Identification of Quantitative Trait Loci Governing

Identification of Quantitative Trait Loci Governing Arthritis Severity and Humoral Responses in the Murine Model of Lyme Disease1 Janis J. Weis,* Barbara A. McCracken,† Ying Ma,* Daryl Fairbairn,† Randall J. Roper,† Tom B. Morrison,* John H. Weis,* James F. Zachary,† R. W. Doerge,‡ and Cory Teuscher† A spectrum of disease severity has been observed in patients with Lyme disease, with ;60% of untreated individuals developing arthritis. The murine model of Lyme disease has provided strong evidence that the genetic composition of the host influences the severity of arthritis following infection with Borrelia burgdorferi: infected C3H mice develop severe arthritis while infected C57BL/6N mice develop mild arthritis. Regions of the mouse genome controlling arthritis severity and humoral responses during B. burgdorferi infection were identified in the F2 intercross generation of C3H/HeNCr and C57BL/6NCr mice. Rear ankle swelling measurements identified quantitative trait loci (QTL) on chromosomes 4 and 5, while histopathological scoring identified QTL on a unique region of chromosome 5 and on chromosome 11. The identification of QTL unique for ankle swelling or histopathological severity suggests that processes under distinct genetic control are responsible for these two manifestations of Lyme arthritis. Additional QTL that control the levels of circulating Igs induced by B. burgdorferi infection were identified on chromosomes 6, 9, 11, 12, and 17. Interestingly, the magnitude of the humoral response was not correlated with the severity of arthritis in infected F2 mice. This work defines several genetic loci that regulate either the severity of arthritis or the magnitude of humoral responses to B. burgdorferi infection in mice, with implications toward understanding the host-pathogen interactions involved in disease development. The Journal of Immunology, 1999, 162: 948 –956.

L

yme disease is caused by tick-transmitted infection with the spirochete Borrelia burgdorferi (1). B. burgdorferi infection is responsible for .90% of vector borne disease in the United States, with 16,461 cases reported to the Center for Diseases Control in 1996 (2). In humans, symptoms can be manifest in a variety of tissues, with great variability in the intensity of disease displayed by different individuals (3). One of the most common late manifestations of disease is arthritis, affecting 60% of individuals not treated at the time of the tick bite. Arthritis is characterized by tendonitis and synovitis and is associated with bacterial invasion of joint tissues (4, 5). Several factors may be responsible for the variability of disease severity in humans, including the characteristics of the infecting strain of B. burgdorferi, possible coinfection with other pathogens, and inherent differences in individual responses to infection (3, 6, 7). Studies with the murine model of Lyme disease have provided evidence that the genetics of the host contributes to the severity of *Department of Pathology, University of Utah, Salt Lake City, UT 84132; †Department of Veterinary Pathobiology, University of Illinois at Urbana-Champaign, Urbana, IL 61802; and ‡Departments of Agronomy and Statistics, Purdue University, West Lafayette, IN 47907 Received for publication June 18, 1998. Accepted for publication September 30, 1998. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by Public Health Service Award AI-43521 to J.J.W. and C.T., AI-32223 to J.J.W., HD21926, HD27275, AI40712, NS36526, and National Multiple Sclerosis Society Grant RG2659 to C.T., and AI-24158 to J.H.W. R.J.R. is supported by National Institute of Health Training Grant T32 GM07283. The project described was also supported in part by an award from the American Lung Association (J.H.W.). The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the American Lung Association. 2 Address correspondence and reprint requests to Dr. Janis J. Weis, Department of Pathology, University of Utah School of Medicine, 50 N. Medical Drive, Salt Lake City, UT 84132. E-mail address: [email protected]

Copyright © 1999 by The American Association of Immunologists

B. burgdorferi-induced arthritis (8). Several substrains of the C3H mice develop severe arthritis 3– 4 wk following infection (8, 9). The arthritis, which is representative of human disease, is reproducibly found in the rear ankle joints, where its development can be monitored by measurement of ankle swelling and by histopathological assessment (10). High levels of B. burgdorferi DNA can be detected in arthritic joint tissues (5). BALB/cAnN and C57BL/6NCr mice are relatively resistant to B. burgdorferi-induced arthritis; however, resistance in these two strains appears to be mediated by distinct mechanisms (11). Arthritis in BALB/cAnN mice is mild when the inoculum dose is small (102–103 B. burgdorferi) but becomes severe at high inoculum doses (104–105). Mild arthritis in BALB/cAnN mice is correlated with 5- to 10-fold fewer spirochetes in joint tissues than found in tissues of severely arthritic mice. Arthritis resistance in C57BL/6NCr mice has distinct characteristics, with mild arthritis seen even in mice infected with very high inoculums of B. burgdorferi (105). Spirochete levels in joint tissues of C57BL/6NCr mice are high at all infectious doses, similar to those found in severely arthritic C3H/HeNCr mice. This suggests that the localized inflammatory responses in infected joint tissues of C57BL/ 6NCr mice are more tightly regulated than in C3H/HeNCr mice. These observations suggest that genetic traits possessed by C57BL/6NCr mice are responsible for more regulated inflammation seen in their joints and that the responsible genetic loci may be identified by quantitative trait assessment in intercross populations bred from C57BL/6NCr and C3H/HeNCr mice. In this study, the mode of inheritance of susceptibility to severe arthritis was analyzed by microsatellite mapping of quantitative trait loci (QTL)3 in the F2 intercross of C3H/HeNCr and C57BL/

3 Abbreviations used in this paper: QTL, quantitative trait loci; LOD, logarithm of the odds favoring linkage; CM, centimorgan.

0022-1767/99/$02.00

The Journal of Immunology 6NCr mice. Several QTL regulating arthritis severity were identified by measurements of rear ankle swelling and by histopathological scoring of joints taken from infected mice. Interestingly, the two measurements of arthritis severity were linked to distinct genomic regions. Additional, distinct QTL were identified that regulate plasma levels of specific and total Igs in infected F2 mice. These results indicate that the severity of arthritis induced by B. burgdorferi infection in mice is regulated by multiple genes, that the magnitude of the humoral response does not regulate arthritis severity, and that distinct genetic loci may be involved in different components of Lyme arthritis development.

Materials and Methods Mice C3H/HeNCr, C57BL/6NCr, and B6C3F1 (C3H/HeNCr 3 C57BL/6NCr) mice were obtained from the National Cancer Institute (Bethesda, MD). B6C3F1 males and females were bred to generate F2 mice. All mice were housed in the Animal Resource Center at the University of Utah Medical Center according to guidelines of the National Institutes of Health for the care and use of laboratory animals.

B. burgdorferi infection Mice between 5 and 7 wk of age were infected by intradermal injection in the shaven back with 2 3 103 of the N40 isolate of B. burgdorferi (provided by Dr. Stephen Barthold, University of California at Davis, CA (11) at passage 3 from an infected mouse) and used at passage 5. Spirochetes were grown in BSK-H medium containing 6% rabbit serum (Sigma, St. Louis, MO), enumerated by counting in a Petroff-Hauser chamber, and diluted with sterile medium. Infection was followed in five male and five female mice of each parental strain (C3H/HeNCr, C57BL/6NCr, and B6C3F1) and in 150 F2 mice (mixture of male and female). Mock-infected groups consisted of five male and five female mice of each parental strain that received an intradermal injection of an equivalent volume of BSK-H.

Detection of B. burgdorferi-specific and total circulating Ig levels Serum samples from infected and control mice were taken from each mouse at the 4-wk sacrifice and analyzed for B. burgdorferi-specific Ab by a quantitative Ab capture ELISA (12). Eleven columns per 96-well plate were coated with 10 mg of B. burgdorferi sonicate per ml. The 12th column was coated with goat anti-mouse IgG, IgM, and IgA (Zymed, South San Francisco, CA) at a concentration of 5 mg/ml. Dilutions of serum were added to wells coated with B. burgdorferi sonicate, and known concentrations of murine IgG and IgM were added to the wells coated with goat anti-mouse Ig to form a standard curve. Unbound samples were removed by washing, and the anti-B. burgdorferi IgG and IgM were detected by the addition of horseradish peroxidase-conjugated Abs to murine IgG or IgM (Boehringer Mannheim, Mannheim, Germany), respectively. Plates were developed by incubation with 0.4 mg of O-phenylenediamine per ml and 0.01% H2O2 (Sigma) and read at an OD of 492 nm with a Vmax 96-well microtest plate spectrophotometer (Molecular Devices, Palo Alto, CA). Comparison of the OD of B. burgdorferi wells with the OD of the standard curve allowed estimation of Ag-specific Ab levels. Sera from mock-infected mice had B. burgdorferi-specific IgG levels ,100 ng/ml, whereas sera from infected mice had specific IgG levels .20 mg/ml. This value was also used to confirm that experimental mice had become infected. The concentration of total circulating IgG and IgM was determined by Ab capture ELISA with plates coated with goat anti-mouse IgG, IgM, and IgA as described above (10). Dilutions of serum were added to wells, and the amount of bound murine Ig was detected by the addition of horseradish peroxidase-conjugated Abs to murine IgG or IgM and read as described above. Ig levels were determined by comparison with a standard of IgG or IgM included on each plate.

Measurement of the ankle joints Rear ankle joints of mice were measured with a metric caliper (Mitutoyo, Tokyo, Japan) at sacrifice. Measurements were taken in the anterior/posterior position, with the ankle extended, through the thickest portion of the ankle (10). Ankle joints from uninfected mice have a diameter of ;3 mm at this position, with measurements up to 4.5 mm observed in severely arthritic mice.

949 Histopathology of ankle joints The rear ankle joint displaying the greatest swelling at the 4-wk sacrifice was taken from each mouse for histological analysis. Samples were fixed in 10% formalin, decalcified, embedded in paraffin, and sections were stained with hematoxylin and eosin. Slides were viewed in a blinded fashion and given a score for severity (11). A score of 41 indicated the greatest severity in tissues of the ankle and tibia and was characterized by a large region of edema with many neutrophils present, thickening of the tendon sheath, and evidence of bone and cartilage abnormalities within the tendon sheath. A score of 11 displayed slight thickening of the tendon sheath, with little edema or neutrophil infiltration. A score of 0 was given to samples indistinguishable from mock-infected controls. Scores of 2 and 3 were assigned to samples with intermediate pathology. Histopathology was independently evaluated by a second pathologist using slightly different criteria for scoring, which placed greater weight on certain aspects of pathology including inflammatory cell infiltrate and bone remodeling. Correlation analysis using the PROC CORR function of SAS found a correlation of 0.7 between the two assignments (13). A second quantification of histopathological severity was obtained by measurement of the thickness of the tendon sheath of the cranial tibial muscle using Image Pro 1.3 software (Media Cybernetics, Silver Spring, MD) and a Leitz Orthoplan microscope (Leitz, Stuttgart, Germany) fitted with a Sony CCD-iris/RGB color video camera (Sony, Tokyo, Japan). The average of four measurements made at approximately equal spacing was used for QTL analysis.

Preparation of DNA from infected tissues Mice were sacrificed at 4 wk following infection, and rear ankle joint tissues and hearts were prepared as previously described (11, 12). Tissues were placed individually in 15-ml polypropylene tubes containing 2.5 ml of a 0.1% collagenase A (Boehringer Mannheim) solution in PBS, pH 7.4. Samples were digested with collagenase for 4 h at 37°C and then mixed with an equal volume of 0.2 mg/ml proteinase K (Boehringer Mannheim) in 200 mM NaCl, 20 mM Tris-HCl, pH 8.0, 50 mM EDTA, and 1% SDS for 16 h at 55°C. DNA was recovered by extraction with an equal volume of phenol:chloroform and precipitation with ethanol. Following digestion with 1 mg/ml DNase-free RNase (Sigma), the DNA samples were subjected to a second extraction and precipitation, and finally resuspended in 1.5 ml dH2O. The average recovery was 200 – 400 mg DNA from each heart and ankle, as determined by absorbency at 260 nm.

Quantification of B. burgdorferi DNA in mouse tissues B. burgdorferi DNA levels were determined in DNA prepared from hearts of infected F2 and parental mice by continuous monitoring PCR using the Light Cycler (Idaho Technologies, Idaho Falls, ID) (14). This technique, which is fully described elsewhere,4 monitors the cycle-by-cycle accumulation of fluorescently labeled product. The cycle at which the product is first detected is used as an indicator of relative starting copy. The chromosomally encoded recA gene was used for quantitation of B. burgdorferi levels in mouse hearts and was normalized relative to 106 copies of a single-copy mouse gene.

Genotyping of F2 mice To characterize the genetic control of arthritis severity and humoral responses in C3H/HeN mice, we constructed a genetic linkage map of the mouse genome using 150 F2 mice of a C3H/HeNCr and C57BL/6NCr intercross. A total of 195 microsatellite markers that discriminated between C3H/HeNCr and C57BL/6NCr alleles were used to perform a 20-cM scan of the mouse genome. Microsatellite primers were either purchased from Research Genetics (Huntsville, AL) or synthesized according to sequences obtained through the Whitehead Institute/Massachusetts Institute of Technology mouse genome database (www.genome.wi.mit.edu/cgi-bin/mouse/ index). Microsatellite typing was done as described (15–17). PCR was performed using a PTC100 Programmable Thermal Controller (MJ Research, Watertown, MA) with the following program: 1 cycle at 94°C for 3 min; 35 cycles at 94°C for 1 min, 55°C for 2 min, 72°C for 2 min; 1 cycle at 72°C for 10 min. The PCR-amplified microsatellite size variants were resolved by electrophoresis on large-format denaturing 7% polyacrylamide gels and visualized by autoradiography.

4 T. B. Morrison, Y. Ma, J. H. Weis, and J. J. Weis. Rapid and sensitive quantification of Borrelia burgdorferi-infected mouse tissues using continuous fluorescent monitoring of PCR. Submitted for publication.

950

GENETIC REGULATION OF MURINE LYME DISEASE

Table I. Summary of quantitative traits associated with B. burgdorferi infection in parental mice

Genotype

C3H/HeNb (10) C57BL/6Nb (10) B6C3F1b (10) a b

Ankle Measurementsa (mm)

Histopathology Scorea (0–41)

Tendon Sheath Thicknessa (mm)

Total IgM (mg/ml)a

4.23 6 0.23

3.7 6 0.48

146.6 6 60.55

3.01 6 0.16

0.8 6 0.71

3.56 6 0.22

1.8 6 0.86

Total IgG (mg/ml)a

Specific IgM (mg/ml)a

B. burgdorferi (No. per 106 mouse cells)a

1.1 6 0.14

5.2 6 0.47

4.5 6 0.41

2117 6 887

67.4 6 17.73

0.97 6 0.18

2.9 6 0.63

2.5 6 1.3

246 6 149

99.0 6 42.15

0.93 6 0.12

3.7 6 0.98

3.4 6 1.2

463 6 254

Traits quantified as described in Materials and Methods. Mean and SD are indicated for each phenotypic trait. Number of mice per group is indicated in parentheses.

Linkage analysis Linkage maps were estimated using the Kosambi map function within the MAPMAKER/EXP computer package (www.genome.wi.mit.edu/ genome_software/other/mapmaker.html) (18). Quantitative trait linkage analysis was performed by interval mapping using MAPMAKER/QTL under the assumptions of a free genetic model (19). Composite interval mapping was performed using QTL-Cartographer (statgen.ncsu.edu/qtlcart/ cartographer.html) (model 6) when the possibility of multiple loci linked on a single chromosome existed (20 –22). Significant markers from the entire genome were used as a control for the genetic background. A window size of 15 cM was used on either side of the test interval to analyze a particular region independently from the rest of the chromosome. Significant linkage of QTL to marker loci for all analyses was determined by permutation threshold analysis using QTL-Cartographer (22, 23). This method of analysis takes into account the specifics of the experimental situation and satisfies the multiple testing issues implicit in the genome (23). For this experiment, 1000 permutations of the actual data were generated to provide the sampling distribution of test statistics under a null hypothesis of no linkage. A permutation was generated by randomly reassigning all of the phenotypes from every mouse to each of the genotypes generated by the microsatellite typing. A new test statistic (LOD score) was found at each test site in each permutation, and the highest test statistic from the entire genome was selected. The highest test statistics for each of the 1000 permutations were then ordered, and the threshold for significant experimentwise linkage was taken at the 95% level (24). Loci showing significant linkage were given designations resulting in the nine proposed disease-modifying loci in this study.

Interaction between marker loci Interaction between marker loci was investigated using a general linear model in SAS (PROC GLM) (13). Models were used to analyze the effects of significant loci on several phenotypes associated with infection. Significant markers for ankle swelling (D5 Mit431 and D4 Mit87), histopathology (D11 Mit120 and D5 Mit312), total IgM (D6 Mit194 and D12 Mit139), specific IgM (D6 Mit200, D11 Mit51, and D17 Mit232), and total IgG (D9 Mit48 and D12 Mit139) were analyzed as independent variables using multiple regression analyses with the appropriate phenotype as the dependent variable (25). To investigate possible interactions between significant marker loci for each phenotype associated with infection, two-locus interaction terms were added to the multiple linear regression models as independent variables. The model was made with the dependent and independent terms, and all model assumptions were verified.

Results General features of murine Lyme disease in the F2 and parental generations C3H/HeNCr, C57BL/6NCr, B6C3F1 (C3H/HeNCr 3 C57BL/ 6NCr), and F2 intercross mice were studied for arthritis severity and humoral immune response following infection with B. burgdorferi (characteristics for parental mice are summarized in Table I). No differences were seen between responses of infected parental male and female mice, allowing data for males and females to be pooled. C3H/HeNCr mice developed severe arthritis when infected with B. burgdorferi, with a mean ankle measurement of 4.23 6 0.23 mm and an average histopathological score of 3.7 6 0.48. Infected C57BL/6NCr mice developed milder arthritis with

ankle measurements of 3.01 6 0.12 mm and histopathological scores of 0.8 6 0.71. Arthritis measurements for these parental strains formed nonoverlapping groups ( p , 0.005 for both ankle measurements and histopathology). Arthritis was intermediate in B6C3F1 mice, with ankle measurements of 3.56 6 0.22 mm and histopathology of 1.8 6 0.86. The complete range of arthritis severity was observed in the F2 population, as is graphically displayed for ankle swelling and tendon sheath thickness in Fig. 1, A and B. In this study, B. burgdorferi-specific IgM levels were twofold greater in C3H/HeNCr mice than in C57BL/6NCr mice at 4 wk following infection (4.5 6 0.41 mg/ml vs 2.5 6 1.3 mg/ml, p , 0.005; Table I). Less obvious differences were also observed in total IgM levels in infected mice of these two strains (1.1 6 0.14 mg/ml for C3H/HeNCr and 0.97 6 0.18 mg/ml for C57BL/6NCr) and in the B. burgdorferi-specific IgG of the parental strains (76 6 13 mg/ml for C3H/HeNCr and 44 6 11 mg/ml for C57BL/6NCr mice). The F2 population of mice displayed a complete range of values for total IgM (0.26 –1.75 mg/ml), B. burgdorferi-specific IgM (0.96 –11.08 mg/ml), and B. burgdorferi-specific IgG (21.5– 101.0 mg/ml). B. burgdorferi levels in tissues of infected mice were determined by quantitative PCR. As found previously, B. burgdorferi levels in ankle joints of C3H/HeNCr and C57BL/6NCr mice were similar (11); thus, levels in joint tissues of F2 mice were not assessed. In contrast, hearts from infected C3H/HeNCr mice had ninefold greater levels of B. burgdorferi DNA than did hearts from C57BL/6NCr mice, suggesting a genetic regulation of spirochete persistence in some tissues (Table I). B. burgdorferi levels in the hearts of F2 mice ranged from 14 to 6158 DNA molecules per 106 mouse cells.

Arthritis severity maps to chromosomes 4, 5, and 11 To characterize the genetic control of arthritis severity and humoral responses in C3H/HeNCr mice, we constructed a genetic linkage map of the mouse genome using 150 F2 mice of a C3H/ HeNCr and C57BL/6NCr intercross. A total of 195 microsatellite markers that discriminated between C3H/HeNCr and C57BL/ 6NCr alleles were used to perform a full genome scan. Intermarker recombination fractions on different chromosomes ranged from 0.037 to 0.107 with a maximum intermarker distance of between 8 and 18 cM. Three parameters were used to assign arthritis severity: measurement of rear ankle joints, histopathological severity, and thickening of tendon sheath. The measurement of rear ankle swelling demonstrated significant results on both chromosomes 4 and 5. The QTL for rear ankle swelling was significantly detected in genomic intervals defined by (D4 Mit171 and D4 Mit172) and (D4 Mit237 and D4 Mit26). The

The Journal of Immunology

951 Histopathological scoring of hematoxylin- and eosin-stained sections from rear ankle joints of F2 mice demonstrated linkage to chromosomes 5 and 11. The genomic region defined by the interval (D5 Mit308 and D5 Mit91) was significantly associated (LOD 3.68) with this quantitative trait (Table II). The thickness of tendon sheaths was also associated (LOD 4.07) with this interval (D5 Mit308 to D5 Mit91) on chromosome 5 (Table II). This locus was designated Bb3, with alleles from C3H/HeNCr mice having a dominant effect on arthritis severity. Histopathological scoring also demonstrated significant association with the interval (D11 Mit350 and D11 Mit120) of chromosome 11. The region demonstrating greatest linkage (LOD 3.91) was termed Bb4. Alleles of Bb4 from the C3H/HeNCr mouse contribute to severe arthritis and are recessive. The slides were scored independently by two individuals, and in both cases linkages to these regions of chromosomes 5 and 11 were obtained. A summary of the linkages to arthritis severity are included in Tables II and III. Quantitative analysis of humoral response to B. burgdorferi in F2 intercross mice

FIGURE 1. Range of arthritis severity in F2 intercross mice. A, Rear ankle joint measurements were made on parental and F2 mice 4 wk following infection with B. burgdorferi, as described in Materials and Methods. The value for the most severely swollen ankle from each animal is shown. Open circles represent values for individual mice, with each parental group comprised of five male and five female mice. The F2 intercross generation consisted of 150 male and female mice. The solid bars indicate the mean ankle measurement for each group. Ankle measurements for mock-infected mice were: C3H/HeNCr, 2.99 6 0.02 mM; C57BL/6NCr, 2.86 6 0.1 mM; and B6C3F1, 2.95 6 0.03 mM (not shown). B, Measurement of the tendon sheath of the cranial tibial muscle was made when present in hematoxylin- and eosin-stained slides of rear ankle joints, as described in Materials and Methods. This structure was visible in ;70% of specimens from parental and F2 mice. Symbols are as in A. Tendon sheath thickness for mock-infected mice were: C3H/HeNCr mice, 24.6 6 9.95 mm; C57BL/6NCr, 20.4 6 7.79 mm; and B6C3F1, 19.9 6 6.76 mm (not shown).

LOD scores for these regions were greatest at 4.22 and 4.79, respectively, and were deemed significant by comparison to 95% experimentwise threshold values estimated for these data (23). Because of the limited capability of interval mapping to detect multiple traits on a chromosome, composite interval mapping was used on chromosome 4 (20, 21). A single significant region between D4 Mit87 and D4 Mit145 was found (maximal LOD 5 4.70, 95% threshold 5 2.53), implying a single locus or region of related genes as being significant. The locus in this region was designated Bb1 for B. burgdorferi-associated arthritis phenotype 1 (Table II). Bb1 represents a C3H/HeNCr allele with a dominant influence on susceptibility to severe arthritis. An additional QTL associated with ankle swelling was detected (LOD 4.07) on chromosome 5 in the interval (D5 Mit24 and D5 Mit98) (Table II). This locus was designated Bb2, and was attributed to C3H/HeNCr alleles with a recessive influence on susceptibility to severe arthritis.

As described above, quantitative differences were found between the humoral immune response of C3H/HeNCr and C57BL/6NCr mice to B. burgdorferi. A locus regulating the elevation in total circulating IgM and in B. burgdorferi-specific IgM was identified on chromosome 6 between markers D6 Mit105 and D6 Mit15. This locus was termed Bb5, and alleles from C57BL/6NCr mice contribute in a recessive manner to lower levels of IgM production (Tables III and IV). A second locus that influences the levels of total circulating IgM and IgG was identified on chromosome 12 between D12 Mit121 and D12 Mit133. This locus was termed Bb6, and alleles from C3H/HeNCr mice contributed to elevated levels of IgG and IgM in a recessive manner. Two additional loci were mapped that influenced B. burgdorferi-specific IgM. Bb7 was identified on chromosome 11 between D11 Mit82 and D11 Mit235, and its effect was identified as a dominant influence leading to lower Ig levels by alleles from C57BL/6NCr mice. Bb8 was identified on chromosome 17 between markers D17 Mit175 and D17 Mit215, and mice heterozygous for Bb8 had the lowest levels of specific IgM. A second locus was also identified that regulated the level of total IgG. This locus, termed Bb9, mapped between D9 Mit48 and D9 Mit73 on chromosome 9. The presence of C57BL/ 6NCr alleles of Bb9 had a dominant effect resulting in reduced levels of total IgG, as summarized in Tables III and IV. Correlation between disease severity and humoral response to B. burgdorferi The outer surface lipoproteins of B. burgdorferi are potent B cell mitogens, suggesting that the degree of humoral response in infected mice could reflect an important aspect of infection influencing disease severity (26). Analyses were performed to determine whether levels of total Ig or B. burgdorferi-specific Ig were correlated with arthritis severity in the F2 population. Statistically significant differences in Ig levels were found in the parental strains, and the entire range of responses were identified in the F2 population. However, there was no significant correlation between arthritis severity (either ankle swelling or histopathological score) and specific or total Ig levels (r2 , 0.18 for all comparisons). Genetic regulation of levels of B. burgdorferi in hearts from F2 mice Assessment of levels of B. burgdorferi DNA in the hearts of F2 mice revealed suggestive linkage to chromosome 4. This linkage was below the threshold of statistical significance and will require further studies for confirmation. It is interesting to note that the

952


Table II. Linkage of QTL controlling arthritis severity in response to B. burgdorferi infection Origin of Susceptibility Alleleg

Locusa

Marker

MGDb

uc

LODd,e

% Variancef

Bb1

D4Mit87 D4Mit26 D4Mit145 D5Mit24 D5Mit431 D5Mit214 D5Mit98 D5Mit308 D5Mit312 D5Mit91 D11Mit350 D11Mit29 D11Mit120 D5Mit308 D5Mit312 D5Mit91

31.4 42.5 44.5 60.0 55.8 70.0 80.0 44.0 50.0 53.0 34.0 40.0 47.0 44.0 50.0 53.0

2 0.121 0.000 2 0.069 0.034 0.088 2 0.042 0.017 2 0.030 0.102 2 0.042 0.017

2 4.03 4.70 2 4.07 4.02 3.95 2 3.59 3.68 2 3.51 3.91 2 4.07 4.04

16.2

C, dominant

12.5

C, recessive

11.0

C, dominant

13.4

C, recessive

16.4

C, dominant

Ankle Swelling

Bb2

Histopathology

Bb3 Bb4

Tendon sheath thickness

Bb3

a

Loci modifying arthritis associated with B. burgdorferi infection have been given the designation Bb1–Bb4. cM assignment for Massachusetts Institute of Technology markers from the Mouse Genome Database. Recombination fractions. d Maximum LOD scores were generated by interval mapping using MAPMAKER/QTL under assumptions of a free genetic model and may fall between actual markers. Composite interval mapping was used to generate LOD scores for the Bb1 locus. e The 95% threshold LOD scores were 3.74, 3.49, and 3.60 for ankle swelling, histopathology, and tendon sheath thickness, respectively. These represent experimentwise permutation-derived threshold values for significance of linkage for QTL (23). f Percent of variance attributable to loci-modifying arthritis associated with B. burgdorferi infection was calculated using MAPMAKER/QTL under the assumptions of a free genetic model. g Indicates which parental allele of loci-modifying arthritis associated with B. burgdorferi infection the phenotype can be attributed. The significance of the differences between genotypic values for each QTL was determined using Tukey’s multiple comparison test (a 5 0.05). C, C3H/HeN. b c

linkage to chromosome 4 is in the region of Bb1, which regulates ankle swelling. However, strong correlation was not observed between spirochete levels in the hearts of F2 mice and ankle swelling (r2 5 0.1). Assessment of interaction of loci associated with infection and disease in F2 mice Because the regulation of arthritis severity and humoral responses were found to be multigenic traits, possible interactions between significant loci for each trait were determined using the most significant marker associated with each QTL (27). Two interactions were suggested: one for ankle swelling measurements between Bb1 and Bb2 on chromosomes 4 and 5; and the second for levels of total IgG between Bb9 and Bb6 on chromosomes 9 and 12. No statistically significant interactions were noted for significant markers for histopathology score, total IgM, specific IgM, or specific IgG.

The significant interaction found for the ankle swelling phenotype was between the markers D4 Mit87 (Bb1) and D5 Mit431 (Bb2), (F 5 6.48, p , 0.001, R2 5 28.35). The interaction between Bb1(D4 Mit87) and Bb2 (D5 Mit431) for ankle swelling was further studied by plotting the genotypes at these two loci (Fig. 2A). F2 mice that were homozygous for C3H/HeNCr alleles at Bb1 (D4 Mit87) or Bb2 (D5 Mit431) had severe ankle swelling, regardless of the genotype at the second locus (Bb2 or Bb1, respectively). Additionally, F2 mice homozygous for C57BL/6NCr alleles at either Bb1 or Bb2 were not protected from severe ankle swelling conferred by the presence of a single C3H/HeNCr allele at the second locus, (Bb2 or Bb1, respectively). F2 mice heterozygous at both loci have increased ankle swelling, supporting an interaction between C3H/HeNCr alleles at the two loci. Significant interaction was also shown for total IgG between the markers D9 Mit48 (Bb9) and D12 Mit139 (Bb6) (F 5 6.80, p ,

Table III. Summary of loci controlling humoral immune response and arthritis severity after B. burgdorferi infection Locus

c

Markers that Flank Interval

D4Mit87–D4Mit145 D5Mit24–D5Mit98

Ankle swelling Ankle swelling

4 @ 31–45 5 @ 60–80

Bb3

5 @ 44–53

D5Mit308–D5Mit91

Bb7 Bb8

Histopathology, tendon sheath thickness Histopathology Total IgM Specific IgM Total IgM Total IgG Specific IgM Specific IgM

11 @ 34–47 6 @ 45–74 6 @ 67–74 12 @ 46–56 12 @ 52–56 11 @ 14–20 17 @ 17–23

D11Mit350–D11Mit120 D6Mit105–D6Mit15 D6Mit59–D6Mit15 D12Mit121–D12Mit133 D12Mit139–D12Mit133 D11Mit82–D11Mit235 D17Mit175–D17Mit215

Bb9

Total IgG

9 @ 34–41

D9Mit48–D9Mit73

Bb6

b

Chromosome @ cMb

Bb1 Bb2

Bb4 Bb5

a

Phenotypesa

Phenotypes associated with humoral response or arthritis after B. burgdorferi infection. Map distances on each chromosome as reported by Mouse Genome Database (www.informatics.jax.org/mgd.html). Human syntenic regions taken from National Center for Biotechnology Information (www.ncbi.nlm.nih.gov/Homology/).

Human Syntenic Regionsc

9q32–q33, 9p21–p23 22q11, 12q22–24.3, 7q11.23, 7q21.3–q22.1 4q11–q21.1 17p11–p13, 17q11–q12 3p24–p26, 3q21–q25, 10q11.2, 12p11–13 14q24.3–q32.3, 14q32.1–q32.3 16p13.3, 5q31.1–q35 19q13.3–13.4, 16p12–13.4, 6p13.2–p21.2 15q21–q25


953

Table IV. Linkage of QTL controlling humoral immune response to B. burgdorferi infection

Total IgM

Marker

MGDb

uc

LODd,e

% Variancef

Bb5

D6Mit105 D6Mit108 D6Mit194 D8Mit338 D6Mit59 D6Mit200 D6Mit15 D12Mit121 D12Mit139 D12Mit133 D6Mit59 D6Mit200 D6Mit15 D11Mit82 D11Mit51 D11Mit110 D11Mit235 D17Mit175 D17Mit230 D17Mit28 D17Mit33 D17Mit232 D17Mit176 D17Mit215 D9Mit48 D9Mit73 D12Mit139 D12Mit133

45.5 48.2 62.5 62.5 67.0 71.0 74.0 46.0 52.0 56.0 67.0 71.0 74.0 14.0 18.0 19.0 20.0 17.7 18.2 7.7 9.8 20.9 22.5 23.3 34.0 41.0 52.0 56.0

2 0.013 0.072 0.015 0.035 0.068 0.010 2 0.144 0.013 2 0.068 0.010 2 0.058 0.008 0.016 2 0.000 0.007 0.003 0.003 0.003 0.007 2 0.080 2 0.013

2 3.82 5.95 6.52 6.51 6.28 5.79 2 4.55 4.71 2 4.92 4.96 2 3.79 3.41 3.70 2 3.68 3.68 3.52 3.78 3.75 3.76 2 3.62 2 3.61

18.9

B, recessive

13.9

C, recessive

14.2

B, dominant

11.8

B, dominant

10.9

Heterozygous effect

11.3

B, dominant

10.5

C, recessive

Bb6 Specific IgM

Bb5 Bb7

Bb8

Total IgG

Origin of Regulatory Alleleg

Locusa

Bb9 Bb6

a

Loci modifying humoral response to Borrelia burgdorferi infection have been given the designation Bb5–Bb9. cM assignment for Massachusetts Institute of Technology markers from the Mouse Genome Database. Recombination fractions. d Maximum LOD scores were generated by interval mapping using MAPMAKER/QTL under assumptions of a free genetic model and may fall between actual markers. e The 95% threshold LOD scores were 3.49, 3.67, and 3.47 for total IgM, specific IgM, and total IgG, respectively. These values represent experimentwise permutation-derived threshold values for significance of linkage for QTL (23). f Percent of variance attributable to loci modifying humoral response to B. burgdorferi infection was calculated using MAPMAKER/QTL under the assumptions of a free genetic model. g Indicates which parental allele of loci modifying the humoral response to B. burgdorferi infection is associated with the phenotype. Significance of the differences between genotypic values for each QTL was determined using Tukey’s multiple comparison test (a 5 0.05). B, C57BL/6; C, C3H/HeN. b c

0.001, R2 5 28.13%). The genotypes for the interacting loci regulating total IgG levels, Bb9 (D9 Mit48) and Bb6 (D12 Mit139), were also plotted and revealed a regulatory effect of the C57BL/ 6NCr allele at Bb9 (D9 Mit48) (Fig. 2B). Those F2 mice with one or two C57BL/6NCr alleles at Bb9 exhibited low levels of total IgG. However, an increase in total IgG was seen if two C3H/ HeNcr alleles were present at Bb6 along with one or two C3H/ HeNCr alleles at Bb9. Interestingly, animals which were heterozygous for the two alleles at Bb6 had a lower level of total IgG than did mice with any other allelic combination at Bb6 or Bb9. This suggests a more complex interaction exists between alleles at these two loci or that there may be an additional (three-way) interaction with a separate locus in the genome.

Discussion The mouse model of Lyme disease has provided valuable insights into many aspects of B. burgdorferi-induced arthritis development. The arthritis, which develops 3– 4 wk after infection of C3H mice, is histopathologically similar to arthritis in humans, with lesions characterized by tendonitis and synovitis, synovial hyperproliferation, edema, and inflammatory cell infiltrate (3, 8). In humans and mice, there is strong evidence that inflammation follows joint tissue invasion by the spirochete (4, 5). The mild arthritis that appears in infected C57BL/6NCr mice is distinguishable from arthritis in C3H/HeNCr mice both by histopathology and by ankle

swelling (Fig. 1 and Table I). The fact that some humans infected with B. burgdorferi develop little to no arthritis while others display severe symptoms suggests that common pathways may regulate the severity and progression of Lyme disease in mice and humans. In this study, arthritis severity was determined 4 wk following infection with B. burgdorferi in 150 mice from an F2 intercross of C3H/HeNCr and C57BL/6NCr mice. Three distinct chromosomal regions that regulate arthritis severity in the F2 population were identified by linkage analysis (summarized in Table III). QTL that regulate the degree of ankle swelling were mapped to chromosomes 4 (Bb1) and 5 (Bb2). Two other QTL that were identified from the histopathological scores or measurements for the ankle joints mapped to chromosomes 5 (Bb3) and 11 (Bb4). Bb2 and Bb3, which were identified by different measurements of arthritis severity, map to adjacent regions of chromosome 5. Interestingly, the identification of QTL unique for ankle swelling on chromosome 4 (Bb1) and unique for histopathological severity on chromosome 11 (Bb4) suggests that different processes are involved in these two arthritis-associated developments. The loci controlling ankle swelling could be expected to regulate edema and inflammatory cell infiltrate, whereas the loci regulating histopathologically determined severity may be required for synovial hyperproliferation and other abnormalities associated with the tendon sheath. This suggests that ankle edema and synovial hyperproliferation are not sequential steps, but rather they are independently

954

FIGURE 2. Interaction between loci regulating responses to infection. A, The interaction between two loci responsible for B. burgdorferi-induced ankle swelling, Bb1 and Bb2, is shown. Genotypes for Bb1, C57BL/6NCr (BL6) and C3H/HeNCr (C3H), are indicated on the x-axis, and ankle measurements are indicated on the y-axis. Symbols for genotypes for Bb2 represent homozygous C3H/HeNCr (open circles), homozygous C57BL/6NCr (open boxes), and heterozygous (filled boxes). B, The interaction between two loci regulating the level of total plasma IgG in B. burgdorferi-infected F2 mice, Bb9 and Bb6, is shown. Genotypes for Bb9, C57BL/6NCr (BL6) and C3H/HeNCr (C3H), are indicated on the x-axis, and IgG concentrations are indicated on the y-axis. Symbols for genotypes for Bb6 represent homozygous C3H/HeNCr (open circles), homozygous C57BL/6NCr (open boxes), and heterozygous (filled boxes).

controlled events contributing to the development of Lyme arthritis. Interestingly, Anguita and colleagues found that Ab neutralization of IL-12 reduces B. burgdorferi-induced histopathology with little effect on ankle swelling (28 –29), consistent with our findings that these two components of Lyme arthritis development are under distinct genetic control (Table II). One of the most puzzling aspects of human Lyme disease has been the presence of a robust immune response in some patients; while in other individuals, infection is associated with a poor serological response (6, 30 –32). Many of the immunogenic molecules expressed by B. burgdorferi during infection of the mammal have been identified as lipoproteins (33–36). Borrelia lipoproteins possess potent B cell mitogen, adjuvant, and proinflammatory properties (26, 37–39). These activities appear to function during infection as a generalized elevation in circulating IgM and IgG have been seen in some patients (30). Infected C3H/HeJ mice also have up to a 20-fold increase in total circulating Igs and have high levels of B. burgdorferi-specific Igs (10). Differences in the levels of total and specific Igs have been documented among inbred strains of mice, suggesting that the magnitude of the humoral response to B. burgdorferi is also subject to genetic control (8). In this study, QTL that regulate total levels of circulating IgM and

GENETIC REGULATION OF MURINE LYME DISEASE IgG in B. burgdorferi-infected mice were identified on chromosomes 6 (Bb5), 9 (Bb9), and 12 (Bb6). QTL that regulate B. burgdorferi-specific IgM were identified on chromosomes 6 (Bb5), 11 (Bb7), and 17 (Bb8). The QTL controlling B. burgdorferi-specific IgM responses on chromosome 17 (Bb8) was the only MHC linkage identified in this study. Although both more severe arthritis and higher levels of total and specific Igs are associated with infection of the C3H/HeNCr parental mice, they are not linked as QTL because these two sets of traits map to different chromosomes. It is noteworthy that the tick component of natural infection has been eliminated from these studies, and that any contribution the tick might make to the regulated expression of B. burgdorferi Ags or to the magnitude or type of immune response cannot be assessed (40). Thus, it is possible that additional QTL might be identified in intercross populations infected by ticks. Further, the presence of Abs to specific Ags was not assessed in this study, but could provide additional information about regulation of immune responses: others have identified MHC regulation of murine serological responses to B. burgdorferi Ags (41). The level of B. burgdorferi DNA present in hearts of F2 mice was determined as an indication of the generalized level of spirochete persistence in tissues of these animals (5, 11, 12). Although a suggestive linkage (90% experimentwise permutation threshold) to the region on chromosome 4 that regulates ankle swelling was detected, there was not a strong correlation between B. burgdorferi levels in hearts and ankle swelling. Also surprising was the lack of correlation between tissue spirochete levels and the magnitude of humoral responses in F2 mice. This suggests that the regulation of the magnitude of humoral responses to B. burgdorferi infection is not dependent on the quantity of spirochetes in tissues. Previous studies on the genetic contribution to arthritis severity have indicated that components of the inflammatory response may be very important in the outcome of disease. Arthritis develops in mice with the scid mutation, indicating that B and T lymphocytes are not required for disease development (42, 43). C3H mice with the scid mutation also develop more severe arthritis than do CD-17 mice (BALB/c congenic) with the scid mutation, further indicating that differences in arthritis severity are not dependent on B and T lymphocytes (42). This is further supported by experiments using both MHC congenic mice and mice genetically deficient in class II MHC molecules. These studies have indicated that class II-dependent responses, and, therefore, CD41 T cells, are not involved in arthritis development or resolution (10, 44, 45). It should be noted that in humans an association has been made between HLA-DR4 alleles and the development of treatment-resistant arthritis in a small subgroup of Lyme patients (46). This appears to be mediated by a distinct mechanism from the treatable, subacute arthritis observed in most patients and studied in mice. Our mapping studies support the prediction that host elements other than the acquired defenses regulate the severity of Lyme arthritis as QTL for arthritis severity were not found within loci encoding Igs, the TCRs, or H-2. The assignment of significant linkage to chromosomal regions also allows speculation about previously identified genes lying within the linkage interval that could be involved in the severity of arthritis development in the mouse (Table V). Several of these genes are directly involved in inflammation or inflammatory control, including: Gro1, the GRO1 oncogene that is involved in chemotaxis of neutrophils, macrophages, and lymphocytes; Pdgfra, the platelet-derived growth factor receptor; and Pafaha, platelet-activating factor acetylhydrolase. Other genes of interest are involved in the responses of PMNs or macrophages to microbial and inflammatory stimuli: Lps, the LPS response gene; Nos2, the inducible nitric oxide synthase gene;


955

Table V. Candidate genes for regulation of arthritis severity Chromosome

4 5

Gene Namea

Lps Gro1 Spp1

11

a

Ibsp Pdgfra Bmp3 Tnfip1 Pafaha Nos2

Description

Regulates response to bacterial LPS GRO1 oncogene; causes chemotaxis of neutrophils, macrophages, and lymphocytes in response to infection Secreted phosphoprotein 1, natural resistance to Rickettsia tsutsugamushi Bone sialoprotein II Platelet-derived growth factor receptor, a-polypeptide Bone morphogenetic protein 3 Tumor necrosis factor-induced protein 1 Platelet-activation factor acetylhydrolase, a subunit Inducible nitric oxide synthase

Genes identified within 3 cM of region meeting threshold of statistical significance.

Tnfip1, TNF-induced protein 1; and Spp1, secreted phosphoprotein 1, which determines innate resistance to Rickettsia tsutsugamushi. C3H mice expressing different alleles of the Lps gene (C3H/HeJ, lpsr and C3H/HeN, lpss) do not display any differences in response to B. burgdorferi infection, including arthritis severity and kinetics, spirochete dissemination and persistence, or humoral response (5, 9). Studies, in which the production of nitric oxide, the product of the enzyme encoded by Nos2, has been blocked in infected mice, have indicated this molecule is not involved in arthritis development or host defense in C3H/He and BALB/c mice (12). However, it is still possible that subtle variations in these genes could contribute to differences in disease severity on some backgrounds. Abnormal bone formation within the tendon sheath is frequently seen in severe Lyme arthritis, and two of the candidate genes are involved in bone development: Ibsp, bone sialoprotein II; and Bmp3, bone morphogenetic protein 3. A cluster of genes encoding important immune and inflammatory regulatory cytokines was identified in a region of chromosome 11 that was suggestive for QTL by both histopathology score and tendon sheath width, but did not reach the required threshold for statistical significance. This cluster includes the genes for IL3, IL-3; IL4, IL-4; IL5, IL-5; and IL13, IL-13. Cytokine neutralization studies in C3H/He, BALB/c, and C57BL/6 mice have indicated that the balance of pro- and anti-inflammatory cytokines can influence the development and severity of arthritis (28, 29); in three reports a protective role for IL-4 in murine Lyme disease was found (47– 49). This same genomic region has recently been identified to regulate development of Th cell responses (50). These findings strongly support the inclusion of the IL-4, IL-5, and IL-13 gene cluster within the candidate group. This study is the first attempt to address the complexity of the genetic regulation of arthritis development induced by B. burgdorferi. Multigenic control of arthritis development was demonstrated, with two loci responsible for joint swelling and two loci responsible for histopathological severity. Additional loci were identified that regulate humoral responses to B. burgdorferi. A more complex interaction between loci may also determine the severity of arthritis and the magnitude of the humoral response. These studies provide an objective estimation that four different genes regulate arthritis severity in mice. This finding will allow further analysis of the actual genes involved in arthritis development and a detailed assessment of the pathways involved, with implication for understanding disease development in humans. The fact that only a portion of humans not treated for early symptoms of Lyme disease go on to develop severe arthritis strongly supports a role for genetic regulation of disease severity in humans. The analysis of syntenic regions in the human genome (Table III) will

provide further opportunity for characterization of candidate genes involved in the regulation of Lyme arthritis in humans.

Acknowledgments We thank JoAnn Eurell for use of the imaging equipment, Carol Lichtensteiger for helpful discussions, Mark Leppert for carefully reading the manuscript, and Nicole Van Buskirk, Brent Smith, and Kathleen Seiler for tissue preparation.

References 1. Burgdorfer, W., A. G. Barbour, S. F. Hayes, J. L. Benach, E. Grunwaldt, and J. P. Davis. 1982. Lyme disease: a tick-borne spirochetosis? Science 216:1317. 2. Center for Disease Control. 1997. Lyme Disease–United States, 1996. Morbidity Mortality Weekly Reports, p. 531. 3. Nocton, J. J., and A. C. Steere. 1995. Lyme disease. Adv. Intern. 40:69. 4. Nocton, J. J., F. Dressler, B. J. Rutledge, P. N. Rys, D. H. Persing, and A. C. Steere. 1994. Detection of Borrelia burgdorferi DNA by polymerase chain reaction in synovial fluid from patients with Lyme arthritis. N. Engl. J. Med. 330:229. 5. Yang, L., J. H. Weis, E. Eichwald, C. P. Kolbert, D. H. Persing, and J. J. Weis. 1994. Heritable susceptibility to severe Borrelia burgdorferi-induced arthritis is dominant and is associated with persistence of large numbers of spirochetes in tissues. Infect. Immun. 62:492. 6. Barbour, A. G., and D. Fish. 1993. The biological and social phenomenon of Lyme disease. Science 260:1610. 7. Pennington, P. M., C. D. Allred, C. S. West, R. Alvarez, and A. G. Barbour. 1997. Arthritis severity and spirochete burden are determined by serotype in the Borrelia turicatae-mouse model of Lyme disease. Infect. Immun. 65:285. 8. Barthold, S. W., D. S. Beck, G. M. Hansen, G. A. Terwilliger, and K. D. Moody. 1990. Lyme borreliosis in selected strains and ages of laboratory mice. J. Infect. Dis. 162:133. 9. Barthold, S. W., D. H. Persing, A. L. Armstrong, and R. A. Peeples. 1991. Kinetics of Borrelia burgdorferi dissemination and evolution of disease after intradermal inoculation of mice. Am. J. Pathol. 139:263. 10. Yang, L., Y. Ma, R. Schoenfeld, M. Griffiths, E. Eichwald, B. Araneo, and J. J. Weis. 1992. Evidence for B-lymphocyte mitogen activity in Borrelia burgdorferi-infected mice. Infect. Immun. 60:3033. 11. Ma, Y., K. P. Seiler, E. J. Eichwald, J. H. Weis, C. Teuscher, and J. J. Weis. 1998. Distinct characteristics of resistance to Borrelia burgdorferi-induced arthritis in C57BL/6N mice. Infect. Immun. 66:161. 12. Seiler, K. P., Z. Vavrin, E. Eichwald, J. Hibbs, Jr., and J. J. Weis. 1995. Nitric oxide production during murine Lyme disease: lack of involvement in host resistance or pathology. Infect. Immun. 63:3886. 13. SAS Institute. 1989. In SAS/STAT User’s Guide, Version 6, Fourth Edition, Vol. 2. SAS Institute, Cary, NC. 14. Wittwer, C. T., M. G. Herrmann, A. A. Moss, and R. P. Rasmussen. 1997. Continuous fluorescence monitoring of rapid cycle DNA amplification. Biotechniques 22:130. 15. Love, J. M., A. M. Knight, M. A. McAleer, and J. A. Todd. 1990. Towards construction of a high resolution map of the mouse genome using PCR-analysed microsatellites. Nucleic Acids Res. 18:4123. 16. Dietrich, W., H. Katz, S. E. Lincoln, H. S. Shin, J. Friedman, N. C. Dracopoli, and E. S. Lander. 1992. A genetic map of the mouse suitable for typing intraspecific crosses. Genetics 131:423. 17. Todd, J. A., T. J. Aitman, R. J. Cornall, S. Ghosh, J. R. Hall, C. M. Hearne, A. M. Knight, J. M. Love, M. A. McAleer, and J. B. Prins. 1991. Genetic analysis of autoimmune type 1 diabetes mellitus in mice. Nature 351:542. 18. Lander, E. S., P. Green, J. Abrahamson, A. Barlow, M. J. Daly, S. E. Lincoln, and L. Newburg. 1987. MAPMAKER: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174.

956 19. Paterson, A. H., E. S. Lander, J. D. Hewitt, S. Peterson, S. E. Lincoln, and S. D. Tanksley. 1988. Resolution of quantitative traits into Mendelian factors by using a complete linkage map of restriction fragment length polymorphisms. Nature 335:721. 20. Zeng, Z. B. 1993. Theoretical basis for separation of multiple linked gene effects in mapping quantitative trait loci. Proc. Natl. Acad. Sci. USA 90:10972. 21. Zeng, Z. B. 1994. Precision mapping of quantitative trait loci. Genetics 136:1457. 22. Basten, C. J., B. S. Weir, and Z.-B. Zeng. 1996. A Reference Manual and Tutorial for QTL Mapping. Department of Statistics, North Carolina State University, Raleigh, NC. 23. Churchill, G. A., and R. W. Doerge. 1994. Empirical threshold values for quantitative trait mapping. Genetics 138:963. 24. Roper, R. J., R. W. Doerge, S. B. Call, K. S. Tung, W. F. Hickey, and C. Teuscher. 1998. Autoimmune orchitis, epididymitis, and vasitis are immunogenetically distinct lesions. Am. J. Pathol. 152:1337. 25. Butterfield, R. J., J. D. Sudweeks, E. P. Blankenhorn, R. Korngold, J. C. Marini, J. A. Todd, R. J. Roper, and C. Teuscher. 1998. New genetic loci that control susceptibility and clinical symptoms of experimental allergic encephalomyelitis in inbred mice. J. Immunol. 161:1860. 26. Ma, Y., and J. J. Weis. 1993. Borrelia burgdorferi outer surface lipoproteins OspA and OspB possess B-cell mitogenic and cytokine-stimulatory properties. Infect. Immun. 61:3843. 27. Paterson, A. H., S. Damon, J. D. Hewitt, D. Zamir, H. D. Rabinowitch, S. E. Lincoln, E. S. Lander, and S. D. Tanksley. 1991. Mendelian factors underlying quantitative traits in tomato: comparison across species, generations, and environments. Genetics 127:181. 28. Anguita, J., S. Samanta, S. W. Barthold, and E. Fikrig. 1997. Ablation of interleukin-12 exacerbates Lyme arthritis in SCID mice. Infect. Immun. 65:4334. 29. Anguita, J., D. H. Persing, M. Rincon, S. W. Barthold, and E. Fikrig. 1996. Effect of anti-interleukin 12 treatment on murine lyme borreliosis. J. Clin. Invest. 97: 1028. 30. Sigal, L. H., A. C. Steere, and J. M. Dwyer. 1988. In vivo and in vitro evidence of B cell hyperactivity during Lyme disease. J. Rheumatol. 15:648. 31. Sigal, L. H. 1997. Lyme disease: a review of aspects of its immunology and immunopathogenesis. Annu. Rev. Immunol. 15:63. 32. Sigal, L. H. 1989. Lyme Disease, 1988: immunologic manifestations and possible immunopathogenic mechanisms. Semin. Arthritis Rheum. 18:151. 33. Fikrig, E., S. W. Barthold, W. Sun, W. Feng, S. R. r. Telford, and R. A. Flavell. 1997. Borrelia burgdorferi P35 and P37 proteins, expressed in vivo, elicit protective immunity. Immunity 6:531. 34. Suk, K., S. Das, W. Sun, B. Jwang, S. W. Barthold, R. A. Flavell, and E. Fikrig. 1995. Borrelia burgdorferi genes selectively expressed in the infected host. Proc. Natl. Acad. Sci. USA 92:4269. 35. Akins, D. R., S. F. Porcella, T. G. Popova, D. Shevchenko, S. I. Baker, M. Li, M. V. Norgard, and J. D. Radolf. 1995. Evidence for in vivo but not in vitro expression of a Borrelia burgdorferi outer surface protein F (OspF) homologue. Mol. Microbiol. 18:507. 36. Ramamoorthy, R., L. Povinelli, and M. T. Philipp. 1996. Molecular characterization, genomic arrangement, and expression of bmpD, a new member of the


37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47. 48.

49. 50.

bmp class of genes encoding membrane proteins of Borrelia burgdorferi. Infect. Immun. 64:1259. Ma, Y., K. P. Seiler, K. F. Tai, L. Yang, M. Woods, and J. J. Weis. 1994. Outer surface lipoproteins of Borrelia burgdorferi stimulate nitric oxide production by the cytokine-inducible pathway. Infect. Immun. 62:3663. Norgard, M. V., L. L. Arndt, D. R. Akins, L. L. Curetty, D. A. Harrich, and J. D. Radolf. 1996. Activation of human monocytic cells by Treponema pallidum and Borrelia burgdorferi lipoproteins and synthetic lipopeptides proceeds via a pathway distinct from that of lipopolysaccharide but involves transcriptional activator NF-kB. Infect. Immun. 64:3845. Sellati, T. J., L. D. Abrescia, J. D. Radolf, and M. B. Furie. 1996. Outer surface lipoproteins of Borrelia burgdorferi activate vascular endothelium in vitro. Infect. Immun. 64:3180. Golde, W. T., and M. C. Dolan. 1995. Variation in antigenicity and infectivity of derivatives of Borrelia burgdorferi, strain B31, maintained in the natural, zoonotic cycle compared with maintenance in culture. Infect. Immun. 63:4795. Golde, W. T., T. R. Burkot, S. Sviat, M. G. Keen, L. W. Mayer, B. J. B. Johnson, and J. Piesman. 1993. The major histocompatibility complex-restricted response of recombinant inbred strains of mice to natural tick transmission of Borrelia burgdorferi. J. Exp. Med. 177:9. Barthold, S. W., C. L. Sidman, and A. L. Smith. 1992. Lyme borreliosis in genetically resistant and susceptible mice with severe combined immunodeficiency. Am. J. Trop. Med. Hyg. 47:605. Schaible, U. E., S. Gay, C. Museteanu, M. D. Kramer, G. Zimmer, K. Eichmann, U. Museteanu, and M. M. Simon. 1990. Lyme borreliosis in the severe combined immunodeficiency (scid) mouse manifests predominantly in the joints, heart, and liver. Am. J. Pathol. 137:811. Fikrig, E., S. W. Barthold, M. Chen, C. H. Chang, and R. A. Flavell. 1997. Protective antibodies develop, and murine Lyme arthritis regresses, in the absence of MHC class II and CD41 T cells. J. Immunol. 159:5682. Ruderman, E. M., J. S. Kerr, S. R. Telford III, A. Spielman, L. H. Glimcher, and E. M. Gravallese. 1995. Early murine Lyme carditis has a macrophage predominance and is independent of major histocompatibility complex class II-CD41 T cell interactions. J. Infect. Dis. 171:362. Kalish, R. A., J. M. Leong, and A. C. Steere. 1993. Association of treatmentresistant chronic Lyme arthritis with HLA-DR4 and antibody reactivity to OspA and OspB of Borrelia burgdorferi. Infect. Immun. 61:2774. Keane-Myers, A., and S. P. Nickell. 1995. Role of IL-4 and IFN-g in modulation of immunity to Borrelia burgdorferi in mice. J. Immunol. 155:2020. Keane-Myers, A., C. R. Maliszewski, F. D. Finkelman, and S. P. Nickell. 1996. Recombinant IL-4 treatment augments resistance to Borrelia burgdorferi infections in both normal susceptible and antibody-deficient susceptible mice. J. Immunol. 156:2488. Matyniak, J. E., and S. L. Reiner. 1995. T helper phenotype and genetic susceptibility in experimental Lyme disease. J. Exp. Med. 181:1251. Gorham, J. D., M. L. Guler, R. G. Steen, A. J. Mackey, M. J. Daly, K. Frederick, W. F. Dietrich, and K. M. Murphy. 1996. Genetic mapping of a murine locus controlling development of T helper 1/T helper 2 type responses. Proc. Natl. Acad. Sci. USA 93:12467.