STAT1 Mutations in Autosomal Dominant Chronic Mucocutaneous Candidiasis Frank L

T h e new england journal o f medicine

original article

STAT1 Mutations in Autosomal Dominant Chronic Mucocutaneous Candidiasis Frank L. van de Veerdonk, M.D., Ph.D., Theo S. Plantinga, Ph.D., Alexander Hoischen, Ph.D., Sanne P. Smeekens, M.Sc., Leo A.B. Joosten, Ph.D., Christian Gilissen, Ph.D., Peer Arts, Ph.D., Diana C. Rosentul, M.Sc., Andrew J. Carmichael, M.D., Chantal A.A. Smits-van der Graaf, M.D., Ph.D., Bart Jan Kullberg, M.D., Ph.D., Jos W.M. van der Meer, M.D., Ph.D., Desa Lilic, M.D., Ph.D., Joris A. Veltman, Ph.D., and Mihai G. Netea, M.D., Ph.D.

Abstr act

Background Chronic mucocutaneous candidiasis (CMC) is characterized by susceptibility to can- From the Departments of Medicine (F.L.V., dida infection of skin, nails, and mucous membranes. Patients with recessive CMC T.S.P., S.P.S., L.A.B.J., D.C.R., C.A.A.S.G., B.J.K., J.W.M.M., M.G.N.), Human Genet- and autoimmunity have mutations in the autoimmune regulator AIRE. The cause of ics (A.H., C.G., P.A., J.A.V.), and Pulmonary autosomal dominant CMC is unknown. Diseases (C.A.A.S.G.), Radboud University Nijmegen Medical Center, and the Nijme- Methods gen Institute for Infection, Inflammation, and Immunity (F.L.V., T.S.P., S.P.S., L.A.B.J., We evaluated 14 patients from five families with autosomal dominant CMC. We D.C.R., C.A.A.S.G., B.J.K., J.W.M.M., M.G.N.) incubated their peripheral-blood mononuclear cells with different combinations of — both in Nijmegen, the Netherlands; and the Department of Dermatology, James stimuli to test the integrity of pathways that mediate immunity, which led to the Cook University Hospital, Middlesbrough selection of 100 genes that were most likely to contain the genetic defect. We used an (A.J.C.), and the Institute for Cellular Medi- array-based sequence-capture assay, followed by next-generation sequencing, to iden- cine, Newcastle University, Newcastle upon Tyne (D.L.) — both in the United Kingdom. tify mutations. Address reprint requests to Dr. van der Meer at the Department of Medicine, Radboud Results University Nijmegen Medical Center, P.O. Box 9101, 6500 HB Nijmegen, the Nether- The mononuclear cells from the affected patients were characterized by poor pro- lands, or at [email protected]. duction of interferon-γ, interleukin-17, and interleukin-22, suggesting that the defect lay within the interleukin-12 receptor and interleukin-23 receptor signaling pathways. The following two groups of authors con- tributed equally to this article: Drs. van de We identified heterozygous missense mutations in the DNA sequence encoding the Veerdonk, Plantinga, and Hoischen; and coiled-coil (CC) domain of signal transducer and activator of transcription 1 (STAT1) Drs. Lilic, Veltman, and Netea. in the patients. These mutations lead to defective responses in type 1 and type 17 This article (10.1056/NEJMoa1100102) was helper T cells (Th1 and Th17). The interferon-γ receptor pathway was intact in these published on June 29, 2011, at NEJM.org. patients. N Engl J Med 2011. Conclusions Copyright © 2011 Massachusetts Medical Society. Mutations in the CC domain of STAT1 underlie autosomal dominant CMC and lead to defective Th1 and Th17 responses, which may explain the increased susceptibility to fungal infection. (Funded by the Netherlands Organization for Scientific Research and others.)

10.1056/nejmoa1100102 nejm.org 1 The New England Journal of Medicine Downloaded from nejm.org on June 30, 2011. For personal use only. No other uses without permission. Copyright © 2011 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

hronic mucocutaneous candidiasis autoimmune hepatitis, and the daughter has auto- (CMC) is a primary immunodeficiency dis- immune hemolysis and pernicious anemia, along Corder that is characterized by susceptibility with antiphospholipid antibodies. The daughter to infection of the skin, nails, and mucous mem- has had pulmonary embolism and Pneumocystis branes by candida species and dermatophytes.1 jirovecii pneumonia with symptomatic cytomega- There are several CMC subtypes: autosomal reces- lovirus infection. The son has extensive dermato- sive autoimmune polyendocrinopathy candidiasis phytosis (Trichophyton rubrum) of the feet but no with ectodermal dystrophy (APECED), autosomal autoimmune phenomena. We characterized the im- dominant CMC with or without thyroid disease, mune responses of peripheral-blood mononuclear and autosomal recessive, isolated CMC. cells (PBMCs) from the three affected family mem- The defect in APECED resides in the autoim- bers, the unaffected mother, and three healthy con- mune regulator AIRE, which has a key role in im- trol subjects. The father has nine unaffected sib- munotolerance.2 The susceptibility to candida in lings (Fig. 1A). We carried out genetic analyses of patients with APECED is attributed to autoantibod- all affected and unaffected family members. ies to interleukin-17 and interleukin-22,3 since type 17 helper T cells (Th17) are crucial for mucosal Families 2, 3, and 4 antifungal immunity.4 Little is known about the We performed immunologic and genetic analyses defects underlying susceptibility to candida in pa- of samples from nine members of three unrelat- tients with autosomal dominant CMC. In a pre- ed families of European descent from the United vious study, we found that Th1–interferon-γ re- Kingdom (Families 2, 3, and 4). All nine patients sponses were defective in patients with autosomal have autosomal dominant CMC with hypothyroid- dominant CMC.5 In a more recent study, investiga- ism. Their characteristics have been reported previ- tors found defective Th17 responses in patients ously9,10 (Table S1 in the Supplementary Appendix). with this disorder.6 Defective recognition of can- One patient in Family 2 has oral squamous-cell dida because of mutations in the dectin-1–CARD9 carcinoma; the father in Family 4 died of esopha- pathway leads to increased susceptibility to fun- geal cancer.9 All patients were screened for AIRE gi,7,8 but the clinical picture is less severe than mutations to rule out the APECED syndrome.11 We in autosomal dominant CMC. We sought the ge- also analyzed samples from unaffected members netic cause of susceptibility to mucocutaneous of Family 4. fungal infection in families with autosomal dominant CMC. Family 5 We analyzed two members with CMC and three Methods unaffected members of a Dutch family (Family 5). Both Patient 1 and her mother had esophageal can- Study Design cer (Table S1 in the Supplementary Appendix). The study was approved by the ethics committee at Radboud University Nijmegen Medical Center and Unaffected Subjects the Newcastle and North Tyneside research ethics We assessed 301 unrelated healthy Dutch control committee. Written informed consent was obtained subjects and 56 healthy British control subjects of from all family members and healthy control sub- European ancestry for the genetic mutations of jects who were evaluated. CMC. We also analyzed an in-house Nijmegen database of 100 exome data sets that were derived Family 1 from healthy subjects of European ancestry with- Family 1 was a nonconsanguineous family of Dutch out signs of CMC. A questionnaire concerning descent in which the father, daughter, and son have the ethnic origin of parents and grandparents of had severe CMC since early childhood (Fig. 1A, and the healthy subjects confirmed their ancestry in the Tables S1 and S2 in the Supplementary Appendix, Netherlands and the United Kingdom. available with the full text of this article at NEJM .org). The patients have severe oropharyngeal Immunologic Studies chronic candidiasis and severe dermatophytosis We incubated PBMCs (5×106 per milliliter), ob- and candidiasis of the feet (Fig. 1B). The father has tained by density centrifugation, at 37°C in 96-well

2 10.1056/nejmoa1100102 nejm.org The New England Journal of Medicine Downloaded from nejm.org on June 30, 2011. For personal use only. No other uses without permission. Copyright © 2011 Massachusetts Medical Society. All rights reserved. STAT1 Mutations in Chronic Mucocutaneous Candidiasis plates (Greiner),12 with or without heat-killed Can- A Family 1 dida albicans (1×106 microorganisms per milliliter [strain UC820]), Escherichia coli lipopolysaccharide (1 ng per milliliter) (Sigma-Aldrich), interleukin-1β (10 ng per milliliter), interleukin-12 (10 ng per milliliter), interleukin-18 (50 ng per milliliter), interleukin-23 (10 ng per milliliter) (R&D Systems), or interferon-γ (1 μg per milliliter) (Boehringer). We used an enzyme-linked immunosorbent assay to P1 measure levels of interleukin-1β, tumor necrosis factor α (TNF-α), interleukin-17, interleukin-22 (R&D Systems), interferon-γ, and interleukin-6 (PeliKine) after 24 and 48 hours of incubation (in the absence of serum) and after 5 days of incuba- P2 P3 tion (in the presence of 10% serum). All experi- ments were performed at least three times. B Clinical Manifestations

Patient 1 Patient 2 Sequence Capture and DNA Sequencing We applied array-based sequence capture followed by next-generation sequencing (454 Life Sciences) to analyze 100 genes from known immunologic pathways (Table S3 in the Supplementary Appen- dix). Details regarding coverage statistics are also provided in Table S4 in the Supplementary Patient 1 Patient 3 Appendix.

Validation of Mutations and Haplotype Analysis To validate the presence of mutations in signal transducer and activator of transcription 1 (STAT1) in affected patients, we amplified their DNA, using a polymerase-chain-reaction (PCR) assay, and se- Figure 1. Pedigree of Family 1 with Autosomal Dominant Chronic Mucocu- quenced the amplified DNA fragments by Sanger’s taneous Candidiasis (CMC) and Clinical Signs in Affected Family Members. method. All coding exons of the coiled-coil (CC) Panel A shows the pedigree of a Dutch family in which three members of domain of STAT1, including exon 10, were ampli- two generations have clinical symptoms characteristic of CMC (black sym- fied and analyzed. AIRE mutations were excluded by bols). Squares indicate male family members, and circles female family 13 members. The affected family members have severe dermatophytosis sequencing the gene as described previously. We and candidiasis of the feet and severe oropharyngeal chronic candidiasis used 250K single-nucleotide-polymorphism (SNP) (Panel B). Patients 1 and 2 also have other autoimmune disorders. arrays (Affymetrix)14 and Sanger sequencing to determine haplotypes associated with the STAT1 mutations. receptor 2 (TLR2), and dectin-1–receptor signaling Results pathways (Fig. 2B).15 In contrast, family mem bers with autosomal dominant CMC had low levels of Immunologic Defects in Family 1 helper T-cell cytokines interferon-γ, interleukin-17, Among unaffected members of Family 1, stimula- and interleukin-22 in PBMCs in response to can- tion of PBMCs with candida produced normal dida stimulation (Fig. 2C, and Fig S1 in the Sup- amounts of the cytokine proteins interleukin-1β, plementary Appendix). interleukin-6, and TNF-α (Fig. 2A, and Fig. S1A in We subsequently investigated the pathways the Supplementary Appendix), along with normal leading to the production of interferon-γ, interleu- activation of toll-like receptor 4 (TLR4), toll-like kin-17, and interleukin-22 by the patients’ T cells.

10.1056/nejmoa1100102 nejm.org 3 The New England Journal of Medicine Downloaded from nejm.org on June 30, 2011. For personal use only. No other uses without permission. Copyright © 2011 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

A B C. albicans LPS Pam3Cys β-glucan Pam3Cys+ β-glucan 10,000 800 40,000

8,000 600 30,000 6,000 400 20,000 4,000 200 10,000 2,000 Interleukin-6 (pg/ml) Interleukin-1β (pg/ml) Interleukin-1β (pg/ml) 0 0 0

Mother Mother Mother ControlControl 1 Control 2 3 PatientPatient 1 Patient 2 3 ControlControl 1 Control 2 3 PatientPatient 1 Patient 2 3 Control 1Control 2Control 3 Patient 1Patient 2Patient 3

C D C. albicans C. albicans Interleukin-12/Interleukin-18 1500 60,000 80,000

60,000 1000 40,000 40,000 500 20,000 20,000 Interferon-γ (pg/ml) Interferon-γ (pg/ml) Interleukin-17 (pg/ml) 0 0 0

Mother Mother Mother ControlControl 1 Control 2 3 PatientPatient 1 Patient 2 3 ControlControl 1 Control 2 3 PatientPatient 1 Patient 2 3 ControlControl 1 Control 2 3 PatientPatient 1 Patient 2 3

E Interleukin-1β/Interleukin-23 Interleukin-1β Interleukin-12 6000 2000 500

400 1500 4000 300 1000 200 2000 500 100 Interferon-γ (pg/ml) Interleukin-6 (pg/ml) Interleukin-22 (pg/ml) 0 0 0

Mother Mother Mother ControlControl 1 Control 2 3 PatientPatient 1 Patient 2 3 ControlControl 1 Control 2 3 PatientPatient 1 Patient 2 3 ControlControl 1 Control 2 3 PatientPatient 1 Patient 2 3

Figure 2. Immunologic Defects in the Affected Members of Family 1. Peripheral-blood mononuclear cells (PBMCs) from two unrelated healthy control subjects, the unaffected spouse of Patient 1 (Control 3), the unaffected mother of Patients 2 and 3, and the three affected family members (Patients 1, 2, and 3) were stimulated with heat-killed Candida albicans (1×106 microorganisms per milliliter) for 24 hours (Panel A); with the toll-like receptor 4 ligand Escherichia coli lipopolysaccharide (LPS) (1 ng per milliliter), the toll-like receptor 2 ligand tripalmitoyl-S-glycerylcysteine (Pam3Cys) (10 μg per milliliter), the dectin-1 ligand β-glucan (10 μg per milliliter), or a combination of β-glucan and Pam3Cys for 24 hours (Panel B); with heat-killed C. albicans (1×106 microorganisms per milliliter) for 5 days (for interleukin-17) and for 48 hours (for interferon-γ) (Panel C); with interleukin-12 (10 ng per milliliter) with or without interleukin-18 (50 ng per milliliter) for 48 hours (Panel D); and with a combination of interleukin-1β (10 ng per milliliter) and interleukin-23 (10 ng per milliliter) for 5 days (for interleukin-22) and with interleukin-1β (10 ng per milliliter) for 24 hours (for interleukin-6) (Panel E). All cytokines were measured by means of an enzyme-linked immunosorbent assay.

The induction of interferon-γ on exposure of these tectable interferon-γ (Fig. 2D), and exposure to cells to a combination of interleukin-12 and in- interleukin-23 and interleukin-1β resulted in no terleukin-18 was partially impaired (Fig. 2D). Ex- detectable interleukin-17 or interleukin-22 pro- posure to interleukin-12 alone resulted in unde- duction (Fig. 2E, and Fig S1B in the Supplemen-

4 10.1056/nejmoa1100102 nejm.org The New England Journal of Medicine Downloaded from nejm.org on June 30, 2011. For personal use only. No other uses without permission. Copyright © 2011 Massachusetts Medical Society. All rights reserved. STAT1 Mutations in Chronic Mucocutaneous Candidiasis tary Appendix). Mitogenic stimulation, however, induced normal interferon-γ production in these Interleukin-12 receptor Interleukin-23 receptor patients (data not shown). Normal production of Interleukin-12 Interleukin-23 receptor β1 interleukin-6 in response to interleukin-1β stim- Interleukin-12 Cytokine Cytokine receptor ulation proved that the pathway that is dependent receptor β1 Interleukin-12 on interleukin-1 receptor type I was intact (Fig. 2E). receptor β2 We concluded that both the interleukin-12 and interleukin-23 pathways are perturbed in affected family members (Fig. 3). Tyk2 Tyk2 P Jak2 P Jak2 T cell P P Missense Mutation in STAT1 STAT3 mutations, which are known to cause the P P P P hyperimmunoglobulin E syndrome and are involved in the interleukin-12 receptor and interleukin-23 receptor signaling pathways, were not pres- STAT1 ent in the three affected members of Family 1. Neither did we find mutations in STAT4, which en- STAT2 codes a protein that is also involved in these sig- STAT3 naling pathways, or in AIRE. On the basis of the STAT4 SOCS1 observed cytokine defects, we selected 100 genes STAT STAT P P SOCS3 encoding proteins that are relevant in interleu- STAT5 SOCS5 kin-12 and interleukin-23 signaling and in Th1 and STAT6 Th17 responses (Table S3 in the Supplementary Appendix). PIAS1 PIAS3 Sequence Analysis of 100 Candidate Genes Using array-based sequence capture followed by next-generation sequencing, we observed an average of 723 variants per sample. Considering the rare nature of the disease and the autosomal dominant pattern of inheritance, we decided to filter these variants against known SNP variants. Of the 723 identified variants, 651 corresponded with known SNPs or were located in a known polymor- phic region. After the exclusion of variants that Interferon-γ Interleukin-22 Interleukin-17 were detected by exome-sequencing projects,16,17 in an in-house database, or through the 1000 Ge- nomes Project, we were left with an average of 38 Figure 3. Molecules Shared by the Interleukin-12–Interleukin-23 Pathway. novel variants per patient (Table S8 in the Supple- The interleukin-12 and interleukin-23 receptor intracellular pathways share the mentary Appendix). Of these variants, only 4 per adapter molecules Tyk2 and Jak2, leading to downstream activation of the patient were nonsynonymous coding variants, STAT, SOCS, and PIAS proteins. The unresponsiveness to cytokines interleukin-12 and interleukin-23 in patients with autosomal dominant chronic COLOR FIGURE and of these, no more than 4 were called in at mucocutaneous candidiasis most likely suggests a defect in one of these least 20% of all reads and were therefore consid- Version 3 06/15/11 adapter proteins. Author van de Veerdonk ered to be heterozygous candidate variants. Fig # 3 Of the 11 candidate variants, only 7 were ob- Title ME served in affected family members. Of these 7 vari- to tryptophan) in the CC domain of STAT1 (Fig. S2 DE BP ants, 3 cosegregated in the three affected members and S3 in the Supplementary Appendix). In each Artist JM AUTHOR PLEASE NOTE: Figure has been redrawn and type has been reset of Family 1 (Table S9 in the Supplementary Ap- affected patient, STAT1 was the only gene with a Please check carefully pendix). These patients each carried a heterozy- heterozygous, nonsynonymous coding variant, Issue date 7/14/11 gous variant of the STAT1 gene in exon 10 mapping mak ing it the most likely candidate gene. One to chromosome 2 (c.820C→T; p.Arg274Trp). The unaffected member of the family (the wife of Pa- mutation predicts an amino acid change (arginine tient 1 and the mother of Patients 2 and 3) and

10.1056/nejmoa1100102 nejm.org 5 The New England Journal of Medicine Downloaded from nejm.org on June 30, 2011. For personal use only. No other uses without permission. Copyright © 2011 Massachusetts Medical Society. All rights reserved. T h e new england journal o f medicine

two unrelated healthy subjects did not carry any that there is a founder effect for each mutation, al- mutation in STAT1. Of all novel, heterozygous, though it is possible that the same mutation oc- nonsynonymous coding variants, those in STAT1 curred more than once on a common haplotype by affected amino acid residues that showed the stron- chance. The latter is corroborated by the fact that in gest conservation over 44 vertebrate species. Family 1, only Patient 1 is affected and nine siblings are unaffected (suggesting but not proving the ex- STAT1 Mutations in Other Affected Persons istence of a de novo mutation). Next we investigated three British families of Eu- ropean descent with autosomal dominant CMC Functional Studies (Fig. 4, and Tables S1 and S10 in the Supplementary Cytokine analysis of PBMCs obtained from patients Appendix). Using the same technique, we observed with CMC in the confirmation study revealed de- a different STAT1 variant (c.800C→T; p.Ala267Val), fects similar to those observed in Family 1: de- which also affected exon 10 in an affected patient. fective interferon-γ production in response to An affected member of another family (Patient 5 in interleukin-12 stimulation and no interleukin-22 Family 2) had very poor coverage for exon 10 of production in response to interleukin-1β and in- STAT1. However, manual read inspection showed terleukin-23 stimulation (Fig. S6A in the Supple- the same variant in two of five reads. In each of the mentary Appendix). In the affected members of affected patients, we confirmed the presence of this Family 1, we found that the addition of interferon-γ mutation on PCR amplification of the DNA se- to E. coli lipopolysaccharide resulted in increased quence encoding the CC domain of STAT1, followed TNF-α production, which showed that the interfer by Sanger sequencing (Fig. 4, and Table S10 in the on-γ signaling pathway was intact (Fig. S6B in the Supplementary Appendix). Supplementary Appendix). Sequence analysis of STAT1 in additional affected patients revealed the Arg274Trp variant in Discussion a member of a third family (Patient 1 in Family 4) and the Ala267Val variant in two patients from a We found that mutations affecting the CC domain second Dutch family with autosomal dominant of STAT1 in patients with autosomal dominant CMC CMC (Patients 1 and 2 in Family 5) (Fig. 4). We did lead to defective Th1 and Th17 responses, char- not identify the Ala267Val or Arg274Trp variant acterized by reduced production of interferon-γ, in 301 unrelated healthy Dutch control subjects or interleukin-17, and interleukin-22; these cytokines in 56 healthy British control subjects. Nor did we are crucial for the antifungal defense of skin and observe such variants in 100 exome data sets from mucosa. Affected members of Family 1 had normal subjects of Dutch ancestry or in the 179 subjects monocyte-dependent cytokine responses but se- of European descent with DNA samples that were verely impaired Th1 and Th17 responses. These sequenced as part of the 1000 Genomes Project.18 findings are in line with both studies showing de- Further analysis of the 100 in-house exomes iden- fective production of interferon-γ, interleukin-17, tified only a single novel missense change in the and interleukin-22 in patients with CMC5,6 or the STAT1 gene (c.314T→C), which we attribute to a hyperimmunoglobulin E syndrome19 and with the mapping artifact caused by an underlying repeat observation that interleukin-17–defective mice are element in exon 5. susceptible to oral candidiasis.4 STAT1 mutations have been described in pa- Haplotype Analysis tients with increased susceptibility to viruses and To determine whether the STAT1 mutations are mycobacteria.20-23 These mutations are located in founder mutations, we performed haplotype anal- regions of STAT1 that encode the Src homology ysis using high-density SNP arrays and Sanger se- 2 (SH2) or DNA-binding domains of the protein quencing of genomic DNA from affected and un- and result in defective signaling of interferon-γ affected members of all the families in our study. and type I interferon–receptor pathways.20-23 A We deduced that the Arg274Trp and Ala267Val mu- mutation in the CC domain of STAT1 (Phe172Ser) tations lie on different haplotypes (Table S11 in the resulted in decreased expression of STAT1 pro- Supplementary Appendix). Furthermore, each mu- tein,24 whereas other mutations in the CC domain tation lies on a distinct haplotype that is common blocked dimerization of nonphosphorylated mol- to all families bearing the mutation, suggesting ecules.25 The mutations affecting the CC domain

6 10.1056/nejmoa1100102 nejm.org The New England Journal of Medicine Downloaded from nejm.org on June 30, 2011. For personal use only. No other uses without permission. Copyright © 2011 Massachusetts Medical Society. All rights reserved. STAT1 Mutations in Chronic Mucocutaneous Candidiasis

Family 2

Male Female P1 CMC with candida CMC with candida and hypothyroidism P2 P3 P4

Family 3 Family 4 Family 5

NT P1 NT

P1 P3 NT NT P1

P2 P2

Figure 4. Confirmation of STAT1 Mutations in Patients with Chronic Mucocutaneous Candidiasis (CMC). Shown are the pedigrees of four additional families with autosomal dominant CMC: three families from the United Kingdom with CMC and thyroid disease (Families 2, 3, and 4) and one Dutch family with CMC and esophageal carcinoma (Family 5). In these families, 11 patients were tested for STAT1 mutations. Patients in Families 2, 3, and 5 were found to have mutation Ala267Val, and those in Family 4 were found to have mutation Arg274Trp. NT denotes not tested. that we observed in patients with autosomal of them carrying the c.800C→T mutation and one dominant CMC exclusively affect Th1 and Th17 carrying the c.820C→T mutation. It is known that responses, possibly by modifying the interaction thyrotropin induces the production of suppressor of STAT1 with its binding partners STAT3 and of cytokine signaling 1 (SOCS1), which in turn can STAT4.26 STAT1/STAT1 homodimers mediate sig- alter STAT1 phosphorylation.28 Thyrotropin may naling by the interferon-γ receptor, which induces act as a cytokine inhibitor in thyroid tissue, and resistance to intracellular microorganisms.26 In mutated STAT1 may hamper the rescue of thyroid terferon-γ–receptor signaling was preserved in the cells by thyrotropin and contribute to hypothyroid- patients with autosomal dominant CMC, which ism. In addition, there may be decreased iodine may explain their normal susceptibility to myco- accumulation, as is the case with Stat1-deficient bacteria and viruses. However, infections other mice.29 Finally, one of the affected Dutch patients than those caused by candida and dermatophytes, (and her deceased mother) and two members of ranging from bacterial chest infections to pneu- the British families, all of whom carried the mocystis and cytomegalovirus infections, occurred Ala267Val mutation, had esophageal or oral car- in the Dutch and British patients. cinoma.9 The loss of function of STAT1 has been Affected members in two of the families with linked to esophageal carcinoma.30,31 the Arg274Trp mutation had autoimmune disor- In conclusion, we found that mutations af- ders. STAT1 and interleukin-17 have been reported fecting the CC domain of STAT1 cause autosomal to be involved in the pathogenesis of autoimmune dominant CMC. Mutant STAT1 probably affects diseases.27 Interleukin-17 seems unlikely to be the the host defense against candida species through cause of these diseases, since it is produced in abnormal Th1 and Th17 responses. These find- low levels by PBMCs from affected patients. Three ings should facilitate the diagnosis of CMC in pa- affected patients have hypothyroidism, with two tients with chronic candidiasis.

10.1056/nejmoa1100102 nejm.org 7 The New England Journal of Medicine Downloaded from nejm.org on June 30, 2011. For personal use only. No other uses without permission. Copyright © 2011 Massachusetts Medical Society. All rights reserved. STAT1 Mutations in Chronic Mucocutaneous Candidiasis

Supported by grants from the Netherlands Organization for Health Research and Development (ZonMW grant 917.66.36, to Scientific Research (to Dr. Netea), the Primary Immunodefi- Dr. Veltman). ciency Association UK (to Dr. Lilic), the Techgene Project funded Disclosure forms provided by the authors are available with by the European Union (Health-F5-2009-223143, to Drs. Veltman the full text of this article at NEJM.org. and Arts), the Aneuploidy Project (LSHG-CT-2006-37627, to Drs. We thank our patients and their families; and Joanne Sedgwick, Hoischen and Veltman), and the Netherlands Organization for Nienke Wieskamp, and Dr. Jolien Tol for their contributions.

References

1. Kirkpatrick CH. Chronic mucocuta- as an additional diagnostic criterion for low virulence viruses in vivo. J Immunol neous candidiasis. J Am Acad Dermatol autoimmune polyendocrine syndrome 2006;176:5078-83. 1994;31:S14-S17. type I. J Clin Endocrinol Metab 2008;93: 22. Dupuis S, Dargemont C, Fieschi C, et 2. Finnish-German APECED Consor- 4389-97. al. Impairment of mycobacterial but not tium. An autoimmune disease, APECED, 12. Netea MG, Gow NA, Munro CA, et viral immunity by a germline human caused by mutations in a novel gene fea- al. Immune sensing of Candida albicans STAT1 mutation. Science 2001;293:300- turing two PHD-type zinc-finger domains. requires cooperative recognition of man- 3. Nat Genet 1997;17:399-403. nans and glucans by lectin and Toll-like 23. Dupuis S, Jouanguy E, Al-Hajjar S, et 3. Puel A, Döffinger R, Natividad A, et receptors. J Clin Invest 2006;116:1642-50. al. Impaired response to interferon-alpha/ al. Autoantibodies against IL-17A, IL-17F, 13. Cetani F, Barbesino G, Borsari S, et beta and lethal viral disease in human and IL-22 in patients with chronic muco- al. A novel mutation of the autoimmune STAT1 deficiency. Nat Genet 2003;33: cutaneous candidiasis and autoimmune regulator gene in an Italian kindred with 388-91. polyendocrine syndrome type I. J Exp Med autoimmune polyendocrinopathy-candi- 24. Hoshino A, Saint Fleur S, Fujii H. 2010;207:291-7. diasis-ectodermal dystrophy, acting in a Regulation of Stat1 protein expression 4. Conti HR, Shen F, Nayyar N, et al. dominant fashion and strongly cosegre- by phenylalanine 172 in the coiled-coil Th17 cells and IL-17 receptor signaling gating with hypothyroid autoimmune domain. Biochem Biophys Res Commun are essential for mucosal host defense thyroiditis. J Clin Endocrinol Metab 2001; 2006;346:1062-6. against oral candidiasis. J Exp Med 2009; 86:4747-52. 25. Zhong M, Henriksen MA, Takeuchi 206:299-311. 14. Xing J, Watkins WS, Zhang Y, With- K, et al. Implications of an antiparallel 5. van der Graaf CA, Netea MG, Drenth erspoon DJ, Jorde LB. High fidelity of dimeric structure of nonphosphorylated IP, te Morsche RH, van der Meer JW, whole-genome amplified DNA on high- STAT1 for the activation-inactivation cy- Kullberg BJ. Candida-specific interferon- density single nucleotide polymorphism cle. Proc Natl Acad Sci U S A 2005;102: gamma deficiency and toll-like receptor arrays. Genomics 2008;92:452-6. 3966-71. polymorphisms in patients with chronic 15. Netea MG, Brown GD, Kullberg BJ, 26. Yu H, Pardoll D, Jove R. STATs in mucocutaneous candidiasis. Neth J Med Gow NA. An integrated model of the rec- cancer inflammation and immunity: a 2003;61:365-9. ognition of Candida albicans by the in- leading role for STAT3. Nat Rev Cancer 6. Eyerich K, Foerster S, Rombold S, et nate immune system. Nat Rev Microbiol 2009;9:798-809. al. Patients with chronic mucocutaneous 2008;6:67-78. 27. Pfitzner E, Kliem S, Baus D, Litterst candidiasis exhibit reduced production of 16. Hoischen A, van Bon BW, Gilissen C, CM. The role of STATs in inflammation Th17-associated cytokines IL-17 and IL-22. et al. De novo mutations of SETBP1 cause and inflammatory diseases. Curr Pharm J Invest Dermatol 2008;128:2640-5. Schinzel-Giedion syndrome. Nat Genet Des 2004;10:2839-50. 7. Ferwerda B, Ferwerda G, Plantinga TS, 2010;42:483-5. 28. Park ES, Kim H, Suh JM, et al. Thy- et al. Human dectin-1 deficiency and mu- 17. Vissers LE, de Ligt J, Gilissen C, et al. rotropin induces SOCS-1 (suppressor of cocutaneous fungal infections. N Engl J A de novo paradigm for mental retarda- cytokine signaling-1) and SOCS-3 in FRTL- Med 2009;361:1760-7. tion. Nat Genet 2010;42:1109-12. 5 thyroid cells. Mol Endocrinol 2000;14: 8. Glocker EO, Hennigs A, Nabavi M, et 18. Durbin RM, Abecasis GR, Altshuler 440-8. al. A homozygous CARD9 mutation in a DL, et al. A map of human genome varia- 29. Kimura HJ, Rocchi R, Landek-Salga- family with susceptibility to fungal in- tion from population-scale sequencing. do MA, et al. Influence of signal trans- fections. N Engl J Med 2009;361:1727-35. Nature 2010;467:1061-73. ducer and activator of transcription-1 9. Koch D, Lilic D, Carmichael AJ. Auto- 19. Milner JD, Brenchley JM, Laurence A, signaling on thyroid morphology and somal dominant chronic mucocutaneous et al. Impaired T(H)17 cell differentia- function. Endocrinology 2009;150:3409- candidiasis and primary hypothyroidism tion in subjects with autosomal dominant 16. complicated by oesophageal carcinoma. hyper-IgE syndrome. Nature 2008;452: 30. Kaganoi J, Watanabe G, Okabe M, et Clin Exp Dermatol 2009;34(8):e818-e820. 773-6. al. STAT1 activation-induced apoptosis 10. Ng WF, von DA, Carmichael AJ, et al. 20. Chapgier A, Boisson-Dupuis S, of esophageal squamous cell carcinoma Impaired T(H)17 responses in patients Jouanguy E, et al. Novel STAT1 alleles in cells in vivo. Ann Surg Oncol 2007;14: with chronic mucocutaneous candidia- otherwise healthy patients with myco- 1405-15. sis with and without autoimmune poly- bacterial disease. PLoS Genet 2006;2(8): 31. Watanabe G, Kaganoi J, Imamura M, endocrinopathy-candidiasis-ectodermal e131. et al. Progression of esophageal carcino- dystrophy. J Allergy Clin Immunol 2010; 21. Chapgier A, Wynn RF, Jouanguy E, et ma by loss of EGF-STAT1 pathway. Can- 126(5):1015.e1-1015.e4. al. Human complete Stat-1 deficiency is cer J 2001;7:132-9. 11. Meloni A, Furcas M, Cetani F, et al. associated with defective type I and II IFN Copyright © 2011 Massachusetts Medical Society. Autoantibodies against type I interferons responses in vitro but immunity to some

8 10.1056/nejmoa1100102 nejm.org The New England Journal of Medicine Downloaded from nejm.org on June 30, 2011. For personal use only. No other uses without permission. Copyright © 2011 Massachusetts Medical Society. All rights reserved.