- Email: [email protected]
Extinction of the Human Leukocyte Antigen Homozygosity Effect After Two Doses of the Measles-Mumps-Rubella Vaccine
Extinction of the Human Leukocyte Antigen Homozygosity Effect After Two Doses of the Measles-Mumps-Rubella Vaccine Jennifer L. St. Sauver, Neelam Dhiman, Inna G. Ovsyannikova, Robert M. Jacobson, Robert A. Vierkant, V. Shane Pankratz, Steven J. Jacobsen, and Gregory A. Poland ABSTRACT: We have reported associations between human leukocyte antigen (HLA) homozygosity and low measles antibody levels after one dose of the measles, mumps, and rubella (MMR) vaccine. Here, we examined associations between HLA homozygosity and immune responses to MMR after two doses of vaccine. We examined associations between HLA homozygosity and measles antibody levels in a group of 178 children (cohort 1) as well as associations between homozygosity and antibody levels and lymphoproliferative responses to MMR in 346 children (cohort 2). In cohort 1, HLA homozygotes and heterozygotes had similar increases in measles antibody levels after a second dose of measles vaccine. In cohort 2, HLA homozygosity was not associated with measles immune measures after two doses of vaccine. Homozygosity at the DPB locus was associated with increased rubella
INTRODUCTION Human leukocyte antigen (HLA) genes play an essential role in the immune response to antigenic proteins. HLA class I gene products bind and present processed antigenic peptides to cytotoxic T cells, whereas the HLA class II gene products bind and present processed antigenic peptides to T helper cells. HLA genes are highly polymorphic, and this diversity is hypothesized to be the result of natural selection, driven by infection with From the Department of Health Sciences Research, Division of Epidemiology ( J.L.S., S.J.J.), Mayo Vaccine Research Group (I.G.O., G.A.P.), Department of Pediatric and Adolescent Medicine (R.M.J.), Department of Health Sciences Research, Division of Biostatistics (R.A.V., S.P.), and Department of Internal Medicine (G.A.P.), Mayo Clinic, Rochester, MN, USA. Address reprint requests to: Dr. Gregory A. Poland, Director, Mayo Vaccine Research Group, 611C Guggenheim Building, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905; Tel: (507) 284-4968; Fax: (507) 266-4716; E-mail: [email protected]. Received January 11, 2004; revised March 11, 2005; accepted March 14, 2005. Human Immunology 66, 788 –798 (2005) © American Society for Histocompatibility and Immunogenetics, 2005 Published by Elsevier Inc.
antibody levels, and homozygosity at the class IA alleles was associated with lower mumps lymphoproliferative response. Homozygosity at increasing numbers of loci was also associated with lower mumps antibody levels and lymphoproliferative response. Therefore, two doses of the MMR vaccine appear to induce sufficient antibody levels and lymphoproliferative responses against measles and rubella, regardless of HLA homozygosity status. However, children who are HLA homozygous may be less protected against mumps compared with children who are heterozygous. Human Immunology 66, 788 –798 (2005). © American Society for Histocompatibility and Immunogenetics, 2005. Published by Elsevier Inc. KEYWORDS: HLA; homozygosity; MMR; humoral immunity; cellular immunity
pathogenic organisms [1, 2]. The diverse HLA proteins bind and present different antigenic and self peptides with varying degrees of efficiency [3, 4]. Therefore, individuals who are heterozygous at HLA loci may have a selective advantage in immune response to antigens because heterozygotes at specific loci may bind and present a wider range of pathogen-derived peptides compared with individuals who are homozygous at the same loci. Both animal and human studies support the concept of the HLA “heterozygote advantage.” For example, Penn et al. have demonstrated that heterozygous mice are less susceptible to infection with multiple strains of Salmonella compared with homozygous mice [5]. In addition, human HLA heterozygotes have been found to have higher antibody levels and more efficient immune responses to a wide range of antigens, including human immunodeficiency virus [6, 7], hepatitis B virus [8, 9], and human T-cell lymphotropic virus [10]. 0198-8859/05/$–see front matter doi:10.1016/j.humimm.2005.03.008
HLA Homozygosity and MMR Immunity
We have previously reported an association between HLA homozygosity and decreased measles antibody levels after a single dose of the measles, mumps, rubella (MMR) vaccine [11]. A second dose of MMR was introduced in 1989 to overcome apparently high rates of primary measles vaccine failure, and protection after two doses of vaccine approaches 95%–98%. However, it is unclear whether low measles immune measures may persist in children with specific risk factors—in this case, HLA homozygosity— even after two doses of vaccine. In addition, the association between HLA homozygosity and mumps and rubella immune measures is currently unknown. To address the underlying question of whether HLA homozygosity was associated with low immune measures after two doses of the MMR vaccine, we took advantage of information obtained from two separate cohorts of children who had participated in two previous studies of vaccine-induced immunity and HLA gene associations. First, we examined a group of 178 children who had participated in our previous study of HLA homozygosity and low measles antibody levels after one dose of measles vaccine (cohort 1). The children in this subgroup who were homozygous for HLA alleles had lower measles antibody levels compared with children who were heterozygous after receiving one dose of measles vaccine [11]. All of these children had since received a second dose of the MMR vaccine; therefore, we determined whether HLA homozygosity was still associated with decreased measles antibody levels, even after receipt of this booster dose of vaccine. Second, we examined a group of 346 children who had participated in a study designed to determine whether HLA gene variants were associated with low measles, mumps, and rubella antibody and T-cell proliferation levels after two doses of the MMR vaccine (cohort 2). In this cohort, we examined the associations between HLA homozygosity and both humoral and cellular measures of immunity to all three of these antigens. MATERIALS AND METHODS Study Populations Cohort 1. Details on the recruitment of the 242 children who had participated in a previous study examining the association between HLA homozygosity and measles antibody levels after one dose of measles vaccine are described in a previous publication [11]. Briefly, we collected serum samples from a volunteer group of 876 Olmsted County, MN, schoolchildren ages 5 to 13 and tested their serum for measles antibody levels. Of these children, we then recruited 72 (87%) of the children who were seronegative for measles antibodies, and 77 (86%) of the children who were serohyperpositive (those whose
789
circulating antibody levels were in the top 10% of the measles antibody range). Finally, we recruited a random sample of the remaining 660 seropositive children (n ⫽ 93). Of these 242 children, 178 (74%) received a second dose of the MMR vaccine. After this second vaccination, a second blood sample was obtained for measurement of measles antibody levels and typing of HLA alleles, as were medical history, history of measles exposure or disease, and demographic information. Cohort 2. Recruitment of the children who participated in the second cohort took place several years after recruitment of the children in cohort 1, and details are available in a previous publication [12, 13]. Briefly, we sent postcards describing our study to an age-stratified random sample of 1,212 parents of children aged 12 to 18 years enrolled in Minnesota Independent School District 535. The postcards described the study, and indicated that study coordinators would be contacting them by letter to determine interest in participating in the study. Seventeen of the children did not meet eligibility criteria, resulting in 1,195 potentially eligible children. Six hundred ninety-four parents (58.1%) did not respond to either the letter or the postcard, and 119 parents (10.0%) we contacted refused to participate. Three hundred eighty-two parents agreed to allow their children to participate in the study, and these children were scheduled for an appointment to obtain a blood sample. Thirty-six children (3.0%) did not appear for the blood draw, leaving 346 final study participants. As the recruitment for both studies took place in the same community, but separated by several years, it was possible that the same children could have participated in both studies. However, overlap between the two groups was minimal; only 24 children participated in both studies. Both studies were approved by the Mayo Clinic Institutional Review Board, and written informed consent was obtained from the parents of all children who participated in the study at the time of study enrollment. All subjects resided in Olmsted County, MN, where no measles, mumps, or rubella viruses have circulated in the community within the children’s lifetimes. All enrolled participants also had medical record documentation of receipt of two doses of the MMR II vaccine (Merck Research, West Point, PA). Antibody Measurement Cohort 1. Measles-specific IgG antibodies after both one and two doses of the measles vaccine were measured by whole virus ELISA (MeasELISA; Biowhittaker). Tests were performed in duplicate and the mean value of the replicates was used for all analyses. The coefficient of variation of this assay in our laboratory was 6.6%. Optical density index values were calculated per the man-
790
ufacturer’s instructions. Index values ⱕ 0.80 units were considered to be seronegative, and values ⱖ1.00 were classified as seropositive. We also classified the top 10% of the index values as serohyperpositive. Cohort 2. Measles, mumps and rubella-specific IgG antibody titers for all serum specimens were determined by the Enzygnost Anti-Masern-, Anti-Parotitis-, and AntiRubella-Virus/IgG enzyme immunoassays, respectively (Dade Behring, Marburg, Germany) according to the manufacturer’s instructions. Briefly, thawed serum samples were added in duplicate to microtiter plates, which contained two parallel wells coated with a test antigen derived from the respective viruses. Optical densities were determined at 450 nm and corrected at 650 nm on a microplate reader (Molecular Devices Corporation, Sunnyvale, CA). The difference in mean absorbance between the viral test antigen and control antigen for each sample was calculated and multiplied by a correction factor (determined by dividing a kit-specific nominal value by the mean of the reference standards) to give the corrected change in absorbance (⌬A). All virus-specific IgG antibody levels (IU/L) were calculated from the antilog of the following formula: log10 ⫽ ␣*⌬A(where ␣ and  are lot-dependent constants). The coefficients of variation for the measles, mumps, and rubella assays in our laboratory were 3.8%, 4.1%, and 4.0%, respectively. Preparation of Peripheral Blood Leukocytes and T-Cell Proliferation Assay T-cell proliferation levels were determined only for the children who comprised the second study cohort. Briefly, peripheral blood mononuclear cells were separated from heparinized blood by Ficoll-Hypaque (Sigma, St. Louis, MO) density gradient centrifugation and washed in RPMI-1640 medium (Celox Laboratories, St. Paul, MN) supplemented with 2 mM L-glutamine, 100 g/ml streptomycin, 100 U/ml penicillin, and 8% fetal calf serum (Life Technologies, Gaithersburg, MD). Viability of isolated cells ranged from 95% to 98% as determined by trypan blue exclusion test. The cellular immune status to MMR vaccines was assessed by an in vitro [3H]-thymidine incorporation assay as previously described [14] with slight modifications. Individual antigen specific T-cell responses were measured by proliferation of fresh peripheral blood mononuclear cells (2 ⫻ 105cells/well) incubated in RPMI-1640 medium, supplemented with 5% autologous sera, in the presence of live attenuated MMR vaccine (25 l vaccine/well). Vaccine diluent in triplicate wells served as unstimulated controls. Phytohemagglutinin (5 g/ml) was used as a positive control for stimulation of cellular proliferation. Lymphocyte proliferation was measured after 4 days by pulsing the cells with [3H]-tritiated thymidine for 18 hours. Cells were then
J.L. St. Sauver et al.
harvested onto glass fiber filters with a 96-well harvesting system (Skatron Instruments, Lier, Norway). The amount of incorporated radioactivity was determined by a liquid scintillation counter (Packard Instrument Company, Boston, MA). We used three replicates of counts per minute (cpm) values for unstimulated cells and three replicates each for T cells stimulated with live vaccines. We calculated the median cpm for unstimulated cells, as well as for cells stimulated with MMR vaccine, individually for each subject. Results were then expressed as antigen-specific stimulation indices (SI), defined as the ratio of the median cpm of antigen-stimulated wells to the median cpm of unstimulated control wells. Stimulation indices of three or higher were considered to be an indicator of a positive cellular immune response. The actual spread of counts was also calculated and is presented as ⌬ cpm (cpm antigen stimulated ⫺ cpm unstimulated). Plots of immune response by assay date identified an upward trend of measured cellular proliferation values over time. We fit polynomial linear regression models to evaluate this apparent association between T-cell proliferation and date of assay. We used the resulting models to remove the observed drift in the measures of cellular immune response. These calibrated values were used in all analyses. No such calibration was necessary with regard to humoral immune response after accounting for differences among lots (as per kit instructions). HLA Typing Cohort 1. DNA was extracted from fresh or frozen blood clots by proteinase K digestion followed by phenol/ chloroform extraction and ethanol precipitation [15]. DNA was typed for HLA class II alleles DRB1, DQA1, DQB1, and DPA1 by polymerase chain reaction by use of sequence-specific primers [16 –20]. Amplifications were performed in a GeneAmp polymerase chain reaction (PCR) system (model 9600, Perkin-Elmer Cetus Instruments). PCR products were separated on 1.5% agarose gels stained with ethidium bromide. DPB1 typing was performed with a DNA sequencing kit (Applied Biosystems.) Class IA and IB alleles were typed by microlymphocytotoxicity assays in the Mayo Tissue Typing Laboratory [21]. Homozygosity was confirmed by either exon sequencing or by typing and sequencing the birth parents of the subjects. Cohort 2. We extracted genomic DNA from frozen blood samples by conventional techniques with the Pyregene extraction kit (Gentra Systems, Minneapolis, MN). HLA class IA locus typing was performed with the Pel-Freez SeCore HLA-A locus sequencing kit (Pel-Freez Clinical Systems, Brown Deer, WI), followed by the Pel-Freez SSP (sequence-specific primer) UniTray typing kit and AmbiSolv, which consisted of specific primer mixes se-
HLA Homozygosity and MMR Immunity
lected to resolve common ambiguities. HLA class IB locus typing was performed with the Pel-Freez reference strand conformation analysis (RSCA) Multi-Dye B locus kit, followed by the ABI B locus sequencing kit, and the Pel-Freez SSP UniTray and AmbiSolv when needed. RSCA samples were amplified on an ABI 377 and analyzed by RSCA Typer software. HLA class I Cw locus typing was performed primarily with the Pel-Freez Cw High Resolution SSP UniTray. Any ambiguities were resolved with the Forensic Analytical C locus sequencing kit and AmbiSolv when needed. All PCR amplifications were carried out on an ABI 377 and analyzed by MatchTools software. For HLA class II typing, we used high-resolution DRB1 sequence-specific primers (SSP) and RSCA, DQB1 SSP, and DPB1 SSP UniTray typing kits with the entire locus on a single tray. PCR reactions were followed by AmbiSolv when needed. Any ambiguities were resolved with the ABI sequencing kits. RSCA samples were performed on an ABI 377 and analyzed by RSCA Typer software. All PCR amplifications were carried out on an ABI 377 and analyzed by MatchTools software. All reactions were run with negative controls, and every 50th PCR reaction was repeated for quality control. Statistical Analysis For each cohort, data were descriptively summarized by using frequencies and percentages for all categorical variables, and by using means, standard deviations, medians, and interquartile ranges for all continuous variables. Cohort 1. Two outcomes were of primary interest in this cohort: humoral immune response (as measured by measles-specific IgG antibodies [optical density units]), after both one and two doses of the measles vaccine. We also considered secondarily the change in immune response before versus after the second dose of vaccine. We compared levels of immune response in homozygous versus heterozygous individuals by linear regression techniques. Variables representing locus-specific homozygosity status were created for each of the seven available loci. We first fit separate models for each of the HLA loci individually. We then created a variable indicating whether a subject was homozygous for at least one of the loci and assessed its relationship with immune response. Because of data skewness, and to allow for comparability with the original homozygosity paper, the original response values for each regression model were replaced with corresponding ranked values. In addition, we accounted for the stratified sampling design by applying weights inversely proportional to a subject’s selection probability. We ran two sets of analyses for each outcome. First, models were fit that included only the homozygosity variables of interest. Subsequent secondary models included the following set of potential confounding variables: age, sex,
791
race, and age at first MMR. Because results were similar in both sets of models, and again to allow for comparability with the original paper, only the unadjusted analyses are presented. Cohort 2. Six outcomes were of primary interest in this cohort. These outcomes reflected humoral immune response (as measured by IgG antibody titers [22]) and cellular immune response (T-cell proliferation as measured by stimulation indices), as induced separately by the MMR vaccine viruses. As with cohort 1, we compared levels of immune response in homozygous versus heterozygous individuals by linear regression techniques. Models assessing the effects of any homozygosity and locus-specific homozygosity were fit by using the same variable definitions and analytic approach as in cohort 1. We then calculated a homozygosity count for each subject. Values of this count could range from 0 to 6, depending on the number of loci for which the subject was homozygous. However, because of sparseness of data, individuals homozygous for five or six loci were grouped with those homozygous for four loci. We used this count variable to assess the possible dose-response relationship between homozygosity and immune response by fitting the count as a one-degree-of-freedom ordinal variable. In contrast to the cohort 1 analyses, the original response values for each regression model were replaced with corresponding logarithmic values, as this was the transformation that best normalized the data. In addition, all linear regression models were adjusted for the following potential confounding variables: age at enrollment, race, sex, age at the first MMR, and age at the second MMR. All statistical tests were two-sided, and all analyses were carried out by SAS software (SAS Institute, Cary, NC). RESULTS Characteristics of the study cohorts were similar; slightly more than half of each group were male (53% and 54%), and the majority of the children were caucasian (95% and 94%). The median ages of the participants at the time of the first and second MMR immunizations were also similar. However, the children in cohort 1 were younger than those in cohort 2, because the time since the second immunization was shorter for the children in cohort 1 compared with cohort 2 (Table 1). Departures from Hardy-Weinberg equilibrium were not observed in the first cohort; however, moderate disequilibrium was observed for the class IC allele (p ⫽ 0.03), and strong disequilibrium was observed for the DQB allele (p ⬍ 0.001). We first examined the group of children who had participated in our previous study examining the association between HLA homozygosity and measles antibody levels after one dose of vaccine [11]. As previously
792
J.L. St. Sauver et al.
TABLE 1 Characteristics of the two study populations Characteristic Sex, n (%) Female Male Race, n (%) White Other Age (months) at first immunization, median (IQR) Age (years) at second immunization, median (IQR) Time (years) since second immunization, median (IQR)
Cohort 1 (n ⫽ 178)
Cohort 2 (n ⫽ 346)
84 (47.2%) 94 (52.8%)
161 (46.5%) 185 (53.5%)
169 (94.9%) 9 (5.1%)
325 (93.9%) 21 (6.1%)
15.5 (15.0–16.5)
15.6 (15.1–17.5)
12.0 (11.0–12.5)
12.1 (11.5–12.6)
2.5 (1.7–3.2)
4.7 (2.8–6.1)
and rubella antibody levels in the second cohort of children. No significant associations were observed between homozygosity at specific alleles and measles or mumps antibody levels (Table 4). However, children who were homozygous at the HLA-DPB locus had higher median rubella antibody levels compared with children who were heterozygous for DPB (p ⫽ 0.0003; Table 4). DPB homozygosity did not differ significantly by child’s sex: 53 (33%) of the girls and 48 (26%) of the boys were homozygous at this locus (2 ⫽ 2.03, p ⫽ 0.15). In the same cohort of children, we also examined the associations between HLA homozygosity at specific loci and lymphoproliferative responses to measles, mumps, and rubella antigens. No associations were observed between HLA homozygosity and measles or rubella lymphoproliferation levels. However, children who were homozygous at the HLA class IA locus had lower mumps-specific lymphoproliferation levels compared with children who were heterozygous at this locus (p ⫽ 0.04). Children who were homozygous at any HLA locus also had lower mumps cellular responses compared with children who were heterozygous at all loci (p ⫽ 0.04; Table 5). Finally, we examined the associations between homozygosity at increasing numbers of HLA loci and measles, mumps, and rubella antibody and lymphocyte proliferation levels. Homozygosity at increasing numbers of HLA loci was associated with decreased mumps antibody levels (test for trend p value ⫽ 0.02; Table 6). Homozygosity at increasing numbers of HLA loci was also associated with decreased mumps lymphoproliferation levels (test for trend p value ⫽ 0.04; Table 7). Homozygosity at increasing numbers of HLA loci was not, however, associated with measles or rubella antibody or lymphoproliferation levels (Tables 6 and 7).
reported, homozygosity at the class IA and DQA alleles was associated with decreased measles antibody levels after one dose of vaccine. After a second dose of vaccine, mean antibody levels in children homozygous at the class IA and DQA alleles remained significantly lower than the mean antibody levels in children who were heterozygous at these loci (Table 2). However, the mean difference in antibody levels among children who were homozygous at the HLA loci was not significantly different from the mean difference in antibody levels among children who were heterozygous at these loci (Table 3). Overall, 169 (95%) of the children were seropositive, 7 (11%) were seroequivocal, and 2 (3%) were seronegative after receiving two doses of the MMR vaccine. We next examined the associations between homozygosity at specific HLA loci and measles, mumps,
TABLE 2 Associations between HLA homozygosity and mean measles antibody levels (optical density units) after one and two doses of the MMR vaccine (cohort 1) Measles antibody levels after one dose of MMR
Measles antibody levels after two doses of MMR
Locus
Heterozygotes, mean (SD)
Homozygotes, mean (SD)
p Valuea
Heterozygotes, mean (SD)
Homozygotes, mean (SD)
p Valuea
Class IA Class IB DPA DPB DQA DQB DRB Any locusb
1.75 (1.42) 1.71 (1.43) 1.66 (1.39) 1.68 (1.40) 1.72 (1.41) 1.68 (1.42) 1.72 (1.42) 1.82 (1.44)
1.31 (1.15) 1.52 (1.12) 1.70 (1.42) 1.77 (1.53) 1.45 (1.34) 1.75 (1.41) 1.62 (1.37) 1.62 (1.41)
0.003 0.43 0.95 0.56 0.07 0.45 0.43 0.05
3.11 (2.42) 3.05 (2.54) 3.10 (2.62) 3.01 (2.34) 3.13 (2.63) 3.00 (2.52) 3.05 (2.56) 3.08 (2.29)
2.68 (2.99) 2.95 (1.91) 3.01 (2.45) 3.10 (2.48) 2.61 (1.78) 3.21 (2.41) 3.05 (2.28) 3.04 (2.49)
0.03 0.73 0.70 0.77 0.03 0.13 0.31 0.75
a b
Analysis based on rank-transformed values; distributions are weighted to account for the sampling scheme. Analysis does not include DPA because of the extreme number of homozygotes at this locus.
HLA Homozygosity and MMR Immunity
793
TABLE 3 Comparisons of changes in antibody EIA values before and after MMR booster in Olmsted County, MN, schoolchildren homozygous or heterozygous at specific HLA loci (cohort 1)a Heterozygote
Homozygote
Locus
Subjects (n)
Mean (SD)
Subjects (n)
Mean (SD)
p Valueb
Class IA Class IB DPA DPB DQA DQB DRB Any locusc
153 165 56 120 144 141 146 63
1.36 (2.18) 1.34 (2.35) 1.44 (2.67) 1.33 (2.32) 1.40 (1.17) 1.31 (2.32) 1.32 (2.34) 1.25 (2.24)
23 13 121 43 31 37 31 94
1.37 (3.05) 1.43 (1.49) 1.31 (2.10) 1.33 (2.14) 1.16 (1.82) 1.46 (2.18) 1.43 (2.02) 1.42 (2.29)
0.54 0.55 0.62 0.68 0.09 0.26 0.38 0.18
a
Distribution estimates are weighted to account for the sampling scheme. Analysis based on rank-transformed values. c Analysis does not include DPA. b
DISCUSSION Our results from two cohorts of children, which used slightly different methods for measuring immunity and for typing HLA alleles, suggest that HLA homozygosity is not strongly associated with decreased measles, mumps, and rubella immune measures after two doses of the MMR vaccine. In fact, in cohort 1, a group of children who had data available after receiving both a first and second dose of the MMR vaccine, HLA homozygous and heterozygous children had similar increases in measles antibody levels after a second dose of measles vaccine. In addition, the vast majority of the children in this study (95%) were seropositive after two doses of vaccine, regardless of HLA homozygosity status. A similar pattern was observed in our second cohort of children, as there was no association between HLA homozygosity and low measles antibody levels after two doses of measles vaccine. However, children who were homozygous at the DPB locus had significantly higher rubella antibody levels compared with children who were heterozygous at the same locus. Children homozygous for class IA alleles also had lower mumps lymphoproliferation levels compared with children who were heterozygous at this locus. In addition, children who were homozygous at increasing numbers of loci had decreasing mumps antibody and lymphoproliferation levels. Measles We found little evidence that either overall homozygosity or homozygosity at specific HLA loci was associated with substantially decreased measles antibody levels or decreased measles lymphoproliferative responses after two doses of the MMR vaccine. Our results differ from our previous study, which suggested that homozygosity at increasing numbers of HLA loci was associated with decreased measles antibody levels [11]. When a subset of
these same children were examined in this study after their second dose of measles vaccine, the association between HLA homozygosity at the class IA and DQA loci and lower measles antibody levels remained. However, the increases in measles antibody levels after two doses of vaccine between HLA homozygotes and heterozygotes were very similar. It is somewhat surprising that homozygotes in the second cohort did not also have lower antibody levels at the class IA and DQA loci, but the antibody assays used in each of these cohorts differed slightly and the results are thus not directly comparable. Instead, it may be best to compare the two cohorts only in terms of seropositivity and seronegativity. In this regard, only a few children from the first cohort remained seronegative or equivocal after two doses of vaccine and failure to become seropositive was not associated with HLA homozygosity. These results are consistent with other studies, which have indicated that 2% to 18% of children may have low measles antibody levels after a single vaccination [23– 27], depending on age at immunization and time since immunization. However, the majority of individuals seronegative after a single measles vaccination become seropositive after a second dose of vaccine. Cohn et al. demonstrated that 70% of children revaccinated with measles vaccine maintained seropositive antibody levels after vaccination [28]. In our own work, we found that over 80% of 130 children who were initially seronegative after their first measles vaccination became measles antibody seropositive after their second immunization [29]. Similarly, McDermott et al. have found that at least some homozygotes initially seronegative for hepatitis B antibodies will seroconvert after revaccination with an appropriate dose of hepatitis B vaccine [22, 30]. Therefore, although HLA homozygosity may initially limit measles
794
TABLE 4 Differences in distribution of continuous measles, mumps, and rubella antibody values (IU/L) between children heterozygous and homozygous for specific loci (cohort 2)a Measles
Mumps
Rubella
Locus
Homozygote (n)
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valueb
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valueb
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valueb
Class IA Class IB Class IC DRB DQB DPB Any locus
75 28 57 46 122 101 251
1528 (779–2677) 1502 (699–2633) 1563 (746–2652) 1562 (779–2599) 1518 (692–2468) 1573 (685–2703) 1573 (584–2498)
1585 (675–2709) 2110 (1252–3525) 1528 (765–2970) 1359 (557–3265) 1593 (763–3042) 1495 (1019–2442) 1543 (765–2762)
0.66 0.08 0.51 0.79 0.59 0.34 0.28
759 (415–1395) 721 (414–1315) 772 (433–1358) 759 (414–1322) 804 (429–1371) 734 (433–1284) 809 (462–1627)
672 (395–1066) 774 (371–1252) 595 (355–970) 676 (332–1083) 676 (362–1189) 711 (385–1315) 699 (391–1251)
0.32 0.63 0.12 0.54 0.10 0.85 0.16
40 (23–62) 39 (21–61) 39 (23–61) 38 (22–61) 37 (22–60) 35 (20–55) 35 (21–57)
36 (20–58) 35 (24–55) 36 (19–59) 40 (22–65) 42 (23–65) 48 (25–76) 39 (22–65)
1.00 0.73 0.84 0.87 0.65 0.0003 0.19
a b
Linear regression analysis. p values are adjusted for age, sex, race, age at first MMR, and age at second MMR. Linear regression p values calculated from log-transformed values. Results significant at p ⱕ 0.05 in bold.
TABLE 5 Differences in distribution of continuous measles, mumps, and rubella lymphoproliferation (SI) values between children heterozygous and homozygous for specific loci (cohort 2) Measles
Mumps
Rubella
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valuea
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valuea
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valuea
Class IA Class IB Class IC DRB DQB DPB Any locus
75 28 57 46 122 101 251
3.6 (2.1–6.2) 3.6 (2.0–6.1) 3.6 (2.1–6.3) 3.7 (2.1–6.3) 3.9 (2.1–6.6) 3.6 (2.1–5.7) 3.9 (2.0–6.4)
3.7 (2.3–6.9) 4.4 (2.5–7.9) 3.7 (2.1–5.8) 3.1 (2.4–5.8) 3.3 (2.4–5.5) 3.8 (2.2–6.8) 3.6 (2.3–6.2)
0.95 0.23 0.88 0.75 0.34 0.16 0.88
4.9 (2.5–10.3) 4.9 (2.5–9.2) 5.0 (2.7–9.6) 4.9 (2.5–9.3) 5.7 (2.6–9.3) 4.6 (2.3–9.3) 5.9 (3.2–13.0)
4.6 (2.2–7.3) 4.4 (1.9–9.0) 3.5 (2.0–8.2) 3.9 (2.0–6.7) 3.8 (2.3–8.7) 5.0 (2.7–8.9) 4.4 (2.3–8.7)
0.04 0.55 0.10 0.52 0.25 0.81 0.04
2.3 (1.5–3.8) 2.3 (1.5–3.5) 2.2 (1.5–3.6) 2.3 (1.5–3.6) 2.4 (1.5–3.8) 2.3 (1.6–3.7) 2.4 (1.4–4.1)
2.3 (1.6–3.1) 3.2 (1.6–4.9) 2.5 (1.6–3.8) 2.1 (1.5–3.6) 2.1 (1.6–3.5) 2.1 (1.5–3.4) 2.3 (1.5–3.5)
0.78 0.07 0.09 0.52 0.27 0.41 0.92
a
Linear regression analysis. p values calculated using log-transformed values; p values adjusted for age at enrollment, age at first MMR vaccination, age at second MMR vaccination, sex, and race. Results significant at p ⱕ 0.05 are bold.
J.L. St. Sauver et al.
Locus
Homozygote (n)
HLA Homozygosity and MMR Immunity
795
TABLE 6 Differences in distribution of continuous measles, mumps, and rubella antibody values (IU/L) among children homozygous for increasing numbers of HLA loci (cohort 2) Measles
Mumps
Rubella
No. of homozygous loci
n
Median (IQR)
p Valuea
Median (IQR)
p Valuea
Median (IQR)
p Valuea
0 1 2 3 ⱖ4
167 109 52 9 9
1573 (584–2498) 1566 (820–2985) 1420 (764–2270) 1645 (823–2976) 2428 (994–3524)
0.19
809 (462–1627) 805 (444–1358) 686 (386–977) 548 (284–914) 752 (459–1618)
0.02
35 (21–57) 41 (23–65) 34 (20–60) 42 (27–69) 52 (24–73)
0.09
a
Linear regression analysis test for trend; calculated using log-transformed values; p values adjusted for age at enrollment, age at first MMR vaccination, age at second MMR vaccination, sex, and race. Results significant at p ⱕ 0.05 are bold.
antibody responses to MMR vaccination, this limitation can apparently be overcome with a second vaccination.
to detect other associations between HLA homozygosity and measures of rubella immunity.
Rubella With one exception, we found no associations between HLA homozygosity and rubella antibody and lymphoproliferation levels. However, children who were homozygous for the HLA-DPB locus had higher mean rubella antibody levels compared with children who were heterozygous for the same locus. This association is in the opposite direction than we had initially predicted, and the clinical significance of the association is unclear. The HLA-DPB*04 allele accounted for approximately 51% of the total DPB alleles observed in our population, and 81% of our homozygotes (data not shown). The DPBw4 molecule has been previously demonstrated to be important in the immune response to the hepatitis B surface antigen, as specific hepatitis B peptides are restricted to DPBw4 [31, 32]. Naturally processed rubella peptides may also be restricted to specific DPB alleles, and homozygosity at these specific DPB loci could theoretically confer an advantage over heterozygotes in response to the rubella vaccine. However, variation in measures of rubella immunity was extremely low in this group of children. Therefore, we may not have had enough power
Mumps Although homozygosity at specific loci was not associated with low mumps antibody levels, homozygosity at increasing numbers of HLA loci was associated with lower mumps antibody levels (test for trend p value ⫽ 0.02). In addition, homozygosity at HLA class IA was associated with a lower median lymphoproliferation level compared with heterozygosity at the same locus (p ⫽ 0.04), and homozygosity at increasing numbers of HLA alleles was associated with decreased lymphoproliferation levels (test for trend p value ⫽ 0.04). These results are in accord with our initial hypothesis, as we expected that HLA homozygosity would be associated with a lower immune response. However, it is surprising that these associations were observed only for a single MMR antigen. We may have detected this association because the distribution of immune responses was more variable for mumps than for either of the other antigens. Forty-three percent of our study population was seronegative or equivocal for mumps antibodies compared with only 14% of measles seronegatives/equivocals and 1% of rubella seronegatives/equivocals. This variability in im-
TABLE 7 Differences in distribution of continuous measles, mumps, and rubella lymphoproliferation (SI) values among children homozygous for increasing numbers of HLA loci (cohort 2)a Measles
Mumps
Rubella
No. of homozygous loci
n
Median SI (IQR)
p Valueb
Median SI (IQR)
p Valueb
Median SI (IQR)
p Valueb
0 1 2 3 ⱖ4
95 131 80 28 12
3.9 (2.0–6.4) 3.6 (2.5–6.1) 3.7 (1.8–6.1) 3.3 (2.5–7.9) 2.9 (2.4–6.3)
0.70
5.9 (3.2–13.0) 4.8 (2.3–9.3) 4.0 (2.2–8.0) 4.8 (2.1–9.0) 3.0 (2.0–5.2)
0.04
2.4 (1.4–4.1) 2.3 (1.5–3.3) 2.4 (1.5–3.7) 2.2 (1.6–3.6) 2.3 (1.5–3.4)
0.72
a b
Values adjusted for age at enrollment, age at first MMR vaccination, age at second MMR vaccination, sex, and race. Results significant at p ⱕ 0.05 are bold. Linear regression analysis test for trend; calculated using log-transformed values.
796
mune response would have increased our power to detect associations with HLA homozygosity. Overall, however, these data also indicate that a significant proportion of our study population may not be protected against mumps infection even after two doses of the MMR vaccine. A strength of our study arises from the availability of data for two cohorts of children that we were able to use to test our hypotheses. Despite the use of slightly different methods for measuring antibody levels and for typing HLA alleles, results were consistent across both groups of children in suggesting that two doses of the MMR vaccine appears to overcome adverse effects of HLA homozygosity on development of measles antibody levels after vaccination. A second strength of our study arises from the data available for the second cohort of children that allowed us to examine measures of both the humoral and the cellular arms of the immune system. Both humoral and cellmediated immunity are essential for conferring complete protection after vaccination with these viral antigens. However, data from other studies suggest that humoral and cellular immune measures (antibody and lymphoproliferation levels) are poorly correlated (data not shown). Therefore, different genetic factors may influence these measures differently, and it was important to examine the association between HLA homozygosity and each of the measures of immunity. A third strength of our study includes our ability to examine measures of the immune response to three viral antigens in the second cohort of children. Previous work indicates that the antibody levels to these vaccine antigens are not highly correlated [11]. These viruses likely produce different antigenic peptides that may be presented differently by the same HLA genes. Therefore, it is likely that HLA associations with rubella antibodies (or lymphoproliferation levels) will differ from those with measles or mumps immune measures, and it is not surprising that the association between HLA-DPB homozygosity and rubella antibodies is not observed with measles or mumps antibodies as well. Limitations of our study include lack of data on the immune measures to mumps and rubella after a single dose of MMR vaccine. Given this limitation, it was not possible to determine whether HLA homozygosity might also have been associated with decreased mumps and rubella immune responses after a single vaccination. Another potential study limitation results from our relatively low participation rate. Overall, about 29% of our initial study population agreed to participate, and it is possible that our study results are biased. However, neither HLA testing nor measles, mumps, or rubella antibody typing are routinely performed in this community. Therefore, subjects were highly unlikely to be aware
J.L. St. Sauver et al.
of their HLA genotype or their measles, mumps, or rubella immune levels. It is possible that parents may have preferentially enrolled their children in one of these studies because they were unsure of their child’s vaccine status and wanted to find out if their child was immune to these viruses. It seems unlikely, however, that parents could have preferentially enrolled children who were also homozygous for specific HLA alleles. Therefore, we think selection bias was minimal in this study; however, results may not be generalizable to other populations with different HLA distributions. A second potential limitation includes our use of EIA tests for measuring antibody levels rather than measurement of neutralizing antibody titers. However, use of neutralizing antibody titers is more labor-intensive than EIA testing, and a variety of commercially available EIA assay results have been previously demonstrated to correlate well with antibody neutralization measures [33– 35]. Therefore, EIA values are a reasonable surrogate for antibody neutralization assays and accurately reflect the measles, mumps, and rubella antibody levels of our study participants. Finally, given that we tested multiple associations, it may not be appropriate to designate a p value of 0.05 as our threshold for statistical significance. Instead, a more conservative threshold might be more appropriate. In conclusion, two doses of the MMR vaccine appear to overcome HLA homozygosity associations with low measles antibody levels observed after one dose of vaccine. Therefore, two doses of the MMR vaccine seem sufficient to induce protective levels of antibodies against both measles and rubella, regardless of HLA homozygosity status, providing further mechanistic evidence to support current public health recommendations for routine administration of two doses of the MMR vaccine. However, children who are HLA homozygous may be less protected against mumps compared with children who are heterozygous, even after two doses of the MMR vaccine. ACKNOWLEDGMENTS We thank Tina Agostini and Dennis Devitt for performing the Pel-Freez HLA typings. We also thank the many children and parents who participated in these studies. This research was funded by NIH grants AI 33144 and AI 48793 and by NIAMS grant AR030582.
REFERENCES 1. Apanius V, Penn D, Slev PR, Ruff LR, Potts WK: The nature of selection on the major histocompatibility complex. Crit Rev Immunol 17:179, 1997. 2. Borghans AM, Beltman JB, De Boer RJ: MHC polymor-
HLA Homozygosity and MMR Immunity
3. 4.
5.
6.
7.
8.
9. 10.
11.
12.
13.
14.
15.
16.
phism under host-pathogen coevolution. Immunogenetics 55:732, 2004. Klein J, Sato A: The HLA system. First of two parts. N Engl J Med 343:702, 2000. Klein J, Sato A: The HLA system. Second of two parts. [Erratum appears in N Engl J Med 2000;343:1504]. N Engl J Med 343:782, 2000. Penn DJ, Damjanovich K, Potts WK: MHC heterozygosity confers a selective advantage against multiple-strain infections. Proc Natl Acad Sci U S A 99:11260, 2002. Carrington M, Nelson GW, Martin MP, Kissner T, Vlahov D, Goedert JJ, Kaslow R, Buchbinder S, Hoots K, O’Brien SJ: HLA and HIV-1: heterozygote advantage and B*35-Cw*04 disadvantage [see comments]. Science 283: 1748, 1999. Tang J, Costello C, Keet IP, Rivers C, Leblanc S, Karita E, Allen S, Kaslow RA: HLA class I homozygosity accelerates disease progression in human immunodeficiency virus type 1 infection. AIDS Res Hum Retroviruses 15: 317, 1999. Thursz MR, Thomas HC, Greenwood BM, Hill AV: Heterozygote advantage for HLA class-II type in hepatitis B virus infection [letter] [published erratum appears in Nat Genet 1998;18:88]. Nat Genet 17:11, 1997. Thursz MR: Host genetic factors influencing the outcome of hepatitis. J Viral Hepat 4:215, 1997. Jeffery KJ, Siddiqui AA, Bunce M, Lloyd AL, Vine AM, Witkover AD, Izumo S, Usuku K, Welsh KI, Osame M, Bangham CR: The influence of HLA class I alleles and heterozygosity on the outcome of human T cell lymphotropic virus type I infection. J Immunol 165:7278, 2000. St Sauver JL, Ovsyannikova IG, Jacobson RM, Jacobsen SJ, Vierkant RA, Schaid DJ, Pankratz VS, Green EM, Poland GA: Associations between human leukocyte antigen homozygosity and antibody levels to measles vaccine. J Infect Dis 185:1545, 2002. Ovsyannikova IG, Jacobson RM, Vierkant RA, Jacobsen SJ, Pankratz VS, Poland GA: The contribution of HLA class I antigens in immune status following two doses of rubella vaccination. Hum Immunol 65:1506, 2004. Ovsyannikova IG, Jacobson RM, Vierkant RA, Jacobsen SJ, Pankratz VS, Poland GA: HLA class II alleles and rubella-specific humoral and cell-mediated immunity following MMR-II vaccination. J Infect Dis 191:515, 2005. Bautista-Lopez N, Ward BJ, Mills E, McCormick D, Martel N, Ratnam S: Development and durability of measles antigen-specific lymphoproliferative response after MMR vaccination. Vaccine 18:1393, 2000. Enrietto PJ, Payne LN, Hayman MJ: A recovered avian myelocytomatosis virus that induces lymphomas in chickens: pathogenic properties and their molecular basis. Cell 352 Pt 1369, 1983. Fugger L, Morling N, Ryder LP, Odum N, Svejgaard A: Technical aspects of typing for HLA-DP alleles using allele-specific DNA in vitro amplification and sequence-
797
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
specific oligonucleotide probes: detection of single base mismatches. J Immunol Methods 129:175, 1990. Kimura A, Dong RP, Harada H, Sasazuki T: DNA typing of HLA class II genes in B-lymphoblastoid cell lines homozygous for HLA. Tissue Antigens 40:5, 1992. Olerup O, Aldener A, Fogdell A: HLA-DQB1 and -DQA1 typing by PCR amplification with sequence-specific primers (PCR-SSP) in 2 hours. Tissue Antigens 41:119, 1993. Olerup O, Zetterguist H: DR “low-resolution” PCR-SSP typing—a correction and an update [letter; comment]. Tissue Antigens 41:55, 1993. Weyand CM, Xie C, Goronzy JJ: Homozygosity for the HLA-DRB1 allele selects for extraarticular manifestations in rheumatoid arthritis. J Clin Invest 89:2033, 1992. Poland GA, Jacobson RM, Schaid D, Moore SB, Jacobsen SJ: The association between HLA class I alleles and measles vaccine-induced antibody response: evidence of a significant association. Vaccine 16:1869, 1998. McDermott AB, Cohen SB, Zuckerman JN, Madrigal JA: Human leukocyte antigens influence the immune response to a pre-S/S hepatitis B vaccine. Vaccine 17:330, 1999. Brunell PA, Weigle K, Murphy MD, Shehab Z, Cobb E: Antibody response following measles-mumps-rubella vaccine under conditions of customary use. JAMA 250:1409, 1983. Mitchell LA, Tingle AJ, Decarie D, Lajeunesse C: Serologic responses to measles, mumps, and rubella (MMR) vaccine in healthy infants: failure to respond to measles and mumps components may influence decisions on timing of the second dose of MMR. Can J Public Health 89:325, 1998. Matson DO, Byington C, Canfield M, Albrecht P, Feigin RD: Investigation of a measles outbreak in a fully vaccinated school population including serum studies before and after revaccination. Pediatr Infect Dis J 12:292, 1993. Boulianne N, De Serres G, Ratnam S, Ward BJ, Joly JR, Duval B: Measles, mumps, and rubella antibodies in children 5– 6 years after immunization: effect of vaccine type and age at vaccination. Vaccine 13:1611, 1995. Johnson CE, Kumar ML, Whitwell JK, Staehle BO, Rome LP, Dinakar C, Hurni W, Nalin DR: Antibody persistence after primary measles-mumps-rubella vaccine and response to a second dose given at four to six vs eleven to thirteen years. Pediatr Infect Dis J 15:687, 1996. Cohn ML, Robinson ED, Faerber M, Thomas D, Geyer S, Peters S, Martin M, Martin A, Sobel D, Jones R, Larkin L, Richert JR: Measles vaccine failures: lack of sustained measles-specific immunoglobulin G responses in revaccinated adolescents and young adults. Pediatr Infect Dis J 13:34, 1994. Poland GA, Jacobson RM, Thampy AM, Colbourne SA, Wollan PC, Lipsky JJ, Jacobsen SJ: Measles reimmunization in children seronegative after initial immunization. JAMA 277:1156, 1997.
798
30. McDermott AB, Madrigal JA, Sabin CA, Zuckerman JN, Cohen SB: The influence of host factors and immunogenetics on lymphocyte responses to Hepagene vaccination. Vaccine 17:1329, 1999. 31. Celis E, Ou D, Otvos L Jr: Recognition of hepatitis B surface antigen by human T lymphocytes: proliferative and cytotoxic responses to a major antigenic determinant defined by synthetic peptides. J Immunol 140:1808, 1988. 32. Celis E, Karr RW: Presentation of an immunodominant T-cell epitope of hepatitis B surface antigen by the HLADPw4 molecule. J Virol 63:747, 1989. 33. Ratnam S, Gadag V, West R, Burris J, Oates E, Stead F, Bouilianne N: Comparison of commercial enzyme immu-
J.L. St. Sauver et al.
noassay kits with plaque reduction neutralization test for detection of measles virus antibody. J Clin Microbiol 33:811, 1995. 34. Cremer NE, Cossen CK, Shell G, Diggs J, Gallo D, Schmidt NJ: Enzyme immunoassay versus plaque neutralization and other methods for determination of immune status to measles and varicella-zoster viruses and versus complement fixation for serodiagnosis of infections with those viruses. J Clin Microbiol 21:869, 1985. 35. de Souza VA, Pannuti CS, Sumita LM, de Andrade Junior HF: Enzyme-linked immunosorbent assay–IgG antibody avidity test for single sample serologic evaluation of measles vaccines. J Med Virol 52:275, 1997.
INTRODUCTION Human leukocyte antigen (HLA) genes play an essential role in the immune response to antigenic proteins. HLA class I gene products bind and present processed antigenic peptides to cytotoxic T cells, whereas the HLA class II gene products bind and present processed antigenic peptides to T helper cells. HLA genes are highly polymorphic, and this diversity is hypothesized to be the result of natural selection, driven by infection with From the Department of Health Sciences Research, Division of Epidemiology ( J.L.S., S.J.J.), Mayo Vaccine Research Group (I.G.O., G.A.P.), Department of Pediatric and Adolescent Medicine (R.M.J.), Department of Health Sciences Research, Division of Biostatistics (R.A.V., S.P.), and Department of Internal Medicine (G.A.P.), Mayo Clinic, Rochester, MN, USA. Address reprint requests to: Dr. Gregory A. Poland, Director, Mayo Vaccine Research Group, 611C Guggenheim Building, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905; Tel: (507) 284-4968; Fax: (507) 266-4716; E-mail: [email protected]. Received January 11, 2004; revised March 11, 2005; accepted March 14, 2005. Human Immunology 66, 788 –798 (2005) © American Society for Histocompatibility and Immunogenetics, 2005 Published by Elsevier Inc.
antibody levels, and homozygosity at the class IA alleles was associated with lower mumps lymphoproliferative response. Homozygosity at increasing numbers of loci was also associated with lower mumps antibody levels and lymphoproliferative response. Therefore, two doses of the MMR vaccine appear to induce sufficient antibody levels and lymphoproliferative responses against measles and rubella, regardless of HLA homozygosity status. However, children who are HLA homozygous may be less protected against mumps compared with children who are heterozygous. Human Immunology 66, 788 –798 (2005). © American Society for Histocompatibility and Immunogenetics, 2005. Published by Elsevier Inc. KEYWORDS: HLA; homozygosity; MMR; humoral immunity; cellular immunity
pathogenic organisms [1, 2]. The diverse HLA proteins bind and present different antigenic and self peptides with varying degrees of efficiency [3, 4]. Therefore, individuals who are heterozygous at HLA loci may have a selective advantage in immune response to antigens because heterozygotes at specific loci may bind and present a wider range of pathogen-derived peptides compared with individuals who are homozygous at the same loci. Both animal and human studies support the concept of the HLA “heterozygote advantage.” For example, Penn et al. have demonstrated that heterozygous mice are less susceptible to infection with multiple strains of Salmonella compared with homozygous mice [5]. In addition, human HLA heterozygotes have been found to have higher antibody levels and more efficient immune responses to a wide range of antigens, including human immunodeficiency virus [6, 7], hepatitis B virus [8, 9], and human T-cell lymphotropic virus [10]. 0198-8859/05/$–see front matter doi:10.1016/j.humimm.2005.03.008
HLA Homozygosity and MMR Immunity
We have previously reported an association between HLA homozygosity and decreased measles antibody levels after a single dose of the measles, mumps, rubella (MMR) vaccine [11]. A second dose of MMR was introduced in 1989 to overcome apparently high rates of primary measles vaccine failure, and protection after two doses of vaccine approaches 95%–98%. However, it is unclear whether low measles immune measures may persist in children with specific risk factors—in this case, HLA homozygosity— even after two doses of vaccine. In addition, the association between HLA homozygosity and mumps and rubella immune measures is currently unknown. To address the underlying question of whether HLA homozygosity was associated with low immune measures after two doses of the MMR vaccine, we took advantage of information obtained from two separate cohorts of children who had participated in two previous studies of vaccine-induced immunity and HLA gene associations. First, we examined a group of 178 children who had participated in our previous study of HLA homozygosity and low measles antibody levels after one dose of measles vaccine (cohort 1). The children in this subgroup who were homozygous for HLA alleles had lower measles antibody levels compared with children who were heterozygous after receiving one dose of measles vaccine [11]. All of these children had since received a second dose of the MMR vaccine; therefore, we determined whether HLA homozygosity was still associated with decreased measles antibody levels, even after receipt of this booster dose of vaccine. Second, we examined a group of 346 children who had participated in a study designed to determine whether HLA gene variants were associated with low measles, mumps, and rubella antibody and T-cell proliferation levels after two doses of the MMR vaccine (cohort 2). In this cohort, we examined the associations between HLA homozygosity and both humoral and cellular measures of immunity to all three of these antigens. MATERIALS AND METHODS Study Populations Cohort 1. Details on the recruitment of the 242 children who had participated in a previous study examining the association between HLA homozygosity and measles antibody levels after one dose of measles vaccine are described in a previous publication [11]. Briefly, we collected serum samples from a volunteer group of 876 Olmsted County, MN, schoolchildren ages 5 to 13 and tested their serum for measles antibody levels. Of these children, we then recruited 72 (87%) of the children who were seronegative for measles antibodies, and 77 (86%) of the children who were serohyperpositive (those whose
789
circulating antibody levels were in the top 10% of the measles antibody range). Finally, we recruited a random sample of the remaining 660 seropositive children (n ⫽ 93). Of these 242 children, 178 (74%) received a second dose of the MMR vaccine. After this second vaccination, a second blood sample was obtained for measurement of measles antibody levels and typing of HLA alleles, as were medical history, history of measles exposure or disease, and demographic information. Cohort 2. Recruitment of the children who participated in the second cohort took place several years after recruitment of the children in cohort 1, and details are available in a previous publication [12, 13]. Briefly, we sent postcards describing our study to an age-stratified random sample of 1,212 parents of children aged 12 to 18 years enrolled in Minnesota Independent School District 535. The postcards described the study, and indicated that study coordinators would be contacting them by letter to determine interest in participating in the study. Seventeen of the children did not meet eligibility criteria, resulting in 1,195 potentially eligible children. Six hundred ninety-four parents (58.1%) did not respond to either the letter or the postcard, and 119 parents (10.0%) we contacted refused to participate. Three hundred eighty-two parents agreed to allow their children to participate in the study, and these children were scheduled for an appointment to obtain a blood sample. Thirty-six children (3.0%) did not appear for the blood draw, leaving 346 final study participants. As the recruitment for both studies took place in the same community, but separated by several years, it was possible that the same children could have participated in both studies. However, overlap between the two groups was minimal; only 24 children participated in both studies. Both studies were approved by the Mayo Clinic Institutional Review Board, and written informed consent was obtained from the parents of all children who participated in the study at the time of study enrollment. All subjects resided in Olmsted County, MN, where no measles, mumps, or rubella viruses have circulated in the community within the children’s lifetimes. All enrolled participants also had medical record documentation of receipt of two doses of the MMR II vaccine (Merck Research, West Point, PA). Antibody Measurement Cohort 1. Measles-specific IgG antibodies after both one and two doses of the measles vaccine were measured by whole virus ELISA (MeasELISA; Biowhittaker). Tests were performed in duplicate and the mean value of the replicates was used for all analyses. The coefficient of variation of this assay in our laboratory was 6.6%. Optical density index values were calculated per the man-
790
ufacturer’s instructions. Index values ⱕ 0.80 units were considered to be seronegative, and values ⱖ1.00 were classified as seropositive. We also classified the top 10% of the index values as serohyperpositive. Cohort 2. Measles, mumps and rubella-specific IgG antibody titers for all serum specimens were determined by the Enzygnost Anti-Masern-, Anti-Parotitis-, and AntiRubella-Virus/IgG enzyme immunoassays, respectively (Dade Behring, Marburg, Germany) according to the manufacturer’s instructions. Briefly, thawed serum samples were added in duplicate to microtiter plates, which contained two parallel wells coated with a test antigen derived from the respective viruses. Optical densities were determined at 450 nm and corrected at 650 nm on a microplate reader (Molecular Devices Corporation, Sunnyvale, CA). The difference in mean absorbance between the viral test antigen and control antigen for each sample was calculated and multiplied by a correction factor (determined by dividing a kit-specific nominal value by the mean of the reference standards) to give the corrected change in absorbance (⌬A). All virus-specific IgG antibody levels (IU/L) were calculated from the antilog of the following formula: log10 ⫽ ␣*⌬A(where ␣ and  are lot-dependent constants). The coefficients of variation for the measles, mumps, and rubella assays in our laboratory were 3.8%, 4.1%, and 4.0%, respectively. Preparation of Peripheral Blood Leukocytes and T-Cell Proliferation Assay T-cell proliferation levels were determined only for the children who comprised the second study cohort. Briefly, peripheral blood mononuclear cells were separated from heparinized blood by Ficoll-Hypaque (Sigma, St. Louis, MO) density gradient centrifugation and washed in RPMI-1640 medium (Celox Laboratories, St. Paul, MN) supplemented with 2 mM L-glutamine, 100 g/ml streptomycin, 100 U/ml penicillin, and 8% fetal calf serum (Life Technologies, Gaithersburg, MD). Viability of isolated cells ranged from 95% to 98% as determined by trypan blue exclusion test. The cellular immune status to MMR vaccines was assessed by an in vitro [3H]-thymidine incorporation assay as previously described [14] with slight modifications. Individual antigen specific T-cell responses were measured by proliferation of fresh peripheral blood mononuclear cells (2 ⫻ 105cells/well) incubated in RPMI-1640 medium, supplemented with 5% autologous sera, in the presence of live attenuated MMR vaccine (25 l vaccine/well). Vaccine diluent in triplicate wells served as unstimulated controls. Phytohemagglutinin (5 g/ml) was used as a positive control for stimulation of cellular proliferation. Lymphocyte proliferation was measured after 4 days by pulsing the cells with [3H]-tritiated thymidine for 18 hours. Cells were then
J.L. St. Sauver et al.
harvested onto glass fiber filters with a 96-well harvesting system (Skatron Instruments, Lier, Norway). The amount of incorporated radioactivity was determined by a liquid scintillation counter (Packard Instrument Company, Boston, MA). We used three replicates of counts per minute (cpm) values for unstimulated cells and three replicates each for T cells stimulated with live vaccines. We calculated the median cpm for unstimulated cells, as well as for cells stimulated with MMR vaccine, individually for each subject. Results were then expressed as antigen-specific stimulation indices (SI), defined as the ratio of the median cpm of antigen-stimulated wells to the median cpm of unstimulated control wells. Stimulation indices of three or higher were considered to be an indicator of a positive cellular immune response. The actual spread of counts was also calculated and is presented as ⌬ cpm (cpm antigen stimulated ⫺ cpm unstimulated). Plots of immune response by assay date identified an upward trend of measured cellular proliferation values over time. We fit polynomial linear regression models to evaluate this apparent association between T-cell proliferation and date of assay. We used the resulting models to remove the observed drift in the measures of cellular immune response. These calibrated values were used in all analyses. No such calibration was necessary with regard to humoral immune response after accounting for differences among lots (as per kit instructions). HLA Typing Cohort 1. DNA was extracted from fresh or frozen blood clots by proteinase K digestion followed by phenol/ chloroform extraction and ethanol precipitation [15]. DNA was typed for HLA class II alleles DRB1, DQA1, DQB1, and DPA1 by polymerase chain reaction by use of sequence-specific primers [16 –20]. Amplifications were performed in a GeneAmp polymerase chain reaction (PCR) system (model 9600, Perkin-Elmer Cetus Instruments). PCR products were separated on 1.5% agarose gels stained with ethidium bromide. DPB1 typing was performed with a DNA sequencing kit (Applied Biosystems.) Class IA and IB alleles were typed by microlymphocytotoxicity assays in the Mayo Tissue Typing Laboratory [21]. Homozygosity was confirmed by either exon sequencing or by typing and sequencing the birth parents of the subjects. Cohort 2. We extracted genomic DNA from frozen blood samples by conventional techniques with the Pyregene extraction kit (Gentra Systems, Minneapolis, MN). HLA class IA locus typing was performed with the Pel-Freez SeCore HLA-A locus sequencing kit (Pel-Freez Clinical Systems, Brown Deer, WI), followed by the Pel-Freez SSP (sequence-specific primer) UniTray typing kit and AmbiSolv, which consisted of specific primer mixes se-
HLA Homozygosity and MMR Immunity
lected to resolve common ambiguities. HLA class IB locus typing was performed with the Pel-Freez reference strand conformation analysis (RSCA) Multi-Dye B locus kit, followed by the ABI B locus sequencing kit, and the Pel-Freez SSP UniTray and AmbiSolv when needed. RSCA samples were amplified on an ABI 377 and analyzed by RSCA Typer software. HLA class I Cw locus typing was performed primarily with the Pel-Freez Cw High Resolution SSP UniTray. Any ambiguities were resolved with the Forensic Analytical C locus sequencing kit and AmbiSolv when needed. All PCR amplifications were carried out on an ABI 377 and analyzed by MatchTools software. For HLA class II typing, we used high-resolution DRB1 sequence-specific primers (SSP) and RSCA, DQB1 SSP, and DPB1 SSP UniTray typing kits with the entire locus on a single tray. PCR reactions were followed by AmbiSolv when needed. Any ambiguities were resolved with the ABI sequencing kits. RSCA samples were performed on an ABI 377 and analyzed by RSCA Typer software. All PCR amplifications were carried out on an ABI 377 and analyzed by MatchTools software. All reactions were run with negative controls, and every 50th PCR reaction was repeated for quality control. Statistical Analysis For each cohort, data were descriptively summarized by using frequencies and percentages for all categorical variables, and by using means, standard deviations, medians, and interquartile ranges for all continuous variables. Cohort 1. Two outcomes were of primary interest in this cohort: humoral immune response (as measured by measles-specific IgG antibodies [optical density units]), after both one and two doses of the measles vaccine. We also considered secondarily the change in immune response before versus after the second dose of vaccine. We compared levels of immune response in homozygous versus heterozygous individuals by linear regression techniques. Variables representing locus-specific homozygosity status were created for each of the seven available loci. We first fit separate models for each of the HLA loci individually. We then created a variable indicating whether a subject was homozygous for at least one of the loci and assessed its relationship with immune response. Because of data skewness, and to allow for comparability with the original homozygosity paper, the original response values for each regression model were replaced with corresponding ranked values. In addition, we accounted for the stratified sampling design by applying weights inversely proportional to a subject’s selection probability. We ran two sets of analyses for each outcome. First, models were fit that included only the homozygosity variables of interest. Subsequent secondary models included the following set of potential confounding variables: age, sex,
791
race, and age at first MMR. Because results were similar in both sets of models, and again to allow for comparability with the original paper, only the unadjusted analyses are presented. Cohort 2. Six outcomes were of primary interest in this cohort. These outcomes reflected humoral immune response (as measured by IgG antibody titers [22]) and cellular immune response (T-cell proliferation as measured by stimulation indices), as induced separately by the MMR vaccine viruses. As with cohort 1, we compared levels of immune response in homozygous versus heterozygous individuals by linear regression techniques. Models assessing the effects of any homozygosity and locus-specific homozygosity were fit by using the same variable definitions and analytic approach as in cohort 1. We then calculated a homozygosity count for each subject. Values of this count could range from 0 to 6, depending on the number of loci for which the subject was homozygous. However, because of sparseness of data, individuals homozygous for five or six loci were grouped with those homozygous for four loci. We used this count variable to assess the possible dose-response relationship between homozygosity and immune response by fitting the count as a one-degree-of-freedom ordinal variable. In contrast to the cohort 1 analyses, the original response values for each regression model were replaced with corresponding logarithmic values, as this was the transformation that best normalized the data. In addition, all linear regression models were adjusted for the following potential confounding variables: age at enrollment, race, sex, age at the first MMR, and age at the second MMR. All statistical tests were two-sided, and all analyses were carried out by SAS software (SAS Institute, Cary, NC). RESULTS Characteristics of the study cohorts were similar; slightly more than half of each group were male (53% and 54%), and the majority of the children were caucasian (95% and 94%). The median ages of the participants at the time of the first and second MMR immunizations were also similar. However, the children in cohort 1 were younger than those in cohort 2, because the time since the second immunization was shorter for the children in cohort 1 compared with cohort 2 (Table 1). Departures from Hardy-Weinberg equilibrium were not observed in the first cohort; however, moderate disequilibrium was observed for the class IC allele (p ⫽ 0.03), and strong disequilibrium was observed for the DQB allele (p ⬍ 0.001). We first examined the group of children who had participated in our previous study examining the association between HLA homozygosity and measles antibody levels after one dose of vaccine [11]. As previously
792
J.L. St. Sauver et al.
TABLE 1 Characteristics of the two study populations Characteristic Sex, n (%) Female Male Race, n (%) White Other Age (months) at first immunization, median (IQR) Age (years) at second immunization, median (IQR) Time (years) since second immunization, median (IQR)
Cohort 1 (n ⫽ 178)
Cohort 2 (n ⫽ 346)
84 (47.2%) 94 (52.8%)
161 (46.5%) 185 (53.5%)
169 (94.9%) 9 (5.1%)
325 (93.9%) 21 (6.1%)
15.5 (15.0–16.5)
15.6 (15.1–17.5)
12.0 (11.0–12.5)
12.1 (11.5–12.6)
2.5 (1.7–3.2)
4.7 (2.8–6.1)
and rubella antibody levels in the second cohort of children. No significant associations were observed between homozygosity at specific alleles and measles or mumps antibody levels (Table 4). However, children who were homozygous at the HLA-DPB locus had higher median rubella antibody levels compared with children who were heterozygous for DPB (p ⫽ 0.0003; Table 4). DPB homozygosity did not differ significantly by child’s sex: 53 (33%) of the girls and 48 (26%) of the boys were homozygous at this locus (2 ⫽ 2.03, p ⫽ 0.15). In the same cohort of children, we also examined the associations between HLA homozygosity at specific loci and lymphoproliferative responses to measles, mumps, and rubella antigens. No associations were observed between HLA homozygosity and measles or rubella lymphoproliferation levels. However, children who were homozygous at the HLA class IA locus had lower mumps-specific lymphoproliferation levels compared with children who were heterozygous at this locus (p ⫽ 0.04). Children who were homozygous at any HLA locus also had lower mumps cellular responses compared with children who were heterozygous at all loci (p ⫽ 0.04; Table 5). Finally, we examined the associations between homozygosity at increasing numbers of HLA loci and measles, mumps, and rubella antibody and lymphocyte proliferation levels. Homozygosity at increasing numbers of HLA loci was associated with decreased mumps antibody levels (test for trend p value ⫽ 0.02; Table 6). Homozygosity at increasing numbers of HLA loci was also associated with decreased mumps lymphoproliferation levels (test for trend p value ⫽ 0.04; Table 7). Homozygosity at increasing numbers of HLA loci was not, however, associated with measles or rubella antibody or lymphoproliferation levels (Tables 6 and 7).
reported, homozygosity at the class IA and DQA alleles was associated with decreased measles antibody levels after one dose of vaccine. After a second dose of vaccine, mean antibody levels in children homozygous at the class IA and DQA alleles remained significantly lower than the mean antibody levels in children who were heterozygous at these loci (Table 2). However, the mean difference in antibody levels among children who were homozygous at the HLA loci was not significantly different from the mean difference in antibody levels among children who were heterozygous at these loci (Table 3). Overall, 169 (95%) of the children were seropositive, 7 (11%) were seroequivocal, and 2 (3%) were seronegative after receiving two doses of the MMR vaccine. We next examined the associations between homozygosity at specific HLA loci and measles, mumps,
TABLE 2 Associations between HLA homozygosity and mean measles antibody levels (optical density units) after one and two doses of the MMR vaccine (cohort 1) Measles antibody levels after one dose of MMR
Measles antibody levels after two doses of MMR
Locus
Heterozygotes, mean (SD)
Homozygotes, mean (SD)
p Valuea
Heterozygotes, mean (SD)
Homozygotes, mean (SD)
p Valuea
Class IA Class IB DPA DPB DQA DQB DRB Any locusb
1.75 (1.42) 1.71 (1.43) 1.66 (1.39) 1.68 (1.40) 1.72 (1.41) 1.68 (1.42) 1.72 (1.42) 1.82 (1.44)
1.31 (1.15) 1.52 (1.12) 1.70 (1.42) 1.77 (1.53) 1.45 (1.34) 1.75 (1.41) 1.62 (1.37) 1.62 (1.41)
0.003 0.43 0.95 0.56 0.07 0.45 0.43 0.05
3.11 (2.42) 3.05 (2.54) 3.10 (2.62) 3.01 (2.34) 3.13 (2.63) 3.00 (2.52) 3.05 (2.56) 3.08 (2.29)
2.68 (2.99) 2.95 (1.91) 3.01 (2.45) 3.10 (2.48) 2.61 (1.78) 3.21 (2.41) 3.05 (2.28) 3.04 (2.49)
0.03 0.73 0.70 0.77 0.03 0.13 0.31 0.75
a b
Analysis based on rank-transformed values; distributions are weighted to account for the sampling scheme. Analysis does not include DPA because of the extreme number of homozygotes at this locus.
HLA Homozygosity and MMR Immunity
793
TABLE 3 Comparisons of changes in antibody EIA values before and after MMR booster in Olmsted County, MN, schoolchildren homozygous or heterozygous at specific HLA loci (cohort 1)a Heterozygote
Homozygote
Locus
Subjects (n)
Mean (SD)
Subjects (n)
Mean (SD)
p Valueb
Class IA Class IB DPA DPB DQA DQB DRB Any locusc
153 165 56 120 144 141 146 63
1.36 (2.18) 1.34 (2.35) 1.44 (2.67) 1.33 (2.32) 1.40 (1.17) 1.31 (2.32) 1.32 (2.34) 1.25 (2.24)
23 13 121 43 31 37 31 94
1.37 (3.05) 1.43 (1.49) 1.31 (2.10) 1.33 (2.14) 1.16 (1.82) 1.46 (2.18) 1.43 (2.02) 1.42 (2.29)
0.54 0.55 0.62 0.68 0.09 0.26 0.38 0.18
a
Distribution estimates are weighted to account for the sampling scheme. Analysis based on rank-transformed values. c Analysis does not include DPA. b
DISCUSSION Our results from two cohorts of children, which used slightly different methods for measuring immunity and for typing HLA alleles, suggest that HLA homozygosity is not strongly associated with decreased measles, mumps, and rubella immune measures after two doses of the MMR vaccine. In fact, in cohort 1, a group of children who had data available after receiving both a first and second dose of the MMR vaccine, HLA homozygous and heterozygous children had similar increases in measles antibody levels after a second dose of measles vaccine. In addition, the vast majority of the children in this study (95%) were seropositive after two doses of vaccine, regardless of HLA homozygosity status. A similar pattern was observed in our second cohort of children, as there was no association between HLA homozygosity and low measles antibody levels after two doses of measles vaccine. However, children who were homozygous at the DPB locus had significantly higher rubella antibody levels compared with children who were heterozygous at the same locus. Children homozygous for class IA alleles also had lower mumps lymphoproliferation levels compared with children who were heterozygous at this locus. In addition, children who were homozygous at increasing numbers of loci had decreasing mumps antibody and lymphoproliferation levels. Measles We found little evidence that either overall homozygosity or homozygosity at specific HLA loci was associated with substantially decreased measles antibody levels or decreased measles lymphoproliferative responses after two doses of the MMR vaccine. Our results differ from our previous study, which suggested that homozygosity at increasing numbers of HLA loci was associated with decreased measles antibody levels [11]. When a subset of
these same children were examined in this study after their second dose of measles vaccine, the association between HLA homozygosity at the class IA and DQA loci and lower measles antibody levels remained. However, the increases in measles antibody levels after two doses of vaccine between HLA homozygotes and heterozygotes were very similar. It is somewhat surprising that homozygotes in the second cohort did not also have lower antibody levels at the class IA and DQA loci, but the antibody assays used in each of these cohorts differed slightly and the results are thus not directly comparable. Instead, it may be best to compare the two cohorts only in terms of seropositivity and seronegativity. In this regard, only a few children from the first cohort remained seronegative or equivocal after two doses of vaccine and failure to become seropositive was not associated with HLA homozygosity. These results are consistent with other studies, which have indicated that 2% to 18% of children may have low measles antibody levels after a single vaccination [23– 27], depending on age at immunization and time since immunization. However, the majority of individuals seronegative after a single measles vaccination become seropositive after a second dose of vaccine. Cohn et al. demonstrated that 70% of children revaccinated with measles vaccine maintained seropositive antibody levels after vaccination [28]. In our own work, we found that over 80% of 130 children who were initially seronegative after their first measles vaccination became measles antibody seropositive after their second immunization [29]. Similarly, McDermott et al. have found that at least some homozygotes initially seronegative for hepatitis B antibodies will seroconvert after revaccination with an appropriate dose of hepatitis B vaccine [22, 30]. Therefore, although HLA homozygosity may initially limit measles
794
TABLE 4 Differences in distribution of continuous measles, mumps, and rubella antibody values (IU/L) between children heterozygous and homozygous for specific loci (cohort 2)a Measles
Mumps
Rubella
Locus
Homozygote (n)
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valueb
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valueb
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valueb
Class IA Class IB Class IC DRB DQB DPB Any locus
75 28 57 46 122 101 251
1528 (779–2677) 1502 (699–2633) 1563 (746–2652) 1562 (779–2599) 1518 (692–2468) 1573 (685–2703) 1573 (584–2498)
1585 (675–2709) 2110 (1252–3525) 1528 (765–2970) 1359 (557–3265) 1593 (763–3042) 1495 (1019–2442) 1543 (765–2762)
0.66 0.08 0.51 0.79 0.59 0.34 0.28
759 (415–1395) 721 (414–1315) 772 (433–1358) 759 (414–1322) 804 (429–1371) 734 (433–1284) 809 (462–1627)
672 (395–1066) 774 (371–1252) 595 (355–970) 676 (332–1083) 676 (362–1189) 711 (385–1315) 699 (391–1251)
0.32 0.63 0.12 0.54 0.10 0.85 0.16
40 (23–62) 39 (21–61) 39 (23–61) 38 (22–61) 37 (22–60) 35 (20–55) 35 (21–57)
36 (20–58) 35 (24–55) 36 (19–59) 40 (22–65) 42 (23–65) 48 (25–76) 39 (22–65)
1.00 0.73 0.84 0.87 0.65 0.0003 0.19
a b
Linear regression analysis. p values are adjusted for age, sex, race, age at first MMR, and age at second MMR. Linear regression p values calculated from log-transformed values. Results significant at p ⱕ 0.05 in bold.
TABLE 5 Differences in distribution of continuous measles, mumps, and rubella lymphoproliferation (SI) values between children heterozygous and homozygous for specific loci (cohort 2) Measles
Mumps
Rubella
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valuea
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valuea
Heterozygote, median (IQR)
Homozygote, median (IQR)
p Valuea
Class IA Class IB Class IC DRB DQB DPB Any locus
75 28 57 46 122 101 251
3.6 (2.1–6.2) 3.6 (2.0–6.1) 3.6 (2.1–6.3) 3.7 (2.1–6.3) 3.9 (2.1–6.6) 3.6 (2.1–5.7) 3.9 (2.0–6.4)
3.7 (2.3–6.9) 4.4 (2.5–7.9) 3.7 (2.1–5.8) 3.1 (2.4–5.8) 3.3 (2.4–5.5) 3.8 (2.2–6.8) 3.6 (2.3–6.2)
0.95 0.23 0.88 0.75 0.34 0.16 0.88
4.9 (2.5–10.3) 4.9 (2.5–9.2) 5.0 (2.7–9.6) 4.9 (2.5–9.3) 5.7 (2.6–9.3) 4.6 (2.3–9.3) 5.9 (3.2–13.0)
4.6 (2.2–7.3) 4.4 (1.9–9.0) 3.5 (2.0–8.2) 3.9 (2.0–6.7) 3.8 (2.3–8.7) 5.0 (2.7–8.9) 4.4 (2.3–8.7)
0.04 0.55 0.10 0.52 0.25 0.81 0.04
2.3 (1.5–3.8) 2.3 (1.5–3.5) 2.2 (1.5–3.6) 2.3 (1.5–3.6) 2.4 (1.5–3.8) 2.3 (1.6–3.7) 2.4 (1.4–4.1)
2.3 (1.6–3.1) 3.2 (1.6–4.9) 2.5 (1.6–3.8) 2.1 (1.5–3.6) 2.1 (1.6–3.5) 2.1 (1.5–3.4) 2.3 (1.5–3.5)
0.78 0.07 0.09 0.52 0.27 0.41 0.92
a
Linear regression analysis. p values calculated using log-transformed values; p values adjusted for age at enrollment, age at first MMR vaccination, age at second MMR vaccination, sex, and race. Results significant at p ⱕ 0.05 are bold.
J.L. St. Sauver et al.
Locus
Homozygote (n)
HLA Homozygosity and MMR Immunity
795
TABLE 6 Differences in distribution of continuous measles, mumps, and rubella antibody values (IU/L) among children homozygous for increasing numbers of HLA loci (cohort 2) Measles
Mumps
Rubella
No. of homozygous loci
n
Median (IQR)
p Valuea
Median (IQR)
p Valuea
Median (IQR)
p Valuea
0 1 2 3 ⱖ4
167 109 52 9 9
1573 (584–2498) 1566 (820–2985) 1420 (764–2270) 1645 (823–2976) 2428 (994–3524)
0.19
809 (462–1627) 805 (444–1358) 686 (386–977) 548 (284–914) 752 (459–1618)
0.02
35 (21–57) 41 (23–65) 34 (20–60) 42 (27–69) 52 (24–73)
0.09
a
Linear regression analysis test for trend; calculated using log-transformed values; p values adjusted for age at enrollment, age at first MMR vaccination, age at second MMR vaccination, sex, and race. Results significant at p ⱕ 0.05 are bold.
antibody responses to MMR vaccination, this limitation can apparently be overcome with a second vaccination.
to detect other associations between HLA homozygosity and measures of rubella immunity.
Rubella With one exception, we found no associations between HLA homozygosity and rubella antibody and lymphoproliferation levels. However, children who were homozygous for the HLA-DPB locus had higher mean rubella antibody levels compared with children who were heterozygous for the same locus. This association is in the opposite direction than we had initially predicted, and the clinical significance of the association is unclear. The HLA-DPB*04 allele accounted for approximately 51% of the total DPB alleles observed in our population, and 81% of our homozygotes (data not shown). The DPBw4 molecule has been previously demonstrated to be important in the immune response to the hepatitis B surface antigen, as specific hepatitis B peptides are restricted to DPBw4 [31, 32]. Naturally processed rubella peptides may also be restricted to specific DPB alleles, and homozygosity at these specific DPB loci could theoretically confer an advantage over heterozygotes in response to the rubella vaccine. However, variation in measures of rubella immunity was extremely low in this group of children. Therefore, we may not have had enough power
Mumps Although homozygosity at specific loci was not associated with low mumps antibody levels, homozygosity at increasing numbers of HLA loci was associated with lower mumps antibody levels (test for trend p value ⫽ 0.02). In addition, homozygosity at HLA class IA was associated with a lower median lymphoproliferation level compared with heterozygosity at the same locus (p ⫽ 0.04), and homozygosity at increasing numbers of HLA alleles was associated with decreased lymphoproliferation levels (test for trend p value ⫽ 0.04). These results are in accord with our initial hypothesis, as we expected that HLA homozygosity would be associated with a lower immune response. However, it is surprising that these associations were observed only for a single MMR antigen. We may have detected this association because the distribution of immune responses was more variable for mumps than for either of the other antigens. Forty-three percent of our study population was seronegative or equivocal for mumps antibodies compared with only 14% of measles seronegatives/equivocals and 1% of rubella seronegatives/equivocals. This variability in im-
TABLE 7 Differences in distribution of continuous measles, mumps, and rubella lymphoproliferation (SI) values among children homozygous for increasing numbers of HLA loci (cohort 2)a Measles
Mumps
Rubella
No. of homozygous loci
n
Median SI (IQR)
p Valueb
Median SI (IQR)
p Valueb
Median SI (IQR)
p Valueb
0 1 2 3 ⱖ4
95 131 80 28 12
3.9 (2.0–6.4) 3.6 (2.5–6.1) 3.7 (1.8–6.1) 3.3 (2.5–7.9) 2.9 (2.4–6.3)
0.70
5.9 (3.2–13.0) 4.8 (2.3–9.3) 4.0 (2.2–8.0) 4.8 (2.1–9.0) 3.0 (2.0–5.2)
0.04
2.4 (1.4–4.1) 2.3 (1.5–3.3) 2.4 (1.5–3.7) 2.2 (1.6–3.6) 2.3 (1.5–3.4)
0.72
a b
Values adjusted for age at enrollment, age at first MMR vaccination, age at second MMR vaccination, sex, and race. Results significant at p ⱕ 0.05 are bold. Linear regression analysis test for trend; calculated using log-transformed values.
796
mune response would have increased our power to detect associations with HLA homozygosity. Overall, however, these data also indicate that a significant proportion of our study population may not be protected against mumps infection even after two doses of the MMR vaccine. A strength of our study arises from the availability of data for two cohorts of children that we were able to use to test our hypotheses. Despite the use of slightly different methods for measuring antibody levels and for typing HLA alleles, results were consistent across both groups of children in suggesting that two doses of the MMR vaccine appears to overcome adverse effects of HLA homozygosity on development of measles antibody levels after vaccination. A second strength of our study arises from the data available for the second cohort of children that allowed us to examine measures of both the humoral and the cellular arms of the immune system. Both humoral and cellmediated immunity are essential for conferring complete protection after vaccination with these viral antigens. However, data from other studies suggest that humoral and cellular immune measures (antibody and lymphoproliferation levels) are poorly correlated (data not shown). Therefore, different genetic factors may influence these measures differently, and it was important to examine the association between HLA homozygosity and each of the measures of immunity. A third strength of our study includes our ability to examine measures of the immune response to three viral antigens in the second cohort of children. Previous work indicates that the antibody levels to these vaccine antigens are not highly correlated [11]. These viruses likely produce different antigenic peptides that may be presented differently by the same HLA genes. Therefore, it is likely that HLA associations with rubella antibodies (or lymphoproliferation levels) will differ from those with measles or mumps immune measures, and it is not surprising that the association between HLA-DPB homozygosity and rubella antibodies is not observed with measles or mumps antibodies as well. Limitations of our study include lack of data on the immune measures to mumps and rubella after a single dose of MMR vaccine. Given this limitation, it was not possible to determine whether HLA homozygosity might also have been associated with decreased mumps and rubella immune responses after a single vaccination. Another potential study limitation results from our relatively low participation rate. Overall, about 29% of our initial study population agreed to participate, and it is possible that our study results are biased. However, neither HLA testing nor measles, mumps, or rubella antibody typing are routinely performed in this community. Therefore, subjects were highly unlikely to be aware
J.L. St. Sauver et al.
of their HLA genotype or their measles, mumps, or rubella immune levels. It is possible that parents may have preferentially enrolled their children in one of these studies because they were unsure of their child’s vaccine status and wanted to find out if their child was immune to these viruses. It seems unlikely, however, that parents could have preferentially enrolled children who were also homozygous for specific HLA alleles. Therefore, we think selection bias was minimal in this study; however, results may not be generalizable to other populations with different HLA distributions. A second potential limitation includes our use of EIA tests for measuring antibody levels rather than measurement of neutralizing antibody titers. However, use of neutralizing antibody titers is more labor-intensive than EIA testing, and a variety of commercially available EIA assay results have been previously demonstrated to correlate well with antibody neutralization measures [33– 35]. Therefore, EIA values are a reasonable surrogate for antibody neutralization assays and accurately reflect the measles, mumps, and rubella antibody levels of our study participants. Finally, given that we tested multiple associations, it may not be appropriate to designate a p value of 0.05 as our threshold for statistical significance. Instead, a more conservative threshold might be more appropriate. In conclusion, two doses of the MMR vaccine appear to overcome HLA homozygosity associations with low measles antibody levels observed after one dose of vaccine. Therefore, two doses of the MMR vaccine seem sufficient to induce protective levels of antibodies against both measles and rubella, regardless of HLA homozygosity status, providing further mechanistic evidence to support current public health recommendations for routine administration of two doses of the MMR vaccine. However, children who are HLA homozygous may be less protected against mumps compared with children who are heterozygous, even after two doses of the MMR vaccine. ACKNOWLEDGMENTS We thank Tina Agostini and Dennis Devitt for performing the Pel-Freez HLA typings. We also thank the many children and parents who participated in these studies. This research was funded by NIH grants AI 33144 and AI 48793 and by NIAMS grant AR030582.
REFERENCES 1. Apanius V, Penn D, Slev PR, Ruff LR, Potts WK: The nature of selection on the major histocompatibility complex. Crit Rev Immunol 17:179, 1997. 2. Borghans AM, Beltman JB, De Boer RJ: MHC polymor-
HLA Homozygosity and MMR Immunity
3. 4.
5.
6.
7.
8.
9. 10.
11.
12.
13.
14.
15.
16.
phism under host-pathogen coevolution. Immunogenetics 55:732, 2004. Klein J, Sato A: The HLA system. First of two parts. N Engl J Med 343:702, 2000. Klein J, Sato A: The HLA system. Second of two parts. [Erratum appears in N Engl J Med 2000;343:1504]. N Engl J Med 343:782, 2000. Penn DJ, Damjanovich K, Potts WK: MHC heterozygosity confers a selective advantage against multiple-strain infections. Proc Natl Acad Sci U S A 99:11260, 2002. Carrington M, Nelson GW, Martin MP, Kissner T, Vlahov D, Goedert JJ, Kaslow R, Buchbinder S, Hoots K, O’Brien SJ: HLA and HIV-1: heterozygote advantage and B*35-Cw*04 disadvantage [see comments]. Science 283: 1748, 1999. Tang J, Costello C, Keet IP, Rivers C, Leblanc S, Karita E, Allen S, Kaslow RA: HLA class I homozygosity accelerates disease progression in human immunodeficiency virus type 1 infection. AIDS Res Hum Retroviruses 15: 317, 1999. Thursz MR, Thomas HC, Greenwood BM, Hill AV: Heterozygote advantage for HLA class-II type in hepatitis B virus infection [letter] [published erratum appears in Nat Genet 1998;18:88]. Nat Genet 17:11, 1997. Thursz MR: Host genetic factors influencing the outcome of hepatitis. J Viral Hepat 4:215, 1997. Jeffery KJ, Siddiqui AA, Bunce M, Lloyd AL, Vine AM, Witkover AD, Izumo S, Usuku K, Welsh KI, Osame M, Bangham CR: The influence of HLA class I alleles and heterozygosity on the outcome of human T cell lymphotropic virus type I infection. J Immunol 165:7278, 2000. St Sauver JL, Ovsyannikova IG, Jacobson RM, Jacobsen SJ, Vierkant RA, Schaid DJ, Pankratz VS, Green EM, Poland GA: Associations between human leukocyte antigen homozygosity and antibody levels to measles vaccine. J Infect Dis 185:1545, 2002. Ovsyannikova IG, Jacobson RM, Vierkant RA, Jacobsen SJ, Pankratz VS, Poland GA: The contribution of HLA class I antigens in immune status following two doses of rubella vaccination. Hum Immunol 65:1506, 2004. Ovsyannikova IG, Jacobson RM, Vierkant RA, Jacobsen SJ, Pankratz VS, Poland GA: HLA class II alleles and rubella-specific humoral and cell-mediated immunity following MMR-II vaccination. J Infect Dis 191:515, 2005. Bautista-Lopez N, Ward BJ, Mills E, McCormick D, Martel N, Ratnam S: Development and durability of measles antigen-specific lymphoproliferative response after MMR vaccination. Vaccine 18:1393, 2000. Enrietto PJ, Payne LN, Hayman MJ: A recovered avian myelocytomatosis virus that induces lymphomas in chickens: pathogenic properties and their molecular basis. Cell 352 Pt 1369, 1983. Fugger L, Morling N, Ryder LP, Odum N, Svejgaard A: Technical aspects of typing for HLA-DP alleles using allele-specific DNA in vitro amplification and sequence-
797
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
specific oligonucleotide probes: detection of single base mismatches. J Immunol Methods 129:175, 1990. Kimura A, Dong RP, Harada H, Sasazuki T: DNA typing of HLA class II genes in B-lymphoblastoid cell lines homozygous for HLA. Tissue Antigens 40:5, 1992. Olerup O, Aldener A, Fogdell A: HLA-DQB1 and -DQA1 typing by PCR amplification with sequence-specific primers (PCR-SSP) in 2 hours. Tissue Antigens 41:119, 1993. Olerup O, Zetterguist H: DR “low-resolution” PCR-SSP typing—a correction and an update [letter; comment]. Tissue Antigens 41:55, 1993. Weyand CM, Xie C, Goronzy JJ: Homozygosity for the HLA-DRB1 allele selects for extraarticular manifestations in rheumatoid arthritis. J Clin Invest 89:2033, 1992. Poland GA, Jacobson RM, Schaid D, Moore SB, Jacobsen SJ: The association between HLA class I alleles and measles vaccine-induced antibody response: evidence of a significant association. Vaccine 16:1869, 1998. McDermott AB, Cohen SB, Zuckerman JN, Madrigal JA: Human leukocyte antigens influence the immune response to a pre-S/S hepatitis B vaccine. Vaccine 17:330, 1999. Brunell PA, Weigle K, Murphy MD, Shehab Z, Cobb E: Antibody response following measles-mumps-rubella vaccine under conditions of customary use. JAMA 250:1409, 1983. Mitchell LA, Tingle AJ, Decarie D, Lajeunesse C: Serologic responses to measles, mumps, and rubella (MMR) vaccine in healthy infants: failure to respond to measles and mumps components may influence decisions on timing of the second dose of MMR. Can J Public Health 89:325, 1998. Matson DO, Byington C, Canfield M, Albrecht P, Feigin RD: Investigation of a measles outbreak in a fully vaccinated school population including serum studies before and after revaccination. Pediatr Infect Dis J 12:292, 1993. Boulianne N, De Serres G, Ratnam S, Ward BJ, Joly JR, Duval B: Measles, mumps, and rubella antibodies in children 5– 6 years after immunization: effect of vaccine type and age at vaccination. Vaccine 13:1611, 1995. Johnson CE, Kumar ML, Whitwell JK, Staehle BO, Rome LP, Dinakar C, Hurni W, Nalin DR: Antibody persistence after primary measles-mumps-rubella vaccine and response to a second dose given at four to six vs eleven to thirteen years. Pediatr Infect Dis J 15:687, 1996. Cohn ML, Robinson ED, Faerber M, Thomas D, Geyer S, Peters S, Martin M, Martin A, Sobel D, Jones R, Larkin L, Richert JR: Measles vaccine failures: lack of sustained measles-specific immunoglobulin G responses in revaccinated adolescents and young adults. Pediatr Infect Dis J 13:34, 1994. Poland GA, Jacobson RM, Thampy AM, Colbourne SA, Wollan PC, Lipsky JJ, Jacobsen SJ: Measles reimmunization in children seronegative after initial immunization. JAMA 277:1156, 1997.
798
30. McDermott AB, Madrigal JA, Sabin CA, Zuckerman JN, Cohen SB: The influence of host factors and immunogenetics on lymphocyte responses to Hepagene vaccination. Vaccine 17:1329, 1999. 31. Celis E, Ou D, Otvos L Jr: Recognition of hepatitis B surface antigen by human T lymphocytes: proliferative and cytotoxic responses to a major antigenic determinant defined by synthetic peptides. J Immunol 140:1808, 1988. 32. Celis E, Karr RW: Presentation of an immunodominant T-cell epitope of hepatitis B surface antigen by the HLADPw4 molecule. J Virol 63:747, 1989. 33. Ratnam S, Gadag V, West R, Burris J, Oates E, Stead F, Bouilianne N: Comparison of commercial enzyme immu-
J.L. St. Sauver et al.
noassay kits with plaque reduction neutralization test for detection of measles virus antibody. J Clin Microbiol 33:811, 1995. 34. Cremer NE, Cossen CK, Shell G, Diggs J, Gallo D, Schmidt NJ: Enzyme immunoassay versus plaque neutralization and other methods for determination of immune status to measles and varicella-zoster viruses and versus complement fixation for serodiagnosis of infections with those viruses. J Clin Microbiol 21:869, 1985. 35. de Souza VA, Pannuti CS, Sumita LM, de Andrade Junior HF: Enzyme-linked immunosorbent assay–IgG antibody avidity test for single sample serologic evaluation of measles vaccines. J Med Virol 52:275, 1997.