|Home | About | Journals | Submit | Contact Us | Français|
Genetic linkage studies of the host response to Leishmania major, the causative agent of cutaneous leishmaniasis, have identified significant genetic complexity in humans and mice. In the mouse model, multiple loci have been implicated in susceptibility to infection, but to date, the genes underlying these loci have not been identified. We now describe the contribution of a novel candidate gene, Fli1, to both L. major resistance and enhanced wound healing. We have previously mapped the L. major response locus, lmr2, to proximal chromosome 9 in a genetic cross between the resistant C57BL/6 strain and the susceptible BALB/c strain. We now show that the presence of the resistant C57BL/6 lmr2 allele in susceptible BALB/c mice confers an enhanced L. major resistance and wound healing phenotype. Fine mapping of the lmr2 locus permitted the localization of the lmr2 quantitative trait locus to a 5-Mb interval comprising 21 genes, of which microarray analysis was able to identify differential expression in 1 gene—Fli1. Analysis of Fli1 expression in wounded and L. major-infected skin and naïve and infected lymph nodes validated the importance of Fli1 in lesion resolution and wound healing and identified 3 polymorphisms in the Fli1 promoter, among which a GA repeat element may be the important contributor.
Cutaneous leishmaniasis (CL) caused by the intracellular protozoan Leishmania major is a parasitic disease transmitted by blood-sucking sandflies. Skin macrophages are infected, and a skin lesion or granuloma develops at the site of infection (31). The severity of human CL varies, with phenotypes ranging from self-healing lesions to persistent lesions lasting years and leading to systemic disease (1). The severity of disease depends on the host's ability to form protective granulomas; in their absence, the parasites spread, resulting in diffuse cutaneous leishmaniasis. Studies of CL in mice have been of immense interest because they allow extensive experimental manipulation and reproduce many of the features of the human disease. To identify factors that modulate lesion resolution in infected individuals, we conducted genetic studies on two inbred mouse strains, C57BL/6 (B6) and BALB/c (Bc), that differ in their susceptibility to L. major. Resistance to L. major, epitomized by B6 mice, is characterized by the rapid healing of skin lesions, while the susceptibility in Bc mice is characterized by lesion chronicity and progression and systemic spread of the parasites.
We previously identified three loci (lmr1, lmr2, and lmr3) mediating lesion resolution during L. major infection (28, 29). Mice congenic for all three resistant lmr alleles [strain C.B6-(lmr1, lmr2)] were generated on the susceptible Bc background and were shown to be more resistant to L. major and more able to heal noninfectious wounds than Bc mice, exhibiting a statistically significant tendency toward the phenotype of the donor strain B6. Furthermore, resistance to L. major was independent of T helper cell responses (12, 13, 30). The conclusions from these studies were that lmr1, -2, and -3 control the rate of wound healing and L. major host response independently of T helper cell responses and that a vigorous wound healing response was required for lesion resolution during L. major infection.
To study the contribution of the lmr2 locus in wound healing and L. major resistance, we generated a C.B6-(lmr2) congenic strain by breeding the resistant B6 lmr2 locus onto the susceptible Bc background. As expected, the C.B6-(lmr2) strain did not completely recapitulate the resistant phenotype observed in B6, implicating the involvement of other genetic regions, such as lmr1 and lmr3 (13). However, we were able to demonstrate that C.B6-(lmr2) mice were more resistant to L. major infection and were better able to heal noninfectious wounds than the parental strain Bc, highlighting the contribution of the lmr2 locus to host resistance and enhanced wound healing.
The current study was undertaken to fine map the lmr2 locus. We show that enhanced wound healing and resistance to L. major is conferred by a 5-Mb region within the B6 lmr2 locus. Microarray profiling identified the differential expression of only one gene within this 5-Mb candidate interval—Fli1. Analysis of Fli1 expression in wounded and L. major-infected skin and naïve and infected lymph nodes and macrophages validated the importance of Fli1 in L. major response and wound healing. Furthermore, analysis of B6 and Bc Fli1 sequences has identified 3 polymorphisms, among which a GAn microsatellite polymorphism in the Fli1 promoter may be an important contributor.
All mice were bred at the Walter and Eliza Hall Institute of Medical Research. lmr2 congenic intervals are on chromosome 9. The most proximal and distal microsatellite markers used for genotyping are listed in Fig. Fig.1.1. The effect of the lmr2 locus is more prominent in female mice, so we present data only for female mice. This project was approved by the Walter and Eliza Hall Institute Animal Ethics Committee (AEC no. 2006.018), and all experiments were carried out in compliance with the Australian National Health and Medical Research Council Code of Practice for the Care and Use of Animals in Research in Australia and in line with the guidelines of the U.S. National Institutes of Health.
The virulent cloned line V121, derived from the L. major isolate LRC-L137 (MHOM/IL/67/JerichoII), was used in all infections. Promastigotes were maintained in vitro at 26°C in the biphasic blood agar culture NNN and used in the stationary phase of growth. The virulence of the parasites was maintained by frequent passaging through mice.
Mice were infected with 106 L. major V121 promastigotes at the base of the tail as previously described (13, 21). All infection experiments were repeated twice, and in each infection experiment, 20 to 30 mice per strain were used. Parasite burdens of the inguinal lymph nodes and skin were quantified at weeks 6 and 12 postinfection by quantitative real-time PCR. The inguinal lymph nodes were chosen because they drain the site of infection and are the main foci of the immune response during L. major infection. The week 6 and 12 time points were selected because week 6 is the peak of infection as observed by the initiation of lesion resolution in B6 mice and week 12 presents the biggest difference in lesion phenotype between Bc and B6 mice. Briefly, inguinal lymph nodes and skin were removed and processed using Trizol reagent (Invitrogen). RNA and DNA were stored at −80°C. DNA was purified using a Wizard DNA isolation kit (Promega). Parasite burdens were quantified using real-time PCR with primers for collagen 1 alpha 1, a single-copy gene in mice (For_GAGAAAGGATCTCCTGGTGCT and Rev_CACCACGTTGTCCAGCAATA), and fatty acid desaturase, a single-copy gene in L. major (For_CAACAACGACGTGATTGACC and Rev_CAACGAACATGTCCAGGATG).
In vivo punch biopsy wound assays were performed as previously described (30). Briefly, mice were anesthetized and full-thickness, 4-mm wounds were made at the base of the tail. The rate of wound healing was quantified by measuring the diameter of the wounds. Histological sections were taken at days 0, 1, 3, 5, 7, 9, and 11 and stained with Masson's Trichrome and hematoxylin and eosin (H&E). The Masson's Trichrome- and H&E-stained sections from three mice at each time point were examined microscopically to determined collagen deposition and organization and density of inflammation. Collagen and inflammatory cells were quantified using the image analysis software Metamorph (Molecular Devices).
Microarray expression profiling of bone marrow-derived macrophages from B6, Bc, and C.B6-(lmr1, lmr2) mice was performed as previously described (30). Briefly, bone marrow from 6 mice each of strains B6, Bc, and C.B6-(lmr1, lmr2) was cultured in the presence of fetal calf serum- and L cell-conditioned medium. After 7 days of culture, mature macrophages were washed and incubated for 4 h with or without L. major promastigotes. The cells were then washed free of unbound parasites and incubated for 24 h before being harvested for RNA preparation. Fifteen micrograms of RNA was prepared for each sample by using RNeasy (Qiagen, Valencia, CA). Samples were biotin labeled and hybridized to murine U74Av2 Genechips (Affymetrix, Santa Clara, CA) following the manufacturer's protocol. The GeneChips were scanned using a Hewlett-Packard GeneArray G2500A scanner.
The intensity values were background corrected, normalized, and summarized with the Bioconductor package software gcrma, using the most reliable and computationally intensive parameter estimation setting (34). The probe set annotation, including gene symbol and chromosomal location, was obtained from the relevant Bioconductor annotation package (35).
Mouse tissue was immersed in RNALater (Ambion) to stabilize RNA for processing at a later date. RNA was extracted using Trizol (Invitrogen) according to the manufacturer's instructions. RNA was stored in RNase-free water at −80°C. The concentration was measured at 260 nm and adjusted to 1 μg/μl.
Two micrograms of total RNA was reverse transcribed according to the manufacturer's instructions using a Superscript III reverse transcriptase kit (Invitrogen) to produce cDNA for subsequent analysis by real-time PCR.
Real-time quantitative PCR was carried out on a Roche LightCycler 480 instrument with 10-μl reaction mixture volumes using a Roche LightCycler master SYBR green I kit. The thermal cycling protocol consisted of an initial incubation at 95°C for 10 min and then 50 cycles at 95°C for 15 s (denaturation), 60°C for 10 s (annealing), and 72°C for 10 s (extension). Fluorescence was acquired at the end of the extension phase. Melting curve analysis was performed to ascertain the specificity of the reaction.
Two microliters of cDNA was used per reaction mixture, and all samples were run in duplicate with a negative (water) control, a positive control (calibrator), and a standard. The relative quantification method was used to calculate the concentrations of unknown samples. Primers for Fli1 were For_AGTGCTGTTGTCGCACCTC and Rev_TTCCTTGACATTCAGTCGTGA. The expression of target genes was normalized to that of the housekeeping genes Hmbs (For_CCTGGTTGTTCACTCCCTGA and Rev_CAACAGCATCACAAGGGTTTT) and Gapdh (For_TGCACCACCAACTGCTTAG and Rev_GGATGCAGGGATGATGTT).
Microsatellite markers were amplified in mice by PCR with a 5′-labeled fluorescent primer and unlabeled reverse primer. Primer sequences for known microsatellites can be found at http://www.informatics.jax.org/javawi2/servlet/WIFetch?page=markerQF. Primer sequences for microsatellites identified by the authors are as follows: Fli1 microsatellite, For_GATCGGGGAGGAAGTCCTAT and Rev_GGAGGGAAGACAAGAGAGAGC; D9Wehi123, For_GAGTGATCTGAACTTCACTGCAA and Rev_TGGGGATGGTTAAAACAATGA; D9Wehi103, For_TGTTCATCCCTCAATCACCA and Rev_TGGATAAGACAAATAGCAATTAAACA; and D9Wehi120, For_TACATGGGGAGCCCTGTAGT and Rev_CAGCCTTTCAGTGGATTCCT. Fluorescence data were collected on an ABI3700 automated sequencer and analyzed using software packages Genescan 3.1 and Genotyper 2 (Applied Biosystems). Genotypes were read manually.
Growth curve data from the infection challenges and punch biopsies were compared between groups using a previously described statistical permutation test (4, 30) which could compare the groups over the entire course of infection/healing (http://bioinf.wehi.edu.au/software/compareCurves/). Two-sided t tests were used to test for differences in parasite burdens, cell quantities, and gene expression levels between two groups.
The microarray data have been submitted to the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo) and are publicly available under accession number GSE17550.
The lmr2 locus spans a region between Mb 0 and 50 on mouse chromosome 9. To fine map the lmr2 locus, we constructed a panel of 8 subcongenic lines derived from C.B6-(lmr2) [strains C.B6-(lmr2.1), C.B6-(lmr2.2), C.B6-(lmr2.4), C.B6-(lmr2.5), C.B6-(lmr2.4a), C.B6-(lmr2.4b), C.B6-(lmr2.4c), and C.B6-(lmr2.4d)] that carry an overlapping tiling array of the B6-derived donor genomic interval spanning the lmr2 locus on the Bc background. These lines are congenic only for the lmr2 interval (Fig. (Fig.1).1). These subcongenic lines were screened for resistance to L. major, and comparisons of mean lesion scores (representative of lesion size) over the 12-week infection period demonstrated that B6 mice were the most resistant, presenting with small nodules early in infection that healed by week 12 postinfection. Bc mice were the most susceptible, with large, severe, nonhealing lesions. The differences between B6 and Bc mice were statistically significant (P = 0.001). Of the eight subcongenic lines created, C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice presented the phenotypes that were most strongly and significantly different from the Bc phenotype. Although C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice did not completely recapitulate the absolute resistance observed in B6 mice, the lesions in these mice were significantly less severe than those of the Bc strain (P = 0.001, P = 0.0001, and P = 0.0001, respectively) (Fig. 2a to c), suggesting that the presence of a 5-Mb interval (Mb 29 to 34, B6 chromosome 9) shared between C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice restrained lesion progression and was associated with resistance. An association between the 5-Mb interval and resistance to L. major was further supported by the more susceptible phenotype observed in C.B6-(lmr2.1), C.B6-(lmr2.2), C.B6-(lmr2.5), C.B6-(lmr2.4a), and C.B6-(lmr2.4d) mice (Fig. 3a to e), which all carry the susceptible Bc haplotype for this interval.
The parasite burdens of the draining lymph nodes (Fig. 2d, e, and f) and skin (Fig. 2g, h, and i) were also measured in congenic mice and compared to those in Bc mice. No significant differences in parasite burdens in the lymph nodes and within the skin lesions were observed between Bc and C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice (Student's t test, P > 0.05 for all comparisons).
We have previously shown that the ability to heal L. major lesions is linked to the ability to heal a noninfectious wound (30). Thus, the kinetics of noninfectious wound healing in the C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) subcongenic lines were examined using reepithelialization following punch biopsy as a measure of wound healing capability. Comparisons of the average rates of wound healing showed that B6 mice healed their in vivo wounds significantly faster than Bc mice (P = 0.0001). Compared to Bc mice, C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice also healed their wounds at significantly higher rates of speed (P = 0.001, P = 0.001, and P = 0.001, respectively), demonstrating, at this level of resolution, a complete association between L. major resistance and the ability to heal a skin wound (Fig. 2j, k, and l). As with L. major infection, the association of the 5-Mb interval with wound healing was further supported by the slower-healing phenotype observed in C.B6-(lmr2.1), C.B6-(lmr2.2), C.B6-(lmr2.5), C.B6-(lmr2.4a), and C.B6-(lmr2.4d) mice (Fig. 3f to j), which all carry the susceptible Bc haplotype across this interval.
The C.B6-(lmr2.4b) and C.B6-(lmr2.4c) congenic lines carry a smaller congenic interval than C.B6-(lmr2.4) mice, and thus, these two congenic lines were utilized for the following analyses. Wound healing is highly temporally coordinated. It begins with an initial inflammatory phase characterized by the recruitment of neutrophils, macrophages, and lymphocytes, followed by a fibroblastic phase characterized by the recruitment of fibroblasts and deposition/organization of collagen fibrils. Persistence of the inflammatory phase during wound repair impairs healing, inhibits wound resolution, and promotes chronicity. Thus, the density and type of inflammatory cells at the wound site were analyzed in H&E-stained skin sections from B6, Bc, C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice at days 0, 7, 9, and 11 after wound induction. At day zero in naïve tissue, all three strains showed identical cell population profiles, with small numbers of inflammatory cells present. As wound closure progressed, the ratio of inflammatory cells to fibroblasts changed in a mouse strain-dependent manner. At day 9, the density of fibroblasts was higher in B6, C.B6-(lmr2.4b), and C.B6-(lmr2.4c) tissue, but equal numbers of inflammatory cells and fibroblasts were observed in Bc tissue (Fig. 4a and c). Thus, at day 9, inflammatory responses had ceased in C.B6-(lmr2.4b) and C.B6-(lmr2.4c) mice but not in Bc mice. Interestingly, the rapid progression from the inflammatory phase to the fibroblastic phase in resistant mice [C.B6-(lmr2.4b), C.B6-(lmr2.4c), and B6 mice] corresponds with the enhanced wound healing and L. major lesion resolution observed in these mice. Differences in the types of inflammatory cells recruited to the wound site were not observed.
We next examined collagen deposition in wounded skin. A dense and organized pattern of collagen fibrils is considered a positive marker of healing, whereas a sparse and disorganized deposition of collagen fibrils is deemed a marker of a poor healing process. Sections of punch biopsy-wounded skin from B6, Bc, C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice at days 0, 7, 9, and 11 after biopsy were therefore examined for their collagen profiles. At day zero in normal tissue, collagen was evenly distributed, showed great depth, and was aligned in an ordered parallel configuration, with fibroblasts being the only cell type present. At day 9, B6, C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice presented more organized and abundant collagen bundles than Bc mice (Fig. 4b and d). Thus, at day 9, C.B6-(lmr2.4b) and C.B6-(lmr2.4c) mice presented collagen profiles that were more similar to the collagen profile of the genetic donor strain B6 than to that of the acceptor strain Bc, indicating that genes within the congenic interval controlled this process. As with the inflammatory cell profile, the abundant and organized deposition of collagen fibrils in resistant mice [C.B6-(lmr2.4b), C.B6-(lmr2.4c), and B6 mice] corresponds with the enhanced wound healing and L. major lesion resolution observed in these mice.
Using the degree of collagen deposition and organization as a tool to assess the efficiency of local healing, the strains can be ordered from the worst to the best healer as follows: Bc → C.B6-(lmr2.4b) → C.B6-(lmr2.4c) → B6. This ranking mirrors the ranking for the degree of resistance to L. major infection as measured by lesion size, as well as the ability to heal in vivo noninfectious wounds as measured by wound size and inflammatory cell profile.
Macrophages play a vital role during wound healing, and during L. major infection, they serve as both the host cell and the destroying effector cell for L. major (15, 23). Moreover, during wound healing, macrophages secrete factors important for wound closure, fibroblast recruitment and activation, and collagen deposition/organization (11, 33, 36). Thus, to identify candidate genes involved in the differential wound healing and L. major disease phenotype observed between congenic and parental mice, we performed a microarray analysis of in vitro-infected and uninfected bone marrow-derived macrophages from Bc, B6, and C.B6-(lmr1, lmr2) mice (GEO accession number GSE17550). We used macrophages from C.B6-(lmr1, lmr2) rather than C.B6-(lmr2.4b) or C.B6-(lmr2.4c) mice to identify host genetic factors within the other lmr loci (lmr1 and lmr3) that were contributing to the differential L. major response observed between the Bc strain and its congenic strains (29). However, only results for the lmr2 locus will be discussed. From over 100 differentially expressed genes identified, 12 were located within the lmr2 locus (Mb 0 to 50, chromosome 9) (Table (Table1).1). Of these, only Fli1 (Mb 32.17 to 32.29, chromosome 9) is present within the 5-Mb candidate interval (Mb 29 to 34, chromosome 9) and is differentially expressed in both C.B6-(lmr1, lmr2) and B6 macrophages relative to its expression in Bc macrophages, thus highlighting its importance in the response to infection. In our mice, the Fli1 transcriptional response to infection, measured by its relative expression levels in infected and uninfected macrophages, was decreased by more than 3-fold in mice carrying the B6 Fli1 allele [B6 and C.B6-(lmr1, lmr2) mice] relative to its expression level in Bc mice. Differential expression of Fli1 in infected and uninfected macrophages was validated independently with quantitative real-time PCR (Fig. (Fig.5a5a).
Perusal of the C.B6-(lmr2.4b) and C.B6-(lmr2.4c) subcongenic lines identified C.B6-(lmr2.4c) as the line with the smallest donor interval and most resistant phenotype, and thus, it was chosen for the experiments described below. To verify the importance of Fli1 in resistance to infection, we measured the gene expression pattern of Fli1 in an in vivo infection system. Fli1 expression in inguinal lymph nodes from naïve, week-6-postinfection (peak of infection), and week-12-postinfection (end of infection) Bc, C.B6-(lmr2.4c), and B6 mice was measured. The inguinal lymph nodes were chosen as the site of analysis because they are the main foci of the immune responses against L. major during infection. Measurement of Fli1 expression in inguinal lymph nodes presented a strain-specific trend in Fli1 expression. Fli1 expression in B6 and C.B6-(lmr2.4c) mice peaked at week 6 postinfection, with significant differences observed between these resistant strains and Bc mice [for Bc versus B6 mice, P = 0.0006, and for Bc versus C.B6-(lmr2.4c) mice, P = 0.002] (Fig. (Fig.5b).5b). In contrast, Fli1 expression in Bc mice was highest at week 12 postinfection. Thus, in resistant mice [B6 and C.B6-(lmr2.4c)], the peak in Fli1 expression in the lymph nodes at week 6 corresponds with the observed phenotypic peak in infection at week 6, in which B6 mice are observed to heal their lesions and lesions in C.B6-(lmr2.4c) mice stabilize. In contrast, lesions in Bc mice at week 6 postinfection continue to increase in size and severity, possibly due to the absence of an increase in Fli1 expression.
We next investigated Fli1 expression locally in the skin, the site of L. major lesions and skin wounds. Analysis of Fli1 expression in L. major-infected skin samples and punch biopsy-wounded skin samples also showed a strain-specific trend in Fli1 expression. B6 and C.B6-(lmr2.4c) mice, which both carry the resistant B6 Fli1 allele, present significantly higher levels of Fli1 in naïve skin than Bc mice [for Bc versus B6 mice, P = 0.0003, and for Bc versus C.B6-(lmr2.4c) mice, P = 0.007]. Fli1 expression in infected skin of B6 and C.B6-(lmr2.4c) mice subsequently decreased throughout the infection period, reaching its lowest at week 12 with significantly lower expression levels observed than in the skin of Bc mice [for Bc versus B6 mice, P = 0.0017, and for Bc versus C.B6-(lmr2.4c) mice, P = 0.33]. In contrast to B6 and C.B6-(lmr2.4c) mice, Fli1 expression in infected skin of Bc mice increased initially and remained constant during the infection period (Fig. (Fig.5c5c).
Similarly, the same trend in Fli1 expression is observed in punch biopsy-wounded skin samples (Fig. (Fig.5d).5d). B6 and C.B6-(lmr2.4c) mice presented significantly higher levels of Fli1 in naïve skin than Bc mice [for Bc versus B6 mice, P = 0.0004, and for Bc versus C.B6-(lmr2.4c) mice, P = 0.0127]. Fli1 expression at this time point was significantly higher in mice carrying the resistant B6 Fli1 allele [B6 and C.B6-(lmr2.4c) mice]. Fli1 expression in the skin of B6 and C.B6-(lmr2.4c) mice subsequently decreased throughout the wound healing period, reaching its lowest at day 11, with significantly lower expression levels observed at day 11 than in wounded skin of Bc mice [for Bc versus B6 mice, P = 0.0186, and for Bc versus C.B6-(lmr2.4c) mice, P = 0.0206]. In contrast to resistant mice, Fli1 expression in wounded skin of Bc mice was maintained at a constant level throughout the 11-day healing period.
We next examined the B6 Fli1 allele [B6 and C.B6-(lmr2.4c) mice] and the Bc Fli1 allele (Bc mice) for sequence polymorphisms that may contribute to the differential disease and gene expression profiles observed between Bc and C.B6-(lmr2.4c) mice. Sequence analysis did not identify any differences within the coding regions of Fli1; however, three polymorphisms in the noncoding regions were identified. A polymorphism in a GAn repeat element in exon 1 was observed. The length of the GAn repeat element was 28 GA repeats in the B6 Fli1 allele and 29 GA repeats in the Bc Fli1 allele (Fig. (Fig.6a).6a). Similarly, a GTn repeat element in the 3′ untranslated region (3′UTR) was also identified (Fig. (Fig.6c).6c). In the B6 Fli1 allele, the length of this element was 22 repeats, and in the Bc Fli1 allele, it was 21 repeats. In addition, a TT deletion event in the 3′UTR of the Bc Fli1 allele was also observed (Fig. (Fig.6b).6b). The polymorphisms were verified with resequencing. Any of the three polymorphisms identified may be the cause of the differential disease and gene expression profiles observed between mice carrying the B6 or Bc Fli1 allele; however, of the three polymorphisms observed, the GAn repeat element is of the most interest. Unlike the GTn repeat element and TT deletion event, the GAn repeat element is 100% conserved between mice and humans and is a known endogenous enhancer of Fli1 expression; the length of this repeat is inversely correlated with Fli1 promoter activity (25).
The mouse model for cutaneous leishmaniasis has been extensively used to elucidate those aspects of the disease which cannot be addressed in humans. This approach has been based on the observation that as a zoonosis in which rodents are the reservoir hosts, Leishmania species infective in humans are also infective in mice. Furthermore, the disease progression in mice often imitates the pattern of lesion development observed in humans and differences in the genetic basis of leishmaniasis in the outbred human population parallel genetic differences observed between inbred mouse strains. This observation thus inspired us to map loci affecting the host response to L. major in mice in the hope that such loci/genes will shed some new light on the mechanisms responsible for host protection in mice and humans. A genome-wide simple sequence length polymorphism (SSLP) scan on the F2 progeny of a C57BL/6 × BALB/c cross identified three loci, lmr1, lmr2, and lmr3, on chromosomes 17, 9, and X, respectively, mediating the host response to L. major (29). This report focuses on the fine mapping of the lmr2 locus.
Through the generation of subcongenic mice and the utilization of lesion scores and rates of wound healing as markers of resistance and susceptibility, we fine mapped the lmr2 loci to a 5-Mb candidate interval in strains C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c). The C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) congenic lines demonstrated a phenotype that was intermediate between the B6 and Bc phenotypes, suggesting the involvement of other loci in L. major resistance. However, the presence of the B6 allele of the 5-Mb candidate interval on the susceptible Bc background of C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice conferred a degree of resistance on an otherwise susceptible strain, highlighting the significant contribution of this interval during L. major infection. Resistance to L. major in the C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) congenic lines was independent of the ability to control parasite numbers, as no significant differences in parasite burdens of the skin and lymph nodes were observed. However, a trend was observed in parasite burdens. C.B6-(lmr2.4), C.B6-(lmr2.4b), and C.B6-(lmr2.4c) mice, which all present an L. major resistance phenotype that is intermediate between the B6 and Bc phenotypes, also present parasite burden levels in week-12-postinfection lymph nodes and skin that are intermediate between those of B6 and Bc mice. Thus, control of lesion growth by the 5-Mb candidate interval may be dependent on the control of parasitemia; however, its effects are minimal and other loci, such as lmr1 and lmr3, may have a more decisive role in this process.
The 5-Mb candidate interval was also associated with an enhanced wound healing phenotype. Enhanced wound healing in C.B6-(lmr2.4b) and C.B6-(lmr2.4c) mice was correlated with abundant collagen deposition and organization and a more rapid inflammatory process than was observed in the background strain Bc. Twenty-one genes are present within this 5-Mb candidate interval, and microarray profiling identified the differential expression of only 1 gene—Fli1. Fli1 is a member of the ETS family of transcription factors, known for their roles in hematopoiesis, vasculogenesis, and embryonic development (7, 18). It acts as a transcriptional activator/repressor in several pathways, is highly expressed in hematopoietic tissues and cells, including the thymus and spleen, and is involved in the maturation of T cells and B cells (16, 20, 27) and, thus, may have an important role to play in the inflammatory response to L. major. Among its various functions, recent studies have shown that Fli1 controls the expression of collagen through its inhibition of connective tissue growth factor expression and, thus, may function in the wound healing pathway (24).
In the microarray analysis, the Fli1 transcriptional response to infection, measured by its relative expression levels in infected and uninfected macrophages, was decreased by more than 3-fold in mice carrying the B6 Fli1 allele. Interestingly, this downregulation of Fli1 in infected macrophages from resistant mice corresponds with the downregulation of Fli1 observed in bone marrow-derived macrophages stimulated with inflammatory response mediators, such as lipopolysaccharide, gamma interferon, and retinoic acid (18). This suggests that downregulation of Fli1 expression is important for the normal function of macrophages against pathogens and that its reduction is required during inflammatory responses against L. major.
The differential expression of Fli1 and its importance during infection were verified in an in vivo infection model. Analysis of Fli1 expression in lymph nodes at week 0, week 6, and week 12 postinfection identified a strain-specific trend in Fli1 expression. Resistant mice [B6 and C.B6-(lmr2.4c) mice] presented a significant increase in Fli1 expression at week 6 postinfection. This increase in expression corresponds with the phenotypic peak of infection at week 6, in which B6 mice begin healing their lesions and lesions in C.B6-(lmr2.4c) mice stabilize. Interestingly, the upregulation of Fli1 expression in the lymph node does not suggest immediate downregulation of Fli1 as would be expected from microarray profiling of macrophages. However, this difference is not unexpected since the lymph node is home to a variety of cell types, such as dendritic cells, neutrophils, B cells, T cells, and macrophages. Of note, upregulation of Fli1 is observed during B cell proliferation/maturation (2), while downregulation of Fli1 is observed during macrophage stimulation (18). Thus, the upregulation of Fli1 observed in infected lymph nodes from B6 and C.B6-(lmr2.4c) mice during the peak of infection corresponds with the increased level of Fli1 observed during B cell proliferation/maturation, a vital process required during immune responses against L. major (17).
Differential Fli1 expression was further validated in L. major-infected skin and punch biopsy-wounded skin. In both experimental assays, a trend in Fli1 expression is observed. B6 and C.B6-(lmr2.4c) mice initially present with significantly high levels of Fli1 which decrease slowly during infectious and noninfectious wound healing. In contrast, Bc mice present with relatively constant levels of Fli1 during the infection and wound healing period. The downregulation of Fli1 expression in the skin of C.B6-(lmr2.4c) and B6 mice indicates that downregulation of Fli1 is required for enhanced lesion resolution and wound healing. It has been previously observed that the downregulation of Fli1 is essential for collagen expression and deposition (9), and thus, the downregulation of Fli1 in the skin of B6 and C.B6-(lmr2.4c) mice during infection and wound healing corresponds with the downregulation of Fli1 required for collagen expression and deposition. Furthermore, the downregulation of Fli1 in the skin of B6 and C.B6-(lmr2.4c) mice correlates with the more abundant and organized collagen deposition observed in the skin of B6 and C.B6-(lmr2.4c) mice. Interestingly, the pattern of Fli1 expression in infected lymph nodes, infected skin, and wounded skin from C.B6-(lmr2.4c) congenic mice is more similar to its pattern in the lmr2 donor strain B6 than to its pattern in the acceptor strain Bc. This observation highlights the contribution and importance of the B6 lmr2/Fli1 allele during L. major infection and wound healing.
Sequence analysis of the B6 and Bc Fli1 allele identified three polymorphic regions. Of the three polymorphisms observed, the GAn repeat element is of the most interest. The repeat element is 100% conserved between mice and humans and has the most relevance in terms of the differential Fli1 expression observed between Bc and C.B6-(lmr2.4c) mice. The GAn repeat element in exon 1 is located directly adjacent to Fli1 regulatory elements, specifically, a GATA/EBS dual element known to act as a transcriptional enhancer (5, 6, 25, 26). The GAn repeat element is a known endogenous enhancer of Fli1, and the length of this repeat is inversely correlated with Fli1 promoter activity, i.e., the longer the repeat, the lower the promoter activity and gene expression (25). In B6 mice, the length of the GAn element is 28 GA repeats, and in Bc mice, it is 29 GA repeats. Hence, the Fli1 GAn repeat element is polymorphic between B6 and Bc mice, with a difference of 1 GA repeat. This difference in the length of the GAn repeat elements in B6 and Bc mice may affect the binding of the ETS/GATA factors to the EBS/GATA dual element and could be the cause of the differential Fli1 expression observed in the skin, lymph nodes, and macrophages, as well as the increased L. major resistance and enhanced wound healing observed in mice carrying the B6 Fli1 allele [C.B6-(lmr2.4c) and B6 mice]. The role and relevance of a single dinucleotide polymorphism in modulating the binding of ETS/GATA factors to the EBS/GATA dual element in Fli1 and the resultant differential expression of Fli1 will require verification; however, it should be noted that polymorphisms in dinucleotide microsatellites are known to mediate the host response to intracellular microorganisms. For example, susceptibility to infection with Leishmania donovani, Mycobacterium avium, Mycobacterium leprae, and Mycobacterium tuberculosis has been linked to a 5′(CA) microsatellite polymorphism in the NRAMP1 promoter (3, 8, 14, 22).
The elucidation of the role of Fli1 during infection is important not only for our understanding of the host response to L. major but also for understanding the host response to Salmonella enterica serovar Typhimurium and Mycobacterium tuberculosis. Linkage analysis of resistance to S. Typhimurium in mice as measured by antibody response was mapped to a locus overlying the lmr2 locus. The linkage peak for this quantitative trait locus mapped approximately between Mb 27 and 70 on mouse chromosome 9 (10). Additionally, linkage analysis of resistance to M. tuberculosis in mice as measured by survival time and loss of body weight was also mapped to a locus, tbs2 (tuberculosis severity 2), overlying the lmr2 locus. The linkage peak for the tbs2 locus mapped approximately between Mb 26 and 44 on mouse chromosome 9 (19, 32). Fli1 is located between Mb 32.2 and 32.4 on mouse chromosome 9. Thus, the involvement of Fli1 in the host response to L. major and the mapping of S. Typhimurium and M. tuberculosis host response loci to the same region as the lmr2 locus and Fli1 present a convincing case for the potential and more general involvement of the lmr2 locus and, in particular, Fli1 in the host response to other intracellular organisms.
We thank Cameron Nowell, Lynn Buckingham, Chrystal Smith, and Michael Durrant for technical assistance and Axel Kallies and Stephen Nutt for reagents.
The work was supported by grants from the U.S. National Institutes of Health and the Australian National Health and Medical Research Council.
The authors have no conflicting financial interests.
Editor: J. H. Adams
Published ahead of print on 5 April 2010.