Unexpected and Variable Phenotypes in a Family with JAK3 Deficiency

DM Frucht1, M Gadina1, GJ Jagadeesh2, I Aksentijevich1, K Takada1, JJH Bleesing3, J Nelson1, LM Muul2, G Perham4, G Morgan5,7, EJA Gerritsen5,8, RF Schumacher6, P Mella6, PA Veys5, TA Fleisher3, ER Kaminski4, LD Notarangelo6, JJ O’Shea1 and F Candotti2 1Arthritis and Rheumatism Branch, National Institute of Arthritis, Musculoskeletal and Skin Diseases, Bethesda, MD, USA; 2Genetics and Molecular Biology Branch, National Human Genome Research Institute, Bethesda, MD, USA; 3Department of Laboratory Medicine, Clinical Center, National Institutes of Health, Bethesda, MD, USA; 4Derriford Hospital, Plymouth, UK; 5Great Ormond Street Hospital, London, UK; 6Department of Pediatrics, University of Brescia, Italy

Mutations of the Janus kinase 3 (JAK3) have been previously described to cause an autosomal recessive variant of severe combined immunodeficiency (SCID) usually characterized by the near absence of T and NK cells, but preserved numbers of B lymphocytes (T-B+SCID). We now report a family whose JAK3 mutations are associated with the persistence of circulating T cells, resulting in previously undescribed clinical presentations, ranging from a nearly unaffected 18-year-old subject to an 8-year-old sibling with a severe lymphoproliferative disorder. Both siblings were found to be compound heterozygotes for the same deleterious JAK3 mutations: an A96G initiation start site mutation, resulting in a dysfunctional, truncated protein product and a G2775(+3)C mutation in the splice donor site sequence of intron 18, resulting in a splicing defect and a predicted premature stop. These mutations were compatible with minimal amounts of functional JAK3 expression, leading to defective cytokine-dependent signaling. Activated T cells in these patients failed to express Fas ligand (FasL) in response to IL-2, which may explain the accumulation of T cells with an activated phenotype and a skewed T cell receptor (TcR) V␤ family distribution. We speculate that residual JAK3 activity accounted for the maturation of thymocytes, but was insufficient to sustain IL-2-mediated homeostasis of peripheral T cells via Fas/FasL interactions. These data demonstrate that the clinical spectrum of JAK3 deficiency is quite broad and includes immunodeficient patients with accumulation of activated T cells, and indicate an essential role for JAK3 in the homeostasis of peripheral T cells in humans. Genes and Immunity (2001) 2, 422–432.

Keywords: JAK3; SCID; IL-2; Fas ligand; lymphoproliferation

Introduction largely accounts for the lack of thymocyte development in mice and humans lacking JAK3.9–11 Additionally, The Janus kinase 3 (JAK3) member of Janus kinases JAK3-deficient humans and mice present with impaired ␥ ␥ associates with the common chain ( c) of the IL-2, 4, 7, NK cell development that is most likely due to defective 9, and 15 receptors and is necessary for normal down- IL-15 signaling, as IL-15 knockout mice have a nearly stream signaling by members of the signal transducer complete absence of peripheral NK cells.12 1,2 and activators of transcription (STAT) family. The cen- We hypothesized that the phenotype of patients with tral role of JAK3 in T cell development is highlighted by partial JAK3 deficiency might be quite different from that patients with JAK3-deficiency and severe combined of patients reported thus far with classical T-B+ SCID. immunodeficiency (SCID) due to the lack of circulating T Partial signaling through the IL-7 receptor in this case + 3–5 cells and presence of dysfunctional B cells (T-B SCID). might be sufficient to sustain thymic development of T Targeted disruption of the Jak3 gene in mice also leads lymphocytes and to allow their appearance in the periph- to profound immunodeficiency, but with defects in both eral circulation where they could manifest consequences 6–8 T and B cell development. The T cell defect of JAK3- of defective signaling in other JAK3-dependent path- deficient humans and mice resembles that of IL-7 and IL- ways. In this regard it is worth noting that mice with ␣ ␣ ␣ 7 receptor chain (IL7R ) null mice and IL-7R deficient targeted disruptions of IL-2 or its receptor ␣ or ␤ chains SCID patients, suggesting that defective IL-7 signaling are characterized by autoimmunity and/or lymphoproliferation, despite the well-established role of IL-2 in cellular proliferation.13–15 This is thought to be a consequence Correspondence: DM Frucht, MD, Cell Biology Laboratory, Building 29B, of the lack of IL-2-mediated apoptotic signals for acti- Room 3NN22 Bethesda, MD 20892, USA. E-mail: fruchtȰcber.fda.gov 7 vated T lymphocytes. Depending on the context, there- Current address: Medical Research Council Laboratories, PO Box fore, IL-2 can have seemingly opposing effects. Although 273, Banjul, The Gambia, West Africa. 8Current address: Department of Pediatrics, Medical Center poorly defined, a role for JAK3 in mediating IL-2-depen- Rijnmond-South, Rotterdam, The Netherlands. dent apoptosis signals is suggested by the presence of Received 19 June 2001; revised and accepted 9 August 2001 activated/memory and auto-reactive T cells in JAK3- Unexpected phenotypes in JAK3 deficiency DM Frucht et al 423 deficient mice16,17 and the development of activated T tration of the gut, bone marrow (Figure 2a), and liver lymphocytes in one patient with JAK3 deficiency.18 These (Figure 2b) was demonstrated which was characterized T cells also are characterized by a limited T cell receptor by a predominance of oligoclonal CD8+ T cells (data not repertoire, consistent with oligoclonal peripheral expan- shown). Attempts to treat the patient’s condition with sion of the few thymocytes that escape the developmen- azathioprine and steroids were unsatisfactory, ultimately tal blockade.17,18 leading to treatment with allogeneic bone marrow trans- We have been screening patients whose immunodefi- plantation (BMT) from his HLA-identical sister (II,3) at ciency status could be due to partial JAK3 deficiency and, age 9. Although engraftment was demonstrated, the in this report, we describe a family carrying mutations patient died from multi-organ failure complicating a sep- leading to severely reduced levels of functional JAK3, but sis episode 5 weeks after BMT. Pre-treatment DNA and allowing residual cytokine signaling and the develop- EBV-B cells were available for analysis. ment of peripheral T cells. Two siblings in this family had variable clinical presentations ranging from overt to Immunological findings very mild immunodeficiency, despite carrying identical The immunological phenotype of patient II,2 on admit- JAK3 mutated alleles. These patients had elevated num- tance to the Clinical Center, NIH is reported in Table 1 bers of activated peripheral lymphocytes, and severe and together with that of her brother (II,4) at 6 months, 4 and progressive multi-organ lymphocytic infiltration was 7 years of age. Both patients II,2 and II,4 demonstrated T observed in one case. Studies of FasL expression in lym- cell lymphopenia primarily due to a reduced number of phocytes from one of these patients showed defective up- CD4+ T cells. In patient II,2, both CD4+ and CD8+ T sub- regulation in response to IL-2 that could help explain sets showed a large reduction in the number of na¨ıve these findings. cells, identified by dual expression of CD45RA and CD62L, a finding consistent with reduced thymic output. In addition, patient II,2 had a high proportion of ␥␦ T Results cells in her peripheral blood, all of which were of the V␦2 subset. This finding was consistent with peripheral Clinical history of patients expansion of these cells, which can develop via extra- Four children resulted from the non-consanguineous thymic pathways.19 Such an expansion was not apparent union of two northern European parents. No biological in patient II,4. One characteristic of T cells in both samples were available from the first male child (Figure patients was the high percentage of T cells that expressed 1, subject II,1) who died at 10 months of age from Pneum- HLA-DR, indicating increased activation compared to ocystis carinii pneumonia (PCP). normal controls. In addition, the NK cell counts in both The medical history of subject II,2 included an unde- patients were extremely low, whereas both B cell num- fined lung infection at age 1, a severe varicella infection bers were within the normal range or elevated (Table 1). and occasional ear and urinary tract infections until age 7. The level of serum immunoglobulin (Ig) in patient II,4 at During laboratory evaluations as potential bone marrow the time of his first hospitalization for PCP indicated a donor for her brother (II,4) she was found to have reduced production of IgG levels (IgG: 290 mg/dl, IgA: reduced T cell counts and was subsequently maintained 27 mg/dl, IgM: 147 mg/dl) for which he was started on on antibiotic prophylaxis. At age 9, she developed recur- monthly infusions of intravenous immunoglobulin rent cutaneous warts that spontaneously disappeared by (IVIG). Serum Ig levels in patient II,2 were consistently age 14. At the time of this study she was 18 years of age normal. No other relatives in the family were known to and in good health. have immunological problems. The second male/fourth child (II,4) had delayed Since proliferation studies performed in 1988 on umbilicus healing and developed PCP at the age of 6 samples from patient II,4 revealed defective T cell pro- months. At age 2 he developed giardiasis and, at age 3, liferation in response to PHA (Stimulation Index = 7.7, presented with pneumococcal pneumonia and a second not shown), we proceeded to determine whether this was PCP. At age 5, he developed severe varicella and, at age the case for the patient II,2 as well. Lymphocytes from 7, a recurrent Giardia lamblia infection accompanied by this patient failed to properly proliferate in response to hepatosplenomegaly and ascites. Lymphocytic infil- PHA and ConA, as well as Candida, tetanus toxoid, and diphtheria toxoid. However, these lymphocytes did proliferate in response to direct stimulation through the TcR, and this proliferation was greatly augmented by IL-2 stimulation and normalized in a 7-day proliferation assay (Table 2).

Analysis of IL-2R signaling In the attempt to characterize defects of cytokine signaling in the patient’s cells, we evaluated whether JAK3 was activated and tranduced signals in a normal manner in response to IL-2. As shown in the upper panel of Figure 3a, JAK3 was phosphorylated in response to IL-2 stimulation in PHA-activated peripheral blood mononuclear cells from the parents (subjects I,1 and I,2) of the Figure 1 Family tree of study family. Four offspring were the result affected children. In contrast, patient II,2 had an of this non-consanguineous union (*genotype analysis not extremely low response, consistent with the very low performed). Pt II,1 died of PCP. detectable levels of JAK3 protein in the immunoprecipit-

Genes and Immunity Unexpected phenotypes in JAK3 deﬁciency DM Frucht et al 424

Figure 2 Histopathology in patient II,4. A bone marrow biopsy section shows an interstitial lymphocytic inﬁltrate composed of small, slightly irregular lymphoid cells (a). A liver biopsy section shows a diffuse lymphocytic inﬁltrate with a sinusoidal distribution, forming occasional dense lobular aggregates (insert) (b).

Table 1 Immunophenotypic proﬁles

Patient II,2 Patient II,4

16 years Reference values 6 months 4 years 7 years Reference values

Lymphocytes 830 1180–2640 2380 1100 1410 1100–11900 T cells 581 (70%) 840–2030 (61–84%) 404 (17%) 616 (56%) 1170 (83%) 900–8000 (39–78%) CD4+ T cells 280 (34%) 480–1340 (33–59%) 119 (5%) 132 (12%) 268 (19%) 300–5500 (23–58%) CD4/CD45RA 17 (2%) 57–420 (5–22%) ——— — CD4/CD45RA/CD62L 11 (1%) 149–682 (13–38%) ——— — CD8+ T cells 160 (19%) 190–720 (11–36%) 214 (9%) 517 (47%) 973 (69%) 300–2300 (11–34%) CD8/CD45RA 38 (5%) 53–413 (4–22%) ——— — CD8/CD45RA/CD62L 22 (3%) 95–330 (5–20%) ——— — ␣␤ T cells 404 (49%) 650–2010 (54–76%) — 682 (62%) —— ␥␦ T cells 122 (15%) 9–166 (1–9%) — 11 (1%) —— CD3+/HLA-DR+ 207 (36%) 58–374 (5–18%) — 396 (36%) — 50–700 (1–14%) NK cells 14 (2%) 120–490 (7–30%) 214 (9%) 11 (1%) 7 (0.5%) 90–1400 (2–26%) B cells 232 (28%) 91–330 (5–16%) 1737 (73%) 429 (39%) — 200–3100 (10–44%)

Results indicate number of cells/␮l and (% of lymphocytes) obtained by ﬂow cytometry as described. —: not available.

ates (Figure 3a, lower panel). Next we evaluated the the region spanning nucleotides 1961–2996. Sequence ability of JAK3 to transduce signals to STAT5. Following analysis revealed that the lower band contained the nor- IL-2 stimulation of PBMC, tyrosine phosphorylation of mally spliced exon 18 sequence, while the upper band STAT5A was clearly detectable in cells from patient II,2, resulted from defective splicing between exons 18 and 19. though substantially reduced compared to her parents This abnormal cDNA included a 87-bp segment of (Figure 3b, upper panel). Expression levels of STAT5A intronic sequence, resulting in a predicted premature and B were similar in the patient II,2 and her parents stop 6 amino acids following the missed splice site at the (Figure 3b, lower panel). intron 18 splice donor sequence. This splicing abnor- mality was not observed in PCR products using cDNA Analysis of JAK3 sequence in cDNA and genomic from healthy controls as a template (Figure 4a, right DNA samples panel). The ﬁndings of abnormal JAK3 expression and function Sequence analysis of the RT-PCR product and the gen- in patient II,2 led us to assess whether mutations existed omic DNA corresponding to this region revealed a 2775 in the JAK3 gene of this family. Initially, cDNA was ana- (+3) G to C mutation in the consensus sequence of the lyzed from PHA-activated PBMC from the patient. As intron 18 splice donor site in both patients II,2 and II,4 seen in the left panel of Figure 4a, a doublet was evident (Figure 4b). Maternal inheritance of this allele was indi- in the PCR product generated following ampliﬁcation of cated, as this mutation was present in genomic DNA

Genes and Immunity Unexpected phenotypes in JAK3 deﬁciency DM Frucht et al 425 Table 2 Lymphocyte proliferation in members of affected family members and healthy control subjects (NL)

I,1 I,2 II,2 NL 1 NL 2

Day 3 Medium 1013 ± 146 484 ± 6 529 ± 58 461 ± 31 778 ± 139 PHA 32970 ± 11331 90379 ± 2722 479 ± 28 134890 ± 3475 78077 ± 11617 ConA 153242 ± 14659 122206 ± 5732 14216 ± 1936 137053 ± 10357 116562 ± 4963 ␣-CD3 154156 ± 3675 187997 ± 7858 63267 ± 7657 132279 ± 10920 111574 ± 4963 ␣-CD3+IL2 192637 ± 8697 212232 ± 16340 96798 ± 4878 149319 ± 3798 150185 ± 2265

Day 7 Medium 1542 ± 988 484 ± 93 252 ± 32 246 ± 81 564 ± 50 Pokeweed Mitogen 37709 ± 6639 79136 ± 11316 1233 ± 329 17062 ± 1078 13173 ± 377 Candida 52759 ± 28374 48590 ± 13624 21609 ± 25557 65625 ± 5334 103448 ± 51606 Tetanus 49555 ± 11237 54353 ± 20782 3511 ± 4517 17426 ± 3259 53399 ± 2721 ␣-CD3 9036 ± 1616 12231 ± 667 3021 ± 364 11192 ± 957 6434 ± 661 ␣-CD3+IL2 23490 ± 944 26674 ± 4359 23187 ± 11919 15816 ± 229 8961 ± 1392

The results are the mean ± s.d. obtained from proliferation assays performed in triplicate.

some residual IL-2-mediated phosphorylation (Figure 5a, upper left panel). Patient II,4 had no detectable JAK3 protein or JAK3 phosphorylation. In addition, the father (I,1), daughter (II,2), and son (II,4) expressed a smaller molecular weight protein that was immunoprecipitated by the C-terminal anti-JAK3 antibody. This smaller protein was not present in the control EBV-B cell line. In the case of patient II,4, this truncated protein was only detectable during anti-phosphotyrosine blotting, likely due to the increased sensitivity of this assay (Figure 5a, upper right panel). In the case of Figure 3 IL-2 receptor signaling analysis in study family. PBMC the father (I,1) and the daughter (II,2), this truncated pro- were isolated from the indicated family members and were either tein was also readily detectable following immunoprecip- + − stimulated with IL-2 for 15 minutes ( ) or left unstimulated ( ). itation or direct Western analysis with the C-terminal Protein lysates were immunoprecipitated with anti-JAK3 antibody, and the resulting products were separated on a SDS-PAGE gel and anti-JAK3 (Figure 5a and 5b, upper panels). Since this transferred. These blots were analyzed for phosphotyrosine (a, smaller protein was not detected with an N-terminal anti- upper panel) or JAK3 expression (a, lower panel). Protein lysates JAK3 (not shown), these findings were consistent with were also immunoprecipitated with anti-STAT5 antibody and ana- the paternally inherited mutation in the JAK3 initiation lyzed for STAT5 expression and IL-2-induced phosphorylation in start site. An alternative translation initiation at a down- a parallel manner (b, upper and lower panels). stream methionine likely accounted for the production of the observed N-terminally truncated JAK3 protein. Inter- from the mother (Figure 4b). To rule out the possibility estingly, this smaller JAK3 protein showed no difference that this nucleotide change was a polymorphism, we in its level of tyrosine phosphorylation following IL-2 screened normal European donors for the presence of this treatment, suggesting that it failed to propagate IL-2- mutation by an HphI restriction digest assay. The mutant dependent signals (Figure 5a, upper left and right restriction pattern was not detected in 100 normal chro- panels). As previously mentioned, the maternally mosomes screened by this method (data not shown). inherited mutations would be predicted to result in a C- Additionally, genomic sequencing analysis led to the terminally truncated protein. However, no evidence was discovery of a paternally-inherited mutation in the trans- found for expression of maternally inherited JAK3 allele lation initiation codon of JAK3 (96A to G). This mutation in either sibling (II,2 or II,4) using an N-terminal anti- was present in the father (subject I,1) and patient II,2 and JAK3 antibody (not shown). II,4 (Figure 4b). Next we examined STAT5A phosphorylation in response to IL-2 stimulation in these cell lines. STAT5A Analysis of JAK3 protein and signaling function in B phosphorylation was low in patient II,2 and absent in cell lines patient II,4 compared to the normal control or their father In an effort to confirm our findings in primary cells and (I,1), although the expression levels of STAT5A were nor- better analyze the JAK3 mutant protein in terms of mal (Figure 5c, upper and lower panels). Thus, the levels expression and function, EBV-transformed B cell lines of STAT5A phosphorylation correlated well with the were prepared from patient II,2 and her father, subject level of normal-sized, functional JAK3 present in these I,1. Additionally, several experiments were performed on cell lines. EBV-transformed B cell lines from patient II,4 before the cell line was permanently lost due to an equipment mal- TcR V␤ usage function. As observed in peripheral lymphocytes, EBV-B We next proceeded to characterize the phenotype of the cells from patient II,2 showed severely reduced T lymphocytes that developed in the surviving affected expression of JAK3 (Figure 5a, lower left panel), with child despite the severe reduction in functional JAK3 pro-

Genes and Immunity Unexpected phenotypes in JAK3 deﬁciency DM Frucht et al 426

Figure 4 JAK3 sequence analysis of study family members. cDNA was prepared from PHA-stimulated PBMC isolated from patient II,2. PCR primers corresponding to the indicated regions of the JAK3 coding sequence were utilized to amplify overlapping regions of the expressed JAK3 gene (left panel). The DNA corresponding to each band of the doublet in the lane representing the 1961–2996 ampliﬁcation was isolated and sequenced. Also a direct comparison of RT-PCR products from ampliﬁcation of 2654–2996 was performed between patient II,2 and a healthy control as indicated (right panel, M: Marker, PT: patient II,2) (a). Genomic DNA from the indicated family members was sequenced from exon 1 (reverse strand is shown, left panels) and exon/intron 18 (right panels). Published cDNA sequence (accession number U09607) was used as a reference for base numbering. Mutations are indicated by the arrow (b). (c). Genomic PCR products generated from subjects heterozygous for the maternally inherited mutation had both cut (259 and 104 base pair fragments) and uncut DNA (363 base pair fragment) following HPHI digestion, whereas PCR products from unaffected controls (NL) were completely digested (M: marker).

tein. Analysis of the TcR V␤ repertoire in patient II,2 (Figure 6a). Additionally, heteroduplex analysis of the V␤ demonstrated that all V␤ families were transcribed by her segments ampliﬁed from a healthy control generated a peripheral blood T lymphocytes. The over-representation smear pattern indicative of highly heterogeneous molecu- of some V␤ families (eg, V␤8), however, suggested pref- lar species,20 whereas discrete bands super-imposed on erential expansion of oligoclonal T cell populations a smear pattern were visible in most of the V␤ families

Genes and Immunity Unexpected phenotypes in JAK3 deﬁciency DM Frucht et al 427

Figure 5 JAK3 Expression and IL-2 signaling in B cell lines. Signal transduction was analyzed in EBV-transformed B cells derived from the indicated patients or a healthy control (NL) unstimulated (−) or stimulated (+) with IL-2. JAK3 tyrosine-phosphorylation (upper left and right panels) and expression (lower left and right panels) are shown following immunoprecipitation with a C-terminal anti-JAK3 antibody and Western blotting (a). JAK3 (upper panel) and STAT5 (lower panel) expression in patient and healthy control EBV-lines are also demonstrated in a Western blot of 100 ␮g of protein from whole cell lysates (b). Protein lysates were also immunoprecipitated with anti-STAT5 a antibody and analyzed for STAT5 expression and IL-2-induced phosphorylation (c, upper and lower panels).

analyzed in patient II,2. Taken together, these data suggest that patient II,2 had a polyclonal T cell repertoire with evidence of oligoclonal expansion (Figure 6a, inset). This possibility was conﬁrmed by sequence analysis of TcR V␤8 and V␤9 clones from patient II,2 that showed a biased utilization of TcR V␤ junctional regions (Table 3).

Defective upregulation of Fas ligand in IL-2- stimulated patient cells The presence of significant numbers of peripheral T cells with an activated phenotype (patients II,2 and II,4) and the development of a multi-organ lymphocytic infiltration in patient II,4, were notable features in this kindred affected with JAK3 deficiency. Given the results of a recent report showing that restoration of normal activation-induced cell death (AICD) in T lymphocytes from Il-2R␤ knockout mice occurred only in association with functional STAT5 signaling,21 we hypothesized that IL-2- dependent upregulation of pro-apoptotic molecules might be JAK3-dependent. One critical pro-apoptotic pathway regulated by IL-2 is Fas/FasL, abrogation of which results in lymphoproliferative disease and autoimmunity in humans.22 Western blot. analysis of FasL expression in normal PBMCs showed low levels of FasL in unstimulated conditions, but greatly increased expression following 24 h of stimulation with IL-2. In contrast, neither unstimulated nor IL-2-treated PBMCs from patient II,2 expressed detectable levels of FasL, indicating that this pathway of upregulation was defective in these cells expressing extremely low levels of functional JAK3 (Figure 6b).

Discussion The requirement for JAK3 in thymic development is Figure 6 TcR and FasL expression analysis. TcR V␤ family usage clearly evident in patients with absolute deficiency in is shown as percentage of total TcR expression in PBMC from JAK3 function, who have nearly undetectable peripheral patient II,2 and a healthy control subject (NL). Heteroduplex analy- T cells.3–5 This phenotype is paralleled in Jak3 knockout sis of selected V␤ families is presented in the inset (a). Analysis of mice which show a profound reduction of early thymic Fas ligand regulation in patient cells. PBMC were cultured in the 6–8,17 + − progenitor cells. Taken together, the phenotype of presence ( ) or absence ( ) of IL-2 for 24 hours. Protein lysates from complete JAK3 deficiency in humans and mice estab- patient II,2 or a healthy control (NL) were analyzed by Western blotting for the expression of ෂ40 kD Fas ligand protein (FasL). The lishes the importance of JAK3 in the maturation of T cells. ෂ75 kD protein observed in all lanes is the result of non-specific Most of the patients described in previous studies lacked (NS) antibody binding (b). detectable JAK3 activity, preventing peripheral T cell

Genes and Immunity Unexpected phenotypes in JAK3 deﬁciency DM Frucht et al 428 Table 3 Amino acid sequence analysis of TcR V␤ junctional regions from PBMC of patient II,2

TcR Vß8 n-D-n TcR ßJ TcR ßC Frequency

SAVYFCASS SGLVTY NEQFFGPGTRLTVL EDLK 11/21 SAVYFCAS RLNATGR DTQYFGPGTRLTVL EDLK 3/21 SAVYFCASS PMGGPSN EKLFFGSGTQLSVL EDLN 1/21 SAVYFCASS ST NEQFFGPGTRLTVL EDLK 1/21 SAVYFCASS LWDWETE DTQYFGPGTRLTVL EDLK 1/21 SAVYFCASS FWGV EAFFGQGTRLTVV EDLN 1/21 SAVYFCASS PGQD EQYFGPGTRLTVT EDLK 1/21 SAVYFCAS RRDRGL ETQYFGPGTRLLVL EDLK 1/21 SAVYFCAS TLGQGASN TGELFGFFGEGSRLTVL EDLK 1/21

TcR Vß9 n-D-n TcR ßJ TcR ßC Frequency

SAVYFCAS TLGQGASN TGELFFGEGSRLTVL EDLK 14/23 SAVYFCASS QDPGN YEQYFGPGTRLTVT EDLK 3/23 SAVYFCASS QTPAEY NEQFFGPGTRLTVL EDLK 2/23 SAVYFCASS VTHWDRES EKLFFGSGTQLSVL EDLK 1/23 SAVYFCASS HRGG YEQYFGPGTRLTVT EDLK 1/23 SAVYFCASS LAGGMGA KNIQYFGAGTRLSVL EDLK 1/23 SAVYFCASS QDLPIAGGFMF YEQYFGPGTRLTVT EDLK 1/23

development, and prohibiting examination of the role of patients, it is unlikely that such rare events would arise JAK3 in the function of peripheral T lymphocytes.3–5,23 in two members of the same family. In addition, revertant The JAK3-deficient patients described here are com- T lymphocytes would be expected to have a strong selec- pound heterozygotes carrying two previously unde- tive advantage and to express JAK3 to levels comparable scribed JAK3 mutations, whose phenotypes indicate a to the heterozygous parents. On the contrary, only barely broader spectrum of disease than previously recognized. detectable levels of JAK3 could be immunoprecipitated Affected individuals in this family express a very small from PBMCs from patient II,2 (Figure 3a), indicating an amount of functional JAK3 protein, while having overall deficiency in JAK3. We therefore conclude that mutations in both alleles of JAK3. This is likely due to most, if not all, T lymphocytes present in peripheral occasional correct recognition of the intron 18 splice blood from patient II,2 have developed despite their donor site in the maternal allele despite the mutation in extremely low levels of JAK3 expression. This would not the consensus sequence. The truncated protein that be entirely unexpected, as patients with attenuated ␥c results from the paternal allele, in contrast, is predictably deficiency develop T cells as well.28–30 non-functional as the N-terminus has been shown pre- Despite being in substantial numbers, the T cells that viously to be required for association with ␥c.24,25 Patient develop in patients from this family are abnormal in II,4 carried the identical mutations as patient II,2, yet no some respects. For example, the T cells that are present expression of functional JAK3 was detected in B cell lines in the periphery of patient II,2 have a skewed T cell reper- generated from this patient. However, the presence of toire. This finding is consistent with peripheral oligo- peripheral T cells in this patient does suggest that low clonal expansion, similar to that observed in Jak3 −/− level JAK3 expression had occurred at least some time mice.16,17,31–34 In addition, the finding of elevated ␥/␦ T during T cell development. Whether a difference in JAK3 cells indicates a reduced thymic output in this patient. expression levels between the siblings resulted in differ- One conclusion that can be made from identification of ences in disease severity is unknown, but it is an JAK3 mutations in these patients is that the phenotype of intriguing possibility. Low and variable levels of JAK3 JAK3 deficiency is more heterogeneous than previously expression could account for the unique features of affec- recognized, ranging from the clinical presentation of ted individuals in this family: T cell counts in the low SCID3–5,23 to mild immunodeficiency (patient II,2). The range, mild immunodeficiency with increased but not developmental effects of JAK3 deficiency vary from the overwhelming susceptibility to infections, and the per- virtual absence of circulating T cells3–5,23 to the develop- sistence of activated T cells. Interestingly, NK cell counts ment of substantial numbers of peripheral T lymphocytes were low in both patients II,2 and II,4, suggesting that (Table 1 and Brugnoni et al18), which may even lead to NK cell development is more dependent on JAK3 sig- lymphoproliferative syndromes as in patient II,4 naling than is T cell development. (Figure 2). This variation occurs even among patients It must be noted that spontaneous reversion of genetic with identical JAK3 mutations, as is evident in the kin- defects in lymphocytes has been demonstrated to be the dred that we describe here. Because of his early death basis of the unexpectedly mild immunological pheno- from PCP, it is tempting to speculate that patient II,1 was types in ADA deficiency and X-linked-SCID patients.26,27 also carrying the JAK3 mutations found in the other The occurrence of genetic reversions in patients II,2 and members of the family. If this very likely possibility were II,4 therefore could have potentially explained their true, the clinical heterogeneity allowed by JAK3 atypical immune phenotypes and the development of mutations within this family would be even greater than peripheral T lymphocytes. Although we have not for- what we currently appreciate. We do not know the expla- mally ruled out the presence of genetic revertants in these nation for this clinical heterogeneity, but speculate that

Genes and Immunity Unexpected phenotypes in JAK3 deficiency DM Frucht et al 429 other genetic factors outside of JAK3 mutations may play expands our understanding of the clinical presentation of a role in the disease phenotype.35 genetic defects in JAK3, as it establishes that JAK3 Another very surprising finding in this study was that deficiency may be present in patients with mild immuno- despite the presence of minimal amounts of JAK3, patient deficiencies and relatively normal peripheral T cell T cells proliferated in response to IL-2 stimulation. More- counts. Not surprisingly, however, JAK3-deficient T cells over, the addition of T cell receptor stimulation to IL- are not fully normal. Interestingly, they lack normal regu- 2 treated cultures apparently normalized proliferation of lation of FasL, likely explaining the persistence of acti- patient cells by 7 days following stimulation (Table 2). vated peripheral T cells in affected patients. Taken Whether decreased IL-2-induced apoptosis plays a role together these data have important implications for the in this normalization is unknown, but possible. Further- evaluation of primary immunodeficiencies, as well as more, IL-2-dependent proliferation occurred despite the establish a new role for JAK3 in IL-2-mediated T cell fact that IL-2-dependent STAT5 phosphorylation was homeostasis in humans. strikingly low. Perhaps activated PBMC produce factors that complement the JAK3 defect in a manner similar to the correcting effect of IL-3 in Jak3-deficient mice.36 Patients and methods Despite the data indicating the importance of STAT5 in IL-2- and JAK3-dependent lymphocyte proliferation,37 it Subjects appears that only a small amount of activation of this The subjects described in this study provided written, signal transduction pathway is required to effect this informed consent for participation in National Institute function. Thus, once T cells progress through thymic of Arthritis and Musculoskeletal and Skin Diseases Proto- maturation even very low levels of functional JAK3 can col 99-AR-0004. Following a medical interview and sustain their proliferation upon activation. physical examination, the patients’ peripheral blood Additionally, the finding that the patients’ cells pro- mononuclear cells (PBMC) were harvested by phleb- liferate in response to IL-2 in a relatively normal manner, otomy or automated apheresis using standard protocols. especially with concomitant T cell receptor engagement, has important implications. It is well known that IL-2 Cell purifications and culture receptor signal transduction pathways lead to apoptosis Both heparinized whole blood and crude apheresis pro- as well as proliferation.38 Indeed the pro-apoptotic role duct underwent Ficoll gradient separation to obtain pur- of IL-2 may supercede its proliferative role in some set- ified PBMC that were enumerated and used fresh or cry- tings, as demonstrated in IL-2 and IL-2 receptor deficient opreserved. PBMC and lymphoblastoid B cell lines mice and humans, whose phenotype is characterized by obtained by immortalization with Epstein–Barr virus lymphoproliferation and autoimmunity.13–15,39 In our (EBV) were cultured in RPMI 1640 medium sup- experiments, JAK3-deficient patients’ T cells did not plemented with 10% fetal calf serum (FCS), 2 mM L-glu- upregulate FasL in response to IL-2 suggesting that this tamine (all from Life Technologies, Grand Island, NY, process is entirely JAK3-dependent. We speculate that USA), penicillin (100 U/ml), streptomycin (1 mg/ml), the inability to upregulate this key pro-apoptotic signal and fungizone (0.25 ␮g/ml, Biofluids, Rockville, MD, may explain the persistence of activated lymphocytes and USA). a general predisposition to uncontrolled proliferation that may have been the basis of the lymphoproliferative syn- Lymphocyte immunophenotyping drome observed in patient II,4. Our data suggest that Anticoagulated (EDTA) fresh peripheral blood samples there may exist different thresholds for IL-2-induced sur- were stained using the whole blood lysis method and vival and apoptosis. Clearly we are limited in our ability analyzed by flow cytometry (Becton Dickinson), as pre- to establish the ramifications of defective IL-2 induced viously described.42 Lymphocyte subpopulations were FasL upregulation by studying the single living affected identified with directly conjugated mouse anti-human patient from this kindred; further studies in patients with monoclonal antibodies (Mabs) specific for the following: JAK3 mutations with similar effects and/or the establish- CD2, CD3, CD4, CD8, TcR␣/␤, CD16/56, CD45RO, ment of appropriate cellular or animal models will be CD28, CD57, CD62L, HLA-DR (Becton Dickinson), informative in this regard. TcR␥/␦ (Endogen, Woburn, MA, USA), CD45RA JAK3 deficiency is being evaluated as a candidate dis- (Beckman Coulter/Immunotech, Brea, CA, USA), CD19, ease for gene therapy.40,41 Our findings suggest that the and CD20 (Becton Dickinson). NK cells were identified achievement of low levels of JAK3 expression through by a combination of anti-CD16 and anti-CD56 (Becton gene transfer may successfully induce T cell development Dickinson), evaluated on CD3-negative lymphocytes. and therefore support the development of gene therapy Peripheral blood lymphocyte counts were obtained in applications for this disease. On the other hand, a degree parallel using a Coulter Counter (Coulter Corporation, of caution comes from the observed lymphocytic infil- Maine, FL, USA). tration syndrome in patient II,4, which may have resulted from failure of T cell homeostasis maintenance mediated T cell proliferation assays by sub-optimal levels of JAK3 expression. In vitro and/or PBMC (2 × 105 per well) were plated in triplicate in RPMI in vivo experiments of human T lymphocyte development 1640 medium (Life Technologies) supplemented with 5% following gene transfer into JAK3-deficient hematopo- human serum, 4 mM L-glutamine, 1% non-essential ietic progenitors may be necessary to gain more infor- amino acids, 1 mM Na pyruvate, and 50 ␮g/ml gentam- mation on the levels of JAK3 expression that can be achi- icin (LifeTechnologies) with or without the following eved in peripheral T cells of subjects who undergo gene stimulants: phytohaemoagglutinin (PHA, 10 ␮g/ml, transfer-based treatments. Difco), concavalin-A, (Con-A, 25 ␮g/ml, Sigma, St Louis, The identification of JAK3 mutations in this family MO, USA), OKT3 (100 ng/ml, Ortho Biotech), IL-2

Genes and Immunity Unexpected phenotypes in JAK3 deficiency DM Frucht et al 430 (100 U/ml, Roche), pokeweed mitogen (PWM, 1:400, Life prepared from the RNA template using a first-strand Technologies), tetanus toxoid (3 LF/ml, Connaught Lab- cDNA kit (Boehringer Mannheim, Indianapolis, IN, oratories, LTD, Toronto, Canada), or Candida albicans USA). Overlapping regions of the JAK3 coding regions (1:30, Hollister Stier Labs, LLC, Spokane, WA, USA). Cul- were PCR-amplified from this cDNA using the follow- ° tures were incubated at 37 C, 5% CO2 for 3 days (PHA, ing primers: Con-A, OKT3, IL-2, OKT3+IL-2) or 7 days (PWM, tetanus, candida) and then pulsed with 1 ␮Ci per well of 3H-thy- JAK3/93 forward:CTCATGGCACCTCCAAGTGAA, midine during the last 6 h of culture. Cells were then har- JAK3/1067 reverse: GTCTGTCCTGGTAACAGTGAC, vested, and incorporated radioactivity was determined JAK3/993 forward: GACATTAGCATCAAGCAGGCC, with a scintillation counter (Wallac, Uppsula, Sweden). JAK3/2022 reverse: GGCCTTTGTCCTCCAGATAGT, JAK3/1961 forward: CTGGAAGCTGCAGGTGGTCAA, TcR V␤ family analysis JAK3/2654 forward: CGTGAAACAGCTGCAGCACAG, Total RNA was isolated from 3–5 × 106 PBMC using the JAK3/2996 reverse: GTCAGCGATCTTGACGTGTGC, PureScript RNA Isolation Kit (Gentra Systems). First JAK3/2888 forward: GCAGATCTGCAAGGGCATGGA, strand cDNA was prepared from 5 ␮g of RNA using the and SuperScript Preamplification System (Gibco-BRL) and JAK3/3517 reverse: ACACAGCCAGTCAACAGAGAC. amplified using an oligonucleotide primer specific for the TcR constant ␤ region (C␤) coupled to a second primer Amplification products were ligated into a TA cloning specific for each of 24 different TcR variable ␤ (V␤) famil- vector (Invitrogen) and sequenced using M13 forward ies. Oligonucleotide sequences for these reactions have and reverse primers on an ABI 377 sequencer (PE been previously published.43 PCR products were sub- Applied Biosystems). jected to electrophoretic separation in 2% agarose and Genomic DNA was obtained using the Wizard DNA visualized by exposure to UV light. The intensity of each purification kit (Promega, Madison, WI, USA), and band was quantified using the NIH Image software sequencing of the entire coding sequence of the JAK3 (http://rsb.info.nih.gov/nih-image). The relative per- gene was performed. To confirm mutations revealed by centage of expression of each TcR V␤ was calculated by cDNA analysis, as well as to obtain sequence information normalizing the intensity of individual families to the at the 5Ј proximal region of the coding sequence 100 ng of sum of the intensity values of all TcR V␤ segments.30 For genomic DNA template was used to amplify exons and heteroduplex analysis, 20 ␮l of the TcR V␤ PCR reactions flanking intronic sequence using the following M13- was heated at 95°C for 5 min and cooled to 50°C for 1 h tailed primers: before being separated using nondenaturating polyacrila- mide gel electrophoresis and visualization by UV transil- JAK3 exon 1 forward: TGTAAAACGACGGCCAGTCC lumination. For analysis of the junctional amino acid TCCAGCACTCCTTTCCATG, sequences of selected TcR V␤ families, TcR V␤ PCR pro- JAK3 exon 1 reverse: AGGAAACAGCTATGACCATG ducts were inserted into the pCR 4-TOPO plasmid using ACTTCTGCAGGCAACTATTCC, the TOPO TA Cloning kit (both from Invitrogen, JAK3 exon 18 forward: TGTAAAACGACGGCC Carlsbad, CA, USA) and sequenced using the T3 primer AGTGCATCCAGGTGCCTGGACATC, on an ABI 373 automated sequencer sequencer (PE JAK3 exon 18 reverse: AGGAAACAGCTATGACCAT applied Biosystems, Foster City, CA, USA). ACGTTCCCAGCCTACCTAAAG.

Immunoprecipitation and Western blotting PCR products were directly sequenced using M13 pri- After 4 h of culture in 2% FCS (for PBMC) or in the mers (Amersham, Piscataway, NJ, USA) on an ABI 377 absence of fetal calf serum (FCS) (for EBV-transformed B sequencer. cells), 5–10 × 107 cells were resuspended in 1 ml of RPMI 1640 medium and stimulated at 37°C with IL-2 (1000 IU) PCR-based mutational analysis screening for 15 min. Following stimulation, lysis in a Triton-X The JAK3 mutation present in intron 18 ablated an HphI buffer, immunoprecipitation, PAGE, and immunoblot- restriction site. JAK3 exon 18 forward and reverse pri- ting were performed as described previously.44 Briefly, mers were used to amplify exon 18 and surrounding lysates were immunoprecipitated with anti-JAK3 (␣- intronic sequences from family members and 50 healthy JAK3)45 or anti-STAT5a (␣-STAT5a)46 rabbit polyclonal controls. PCR products were digested overnight at 37°C antibodies and immunoblotted with anti-phosphotyros- with HphI (New England Biolabs, Beverly, MA, USA) and ine mouse monoclonal Ab (␣-PY, 4G10, Upstate Biotechn- analyzed by gel electrophoresis. ology, Lake Placid, NY, USA) and with the original immunoprecipitation antibody as described.44 Detection Histopathology of immunocomplexes was then performed by enhanced Microtome sections (3–4 ␮m-thick) were obtained from chemiluminescence (ECL, Amersham Life Science, paraffin-embedded bone marrow and percutaneous liver Arlington Heights, IL, USA). For direct Western blot biopsy samples and stained with hematoxylin and eosin analysis of cell lysates, 100 ␮g of protein was subjected to for morphological examination. Slides were examined PAGE and immunoblotted using the ␣-JAK3 or ␣-STAT5 using an Olympus BX41 microscope and photographed antibodies as described.5 at magnification of ×400 and ×600 using an Olympus DP11 digital camera. Sequence analysis RNA was prepared from 2 × 106 PBMC stimulated with Analysis of Fas ligand expression PHA for 24 h using RNA STAT-60 kit (Tel-Test Incorpor- Isolated PBMC were incubated with or without IL-2 ated, Friendswood, TX, USA). Subsequently, cDNA was (1000 U/ml) for 24 h. Protein lysates were electrophor-

Genes and Immunity Unexpected phenotypes in JAK3 deficiency DM Frucht et al 431 esed on a 12.5% PAGE gel. Western blotting was perfor- 20 Sottini A, Quiros-Roldan E, Albertini A, Primi D, Imberti L. med as previously described44 using an IgG1 monoclonal Assessment of T-cell receptor beta-chain diversity by heterodu- anti-human FasL (G247–4, Pharmingen, San Diego, CA, plex analysis. Hum Immunol 1996; 48:12–22. USA) at a dilution of 1:1000. 21 Van Parijs L, Refaeli Y, Lord JD, Nelson BH, Abbas AK, Balti- more D. Uncoupling IL-2 signals that regulate T cell proliferation, survival, and Fas-mediated activation-induced cell death. Acknowledgements Immunity 1999; 11: 281–288. The authors would like to thank Drs Silvia Giliani, Donna 22 Fischer A, Rieux-Laucat F, Le Deist F. Autoimmune lymphoproliferative syndromes (ALPS): models for the study of peripheral M Krasnewich, Makoto Otsu, and Shepherd Schurman tolerance. Rev Immunogenet 2000; 2:52–60. for helpful discussions, and Dr Stefania Pittaluga for her 23 Buckley RH, Schiff RI, Schiff SE et al. Human severe combined advice in evaluating histological samples. immunodeficiency: genetic, phenotypic, and functional diversity in one hundred eight infants. J Pediatr 1997; 130: 378–387. References 24 Chen M, Cheng A, Chen YQ et al. The amino terminus of JAK3 is necessary and sufficient for binding to the common gamma 1 Leonard WJ, O’Shea JJ. Jaks and STATs: biological implications. chain and confers the ability to transmit interleukin 2-mediated Annu Rev Immunol 1998; 16: 293–322. signals. Proc Natl Acad Sci USA 1997; 94: 6910–6915. 2 Ihle JN, Thierfelder W, Teglund S et al. Signaling by the cytokine 25 Cacalano NA, Migone TS, Bazan F et al. Autosomal SCID caused receptor superfamily. Ann NY Acad Sci 1998; 865:1–9. by a point mutation in the N-terminus of Jak3: mapping of the 3 Macchi P, Villa A, Giliani S et al. Mutations of Jak-3 gene in Jak3-receptor interaction domain. Embo J 1999; 18: 1549–1558. patients with autosomal severe combined immune deficiency 26 Hirschhorn R, Yang DR, Puck JM, Huie ML, Jiang CK, Kurland- (SCID). Nature 1995; 377:65–68. sky LE. Spontaneous in vivo reversion to normal of an inherited 4 Russell SM, Tayebi N, Nakajima H et al. Mutation of Jak3 in a mutation in a patient with adenosine deaminase deficiency. Nat patient with SCID: essential role of Jak3 in lymphoid develop- Genet 1996; 13: 290–295. ment. Science 1995; 270: 797–800. 27 Stephan V, Wahn V, Le Deist F et al. Atypical X-linked severe 5 Candotti F, Oakes SA, Johnston JA et al. Structural and func- combined immunodeficiency due to possible spontaneous tional basis for JAK3-deficient severe combined immunodefi- reversion of the genetic defect in T cells. N Engl J Med 1996; 335: ciency. Blood 1997; 90: 3996–4003. 1563–1567. 6 Thomis DC, Gurniak CB, Tivol E, Sharpe AH, Berg LJ. Defects in 28 Schmalstieg FC, Leonard WJ, Noguchi M et al. Missense B lymphocyte maturation and T lymphocyte activation in mice mutation in exon 7 of the common gamma chain gene causes a lacking Jak3. Science 1995; 270: 794–797. moderate form of X-linked combined immunodeficiency. J Clin 7 Nosaka T, van Deursen JM, Tripp RA et al. Defective lymphoid Invest 1995; 95: 1169–1173. development in mice lacking Jak3. Science 1995; 270: 800–802. 29 Sharfe N, Shahar M, Roifman CM. An interleukin-2 receptor 8 Park SY, Saijo K, Takahashi T et al. Developmental defects of gamma chain mutation with normal thymus morphology. J Clin lymphoid cells in Jak3 kinase-deficient mice. Immunity 1995; 3: Invest 1997; 100: 3036–3043. 771–782. 30 Mella P, Imberti L, Brugnoni D et al. Development of autologous 9 Peschon JJ, Morrissey PJ, Grabstein KH et al. Early lymphocyte T lymphocytes in two males with X-linked severe combined expansion is severely impaired in interleukin 7 receptor- immune deficiency: molecular and cellular characterization. Clin deficient mice. J Exp Med 1994; 180: 1955–1960. Immunol 2000; 95:39–50. 10 von Freeden-Jeffry U, Vieira P, Lucian LA, McNeil T, Burdach 31 Gozalo-Sanmillan S, McNally JM, Lin MY, Chambers CA, Berg SE, Murray R. Lymphopenia in interleukin (IL)-7 gene-deleted LJ. Cutting edge: two distinct mechanisms lead to impaired T mice identifies IL-7 as a nonredundant cytokine. J Exp Med 1995; 181: 1519–1526. cell homeostasis in Janus kinase 3- and CTLA-4-deficient mice. 11 Puel A, Ziegler SF, Buckley RH, Leonard WJ. Defective IL7R J Immunol 2001; 166: 727–730. expression in T(−)B(+)NK(+) severe combined immunodefi- 32 Baird AM, Thomis DC, Berg LJ. T cell development and acti- ciency. Nat Genet 1998; 20: 394–397. vation in Jak3-deficient mice. J Leukoc Biol 1998; 63: 669–677. 12 Kennedy MK, Glaccum M, Brown SN et al. Reversible defects 33 Thomis DC, Berg LJ. Peripheral expression of Jak3 is required to in natural killer and memory CD8 T cell lineages in interleukin maintain T lymphocyte function. J Exp Med 1997; 185: 197–206. 15-deficient mice. J Exp Med 2000; 191: 771–780. 34 Thomis DC, Lee W, Berg LJ. T cells from Jak3-deficient mice 13 Suzuki H, Kundig TM, Furlonger C et al. Deregulated T cell have intact TCR signaling, but increased apoptosis. J Immunol activation and autoimmunity in mice lacking interleukin-2 1997; 159: 4708–4719. receptor beta. Science 1995; 268: 1472–1476. 35 Vockley J, Rinaldo P, Bennett MJ, Matern D, Vladutiu GD. Syn- 14 Willerford DM, Chen J, Ferry JA, Davidson L, Ma A, Alt FW. ergistic heterozygosity: disease resulting from multiple partial Interleukin-2 receptor alpha chain regulates the size and content defects in one or more metabolic pathways. Mol Genet Metab of the peripheral lymphoid compartment. Immunity 1995; 3: 2000; 71:10–18. 521–530. 36 Brown MP, Nosaka T, Tripp RA et al. Reconstitution of early 15 Schorle H, Holtschke T, Hunig T, Schimpl A, Horak I. Develop- lymphoid proliferation and immune function in Jak3-deficient ment and function of T cells in mice rendered interleukin-2 mice by interleukin-3. Blood 1999; 94: 1906–1914. deficient by gene targeting. Nature 1991; 352: 621–624. 37 Moriggl R, Topham DJ, Teglund S et al. Stat5 is required for IL- 16 Saijo K, Park SY, Ishida Y, Arase H, Saito T. Crucial role of Jak3 2-induced cell cycle progression of peripheral T cells. Immunity in negative selection of self-reactive T cells. J Exp Med 1997; 185: 1999; 10: 249–259. 351–356. 38 Lenardo M, Chan KM, Hornung F et al. Mature T lymphocyte 17 Baird AM, Lucas JA, Berg LJ. A profound deficiency in thymic apoptosis-immune regulation in a dynamic and unpredictable progenitor cells in mice lacking Jak3. J Immunol 2000; 165: antigenic environment. Annu Rev Immunol 1999; 17: 221–253. 3680–3688. 39 Sharfe N, Dadi HK, Shahar M, Roifman CM. Human immune 18 Brugnoni D, Notarangelo LD, Sottini A et al. Development of disorder arising from mutation of the alpha chain of the autologous, oligoclonal, poorly functioning T lymphocytes in a interleukin-2 receptor. Proc Natl Acad Sci USA 1997; 94: 3168– patient with autosomal recessive severe combined immuno- 3171. deficiency caused by defects of the Jak3 tyrosine kinase. Blood 40 Candotti F, Oakes SA, Johnston JA, Notarangelo LD, O’Shea JJ, 1998; 91: 949–955. Blaese RM. In vitro correction of JAK3-deficient severe com- 19 McVay LD, Carding SR. Generation of human gammadelta T- bined immunodeficiency by retroviral-mediated gene transduc- cell repertoires. Crit Rev Immunol 1999; 19: 431–460. tion. J Exp Med 1996; 183: 2687–2692.

Genes and Immunity Unexpected phenotypes in JAK3 deﬁciency DM Frucht et al 432 41 Bunting KD, Sangster MY, Ihle JN, Sorrentino BP. Restoration and activation of the Jak-3 Janus kinase in response to interleu- of lymphocyte function in Janus kinase 3-deﬁcient mice by retro- kin-2. Nature 1994; 370: 151–153. viral-mediated gene transfer. Nat Med 1998; 4:58–64. 45 Kawamura M, McVicar DW, Johnston JA et al. Molecular clon- 42 Sneller MC, Wang J, Dale JK et al. Clinical, immunologic, and ing of L-JAK, a Janus family protein-tyrosine kinase expressed genetic features of an autoimmune lymphoproliferative syn- in natural killer cells and activated leukocytes. Proc Natl Acad drome associated with abnormal lymphocyte apoptosis. Blood Sci USA 1994; 91: 6374–6378. 1997; 89: 1341–1348. 46 Liu X, Robinson GW, Gouilleux F, Groner B, Hennighausen L. 43 Rowen L, Koop BF, Hood L. The complete 685-kilobase DNA Cloning and expression of Stat5 and an additional homologue sequence of the human beta T cell receptor locus. Science 1996; (Stat5b) involved in prolactin signal transduction in mouse 272: 1755–1762. mammary tissue. Proc Natl Acad Sci USA 1995; 92: 8831–8835. 44 Johnston JA, Kawamura M, Kirken RA et al. Phosphorylation

Genes and Immunity