Beth R. Larrabee

Fine Mapping Causal Variants with an Approximate Bayesian Method Using Marginal Test Statistics

Two recently developed fine-mapping methods, CAVIAR and PAINTOR, demonstrate better performance over other fine-mapping methods. They also have the advantage of using only the marginal test statistics and the correlation among SNPs. Both methods leverage the fact that the marginal test statistics asymptotically follow a multivariate normal distribution and are likelihood based. However, their relationship with Bayesian fine mapping, such as BIMBAM, is not clear. In this study, we first show that CAVIAR and BIMBAM are actually approximately equivalent to each other. This leads to a fine-mapping method using marginal test statistics in the Bayesian framework, which we call CAVIAR Bayes factor (CAVIARBF). Another advantage of the Bayesian framework is that it can answer both association and fine-mapping questions. We also used simulations to compare CAVIARBF with other methods under different numbers of causal variants. The results showed that both CAVIARBF and BIMBAM have better performance than PAINTOR and other methods. Compared to BIMBAM, CAVIARBF has the advantage of using only marginal test statistics and takes about one-quarter to one-fifth of the running time. We applied different methods on two independent cohorts of the same phenotype. Results showed that CAVIARBF, BIMBAM, and PAINTOR selected the same top 3 SNPs; however, CAVIARBF and BIMBAM had better consistency in selecting the top 10 ranked SNPs between the two cohorts. Software is available at https://bitbucket.org/Wenan/caviarbf.

Race and sex-based differences in cytokine immune responses to smallpox vaccine in healthy individuals

We assessed the effects of sex, race and ethnicity on smallpox vaccine-induced immune responses in 1071 armed forces members after primary Dryvax(®) smallpox vaccination, including 790 males and 281 females; 580 Caucasians, 217 African-Americans, and 217 Hispanics. Analysis of vaccinia-specific cytokine responses revealed that Caucasians had higher total IFNγ ELISPOT responses (median 57 spot-forming units/SFUs per 200,000 cells, p=0.01) and CD8(+)IFNγ ELISPOT responses (12 SFUs, p<0.001) than African-Americans (51 and 4 SFUs, respectively) and Hispanics (47 and 8 SFUs, respectively). Similarly, Caucasians secreted higher levels of vaccinia-specific IL-2 (p=0.003) and IFNα (p<0.001) compared to other racial/ethnic groups. Males had higher total IFNγ ELISPOT responses (median 55 SFUs) compared to females (41 SFUs, p<0.001). We observed statistically significant sex-related differences in the secretion of IL-2 (p<0.001), IL-1β (p<0.001) and IL-10 (p=0.017). These data suggest that vaccinia-specific cytokine responses following primary smallpox vaccination are significantly influenced by race and sex of vaccinees.

Leptin and leptin-related gene polymorphisms, obesity, and influenza A/H1N1 vaccine-induced immune responses in older individuals.

Obesity is a risk factor for complicated influenza A/H1N1 disease and poor vaccine immunogenicity. Leptin, an adipocyte-derived hormone/cytokine, has many immune regulatory functions and therefore could explain susceptibility to infections and poor vaccine outcomes. We recruited 159 healthy adults (50-74 years old) who were immunized with inactivated TIV influenza vaccine that contained A/California/7/2009/H1N1 virus. We found a strong correlation between leptin concentration and BMI (r=0.55, p<0.0001), but no association with hemagglutination antibody inhibition (HAI), B-cell, or granzyme B responses. We found a slight correlation between leptin concentration and an immunosenescence marker (TREC: T-cell receptor excision circles) level (r=0.23, p=0.01). We found eight SNPs in the LEP/LEPR/GHRL genes that were associated with leptin levels and four SNPs in the PTPN1/LEPR/STAT3 genes associated with peripheral blood TREC levels (p<0.05). Heterozygosity of the synonymous variant rs2230604 in the PTPN1 gene was associated with a significantly lower (531 vs. 259, p=0.005) TREC level, as compared to the homozygous major variant. We also found eight SNPs in the LEP/PPARG/CRP genes associated with variations in influenza-specific HAI and B-cell responses (p<0.05). Our results suggest that specific allelic variations in the leptin-related genes may influence adaptive immune responses to influenza vaccine.

Cytokine gene polymorphisms and progression-free survival in classical Hodgkin lymphoma by EBV status: Results from two independent cohorts

BACKGROUND Cytokines are important immune mediators of classical Hodgkin lymphoma (CHL) pathogenesis, and circulating levels at diagnosis may help predict prognosis. Germline single nucleotide polymorphisms (SNPs) in immune genes have been correlated with cytokine production and function. METHODS We investigated whether selected germline SNPs in IL10 (rs1800890, rs1800896, rs1800871, rs1800872), TNFA (rs1800629), IL6 (rs1800795), ILRN (rs419598), INFG (rs2430561) and CCL17 (rs223828) were associated with circulating levels of related cytokines at diagnosis and progression-free survival (PFS) in CHL. Patients were from France (GELA, N=464; median age=32years) and the United States (Iowa/Mayo Specialized Program Of Research Excellence [SPORE], N=239; median age=38years); 22% of 346 CHL cases with EBV tumor status were positive. RESULTS There was no association with any of the SNPs with cytokine levels. Overall, there was no association of any of the SNPs with PFS. In exploratory analyses by EBV status, TNFA rs1800629 (HRAA/AG=2.41; 95%CI, 1.17-4.94) was associated with PFS in EBV-negative GELA patients, with similar trends in the SPORE patients (HRAA/AG=1.63; 95%CI, 0.61-4.40). In a meta-analysis of the two studies, TNFA (HRAA/AG=2.11; 95%CI, 1.18-3.77; P=0.01) was statistically significant, and further adjustment for the international prognostic system did not alter this result. CONCLUSIONS This study showed that germline variation in TNFA was associated with CHL prognosis for EBV-negative patients, which will require confirmation. These results support broader studies on the differential impact of genetic variation in immune genes on EBV-positive vs. EBV-negative CHL pathogenesis.

High-throughput assay optimization and statistical interpolation of rubella-specific neutralizing antibody titers.

ABSTRACT Rubella remains a social and economic burden due to the high incidence of congenital rubella syndrome (CRS) in some countries. For this reason, an accurate and efficient high-throughput measure of antibody response to vaccination is an important tool. In order to measure rubella-specific neutralizing antibodies in a large cohort of vaccinated individuals, a high-throughput immunocolorimetric system was developed. Statistical interpolation models were applied to the resulting titers to refine quantitative estimates of neutralizing antibody titers relative to the assayed neutralizing antibody dilutions. This assay, including the statistical methods developed, can be used to assess the neutralizing humoral immune response to rubella virus and may be adaptable for assessing the response to other viral vaccines and infectious agents.

Statistical Methods for Testing Genetic Pleiotropy

Genetic pleiotropy is when a single gene influences more than one trait. Detecting pleiotropy and understanding its causes can improve the biological understanding of a gene in multiple ways, yet current multivariate methods to evaluate pleiotropy test the null hypothesis that none of the traits are associated with a variant; departures from the null could be driven by just one associated trait. A formal test of pleiotropy should assume a null hypothesis that one or no traits are associated with a genetic variant. For the special case of two traits, one can construct this null hypothesis based on the intersection-union (IU) test, which rejects the null hypothesis only if the null hypotheses of no association for both traits are rejected. To allow for more than two traits, we developed a new likelihood-ratio test for pleiotropy. We then extended the testing framework to a sequential approach to test the null hypothesis that k+1 traits are associated, given that the null of k traits are associated was rejected. This provides a formal testing framework to determine the number of traits associated with a genetic variant, while accounting for correlations among the traits. By simulations, we illustrate the type I error rate and power of our new methods; describe how they are influenced by sample size, the number of traits, and the trait correlations; and apply the new methods to multivariate immune phenotypes in response to smallpox vaccination. Our new approach provides a quantitative assessment of pleiotropy, enhancing current analytic practice.

HLA Genotypes and Rubella Vaccine Immune Response: Additional Evidence

Recent population-based studies have demonstrated the genetic heritability of rubella vaccine response and assessed that the HLA system may explain about 20% of the inter-individual variance in humoral immune response to this vaccine. Our earlier studies compared HLA allelic associations with rubella vaccine-specific antibodies between two smaller cohorts of healthy Rochester, MN, children (346 and 396 subjects) after two doses of rubella-containing vaccine. This study found that specific HLA alleles were consistently associated with rubella-specific antibody titers (B*27:05, DPA1*02:01, and DPB1*04:01 alleles). The current study examined HLA associations in an independent larger cohort of 1012 healthy San Diego, CA, subjects (age 19-40 years) after rubella vaccine in order to replicate our previous findings in the Rochester subjects. Two HLA associations of comparable magnitudes were consistently observed between B*27:05 (median NT50 Rochester cohort 48.9, p=0.067; San Diego cohort 54.8, p=0.047) and DPB1*04:01 (median NT50 Rochester cohort 61.6, p<0.001; San Diego cohort 70.8, p=0.084) alleles and rubella virus-neutralizing antibody titers. Additional HLA alleles resulted in consistent effects on IL-6 production in both cohorts, but did not meet criteria for statistical significance. Our data suggest these HLA alleles play a role in rubella vaccine-induced immunity and provide the basis for future studies that may explain the mechanism(s) by which these HLA polymorphisms affect immune responses to rubella vaccine.

Single-nucleotide polymorphism associations in common with immune responses to measles and rubella vaccines

Single-nucleotide polymorphisms (SNPs) in candidate immune response genes were evaluated for associations with measles- and rubella-specific neutralizing antibodies, interferon (IFN)-γ, and interleukin (IL)-6 secretion in two separate association analyses in a cohort of healthy immunized subjects. We identified six SNP associations shared between the measles-specific and rubella-specific immune responses, specifically neutralizing antibody titers (DDX58), secreted IL-6 (IL10RB, IL12B), and secreted IFN-γ (IFNAR2, TLR4). An intronic SNP (rs669260) in the antiviral innate immune receptor gene, DDX58, was significantly associated with increased neutralizing antibody titers for both measles and rubella viral antigens post-MMR vaccination (p values 0.02 and 0.0002, respectively). Significant associations were also found between IL10RB (rs2284552; measles study p value 0.006, rubella study p value 0.00008) and IL12B (rs2546893; measles study p value 0.005, rubella study p value 0.03) gene polymorphisms and variations in both measles- and rubella virus-specific IL-6 responses. We also identified associations between individual SNPs in the IFNAR2 and TLR4 genes that were associated with IFN-γ secretion for both measles and rubella vaccine-specific immune responses. These results are the first to indicate that there are SNP associations in common across measles and rubella vaccine immune responses and that SNPs from multiple genes involved in innate and adaptive immune response regulation may contribute to the overall human antiviral response.

Associations between single nucleotide polymorphisms in cellular viral receptors and attachment factor-related genes and humoral immunity to rubella vaccination

Background Viral attachment and cell entry host factors are important for viral replication, pathogenesis, and the generation and sustenance of immune responses after infection and/or vaccination, and are plausible genetic regulators of vaccine-induced immunity. Methods Using a tag-SNP approach in candidate gene study, we assessed the role of selected cell surface receptor genes, attachment factor-related genes, along with other immune genes in the genetic control of immune response variations after live rubella vaccination in two independent study cohorts. Results Our analysis revealed evidence for multiple associations between genetic variants in the PVR, PVRL2, CD209/DC-SIGN, RARB, MOG, IL6 and other immune function-related genes and rubella-specific neutralizing antibodies after vaccination (meta p-value <0.05). Conclusion Our results indicate that multiple SNPs from genes involved in cell adhesion, viral attachment, and viral entry, as well as others in genes involved in signaling and/or immune response regulation, play a role in modulating humoral immune responses following live rubella vaccination.

FCGR2A and FCGR3A polymorphisms in classical Hodgkin lymphoma by Epstein–Barr virus status

We read with interest the study of Diamantopoulos et al. published in Leukemia and Lymphoma in which the authors analyzed the correlation between latent EBV infection and an FCGR2A polymorphism among 40 Greek-Caucasians with leukemic low-grade B-cell lymphoma, mainly chronic lymphocytic leukemia (N=23, 57.5%) and splenic marginal zone lymphoma (N=11, 27.5%) [1]. Patients were considered as EBV-positive based on the detection of the BXLF-1 gene by PCR in a peripheral blood sample (viral load, 2164.2 copies per mL, range 79–15,600). The viral mRNA LMP1 and the EBV serological profile of the patients were also investigated. The correlation of the EBV expression with the germline FCGR2A single nucleotide polymorphism (SNP) rs1801274 found that the proportion of FCGR2A R patients in EBV+ patients (84.2%) was significantly higher than in EBV− patients (26.6%) (p = 0.001), with the following FCGR2A genotype distribution: 3 HH (19%), 10 HR (62%) and 3 RR (19%) for patients with a positive EBV load compared to 15 HH (71%), 4 HR (19%) and 2 RR (10%) for patients with a negative EBV load. The correlation between FCGR2A genotypes and LMP1 expression showed similar results [1]. The FCGR2A SNP rs1801274 is a non-synonymous SNP resulting in a histidine to arginine substitution at position 131, and FCγRIIA 131H variants have a higher affinity to human immunoglobulin (Ig) G1, G2 and G3 [2]. Diamantopoulos et al. hypothesized that NHL patients harboring FCγRIIA 131HH variants have better control of latency for EBV infection than patients with FCγRIIA 131R [1]. They also hypothesized that FCGR2A SNP probably does not interfere with the EBV acquisition, as the majority of the patients were seropositive for EBV based on anti-VCA-IgG antibody positivity (N=36, 90%). Intrigued by the results of this study [1], we investigated the role of FCGR2A SNP as well as the FCGR3A SNP (rs396991) in a cohort of 239 newly diagnosed Classical Hodgkin Lymphoma (CHL) patients age 18 years and older prospectively enrolled from 2002 to 2009 in the University of Iowa/Mayo Clinic SPORE (Specialized Program of Research Excellence) Molecular Epidemiology Resource [3]. In this cohort, 202 patients were Caucasian, and the remainder were non-Caucasian (n=6) or unknown (n=31). All were negative for human immunodeficiency virus. A peripheral blood sample was collected from all patients at diagnosis and DNA for genotyping was extracted using standard protocols. The FCGR2A SNP was genotyped as part of a larger project using a custom Illumina Infinium array (Illumina, San Diego, CA) and the FCGR3A SNP was genotyped using a custom designed pyrosequencing assay. This study was approved by the Human Subjects Institutional Review Board at Mayo Clinic and the University of Iowa for the SPORE study and all patients provided written consent for participation. Clinical characteristic of the series and FCGR2A and FCGR3A genotyping are presented in Table I. The EBV status of the tumor was established by EBER in situ hybridization for 104/239 patients (48%) with available tissue, and was positive for 23 (22%) and negative for 81 (78%) patients. Table I Clinical characteristics, FCGR2A and FCGR3A genotyping of classical Hodgkin lymphoma patients. Consistent with the observations of Diamantopoulos et al., we found a higher proportion of FCGR2A R patients in EBV+ CHL (N=20, 87%) compared to EBV− CHL (N=55, 68%) (p = 0.06). The FCGR2A HH, HR and RR genotype distributions were 26 (32%), 37 (46%) and 18 (22%) in EBV+ CHL compared to 3 (13%), 11 (48%), 9 (39%) in EBV− CHL (p = 0.04). To better decipher the specific correlation between FCGR2A and EBV-related CHL, we also analyzed as a “genetic control” the non-synonymous SNP rs396991 in FCGR3A, which is located on the same region of chromosome 1. We found that the FCGR3A genotype distribution was similar among EBV+ and EBV− CHL (Table I), with a minor allelic frequency (MAF, FCGR3A V allele) of 41% in EBV+ and 37% in EBV− CHL. Among 1521 controls included in SPORE, the FCGR2A HH, HR and RR genotype distribution was 387 (25%), 736 (48%) and 398 (26%), which is close to the distribution observed in EBV− CHL. The FCGR3A genotype distribution in controls was FCGR3A VV (N=647, 11%), VF (N=644, 44%), FF (N=167, 44%) with a MAF of 34% which was similar to the CHL patients overall or by EBV status. We conclude that EBV+ CHL patients in our study had a higher frequency of the low affinity FCγRIIA 131R which suggests an influence of the humoral response to EBV infection by FCγRIIA. CHL is one of the EBV-related lymphomas. The virus is present in Hodgkin and Reed-Sternberg cells in 20% to 50% of CHL cases in Western countries. The detection of EBV-encoded small RNAs (EBER) in malignant cells is considered to be the gold standard to determine the EBV status of the tumor [4]. EBV-related CHL typically expresses three viral proteins, EBV nuclear antigen 1 (EBNA-1), LMP1 and LMP2A. After an EBV infection, the number of infected B-cells decreases due to EBV-specific cytotoxic T-lymphocytes, and this immune response is thought to be influenced by host genetics (e.g., HLA class I region) [5]. Regarding the humoral response, the serologic profile after EBV infection is characterized first by the occurrence of anti-VCA, anti-EA and anti-EBNA-2 followed by the appearance of anti-EBNA-1 and a decrease of EBNA-2 [6]. It has been well documented that a symptomatic EBV infection is a risk factor for EBV+ CHL, but not for EBV− CHL [7]. Levin et al. recently reported that before EBV+ but not EBV− HL diagnosis, an aberrant serologic profile is observed with a reduced anti-EBNA-1 to EBNA-2 ratio, which is characteristic of a defective response to latent EBV infection [8]. In the latter study, higher anti-EBV VCA IgG antibody titers were observed before the EBV+ HL diagnosis, which could be associated with an increase in EBV replication [8]. Interestingly, in HL patients with active disease, higher EBNA-1 antibody titers were correlated with lower plasma EBV-DNA concentration, suggesting functional activities of these antibodies, although not sufficient for complete control of EBV [9]. The FCGR2A SNP influences bacterial infection susceptibility and also autoimmune disease in relation to the different capacities of FCγRIIA allotypes to clear immune complexes [10]. For viral infections, one study showed that FCγRIIA 131RR HIV subjects progressed faster than FCγRIIA 131H carriers to a CD4+ cell count below 200/mm3, and this impact on CD4+ cell count was independent of viral load [11]. They also observed that HIV-1 immune complexes were more efficiently internalized by FCγRIIA 131HH compared to RR subject monocytes, probably due to the presence of IgG2 in the immune complex, which have a low affinity for FCγRIIA 131RR [11]. While, FCγR-mediated inhibition seems important during HIV infection [12], to our knowledge, no study reported a similar association between EBV and FCGR2A genotype. We were also able to investigate whether the FCGR2A SNP influences CHL prognosis in the SPORE cohort. The median follow-up of whole cohort was 5 years (range, 0.04–9.75) and the 6-year event free survival (EFS) was 72.9% (95%CI, 64.8–79.7). EBV− patients had a higher EFS rate compared to EBV+ CHL in the SPORE cohort (71.9% [58–89] vs 29.0% [7–100], p = 0.007), noting that the EBV+ group was based on only 23 cases and 11 events. Using an ordinal (per R allele) model, FCGR2A did not influence the EFS of the whole series (Hazard ratio [HR]=0.91; 95%CI, 0.64–1.31; p = 0.63), EBV− CHL (HR=0.91; 95%CI, 0.48–1.74; p = 0.77) or EBV+ CHL (HR=0.77; 95%CI, 0.27–2.17; p = 0.62). No difference in EFS was observed when testing a dominant model for FCGR2A or for the FCGR3A SNP (data not shown). However, our study was limited by the low number of patients, especially in EBV+ group. Nevertheless, these first results suggest that the FCGR2A SNP is unlikely to have a role in the control of the disease. These exploratory results need to be validated in a larger series of CHL with EBV tumor status, as well as directly correlate genotypes with plasma EBV DNA load, as was done by Diamantopoulos et al [1]. While, we did not have EBV load, a recent study showed the majority of EBER positive HL had a high EBV DNA load in plasma [9]. In addition, it was suggested by Diamantopoulos et al. that FCGR2A SNP probably does not affect the risk of acquisition of EBV infection [1]. It would be interesting in the context of HL to consider the correlation between FCGR2A SNP and a complete EBV antibody profile before the diagnosis of HL in epidemiologic cohorts with banked pre-disease specimens [8] and a global screening of humoral immune responses to EBV [13]. A recent study showed that patients with a persistent or reappearance of EBV in plasma after initiation of treatment had an unfavorable outcome [14]. Whether the FCGR2A SNP impacts EBV load during the course of the treatment should also been considered. In conclusion, we observed that a higher proportion of EBV+ CHL patients carried FCGR2A RR genotype compared to EBV− CHL and unaffected controls, supporting a potential role of this SNP in the control of EBV-latent infection.