Signal transducer and activator of transcription 5B deficiency due to a novel missense mutation in the coiled-coil domain

A novel homozygous missense mutation in the coiled-coil domain of STAT5B causes an immunodeficiency syndrome characterised by short stature due to GH insensitivity, immune dysregulation and hypothyroidism but currently without pulmonary disease or T-cell lymphopenia.

Signal transducer and activator of transcription 5B deficiency due to a novel missense mutation in the coiled-coil domain To the Editor: We report a 17-year-old boy who presented with growth hormone (GH)-refractory growth failure (Fig 1, A), severe eczema (Fig 1, B), and autoimmunity. He was born to consanguineous parents at term with normal gestational weight (3.4 kg) and length (51 cm). He initially presented with autoimmune thyroiditis at age 4 years, with symptoms of lethargy, poor growth, delayed bone age, and constipation (Fig 1, C), confirmed by high titers of antithyroid peroxidase antibodies (>1000 IU/L; normal range, <34 IU/L) and antithyroglobulin antibodies (680 IU/L; normal range, <100 IU/L). He was treated with L-thyroxine (levothyroxine), which led to some improvement in his weight and growth. He also suffered from iron-deficiency anemia, which was treated with iron (III) hydroxide polymaltose (Maltofer), and ichthyosis from infancy, and developed atopic dermatitis at the age of 7 years. At age 9 years, he lost his head-hair, eyebrows, and eyelashes, and alopecia persisted despite treatment with steroids (prednisolone) on several occasions; this also failed to resolve his ongoing dermatitis. Three years later he was started on a gluten-free diet after antigliadin antibodies were detected in his blood; however, he had no anti-tissue transglutaminase antibodies and did not meet the clinical criteria of celiac disease. Furthermore, because there was no symptomatic improvement, the gluten-free diet was stopped after 2 years. Meanwhile, he maintained poor linear growth and delayed puberty. He had normal concentrations of basal GH, but low levels of insulin-like growth factor-1 (IGF-1) (Fig 1, D), IgA (see Table E1 in this article's Online Repository at www.jacionline.org), and vitamin D. In 3 GH provocation tests he had peak GH levels of less than 10 ng/mL (insulin, 7.35 ng/mL; glucagon, 3.17 ng/mL; arginine, 4.88 ng/mL), which is a response characteristic rather of GH deficiency than of GH insensitivity. However, therapeutic trial with GH for 2 years did not improve growth velocity, and IGF-1 remained below the normal limits throughout the follow-up. Treatment with cyclosporine was started when he was 14 years old and improved the eczema (Fig 1, B) and total alopecia, with some relapse of skin rash/thickened dry skin and alopecia after 1 year. Now he is in a stable condition on cyclosporine therapy, with greatly improved skin status and progress into puberty, although he is still small for his age. Hypothyroidism is well compensated by L-thyroxine therapy, but he has chronic recurring alopecia.
Assessment of his immunological status revealed a slight reduction in naive CD4 1 T and B cells, a striking increase in class-switched memory B cells, abundant activated HLA-DR 1 T cells, low levels of IgA and IgM, and increased levels of IgE (Table E1).
In an attempt to identify a presumed autosomal-recessive cause of this striking phenotype, we undertook whole-exome sequencing and found a novel homozygous missense mutation in the coiled-coil domain (CCD) of the gene STAT5B (signal transducer and activator of transcription [STAT] 5B; Fig 2, A). The nucleotide change c.452T>C, translating to p.L151P, was confirmed by Sanger sequencing and was found to be heterozygous in both parents and wild-type in the healthy sister (Fig 2, B). The mutation is predicted to be deleterious by several prediction programs (Fig 2, C). Indeed, our patient's phenotype was very compatible with that previously described for STAT5B deficiency. STAT5 functions downstream of cytokines such as IL-2, IL-4, IL-7, IL-9, IL-13, IL-15, IL-21, erythropoietin, thrombopoietin, and GH by translocating to the nucleus upon receptor engagement and activating transcription of growth factors, transcription factors, and other effector molecules. Even though STAT5A and STAT5B share greater than 90% homology, they have nonredundant as well as redundant functions. 1 The GH receptor preferentially uses STAT5B to induce expression of IGF-I, which promotes skeletal development and adipogenesis. 2 STAT5B also regulates expression of the transcription factor FOXP3 and the high-affinity IL-2 receptor subunit CD25, both key molecules in the development and function of regulatory T (Treg) cells. 1 Thus, human STAT5B deficiency is characterized by short stature due to GH insensitivity, and immune dysregulation typically manifest as eczema, chronic diarrhea, recurrent infections, and autoimmune diathesis. 2 STAT5B-deficient patients have normal levels of GH, but severely reduced serum levels of IGF-I, and are refractory to GH treatment. Other characteristic features are chronic lung diseases such as recurrent pneumonia, lymphocytic interstitial pneumonitis, and/or lung fibrosis, as well as moderate T-cell lymphopenia. 2,3 Since the first mutation in STAT5B was reported in 2003, 9 more patients with autosomal-recessive 2 and 13 individuals with autosomal-dominant 4,5 STAT5B deficiency were identified (Fig2, A). Homozygous mutations were loss-of-function, with 5 frameshift or nonsense mutations and 2 missense mutations in the SH2 domain. Both of these missense mutations cause aberrant folding of STAT5B, leading to protein aggregation, reduced expression levels, and impaired transcriptional activity. 6,7 AD STAT5B deficiency with variable penetrance was caused by heterozygous missense mutations in the CCD or the DNA-binding domain, with intact expression of a mutant protein with dominant-negative activity (Fig 2, A; see Table E2 in this article's Online Repository at www.jacionline.org). 4,5 The CCD is important for protein-protein interactions and nuclear import. 8 To explore the pathogenicity of our patient's p.L151P mutation, we first checked for STAT5B protein expression by

Patient
Healthy controls immunoblotting, showing normal levels in patient EBV-B cells compared with control (Fig 2, D). To test whether the mutant STAT5B is functionally active, we stimulated PBMCs from the patient and healthy controls with recombinant human IL-2 for 0 to 90 minutes and measured STAT5 phosphorylation by flow cytometry. Compared with 3 controls, the patient showed reduced STAT5 phosphorylation (Fig 2, E). Because the antibody used recognizes phosphorylated STAT5A and STAT5B, it is possible that the remaining signal entirely reflects phosphorylated STAT5A. A similar result was obtained using patient and control EBV-B cells (data not shown). We next investigated Treg-cell numbers and function, which have been shown to be reduced in patients with STAT5B deficiency. 2 Compared with the travel control, the CD4 1 CD25 hi CD127 2 Treg-cell population was reduced by about half in the patient's peripheral blood (Fig 2, F). Treg-cell suppressive capacity was assessed with Treg cells from the patient and a travel control (blood shipped for 36 hours) and an in-house control (fresh blood) (see Fig E1, A, in this article's Online Repository at www.jacionline.org). Patient's Treg cells showed significantly reduced suppressive capacity compared with those of the in-house control (P 5 .0021), but not the travel control (Fig E1,  B). Thus, unfortunately, no conclusion about Treg-cell function can be drawn. T-cell proliferation to all mitogens measured was in the normal range (Fig E1, C), albeit perhaps a lower response in the patient was more pronounced for IL-2 (21,894 counts per minute patient vs 47,107 counts per minute control) than for other mitogens. The significance of this finding is not clear, but could relate to a defect in STAT5B signaling downstream of the IL-2 receptor. How a missense mutation in the CCD affects phosphorylation of Y694 in the transcriptional activation domain is interesting to consider. It has been shown for Stat3 that the a1 helix of the CCD, through intramolecular interactions, is crucial for IL-6-induced recruitment of Stat3 to the IL-6 receptor and subsequent Stat3 phosphorylation, nuclear translocation, and DNA binding. 9 Because of high structural conservation between STAT proteins, a similar mechanism might apply to the role of STAT5B's CCD in the recruitment of STAT5B to the IL-2 receptor and subsequent phosphorylation. Interestingly, another STAT5B CCD mutation, Q206R, also inhibits IL-2-induced STAT5B-Y694 phosphorylation. 4 This heterozygous mutant has a dominant-interfering effect on STAT5B transcriptional function and leads to a phenotype of autoimmunity, lymphoproliferation, granulocytosis, and hypogammaglobulinemia (patient 1), or multiple sclerosis, arthritis, and recurrent infections (patient 2), but no GH insensitivity, delayed puberty, eczema, or pulmonary disease (see Table E2 in this article's Online Repository at www.jacionline.org). 4 In contrast, a further heterozygous STAT5B CCD mutant, Q177P, is robustly phosphorylated, but exerts a dominant-negative effect by abrogating STAT5B nuclear import. 5 This mutation results in GH-insensitive growth failure, delayed puberty, reduced IGF-1 levels, and severe eczema, as seen in our L151P patient and the 4 patients with nonsense and frameshift mutations in the CCD. 2,5 Both our patient and the Q177P index patient had normal instead of elevated levels of prolactin, and raised IgE levels were found in both Q177P and 1 R152X patient (Table E2).
In summary, this is the first report of a homozygous hypomorphic missense mutation in the STAT5B CCD leading to functional defects and a phenotype of STAT5B deficiency. This included GH-refractory growth failure, severe eczema, and autoimmune disease but to date without pulmonary problems, severe infections, or T-cell lymphopenia. It would have been ideal to treat the GH insensitivity with IGF-1, but unfortunately, that was not available for the patient. We observed a gratifying clinical response of the atopic dermatitis to cyclosporine therapy, coinciding with improved nutritional status, growth, and general well-being. However, the long-term outlook remains guarded. Alternative treatment strategies for immune dysregulation one might consider if symptoms deteriorated include sirolimus therapy or even stem cell transplantation.
For detailed methods, please see the Methods section in this article's Online Repository at www.jacionline.org.

Study subjects
Patients and their relatives provided written informed consent to participate in research protocols approved by the Newcastle and North Tyneside 1 Research Ethics Committee. Whole blood samples were obtained from these individuals, and genomic DNA was isolated using the DNeasy or QIAamp DNA mini kit (Qiagen, Manchester, UK).

PCR and sequencing analysis
Specific primers were designed in Primer3web version 4.0.0 (http://bioinfo. ut.ee/primer3/). Primer sequences are available on request. Capillary sequencing was performed according to standard methods. Sequences were aligned with the consensus coding sequence (human genome assembly 38) in nucleotide BLAST (http://blast.ncbi.nlm.nih.gov/blast/). ChromasLite Version 2.1.1 was used for visualization of the sequences.

Whole-exome sequencing
Whole exome of the patient was captured using the Agilent SureSelect V5 kit and a library subsequently prepared for paired-end sequencing. This was carried out on an Illumina NextSeq instrument, on a high-output flowcell by OGT (Oxford Gene Technology, Begbroke, UK). Sequencing reads were assessed for quality using the FastQC and MultiQC tools. E1,E2 Reads were aligned to the hg19 assembly of the human reference genome, and prepared for variant calling. Exomes were analyzed according to GATK best practices, E3 as follows: the raw alignment was realigned around indels, quality scores recalibrated (Base Quality Score Recalibration [BQSR]), and duplicates marked using the Picard suite. Genotyping was performed using a combination of the HaplotypeCaller from GATK in the gVCF mode, and JointCalling program leveraging a catalogue of exome samples from other patients with primary immunodeficiency. To avoid false-positive calls, Variant Quality Score Recalibration (VQSR) was performed on the raw variant callsets for indels and single nucleotide polymorphisms. Downstream analysis was carried out using the Web-based analytics application Ingenuity Variant Analysis (Qiagen).

Cell lines
EBV-immortalized B-cell lines were generated from PBMCs after Ficollpurification from whole blood (EDTA). Cells were incubated with EBVcontaining medium for 1 hour initially, and then in the presence of 1 mg/mL cyclosporine A for a further 24 hours at 378C, 5% CO 2 . Cells were then transferred to fresh 10% FCS/cyclosporine A-containing RPMI and incubated until foci of immortalized cells appeared.

Lymphocyte proliferation
PBMCs were cultured for 3 days with the respective mitogens, and proliferative responses were measured as counts per minute of tritiated thymidine incorporated during cell division. Cultures were performed in triplicate, with results expressed as the means of the triplicates. The stimulatory index is calculated as stimulated over unstimulated (control) counts.