Haptoglobin promoter polymorphism rs5472 as a prognostic biomarker for peptide vaccine efficacy in castration-resistant prostate cancer patients

Abstract

Personalized peptide vaccination (PPV) is an attractive approach to cancer immunotherapy with strong immune-boosting effects conferring significant clinical benefit. However, as with most therapeutic agents, there is a difference in clinical efficacy among patients receiving PPV. Therefore, a useful biomarker is urgently needed for prognosticating clinical outcomes to preselect patients who would benefit the most from PPV. In this retrospective study, to detect a molecular prognosticator of clinical outcomes for PPV, we analyzed whole-genome gene expression profiles of peripheral blood mononuclear cells (PBMCs) in castration-resistant prostate cancer (CRPC) patients before administration of PPV. Cox regression analysis revealed that mRNA expression of myeloperoxidase, haptoglobin, and neutrophil elastase was significantly associated with overall survival (OS) among vaccinated CRPC patients (adjusted P < 0.01). By promoter sequence analysis of these three genes, we found that rs5472 of haptoglobin (HP), an acute-phase plasma glycoprotein, was strongly correlated to OS of vaccinated CRPC patients (P = 0.0047, hazard ratio 0.47; 95 % confidence interval 0.28–0.80). Furthermore, both HP mRNA expression in PBMCs and protein level in plasma of CRPC patients before administration of PPV exhibited rs5472 dependence (P < 0.001 for mRNA expression and P < 0.05 for protein level). Our findings suggest that rs5472 may play an important role in the immune response to PPV via regulation of HP. Thus, we concluded that rs5472 is a potential prognostic biomarker for PPV.

Introduction

Immunotherapeutic vaccination is currently recognized as the fourth option in cancer therapy following surgery, chemotherapy, and radiation therapy. We have developed a novel immunotherapeutic approach, “personalized peptide vaccination (PPV),” in which a maximum of four human leukocyte antigen (HLA) class IA-restricted peptides from a pool are selected for vaccination on the basis of both HLA class IA type and preexisting host immunity before vaccination [1, 2]. In clinical studies of PPV carried out over the past decade, we have demonstrated promising results in multiple advanced cancers, such as malignant glioma, glioblastoma multiforme, colorectal, pancreatic, and gastric cancers [1–5]. However, as with most therapeutic agencies, there is a difference in clinical efficacy and immune response among cancer patients receiving PPV. Therefore, prognostic biomarkers are urgently needed to select patients who would be benefit most from PPV during patient treatment selection [1].

Ongoing research has produced a number of reported pretreatment biomarkers for cancer vaccinations. Harrop et al. [6] analyzed blood biochemical and hematological factors to identify pretreatment markers predicting benefit of the cancer vaccine MVA-5T4. Blood is well recognized as a surrogate biopsy material with easy sampling and may reflect physiological, pathological, and immune events occurring in different tissues of the body [7–9]. Although this report contained interesting observations of the relationship between blood factors and the immune response to this cancer vaccine, the approach was not comprehensive and investigated a limited number of blood factors. There have been more comprehensive studies attempting to identify biomarkers of clinical outcome in cancer immunotherapy, such as gene expression profile analyses, but these focused on tumors [9]. More broadly tuned than tumor gene expression profiles, blood profiles hold promise as potential sources of the genomic signatures predictive of the response to immunotherapy [9].

To identify novel blood prognostic biomarkers for PPV, we had previously analyzed the gene expression profiles of peripheral blood mononuclear cells (PBMCs) in patients with castration-resistant prostate cancer (CRPC) before administration of PPV [10]. In that initial study, CRPC patients were administered 3 or 4 right peptides selected from the 26 kinds of peptide candidates in consideration of the preexisting host immunity before vaccination, assessed by titers of immunoglobulin (Ig)G specific to each candidate. Then, we compared gene expression profiles of PBMC in CRPC patients, who survived for >900 days (long-term survivors, n = 20) or died within 300 days (short-term survivors, n = 20) after treatment with PPV. Nineteen genes known to be preferentially expressed in granulocytes were identified as differentially expressed between long-term and short-term survivors, but that study was of a limited number of patients. Here, to improve statistical power for detection and identify genetic variations of a prognostic biomarker, we re-analyzed gene expression profiles with an expanded sample size of 112 CRPC patients including the initial 40 in our previous study and performed promoter sequence analysis. Haptoglobin (HP), an acute-phase plasma glycoprotein synthesized mainly in the liver, was identified as one of the prognostic gene signatures. HP mRNA expression of PBMC before vaccination was significantly lower in CRPC patients with good vaccine response than those with poor response. This gene was identified in our previous study and quantitatively validated in the small number of patients [10, 11]. In humans, the HP locus is polymorphic, with two codominant alleles referred to as HP1 and HP2 that yield three distinct phenotypes (HP1-1, HP2-1, and HP2-2). There are many reports of the association of HP phenotype with various diseases such as cancer, cardiovascular disease, infection disease, and neurological disease [12], and with the immune response to hepatitis B vaccination and typhoid vaccination [13, 14], whereas there is little information about clinical importance and relevance of HP promoter polymorphism. In this study, we analyzed promoter sequence of HP in vaccinated CRPC patients and identified five promoter polymorphisms. We evaluated them as a prognostic biomarker of clinical benefit for PPV.

Materials and methods

Patients

PBMC or plasma samples were obtained from totaling 117 patients with metastatic CRPC who were positive for human leukocyte antigen (HLA)-A2, A24, A3 super type (A3, A11, A30, A31 and A33), or A26 and enrolled in phase I, I/II, and II clinical trials for PPV between February 2001 and April 2008. These studies were approved by the ethics review committee at the participating hospitals in Japan (Kurume University Hospital, Kinki University Hospital, Okayama University Hospital, and Nara Medical University Hospital). Since this study is retrospective design, PBMC or plasma samples were not entirely available for all three types of analysis (microarray, protein abundance, and promoter polymorphism) due to the unavailability of samples or low quality data. Finally, 112, 61, and 73 samples were analyzed by microarray, protein abundance, and promoter polymorphisms analysis, respectively (58 samples were analyzed by all three types of analysis). The detailed clinical trial methods have been described in our previous reports [5, 10].

Blood samples

PBMCs and plasma were used for measurement of gene expression profiles and soluble factors, respectively. Since this was a retrospective study with limited availability of patient samples, analysis of both PBMCs and plasma was not performed for all patients (i.e., promoter polymorphism and protein-level analysis).

RNA isolation from PBMCs

PBMCs were prepared from 20 mL of peripheral blood by density gradient centrifugation using Ficoll-Paque (GE Healthcare Life Sciences, Uppsala, Sweden). All samples were cryopreserved until RNA extraction. Total RNA was isolated using TRIzol LS reagent (Invitrogen, Carlsbad, Calif) and purified using RNeasy Mini Kit (Qiagen, Valencia, Calif), according to the manufacturer’s instructions. Quality and integrity of the purified total RNA were confirmed using an Agilent 2100 bioanalyzer (Agilent Technologies, Palo Alto, Calif) and Nanodrop ND-1000 (Thermo Fisher Scientific, Wilmington, Del). Prevaccination PBMCs in 112 selected patients were analyzed by DNA microarray.

Microarray experiment and analysis

RNAs were amplified, labeled, and hybridized by Illumina HumanWG-6 v3.0 Expression BeadChip (Illumina, San Diego, Calif) according to the manufacturer’s instructions. Microarray raw data were extracted by using BeadStudio v3.0 software (Illumina). In the preprocessing procedure, raw data were normalized by neqc [15], and its batch effect (day of experiment) was adjusted by ComBat [16], as implemented in the limma and sva Bioconductor (http://www.bioconductor.org) package, respectively. Low expression probes that might represent transcriptional noise and exhibit insufficient reliability were filtered out. In this study, we removed any probe with a detection P value computed by BeadStudio of >0.01 for >50 % of the all experimental samples. Accordingly, 11,412 out of 48,803 probes remained and were subsequently analyzed. An updated Illumina probe set annotation by ReMOAT was used [17]. Gene set analysis was performed using GeneSetDB which is a comprehensive meta-database, statistical, and visualization framework for gene set analysis [18]. An adjusted P value of 0.05 was set to be the level of significant gene sets with the additional requirement of at least three genes of annotated biological functions. All microarray data were deposited in the National Center for Biotechnology Information’s Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) with GEO series accession number GSE53922.

Promoter polymorphism analysis

Genomic DNA was extracted from PBMCs using TRIzol (Invitrogen, Carlsbad, Calif) and purified by ethanol precipitation. The promoter region from the transcriptional start site (+1) to approximately −1000 bp upstream of each gene was amplified using the primers 5′-TCAGTGTCACCATGATTATCCA-3′ (HP forward), 5′-GATTTAACACACTAAGCCCTTTGG-3′ (HP reverse), 5′-TCACTGAAGTTCTCCTGAAGAGG-3′ (MPO forward), 5′-AGTCATCCTGTGGGAAGAGC-3′ (MPO reverse), and 5′-TTCTCTGGTGCACCTTCAAC-3′ (ELANE forward), 5′-TGCTCAAGTTCGTTGTCTTCA-3′ (ELANE reverse). For sequencing, primary PCR products were cloned by TA cloning and sequenced using M13 standard primers. The cloned PCR products isolated by TA cloning were treated with ExoSAP-IT (Amersham Pharmacia Biotech, Piscataway, NJ, USA) to degrade single-strand DNA and deoxyribonucleotide triphosphate. All sequences were analyzed using an Applied Biosystems 3130 sequencer (PE Applied Biosystems). Promoter sequences were analyzed in 73 selected patients for HP and 32 selected patients for other genes as preliminary analysis. HP promoter polymorphism rs5472 was also confirmed by TaqMan assay.

Determination of HP levels in plasma samples

The level of HP in plasma samples from 61 patients was determined by the sandwich ELISA as described by Shimada et al. [19] with slight modifications; i.e., the detection antibody (AbD Serotec, Oxford, UK), blocking reagent (Blocking One, Nacalai Tesque, Kyoto, Japan), and detection reagent for peroxidase enzyme activity (ELISA POD substrate TMB kit, Nacalai Tesque) were different from those in the original method.

Statistical analysis

A univariate Cox proportional hazard model was applied to assess genes whose individual mRNA expression was associated with OS. The calculated P value was adjusted by multiple testing correction using the Benjamini and Hochberg method [20]. An adjusted P value of 0.01 was set to be the level of significance for further promoter polymorphism analysis. The relationships between clinical outcome (OS) and genotype based on promoter polymorphisms were analyzed by Fisher’s exact test, Kaplan–Meier survival curves with log-rank test, and Cox proportional hazard model. Statistical tests carried out during HP mRNA expression and protein level analysis used t test. All statistical analyses including the microarray preprocessing were performed by R software version 3.0.1 (http://www.r-project.org).

Results

Differentially expressed genes in CRPC patients associated with prognosis of PPV

Gene expression profiles of PBMCs were obtained from 112 CRPC patients before administration of PPV. We identified 35 genes as prognostic gene signatures whose expression was significantly associated with OS at the level of P < 1.0E−04 (adjusted P = 0.03) using a univariate Cox proportional hazard model (Table 1). Among 35 genes, 17 genes and 18 genes were overexpressed and underexpressed in vaccinated CRPC patients with favorable clinical outcome (OS longer than or equal to 365 days), respectively, or vice versa (Fig. 1).

Table 1.

Signature genes whose mRNA expression highly associated with OS

Gene symbol	Gene name	P	Adjusted P
MPO	Myeloperoxidase	1.6E−07	1.7E−03
HP	Haptoglobin	3.0E−07	1.7E−03
ELANE	Elastase, neutrophil expressed	1.9E−06	7.4E−03
BX114951	BX114951	3.8E−06	1.1E−02
CTSG	Cathepsin G	5.9E−06	1.4E−02
PSIP1	PC4 and SFRS1 interacting protein 1	1.9E−05	2.5E−02
WWP1	WW domain containing E3 ubiquitin protein ligase 1	1.9E−05	2.5E−02
AMY2A	Amylase, alpha 2A (pancreatic)	2.0E−05	2.5E−02
C3orf17	Chromosome 3 open reading frame 17	2.1E−05	2.5E−02
MAN2A1	Mannosidase, alpha, class 2A, member 1	2.2E−05	2.5E−02
HBQ1	Hemoglobin, theta 1	2.8E−05	2.5E−02
DEFA4	Defensin, alpha 4, corticostatin	3.1E−05	2.5E−02
AHSP	Alpha hemoglobin stabilizing protein	3.4E−05	2.5E−02
STRBP	Spermatid perinuclear RNA binding protein	3.8E−05	2.5E−02
HBM	Hemoglobin, mu	4.1E−05	2.5E−02
SP4	Sp4 transcription factor	4.2E−05	2.5E−02
HBD	Hemoglobin, delta	4.6E−05	2.5E−02
GARNL1	GTPase-activating Rap/Ran-GAP domain-like 1	4.6E−05	2.5E−02
RNF219	Ring finger protein 219	4.8E−05	2.5E−02
NR3C2	Nuclear receptor subfamily 3, group C, member 2	5.0E−05	2.5E−02
ZZZ3	Zinc finger, ZZ-type containing 3	5.1E−05	2.5E−02
MARCH2	Membrane-associated ring finger (C3HC4) 2	5.2E−05	2.5E−02
KIAA0947	KIAA0947 protein	5.3E−05	2.5E−02
ABLIM1	Actin binding LIM protein 1	5.4E−05	2.5E−02
CMPK1	Cytidine monophosphate kinase 1, cytosolic	5.7E−05	2.6E−02
SLC7A6	Solute carrier family 7, member 6	5.9E−05	2.6E−02
LCN2	Lipocalin 2	6.4E−05	2.7E−02
BPI	Bactericidal/permeability-increasing protein	7.6E−05	3.1E−02
NR2C1	Nuclear receptor subfamily 2, group C, member 1	8.4E−05	3.1E−02
TSPAN5	Tetraspanin 5	8.6E−05	3.1E−02
GRPEL2	GrpE-like 2, mitochondrial (E. coli)	9.2E−05	3.1E−02
TMOD1	Tropomodulin 1	9.2E−05	3.1E−02
PMAIP1	Phorbol-12-myristate-13-acetate-induced protein 1	9.3E−05	3.1E−02
SERF2	Small EDRK-rich factor 2	9.3E−05	3.1E−02
CA1	Carbonic anhydrase I	9.4E−05	3.1E−02

Fig. 1.

Hierarchical clustering of 35 signature genes in 112 CRPC patients receiving PPV. The main color scheme shows Z score of signal intensity of each probe. Blue and orange bars in the column dendrogram show patients with OS ≥ 365 days and OS <365 days, respectively

To functionally characterize these prognostic gene signatures, GeneSetDB was performed using Gene Ontology, pathway databases, and disease/phenotype databases. There were 19 significant gene sets at the level of adjusted P < 0.05 (Fig. 2): Oxygen transport-related gene sets such as hemoglobin complex (GO:0005833), oxygen transporter activity (GO:0005344), oxygen binding (GO:0019825), heme binding (GO:0020037) including AHSP, HBD, HBM, and HBQ1; defense response-related gene sets such as negative regulation of growth of symbiont in host (GO:0044130), defense response to fungus (GO:0050832), and increased susceptibility to fungal infection including CTSG, DEFA4, ELANE, and MPO; immune system-related gene sets such as immune system disorder and anaphylactoid reaction including ELANE, HP, and MPO; and hematopoietic-related gene sets such as abnormal myelopoiesis, abnormal erythrocyte morphology, anemia, heparin binding (GO:0008201) including AHSP, CTSG, ELANE, LCN2, MAN2A1, and TMOD1 (Fig. 2). Interestingly, ELANE, HP, and MPO, which involve the immune system, were the top three highly ranked genes based on Cox proportional analysis with adjusted P of <0.01 (Table 1). These genes are less expressed in CRPC patients with good prognosis than patients with poor prognosis before administration of PPV (Fig. 1).

Fig. 2.

Functional analysis of 35 genes (listed in Table 1) using GeneSetDB (adjusted P value <0.05). Bars represent the proportion of genes involved in each category for which statistical significance and corresponding gene names are shown in brackets

Promoter polymorphism of the MPO, HP, and ELANE

We further focused on three genes (MPO, HP, and ELANE) and sequenced their promoter regions to investigate whether there is a promoter polymorphism in association with OS of vaccinated CRPC patients. First, we performed a preliminary evaluation with a small number of CRPC patients receiving PPV: 17 patients with OS more than 900 days (long-term survivors) and 15 patients with OS < 300 days (short-term survivors). Table 2 shows the frequency of promoter polymorphism of each gene in long-term and short-term survivors. There were two and five promoter polymorphisms (totally seven polymorphisms) in the promoter regions of MPO and HP, respectively, but there was none in the ELANE promoter region. Six out of seven polymorphisms are registered in dbSNP (http://www.ncbi.nlm.nih.gov/SNP/), but the other is not, and may be a novel polymorphism or rare variation. Among the promoter polymorphisms listed in Table 2, Fisher’s exact test revealed that only the HP promoter polymorphism −55A>G (dbSNP ID: rs5472) bears a statistically significant relationship to clinical outcome in CRPC patients receiving PPV (P < 0.001).

Table 2.

Promoter polymorphism frequency of haptoglobin (HP) and myeloperoxidase (MPO)

Gene	Polymorphism	Genotype	Short-term survivors (n = 17)	Long-term survivors (n = 15)	P	dbSNP ID
HP	−242C>T	CC	15	13	1	rs5467
		CT	2	2
HP	−233G>C	GG	16	15	1	na
		GC	1	0
HP	−191T>G	TT	16	14	1	rs5468
		TG	1	1
HP	−61A>C	AA	15	15	0.4895	rs5471
		AC	2	0
HP	−55A>G	AA	6	0	<0.001	rs5472
		AG	11	9
		GG	0	6
MPO	−784G>A	GG	11	13	0.3	rs34565045
		GA	5	2
		AA	1	0
MPO	−748A>T	AA	15	13	1	rs8079006
		AT	2	2
		TT	0	0

rs5472 is predominant in all populations [12, 21]. There are three genotypes arising from the −55A>G locus (A/A, A/G, and G/G). In order to confirm a preliminary evaluation as the prognostic utility of rs5472, we additionally sequenced HP promoter region of 73 vaccinated CRPC patients in total. Figure 3 shows Kaplan–Meier survival curves of 73 CRPC patients with respect to the three genotypes. The A/G or G/G genotypes had significant positive impact on OS of vaccinated CRPC patients compared to the A/A genotype with P = 0.0047 (hazard ratio 0.47; 95 % confidence interval 0.28–0.80).

Fig. 3.

Kaplan–Meier survival curves for CRPC patients stratified by rs5472 genotypes. Yellow line shows patients with A/G (n = 30) or G/G genotypes (n = 13); blue line shows patients with A/A genotype (n = 30). P = 0.0047

Association of rs5472 with HP mRNA expression in PBMC and protein level in plasma

We next investigated whether rs5472 modulates HP mRNA expression in PBMC and/or protein level in plasma in vaccinated CRPC patients before administration of PPV. We observed HP mRNA expression in PBMC from microarray data and protein levels in plasma from ELISA assay. Microarray data show that mRNA expression of HP was significantly differentially expressed among these three genotypes. As shown in Fig. 4a, it was preferentially high in CRPC patients with a rank order of A/A, A/G, and G/G genotype (P = 0.0023 in A/A vs A/G, P = 2.0E−08 in A/A vs G/G, P = 3.4E−04 in A/G vs G/G). A similar result was obtained for protein level from ELISA assay. HP levels in plasma from CRPC patients with G/G genotype were lower than those from patients with A/A (P = 0. 0086) and A/G genotype (P = 0.011), but they were not significantly different between CRPC patients with A/A and A/G genotype (P = 0.78, Fig. 4b).

Fig. 4.

Association of three genotypes stratified by rs5472 with HP mRNA expression in PBMC and protein level in plasma. a Bar plots of mRNA expression from microarray in patients with A/A (n = 28, mean 7.4, SE 0.2), A/G (n = 28, mean 6.6, SE 0.2), G/G (n = 13, mean 5.5, SE 0.2) genotype, respectively. b Bar plots of protein level by ELISA in patients with A/A (n = 27, mean 252.9, SE 21.9), A/G (n = 23, mean 238.9, SE 33.3), G/G (n = 11, mean 131.6, SE 0.2), respectively. Bar plots represent mean and SE. ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05

Discussion

In this study, we identified and ranked 35 genes whose individual mRNA expression in PBMCs is significantly associated with clinical outcome of CRPC patients receiving PPV from microarray analysis. We also found that the only promoter polymorphism −55A>G locus (rs5472) located in HP, which is the second-ranked gene in microarray analysis, was significantly correlated to clinical outcome for PPV. Interestingly, our data demonstrate that HP level in blood exhibited rs5472 dependence. Among three genotypes, A/A, A/G, and G/G, arising from rs5472, vaccinated CRPC patients with the AA genotype have significantly lower OS rates and show higher expression levels of HP than patients with the AG or GG genotypes. These results suggest that rs5472 may have the potential role of HP elevation in peripheral blood playing an important role in the immune response to PPV and rs5472 could be a possible prognosis biomarker for PPV. Recently, our group has demonstrated that rs5472 has significant impacts on serum HP level in a Japanese population [22]. In Japanese, subjects with A/A genotype at rs5472 have higher serum HP levels than subjects with A/G or G/G genotype. Taken together with current study, rs5472 can be considered as a key determinant of serum HP levels.

HP has a wide variety of functions in biological systems, such as binding to free hemoglobin to prevent iron loss and kidney damage during hemolysis, protection against oxidative damage/toxic radical, inhibition of nitric oxide, angiogenesis, and others [12]. To date, several studies have reported that HP level in peripheral blood is elevated in cancer patients including gastric, liver, non-small cell lung, and pancreatic cancer compared to benign tumors or normal subjects [23–26]. Importantly, in epithelial ovarian cancer, the levels of serum HP were significantly correlated to tumor staging/grading according to Federation of Gynecology and Obstetrics (FIGO) staging, and patients with higher HP concentrations had a significantly worse survival probability [27]. Our data demonstrated in a manner similar to these previous reports that HP mRNA expression in CRPC patients was increased in comparison with normal subjects (P < 0.01, data not shown) and vaccinated CRPC patients with poor prognosis show higher HP mRNA expression.

Interestingly, Elpek et al. [28] recently reported that HP mRNA and protein are highly expressed in mouse myeloid-derived suppressor cells (MDSC). In addition, the authors showed the correlation between serum HP levels and the frequency of Gr1 (granulocyte differentiation antigen 1) ^hi cells subsets of MDSC in blood, suggesting its potentiality as a diagnostic biomarker for tumors characterized by the accumulation of myeloid cell. Surprisingly, mRNA of ELANE and MPO, which are two of top three highly ranked genes in this study, also highly expressed in mouse MDSC. These genes are related to granulocyte differentiation. MDSCs are immunosuppressive cells promoting tumor immune evasion and tumor progression and are one of the major factors down-regulating the efficacy of cancer immune therapy [29]. Recent clinical study reported that the depletion of MDSC improved the immune response to cancer vaccine in patients with small cell lung cancer [30]. Therefore, it is possible that HP and other molecules derived from MDSCs in PBMC influence the prognosis of PPV. We need to further characterize the functional roles of HP at the molecular level in future experiments to confirm this hypothesis.

In terms of genetic polymorphism of HP, it is widely known that there is an association between HP phenotypes and disease susceptibility and clinical prognosis in diabetes, atherosclerosis, and cardiovascular diseases [12]. The phenotype of HP in cancer is also well studied that it is a risk factor for cancer development and is associated with prognosis in glioblastoma, gastric, pancreatic, ovarian cancer, and so on [31]. However, to our knowledge, there are three reports that HP promoter polymorphisms are associated with disease susceptibility and clinical prognosis. All of them were reported for African. HP promoter polymorphisms −61A>C (rs5471) and/or −101C>G (rs5470) are associated with malaria protection and risk of active trachoma in Gambian children and ahaptoglobinaemia and hypohaptoglobinaemia in Ghana [32–34]. No reports have described the evaluation of rs5472 in the context of disease susceptibility and clinical prognosis. Therefore, the present study is the first report of the association of rs5472 with a specific clinical prognosis.

In conclusion, using gene expression profiles and promoter sequence analyses of PBMCs in CRPC patients before administration of PPV, we observed that HP mRNA expression and its promoter polymorphism rs5472 are significantly associated with OS of vaccinated CRPC patients. To further elucidate our findings, it is necessary to conduct larger-scale cohort analysis. Moreover, since this is the retrospective study lacking control samples, who were not administrated with PPV, our conclusion is confined to CRPC patients receiving PPV. Therefore, the possibility that rs5472 may not only involve the PPV efficacy in CRPC patients, but also generally involve the prognosis for CRPC patients still remains. In the future, we will further investigate the potential applicability of rs5472 as a prognostic biomarker for PPV in other cancers, other immunotherapy, chemotherapy, and prognosis of diseases.

Abbreviations

CRPC

Castration-resistant prostate cancer

HP

Haptoglobin

PPV

Personalized peptide vaccination

Compliance with ethical standards

Conflict of interest

The authors have declared no conflicts of interest.

Funding

Kyogo Itoh received research fund from Taiho Pharmaceutical Co., Ltd.