First Identification of Biallelic Inherited DUOX2 Inactivating Mutations as a Cause of Very Early Onset Inflammatory Bowel Disease.

By catalyzing generation of hydrogen peroxide (H202) in enterocytes, DUOX2 participates in innate defense against intestinal microbes.1Ha E.M. et al.Science. 2005; 310: 847-850Crossref PubMed Scopus (581) Google Scholar, 2Flores M.V. et al.Biochem Biophys Res Commun. 2010; 400: 164-168Crossref PubMed Scopus (76) Google Scholar One monoallelic exonic variant of DUOX2 was recently found in 15 out of 19 Crohn disease-affected patients in a large Ashkenazi family.3Levine A.P. et al.Gastroenterology. 2016; 151: 698-709Abstract Full Text Full Text PDF PubMed Scopus (42) Google Scholar Herein, we report the first identification of biallelic, inherited mutations of DUOX2 as a Mendelian cause of very early onset inflammatory bowel disease (VEO-IBD) in a boy born from unrelated parents of Italian origin. The patient developed pancolitis with bloody diarrhea at the age of 3. Histology showed dense lymphoplasmocytic and eosinophilic infiltrates, mucin depletion, architectural distortion and crypt abscesses without epitheloid granuloma (Figure 1A). No extragastrointestinal manifestation was reported. Mesalamine and systemic steroids induced symptom resolution, but clinical relapse occurred with steroid withdrawal. Thiopurines allowed complete remission but were discontinued after 1 month because of increased pancreatic enzymes. He is currently in remission under mesalamine. A known molecular defect was excluded using targeted sequencing of 66 VEO-IBD–associated genes. Whole exome sequencing performed on patient and parents identified compound heterozygous missense variants in DUOX2, both confirmed by Sanger sequencing (Figure 1B, C). The first variant substituted proline by serine (NM_014080.4:c.1825G>A; Minor Allele Frequency: 2.7 × 10−4 P609S) within the first transmembrane domain (Figure 1D). The second missense variant replaced arginine with histidine (NM_014080.4:c.857G>A; R286H) in the peroxidase homology domain of DUOX2 (Figure 1D). R286H was absent from both public and Imagine databases. Multiple sequence alignment showed that mutations affected evolutionarily conserved residues (Supplementary Figure 1A), and both amino acid changes were predicted to be damaging. Membrane targeting and catalytic activity of the DUOX enzyme complexes requires heterodimerization of DUOX2 with DUOXA2.4Luxen S. et al.J Cell Sci. 2009; 122: 1238-1247Crossref PubMed Scopus (70) Google Scholar When expressing HA-tagged WT DUOX2 or mutants in H661-DUOXA2 cells, comparable amounts of WT DUOX2, R286H and P609S proteins were detected in total cell lysates (Supplementary Figure 1B). In contrast, flow cytometry revealed reduced surface expression of P609S compared to WT DUOX2, whereas R286H was undetectable (Figure 1E). Overall decreased surface expression of DUOX2 was observed when the mutants were expressed at 1:1 ratio (Figure 1E), suggesting that DUOX2 mutants were retained in the endoplasmic reticulum.4Luxen S. et al.J Cell Sci. 2009; 122: 1238-1247Crossref PubMed Scopus (70) Google Scholar Staining of DUOX2 carried out on patient's biopsies showed markedly decreased protein expression compared with controls (Figure 1F). H2O2 generation upon thapsigargin stimulation was undetectable in H661-DUOXA2 cells expressing R286H and partially but significantly decreased in cells expressing P609S (Figure 1G) or coexpressing the 2 mutants at 1:1 ratio. Overall, our data provide evidence that rare, inherited, biallelic mutations in DUOX2 were the molecular cause of this VEO-IBD case. Both mutations are reminiscent of monoallelic or biallelic DUOX2 mutations that cause either transient or permanent congenital hypothyroidism.5Grasberger H. Mol Cell Endocrinol. 2010; 322: 99-106Crossref PubMed Scopus (83) Google Scholar Of note, thyroid function was normal in the affected boy, a finding consistent with the highly variable penetrance of DUOX2 mutations. In contrast, the child developed severe intestinal inflammation. DUOX2 is expressed in all gut segments where it is up-regulated by microbiota6Grasberger H. et al.Gastroenterology. 2015; 149: 1849-1859Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar and inflammation-derived signals.7Haberman Y. et al.J Clin Invest. 2014; 124: 3617-3633Crossref PubMed Scopus (332) Google Scholar Very rare monoallelic mutations of DUOX2 were also reported as risk factor for VEO-IBD in 2 patients.8Hayes P. et al.Cell Mol Gastroenterol Hepatol. 2015; 1: 489-502Abstract Full Text Full Text PDF PubMed Scopus (91) Google Scholar Together with our data, these observations strengthen the hypothesis that, alike in congenital hypothyroidism, both monoallelic and biallellic DUOX2 mutations can increase susceptibility to intestinal inflammation. The important function of DUOX2 was first evidenced in invertebrates. Lack of the duox2 homolog in Drosophila gut or knock-down of duox in Zebrafish increased susceptibility against a broad spectrum of enteropathogens,1Ha E.M. et al.Science. 2005; 310: 847-850Crossref PubMed Scopus (581) Google Scholar, 2Flores M.V. et al.Biochem Biophys Res Commun. 2010; 400: 164-168Crossref PubMed Scopus (76) Google Scholar pointing out a nonredundant antibacterial role of DUOX2. Altogether these observations indicate that DUOX2-mediated H2O2 production plays a crucial role in controlling overgrowth of bacteria at the intestinal surface and preserving intestinal homeostasis. We suggest that the biallelic, inactivating DUOX2 mutations described here might license overgrowth of mucosa-associated bacteria, which triggers colitis. Our findings expand the clinical phenotype and inheritance mechanism of DUOX2 defects, and indicate that monoallelic and biallelic DUOX2 mutations should be tested for in genetic screens as a monogenic cause of IBD. The authors thank K. O’Neill and A. Singh for technical support. Informed written consent for genetic study was obtained from the family in accordance with the Helsinki Declaration. Biological evaluation showed anemia (hemoglobin 10 g/dL), increased sedimentation rate (60 mm/h), high fecal calprotectin (560 μg/g), moderately increased lymphocyte counts with normal lymphocytes distribution, and normal serum immunoglobulins, but presence of cytoplasmic anti-neutrophil cytoplasmic antibodies. Nitroblue tetrazolium showed normal production of reactive oxygen species by neutrophils. Genomic DNA from peripheral blood cells was isolated using the QIAamp DNA Blood Mini Kit (Qiagen, Courtaboeuf, France) according to manufacturer's instructions. Mutations in 66 genes previously associated with VEO-IBD were first screened out by next generation sequencing (Charbit-Henrion et al, submitted). The DNA from the patient and his parents were analyzed by whole exome sequencing using the 58Mb V6 version of the SureSelect Exome kit (Agilent, Santa Clara, CA) and a HiSeq2500 (Illumina, San Diego, CA). High-quality exome data were obtained for all the samples analyzed: the mean depth of coverage of the exonic targeted bases was superior to 144X with >97% of the targeted bases covered by a minimum of 30 independent reads. Annotated data were analyzed with in-house software (Polyweb). Candidate variants were ranked by filtering out conservative and known polymorphisms reported in public databases (Human Gene Mutation Database, NHLBI GO Exome Sequencing Project, 1000 Genomes Project, Exome Aggregation Consortium, database SNP), which resulted in 8736 rare variants. From this list, 252 variants were inherited in an autosomal recessive and X-linked manner, and 244 were further screened out because already present in Imagine database (8000 exomes), and thus considered unreported polymorphisms. Potential pathogenicity of each protein-coding variant was evaluated using OMIM (Online Mendelian Inheritance in Man), evolutionary conservation and prediction tools (SIFT, PolyPhen version 2.2.2, Mutation Taster, CADD). Identified variants were then filtered, retaining only missense and nonsense, with a CADD score greater than or equal to the MSC1Itan Y. et al.Nat Methods. 2016; 13: 109-110Crossref PubMed Scopus (184) Google Scholar narrowing down this list to compound heterozygous mutations in DUOX2. No mutation was found in DUOX1, DUOXA1, or DUOXA2 genes. Crohn disease-associated exonic variants (ATG16L1 T300A, IL23R R381Q, FUT2 W134X, NOD2 R702W G908R L1007fsinC, and CSF2RB S709LX22) were not detected by whole exome sequencing.3Martyn K.D. et al.Cell Signal. 2006; 18: 69-82Crossref PubMed Scopus (616) Google Scholar Overall, in silico analysis suggested that the patient had autosomal recessive DUOX2 deficiency. Genomic DNA was amplified by standard techniques using oligonucleotide primers flanking the exon 6 and exon 15 of DUOX2 respectively (forward 5'- CACCCTGCCCTCAACCTAAG -3', reverse 5'- GCTTCTCTCTGGAGTGGTGG- 3') and (forward 5'-TCTGGTTCCGCTACCACAAC- 3', reverse 5'- CCCTCAAATCCCCTGGTTCC- 3') using Ampli TaqDNA Polymerase (Applied Biosystems, Foster City, CA). Polymerase chain reaction products were purified with the QIAquick PCR Purification kit (Qiagen) and Sanger sequenced using the same primers by Eurofins on the Genomic Platform of Université Paris Descartes. NCI-H661 cells (deprived of all known NADPH oxidases) were cultivated in RPMI 1640 medium with 10% fetal bovine serum. Both variants, c.857 G>A, p.Arg286His and c.1825 C>T, p.Pro609Ser, were introduced separately into HA-tagged DUOX2 in pcDNA3.113 by site-directed mutagenesis and were verified by sequencing. DUOX2 wild-type and missense variants were transiently cotransfected with DUOXA2 into H661 cells using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA) for 48 hours. To replicate to some extent the compound heterozygosity of the patient, both mutants were additionally transfected at a 1:1 ratio. Proteins were extracted in RIPA buffer, separated by SDS-page and immunoblotted with HA antibody (Covance Laboratories, Princeton, NJ), DUOX1/2,2Pacquelet S. et al.J Biol Chem. 2008; 283: 24649-26458Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar and actin (Sigma-Aldrich, St. Louis, MO), followed by secondary horseradish peroxidase-conjugated anti-rabbit (Southern Biotech) or anti-mouse antibody (Cell Signaling, Danvers, MA). Proteins were visualized using ECL reagent (Pierce, Waltham, MA). H2O2 production by DUOX2, stimulated through the addition of 1 μmol/L of thapsigargin, was measured using the homovanillic acid assay.3Martyn K.D. et al.Cell Signal. 2006; 18: 69-82Crossref PubMed Scopus (616) Google Scholar H2O2 production was standardized to H2O2 standard curves and cell lysate protein concentration measured using the BCA assay. Empty vector transfection served as a control. Transfected NCI-H661 cells were incubated with HA antibody in FACS buffer on ice for 30 minutes without permeabilization. After incubation with anti-mouse Alexa Fluor 647, cells were fixed in 1.5% PFA and analyzed using an Accuri C6 flow cytometer (BD Biosciences, San Jose, CA). Intestinal biopsies were fixed in 4% formaldehyde and embedded in paraffin. Sections were stained with hematoxylin and eosin. Images were acquired using a Nikon Eclipse 80i microscope with Plan-Apochromat 10×/0.45 DIC N1 objective and EOS 600D camera. For immunofluorescence study, sections on slides were deparaffinized with xylene and rehydrated with different percentages of ethanol. Slides were placed in citrate buffer containing 0.05% Tween 20 for 25 minutes at 95°C and then washed with ice-cold phosphate-buffered saline (PBS). Slides were incubated in blocking buffer (5% fetal bovine serum, 3% bovine serum albumin in PBS) for 1 hour at room temperature and then incubated with anti-DUOX1/2 antibody (1:600) at 4°C overnight, washed with PBS, and then incubated with anti-rabbit Alexa Fluor 488 for 1 hour at room temperature. After a final series of PBS washes, the tissue sections were stained with 300 μmol/L DAPI for 10 minutes before mounting. Images were acquired using a Zeiss LSM 700 microscope with a Plan-Apochromat 20°/0.75 DIC N2 objective. Experiments were conducted in triplicate with 3 biological repeats (n = 3), followed by an unpaired Student’s t test.