Lack of Association between the Tlr4 (Lps^d/Lps^d) Genotype and Increased Susceptibility to Escherichia coli Bladder Infections in Female C3H/HeJ Mice

ABSTRACT

Toll-like receptor 4 is thought to have a primary role in host defense against Escherichia coli bladder colonization, based on mouse models of urinary tract infection using C3H/HeJ female mice. This strain carries a point mutation in the Tlr4 gene, which renders the mice unresponsive to lipopolysaccharide (LPS) and thus limits the bladder inflammatory response and infection resolution. The importance of Tlr4 as the sole genetic determinant of resistance or susceptibility can be questioned, however, by the observation that C3H/HeOuJ female mice with a functional Tlr4 do not effectively resolve E. coli bladder infections. The present study further examined this inconsistency by investigating the association of Tlr4 Lps^d and Lpsⁿ alleles with bladder infection susceptibility by using genetic crosses of C3H/HeJ mice with Tlr4 (Lpsⁿ/Lpsⁿ) or (Lpsⁿ/Lps^d) mice. Heterozygous offspring of C3H/HeJ (Lps^d/Lps^d) × BALB/cAnN (Lpsⁿ/Lpsⁿ) mice successfully resolved bladder infections induced by a uropathogenic E. coli strain, while heterozygous mice from a C3H/HeJ (Lps^d/Lps^d) × C3H/HeOuJ (Lpsⁿ/Lpsⁿ) cross had severe infections. A backcross of C3H/HeJ (Lps^d/Lps^d) with (BALB/cAnN × C3H/HeJ)F₁ (Lpsⁿ/Lps^d) produced mice that were either resistant or susceptible to E. coli bladder infections and had Lps^d/Lps^d or Lpsⁿ/Lps^d Tlr4 genotypes. The Lps^d/Lps^d or Lpsⁿ/Lps^d genotypes were present in individual mice with unresolved bladder infections, and the Lps^d/Lps^d genotype was found in infection-resistant mice. These results indicate that at least one gene other than Tlr4 strongly influences susceptibility to E. coli bladder infections in C3H/HeJ mice.

IMPORTANCE

We have previously demonstrated that mouse strains with either a functional or nonfunctional Tlr4 were not able to resolve induced Escherichia coli bladder infections and that a chromosomal site distinct from Tlr4 was associated with an inability to resolve bladder infections in C3H/HeJ mice. The present study has further investigated the relevance of Tlr4 in bladder infection resolution by defining the Tlr4 alleles present in offspring of genetic crosses of C3H/HeJ mice with infection-resistant and -susceptible inbred strains. The results of these experiments showed that mice with a normal Tlr4 on different genetic backgrounds were not able to clear E. coli bladder infections and that animals with a defective Tlr4 could successfully resolve infections. These results strongly imply the presence of a gene other than in Tlr4 as an important genetic determinant of infection resistance/susceptibility in C3H/HeJ and other inbred mouse strains used in mouse models of infectious diseases.

Introduction

The importance of Tlr4 in host defense against Escherichia coli urinary tract infection (UTI) has been extensively studied in mouse models where the course of infection and immune responses can be studied in animals with normal and defective Tlr4 alleles (1, 2). Mice from the C3H/HeJ strain have a point mutation in the Tlr4 gene that makes Tlr4-bearing cells unresponsive to lipopolysaccharide (LPS), and the animals are unable to clear E. coli from the bladder up to 14 days after intravesical inoculation (2, 3). Because uroepithelial cells in C3H/HeJ mice do not secrete proinflammatory cytokines normally initiated through the Tlr4, there is a markedly reduced neutrophil influx into the infected bladder. Mouse strains with a functional Tlr4 such as BALB/cAnN, C3H/HeN, and C57BL/6 resolve similarly induced infections within several days and exhibit pyuria (4) and inflammatory cells in the bladder wall (3). These studies imply that a defective Tlr4 is the primary reason for failure of C3H/HeJ mice to resolve E. coli bladder infections.

In contrast to the studies cited above, further investigations with C3H/HeJ mice and the closely related C3H/HeOuJ strain have raised the question of whether Tlr4 is the sole genetic determinant of bladder infection susceptibility in mouse models of E. coli UTI. First, we observed that Tlr4-normal, C3H/HeOuJ female mice failed to resolve induced E. coli bladder infections in spite of an intense bladder inflammatory response, indicating that one or more genes other than Tlr4 play a significant role in resistance to infection (3). Second, an investigation into the inheritance of susceptibility to E. coli bladder infection using C3H/HeJ, BALB/cAnN, and (C3H/HeJ × BALB/cAnN)F₁ mice showed that infection susceptibility was a recessive trait (5). Furthermore, statistical analysis indicated that one or more genes in C3H/HeJ mice were responsible for their inability to clear bladder infections. Third, in a genetic linkage analysis using BALB/cAnN and C3H/HeJ mice, we identified a quantitative trait locus (QTL), Becis1, that was significantly associated with the inability of C3H/HeJ female mice to resolve E. coli bladder infections (6). The location of Becis1 was estimated to be at 29.0 cM on chromosome 4, which is near the Tlr4 locus at 33.0 cM, suggesting that two distinct genes may contribute to host defense against bladder infection.

The present studies have sought to determine whether the defective Tlr4 allele present in C3H/HeJ mice could be consistently associated with increased susceptibility or resistance to E. coli bladder infections in mice that are homozygous or heterozygous for the defective and normal alleles in different genetic backgrounds. To assess the relationship between the Tlr4 genotype and resistance or susceptibility to severe E. coli bladder infection in female mice, two sets of experiments investigated E. coli bladder infection intensities in mice with various combinations of the normal and defective Tlr4 alleles, Lpsⁿ and Lps^d, respectively. The first experiments utilized inbred mice that were homozygous for Lpsⁿ or Lps^d or hybrids that were heterozygous Lpsⁿ/Lps^d. A subsequent study investigated the association between the Tlr4 genotype and infection phenotype in a backcross population where animals carried the Lps^d/Lps^d or Lpsⁿ/Lps^d genotypes in combination with other genes present in the parental strains. All experiments measured bladder infection intensities 10 days after intravesical inoculation with 2 × 10⁸ uropathogenic E. coli cells of strain 1677 using a protocol (7) approved by the Animal Use and Care Committee of the University of Wisconsin School of Medicine and Public Health. E. coli strain 1677 was isolated from a woman with a febrile UTI. Strain 1677 belongs to the B2 phylotype and has the O6 serotype and many virulence genes and markers commonly associated with uropathogenic E. coli, such as type 1 and p fimbriae and the genes iha, hlyA, sat, fyuA, iutA, ompT, chuA, yjaA, malX, and traT.

The results in Table 1 show clear differences in the abilities of the BALB/cAnN, C3H/HeJ, C3H/HeOuJ strains and hybrids of these mice to resolve E. coli bladder infections, implying that the Tlr4 Lps^d allele is not the primary genetic factor in infection susceptibility. Female BALB/cAnN mice (Lpsⁿ/Lpsⁿ) were able to effectively resolve their bladder infections by 10 days after inoculation, while C3H/HeJ (Lps^d/Lps^d) mice remained severely infected at this time point. Mice that were Lpsⁿ/Lps^d on a background of BALB/cAnN and C3H/HeJ alleles were also resistant to infection, suggesting that the Tlr4 Lpsⁿ allele confers protection. The importance of a functional Tlr4 can be questioned, however, by the failure of C3H/HeOuJ (Lpsⁿ/Lpsⁿ) mice to effectively resolve infections. Moreover, females from a cross between C3H/HeJ (Lps^d/Lps^d) and C3H/HeOuJ (Lpsⁿ/Lpsⁿ) mice were also not able to effectively resolve E. coli bladder infections within 10 days. Thus, there are very likely to be alleles of genes other than Tlr4 that significantly lower resistance to E. coli bladder infection in both of these C3H strains.

TABLE 1 .

Bladder infection intensities and Tlr4 genotypes of female inbred mice and hybrids following intravesical inoculation with E. coli

Inbred mouse strain or hybrida	No. of mice tested	Tlr4 genotype	CFU in bladderb
Inbred
BALB/cAnN	10	Lpsn/Lpsn	70 (1.84 ± 0.20) ABC
C3H/HeJ	6	Lpsd/Lpsd	533,921 (5.73 ± 0.60) AD
C3H/HeOuJ	6	Lpsn/Lpsn	489,778 (5.69 ± 0.20) BD
Hybrid
C3H/HeJ × BALB/cAnN	15	Lpsn/Lpsd	44 (1.64 ± 0.09) EC
C3H/HeJ × C3H/HeOuJ	7	Lpsn/Lpsd	248,427 (5.40 ± 0.04) E

^a Female mice from specific inbred strains (BALB/cAnN, C3H/HeJ, and C3H/HeOuJ) and hybrids bred from these strains [(C3H/HeJ × BALB/cAnN) and (C3H/HeJ × C3H/HeOuJ)].

^b Female mice were inoculated intravesically with 1 × 10⁸ E. coli 1677 and euthanized 10 days later to assay for the total number of CFU in the bladder. Values are geometric means with mean log₁₀ (CFU/mg of tissue) ± standard errors of the means in parentheses. Statistical analysis was conducted in two steps using SAS/STAT software, version 9.1 (SAS Institute, Inc.). The log-transformed CFU data were first analyzed by analysis of variance to estimate variability of values within and between experimental groups and then by Fisher’s protected least significant differences test to determine significant differences between groups. P values less than 0.05 were considered statistically significant. For comparisons between groups indicated by the letters A, B, and E, P is <0.001. For comparisons between groups indicated by the letters C and D, P is >0.05.

The relevance of the normal and defective Tlr4 alleles in resolution of E. coli bladder infections was further investigated in female mice from a backcross of C3H/HeJ with (BALB/cAnN × C3H/HeJ)F₁. These backcross mice were either heterozygous for BALB/cAnN and C3H/HeJ alleles or homozygous for C3H/HeJ alleles at individual loci, including Tlr4. All mice were inoculated with E. coli and assessed for bladder CFU 10 days later as described above. The Tlr4 genotypes of mice with less than 1,000 or greater than 100,000 CFU/mg of tissue were analyzed for associations between Tlr4 and infection intensity. Using the single nucleotide polymorphism that distinguishes the Lpsⁿ and Lps^d Tlr4 alleles (8), we genotyped 140 backcross mice for the Tlr4 normal and defective alleles by a previously described method (9). Genomic DNA was prepared from the spleen of each of mouse with the Gentra Puregene tissue kit (Qiagen). A DNA fragment covering the SNP region was generated by PCR using forward (5′ GCTTTCACCTCTGCCTTCAC3′) and reverse (5′ATAACCTTCCGGCTCTTGTG3′) primers (Integrated DNA Technologies); RedTaq polymerase (Sigma, Chemical Co.); and 35 cycles of 94°C for 30 s, 55°C for 60 s, and 72°C for 45 s with final extension at 72°C for 10 min. The PCR product was purified with the Wizard PCR cleanup system (Promega). To differentiate the Lpsⁿ and Lps^d alleles, the reaction product was digested with Hsp92II restriction endonuclease (New England Biolabs), followed by electrophoresis on a 2.5% agarose gel. The restriction enzyme cleaves DNA with the Lps^d allele to yield two fragments distinguishable from the undigested DNA with the Lpsⁿ allele. These results are presented in Table 2. Twenty-one of the backcross mice had severe bladder infections 10 days after inoculation with E. coli. Six of these animals were Lpsⁿ/Lps^d heterozygous and had infection intensities that were equivalent to those of either the C3H/HeJ or C3H/HeOuJ strains shown in Table 1. All of the remaining 119 backcross mice resolved their bladder infections as effectively as BALB/cAnN mice. Of these infection-resistant mice, 48 were Lps^d/Lps^d homozygous, and 71 were Lpsⁿ/Lps^d heterozygous. The results showing successful resolution of infections in mice with two copies of the Lps^d allele are thus inconsistent with those from C3H/HeJ mice shown in Table 1.

TABLE 2 .

Bladder infection intensities and Tlr4 genotypes of C3H/HeJ female backcross mice following intravesical inoculation with E. coli

Backcross mice testeda	No. of mice tested	Tlr4 genotypeb	CFU in bladderc
(BALB/cAnN × C3H/HeJ) × C3H/HeJ	15	Lpsd/Lpsd	468,814 (5.67 ± 0.10) AB
(BALB/cAnN × C3H/HeJ) × C3H/HeJ	6	Lpsn/Lpsd	322,775 (5.51 ± 0.30) AC
(BALB/cAnN × C3H/HeJ) × C3H/HeJ	48	Lpsd/Lpsd	76 (1.88 ± 0.07) BCD
(BALB/cAnN × C3H/HeJ) × C3H/HeJ	71	Lpsn/Lpsd	32 (1.51 ± 0.06) BCD

^a Female mice from a backcross of an infection-susceptible C3H/HeJ mouse to infection-resistant (BALB/cAnN × C3H/HeJ)F₁.

^b Mice were genotyped to determine if they were homozygous or heterozygous for the Tlr4-normal (Lpsⁿ) and Tlr4-defective (Lps^d) alleles, as defined by a single nucleotide polymorphism in the alleles.

^c Female mice were inoculated intravesically with 1 × 10⁸ E. coli cells and euthanized 10 days later to assay for the total number of CFU in the bladder. Values are geometric means with mean log₁₀ (CFU/mg of tissue) ± standard errors of the means in parentheses. Statistical analysis was conducted in two steps with SAS/STAT software, version 9.1 (SAS Institute, Inc.). The log-transformed CFU data were first analyzed by analysis of variance to estimate variability of values within and between experimental groups and then by Fisher’s protected least significant differences test to determine significant differences between groups. P values less than 0.05 were considered statistically significant. For comparisons between the groups indicated by the letters A and D, P is >0.05. For comparisons between the groups indicated by the letters B and C, P is <0.001.

A functional Tlr4 receptor, as specified by the Tlr4 Lpsⁿ allele, has been proposed to play a central role in host defense against E. coli bladder infections in inbred strains of mice through initiation of local inflammatory responses (1, 10). These studies have noted impaired infection resolution in C3H/HeJ (Lps^d/Lps^d) mice compared to successful resolution in C3H/HeN (Lpsⁿ/Lpsⁿ) mice. Several of the findings reported here, however, question whether Tlr4 is the gene solely responsible for determining host resistance to bladder infection. First, C3H/HeOuJ (Lpsⁿ/Lpsⁿ) mice are comparable to C3H/HeJ (Lps^d/Lps^d) mice in their inability to resolve E. coli bladder infections. Second, when the Tlr4 Lpsⁿ allele was introduced into C3H/HeJ by a cross with C3H/HeOuJ, the hybrid mice are unable to clear infections. This observation is in contrast to results obtained in a cross of BALB/cAnN (Lpsⁿ/Lpsⁿ) mice with C3H/HeJ (Lps^d/Lps^d) mice in which the hybrid mice successfully resolved bladder infections. Third, heterozygous Lpsⁿ/Lps^d mice from a C3H/HeJ backcross to (C3H/HeOuJ × BALB/cAnN)F₁ mice had severe E. coli bladder infections similar to C3H/HeJ mice or the (C3H/HeOuJ × C3H/HeJ) hybrid. Fourth, homozygous Lps^d/Lps^d backcross mice were able to successfully resolve induced E. coli bladder infections. Thus, there is strong evidence that alleles of a gene or genes other than Tlr4 strongly affect resistance or susceptibility to E. coli bladder infections in C3H/HeJ, C3H/HeOuJ, and BALB/cAnN mice.

One model to reconcile findings in this study would be one in which a UTI-associated gene with at least two alleles is present in C3H/HeJ, C3H/HeOuJ, and BALB/cAnN mice. An allele conferring resistance to E. coli bladder infections would be homozygous in the BALB/cAnN strain. An allele associated with an impaired ability to resolve infections would be homozygous in C3H/HeJ, C3H/HeOuJ, and (C3H/HeOuJ × C3H/HeJ) mice. Because heterozygous offspring of a cross between BALB/cAnN and C3H/HeJ mice were able to effectively clear induced E. coli bladder infections, infection resistance would be considered a dominant trait. The infection-susceptible allele in this model would be expected to be homozygous in C3H/HeJ backcross mice with severe bladder infections, while at least one copy of the infection-resistant allele from the BALB/cAnN parental strain would be in backcross mice who successfully resolved infections.

Although the UTI-associated gene proposed in the above model is not currently known, our previous genetic studies have identified the Becis1 QTL as significantly associated with unresolved bladder infections in C3H/HeJ mice (6). This locus maps to chromosome 4 at 29 cM, and the as yet unidentified gene inferred by the Becis1 QTL is the most likely candidate to account for the results in the present study. The Becis1 gene is very likely different from Tlr4 based on chromosome location and the inconsistent associations of the Tlr4 Lps^d allele with high bladder infection intensities. In addition, there are no currently defined genes near Becis1 that are directly involved with either innate or adaptive immune responses or appear to be related to host factors affecting bladder colonization by E. coli. The gene inferred by Becis1 will thus need to be defined and characterized.

Another candidate gene to consider in explaining the current results is Tlr11, which has been noted to play an important role in host defense against E. coli UTIs in mice (11). As currently defined, however, Tlr11 is thought to be located on chromosome 14 and does not coincide with the Becis1 QTL site. A recent report on the evolution of the Toll-like receptor gene family has proposed that mouse Tlr11 should be renamed Tlr12 based on similarities between the Tlr11 and Tlr12 sequences (12). If the two genes are identical or alleles of a single gene, then Tlr11/12 would be located at the current site of Tlr12 on chromosome 4. The Tlr12 gene has not been genetically mapped to a specific site but is syntenic on chromosome 4 (13), and physical mapping places the gene at bp 128292690 to 128295863 (13). The physical mapping would thus indicate a chromosomal location at approximately 61 cM and thus not near either Becis1 or Tlr4.

The present study has investigated the significance of Tlr4 in resolution of E. coli bladder infections in C3H/HeJ mice or mice derived from genetic crosses with C3H/HeJ mice. Although C3H/HeJ (Lps^d/Lps^d) mice have been used as a model to study host defense mechanisms against E. coli UTIs and have shown the importance of innate immune responses in infection resolution, the results presented here and in other studies support the view that Tlr4 is not the sole determinant of resistance or susceptibility to E. coli bladder infections in mice. Rather, there is good evidence that another gene on chromosome 4 has alleles that strongly influence whether an induced E. coli bladder infection will be successfully resolved.

These studies have important implications for investigations using C3H/HeJ mice to study the role of Tlr4 in resolution of infections in organ systems other than the urinary tract. For example, comparisons between C3H/HeJ and Tlr4-normal strains have provided model systems with which to study host defense mechanisms against Neisseria meningitidis bacteremia (14), Haemophilus influenzae pulmonary infections (15), and lethal Salmonella enterica serovar Typhimurium infection (16). Although the defective Tlr4 likely plays a role in susceptibility to bacterial infections in models using C3H/HeJ mice, the studies reported here strongly indicate the presence of multiple genetic factors in host defense against infections caused by E. coli and, potentially, other types of bacteria. This view is consistent with the multigenic nature of susceptibility or resistance to infections documented in several mouse models (17). Thus, it is important to consider genes acting individually or together with Tlr4 when using C3H/HeJ mice in models of infectious diseases.