| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Developmental Immunology, Max-Planck Institute of Immunobiology, Freiburg, Germany;
Exelixis Germany, Tübingen, Germany;
Department of Genetics, Max-Planck Institute of Developmental Biology, Tübingen, Germany; and
Department of Cell Biology, Complutense University, Madrid, Spain
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
(6) genes, it was later found to give rise to a complex set of protein isoforms that are generated following alternative splicing of pre-mRNA (3). The Ikaros protein is characterized by the presence of six zinc finger domains. The first four zinc fingers engage in specific DNA binding and are variously contained in the different Ikaros isoforms; the last two fingers are required for protein-protein interactions and engage in homo- and heterodimerizations with Ikaros family members (7). In the mouse, several alleles of the Ikaros gene have been generated. The removal of the last exon of the gene (encoding the last two zinc fingers) is considered a null mutation and causes lack of B and fetal T cell development (8). A dominant-negatively acting version of Ikaros, generated by removal of the first two zinc fingers additionally inhibits development of adult T and NK cell development and leads to tumorigenesis (9). A mouse strain with low levels of the Ikaros protein was recently generated by removal of the second exon of the gene; this mutant protein retains the zinc fingers of the wild-type protein and supports residual B cell development reminiscent of a hypomorph (10). A single base substitution disrupting the third zinc finger of the Ikaros protein was recently identified from a library of ethylnitrosourea (ENU)7 mutagenized mice; this led to widespread failure of hemolymphoid differentiation (11) due to the formation of nonfunctional protein complexes. Homologs of the Ikaros transcription factor gene have been identified in the genome of all vertebrates studied to date, including zebrafish (12). It is unknown, however, whether this structural conservation also extends to evolutionarily conserved functions. The zebrafish has been recently advocated as a suitable model to investigate the genetic basis of lymphocyte and lymphoid organ development (13, 14, 15), but information about these processes in the fish is still scarce. Genetic analysis of zebrafish development has become possible through the application of forward genetic screens following ENU (16, 17) or insertional (18) mutagenesis. To develop the zebrafish model as a useful addition to the tool box of immunological research, it is important to establish not only the extent of similarities but also the degree of uniqueness of its immune system compared with the much better-studied mammalian model systems. Given the central role that Ikaros plays in lymphocyte development in mammals, we considered it worthwhile to establish the function of Ikaros in zebrafish lymphopoiesis. To this end, we made use of our collection of zebrafish mutant lines that were derived from a previously described pilot screen (19) and the recent Tübingen 2000 screen (M. Schorpp et al., manuscript in preparation).
In this study, we describe the phenotype of zebrafish homozygous for a recessive ikaros allele containing a nonsense mutation in the last coding exon, similar to the situation in a loss-of-function mutation in the mouse. Our results define larval and adult stages of lymphopoiesis in the zebrafish, provide genetic evidence for two different B cell lineages, and reveal remarkable similarities of fish and mammalian lymphocyte development.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Adult zebrafish males of the Tübingen line were incubated in buffered E3 medium containing ENU as previously reported (16). After mutagenesis, males were repeatedly mated to normal females to generate over 10,000 F1 fish. Such F1 fish were crossed to each other to generate F2 families that underwent brother-sister crosses to generate homozygous larvae in the resulting F3 clutches. In total, F3 clutches of 4,584 F2 families, representing 4,253 mutagenized haploid genomes, were screened. This represents 
1.5 times more mutagenized haploid genomes than in the first Tübingen large-scale screen (16).
Twenty to 30 larvae of each F3 clutch were screened via rag1 in situ hybridization. Based on the results on a small-scale pilot screen (19), we used the extent and pattern of rag1 staining in larvae at 120 h postfertilization (hpf) as a suitable marker to detect alterations in thymus development. At this time point in larval development, rag1 staining is confined to the thymus rudiment (19, 20). rag1 staining was expected to be reduced or absent in case of disturbed lymphoid development and also in case of aberrant development of the thymic stromal microenvironment and secondary impairment of thymopoiesis. A clutch was considered positive when 20–30% of the larvae showed a similar alteration in the rag1 expression pattern or intensity.
To confirm and recover mutations, F2 pairs producing putative mutants among their F3 offspring were crossed out, and the resulting new F3 families underwent the same inbreeding and screening procedure.
To determine whether two mutations causing similar phenotypes reside in the same or in two different genes, complementation analyses were performed, crossing a heterozygous fish of one mutation with a heterozygous fish of the other mutation. Allelic mutations fail to complement each other in transheterozygous embryos, which show the mutant phenotype like homozygotes of either allele. If the mutations are in different genes, the double heterozygous offspring show a wild-type phenotype.
A total of 141 mutants were detected in the primary analysis; of these, 92 were not yet recovered for in-depth analysis. The remaining 49 mutants were initially classified based on the results of rag1 in situ hybridization and gross morphology of fish. The first group consists of mutants that do not exhibit any obvious abnormality other than lack of or reduced rag1 staining. The second group of fish additionally displays developmental abnormalities, some with craniofacial defects of varying degrees.
Using molecular probes, all mutants were subsequently analyzed for potential abnormalities of hemopoietic cells, development of pharyngeal endoderm and ectoderm, and structures derived from neural crest at various time points during the first 5 days of embryonic development. Differentiation of hemopoietic cells in the intermediate cell mass, a site of embryonic blood formation, was assessed by staining with scl, a gene that specifies hemopoietic and vascular progenitor cells (21), gata1, a gene required for red cell development (22), and ikaros, as a putative marker of lymphoid progenitors in zebrafish (12). The arrival and early differentiation of T cell progenitors in the thymic rudiment was assessed by staining with ikaros, ccr9, the zebrafish homolog of the mammalian chemokine receptor 9, a marker of early T cells in the mouse (23, 24), L-plastin, a marker of the myeloid lineage (25), rag1, a marker of immature lymphoid cells rearranging their Ag receptor loci (19, 20, 26), TCR 
(tcr), and (tcr) (see Results), as markers for 
and 
T cells, respectively. Development of the pharyngeal arches was analyzed using gcm2, the zebrafish homolog of the mouse Gcm2 gene as a marker for pharyngeal ectoderm (27), and foxn1 (28), the zebrafish homolog of the mouse Foxn1 gene that is required for differentiation of thymic epithelial cells (29) and expressed in endodermal derivatives. Neural crest development was assessed by dlx2 expression (30) and cartilage formation by Alcian blue stainings.
Genomic localization of zebrafish mutations was determined using the Tübingen marker set for genome scans (version 4) on F2 Tübingen x Wik crosses of the mutant carriers. Primer sequences are available from the MGH website 
http://zebrafish.mgh.harvard.edu
. The ikaros gene, which in the HS mapping panel has been linked to z5395; in the MGH mapping panel, z5395 and z10070 map to linkage group 13 at 40 cM; cf ZFIN database at http://zfin.org/cgi-bin/webdriver?MIval=aa-newmrkrselect.apg). To confirm linkage to the ikaros gene, an intragenic marker was used (Table I).
|
For initial screening, F3 clutches were incubated in the presence of 0.25 mM 1-phenyl-2-thiourea (Sigma-Aldrich) to inhibit melanin synthesis, and fixed in 4% paraformaldehyde/PBS at 120 hpf. Whole mount in situ hybridization (WISH) was conducted in specially designed 48-well plates (Aldinger) with digoxigenin-labeled probes for growth hormone (31), and rag1, a marker for thymic T cells (20, 26), following standard protocols (19). Details for these and all other probes used in this study can be found in Table II. Analysis by electron microscopy followed earlier procedures (20).
|
To analyze igµ rearrangements in genomic DNA, a published protocol was used (32); ig
rearrangement assays used primers listed in Table I. VDJC containing cDNAs from igµ, ig, tcr, tcr genes were amplified using primers listed in Table I. Fragments were cloned and sequenced. TUNEL assays were performed using Roche in situ cell death detection kit; the wound healing and proliferation assays were performed by incubation with BrdU at a concentration of 150 µg/ml in fish water and the incorporation was assayed using the Roche BrdU Labeling and Detection kit II. All analyses were performed for at least two animals of each genotype; since no significant interindividual differences were found, the results per genotype were pooled.
Data mining
The zebrafish ccr9 gene was identified using the human CCR9 coding region sequence (accession NM_031200) as a query sequence against the zebrafish whole genome sequence (WGS) trace archive using the megablast algorithm. Using these initial genomic sequences that covered exon 3 of zebrafish ccr9, RT-PCR primers were designed for 5'-RACE to obtain cDNA sequence information (data not shown). This resulted in 150 bp of additional sequence at the 5'-end, which is contained in two exons, as determined by blastn comparisons against the WGS trace archive. The initiation codon is found in exon 1. The location of the poly(A) site has not been determined.
The tcr C region gene was identified using the flounder tcr C region domain (nt 18922–19064 and 19137–19376 in accession number AL596139) as a query sequence against the zebrafish WGS trace archive using the blastn function of the megablast algorithm. Using these initial genomic sequences, a sequenced bacterial artificial chromosome (BAC) clone, assigned to linkage group 2, was identified that contained three D, 2 J, the C C region exons and numerous presumptive J elements downstream of C (accession number BX681417.10). RT-PCR primers were designed to link the first and last C exons and revealed an additional exon not predicted by gene structure algorithms. The BAC sequence does not contain V elements for 190 kb upstream of D1, indicating that V/V elements may be in inverted configuration downstream of C as observed in Tetraodon nigroviridis (33) and other teleosts. Indeed, a contig of three BACs joins J, C, and V/V elements, supporting this assumption. So far, a total of five V sequences (contained in DD332, B14, B22, B35, B38 cDNA clones) were identified spliced to C sequences (data not shown for B14, B22, B38; for B35, see Fig. 7; the VDD332 element is contained in trace sequence zfish44910-13e08.p1k and BAC clone BUSM1-257N2, accession number AL928815.15). These V element sequences (with the exception of VDD332) have previously been annotated as V sequences (the sequences for B14 and B22 were too short to allow an unambiguous assignment to a specific V element); B38 has been annotated as VI-14 (cf accession AL591674); B35 has been annotated as VII-36 (cf accession AL592550.11).
|
C region genes were identified using the various tcr cDNA or expressed sequence tag sequences derived from other teleost fish sequences as query sequences against the zebrafish WGS trace archive using the blastn function of the megablast algorithm. These initial genomic sequences were assembled into contigs and manually curated for consistency; the assembly process yielded a single contig for C1 (with some uncertainty because of a tetranucleotide repeat region between exons 3 and 4), and two contigs for C2. Some of these sequences are contained in the sequence of contig assembly CAAK01000103, which also contains V elements. The assembled sequence of CAAK01000103 (version of January 19, 2005) awaits expert curation and is probably not correct, because some sequences of C region genes are in incorrect order possibly due to many repeats found in this sequence. Upstream of C1, one D element (D1.1) and several J elements (J1.1 to J1.23) were identified (see Table IV). Five further putative J1 elements were found in cDNAs and confirmed by the presence of equivalent genomic sequences in WGS trace archive; a further three J1 elements were detected in cDNAs but not in genomic sequences. Upstream of C2, no D, but a single J, element (J2.1; see Table IV) was identified. One other J element (possibly also belonging to the J2 cluster as judged from the sequence of cDNAs containing this element) was identified in assembly sequence CAAK01003749. This arrangement of J elements is reflected in cDNA sequences that generally have the structure V-D1-J1-C1 and V-J2-C2 (data not shown).
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
The phenotype of mice with a truncated Ikaros gene indicates that the null phenotype of this gene is characterized by the complete lack of fetal thymopoiesis. In our collection of ENU-generated zebrafish mutants, several lines exhibited a complete lack of thymopoiesis as determined by WISH with rag1-specific probes in the absence of other developmental defects at 5 days postfertilization (dpf). In zebrafish, the larval stage is commonly considered to last for the first 2 wk, followed by an adolescent period up to 3 mo of age. To identify potential ikaros mutants among this subgroup, we tested for genetic linkage of mutations to chromosome (linkage group) 13, to which ikaros has previously been assigned. Segregation analysis using a panel of simple sequence length polymorphism markers spanning the entire zebrafish genome localized the mutation in line IT325 between markers z5395 and z13250 and very close to marker z10070 (zero recombination events in 168 meioses initially analyzed). Because of the close genetic linkage between the ikaros locus and the map position of the mutation in IT325, the ikaros gene was sequenced in this line. We identified a nonsense mutation at the beginning of the last coding exon of the ikaros gene (allele designation t24980). This exon is equivalent to exon 7 in the mouse Ikaros gene. A C>T transition converts a CAA (Q360) codon to a TAA (termination) codon. The predicted truncated Ikaros protein encoded by the t24980 allele is 177 aa shorter than the wild-type protein, leaving intact the putative bipartite activation domain, but removing the two C-terminal zinc fingers implicated in essential protein-protein interactions (7) (Fig. 1A). To ascertain that the phenotype observed in IT325 mutants indeed segregated with the identified ikaros mutation, we developed an intragenic polymorphic marker (ikaros001, located in the third intron of the ikaros gene; see Table I). It showed tight genetic linkage (zero recombinants in 1020 meioses) and all mutants analyzed were homozygous for the nonsense mutation. This strongly suggests that the phenotype in IT325 mutants is caused by the mutation in the ikaros gene. This conclusion was supported by the finding of reduced numbers of rag-1-positive lymphocytes in the thymus at 4 dpf in ikaros morphants (Table III). Collectively, these experiments indicated that the phenotype in IT325 mutants was caused by the nonsense mutation in the ikaros gene.
|
|
Fish homozygous for the t24980 allele lacked rag1 and ikaros expression in the thymus up to the first 8 dpf as determined by WISH (Fig. 1B). Likewise, ccr9 expression in the thymic rudiment (Fig. 2A), which in the mouse is a marker of early T cell progenitors (23) was absent (see Fig. 3 for characterization of the zebrafish ccr9 gene, and Table II for information about all other probes used for in situ hybridization). The apparent lack of lymphocyte progenitors was confirmed by WISH with probes specific for the C regions of the tcr2 and tcr genes (Fig. 4A), and RT-PCR analyses for these genes (data not shown). A general hemopoietic defect in ikaros mutants was ruled out by the normal expression patterns of gata-1 and scl (Fig. 2, B and C) at 24 hpf. Furthermore, the thymic microenvironment was not disturbed in the mutants as shown by normal expression patterns for gcm2, foxn1 (Fig. 2A), and dlx2 and the normal appearance of cartilage in branchial arches (Fig. 2D). The foxn1 hybridization pattern in t24980 mutants (Fig. 2A) was indicative of a compact arrangement of epithelial cells, presumably because no hemopoietic progenitors were present in the thymic rudiment that disperse the resident epithelial cells. This finding was supported by the results of electron microscopic studies. The wild-type thymus was covered by pharyngeal epithelium and contains numerous lymphoblasts and lymphocytes among thymic epithelial cells; in the mutant thymus, the lymphoid cells were completely missing, whereas the epithelial structures appeared normal (Fig. 4B). Collectively, these studies are compatible with the notion that the t24980 allele affects T lymphocytic differentiation at some prethymic stage. Heterozygous animals analyzed according to the above criteria were completely normal up to 8 dpf, suggesting the recessive nature of this allele.
|
|
|
and loci in zebrafish
The analysis of zebrafish lymphopoiesis is complicated by the fact that no cell surface markers are available that facilitate such analyses in mouse and human systems. Although a recently developed transgenic zebrafish line facilitates gross analyses of the entire T lineages (34), it does not allow the detailed analysis of B and T cell subsets required for the present experiments. Therefore, we have developed specific reagents to examine tcr and tcr expressing cells by in situ hybridization and sequence analysis of their Ag receptor gene rearrangements.
Using a combination of database mining and cDNA cloning (see Materials and Methods and Figs. 5–7), an initial characterization of zebrafish tcr and tcr loci was conducted. The tcr C region gene of zebrafish was identified using the flounder tcr C region domain as a query sequence against the zebrafish WGS trace archive. Using these initial genomic sequences, a sequenced BAC clone was identified that contained three D, two J, the three C C region exons, and numerous presumptive J elements downstream of C (Figs. 5 and 6). Some V sequences identified in tcr cDNAs were also found in tcr cDNAs, indicating that tcr and tcr shared the same V elements (data not shown). Because V/V and J/C elements were linked to each other in inverted orientation, V/V elements may be located, in an inverted configuration, downstream of C as observed in T. nigroviridis (33) and other teleosts.
|
|
genes, and several putative V elements. These cDNA sequences were localized to several partially sequenced BAC clones; further analysis suggested that the two tcr regions are tandemly arranged, with a single D element (designated D1.1) associated with the first complex. Upstream of the four C1 exons, >20 J1 elements were identified in the available (incomplete) contig sequence; further J1 elements were identified in the cDNA clones investigated in this study (Table IV). The cDNAs containing the C2 gene lacked readily identifiable D sequences and contained either of two J elements (J2.1 and J2.2). In contrast to C1, the C region of tcr2 was found to be encoded in three exons. During ontogeny, complete (VJC) tcr transcripts were first detected with C2 sequences (beginning 4 dpf), whereas VDJC transcripts with C1 sequences were not detected up to 7 dpf. In adult tissues, both forms were coexpressed. Complete VDJC tcr transcripts appeared at 5 dpf. Incomplete (sterile) transcripts lacking V elements were found considerably earlier (tcr2 and tcr from 2 dpf onward; data not shown).
Collectively, these studies provided specific reagents to examine tcr and tcr expressing cells by in situ hybridization and sequence analysis of VDJ rearrangements.
T cell development in ikarost24980 homozygous mutants
Interestingly, homozygous mutants survived for >17 mo (the latest time point of analysis) under nonsterile conditions and were fertile. Because it appeared unlikely that lymphopenic fish would survive for such long periods, we examined the possibility that lymphocyte development recovered at later stages of development in ikaros mutants. To this end, we examined lymphocyte development in the thymus by in situ hybridization on tissue sections. Early larval stages of wild-type fish showed an intense staining with the rag1 probe, while mutants lacked evidence for thymopoiesis by this criterion (Figs. 1 and 4). By contrast, adolescent mutant thymus, albeit smaller, contained a large number of rag1-positive cells as well as tcr- and tcr-expressing cells (Fig. 8A). The first few rag1-positive cells were detected in the thymus of ikaros mutants at the beginning of adolescence, at 14 dpf (data not shown). electron microscopic studies confirmed the overall similarity of thymic structures between wild-type and mutant fish at these stages (data not shown). This pattern was maintained until 12 mo of age and beyond (data not shown). This indicated that in contrast to adolescent and adult T cell development, larval T cell development was dependent on ikaros function, as previously observed in mice.
|
1 and tcr cDNAs. The majority of VDJC transcripts of tcr1 and tcr genes amplified from wild-type and mutant fish emanated from productive rearrangements (Fig. 8, B and D). The proportion of unique tcr1 sequences among the pool of sequenced cDNAs was significantly higher in wild-type as compared with mutant fish; the same was true for tcr cDNAs (Fig. 8, C and E). We noted, however, that the tcr1 repertoire was more severely restricted than that of tcr clones (compare Fig. 8, C and E; 
2 = 6.31; p = 0.012). To explore further possible abnormalities of T cell repertoire formation, we examined the pattern of V and J usage (because of the many available J1 elements, a similar analysis for tcr1 rearrangements was hampered by insufficient statistical power). Whereas in the wild-type, J1 was used 16 times and J2 17 times all with the V35 element, ikaros mutants used J1 only 4 and J2 29 times (2 = 8.88; p = 0. 0003), and the Vd35 element only 23 of 34 times (the remainder using V35a, see Fig. 6) (2 = 12.77; p = 0.0004; data not shown). This clearly showed that T cell selection in the thymus was abnormal in ikaros mutants.
To explore the biological relevance of these findings further, we repeated the analysis with cDNAs obtained from kidney (an important primary and secondary lymphoid tissue in teleosts) of adult fish (9 mo or older). The results indicated that the severe restriction of the tcr1 repertoire persisted in the mutants (Fig. 8C), whereas the repertoire of tcr clones had become indistinguishable between wild type and mutant (Fig. 8E). These results indicated that in the absence of normal ikaros function, development of tcr1-expressing T cells was more severely disturbed than that of tcr-expressing T cells. In both wild-type and ikaros mutants, V35 and J2 elements dominated in the sequenced clones, indicating that the skewed T cell repertoire in the thymus was normalized in peripheral tissues. It also suggested that, at least for tcr-expressing T cells in the periphery, abnormal proliferation did not occur. Collectively, our analyses indicate that thymopoiesis occurs with reduced efficiency in mutant fish with only few cells completing T cell maturation, followed by a certain degree of homeostatic proliferation in the periphery.
Two B cell lineages in zebrafish
Recent reports have identified a second Ig H chain isotype, ig, in zebrafish (35) and rainbow trout (36). It has been proposed that it may be expressed in a separate lineage of B cells (36). Indeed, the structure of the igh locus (35) (Fig. 9A) lends itself to a genetically programmed lineage choice; rearrangement of a VH segment to Dµ/Jµ elements deletes the intervening DJ elements, much like the situation of the tandemly arranged tcr/tcr loci whose rearrangement patterns guide the development of and T cells, respectively.
|
in wild-type and ikaros mutants at 
7 mo. Using RT-PCR, we failed to amplify VHDJC cDNAs in mutant fish, using primers specific for all VH families (37) tested, whereas VHDµJµCµ cDNAs were readily amplified in wild-type and mutants (Fig. 9B). Surprisingly, VHDJ rearrangements could not be amplified even from total genomic DNA of kidney marrow in ikaros mutants; again, VHDµJµ rearrangements were readily amplified (Fig. 9B). This indicated that ig-using B cells were absent in ikaros mutant fish and suggested that ig-expressing B cells required ikaros function for development and/or maintenance; furthermore, the absence of rearranged ig alleles indicates that the choice of differentiating along the igµ or ig pathways is not stochastic.
Next, we determined the presence of VHDµJµ- and VHDJ-rearrangements in wild-type and mutant fish at earlier stages of development. In wild-type fish, VHDµJµ rearrangements were first detected 3 wk postfertilization, whereas VHDJ rearrangements were first detected 1 wk later. In ikaros mutants, VHDµJµ rearrangements were first detected 4 wk postfertilization, albeit in fewer copies per embryo than in wild types (data not shown). VHDJ rearrangements were never observed in mutant fish. Collectively, these studies indicated that the development of B cells using igµ is delayed but not completely abrogated in ikaros mutants.
Abnormal development of igµ-expressing B cells in ikarost24980 homozygous mutants
B cells represent a large part of lymphoid cells in the head kidney, the fish equivalent of the mammalian bone marrow. igµ cDNA sequences can be readily amplified from both wild-type and mutant tissues using primers specific for all VH families tested (37). For in-depth sequence analysis in our mutants, we arbitrarily chose cDNAs containing VH1 family genes. Limited analysis of other VH families revealed qualitatively similar results (data not shown). Most igµ cDNA sequences obtained by RT-PCR using VH1- and Cµ-specific primers, respectively, from wild-type and mutant tissues were in-frame (productive) across the V-D-J-C junctions (Fig. 9C). By contrast, the fraction of productively rearranged igµ alleles amplified from genomic DNA was lower (Fig. 9C), suggesting the presence of nonsense-mediated decay of mRNAs with premature termination codons in B cells of zebrafish. A second observation was that the fraction of productively rearranged alleles in genomic DNA was significantly lower in mutant tissues (0.15 in mutant vs 0.75 in wild type). The overrepresentation of nonproductive rearrangements in mutant tissue (2 = 64.75; p < 0.0001) could only be explained by the presence of cells with µ–/o and µ–/– genotypes, because, in the presence of allelic exclusion, igµ+ cells can only be µo/+ or µ–/+ (µ°; no rearrangement at igm; µ–; nonproductive rearrangement at igm; µ+; productive rearrangement at igm) (see Fig. 9D).
This suggested that in the absence of normal ikaros function, cells with nonproductive rearrangements were maintained, whereas they were eliminated in normal tissues. Therefore, it was necessary to exclude a global defect in apoptosis and/or proliferation of hemopoietic cells in the absence of normal ikaros function. To this end, the number of apoptotic cells in head kidney tissues was determined by TUNEL staining at 8 wk postfertilization. The results showed a slight (25%) increase in apoptosis in ikaros mutants (p = 0.01; 2 test, data not shown). Although in these experiments we were unable to differentiate among hemopoietic cells of different cell lineages, the results nevertheless excluded some general defect in apoptosis pathways as a result of loss of normal ikaros function. Next, we examined the general proliferation propensity of hemopoietic tissues in the head kidney of wild-type and ikaros mutants at 8 wk postfertilization. Proliferation was analyzed in two different ways. First, we examined the expression of pcna on sections of head kidney by RNA in situ hybridization to measure the number of proliferating cells. Mutant tissues showed a slight (20%) reduction of proliferating cells (p = 0.002; data not shown). In a second experiment, 7-wk-old fish were exposed to BrdU in their tank water, the tissues were fixed, and sections were developed for BrdU staining. The results after 5 days of BrdU treatment indicated that the number of BrdU-positive cells (both high and low intensity staining) accumulating during this time period is about the same for wild-type and mutant tissues (data not shown). When fish were exposed to BrdU for 5 days, and then kept for an additional 2 days in the absence of BrdU before analysis, the number of strongly stained nuclei was reduced compared with samples without chase. However, there was no difference in the number of BrdU-positive cells between wild-type and mutant tissues, indicating that mutant cells did not proliferate faster than normal ones (data not shown). Finally, we repeated the BrdU-labeling experiment (10 day labeling period) with 7-mo-old fish. Here, a slight (20%) reduction of labeled cells was observed in mutants (p = 0.01, data not shown), suggesting that overall capacity of proliferation in hemopoietic tissues was somewhat reduced at this age. Collectively, however, our data excluded major changes in apoptosis and proliferation in the absence of normal ikaros function.
We next examined the repertoire of igµ sequences represented in productively and nonproductively rearranged igµ alleles. As shown in Fig. 10, A and B, the sequence diversities differed significantly between wild-type and mutant fish, both for productively and nonproductively rearranged alleles. However, because productively rearranged alleles dominate in the wild-type (Fig. 9C), nonproductively rearranged alleles can be considered passenger sequences in cells with a µ–/+ genomic configuration at the igh locus (Fig. 9D). In mutant cells, productively rearranged alleles were the minority and hence did not contribute significantly to the properties of the entire cell population (Fig. 9C). Therefore, with regard to sequence diversity, the most informative comparison was that between the sequence diversities of the dominant types of alleles, the productively rearranged wild-type and nonproductively mutant alleles. The proportion of unique sequences was significantly smaller in nonproductively rearranged alleles of mutant fish (2 = 18.81; p < 0.0001). A qualitatively similar result was observed for sequences obtained from cDNA, although the low numbers obtained from mutant tissues reduced the power of the statistical analysis. This result strongly suggested the presence of oligoclonal immature igµ– B cell populations in the mutant kidney tissue that outnumbered igµ+ cells. To verify these predictions, we directly examined the presence of igµ+ cells in wild-type and mutant tissues by in situ hybridization and determined the number of total B cells by use of pax5 expression, as a marker of B lineage identity (38). In wild-type tissue, igµ+ and pax5+ cells were present in comparable numbers; in mutant tissue, the ratio of igµ+ to pax5+ cells was much lower, suggesting that immature igµ– cells predominate among pax5+ B cells (Fig. 10C).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
The mutation we describe here is similar but not identical to that predicted for the null allele in mouse (8); the mouse allele was generated by deletion of the entire C-terminal exon that contains a bipartite activation domain and two zinc fingers required for protein-protein interactions (8). The ikaros allele described here differs from this mouse mutant in that the activation domain is retained in the zebrafish, while the two zinc fingers are also missing. At present, we have no evidence indicating that this allele acts in a dominant-negative fashion, although more detailed analyses are required to firmly establish this conclusion.
Collectively, our results suggest that the t24980 mutation generates a null (or a hypomorphic) allele of the zebrafish ikaros gene. Our data indicate that T cell development in the zebrafish proceeds in at least two phases. We have characterized the tcr and tcr loci as a tool to assess the development of these two major T cell lineages. Larval development of tcr- and tcr-expressing T cells is absolutely dependent on normal ikaros function. Interestingly, thymopoiesis resumes at later stages (beginning in adolescence), although the T cell repertoire in the thymus of ikaros mutants is abnormal and less diverse than in wild-type siblings. A qualitatively similar observation has also been made in mice with an Ikaros null mutation, where some subsets of T cells are missing and others as well as T cells exhibit signs of restricted repertoire (8). In our zebrafish mutants, low efficiency of thymopoiesis, biased VDJ rearrangements, aberrant selection of the resulting TCRs, and a low efficiency of rearrangements per se may all contribute to a biased repertoire. It also appears that tcr-expressing T cells are more severely affected than tcr-expressing cells. An oligoclonal T cell repertoire, defective T cell selection, and impaired CD4 vs CD8 lineage decisions have been found in the thymus of Ikarosnull mice (40), supporting the view that the functions of Ikaros in T cell development are similar in mouse and fish.
In the mouse, B cell development does not recover in adult Ikaros mutants (8). In zebrafish, this is only true for the igz-expressing lineage of B cells, while cells expressing the igm isotype recover. At present, we do not know whether the recovery of igm-expressing B cells is due to the particular nature of the ikarost24980 allele or a species-specific difference between zebrafish and mouse. We note, however, that a hypomorphic allele of Ikaros has been shown to allow residual B cell differentiation in the mouse (10). If rearrangement at the igz/m locus were entirely stochastic, a certain fraction of cells expressing igm should also harbor nonproductive igz rearrangements. By contrast, if rearrangement at the igz/m locus is instructive, each cell lineage should rearrange either z or m. The latter is compatible with our results, suggesting that the development of B cells expressing the z and m isotypes can be genetically separated with respect to the requirement of normal ikaros function.
Ikaros likely functions at different levels of T and B lymphocyte differentiation. First, it regulates the provision of lymphoid progenitors. For larval lymphopoiesis, this function is essential, whereas at later stages of development, ikaros function is partially dispensable. Second, ikaros appears to participate or be required in lineage decisions among lymphoid subtypes, namely - vs -expressing T cells and ig- vs igµ-expressing B cells. Third, it appears to play a role in the regulation of Ag receptor rearrangements as exemplified by the biased usage of V and J elements in tcr rearrangements.
Remarkably, no excessive deaths were observed among mutant fish up to 17 mo of age, suggesting that even a severely restricted Ag receptor repertoire does not impair essential surveillance functions such as specific immune defense and wound repair (data not shown). Interestingly, the oligoclonal repertoire does not lead to overt lymphoma or leukemia and no indolent disease was observed in three 1-year-old mutant fish using complete histological surveys; this may be due to residual activity of the truncated Ikaros transcription factor suppressing uncontrolled proliferation and/or due to the fact that secondary genetic lesions occur at low frequency, if at all, in our mutant fish.
In conclusion, our analysis of the ikarost24980 mutant is the first report of the isolation of a lymphocyte mutant in the zebrafish model based on the presence of a specific phenotype. This suggests that the mutagenesis screens conducted in our and other laboratories (13, 14) may reveal further details of the genetic requirements of zebrafish lymphocyte development. Reassuringly, our data indicate that the overall genetic requirements for lymphopoiesis in zebrafish are very similar to those of mammals. This validates the zebrafish system as a tool to identify novel genes involved in early lymphoid development. Lastly, the analytical tools described here will enable a more detailed analysis of other lymphocytic mutants that have been isolated in our and other screens.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported by the Deutsche Forschungsgemeinschaft (SFB620-A8). 
2 M.S. and M.B. contributed equally to this work.
3 Current address: Department of Genetics, Max-Planck Institute of Developmental Biology, D-72076 Tübingen, Germany.
4 Current address: Hertie-Institut für Klinische Hirnforschung, D-72076 Tübingen, Germany.
5 A list of members of the Tübingen 2000 Screen Consortium and the Freiburg Screening Group and their affiliations appears in Acknowledgments at the end of the paper.
6 Address correspondence and reprint requests to Dr. Thomas Boehm, Department of Developmental Immunology, Max-Planck Institute of Immunobiology, D-79108 Freiburg, Germany. E-mail address: boehm{at}immunbio.mpg.de
7 Abbreviations used in this paper: ENU, ethylnitrosourea; hpf, hours postfertilization; WGS, whole genome sequence; dpf, days postfertilization; BAC, bacterial artificial chromosome; WISH, whole mount in situ hybridization.
Received for publication November 22, 2005. Accepted for publication June 5, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
/ linkage in the teleost fish Tetraodon nigroviridis. Genomics 79: 241-248. [Medline]This article has been cited by other articles:
|
|
 |
|
  O. W. Stockhammer, A. Zakrzewska, Z. Hegedus, H. P. Spaink, and A. H. Meijer Transcriptome Profiling and Functional Analyses of the Zebrafish Embryonic Innate Immune Response to Salmonella Infection J. Immunol., May 1, 2009; 182(9): 5641 - 5653. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. B. John, S. Yoong, and A. C. Ward Evolution of the Ikaros Gene Family: Implications for the Origins of Adaptive Immunity J. Immunol., April 15, 2009; 182(8): 4792 - 4799. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Soza-Ried, C. C. Bleul, M. Schorpp, and T. Boehm Maintenance of Thymic Epithelial Phenotype Requires Extrinsic Signals in Mouse and Zebrafish J. Immunol., October 15, 2008; 181(8): 5272 - 5277. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Carradice and G. J. Lieschke Zebrafish in hematology: sushi or science? Blood, April 1, 2008; 111(7): 3331 - 3342. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. W. Glenney and G. D. Wiens Early Diversification of the TNF Superfamily in Teleosts: Genomic Characterization and Expression Analysis J. Immunol., June 15, 2007; 178(12): 7955 - 7973. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
