Cryptogenic multifocal ulcerating stenosing enteritis associated with homozygous deletion mutations in cytosolic phospholipase A2-α
- Matthew A Brooke1,
- Hilary J Longhurst2,
- Vincent Plagnol3,
- Nicholas S Kirkby4,5,
- Jane A Mitchell5,
- Franz Rüschendorf6,
- Timothy D Warner4,
- David P Kelsell1,
- Thomas T MacDonald1
- 1Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London, UK
- 2Department of Clinical Immunology, Barts Health NHS Trust, London, UK
- 3Department of Genetics, University College London, London, UK
- 4William Harvey Research Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London, UK
- 5National Heart and Lung Institute, Imperial College, London, UK
- 6Max-Delbrück-Center for Molecular Medicine, Berlin, Germany
- Correspondence to Professor Thomas T MacDonald, Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, Newark St, London E1 4AD, UK;
- Received 12 October 2012
- Revised 20 November 2012
- Accepted 21 November 2012
- Published Online First 25 December 2012
Objective Cryptogenic multifocal ulcerating stenosing enteritis (CMUSE) is an extremely rare, but devastating, disease of unknown aetiology. We investigated the genetic basis of this autosomal recessive condition in a pair of affected siblings who have 40-year histories of catastrophic gastrointestinal and extraintestinal disease.
Design Genome-wide single-nucleotide polymorphism homozygosity mapping in the two affected family members combined with whole-exome sequencing of one affected sibling. This was followed by confirmatory Sanger sequencing of the likely disease-causing sequence variant and functional studies in affected and unaffected family members.
Results Insertion/deletion variation analysis revealed the presence of a homozygous 4 bp deletion (g.155574_77delGTAA) in the PLA2G4A gene, located in the splice donor site directly after exon 17 (the penultimate exon) of the gene in both affected siblings. This introduces a frameshift of 10 amino acids before a premature stop codon (p.V707fsX10), which is predicted to result in the loss of 43 amino acids (residues 707–749) at the C-terminus of cytosolic phospholipase A2-α (cPLA2α). cPLA2α protein expression was undetectable in the gut of both siblings, with platelet aggregation and thromboxane A2 production, as functional assays for cPLA2α activity, grossly impaired.
Conclusions We have identified mutations in PLA2G4A as a cause of CMUSE in two affected siblings. Further studies are needed to determine if mutations in this gene are also responsible for disease of a similar phenotype in other cases.
Significance of this study
What is already known about this subject?
Non-steroidal anti-inflammatory drug (NSAID) use leads to upper gastrointestinal ulceration.
Cryptogenic multifocal ulcerating stenosing enteritis is a disease of unknown origin associated with severe intestinal ulceration.
A previous study of a single case with onset of intestinal ulceration in early middle age was associated with compound heterozygote mutation in the gene encoding cytoplasmic phospholipase A2-α.
What are the new findings?
We describe two siblings with severe, early-onset cryptogenic multifocal ulcerating stenosing enteritis.
We found a homozygous 4 bp deletion (g.155574_77delGTAA) in the PLA2G4A gene, located in the splice donor site directly after exon 17 (the penultimate exon) of the gene in both siblings, which introduces a frameshift of 10 amino acids before a premature stop codon (p.V707fsX10).
Functionally, platelet aggregation and thromboxane A2 production were the same as in platelets from normal individuals treated with aspirin.
How might it impact on clinical practice in the foreseeable future?
Patients with intestinal ulceration in the absence of NSAID use should be screened for mutations in cPLA2α.
There have been a number of genome-wide association studies (GWAS) linking common genetic variation with more prevalent gastrointestinal inflammatory conditions such as inflammatory bowel disease (IBD) and coeliac disease. Genetic studies on much rarer, highly penetrant conditions of unknown aetiology, usually presenting with gastrointestinal symptoms early in life have also been highly informative. For example, the disease formerly known as autoimmune enteropathy is now known to be caused by mutations in the FOXP3 gene, which result in a defect in immune regulation.1 ,2 Mutations in the IL-10R gene also result in a very severe IBD in children, which can, importantly, be cured by a bone marrow transplant,3 while polymorphisms in this same gene and its receptor have been associated by GWAS with early-onset ulcerative colitis4 ,5 and Crohn's disease.6 We have also recently identified a pair of siblings, who presented as infants with diarrhoea and skin lesions, as knockouts for ADAM17, the tumour necrosis factor-α- converting enzyme.7
Recurrent small bowel ulceration and stenosis of unknown aetiology (cryptogenic multifocal ulcerous stenosing enteritis (CMUSE)) is an extremely rare condition, and has thus far eluded pathophysiological explanation.8 ,9 The lesions in patients with CMUSE resemble those associated with use of non-steroidal anti-inflammatory drugs (NSAIDS), a well-known cause of gastric10 ,11 and small bowel12 ,13 ulcers, and in extreme cases small bowel strictures and multiple concentric stenoses of the intestinal lumen (so-called ‘diaphragm disease’).14–17 NSAIDs are inhibitors of the cyclo-oxygenase (COX-1 and -2) enzymes, which function in the synthesis of prostaglandins, using free arachidonic acid (AA) as a substrate.18 Reduction in prostaglandin production is considered primarily responsible for NSAID-associated enteritis, because of the role of prostaglandins in the regulation of gastrointestinal blood flow, cytoprotection and intestinal mucus secretion.12 ,19
Here, we describe two siblings of Serbian origin who have a 40-year history, beginning in early childhood, of severe peptic and upper small-intestinal ulceration, extensive small-intestinal stricturing, fibrosis and fistulae, and multiple severe extraintestinal complications. In each patient, we identified a homozygous 4 bp deletion in PLA2G4A, which maps in a large region of shared homozygosity on chromosome 1, which results in the loss of the C-terminal region of its protein product, cytosolic phospholipase A2-α (cPLA2α), and, putatively, the abrogation of its enzymatic function.
The parents of the affected siblings are not known to be consanguineous, although they both come from a small, isolated community in Serbia.
The first affected sibling (male, born 1963) presented at the age of 4 years with severe peptic ulceration and bleeding. This was initially treated medically, but vagotomy and gastroenterostomy were required by the age of 7 years after the patient developed pyloric stenosis and duodenal ulceration. This was then followed by surgery for a gangrenous terminal ileum, secondary to volvulus and multiple adhesions. At the age of 13, the patient required surgery for a perforated gastric ulcer. In 1982, he underwent a partial gastrectomy for a large gastric ulcer, and in 1986 he had a fibrous polyp removed and was noted to have multiple gastric ulcers. In the 1990s, he was found to have a small gastric lumen, generalised gastritis, and Barrett's oesophagus with histological gastric metaplasia, but a normal duodenum. In 1993, he underwent laparotomy for revision of a retrocolic gastrojejunostomy, at which time multiple adhesions were seen. In 1996, he presented with a stenotic duodenum. By 2001, symptoms of dysphagia and abdominal pain caused by partial bowel obstruction were severe, and he was malnourished with low albumin, low vitamin B12 and peripheral oedema. In September 2008, he underwent endoscopic retrograde cholangiopancreatography and was noted to have a thickened, tortuous, dilated common bile duct, suggestive of fibrosis, and ulceration at the biliary sphincter. In 2011, he was admitted to hospital with a gastric and small bowel stricture and peritoneal adhesions and underwent gastroplasty, jejunojejunal anastomosis and adhesiolysis. One month after surgery, he was admitted with worsening malnutrition, hypoalbuminaemia and oedema. He had multiple liver abscesses and an enterocutaneous fistula. He has had iron deficiency anaemia since 1996, with haemoglobin levels as low as 5 g/dl. He was diagnosed with type 2 diabetes in 2010 and has developed severe peripheral neuropathy. He also has osteoporosis. Colonoscopy in 2010 showed a normal large intestine and terminal ileum.
The second affected sibling (female, born 1966) presented at the age of 2 years with peptic ulceration, bleeding and pyloric stenosis, which required pyloroplasty and selective vagotomy. In 1980, she underwent gastrojejunostomy for further peptic ulceration and duodenal stenosis, and in 1996 was diagnosed with pernicious anaemia. In 2000, gastroscopy showed oesophageal and gastric ulcers, a tight ulcerated pyloric canal, and a chronic duodenal ulcer. In the same year, she experienced Salmonella enteriditis infection and underwent surgery for ileal perforation. In addition in 2000, she experienced two episodes of severe Campylobacter enteritis and was diagnosed with xanthogranulomatous pyelonephritis on renal biopsy; she continues to have chronic renal failure. In 2001, she experienced another perforation of the ileum and an ileoileal fistula, complicated by Candida septicaemia, resistant staphylococcal chest infection, acute respiratory distress syndrome and acute renal failure requiring haemofiltration. Later that year, she required drainage of a pelvic collection. In 2009, she required a transverse colon resection for volvulus. In 2010, she was malnourished with hypoalbuminaemia and oedema, and started total parenteral nutrition. Also in 2010, she developed biliary reflux with severe oesophagitis and strictures. She required repeated oesophageal dilatation in 2010–2011, but, despite this, developed bronchiectasis secondary to recurrent aspiration pneumonia. She underwent a revision gastroenterostomy in 2001. Further complications include endometriosis (1990), gall stones (2006), left ventricular concentric hypertrophy with a small cavity, and a fibrotic, unstable bladder with carbapatite stones, requiring ureteric stents (both 2009). Furthermore, she is infertile.
Symptoms in both siblings have progressed despite maximal medical therapy—most recently with high-dose proton pump inhibitors—including eradication of Helicobacter pylori. Short courses of moderate-dose corticosteroids were unhelpful in the male sibling, and have not been evaluated in the female.
The siblings’ mother, father and brother (all heterozygous carriers of the mutation) are healthy and have no family history of similar symptoms, with the single exception of a peptic ulcer in the father's middle age, which resolved with a typical clinical course.
Ethics and consent
Written informed consent was obtained from all family members. This study was approved by the South East NHS Research Ethics Committee and was performed according to the Declaration of Helsinki Principles.
Whole-genome single-nucleotide polymorphism (SNP) array analysis of the two affected siblings was performed using the Human Omni2.5Quad V.1.0 to identify shared regions of homozygosity. This array provides more than 2.4 M SNP with a physical position genome-wide. At first, a linkage analysis was carried out using the program Merlin20 with a reduced marker set of 64 983 SNPs, a hypothetical pedigree assuming consanguinity with a cousin marriage of 2nd degree, allele frequencies from a European population, a genetic map adjusted to Rutgers map v2,21 and a recessive genetic model with complete penetrance (see online supplementary materials). To narrow down recombination events, we used all 2.4 M markers from the array. With a self-written Perl script, we identified all regions where both affected siblings were homozygous on the same allele (see online supplementary materials). In tandem, whole-exome sequencing was performed in the affected female using a SureSelect Human all Exon 50 Mb kit (Agilent Technologies, Santa Clara, California, USA) and sequenced on an Illumina HiSeq. Sequencing reads were aligned to the hg19 build of the human reference genome using the software novoalign (http://www.novocraft.com). SNP and indel calling were performed using samtools V.0.18 and were annotated using the software ANNOVAR.22 Candidate variants were filtered on the basis of function (as predicted by ANNOVAR), and 1000 Genomes (http://www.1000genomes.org) and NLHBI exome sequencing project (http://evs.gs.washington.edu/EVS/) frequencies. Rare variants mapping to the shared regions of homozygosity were investigated further. PCR and Sanger sequencing were performed in the two affected individuals, their unaffected sibling and both parents, to confirm the segregation of the mutation with disease, using an ABI PRISM 3130xl sequencer and the primer pair PLA2-F and PLA2-R, designed specifically against the mutation site. Sequence traces were analysed using Chromas Lite software and sequences were aligned using the Multalin multiple alignment tool (multalin.toulouse.inra.fr/multalin). Details of the cPLA2α protein and active site domain sequences were obtained from the NCBI Protein (Accession Number CAB42689.2) and Conserved Domain ( Accession Number cd07200) databases, respectively. Details of cPLA2α protein sequences for non-human species were also obtained from the NCBI Protein database; these were aligned and compared using the ClustalW2 multiple sequence alignment tool (http://www.ebi.ac.uk/tools/msa/clustalw2) and visualised using the Jalview multiple alignment editor (http://www.jalview.org).
Immunofluorescence and western blotting
Immunofluorescence and western blotting studies were performed using antibodies against cPLA2α (Ab58375; Abcam, Cambridge, UK), using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a loading control (Ab9485; Abcam). For cPLA2α staining, paraffin-embedded small bowel sections from the two affected siblings and unaffected controls were dewaxed, antigen-retrieved using citrate buffer, and blocked with 3% serum in phosphate-buffered saline and avidin/biotin blocking solution (Vector Laboratories, Burlingame, California, USA). The sections were then incubated with rabbit cPLA2α antibody (Abcam) in 1:100 dilutions overnight at 4°C. Sections were then further incubated with anti-rabbit IgG conjugated with fluorescein isothiocyanate for 1 h and mounted using mounting medium containing the nuclear stain 4′6-diamidino-2-phenylindole. Stained tissues were imaged using a Zeiss laser confocal microscopy 710 (Carl Zeiss, Welwyn Garden City, UK), and images were processed using Adobe Photoshop CS (Adobe Systems, San Jose, California, USA).
For western blotting studies, peripheral blood mononuclear cells (PBMCs) were isolated, following the manufacturer's instructions, from whole-blood samples taken from the two affected siblings and their unaffected father using Ficoll Paque premium isolation medium (GE Healthcare Life sciences, Little Chalfont, Buckinghamshire, UK). Isolated PBMCs were lysed, and western blotting was performed on the lysates using antibodies against cPLA2α and GAPDH.
cPLA2α activity studies
Platelet aggregation and release reactions were determined as reported previously.23 ,24 Briefly, blood was collected by venepuncture into trisodium citrate (Vacutainer367 691; BD Diagnostics, UK). Platelet-rich plasma (PRP) was obtained by centrifugation at 175 g for 15 min at 25°C. Platelet-poor plasma was obtained by centrifugation of PRP at 15000 g for 2 min. All experiments were completed within 2 h of blood collection.
For light transmission aggregometry, responses to arachidonic acid (AA) (1 mM; Sigma, UK), collagen (0.3–3 μg/ml; Nycomed, Austria) or adenosine diphosphate (ADP) (5 μM; Chronolog, UK) were measured in a Bio/Data PAP-8E turbidometric aggregometer, with platelet aggregation determined as percentage change in absorbance. At the end of platelet aggregation monitoring, cyclo-oxygenase (COX) activity was halted by the addition of 1 mmol/l diclofenac (Sigma, UK), the samples were centrifuged at 1300 g for 10 min at 4°C, and the supernatants removed and frozen. Plasma thromboxane B2 (TXB2) levels, as a surrogate for thromboxane A2 (TxA2) production, were determined using a selective, competitive enzyme immunoassay (Cayman Chemical, USA).
For 96-well plate aggregometry (Optimul),25 ,26 PRP was added to clear half-area 96-well microplates containing lyophilised platelet agonists: AA (0.03–1 mM), ADP (0.005–40 μM), collagen (0.01–40 μg/ml), epinephrine (0.0004–10 μM; Chronolog, UK), U46619 (aTxA2, receptor agonist; 0.005–40 μM; Cayman Chemical) or ristocetin (0.14–4 mg/ml; Helena Bioscience, UK) (final concentration) or vehicle. Plates were then vigorously mixed (1200 rpm, 37°C; BioShake IQ, Q Instruments, Germany) for 5 min, and absorbance at 595 nm was measured using a standard absorbance microplate reader (Sunrise, Tecan, Switzerland). Platelet aggregation was calculated as percentage change in absorbance.
To measure adenosine triphosphate (ATP) secretion, PRP was activated with collagen (0.1–30 μg/ml) or ADP (5 μM) and vigorous mixing (1200 rpm, 37°C; BioShake IQ) in white 96-well microplates. After 2 min, Chrono-lume reagent (1:5, v/v; Chronolog, UK) was added, and plates were incubated for a further 2 min before the luminescence of each well was read in a multimode plate reader (Mithras LB940, Berthold Technologies, Germany). The luminescence of platelet-poor plasma was subtracted from all readings, and ATP concentrations determined by comparison with the luminescence of wells containing a known quantity of ATP (Chronolog, UK).
In all platelet studies, responses of sibling platelets were compared with those prepared from healthy volunteers, with and without in vitro treatment with aspirin (30 μM; Sigma, UK). ATP and TXB2 release data were normalised to the platelet count in PRP. Statistical analyses were performed in Prism V.5.0 (GraphPad software).
SNP homozygosity mapping and exome sequencing reveals a homozygous deletion in PLA2G4A in affected individuals
To identify the genetic cause of this condition, a combination of whole-genome SNP array analysis and exome sequencing was performed. The linkage analysis with the less dense marker set revealed two long stretches with the maximal logarithm of the odds score of 2.38. On chromosome 1, the markers rs6425457 and rs4607826 border a region of 9.5 Mb, and on chromosome 10 the region length is 5.42 Mb between rs11201179 and rs12265445 (see online supplementary material). Homozygosity mapping with the full marker set identified the same two regions on the autosome as the longest stretches. The region on chromosome 1 contains 6765 homozygous SNPs and spans 8.6 Mb. The first heterozygote markers on each side defining the recombination events are kgp4929831 and kgp15471158 (178 785 646 Mb and 187 437 747 Mb, NCBI build 37). On chromosome 10, a set of 4169 adjacent markers were homozygous, spanning 5.07 Mb with restricting markers kgp7926438 and rs12782553 (86 578 120 Mb and 91 649 717 Mb). Insertion/deletion variation analysis of the exome data from the affected female subject revealed the presence of a homozygous 4 bp deletion (g.155574_77delGTAA) in the PLA2G4A gene, located in the splice donor site directly after exon 17 (the penultimate exon) of the gene. PLA2G4A (Accession Number NG_012203.1) maps within the common region of homozygosity on chromosome 1q25. The presence of this deletion was confirmed by Sanger sequencing, with the deletion found to be in homozygosity in both affected siblings. Both parents and their unaffected brother were heterozygous for the deletion (figure 1A, B). This deletion has not been described before, nor is it present in the dbSNP or 1000 Genome databases. Bioinformatic analysis predicts that this deletion results in the loss of the affected splice donor site, and introduces a frameshift of 10 amino acids before a premature stop codon (p.V707fsX10). This would result in the loss of 43 amino acids (residues 707–749) at the C-terminus of cPLA2α (illustrated on cPLA2α structure models in online supplementary figure S1). The deleted region contains residues highly conserved across species (figure 1C), with many of the deleted residues forming part of the cPLA2α catalytic domain (which encompasses residues 144–73027). Furthermore, the deleted section contains a regulatory site (Ser-727) at which phosphorylation is required for cPLA2α activity.28 This analysis therefore suggested that the observed deletion would be expected to ablate cPLA2α enzymatic function. With the exception of a single report of cPLA2α loss-of-function associated with compound heterozygous point mutations in PLA2G4A—to be discussed below—mutations or common variations in PLA2G4A have not previously been associated with any pathological condition, in the gastrointestinal tract or elsewhere.
cPLA2α protein cannot be detected in affected patients’ gut or PBMCs
Immunohistochemical analysis of small bowel biopsy specimens from the two affected individuals and an unaffected, unrelated normal control individual, using an antibody against cPLA2α, showed significant expression of cPLA2α in the small bowel epithelium of the unaffected individual (figure 2A). However, cPLA2α expression could not be detected in the epithelium of the two affected siblings (figure 2B, C), with the level of staining observed comparable to that seen in negative control specimens. Furthermore, western blotting analysis of PBMCs isolated from the two affected siblings and their unaffected father showed an absence of cPLA2α expression in the two affected individuals, in contrast with their unaffected father (figure 2D).
Platelet aggregation and TxA2 production are significantly abrogated in affected patients
To assess the effect of the cPLA2α mutations on its enzymatic function, aggregation of platelets in the plasma of the affected individuals—a process dependent on the production of the cPLA2α downstream product TxA2—was measured. Platelet aggregation in response to stimulation by collagen was significantly reduced in the two affected individuals compared with controls. Indeed, platelet aggregation in the affected individuals was comparable to that of aspirin-treated platelets from a healthy, unrelated volunteer (figure 3A). However, the addition of 1 mM AA to the affected patients’ PRP resulted in a level of aggregation in the affected individuals equal to that of a healthy volunteer, a response that was not seen in aspirin-treated platelets from healthy volunteers (figure 3B). This illustrates that the TxA2 synthesis pathway downstream of cPLA2α-dependent AA production remains unaffected in these individuals.
Platelet aggregometry further demonstrated significantly reduced platelet aggregation in the affected individuals in response to collagen or adrenaline (table 1 and figure 3), with a less marked reduction in aggregation observed in response to ADP. For collagen and adrenaline stimulation, the results were again comparable to those seen in aspirin-treated platelets from healthy volunteers. A synthetic TxA2 analogue (U46619) almost completely abolished the reduction in platelet aggregation, illustrating that no dysfunction exists in either the thromboxane receptor itself or its downstream signalling pathway. Furthermore, platelet agglutination, stimulated in a TxA2-independent manner by the antibiotic, ristocetin, was virtually identical with control levels in affected individuals. Once again, treatment with AA abolished the reduction in platelet aggregation seen in affected individuals, a response again not observed in aspirin-treated platelets from healthy volunteers.
TxA2 produced by affected individuals’ platelets was also measured, via levels of its inactive metabolite, TxB2 (to which TxA2 is converted with a half-life of ∼30 s). Once more, this revealed significantly reduced TxA2 production by affected individuals’ platelets compared with controls in response to collagen, with the reduction entirely abolished by 1 mM AA (table 2).
As a final measure of platelet function, dense granule secretion was measured by means of ATP release from PRP. Again, the platelets of the two affected individuals showed significantly reduced function compared with controls in response to collagen (table 3), although there was no substantial difference in response to ADP (not shown).
We describe a pair of siblings with a catastrophic 40-year clinical history of intestinal and extraintestinal disease. A combination of genome-wide SNP homozygosity mapping and whole-exome sequencing revealed the siblings to be homozygous for a 4 bp deletion in the PLA2G4A gene, which encodes the enzyme cPLA2α,29 which we predict will eliminate enzymatic activity. Functional studies revealed absent production of the eicosanoid, TxA2, and severely impaired platelet aggregation in response to collagen, which was restored by the addition of exogenous AA—the enzymatic product of cPLA2α.
A single case of ileojejunal ulceration, accompanied by gastrointestinal blood loss, anaemia and impaired eicosanoid biosynthesis, has previously been reported to be associated with compound heterozygous single-base-pair mutations in PLA2G4A.24 Interestingly, this patient only developed disease requiring surgery in the 4th decade of life, although milder symptoms of gastrointestinal ulceration had been ‘lifelong’. In addition, unlike the siblings we describe here, who have severe duodenal, gastric and oesophageal disease, the compound heterozygote had a normal upper bowel. Although in the course of their disease, both affected siblings had ileal disease, this may have been secondary to volvulus in the affected male and Salmonella infection in the affected female. We have not yet carried out sufficient detailed analysis to determine the magnitude of the defect in production of all eicosanoids in our patients, although we would predict—on the basis of the nature of the deletion, the protein expression pattern observed, and the substantial defects in TxA2 production and platelet aggregation—that there should be no cPLA2α activity. However, the previously described compound heterozygote appeared to retain some degree of phospholipase activity, which may result from some residual enzymatic activity of his cPLA2α—perhaps reflecting incomplete abolition of cPLA2α activity by one or both heterozygous point mutations observed in this individual, in contrast with the complete loss of function observed here—or due to the activity of other phospholipases, and which may account for his milder phenotype compared with the siblings we describe here. It would, however, be of extreme interest to analyse other case series of patients with CMUSE,6 to determine if they also have mutations in PLA2G4A and to accumulate sufficient cases to correlate disease type with mutation status.
cPLA2α is a widely expressed, Ca2+-dependent enzyme, which functions to specifically catalyse the release of AA from membrane phospholipids, via the hydrolysis of sn-2 ester bonds.30 Expressed in the vast majority of cell types, cPLA2α activity, and therefore intracellular AA production, is tightly regulated by intracellular Ca2+ concentration and phosphorylation of the enzyme itself by mitogen-activated protein kinases and MNK1-related kinases31 at a number of residues. AA serves as the substrate for a spectrum of enzymes involved in the synthesis of eicosanoids, including COX-1 and -2, lipoxygenases and cytochrome P450 expoxygenase. cPLA2α-mediated AA release is thus a rate-limiting step in eicosanoid production, and cPLA2α activity therefore plays an important role in the wide variety of physiological and pathological processes affected by eicosanoids (reviewed by Ghosh et al32). Mice deficient in cPLA2α show normal development and lifespan, but defects in reproduction,33 parturition,34 renal function29 and the allergic response,30 alongside exaggerated heart and striated muscle growth35 and the presence of numerous ulcerative lesions of the small intestine.36 These effects can be associated with particular eicosanoid pathways. For instance, mice deficient in COX-2 showed impaired inflammatory responses, impaired resolution of inflammation, impaired renal development and impaired female reproduction.37 ,38
As the siblings described here had undetectable cPLA2α protein in their intestine, it is very likely that they have a global defect in their ability to produce AA from membrane phospholipids, and consequently have a systematic impairment in prostaglandin and leukotriene production. It is also highly likely that, in the absence of prostaglandins, the pathophysiology of the disease of these individuals mirrors that seen in individuals with enteropathy associated with long-term NSAID use, in that the oesophageal, gastric and duodenal mucosa becomes very sensitive to injury, perhaps mediated by acid, dietary components or bacteria. Furthermore, these individuals would be expected to be deficient in a number of COX-2 products, such as prostaglandin D239 and 15-deoxy-Δ12 ,14-prostaglandin J240—as well as the lipoxygenase-produced lipoxins41—which have key anti-inflammatory and pro-resolution roles. Prostaglandin D2, for example, has been shown to be specifically upregulated in the long-term remission of ulcerative colitis,42 illustrating a gastrointestinal-protective effect, which would presumably be lacking in these individuals. In this way, mucosal injury can initiate a vicious cycle of ulceration and fibrosis as the lesions persist in the presence of impaired resolution.
It is difficult to determine whether the systemic disease seen in these patients was secondary to the problems observed in their gastrointestinal tract, or was a separate phenomenon. However, the male sibling has clear pathologies of his biliary tree, accompanied by liver abscesses and perhaps liver fibrosis. This damage is most likely mechanical in nature, reflecting the increased susceptibility of the bile duct and related tissues to stress in the absence of eicosanoid-mediated protection—although it may also be associated with the role of cPLA2α in protecting the liver from injury.43 The female sibling, meanwhile, has developed pernicious anaemia—perhaps as a result of intrinsic factor deficiency due to gastric damage, bladder problems and endometriosis, the pathogenesis of which is not known.
The female sibling in particular has also suffered repeatedly from infectious diseases, including infection by Candida albicans, Campylobacter and Salmonella, as well as repeated, severe Staphylococcus infections. This may be indicative of the role cPLA2α is known to play in immunity—for example, cPLA2α is necessary for efficient neutrophil-mediated bacterial killing,44 which some reports have suggested may be related to a requirement for cPLA2α in the activation of NADPH oxidase, and thus the generation of reactive oxygen species.45 ,46 Conditions related to neutrophil killing disorders in humans (such as chronic granulomatous disease) often present with abscesses of the skin, tissues and organs, suggesting an explanation for the presence of such lesions in this pair of siblings.47 ,48 Furthermore, immunity to intracellular pathogens such as Chlamydia trachomatis requires cPLA2α, since, in its absence, production of type I interferon is deficient.49 In addition, Candida albicans has been shown to rapidly upregulate cPLA2α in macrophages, suggesting that its absence may compromise innate immunity to Candida.50 Furthermore, cPLA2α is also involved in the induction, by proinflammatory cytokines, of intercellular adhesion molecule 1 expression on the surface of endothelial cells, suggesting that cPLA2α-deficient patients may be affected by impaired mobilisation and transmigration of leucocytes from the bloodstream into tissues to deal with infectious agents.51
In conclusion, we have identified mutations in the PLA2G4A gene as the probable cause of CMUSE of over 40 years standing in two siblings. Further studies are needed to fully characterise the biochemical consequences of these mutations and to determine if mutations in the PLA2G4A gene are present in other cases with a similar phenotype.
MAB and HJL are considered to be joint first authors.
Contributors MAB carried out the Sanger sequencing, immunofluorescence and western blotting and cowrote the paper with TTM and DPK. HJL referred the patients to TTM and provided the clinical histories. VP and FR carried out the bioinformatics of exome and SNP array data, respectively. NSK, JAM and TDW carried out all the platelet assays. DPK supervised the genetic and immunohistochemical aspects of the work. TTM cowrote the paper with MAB and DPK, and supported the exome sequencing.
Funding Supported by grants from Barts and the London Charity (to DPK) and the European Commission Seventh Framework Programme ‘Intestinal Proteases: Opportunity for Drug Discovery grant FP7-Health-2007-A’ (to TTM).
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.