Increased frequency of canine distemper virus‐specific antibodies in multiple sclerosis

Abstract Background and purpose Canine distemper virus (CDV) is a candidate agent in the etiology of multiple sclerosis (MS). Elevated anti‐CDV levels were previously found in the sera from MS patients compared with controls. We now investigated whether there was an age‐related association with the presence of antibodies specific to CDV‐hemagglutinin (H) protein in MS. Methods Sera from patients with MS, other neurological diseases, and inflammatory and/or autoimmune diseases, and healthy individuals were screened for anti‐CDV in an ELISA using linear peptides of the CDV‐H protein as antigen. Antibody levels to measles and varicella‐zoster virus were measured and served as controls. Results Analysis of the new cohort of MS patients and controls confirmed our initial finding of elevated anti‐CDV‐H levels in MS patients. An increase in measles but not varicella‐zoster virus antibody levels was found in MS patients compared with healthy controls. Data from the new cohort of patients and controls were combined with data from the original study and analyzed with respect to age distribution of anti‐CDV IgG. Mean CDV antibody levels were significantly elevated in each decade from 20 to 50 years of age in MS compared with healthy and disease controls. Antibody levels to measles virus were not consistently elevated during this age span. A striking relationship (p < .0001, odds ratio = 5.0) was observed between elevated anti‐CDV‐H levels and diagnosis of MS. Conclusions The finding that anti‐CDV levels are elevated in MS patients of all ages provides substantial evidence of a strong association between elevated anti‐CDV and MS.

Serological studies searching for CDV-specific antibodies (Abs) were hampered due to the high degree of homology between CDV and the closely related measles virus (MV). We devised a peptide-based enzyme-linked immunoabsorbent assay (P ELISA) using peptide sequences predicted to contain CDV-specific antigenic determinants to search for CDV Abs in animal and human sera (Rohowsky-Kochan et al., 1995). Initial studies demonstrated that sera from CDV-immunized animals reacted with the CDV peptides in P-ELISA, whereas sera from both animals and humans infected with MV did not react with the CDV peptides (Rohowsky-Kochan et al., 1995). A threefold increase in CDV antibody titers was observed in MS patients compared to both patients with other neurological disorders and healthy controls. These results provided the first unconfounded serological evidence that MS patients have elevated anti-CDV levels in support of the hypothesis that MS, in some instances, may be triggered by this neurotropic dog virus.
We now investigated whether there was an age-related association with the presence of CDV-H Abs in MS. To perform this analysis, we measured CDV H Abs in a new (previously not tested) cohort of patients with MS, inflammatory and/or autoimmune disorders (IAD), and other neurological diseases (OND) in order to increase the sample size for each age group. We determined antibody levels to an analogous peptide from the MV-H protein and to varicella-zoster virus (VZV). We show that anti-CDV-H levels were significantly higher in each decade from 20 to 50 years of age in MS patients compared with healthy and disease controls. This study extends and confirms our initial report that the highest levels of anti-CDV-H are found in MS patients.

| Selection and synthesis of CDV and MV peptides
CDV-H l, CDV-H2, and CDV-H3 peptides correspond to the most hydrophilic and presumed immunogenic regions of the CDV-H protein as assessed by the Hopp and Woods algorithm (Hopp & Woods, 1981). The sequence of the CDV peptides is as follows: CDV-H l (residues 234-248)-LVPDDIEREFDTREI; CDV-H2 (residues 365-380)-ALASEKQEEKGCLES; and CDV-H3 (residues 70-84)-FSRLLKEDMEKSEAV. There were no structural relationships between the three CDV-H peptides. A control peptide (16 amino acids in length) corresponding to a predicted antigenic determinant of the MV-H protein (residues 369-384) was also selected using this algorithm. The sequence of MVH2 is TTRTDDKLRMETCFQQ.
MV-H2 has only a single amino acid identity with its homologous CDV-H2 peptide. All four peptides were synthesized by Peptides International, Inc., Louisville, KY. The three CDV-H peptides were >95% pure, while the MV-H2 peptide was >80% pure as determined by analytical high-pressure liquid chromatography.

| Subjects
Research protocols were approved by the Institutional Review Board in accordance with regulations mandated by the Department of Health and Human Services and the Declaration of Helsinki. Informed consent was obtained from each subject. Sera were obtained from 94 randomly selected patients (mean age was 37.5 years; 61 females: 33 males) with clinically definite MS (Poser et al., 1983). The MS patients had relapsing-remitting or secondary progressive disease. The demographics of the patients and controls in the present study were similar to those reported in the original study (Rohowsky-Kochan et al., 1995). The mean duration of MS for 73 patients was 8.8 yr.
Samples were collected over 9 years, handled identically, and stored at −70°C until testing. None of the MS patients or disease and healthy controls had been previously tested for levels of anti-CDV-H.

| Detection of viral antibodies in human sera
CDV and MV peptide-specific Abs in sera were detected by P-ELISA as described (Rohowsky-Kochan et al., 1995). Briefly, 96-well U-bottom microtiter plates were coated overnight with CDV-H1, CDV-H2, CDV-H3, or MV-H2 peptides, washed, and nonspecific sites were blocked by incubating with 2% fetal bovine serum for 1 hr.
Sera (1:10 dilution) were added and incubated for 2 hr at 23°C, plates were washed, and an anti-human IgG horseradish peroxidase conjugate was added. After 1-hr incubation in the dark, a substrate solution-2, 2-azino-di (3-ethylbenzthiazoline sulfonic acid) (ABTS)-was added, followed by a brief incubation in the dark, and termination of the reaction with 1% sodium dodecyl sulfate (SDS). Absorbance values (405 nm) were read blindly on an automated microplate reader (BioTek Instruments) and corrected by subtracting optical densities from wells with 1% bovine serum albumin (BSA). Each serum was tested twice in separate assays at randomly assigned positions in the tray. The average of the two absorbance values was used. To control for assay variability, a serum with a high titer of anti-CDV-H and a nonreactive serum were tested in each assay. The assay was performed blindly with no biased selection of MS patient samples.
Equal numbers of MS and control sera were assayed simultaneously.
No diagnosis was written on the test tube or protocol sheet. A serum was considered positive for CDV or MV peptide antibodies if its absorbance value was two SDs above the mean absorbance value for HC. VZV-specific antibodies in human sera were measured using a commercially available ELISA directed against a varicella-zoster virus antigen (Wampole Laboratories) as per the manufacturer's instructions and processed similar to the P-ELISA.

| Determination of total serum IgG levels
Microtiter trays were coated with rabbit anti-human IgG antibody (25 ug/ml) (Jackson ImmunoResearch) in 0.05M carbonate buffer, pH-9.6, for 48 hr at 4°C. Trays were blocked for 24 hr at 4°C with 1.0% dry milk/phosphate-buffered saline and dilutions of human IgG standards (Chemicon) or test sera were plated in triplicate and incubated for 2 hr. After washing, a rabbit anti-human-lgG horseradish peroxidase conjugate (1:3000, Jackson ImmunoResearch) was added for 1 hr. The reactions were visualized by the addition of ABTS followed by 1% SDS to stop the reaction. Absorbance at 405nm was read on a microplate reader. IgG concentrations were determined by extrapolation from the standard curve plot.

| Purification of IgG from human sera
IgG was isolated from human sera using the Affi-Gel ® ID Protein A MAPS II Kit (Bio-Rad) as per manufacturer's instructions. Briefly, sera were diluted 1:1 with binding buffer and loaded on a protein A-bound gel column. The column was washed with 10-bed volumes of binding buffer till the serum protein peak was eluted, and the IgG was eluted using 5-to 7-bed volumes of elution buffer. IgGcontaining fractions were pooled, dialyzed overnight at 4°C with distilled H 2 0, lyophilized, reconstituted with phosphate-buffered saline, and stored at −70°C.

| Statistical analysis
Nonparametric rank-sum (Kruskal-Wallis and Wilcoxon) tests were performed to compare the distributions of absorbance values of CDV, MV, or VZV Abs in MS, IAD, and OND patients as well as HC.
The criterion for significance was two-tailed p < .05, adjusted for multiple comparisons by Bonferroni's inequality. Correlation between the three CDV peptide antibodies and MV antibodies was determined by Pearson's correlation test. The degree of association between positivity for anti-CDV-H and MS diagnosis was expressed as an odds ratio (OR) and evaluated by Fisher's exact test.
A significant difference was observed in the levels of anti-CDV-H2 (p < .004) and anti-CDV-H3 (p < .002) among patients with MS, IAD, OND, and HC (Table 1). Significantly elevated levels of anti-CDV-H2 were found in MS patients compared with HC, IAD, and OND patients. CDV-H3 levels were significantly elevated in MS patients compared with IAD patients and tended (p = .06) to be higher compared with both OND and HC. Twenty-two percent of MS patients had elevated CDV-H2 Abs compared with 2% of IAD patients, and 4% of OND patients and HC, whereas 19% of MS patients had elevated CDV-H3 Abs compared with 2% of IAD patients, 4% of OND patients, and 5% of HC.
A significant difference (p < .011) was seen in the levels of MV-H2 Abs among patients with MS, IAD, OND, and HC. MS patients exhibited significantly higher levels of anti-MV-H2 than HC and OND patients but not IAD patients (Table 1). Fifteen percent of MS patients had higher anti-MV-H2 levels compared with 5% and 2% of healthy and OND controls, respectively.

| Levels of VZV antibodies
There were no significant differences in the levels of anti-VZV between MS patients (1.208 ± 0.32, n = 57) and HC (1.311 ± 0.30, n = 34). Similar levels of VZV-specific Abs were noted in MS patients and HC who had elevated levels of anti-CDV-H and those who were negative for distemper antibodies. Levels of VZV-specific Abs were similar in the IAD and OND controls tested (data not shown).

| Association between CDV-H antibodies and MS diagnosis
We examined the relationship between the presence of elevated after controlling for anti-MV, the OR increased to 8.1 (95% CI = [3.1, 20.9]). In comparison, an OR of 2.9 (95% confidence interval [1.3, 6.9], p < .008) was observed between elevated anti-MV levels and MS diagnosis. In sum, there is an increased risk of MS (OR >5) when antibodies to all three CDV peptides are elevated (higher than 2 SD over the distribution in the healthy individuals) and an even greater risk (OR >7) when antibodies to CDV-H1 are elevated, and this risk is independent of the effect of the measles virus. There was no correlation between anti-CDV-H levels and disease duration, age at MS onset, age, or sex.

| Combined results from two independent studies
Similar levels of significance and degree of associations observed in the present and original studies justified combining and reanalyzing the data. A homogeneity p value of .77 and .83 (with and without OND controls, respectively) for the comparison of the odds ratio of the two studies provided a high degree of confidence that pooling the data was an appropriate analysis choice. For this analysis, IAD patients were grouped together with the OND patients and will be referred to as patients with other diseases (OD). A significant difference (p < .0001) was observed in mean antibody levels to each of the three CDV peptides among MS patients, OD patients, and HC. CDV-H1, CDV-H2, and CDV-H3 levels in MS patients were significantly increased compared with HC or OD patients (Table 2). CDV-H l, CDV-H2, and CDV-H3 IgG levels were elevated in 27%, 25%, and 21% of MS patients, respectively, 6%, 4%, and 5% of HC and 8%, 4%, and 5% of OD patients. Antibodies to the MV-H2 peptide were elevated in 23% of MS patients, 21% of HC, and 17% of OD patients.
The majority of positive patient and control sera reacted with all three CDV peptides, although sera that reacted with one or two out of three CDV peptides were identified. There are strong correla-  the 41-to 50-year-old group (Figure 2b). For the 51-to 60-year-old group, there was a trend for anti-CDV-H1 to be higher in MS patients compared with both control groups. Levels of anti-CDV-H2 and anti-CDV-H3 also remained stable and elevated over the wide age span (data not shown).

| Total IgG levels in sera
There was no significant difference in the IgG content between MS (11.4 ± 5.3 mg/ml) and OND patients (12.6 ± 10.0 mg/ml) or HC (13.7 ± 4.7 mg/ml). Comparison of IgG levels in individuals who had CDV-H Abs and those who were negative for CDV Abs did not show any significant differences in serum IgG levels in MS patients or the controls (data not shown).

| Analysis of binding of purified IgG
To further show that we were measuring binding of IgG-specific antibodies, we purified IgG from the sera of several MS patients and HC. Three sera contained anti-CDV H1, while three sera were negative for anti-CDV-H1. Serial dilution of sera and the corresponding purified IgG sample showed comparable absorbance values for binding to CDV H1 peptide in ELISA (Figure 3).

| D ISCUSS I ON
In the present study, we showed that anti-CDV-H levels were signifi- Increased B-cell immune reactivity to a variety of viruses has been described in MS (Tselis, 2011). Consistent with published studies, we observed elevated MV Abs in MS patients compared to patients with OND and HC (Adams & Imagawa, 1962;Brody et al., 1972;Haire et al., 1973). However, we found similar anti-MV levels in MS and AID patients, suggesting that such Abs may be a secondary consequence of an inflammatory response. The ele- CDV-H antibody levels do not appear to be affected by immunomodulatory agents such as steroids or interferon-β (Betaseron®).
We observed no significant difference in absorbance values when Exposure to animals, specifically dogs, may be linked with an increased risk of MS, although conflicting results have been reported in support of this association (Anderson et al., 1984;Bansil et al., 1997;Bauer & Wikstrom, 1977;Cook & Dowling, 1977;Flodin et al., 1988;Hernán et al., 1999;Hughes et al., 1980;Jotkowitz, 1977;Landtblom et al., 1993;Mititelu et al., 1986;Norman et al., 1983;Read et al., 1982;Siejka et al., 2016;Warren et al., 1982). Significantly more exposure of MS patients than controls to dogs with a CDV-like illness has been reported (Anderson et al., 1984;Warren et al., 1982), and others noted a trend that did not reach statistical significance toward more exposure to dogs with a CDV-like illness in MS patients (Bauer & Wikstrom, 1977;Hughes et al., 1980;Read et al., 1982). A recent study showed that exposure to dogs in early adolescence was associated with an increased risk of MS (Jong et al., 2019). In our studies, we collected and analyzed sera from patients and controls without Molecular mimicry between a virus and brain antigen may trigger an autoimmune response resulting in demyelination (Fujinami & Oldstone, 1985 with high titers (1:320 and 1:640, respectively) of anti-EBNA (Levin et al., 2005). In a longitudinal follow-up, a fourfold increase in anti-EBNA titers was associated with a risk of 3.0 (Levin et al., 2005).

| CON CLUS ION
The results of this study provide plausible, significant serological evidence corroborating epidemiologic data suggesting that CDV may be associated with MS, in at least some MS patients. Individuals with antibodies to the three CDV peptides have a five times greater chance of having MS than individuals in whom such antibodies were not detected and the risk of MS is independent of the effect of measles. The finding that anti-CDV levels are elevated in MS patients of all ages provides substantial evidence of a strong association between elevated anti-CDV and MS.

ACK N OWLED G M ENTS
We thank Traci Cioppa for her technical assistance, Annette Jotkowitz for helping obtain MS sera samples and Joan Skurnick, PhD. for performing the initial statistical analysis. We are grateful to Dr. Benjamin Blumberg for critically reading the manuscript and helpful suggestions. We are indebted to Dr. Marinos Dalakas for providing us with sera from patients with neurological and autoimmune diseases. We also thank Dr. Moore and her laboratory in the Department of Pathology, Immunology and Laboratory Medicine (Rutgers University, NJMS) for performing the VZV-antibody ELISA.

CO N FLI C T O F I NTE R E S T
The authors declare no financial or other conflicts of interests.

AUTH O R CO NTR I B UTI O N
Dr. Rohowsky-Kochan conceived the idea for this paper, designed all the experiments, analyzed the data, and wrote the original and

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/brb3.1920.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.