0.086	1.660	0.931 $\sim$ 2.960
(CT $+$ TT vs. CC)
rs4143815 C>G	0.822	0.272	9.151	0.002	2.274	1.336 $\sim$ 3.873	0.116	0.319	0.132	0.717	1.123	0.601 $\sim$ 2.099
(CG $+$ GG vs. CC)
PD1 mRNA expression	$-$ 0.222	0.070	10.123	0.001	0.801	0.699 $\sim$ 0.918	$-$ 0.212	0.091	9.865	0.004	0.510	0.167 $\sim$ 0.952
PDL1 mRNA expression	$-$ 0.214	0.040	29.326	$<$ 0.001	0.807	0.747 $\sim$ 0.872	$-$ 0.273	0.069	15.683	$-$ 0.001	0.761	0.665 $\sim$ 0.871

Notes: FIGO, Federation of Gynecology and Obstetrics; values in bold font is to underscore statistical significance ( $P<$ 0.05).

4. Discussion

In this study, one of the most striking results was PD-1 rs2227982 C>T can affect the PD-1 expression in PBMCs to be involved in the occurrence and development of ovarian cancer. Currently, there still exists controversy about the role of PD-1 rs2227982 C>T in diseases, taken an example by Piskin et al., the frequency of T alleles in the PD-1 rs2227982 polymorphism was higher in control group than patients with subacute sclerosing panencephalitis (SSPE) [28], suggesting a protective role of T allele in SSPE. Dong W and his group carried out a meta-analysis stating that no significant association was observed between rs2227982 polymorphism and cancer risks in all genetic models and alleles [19]. Nevertheless, our results showed that the T allele within the PD-1 rs2227982 C>T significantly increased the risk for ovarian cancer, and such carriers of ovarian cancer patients had a higher FIGO stage but a lower differentiation grade. Consistently, both Qian and Yang teams demonstrated that the single nucleotide polymorphism (SNP) rs2227982 of the PD-1 gene may confer susceptibility to the development of ankylosing spondylitis, and the carriage of the TT genotype can enhance the thoracolumbar kyphosis severity in ankylosing spondylitis patients [29, 30]. And patients with dilated cardiomyopathy presented higher frequencies of TT genotype and T allele of rs2227982, as revealed by Yang et al. [31], indicating T allele of rs2227982 C>T was a risk factor for ankylosing spondylitis, dilated cardiomyopathy, as well as ovarian cancer and other diseases, despite of different mechanisms in different diseases. To our knowledge, gene mutation can affect diseases progression via altering its transcriptional level, for example, the NF- $\kappa$ B mRNA level in cancer tissue was strongly related to the NF- $\kappa$ B -94 ins/del ATTG polymorphism in patients with ovarian cancer [32]. Similar to our findings, the PD-L1 expression was significantly higher in patients carrying PD-1 rs2227982 C>T mutant homozygote than in heterozygote carriers, suggesting that rs2227982 may affect PD-L1expression to induce ovarian cancer. Furthermore, there is evidence pointed out that an amino acid substitution from alanine to valine caused by the genetic variation in rs2227982 may change the structure and function for PD-1 and thereby affect the progression of different diseases [21], but the specific mechanism needs to be further studied.

Furthermore, another important finding was that patients carrying PD-L1 rs4143815 C>G C/G, G/G and CG+GG genotypes and G allele were in increased risk for ovarian cancer, and this SNP was significantly correlated with histological type and differentiation grade. Multiple researchers have found that the PD-L1 rs4143815 G allele carriers had an obvious elevation with the risk of numbers of diseases as compared with C allele carriers, such as gastric cancer [23] and Type I diabetes [33], which were consistent with our results, possibly because G allele in the SNP rs4143815 of PD-L1 can reduce luciferase activity via influencing miRNA binding, exhibiting a reduced PD-L1 expression [34, 35]. What’s more, Kataoka et al. reported structural variations disrupted PD-L1 3’-UTR and thereby resulted in abnormal PD-L1 expression, which capacitates immune evasion in various cancers [36]. Therefore, we conducted qRT-PCR to determine the transcription level of PD-L1 and found that PD-L1 expression was significantly decreased in PBMCs of patients carried with rs4143815 mutant genotype, indicating that PD-L1 rs4143815 C>G can potentially regulate PD-L1 expression in ovarian cancer.

PD-1/PD-L signaling pathway is one of the most important pathways inhibiting the tumor immune response, and one previous study indicated that PD-L1 expression on tumor cells or antigen-presenting cells (APCs) in the tumor microenvironment could promote tumor growth and induce the apoptosis of T cells of PD-1 with antitumor activity, further leading to poor-prognosis and therapy-resistant, such as esophagus cancer [37] and renal cell carcinoma [38]. Of note, in spite of their well-known immune-modulatory function, the high expression of PD-1/PD-L also suggested a good prognosis in pancreatic cancer [14], thymic cancer [39], and malignant melanoma [40], and there is a possible explanation that PD-1 and PD-L1 is in compensatory up-regulation in a microenvironment which may threaten tumors by means of an active immune response [16] Moreover, this study also revealed that PFS and OS in ovarian cancer patients with high PD-1 and PD-L1 expression were significantly higher than those patients with low expression, and high FIGO stage and low expressions of PD-1 and PD-L1 were independent risk factors for the prognosis of patients with ovarian cancer. As illustrated by Darb-Esfahani et al., high PD-1 and PD-L1 expression levels are indeed indicators of a good prognosis in ovarian cancer [16]. but PD-L1 expression was found to exert a negative prognostic effect in ovarian cancer cells in Hamanishi et al. article [17], which might be partially associated with only 40% of serous carcinomas cases investigated in this study, while our study enrolled 65% patients with serous ovarian cancer. Besides, PFS and OS were significantly higher in patients carrying wild homozygote genotypes of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G than mutant genotype carriers. Considering the above, we reasonably speculated that PD-1/PD-L1 mutation may lead to aberrant gene expression, and consequently influencing the prognosis of patients with ovarian cancer.

Taken together, PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G mutation increased the risk of ovarian cancer, which was associated with down-regulated PD-1/PD-L1 expression in PBMCs. Moreover, patients carried with the wild homozygous genotype of rs2227982 C>T and rs4143815 C>G had significantly higher PFS and OS than mutant genotype carriers, and higher expressions of PD-1 and PD-L1 indicated a better prognosis with higher PFS and OS. However, PD-1/PD-L1 expressions in cancer cells and tumor-infiltrating lymphocytes (TILs) as well as other SNPs in PD-1/PD-L1 gene (especially rs2297136 and rs2282055) were not determined due to the limited time and fees, which needs to be further explored in the following experiments.

Footnotes

Acknowledgments

The authors appreciate the reviewers for their useful comments in this paper.

Conflict of interest

The authors declare no conflicts of interest.

References

Deng

Wang

and Du

, Androgen receptor gene CAG repeat polymorphism and ovarian cancer risk: A meta-analysis, Biosci Trends (2017).

Torre

L.A.

Bray

Siegel

R.L.

Ferlay

Lortet-Tieulent

and Jemal

, Global cancer statistics, 2012, CA Cancer J Clin 65 (2015), 87–108.

Romanowicz

Michalska

M.M.

Samulak

Malinowski

Szaflik

Bienkiewicz

and Smolarz

, Association of R156R single nucleotide polymorphism of the ERCC2 gene with the susceptibility to ovarian cancer, Eur J Obstet Gynecol Reprod Biol 208 (2017), 36–40.

Sundar

Neal

R.D.

and Kehoe

, Diagnosis of ovarian cancer, BMJ 351 (2015), h4443.

Qiu

Wang

Guo

Liu

Wang

Yuan

and Shi

, HOTAIR rs920778 polymorphism is associated with ovarian cancer susceptibility and poor prognosis in a Chinese population, Future Oncol 13 (2017), 347–355.

Supic

Kozomara

Zeljic

Jovic

and Magic

, Prognostic value of the DNMTs mRNA expression and genetic polymorphisms on the clinical outcome in oral cancer patients, Clin Oral Investig 21 (2017), 173–182.

Shi

T.Y.

Jiang

Yin

Wang

M.Y.

K.D.

Shao

Z.M.

Sun

M.H.

Zang

and Wei

, Polymorphisms in the kinesin-like factor 1 B gene and risk of epithelial ovarian cancer in Eastern Chinese women, Tumour Biol 36 (2015), 6919–6927.

Okazaki

and Honjo

, PD-1 and PD-1 ligands: from discovery to clinical application, Int Immunol 19 (2007), 813–824.

Zhou

T.C.

Sankin

A.I.

Porcelli

S.A.

Perlin

D.S.

Schoenberg

M.P.

and Zang

, A review of the PD-1/PD-L1 checkpoint in bladder cancer: From mediator of immune escape to target for treatment, Urol Oncol 35 (2017), 14–20.

10.

Tang

Wang

Jiang

Liu

Chen

and Kang

, Programmed death-1 (PD-1) rs2227981 C>T polymorphism is associated with cancer susceptibility: a meta-analysis, Int J Clin Exp Med 8 (2015), 22278–22285.

11.

Clave

Casadevall

Pijuan

Rodriguez-Rivera

Garcia-Moreno

Taus

Albanell

Espinet

Arriola

and Salido

, 4P Co-amplifications of PD-L1/PD-L2 genes (9p24.1) and PD-L1 protein expression in NSCLC patients, J Thorac Oncol 11 (2016), S58.

12.

Zhang

G.X.

J.J.

Liu

Chen

Y.J.

and Zhang

, Expression of PD-1/PD-L1 in triple-negative breast carcinoma and its significance, Zhonghua Bing Li Xue Za Zhi 46 (2017), 20–24.

13.

Mori

Motoi

Ninomiya

Matsuura

Nakao

Mun

Okumura

Nishio

Morikawa

and Ishikawa

, High expression of programmed cell death 1 ligand 1 in lung adenocarcinoma is a poor prognostic factor particularly in smokers and wild-type epidermal growth-factor receptor cases, Pathol Int 67 (2017), 37–44.

14.

Wang

Lin

Cui

Han

Jiao

Meng

and Wang

, Prognostic value and clinicopathological features of PD-1/PD-L1 expression with mismatch repair status and desmoplastic stroma in Chinese patients with pancreatic cancer, Oncotarget 8 (2017), 9354–9365.

15.

Ning

X.H.

Gong

Y.Q.

S.M.

Wang

J.Y.

Peng

S.H.

Chen

J.C.

Liu

J.Y.

N.N.

Guo

Y.L.

and Gong

, Higher programmed cell death 1 ligand 1 (PD-L1) mRNA level in clear cell renal cell carcinomas is associated with a favorable outcome due to the active immune responses in tumor tissues, Oncotarget 8 (2017), 3355–3363.

16.

Darb-Esfahani

Kunze

C.A.

Kulbe

Sehouli

Wienert

Lindner

Budczies

Bockmayr

Dietel

Denkert

Braicu

and Johrens

, Prognostic impact of programmed cell death-1 (PD-1) and PD-ligand 1 (PD-L1) expression in cancer cells and tumor-infiltrating lymphocytes in ovarian high grade serous carcinoma, Oncotarget 7 (2016), 1486–1499.

17.

Hamanishi

Mandai

Iwasaki

Okazaki

Tanaka

Yamaguchi

Higuchi

Yagi

Takakura

Minato

Honjo

and Fujii

, Programmed cell death 1 ligand 1 and tumor-infiltrating CD8+ T lymphocytes are prognostic factors of human ovarian cancer, Proc Natl Acad Sci U S A 104 (2007), 3360–3365.

18.

Zhang

Zhao

Huang

and Yu

, The association between polymorphisms in the PDCD1 gene and the risk of cancer: A PRISMA-compliant meta-analysis, Medicine (Baltimore) 95 (2016), e4423.

19.

Dong

Gong

Shi

Xiao

Zhang

and Peng

, Programmed Cell Death-1 Polymorphisms Decrease the Cancer Risk: A Meta-Analysis Involving Twelve Case-Control Studies, PLoS One 11 (2016), e0152448.

20.

Zhou

R.M.

Wang

Huang

Cao

S.R.

and Shan

B.E.

, Association of programmed death-1 polymorphisms with the risk and prognosis of esophageal squamous cell carcinoma, Cancer Genet 209 (2016), 365–375.

21.

Ren

H.T.

Y.M.

Wang

X.J.

Kang

H.F.

Jin

T.B.

X.B.

Liu

X.H.

Wang

Liu

Yao

Q.L.

and Dai

Z.J.

, PD-1 rs2227982 Polymorphism Is Associated With the Decreased Risk of Breast Cancer in Northwest Chinese Women: A Hospital-Based Observational Study, Medicine (Baltimore) 95 (2016), e3760.

22.

Tang

Chen

Sun

and Kang

, Programmed death-1 (PD-1) polymorphism is associated with gastric cardia adenocarcinoma, Int J Clin Exp Med 8 (2015), 8086–8093.

23.

Tao

L.H.

Zhou

X.R.

F.C.

Chen

Meng

F.Y.

Mao

Hua

Zhang

H.J.

Wang

W.P.

and Chen

W.C.

, A polymorphism in the promoter region of PD-L1 serves as a binding-site for SP1 and is associated with PD-L1 overexpression and increased occurrence of gastric cancer, Cancer Immunol Immunother 66 (2017), 309–318.

24.

Wang

Mao

Zhou

Sun

Liu

Chen

Hua

and Zhang

, A miR-570 binding site polymorphism in the B7-H1 gene is associated with the risk of gastric adenocarcinoma, Hum Genet 132 (2013), 641–648.

25.

Nomizo

Ozasa

Tsuji

Funazo

Yasuda

Yoshida

Yagi

Sakamori

Nagai

Hirai

and Kim

Y.H.

, Clinical Impact of Single Nucleotide Polymorphism in PD-L1 on Response to Nivolumab for Advanced Non-Small-Cell Lung Cancer Patients, Sci Rep 7 (2017), 45124.

26.

The Helsinki Declaration of the World Medical Association (WMA). Ethical principles of medical research involving human subjects, Pol Merkur Lekarski 36 (2014), 298–301.

27.

Ishioka

Sagae

Terasawa

Sugimura

Nishioka

Tsukada

and Kudo

, Comparison of the usefulness between a new universal grading system for epithelial ovarian cancer and the FIGO grading system, Gynecol Oncol 89 (2003), 447–452.

28.

Piskin

I.E.

Calik

Abuhandan

Kolsal

Celik

S.K.

and Iscan

, PD-1 gene polymorphism in children with subacute sclerosing panencephalitis, Neuropediatrics 44 (2013), 187–190.

29.

Qian

B.P.

Wang

X.Q.

Qiu

Jiang

M.L.

and Jiang

, An exon polymorphism of programmed cell death 1 gene is associated with both the susceptibility and thoracolumbar kyphosis severity of ankylosing spondylitis in a Chinese Han population, J Orthop Sci 18 (2013), 514–518.

30.

Yang

Zou

Bai

and Li

, The programmed cell death 1 gene polymorphisms (PD 1.3 G/A, PD 1.5 C/T and PD 1.9 C/T) and susceptibility to ankylosing spondylitis: a meta-analysis, J Orthop Sci 20 (2015), 55–63.

31.

Yang

Peng

Xie

Zhou

and Rao

, Association of programmed cell death 1 gene polymorphisms with dilated cardiomyopathy in Chinese Han population, Zhonghua Yi Xue Yi Chuan Xue Za Zhi 31 (2014), 362–366.

32.

Huo

Z.H.

Zhong

H.J.

Zhu

Y.S.

Xing

and Tang

, Roles of functional NFKB1 and beta-TrCP insertion/deletion polymorphisms in mRNA expression and epithelial ovarian cancer susceptibility, Genet Mol Res 12 (2013), 3435–3443.

33.

Pizarro

Garcia-Diaz

D.F.

Codner

Salas-Perez

Carrasco

and Perez-Bravo

, PD-L1 gene polymorphisms and low serum level of PD-L1 protein are associated to type 1 diabetes in Chile, Diabetes Metab Res Rev 30 (2014), 761–766.

34.

Lee

S.Y.

Jung

D.K.

Choi

J.E.

Jin

C.C.

Hong

M.J.

S.K.

Kang

H.G.

Lee

W.K.

Seok

Lee

E.B.

Jeong

J.Y.

Shin

K.M.

Yoo

S.S.

Lee

Cha

S.I.

Kim

C.H.

and Park

J.Y.

, Functional polymorphisms in PD-L1 gene are associated with the prognosis of patients with early stage non-small cell lung cancer, Gene 599 (2017), 28–35.

35.

Wang

Sun

Liu

Yang

Zhu

Chen

Tian

Zhou

Mao

Zhang

Jiang

Hua

Chen

and Zhang

, A frequent somatic mutation in CD274 3’-UTR leads to protein over-expression in gastric cancer by disrupting miR-570 binding, Hum Mutat 33 (2012), 480–484.

36.

Kataoka

Shiraishi

Takeda

Sakata

Matsumoto

Nagano

Maeda

Nagata

Kitanaka

Mizuno

Tanaka

Chiba

Ito

Watatani

Kakiuchi

Suzuki

Yoshizato

Yoshida

Sanada

Itonaga

Imaizumi

Totoki

Munakata

Nakamura

Hama

Shide

Kubuki

Hidaka

Kameda

Masuda

Minato

Kashiwase

Izutsu

Takaori-Kondo

Miyazaki

Takahashi

Shibata

Kawamoto

Akatsuka

Shimoda

Takeuchi

Seya

Miyano

and Ogawa

, Aberrant PD-L1 expression through 3’-UTR disruption in multiple cancers, Nature 534 (2016), 402–406.

37.

Ohigashi

Sho

Yamada

Tsurui

Hamada

Ikeda

Mizuno

Yoriki

Kashizuka

Yane

Tsushima

Otsuki

Yagita

Azuma

and Nakajima

, Clinical significance of programmed death-1 ligand-1 and programmed death-1 ligand-2 expression in human esophageal cancer, Clin Cancer Res 11 (2005), 2947–2953.

38.

McDermott

D.F.

Drake

C.G.

Sznol

Choueiri

T.K.

Powderly

J.D.

Smith

D.C.

Brahmer

J.R.

Carvajal

R.D.

Hammers

H.J.

Puzanov

Hodi

F.S.

Kluger

H.M.

Topalian

S.L.

Pardoll

D.M.

Wigginton

J.M.

Kollia

G.D.

Gupta

McDonald

Sankar

Sosman

J.A.

and Atkins

M.B.

, Survival, Durable Response, and Long-Term Safety in Patients With Previously Treated Advanced Renal Cell Carcinoma Receiving Nivolumab, J Clin Oncol 33 (2015), 2013–2020.

39.

Yokoyama

Miyoshi

Nakashima

Shimono

Hashiguchi

Mitsuoka

Takamori

Akagi

and Ohshima

, Prognostic Value of Programmed Death Ligand 1 and Programmed Death 1 Expression in Thymic Carcinoma, Clin Cancer Res 22 (2016), 4727–4734.

40.

Kluger

H.M.

Zito

C.R.

Barr

M.L.

Baine

M.K.

Chiang

V.L.

Sznol

Rimm

D.L.

Chen

and Jilaveanu

L.B.

, Characterization of PD-L1 Expression and Associated T-cell Infiltrates in Metastatic Melanoma Samples from Variable Anatomic Sites, Clin Cancer Res 21 (2015), 3052–3060.

Correlation of PD-1/PD-L1 polymorphisms and expressions with clinicopathologic features and prognosis of ovarian cancer

Abstract

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSION:

Keywords

1. Introduction

Table 1 Comparison of baseline characteristics between case group and control group

2.1 Ethics statement

2.2 Study subjects

2.3 Quantitative reverse transcription-PCR (qRT-PCR)

Table 2 Primers and PCR conditions for the amplification of the analyzed genes

2.5 Follow-up

2.6 Statistical analysis

3. Results

3.1 Association of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G polymorphisms with the risk of ovarian cancer and its clinicopathologic features

Table 3 Association of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G polymorphisms with the risk of ovarian cancer

Table 5 Correlation of PD-1/PD-L1 expressions with clinicopathologic features in PBMCs of patients with ovarian cancer

Table 6 Risk factors for prognosis of patients with ovarian cancer by Cox regression model

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Comparison of baseline characteristics between case group and control group

Table 2
Primers and PCR conditions for the amplification of the analyzed genes

Table 3
Association of PD-1 rs2227982 C>T and PD-L1 rs4143815 C>G polymorphisms with the risk of ovarian cancer

Table 5
Correlation of PD-1/PD-L1 expressions with clinicopathologic features in PBMCs of patients with ovarian cancer

Table 6
Risk factors for prognosis of patients with ovarian cancer by Cox regression model