Abstract

Methods

Results

Disease Manifestations

We identified three independent families of European ancestry with cold urticaria and variable immune defects (Table 1, and Table S1 in the Supplementary Appendix). The cold urticaria developed at a very young age and was lifelong; the cold-related symptoms in one of these families have been described previously.⁷ The study subjects differed from patients with typical cold urticaria in that they had negative results on skin testing with ice-cube and cold-water immersion. However, they had positive results on skin testing for evaporative cooling (Fig. 1A, and Fig. S1 in the Supplementary Appendix) and generalized exposure to cold air, findings that are consistent with their reports that symptoms were often elicited by cold wind rather than contact with cold objects.

Open in a separate window

Figure 1

Clinical and Immunologic Manifestations in Patients with Phospholipase Cγ₂–Associated Antibody Deficiency and Immune Dysregulation (PLAID)

Panel A shows the results of an evaporative cooling test in one subject. Cold urticaria was provoked with droplets of ethanol (E) or air-blown water (A) but not with droplets of unblown water (W) or covered water (C). In Panel B, blood leukocyte counts are shown for 18 adult subjects with PLAID. The tops and the bottoms of the boxes represent a 2-SD range above and below the mean values in healthy control subjects, and the horizontal lines represent the median value among subjects with PLAID. NK denotes natural killer. Panel C shows the mean frequency of IgA or IgG antibody–secreting cells, measured by means of enzyme-linked immunospot assay (ELISPOT) after 4 days of expansion in culture. ELISPOT data were normalized to an equal number of B cells on the basis of the proportion of B cells in peripheral-blood mononuclear cells at the start of culture. The T bars indicate standard errors for five subjects from each group. In Panel D, Igκ secondary recombination in transitional B cells is shown for 12 healthy donors (HD), 4 subjects with PLAID, and 4 subjects with X-linked agammaglobulinemia (XLA), indicated by the frequency of Jκ region usage, as determined by single-cell Igκ sequence analysis of data from subjects who were pooled according to diagnosis. The increased usage of the terminal Jκ4 and Jκ5 genes, as seen in both PLAID and XLA, indicates impaired termination of secondary recombination.

Table 1

Summary of the Clinical Manifestations of Phospholipase Cγ₂–Associated Antibody Deficiency and Immune Dysregulation in the Subjects

Clinical Manifestation	Frequency no./total no. (%)
Cold urticaria	27/27 (100)
Recurrent sinopulmonary infection	12/27 (44)
Antibody deficiency*	15/20 (75)
Common variable immunodeficiency	3/27 (11)
Symptomatic autoimmune disease†	7/27 (26)
Positive test for antinuclear antibodies‡	13/21 (62)
Symptomatic allergic disease	15/27 (56)

Open in a separate window

^*Antibody deficiency was defined as a serum IgM, IgG, or IgA level that was more than 2 SD below the age-specific mean.

^†Symptomatic autoimmune disease was defined as the presence of any autoimmune disease, including vitiligo, autoimmune thyroiditis, inflammatory arthritis, and undifferentiated connective-tissue disease.

^‡The test was weakly positive (1 to 3 enzyme-linked immunosorbent assay [ELISA] units) in 12 subjects and strongly positive (>3 ELISA units) in 1 subject.

Of the 27 subjects who were tested, 26 had immunologic abnormalities in addition to cold urticaria (Table S1 in the Supplementary Appendix). We detected antibody deficiency in 75% of subjects who were tested, with 56% reporting a history of recurrent infection; 3 subjects had received treatment with intravenous immune globulin for severe recurrent infection and hypogammaglobulinemia. Autoantibodies or autoimmune disease was present in 56% of subjects. Granulomatous disease was observed in 7 subjects from 2 unrelated families; 3 of these subjects were born with cutaneous lesions on the fingers, nose, and ears that disappeared during infancy. The remaining 4 subjects had persistent granulomatous skin disease, and 1 of these subjects had disfiguring granulomata with soft-palate and laryngeal involvement (Fig. S2 in the Supplementary Appendix).

Common laboratory findings in affected subjects included low levels of serum IgM and IgA, depressed levels of circulating CD19+ B cells and IgA+ and IgG+ class-switched memory B cells, and low levels of natural killer cells, whereas IgE levels were elevated in a majority of the subjects (Fig. 1B, and Table S1 and Fig. S3 in the Supplementary Appendix). Neutrophil, monocyte, eosinophil, and basophil counts were normal, as were numbers of naive and memory T cells (Fig. S3 in the Supplementary Appendix).

Of the 21 subjects who were tested for the presence of antinuclear antibodies, 13 (62%) had positive results (Table 1, and Table S1 in the Supplementary Appendix). B cells were characterized by poor in vitro expansion and reduced classswitched–antibody production after stimulation with Staphylococcus aureus Cowan I (SAC) and unmethylated cytosine–guanine dinucleotide (CpG)– containing oligonucleotide (Fig. 1C, and Fig. S4 in the Supplementary Appendix). Secondary recombination in CD19+CD10^highIgM^highCD27– transitional B cells was increased, as shown by the skewing of light-chain gene usage toward the terminal elements, a characteristic that is also observed in Bruton's tyrosine kinase deficiency in patients with X-linked agammaglobulinemia,¹⁷ a disease that is associated with severe impairment in B-cell signaling (Fig. 1D, and Fig. S5 in the Supplementary Appendix). Unlike patients with X-linked agammaglobulinemia, who have few peripheral-blood mature B cells, subjects with PLCG2 mutations have a substantial mature B-cell compartment, which may lead to an increased occurrence of autoimmunity.

Linkage and Mutational Analysis

Linkage analysis based on the cold urticaria phenotype in Family 1 identified a single linkage interval of 7.7 Mb on chromosome 16q21 with a logarithm of odds (LOD) score of 4.2 (Fig. 2, and Fig. S6 in the Supplementary Appendix). Whole-genome sequencing of one affected member of Family 1 did not identify any novel coding or splice-site single-nucleotide variants, insertions, or deletions within this interval. The linkage interval was narrowed by an independent analysis of Family 2 (Fig. S7 in the Supplementary Appendix), which identified a 12.6-Mb haplotype that segregated with disease status and overlapped the linkage interval in Family 1 by 3.5 Mb (Fig. 2). Among the 24 genes contained within the narrowed candidate interval, we identified PLCG2 as the primary candidate gene (Table S2 in the Supplementary Appendix). The encoded protein, PLCγ₂, is a member of the phosphoinositide-specific phospholipase C family. Members of this family propagate a wide range of extracellular signals by producing diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP₃) through hydrolysis of a phospholipid, phosphatidylinositol 4,5-bisphosphate (PIP₂).¹⁸ The IP₃ that is produced is responsible for calcium release from the endoplasmic reticulum, a critical event in cellular activation. There are two members of the PLCγ family, PLCγ₁ and PLCγ₂. Although PLCγ₁ is expressed in many cell types, the Racresponsive PLCγ₂ isoform predominates and appears to have nonredundant functions in B cells¹⁹ and natural killer cells,²⁰ both of which had abnormalities in the three families in our study.

Open in a separate window

Figure 2

Genomic Deletions in PLCG2 Encoding Phospholipase Cγ₂ in the Study Subjects

Panel A shows the pedigrees of three families with PLAID. Solid symbols indicate affected subjects, and open symbols unaffected relatives. Squares indicate male subjects, and circles female subjects. Slashes indicate deceased subjects. The starred member in Family 3 died at the age of 1 year from pneumonia. Panel B shows the candidate interval on the long arm of chromosome 16, defined by the intersection of a linkage interval in Family 1 and candidate interval in Family 2. Three distinct PLCG2 deletions were identified in the three families (red horizontal bars).

Direct sequencing of cDNA from peripheral-blood mononuclear cells from members of Family 1 identified PLCG2 transcripts lacking exon 19 on one allele, a mutation that was present only in affected family members. Amplification of the genomic segment of PLCG2 that is flanked by exons 18 and 20 revealed a heterozygous deletion that cosegregated with cold urticaria (Fig. S8 in the Supplementary Appendix). By directly sequencing PCR products from overlapping segments flanking exon 19, we identified a deletion of 5.9 kb in PLCG2 (Fig. S8 in the Supplementary Appendix).

Using this same approach, we identified additional heterozygous deletions of PLCG2 in Families 2 and 3 (Fig. S8 in the Supplementary Appendix). In Family 2, we identified PLCG2 transcripts that lacked exons 20 through 22, caused by an 8.2-kb deletion. In Family 3, we identified PLCG2 transcripts that also lacked exon 19 but were caused by a 4.8-kb deletion with breakpoints distinct from those found in Family 1. These family-specific deletions were not detected in 200 healthy subjects of Northern European ancestry. A post hoc analysis of the whole-genome sequence from 1 member of Family 1 confirmed the 5.9-kb deletion that was initially discovered on Sanger sequencing (see the Methods section in the Supplementary Appendix). We propose the term PLAID (PLCγ₂-associated antibody deficiency and immune dysregulation) to describe this unique constellation of clinical, genetic, and functional findings.

Functional Studies of Mutant PLCγ₂

Each of the three deletions that were identified in these families involved the C-terminal Src-homology 2 (cSH2) domain of PLCγ₂ (Fig. 3A), which is autoinhibitory and normally prevents constitutive enzymatic function.^21,22 Transfection of COS-7 cells with PLCG2 expression constructs with a deletion of the full cSH2 domain, a deletion of exon 19 (Δ19), and a deletion of exons 20 through 22 (Δ20–22) produced equivalent PLCγ₂ protein expression (Fig. S9 in the Supplementary Appendix) but elevated basal and Rac-activated phospholipase activity, as compared with a nonmutant PLCγ₂ expression construct (Fig. 3A).

Open in a separate window

Figure 3

Effects of PLAID-Associated PLCG2 Deletions, Including Increased Phospholipase Activity, Deceased Cellular Activation at Physiologic Temperatures, and Enhanced Cellular Activation at Subphysiologic Temperatures

Panel A shows enzymatic activity of phospholipase Cγ₂ (PLCγ₂) mutants. COS-7 cells were transfected with DNA encoding either wild-type PLCG2 or PLCG2 with deletions of the C-terminal Src homology 2 (cSH2) domain, exon 19 (Δ19), and exons 20 through 22 (Δ20–22) (at left). PLCγ₂ activity was assayed by quantification of [³H] inositol phosphates generated by the respective enzymes (at right). Basal activity was quantified in COS-7 cells transfected with PLCG2 alone, whereas the Rac-activated condition included cotransfection with a Rac2 V12 expression construct. The I bars indicate 1 SD above and below the mean release of inositol phosphate for each condition. Data are representative of three independent experiments. C2 denotes calcium-binding C2 domain, EF EF hand motif, nSH2 N-terminal SH2 domain, PH pleckstrin homology domain, SH3 SH3 domain, sPH split pleckstrin homology domain, X-box × catalytic domain, and Y-box Y catalytic domain. Panel B shows cytoplasmic calcium content (as measured by FLUO-4 staining) in natural killer (NK) cells before and after surface-receptor cross-linking (arrow indicates time of administration) in two subjects with PLAID and one control subject. Panel C shows degranulation of NK cells after incubation with sensitive target cells, as measured by cell-surface expression of CD107a as a percent of the maximum CD107a expression observed in control NK cells. Panel D shows the effect of temperature reduction, in the absence of surface stimulation, on cytosolic calcium content in B cells sorted from a case subject and a control subject, as measured on confocal microscopy. Data points reflect the average mean fluorescence intensity of 100 individually imaged cells, and I bars indicate standard errors. Data are representative of three independent experiments. One asterisk indicates a P value of 0.01 to 0.05, two asterisks indicate a P value of 0.001 to less than 0.01, and three asterisks indicate a P value of less than 0.001. Panel E shows mast-cell degranulation, as indicated by the surface expression of lysosomal-associated membrane protein 2 (LAMP2) and measured by means of flow cytometry at 37°C (red) and 20°C (blue) in Laboratory of Allergic Diseases 2 (LAD2) human mast cells transfected with wild-type and deleted forms of PLCG2.

Discussion

In three families with a dominantly inherited syndrome of cold urticaria and pleiotropic immune dysregulation, we found that genomic deletions in PLCG2 were responsible for this unique phenotype. Diminished receptor-mediated signaling in PLAID B cells resulted in abnormal activation, class-switch recombination, and receptor editing, which have been associated with antibody deficiency, recurrent infection, and impaired central tolerance. Furthermore, the signaling defect in PLAID was temperature-sensitive, as shown by the enhanced signaling and cellular activation observed at subphysiologic temperatures, even in the absence of surface-receptor ligation in B cells and mast cells expressing mutant PLCγ₂.

The PLAID-associated PLCG2 deletions affected the regulatory region, including the autoinhibitory cSH2 domain, which normally couples the enzymatic activity of PLCγ₂ to upstream pathways.²¹ PLAID-associated deletions resulted in the failure of autoinhibition and constitutive phospholipase activity, as measured by standard enzymatic assays. These data are consistent with the hypothesis that, at rest, cSH2 obstructs the PLCγ₂ catalytic site and that this inhibitory interaction is relieved by phosphorylation of critical tyrosine residues, which facilitates a conformational change in PLCγ₂, thereby exposing the catalytic site.^21-23

Nonetheless, B cells and natural killer cells from subjects with PLAID clearly had decreased PLCγ₂-dependent signaling and function, with responses similar to those observed in Plcg2-deficient mice.¹⁹ This apparently paradoxical loss of downstream function may be the direct result of chronic signaling, just as chronic B-cell–receptor stimulation results in calcium currents of diminished amplitude, alteration of signaling cascades, and ultimately, proliferative anergy.^24-26 Although the specific mechanisms that underlie the reduction in PLCγ₂-mediated signal transduction remain to be elucidated, there are several possible explanations. Increased phospholipase activity may lead to depletion of the substrate PIP₂ proximate to PLCγ₂, resulting in impaired IP₃ production and IP₃-mediated calcium release.²⁷ Alternatively, excessively high concentrations of the products of PLCγ₂ may induce feedback-mediated down-regulation of the distal signaling pathways.²⁸ Both these mechanisms suggest that phospholipase inhibitors²⁹ might be used to treat PLAID. It is also possible that the deleted regions have other functions that are critical for PLCγ₂-mediated signaling, in addition to the phospholipase activity.³⁰

Cold urticaria in patients with PLAID is characterized by mast-cell degranulation after cold stimulation, which is similar to the lesions reported in other types of cold urticaria.^7,9 Previous serum transfer experiments have suggested that causative autoantibodies may underlie some cases of cold urticaria.^9,31 However, in our study, B cells and mast cells expressing the mutant forms of PLCγ₂ had increased cellular activation and effector function at subphysiologic temperatures without receptor stimulation, indicating that the altered function of mutant PLCγ₂ was responsible for cold urticaria in the subjects in a dominant, cell-intrinsic fashion.

There are two gain-of-function Plcg2 mouse models, but they differ from PLAID in that these mice lack constitutively activated PLCγ₂. In addition, the mouse cells are subject to abnormally high calcium flux after surface-receptor ligation, and all the mice have severe inflammatory arthritis.^32,33 Furthermore, the autoimmunity in subjects with PLAID and in one of these mouse models (Plcg2^Ali5 mice) appears to develop through different mechanisms, given that the induction of intracellular signaling by receptor stimulation is reduced in PLAID B cells but increased in Plcg2^Ali5 B cells. Although cold urticaria is not manifested in either of these mouse models, Plcg2^Ali5 mice have inflammatory dermatitis on the coolest body parts (the ears and distal extremities), resembling the distribution of cutaneous granulomatous lesions in several subjects with PLAID in our study.

Common variable immunodeficiency (CVID) is a prototypic antibody-deficiency disorder, which is also associated with autoimmunity. Although CVID was present in only three subjects in our study, there is substantial phenotypic overlap between PLAID and CVID. In addition to antibody deficiency and autoimmunity, common features of the two disorders include granulomatous disease,¹² diminished class-switched memory B cells,¹⁷ impaired B-cell calcium flux,³⁴ and low numbers of natural killer cells.³⁵ Like PLAID, CVID is associated with mutations that affect proteins involved in B-cell activation.^36-40 Given the overlap of pathways and phenotypes between the two diseases, it seems likely that an understanding of PLAID will provide mechanistic insights into the pathogenesis of CVID and CVID-associated symptoms.

The deletions that we discovered in the three families occur in a region of PLCG2 that is rich in repetitive elements known to facilitate aberrant recombination events. Indeed, five of the six deletion breakpoints occurred within repetitive elements, suggesting that sporadic cases of PLAID may be caused by similar de novo or even somatic mutations in PLCG2.

PLAID-associated PLCG2 mutations and the resulting phenotypes provide a rare example of a monogenic disease defined by an atopic phenotype. The effects of the mutant forms of PLCγ₂ in this complex disorder illuminate vital elements of phospholipase-mediated signaling, including its role in leukocyte function, host defense, and self-tolerance. In addition, such effects illuminate how novel genetic variants can cause atypical temperature responses by altering the balance among complex biologic pathways.

Supplementary Material

Acknowledgments

Appendix

The authors' affiliations are as follows: the Inflammatory Disease Section, National Human Genome Research Institute (M.J.O., E.F.R., I.A., D.L.K.); the Laboratories of Allergic Diseases (G.S., S.D., H.K., G.C., M.-Y.J., A.M.G., D.D.M., C.N., M.O., L.W., K.S., J.D.M.), Clinical Infectious Diseases (A.F.F., S.M.H.), Systems Biology (P.T.-P., N.S.), Immunogenetics (H.S.K., E.O.L.), and Immunoregulation (J.H., S.M.) and the Vaccine Research Center (D.C.D.), National Institute of Allergy and Infectious Diseases; and the Critical Care Medicine Department, Clinical Center (P.T.-P.) — all at the National Institutes of Health, Bethesda, MD; the Department of Structural and Molecular Biology, Division of Biosciences, University College London, London (T.D.B., R.W.B., M.S.M., M.K.); the Department of Immunobiology, Yale University School of Medicine, New Haven, CT (N.R., E.M.); the Department of Medicine, Graduate School, Medical Research Center, University of Ulsan, Ulsan, South Korea (H.S.K.); the Division of Allergy, Immunology, and Rheumatology, Department of Pediatrics and Medicine, University of California San Diego (C.G., H.M.H.), and Rady Children's Hospital of San Diego and Ludwig Institute of Cancer Research (H.M.H.) — all in San Diego; the University of Colorado Health Sciences Center, Aurora (A.A.W.); the Department of Human Genetics, University of California Los Angeles, Los Angeles (H.L., S.F.N.); and the Center for Human Genome Variation, Duke University School of Medicine, Durham, NC (K.V.S., E.T.C., D.B.G.).

Footnotes

References