Introduction

Multiple myeloma (MM) is a lymphoproliferative disorder characterized by the proliferation of malignant plasma cells, lytic bone lesions, and the presence of monoclonal immunoglobulins in serum and/or urine. The incidence of MM increases with age; it is more common in the elderly and is rare in patients under 40 years of age [1]. Recently, novel agents, including immunomodulatory drugs (IMiDs) such as thalidomide and proteasome inhibitors such as bortezomib, have been used for treating this disease [2]. These agents have improved the prognosis of MM patients; however, MM remains an incurable disease. Moreover, the interactions between MM cells and their bone marrow (BM) microenvironment such as immune cells, endothelial cells, and mesenchymal stromal cells play important roles in myeloma cell growth and survival, and the development of resistance to therapy [3,4,5].

Immune checkpoint pathways are critical modulators of the immune system, allowing the initiation of the immune response and preventing the onset of autoimmunity. Programmed cell death protein-1 (PD-1, encoded by PDCD1) and cytotoxic T lymphocyte-associated antigen-4 (CTLA-4, encoded by CTLA4) play central roles in immune checkpoint pathways. PD-1 is expressed in the activated T cells and has been reported to suppress activation of lymphocytes and cytokine production by interacting with its ligands, PD-1 ligand-1 (PD-L1, encoded by PDCD1LG1) and PD-1 ligand-2 (PD-L2) [6, 7]. CTLA-4 is homologous to CD28 and a competitive antagonist for B7 (CD80 and CD86) on the surface of antigen-presenting cells. CTLA-4 has greater affinity for B7 than CD28 and is responsible for T cell inactivation. In some cancers, these immune checkpoint molecules often inhibit the anti-tumor immune response. Compared to healthy controls, overexpression of CTLA-4 was observed in the BM of MM patients [8]. Moreover, PD-L1 molecules on myeloma cells not only induce T cell down-regulation, but also enhance aggressive characteristics of myeloma cells that include high proliferative ability and drug resistance [9]. However, the effect of PD-1-blocking antibody (nivolumab) was insufficient in MM patients compared with other B cell malignancies in phase I trial [10]. Thus, further understanding of the association between MM and immune checkpoint molecules is essential.

The production of PD-1 and CTLA-4 is strongly influenced by genetic factors. Single-nucleotide polymorphisms (SNPs) of the PDCD-1, PD-L1, and CTLA-4 genes (Fig. 1) are implicated in the pathogenesis of some cancers such as colon cancer, breast cancer, non-small cell lung cancer, and esophageal squamous cell carcinoma [11,12,13,14]. However, to our knowledge, there has been no study reporting the association among the SNPs in the immune checkpoint genes and MM. We investigated the role of PDCD1, PDCD1LG1, and CTLA4 SNPs in MM pathogenesis and the susceptibility to and clinical features of MM.

Fig. 1
figure 1

The PDCD1, PDCD1LG1, and CTLA4 single-nucleotide polymorphisms and their interactions. The coding region is represented by a filled black box. a Single-nucleotide polymorphisms in PDCD1. b Single-nucleotide polymorphisms in PDCD1LG1. c Single-nucleotide polymorphisms in CTLA4

Full size image

Materials and methods

Patient characteristics

A total of 124 patients with MM and 211 healthy controls (both groups from Japan) were enrolled in this study. Patients with MM were diagnosed according to the International Myeloma Working Group Classification (2003) between 1994 and 2010. This study was approved by the Institutional Review Board of Gunma University Hospital, Japan (Approval #160007). Various clinical characteristics of all the patients were recorded (Table 1).

Table 1 Characteristics of patients with multiple myeloma
Full size table

PDCD1, PDCD1LG1, and CTLA4 genotyping

To detect PDCD1 SNPs (rs36084323, rs41386349, and rs2227982), PDCD1LG1 SNPs (rs2297136), and CTLA4 SNPs (rs733618, rs11571316, rs231775, and rs3087243), we used the PCR–restriction fragment length polymorphism method (PCR–RFLP). PDCD1LG1 SNP (rs4143815) was genotyped using the TaqMan allelic discrimination real-time PCR method. Genomic DNA was isolated from whole blood using a DNA extraction kit (Qiagen GmbH, Hilden, Germany).

For analyzing the PDCD1 variants, the forward primer 5′-GGGAAGAAGGTCAAGGCTGG-3′ and the reverse primer 5′-CCACTCCCATTCTGTCGGAG-3′ for rs36084323, the forward primer 5′-CAGCAACCTCAATCCCTAAAGC-3′ and the reverse primer 5′-GAAATCCAGCTCCCCATAGTCC-3′ for rs41386349, and the forward primer 5′-GGACAGCTCAGGGTAAGCAG-3′ and the reverse primer 5′-GAAATCCAGCTCCCCATAGTCC-3′ for rs2227982 were used. For analyzing the PDCD1LG1 variants, the forward primer 5′-TGAAAGTATCAAGGTCTCCCTCC-3′ and the reverse primer 5′-GGGTTTTCCAGGATATCATGTAAGG-3′ for rs2297136 were used. For analyzing the CTLA4 variants, the forward primer 5′-CAAGCTTTGTCCTGTGACCA-3′ and the reverse primer 5′-AAGCGCCAACAAGCAATAAC-3′ for rs733618, the forward primer 5′-GTCCTGTGACCATAATGAACTCTTC-3′ and the reverse primer 5′-TTTCTGACCTGCCTGTTTTCTATAC-3′ for rs11571316, the forward primer 5′-CTCTACTTCCTGAAGACCTGAACAC-3′ and the reverse primer 5′-ATTCATGAAGCCCCTACTAAATACC-3′ for rs231775, and the forward primer 5′-TTTCTGAAAATTAACACTGCTTGTG-3′ and the reverse primer 5′-ACTGTAATGCCTGTGATAGTTGAGC-3′ for rs3087243 were used. The PCR products were digested with the restriction enzyme NciI (for PDCD1 rs36084323), BstUI (for PDCD1 rs41386349), DrdI (for PDCD1 rs2227982), PspOMI (for PDCD1LG1 rs2297136), ApeKI (for CTLA4 rs733618), HpyAV (for CTLA4 rs11571316), BbvI (for CTLA4 rs231775), and HpyCH4 (for CTLA4 rs3087243). The products were analyzed on 2% agarose gels by electrophoresis. To confirm the accuracy of PCR–RFLP, the amplification products of several individuals were sequenced using an ABI Prism Genetic Analyzer (Applied Biosystems, Foster City, CA, USA).

Statistical analysis

All statistical analyses were performed with the IBM SPSS software package ver. 24 (IBM, Armonk, NY, USA). Genotype and allele frequencies of PDCD1, PDCD1LG1, and CTLA4 were compared between the healthy controls and patients with MM using the Chi-square test. Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated for each analysis. The clinical parameters of the myeloma patients were compared with PDCD1, PDCD1LG1, and CTLA4 polymorphisms using the independent t-test for continuous variables and the Chi-square test for categorical variables. Overall survival (OS) was defined as the interval from the date of diagnosis to the date of death or the last clinical appointment. OS was estimated by the Kaplan–Meier method and was compared using the log-rank test. P < 0.05 was considered statistically significant.

Results

Clinical characteristics of patients with MM

The clinical characteristics of patients with MM are shown in Table 1. Of the 124 MM patients, 66 were men (53.2%) and 58 were women (46.8%). Their median age at diagnosis was 65.9 years (range 34.2–83.3 years). The immunoglobulin subtype was IgG for 70 patients (56.4%), IgA for 24 patients (19.4%), IgD for 3 patients (2.4%), Bence Jones protein for 25 patients (20.2%), and non-secretory type for 2 patients (1.6%). According to the International Staging System (ISS), 40 patients (32.3%) were classified as stage I, 41 (33.1%) were classified as stage II, and 43 (34.7%) were classified as stage III. Forty-six patients (37.9%) were treated with thalidomide and/or bortezomib.

Distribution of genotypes and allele frequencies among the healthy controls and patients with MM

The distributions of genotype and allele frequencies of the SNPs are summarized in Tables 2 and 3. There were no significant differences between the controls and the MM patients in PDCD1, PDCD1LG1, and CTLA4 SNPs. To analyze the haplotypes of PDCD1, PDCD1LG1, and CTLA-4 polymorphisms, we used Haploview software (BROAD institute, Cambridge, MA, UK). The patients with MM had a significantly higher frequency of the PDCD1 GCC/GCC haplotype (rs36084323/rs41386349/rs2227982) compared with the healthy controls (10.5% vs. 4.3%, odds ratio = 2.63, 95% confidence interval = 1.09–6.34, P = 0.027) (Table 2). When we compared the distribution of each genotype according to sex and age among the control group and MM patients as a whole, no significant differences were observed in the genotype or in the allele frequencies.

Table 2 Genotype and haplotype distributions of PDCD1 and PDCD1LG1 polymorphisms
Full size table
Table 3 Genotype and haplotype distributions of CTLA4 polymorphisms
Full size table

Association of PDCD1, PDCD1LG1, and CTLA4 polymorphisms with clinical variables and prognosis of patients with MM

We summarized the association of PDCD1, PDCD1LG1, and CTLA4 polymorphisms with the clinical variables of patients with MM in Tables 4, 5, and 6. Compared with the CT and TT types, the PDCD1 rs2227982 CC genotype was associated significantly with a higher frequency of bone lesions (37 [86.0%] vs. 56 [69.1%]; P = 0.038) (Table 4). Patients with PDCD1LG1 rs2297136 TT and TC types (high-expression types) showed lower albumin level than those with CC genotype (mean ± standard deviation (SD), 3.84 ± 0.65 vs. 4.36 ± 0.54 mg/dL; P = 0.040) (Table 5). In addition, the PDCD1LG1 rs4143815 CC and CG types (high-expression types) were associated significantly with higher frequency of patients who were treated with thalidomide and/or bortezomib (41 [41.8%] vs. 5 [20.0%]; P = 0.044) (Table 5). However, there was no statistical significance between CTLA4 polymorphisms and clinical variables of patients with MM. In addition, the rate of ISS high-risk patients was not significantly different for all PDCD1, PDCD1LG1, and CTLA4 SNPs. However, the R-ISS could not be evaluated because our patients were diagnosed between 1994 and 2010 and the chromosome analysis/FISH was done in a small number of patients. Moreover, no significant differences in frequencies of patients with the PDCD1, PDCD1LG1, and CTLA4 haplotypes were observed in clinical variables.

Table 4 Clinical characteristics of MM patients according to PDCD1 polymorphisms
Full size table
Table 5 Clinical characteristics of MM patients according to PDCD1LG1 polymorphisms
Full size table
Table 6 Clinical characteristics of MM patients according to CTLA4 polymorphisms
Full size table

Subsequently, we also examined the effect of PDCD1, PDCD1LG1, and CTLA4 polymorphisms on the OS in patients with MM (Fig. 2). There were no significant differences between all the polymorphisms and OS.

Fig. 2
figure 2

a Overall survival (OS) of MM patients according to the PDCD1 rs36084323 genotypes. The median survival time of patients with the GG genotype and GA and GG genotypes was 62.1 and 65.0 months, respectively (P = 0.42). b The OS of MM patients according to the PDCD1 rs41386349 genotypes. The median survival time of patients with the CC genotype and CT and TT genotypes was 71.0 and 54.1 months, respectively (P = 0.52). c OS of MM patients according to the PDCD1 rs2227982 genotypes. The median survival time of patients with the CC genotype and CT and TT genotypes was 42.4 and 65.0 months, respectively (P = 0.18). d OS of MM patients according to the PDCD1 haplotype. The median survival time of patients with the GCC/GCC haplotype and the other genotypes was 86.1 and 62.1 months, respectively (P = 0.73). e The OS of MM patients according to the PDCD1LG1 rs2297136 genotypes. The median survival time of patients with the TT and TC genotypes and CC genotype was 65.5 and 42.4 months, respectively (P = 0.57). f The OS of MM patients according to the PDCD1LG1 rs4143815 genotypes. The median survival time of patients with the CC and CG genotypes and GG genotype was 69.5 and 42.4 months, respectively (P = 0.72). g The OS of MM patients according to the PDCD1LG1 haplotype. The median survival time of patients with the other genotypes and CG/CG haplotype was 69.5 and 42.4 months, respectively (P = 0.93). h The OS of MM patients according to the CTLA4 rs733618 genotype. The median survival time of patients with the AA genotype and AG and GG genotypes was 69.5 and 59.8 months, respectively (P = 0.31). i The OS of MM patients according to the CTLA4 rs11571316 genotypes. The median survival time of patients with the GG genotype and AA and AG genotypes was 70.0 and 61.0 months, respectively (P = 0.82). j The OS of MM patients according to the CTLA4 rs231775 genotypes. The median survival time of patients with the AA genotype and AG and GG genotypes was 61.0 and 65.0 months, respectively (P = 0.36). k The OS of patients with MM according to the CTLA4 rs3087243 genotypes. The median survival time of patients with the AA genotype and AG and GG genotypes was 49.7 and 67.9 months, respectively (P = 0.89). l The OS of patients with MM according to the CTLA4 haplotype. The median survival time of patients with the AAAA/AAAA haplotype and the other haplotypes was 49.7 and 67.9 months, respectively (P = 0.89)

Full size image

Discussion

Both PDCD1 and CTLA4 are located on chromosome 2, and PDCD1LG1 is located on chromosome 9. Previous studies have shown that the SNPs in these genes influenced the expression of PD-1, PD-L1, or CTLA-4 (Fig. 1) [15,16,17,18,19,20]. A luciferase assay of the human embryonic kidney 293 (HEK293) cells showed that the promoter activity of the PDCD1 rs36084323 A allele was lower than that of the PDCD1 rs36084323 G allele [16]. Using the same method, Zheng et al. showed that the PDCD1 rs41386349 C/T in the putative enhancer-like region created a negative element in PDCD1 and that the PDCD1 rs41386349 T allele had lower transcriptional activity of PD-1 in human T cells than the PDCD1 rs41386349 C allele (Fig. 1a) [17]. The PDCD1 rs41386349 C/T results in an amino acid substitution from alanine (C allele) to valine (T allele) in exon 5 (A215V). Exon 5 in PDCD1 encodes the cytoplasmic domain of PD-1, including the immunoreceptor tyrosine-based inhibitory motif (ITIM) and the immunoreceptor tyrosine-switch motif (ITSM). The PDCD1LG1 rs2297136 and rs4143815 were located in the binding sites of miR-324-5p and miR-570, respectively. The PDCD1LG1 rs2297136 C allele and rs4143815 G allele are complementary sequence to the bound micro-RNA and may affect the PD-L1 expression [19, 20]. The CTLA4 rs733618 A/G and rs11571316 A/G alleles are located in the promoter region. The CTLA4 rs733618 A allele showed an increased transcriptional activity in the luciferase assay performed on the JURKAT cell line (CD4+ T cell line) [18]. The CTLA4 rs231775 A/G leads to an amino acid substitution from threonine (A allele) to alanine (G allele) in the leader peptide. Anjos et al. reported a higher expression of the CTLA4 rs231775 A allele cell surface compared with the CTLA4 rs231775 G allele by quantitative confocal microscopy [15]. The CTLA4 rs3087243 A/G is located in the 3′-untranslated region of the gene. In addition, activated human T cells containing the CTLA4 rs3087243 GG genotype had lower expression levels of soluble CTLA4 mRNA, a transcript variant that lacks the transmembrane domain.

A higher frequency of the PDCD1 GCC/GCC haplotype was observed in patients with MM compared to the healthy controls in this study. The PDCD1 GCC/GCC haplotype includes the higher expression alleles of rs36084323 (G allele) and rs41386349 (T allele). The expression of PD-1 was observed in the circulating T cells isolated from advanced MM patients, whereas the expression of PD-1 on circulating T cells was reduced in patients who achieve a minimal disease state following high-dose chemotherapy [21]. Moreover, Benson et al. showed that NK cells from MM patients express PD-1 and down-modulate the anti-myeloma cell effect of NK cells [22]. Although binding of myeloma PD-L1 to PD-1 did not directly affect the proliferation of myeloma cells [23], the inhibition of cancer immunity by PD-1/PD-L1 axis may lead to an environment favoring myeloma cells. Our results suggest that the PD-1 high-expression haplotype is implicated in susceptibility to MM. On the other hand, Karabon et al. reported that the CTLA4 rs231775 G allele and rs3087243 G allele were related to susceptibility to multiple myeloma in Polish population [24]. However, there were no significant differences in the CTLA4 SNPs between MM patients and healthy individuals in our study. According to 1000 Genomes Project resource (http://www.internationalgenome.org/), the distribution of PDCD1, PDCD1LG1, and CTLA4 polymorphisms varies by race. The present study showed that the PDCD1LG1 and CTLA4 SNPs have no association with the susceptibility to multiple myeloma in Japanese population.

A previous study showed that the PD-L1 expression on plasma cells from MM patients with ISS stage II and III tended to be higher than that from the MM patients with ISS stage I [9]. In addition, PD-L1 expression levels on myeloma cells from the same patients were often up-regulated when the patients relapsed or became refractory to anti-MM chemotherapy compared to the levels at initial diagnosis [9]. Lee et al. reported that the expression levels of PD-1 on CD4+ T cells from MM patients in refractory state after chemotherapy were significantly higher than those in both at diagnosis and at complete remission after chemotherapy [25]. PD-L1+ human myeloma cell lines had higher proliferative potential compared with PD-L1− myeloma cell lines [26]. The present study showed that PDCD1LG1 polymorphism was not associated with ISS high risk; however, the PDCD1LG1 rs2297136 TT and TC types (high-expression type) showed low serum albumin levels. In addition, the PDCD1 rs2227982 CC genotype showed a higher frequency of bone lesion in this study. Nagahama et al. reported the association between PD-1 and osteoclastogenesis using PD-1-deficient mice [27]. Furthermore, the expression levels of CTLA-4 on CD4+ T cells were higher in MM patients with R-ISS stage III than those with R-ISS stage II disease [25]. However, the CTLA4 polymorphisms had no association with ISS stages of MM in our study. Our results suggested that PDCD1 and PDCD1LG1 polymorphisms might contribute to the poor prognosis due to decreased albumin levels and bone lesions.

Previous studies showed the high expression of PD-L1 was associated with an increased risk of disease progression and short overall survival in MM patients [28,29,30]. Binding of myeloma PD-L1 to PD-1 induced resistance to apoptosis that was induced by melphalan and the proteasome inhibitor bortezomib via the PI3K/AKT pathway [23]. Qin et al. reported that the CTLA4 rs733618 GG genotype reduced the progression-free survival and the overall survival of MM patients who received bortezomib-based regimens [31]. However, there were no significant differences between PDCD1, PDCD1LG1, and CTLA4 polymorphisms and the prognosis of MM patients in this study. Moreover, previous studies have reported that IMiDs and proteasome inhibitors affect the expression of PD-1, PD-L1, and CTLA-4. Lenalidomide, one of IMiDs, has been shown to affect the PD-1/PD-L1 axis by down-regulating the expression of PD-L1 in MM cells [22, 32] and by decreasing the expression of PD-1 on T cells from MM patients [32, 33]. On the other hand, proteasome inhibitor bortezomib has been shown to increase PD-L1 and PD-L2 levels on MM cells and up-regulate CTLA-4 in normal T cells [34, 35]. The majority of our patients received conventional chemotherapy because the patients were diagnosed between 1994 and 2006. We also examined the effect of PDCD1, PDCD1LG1, and CTLA4 polymorphisms on the OS in patients treated with thalidomide and/or bortezomib (n = 47). However, no significant differences were observed in those polymorphisms (data not shown). In addition, there were also no statistical significance differences between PDCD1, PDCD1LG1, and CTLA4 polymorphisms and OS in patients treated with conventional therapy only (data not shown). The present study suggests that the PDCD1, PDCD1LG1, and CTLA4 are not implicated in prognosis of MM in the Japanese population.

Conclusion

Our study indicates that the PDCD1 haplotype is associated with a susceptibility to MM. The PDCD1 rs2227982 and PDCD1LG1 rs2297136 affect the clinical features of multiple myeloma patients. In addition, there is no effect of SNPs in CTLA4. However, there are limitations to the interpretation of the results in this study because the sample size was relatively small with a total of 124 MM patients. Therefore, further investigations with larger sample sizes are needed to corroborate our results.