Previous Article | Next Article ![]()
Journal of Virology, September 2005, p. 11517-11522, Vol. 79, No. 17
0022-538X/05/$08.00+0 doi:10.1128/JVI.79.17.11517-11522.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Anne U. Jackson,4
Gwen N. Hirsch,3,
and
Katherine R. Spindler1*
Department of Microbiology and Immunology,1 Department of Biostatistics, University of Michigan, Ann Arbor, Michigan 48109,4 Department of Anatomy and Neurobiology, University of Tennessee Health Science Center, Memphis, Tennessee 38163,2 Department of Genetics, University of Georgia, Athens, Georgia 306063
Received 6 May 2005/ Accepted 10 June 2005
|
|
|---|
|
|
|---|
Inbred strains of mice with different susceptibilities to MAV-1 infection have been identified previously (14, 33). SJL/J mice are highly susceptible to MAV-1 infection, with a 50% lethal dose (LD50) more than 4 log units lower than those for other strains of mice, including C3H/HeJ, BALB/cJ, 129/J, and C57BL/6J (33). In vivo studies with immunodeficient mice have demonstrated the importance of innate and adaptive immunity in response to MAV-1 infections. The presence of T cells contributes to immunopathology during the acute phase of infection and to clearance of virus in long-term infection (25). B cell-deficient mice are highly susceptible to MAV-1 infections and die very early, at 6 to 9 days postinfection (p.i.), of a disseminated infection (26). T cell-independent production of neutralizing antiviral immunoglobulin M is a critical factor for control of MAV-1 infection. Infection of primary cells of susceptible and resistant mice ex vivo gives equivalent yields of virus, and sublethal irradiation renders resistant mice susceptible (33). These results suggest that host immune response factors play a role in susceptibility to MAV-1.
We have compared the adaptive immune responses that have been characterized as being important for control of MAV-1 infection (25, 26), as well as NK cell function and induction of cytokines and chemokines by infection, in susceptible and resistant mice (A. Welton and K. Spindler, unpublished data). To date, we have not identified physiological differences between susceptible and resistant strains have been identified by using these approaches. In this report, we show that MAV-1 susceptibility is linked to mouse chromosome 15 (Chr 15) and unlinked to the H-2 major histocompatibility locus. We used a genomewide search strategy, quantitative trait locus (QTL) detection, to identify host factors associated with susceptibility to MAV-1. This method is advantageous because it does not require prior hypotheses of the pathogenic mechanism and can therefore reveal novel pathways. We measured the susceptibility phenotypes of 192 backcross progeny. The analysis identified a major QTL on distal Chr 15 with a highly significant logarithm of odds (LOD) score of 21.2 and a minor QTL on Chr 5 with a suggestive LOD score of 1.2, based on 5,000 permutations of the mapping analysis. A LOD score of 21 means that the odds are 1021:1 in favor of linkage. The Chr 15 locus alone accounts for 40% of the trait variance between SJL/J and BALB/cJ mice. A single major locus of this influence is readily amenable to positional cloning strategies.
Susceptibility to MAV-1.
In LD50 assays, SJL/J mice are susceptible to MAV-1 infections whereas mice of other inbred strains, including BALB/cJ, C3H/HeJ, and A/J, are resistant (33). SJL/J mice infected with 100 PFU of MAV-1 have high viral loads in the brain 8 days p.i. (
5 x 105 PFU/g), whereas resistant C3H/HeJ mice have low viral loads (<2 x 103 PFU/g). Mice of two additional strains determined to be resistant according to LD50s, A/J and BALB/cJ, also had low viral loads in the brain 8 days p.i. with a dose of 100 PFU (Fig. 1). Because LD50 assays cannot be used for mapping studies, in which susceptibilities of individual progeny mice need to be determined, we used brain viral loads as a quantitative measure of virus susceptibility in the studies reported here.
![]() View larger version (13K): [in a new window] |
FIG. 1. (A and B) Susceptibility of H-2 congenic mice to MAV-1. Mice of the indicated strains and H-2 haplotypes were infected i.p. with 100 PFU of MAV-1. Brains were collected at 8 days p.i., and homogenates were assayed for viral loads via plaque assay. The asterisk and dotted line indicate the detection confidence of the assay; values below this line correspond to fewer than 20 plaques/plate (33). Each symbol represents an individual mouse. Panels A and B show results from independent experiments. (C) Susceptibility of F1 mice to MAV-1. Brain titers were determined for the indicated parental and F1 strains. LD50s indicated below panel C for SJL/J and BALB/cJ mice were reported in reference 33 and determined similarly for (BALB/cJ x SJL/J)F1 mice (CSJLF1/J). Means of the log titers were compared using a two-tailed t test, assuming equal variance, and the P values are indicated.
|
To investigate the inheritance of susceptibility, we tested whether (BALB/cJ x SJL/J)F1 (CSJLF1/J; Jackson Laboratory) mice were susceptible to MAV-1 by both LD50 assay and titration of virus from infected brain homogenates by plaque assay. The LD50 of the F1 mice was 102.3 PFU, between that of mice of the BALB/cJ (>104.4) and SJL/J (100.32) parental strains (Fig. 1C, bottom). The plaque assay results shown at the top of Fig. 1C are similar to results of the LD50 assay. The virus load in the F1 mice was significantly higher than that in mice of the resistant BALB/cJ parent strain (P < 0.0001) and was lower than that of mice of the susceptible SJL/J parent strain (P = 0.007). These results demonstrate that susceptibility is semidominant and are consistent with its being a quantitative trait.
Antigen capture ELISA for detection of MAV-1 viral particles. Mapping susceptibility to a small genome interval requires analysis of a large number of mice. The size of the QTL interval identified is inversely proportional to the number of recombinant mice generated. The phenotypic assay of susceptibility was the rate-limiting step in our analysis. Therefore, we developed a rapid, reliable, and inexpensive phenotypic screen for viral load, a capture enzyme-linked immunosorbent assay (ELISA) (39), and compared it to our plaque assay method. Briefly, in the ELISA, microtiter plates were coated with rabbit polyclonal anti-MAV-1 antiserum (20), washed, and then blocked with 1% bovine serum albumin. They were then washed and incubated with MAV-1-infected mouse brain homogenates prepared as previously described (33). The plates were washed and then incubated with mouse polyclonal anti-MAV-1 antiserum. The wells were washed, and the mouse antiserum was detected with a horseradish peroxidase-conjugated anti-mouse immunoglobulin G followed by horseradish peroxidase detection and assayed in a microtiter plate reader. ELISA values (optical densities at 450 nm) were calculated by subtracting the mean value for mock-infected brain homogenates, which were considered as background.
The capture ELISA and plaque assay methods were used to analyze a small subset of female backcross mice and corresponding parental controls. Plaque assay results are shown in Fig. 2A and capture ELISA results for the same mice are shown in Fig. 2B. We compared the two methods by regression analysis (Fig. 2C). There was a high positive correlation between the results of the plaque assay and those of the capture ELISA (r2 = 0.80). A similar regression analysis was carried out with plaque assay and capture ELISA data on infected male backcross mouse brain homogenates, and the regression again showed an excellent correlation between results of the two assays (data not shown). Therefore, we analyzed subsequent viral loads using the capture ELISA.
![]() View larger version (12K): [in a new window] |
FIG. 2. Backcross mice (3 to 6 weeks old) were infected i.p. with 100 PFU of MAV-1, and brain homogenates were prepared from mice harvested 8 days p.i. Each symbol represents an individual mouse. Circles, BALB/cJ; triangles, backcross mice; squares, F1 mice. (A) The results of a single plaque assay determination on backcross females and controls are shown. A replicate assay was done on these females, and similar results were obtained (data not shown). Horizontal lines indicate mean values for controls. (B) The results of a capture ELISA on the same mice assayed for panel A are shown. Each mouse brain homogenate was assayed in triplicate, and each symbol represents the average of the triplicate measurements per homogenate. A replicate capture ELISA was done, and similar results were obtained (data not shown). The data were log transformed in order to make statistical comparisons between the plaque assay and capture ELISA methods. OD, optical density. (C) The plaque assay and capture ELISA data from panels A and B were compared in a regression analysis. The line represents the best-fit linear regression.
|
![]() View larger version (16K): [in a new window] |
FIG. 3. MAV-1 loads in infected brains of 192 backcross progeny, determined by capture ELISA and plotted in rank order. Each mouse brain homogenate was assayed in triplicate, and each bar represents the average of the triplicate measurements per homogenate. Bars falling below zero indicate homogenate ELISA values for infected mice that were less than uninfected control values. Error bars indicate standard deviation. OD450, optical density at 450 nm.
|
QTL analysis involves associating individual genotypes at locations across the genome with a quantitative phenotypic variable, in this case, viral load in the brain as measured by capture ELISA following infection. QTL analysis was performed with R/qtl software, a QTL mapping package for the statistical software R (5). Quantitative trait data for each mouse were entered as the average ELISA (optical density at 450 nm) value of triplicate measurements. These data were then used to determine linkages to genome location across all mouse chromosomes.
An association of susceptibility with markers on Chrs 5 and 15 was observed using the R/qtl imputation method and multilocus models on the data from 192 backcross mice genotyped with 65 markers. A one-way scan for QTLs using the imputation method with 256 imputations on a 5-cM grid (31) showed that viral load was determined by a single major effect QTL on Chr 15 and a small effect modifier on Chr 5 (Fig. 4A). We detected the locus of greatest effect on distal Chr 15 (Fig. 4B). The peak LOD of 21.2 corresponds to 45 cM on the 5-cM grid, between D15Mit270 and D15Mit70. The percentage of variance explained by the QTL on Chr 15 analyzed alone was 39.9% (F[1, 190] = 126.25, P = 8.6e23). The 2 LOD confidence interval (CI) for the Chr 15 QTL was estimated to be from 39 to 54 cM (based on the map position data in Fig. 4B) from the results of a chromosome-specific scan using an imputed 1-cM grid. This 2 LOD CI corresponds roughly to an interval from
65 to 78 Mb on the physical map, based on estimation and interpolation from the flanking marker coordinates. The locus on Chr 5 had an additive effect on the phenotype (data not shown). Multiple-QTL modeling analysis of the data using R/qtl (4) indicated that in a model consisting of the loci on Chrs 15 and 5, 41.7% of the trait variance was explained (F[2, 189] = 67.66, P = 6.9e23).
![]() View larger version (13K): [in a new window] |
FIG. 4. (A) One-way genomewide scan for QTLs using R/qtl with a 1-cM grid. The analysis is for 192 backcross animals and 65 markers. Tick marks just above the x axis represent the 65 SSLP markers used in the genome scan. (B) Interval map for Chr 15 as determined using R/qtl. Note that the map positions indicated are based on the recombination frequencies calculated from the backcross data, and the most proximal Chr 15 marker in our scan, D15Mit265, shown at 0 cM, maps to 13 Mb on the physical map. The physical locations of all three markers, based on the Ensembldatabase, release v30.33f (archive date, 22 March 2005; http://apr2005.archive.ensembl.org/Mus_musculus), are indicated at the top in parentheses. The 2 LOD CI for the Chr 15 QTL is from 39 to 54 cM on this interval map, corresponding to 65 to 78 Mb on the physical map. For both panels A and B, the solid, dashed, and dotted horizontal lines indicate the highly significant (P < 0.001), significant (P < 0.05), and suggestive (P < 0.63) empirical thresholds, respectively, determined by performing 5,000 permutations in the one-way scan.
|
Candidate genes for MAV-1 susceptibility.
The locus on proximal Chr 15 represents the most significant determinant of MAV-1 susceptibility in the backcross population, contributing 39.9% of the phenotypic variation. The 2 LOD CI from 65 to 78 Mb on Chr 15 includes at least 250 known genes, based on data from Swiss-Prot, TrEMBL, mRNA, and RefSeq databases (found at the University of CaliforniaSanta Cruz genome browser site, http://www.genome.ucsc.edu/). Several of the genes between 65 and 78 Mb are of particular interest due to involvement with the immune response. These include the Ly6 complex, consisting of
11 gene family members (75 to 76 Mb). The Ly6 superfamily gene members encode cell surface glycoproteins expressed on discrete hematopoietic populations and are involved in lymphocyte adhesion, T-cell immune responses, and the interferon response (reviewed in references 15 and 29). The heat shock factor 1 gene (Hsf1) maps to 77 Mb and encodes the major heat shock transcription factor essential for control of heat shock protein induction in response to environmental stimuli. Mice disrupted for Hsf1 expression have altered immune and inflammatory responses (18, 40, 41). The Src-like adapter protein 1 gene (Sla1) maps to 67 Mb and encodes an adapter protein that regulates T-cell receptor signaling. In the region from 65 to 78 Mb, there are other genes whose functions have not yet been demonstrated that have homologies to genes encoding proteins involved in the immune response, including tyrosine phosphatase and kinase, G-coupled protein receptors, and vacuolar protein-sorting proteins. Fine mapping of the Chr 15 interval using additional recombinant mice and more closely spaced markers will enable us to generate a list of candidate genes that can be tested for correlation with MAV-1 susceptibility. It is possible that fine mapping will reveal that the Chr 15 interval contains two or more QTLs contributing to MAV-1 susceptibility.
There have been a number of studies on mouse susceptibility to animal viruses (reviewed in reference 3), including retroviruses (1, 2, 17, 34), poxviruses (10), polyomaviruses (23, 37), herpesviruses (11), flaviviruses (24, 28), alphaviruses (35, 36), and rhabdoviruses (19). However, the specific genes involved in susceptibility have been identified for only several viruses (6, 13, 22, 24, 28, 36). There have been few studies on susceptibility to human viruses and none on susceptibility to human adenoviruses. Study of the tractable MAV-1 model system has identified one major effect QTL for susceptibility to MAV-1 infection. It is likely that identification of the specific genes involved will provide important insight into mechanisms of virus-host interaction in infectious disease.
Supplemental material for this article may be found at http://jvi.asm.org/. ![]()
Present address: Department of Horticulture, University of Georgia, Athens, Georgia. ![]()
Present address: Department of Animal and Dairy Science, University of Georgia, Athens, Georgia. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2010 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»