Genetic Control of Bordetella pertussis Infection: Identification of Susceptibility Loci Using Recombinant Congenic Strains of Mice
http://www.100md.com
感染与免疫杂志 2005年第2期
Laboratory for Vaccine-Preventable Diseases
Laboratory of Toxicology, Pathology, and Genetics
Computerization and Methodological Consultancy Unit, National Institute of Public Health and the Environment (RIVM), Bilthoven, The Netherlands
Department of Molecular and Cellular Biology, Roswell Park Cancer Institute, Buffalo, New York
ABSTRACT
Susceptibility to and severity of Bordetella pertussis infection in infants and children vary widely. The spectrum of clinical symptoms ranges from subclinical infection to mild disease, severe whooping cough, and death. The aims of this study were to examine genetic susceptibilities of mice to B. pertussis and to identify genetic loci in the mouse genome that are involved in susceptibility to B. pertussis infection. For this purpose we screened two sets of recombinant congenic strains (RCS) of mice, HcB and CcS, for differences in the numbers of bacteria in the lung 7 days after inoculation. For both CcS and in HcB mice, a wide range in numbers of bacteria in the lung was found, suggesting that the course of infection is under multigenic control. From both RCS sets of mice, we selected one strain to identify possible susceptibility loci in F2 hybrid mice. The degree of lung colonization 7 days postinoculation in these F2 mice was evaluated in relation to genetic markers by linkage analysis. We found three novel loci that are involved in the control of B. pertussis infection. One locus, designated B. pertussis susceptibility locus 1 (Bps-1), was identified on chromosome 12. The presence of the C57BL/10 genome on this locus instead of the C3H genome significantly decreased the number of B. pertussis bacteria in the lung. Bps-1 has a dominant-positive effect on the clearance of B. pertussis from the lung. The function of most genes in this region is unknown. Two other loci, Bps-2 and Bps-3, showed genetic interaction and are located on chromosomes 5 and 11. We aim to identify the gene(s) in these regions which modify susceptibility to B. pertussis.
INTRODUCTION
Bordetella pertussis isa gram-negative bacterium that causes the respiratory disease known as whooping cough or pertussis. Worldwide, this bacterial agent causes some 20 million to 40 million cases of pertussis and an estimated 300,000 deaths each year (36).
The bacterium enters the airways via aerosol droplets and attaches to the epithelium of the upper respiratory tract. After a 7-to-10-day incubation period, the first symptoms of the disease, which are similar to those of the common cold, may be observed. One to two weeks later, more serious symptoms can occur, such as the typical "whooping cough" (24). In spite of worldwide vaccination since the 1950s, the incidence of pertussis is increasing again (4). Several causes have been suggested for the resurgence of pertussis, including improved diagnostics and surveillance, waning immunity, and the emergence of escape variants (26, 27).
The clinical course of B. pertussis infection varies widely. Knowledge about host genetic and immunological factors that influence susceptibility and severity of the infection may lead to the identification of new approaches for prevention or treatment of infectious diseases (14). Yet knowledge about human genetic factors that influence B. pertussis infection is still very limited. A number of studies provide clues for the role of host genes in susceptibility to B. pertussis infection. The genetic makeup of mouse strains affected the immune response to B. pertussis (2, 21, 25). This was also confirmed in studies using the respiratory B. pertussis infection model with knockout mice, where numerous genes, such as those for CR3, CD32, CD32, FcR, gamma interferon, interleukin 4, and immunoglobulin, have been knocked out to establish their involvement in the pathogenesis of B. pertussis infection (12, 25). Recently a mutation in Toll-like receptor 4 (Tlr-4) was identified as a major factor that influences the course of B. pertussis infection in mice (13, 22).
In human cell lines, differences in expression between B. pertussis-treated cells and untreated cells have been found for a number of genes. Upregulated genes encoded cytokines, chemokines, antiapoptotic factors, and nuclear factor of B (NF-B), whereas downregulated genes encoded DNA-binding proteins and cellular adhesin molecules (3, 30).
In general, to identify unknown genes involved in the course of complex diseases, quantitative trait locus (QTL) mapping studies with humans or animals have been used (7). A QTL is a polymorphic locus which contains alleles that differentially affect the expression of a continuously distributed phenotypic trait. QTL mapping is a phenotype-driven approach to identifying genes affecting a phenotype. As such, it permits the discovery of new genes and contrasts with gene-driven approaches, such as use of knockout mice, which allow the study of genes with known function (5, 9).
There are several ways to map a QTL. One approach, which reduces the genetic complexity of the mouse genome by 90%, uses so called recombinant congenic strains (RCS) of mice (8). This approach also enables the identification of possible low-penetrance genes and their interactions (7, 10). RCS are derived from two different inbred strains, the so-called background and donor strains. After two backcrosses and intercrossing, a set of RCS is created, with each strain containing 12.5% of the donor genome differently distributed across the background genome (8). The approximate distribution of these chromosomal regions of the donor strain is called strain distribution pattern (SDP). A more detailed description, including an example of such an SDP, was provided by P. Demant and colleagues (11, 31).
The aims of this study were to examine first whether RCS of mice show genetic differences in susceptibility to B. pertussis and second whether we could identify one or more genetic loci responsible for such differences. We used the number of bacteria in the lung 1 week after inoculation to define the phenotype and microsatellite markers to define the genotype. Using this approach, we identified a locus on chromosome 12, designated B. pertussis susceptibility locus 1 (Bps-1), and two interacting loci on chromosome 5 and 11, designated Bps-2 and -3, which influence the number of bacteria in the lung 1 week after inoculation.
MATERIALS AND METHODS
Experimental design. We examined the course of B. pertussis infection in 12 different CcS/Dem strains and 21 HcB/Dem strains. Approximately 10 mice of each strain, i.e., 145 CcS mice and 170 HcB mice, were tested to determine the number of bacteria in the lung 1 week postinoculation. Two F2 hybrid generations of mice generated from two different recombinant congenic strains, 211 (CcS4 x BALB/c)F2 mice and 230 (HcB28 x C3H)F2 mice, were subsequently phenotyped as described below.
Due to logistical limitations, we inoculated maximally 100 mice per day and combined the results. To test the reproducibility of the infection model, on several days the experimental groups contained BALB/c control mice that were inoculated in the same way. The original RCS mice were examined in four experiments, and the F2 hybrid mice were examined in eight experiments.
Animals. Only female mice were used for the infection experiments to increase the reproducibility. The RCS of mice were derived as described in previous publications (6, 10). HcB/Dem (referred to as HcB) strains are derived from the mouse strains C3H/DISnA (referred to as C3H) as background and C57Black/10ScSnA (referred to as C57BL/10) as donor. The CcS/Dem (referred to as CcS) strains are derived form the mouse strains BALB/cHeJ (referred to as BALB/c) as background and STS/A (referred to as STS) as donor as described previously (6, 10).
Two hundred thirty-two Hcb28 F2 hybrid mice were generated by crossing HcB28 to C3H and subsequently intercrossing their F1 progeny. Similarly, 211 Ccs4 F2 hybrid mice were generated by crossing CcS4 and BALB/c mice and intercrossing their F1 progeny. All mice were acclimatized at our animal testing facility for at least 1 week after transport before the start of the experiments. Mice received a standard laboratory chow (SRM-A; Hope Farms, Woerden, The Netherlands) and tap water ad libitum. All animal experiments were approved by the Institute's Animal Ethics Committee.
Infection experiments. In this study, we used the number of viable B. pertussis bacteria in the lung 1 week after inoculation to define the phenotype. This infection protocol was described previously (15, 35). Briefly, female mice were intranasally inoculated with 2 x 107 CFU of B. pertussis strain B213 after being anesthetized with diethyl ether or enflurane. Seven days after inoculation, mice were sacrificed and the lungs were collected in Verwey medium (33). The lungs were homogenized in Verwey medium and diluted 10 and 1,000 times. The numbers of CFU in these dilutions were determined by plating on Bordet Genou agar supplemented with 15% sheep blood and 30 μg of streptomycin/ml. Plates were incubated for 3 days at 35°C.
Genotyping. Microsatellite markers that have been used to construct the SDPs were also applied to define the genotypes of the F2 hybrid mice. A selection of these SDPs is presented by Jackson Laboratories (28). A more detailed SDP for the HcB series of mice was recently obtained (P. Demant, unpublished work). A schematic representation of SDPs of the two RCS of mice used in this study is presented in Fig. 1. In an F2 generation constructed of RCS, only markers present in regions in which the parental strain contains donor genome are informative.
Genomic DNA was isolated from mouse tails using the DNeasy tissue kit (QIAGEN). Strain CcS4 carries the genetic material of STS origin on eight segments on seven chromosomes as described previously (11, 31). For genotyping, we selected 15 microsatellite markers in these donor regions, D5Mit179, D6Mit109, D11Mit151, D11Mit51, D11Mit139, D11Mit28, D11Mit36, D11Mit122, D11Mit61, D11Mit49, D12Mit37, D15Mit121, D15Mit1, D15Mit3, and D15Mit37. Strain HcB28 also carries the genetic material of C57BL/10 origin on eight segments on seven chromosomes (11, 31). We genotyped these segments using 13 microsatellite markers: D7Mit294, D7Mit350, D7Mit330, D8Rivm46, D9Mit260, D9Mit182, D9Mit82, D11Rivm263, D12Mit167, D12Mit263, D15Mit68, D15Mit107, D17Mit64. Six additional flanking markers were selected around a region of special interest: three microsatellite markers, D12Mit53, D12Mit133, and D12Rivm144, and three single-nucleotide polymorphism markers, S12Rivm101, S12Rivm102, and S12Rivm104. The sequences of all primers except the RIVM markers (Table 1) were obtained from the mouse genome database of the Massachusetts Institute of Technology (MIT) (23).
DNA was amplified in a 10-μl PCR reaction with 5 μl of Hotstar 5x Mastermix (QIAGEN), 1.0 μM (each) primer, and approximately 2 mM tail DNA. Amplification was performed with a GeneAmp PCR System 9700 (Applied Biosystems), according to the following scheme: an initial 15 min at 95°C to denaturize the DNA and to activate the Hotstar Taq, followed by 30 cycles of 45 s at 94°C for denaturizing, 45 s at 57°C for annealing, 1 min at 70°C for elongation, and finally 10 min at 72°C for elongation. PCR products were stored at 4°C until further use. 6-Carboxyfluorescein-labeled microsatellite primer sets were used (Isogen Life science, Maarssen, The Netherlands), and fragment sizes were determined with a 3700 Capillary DNA sequencer-genotyper system (Applied Biosystems), using Genotyper software (Applied Biosystems). The three single-nucleotide-polymorphism markers were analyzed by restriction fragment length polymorphism assay on a 2.5% agarose gel (Table 1). Reaction conditions were used according to the manufacturer's instructions (New England Biolabs).
Statistical analysis. The statistical differences in phenotypes (number of CFU in the lung) between the different RCS of mice were examined by analysis of variance (ANOVA) (SPSS) and tested with the Student-Newman-Keuls test for multiple comparisons. To stabilize variances and to obtain approximately normal distributions, the numbers of CFU were square root (sqrt) transformed. For the F2 mice, linkage between numbers of CFU in the lung and the genotypes and their effect on the total phenotypic variation were calculated by an ANOVA with genotypes as a fixed factor and numbers of CFU as a dependent variable. To correct for the influence of experiment, experiment was included as a random factor. All single markers and all pairs of nonlinked markers were tested for linkage with a marker or interaction between markers. Interaction, or epistasis, is defined as the combined effect of two or more genes on a phenotype that could not have been predicted as the sum of their separate effects (10, 29). Linkage is presented as P value and log of the odds (LOD) score, which was calculated as negative log of significance (P value).
All markers and interactions were tested at the level of 0.05 (P < 0.05). P values were corrected for multiple comparisons using the formula (18, 19)
where μ(T) is the desired corrected P value, C is the number of chromosomes segregating in the cross (for Hcb28 and Ccs4, C = 7; Fig. 1), is the crossover rate (1.5 for a F2 hybrid generation), G is the genome length of the segregating part of the donor genome in morgans (12.5% of the whole mouse genome of 16 morgans is 2). T2 is the threshold, the F value from ANOVA for the observed P value is used as T2, and T is the observed uncorrected P value.
The estimated effect of a linked locus on the total of the observed phenotypic variation was presented as r2. To test the reproducibility of the infection protocol, we performed a t test with numbers of CFU of the control mice inoculated in different experiments.
To test a normal distribution of genotypes per experiment, we used a 2 test.
RESULTS
Reproducibility of the infection model. To test the reproducibility of the infection model, we performed several experiments with BALB/c mice. The number of bacteria in the lungs of the control mice, 1 week postinoculation, was similar regardless of the day the experiment was performed (data not shown). Because there was no significant difference between any of the control groups (P > 0.05), we combined the results of all experiments.
Differences between the congenic strains in B. pertussis infection of the lung. We tested 12 different CcS strains to determine the number of viable bacteria (CFU) in the lung 1 week postinoculation. Both for CcS mice and for HcB mice, a wide range in bacterial numbers was found. Results are shown in Fig. 2. Significant differences in lung colonization were observed between the strains of mice. Horizontal lines connect groups of mice that are mutually not significantly different according to the Student-Neuman-Keuls test. The number of CFU varied from 1.6 x 106 to 9.0 x 106 CFU per lung (Fig. 2). From these strains we selected CcS4 as the most representative strain for further breeding.
Similar experiments were performed with 21 different HcB strains (Fig. 2). The differences in lung colonization for Hcb mice were more pronounced compared to that for CcS mice and ranged from 1.4 x 105 to 6.8 x 106 CFU per lung. From these strains we selected HcB28 as the most resistant strain (HcB29 was not a suitable strain for breeding). We also observed pathological differences in the lungs, ranging from macroscopically healthy lungs to severe lung edema (data not shown).
Identification of new susceptibility loci. Based on the data presented above, we selected the CcS4 and HcB28 strains for subsequent F2 hybrid experiments to identify susceptibility loci. Both strains were used to generate an F2 hybrid generation of approximately 200 mice. F2 mice were inoculated with B. pertussis, and the number of CFU in the lung was determined after 7 days (Fig. 3). The numbers of CFU per lung for the F2 hybrid generation (CcS4 x BALB/c) ranged from 1.0 x 102 (detection limit) to 1.8 x 107, and the average for this group was 3.7 x 106. The numbers of CFU per lung for the F2 hybrid generation (Hcb28 x C3H) ranged from 1.0 x 102 to 3.4 x 106, and the average for this group was 5.0 x 105 CFU per lung. All mice were individually genotyped, and we compared genotypes with phenotypes to identify possible QTLs by linkage analysis.
In the F2 mice, generated by crossing HcB28 and C3H mice, we found linkage between the number of CFU and a region on chromosome 12. In Fig. 4, the LOD diagram is shown for the telomere region of this chromosome. We found maximum linkage between the number of CFU and the locus defined by S12Rivm102 (102.4 Mb) with a LOD score of 4.6. This LOD score and the corresponding P value (P = 0.000025) are statistically significant, also after correction for multiple comparisons as described in the statistical section (P = 0.0019). However, this result does not withstand correction for experiment as a random factor (P = 0.36). Although the experimental conditions are highly reproducible and the distribution of genotypes per experiment appeared to be normal according to the 2 test (P = 0.489), the day of the experiment has influence on the established linkage. We therefore consider these results "suggestive linkage" instead of significant linkage, and they warrant further investigation. Almost 10 percent (9.7%) of the variation in the number of bacteria in the experiments could be ascribed to the found locus (r2 = 0.097). We propose to call the locus B. pertussis susceptibility locus 1 (Bps-1). In Table 2, means of numbers of CFU for F2 mice are displayed per genotype. The mean differences are significantly different at the 0.05 level between C3H and heterozygotes and between C3H and C57BL/10, but there is no significant difference between heterozygotes and C57BL/10 mice. This indicates that the presence of C57BL/10 DNA has a dominant-positive effect on the clearance of B. pertussis from the lung.
In the F2 mice generated by crossing CcS4 with BALB/c, we did not find any linkage between individual markers and the number of CFU. However, we found linkage between the numbers of CFU and the interacting markers D5Mit179 (chromosome 5) and D11Mit122 (chromosome 11) with a LOD score of 2.92 (Fig. 5). This LOD score and the corresponding P value (P = 0.010) are statistically significant, but not after correction for multiple comparisons as described in the statistical section (P = 0.28). However, when the experiment is included as a random factor, the significance increases (P = 0.089). We therefore consider these results suggestive linkage, instead of significant linkage, and they warrant further investigation.
Seven percent (7.0%) of the variation in the number of bacteria in the experiments could be ascribed to these interacting loci (r2 = 0.070). We propose to call these loci Bps-2 and Bps-3. In Table 3, means of numbers of CFU for F2 mice are displayed per interaction of D5Mit179 and D11Mit122 (the marker combination with the highest linkage) genotypes. When both markers carry the BALB/c allele instead of STS, the number of CFU in the lung is relatively high.
DISCUSSION
Pertussis is reemerging, despite high vaccination coverage (26). Further, pertussis morbidity and mortality is especially high for babies that are too young to be vaccinated (32). Treatment of pertussis with antibiotics is only effective if initiated early during infection. Thus, novel therapeutic or preventive treatments of pertussis are required.
Elucidation of genetic differences in susceptibility to complex diseases can result in improved knowledge of pathways involved in the pathogenesis of these diseases. Identification of genes which affect susceptibility to infectious diseases is important in unraveling pathogen-host interactions. Furthermore, identification of such genes may lead to novel therapeutics. Novel therapeutics are especially important for infectious diseases which are difficult to control by vaccination, such as tuberculosis, pertussis, and AIDS, and/or by antibiotics, such as diseases caused by multiresistant bacteria.
To identify genes that affect the course of B. pertussis infection, we tested two sets of recombinant congenic strains, the STS and HcB sets. This approach has been applied successfully before in identifying Sst1 as one of the loci in mice affecting susceptibility to Mycobacterium tuberculosis (16, 17). The same approach was also successful in identifying a series of loci (Lmr) each associated with a different combination of pathological and immunological reactions to Leishmania major (1, 20, 34).
In this study we studied the effect of genetic differences in the mouse on lung colonization by B. pertussis. Recently a mutation in Toll-like receptor 4 (Tlr-4) was identified as a major factor that influences the course of B. pertussis infection in mice (13, 22). The gene coding for Tlr4 is positioned on mouse chromosome 4 (33 centimorgans [cM]). The strains of mice we applied here do not have documented mutations in Tlr4, and this gene therefore did not influence the variation observed in our experiments.
We tested two sets of RCS of mice and selected one strain from each set for generation of F2 hybrid generations to identify susceptibility loci. For both sets of RCS, a wide range was observed in the numbers of CFU recovered from the lungs. The phenotypic variation in these strains of mice, which differ in only 12.5% of their genome, indicates that genetic factors do influence B. pertussis pathogenesis. In addition, the range of variation between these strains suggests that the course of pertussis is controlled by multiple genes and that the genetics of B. pertussis infection is complex and controlled by multiple QTLs.
To identify loci potentially responsible for the observed differences in lung colonization, we generated F2 hybrid generations of approximately 200 mice. As shown in Fig. 3, there are large but similar variations in bacterial colonization in F2 mice generated from CcS4 and HcB28. Based on the SDPs, we genotyped all mice from both F2 hybrids and compared the genotypes with the degree of lung colonization. For a classical mouse F2 intercross generation (whole-genome segregating), the following critical values for establishing linkage are used: a LOD score below 2.8 argues against linkage, while a LOD score of 4.3 or higher indicates linkage (18, 19). For an RCS F2 hybrid generation, the critical values are lower because only 12.5% of the genome is segregating. Therefore, all P values are corrected for multiple comparisons as described in the statistical analysis section.
In the F2 mice generated from the HcB28 mice, we initially found linkage between lung colonization and one marker, D12Mit133 (chromosome 12), with a LOD score of 2.7. We subsequently tested several flanking markers and found a maximum linkage between number of CFU and S12Rivm102 with a LOD score of 4.6. This finding could explain almost 10% of the variation in CFU observed with the F2 mice generated from HcB28 (Fig. 4). This locus, which we propose to designate B. pertussis susceptibility locus 1 (Bps-1), is positioned near the telomere on chromosome 12 (approximately 102 Mb) and influences the number of bacteria present in the lung 1 week after inoculation. As is shown in Table 2, the presence of C57BL/10 DNA on this allele instead of C3H DNA has a dominant-positive effect on the clearance of bacteria from the lung.
For the F2 mice generated by crossing CcS4 with BALB/c, we did not find any linkage between individual markers and the number of CFU. However, we found linkage between the number of CFU and the interacting markers D5Mit179 (chromosome 5) and D11Mit122 (chromosome 11) with a LOD score of 2.92 (Fig. 5). Some (7.0%) of the variation in the number of bacteria in this experiment could be ascribed to these interacting loci (r2 = 0.070). We propose to call these loci Bps-2 and Bps-3. In Table 3, means of CFU for F2 mice are displayed per interaction of D5Mit179 and D11Mit122 (the marker combination with the highest linkage) genotypes. These linked regions are homologous to the human loci 7q21-q22 (D5Mit179) and 17p11-q12 (D11Mit122). These regions contain numerous genes, including genes encoding immunological functions, but the regions must be narrowed down before focusing on individual genes is possible.
The significance of the results described here may not withstand all statistical corrections, but nevertheless, the results are promising enough for future investigation. We consider the three novel loci to be in suggestive linkage. Because, according to the 2 test, the genotypes on the Bps-1 locus are normally distributed (P = 0.489), the relevance of the Bps-1 locus is underlined.
Because the wide range of variation found in the RCS mice suggests a multigenic control of B. pertussis infection, we did not expect to find only three loci influencing our phenotype. It is therefore likely that the combined effect of other loci on this phenotype is too subtle to detect in these experiments. Presumably, because B. pertussis infection is controlled by multiple genes, the extremes in phenotypes must be larger for identification of subtle genetic effects. Although all mice were raised and housed in the same facility under carefully controlled conditions, the contributions of other factors besides genetics cannot be excluded completely as complicating factors in the differences in bacterial load between mice. In addition, it would probably be necessary to increase the number of F2 animals to detect significant QTLs.
Based on the data presented in Fig. 4, we selected the region between 100 and 108 Mb around Bps-1 on chromosome 12 for future analysis. This region is similar to the human locus 14q32. It contains 84 different genes, including genes involved in immunity, such as tumor necrosis factor-associated factor 3 and microtubule affinity-regulating kinase 3. However, the majority of genes in this region have an unknown function (RIKEN cDNA). In future experiments we aim to identify the gene(s) that modifies the susceptibility to B. pertussis by a combination of positional cloning and expression analysis.
In conclusion, we screened two sets of RCS of mice for susceptibility to B. pertussis infection and found a wide range in bacterial numbers in the lung at 1 week postinoculation. This indicates multigenic control of the B. pertussis infection. We identified one locus located on chromosome 12, which we designated as Bps-1, and two interacting loci on chromosomes 5 and 11, designated Bps-2 and -3, which influence the number of bacteria in the lung 1 week after inoculation. The presence of C57BL/10 DNA in Bps-1 instead of C3H DNA has a dominant-positive effect on the clearance of bacteria from the lung.
ACKNOWLEDGMENTS
We thank Hans van Oirschot and Silke David for helping with the animal experiments. We also thank all biotechnicians of our animal facility, especially Henk Gielen, for useful help in performing the animal experiments.
REFERENCES
1. Badalova, J., M. Svobodova, H. Havelkova, V. Vladimirov, J. Vojtiskova, J. Engova, T. Pilcik, P. Volf, P. Demant, and M. Lipoldova. 2002. Separation and mapping of multiple genes that control IgE level in Leishmania major infected mice. Genes Immun. 3:187-195.
2. Barnard, A., B. P. Mahon, J. Watkins, K. Redhead, and K. H. Mills. 1996. Th1/Th2 cell dichotomy in acquired immunity to Bordetella pertussis: variables in the in vivo priming and in vitro cytokine detection techniques affect the classification of T-cell subsets as Th1, Th2 or Th0. Immunology 87:372-380.
3. Belcher, C. E., J. Drenkow, B. Kehoe, T. R. Gingeras, N. McNamara, H. Lemjabbar, C. Basbaum, and D. A. Relman. 2000. The transcriptional responses of respiratory epithelial cells to Bordetella pertussis reveal host defensive and pathogen counter-defensive strategies. Proc. Natl. Acad. Sci. USA 97:13847-13852.
4. Cherry, J. D. 1999. Epidemiological, clinical, and laboratory aspects of pertussis in adults. Clin. Infect. Dis. 28(Suppl.)2:S112—S117.
5. Churchill, G. 2004. QTL mapping overview. [Online.] http://www.jax.org/staff/churchill/labsite/research/qtl/. Accessed 13 April 2004.
6. Demant, P. 1992. Genetic resolution of susceptibility to cancer-new perspectives. Semin. Cancer Biol. 3:159-166.
7. Demant, P. 2003. Cancer susceptibility in the mouse: genetics, biology and implications for human cancer. Nat. Rev. Genet. 4:721-734.
8. Demant, P., and A. A. Hart. 1986. Recombinant congenic strains-a new tool for analyzing genetic traits determined by more than one gene. Immunogenetics 24:416-422.
9. Frankel, W. N. 1995. Taking stock of complex trait genetics in mice. Trends Genet. 11:471-477.
10. Frankel, W. N., and N. J. Schork. 1996. Who's afraid of epistasis Nat. Genet. 14:371-373.
11. Groot, P. C., C. J. Moen, W. Dietrich, J. P. Stoye, E. S. Lander, and P. Demant. 1992. The recombinant congenic strains for analysis of multigenic traits: genetic composition. FASEB J. 6:2826-2835.
12. Hellwig, S. M., H. F. van Oirschot, W. L. Hazenbos, A. B. van Spriel, F. R. Mooi, and J. G. van de Winkel. 2001. Targeting to Fc receptors, but not CR3 (CD11b/CD18), increases clearance of Bordetella pertussis. J. Infect. Dis. 183:871-879.
13. Higgins, S. C., E. C. Lavelle, C. McCann, B. Keogh, E. McNeela, p. Byrne, B. O'Gorman, A. Jarnicki, P. McGuirk, and K. H. Mills. 2003. Toll-like receptor 4-mediated innate IL-10 activates antigen-specific regulatory T cells and confers resistance to Bordetella pertussis by inhibiting inflammatory pathology. J. Immunol. 171:3119-3127.
14. Kimman, T. 2001. Genetics of infectious disease susceptibility. Kluwer Academic Publishers, Dordrecht, The Netherlands.
15. King, A. J., G. Berbers, H. F. van Oirschot, P. Hoogerhout, K. Knipping, and F. R. Mooi. 2001. Role of the polymorphic region 1 of the Bordetella pertussis protein pertactin in immunity. Microbiology 147:2885-2895.
16. Kramnik, I., P. Demant, and B. B. Bloom. 1998. Susceptibility to tuberculosis as a complex genetic trait: analysis using recombinant congenic strains of mice. Novartis Found. Symp. 217:120-131.
17. Kramnik, I., W. F. Dietrich, P. Demant, and B. R. Bloom. 2000. Genetic control of resistance to experimental infection with virulent Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 97:8560-8565.
18. Lander, E., and L. Kruglyak. 1995. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat. Genet. 11:241-247.
19. Lander, E. S., and N. J. Schork. 1994. Genetic dissection of complex traits. Science 265:2037-2048.
20. Lipoldova, M., M. Svobodova, M. Krulova, H. Havelkova, J. Badalova, E. Nohynkova, V. Holan, A. A. Hart, P. Volf, and P. Demant. 2000. Susceptibility to Leishmania major infection in mice: multiple loci and heterogeneity of immunopathological phenotypes. Genes Immun. 1:200-206.
21. Mahon, B. P., B. J. Sheahan, F. Griffin, G. Murphy, and K. H. Mills. 1997. Atypical disease after Bordetella pertussis respiratory infection of mice with targeted disruptions of interferon-gamma receptor or immunoglobulin mu chain genes. J. Exp. Med. 186:1843-1851.
22. Mann, P. B., M. J. Kennett, and E. T. Harvill. 2004. Toll-like receptor 4 is critical to innate host defense in a murine model of bordetellosis. J. Infect. Dis. 189:833-836.
23. Massachusetts Institute of Technology. 2004. Look up STSs by name. [Online.] http://www.broad.mit.edu/cgi-bin/mouse/sts_infodatabase=mouserelease. Accessed 28 April 2004.
24. Mattoo, S., A. K. Foreman-Wykert, P. A. Cotter, and J. F. Miller. 2001. Mechanisms of Bordetella pathogenesis. Front. Biosci. 6:E168-E186.
25. Mills, K. H., M. Ryan, E. Ryan, and B. P. Mahon. 1998. A murine model in which protection correlates with pertussis vaccine efficacy in children reveals complementary roles for humoral and cell-mediated immunity in protection against Bordetella pertussis. Infect. Immun. 66:594-602.
26. Mooi, F. R., I. H. Van Loo, and A. J. King. 2001. Adaptation of Bordetella pertussis to vaccination: a cause for its reemergence Emerg. Infect. Dis. 7:526-528.
27. Mooi, F. R., H. van Oirschot, K. Heuvelman, H. G. van der Heide, W. Gaastra, and R. J. Willems. 1998. Polymorphism in the Bordetella pertussis virulence factors P.69/pertactin and pertussis toxin in The Netherlands: temporal trends and evidence for vaccine-driven evolution. Infect. Immun. 66:670-675.
28. Mouse Genome Informatics. 2004. Recombinant congenic strain distribution patterns query form. [Online.] http://www.informatics.jax.org/searches/rcsetb_form.shtml. Accessed 28 April 2004.
29. Nadeau, J. H. 2001. Modifier genes in mice and humans. Nat. Rev. Genet. 2:165-174.
30. Nickerson, C. A. 2001. Altered gene expression in response to Bordetella pertussis infection. Trends Microbiol. 9:57-58.
31. Stassen, A. P., P. C. Groot, J. T. Eppig, and P. Demant. 1996. Genetic composition of the recombinant congenic strains. Mamm. Genome 7:55-58.
32. Tanaka, M., C. R. Vitek, F. B. Pascual, K. M. Bisgard, J. E. Tate, and T. V. Murphy. 2003. Trends in pertussis among infants in the United States, 1980-1999. JAMA 290:2968-2975.
33. Verwey, W. F., E. H. Thiele, D. N. Sage, and L. T. Suchardt. 1949. A simplified liquid culture medium for the growth of Haemophilus pertussis. J. Bacteriol. 58:127-134.
34. Vladimirov, V., J. Badalova, M. Svobodova, H. Havelkova, A. A. Hart, H. Blazkova, P. Demant, and M. Lipoldova. 2003. Different genetic control of cutaneous and visceral disease after Leishmania major infection in mice. Infect. Immun. 71:2041-2046.
35. Willems, R. J., J. Kamerbeek, C. A. Geuijen, J. Top, H. Gielen, W. Gaastra, and F. R. Mooi. 1998. The efficacy of a whole cell pertussis vaccine and fimbriae against Bordetella pertussis and Bordetella parapertussis infections in a respiratory mouse model. Vaccine 16:410-416.
36. World Health Organization. 2003. Pertussis vaccine. [Online.] http://www.who.int/vaccines/en/pertussis.shtml. Accessed 17 Oct. 2003.(H. A. Banus, H. J. van Kr)
Laboratory of Toxicology, Pathology, and Genetics
Computerization and Methodological Consultancy Unit, National Institute of Public Health and the Environment (RIVM), Bilthoven, The Netherlands
Department of Molecular and Cellular Biology, Roswell Park Cancer Institute, Buffalo, New York
ABSTRACT
Susceptibility to and severity of Bordetella pertussis infection in infants and children vary widely. The spectrum of clinical symptoms ranges from subclinical infection to mild disease, severe whooping cough, and death. The aims of this study were to examine genetic susceptibilities of mice to B. pertussis and to identify genetic loci in the mouse genome that are involved in susceptibility to B. pertussis infection. For this purpose we screened two sets of recombinant congenic strains (RCS) of mice, HcB and CcS, for differences in the numbers of bacteria in the lung 7 days after inoculation. For both CcS and in HcB mice, a wide range in numbers of bacteria in the lung was found, suggesting that the course of infection is under multigenic control. From both RCS sets of mice, we selected one strain to identify possible susceptibility loci in F2 hybrid mice. The degree of lung colonization 7 days postinoculation in these F2 mice was evaluated in relation to genetic markers by linkage analysis. We found three novel loci that are involved in the control of B. pertussis infection. One locus, designated B. pertussis susceptibility locus 1 (Bps-1), was identified on chromosome 12. The presence of the C57BL/10 genome on this locus instead of the C3H genome significantly decreased the number of B. pertussis bacteria in the lung. Bps-1 has a dominant-positive effect on the clearance of B. pertussis from the lung. The function of most genes in this region is unknown. Two other loci, Bps-2 and Bps-3, showed genetic interaction and are located on chromosomes 5 and 11. We aim to identify the gene(s) in these regions which modify susceptibility to B. pertussis.
INTRODUCTION
Bordetella pertussis isa gram-negative bacterium that causes the respiratory disease known as whooping cough or pertussis. Worldwide, this bacterial agent causes some 20 million to 40 million cases of pertussis and an estimated 300,000 deaths each year (36).
The bacterium enters the airways via aerosol droplets and attaches to the epithelium of the upper respiratory tract. After a 7-to-10-day incubation period, the first symptoms of the disease, which are similar to those of the common cold, may be observed. One to two weeks later, more serious symptoms can occur, such as the typical "whooping cough" (24). In spite of worldwide vaccination since the 1950s, the incidence of pertussis is increasing again (4). Several causes have been suggested for the resurgence of pertussis, including improved diagnostics and surveillance, waning immunity, and the emergence of escape variants (26, 27).
The clinical course of B. pertussis infection varies widely. Knowledge about host genetic and immunological factors that influence susceptibility and severity of the infection may lead to the identification of new approaches for prevention or treatment of infectious diseases (14). Yet knowledge about human genetic factors that influence B. pertussis infection is still very limited. A number of studies provide clues for the role of host genes in susceptibility to B. pertussis infection. The genetic makeup of mouse strains affected the immune response to B. pertussis (2, 21, 25). This was also confirmed in studies using the respiratory B. pertussis infection model with knockout mice, where numerous genes, such as those for CR3, CD32, CD32, FcR, gamma interferon, interleukin 4, and immunoglobulin, have been knocked out to establish their involvement in the pathogenesis of B. pertussis infection (12, 25). Recently a mutation in Toll-like receptor 4 (Tlr-4) was identified as a major factor that influences the course of B. pertussis infection in mice (13, 22).
In human cell lines, differences in expression between B. pertussis-treated cells and untreated cells have been found for a number of genes. Upregulated genes encoded cytokines, chemokines, antiapoptotic factors, and nuclear factor of B (NF-B), whereas downregulated genes encoded DNA-binding proteins and cellular adhesin molecules (3, 30).
In general, to identify unknown genes involved in the course of complex diseases, quantitative trait locus (QTL) mapping studies with humans or animals have been used (7). A QTL is a polymorphic locus which contains alleles that differentially affect the expression of a continuously distributed phenotypic trait. QTL mapping is a phenotype-driven approach to identifying genes affecting a phenotype. As such, it permits the discovery of new genes and contrasts with gene-driven approaches, such as use of knockout mice, which allow the study of genes with known function (5, 9).
There are several ways to map a QTL. One approach, which reduces the genetic complexity of the mouse genome by 90%, uses so called recombinant congenic strains (RCS) of mice (8). This approach also enables the identification of possible low-penetrance genes and their interactions (7, 10). RCS are derived from two different inbred strains, the so-called background and donor strains. After two backcrosses and intercrossing, a set of RCS is created, with each strain containing 12.5% of the donor genome differently distributed across the background genome (8). The approximate distribution of these chromosomal regions of the donor strain is called strain distribution pattern (SDP). A more detailed description, including an example of such an SDP, was provided by P. Demant and colleagues (11, 31).
The aims of this study were to examine first whether RCS of mice show genetic differences in susceptibility to B. pertussis and second whether we could identify one or more genetic loci responsible for such differences. We used the number of bacteria in the lung 1 week after inoculation to define the phenotype and microsatellite markers to define the genotype. Using this approach, we identified a locus on chromosome 12, designated B. pertussis susceptibility locus 1 (Bps-1), and two interacting loci on chromosome 5 and 11, designated Bps-2 and -3, which influence the number of bacteria in the lung 1 week after inoculation.
MATERIALS AND METHODS
Experimental design. We examined the course of B. pertussis infection in 12 different CcS/Dem strains and 21 HcB/Dem strains. Approximately 10 mice of each strain, i.e., 145 CcS mice and 170 HcB mice, were tested to determine the number of bacteria in the lung 1 week postinoculation. Two F2 hybrid generations of mice generated from two different recombinant congenic strains, 211 (CcS4 x BALB/c)F2 mice and 230 (HcB28 x C3H)F2 mice, were subsequently phenotyped as described below.
Due to logistical limitations, we inoculated maximally 100 mice per day and combined the results. To test the reproducibility of the infection model, on several days the experimental groups contained BALB/c control mice that were inoculated in the same way. The original RCS mice were examined in four experiments, and the F2 hybrid mice were examined in eight experiments.
Animals. Only female mice were used for the infection experiments to increase the reproducibility. The RCS of mice were derived as described in previous publications (6, 10). HcB/Dem (referred to as HcB) strains are derived from the mouse strains C3H/DISnA (referred to as C3H) as background and C57Black/10ScSnA (referred to as C57BL/10) as donor. The CcS/Dem (referred to as CcS) strains are derived form the mouse strains BALB/cHeJ (referred to as BALB/c) as background and STS/A (referred to as STS) as donor as described previously (6, 10).
Two hundred thirty-two Hcb28 F2 hybrid mice were generated by crossing HcB28 to C3H and subsequently intercrossing their F1 progeny. Similarly, 211 Ccs4 F2 hybrid mice were generated by crossing CcS4 and BALB/c mice and intercrossing their F1 progeny. All mice were acclimatized at our animal testing facility for at least 1 week after transport before the start of the experiments. Mice received a standard laboratory chow (SRM-A; Hope Farms, Woerden, The Netherlands) and tap water ad libitum. All animal experiments were approved by the Institute's Animal Ethics Committee.
Infection experiments. In this study, we used the number of viable B. pertussis bacteria in the lung 1 week after inoculation to define the phenotype. This infection protocol was described previously (15, 35). Briefly, female mice were intranasally inoculated with 2 x 107 CFU of B. pertussis strain B213 after being anesthetized with diethyl ether or enflurane. Seven days after inoculation, mice were sacrificed and the lungs were collected in Verwey medium (33). The lungs were homogenized in Verwey medium and diluted 10 and 1,000 times. The numbers of CFU in these dilutions were determined by plating on Bordet Genou agar supplemented with 15% sheep blood and 30 μg of streptomycin/ml. Plates were incubated for 3 days at 35°C.
Genotyping. Microsatellite markers that have been used to construct the SDPs were also applied to define the genotypes of the F2 hybrid mice. A selection of these SDPs is presented by Jackson Laboratories (28). A more detailed SDP for the HcB series of mice was recently obtained (P. Demant, unpublished work). A schematic representation of SDPs of the two RCS of mice used in this study is presented in Fig. 1. In an F2 generation constructed of RCS, only markers present in regions in which the parental strain contains donor genome are informative.
Genomic DNA was isolated from mouse tails using the DNeasy tissue kit (QIAGEN). Strain CcS4 carries the genetic material of STS origin on eight segments on seven chromosomes as described previously (11, 31). For genotyping, we selected 15 microsatellite markers in these donor regions, D5Mit179, D6Mit109, D11Mit151, D11Mit51, D11Mit139, D11Mit28, D11Mit36, D11Mit122, D11Mit61, D11Mit49, D12Mit37, D15Mit121, D15Mit1, D15Mit3, and D15Mit37. Strain HcB28 also carries the genetic material of C57BL/10 origin on eight segments on seven chromosomes (11, 31). We genotyped these segments using 13 microsatellite markers: D7Mit294, D7Mit350, D7Mit330, D8Rivm46, D9Mit260, D9Mit182, D9Mit82, D11Rivm263, D12Mit167, D12Mit263, D15Mit68, D15Mit107, D17Mit64. Six additional flanking markers were selected around a region of special interest: three microsatellite markers, D12Mit53, D12Mit133, and D12Rivm144, and three single-nucleotide polymorphism markers, S12Rivm101, S12Rivm102, and S12Rivm104. The sequences of all primers except the RIVM markers (Table 1) were obtained from the mouse genome database of the Massachusetts Institute of Technology (MIT) (23).
DNA was amplified in a 10-μl PCR reaction with 5 μl of Hotstar 5x Mastermix (QIAGEN), 1.0 μM (each) primer, and approximately 2 mM tail DNA. Amplification was performed with a GeneAmp PCR System 9700 (Applied Biosystems), according to the following scheme: an initial 15 min at 95°C to denaturize the DNA and to activate the Hotstar Taq, followed by 30 cycles of 45 s at 94°C for denaturizing, 45 s at 57°C for annealing, 1 min at 70°C for elongation, and finally 10 min at 72°C for elongation. PCR products were stored at 4°C until further use. 6-Carboxyfluorescein-labeled microsatellite primer sets were used (Isogen Life science, Maarssen, The Netherlands), and fragment sizes were determined with a 3700 Capillary DNA sequencer-genotyper system (Applied Biosystems), using Genotyper software (Applied Biosystems). The three single-nucleotide-polymorphism markers were analyzed by restriction fragment length polymorphism assay on a 2.5% agarose gel (Table 1). Reaction conditions were used according to the manufacturer's instructions (New England Biolabs).
Statistical analysis. The statistical differences in phenotypes (number of CFU in the lung) between the different RCS of mice were examined by analysis of variance (ANOVA) (SPSS) and tested with the Student-Newman-Keuls test for multiple comparisons. To stabilize variances and to obtain approximately normal distributions, the numbers of CFU were square root (sqrt) transformed. For the F2 mice, linkage between numbers of CFU in the lung and the genotypes and their effect on the total phenotypic variation were calculated by an ANOVA with genotypes as a fixed factor and numbers of CFU as a dependent variable. To correct for the influence of experiment, experiment was included as a random factor. All single markers and all pairs of nonlinked markers were tested for linkage with a marker or interaction between markers. Interaction, or epistasis, is defined as the combined effect of two or more genes on a phenotype that could not have been predicted as the sum of their separate effects (10, 29). Linkage is presented as P value and log of the odds (LOD) score, which was calculated as negative log of significance (P value).
All markers and interactions were tested at the level of 0.05 (P < 0.05). P values were corrected for multiple comparisons using the formula (18, 19)
where μ(T) is the desired corrected P value, C is the number of chromosomes segregating in the cross (for Hcb28 and Ccs4, C = 7; Fig. 1), is the crossover rate (1.5 for a F2 hybrid generation), G is the genome length of the segregating part of the donor genome in morgans (12.5% of the whole mouse genome of 16 morgans is 2). T2 is the threshold, the F value from ANOVA for the observed P value is used as T2, and T is the observed uncorrected P value.
The estimated effect of a linked locus on the total of the observed phenotypic variation was presented as r2. To test the reproducibility of the infection protocol, we performed a t test with numbers of CFU of the control mice inoculated in different experiments.
To test a normal distribution of genotypes per experiment, we used a 2 test.
RESULTS
Reproducibility of the infection model. To test the reproducibility of the infection model, we performed several experiments with BALB/c mice. The number of bacteria in the lungs of the control mice, 1 week postinoculation, was similar regardless of the day the experiment was performed (data not shown). Because there was no significant difference between any of the control groups (P > 0.05), we combined the results of all experiments.
Differences between the congenic strains in B. pertussis infection of the lung. We tested 12 different CcS strains to determine the number of viable bacteria (CFU) in the lung 1 week postinoculation. Both for CcS mice and for HcB mice, a wide range in bacterial numbers was found. Results are shown in Fig. 2. Significant differences in lung colonization were observed between the strains of mice. Horizontal lines connect groups of mice that are mutually not significantly different according to the Student-Neuman-Keuls test. The number of CFU varied from 1.6 x 106 to 9.0 x 106 CFU per lung (Fig. 2). From these strains we selected CcS4 as the most representative strain for further breeding.
Similar experiments were performed with 21 different HcB strains (Fig. 2). The differences in lung colonization for Hcb mice were more pronounced compared to that for CcS mice and ranged from 1.4 x 105 to 6.8 x 106 CFU per lung. From these strains we selected HcB28 as the most resistant strain (HcB29 was not a suitable strain for breeding). We also observed pathological differences in the lungs, ranging from macroscopically healthy lungs to severe lung edema (data not shown).
Identification of new susceptibility loci. Based on the data presented above, we selected the CcS4 and HcB28 strains for subsequent F2 hybrid experiments to identify susceptibility loci. Both strains were used to generate an F2 hybrid generation of approximately 200 mice. F2 mice were inoculated with B. pertussis, and the number of CFU in the lung was determined after 7 days (Fig. 3). The numbers of CFU per lung for the F2 hybrid generation (CcS4 x BALB/c) ranged from 1.0 x 102 (detection limit) to 1.8 x 107, and the average for this group was 3.7 x 106. The numbers of CFU per lung for the F2 hybrid generation (Hcb28 x C3H) ranged from 1.0 x 102 to 3.4 x 106, and the average for this group was 5.0 x 105 CFU per lung. All mice were individually genotyped, and we compared genotypes with phenotypes to identify possible QTLs by linkage analysis.
In the F2 mice, generated by crossing HcB28 and C3H mice, we found linkage between the number of CFU and a region on chromosome 12. In Fig. 4, the LOD diagram is shown for the telomere region of this chromosome. We found maximum linkage between the number of CFU and the locus defined by S12Rivm102 (102.4 Mb) with a LOD score of 4.6. This LOD score and the corresponding P value (P = 0.000025) are statistically significant, also after correction for multiple comparisons as described in the statistical section (P = 0.0019). However, this result does not withstand correction for experiment as a random factor (P = 0.36). Although the experimental conditions are highly reproducible and the distribution of genotypes per experiment appeared to be normal according to the 2 test (P = 0.489), the day of the experiment has influence on the established linkage. We therefore consider these results "suggestive linkage" instead of significant linkage, and they warrant further investigation. Almost 10 percent (9.7%) of the variation in the number of bacteria in the experiments could be ascribed to the found locus (r2 = 0.097). We propose to call the locus B. pertussis susceptibility locus 1 (Bps-1). In Table 2, means of numbers of CFU for F2 mice are displayed per genotype. The mean differences are significantly different at the 0.05 level between C3H and heterozygotes and between C3H and C57BL/10, but there is no significant difference between heterozygotes and C57BL/10 mice. This indicates that the presence of C57BL/10 DNA has a dominant-positive effect on the clearance of B. pertussis from the lung.
In the F2 mice generated by crossing CcS4 with BALB/c, we did not find any linkage between individual markers and the number of CFU. However, we found linkage between the numbers of CFU and the interacting markers D5Mit179 (chromosome 5) and D11Mit122 (chromosome 11) with a LOD score of 2.92 (Fig. 5). This LOD score and the corresponding P value (P = 0.010) are statistically significant, but not after correction for multiple comparisons as described in the statistical section (P = 0.28). However, when the experiment is included as a random factor, the significance increases (P = 0.089). We therefore consider these results suggestive linkage, instead of significant linkage, and they warrant further investigation.
Seven percent (7.0%) of the variation in the number of bacteria in the experiments could be ascribed to these interacting loci (r2 = 0.070). We propose to call these loci Bps-2 and Bps-3. In Table 3, means of numbers of CFU for F2 mice are displayed per interaction of D5Mit179 and D11Mit122 (the marker combination with the highest linkage) genotypes. When both markers carry the BALB/c allele instead of STS, the number of CFU in the lung is relatively high.
DISCUSSION
Pertussis is reemerging, despite high vaccination coverage (26). Further, pertussis morbidity and mortality is especially high for babies that are too young to be vaccinated (32). Treatment of pertussis with antibiotics is only effective if initiated early during infection. Thus, novel therapeutic or preventive treatments of pertussis are required.
Elucidation of genetic differences in susceptibility to complex diseases can result in improved knowledge of pathways involved in the pathogenesis of these diseases. Identification of genes which affect susceptibility to infectious diseases is important in unraveling pathogen-host interactions. Furthermore, identification of such genes may lead to novel therapeutics. Novel therapeutics are especially important for infectious diseases which are difficult to control by vaccination, such as tuberculosis, pertussis, and AIDS, and/or by antibiotics, such as diseases caused by multiresistant bacteria.
To identify genes that affect the course of B. pertussis infection, we tested two sets of recombinant congenic strains, the STS and HcB sets. This approach has been applied successfully before in identifying Sst1 as one of the loci in mice affecting susceptibility to Mycobacterium tuberculosis (16, 17). The same approach was also successful in identifying a series of loci (Lmr) each associated with a different combination of pathological and immunological reactions to Leishmania major (1, 20, 34).
In this study we studied the effect of genetic differences in the mouse on lung colonization by B. pertussis. Recently a mutation in Toll-like receptor 4 (Tlr-4) was identified as a major factor that influences the course of B. pertussis infection in mice (13, 22). The gene coding for Tlr4 is positioned on mouse chromosome 4 (33 centimorgans [cM]). The strains of mice we applied here do not have documented mutations in Tlr4, and this gene therefore did not influence the variation observed in our experiments.
We tested two sets of RCS of mice and selected one strain from each set for generation of F2 hybrid generations to identify susceptibility loci. For both sets of RCS, a wide range was observed in the numbers of CFU recovered from the lungs. The phenotypic variation in these strains of mice, which differ in only 12.5% of their genome, indicates that genetic factors do influence B. pertussis pathogenesis. In addition, the range of variation between these strains suggests that the course of pertussis is controlled by multiple genes and that the genetics of B. pertussis infection is complex and controlled by multiple QTLs.
To identify loci potentially responsible for the observed differences in lung colonization, we generated F2 hybrid generations of approximately 200 mice. As shown in Fig. 3, there are large but similar variations in bacterial colonization in F2 mice generated from CcS4 and HcB28. Based on the SDPs, we genotyped all mice from both F2 hybrids and compared the genotypes with the degree of lung colonization. For a classical mouse F2 intercross generation (whole-genome segregating), the following critical values for establishing linkage are used: a LOD score below 2.8 argues against linkage, while a LOD score of 4.3 or higher indicates linkage (18, 19). For an RCS F2 hybrid generation, the critical values are lower because only 12.5% of the genome is segregating. Therefore, all P values are corrected for multiple comparisons as described in the statistical analysis section.
In the F2 mice generated from the HcB28 mice, we initially found linkage between lung colonization and one marker, D12Mit133 (chromosome 12), with a LOD score of 2.7. We subsequently tested several flanking markers and found a maximum linkage between number of CFU and S12Rivm102 with a LOD score of 4.6. This finding could explain almost 10% of the variation in CFU observed with the F2 mice generated from HcB28 (Fig. 4). This locus, which we propose to designate B. pertussis susceptibility locus 1 (Bps-1), is positioned near the telomere on chromosome 12 (approximately 102 Mb) and influences the number of bacteria present in the lung 1 week after inoculation. As is shown in Table 2, the presence of C57BL/10 DNA on this allele instead of C3H DNA has a dominant-positive effect on the clearance of bacteria from the lung.
For the F2 mice generated by crossing CcS4 with BALB/c, we did not find any linkage between individual markers and the number of CFU. However, we found linkage between the number of CFU and the interacting markers D5Mit179 (chromosome 5) and D11Mit122 (chromosome 11) with a LOD score of 2.92 (Fig. 5). Some (7.0%) of the variation in the number of bacteria in this experiment could be ascribed to these interacting loci (r2 = 0.070). We propose to call these loci Bps-2 and Bps-3. In Table 3, means of CFU for F2 mice are displayed per interaction of D5Mit179 and D11Mit122 (the marker combination with the highest linkage) genotypes. These linked regions are homologous to the human loci 7q21-q22 (D5Mit179) and 17p11-q12 (D11Mit122). These regions contain numerous genes, including genes encoding immunological functions, but the regions must be narrowed down before focusing on individual genes is possible.
The significance of the results described here may not withstand all statistical corrections, but nevertheless, the results are promising enough for future investigation. We consider the three novel loci to be in suggestive linkage. Because, according to the 2 test, the genotypes on the Bps-1 locus are normally distributed (P = 0.489), the relevance of the Bps-1 locus is underlined.
Because the wide range of variation found in the RCS mice suggests a multigenic control of B. pertussis infection, we did not expect to find only three loci influencing our phenotype. It is therefore likely that the combined effect of other loci on this phenotype is too subtle to detect in these experiments. Presumably, because B. pertussis infection is controlled by multiple genes, the extremes in phenotypes must be larger for identification of subtle genetic effects. Although all mice were raised and housed in the same facility under carefully controlled conditions, the contributions of other factors besides genetics cannot be excluded completely as complicating factors in the differences in bacterial load between mice. In addition, it would probably be necessary to increase the number of F2 animals to detect significant QTLs.
Based on the data presented in Fig. 4, we selected the region between 100 and 108 Mb around Bps-1 on chromosome 12 for future analysis. This region is similar to the human locus 14q32. It contains 84 different genes, including genes involved in immunity, such as tumor necrosis factor-associated factor 3 and microtubule affinity-regulating kinase 3. However, the majority of genes in this region have an unknown function (RIKEN cDNA). In future experiments we aim to identify the gene(s) that modifies the susceptibility to B. pertussis by a combination of positional cloning and expression analysis.
In conclusion, we screened two sets of RCS of mice for susceptibility to B. pertussis infection and found a wide range in bacterial numbers in the lung at 1 week postinoculation. This indicates multigenic control of the B. pertussis infection. We identified one locus located on chromosome 12, which we designated as Bps-1, and two interacting loci on chromosomes 5 and 11, designated Bps-2 and -3, which influence the number of bacteria in the lung 1 week after inoculation. The presence of C57BL/10 DNA in Bps-1 instead of C3H DNA has a dominant-positive effect on the clearance of bacteria from the lung.
ACKNOWLEDGMENTS
We thank Hans van Oirschot and Silke David for helping with the animal experiments. We also thank all biotechnicians of our animal facility, especially Henk Gielen, for useful help in performing the animal experiments.
REFERENCES
1. Badalova, J., M. Svobodova, H. Havelkova, V. Vladimirov, J. Vojtiskova, J. Engova, T. Pilcik, P. Volf, P. Demant, and M. Lipoldova. 2002. Separation and mapping of multiple genes that control IgE level in Leishmania major infected mice. Genes Immun. 3:187-195.
2. Barnard, A., B. P. Mahon, J. Watkins, K. Redhead, and K. H. Mills. 1996. Th1/Th2 cell dichotomy in acquired immunity to Bordetella pertussis: variables in the in vivo priming and in vitro cytokine detection techniques affect the classification of T-cell subsets as Th1, Th2 or Th0. Immunology 87:372-380.
3. Belcher, C. E., J. Drenkow, B. Kehoe, T. R. Gingeras, N. McNamara, H. Lemjabbar, C. Basbaum, and D. A. Relman. 2000. The transcriptional responses of respiratory epithelial cells to Bordetella pertussis reveal host defensive and pathogen counter-defensive strategies. Proc. Natl. Acad. Sci. USA 97:13847-13852.
4. Cherry, J. D. 1999. Epidemiological, clinical, and laboratory aspects of pertussis in adults. Clin. Infect. Dis. 28(Suppl.)2:S112—S117.
5. Churchill, G. 2004. QTL mapping overview. [Online.] http://www.jax.org/staff/churchill/labsite/research/qtl/. Accessed 13 April 2004.
6. Demant, P. 1992. Genetic resolution of susceptibility to cancer-new perspectives. Semin. Cancer Biol. 3:159-166.
7. Demant, P. 2003. Cancer susceptibility in the mouse: genetics, biology and implications for human cancer. Nat. Rev. Genet. 4:721-734.
8. Demant, P., and A. A. Hart. 1986. Recombinant congenic strains-a new tool for analyzing genetic traits determined by more than one gene. Immunogenetics 24:416-422.
9. Frankel, W. N. 1995. Taking stock of complex trait genetics in mice. Trends Genet. 11:471-477.
10. Frankel, W. N., and N. J. Schork. 1996. Who's afraid of epistasis Nat. Genet. 14:371-373.
11. Groot, P. C., C. J. Moen, W. Dietrich, J. P. Stoye, E. S. Lander, and P. Demant. 1992. The recombinant congenic strains for analysis of multigenic traits: genetic composition. FASEB J. 6:2826-2835.
12. Hellwig, S. M., H. F. van Oirschot, W. L. Hazenbos, A. B. van Spriel, F. R. Mooi, and J. G. van de Winkel. 2001. Targeting to Fc receptors, but not CR3 (CD11b/CD18), increases clearance of Bordetella pertussis. J. Infect. Dis. 183:871-879.
13. Higgins, S. C., E. C. Lavelle, C. McCann, B. Keogh, E. McNeela, p. Byrne, B. O'Gorman, A. Jarnicki, P. McGuirk, and K. H. Mills. 2003. Toll-like receptor 4-mediated innate IL-10 activates antigen-specific regulatory T cells and confers resistance to Bordetella pertussis by inhibiting inflammatory pathology. J. Immunol. 171:3119-3127.
14. Kimman, T. 2001. Genetics of infectious disease susceptibility. Kluwer Academic Publishers, Dordrecht, The Netherlands.
15. King, A. J., G. Berbers, H. F. van Oirschot, P. Hoogerhout, K. Knipping, and F. R. Mooi. 2001. Role of the polymorphic region 1 of the Bordetella pertussis protein pertactin in immunity. Microbiology 147:2885-2895.
16. Kramnik, I., P. Demant, and B. B. Bloom. 1998. Susceptibility to tuberculosis as a complex genetic trait: analysis using recombinant congenic strains of mice. Novartis Found. Symp. 217:120-131.
17. Kramnik, I., W. F. Dietrich, P. Demant, and B. R. Bloom. 2000. Genetic control of resistance to experimental infection with virulent Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 97:8560-8565.
18. Lander, E., and L. Kruglyak. 1995. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat. Genet. 11:241-247.
19. Lander, E. S., and N. J. Schork. 1994. Genetic dissection of complex traits. Science 265:2037-2048.
20. Lipoldova, M., M. Svobodova, M. Krulova, H. Havelkova, J. Badalova, E. Nohynkova, V. Holan, A. A. Hart, P. Volf, and P. Demant. 2000. Susceptibility to Leishmania major infection in mice: multiple loci and heterogeneity of immunopathological phenotypes. Genes Immun. 1:200-206.
21. Mahon, B. P., B. J. Sheahan, F. Griffin, G. Murphy, and K. H. Mills. 1997. Atypical disease after Bordetella pertussis respiratory infection of mice with targeted disruptions of interferon-gamma receptor or immunoglobulin mu chain genes. J. Exp. Med. 186:1843-1851.
22. Mann, P. B., M. J. Kennett, and E. T. Harvill. 2004. Toll-like receptor 4 is critical to innate host defense in a murine model of bordetellosis. J. Infect. Dis. 189:833-836.
23. Massachusetts Institute of Technology. 2004. Look up STSs by name. [Online.] http://www.broad.mit.edu/cgi-bin/mouse/sts_infodatabase=mouserelease. Accessed 28 April 2004.
24. Mattoo, S., A. K. Foreman-Wykert, P. A. Cotter, and J. F. Miller. 2001. Mechanisms of Bordetella pathogenesis. Front. Biosci. 6:E168-E186.
25. Mills, K. H., M. Ryan, E. Ryan, and B. P. Mahon. 1998. A murine model in which protection correlates with pertussis vaccine efficacy in children reveals complementary roles for humoral and cell-mediated immunity in protection against Bordetella pertussis. Infect. Immun. 66:594-602.
26. Mooi, F. R., I. H. Van Loo, and A. J. King. 2001. Adaptation of Bordetella pertussis to vaccination: a cause for its reemergence Emerg. Infect. Dis. 7:526-528.
27. Mooi, F. R., H. van Oirschot, K. Heuvelman, H. G. van der Heide, W. Gaastra, and R. J. Willems. 1998. Polymorphism in the Bordetella pertussis virulence factors P.69/pertactin and pertussis toxin in The Netherlands: temporal trends and evidence for vaccine-driven evolution. Infect. Immun. 66:670-675.
28. Mouse Genome Informatics. 2004. Recombinant congenic strain distribution patterns query form. [Online.] http://www.informatics.jax.org/searches/rcsetb_form.shtml. Accessed 28 April 2004.
29. Nadeau, J. H. 2001. Modifier genes in mice and humans. Nat. Rev. Genet. 2:165-174.
30. Nickerson, C. A. 2001. Altered gene expression in response to Bordetella pertussis infection. Trends Microbiol. 9:57-58.
31. Stassen, A. P., P. C. Groot, J. T. Eppig, and P. Demant. 1996. Genetic composition of the recombinant congenic strains. Mamm. Genome 7:55-58.
32. Tanaka, M., C. R. Vitek, F. B. Pascual, K. M. Bisgard, J. E. Tate, and T. V. Murphy. 2003. Trends in pertussis among infants in the United States, 1980-1999. JAMA 290:2968-2975.
33. Verwey, W. F., E. H. Thiele, D. N. Sage, and L. T. Suchardt. 1949. A simplified liquid culture medium for the growth of Haemophilus pertussis. J. Bacteriol. 58:127-134.
34. Vladimirov, V., J. Badalova, M. Svobodova, H. Havelkova, A. A. Hart, H. Blazkova, P. Demant, and M. Lipoldova. 2003. Different genetic control of cutaneous and visceral disease after Leishmania major infection in mice. Infect. Immun. 71:2041-2046.
35. Willems, R. J., J. Kamerbeek, C. A. Geuijen, J. Top, H. Gielen, W. Gaastra, and F. R. Mooi. 1998. The efficacy of a whole cell pertussis vaccine and fimbriae against Bordetella pertussis and Bordetella parapertussis infections in a respiratory mouse model. Vaccine 16:410-416.
36. World Health Organization. 2003. Pertussis vaccine. [Online.] http://www.who.int/vaccines/en/pertussis.shtml. Accessed 17 Oct. 2003.(H. A. Banus, H. J. van Kr)