Immunodeficiency and disseminated mycobacterial infection associated with homozygous nonsense mutation of IKKβ☆

A, Genealogical tree for the kindred. Arrow highlights index case. Asterisks denote family members who died within the first 2 months of life with febrile illnesses. B, Histology showing granuloma in skin by hematoxylin and eosin staining.

investigations demonstrated normal proportions and numbers of naive and memory T, B, and natural killer cells, normal T-cell repertoire, and T-cell proliferation to PHA (Table I). Expression of MHC class II molecules was preserved. Neutrophil number, respiratory burst, and integrin expression were normal, as was phenotyping for monocytes and dendritic cell populations. Immunoglobulin levels had been documented to be low on repeated sampling at age 3 months (IgG 1.2 g/L, IgA < 0.05 g/L, and IgM < 0.05 g/L) and therefore immunoglobulin replacement had already been commenced. Liver function test results were normal (including aspartate transaminase, gamma-glutamyl transpeptidase, alkaline phosphatase, albumin, total protein, and total bilirubin levels). Whole blood cytokine assays on 2 occasions demonstrated severely impaired production of IFN-g to all stimuli, absent production of IL-17, and a markedly reduced production of the proinflammatory cytokines, TNF-a and IL-6, in response to a range of Toll-like receptor ligands (Fig 1, C). Stimulated IL-12 production was also reduced and not significantly rescued by addition of exogenous IFN-g. The pattern of cytokine response and clinical features were not consistent with a classical IL-12/INF-g pathway defect but would be consistent with a defect in the NF-kB pathway.
Our patient 7 was treated with ethambutol, rifampicin, isoniazid, and pyridoxine as well as prophylactic azithromycin, fluconazole, and intravenous immunoglobulin in view of her immunodeficiency. INF-g treatment was added at 50 mg/m 2 and gradually increased to 200 mg/m 2 . 7 This was not well tolerated, and she developed ulcerating skin inflammation and increasing hepatosplenomegaly necessitating low-dose steroid treatment (0.5 mg/kg) and therefore IFN-g treatment was withdrawn. She continued to have recurrent episodes of fever and increased inflammatory markers that most likely reflected her BCGosis. She subsequently developed a progressive respiratory distress, was admitted to intensive care for noninvasive ventilation, and, unexpectedly, died from a massive gastrointestinal and pulmonary hemorrhage after a surgical central line insertion at age 25 months.
To identify the causative mutation, we used exome sequencing. Blood sample of the patient was obtained with informed consent from the parents in accordance with the Declaration of Helsinki and with approval from the ethics committees (04/Q0501/119 and 06/Q0508/16). Library preparation, exome capture, and sequencing have been done according to the manufactures instructions. For exome target enrichment, Agilent SureSelect 50 Mb kit was used. Sequencing was done using Illumina HiSeq with 94 bp paired-end reads. In the exome data, we found 22,754 single-nucleotide variants and small insertions/deletions, including 172 very rare ones, that is, those not seen in the 1000 Genomes database (April 2012 data release) and our internal databases. Five of these variants were homozygous (see Table E1 in this article's Online Repository at www.jacionline.org). Among these we identified a homozygous nonsense mutation c.321C>A in the IKBKB gene that encodes the IKKb protein, leading to a premature stop codon p.Y107X (Fig 1, D; see Fig  E2, A, in this article's Online Repository at www.jacionline. org). In the exome sequence data, we found no mutations in genes encoding IkBa, NEMO, or other proteins of the NF-kB pathway. Because gene IKBKG that encodes NEMO was poorly covered in the exome data, we studied it by Sanger sequencing and again found no mutations. We then confirmed the IKBKB mutation by Sanger sequencing and showed that both parents were heterozygous carriers of this mutation (Fig E2, B). We then purified proteins from PBMCs and demonstrated that the patient completely lacked IKKb expression (Fig 1, E). We used polyclonal rabbit anti-IKKb antibody (Cell Signaling Technology cat no. 2678) produced by immunizing animals with a synthetic peptide corresponding to residues surrounding Leu570 of the 756 amino acid IKKb protein. Therefore, presence of a shorter IKKb protein that potentially could be generated after reinitiation of translation seems unlikely. Further functional experiments were not possible, because no patient-derived cell lines were available after the patient's death. Given that functional reconstitution experiments were impossible, we cannot formally prove that the IKKb deficiency has caused the immune phenotype of the patient, although it is very likely in the absence of other candidate loss-of-function mutations. However, additional unknown modifier genetic or nongenetic factors may have affected the severity of the patient's phenotype.
Here we demonstrate that an autosomal-recessive human IKKb deficiency has a clinical presentation, which resembles both hypermorphic IkBa and hypomorphic NEMO mutations, with a prominent susceptibility to mycobacteria and an increased frequency of other bacterial infections. No loss-of-function alleles in the IKBKB gene were previously found in up to 6500 healthy subjects (National Heart, Lung, and Blood Institute exome variant server, accessed May 2013), suggesting that such mutations are likely to be pathogenic. Recently, a conference abstract has been published describing a group of patients with severe combined immunodeficiency with agammaglobulinemia that were associated with a frameshift mutation in the IKBKB gene, which apparently led to an abolished IKKb expression. 8 In contrast to our patient, they had reduced numbers of natural killer cells and manifested with invasive bacterial and viral, but not mycobacterial, infections. Furthermore, no ectodermal dysplasia was reported, which distinguishes them from our patient and most patients with NEMO and IkBa mutations. 5 The reasons for such discrepancies should be clarified when a more detailed description of those patients is published.
The IKKb-deficient mouse embryos die because of liver degeneration and apoptosis. [9][10][11] In contrast, our data suggest that humans with the IKKb deficiency can be viable. This is potentially explained by the compensatory mechanisms involving IKKa and the noncanonical pathway of NF-kB activation that are unlikely to be affected by the IKKb deficiency. Interestingly, while NF-kB activation in cells of the IKKb-deficient mice was significantly impaired, it was not completely abolished, as residual signaling was mediated by IKKa. 9,10 In the absence of IKKb, IKKa can make homodimers, which can bind NEMO and phosphorylate IkBa, thus partially compensating for the loss of the IKKa/IKKb heterodimers. 9, 12 We did not observe significant biochemical liver disease in our patient, which may also relate to compensatory mechanisms or result from species variation.
The severe granulomatous reaction seen in our patient may be a feature of systemic infection. 13 However, altered keratinocyte function may additionally cause skin inflammation because mice with epidermis-specific deletion of IKKb develop a severe TNF-mediated inflammatory skin disease. 14 Our patient had chronic diarrhea, similar to most of the patients with the IkBa mutations. This is consistent with the importance of the NF-kB pathway for mucosal immunity, maintaining intestinal epithelia integrity and homeostasis in the gut. 15 In line with other defects of NF-kB signaling, hypogammaglobulinemia and ectodermal dysplasia, characterized by peg teeth, were features of IKKb deficiency in our case. The immunological phenotype was characterized by a profound defect in the production of proinflammatory cytokines despite intact development of all immune cell lineages. Although the clinical spectrum of the condition will be clarified by the identification of other affected families, the severity of immunodeficiency in our patient and failure to respond to conservative therapy suggest that bone marrow transplantation should be considered for children with complete IKKb deficiency.     DOCK8 deficiency is characterized clinically by recurrent viral, fungal, and sinopulmonary infections. [1][2][3][4] These patients also have severe atopic dermatitis, allergies, and an increased likelihood of developing cancer. DOCK8 deficiency affects both innate and adaptive immunity and causes combined immunodeficiency.
We present a patient who fulfilled clinical and immunological parameters for HIES (see the Case report section in this article's Online Repository at www.jacionline.org). Mutations in the signal transducer and activator of transcription 3 gene were ruled out. At this stage, a DOCK8 deficiency was suspected. Using comparative genome hybridization, a heterozygous deletion of exons 2 to 8 was identified in the DOCK8 gene (see Fig E1, A, in this article's Online Repository at www.jacionline.org). Sequence analysis also demonstrated a heterozygous G > A mutation at position 15 in the 59 splice donor site of intron 17 (IVS1715 G>A) deleting the exon 17 (Genbank KC736820). This alteration corresponded to a deletion in the conserved DHR1 domain of the protein (Fig E1, B and C) and was absent in 50 healthy donors. Some normal DOCK8 transcripts were present, indicating that the exon 17 splicing mutation did not abrogate normal processing of this allele (Fig E1, D). DOCK8 transcript expression was lower in the patient than in controls in PBMCs, as well as CD4 1 and natural killer (NK) cells ( Fig  E1, E); however, the DOCK8 protein was clearly detected in the patient by Western blot, albeit at reduced levels. On the basis of these findings, we conclude that this is a new case of partial DOCK8 deficiency (Fig E1, F).
Abnormalities in the T-cell compartment have been described in classical DOCK8 deficiency, 2-4 and we decided to assess the impact of a partial DOCK8 deficiency on T-cell phenotype and function. Our patient showed defective thymopoiesis with an absence of T-cell receptor rearrangement excision circles, reduced recent thymic emigrants, a restricted T-cell repertoire, and decreased naive CD4 1 T cells in PBMCs (Table I; see Fig  E2, A, B, and C, and Fig E3, A, in this article's Online Repository at www.jacionline.org). Furthermore, there was an increase in plasma levels of IL-7 (Table I), which could reflect a compensatory attempt to boost the expansion of recent thymic emigrants cells to overcome the depletion of peripheral CD4 1 T cells. Interestingly, in spite of naive CD4 1 T-cell lymphopenia, the patient's T-cell proliferation was unaffected and similar to that in controls (Fig E3, B). This could be related to the nonexhausted phenotype showed by the patient, with normal levels of T-effector memory CD45RA (TEMRA) CD8 1 T cells (Fig E3, A), in contrast to classical loss-of-function/expression of DOCK8 in patients