Genetic Analysis of the Influence of Pertussis Toxin on Experimental Allergic Encephalomyelitis Susceptibility: An Environmental Agent Can Override Genetic Checkpoints

Animals and immunizations

(B10.S/DvTe × SJL/J) × B10.S/DvTe backcross mice (BC1) were generated continuously over 12 mo from the same pool of B10.S/DvTe and (B10.S/DvTe × SJL/J) F₁ hybrid breeding stock. Inoculation of groups of mice, ranging in age from 6 to 12 wk, was staggered over the same period. Mice were injected s.c. at two sites on the posterior flank (0.15 ml/injection site) with 1.0 mg of SJL/J spinal cord homogenate (SCH), in 0.15 ml PBS, emulsified with an equal volume of CFA (16). A booster injection of SJL/J SCH + CFA, prepared in the same manner as the primary inoculum, was given on day 7. PTX-treated mice received 10 μg of crude PTX (17) i.p. at the time of inoculation and 5 μg 24–48 h later. Ninety-two mice were immunized using SCH, CFA, and PTX (PTX⁺) and 122 mice were immunized with SCH and CFA alone (PTX⁻). Mice were monitored daily for symptoms and graded from 0 to 4 as follows: 0, no clinical expression of disease; 1, floppy tail without hind limb weakness; 2, hind limb weakness with or without flaccid tail; 3, hind leg paralysis and floppy tail; and 4, hind leg paralysis accompanied by a floppy tail and urinary or fecal incontinence. Animals progressing to a score of 4 were euthanized. Animals were observed for 30 or 60 days, at which time they were sacrificed and their brains and spinal cord (SC) were retrieved for histological analysis. Severity of disease among affected animals was analyzed using a severity index generated by averaging the clinical scores for each animal over the number of days that it exhibited clinical symptoms. Liver tissue was collected for DNA isolation.

Histological evaluation

Brain and SC samples were dissected from calvaria and vertebral columns, respectively, and fixed by immersion in 10% phosphate-buffered Formalin (pH 7.2) at 4°C. Following adequate fixation, brain and SC were trimmed and representative transverse sections embedded in paraffin, sectioned at 5 μm, and mounted on glass slides. Sections were stained with hematoxylin and eosin for routine evaluation and luxol fast blue-periodic acid Schiff reagent for demyelination. Representative areas of the brain and SC, selected for histological evaluation, were chosen based on previous studies (18, 19). These included brain stem, cerebrum, and the cervical, thoracic, and lumbar segments of the SC. Scoring was based on a semiquantitative assessment for the following individual variables: 1) severity of the lesion as represented by each component of the histological assessment; 2) extent and degree of myelin loss and tissue injury (swollen axon sheaths, swollen axons, and reactive gliosis); 3) number of neutrophils/eosinophils comprising the inflammatory exudate; and 4) number of lymphocytes/macrophages comprising the inflammatory exudate. A score was assigned to the entire brain and to the entire SC based on a subjective numerical scale ranging from 0 to 5. A score of zero indicates no lesion, myelin loss, or inflammatory response was present in the tissue evaluated. The remaining scores were as follows: 1, minimal; 2, mild; 3, moderate; 4, marked; and 5, severe lesions, myelin loss, or inflammatory response present in the tissue.

Genetic analysis

Genomic DNA was isolated from liver tissue by standard techniques and used as template for PCR amplification using radiolabeled primers as described previously (16). Primers for microsatellite markers were purchased from Research Genetics (Huntsville, AL). Alleles at these markers were resolved on denaturing polyacrylamide gels and read from autoradiographic films directly into a MapManager QT database (20). Linkage maps were generated for the PTX⁺ and PTX⁻ groups using MapManager QT. Linkage of marker loci to disease susceptibility was tested by χ² in 2 × 2 contingency tables. Quantitative trait loci (QTL) were identified by interval mapping using model 3 of the Zmapqtl program in QTL Cartographer version 1.13 (21). Traits used for interval mapping included highest score seen during the course of EAE, severity index, monocyte/lymphocyte infiltration, suppuration, and demyelination of brain or SC. Critical values for declaration of significant linkage were determined by the permutation threshold theory (22). For this experiment, 1000 permutations of the data were generated to provide a sampling distribution of the χ² or the likelihood ratio test (LRT) statistics under the null hypothesis of no linkage. Each permutation was done by randomly shuffling and reassigning trait values among the BC1 animals, while holding the genotypic information fixed. Linkage analysis was then done for each permuted data set, and the distribution of the maximum test statistic from each of the 1000 permutations was used to set experimentwise significance thresholds. Linkages were reported if the test statistic for a particular marker locus was greater than the experimentwise critical value calculated for α = 0.10 (suggestive) or α = 0.05 (significant) for that trait. Significant linkages, or suggestive linkages that replicate the findings of a previous study, were given eae locus designations.

Interaction between marker loci

To test the hypothesis that significant QTL interact, we used a regression model containing significant marker loci and two-locus interaction variables between significant marker loci. Regression models with dummy coding for marker loci were analyzed in SAS using PROC REG (23). Significance of the overall model was assessed using an F statistic, and significance of interaction variables was assessed using a T statistic (24).

Incidence of EAE in the two treatment groups

Signs of EAE were more frequent, and the latency to onset was significantly shorter, in both males and females in the PTX⁺ group (69% affected) compared with the PTX⁻ group (43% affected) (Table I⇓). Although the incidence of disease was higher in PTX⁺ mice, there was no difference in susceptibility between males and females within either group. However, disease as measured by average peak score (or by a severity index, data not shown) was less severe in the PTX⁺ group, especially in females (Table I⇓). A substantial cohort of mice in the PTX⁻ group developed neuropathology in the absence of detectable clinical signs. We have recently termed this type of disease benign EAE (25). Benign EAE was prominent in PTX⁻ mice: 40% clinically unaffected PTX⁻ mice showed monocytes and lymphocytes in their SCs without any outward signs of EAE, whereas only 16% of unaffected PTX⁺ mice had this phenotype. This was not influenced by sex.

Neuropathology in the two treatment groups

SC and brain samples were obtained from and scored for the occurrence and degree of mononuclear infiltration, suppuration, and demyelination. The character and distribution of CNS histopathological lesions observed in this study were consistent with those reported in previous studies (18, 19, 25). Inflammatory responses in susceptible animals ranged from those with a predominantly neutrophilic response admixed with a smaller monocytic/lymphocytic component to those with a predominantly monocytic/lymphocytic response. CNS tissue response ranged from mice with no tissue injury to those with loss of myelin, reactive gliosis, swollen axon sheaths, and swollen axons. In all mice with histopathology, the inflammatory response had a perivascular distribution that was predominantly observed in the meninges and in the white matter. In the SC, predilection for the nerve root entry zone was observed, as reported previously (19).

When mice from the two treatment groups were examined for the presence of histopathology in the CNS (brain or SC), the PTX⁺ group again had a higher incidence of affected mice than the PTX⁻ group for either CNS site (data not shown). Differences between males and females were not observed for histopathological traits in the brain or SC of PTX⁺ mice. In contrast, female mice in the PTX⁻ group had significantly greater brain involvement than males (Table II⇓). Because animals were not all euthanized at the same time points subsequent to the manifestation of their disease, overall between-group comparisons could not be made for histological traits. However, we could analyze a subset of EAE-susceptible mice from each treatment group (51 PTX⁺ and 37 PTX⁻) that were sacrificed within 1 wk of their last clinical sign, because the lesions in these mice would be expected to be comparable. Surprisingly, the extent of demyelination at either CNS site was substantially less severe in the PTX⁺ than in the PTX⁻ subset (the mean value for SC demyelination in these PTX⁺ mice was 1.39 ± 0.7 and for the PTX^- mice was 2.32 ± 0.63, p < 1 × 10⁻⁸; the mean value for brain demyelination in PTX⁺ mice was 0.25 ± 0.44 and for the PTX⁻ subset was 0.89 ± 0.88, p = 0.0002).

Genetic control of EAE

Genome scanning was performed on both the PTX⁺ and PTX⁻ cohorts. We analyzed both the binary trait of EAE incidence (or susceptibility) and the quantitative traits of severity and histopathology. For PTX⁺ mice, linkage to the binary trait of susceptibility was found in an interval that is marked by D9 Mit105 (χ² values were 11.7 and 9.3, indicating probability of linkage at α = 0.05 and α = 0.1, respectively). This effect came largely from male mice (χ² = 17.0). A QTL designated eae9 has previously been identified in the same interval in an F₂ intercross with the same parental strains (16). In the PTX⁻ cohort, linkage to susceptibility was found to markers on chromosome 8, peaking at D8 Mit190 (χ² = 18.6, α = 0.1).

Loci that showed evidence of linkage to severity or histopathological signs of EAE are given in Table III⇓. In the PTX⁺ cohort, we again found linkage to a single major QTL (eae9) between D9 Mit22 and D9 Mit24. The SJL-derived allele at eae9 is associated with increased severity. Stratification by sex revealed that one phenotype of eae9 is to increase monocyte/lymphocyte infiltration in the SC in males. No linkage to any other locus, including previously identified eae-m loci, was seen in PTX⁺ mice. In contrast, five linkages were seen in the PTX⁻ cohort, including two QTL previously shown to play a role in this strain combination (16, 25). The most significant QTL in the PTX⁻ cohort is on chromosome 8 between D8 Mit3 and D8 Mit31. Susceptibility, severity, and SC demyelination are associated with the B10.S-derived allele at this locus, which we have designated eae14. In addition to eae14, a locus on chromosome 11 that colocalizes with eae7 (16) contributed to severity in this cohort. As in the F₂ intercross between these same strains, sex-specific linkages were seen in the PTX⁻ mice. Females from the PTX⁻ group showed linkage to chromosome 10 for severity of clinical signs in an interval between D10 Mit126 and D10 Mit10. We have designated this locus as eae17. A QTL controlling severity of mononuclear infiltrates in the SC was mapped to a broad region of chromosome 16. Increased severity of inflammation was associated with inheritance of the SJL allele at this locus, which was previously designated as eae11 (25). In PTX⁻ males, a strong QTL was mapped to chromosome 18. The SJL-derived allele at this locus (eae18) resulted in greater infiltration and demyelination in the SC. No QTL for brain histopathology met experimentwise cutoffs for suggestive (α = 0.10) or significant (α = 0.05) linkage in either cohort.

Multiple linear regression analysis

Multiple linear regression was used to test the hypothesis that significant QTL interact. Significant marker loci for disease severity (D8 Mit190, D11 Mit98, and D10 Mit42), SC demyelination (D8 Mit190 and D18 Mit3), and SC monocytic/lymphocytic infiltration (D8 Mit190, D16 Mit50, and D18 Mit3) were analyzed as independent variables in multiple linear regression analyses with the appropriate phenotype as the dependent variable. To investigate possible interactions between significant marker loci, two-locus interaction terms were added to the regression models as independent variables. None of the interaction terms was significant (p > 0.05). Without interaction variables, statistical significance was achieved for disease severity (F = 13.59, p < 0.0001), SC demyelination (F = 10.54, p < 0.0001), and SC monocytic/lymphocytic infiltration (F = 6.31, p = 0.0005).