Effect of HLA genotype on intravesical recurrence after bacillus Calmette–Guérin therapy for non-muscle-invasive bladder cancer

Abstract

The intravesical administration of bacillus Calmette–Guérin (BCG) is widely used to control the intravesical recurrence of non-muscle-invasive bladder cancer (NMIBC). This study aimed to reveal the effects of zygosity on human leukocyte antigen (HLA) genes and individual HLA genotypes on intravesical recurrence after intravesical BCG therapy for NMIBC. This study included Japanese patients who had received intravesical BCG for NMIBC. HLA genotyping of HLA-A, B, C, and DRB1 was performed. The effect of HLA zygosity and HLA genotype on intravesical recurrence was evaluated. Among 195 patients, those homozygous for the HLA-B supertype were more likely than those heterozygous for the HLA-B supertype to experience intravesical recurrence by univariate analysis (hazard ratio [HR], 95% confidence interval [CI]; 1.87, 1.14–3.05, P = 0.012) and multivariate analysis (HR, 95% CI; 2.26, 1.02–5.01, P = 0.045). Patients with B07 or B44 had a decreased risk of intravesical recurrence by univariate analysis (HR, 95% CI; 0.43, 0.24–0.78, P = 0.0056) and multivariate analysis (HR, 95% CI; 0.36, 0.16–0.82, P = 0.016). This study suggests the importance of the diversity and specificity of HLA-B loci in the antitumor effect of BCG immunotherapy for NMIBC. These findings may contribute to the delineation of risk strata for BCG therapy and improve the medical management of NMIBC.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00262-021-03032-0.

Introduction

About 70% of bladder cancers are non-muscle-invasive bladder cancer (NMIBC) [1, 2]. NMIBC is usually treated by transurethral resection (TUR) followed by intravesical chemotherapeutic agents or bacillus Calmette–Guérin (BCG) dependent on risk classification based on clinicopathological parameters [1, 2]. Although intravesical BCG therapy has excellent therapeutic efficacy, approximately 50% of NMIBC cases experience intravesical recurrence [3, 4]. Clinicopathological parameters including age, gender, history of bladder cancer, number of tumors, tumor size, tumor grade, tumor stage, and presence of carcinoma in situ (CIS) are known predictors of intravesical recurrence after BCG therapy [5, 6]. Several risk calculators including the European Organization for Research and Treatment of Cancer (EORTC) risk table and Spanish Urological Club for Oncological Treatment (CUETO) scoring model have been developed that incorporate these risk factors [2, 7]. However, the predictive power of these models, which are based solely on clinicopathological parameters, is not sufficient, and there is a need to identify new prognostic biomarkers.

Human leukocyte antigens (HLA) are expressed on a variety of cells, including cancer cells and immune cells. HLA promote a series of signals that activate immune responses by presenting processed antigens to T lymphocytes [8]. HLA class I (HLA-I) is expressed on many cell types and is recognized by CD8⁺ T cells whereas HLA class II (HLA-II) is expressed mainly on immune cells such as antigen-presenting cells and is recognized by CD4⁺ helper T cells and regulatory T cells [8]. HLA is characterized by a high degree of polygenicity and polymorphism, and is classified into subclasses including serological specificities and alleles according to its characteristics. In addition, supertypes have been defined for HLA-I especially for HLA-A and HLA-B loci according to their peptide-binding repertoires [9]. Polymorphisms at these HLA loci result in different peptide affinities and allow for the presentation of a diverse peptide repertoire. Recently, it was reported that HLA gene zygosity, divergence, and individual HLA genotypes were associated with the prognosis of checkpoint blockade immunotherapy [10–14]. Although the antitumor mechanism of BCG is poorly understood, BCG infection of urothelial cells and bladder cancer cells, induction of immune responses via T cells, macrophages and cytokines, and the development of antitumor effects are important steps in intravesical BCG therapy [5, 15]. Therefore, the host immune response associated with various factors including HLA genotype may affect the anticancer effect of intravesical BCG on NMIBC. Accordingly, we investigated whether HLA gene zygosity and individual HLA genotypes were associated with intravesical recurrence after BCG therapy for NMIBC.

Patients and methods

Patients and treatment

Japanese patients (n = 195) who received intravesical BCG therapy for Ta or T1 NMIBC at Akita University Hospital (Akita, Japan), Yamaguchi University Hospital (Ube, Japan), University of Occupational and Environmental Health (Kitakyushu, Japan), and Kyushu University Hospital (Fukuoka, Japan) between 1990 and 2019 were included as previously described [16, 17]. Patients who had received BCG therapy for pure CIS were excluded. Written informed consent was obtained from all patients, and the study was conducted in accordance with the principles described in the Declaration of Helsinki and the ethical guidelines for epidemiological studies established by the Japanese government. The review committee of each institution approved this study.

Pathological evaluation of bladder cancer was based on the 1976 World Health Organization grading system [18]; T-stage was determined according to the unified TNM criteria based on the pathological results of TUR [19]. Of all patients treated primarily with TUR, 53 underwent a second TUR. BCG (Tokyo 172 strain [n = 153] or Connaught strain [n = 42]) was dissolved in 40 ml of saline and administered weekly for 6–10 cycles, as determined by the physician based on adverse events or patient refusal. Maintenance therapy with BCG was not administered because these were not standard when patients were enrolled or because physicians did not recommend. Patients were followed up periodically after TUR with interview, urinalysis, urine cytology, and cystoscopy. The time to intravesical recurrence was defined as the time from the date of the last pathological diagnosis of bladder cancer to the date of pathological diagnosis of intravesical recurrence.

Serum sample collection and HLA typing

Genomic DNA was extracted from whole blood samples and stored at − 80 °C prior to analysis. Genotyping of 177 samples for HLA-I (A, B, C) and HLA-II (DRB1) alleles was performed by polymerase chain reaction with a sequence-specific oligonucleotide probe using WAKFlow® HLA typing kits (Wakunaga Pharmaceutical Co., Hiroshima, Japan) and LABType™ SSO (One Lambda, Inc., Canoga Park, CA, USA) as described previously [20]. The HLA-I (A, B, C) and HLA-II (DRB1) genotypes of the remaining 18 samples were determined by imputation as described previously using Japonica Array version 2 (Toshiba, Tokyo, Japan) [21]. Supertypes of HLA-A and HLA-B loci were classified according to a method previously described [9].

Calculation of patient HLA evolutionary divergence (HED)

The HED was calculated as previously described [22]. Divergence between allele sequences was calculated using the Grantham distance, which is a quantitative one-to-one distance that accounts for the physicochemical properties of amino acids and therefore the functional similarity between sequences [23]. For a particular HLA-I or HLA-DRB1, the peptide-binding domain sequences of each allele were aligned [24] and the Grantham distance was calculated as the sum of amino acid differences (accounting for the biochemical composition, polarity, and volume of each amino acid) along the sequence of the peptide-binding domain according to the formula reported by Grantham et al. [23].

$$\text{Grantham}\text{distance}=\mathrm{\Sigma }{D}_{\mathit{ij}}=\mathrm{\Sigma }{\left[\alpha {\left(c_i-c_j\right)}^2+\beta {\left(p_i-p_j\right)}^2+\gamma {\left(v_i-v_j\right)}^2\right]}^{1/2}$$ 1

where i and j are two homologous amino acids at a given position in the alignment and D is the Grantham distance between them. c, p, and v represent the composition, polarity, and volume of the amino acids, respectively. α, β, and γ are constants, and all values were taken from the original study [25]. The final Grantham distance was calculated by normalizing the value in Eq. (1) by the length of the alignment between the peptide-binding domains of the two alleles of a particular HLA-I or HLA-DRB1 genotype.

Statistical analyses

All statistical analyses were performed using JMP14 software (SAS Institute, Cary, NC, USA). Survival analysis was performed using the Kaplan–Meier method and the log-rank test. Univariate and multivariate analyses were performed using the Cox hazard proportional model, and hazard ratios (HRs) were estimated with 95% confidence intervals (CIs). Harrell’s C-index was calculated using Stata version17 (College Station, TX, USA) as described previously [26, 27]. P values were all two-sided, and a P value < 0.05 was considered statistically significant.

Results

Homozygosity of HLA genotypes

In this study, 195 Japanese patients who received intravesical BCG for NMIBC were included, and their clinicopathological characteristics according to the classification in the CUETO scoring model are shown in Table 1. The median follow-up of men without intravesical recurrence at the time of censoring was 35.7 months (interquartile range 16.9–66.6 months). During the observation period, 75 patients (38.5%) experienced intravesical recurrence. The median survival without intravesical recurrence was 80.2 months (95% CI, 59.0–135.0 months).

Table 1.

Patient characteristics

Variable	n = 195
Age, years, n (%)
< 60	26 (13.4%)
60–70	79 (40.7%)
> 70	89 (45.9%)
NA	1
Gender, n (%)
Male	159 (81.5%)
Female	36 (18.5%)
History of recurrence, n (%)
Primary	111 (72.1%)
Recurrent	43 (27.9%)
NA	41
Tumor number, n (%)
≤ 3	117 (83.0%)
> 3	24 (17.0%)
NA	54
Grade, n (%)
Grade 1	10 (5.2%)
Grade 2	44 (22.7%)
Grade 3	140 (72.2%)
NA	1
T category, n (%)
Ta	88 (45.1%)
T1	107 (54.9%)
Carcinoma in situ, n (%)
Absent	115 (59.0%)
Present	80 (41.0%)

NA not available

In this cohort, we investigated the association between HLA gene homozygosity and the risk of intravesical recurrence. When HLA gene allele and serological specificity homozygosity were examined, no genotypes were associated with intravesical recurrence in univariate analysis (Supplementary Table 1). Similarly, no genotypes were associated with intravesical recurrence in multivariate analysis when adjusted for the following parameters: age, gender, number of tumors, tumor size, history of bladder cancer, tumor stage, tumor grade, and presence of CIS (Supplementary Table 1). We then examined the association between HLA supertype homozygosity and intravesical recurrence and found that patients homozygous for the HLA-B supertype had a higher risk of intravesical recurrence than those heterozygous for the HLA-B supertype in the univariate analysis (HR, 95% CI; 1.87, 1.14–3.05, P = 0.012) and multivariate analysis (HR, 95% CI; 2.26, 1.02–5.01, P = 0.045) (Table 2). However, homozygosity for HLA-A supertype was not associated with the risk of intravesical recurrence in the univariate (HR, 95% CI; 0.65, 0.37–1.13, P = 0.13) and multivariate (HR, 95% CI; 0.87, 0.41–1.83, P = 0.71) analyses (Table 2). Similarly, the Kaplan–Meier curve showed inferior intravesical recurrence-free survival in patients homozygous for the supertype of the HLA-B locus, whereas intravesical recurrence-free survival was similar for the homozygous and heterozygous supertypes of the HLA-A locus (Fig. 1). When the association between HED at the HLA-I or HLA-DRB1 locus and intravesical recurrence was analyzed, only low HED at the HLA-B locus was associated with inferior recurrence-free survival (Fig. 2).

Table 2.

Associations of HLA supertype homozygosity with recurrence after intravesical bacillus Calmette–Guérin therapy

		Univariate			Multivariate††
	Frequency, n (%)	HR†	95% CI	P value	HR†	95% CI	P value
Homozygosity of HLA-A supertype
All	45 (23.8%)	0.65	0.37–1.13	0.13	0.87	0.41–1.83	0.71
A01	2 (1.1%)	0.65	0.086–4.84	0.67	1.97	0.24–16.6	0.53
A02	11 (5.8%)	0.96	0.41–2.25	0.92	1.10	0.23–5.35	0.90
A03	13 (6.9%)	0.94	0.38–2.33	0.89	0.95	0.31–2.85	0.92
A24	19 (10.1%)	0.34	0.12–0.95	0.039*	0.60	0.18–2.05	0.42
Homozygosity of HLA-B supertype
All	44 (23.3%)	1.87	1.14–3.05	0.012*	2.26	1.02–5.01	0.045*
B07	26 (13.8%)	1.57	0.85–2.90	0.15	1.57	0.56–4.40	0.39
B44	8 (4.2%)	1.49	0.46–4.80	0.50	2.03	0.41–9.95	0.38
B62	9 (4.8%)	3.80	1.78–8.10	0.0006*	5.73	1.82–18.0	0.0028*

^*Statistically significant. HR hazard ratio; CI confidence interval

^†Reference is a carrier of heterozygosity

^††Adjusted for age, gender, number of tumors, tumor size, prior history of bladder cancer, tumor stage, tumor grade, and presence of carcinoma in situ

Fig. 1.

Kaplan–Meier survival analysis of intravesical recurrence-free survival stratified by HLA supertype zygosity. A, B Intravesical recurrence-free survival stratified by HLA-A supertype (A) and HLA-B supertype (B) zygosity

Fig. 2.

Kaplan–Meier survival analysis of intravesical recurrence-free survival stratified by HLA evolutionary divergence (HED) at the HLA locus. A–D Intravesical recurrence-free survival stratified by HED at the HLA-A locus (A), HLA-B locus (B), HLA-C locus (C), and HLA-DRB1 locus (D). The threshold was set as the median: high > median; low ≤ median

Subsequently, homozygosity of a specific supertype at the HLA-A and HLA-B loci was analyzed. As shown in Table 4, patients homozygous for the HLA-A24 supertype at the HLA-A locus were associated with a decreased risk of intravesical recurrence in the univariate analysis (HR, 95% CI; 0.34, 0.12–0.95, P = 0.039), but not the multivariate analysis (HR, 95% CI; 0.60, 0.18–2.05, P = 0.42). Furthermore, homozygosity of HLA-B62 supertype at the HLA-B locus was associated with the increased risk of intravesical recurrence compared with heterozygosity in the univariate (HR, 95% CI; 3.80, 1.78–8.10, P = 0.0006) and multivariate analyses (HR, 95% CI; 5.73, 1.82–18.0, P = 0.0028) (Table 2).

Table 4.

Association between HLA supertype combination and recurrence after intravesical bacillus Calmette–Guérin therapy

		Univariate			Multivariate††
	Frequency, n (%)	HR†	95% CI	P value	HR†	95% CI	P value
Combination of HLA-A supertype (n = 189)
A01 or A01A03	45 (23.8%)	0.90	0.53–1.54	0.71	1.65	0.78–3.52	0.19
A01 or A02	118 (62.4%)	1.13	0.70–1.83	0.60	1.13	0.57–2.25	0.72
A01 or A03	124 (65.6%)	1.46	0.88–2.41	0.14	1.44	0.72–2.87	0.30
A01 or A24	132 (69.8%)	0.76	0.47–1.22	0.26	1.12	0.50–2.51	0.79
A01A03 or A02	85 (45.0%)	1.23	0.78–1.94	0.36	0.80	0.40–1.57	0.51
A01A03 or A03	97 (51.3%)	1.51	0.94–2.42	0.087	1.07	0.58–2.01	0.82
A01A03 or A24	108 (57.1%)	0.79	0.50–1.24	0.30	0.78	0.39–1.55	0.48
A02 or A03	149 (78.8%)	1.91	1.002–3.63	0.049*	0.95	0.41–2.19	0.91
A02 or A24	157 (83.1%)	1.19	0.64–2.21	0.58	0.84	0.39–1.81	0.65
A03 or A24	165 (87.3%)	0.86	0.47–1.57	0.62	0.71	0.29–1.71	0.44
Combination of HLA-B supertype (n = 189)
B07 or B27	136 (72.0%)	0.73	0.45–1.19	0.21	0.82	0.41–1.67	0.59
B07 or B44	164 (86.8%)	0.43	0.24–0.78	0.0056*	0.36	0.16–0.82	0.016*
B07 or B62	163 (86.2%)	1.00	0.52–1.90	0.99	0.67	0.26–1.75	0.41
B27 or B44	114 (60.3%)	0.92	0.58–1.46	0.71	0.98	0.49–2.00	0.97
B27 or B62	100 (52.9%)	0.90	0.57–1.42	0.64	0.88	0.45–1.71	0.70
B44 or B62	142 (75.1%)	0.80	0.48–1.33	0.39	0.80	0.36–1.79	0.59

^*Statistically significant. HR hazard ratio; CI confidence interval

^†Reference is a carrier of other genotypes

^††Adjusted for age, gender, number of tumors, tumor size, prior history of bladder cancer, tumor stage, tumor grade, and presence of carcinoma in situ

Individual HLA genotypes

Next, we examined the association between individual HLA genotypes and intravesical recurrence. HLA-C and HLA-DRB1 serological specificities were examined, and the Cw5 serological specificity of HLA-C was associated with an increased risk of intravesical recurrence in the univariate analysis (HR, 95% CI; 4.35, 1.04–18.1, P = 0.044), but not the multivariate analysis (HR, 95% CI; 2.54, 0.50–12.9, P = 0.26) (Supplementary Table 2). The DR7 serological specificity of HLA-DRB1 was associated with an increased risk of intravesical recurrence in the univariate analysis (HR, 95% CI; 5.47, 1.20–25.0, P = 0.029), but not the multivariate analysis (HR, 95% CI; 2.78, 0.87–8.87, P = 0.083) (Supplementary Table 2).

We then examined the association between HLA supertypes and intravesical recurrence and found that no single supertype was associated with a risk of intravesical recurrence (Table 3). Therefore, we examined the use of a combination of supertypes for each HLA locus. As shown in Table 4, patients with HLA-A02 or HLA-A03 were associated with an increased risk of intravesical recurrence in the univariate analysis (HR, 95% CI; 1.91, 1.002–3.63, P = 0.049), but not the multivariate analysis (HR, 95% CI; 0.95, 0.41–2.19, P = 0.91). Conversely, patients with HLA-B07 or HLA-B44 were associated with a decreased risk of intravesical recurrence in the univariate (HR, 95% CI; 0.43, 0.24–0.78, P = 0.0056) and multivariate analyses (HR, 95% CI; 0.36, 0.16–0.82, P = 0.016) (Table 4). Similarly, Kaplan–Meier curves showed inferior intravesical recurrence-free survival in patients with the HLA-A supertypes HLA-A02 or HLA-A03, but better intravesical recurrence-free survival in patients with the HLA-B supertypes HLA-B07 or HLA-B44 (Fig. 3).

Table 3.

Association between HLA supertype and recurrence after intravesical bacillus Calmette–Guérin therapy

		Univariate			Multivariate††
	Frequency, n (%)	HR†	95% CI	P value	HR†	95% CI	P value
HLA-A supertype (n = 189)
A01	43 (22.8%)	0.91	0.64–1.90	0.72	1.75	0.82–3.74	0.15
A01A03	2 (1.1%)	0.94	0.13–6.81	0.96	–	–	–
A02	84 (44.4%)	1.28	0.81–2.01	0.29	0.83	0.42–1.63	0.58
A03	96 (50.8%)	1.46	0.91–2.33	0.12	1.07	0.58–2.01	0.82
A24	106 (56.1%)	0.79	0.50–1.24	0.31	0.81	0.41–1.61	0.55
HLA-B supertype (n = 189)
B07	117 (61.9%)	0.73	0.86–2.18	0.18	0.70	0.36–1.38	0.31
B27	30 (15.9%)	1.00	0.54–1.86	1.00	1.39	0.56–3.45	0.48
B44	90 (47.6%)	0.83	0.52–1.32	0.42	0.79	0.40–1.57	0.50
B62	80 (42.3%)	0.96	0.60–1.54	0.88	0.85	0.44–1.65	0.64

HR hazard ratio; CI confidence interval

^†Reference is a carrier of other genotypes

^††Adjusted for age, gender, number of tumors, tumor size, prior history of bladder cancer, tumor stage, tumor grade, and presence of carcinoma in situ

Fig. 3.

Kaplan–Meier survival analysis of intravesical recurrence-free survival stratified by HLA supertype. A, B Intravesical recurrence-free survival stratified by HLA-A supertype (A) and HLA-B supertype (B)

Risk stratification by HLA-B genotype in combination with CUETO scoring model

Finally, we aimed to improve prognostic ability of CUETO scoring model by incorporating HLA-B genotype. As expected, CUETO scoring model could prognosticate intravesical recurrence after BCG treatment (Supplementary Fig. 1). The risk of tumor recurrence was higher in those with 7–9 (HR, 95% CI; 2.63, 1.04–6.64, P = 0.040) and 10–16 (HR, 95% CI; 3.12, 1.05–9.29, P = 0.041), but not 5–6 (HR, 95% CI; 1.26, 0.45–3.55, P = 0.66), compared to those with 0–4. Harrell’s C-index for intravesical recurrence by CUETO scoring model was 0.61 (95% CI, 0.52–0.69). When intravesical risk was classified into three risk categories using zygosity for the HLA-B supertype (heterozygosity vs homozygosity) or HLA-B supertype (HLA-B07 or HLA-B44 vs others) in combination with CUETO scoring model (0–6 vs 7–16), Harrell’s C-indices were improved to 0.62 (95% CI, 0.54–0.71) or 0.65 (95% CI, 0.57–0.73), respectively.

Treatment-emerged adverse events which resulted in discontinuation of BCG treatment were observed in 22 patients (11.3%). HLA zygosity or individual HLA genotype was not associated with treatment-emerged adverse events (data not shown).

Discussion

Previous studies have shown that a reduced expression of HLA-I and a loss of heterozygosity of the HLA-I gene are associated with reduced recurrence-free survival in BCG therapy and anti-PD-1 immunotherapy [28–31]. In the present study, we investigated the association between HLA genotype and intravesical recurrence after BCG therapy and found that homozygosity at the HLA-B locus and a specific combination of supertypes at the HLA-B locus were robustly associated with intravesical recurrence, which suggests the importance of the HLA-I genotype at the HLA-B locus in response to BCG therapy.

Divergent alleles have been proposed to be advantageous in immune responses to a variety of antigens [32]. In line with this idea, heterozygosity of the HLA-I gene was reported to be associated with an improved prognosis in checkpoint blockade immunotherapy [10, 13, 14]. Chowell et al. reported that heterozygosity at all HLA-I loci (A, B, and C) led to improved overall survival (OS) in patients with melanoma and non-small cell lung cancer treated with anti-CTLA-4 or anti-PD-1 therapy [10]. Similarly, heterozygosity for HLA-I, but not HLA-II, was associated with improved OS in non-small cell lung cancer patients following anti-PD-1/PD-L1 therapy [14]. Furthermore, it was recently reported that HED at the HLA-A and HLA-B loci correlated with the beneficial effects of cancer immunotherapy [11]. Similar to these reports, we found that heterozygosity and high HED at the HLA-B locus were associated with reduced intravesical recurrence after immunotherapy with BCG, which suggests an enhanced anticancer effect of BCG therapy. However, there was no association between the zygosity of other HLA loci and intravesical recurrence after BCG treatment. This differential effect of distinct HLA loci may be explained by previous reports that HLA-B is expressed at higher levels on the cell surface, is more divergent, and binds to a greater variety of peptides compared with HLA-A and HLA-C [8, 11]. Thus, homozygosity for HLA-B may reduce its affinity for a variety of cancer antigens and a markedly reduced immune response to tumor cells due to recognizing a smaller variety of HLA alleles, resulting in a reduced anticancer effect of immunotherapy. Taken together, homozygosity at the HLA-B locus appeared to be associated with a decreased antitumor effect of immunotherapy, but zygosity at other loci should be evaluated further in future studies. However, in the present study, no association was found between allele and serological specificity at the HLA loci in contrast to supertype at HLA-B locus. This result may be reasonable, given that the supertype more accurately reflects binding affinity to an antigen than allele or serological specificity.

When the individual genotypes of the HLA locus were examined, the B07 or B44 supertype of the HLA-B locus was associated with reduced intravesical recurrence, which indicated a superior anticancer effect of immunotherapy with BCG. The HLA-B07 and HLA-B44 supertypes may have higher affinities to cancer antigens associated with bladder cancer and BCG-specific immunogenic antigens. Indeed, B*35:01, a B07 supertype, was reported to have a high affinity to Mycobacterium tuberculosis-derived epitopes [33]. Interestingly, the HLA-B44 supertype was reported to be associated with favorable OS in melanoma patients treated with checkpoint block immunotherapy [10]. The common association of the HLA-B44 supertype with a favorable outcome for checkpoint block immunotherapy or BCG therapy suggests the HLA-B44 supertype may exert a stronger antitumor effect by enriching the presentation of neoantigens with a high affinity to the HLA-B44 supertype [10]. Conversely, homozygosity of HLA-B62 showed a prominent detrimental anticancer effect during BCG treatment, which suggests that inferior immunological reactions to cancer antigens associated with bladder cancer and BCG-specific immunogenic antigens occur with HLA-B62 during BCG treatment.

The present study suggested that supertype zygosity and individual supertype at the HLA-B locus determine the prognosis of intravesical recurrence in intravesical BCG therapy. There are various immunological mechanisms involved during BCG therapy [5, 15]. The importance of the HLA-B locus in BCG therapy supports the idea that the cytotoxicity of CD8⁺ T cells has a major role in the anticancer effect of BCG on NMIBC. Furthermore, the prognostic value of HLA genotype may contribute to the refinement of treatment strategies according to risk groups. For high-risk patients, a more intensive treatment strategy using immune checkpoint inhibitors such as PD-1 inhibitors or alternative therapies such as the development of recombinant BCG strains may be necessary to overcome BCG resistance [34, 35]. However, because this study was exploratory, further investigation is needed for validation.

The present study had several limitations. First, this study included only Japanese patients, which may be an obstacle to its application to other ethnic groups. Second, information on smoking history, performance status, and comorbidity status was not available. In addition, this study didn’t investigate biological mechanism on an association between HLA type and phenotype. Finally, the loss of heterozygosity, HLA expression levels, and the number of mutations in the tumor were not investigated. Recently, somatic mutations in HLA-I genes were reported; the frequency of non-silent mutation in bladder cancer was less than 5% [36, 37]. In addition, HLA-I expression in tumor is reduced in some bladder cancer [28, 29]. Therefore, their impact of HLA sequence in tumor might not so large. However, further investigation in tumor may be important using a novel technology.

Conclusions

In conclusion, this study suggests the importance of the diversity and specificity of HLA-B locus in the antitumor effect of BCG immunotherapy for NMIBC. These findings may contribute to improved risk stratification in BCG therapy and lead to the better medical management of NMIBC, although further study is required.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ contributions

MS and SN contributed to conception and design. MK, NF, TT, SN, and MS collected and assembled data. SN, YY, NF, AT, KN, and MS provisioned study material or patients. TT, SU, and MS analyzed and interpreted data. MS contributed to manuscript writing. TT and SU contributed to co-writing the manuscript. NF, SN, AT, KN, TH, HM, and ME critically reviewed the manuscript. ME contributed to supervision. All authors reviewed the manuscript and gave final approval of the manuscript.

Funding

This work was partly supported by grants (Nos. 21K06592, 19K18551) from the Japan Society for the Promotion of Science, Tokyo, Japan, and the Nakatomi Foundation, Tokyo, Japan.

Declarations

Conflict of interest

Shintaro Narita received honoraria from Janssen Pharmaceutical, Bayer Yakuhin, AstraZeneca, Takeda Pharmaceutical, Sanofi, and Astellas Pharma. Tomonori Habuchi received honoraria from Janssen Pharmaceutical, Takeda Pharmaceutical, Astellas Pharma, Daiichi Sankyo, AstraZeneca, Sanofi, and Bayer Yakuhin, and research funding support from Takeda Pharmaceutical, Astellas Pharma, Daiichi Sankyo, Sanofi, and Bayer Yakuhin. Masatoshi Eto received honoraria from Ono Pharmaceutical, Takeda Pharmaceutical, Novartis Pharma, Pfizer, Bristol-Myers Squibb, Janssen Pharmaceutical, MSD, and Merck Biopharma, and research funding support from Sanofi, Bayer Yakuhin, Astellas Pharma., Ono Pharmaceutical., and Takeda Pharmaceutical. Masaki Shiota received honoraria from Janssen Pharmaceutical, AstraZeneca, and Astellas Pharma, and research funding support from Daiichi Sankyo.

Ethical approval and ethical standards

The study was conducted under review board at Kyushu University (727-01).

Informed consent

Written informed consent was obtained from all patients.