Human Cytokine Genetic Variants Associated With HBsAg Reverse Seroconversion in Rituximab-Treated Non-Hodgkin Lymphoma Patients

Supplemental Digital Content is available in the text

Abstract

Hepatitis B virus (HBV) reactivation has been noted in HBV surface antigen (HBsAg)-seronegative patients with CD20⁺ B-cell non-Hodgkin lymphoma (NHL) undergoing rituximab treatment. Clinically, hepatitis flares are usually associated with the reappearance of HBsAg (reverse seroconversion of HBsAg, HBV-RS). It is unclear whether human genetic factors are related to rituximab-associated HBV reactivation.

Unvaccinated HBsAg-seronegative adults (n = 104) with CD20⁺ NHL who had received rituximab-containing therapy without anti-HBV prophylaxis were enrolled. Eighty-nine candidate single nucleotide polymorphisms (SNPs) of 49 human cytokine genes were chosen and were analyzed using the iPLEX technique. Competing risk regression was used to identify the factors associated with HBV-RS.

Participants had a median age of 66.1 years and 56.7% were male (n = 59). The anti-HBs and anti-HBc positivity rates were 82.4% and 94.1%, respectively, among patients for whom data were available (approximately 81%). A mean of 7.14 cycles of rituximab therapy were administered, and a total of 14 (13.4%) patients developed HBV-RS. Nine SNPs showed significant differences in frequency between patients with or without HBV-RS: CD40 rs1883832, IL4 rs2243248 and rs2243263, IL13 rs1295686, IL18 rs243908, IL20 rs1518108, and TNFSF13B rs12428930 and rs12583006. Multivariate analysis showed that ≥6 cycles of rituximab therapy, IL18 rs243908, and the IL4 haplotype rs2243248∼rs2243263 were independently associated with HBV-RS. The IL4 haplotype rs2243248∼rs2243263 was significantly associated with HBV-RS regardless of anti-HBs status.

Polymorphisms in human cytokine genes impact the risk of rituximab-associated HBV-RS.

INTRODUCTION

Rituximab, an anti-CD20 monoclonal antibody, is used to treat patients with CD20-positive B-cell non-Hodgkin lymphoma (CD20⁺ NHL)¹ or rheumatoid arthritis.² Hepatitis B virus (HBV) reactivation has been noted in hepatitis B virus surface antigen (HBsAg)-seronegative patients with CD20⁺ NHL, and at 10% risk of reverse seroconversion of hepatitis B virus surface antigen (HBV-RS).^3–7 Clinically, hepatitis flares are frequently associated with the reappearance of HBsAg (i.e., HBV-RS).⁵ Among the risk factors for HBV-RS in HBsAg-seronegative patients with CD20⁺ NHL, HBV serological status prior to rituximab therapy, including antibody to hepatitis B virus core antigen (anti-HBc) seropositivity and antibody to hepatitis B virus surface antigen (anti-HBs) seronegativity, has been shown to be associated with a significantly increased risk in some reports.^5,8 A higher number of cycles of rituximab therapy has also been highlighted as a risk factor.^6,7 However, the anti-HBc and anti-HBs seropositivity rate is relatively high in unvaccinated HBsAg-seronegative adults in HBV-hyperendemic areas,^3,4,6–9 for example, at least 70% in Taiwan,^4–7,9 which was the 1st country to initiate universal HBV vaccination, in 1984.¹⁰ Therefore, it is necessary to investigate whether human genetic factors are related to rituximab-associated HBV reactivation.

Few studies investigated the association between genetic background and the adverse effects of rituximab therapy. Rossi et al¹¹ evaluated 19 single nucleotide polymorphisms (SNPs) in 106 patients with diffuse large B cell lymphoma who underwent treatment with rituximab combination with cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP) and found that a variant of NAD(P)H oxidase subunits, NCF4 rs1883112, was an independent predictor against hematologic, infectious, and cardiac toxicities. However, the candidate genes were chosen mainly based on their metabolic involvement with the chemotherapeutic agents, rather than with rituximab. It remains unclear whether human genetic factors are related to rituximab-associated HBV reactivation. However, because the pathogenesis of HBV infection is mainly immune-mediated,¹² HBV reactivation and HBV-RS might be associated with human genetic factors responsible for immune responses.

As few previous studies have examined this issue, this pilot study aimed to analyze SNPs of candidate genes. The genes involved in T-cell immune responses are known to be involved in both HBV reactivation and rituximab therapy, and they were chosen for analysis, for several reasons. First, the degree of T-cell immune response and the interaction of several cytokines are known to influence seroconversion, severity, and chronicity in HBV infection.^12,13 Several studies have investigated polymorphisms of the genes encoding these cytokines, including tumor necrosis factor (TNF) and interferon gamma (INF γ).^14,15 Second, evidence from a mouse model¹⁶ and from patients with various underlying diseases indicated that rituximab-induced B-cell depletion may also influence T-cell immune responses. Hilchey et al¹⁷ found that rituximab killed follicular lymphoma (FL) cells via the elicitation of an FL-specific T-cell response. Rituximab caused reversion of the T-cell immune response in patients with immune thrombocytopenic purpura,¹⁸ and patients with systemic lupus erythematosus and rheumatoid arthritis.¹⁹ However, rituximab-induced B-cell depletion might exacerbate T-cell-dependent immune-mediated diseases²⁰ and is known to impair the vaccine response.^21,22 These findings implied that in CD20⁺ NHL patients, rituximab therapy would have damaged the balance between T-cell immune responses and occult HBV infection harbored within hepatocytes, finally resulting in HBV reactivation, HBV-RS, and hepatitis flare.

In this study, we evaluated human genetic variants that might be responsible for anti-HBV immune responses and the adverse effects of rituximab through mass screening of 89 SNPs among 49 candidate genes.

PATIENTS AND METHODS

Patients and Data Collection

For consecutive adult patients (≥18 years of age) with the diagnosis of CD20⁺ NHL between January 2000 and December 2010 at Taipei Veterans General Hospital,⁷ those who were seronegative for HBsAg at diagnosis, received rituximab alone or combined with chemotherapy, and had genomic DNA archived were eligible for the study. Patients were excluded if any of the following criteria were met: seropositive for hepatitis C virus (HCV) or human immunodeficiency virus at diagnosis, HBV vaccination, prophylactic use of anti-HBV agents, and organ or hematopoietic stem cell transplantation administered before the event of HBV reactivation or at last follow-up.

The collected data included demographics, NHL subtypes and treatments, liver function tests, HBV serology tests (HBsAg, anti-HBs, and anti-HBc) and viral load, and outcome. The HBV-RS event was identified, and additional data on liver function, viral serology, and virological tests were collected. This study was approved by the Institutional Review Board of the Taipei Veterans General Hospital.

Rituximab-Containing Therapy and Monitoring of HBV Reactivation

The induction therapy for patients with aggressive CD20⁺ NHL included 6 to 8 cycles of rituximab (375 mg/m² per cycle) in combination with CHOP or a CHOP-like regimen every 3 weeks.¹ At diagnosis, serological tests for HIV, HBsAg, and anti-HCV were routinely performed, and testing for anti-HBs and anti-HBc were performed more frequently subsequent to the late 2000s. Liver function tests were performed before every cycle of rituximab-containing therapy, and during the follow-up period, including alanine aminotransferase (normal ranges <40 IU/L), aspartate aminotransferase, alkaline phosphate, and total bilirubin.^4,7 For clinical suspicion of HBV reactivation during and after rituximab-containing therapy, HBsAg and anti-HCV were tested again, and, in addition, quantitative measurement of HBV viral load in the serum was performed, using the COBAS Taqman HBV test (Roche Diagnostics, Switzerland), with a detection limit of 12 IU/mL.^4,7 HBeAg (and anti-HBe) testing was usually performed when the results of the HBsAg test were positive.^4,23

In this study, HBV-RS was used to represent HBV reactivation and was defined as reappearance of HBsAg in the serum, and an HBV viral load >2000 IU/mL if applicable.^4,7 Hepatitis flare preceded or accompanied by HBV-RS was defined if serum alanine aminotransferase level was above 100 IU/L.^4,7 The cycle number of rituximab-containing therapy was recorded until HBV-RS was noted, or the final follow-up if HBV-RS did not develop. The overall survival (OS) was defined as the time between the beginning of rituximab-containing therapy and the final follow-up.⁷

Selection and Genotyping of Candidate Genes and SNPs

This pilot study was part of the Genetic Polymorphisms Influencing the Impact of HBV Infection on CD20⁺ NHL Patients study, in terms of those relevant to toxicities and susceptibility. Candidate genes and SNPs driving immune responses were chosen and analyzed via 3 steps. First, except for SNPs of NCF4,¹¹ candidate genes and their SNPs were chosen based on a comprehensive literature review of articles relevant to immune response. The articles were collected via a PubMed search on Sep. 30, 2012 using different combinations of several keywords, including the combination of “lymphoma, susceptibility, and polymorphism” and “(cytokine, or chemokine, or innate immunity);” the combination of “hepatitis B virus and polymorphism” and “(cytokine, or chemokine, or innate immunity).” Those SNPs with sufficient information, including statistical significance, were chosen. Second, SNPs with a known allele frequency <5% in Asian or Chinese populations were excluded, based on data in the National Center for Biotechnology Information SNP database (http://www.ncbi.nlm.nih.gov/snp/ ). Finally, only those SNPs for which primers were available for the iPLEX genotyping technique (Sequenom, Inc., MI) were analyzed. A total of 89 SNPs of 49 genes were analyzed in 3 multiplex reactions (supplemental Table 1).

Genomic DNA was used for genotyping and was mainly collected from the peripheral blood or bone marrow of CD20⁺ NHL patients at diagnosis. Genotyping of SNPs was performed using iPLEX SNP Genotyping (Sequenom, Inc., MI) at the National Center for Genomic Medicine, Taiwan. The iPLEX assay can detect sequence differences at the single nucleotide level based on the primer extension process, in which the primer is extended dependent upon the template sequence, resulting in allele-specific differences in mass between extension products.²⁴

Statistical Analysis

Patients with and without HBV-RS were categorized as the case and control subgroups, respectively. Clinical variables were compared between the case and control subgroups. The Chi-squared test (or Fisher exact test when applicable), Student's t-test (or one-way analysis of variance when applicable), and the Mann–Whitney U-test were applied for categorical, parametrically, and nonparametrically continuous variables between subgroups, respectively.

Allele and genotype frequencies of SNPs were calculated, and those with a call rate of <90% that deviated from the Hardy–Weinberg equilibrium (with a cut-off value of P < 0.05) in the control subgroup, or that were monomorphic were excluded from subsequent association analysis. Using HBV-RS as a binary response variable, an association analysis of single SNPs with HBV-RS was performed using logistic regression, and results are stated using odds ratios (ORs) with a relative 95% confidence interval (CI). The best inheritance model (codominant, dominant, recessive, over-dominant, or log-additive) was selected, according to values for P, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). Furthermore, those SNPs with a P-value near 0.05 in the single SNP association analysis were further analyzed in terms of the presence of linkage disequilibrium (LD) and haplotype. For LD, matrices with selected statistics (D, D′, Pearson r, and associated P-values) are shown. The association of haplotype with HBV-RS was shown as the global score test.²⁵

Finally, clinical variables, genotypes, and haplotypes that differed significantly between the case and control subgroups were considered to be potential factors influencing the development of rituximab-associated HBV-RS, and were further evaluated using regression analysis. With “death in the absence of HBV-RS” considered as a competing event, the cumulative incidence of HBV-RS was calculated, and the Fine and Gray model was used to assess the impact of different variables on the cumulative incidence of HBV-RS,⁷ and reported using subhazard ratios (SHRs) with a relative 95% CI. Finally, the multivariate regression analysis was used to identify factors that were independently influencing the development of rituximab-associated HBV-RS, after adjusting for the factors that had P < 0.1 in univariate analyses. A P-value <0.05 was thought statistically significant. The testing for Hardy–Weinberg equilibrium, linkage disequilibrium, association, and descriptive analysis of SNPs were performed using the SNPstata web site (http://bioinfo.iconcologia.net/SNPstats).²⁶ Other analyses were done using the Statistical Package for the Social Sciences (SPSS version 17.0), Stata version 11.0 (Stata Corp., College Station, TX), and XLSTAT (Addinsoft SARL, Paris, France).

RESULTS

Patient Characteristics According to the Presence of HBV-RS

As shown in Table 1, 104 unvaccinated HBsAg-seronegative patients with CD20⁺ NHL were enrolled. The median age was 66.1 years and more patients were male than female (n = 59 males, 56.7%). The major histological subtypes were diffuse large B cell lymphoma and FL, which accounted for 66.7% (n = 59) and 23.1% (n = 24) of patients, respectively. Information on anti-HBs and anti-HBc status was available for 81.7% and 82.7% of patients, respectively, with respective positive rates of 82.4% and 94.1%.

TABLE 1.

Characteristics of 104 Unvaccinated HBsAg-Seronegative Patients With CD20⁺ NHL According to the Development of HBsAg Reverse Seroconversion

At time of last follow-up or HBV-RS, a mean of 7.14 cycles of rituximab therapy had been administered, and approximately 70% of patients had received >6 cycles. A total of 14 (13.4%) patients developed HBV-RS, and all of those for whom data were available were seropositive for anti-HBc (10/10). Eleven (79%) patients experienced hepatitis flare. Compared with their counterparts, patients with HBV-RS had a lower rate of anti-HBs-positivity (58.3% vs 86.3%, P = 0.033), and more cycles and longer durations of rituximab therapy. No significant differences in OS were observed between patients with and without HBV-RS.

Candidate SNPs Associated With HBV-RS

The allele and genotype frequencies of the 89 SNPs were analyzed by iPLEX assay and are listed in supplemental Table 2. For 84 of the SNPs, genotyping had a call rate of 90.38% to 100%; only 5 SNPs (CXCL8 rs4073, CXCR4 rs2228014, IL1B rs1143643, IL12RB1 rs12564159, and IL18 rs187238) had a low call rate. Three SNPs (CARD15/NOD2 rs2066844, NLRC5 rs9936269, and TLR2 rs5743708) were monomorphic. In addition, the genotype frequencies of 7 SNPs deviated from Hardy–Weinberg equilibrium, including CD40LG rs3092933, FCGR2A rs10800309, IFNL3 rs8099917 and rs12980275, IFNL4(IL28B) rs12979860, IL12A rs485497, and TNFSF13B rs1224141. The remaining 74 SNPs were included in the HBV-RS association analysis (supplemental Table 3). Finally, without correction of multiple comparisons, 9 candidate SNPs showed a possible association with HBV-RS, with the cut-off value of uncorrected P < 0.06, including CCL5 rs2107538, CD40 rs1883832, IL4 rs2243248 and rs2243263, IL13 rs1295686, IL18 rs243908, IL20 rs1518108, TNFSF13B rs12583006, and rs12428930 (Table 2). Of these, the IL4 rs2243248 T/G and rs2243263 C/G genotypes were present in 20% (n = 18) and 23.3% (n = 21) of the total patients, respectively, none of these developed HBV-RS (Table 2).

TABLE 2.

Candidate Single Nucleotide Polymorphisms Associated With HBsAg Reverse Seroconversion After Rituximab Therapy

Haplotypes Associated with HBV-RS

The LD analysis of these 9 candidate SNPs showed LD between IL4 rs2243248 and IL4 rs2243263 (r² = 0.9178) and between TNFSF13B rs12583006 and TNFSF13B rs12428930 (r² = 0.9813) (supplemental Tables 4 and 5). The 18 patients (17.3% of 104) who had IL4 rs2243248 T/G and did not develop HBV-RS also had IL4 rs2243263 C/G (supplemental Tables 3 and 5). Of note, the IL4 rs2243248∼rs2243263 haplotype was significantly associated with the development of rituximab-associated HBV-RS (global P = 0.033) (Table 3).

TABLE 3.

Haplotype Frequencies and Association Analysis of IL4 and TNSSF13B According to the Development of HBsAg Reverse Seroconversion

Competing-Risks Regression Analysis of the Association of Clinical Factors, SNPs, and Haplotypes With HBV-RS

When “death in the absence of HBV-RS” was considered as the competing risk, the cumulative incidence of HBV-RS was estimated to be 2%, 5%, and 13.3% at 6 months, 12 months, and 24 months after rituximab therapy, respectively (Figure 1).

FIGURE 1.

Cumulative incidence of rituximab-associated reverse seroconversion of HBsAg in 104 unvaccinated patients with CD20⁺ NHL, with “death in the absence of HBsAg reverse seroconversion” as the competing risk. The P value was calculated by competing risk analysis based on the Fine and Gray model. CD20⁺ NHL = CD20-positive B-cell non-Hodgkin lymphoma, HBsAg = hepatitis B virus surface antigen.

In the univariate regression analysis (Table 4), the factors found to be independently associated with the development of HBV-RS were anti-HBs status, cycles of rituximab therapy (≥6 cycles), the SNPs IL4 rs2243248 (Figure 2A) and rs2243263 (Figure 2B), IL13 rs1295686 (Figure 2C), IL18 rs243908 (Figure 2D), IL20 rs1518108 (Figure 2E), and the haplotype IL4 rs2243248∼rs2243263 (Figure 3A). In addition, the SNPs TNFSF13B rs12428930 and rs12583006, and the haplotype TNFSF13B rs12428930∼rs12583006 (Figure 3B) showed a trend of association with HBV-RS.

TABLE 4.

Competing Risk Regression for Predictors of HBsAg Reverses Seroconversion After Rituximab Therapy^∗

FIGURE 2.

Cumulative incidence of rituximab-associated reverse seroconversion of HBsAg in 104 unvaccinated patients with CD20⁺ NHL, with “death in the absence of HBsAg reverse seroconversion” as the competing risk, according to SNPs (A) IL4 rs2243248, (B) IL4 rs2243263, (C) IL13 rs1295686, (D) IL18 rs243908, and (E) IL20 rs1518108. The P value was calculated by competing risk analysis based on the Fine and Gray model. CD20⁺ NHL = CD20-positive B-cell non-Hodgkin lymphoma, HBsAg = hepatitis B virus surface antigen, SNP = single nucleotide polymorphism.

FIGURE 3.

Cumulative incidence of rituximab-associated reverse seroconversion of HBsAg in 104 unvaccinated patients with CD20⁺ NHL, with “death in the absence of HBsAg reverse seroconversion” as the competing risk, according to haplotypes (A) IL4 rs2243248∼rs2243263 and (B) TNFSF13B rs12428930∼rs12583006. The P value was calculated by competing risk analysis based on the Fine and Gray model. CD20⁺ NHL = CD20-positive B-cell non-Hodgkin lymphoma, HBsAg = hepatitis B virus surface antigen.

In the multivariate analysis adjusted for gender and histological subtype of all patients, the remaining independent factors associated with HBV-RS were cycles of rituximab therapy (≥6 cycles), the SNP IL18 rs243908, and the haplotype IL4 rs2243248∼rs2243263 (Table 4). The SNP IL20 rs1518108 showed a trend of association (P = 0.076).

IL4 Haplotype Was Not Related to Anti-HBs Status in the Risk of HBV-RS

As 18.3% of patients were missing data for anti-HBs, anti-HBs was not included in the multivariate analysis. Subgroup analysis was performed to determine whether the IL18 SNP and the IL4 haplotype were associated with the risk of HBV-RS, according to anti-HBs status at diagnosis. In this analysis, IL18 rs243908 was no longer significantly associated with HBV-RS in the anti-HBs seropositive and negative subgroups (see Figure 4A [SHR = 2.512; 95% CI, 0.574–11.001; P = 0.222] and 4B [SHR = 1.088; 95% CI, 0.206–5.749; P = 0.921] for the seropositive and seronegative subgroups, respectively). It should be noted that the association of IL4 haplotype rs2243248∼rs2243263 with HBV-RS remained significant in both the anti-HBs seropositive and seronegative subgroups (Figure 4C and D for the seropositive [SHR = 1.51e-19; P ≈ 0] and seronegative groups [SHR = 1.76e-19; P ≈ 0], respectively).

FIGURE 4.

Cumulative incidence of rituximab-associated reverse seroconversion of HBsAg in 86 unvaccinated patients with CD20⁺ NHL, with “death in the absence of HBsAg reverse seroconversion” as the competing risk, according to the anti-HBs serostatus at diagnosis and IL18 SNP rs243908 (A, B), and IL4 haplotype rs2243248∼rs2243263 (C, D). The P value was calculated by competing risk analysis based on the Fine and Gray model. anti-HBs = antibody to hepatitis B virus surface antigen, CD20⁺ NHL = CD20-positive B-cell non-Hodgkin lymphoma, HBsAg = hepatitis B virus surface antigen, SNP = single nucleotide polymorphism.

DISCUSSION

To our knowledge, this pilot study was the 1st to explore the potential association of SNPs and haplotypes of human cytokine genes with rituximab-associated HBV-RS. The findings are particularly noteworthy because the regression analysis employed clinical factors, and because the association with a higher number of cycles of rituximab (≥6) noted here is similar to results published in our recent report.^4,7 First, both SNPs of IL4, rs2243248 and rs2243263, as well as the IL4 haplotype rs2243248∼rs2243263 were found to be significantly associated with the development of rituximab-associated HBV-RS. IL4 is produced by type 2 helper T (Th2) cells and can costimulate B-cell proliferation.²⁷ For possible mechanisms by which IL4 influences rituximab-associated HBV-RS, IL4 could suppresses the expression and the replication of HBV in the hepatocellular carcinoma cell line Hep3B,²⁸ but rituximab was shown to cause a 12-fold peak-decrease of IL4 mRNA in whole blood.²⁹ A lower IL4 level was associated with a lower anti-HBs level after HBV vaccination in HIV patients.³⁰ Total 6 SNPs of IL4 were genotyped in this study (supplemental Table 1), including rs2243248 located in the promoter region, rs2070874 in the 5′ untranslated region, rs2227284, rs2243263, and rs2243268 all in the intron region, and rs2243291 in 3′ untranslated region. The allele frequencies of IL4 rs2243248 and rs2243263 and the strong LD between them noted here are concordant with previous results.³¹IL4 rs2243248 might influence the mRNA transcription of IL4, similar to the way of IL4 rs2243250 (−590C>T),^32–34 which is also located in the promoter region, but was not included in the present analysis owing to technical limitations of iPLEX. On the other hand, IL4 rs2243263 might participate in the regulation of IL4 expression via binding with CTCF transcription factors.³⁵ Chen et al³⁶ reported that the IL4 rs2243248 G allele was associated with nonresponse to HBV vaccine; however, it was found to be associated with a lower incidence of HBV-RS in our study (Table 4). In addition, IL4 rs2243248 was associated with an increased the likelihood of high lytic antibody titers for human herpesvirus 8 infection.³⁷ Consequently, we speculated that CD20⁺ NHL patients with different IL4 rs2243248∼rs2243263 haplotypes will have different serum IL4 levels and differing T-cell responses during rituximab therapy, thus influencing the development of HBV-RS.

Second, IL18 rs243908 was found to be associated with the development of HBV-RS in our patients (Table 4). As previously described,³⁸ IL18 is produced not only by immune cells but also by nonimmune cells. IL18 enhances the IL12-driven Th1 immune response, but it can also stimulate Th2 immune responses in the absence of IL12. IL18 could also influence HBV replication by activating resident intrahepatic natural killer cells and natural killer T cells to produce gamma interferon (IFN-γ), and by inducing IFN-α/β production in the liver.³⁹ Two IL18 SNPs – rs243908 and rs187238 were genotyped in this study (supplemental Table 1), based on the association of lymphoma⁴⁰ and HBV infection,⁴¹ respectively. IL18 rs187238 located in the promotor region of IL18 was found to be associated with the transcriptional activity of the IL18 promoter,⁴² and even with serum IL18 levels.⁴³IL18 rs187238 and several other IL18 SNPs were reported to be associated with chronic HBV infection,⁴⁴ the development of hepatocellular hepatoma, and the clearance of HBV.⁴⁵ Unfortunately, we were unable to include rs187238 in our analysis because it showed a low call rate (supplemental Tables 1 and 2). On the other hand, IL18 rs243908 was also located in 5′ untranslated region, and the relationship with IL18 secretion was not understood. It may potentially be associated with HBV infection, as it showed modest LD with IL18 rs187238 (r² = 0.809). The association of IL18 rs243908 with HBV-RS was not found to be significant in the subgroup analysis according to anti-HBs serostatus (Figure 4A and B). The loss of significance was probably caused by the relatively low number of patients in each subgroup; its impact on HBV-RS might have been overshadowed by the effect of the anti-HBs serostatus group.

In addition, the multivariate analysis showed that IL20 rs1518108 had a trend of association with the development of HBV-RS (Table 4). IL20, which belongs to the IL10 family, is a potent immunomodulatory cytokine, and the selection of IL20 rs1518108 here was based on the potential association with HBV infection.⁴⁵ Similarly to the report published by Truelove et al,⁷ which stated that the IL20 rs1518108 C/T variant is associated with chronic HBV infection in African American, we demonstrated that patients with the variants C/T-T/T tended to have a higher risk of HBV-RS than those with the C/C variant (Table 4). Because all of the above cytokines are closely regulated and correlated with each other during infection in humans, it is possible that all of the IL18 and IL20 SNPs and the IL4 haplotypes are participating in the process of HBV-RS. However, IL4 haplotype may serve as a surrogate marker to represent host immune status and susceptibility to HBV-RS during rituximab treatment in CD20⁺ NHL patients with resolved hepatitis B.

Our study has a few limitations. First, this pilot study was part of the Genetic Polymorphisms Influencing the Impact of HBV Infection on CD20⁺ NHL Patients project, in terms of those relevant to toxicity and susceptibility, and candidate genes were selected from those relevant to lymphoma susceptibility and those relevant to HBV infection. Identification of the underlying mechanisms of the SNPs and haplotypes identified here is beyond the scope of this study, but should be investigated in future. Second, the retrospective design, the heterogeneity of patients, and the relatively small number of HBV-RS patients might obscure the potential significance of some SNPs. A larger number of patients would not only improve the degree of precision for these SNPs and haplotypes with a trend of association in this study, but also help us to further analyze the relationship between these SNPs and treatment outcome or drug toxicity. In addition, the association of these SNPs and haplotypes with HBV-RS should also be examined in CD20⁺ NHL patients with other risk factors recently reported,^6,7 including anti-HCV seropositivity, histological subtype of posttransplant lymphoproliferative disorders, and succeeding hematopoietic stem cell transplantation. Third, we did not analyze the impact of baseline HBV DNA, because it was not routinely tested for our patients of this study. Our previous report showed that undetectable HBV viral load before chemotherapy did not confer reactivation-free status of rituximab treated lymphoma patients.⁴ Finally, the other factor – the intensity of combination chemotherapy (e.g., CHOP) was not analyzed here, because the risk of HBV-RS in CD20⁺ NHL patients with resolved hepatitis B treated with CHOP alone was very low.³

In conclusion, this pilot study was the 1st to provide evidence that the development of rituximab-associated HBV-RS is likely to be influenced by host genetic background in lymphoma patients. Our findings provide a clinical indication to further stratify the risk of HBV-RS in CD20⁺ NHL patients with resolved hepatitis B before administering rituximab treatment.