Abstract
We performed a genome-wide association study (GWAS) of antibody levels in a multi-ethnic group of 1,071 healthy smallpox vaccine recipients. In Caucasians, the most prominent association was found with promoter SNP rs10489759 in the LOC647132 pseudogene on chromosome 1 (p=7.77 × 10-8). In African-Americans, we identified eight genetic loci at p< 5 × 10-7. The SNP association with the lowest p-value (rs10508727, p=1.05 × 10-10) was in the Mohawk homeobox (MKX) gene on chromosome 10. Other candidate genes included LOC388460, GPR158, ZHX2, SPIRE1, GREM2, CSMD1, and RUNX1. In Hispanics, the top six associations between genetic variants and antibody levels had p-values less than 5 × 10-7, with p=1.78 × 10-10 for the strongest statistical association (promoter SNP rs12256830 in the PCDH15 gene). In addition, SNP rs4748153 in the immune response gene PRKCQ (protein kinase C, theta) was significantly associated with neutralizing antibody levels (p=2.51 × 10-8). Additional SNP associations in Hispanics (p ≤3.40 × 10-7) were mapped to the KIF6/LOC100131899, CYP2C9, and ANKLE2/GOLGA3 genes. This study has identified candidate SNPs that may be important in regulating humoral immunity to smallpox vaccination. Replication studies, as well as studies elucidating the functional consequences of contributing genes and polymorphisms, are underway.
Keywords: GWAS, Smallpox Vaccine, Vaccinia Virus, Humoral Immunity, Immunogenetics, SNPs
Introduction
Vaccination is the only defense against smallpox (variola major), an infectious disease with approximately a 30% mortality rate [1]. Despite eradication, the disease is “feared as a potential agent of bioterrorism because of its lethality, transmissibility, and the lack of known treatment” [2]. Therefore, immunity to smallpox after vaccination is an important issue for biodefense and for advancing our understanding of the immunogenetic regulation of the immune response. Immune response to smallpox vaccination can be highly variable among individuals. Genetic variations influence adaptive immune responses and play important roles in defining the type of host response generated by a vaccine. Associations between gene polymorphisms (SNPs) and variations in adaptive immune responses to smallpox vaccine are poorly understood. Closer study of the individual immune and genetic factors related to vaccinia virus-induced immunity is needed.
Smallpox vaccine (live vaccinia virus) has one of the highest complication rates among all the vaccines currently in use [3]. Previous studies of genetic predisposition for local and systemic adverse events (mostly fever) following primary smallpox vaccination revealed genetic variants (specific haplotypes in the IL1 and IL18 genes) associated with these adverse events [4]. Further, associations between adverse events (i.e., fever) after smallpox vaccine and polymorphisms in the 5,10-methylenetetrahydrofolate reductase (MTHFR), interferon regulatory factor-1 (IRF1), and IL4 genes were found in two independent studies [5].
Host genetics has been demonstrated to play a role in the variation in vaccine-induced immunity [6]. A number of human leukocyte antigen (HLA) alleles have been recently found to play a role in smallpox vaccine-induced immunity [7]. For example, B*4403 and B*4801 alleles are associated with lower neutralizing antibody titers, whereas alleles DQB1*0302 and DQB1*0604 are linked to higher antibody titers in individuals immunized with smallpox vaccine [7]. Another example is the association found between individual SNPs and haplotypes in the genes coding for IL18 and IL18R1 and vaccinia antibody titers [8]. These genetic associations were discovered using a candidate-gene approach.
Genome-wide association studies (GWAS), including population-based vaccination studies, are a powerful approach for discovery of novel genetic variants and links with immunity [9;10]. To identify additional host genetic factors associated with variations in humoral immune response to smallpox vaccine, we conducted a GWAS of smallpox vaccine in African-American, Caucasian, and Hispanic population samples and examined the association between SNPs and post-vaccination antibody titers. We hypothesized that other genes, beside HLA, cytokine, and cytokine receptor genes, may also be associated with smallpox vaccine-induced humoral immunity.
Materials and Methods
Study subjects
As previously described, our study cohort comprised a sample of 1,076 healthy subjects (age 18 to 40 years) who participated in both the US Department of Health and Human Services civilian healthcare worker smallpox immunization program at Mayo Clinic in Rochester, MN, and the smallpox immunization program at the US Naval Health Research Center (NHRC) in San Diego, CA [7;8;11]. Out of 1,076 subjects, 1,071 subjects had vaccinia neutralizing antibody and genotyping data available for this report. All study subjects received a single dose of live virus Dryvax vaccine (Wyeth Laboratories) at least one month, but no more than four years, earlier and had a documented vaccine “take,” development of a pustule, at the vaccination site. The Institutional Review Boards of both Mayo Clinic and NHRC approved the study, and written informed consent from each subject was obtained before enrollment.
Neutralizing antibody assay
We utilized a vaccinia-specific neutralization assay using β-galactosidase expressing vaccinia virus, as previously described [12;13]. Each serum sample was tested at least three times. Results are defined as the serum threshold dilution that inhibits 50% of virus activity (ID50), (estimated using the M estimation approach introduced by Huber) [14], which is robust to outliers and is implemented in the ROBUSTREG procedure of the SAS software package (Cary, NC). The coefficient of variation for this assay in our laboratory was 6.9%.
Genotyping and quality control
Infinium HumanHap650 Y BeadChip arrays were used to genotype SNPs in self-declared African-American subjects, as well as those who marked that they did not know their race or left the categories unmarked and a sampling of those who declared that they were of another race. All other subjects, the majority of whom were Caucasian, were genotyped using the Infinium HumanHap550 BeadChip array. DNA samples underwent whole genome amplification, fragmentation and hybridization onto each BeadChip, which were imaged on an Illumina BeadArray reader. Genotype calls based on clustering of the raw intensity data were made using the genotyping module of the BeadStudio 2 software. Genotype data on SNPs were generated by BeadStudio and transferred electronically to a server from which data were exported into SAS for further analysis. Quality control checks included genotyping reproducibility, gender checks, SNP and subject call rate cutoffs of > 0.95, elimination of monomorphic SNPs, and a Hardy-Weinberg Equilibrium (HWE) check, leaving 1,000 subjects for analysis. We accounted for potentially heterogeneous allele frequencies across racial groups by running race-stratified HWE tests, similar in spirit to that proposed by Schaid et al. [15].
Principal components analysis of race and ethnicity
The study set for this investigation was genetically diverse. We selected 22,863 SNPs with > 99% call rates, with inter-SNP distances of at least 100kb intervals, and used these SNPs in the principal components approach implemented in the Eigenstrat software package to assess population structure [16]. An initial set of genetic principal components was generated based on all eligible subjects in order to initially assess population genetic structure, both within and among racial groups. Because study subjects were racially diverse, and because they did not all provide a conclusive racial self-declaration, we utilized the genotype data from these SNPs to genetically classify individuals into Caucasian and African-American subgroups using a clustering approach similar to the one incorporated in the Structure program [16], but that utilized more potential axes of genetic variation in order to more completely capture the genetic complexity among study participants. These groups were compared to the available self-declared race data and the agreement between self-declared race and genetically classified racial grouping was significant. Of the 1,071 subjects with data in this report, 217 (20%) were classified as African-American and 580 (54%) as Caucasian. Two hundred and seventy-four (26%) subjects identified themselves as being multiracial or other.
Statistical methods
The statistical methods described here are similar to those carried out for our previous genetic association publications [7;8;11]. Serum antibody titers were tested at least three times for each individual and a subject-specific ID50 value was obtained for summary purposes using the median of these multiple measures. The neutralizing antibody ID50 levels along with other demographic and clinical variables, were summarized across individuals using frequencies and percentages for categorical variables, and medians and inter-quartile ranges for continuous variables.
Associations between neutralizing antibody titers and SNP genotypes were assessed using linear regression models with separate analyses performed within the Caucasian-, African- and Hispanic-American genetic subgroups. In the primary genome-wide analyses, SNPs were modeled according to an ordinal genotypic effect using a covariate that represented the number of copies of the minor allele carried by each individual. In contrast to the descriptive analyses, which used a single summary measure of ID50, formal statistical assessments used repeated measures analyses that modeled all of the multiple observations per subject while accounting for intra-subject correlations using generalized estimating equations (GEEs). All statistical tests of SNP-phenotype association adjusted for the following set of demographic and clinical variables: gender; age at blood draw (quartiles); time from smallpox immunization to blood draw (quartiles); time from blood draw to assay (quartiles); shipping temperature of the sample (frozen or ambient); time of year when the sample was shipped (warm-weather months April-September vs. cold-weather months October-March); and the first four eigenvectors obtained from within-race Eigenstrat population stratification analyses. Additionally, adjustments were made for assay operator as a covariate due to systematic differences among the different operators. A logarithmic transformation of the ID50 values was used in all linear regression analyses in order for the data to conform to statistical assumptions. After performing tests of significance for each SNP within racial groups, we constructed race-specific plots that compared the observed distribution of p-values to what is expected under the null hypothesis of no genetic associations being present. These plots are typically referred to as quantile-quantile, or “q-q” plots, as they plot the observed –log10(p-values) against values predicted from the ordered values from the reference distribution under the null hypothesis. Additionally, as there was evidence that the p-values might have been on average too small, genomic control lambda values were calculated from these observed-to-expected comparisons. These lambda values were then used to assess and correct for any inflation in the levels of significance, as described by Devlin and Roeder [17]. All statistical tests were performed to test a two-sided statistical alternative hypothesis. All genome-wide analyses were carried out using the R statistical software package [18].
The currently accepted p-value threshold to declare a SNP to be genome-wide significant is <5 × 10-8. Because our current study is comprised of fewer than 1,000 subjects in each of the race groups, and because we plan to replicate our findings in a subsequent study, we have chosen the less stringent, but still conservative significance threshold of <5 × 10-7 to consider a SNP to be of sufficient interest to highlight in this report. Additionally, in supplemental tables we include in results from SNPs with p-values less than 1×10-5 for further reference, as these SNPs will all be included in our follow-up replication study. Using these thresholds, several significant SNPs were identified for which few subjects were homozygous for the minor allele. To determine if the results from these individuals exerted undue influence on the overall results, their data were combined with that from subjects with one copy of the minor allele, and an additional series of models were run with the same adjusting factors as the ordinal model. The resulting estimates from these sensitivity analyses were compared to the per-allele estimates from the original analyses to verify their similarity.
Results
Demographic characteristics
A total of 1,071 healthy male and female subjects (18-40 years old) who had received a single dose of smallpox vaccine participated in our study, as previously described [7]. Of these, 790 (74%) were male and 281 (26%) were female. The average age at enrollment was 24 (interquartile range [IQR], 18-40) years, and the median time from smallpox vaccination to blood draw was 15.3 (IQR, 9.0-33.9) months. The median for vaccinia-specific neutralizing antibody ID50 values for all subjects was 133 (IQR, 79-206) [7]. Further summaries from the groups of subjects who comprised the racial and ethnic groups included in our analyses are included in Table 1.
Table 1.
Demographic characteristics and antibody levels in the study subjects
| African-American | Caucasian-American | Hispanic-American | ||||
|---|---|---|---|---|---|---|
| N | % | N | % | N | % | |
| Gender | ||||||
| Male | 174 | 80.2 | 422 | 72.8 | 150 | 69.1 |
| Female | 43 | 19.8 | 158 | 27.2 | 67 | 30.9 |
| Self-Declared Race | ||||||
| African American | 184 | 84.8 | 0 | 0.0 | 0 | 0.0 |
| Caucasian | 0 | 0.0 | 531 | 91.6 | 39 | 18.0 |
| More than one race | 19 | 8.8 | 34 | 5.9 | 32 | 14.7 |
| Other or do not know | 14 | 6.5 | 15 | 2.6 | 146 | 67.3 |
| Self-Declared Ethnicity | ||||||
| Not Hispanic | 179 | 82.5 | 544 | 93.8 | 25 | 11.5 |
| Hispanic | 14 | 6.5 | 24 | 4.1 | 188 | 86.6 |
| Do not know | 24 | 11.1 | 12 | 2.1 | 4 | 1.8 |
| Age Category | ||||||
| 19-21 | 49 | 22.6 | 133 | 22.9 | 48 | 22.1 |
| 22-23 | 39 | 18.0 | 156 | 26.9 | 54 | 24.9 |
| 24-27 | 54 | 24.9 | 167 | 28.8 | 64 | 29.5 |
| 28-40 | 75 | 34.6 | 124 | 21.4 | 51 | 23.5 |
| Time from Immunization to Blood Draw | ||||||
| 1-8 months | 53 | 24.4 | 149 | 25.6 | 49 | 22.6 |
| 9-14 months | 46 | 21.2 | 147 | 25.3 | 58 | 26.7 |
| 15-33 months | 44 | 20.3 | 161 | 27.8 | 60 | 27.6 |
| 34-48 months | 74 | 34.1 | 123 | 21.2 | 50 | 23.0 |
| Mean | St. Dev. | Mean | St. Dev. | Mean | St. Dev. | |
| Antibody Level (ID50) a | 168.6 | 118.7 | 162.5 | 128.8 | 153.5 | 100.5 |
| Median | IQR | Median | IQR | Median | IQR | |
| 132.5 | 80.1-226.6 | 131.2 | 76.9-202.2 | 135.5 | 80.0-190.9 | |
St. Dev, standard deviation; IQR, interquartile range.
Values are in ID50 values as measured by the β-galactosidase based vaccinia-specific neutralization assay.
Genetic associations with humoral immune responses
Analysis of the genome-wide association data identified a number of SNPs associated with variations in antibody levels at p<5 × 10-7. The q-q plots of the observed versus expected –log10(p-values) for neutralizing antibody levels are depicted in Figure 1, separately by racial/ethnic group (African-Americans, Caucasians and Hispanics). These plots displayed a slight inflation in the number of significant tests, therefore the race/ethnicity-specific inflation factors were used to correct the SNP-specific p-values. The Manhattan plots shown in Figure 2 illustrate the regions where significant SNPs are located. While these findings are encouraging, the need to formally replicate these observations remains.
Figure 1.
Quantile-quantile plots of the expected (x-axis) and observed (y-axis) -log10(p-value) in smallpox vaccine response GWAS. A) Results for the African-American cohort and neutralizing antibody outcome. B) Results for the Caucasian cohort and neutralizing antibody outcome, and C) Results for Hispanic cohort and neutralizing antibody outcome. If no SNPs are associated with the outcome, and if no biases are unaccounted for, the points would be expected to lie along the plotted diagonal line. Of note, several SNPs reached high levels of significance with p-values less than 10-10.
Figure 2.
Manhattan plot summaries of GWAS results for humoral immune response outcome. The x-axis displays the –log10 of the p-value for each SNP association and the y-axis displays the chromosomes in alternating black and gray. A) Results for the African-American cohort and neutralizing antibody outcome. B) Results for the Caucasian cohort and neutralizing antibody outcome, and C) Results for Hispanic cohort and neutralizing antibody outcome. Chromosomal distribution and p-values of each SNP association with the neutralizing antibody levels are shown.
In Caucasians, a promoter SNP, rs10489759, from the LOC647132 DEAH (Asp-Glu-Ala-His) box polypeptide 29 pseudogene on chromosome 1 demonstrated a significant association with vaccinia-specific antibody levels (p=7.77 × 10-8) (Table 2). In addition, ten SNPs in either immune function or other genes (including the calcium channel-related gene, CACNA2D3, transmembrane and tetratricopeptide repeat containing 2, TMTC2, and serotonin-related transcription factor for the fifth Ewing variant, FEV) were found to be associated (at a significance level of p<1 × 10-5) with humoral immune responses in Caucasians (see Supplemental Table 1).
Table 2.
Single-nucleotide polymorphisms (SNPs) associated with vaccinia neutralizing antibody responses in Caucasians
| SNP ID | Chromosome | Gene | Location | Distance from Genea | MAF | Genotypeb | Nb | Median Ab level ID50 (IQR)c | p-valued |
|---|---|---|---|---|---|---|---|---|---|
| rs10489759 | 1 | LOC647132 | 3′downstream | 369241 | 12.0 | GG | 505 | 133.6 (79.3, 207.5) | 7.77E-08 |
| GA | 12 | 59.6 (52.9, 74.8) | |||||||
| AA | 0 | -- |
MAF, minor allele frequency; Ab, antibody; ID50, serum dilution that inhibits 50% of virus activity; IQR, interquartile range.
A p-value threshold of 5 × 10-7 was used.
--No subject in that genotype.
Location of the SNP relative to the gene or distance to nearest gene.
Values are presented as homozygous major allele/heterozygous/homozygous minor allele.
IQR, interquartile range. Values are in ID50 values as measured by the β-galactosidase based vaccinia-specific neutralization assay.
Ordinal p-value from the analysis of variance adjusting for age, gender, time from immunization to blood draw, time from blood draw to assay, season and temperature in which samples were shipped, operator for the assay, and the first four principal components.
The GWAS analysis in African-Americans identified eight top scoring associations between SNPs and vaccinia antibody levels with p< 5 × 10-7 (Table 3). The SNP association with the lowest p-value (rs10508727, p=1.05 × 10-10) was in the Mohawk homeobox (MKX) gene on chromosome 10. Two promoter SNPs, rs10503951 (p=3.38 × 10-9) and rs12775535 (p=3.97 × 10-9) within the LOC388460 and GPR158 (G protein-coupled receptor 158) genes, respectively, were strongly associated with antibody levels in African-Americans. Five other intronic SNP associations (p≤4.79 × 10-7) in African-Americans were found in the ZHX2 (zinc fingers and homeoboxes 2), SPIRE1 (spire homolog 1), GREM2 (gremlin 2), CSMD1 (CUB and Sushi multiple domains 1), and RUNX1 (runt-related transcription factor 1) genes. Further, eight additional SNPs, including a promoter rs3804795 in the immune function IL5RA (interleukin 5 receptor alpha) gene, were found to be associated (at a significance level of p<1 × 10-6) with humoral immune responses in the GWAS in African-Americans (Supplemental Table 2). We have also identified 82 SNPs associated (at a significance level of p<1 × 10-5) with variations in antibody levels in African-Americans (data not shown due to space constraints).
Table 3.
Single-nucleotide polymorphisms (SNPs) associated with vaccinia neutralizing antibody responses in African-Americans
| SNP ID | Chromosome | Gene | Location | Distance from Genea | MAF | Genotypeb | Nb | Median Ab level ID50 (IQR)c | p-valued |
|---|---|---|---|---|---|---|---|---|---|
| rs10508727 | 10 | MKX | intron | 0 | 12.0 | GG | 192 | 127.7 (78.1, 225.1) | 1.05E-10 |
| GA | 12 | 257.9 (233.2, 295.3) | |||||||
| AA | 0 | -- | |||||||
| rs10503951 | 8 | LOC388460 | 3′downstream | 51218 | 12.0 | AA | 192 | 127.7 (78.1, 226.3) | 3.38E-09 |
| AG | 12 | 282.5 (222.5, 376) | |||||||
| GG | 0 | -- | |||||||
| rs12775535 | 10 | GPR158 | 3′downstream | 27721 | 16.0 | GG | 184 | 126.6 (77.3, 225.3) | 3.97E-09 |
| GA | 16 | 226.8 (168.7, 302.8) | |||||||
| AA | 0 | -- | |||||||
| rs10108684 | 8 | ZHX2 | intron | 0 | 12.0 | GG | 191 | 138.0 (81.4, 234.6) | 1.36E-08 |
| GA | 12 | 80.5 (64.9, 93.2) | |||||||
| AA | 0 | -- | |||||||
| rs9959145 | 18 | SPIRE1 | intron | 0 | 20.0 | CC | 180 | 127.0 (79.7, 210.7) | 2.53E-08 |
| CA | 20 | 297.2 (213.8, 348.0) | |||||||
| AA | 0 | -- | |||||||
| rs10495471 | 1 | GREM2 | intron | 0 | 28.0 | AA | 177 | 142.3 (81.0, 249.8) | 1.57E-07 |
| AG | 24 | 94.2 (65.0, 128.4) | |||||||
| GG | 2 | 46.7 (46.5, 46.9) | |||||||
| rs17070309 | 8 | CSMD1 | intron | 0 | 56.0 | GG | 152 | 117.4 (76.2, 210.7) | 4.65E-07 |
| GA | 48 | 169.0 (125.2, 274.8) | |||||||
| AA | 4 | 358.2 (318.3, 410.5) | |||||||
| rs2834655 | 21 | RUNX1 | intron | 0 | 134.0 | GG | 90 | 102.0 (67.0, 191.0) | 4.79E-07 |
| GA | 92 | 161.8 (90.9, 261.0) | |||||||
| AA | 21 | 161.8 (130.5, 292.3) |
MAF, minor allele frequency; Ab, antibody; ID50, serum dilution that inhibits 50% of virus activity; IQR, interquartile range.
A p-value threshold of 5 × 10-7 was used. A sensitivity analysis p-value = 1.75E-07 for SNP rs10495471. The sensitivity analysis p-value = 1.61E-05 for SNP rs17070309.
--No subject in that genotype.
Location of the SNP relative to the gene or distance to nearest gene.
Values are presented as homozygous major allele/heterozygous/homozygous minor allele.
IQR, interquartile range. Values are in ID50 values as measured by the β-galactosidase based vaccinia-specific neutralization assay.
Ordinal p-value from the analysis of variance adjusting for age, gender, time from immunization to blood draw, time from blood draw to assay, season and temperature in which samples were shipped, operator for the assay, and the first four principal components
SNPs that displayed significant associations (p<5 × 10-7) with vaccinia antibody levels in Hispanics are shown in Table 4. The top five associations between genetic variants and antibody levels had p-values less than 3.40 × 10-7, with p=1.78 × 10-10 for the strongest statistical association (promoter SNP rs12256830 in the PCDH15 [protocadherin 15] gene). Another intronic SNP, rs4748153 in the PRKCQ (protein kinase C, theta) gene was significantly associated with neutralizing antibodies (p=2.51 × 10-8). Additional SNP associations in Hispanics (p≤ 3.40 × 10-7) were mapped to the KIF6 (kinesin family member 6)/LOC100131899 (pseudogene), CYP2C9 (cytochrome P450), and ANKLE2 (ankyrin repeat and LEM domain containing 2)/GOLGA3 genes. In Hispanics, we also identified several genes/SNPs (including the CYP2C8, CRBN and SNX18 genes) that demonstrated an association p-value <1 × 10-6 that require replication (Supplemental Table 3). In addition, 55 SNPs were found to be associated (p<1 × 10-5) with humoral immune responses in Hispanics (data not shown).
Table 4.
Single-nucleotide polymorphisms (SNPs) associated with vaccinia neutralizing antibody responses in Hispanics
| SNP ID | Chromosome | Gene | Location | Distance from Genea | MAF | Genotypeb | Nb | Median Ab level ID50 (IQR)c | p-valued |
|---|---|---|---|---|---|---|---|---|---|
| rs12256830 | 10 | PCDH15 | 3′downstream | 196117 | 12.0 | GG | 188 | 137.3 (86.9, 201.3) | 1.78E-10 |
| GA | 12 | 54.9 (51.6, 74.5) | |||||||
| AA | 0 | -- | |||||||
| rs4748153 | 10 | PRKCQ | Intron | 0 | 26.0 | GG | 174 | 122.1 (75.4, 173.0) | 2.51E-08 |
| GA | 26 | 205.4 (155.9, 306.4) | |||||||
| AA | 0 | -- | |||||||
| rs9380880 | 6 | KIF6; LOC100131899 | Intron; 3′downstream | 0; 6263 | 11.0 | GG | 189 | 137.8 (84.7, 201.0) | 1.24E-07 |
| GA | 11 | 79.0 (61.1, 101.4) | |||||||
| AA | 0 | ||||||||
| rs2860975 | 10 | CYP2C9 | 3′downstream | 17787 | 196.0 | CC | 52 | 148.5 (114.8, 252.8) | 1.80E-07 |
| CA | 98 | 147.6 (85.9, 200.1) | |||||||
| AA | 49 | 88.3 (54.1, 139.1) | |||||||
| rs12282 | 12 | ANKLE2 or GOLGA3 | 5′upstream; 3′UTR | 7358; 0 | 17.0 | GG | 181 | 126.9 (76.4, 188.8) | 3.40E-07 |
| GA | 15 | 172.6 (149.4, 331.1) | |||||||
| AA | 1 | 306.9 (306.9, 306.9) |
MAF, minor allele frequency; Ab, antibody; ID50, serum dilution that inhibits 50% of virus activity; IQR, interquartile range.
A p-value threshold of 5 × 10-7 was used. A sensitivity analysis p-value = 2.66E-06 for SNP rs12282.
--No subject in that genotype.
Location of the SNP relative to the gene or distance to nearest gene.
Values are presented as homozygous major allele/heterozygous/homozygous minor allele.
IQR, interquartile range. Values are in ID50 values as measured by the β-galactosidase based vaccinia-specific neutralization assay.
Ordinal p-value from the analysis of variance adjusting for age, gender, time from immunization to blood draw, time from blood draw to assay, season and temperature in which samples were shipped, operator for the assay, and the first four principal components.
Discussion
In this study we identified polymorphisms in both immune response and other genes, such as: MKX gene (a regulator of collagen expression and tendon differentiation) [19]; GPR158 (G protein-coupled receptor 158 gene that is a susceptibility locus for Alzheimer disease and esophageal cancer) [20]; a transcriptional repressor and regulator of alpha-fetoprotein metabolism, ZHX2 (zinc fingers and homeoboxes 2 gene) [21-23]; and CSMD1 (CUB and Sushi domain-containing protein 1, which is a potential tumor suppressor), which are associated with variations in antibody results in African-Americans, as well as nonfunctional pseudogenes with unknown protein-coding abilities, such as LOC647132 and LOC388460 associated with variations in antibody levels in the Caucasian and African-American subjects, respectively. We also found polymorphisms in novel genes with no previously identified immune function, such as SPIRE1 (the spire homolog 1 conserved gene is involved in actin organization and intracellular transport) [24] and GREM2 (gremlin 2 gene mapped to chromosome 1q43) associated with variations in antibody levels in African-American subjects.
In addition, we identified several polymorphisms in immune response genes, such as: PCDH15 (protocadherin-related 15 or deafness gene that is also a susceptibility locus for both Usher syndrome and lipid abnormalities) [25-27]; PRKCQ (protein kinase C theta type with known T cell activation and IL-4 regulation properties) [28]; KIF6 (kinesin-like protein 6 gene that is involved in cardiovascular disease) [29;30]; and CYP2C9 (cytochrome P450 enzyme gene that is important for drug metabolism) [31]. These polymorphisms are associated with variations in vaccinia antibody levels in Hispanics, as well as novel genes with no previously identified immune function, such as the ANKLE2 (ankyrin repeat and LEM domain containing 2) gene. These results require replication and studies designed to validate the results and elucidate functional mechanisms behind the immune phenotypes.
As can be seen in Tables 2-4 and Figures 1-2, there are SNPs on chromosome 10 that have very small p-values (rs10508727, p=1.05 × 10-10 for African-Americans and rs12256830, p=1.78 × 10-10 for Hispanics) that require further study. As a starting point, the following gene SNPs are top candidates for further study given their known role in immune function: SPIRE1 (rs9959145), PRKCQ (rs4748153), GPR158 (rs12775535), ZHX2 (rs10108684), and CYP2C9 (rs2860975). It has been recently shown that the signaling lymphocytic activation molecule (SLAM)-mediated increase in IL-4 secretion in CD4+ T cells requires protein kinase C theta (PRKCQ) [28]. In our Hispanic subgroup, the presence of a homozygous major allele genotype for PRKCQ rs4748153 was associated with a decrease in vaccinia-specific antibody response, and this association was highly significant (p=2.51 × 10-8). One possible mechanism behind the association between PRKCQ and antibody response might be the influence of T cell receptor (TCR)-modulated IL-4 secretion by protein kinase C theta and concurrent up- or down-regulation of antibody production following vaccinia virus exposure [28]. Other immunogenetic studies have examined the role of genetic polymorphisms in smallpox vaccine responses. We previously discovered significant associations with HLA alleles, SNPs, and haplotypes in IL18 and IL18R1 genes and humoral immunity after a single dose of smallpox vaccination [7;8]. Reif et al. [5] identified a number of SNPs in the IRF-1 and MTHFR genes associated with adverse events after smallpox vaccination. Similarly, Stanley et al. [4] examined genetic associations with the development of fever after smallpox immunization and identified haplotypes in the IL1 gene complex and the IL18 gene that may predict the development of fever following vaccination. Both of these studies focused on specific SNPs within candidate genes. Of note, GWAS analysis in Caucasians, African-Americans and Hispanics revealed that our previously reported candidate IL18/IL18R1 gene SNPs [8] associated with antibody levels (Caucasian cohort: IL18R1 rs1035130, p=0.0006, African-American cohort: IL18R1 rs1035130, p=0.01; IL18 rs2043055, p=0.004; IL18 rs5744280, p=0.004; IL18R1 rs3771166, p=0.006, Hispanic cohort: IL18R1 rs1035130, p=0.049) did not reach genome-wide significance (p<5 × 10-7). To the best of our knowledge, the data reported here represent the first genome-wide SNP association study conducted on the antibody response to smallpox vaccine. As seen in Tables 2-4, our data indicate that many SNPs in multiple genes show significant associations with vaccinia-specific neutralizing antibody levels following smallpox vaccination. Examples include: 1) the presence of several promoter SNPs, which most likely tag causal SNPs, mapping to translationally controlled pseudogenes or gene deserts (LOC647132, LOC388460, andKIF6/LOC100131899), all associated with increased antibody responses in Caucasians, African-Americans and Hispanics, respectively, 2) Major alleles for SNPs in the MKX, GPR158, SPIRE1, CSMD1, and RUNX1 genes associated with decreased antibody levels in African-American subjects, and 3) Major alleles for SNPs in the PCDH15, KIF6, and CYP2C9 genes associated with increased antibody levels in Hispanics. While a small number of the genes found in our study have known immune functions, many do not. An essential next step is replication of these findings in order to eliminate false-positives and validate key findings so they may be further investigated in functional and mechanistic studies. We are currently conducting a GWAS replication study in a new replication cohort of 1,300 subjects following smallpox vaccination. In our replication study we will also carry forward a number of SNPs that did not reach our significance threshold (p-values <5 × 10-7) in order to evaluate smaller effects which were not detectable after our multiple testing correction for the many statistical tests performed in this first genome-wide assessment. Among these are SNPs in the calcium-gated channel (CACNA2D3), interleukin 5 receptor alpha (IL5RA), RAB10 (a member of the RAS oncogene family), cereblon (CRBN), serotonon-linked fifth Ewing variant (FEV) and other genes with known or unknown function (Supplemental Tables 1-3). Notably, we found that a polymorphism in the CACNA2D3 calcium-gated channel gene was associated with an allele dose-related increase (1.5-fold) in antibody levels (rs 4955826, p=4.9 × 10-6) in Caucasians. However, this association did not reach our pre-specified level of significance (p<5 × 10-7). Since the CACNA2D3, CACNA1E, and CACNA1H gene products have a possible role in the mechanisms of hypertension and blood pressure regulation [32;33], genetic polymorphisms that alter this gene's expression could potentially have downstream effects on cardiovascular function. A rare but potentially harmful side effect of the smallpox vaccine is myopericarditis – inflammation of the heart muscle and/or its lining [34]. Genetic factors that affect a person's risk of developing myopericarditis are not known. It is plausible that polymorphisms in the CACNA2D3 and other genes and pathways have a biologic role in cardiac abnormalities. In African-Americans, we also found several host genetic polymorphisms that exhibited associations at lower levels of significance. We found SNPs in the RBM24 gene (that is important for cardiovascular development) [35], mitogen activated protein kinase (MAPK) signaling pathway-associated gene (such as RASAL2), VENT homeobox pseudogene 5 (VENTXP5), and IL5RA gene. The precise role of some of these genetic loci in vaccinia-induced humoral immunity is unclear, while IL-5-induced B-cell proliferation and immunoglobulin production is a well-known factor [36]. IL-5 binds to the alpha chain of IL-5R and is linked with Janus Kinase 2 (JAK2), which is critical in IL-5 signal transduction [37]. Importantly, we found a promoter SNP in the IL5RA gene associated with an allele dose-related increase (2.3-fold) in antibody levels (rs3804795, p=9.5 × 10-7) in African-Americans. Similarly, in Hispanics we found gene polymorphisms that demonstrated associations that did not quite reach statistical significance with vaccinia-induced humoral immunity: a member of the cytochrome P450 enzyme family (CYP2C8) that is involved in drug metabolism and lipid synthesis, mental retardation-related gene cereblon (CRBN) [38], sorting nexin 18 gene (SNX18) that is involved in intracellular trafficking, and other genes.
Several strengths and limitations of this study must be acknowledged. This is the first genome-wide study of responses to smallpox vaccine involving a cohort of more than 1,000 recently vaccinated subjects. The major strengths of our study design include the recruitment of a large number of subjects who received their smallpox vaccination months to years earlier and had a documented smallpox vaccine vesicular “take.” A standardized and sensitive neutralization assay was used to characterize humoral responses [12;13]. The primary limitation of this study is the limited samples sizes for the primary analyses within race/ethnic categories. The small sample sizes may be responsible for the differences in the numbers of SNPs reaching significance thresholds among the Caucasian, African-American and Hispanic cohorts and could result in the presence of excess false positive results in the smaller study groups. However, the differences may be due to other causes such as differences in minor allele frequency. The number of African-American subjects (n=217) included in this smallpox vaccine immunogenetics study is only moderate in size for a GWAS study, and the number of Hispanics was comparable to this. Nevertheless, in both African-Americans and Hispanics we found significant associations that were not perturbed by sensitivity analyses in eight and five genes, respectively, with both known and unknown immune functions. We also utilized genomic data to summarize the degree of population admixture present among our study subjects, and to classify individuals who did not self-declare conclusive racial/ethnic categories. The agreement between self-declared race/ethnicity and genomically-defined race/ethnicity was high. Furthermore, we assessed and controlled for residual population admixture in our analyses of race/ethnic subgroups. The functional effects of these polymorphisms must be further examined to better understand genetic contribution to inter-individual variability in smallpox vaccine-induced humoral immune response.
Multiple statistical tests were performed for this genome-wide association study; therefore, it is likely that a number of false positive findings have been identified, and it is possible that the likelihood for this is higher in our smaller study cohorts. Two points mitigate this limitation. First, we employed a stringent level of significance (p<5 × 10-7) that accounts for the large number of statistical tests performed. Second, we have evaluated several targeted SNPs in the IL18 and IL18R1 SNPs and have found that the magnitude of the associations have been similar, although the levels of statistical significance differed among the race/ethnicity groups. Nonetheless, it is certainly biologically plausible that SNPs associated with variations in immune response may vary between racial/ethnic populations.
In conclusion, this is the first GWAS study that provides evidence for a genetic contribution to inter-individual variation in antibody response after smallpox vaccination. Molecular pathways for the effect of these identified set of genes remain to be discovered. Replication studies, as well as studies elucidating the functional mechanisms of causative genes and genetic polymorphisms, are underway. Ultimately, this knowledge will heighten our understanding of immune mechanisms and define biomarkers of immunity that can help in optimizing the development of new vaccines, diagnostic tests, and therapeutics to protect humans from smallpox.
Supplementary Material
Highlights.
Vaccination is the only defense against smallpox.
Smallpox vaccine has one of the highest complication rates among all the vaccines currently in use.
This is the first genome-wide study of responses to smallpox vaccine involving a large cohort of recently vaccinated subjects.
In Caucasians, the most prominent association was found with promoter SNP in the LOC647132 pseudogene, p=7.77 × 10-8.
In African-Americans, we identified eight genetic loci at p<5 × 10-7.
In Hispanics, the top six associations between genetic variants and antibody levels had p-values <5 × 10-7.
Acknowledgments
We thank the Mayo Clinic Vaccine Research Group in Rochester, Minnesota, the Naval Health Research Center in San Diego, California and the subjects who participated in our studies. We thank Robert A. Vierkant and David A. Watson for their assistance in statistical analysis and Julie M. Cunningham and the Mayo Advanced Genomic Technology Center for assistance with genotyping. We gratefully acknowledge Caroline L. Vitse for her editorial assistance in preparing this article.
Funding
This project was funded by federal funds from the National Institute of Allergies and Infectious Diseases, National Institutes of Health, Department of Health and Human Services, under Contract No. HHSN266200400065C.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of interest statement
None reported.
References
- 1.Bourzac K. Smallpox: historical review of a potential bioterrorist tool. Journal of Young Investigators. 2002;(3) [Google Scholar]
- 2.Henderson DA, Inglesby TV, Bartlett JG, Ascher MS, Eitzen E, Jahrling PB, et al. Smallpox as a biological weapon: medical and public health management. JAMA. 1999 Jun 9;281(22):2127–37. doi: 10.1001/jama.281.22.2127. [DOI] [PubMed] [Google Scholar]
- 3.Cono J, Casey CG, Bell DM. Smallpox vaccination and adverse reactions. Guidance for clinicians. MMWR. 2003 Feb 21;52(RR-4):1–28. [PubMed] [Google Scholar]
- 4.Stanley SL, Jr., Frey SE, Taillon-Miller P, Guo J, Miller RD, Koboldt DC, et al. The immunogenetics of smallpox vaccination. J Infect Dis. 2007 Jul 15;196(2):212–9. doi: 10.1086/518794. [DOI] [PubMed] [Google Scholar]
- 5.Reif DM, McKinney BA, Motsinger AA, Chanock SJ, Edwards KM, Rock MT, et al. Genetic basis for adverse events after smallpox vaccination. J Infect Dis. 2008 May 2;198(1):16–22. doi: 10.1086/588670. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Poland GA, Ovsyannikova IG, Jacobson RM, Smith DI. Heterogeneity in vaccine immune response: the role of immunogenetics and the emerging field of vaccinomics. Clin Pharmacol Ther. 2007 Dec;82(6):653–64. doi: 10.1038/sj.clpt.6100415. [DOI] [PubMed] [Google Scholar]
- 7.Ovsyannikova IG, Vierkant RA, Pankratz VS, Jacobson RM, Poland GA. Human leukocyte antigen genotypes in the genetic control of adaptive immune responses to smallpox vaccine. J Infect Dis. 2011 Jun;203(11):1546–55. doi: 10.1093/infdis/jir167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Haralambieva IH, Ovsyannikova IG, Dhiman N, Kennedy RB, O'Byrne M, Pankratz VS, Jacobson RM, Poland GA. Common SNPs/haplotypes in IL18R1 and IL18 genes are associated with variations in humoral immunity to smallpox vaccination in Caucasians and African-Americans. J Infect Dis. 2011 Aug 1;204(3):433–41. doi: 10.1093/infdis/jir268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wellcome Trust Case Control Consortium Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature. 2007 Jun 7;447(7145):661–78. doi: 10.1038/nature05911. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hirschhorn JN, Daly MJ. Genome-wide association studies for common diseases and complex traits. Genetics. 2005 Feb 6;(2):95–108. doi: 10.1038/nrg1521. [DOI] [PubMed] [Google Scholar]
- 11.Kennedy RB, Ovsyannikova IG, Pankratz VS, Vierkant RA, Jacobson RM, Ryan MA, Poland GA. Gender effects on humoral immune responses to smallpox vaccine. Vaccine. 2009 May 26;27(25-26):3319–23. doi: 10.1016/j.vaccine.2009.01.086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Manischewitz J, King LR, Bleckwenn NA, Shiloach J, Taffs R, Merchlinsky M, et al. Development of a novel vaccinia-neutralization assay based on reporter-gene expression. J Infect Dis. 2003 Aug 1;188(3):440–8. doi: 10.1086/376557. [DOI] [PubMed] [Google Scholar]
- 13.Kennedy R, Pankratz VS, Swanson E, Watson D, Golding H, Poland GA. Statistical approach to estimate vaccinia- specific neutralizing antibody titers using a high throughput assay. Clin Vaccine Immunol. 2009 Jun 17;16(8):1105–12. doi: 10.1128/CVI.00109-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Huber PJ. Robust Regression: Asymptotics, Conjectures and Monte Carlo. The Annals of Statistics. 1973;1(5):799–821. [Google Scholar]
- 15.Schaid DJ, Batzler AJ, Jenkins GD, Hildebrandt MA. Exact tests of Hardy-Weinberg equilibrium and homogeneity of disequilibrium across strata. Am J Hum Genet. 2006 Dec;79(6):1071–80. doi: 10.1086/510257. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006 Aug;38(8):904–9. doi: 10.1038/ng1847. [DOI] [PubMed] [Google Scholar]
- 17.Devlin B, Roeder K. Genomic control for association studies. Biometrics. 1999 Dec;55(4):997–1004. doi: 10.1111/j.0006-341x.1999.00997.x. [DOI] [PubMed] [Google Scholar]
- 18.R Software (computer program) R Development Core Team . R: A language and environment for statistical computing. R Foundation for Statistical Computing; Vienna, Austria: 2011. ISBN 3-900051-07-0, URL http://www.R-project.org/ [Google Scholar]
- 19.Ito Y, Toriuchi N, Yoshitaka T, Ueno-Kudoh H, Sato T, Yokoyama S, et al. The Mohawk homeobox gene is a critical regulator of tendon differentiation. Proc Natl Acad Sci U S A. 2010 Jun 8;107(23):10538–42. doi: 10.1073/pnas.1000525107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Oka D, Yamashita S, Tomioka T, Nakanishi Y, Kato H, Kaminishi M, et al. The presence of aberrant DNA methylation in noncancerous esophageal mucosae in association with smoking history: a target for risk diagnosis and prevention of esophageal cancers. Cancer. 2009 Aug 1;115(15):3412–26. doi: 10.1002/cncr.24394. [DOI] [PubMed] [Google Scholar]
- 21.Kawata H, Yamada K, Shou Z, Mizutani T, Yazawa T, Yoshino M, et al. Zinc-fingers and homeoboxes (ZHX) 2, a novel member of the ZHX family, functions as a transcriptional repressor. Biochem J. 2003 Aug 1;373(Pt 3):747–57. doi: 10.1042/BJ20030171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Peterson ML, Ma C, Spear BT. Zhx2 and Zbtb20: novel regulators of postnatal alpha-fetoprotein repression and their potential role in gene reactivation during liver cancer. Semin Cancer Biol. 2011 Feb;21(1):21–7. doi: 10.1016/j.semcancer.2011.01.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gargalovic PS, Erbilgin A, Kohannim O, Pagnon J, Wang X, Castellani L, et al. Quantitative trait locus mapping and identification of Zhx2 as a novel regulator of plasma lipid metabolism. Circ Cardiovasc Genet. 2010 Feb;3(1):60–7. doi: 10.1161/CIRCGENETICS.109.902320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kerkhoff E, Simpson JC, Leberfinger CB, Otto IM, Doerks T, Bork P, et al. The Spir actin organizers are involved in vesicle transport processes. Curr Biol. 2001 Dec 11;11(24):1963–8. doi: 10.1016/s0960-9822(01)00602-9. [DOI] [PubMed] [Google Scholar]
- 25.Aller E, Jaijo T, Garcia-Garcia G, Aparisi MJ, Blesa D, Díaz-Llopis M, et al. Identification of large rearrangements of the PCDH15 gene by combined MLPA and a CGH: large duplications are responsible for Usher syndrome. Invest Ophthalmol Vis Sci. 2010 Nov;51(11):5480–5. doi: 10.1167/iovs.10-5359. [DOI] [PubMed] [Google Scholar]
- 26.Huertas-Vazquez A, Plaisier CL, Geng R, Haas BE, Lee J, Greevenbroek MM, et al. A nonsynonymous SNP within PCDH15 is associated with lipid traits in familial combined hyperlipidemia. Hum Genet. 2010 Jan;127(1):83–9. doi: 10.1007/s00439-009-0749-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Jacobson SG, Cideciyan AV, Aleman TS, Sumaroka A, Roman AJ, Gardner LM, et al. Usher syndromes due to MYO7A, PCDH15, USH2A or GPR98 mutations share retinal disease mechanism. Hum Mol Genet. 2008 Aug 1;17(15):2405–15. doi: 10.1093/hmg/ddn140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Cannons JL, Wu JZ, Gomez-Rodriguez J, Zhang J, Dong B, Liu Y, et al. Biochemical and genetic evidence for a SAP-PKC-theta interaction contributing to IL-4 regulation. J Immunol. 2010 Sep 1;185(5):2819–27. doi: 10.4049/jimmunol.0902182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Iakoubova O, Shepherd J, Sacks F. Association of the 719Arg variant of KIF6 with both increased risk of coronary events and with greater response to statin therapy. J Am Coll Cardiol. 2008 Jun 3;51(22):2195–6. doi: 10.1016/j.jacc.2008.02.061. [DOI] [PubMed] [Google Scholar]
- 30.Iakoubova OA, Sabatine MS, Rowland CM, Tong CH, Catanese JJ, Ranade K, et al. Polymorphism in KIF6 gene and benefit from statins after acute coronary syndromes: results from the PROVE IT-TIMI 22 study. J Am Coll Cardiol. 2008 Jan 29;51(4):449–55. doi: 10.1016/j.jacc.2007.10.017. [DOI] [PubMed] [Google Scholar]
- 31.Rettie AE, Jones JP. Clinical and toxicological relevance of CYP2C9: drug-drug interactions and pharmacogenetics. Annu Rev Pharmacol Toxicol. 2005;45:477–94. doi: 10.1146/annurev.pharmtox.45.120403.095821. [DOI] [PubMed] [Google Scholar]
- 32.Hayashi K, Wakino S, Sugano N, Ozawa Y, Homma K, Saruta T. Ca2+ channel subtypes and pharmacology in the kidney. Circ Res. 2007 Feb 16;100(3):342–53. doi: 10.1161/01.RES.0000256155.31133.49. [DOI] [PubMed] [Google Scholar]
- 33.Adeyemo A, Gerry N, Chen G, Herbert A, Doumatey A, Huang H, et al. A genome-wide association study of hypertension and blood pressure in African Americans. PLoS Genet. 2009 Jul;5(7):e1000564. doi: 10.1371/journal.pgen.1000564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Halsell JS, Riddle JR, Atwood JE, Gardner P, Shope R, Poland GA, et al. Myopericarditis following smallpox vaccination among vaccinia-naive US military personnel. JAMA. 2003;289:3283–9. doi: 10.1001/jama.289.24.3283. [DOI] [PubMed] [Google Scholar]
- 35.Maragh S, Miller RA, Bessling SL, McGaughey DM, Wessels MW, de Graaf B, et al. Identification of RNA binding motif proteins essential for cardiovascular development. BMC Dev Biol. 2011 Oct 19;11(1):62. doi: 10.1186/1471-213X-11-62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Dubucquoi S, Desreumaux P, Janin A, Klein O, Goldman M, Tavernier J, et al. Interleukin 5 synthesis by eosinophils: association with granules and immunoglobulin-dependent secretion. J Exp Med. 1994 Feb 1;179(2):703–8. doi: 10.1084/jem.179.2.703. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Takaki S, Kanazawa H, Shiiba M, Takatsu K. A critical cytoplasmic domain of the interleukin-5 (IL-5) receptor alpha chain and its function in IL-5-mediated growth signal transduction. Mol Cell Biol. 1994 Nov;14(11):7404–13. doi: 10.1128/mcb.14.11.7404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Dijkhuizen T, van Essen T, van der Vlies P, Verheij JB, Sikkema-Raddatz B, van der Veen AY, et al. FISH and array-CGH analysis of a complex chromosome 3 aberration suggests that loss of CNTN4 and CRBN contributes to mental retardation in 3pter deletions. Am J Med Genet A. 2006 Nov 15;140(22):2482–7. doi: 10.1002/ajmg.a.31487. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.