Mutation of Jak3 in a Patient with SCID: Essential Role of Jak3 in Lymphoid Development

Males with X-linked severe combined immunodeficiency (XSCID) have defects in the common cytokine receptor γ chain (γc) gene that encodes a shared, essential component of the receptors for interleukin-2 (IL-2), IL-4, IL-7, IL-9, and IL-15. The Janus family tyrosine kinase Jak3 is the only signaling molecule known to be associated with γc, so it was hypothesized that defects in Jak3 might cause an XSCID-like phenotype. A girl with immunological features indistinguishable from those of XSCID was therefore selected for analysis. An Epstein-Barr virus (EBV)-transformed cell line derived from her lymphocytes had normal γc expression but lacked Jak3 protein and had greatly diminished Jak3 messenger RNA. Sequencing revealed a different mutation on each allele: a single nucleotide insertion resulting in a frame shift and premature termination in the Jak3 JH4 domain and a nonsense mutation in the Jak3 JH2 domain. The lack of Jak3 expression correlated with impaired B cell signaling, as demonstrated by the inability of IL-4 to activate Stat6 in the EBV-transformed cell line from the patient. These observations indicate that the functions of γc are dependent on Jak3 and that Jak3 is essential for lymphoid development and signaling.

an increase in the ratio of CD4+ cells to CD8+ cells. This observation suggests that cytokine signaling may be more critical for the maturation or survival of CD8 lineage T cells than CD4 lineage T cells.
Previous biochemical studies of Jak3 indicated an association with IL-2 receptor signaling (1,2). Furthermore, inhibitors of IL-2 receptor signaling block T cell proliferative responses (14). Our results suggest that Jak3 is essential for IL-2 receptor signaling in primary T cells. However, Jak3 may not be the only tyrosine kinase required for IL-2 receptor signaling, because both Jakl and p56"k associate with the IL-2R f chain (2,15). Splenic T cells from -/mice, when stimulated, produce much less IL-2 than wild-type splenic T cells. This reduction in IL-2 secretion may result from a nonresponsiveness, or anergy, induced in the -/-T cells as a result of prior T cell receptor stimulation in the absence of Jak3. This possibility is compatible with previous studies showing that activation of Jak3 is associated with the prevention of anergy (16). Consistent with this, we also observed increased expression of activation markers on the vast majority of -/splenic T cells, suggesting a prior T cell receptor activation event. One possibility is that most T cells routinely encounter a T cell receptor stimulation signal, perhaps as a component of the mechanism inducing thymic emigration (17), and that -/-T cells are deficient in returning to the normal resting state. These unusual aspects of T cell development and differentiation in -/mice indicate previously undescribed functions of cytokine receptor signaling pathways. W. L. Farrar, J. Biol. Chem. 269,19136 (1994); S. G. Rane and E. P. Reddy, Oncogene 9, 2415Oncogene 9, (1994; T. Takahashi and T. Shirasawa, FEBS Lett. 342, 124 (1994); C. B. Gurniak and L. J. Berg, in preparation. 4. J. J. Peschon et al., J. Exp. Med. 180, 1955 Males with X-linked severe combined immunodeficiency (XSCID) have defects in the common cytokine receptor -y chain (-y) gene that encodes a shared, essential component of the receptors for interleukin-2 (IL-2), IL-4, IL-7, IL-9, and IL-15. The Janus family tyrosine kinase Jak3 is the only signaling molecule known to be associated with yC, so it was hypothesized that defects in Jak3 might cause an XSCID-like phenotype. A girl with immunological features indistinguishable from those of XSCID was therefore selected for analysis. An Epstein-Barr virus (EBV)-transformed cell line derived from her lymphocytes had normal -yc expression but lacked Jak3 protein and had greatly diminished Jak3 messenger RNA. Sequencing revealed a different mutation on each allele: a single nucleotide insertion resulting in a frame shift and premature termination in the Jak3 JH4 domain and a nonsense mutation in the Jak3 JH2 domain. The lack of Jak3 expression correlated with impaired B cell signaling, as demonstrated by the inability of IL-4 to activate Stat6 in the EBV-transformed cell line from the patient. These observations indicate that the functions of yc are dependent on Jak3 and that Jak3 is essential for lymphoid development and signaling.
The -yc chain is an essential signaling component of receptors for IL-2 (1), IL-4 (2, 3), IL-7 (4, 5), IL-9 (6, 7), and IL-15 (8). Defects in this chain cause XSCID (9, 10), a disease characterized by impaired function of B cells and a complete or almost complete deficiency ofT cells (10). Thus, XSCID occurs as a result of inactivation of numerous cytokine signaling pathways, an observation compatible with the severe phenotype of XSCID (9, 10) and of ,yc-deficient mice (11). IL-2, IL-4, IL-7, IL-9, and IL-15 all activate the same Janus family kinases, jakl and Jak3 (6,12). In each case, Jakl associates with the receptor chain that SCIENCE * VOL. 270 * 3 NOVEMBER 1995 plays the major role in determining both the cytokine binding specificity (6,13) and the STAT (signal transducers and activators of transduction) proteins that are activated (14,15). Jak3 is primarily associated with y,, and as no other signaling molecules have been identified that associate with yc, we hypothesized that the phenotype resulting from defects in Jak3 might be indistinguishable from that resulting from defects in yz, and that the key role of yc may be to bring Jak3 into proximity with the primary binding chain and its associated signaling molecules (6).
Although true XSCID (-yC-I-) females 797 m could be bom [for example, to XSCID carrier (,Y:+/-) females and healthy (yc+/rY) fathers in whom a -yc mutation occurred during spermatogenesis], such individuals would be expected to be quite rare. Because the gene for Jak3 is not on the X chromosome (16), we hypothesized that females with clinical and immunological features typical of XSCID might instead have defects in Jak3. A girl with features typical of XSCID was identified (Table 1, patient AP), and an EBVtransformed B cell line was generated from her peripheral blood lymphocytes (PBLs). We then examined protein lysates from these cells for yc expression (Fig. 1A). In contrast to EBV-transformed cells from a lyc-deficient XSCID patient, AP had yc expression equivalent to the normal control EBV-transformed cell line from individual PN. AP also had normal expression of Jakl (Fig. 1B); however, AP did not express detectable amounts of Jak3 protein (Fig. iC). Protein immunoblot analysis of lysates from EBV-transformed cells derived from the parents (GP and JP, both of whom were healthy) showed substantial Jak3 expression (Fig. 1D). The lack of Jak3 in AP was explained by the observation that Jak3 mRNA was undetectable in AP cells by Northern (RNA) blot analysis (Fig. 1E) in contrast to -yc mRNA, which was normal (Fig. 1F).
Despite the absence of Jak3 mRNA as detected by Northern blot, trace amounts of Jak3 RNA were identified by reverse transcriptase-polymerase chain reaction (RT-PCR) after 40 cycles of amplification, allowing us to sequence the entire coding region of jak3 from both AP and normal controls. Consistent with this girl's form of SCID being an autosomal trait in which a different mutation was inherited from each parent, she had mutations at two positions (ml and m2, Fig. 2A) (17). A single base insertion of a 0 nucleotide was found in the JH4 domain at nucleotide 1172 (amino acid 391) that resulted in a frame shift and subsequent stop codon at amino acid 408 (ml, Fig. 2B), and a point mutation was found at nucleotide 1695 (amino Table 1. Comparison of clinical and immunological parameters of patient AP and normal control individuals. AP was healthy until age 4 months and then developed recurrent otitis media, frequent diarrhea, poor weight gain, and two episodes of monilial dermatitis. At age 9.5 months, she developed a right upper lobe pneumonia and was found to have hypogammaglobulinemia, lymphopenia, and a lack of visible or palpable Iymphoid tissue. At 10 months of age, she was referred to Duke University Medical Centerfor possible bone marrow transplantation. Immunologic evaluation revealed normal serum immunoglobulin M (IgM) but no detectable serum IgG or IgA. Her total lymphocyte count was 2,91 2/mm3 (normal >4,500/mm3), of which 89% were CD20+ B cells and 7% were CD3+ T cells. Her lymphocytes exhibited low responses to mitogens and allogenic cells in (3H]thymidine incorporation assays. Although 1% of her peripheral blood tymphocytes was CD16+, natural killer (NK) cell functional studies revealed no killing of K562 erythroleukemia cells, analogous to the absence of NK cells in XSCID. For all assays, n = 167 fornormal controls. N.d., not detected; PHA, phytohemagglutinin; Con A, concanavalin A; PWM, pokeweed mltogen.  (19). Thus, it was assumed that the amount of RNA derived from the mutant allele would be much less than that derived from the wild-type allele in AP's parents, so genomic DNA from EBV-transformed cells from the mother (JP) and father (GP) was amplified. Direct sequencing indicated a homozygous TGC codon in PN and GP, whereas AP and JP were both heterozygous for TGC and TGA at this position, indicating that the stop codon mutation (m2) was inherited from the mother, and the insertion mutation (ml) was inherited from the father (i8).

Concentration of serum Ig
The correlation between the lack of Jak3 expression and greatly diminished T cells in this patient indicates a role for Jak3 in T cell development. This diminished immunoglobulin synthesis makes sense in view of the SCIENCE * VOL. 270 * 3 NOVEMBER 1995 diminished T cell help in the patient. To determine whether there was also an intrinsic B cell defect, we investigated whether Stat6 could be activated by IL-4 in the EBV-transformed B cell line from AP. Stat6 was activated by IL-4 in the PN but not in the AP EBV line (Fig. 3A), demonstrating an apparent defect in B cell signaling and that Jak3 was required for Stat6 activation in these cells. Stat6 is ubiquitously expressed (14), and the lack of Stat6 activation by ILA was not due to a lack of Stat6 protein in AP cells (Fig.  3B). We also observed the same defect in Stat6 activation by IL-4 in XSCID cell lines (18). Thus, the defects in B cell function seen in the patient AP and in XSCID patients are probably due to deficiencies in B cell signaling as well as to deficiencies in T cell help. It is interesting that B cell numbers are normal or increased in humans with XSCID and Jak3 deficiency, but are essentially absent in mice with y1 deficiency (11) and in young (4to 6-week-old) mice with Jak3 deficiency (20). This is probably because IL-7 is an essential pre-B cell growth factor in mice but not in humans (21).  The identification of defective Jak3 expression in a case of SCID allows a number of conclusions to be drawn. First, the diminished number of T and natural killer (NK) cells indicates the important role played by Jak3 in the development of these lineages. Second, because the phenotype of jak3-deficient SCID is indistinguishable from that of XS-CID, most if not all -y.-dependent signals may be dependent on Jak3. Third, the observation that the Jak3 deficiency was not more severe than XSCID suggests that Jak3 may not be essential for signaling from cytokines other than those whose receptors contain yc. The cell type-restricted expression of Jak3 (22), and the fact that Jak3 appears only to be required for signaling through yc-containing receptors, is in contrast to other Jak family kinases, which are ubiquitously expressed and associate with multiple cytokine receptors. Fourth, the lack of Stat6 activation by IL-4 in AP cells indicates that Jak3 plays an important role in STAT protein activation; indeed, it is possible that the loss of STAT protein activation significantly contributes to the phenotype in XSCID and Jak3-deficient SCID. Because STAT activation is only one type of signal induced by IL-2, IL-4, IL-7, IL-9, and IL-15, it will be interesting to determine the role of Jak3 in activating the other signaling pathways.
The possibility of using pharmacological agents that block -yc-Jak3 association as immunosuppressants has been suggested (6). The current study further suggests that any agents that inactivate Jak3 function may be potent immunosuppressants. Moreover, the identification of Jak3 deficiency as the molecular basis for some autosomal recessive cases of SCID will allow the development of diagnostic procedures for identification of Jak3-deficient SCID patients and carriers and investigation into the possibility of gene therapy for Jak3-deficient SCID patients, analogous to that being developed for XSCID (23). . It should be noted that some transcripts from the patient contained insertions (67, 78, or 1 14 nucleotides) after nucleotide 1441. These insertions appeared to arise by altemative splicing because they represent intronic sequences beginning 230, 194, or 183 nucleotides 3' of the exon-intron boundary at nucleotide 1441 and extend to a common endpoint. Although the insertions contain in-frame stop codons, the altemative splicing does not represent a diseasecausing difference in AP as no mutations in the intron or surrounding exon sequence were found, and these splice variants were also found in other RNA samples this question and to determine whether Jak3 is critical for signaling in other lineages, we developed mice lacking Jak3.
We constructed a Jak3 targeting vector to disrupt the first coding exon, which created a null allele for expression (Fig. 1A). The construct was electroporated into E14 embryonic stem (ES) cells, and four independent clones with normal karyotypes were injected into C57BL/6 blastocysts to create chimeric mice. Chimeric mice from two clones transmitted the targeted allele, and heterozygous mice from the two clones were separately bred to create homozygous mutant (-/-) mice. Genotyping of 59 progeny yielded 16 wild-type (+/+) mice, 30 heterozygous (+/-) mice, and 13 -/mice; this correspondence to the expected ratios ( 1: 2:1 ) indicated the lack of an effect on embryonic development. The -/mice were indistinguishable from their littermates and thrived comparably under specific pathogen-free conditions. No Jak3 protein was detectable in splenic, thymic, or bone marrow extracts from the -/mice (6).
Morphologically, the lymphoid tissues of the -/mice were markedly different from Wild-type Jak3 Maps of the Jak3 locus (top), the targeting con-B H EEE Targetingvector struct in pBluescript 11 Hyg t0 thymidine kinase (tk) gene cassette, containing the same promoter as in the Hyg cassette, was inserted into the 3' end of the targeting vector for negative selection. The targeting vector contained 5.8-and 3.2-kb Jak3 fragments. Electroporation of the linearized plasmid into 129derived El 4 ES cells and screening by Southern (DNA) blot analysis for homologous recombination were as described (14); the 3' flanking probe used in Southern blot analysis is shown as a bar. The efficiency of homologous recombination was 96%. (B) Southern blot analysis of mouse tail DNA. Genomic DNA from +/+ mice (lane 1), +/-mice (lanes 2 through 5, 7, and 9), and -/mice (lanes 6 and 8) was digested with Hind IlIl and probed with a 1.4-kb Eco RI-Hind IlIl fragment. The 5.0-and 7.0-kb bands represent the wild-type and mutated alleles, respectively. When the blot was rehybridized with a Hyg probe, only the 7.0-kb band was detected (15). Bam Hi digests probed with another 3' probe and Eco RI digests probed with a 0.85-kb Sst complementary DNA probe containing the first ATG further confirmed appropriate homologous recombination (15). SCIENCE * VOL. 270 * 3 NOVEMBER 1995 ..l