rss
J Med Genet 39:537-545 doi:10.1136/jmg.39.8.537
  • Review article

Clinical and molecular features of the immunodysregulation, polyendocrinopathy, enteropathy, X linked (IPEX) syndrome

  1. R S Wildin1,
  2. S Smyk-Pearson1,
  3. A H Filipovich2
  1. 1Department of Molecular and Medical Genetics, Oregon Health Sciences University, Mailcode MP350, 3181 SW Sam Jackson Park Road, Portland, OR 97201-3098, USA
  2. 2Division of Hematology/Oncology, Children’s Hospital Medical Center, Cincinnati, OH 45229-3039, USA
  1. Correspondence to:
 Dr R S Wildin, 6140 SW 41st Avenue, Portland, OR 97221-3344, USA;
 wild{at}alum.mit.edu

    Abstract

    Immunodysregulation, polyendocrinopathy, enteropathy, X linked (IPEX, OMIM 304790) is a rare, recessive disorder resulting in aggressive autoimmunity and early death. Mutations in FOXP3 have been identified in 13 of 14 patients tested. Research in the mouse model, scurfy, suggests that autoimmunity may stem from a lack of working regulatory T cells. We review published reports regarding the genetics, clinical features, immunology, pathology, and treatment of IPEX. We also report three new patients who were treated with long term immunosuppression, followed by bone marrow transplantation in two. IPEX can be differentiated from other genetic immune disorders by its genetics, clinical presentation, characteristic pattern of pathology, and, except for high IgE, absence of substantial laboratory evidence of immunodeficiency. While chronic treatment with immunosuppressive drugs may provide temporary benefit for some patients, it does not cause complete remission. Remission has been observed with bone marrow transplantation despite incomplete engraftment, but the long term outcome is uncertain.

    The syndrome of immunodysregulation, polyendocrinopathy, enteropathy, X linked (IPEX) is a rare X linked recessive disorder of immune regulation resulting in the expression of multiple autoimmune disorders. Protean features include early onset type 1 diabetes mellitus (T1DM), severe enteropathy, eczema, anaemia, thrombocytopenia, and hypothyroidism. IPEX is usually lethal in infancy or childhood. Since there are reports of similar patients under many different names, the Human Genome Organization (HUGO) recently established “IPEX” as the official designation for this condition. Mutations in the FOXP3 gene in Xp11.23, encoding a protein called scurfin, were recently identified in several patients, thus establishing a genetic aetiology.

    The clinical and genetic features of the syndrome and the options for management have not been extensively reviewed. The purpose of this paper is to describe three additional patients with IPEX and their associated mutations, and to review published reports regarding the clinical and molecular features of the disease.

    NEW IPEX CASES

    Three cases of IPEX were diagnosed and managed at the Children’s Hospital Medical Center in Cincinnati, Ohio between 1986 and 2001.

    Case 1

    The clinical diagnosis of IPEX was made at 13 years of age after a review of the past medical history and an immunological evaluation. This patient developed T1DM at approximately 3 months of age and was insulin dependent for his entire life. Symptoms of diarrhoea and failure to thrive began during the first year of life and led to a biopsy confirmed diagnosis of coeliac disease at 18 months of age. The patient initially showed improvement with a gluten free diet, but subsequently relapsed and his enteropathy was reclassified as ulcerative colitis, and later as autoimmune when anti-enterocyte antibodies were detected. During the last five years of his life the patient showed no increase in stature or substantial weight gain. He was maintained on intravenous nutrition and partial elemental enteral feeds, prednisone, CSA, and later tacrolimus, the latter treatment showing more substantial benefit. During the year before his bone marrow transplant (BMT), he had experienced intractable hypertension with encephalopathy, probably resulting in a residual left posterior temporal lesion identified by MRI scan of the brain, and life threatening central line sepsis.

    The patient’s maternal grandfather had a life long history of ulcerative colitis and severe autoimmune haemolytic anaemia, both of which contributed to his death at the age of 36 having fathered a healthy daughter and son. The patient was the oldest of six living children. His parents, two sisters, and three brothers are in good health; his mother experienced one miscarriage, gender unknown.

    Believing that the patient suffered from a genetically determined X linked immunoregulatory disorder, allogeneic BMT from his 11 year old HLA matched sister was proposed. Pre-transplant evaluation showed only mild atrophy of the pancreas on abdominal CT scan. The patient underwent pre-transplant conditioning with IV cyclophosphamide, 1200 cGY total body irradiation, and rabbit antithymocyte globulin. The allogeneic marrow was partially depleted of T cells. Post-transplant prophylaxis for graft v host disease (GvHD) consisted of low dose steroids and CSA. Engraftment occurred on day 10 post-BMT and 100% donor chimerism was documented. Initial post-transplant complications were minor, consisting of neutropenic fever and mild haemorrhagic cystitis. The patient experienced significant decrease in stool frequency and volume beginning approximately three weeks after grafting and showed initial weight gain. Hyperalimentation was discontinued before discharge to his home state. Follow up endoscopy showed no significant abnormalities. Three months after transplant, donor engraftment had declined to approximately 50%, where it remained until the time of death. The patient was readmitted for the last time five months post-transplant with recurrent diarrhoea, fevers, and progressive marrow failure eventually leading to a diagnosis of adenovirus infection. He developed Gram negative pneumonia and died quietly on day +194 post-BMT of respiratory insufficiency.

    Sequencing of genomic DNA showed a missense mutation in the winged helix domain of the predicted scurfin protein (table 1).

    Table 1

    FOXP3 sequence variations found in genomic DNA from IPEX patients

    Case 2

    This patient, currently 5 years old, is the only child born after an uncomplicated pregnancy. He weighed 4252 g at birth. He developed otitis media during the first month of life, followed by diarrhoea, and presented in diabetic ketoacidosis at 6 weeks of age. Anti-islet cell antibodies were documented. While in hospital, the patient developed interstitial pneumonitis. Lung biopsy and endoscopic GI biopsies showed CMV. In infancy, the CT scan of the pancreas was within normal limits. After resolution of the CMV infection with antiviral therapy, diarrhoea persisted. GI biopsies showed villous atrophy and chronic inflammation. Steroid therapy was initiated for a presumptive diagnosis of autoimmune enteritis. Episodic diarrhoea has continued throughout the patient’s life, but he has been growing on oral intake alone. The family reported no similarly affected relatives.

    Around 1 year of age lymphadenopathy, hepatosplenomagaly, and eczema appeared. Owing to massive lymphadenopathy and symptoms of obstructive apnoea the patient underwent tonsillectomy, adenoidectomy, lymph node biopsy, and bone marrow biopsy at 20 months of age. These showed Epstein-Barr virus (EBV) infection. The patient subsequently recovered spontaneously and currently maintains protective convalescent anti-EBV viral capsid antigen and anti-Epstein-Barr nuclear antigen titres. Hypothyroidism was diagnosed and treated at 2 years of age. At 2½12 years the patient developed severe, relapsing, autoimmune haemolytic anaemia, which was initially managed with corticosteroids. Owing to complications of severe hypertension and cardiomegaly, monthly rituximab (anti-CD20 antibody) therapy was initiated 15 months ago. This has permitted substantial weaning of the steroids without recrudescence of haemolysis. With reduction in steroid therapy, systemic eczema and massive lymphadenopathy have returned. The patient has experienced one brief episode of immune neutropenia, and has suffered with chronic sinus and ear infections requiring frequent antibiotics, surgery, and, more recently, empirical IV immunoglobulin G (IgG). In the past, he has shown adequate responses to all immunisations.

    DNA sequencing showed no FOXP3 coding region mutations, but a previously reported population variant in intron 91 was observed.

    Case 3

    The clinical diagnosis of IPEX was made at 9 years of age after a review of the clinical history and an immunological evaluation. The patient presented with enteritis and T1DM at 2 months of age. His older brother had died at 15 months of age of infectious complications secondary to immunosuppression following a similar clinical presentation. The patient had life long problems with diabetes and enteropathy requiring chronic immunosuppression with steroids, CSA, and tacrolimus. In addition, he developed debilitating polyarticular arthritis that was treated with rofecoxib, methotrexate, and infliximab, and suffered chronic idiopathic thrombocytopenic purpura during the last two and a half years of his life, treated with steroids, high dose IV IgG, and rituximab. Hepatomegaly and mild hepatitis with focal hepatosteatosis were present on liver biopsy. In addition to chronic failure to thrive, the patient showed evidence of progressive renal insufficiency and hypertension, believed to be secondary to the chronic dependence on CSA and tacrolimus. At 9 years, height and weight were further than 3 standard deviations below the mean for age.

    Because of his deteriorating course, the patient’s parents consented to treatment with a matched unrelated donor transplant. As in case 1, the pre-transplant conditioning consisted of IV cyclophosphamide, total body irradiation, and rabbit antithymocyte globulin. The marrow was partially T cell depleted. The patient showed prompt engraftment and 100% donor peripheral cells initially. The transplant was complicated by haemorrhagic cystitis and a number of infections, including reactivation of cytomegalovirus (CMV), first noted on the day of transplant. Despite these problems, the patient’s stool output normalised for a while and he tolerated full enteral feedings. One month post-transplant, colon biopsies showed only “slight chronic architectural distortion, most likely a residue of the chronic inflammatory bowel disease antecedent to the BMT”. Arthritis resolved despite a decline in peripheral donor chimerism to 70%. Approximately three months post-grafting the patient was readmitted to the intensive care unit with respiratory distress thought to result from infection and fluid overload, and he died on post-BMT day 94. At necropsy, an EBV related lymphoproliferative disorder involving the lung and all other major organs was observed. Chronic sclerosing pancreatitis with absent pancreatic islets, CSA glomerulopathy, and testicular atrophy were also noted.

    His FOXP3 DNA sequence incorporated an in frame, three nucleotide deletion expected to cause the deletion of a single scurfin amino acid residue in a predicted coiled coil domain upstream of the winged helix domain. A previously described silent variation at serine 181 was also detected.

    BACKGROUND

    Incidence of IPEX

    Based on clinical experience and the limited number of published reports, IPEX is probably extremely rare. No estimates of incidence have been proposed. However, new mutations or cases without a family history or with variant presentation are likely to have been under-reported, so the true incidence may be higher than is currently perceived. Males appear to be affected in the expected Mendelian proportions in family reports, but no affected females have been reported. Obligate carrier females are healthy.

    Animal model

    The scurfy mouse (sf) is a natural mutant resembling IPEX. Scurfy mice show X linked recessive inheritance of scaly skin, runting, progressive anaemia, thrombocytopenia, leucocytosis and lymphadenopathy, hypogonadism in males, apparent infection, diarrhoea, gastrointestinal bleeding, cachexia, and death by 3-4 weeks of age.2,3 Several elegant immunology studies have shown that CD4+ T cells mediate scurfy disease.4,5 Scurfy T cells are hyper-responsive to activation via the T cell receptor (TCR), have a reduced requirement for co-stimulation via CD-28, and are resistant to CSA.6 Initiation of illness appears to require antigen stimulation via the TCR.7 High levels of circulating cytokines are present in scurfy mice, especially in the skin.8 In particular, raised interleukin (IL)-2, IL-4, IL-5, IL-6, IL-10, interferon-γ, and tumour necrosis factor-α have been documented.5,8

    The gene mutated in scurfy was recently obtained by positional cloning and named Foxp3. Located near the centromere of the mouse X chromosome,3,9Foxp3 encodes a 429 residue protein, scurfin.10 This novel protein has at the carboxy-terminus a conserved domain belonging to the forkhead class of winged helix transcription factors and a zinc finger motif in the middle third of the sequence. These features suggest it has DNA binding properties and may function in the regulation of gene transcription. The scurfy mutation is a 2 bp insertion that results in the loss of the carboxy-terminal half of the protein including the winged helix domain.

    The functions of normal scurfin and the molecular mechanisms by which mutated versions cause disease are not wholly known. Recent molecular analysis of human scurfin function indicates that it binds DNA elements in the IL-2 promoter and granulocyte macrophage colony stimulating factor enhancer near nuclear factor of activated T cells (NFAT) sites and is capable of repressing transcription of these genes and reducing IL-2 expression in an activated CD4+ T cell line. These functions, as well as nuclear localisation, do not occur in the absence of the forkhead domain.11 However, scurfy mice can be rescued with a single injection of normal T cells (Wildin et al, submitted), suggesting that Foxp3 deficiency results primarily in the lack of functional regulatory T cells, rather than in the failure to repress inflammatory cytokine expression.

    GENETICS OF IPEX

    Genetic mapping in a few large families isolated the human IPEX locus to Xp11.3-q13.3.12–14 The clinical similarities between IPEX and the scurfy mouse and the synteny of their respective map locations recently led us and others to seek mutations in the human orthologue (JM2, FOXP3) of the mouse gene in several unrelated IPEX patients.12,15–17 Coding region mutations consistent with a disruption of function were found in 10 of 11 families (table 1). The eleventh family has recently been shown to have a mutation in the 3` untranslated region of FOXP3 (see below). In addition, two of the patients reported here have mutations predicted to alter the amino acid sequence of the scurfin protein. Thus, more than 90% of IPEX families tested have mutations in FOXP3.

    Table 1 lists the known variations in IPEX probands. None of the disease associated mutations has been seen in a screen of >200 ethnically diverse people.16

    Several of the IPEX mutations occur in the winged helix domain of scurfin in ways that have been predicted to alter a putative DNA binding activity of this motif.16 Others truncate the protein leaving it without a winged helix domain or extend the carboxy-terminus.15 The former may prevent nuclear transport and/or eliminate putative transcriptional repression,11 while the latter probably interferes with folding or positioning of the winged helix. Thus, loss or functional alteration of the winged helix domain can cause IPEX in the recessive, hemizygous state. Other mutations, including those in cases 1 and 3, suggest that other scurfin domains are also essential for preventing autoimmunity.

    Two patients with clinical disease consistent with IPEX have not yielded mutations in the coding region of FOXP315,16 (case 2). The absence of a FOXP3 coding mutation despite good mapping data to this region in the large family reported by Powell et al18 suggests that regulatory or conditional mutations may occur outside the FOXP3 coding regions.16 Indeed, Bennett et al19 recently identified a mutation in the first canonical polyadenylation signal following the final coding exon of FOXP3 in this family. This change is associated with a reduction in FOXP3 mRNA expression, presumably because of non-specific degradation of the resulting aberrant RNA species. A non-coding mutation in our case 2 has not been excluded.

    Females who are obligate carriers for IPEX associated mutations lack the skewed pattern of X inactivation that is seen in the peripheral lymphocytes of some X linked immunodeficiency syndromes (R S Wildin, unpublished data).

    CLINICAL REVIEW

    Clinical features

    Table 2 summarises reported cases with clinical and/or molecular diagnosis of IPEX and patients the authors believe resemble IPEX. In general, cases were selected when the combination of both early T1DM and severe enteropathy existed in a male or among males in a family where X linked recessive inheritance was possible. Many of these cases have not previously been linked or associated as a single disorder. At least seven additional families meeting these criteria have been summarised in two series of protracted diarrhoea of infancy and are not reproduced here.20,21

    Table 2

    Summary of IPEX or IPEX-like cases

    The initial presenting signs, where stated, are listed in table 3. Onset of signs and symptoms usually occurred in the perinatal or infancy period. Later onset was occasionally described, particularly in the family reported by Powell et al,18 where one member first presented in adulthood. The family history of our case 1 suggests that the maternal grandfather may also have been affected. Thus, a diagnosis of IPEX should not be ruled out on the basis of age.

    Table 3

    Initial presenting features

    The most common clinical features were T1DM and enteropathy manifesting as secretory diarrhoea or ileus. In diabetic patients, euglycaemia was often elusive despite careful insulin therapy. T1DM is probably the result of inflammatory destruction of the islet cells rather than islet cell agenesis, as previously suggested.22,23 Diarrhoea was sometimes present before feeding, but generally appeared or worsened after the initial feeding and often failed to abate with enteric rest. Institution of a gluten free diet rarely altered the course of enteropathy, although two patients responded well,24,25 and two gained mild or transient benefit18 (case 1).

    Other clinical features appeared less frequently or were omitted from clinical descriptions. These included eczema (or exfoliative or atopic dermatitis), thrombocytopenia, Coombs positive anaemia, and lymphadenopathy. Several patients were hypothyroid. The initial presentation sometimes included metabolic acidosis, leading to evaluation for inherited metabolic disorders. Polyarticular arthritis, asthma, and ulcerative colitis occurred in a non-diabetic survivor in the family of Powell et al,18 glomerulonephropathy was diagnosed in three cases,18,26,27 and interstitial nephritis independent of medication in three.27,28 Arthritis also occurred in our case 3 and ulcerative colitis in case 1. Hypotonia and muscle atrophy was only noted in a single case.22

    Sepsis, especially catheter related sepsis, and other serious infections including peritonitis, pneumonitis, and septic arthritis, complicated treatment in several cases14,18,22,27–32 (case 1). The most common organisms were Enterococcus and Staphylococcus species. Sepsis occasionally occurs in the absence of immunosuppressive drugs.22,28 These increased infections may relate to a primary defect in immune regulation, as postulated by Powell et al,18 or arise from autoimmune neutropenia or the use of immunosuppressive drugs. In addition, severe enteropathy and eczema could facilitate bacterial entry via the gut and skin, and malnutrition may decrease the effectiveness of the immune system and contribute to severe infection.

    Growth retardation, which may begin prenatally, and cachexia are striking features of the syndrome. This may be the result of T1DM and enteropathy, but given its persistence despite adequate parenteral nutrition and exogenous insulin, it more likely represents inanition related to chronic illness and cytokine excess.

    The causes of death for reported cases are listed in table 4. Haemorrhage, sepsis, intractable diarrhoea, and diabetic complications are the most frequent. Death from reactions related to acute stimulation of the immune system were present in the family reported by Powell et al,18 which exhibits an attenuated phenotype and a non-coding mutation. Of all reported patients, only six are clearly stated to have survived14,18,24,33–35 (case 2) and more than half were treated long term with immunosuppressants.

    Table 4

    Causes of death (where stated)

    Laboratory findings

    Laboratory abnormalities consistent with T1DM, severe enteropathy, hypothyroidism, and cytopenias are common. Apart from these, clinical laboratory testing has shown few consistent diagnostic findings and shed only a little light on the underlying pathophysiology.

    Karyotypes were normal in six patients.22,24,30,32,33,36 Urinary acids were normal in one,24 but raised glycine and serine were noted in another,29 and non-selective amino aciduria was found in still another.30 Hypocalcaemia with normal or raised parathyroid hormone was present in two cases,18,22 suggesting some parathyroid hormone resistance.

    Serum cortisol, somatostatin, gastrin releasing peptide, calcitonin, substance P, neurotensin, adrenalin, noradrenaline, cholecystokinin, pepsinogen I, and urine 5-hydroxyindoleacetic acid were normal in one or two patients.32 Mild increases of gastrin, vasoactive intestinal peptide, pancreatic peptide, and urinary vanillylmandelic and homovanillic acids were variably found.32 Biochemical evidence of hepatic parenchymal disease and cholestasis was common14,22,24,29,32,35,37,38 (case 3).

    T lymphocyte subsets and CD4+/CD8+ ratios were normal in most,14,24,29,32,37,38 although slight increases in CD4+ cells were reported by Satake et al.33 T cell stimulations with phytohaemagglutinin, streptokinase-dornase, poke weed mitogen, concanavalin-A, OKT3, and Candida were slightly low or normal,18,29,33,38 although Shigeoka et al39 reported increased proliferation with phytohaemagglutinin, along with increased IL-2 production from CD-4 positive and total T cells in a member of the family reported by Powell et al.18 This patient showed mildly increased T cell apoptosis, increased numbers of CD4+ T cells at birth, and increases in T cell activation markers associated with episodes of clinical deterioration (Bakke and Wildin, submitted). In one family, one boy had markedly diminished T cell responses, while another had hypogammaglobulinaemia and no response to immunisation with diphtheria and tetanus toxoid.14 A third affected boy had normal antibody responses to protein immunogens, but defective response to Pneumovax™. Post-immunisation titres of antibodies to tetanus toxoid and poliovirus were normal in one boy.38 Lymphocyte mitogen stimulation resulted in exaggerated expression of IL-4, IL-5, IL-10, and IL-13, and reduced expression of interferon-γ.38

    CD16 positive natural killer cells were high in one case,29 and CD20 positive cells and NKH-1 positive cells were raised in another.33 Class II-DR expressing T cells were raised.33,39 Neutrophil chemotaxis, nitroblue tetrazolium reduction, and myeloperoxidase activity were also normal.14,18,29 Complement was slightly low or normal.14,29,30,37 IgG, IgM, and IgA levels were generally normal, but IgE was often raised, sometimes dramatically.12,14,24,27–30,33,36–38 Persistent or periodic eosinophilia was frequent.14,18,24,29,30,36 Allergo-absorbent and skin prick tests for immediate hypersensitivity were consistent with heightened allergic response.12,28 Neutropenia was present in at least two cases40 (case 2). A detailed summary of immune function testing in the three patients described here is being prepared for publication elsewhere. No unusual findings in lymphocyte subsets, mitogen responses, or immunoglobulin levels were identified.

    Autoantibodies have been identified in many patients but are absent in others.26,27,32,37 Positive reports include antibodies against pancreatic islets14,29,30,32,35,37,38,41 (case 2), insulin,36,37 glutamic acid decarboxylase (GAD),36,38 thyroid (anti-microsomal and anti-thyroglobulin),14,18,24,33,40,41 smooth muscle,18,33,35,37 blood group B,18 intestinal epithelium (enterocytes24,37 rat and human small intestine and colon,35 human jejunum,33,40 human duodenum, jejunum, and colon,41 and rabbit colon and small intestine and human rectum27 (case 1), and reticulin and/or gliadin14,35). Kobayashi et al28,42 found circulating autoantibodies to a novel 75 kDa gut and kidney specific antigen, AIE-75, in the family first reported by Satake et al,33 but AIE-75 specific antibodies were not detected by Baud et al.38 Thyroid function remained normal in some patients despite the presence of antithyroid antibodies. In one case, anti-insulin antibodies developed only after exogenous insulin therapy.37 Antinuclear antibody has been either absent or present at low titre.14,18,27,37,38 Late appearance of antibodies to proximal renal tubules, keratinocytes, pancreatic ascini and canals, and adrenals has also been reported.37 Cilio et al36 reported a diabetes susceptibility HLA allele inherited from the father (DQB1*0201, DR3) in a severely affected infant.

    Pathology

    The pathological findings in IPEX are summarised in table 5. By far the most striking and consistent feature is the absence of normal small bowel mucosa and the presence of inflammatory cells in the lamina propria and/or submucosa. Involvement of the large intestine is not uncommon. Inflammatory infiltrates are observed in multiple organs. Foci of inflammation are often seen in the pancreas and islet cells have been reduced or absent in most cases with T1DM. Cholestasis and hepatic fatty change are also common, though the cause in many cases may be chronic parenteral nutrition rather than the underlying autoimmune disorder. Pathology in lymphoid organs is variable, and evidence of bleeding or haemorrhage is not unusual. Histology of the skin shows immune cell infiltration with other changes that are characterised as psoriaform hyperplasia or eczematoid. In one case, IgG and complement component C3 deposition were associated with bullae.14,29 Renal pathology may include tubulointerstitial nephritis, focal tubular atrophy, membranous glomerulopathy, and irregular granular immune deposits in glomeruli and tubular basement membranes.27,28

    Table 5

    Pathology of examined cases14,18,22–25,27–33,35–38,40,41

    Treatment

    Many interventions have been attempted (table 6). Most are ineffective, but two approaches show promise: immunosuppression and BMT. Supportive measures, such as parenteral nutrition and red blood cell and platelet transfusions, are often necessary.

    Table 6

    Attempted treatments

    Immunosuppression

    Chronic immunosuppression has proven partially effective in some patients, but ineffective in others. Regimens that include tacrolimus have shown significant effect, but its prolonged use is limited by its direct toxicity. Furthermore, even those with prolonged survival continue to have significant clinical disease. CSA seemed to benefit some members of the kindred reported by Powell et al,18 where one member still survives on chronic CSA (Buist and Wildin, unpublished data). CSA alone may have prolonged life in a few other cases but generally did not prevent a fatal outcome. Following eventual failure of oral CSA in case 1 of Satake et al,33 Kobayashi et al34,43 reported improvement with oral tacrolimus plus dexamethasone. This patient survives, but suffers from renal tubular disease, osteoporosis, a steroid induced cataract, and growth failure.28,43 Case 1 of Ferguson et al14 had a similar response to tacrolimus, and his skin responded to a topical combination of CSA, prednisone, and dapsone. The patients reported here survived for many years with chronic immunosuppressive medications, including CSA, tacrolimus, methotrexate, corticosteroids, infliximab, and rituximab, but also developed progressive toxicities. These included severe hypertension, renal insufficiency, cardiac hypertrophy, and sepsis related to indwelling catheters. Psychiatric complications including clinical depression and conduct disturbance required medical intervention in the two older patients.

    T cells from scurfy mice are resistant to CSA suppression,10 suggesting that it and other agents inhibiting TCR signalling by the same mechanism may not be good choices for modulating the abnormally high reactivity of the mutant T cells. The limited success of CSA and tacrolimus in humans is consistent with this idea. Experience with newer, more potent immunosuppressive medications is lacking.

    Chronic immunosuppression also increases the chance of severe or opportunistic infections. It is possible that the disordered immune function of IPEX patients contributes to the development of infections in the context of immunosuppressive medications.

    Bone marrow transplantation

    HLA identical sib BMT has been attempted in at least three cases. One boy received HLA identical bone marrow at 4 months of age from his sister, who was not a carrier.38 Remission began during the conditioning phase and continued after the transplant. Enteropathy, T1DM, and eczema all resolved. He remained disease free for 29 months, then developed a lymphoproliferative, haemophagocytic syndrome and died suddenly. Engraftment studies showed host/donor chimerism throughout remission with a donor contribution of 3-30% among peripheral blood cell subsets.

    Two of the affected boys reported here received BMT as a final resort after developing life threatening complications of long term immunosuppression medications. Although both died from infectious complications of their transplants, their enteropathy, arthritis, immune thrombocytopenic purpura, and eczema showed marked improvement. In addition, a decreased insulin requirement was intermittently observed in both cases despite a long history of insulin dependence.

    These mixed results suggest that BMT may, with further experience, become an effective treatment, especially if performed early in the course of clinical manifestations. However, this option should be used cautiously and selectively until long term survival potential with BMT is proven. The discovery of the underlying genetic defect in IPEX permits us to perform molecular diagnosis in newborn boys at risk and to contemplate presymptomatic BMT. It is possible that this approach would improve the long term outcome.

    Other treatments

    Peake et al29 reported some improvement of diarrhoea with fresh frozen plasma and pancreatic enzymes. Seidman et al40 reported marked but transient improvement in enteropathy with human colostrum given at 18 months of age.

    DISCUSSION

    Diagnosis and differential diagnosis

    The diagnosis of IPEX rests on the clinical presentation, the family history, and the elimination of other diagnoses with similar presentation. IPEX displays clinical overlap with a number of genetic disorders. Table 7 proposes features that may help clinicians differentiate them from IPEX. Non-genetic disorders, especially pre- or postnatal viral infections, may present in a similar fashion. These should be considered and appropriately managed until ruled out.

    Table 7

    Syndromes with overlapping features

    Clinical variation among and within families

    Families reported by Powell et al,18 Satake et al,33 and perhaps Ferguson et al14 differ from the majority of reports in that the disease is sometimes compatible with survival, and appears to show some response to immunosuppression. Members of the kindred reported by Powell et al18 had later ages of onset and showed episodic, rather than persistent diarrhoea. Many of the deaths were associated with first immunisation, viral infection, or other exogenous immune stimulating events. Those surviving past infancy manifested arthritis, glomerulonephritis, ulcerative colitis, hypertension, recurrent infections, sarcoidosis, and peripheral nerve sensory neuropathy.18 Similarly, the brothers reported by Satake et al33 lacked overt T1DM, although two had glucose intolerance and the diarrhoea responded to strong immunosuppression in one.33 The current cases also showed a clinical response to chronic immunosuppression and a longer lifespan. It is not yet possible to identify a reliable genotype-phenotype relationship for the observed variations, but the mutation affecting mRNA stability found in the family reported by Powell et al18 might explain the variable, sometimes attenuated phenotype.

    The patient reported by Hattevig et al25 had neonatal T1DM and recurrent diarrhoea; the latter eventually responded to withdrawal of dietary gluten and the patient survived. He is included here because of the combination of neonatal illnesses, the similarity of gut pathology, and a pedigree that is complicated, but consistent with an X linked recessive contribution. One member of the family reported by DiRocco and Marta24 also did well on a gluten free diet. In several other cases, gluten could be excluded as a contributing agent. Of interest, symptomatic and asymptomatic coeliac disease may occur at increased frequency among children with the common form of juvenile T1DM.44

    In contrast, the remaining families and individual cases had early, even prenatal onset, were almost uniformly fatal, and immunosuppression, when used, did not prevent death. It may prove clinically useful to distinguish disease variants. However, it remains to be seen whether the differences are the result of chance, alternate genetic mechanisms, variably severe mutations of FOXP3 or its regulatory regions, or modifying genes such as HLA. Environmental influences and variable management may also influence outcomes.

    In addition to clinical variation among families, presentation and expression varies considerably within families. Enteropathy may be present in the absence of T1DM22,24,29 and vice versa. Indeed, several reports of X linked autoimmune enteropathy (AIE) strongly resemble IPEX with or without T1DM (Goulet’s types Ia and Ib, respectively20), but allelism has not yet been shown. The age of onset and the prevalence of atopic dermatitis, thrombocytopenia, and haemolytic anaemia vary within families as well.

    Immunology and pathophysiology

    The current cases add substantially to the limited knowledge of the immune milieu in IPEX patients. Remarkably few abnormalities are found, and none supports the presence of an underlying deficiency of immunoprotective cells or molecules. Instead, the pathological evidence suggests a hyperactive immune or immune dysregulated state that results in destruction or compromise of various essential organs, as well as secondary effects such as cytokine mediated cachexia. This concept is supported by the available data from the scurfy mouse, where abnormal T cell reactivity appears central to the development of autoimmunity and cytokine excess.

    In this context, the therapeutic efficacy of partial chimerism for normal bone marrow is of particular interest. This observation suggests that the abnormally reactive T cells can be made quiescent by the addition of normal haematopoietic cells. Patel45 alluded to a similar effect in scurfy mice. The phenomenon is presumably also responsible for the absence of disease in female heterozygotes. These females are probably mosaic by Lyonisation for cells expressing only the normal FOXP3 allele and cells expressing only the IPEX associated allele, yet the latter fail to cause disease. One of us has recently shown a therapeutic effect of normal T cells in the scurfy mouse, suggesting that scurfy T cells are susceptible to regulation by regulatory T cells, but are not themselves capable of providing T regulation (R S Wildin, submitted).

    Acknowledgments

    Portions of this work were supported by the following grants: National Institutes of Health R29 DK47278, NIDDK R21-DK60207, American Society of Nephrology R29 Supplement, and The Tartar Foundation (RSW); AHF holds the Ralph J Stolle Chair in Clinical Immunology at the Children’s Hospital Research Foundation, Cincinnati, Ohio. We thank Neil Buist, Berkely Powell, Richard Stenzel, and Anthony Bakke for helpful discussions. We thank the following Children’s Hospital Medical Center of Cincinnati physicians for their involvement in the care of the three patients described here: Drs James Henbi and John Bucuvalus, Pediatric Gastroenterology; Dr Larry Dolan, Pediatric Endocrinology; Dr Murray Passo, Pediatric Rheumatology; Drs Brett Leochelt and Jaqueline Weirmaa, Pediatric Hematology/Oncology; Dr Frederic Strite, Pediatric Nephrology. We also thank referring physicians Dr Richard Hayes, W Virginia, Dr Jerry Barbosa, Florida, and Dr William Gerhardt, Cincinnati, Ohio.

    REFERENCES