Background
Identification of genes that regulate susceptibility to arthritis is essential in the selection of molecular targets for therapeutic application. The primary locus regulating collagen induced arthritis (CIA) was found to be within the major histocompatibility complex (MHC) loci [
1‐
5]. This was expected because it was established that CIA in both rats and mice is MHC-linked [
6‐
10]. Tremendous research has been done on the study of molecular mechanism of CIA model [
11‐
15]. However, a considerable contribution to the arthritis is caused by a non-MHC-linked molecular mechanism [
16‐
23]. Many non-MHC-linked loci have been identified [
19,
22,
23], yet the causal genes for these QTL have not been known. Understanding of genes that regulate susceptibility to non-MHC-linked arthritis is essential for—the selection of molecular targets for therapy for non MHC-linked arthritis.
BALB/c and DBA/1 are important pairs of mouse strains. BALB/c and DBA/1 and C57B6/J mice share the same MHC (H-2d) haplotype; however, at the whole genome level they are genetically distant [
17,
24]. In the mouse model, with a standard protocol, CIA is induced in DBA/1 and B6 [
3,
10,
11,
25]. Under the same standard protocol, CIA could not be induced in BALB/c mice. The spontaneous arthritis (SAD) occurs in interleukin-1 (IL-1) receptor antagonist (IL-1rn) -deficient mice, which is dependent on non-MHC genetic bases [
25,
26].
Over the last decade, we have been studying SAD in
IL-1ra-deficient mice [
25‐
29]. To better understand the pathogenesis of SAD, initially we used classical genetic techniques and bred susceptible and resistant mice to obtain an F2 generation and identified QTL associated with arthritis susceptibility [
26]. We obtained evidence for potential QTL on chromosomes 1, 6, 11, 12, and 14. The QTL on the chromosome 1 was the major regulator for the SAD. To confirm the importance of the QTL and to identify potential candidate genes within it, we conducted speed congenic breeding to transfer the QTL region from DBA/1 mice that are resistant to spontaneous arthritis into BALB/c
-/- which are susceptible [
29]. Our congenic breeding was successful in identifying a QTL associated with the development of spontaneous arthritis. When a fragment of DNA from the DBA/1 strain was introduced onto a BALB/c background, arthritis was delayed in onset and was less severe. Using the congenic strains that we developed, the genomic region of the originally identified QTL was redefined into a region that is downstream from the peak region of our original mapping [
26,
27,
29].
Microarrays have been used for the analysis of whole genome expression profiles for the more than a decade. They have developed as a mature technology not only in the producing of gene expression profiles but also in the analysis of differentially expressed genes. One of the useful applications of microarray-generated whole genome expression profiling is the identification of candidate genes that regulate a specific disease trait or a molecular pathway [
28,
30‐
34].
According to the Ensembl database, the transferred region from DBA/1-/- to BALB/c-/- is between D1Mit110 and D1Mit209 on chromosome 1 with a size of 23.73 Mb which contains 320 genetic elements. Among those genes, 115 are identified as genes relevant to arthritis and its potential pathways [
29]. Although the number of genes seems large and to some degree difficult to specifically target candidate genes, comparison of the data from congenic strains provide a defined genomic region for their localization. The use of congenic strains enable us to analyze the effect of the transferred fragment within QTL region on global gene expression. Previously, we have been using the microarray tools for the study of differential gene expression comparing diseased knockout and normal wild type mice [
28,
32,
33]. In this study, we compared gene expression profiles of the whole genome of the congenic strain and parental strains.
Discussion
Our multiple comparisons confirmed that the decreased expression levels of genes in T-cell receptor beta chain and Ifi cluster in the congenic strain are the major change that potentially leads to the increased resistance to SAD in the congenic strain in comparing to that of BALB/c-/-. The expression levels of these genes are similar to that in DBA and DBA/1-/-, which are resistant to the SAD. The expression of these genes is lower than that of BALB/c and BALB/c-/-, which are susceptible to SAD. There are considerable number of genes that are differentially expressed between the congenic strain and three other parental strains, the DBA, -/-DBA/1-/-and BALB/c because of the different genomic background and mutation of Il1r. However, there is only one difference between the congenic strain and BALB/c-/-. The difference is that in congenic mice, the genome region of QTL on chromosome 1 is replaced by the fragment from DBA while the entire remainder of the genome of BALB/c-/-are from the BALB/c strain. The small number of candidate genes derived from our analyses pave the way to final determine the causal gene in the congenic strain.
Our preliminary data indicated the
Ifi gene 200 cluster contains the favorite candidate gene for the SAD resistance in the congenic strain. Among differentially expressed genes between congenic and BALB/c
-/-, four genes in the
Ifi gene 200 cluster are differentially expressed and are located in transferred genomic fragments in congenic strains. These genes are located in the transferred QTL genomic fragment on mouse chromosome 1, which is between D1Mit110 (167758517-167758653) and D1Mit209 (191493187-191493284). Our earlier analysis indicated that four candidate genes:
Fcer1g, Fcgr3, Ifi202b, and
Kmo are among the differentially expressed genes [
29,
34]. Our current analysis indicates that the other three genes,
Fcer1g, Fcgr3, and
Kmo, were not differentially regulated in congenic strains compared to BALB/c
-/- that is susceptible to spontaneous arthritis. Interferon-activatable protein has been reported as an immune suppressor [
36,
38‐
40]. For example,
Ifi202 is known as an interferon-inducible lupus susceptibility gene [
36,
38]. Silencing of
Ifi202b in tolerized CD8+ Ti cells abrogated the suppressive capacity of these cells. Recently the role of
Ifi204 in the response to bacterial infection has been investigated [
39]. It is expected that in the near future much broad function of
Ifi family genes in the infection and immune innate immune responses will be identified [
40]. Accordingly, it is very important to further investigate their candidacy.
The regulatory role of the T-cell receptor beta chain in the susceptibility of arthritis is well known [
41‐
43]. Current information indicates that genes of T-cell receptor beta chain are not located in the transferred genomic fragment of the QTL region in the congenic strain. However, it is important to investigate if the genes between T-cell receptor beta chain family and
Ifi family interact to down regulate SAD in the congenic strain.
In addition to the T-cell receptor beta chain and Ifi genes, we also detected the differential expression of Lefty1 in the QTL region. Early study indicated that Lefty1, Lefty2, and Nodal are expressed on the left side of developing mouse embryos and are implicated in L-R determination. Recently, because of the unique ability of Lefty to reroute the cell fate by making cells nonresponsive to diverse differentiation factors such as Nodal or other signals that require Egfcfc as a co-receptor, Lefty has become regarded as a determinant of cell stemness or differentiative events. There is no report for its role in arthritis, inflammation or infection. At present, we do not have enough information to explain its role in the congenic strains.
Our study was done with three replicates, each from female mice from each strain. The data may reflect a bias due to the use of female mice and may not represent that from male mice. Sex hormones have been shown to differentially regulate expression of
Ifi202 [
44]. However, all the strains are measured at the same age stge. Therefore, the differences at least reflect the genomic variation in the female mice between congenic and other strains. We used the spleen as the organ for the RNA extraction and whole genome gene profiling [
28]. Although spleen as an organ important for immune, the phenotype of SAD in these mice were based on the joints. Therefore, we do not rule out that possibility that the data may not reflect what all happened in the joints. However, we believe that the molecular mechanism detected from this study is important to elucidate the causal gene for the congenic strain. Further study to directly link the candidate to the disease is necessary.
The “interferon-activatable protein” was labeled in the Illumina chip for a probe of LOC100040462. However, it lately has been validated as a transcription factor called
Mndal, which is also a member of
Ifi family gene. Interestingly, Zhang et al. in their paper indicated that
Mndal is absent from DBA/2j [
45]. We assume that DBA/1, the strain we used in this study, is also absent. The increased level of expression level of
Mndal in this case perhaps represents the
Mndal gene from absent to present. However, the expression level from the DBA/1j genome detected by this probe is above the none-expression level (Absolute value over 118) while the expression level from the Balb/c genome is twice higher than that of the DBA/1j genome. As to its potential effect, it could be to interact with other genes in the
Ifi family or to affect the expression level of heat hsp105/110hDa and Igkc2-112, which are only two known genes which showed the decreased expression between the congenic strain and the most closest parental strain, DBA-/-. However, because of the differential expression level of other genes in
Ifi family, whether
Mndal affect the expression of hsp105/110hDa and Igkc2-112 need further study.
Our data indicated that many differentially expressed among those mouse strains are not associated with Il1-rn. This result is not completely unexpected. First of all, SAD occurs in the mouse strain that is lack of expression of Il1-rn. One would predict that the gene expression profile of such a mouse strain does not reflect the pathways that involves the Il1-rn. Secondly, the causal genes derived from the comparison between the congenic strain BALB.D1-1 and BALB/c-/- are under the genomic background of lack of Il1-rn. Thus, the causal genes are in fact acting when there is a lack of Il1-rn. It is important to notice that with standard protocol, CIA has been induced in DBA/1 and B6 [
3‐
5]. Under standard protocol, CIA could not be induced in BALB/c mice without modification. The SAD in interleukin-1 (IL-1) receptor antagonist (IL-1rn) -deficient mice is dependent on non-MHC (major histocompatibility complex) genes in mice. It is well-known that Il1-rn has played an important role in the arthritis. The role of
Ifi family genes in SAD may act in the population that is non-MHC and lack of role of Il1-rn. On the other hand, the lack of association in the gene expression does not rule out the protein-protein interaction. The protein products of these two gene families may be directly interact or binding to each other.
Acknowledgements
This work was partially supported funding from The first hospital of Qiqihare City and merit grants (I01 BX000671 to WG, 1IPIBX001607 to JS) from the Department of Veterans Affairs and the Veterans Administration Medical Center in Memphis, TN, USA and National Natural Science Foundation of China (Project 81372996 to YJ; Project 81171679 to YHC), P.R. China.