|
|
Background The hyper-IgE syndrome (or Job's syndrome) is a rare disorder of immunity and connective tissue characterized by dermatitis, boils, cyst-forming pneumonias, elevated serum IgE levels, retained primary dentition, and bone abnormalities. Inheritance is autosomal dominant; sporadic cases are also found.
Methods We collected longitudinal clinical data on patients with the hyper-IgE syndrome and their families and assayed the levels of cytokines secreted by stimulated leukocytes and the gene expression in resting and stimulated cells. These data implicated the signal transducer and activator of transcription 3 gene (STAT3) as a candidate gene, which we then sequenced.
Results We found increased levels of proinflammatory gene transcripts in unstimulated peripheral-blood neutrophils and mononuclear cells from patients with the hyper-IgE syndrome, as compared with levels in control cells. In vitro cultures of mononuclear cells from patients that were stimulated with lipopolysaccharide, with or without interferon-
Conclusions Mutations in STAT3 underlie sporadic and dominant forms of the hyper-IgE syndrome, an immunodeficiency syndrome involving increased innate immune response, recurrent infections, and complex somatic features.
, had higher tumor necrosis factor 
levels than did identically treated cells from unaffected persons (P=0.003). In contrast, the cells from patients with the hyper-IgE syndrome generated lower levels of monocyte chemoattractant protein 1 in response to the presence of interleukin-6 (P=0.03), suggesting a defect in interleukin-6 signaling through its downstream mediators, one of which is STAT3. We identified missense mutations and single-codon in-frame deletions in STAT3 in 50 familial and sporadic cases of the hyper-IgE syndrome. Eighteen discrete mutations, five of which were hot spots, were predicted to directly affect the DNA-binding and SRC homology 2 (SH2) domains.
Eosinophilia, eczema, and elevated IgE levels focused attention on the T helper cell Th1/Th2 profile,8 but no primary defect was uncovered. Cytogenetic and linkage data suggested the presence of a locus on chromosome 4q, but no disease gene was found there.9 Microarray-based approaches have yielded inconsistent results.10,11 We therefore aimed to comprehensively describe the phenotypes of affected persons in a large cohort and to assay gene expression and in vitro cytokine production to identify relevant pathways.
Methods
Patients and Controls
Patients suspected to have the hyper-IgE syndrome, or their parents, and their relatives gave written informed consent. These patients and relatives were enrolled in studies approved by the institutional review board of the National Institute of Allergy and Infectious Diseases. We collected blood samples while patients were clinically well and were not receiving corticosteroids or nonsteroidal antiinflammatory drugs. An empirical scoring system was used to calculate the risk of carrying the hyper-IgE syndrome trait on the basis of an aggregation of diagnostic features.9 We arbitrarily categorized hyper-IgE syndrome scores of 0 to 15 as unaffected, 16 to 39 as possibly affected, 40 to 59 as probably affected, and 60 or more as definitely affected.9 We analyzed by microarray blood samples obtained from probands who were definitely affected (had scores of 60 or more) and were over the age of 16 years. We also performed cytokine phenotyping on these samples, in addition to those from persons with familial and sporadic hyper-IgE syndrome (which included possibly and probably affected persons). Samples from 50 patients and 48 family members were available for sequence analysis. The race or ethnic group of patients was determined by a physician author and by self report (for details, see the Supplementary Appendix, available with the full text of this article at www.nejm.org). Control DNA was collected according to several approved protocols from unaffected persons: 125 whites, 15 Hispanics, 4 blacks, 2 Asians, 1 white Hispanic, 1 black Hispanic, and 10 persons of unknown ancestry. Race or ethnic group of these control subjects was self-reported.
Microarray Analysis
We isolated polymorphonuclear leukocytes and peripheral-blood mononuclear cells from venous blood samples and carried out phagocytosis and microarray assays as previously described.12,13 We labeled samples, allowed for hybridization of complementary RNA to microarrays (HU133 Plus 2.0, Affymetrix), and scanned the microarrays according to standard Affymetrix protocols. For each time point, we analyzed cells from 9 control donors of peripheral-blood mononuclear cells or 10 control donors of polymorphonuclear leukocytes and 7 patients with the hyper-IgE syndrome, using a separate oligonucleotide array for each donor. One sample of stimulated control polymorphonuclear leukocytes for the 360-minute time point was lost owing to technical error.
We analyzed the gene-expression data by using Microarray Suite software, version 5.0 (Affymetrix), and GeneSpring software, version 5.0 (Silicon Genetics), as described previously,12,13 but with modifications. We defined genes as being differentially expressed when all three of the following criteria were met: the gene was identified as "present" by the Microarray Suite software in 8 of the 9 samples of peripheral-blood mononuclear cells or 9 of the 10 samples of polymorphonuclear leukocytes from control subjects and in 6 of the 7 samples of cells from patients, the difference in gene expression between the samples from control subjects and those from patients was twofold or greater, and the normalized signal intensity was at least 100. We carried out principal-component analysis using software from Partek and compared the average difference intensity of transcript levels in cells from control subjects and patients with the use of Student's t-test. Complete microarray data are provided, with analysis, in Tables 1, 2, 3, 4, and 5 of the Supplementary Appendix. Raw data are available at the Gene Expression Omnibus Web site (www.ncbi.nlm.nih.gov/geo/) under accession number GSE8507 [NCBI GEO] . We visualized the implicated biochemical pathways using Ingenuity Pathways Analysis software (Ingenuity Systems).
Cytokines
Peripheral-blood mononuclear cells were prepared with the use of density centrifugation from heparinized whole blood, as previously described.14 Cells were stimulated for 48 hours at 37°C in 5% carbon dioxide with phytohemagglutinin (1:100), phytohemagglutinin with interleukin-12 (10 ng per milliliter), lipopolysaccharide (200 ng per milliliter), lipopolysaccharide with interferon- (1000 IU per milliliter), or interleukin-6 (40 ng per milliliter). The supernatants were harvested and stored at –20°C until cytokine levels could be measured with the use of plate-based methods (Meso Scale Discovery) or bead-based methods (BioRad Luminex), according to the manufacturers' recommendations. The cytokine levels were compared with the use of a Mann–Whitney U test (GraphPad Software), with P values of less than 0.05 considered to indicate statistical significance.
Sequencing
Complementary DNA (cDNA) was prepared from peripheral-blood mononuclear cells and neutrophils of unrelated probands and control subjects and was sequenced for the signal transducer and activator of transcription 3 gene (STAT3) (for a list of primers, see the Methods in the Supplementary Appendix). Sequencing was performed on an Applied Biosystems 3100 sequencer and analyzed using Sequencher software (Gene Codes). We also sequenced cDNA from peripheral-blood leukocytes or transformed lymphoblastoid cell lines or genomic DNA of relatives. For sequencing of genomic DNA, polymerase-chain-reaction primers (available in the Supplementary Appendix) were chosen according to the proband's mutation. GenBank recognizes three human isoforms of STAT3: NP_644805.1 [GenBank] , NP_003141.2 [GenBank] , and NP_998827.1 [GenBank] . The amino acid positions reported here are for the isoform NP_644805.1 [GenBank] .
Flow Cytometry for Detection of Phosphorylated STAT3
Peripheral-blood mononuclear cells from healthy adults or from patients with the hyper-IgE syndrome were isolated with the use of a Ficoll gradient. Peripheral-blood mononuclear cells (1x106) were incubated in 100 µl of phosphate-buffered saline with 2% fetal-calf serum, either without stimuli or with interleukin-6 (10 ng per milliliter) (R&D Systems), for 25 or 60 minutes. Cells were fixed, permeabilized, and stained, as previously described.15 An Alexa Fluor 647 conjugated phospho-STAT3 antibody against phospho-S727 (BD Biosciences) was used to identify intracellular or intranuclear phospho-STAT3. Mouse IgG1 (BD Biosciences) was used as isotype control. Analysis was carried out on a FACSCalibur flow cytometer (BD Biosciences) with Cell Quest software (BD Biosciences).
Bioinformatics Analysis of Missense Changes in STAT3
For the missense changes detected, we used bioinformatics methods encoded in the programs Sorting Intolerant from Tolerant (SIFT)16 and PolyPhen.17 SIFT reports the probability that the substitution is deleterious by predicting whether the substitution is or is not probably damaging. PolyPhen reports a score, with putatively benign changes scored from 0.00 to 1.50, possibly damaging changes scored from 1.51 to 2.00, and probably damaging changes scored as 2.01 or greater.
Results
Gene-Expression Profiling of Leukocytes
We found greater levels of expression of proinflammatory molecules in unstimulated and phagocytosis-stimulated leukocytes of patients with the hyper-IgE syndrome than in those of controls (Figure 1 in the Supplementary Appendix and Tables 1, 2, 3, and 4 in the Supplementary Appendix). Thirty transcripts encoding proteins related to interferon signal transduction or typically induced by interferon showed elevated levels of expression in polymorphonuclear leukocytes from patients with the hyper-IgE syndrome (Table 1 in the Supplementary Appendix). Known or putative genes involved in host defense were up-regulated in polymorphonuclear leukocytes and peripheral-blood mononuclear cells from the patients (Figure 1C in the Supplementary Appendix and Tables 1, 2, and 3 in the Supplementary Appendix). After phagocytosis, 15 genes encoding proteins related to interferon signal transduction or induced by interferon showed elevated levels of expression in polymorphonuclear leukocytes from patients, as compared with those from controls (Figure 2 in the Supplementary Appendix and Table 4 in the Supplementary Appendix). Taken together, these data suggest that cytokine-related signal transduction, such as that involving the interferons and STATs, is altered in patients with the hyper-IgE syndrome (Figure 1D in the Supplementary Appendix).
Cytokine Expression in Leukocytes
Consistent with the microarray data, cytokine protein levels were elevated in the supernatants of leukocytes from patients with the hyper-IgE syndrome as compared with those of controls, depending on the stimulus (Table 1). For example, tumor necrosis factor (TNF-) levels were higher in the supernatants of leukocytes from patients with the hyper-IgE syndrome than in those from controls, after the leukocytes were stimulated by the toll-like receptor 4 agonist lipopolysaccharide alone (P=0.003) or by lipopolysaccharide and interferon- (P=0.03). TNF- levels were also elevated in the supernatants from leukocytes of patients after exposure to the toll-like receptor 2 agonist heat-killed Listeria monocytogenes (P<0.001) or to S. aureus Cowan strain (P=0.001). Similarly, stimulation with lipopolysaccharide or lipopolysaccharide and interferon- resulted in elevated interleukin-12p70 production in patients with the hyper-IgE syndrome, as compared with controls (P=0.001 and P=0.003, respectively). In response to the nonspecific T-cell mitogen phytohemagglutinin or phytohemagglutinin and interleukin-12, interferon- levels were also elevated in the patients (P=0.03 for both comparisons). In contrast, we found diminished production of monocyte chemoattractant protein 1 in leukocytes from patients, as compared with control leukocytes, after exposure to interleukin-6 (P=0.03) (Table 1).
|
STAT3 Mutations in the Hyper-IgE Syndrome
Expression levels of the interleukin-6 receptor (IL6R or CD126) and the interleukin-6 signal transducer (IL6ST, CD130, or gp130) were normal on cells from patients with the hyper-IgE syndrome (data not shown). We inferred that Janus kinase 1 was intact in cells from patients, because it mediates TNF- production in response to interferon-. STAT3 is physically associated with gp130, and its deletion in animal models and inhibition in tumor-cell lines is associated with increased production of interferon- and TNF-.18 We therefore sequenced STAT3 in DNA samples obtained from 50 persons affected with the hyper-IgE syndrome and 48 unaffected relatives from 35 independent and unrelated white, Asian, Hispanic, or black families (Table 2). Seven patients had de novo STAT3 mutations, 17 had familial transmission of STAT3 mutations, and 26 had sporadic STAT3 mutations (defined as those for which the parents were reported to be clinically unaffected and, in cases in which DNA was available for analysis, did not carry a STAT3 mutation), 1 of whom had chimerism for the STAT3 mutation. All STAT3 mutations localized to regions encoding the DNA-binding or SH2 domains; several mutational "hot spots" were apparent (Figure 1). For example, 13 mutations from unrelated patients affected the arginine (R) residue at position 382 in the DNA-binding domain. Several other residues in the DNA-binding domain — phenylalanine (F) at position 384, arginine (R) at position 423, and valine (V) at position 463 — were each affected by at least two de novo mutations. The SH2 domain has a hot spot at position 637; we found seven mutations from unrelated patients that involved the valine (V) residue normally found at this position. All mutations are missense or in-frame deletions, consistent with protein expression, and affect residues included in all splice variants of STAT3.19 We did not find any of the mutations in the 158 unaffected control subjects of various ancestral origins.
|
|
|
Thirteen of the 16 missense mutations identified are predicted to be possibly or probably damaging to STAT3 function according to at least one of the two methods (Table 6 in the Supplementary Appendix). R423Q, a recurrent mutation, is not predicted to be damaging but is clearly associated with disease in Families J007 and J016.
Functional Studies
Phosphorylation and expression of STAT3, as determined with the use of flow cytometry, was equivalent in the leukocytes from patients and those from controls (Figure 3 in the Supplementary Appendix). The trafficking of STAT3 from the cytoplasm into the nucleus of leukocytes from patients was grossly normal (data not shown). Unlike in mouse models of STAT3 mutation,20 superoxide production by neutrophils from patients in response to opsonized zymosan and N-formyl-methionyl-leucyl-phenylalanine and phorbol myristate acetate was normal, as was killing of staphylococci by neutrophils in vitro (data not shown). Although interferon- and related proinflammatory cytokines prolong neutrophil survival, and our microarray suggested that this should be the case,21 the rates of spontaneous and phagocytosis-induced apoptosis did not differ between polymorphonuclear leukocytes from patients with the hyper-IgE syndrome and those from controls (data not shown). This is consistent with an inability of mutant STAT3 to promote cytokine-mediated delay of neutrophil apoptosis.22
Discussion
STAT3 mutation is the predominant cause of sporadic and familial hyper-IgE syndrome, although other genomic loci may also be involved. Within the gene encoding STAT3, we found mutational hot spots, indicated by independent mutations that affect the same codons, in the evolutionarily conserved SH2 and DNA-binding domains of STAT3. These mutations are predicted not to affect protein levels of STAT3 and are consistent with the survival of those carrying the mutations. In support of STAT3 as the cause of the hyper-IgE syndrome, Minegishi et al. found STAT3 mutations affecting the DNA-binding domain in 8 of 15 patients examined.23 Therefore, all mutations associated with the hyper-IgE syndrome identified to date are restricted to the DNA-binding or SH2 regions of STAT3. These mutations are likely to preserve protein levels, phosphorylation, and nuclear localization. Mice with half the normal dose of Stat3 seem unaffected, suggesting that haploinsufficiency is an unlikely pathogenic mechanism.24
A patient with a newly discovered primary immunodeficiency with mild IgE elevation due to homozygous recessive tyrosine kinase 2 deficiency has also been reported,25 but this patient differs from our patients with classic hyper-IgE syndrome in the pattern of inheritance, susceptibility to viral infection, the degree of IgE elevation, and the absence of nonimmune features. Patients with the syndrome reported as autosomal recessive hyper-IgE syndrome also differ from those with the classic hyper-IgE syndrome described here in that they have fewer pulmonary infections and pneumatoceles, more viral infections, and vasculitis.26 In keeping with these clinical differences, an index patient with autosomal recessive hyper-IgE syndrome from our original report26 had no STAT3 mutations (data not shown).
STAT3 mutations in the hyper-IgE syndrome clarify many disparate aspects of the syndrome that have puzzled investigators. Targeted mutations in mice show specific roles of Stat3 in organogenesis, organ preservation, and organ-specific inflammation, distinct from hematopoiesis. Stat3 deficiency specific to the pulmonary epithelium in mice exposed to hyperoxia causes excessive lung inflammation and airspace enlargement,27 and lung expression of Stat3 is critical for repression of the inflammatory response to lipopolysaccharide20; these observations are consistent with the development of pneumatoceles after infection in patients with the hyper-IgE syndrome. Myeloid-specific Stat3 deficiency causes increased myelopoiesis and eosinophilia in mice,28 and eosinophilia is a consistent feature of the hyper-IgE syndrome.3 Hematopoiesis-specific Stat3 deletion in the mouse is associated with osteoclast generation and osteopenia,29 phenotypes also found in persons with the hyper-IgE syndrome.30 Ligand binding of the receptor for interleukin-22 (composed of IL-22R1 and IL-10R2), expressed predominantly on epithelial cells (including those making up the skin and lung),31 invokes STAT3-mediated expression of 
-defensins.32 Impaired defensin production in the skin and lung may explain the predilection for skin and lung abscesses in patients with the hyper-IgE syndrome. The depletion of Stat3 in mouse cardiac myocytes is associated with increased production of TNF- by these cells and with cardiac inflammation and dysfunction,33 consistent with coronary-artery aneurysms recently described in adult patients with the hyper-IgE syndrome.5 Similarly, a deficiency of Stat3 in the mouse brain is associated with increased inflammation, demyelination, and astrocytosis in response to nerve injury,34 consistent with the parenchymal brain lesions found in persons with the hyper-IgE syndrome.6
STAT3 up-regulates myeloid adhesion,35 expression of PU.1,36 expression of secondary granule proteins in neutrophils,37 the interleukin-23 receptor, the generation of interleukin-17–producing CD4+ T lymphocytes,38 and the antiinflammatory effects of interleukin-10.18 In contrast, STAT3 down-regulates expression of T-bet, GATA3, IL12Rb2, and IFN,38 as well as the formation of osteoclasts.29 Therefore, STAT3 deficiency leads to the up-regulation of many cytokines secreted by type 1 helper T cells (e.g., interferon- or TNF-), and down-regulation of the inflammatory and antiinflammatory responses governed by interleukin-6 and interleukin-10.20 Patients with the hyper-IgE syndrome have retarded and inadequate inflammatory responses in the skin, leading to cold abscesses, and destructive inflammation in the lung, leading to pneumatoceles. The extraordinary elevation of the IgE level in persons with the disorder, seen from birth through adulthood and uncorrelated with eosinophilia,3 may reflect the known role of STAT3 in mediating interleukin-21 receptor signaling, since interleukin-21 receptor –knockout mice have elevated IgE levels.39
In conclusion, the newly recognized genetic cause of the hyper-IgE syndrome — STAT3 mutation — affects complex, compartmentalized somatic and immune regulation. The discovery of this genetic cause opens new doors to understanding organ-specific infection, inflammation, and therapy.
Supported in part by the Intramural Research Program of the National Institutes of Health, the National Institute of Allergy and Infectious Diseases, the National Human Genome Research Institute, the National Cancer Institute, and the National Library of Medicine and by a European Commission Marie Curie Excellence Grant (MEXT-CT-2006-042316, to Dr. Grimbacher). This project has been funded in part with federal funds from the National Cancer Institute, National Institutes of Health, under contract N01-CO-12400.
The views expressed are those of the authors and do not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
No potential conflict of interest relevant to this article was reported.
We thank Daniel E. Sturdevant of the National Institute of Allergy and Infectious Diseases for assisting with the statistical analysis of microarray data.
Source Information
From the National Institute of Allergy and Infectious Diseases, Bethesda, MD (S.M.H., H.Z.E., A.P.H., G.U., N.B., A.F.F., A.D., V.L.A., D.N.D., P.A.W., H.L.M., J.I.G.), and Hamilton, MT (F.R.D., S.D.K., A.R.W., J.M.V., J.M.M.); the National Human Genome Research Institute (J.D.), the National Cancer Institute (M.L.T.), and the National Center for Biotechnology Information (A.A.S.) — all in Bethesda, MD; Science Applications International Corporation–Frederick, National Cancer Institute at Frederick, Frederick, MD (D.B.K.); the Center for Drug Evaluation and Research, Food and Drug Administration, Rockville, MD (D.M.F.); Royal Free Hospital and University College London, London (C.W., B.G.); and the University of California at San Francisco, San Francisco (J.M.P.).
Drs. Puck and Grimbacher contributed equally to this article.
This article (10.1056/NEJMoa073687) was published at www.nejm.org on September 19, 2007.
Address reprint requests to Dr. Holland at the National Institute of Allergy and Infectious Diseases, National Institutes of Health, CRC B3-4141 MSC 1684, Bethesda, MD 20892, or at smh{at}nih.gov.
References
-interferon and tumor necrosis factor- by circulating T cells from patients with the hyper-immunoglobulin E syndrome. J Clin Invest 1989;84:1830-1835. [ISI][Medline], interleukin-12, and tumor necrosis factor- in human interferon- receptor 1 deficiency. J Infect Dis 1998;178:1095-1104. [ISI][Medline]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
STAT3 Mutation in the Original Patient with Job's Syndrome
Renner E. D., Torgerson T. R., Rylaarsdam S., Añover-Sombke S., Golob K., LaFlam T., Zhu Q., Ochs H. D.
Extract |
Full Text |
PDF
N Engl J Med 2007;
357:1667-1668, Oct 18, 2007.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
T), R382Q (1145G