Mouse models are widely used for studying polygenic traits underlying various diseases to identify the low penetrance disease susceptibility genes. Classical genetic studies using cross breeding of mouse strains with differing susceptibilities to different diseases have identified chromosomal regions associated with predisposition to lung tumours, lung injury, disease resistance, and seizure.^1,2 The process, called quantitative trait loci (QTL) mapping, involves the use of evenly spaced polymorphic DNA markers to create landmarks across each chromosome to find correlations between marker alleles and the phenotypic variation.³ Inbred mouse strains vary in their susceptibility to lung tumours, lung injury, disease resistance, and seizures, with two extreme strains being A/J (susceptible) and C57BL/6J (B6, resistant).^1,2 As the mouse genome sequences for both the A/J and B6 strains of mice have been completed, identification of candidate genes responsible for disease related QTL, mapped using these two strains of mice as parental strains, becomes feasible by undertaking single nucleotide polymorphism (SNP) analysis.

One set of mouse QTL is associated with genetic predisposition to lung tumour. In humans, the wide variety of carcinogens and varying degrees of exposure make identifying the predisposing genes difficult, but in a mouse model, such confounding variables can be controlled. There is evidence that, although lung cancer incidence is largely associated with environmental factors such as smoking or occupational exposure, genetic components are involved in lung cancer development. Genetic linkage studies using various strains of inbred mice mapped pulmonary adenoma susceptibility (Pas), pulmonary adenoma resistance (Par), and susceptibility to lung cancer (Sluc) loci.¹ Recently, we have mapped three major QTL, named mouse lung tumour (MLT) 1–3, in 25 strains of mice with known susceptibility to lung cancer development, using a whole genome LD analysis with 5638 genetic markers.⁴ MLT1 locates between D4Mit50 and D4Mit361, MLT2 between D6Mit15 and D6Mit304, and MLT3 from D8Mit127 to D8Mit69. SNP analysis for candidate genes at these three lung tumour related loci was an objective of the present study.

The second set of mouse QTL is associated with genetic susceptibility to acute lung injury. Acute lung injury is a syndrome characterised by increased alveolar–capillary permeability and hypoxaemia. The most severe cases of acute lung injury culminate in respiratory distress, a condition that remains a serious clinical challenge in critical care medicine because of its high incidence and substantial mortality.^5,6 Two extreme inbred mice strains—A/J (susceptible) and C57BL/6J (resistant)—vary in their susceptibility to acute lung injury.^7,8 Four QTL on mouse chromosome 11, 13, 17, and 6 were identified as acute lung injury susceptibility loci, and were designated as Ali1, Ali2, Ali3, and Ali4, respectively.^8,9 SNP analyses between A/J and C57BL/6J mice for candidate genes in these regions were conducted in this study.

The third set of mouse QTL is associated with genetic resistance to trypanosomiasis. Trypanosomiasis, or sleeping sickness, is a re­emerging public health problem of epidemic proportions in many parts of rural Africa. It is caused by a subspecies of Trypanosoma brucei and is transmitted by tsetse flies.¹⁰ Multigene control of variation in susceptibility to trypanosomiasis has been mapped in mice, where the C57BL/6J strain is relatively resistant and the A/J strain is susceptible.^11,12 Kemp et al reported three QTL of trypanosomiasis resistance located on mouse chromosomes 17, 5, and 1 using the intercross of resistant C57BL/6J mice and susceptible A/J or BALB/c mice.¹³ The loci were named Tir1, Tir2, and Tir3, respectively. Among these three QTLs, Tir1 and Tir2 were mapped using the intercross of A/J strain and C57BL/6J strain, whereas Tir3 was mapped using the intercross of C57BL/6J strain and BALB/c strain. Later, the Tir3 locus located on mouse chromosomes 1 was fine mapped by using G6 generation of advanced intercross lines between A/J and C57BL/6J mice.¹⁴ On mouse chromosomes 1, they found three QTL peaks which were named Tir3a, Tir3b, and Tir3c.¹⁴ These regions are examined for potential candidates in the present study.

The fourth set of mouse QTL is associated with genetic susceptibility to seizures. A seizure is a brief change in behaviour; it is the uncontrolled hypersynchronous electrical discharges of neurones in the brain that interfere with normal function.¹⁵ Epilepsy is characterised by recurrent seizures and refers to a collection of disorders that affect 1–2% of the world’s population.^16,17 Epilepsy genes fall into several quite distinct classes including those in which mutations cause abnormal brain development, progressive neurodegeneration, disturbed energy metabolism, or dysfunction of ion channels.¹⁸ To define the genetic contributions affecting individual differences in seizure threshold, Gershenfeld et al mapped quantitative trait loci for seizures induced by the intraperitoneal injection of methyl-β-carboline-3-carboxylate (β-CCM), a GABA_A(γ-aminobutyric acid) receptor inverse agonist and convulsant.¹⁹ By using backcross and intercross populations between susceptible A/J strains and resistant C57BL/6J strain, they reported that the QTLs of seizure susceptibility loci were located on mouse chromosomes 10, 4, and 7, and were designated as Exq1, Bis1, and Bis4, respectively. In this study, these regions are further characterised for candidate genes using SNP analysis.

Genomic DNA contains a variety of polymorphisms, such as single nucleotide substitutions, insertion/deletions, and nucleotide repeat motifs. SNPs are changes in a single base at a specific position in the genome, in most cases with two alleles.²⁰ They are useful as marker in population genetics and evolutionary studies.^21,22 Depending on where an SNP occurs, it may have different consequences at the phenotypic level. SNPs in the regulatory regions of genes might influence the risk of common disease. SNPs in the 3′-UTR region may alter the stability of the mRNA by changing binding sites or secondary structure, thus making it more or less likely to be degraded. An SNP in the 5′ region may change binding sites and thereby modify the affinity for a transcription factor. Non-sense SNPs can introduce a premature stop-codon resulting in a truncated polypeptide, and often result in a loss of function. Missense mutations cause amino acid alterations that can be significant if the properties of the new amino acid (charge, polarity, and so on) are different from the one it replaced.

Over the past few years, SNPs have been proposed as the next generation of markers for the identification of genes associated with complex diseases.^23–25 As A/J and C57BL/6J mice vary significantly in their susceptibility to lung tumour development, lung injury, disease resistance, and seizures, detailed SNP analysis between A/J and C57BL/6J mice is ideal for identifying potential candidate genes for relevant QTLs. This approach has been aided greatly by the fact that the genome sequences of both strains have been sequenced and assembled.

METHODS

Approach

Markers flanking each QTL were found in the Celera (Rockville, Maryland, USA) mouse genome database (www.celera.com). These markers were used to identify the reference DNA positions within the assembled genome for our database query. The Celera mouse SNP reference database was interrogated for SNPs within each QTL and then filtered to keep only SNPs between A/J and C57BL/6J mice (A-B SNPs).

QTL associated with lung tumour development

Using linkage disequilibrium (LD) analysis of the whole genome, employing 5658 genetic markers in 26 strains of mice, we previously mapped three major loci: MLT1 near D4Mit2, MLT2 near D6Mit26, and MLT3 near D8Mit205, all showing highly significant LDs.⁴ As a gradual decrease in LD surrounding each of the peak markers (D4Mit2, D6Mit26, and D8Mit205) was evident, we selected the following MLT candidate regions for SNP analysis: MLT1, located between D4Mit50 at 5.4 cMto D4Mit361 at 6.75 cM; MLT2, limited between D6Mit15 at 74 cM and D6Mit304 at 75 cM; and MLT3, the region from D8Mit127 at 29 cM to D8Mit69 at 31 cM (fig 1A).

QTL associated with acute lung injury

As shown in fig 2A, Ali1 is located from D11Mit245 at 44.8 cM to D11Mit10 at 63 cM; Ali2 is located from D13Mit236 at 4 cM to D13Mit209 at 37 cM; Ali3 is located from D17Mit218 at 43 cM to D17Mit76 at 54.6cM; and Ali4 is located from D6Mit352 at 20.4cM to D6Mit188 at 32.5 cM (fig 2A).^8,9

QTL associated with resistance to trypanosomiasis

In Tir1 and Tir2 we selected DNA makers relative to LOD scores of 6.0 and 3.0, respectively,¹³ with a focus on regions from 16.4 cM to 29.4 cM and from 41 cM to 57 cM. In Tir3, we selected those makers within about 5 cM around, respectively, Tir3a, Tir3b, and Tir3c when the LOD score was over 4.0.¹⁴ The regions are from 58.7 cM to 74.3 cM (in Tir3a and Tir3b) and from 91.3 cM to 101 cM (in Tir3c)(fig 3A).

QTL associated with seizure

In Exq1, we combined the two QTLs on mouse chromosome 10 of the intercross F2 population and the backcross N2 population and selected those makers with a LOD score more than 3.0 for regions from D10Mit68 at 51.5 cM to D10Mit297 at 70 cM. In QTL Bis1 derived from backcross N2 population, we selected regions from D4Mit204 at 61.9 cM to D4Mit127 at 77.5 cM with a LOD score over 2.2. In QTL Bis4 derived from intercross F2 population, we selected those DNA makers with a LOD score over 3.2 for the region from D7Mit117 at 11 cM to D7Mit145 at 26.4 cM (fig 4A).

Data filtering

Additional filtering of the A-B SNPs was done to increase confidence in the positive results. With screening, we sought to reduce the false negative rate and thereby to be inclusive, but owing to the tremendous number of SNPs (table 1) we needed also to minimise the false positive rate. To accomplish this, A-B SNPs that met the following criteria were selected for further analysis:(1) A/J and C57BL/6J that were polymorphic for the SNP;(2) SNPs appearing in the coding region, the 5′-regulatory region, or the 3′-untranslated region;(3) SNPs appearing in known genes;(4) exclusion of intronic, intergenic, and silent SNPs. We focused on those SNPs within the coding region, although there is a downside of possibly filtering out some genes responsible for the QTLs. This was done for two reasons: first, the function of the intronic and intergenic sequence is a matter of research beyond the scope of this study; second, the number of these SNPs made their analysis currently untenable. For the resulting A-B SNPs in genes, we chose those genes that might be a plausible link to the aetiology or pathophysiology of mouse lung tumour development, acute lung injury, trypanosomiasis, and seizures, respectively, for further analysis by undertaking sequence verification.

SNP annotation

SNP IDs were used to link SNPs with genes in which they occur and Celera’s mechanism to link its genes with public sequences was used for reporting public accession numbers and description from the Celera database and/or GenBank.

Sequence verification of SNPs

A-B SNPs were sequence verified. Briefly, the sequence for the candidate genes was downloaded from the Celera database and polymerase chain reaction (PCR) primers flanking the SNPs designed such that approximately 200 base pair (bp) fragments were produced. PCR was done to amplify DNA harvested from normal lungs of A/J and C57BL/6J mice. PCR products were resolved on 1.2% ethidium bromide stained agarose gels and purified using QIAquick gel extraction kits (Qiagen, Hilden, Germany). Automated sequencing was undertaken using dideoxy terminator cycle sequencing kits (Applied Biosystems, Foster City, California, USA) and Applied Biosystems model 377 DNA sequencers (Perkin-Elmer, Foster City).

Confirmation of candidate genes

After sequencing analysis, we confirmed the SNPs from A/J and C57BL/6J mice by using software of SEQUENCHER (version 4.0.5, Gene Codes Corporation, Ann Arbor, Michigan, USA). The confirmed SNPs were compared with those public sequences of known genes in GenBank, correspondingly using online software of “BLAST 2 SEQUENCES”(http://www.ncbi.nlm.nih.gov/blast/bl2seq/bl2.html). The reading frame was established and suitable candidates were those where the SNP could result in an amino acid alteration and where there was a plausible link to the aetiology or pathology of each disease.

Confirmation of potential functional domains

After the confirmation of amino acid alteration in candidates, the codon positions were checked with NCBI Conserved Domain Database by using the amino acid sequence of candidates (http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml). When the region of amino acid changed codon matched a known functional conserved domain, the domain was reported. If the region matched several known conserved domains, the highest score value domain was reported.

RESULTS

Candidate genes for lung tumour QTL

In this study, all nine confirmed SNPs in nine different genes within the MLT1 locus were of missense type. The human cytogenetic positions of these SNPs in MLT1 were located at 8q and 6q, respectively (table 2). Two of the candidates for MLT1 are Cdh17 and GPR63. Cdh17, locating at 8q22.1—also known as liver-intestine (LI) cadherin—represents a novel member of the cadherin superfamily. The amino acid at codon 644 was changed from TAC (tyrosine) in A/J mice to TCC (serine) in C57BL/6J mice for the Cdh17 gene (fig 1, panels B and C;table 2). GPR63 (G protein coupled receptor 63, located at human cytogenetic 6q16.1–q16.3)—also known as PSP24B—shares 57% identity with Xenopus PSP24.²⁶ At codon 407 of the GPR63 gene, the amino acid of AGT (serine) in A/J mice was substituted by AGG (arginine) in C57BL/6J mice (fig 1C;table 2). SNPs confirmed in MLT2 were located on human cytogenetic 12p11–p12 (table 2). LOC272322 is annotated as BMAL2 gene (brain-muscle ARNT-like protein 2). In BMAL2, the missense polymorphisms gave rise to amino acid change at codon 551 (ACA/ATA)(table 2). At codon 551, threonine in A/J mice was changed to isoleucine in C57BL/6J mice (fig 1, panels B and C). Another gene in MLT2 is Mrps35 (mitochondrial ribosomal protein S35). In MLT3, we found only one SNP in ADAM29 (a disintegrin and metalloproteinase domain 29). Overall, we identified 13 candidate genes for mouse MLT QTLs.

Candidate genes for acute lung injury QTL

As shown in table1, a total of 178 SNPs in 88 genes was found and those genes with potential functional association with acute lung injury were chosen for sequence verification. However, only 24 SNPs in 22 genes were confirmed by re-sequencing (table 3).

In the Ali1 locus, 15 SNPs in 13 genes have been sequence verified. Of the 13 genes, five—RAD51l3 (RAD51-like 3)(fig 2, panels B and C), Slfn2 (schlafen 2), Mpo (myeloperoxidase), Akap (A kinase anchor protein), and Map3k14 (mitogen activated protein kinase kinase kinase 14)—were considered candidates because of their likely link to acute lung injury (table 3).

In Ali2, four SNPs in four genes were identified. After comparing with the public sequences of known genes and excluding those without a plausible link to acute lung injury, two genes—Fgfr4 (fibroblast growth factor receptor 4) and Gprk6 (G protein coupled receptor kinase 6)—were considered candidates (table 3;fig 2C).

In Ali3, four SNPs were found in three genes. Among these, SNP mCV23088765 and mCV23090856 were found in Ltbp1 (latent transforming growth factor binding protein 1). SNP mCV23088765 could not be confirmed because of the extremely high GC content of the flanking sequence, but SNP mCV23090856 was confirmed. The SNP results in a valine alteration in A/J mice and a methionine alteration in C57BL/6J mice (fig 2C). The other candidate gene in this Ali3 locus is Xdh (xanthine dehydrogenase)(table 3).

In the Ali4 locus, only two SNPs were identified in two genes—Imap38 (immunity associated protein, 38 kDa) and Fin15 (fibroblast growth factor inducible 15)(fig 2, panels B and C); these were confirmed and are regarded as the candidate genes for acute lung injury (table 3).

Candidate genes for trypanosomiasis QTL

Among the 83 SNPs in 49 genes filtered, 32 SNPs in 26 genes were sequence verified. The 32 A-B SNPs are enumerated in table 4. In QTLs of Tir1 and Tir3, SNPs in the coding region between A/J and C57BL/6J mice were of missense type. There were no SNPs found in the coding regions in known genes except three silent mutations in Tir2 (table 4). Representative amino acid alterations are presented in fig 3, panels B and C. In the Tir1 locus, 21 SNPs in 16 genes were sequence verified. After sequence verification and examination of the open reading frames, 10 genes—Daxx, Tapbp, Abcb2, Abcb3, Psmb8, H2-Eb1, Bat2, Ltb, Rnf23, and Notch4—were considered as the candidate genes (table 4). On mouse chromosome 1, three QTLs (Tir3a, b, c) are responsible for mouse resistance to trypanosomiasis.¹⁴ We categorised these as two continuous regions on chromosome 1 as shown in fig 3A. In this locus, we verified a total of eight SNPs in eight genes, and five genes—Ddr2, Fcgr2b, Ly9, CD84, and Fcer1a—were identified as the candidates (table 4).

Candidate genes for seizure QTL

After filtering the SNPs, 33 SNPs in 23 genes were identified. Twenty one SNPs in 16 genes were confirmed. Among these 16 genes, those related to seizure were considered as the candidates. The A-B SNPs listed in table 5 are filtered from the first query in the Celera database with the criteria described above. In Exq1, Bis1, and Bis4, SNPs in the coding regions between A/J and C57BL/6J mice were of mis-sense type as shown in table 5. In the Exq1 locus, four SNPs in four genes were sequence verified. After examination of the open reading frames, two genes (Lta4h and Iltif) were considered as candidates because of their potential link to seizure (table 5;fig 4C). In the Bis1 locus, 15 SNPs in 10 genes were sequence verified. Five genes (Cnr2, Htr1d, C1qb, Hspg2, and Kif17) were considered as the candidates (table 5). In the Bis4 locus, two SNPs in two genes, Ccne1 and Otog, were confirmed (table 5). However, the relation between these two genes and seizures is not clear at present.

After confirmation of the amino acid alteration and the position of the changed codon in each candidate gene, the changed codon regions were checked to see if they were within potential functional domains by using the NCBI conserved domain database. Most of the candidates for the four diseases described above were within potential functional domains (table 6).

DISCUSSION

In this study, we systematically analysed single nucleotide polymorphisms in QTLs for four different diseases: lung tumours, lung injury, trypanosomiasis, and seizures. These QTLs were mapped by using A/J and C57BL/6J strains of mice as parental strains. Using this approach, we identified several candidates for each QTL (tables 2–5). The present study will facilitate the identification of modifier genes to lung tumour development, lung injury, trypanosomiasis, and seizures. The results from our study show the synergies between classical genetic studies and genomic tools such as the SNP database in dissecting the genetic basis of disease in mouse models.

Lung tumours

MLT QTL were mapped through LD analysis of the whole mouse genome using 5658 genetic markers in 25 strains of mice, resulting in three major QTLs: MLT1 (D4Mit2), MLT2 (D6Mit26), and MLT3 (D8Mit205)(fig 1A). We have identified 13 candidate genes for MLT1-3. Two candidates—Cdh17 and GPR63—were identified for the MLT1 locus. Cdh17, a liver-intestine (LI) cadherin, contains a tyrosine to serine change at codon 644. Expression of Cdh17 appears to correlate with tumour differentiation—for example, in ductal adenocarcinoma of the pancreas, LI-cadherin is highly expressed in well differentiated carcinoma but expression is reduced in less well differentiated areas and in poorly differentiated carcinomas.²⁷ Similarly, LI-cadherin is a marker for gastric intestinal metaplasia and well differentiated adenocarcinomas.²⁸ GPR63 (G protein coupled receptor 63) contains a valine to glycine change at codon 407. GPR63 (also known as PSP24B) shares 57% identity with Xenopus PSP24.²⁶ Xenopus PSP24 was originally identified as a functional lysophosphatidic acid (LPA) receptor. LPA may increases angiogenesis through vascular endothelial growth factor (VEGF), and also increases the level of cyclin D1 in ovarian cancer cells, increasing their proliferation.²⁹

MLT2 (from 74 cM to 75cM on chromosome 6) was very close to the previously mapped pulmonary adenoma susceptibility 1 locus (Pas1, located at 72.2 cM on chromosome), a major locus affecting inherited predisposition to the development of lung tumours. SNPs that were confirmed in MLT2 include LOC272322 and Mrps35 (mitochondrial ribosomal protein S35). The LOC272322, containing a threonine to isoleucine change at codon 551, is considered to be the BMAL2 gene (brain-muscle ARNT-like protein 2). It belongs to the bHLH-PAS superfamily.³⁰ Members of the bHLH-PAS family are transcription factors that contain a basic helix-loop-helix (bHLH) DNA recognition motif located N-terminal to a PAS domain and composed of two imperfect direct repeats. Several lines of evidence indicate that proteins in the bHLH-PAS superfamily are involved in the regulation of cell growth and differentiation. Also, the BMAL2 gene is downregulated in hepatocellular carcinoma, and overexpression of antisense BMAL2 RNA promotes cell proliferation.³⁰ Another confirmed known gene in MLT2 is Mrps35, which is reported to be overexpressed in testicular germ cell tumours.³¹

Only one SNP was verified for MLT3: ADAM29 (a disintegrin and metalloproteinase domain 29), containing an amino acid changing SNP in its coding sequence. Members of the ADAM family have been implicated in various important biological processes involving cell–cell and cell–matrix interactions, including fertilisation, muscle development, and neurogenesis. ADAM29 may function as a regulator of tumour metastasis.

Acute lung injury

In this study we identified several candidates for Ali1–4: Rad51I3, Mpo, Gprk6, and Xdh are candidates for mediating effects of early stage acute lung injury, while Fgfr4, Ltbp1, and Fin15 may play an important role in tissue repair after acute lung injury. We identified two novel candidates: Fgfr4 for Ali2 and Fin15 for Ali4.

Ali1 is located at 2 cM distal to D11Mit179⁸(fig 2A). We found 95 SNPs in 45 genes in Ali1. Among those genes that are likely to associate with acute lung injury, 15 SNPs in 13 genes were confirmed (table 3). Rad51l3 proteins—which belong to the Rad51 paralogs—are required for homologous recombinational repair (HRR) in vertebrates.³² Purified RAD51l3 protein possesses single stranded DNA binding activity and DNA stimulated ATPase activity.³³ Thus this gene may play an important role in the repair of DNA damage induced by chemical irritants in acute lung injury. Mpo (myeloperoxidase) is released from cytoplasmic granules of neutrophils and monocytes by a degranulation process, and reacts with the H₂O₂ formed by the respiratory burst to form a complex that can oxidise a large variety of substances.³⁴ It is expressed in neutrophils recruited to the lung after chemical or immunological insults. Akap was considered to be a “transduceosome” by acting as an autonomous multivalent scaffold that assembles and integrates signals derived from multiple pathways.^35–37 Map3k (mitogen activated protein kinase kinase kinase) activates both the SEK1-JNK and MKK3/6-p38 MAP kinase pathways and constitutes a pivotal signalling pathway in cytokine and stress induced apoptosis.³⁸ Slfn is a new family of growth regulatory genes that affect thymocyte development and guide both cell growth and T cell development.³⁹ As the genes of Rad51, Mpo, Akap, Map3k14, and Slfn2 have a plausible link with inflammation from multiple causes, they are considered candidates for acute lung injury.

Ali2 is located 2 cM distal to D13Mit59⁸(fig 2A), near the gene encoding the antioxidant enzyme glutathione peroxidase (Gpx), which catalyses the production of uric acid, a potent antioxidant found in airway secretions. Our results showed that the likely candidates are fibroblast growth factor receptor 4 (Fgfr4) and G protein coupled receptor kinase 6 (Gprk6). The fibroblast growth factors and the corresponding receptors are implicated in the regulation of epithelial cell proliferation and differentiation. Gprk6 is highly expressed in peripheral blood leucocytes and in some myeloid and lymphoid cell lines.^40–42 Gprk6 may be implicated in acute lung injury by regulating the inflammatory response.

Ali3 is located on chromosome 17⁸(fig 2A). Candidates for Ali3 include Xdh and Ltbp1. The expression of the Xdh gene is regulated in a cell specific manner and is markedly affected by inflammatory cytokines, steroids, and physiological events such as hypoxia.⁴³ Ltbp1 is latent transforming growth factor β binding protein 1. Transforming growth factor (TGF)β cytokines are a multifunctional family that exert a wide variety of effects on both normal and transformed mammalian cells, including cell invasion, tissue remodelling, and wound healing. Latency associated proteins and latent TGFβ binding proteins regulate the secretion and activation of this cytokine.⁴⁴ TGFβ has been most thoroughly studied during the late phases of tissue repair, where it plays a critical role in the development of pulmonary fibrosis.⁴⁵ However, in a recent study, expression levels of several TGFβ inducible genes were found to be dramatically increased as early as two days after the induction of injury.⁴⁶ It has been reported that pharmacological inhibition of TGFβ protected wild-type mice from pulmonary oedema induced by bleomycin or E coli endotoxin.⁴⁷ Thus Ltbp1 may function as a critical mediator in pulmonary repair after acute lung injury.

Ali4 is located on chromosome 6⁹(fig 2A). Candidates for this region are Imap38 (immunity associated protein, 38 kDa) and Fin15 (fibroblast growth factor inducible 15). Imap38 expression is found almost exclusively in the spleen, and Imap-like genes in mice are associated with proliferative and apoptotic events, suggesting a role in the control of cell death/survival.⁴⁸ The fibroblast growth factors (FGFs) play important roles in multiple physiological functions, including angiogenesis, mitogenesis, pattern formation, cellular differentiation, metabolic regulation, tissue repair, and oncogenesis. Although the function of Fin15 is unknown, the FGF-15 gene is expressed in a regionally restricted pattern in the developing nervous system, and may play an important part in regulating cell division and patterning within specific regions of the embryonic brain, spinal cord, and sensory organs.⁴⁹ Thus Fin15 may play a major role in tissue repair after acute lung injury.

Trypanosomiasis

We identified several candidates for Tir loci, including Daxx, MHC super family genes, CD84, Ly-9, Fcgr2b, Notch4, and Ddr2. Daxx (death associated protein) is a Fas binding protein, mainly located in the nucleus, functioning as a transcriptional regulator. It has been reported to mediate the Fas/JNK dependent signals in the cytoplasm as an apoptosis enhancer⁵⁰ inducing T cell death.⁵¹ The protozoan parasite Trypanosoma cruzi causes a persistent, lifelong infection. During acute Trypanosoma cruzi infection in mice, many leucocytes undergo apoptosis. Fas induced apoptosis had been implied in modulation of the immune response against Trypanosoma cruzi by interfering with cytokine and NO (nitric oxide) production during the acute phase of the infection.^52–55 Thus Daxx may play an important role in trypanosomiasis by its apoptosis inducing effect.

MHC (major histocompatibility complex) encodes various genes that are essential for immune function.⁵⁶ Notch4 is a member of the Notch family of transmembrane receptors that is expressed primarily on endothelial cells. The Notch signalling pathway has an essential role in regulating embryonic vascular morphogenesis and remodelling⁵⁷; expression of an activated form of Notch4 in embryonic vasculature leads to abnormal vessel structure and patterning and involves the modulation of the endothelial cell phenotype associating with vessel patterning and remodelling in phases of vascular development.⁵⁸ We speculate that Notch4 expression could affect vessel penetrability in the blood–brain barrier, allowing trypanosomes to invade the central nervous system in the second stage of the disease.

CD84 is a member of the CD2 subset of the Ig superfamily of cell surface molecules. It functions as a homophilic adhesion molecule and enhances interferon (IFN) secretion.⁵⁹ IFN is known to help interleukin-12 in reducing parasitaemia and prolonging survival time after acute infection.⁶⁰ In trypanosoma infection, the CD84 molecule may play a protective role in the acute phase.

Immunoglobulins have an important function in the immune system. We identified two genes of Fcer1a and Fcgr2b related to immunoglobulins in Tir3 region. Fcer1a (Fc receptor, IgE, high affinity, α polypeptide) is one subunit of the IgE receptor. The IgE receptor plays a central role in allergic disease, coupling allergen and mast cell to initiate the inflammatory and immediate hypersensitivity responses. The initiation of IgE mediated allergic responses requires the binding of IgE antibody to its high affinity receptor, Fcer1a.⁶¹ Fcgr2b (Fc receptor, IgG, low affinity 2b) molecules have been reported to serve as a negative feedback regulator for B cell Ag receptor elicited activation of B cells; any impaired Fcgr2b function is possibly related to aberrant B cell activation.⁶² Ly-9 (lymphocyte antigen 9) is a novel member of a subgroup of the immunoglobulin superfamily. Though the function of Ly-9 was unknown, other subgroups of lymphocyte antigen members are involved in adhesion reactions between T lymphocytes and accessory cells.⁶³

Two Fc receptors and Ly-9 appear to be associated with the responses in the acute phase of infection. Ddr2 (discoidin domain receptor 2) is a tyrosine kinase receptor expressed in mesenchymal tissues, the ligand of which is fibrillar collagen. Olaso et al⁶⁴ reported that proliferation of Ddr2^(−/−) fibroblasts was significantly decreased compared with Ddr2^(+/−) cells. Thus Ddr2 may be involved in the diffuse myocardial fibrosis of chronic Chagasic cardiomyopathy.

Seizures

Epilepsy syndromes have been classified into more than 40 distinct types. During the past few years, several genes have been associated with epilepsy in human families and mouse models.⁶⁵ In this study we identified seven genes as the candidates. In Exq1 loci, two genes appear to be candidate genes (table 5). Lta4h (leukotriene A4 hydrolase) is a bifunctional zinc metalloenzyme that catalyses the final and rate limiting step in the biosynthesis of leukotriene B4, a classical chemoattractant and immune modulating lipid mediator.⁶⁶ Iltif is interleukin (IL)-10 related T cell derived inducible factor, also named interleukin-22. IL-22 is produced by T cells and induces the production of acute phase reactants in vitro and in vivo, suggesting a role in inflammation.⁶⁷ In the past few years, an autoimmune mechanism has emerged in a rare form of human epilepsy, Rasmussen’s encephalitis.¹⁶ Access of IgG to neuronal epitopes in the central nervous system may trigger complement mediated neuronal damage and contribute to the pathogenesis of epilepsy.¹⁶ Perhaps the two genes Lta4h and Iltif can increase neuronal hyperexcitability by an autoimmune mechanism.

In Bis1 loci, five genes were found to be potential candidates. Cnr2 is G protein coupled peripheral cannabinoid receptor-2 (also referred to as Cb2). Cannabinoids have been shown to affect immune responses, acting on different populations of immune cells and also to inhibit T-type Ca²⁺ channels.⁶⁸ Cannabinoid acts to inhibit seizure spread in the CNS by an action on GABA,⁶⁹ and can be used as an antiepileptic.^70,71 The SNP identified in the present study may reduce the antiepileptic role of Cnr2 and increase the sensitivity of A/J strain mice to seizures.

Htr1d (5-hydroxytryptamine receptor 1D) was considered another candidate gene for its plausibility as a regulator of seizure threshold.¹⁹ It is a known G protein coupled receptor and is expressed in the human cerebral cortex; it can couple to inhibition of adenylate cyclase activity.⁷² Through this mechanism of inhibiting adenylate cyclase activity, the anticonvulsant galnon (a galanin receptor agonist) reduces the severity and increases the latency of pentylenetetrazole induced seizures in mice.⁷³

C1qb (complement component C1q, B chain) mediates complement activation through the classical pathway and plays an important role in the development of antibody responses.^74,75 It may trigger complement mediated neuronal damage and contribute to the pathogenesis of seizures by allowing IgG access to neuronal epitopes in the CNS.¹⁶

Hspg2 (heparan sulphate proteoglycan of basement membrane) encodes perlecan (a ubiquitous heparan sulphate proteoglycan). Perlecan has an important role in neuromuscular function.⁷⁶ Mutation in Hspg2 can increase muscle hyperexcitability by modulation of ion channel expression or function through their interaction with perlecan.⁷⁷

Kif17 (kinesin family member 17) is a neurone specific molecular motor in neuronal dendrites. It binds and transports the NR2B subunit of N-methyl-D-aspartate receptors (NMDARs).⁷⁸ NMDARs play an important role in synaptic plasticity and neuronal morphogenesis.^78,79 Altering the expression of NMDARs in both the dentate gyrus and the subiculum can affect tissue excitability and seizure activity in seizure sensitive gerbils.⁸⁰ Moreover, presynaptic NMDARs can facilitate glutamate release onto principal neurones in the entorhinal cortex.⁸¹ Excessive activation of glutamate receptors can destroy cortical neurones through an excitotoxic mechanism. The damaged cortex may trigger axonal sprouting and the formation of new circuits with increased excitability to seizure. Through this mechanism, Kif17 may play an important role in seizure sensitivity in mice.

Conclusions

We carried out SNP analysis based on mapped QTLs for susceptibility to diseases in animal models. Our results indicate that the combination of QTL mapping and SNP analysis is a powerful filtering tool to narrow down the gene list in QTL. Further mRNA expression screening of the candidates, for example by real time PCR, will be further improve the identification of candidate genes.