Large deletions and point mutations involving the dedicator of cytokinesis 8 (DOCK8) in the autosomal-recessive form of hyper-IgE syndrome

Article Outline

• Abstract
• Methods
  • Patients and controls
  • Methods
• Results
  • Search for copy number variations, microdeletions, and homozygosity
  • Positional identification of candidate genes
  • Mutation detection in DOCK8
  • DOCK8 deficiency impairs T-cell activation
  • Clinical phenotype of DOCK8 deficiency
• Discussion
• Acknowledgment
• Methods
  • Homozygosity mapping and copy number analysis
  • PCR and sequence analysis
  • Immunoblotting
  • Proliferation studies
  • CFSE dilution studies
• Fig E1. 
• Fig E2. 
• Table E1. 
• References
• References
• Copyright

Background

The genetic etiologies of the hyper-IgE syndromes are diverse. Approximately 60% to 70% of patients with hyper-IgE syndrome have dominant mutations in STAT3, and a single patient was reported to have a homozygous TYK2 mutation. In the remaining patients with hyper-IgE syndrome, the genetic etiology has not yet been identified.

Objectives

We aimed to identify a gene that is mutated or deleted in autosomal recessive hyper-IgE syndrome.

Methods

We performed genome-wide single nucleotide polymorphism analysis for 9 patients with autosomal-recessive hyper-IgE syndrome to locate copy number variations and homozygous haplotypes. Homozygosity mapping was performed with 12 patients from 7 additional families. The candidate gene was analyzed by genomic and cDNA sequencing to identify causative alleles in a total of 27 patients with autosomal-recessive hyper-IgE syndrome.

Results

Subtelomeric biallelic microdeletions were identified in 5 patients at the terminus of chromosome 9p. In all 5 patients, the deleted interval involved dedicator of cytokinesis 8 (DOCK8), encoding a protein implicated in the regulation of the actin cytoskeleton. Sequencing of patients without large deletions revealed 16 patients from 9 unrelated families with distinct homozygous mutations in DOCK8 causing premature termination, frameshift, splice site disruption, and single exon deletions and microdeletions. DOCK8 deficiency was associated with impaired activation of CD4⁺ and CD8⁺T cells.

Conclusion

Autosomal-recessive mutations in DOCK8 are responsible for many, although not all, cases of autosomal-recessive hyper-IgE syndrome. DOCK8 disruption is associated with a phenotype of severe cellular immunodeficiency characterized by susceptibility to viral infections, atopic eczema, defective T-cell activation and T_h17 cell differentiation, and impaired eosinophil homeostasis and dysregulation of IgE.

Key words: Autosomal recessive hyper-IgE syndrome, human gene mutation, DOCK8, primary immunodeficiency, molluscum contagiosum, recurrent infection, T cells, T_H17 cells, eosinophils, IgE regulation, copy number variations, genomic deletions

Abbreviations used: AR, Autosomal-recessive, CFSE, Carboxyfluorescein succinimidyl ester, CNV, Copy number variation, DHR, Dedicator of cytokinesis homology region, DOCK, Dedicator of cytokinesis, GEF, Guanine nucleotide exchange factor, HIES, Hyper-IgE syndrome, NIH, National Institutes of Health, NK, Natural killer, STAT3, Signal transducer and activator of transcription 3, WASP, Wiskott-Aldrich syndrome protein

The hyper-IgE syndromes (HIES; also called Job syndrome; OMIM: Online Mendelian Inheritance in Man #147060 and #243700) are rare primary immunodeficiencies (estimated prevalence <1:1 million) characterized by the clinical triad of recurrent staphylococcal skin abscesses, recurrent pneumonia, and serum IgE levels >10 times the upper norm. HIESs usually manifest in childhood and have a highly variable expressivity.¹, ² Although most cases are sporadic, both autosomal-dominant and autosomal-recessive (AR) inheritance have been described. The predominant form is autosomal-dominant HIES caused by heterozygous, dominant-negative mutations in signal transducer and activator of transcription 3 (STAT3).³ In this form of HIES, extraimmune manifestations occur, including skeletal abnormalities such as retained primary teeth and a typical facial appearance. A scoring system for these findings has been developed at the National Institutes of Health (NIH)⁴ and refined to a STAT3 score by Woellner et al.⁵ In contrast, AR-HIES is characterized by recurrent viral and bacterial infections, extreme eosinophilia, and elevated IgE without skeletal or dental abnormalities.⁶ The genetic causes of AR-HIES are largely unknown. Minegishi et al⁷ reported a monogenetic defect in the cytoplasmatic tyrosine kinase Tyk2 in a single patient with clinical features of AR- HIES who had consanguineous parents. In a follow-up study, however, additional patients with AR-HIES did not have mutations in the TYK2 gene.⁸ Recently a second patient with a related phenotype has been found to be deficient in Tyk2 (J.-L. Casanova, personal communication, October, 2009).

Autosomal-recessive HIES has been described predominantly in consanguineous families from Turkey. We investigated, by genome-wide homozygosity mapping and copy number analysis, 16 patients from 14 families with AR-HIES, defined as a positive NIH HIES score, and absence of significant skeletal findings. Eleven additional patients from six families were analyzed after the candidate gene had been identified.

Homozygosity mapping is a method to localize a disease-associated recessive genotype by searching for homozygous haplotypes in consanguineous families; the underlying assumption is that the recessive mutant allele is identical by descent in affected subjects.⁹ If the founder mutation is relatively recent, the causative mutation is likely to be found within the largest stretches of homozygosity. The genotyping required for this approach may be accomplished by using high-density oligonucleotide arrays, which have the additional benefit of providing data on copy number at each single nucleotide polymorphism locus on the array. Copy number variations (CNVs) occur as a result of deletions and insertions of variable size in the genome.¹⁰ CNVs are common and may be polymorphic in the population or arise de novo in individuals.

Our analysis has demonstrated an AR-HIES genomic locus and made possible detection of mutations in the dedicator of cytokinesis (DOCK)–8 as the major defect in AR-HIES.

Back to Article Outline

Methods 

Patients and controls 

The phenotypes of the patients with AR-HIES analyzed are shown in Table I. Whole blood samples were taken from patients, family members, and healthy volunteers with informed consent. We analyzed DNA from 20 families suspected of having AR-HIES on the basis of clinical assessment. Criteria for AR-HIES were defined as elevated IgE, eczema, hypereosinophilia, and significant infections (particularly with molluscum contagiosum and herpes family viruses; Fig 1). In three kindreds, there were five deceased affected siblings, but all parents were unaffected. There were 13 families from Turkey, 2 from Iran, and 1 each from Lebanon, Oman, Mexico, Italy, and Ireland. All affected individuals had an NIH HIES score⁴ > 20. In family AHR019 an elder sibling died following seizures and developing a coma, both characteristic of AR-HIES. Clinical data on families ARH011 and ARH015 were previously published by Renner et al,⁶ family ARH011 has also been reported by Zhang et al,¹¹ family ARH010 is family 18 in the report by Grimbacher et al,⁴ and patients ARH001 to ARH004 and ARH006 to ARH009 were reported by Al Khatib et al.¹²

Table I. Characteristics of patients with AR-HIES with and without detected DOCK8 mutations

CMV, Cytomegalovirus; CNS, central nervous system; HSV, herpes simplex virus; HPV, human papilloma virus; NA, not available; PML, progressive multifocal leukoencephalopathy.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 1. 

  Clinical findings in patients with DOCK8 mutations. A-C, Severe molluscum contagiosum burden of patient ARH014. H, Molluscum infection of patient ARH012. D-F, Severe dermatitis in patients ARH010.8 and ARH010.9. G, Severe oral papilloma virus infection of patient ARH009. I (MRI) and J (diffusion scan) document the cause of death in patient ARH003, who developed an undefined form of encephalitis.

Methods 

A detailed description of the methods can be found in this article's Methods section in the Online Repository at www.jacionline.org, including homozygosity mapping and copy number analysis, PCR and sequence analysis, immunoblotting, and proliferation and carboxyfluorescein succinimidyl ester (CFSE) dilution studies.

Back to Article Outline

Results 

Search for copy number variations, microdeletions, and homozygosity 

Representational oligonucleotide microarray analysis for CNVs was performed on 8 index patients with AR-HIES previously identified among a cohort of Turkish patients¹² and a singleton patient from Italy with HIES (Fig 2, A). In the subtelomeric region of chromosome 9p, homozygous deletions (copy number = 0) were identified in 4 patients and a large compound heterozygous deletion was identified in the Italian patient with AR-HIES. In 1 additional patient hemizygosity was identified at this locus. For most patients the deletion extended from the most terminal p locus to within the DOCK8 gene (Fig 2, B). Homozygous deletions were confirmed by PCR of the affected segments of DOCK8 (see this article's Fig E1 in the Online Repository at www.jacionline.org). The parental origin of the deletion was investigated and is shown in this article's Fig E2 in the Online Repository at www.jacionline.org.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 2. 

  A, Representational oligonucleotide microarray analysis data demonstrating copy number abnormalities consistent with subtelomeric deletions of 9p in AR-HIES. Individuals ARH001 to ARH004 have homozygous deletions, ARH005 has a compound heterozygous deletion, and ARH006 has a heterozygous deletion. The remaining subjects do not have demonstrable deletions. Genome-wide single nucleotide polymorphism Nsp 250k arrays were used for subjects ARH001 to ARH009. B, Deletions and homozygous intervals and known and predicted genes at the terminus of chromosome 9p for patients with HIES. C9orf66 is an open reading frame, and FAM138C is a noncoding RNA gene. FOXD4 is a transcription factor, and CBWD has a cobalamin binding domain and nuclease function. DOCK8 is described in the text.

To obtain additional evidence that the terminal region of chromosome 9p is associated with AR-HIES, homozygosity mapping was performed on the four patients that did not have biallelic deletions and on further seven affected subjects. All eleven subjects were identified to have homozygous regions on distal chromosome 9p (data not shown).

Positional identification of candidate genes 

Taken together, all 16 affected probands had either deletions or extended homozygous haplotypes at 9p24.3, a region encoding at least 4 genes, the largest of which is DOCK8. The other known genes in this interval include forkhead box D4 (FoxD4), COBW domain containing 1 (CBWD1), and family with sequence similarity 138, member C (FAM138C, a non-coding RNA gene). In addition, there are 4 hypothetical genes and 1 open reading frame present in this region. The majority of the deletions described in our families (ARH001-ARH006) deleted all of these genes as well as disrupting DOCK8. Because individuals ARH003 and ARH005 had only DOCK8 involved in biallelic deletions, we focused on DOCK8 for further analysis.

Mutation detection in DOCK8 

Analysis of DOCK8 genomic and cDNA sequences in individuals with no biallelic deletions detectable by CNV analysis revealed a number of mutations and smaller deletions in 16 patients from 9 unrelated families of 15 families sequenced.

The lack of exonic PCR products in families ARH014, ARH015, and ARH016 suggests deletions including parts of the DOCK8 gene (Fig 3). The single affected individual in family ARH014 showed normal PCR amplification products for exons 3 to 10, a shorter PCR product for exon 11, and no amplification products for exons 12, 13, 15, 16, 26, and 46 to 48 (Fig 3, A and B). We conclude that a large homozygous genomic deletion begins in exon 11 and extends beyond the DOCK8 gene. Likewise, patients from families ARH015 and ARH016 show a deletion up to and including exon 26 and 25, respectively (Fig 3, C-F), deleting the first half of the gene.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 3. 

  Exonic deletions in DOCK8. Pedigrees (A, C, E). Squares, males; circles, females. Filled symbols, patients; slashes, deceased individuals. The lack of PCR products from patients' DNA compared with control DNA suggests exonic deletions (B, D, F). cntrl, Control.

In family ARH010 (Fig 4, A), we found a homozygous splice donor site mutation (gta→tta) after exon 25 (Fig 4, B) that leads to skipping of exon 25 in the DOCK8 cDNA (Fig 4, C). PCR amplification of exons 22 to 27 from cDNA revealed 2 products, 1 normal for wild-type individual ARH010.5 and 1 faster migrating because of a lack of exon 25 sequence for the patients, whereas both products were obtained for heterozygous family members (Fig 4, C). Sequence analysis confirmed that exon 25 was missing from mRNA transcribed from the mutated allele (Fig 4, C). The mutation perfectly segregated with the phenotype in this family (Fig 4, A), with all affected subjects homozygous, and the mother as well as the healthy sister of the affected girls heterozygous for the mutation. Family members heterozygous or homozygous for the wild-type allele had no signs of HIES. We analyzed this specific splice site mutation in DNA from 148 healthy controls and found no evidence of this sequence variant, indicating that it is not a polymorphism. Exon 25 is made up of 150 base pairs; hence its excision leads to an in-frame deletion of 50 amino acids within the DOCK8 protein that are not part of either of the 2 predicted functional domains (UniProtKB/Swiss-Prot: DOCK8_HUMAN, Q8NF50; Fig 5).

• 
  • View Large Image
  • Download to PowerPoint

• Fig 4. 

  Mutations in DOCK8. Pedigrees (A, D, I, L, O, Q). Squares, males; circles, females. Filled symbols, patients; slashes, deceased individuals. Point mutations in the splice site (B, P) or within exons (E, J). Stop codon caused by point mutation (K) or frameshift (H, M). Generation of a cryptic splice site leading to a 4-bp deletion (F, G). Exon skipping shown by cDNA sequencing (C, M, S) and PCR (C, R). Lack of PCR products from genomic DNA suggesting exonic deletions (N, R).

• 
  • View Large Image
  • Download to PowerPoint

• Fig 5. 

  Cartoon showing the predicted impact of the mutations on DOCK8 protein expression. In 3 families, the mutation results in a truncated protein affecting both DHR domains, whereas in 2 families, the truncated protein lacks the DHR-1 domain. In family ARH009, 53 amino acids are missing within the DHR-2 domain, and in family ARH0010, 50 amino acids are missing in between the 2 DHR domains. In family ARH013, DOCK8 is truncated at the C-terminus.

Sequence analysis of family ARH011 revealed a homozygous point mutation (A→G) at position 133 in exon 12 of DOCK8 that segregated with HIES (Fig 4, D), resulting in substitution of arginine for lysine at position 473 (K473R; Fig 4, E). However, in the cDNA of patient ARH011.4, we found a 4-bp deletion directly after the point mutation, which was not seen on the genomic level (Fig 4, F). To explain this finding, the point mutation must create an upstream cryptic splice donor site, described as CAG/gc site (http://www.life.umd.edu/labs/mount/RNAinfo/consensus.html), which may be used instead of or in preference over the wild-type sequence CAG/gt at the 3′-exon/intron boundary (Fig 4, G) and causes the 4-bp deletion in the cDNA. The deletion creates a frameshift followed by a premature stop codon (Fig 4, H), predicted to result in the expression of a truncated protein lacking the end of the dedicator of cytokinesis homology region (DHR)–1 and the whole DHR-2 domain (Fig 5). The residual full-length protein expression reported in this family by Zhang et al¹¹ suggests that the cryptic splice site is used in preference over, rather than instead of, the wild-type splice site. We sequenced exon 12 of DOCK8 in 91 healthy controls and found no evidence that this is a polymorphism.

In the only patient of family ARH012, genomic DNA sequencing revealed a homozygous point mutation (A→T) at position 70 in exon 7 of DOCK8 that segregated with HIES (Fig 4, I), resulting in the premature stop codon TAG at amino acid position 271 (Fig 4, J and K). Thus, the mutation will cause a truncated form of the DOCK8 protein lacking both DHR domains (Fig 5). We sequenced exon 7 of DOCK8 in 115 healthy controls and found no evidence that this is a polymorphism.

In family ARH013, exon 46 of DOCK8 was absent in the cDNA of both patients (Fig 4, M). In contrast with the parents' DNA, it was impossible to amplify exon 46 with specific primers by PCR from the patients' genomic DNA (Fig 4, N). Sequence analysis of the parents' PCR products showed no mutation within exon 46 or the flanking splice sites (data not shown). We therefore conclude that the patients harbor a homozygous genomic deletion of exon 46. The lack of exon 46 with its 107 base pairs will lead to a frameshift and cause a truncated protein (Fig 5).

In patient ARH008.3, a homozygous missense mutation replaced the canonical G in the exon 17 splice acceptor site to a C (IVS16-1 G→C; Fig 4, P). The mutation segregated with HIES (Fig 4, O). The failure to amplify cDNA beyond exon 11 (data not shown) suggested that only proximal shorter messages were expressed. Longer messages would probably be rendered unstable because of exon skipping, leading to out-of-frame sequences and premature termination, which in turn would trigger nonsense mediated degradation. Consistent with this view, the parental cDNA sequences were normal, suggesting that abnormal messages directed by the mutant allele were lost.

Sequencing of DOCK8 cDNA in the proband ARH009 revealed skipping of exon 37 (Fig 4, S). This deletion is predicted to result in an in-frame deletion of 53 amino acids. Although genomic sequences of exon 36 and 38 and their surrounding intronic junctions were normal in the proband, attempts to amplify exon 37 and its associated splice junctions were unsuccessful (Fig 4, R). Normal genomic exon 37 sequences were amplified from both parents, but cDNA analysis revealed 2 species, 1 normal and 1 faster migrating because of a lack of exon 37 sequence. These results are consistent with homozygous deletion of exon 37 in the proband, inherited from parents heterozygous for the same deletion (Fig 4, Q).

Finally, sequencing of genomic DNA of proband ARH006, who had a heterozygous deletion in the DOCK8 locus, failed to reveal exonic mutations, suggesting the presence of a cryptic mutation in the undeleted allele.

DOCK8 deficiency impairs T-cell activation 

To establish the functional consequences of DOCK8 mutations, we first analyzed the expression of DOCK8 protein in PBMCs of probands, their family members, and control subjects. Immunoblot analysis of PBMC lysates of control subjects and family members, carried out by using an anti-human DOCK8 antibody, revealed a major immunoreactive band at about 180 kilo Dalton (kDa) (Fig 6, A, arrowheads) and several less intensely staining smaller bands. In contrast, these bands were missing in 2 patients with documented mutations in DOCK8, including 1 with a splice junction mutation leading to early premature termination (ARH008) and another with an in-frame exon 37 deletion (ARH009). These results demonstrate a deleterious effect of these DOCK8 mutations on protein expression. Interestingly, another proband (ARH007) with no identified DOCK8 mutations also had absent DOCK8 immunoreactive bands, consistent with absent protein expression. This may be indicative of a hitherto undetected deleterious mutation affecting DOCK8 in the nonexonic regions of the gene or a mutation in another gene that regulates DOCK8 expression and/or stability.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 6. 

  DOCK8 deficiency impairs T-cell activation. A, DOCK8 protein expression in probands, family members, and control samples. Lysates from PBMCs, normalized for protein content, were analyzed by immunoblotting with an anti-human DOCK8 antibody. A dominant band of about 180 Kilo Dalton (arrowhead) and several smaller isoforms were detected in control and family members samples but not in those of the probands. As loading control, the blots were re-probed with an anti-human Tubulin antibody. B, Proliferative responses of PBMCs to anti-CD3 mAb treatment (n = 2-5/group; ^∗P = .02). C, DOCK8 deficiency impairs the activation of both CD4⁺ and CD8⁺ T cells. PBMCs were loaded with CFSE and stimulated with anti-CD3 + anti-CD28 mAb for 3 days. Gated populations of mAb-stimulated CD3⁺, CD4⁺, and CD8⁺ T cells were analyzed for CFSE fluorescence intensity (solid lines) and compared with the respective unstimulated cell population (shaded area). CNTL, Control.

We next analyzed T-cell proliferative responses in probands with documented DOCK8 defects compared with those of control subjects. Results revealed a profound defect in anti-CD3 mAb–induced lymphoproliferation in the probands when compared with controls as measured by ³H-thymidine incorporation into proliferating cells (Fig 6, B). Similarly, proband T cells that had been loaded with the chromophore precursor CFSE before stimulation with anti-CD3 + anti-CD28 mAbs failed to dilute their immunofluorescence signal. Such a dilution is a marker of cell proliferation and was readily detected in similarly treated control T cells (Fig 6, C). The proliferative defect affected both the CD4⁺ and CD8⁺ subsets, consistent with a prominent role for DOCK8 in T-cell activation.

Clinical phenotype of DOCK8 deficiency 

All but 2 patients with mutations in DOCK8 had upper respiratory tract infections, and all but 3 had recurrent pneumonia on as many as 20 occasions (Table I). Four patients developed bronchiectasis. Seventeen of 21 (81%) patients had skin abscesses, and 17 of 21 (81%) patients had a severe susceptibility to recurrent and partially mutilating viral infections mainly caused by herpes simplex virus or molluscum contagiosum. Candidiasis was also very prevalent, with 17 of 21 (81%) patients affected. A very severe form of atopic dermatitis, often colonized with Staphylococcus aureus was seen in all 21 patients. Seven of 21 patients had (33%) asthma, and 10 of 21 (48%) patients had multiple food allergies, with 6 patients also having environmental allergies. Cerebral features associated with DOCK8 mutations were central nervous system vasculitis (2 patients), brain infarction (1 patient), meningitis (4 patients), and JC virus-associated progressive multifocal leukoencephalopathy (2 patients). One patient had Burkitt lymphoma, and 1 patient had autoimmune hemolytic anemia.

DOCK8 deficiency was associated with elevated serum IgE (range, 2830-90,910 IU/mL; median, 20,225 IU/mL; normal < 100 IU/mL) and eosinophilia (range, 290-37,880 cells/μL; median, 5675 cells/μL). The remainder of the differential blood count was normal in most of the patients. Total T-cell numbers were within the normal range in 8 of 13 patients (62%) and decreased in 5 of 13 patients (38%). CD4⁺ T cells were decreased in 7 of 13 patients (54%), whereas CD8⁺ T cells were decreased in 3 of 12 patients (25%) and within the normal range in 7 of 12 patients (58%). In contrast with the data reported by Zhang et al,¹¹ B cells were within the normal range in 6 of 13 patients (46%) and increased in 7 of 13 patients (54%). Natural killer (NK) cells were within the normal range in most (11/12; 92%) patients. IgG serum levels were within the normal range in 11 of 13 (85%) of patients with mutations in DOCK8. Eight of 13 patients (62%) had normal IgA. In contrast, 10 of 13 (77%) of patients had decreased levels of IgM.

Back to Article Outline

Discussion 

In 20 families with 27 patients with AR-HIES, a total of 21 patients had biallelic deletions or intragenic small mutations involving DOCK8. Four patients harbored large homozygous deletions, 1 had a large compound heterozygous deletion, 7 had exonic deletions, and 9 had homozygous point mutations predicted to impair DOCK8 protein expression or function.

The fact that we were not able to demonstrate exonic mutations in the remaining 6 AR-HIES families may be a result of difficulties in detecting e.g. compound heterozygous mutations in a 48-exon gene. Moreover, mutations may lie in intronic regions or in the regulatory regions of the gene. We will pursue this by in-depth analysis of patients' cDNA and protein expression.

By genomic sequence analysis, Griggs et al¹³ had mapped DOCK8 to chromosome 9p24. They determined that DOCK8 contains 47 exons spanning 190 kb; more recently, however, the Ensembl database (http://www.ensembl.org) provides evidence for 10 different transcripts and 48 exons. Northern blot analysis detected DOCK8 expression in human placenta, lung, kidney, and pancreas and to a lesser degree in brain, heart, and skeletal muscle.¹⁴

Dedicator of cytokinesis 8 is 1 of 11 members of the DOCK protein family.¹⁵ DOCK8 interacts with GTP-bound and GDP-bound forms of Cell division control protein 42 homolog (Cdc42) and ras-related C3 botulinum toxin substrate 1 (Rac1) and the Cdc42 family members TCL and TC10 in a yeast 2-hybrid assay.¹⁴ The protein contains Dock homology region 1 (also known as -zizimin homology 1) and DHR-2 (CZH2) domains, characteristic of DOCK180-related proteins. In some DOCK180-related family members, these domains bind phospholipid and carry out guanine nucleotide exchange factor (GEF) function, respectively.¹⁵, ¹⁶ Although the GEF activity of DOCK8 remains to be determined, the high DHR-2 domain homology of the DOCK-C subfamily suggests that like orthologs DOCK6 and DOCK7, DOCK8 may be a GEF specific for Cdc42 and/or Rac1¹⁷ (Fig 7).

• 
  • View Large Image
  • Download to PowerPoint

• Fig 7. 

  Hypothetical function of DOCK8. DOCK8 is likely to function as a GEF for the Rho-GTPases Cdc42 and Rac1, turning them into the active, GTP-bound form on receptor engagement (eg, receptor tyrosine kinases, antigen receptors, and adhesion receptors). An unknown protein possibly stabilizes the interaction of DOCK8 with Cdc42 and Rac1. GTPase activation induces dynamic filamentous actin rearrangements and lamellipodia formation, possibly via WASP, leading to cell growth, migration, and adhesion. Given the clinical phenotype of the AR-HIES patients with DOCK8 deficiency, we propose an important role of DOCK8 in T-cell actin dynamics, which might be important for the formation of the immunologic synapse, leading to T-cell activation, proliferation, and differentiation. TCR, T-cell receptor.

Dedicator of cytokinesis family proteins play roles in regulation of cell migration, morphology, adhesion, and growth. Immunofluorescence localized transiently transfected and endogenous DOCK8 to the cytoplasm at cell edges forming lamellipodia, which increased when cells were treated with platelet-derived growth factor (PDGF) or FCS. Transient transfection of a C-terminal fragment of DOCK8 containing the DHR-2 domain resulted in formation of vesicular structures, suggesting that DOCK8 may play a role in the organization of filamentous actin¹⁴ (Fig 7).

Patients with AR-HIES have atopic eczema, severe susceptibility to viral infections, and deficient T_H17 cell function.¹² Our present studies demonstrate that DOCK8 deficiency impairs CD4⁺ and CD8⁺ T-cell proliferative responses, consistent with a critical role for this protein in T-cell activation and effector functions. It can be speculated, based on the protein domain structure and its conservation across species, that DOCK8 may fulfill an important cytoskeletal function relevant to T-cell activation, such as in the formation of the immunologic synapse (see this article's Table E1 in the Online Repository at www.jacionline.org). Other immune deficiencies, such as the Wiskott-Aldrich syndrome and severe combined immunodeficiency caused by lack of Coronin-1A, also involve deficient function of important cytoskeleton-regulating proteins.¹⁸, ¹⁹

The mechanisms by which abnormal T-cell effector functions develop in DOCK8-deficient patients, including impaired T_H17 differentiation, remains to be established. However, a role for DOCK8 in the maintenance of memory T_H17 cells can be inferred, because it was that component of the T_H17 response that appeared most affected in DOCK8-deficient subjects.¹² It is conceivable that T_H17 deficiency in AR-HIES may be symptomatic of a more widespread derangement of T_H-cell differentiation, an eventuality that is currently under investigation. The presence of atopic dermatitis in AR-HIES suggests the action of immune dysregulatory mechanisms, although an additional contribution from a defective barrier function may be involved. Overall, we propose that DOCK8 may play a critical role in cytoskeletal organization, and that its deficiency might result in impaired T-cell activation and effector responses.

Two unrelated patients have been previously described with mental retardation and developmental disability associated with heterozygous disruption of DOCK8 by deletion and by a translocation breakpoint, respectively. Immunologic abnormalities were not described in these patients. Mapping of the critical region shared by the 2 patients showed truncation of the longest isoform of DOCK8. However, these patients had single-copy deletions of DOCK8 along with disruptions of other genes,¹³, ²⁰ which may explain why their phenotype differed from those of our patients.

After the initial submission of our work, Zhang et al¹¹ published homozygous and compound heterozygous mutations in DOCK8 in a cohort of 12 patients from 8 families. One of their families (#8 in Zhang et al¹¹) is family ARH011 in our article. They described the phenotype as a “combined T- B- NK- immunodeficiency” and presented evidence of decreased T, B, and NK–cell numbers. Our assessment of these published cases would suggest that they also have an AR form of HIES, given that the major diagnostic criteria of elevated IgE, eosinophilia, recurrent pneumonia, and skin eczema together with a susceptibility to viral infections were fulfilled in all but 1 patient. However, in contrast with the findings by Zhang et al,¹¹ studies on the cohort presented herein demonstrate a selective decrease in CD4⁺ T-cell numbers in many but not all patients. The CD8⁺ T-cell population is less affected, whereas B and NK cells are usually within the normal range. Furthermore, whereas Zhang et al¹¹ found a selective defect in CD8⁺ T-cell activation, our findings are consistent with a more comprehensive T-cell activation defect involving both the CD4⁺ and CD8⁺ T-cell subsets. Such a global defect in T-cell activation may explain the severe clinical phenotype of AR-HIES.

Dedicator of cytokinesis 8 deficiency is characterized by recurrent infections of the upper and lower respiratory tract; susceptibility to severe, recurrent, and mutilating viral infections (especially by molluscum contagiosum and herpes viruses); a severe dermatitis that resembles atopic dermatitis and may often be superinfected; elevated IgE levels; and eosinophilia. Other clinical features such as asthma, allergies, central nervous system symptoms, or autoimmune phenomena are variably associated. Whether there is a genotype-phenotype correlation with regard to the size of the deletion or whether other deleted genes next to DOCK8 contribute to the phenotype needs to be determined in future studies. It is remarkable that the typical feature of cyst-forming pneumonia (pneumatoceles), as frequently seen in STAT3-mutated HIES, is not typical for DOCK8 deficiency and can be used as a discriminating feature.

Subtelomeric deletions often arise at sites of interspersed repetitive genomic sequences such as Alu, long interspersed elements, long terminal repeats, and simple tandem repeats.²¹, ²² The region surrounding DOCK8 is particularly rich in these sequences, with 264 such sequences in the region from the p terminus to the end of DOCK8 ( National Center for Biotechnology Information Alu database). In particular, there are Alu-Jb sequences near the breakpoints of our patients as mapped by microarray analysis. Abnormal recombination or transposable element activity could contribute to the high frequency of deletions seen. Careful analysis of the breakpoints of these deletions will be helpful in delineating which functions of DOCK8 are affected by the deletions, potentially shedding light on the role of DOCK8 in immune and integument function.

Back to Article Outline

We thank Cindy Ng and Anupama Rambhatla for technical support, Jennifer Birmelin for collecting healthy control samples, Erik Glocker for sequencing controls, Alejandro A. Schäffer and Michael Gertz for help with homozygosity mapping, and Othmar Engelhardt for critical reading of the manuscript.

Back to Article Outline

Methods 

Homozygosity mapping and copy number analysis 

Genomic DNA samples from affected individuals and parents were amplified and hybridized to Affymetrix 250 k Nsp or 6.0 SNP Mapping Arrays (Affymetrix, San Jose, Calif) according to the manufacturer's standard protocol. Call rates ranged from 88% to 93%. This resulted in 20 affected individuals available for analysis. The arrays were analyzed with the crlmm package v. 1.0 using R v. 2.9.2.^E1 For the purpose of homozygosity mapping, single nucleotide polymorphisms in which there was a low confidence call (<0.99 by crlmm) were excluded from analysis. Replication analysis showed that this reduced the genotyping error rate to <1% (data not shown). After preprocessing, 65,600 SNPs were available for homozygosity mapping, which was performed by using PLINK v.1.06.^E2 Extended stretches of homozygosity were presumed to be identical by descent in consanguineous matings.

Copy number analysis was performed by comparing signal intensities from family members to a single standard reference individual as previously reported.^E3 Briefly, the subject and reference arrays were normalized, and the sums of signal intensities from both alleles were compared after adjusting for effects of PCR efficiency. The log base 2 ratio of the subject to the reference was plotted against physical position on the chromosome, and CNV regions of reduced or expanded copy number were identified by circular binary segmentation.^E4, ^E5 All CNV analysis was performed by using the DNAcopy package v. 1.16 and R v. 2.9.2. Regions of apparent low copy number were confirmed by quantitative PCR of coding exons from genes in the CNV regions.

PCR and sequence analysis 

Genomic DNA was extracted from whole blood by using a Gentra Puregene purification kit (Qiagen, Crawley, United Kingdom). For some patients and family members, total RNA was prepared from blood, and cDNA was synthesized as previously described. Exons and flanking intron/exon boundaries from DOCK8 were amplified from genomic DNA by PCR according to standard protocols with Taq polymerase from PeqLab (Fareham, United Kingdom). Primer sequences are available on request. PCR products were purified by using shrimp alkaline phosphatase (Promega, Madison, Mich) and Exonuclease I (Thermo Scientific, Waltham, Mass); DNA was sequenced with the ABI PRISM BigDye Terminator kit V3.1 (Applied Biosystems, Foster City, Calif), the 3130xl Applied Biosystems Genetic Analyzer, DNA Sequencing Analysis software, version 5.2 (Applied Biosystems), and Sequencher, version 4.8 (Gene Codes Corp, Ann Arbor, Mich).

Immunoblotting 

Cells were lysed in 1% NP-40 Nonidet P-40, 150 mmol/L NaCl, 25 mM/L HEPES, pH 7.4, and cell lysis buffer supplemented with protease inhibitors (Roche, Indianapolis, IN). Whole cell lysates (50 μg/lane) were resolved by SDS-PAGE and transferred to nitrocellulose filters. The latter were blocked with 5% nonfat milk, then immunoblotted with polyclonal rabbit anti-DOCK8 (Sigma-Aldrich, St. Louis, MO). The blots were developed by using horseradish peroxidase–conjugated secondary antibodies and enzyme-linked chemiluminescence, and exposed to film.

Proliferation studies 

PBMCs were resuspended at 1 × 10⁶ cells/mL in RPMI medium supplemented with 10% FCS. The cells were seeded in 96-well plates at 2 × 10⁵ cells/well and were left otherwise untreated or were stimulated with 1 μg/mL anti-CD3 (OKT3, Ortho Biotech). After 3 days, the cell cultures were pulsed with ³H-thymidine at 0.4 μCi and analyzed for radioisotype incorporation by using a β-counter.

CFSE dilution studies 

PBMCs were suspended at 10⁷/mL in PBS and loaded with 5 μmol/L CFSE (Invitrogen) for 15 min in the dark, washed, and resuspended at 2 × 10⁶/mL in RPMI/10% FCS. CFSE-labeled cells were placed in culture and treated either with 50 U/mL recombinant human IL-2 (Peprotech) (control cells) or stimulated with anti-CD3 plus anti-CD28 mAb coated beads (Miltenyi Biotech) for 3 days. The cells were then harvested and analyzed for CFSE fluorescence by flow cytometry. T-cell subsets were concurrently identified using phycoerythrin–anti-CD4 and APC–anti-CD8 antibodies (BD Biosciences, San Jose, California).

Back to Article Outline

Fig E1. 

• 
  • View Large Image
  • Download to PowerPoint

• PCR identification of homozygous deletions affecting DOCK8 in patients with AR-HIES with CNV abnormalities at 9p. A, The deletion in ARH002 ends between E25 and E26, whereas that of ARH003 ends between E32 and E33. B, The deletion in ARH001 ends between E3 and E4. All 3 deletions were found to extend proximally beyond exon 1, but their precise 5′’ edges have not been established. Note that ARH006, who has a heterozygous deletion at 9p, successfully amplifies the respective exons, as would be expected given his possession of an undeleted copy of DOCK8. CNTL, Control.

Back to Article Outline

Fig E2. 

• 
  • View Large Image
  • Download to PowerPoint

• Heritability of copy loss in 4 families with AR-HIES. ARH004 shows typical mendelian inheritance of a deletion. ARH006 demonstrates partial repair of the CNV, presumably by gene conversion.^E6 ARH001 shows a de novo homozygous deletion. Parentage for ARH001 was confirmed by SNP inheritance. No mendelian errors were found in 44,645 genotypes surveyed. ARH007 is a patient without a deletion but with extensive homozygosity at this locus.

Back to Article Outline

Table E1. 

Proteins similar to DOCK8 are found in a variety of euteleosts, suggesting evolutionary conservation in greater than 90% of vertebrate organisms

The table gives the percentage protein or DNA homology between Homo sapiens and the listed model organisms, as well as the number of nucleotide substitutions per site (d) and the ratio of nonsynonymous to synonymous substitutions (dN/dS or dNR/dNC). HomoloGene, http://www.ncbi.nlm.nih.gov/homologene/, accessed 10/15/09.

Back to Article Outline

References 

1. Grimbacher B, Holland SM, Gallin JI, Greenberg F, Hill SC, Malech HL, et al. Hyper-IgE syndrome with recurrent infections—an autosomal dominant multisystem disorder. N Engl J Med. 1999;340:692–702
   • View In Article
   • MEDLINE
   • CrossRef
2. Grimbacher B, Holland SM, Puck JM. Hyper-IgE syndromes. Immunol Rev. 2005;203:244–250
   • View In Article
   • MEDLINE
   • CrossRef
3. Holland SM, DeLeo FR, Elloumi HZ, Hsu AP, Uzel G, Brodsky N, et al. STAT3 mutations in the hyper-IgE syndrome. N Engl J Med. 2007;357:1608–1619
   • View In Article
   • CrossRef
4. Grimbacher B, Schaffer AA, Holland SM, Davis J, Gallin JI, Malech HL, et al. Genetic linkage of hyper-IgE syndrome to chromosome 4. Am J Hum Genet. 1999;65:735–744
   • View In Article
   • MEDLINE
   • CrossRef
5. Woellner C, Gertz EM, Schäffer AA, Lagos M, Perro M, Glocker EO, et al. Mutations in the signal transducer and activator of transcription 3 (STAT3) and diagnostic guidelines for the Hyper-IgE syndrome. J Allergy Clin Immunol. 2009;in press
   • View In Article
6. Renner ED, Puck JM, Holland SM, Schmitt M, Weiss M, Frosch M, et al. Autosomal recessive hyperimmunoglobulin E syndrome: a distinct disease entity. J Pediatr. 2004;144:93–99
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (448 KB)
   • CrossRef
7. Minegishi Y, Saito M, Morio T, Watanabe K, Agematsu K, Tsuchiya S, et al. Human tyrosine kinase 2 deficiency reveals its requisite roles in multiple cytokine signals involved in innate and acquired immunity. Immunity. 2006;25:745–755
   • View In Article
   • MEDLINE
   • CrossRef
8. Woellner C, Schaffer A, Puck JM, Renner ED, Knebel C, Holland SM, et al. The hyper IgE syndrome and mutations in TYK2. Immunity. 2007;26:535
   • View In Article
   • MEDLINE
   • CrossRef
9. Kristiansson K, Naukkarinen J, Peltonen L. Isolated populations and complex disease gene identification. Genome Biol. 2008;9:109
   • View In Article
   • CrossRef
10. Iafrate AJ, Feuk L, Rivera MN, Listewnik ML, Donohoe PK, Qi Y, et al. Detection of large-scale variation in the human genome. Nat Genet. 2004;36:949–951
    • View In Article
    • MEDLINE
    • CrossRef
11. Zhang Q, Davis JC, Lamborn IT, Freeman AF, Jing H, Favreau AJ, et al. Combined immunodeficiency associated with DOCK8 mutations. N Engl J Med. 2009 Sep 23;[Epub ahead of print]
    • View In Article
12. Al Khatib S, Keles S, Garcia-Lloret M, Karakoc-Aydiner E, Reisli I, Artac H, et al. Defects along the T(H)17 differentiation pathway underlie genetically distinct forms of the hyper IgE syndrome. J Allergy Clin Immunol. 2009;124:342–348
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (308 KB)
    • CrossRef
13. Griggs BL, Ladd S, Saul RA, DuPont BR, Srivastava AK. Dedicator of cytokinesis 8 is disrupted in two patients with mental retardation and developmental disabilities. Genomics. 2008;91:195–202
    • View In Article
    • CrossRef
14. Ruusala A, Aspenstrom P. Isolation and characterisation of DOCK8, a member of the DOCK180-related regulators of cell morphology. FEBS Lett. 2004;572:159–166
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (2004 KB)
    • CrossRef
15. Cote JF, Vuori K. Identification of an evolutionarily conserved superfamily of DOCK180-related proteins with guanine nucleotide exchange. J Cell Sci. 2002;115:4901–4913
    • View In Article
    • MEDLINE
    • CrossRef
16. Yang J, Zhang Z, Roe SM, Marshall CJ, Barford D. Activation of Rho GTPases by DOCK exchange factors is mediated by a nucleotide sensor. Science. 2009;325:1398–1402
    • View In Article
    • CrossRef
17. Miyamoto Y, Yamauchi J, Sanbe A, Tanoue A. Dock6, a Dock-C subfamily guanine nucleotide exchanger, has the dual specificity for Rac1 and Cdc42 and regulates neurite outgrowth. Exp Cell Res. 2007;313:791–804
    • View In Article
    • MEDLINE
    • CrossRef
18. Bouma G, Burns SO, Thrasher AJ. Wiskott-Aldrich syndrome: immunodeficiency resulting from defective cell migration and impaired immunostimulatory activation. Immunobiology. 2009;214:778–790
    • View In Article
    • CrossRef
19. Shiow LR, Roadcap DW, Paris K, Watson SR, Grigrova IL, Lebet T, et al. The actin regulator coronin 1A is mutant in a thymic egress–deficient mouse strain and in a patient with severe combined immunodeficiency. Nat Immunol. 2008;9:1307–1315
    • View In Article
    • CrossRef
20. MacDermot KD, Hulten M. Female with hypohidrotic ectodermal dysplasia and de novo (X;9) translocation: clinical documentation of the AnLy cell line case. Hum Genet. 1990;84:577–579
    • View In Article
    • MEDLINE
21. Stankiewicz P, Lupski JR. Genome architecture, rearrangements and genomic disorders. Trends Genet. 2002;18:74–82
    • View In Article
    • MEDLINE
    • CrossRef
22. Yatsenko SA, Brundage EK, Roney EK, Cheung SW, Chinault AC, Lupski JR. Molecular mechanisms for subtelomeric rearrangements associated with the 9q34.3 microdeletion syndrome. Hum Mol Genet. 2009;18:1924–1936
    • View In Article
    • CrossRef

Back to Article Outline

References 

1. Carvalho B, Speed TP, Irizarry RA. Exploration, normalization, and genotype calls of high density oligonucleotide SNP array data. Biostatistics. 2007;8:485–499
2. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MAR, Bender D, et al. PLINK: a toolset for whole-genome association and population-based linkage analysis. Am J Hum Genet. 2007;81:559–575
3. Stanczak CM, Chen Z, Nelson SF, Suchard M, McCabe ER, McGhee S. Representational oligonucleotide microarray analysis (ROMA) and comparison of binning and change-point methods of analysis: application to detection of del22q11.2 (DiGeorge) syndrome. Hum Mutat. 2008;29:176–181
4. Olshen AB, Venkatraman ES, Lucito R, Wigler M. Circular binary segmentation for the analysis of array-based DNA copy number data. Biostatistics. 2004;5:557–572
5. Venkatraman ES, Olshen AB. A faster circular binary segmentation algorithm for the analysis of array CGH data. Bioinformatics. 2007;15:657–663
6. Duret L, Galtier N. Biased gene conversion and the evolution of mammalian genomic landscapes. Annu Rev Genom Human Genet. 2009;10:285–311