KS Denianke 1, IJ Frieden 2,3, MJ Cowan 3, ML Williams 2,3 and TH McCalmont 4

Transcription

1 (2001) 28, Nature Publishing Group All rights reserved /01 $ Severe combined immunodeficiency Cutaneous manifestations of maternal engraftment in patients with severe combined immunodeficiency: a clinicopathologic study KS Denianke 1, IJ Frieden 2,3, MJ Cowan 3, ML Williams 2,3 and TH McCalmont 4 1 School of Medicine, 2 Department of Dermatology, 3 Department of Pediatrics, and 4 Dermatopathology Section, Departments of Pathology and Dermatology, University of California, San Francisco, CA, USA Summary: SCID is a heterogeneous group of disorders characterized by defective T cell and B cell function. Eczematous and morbilliform eruptions are common, and graft-versus-host disease (GVHD) due to maternal engraftment has been documented. We sought to better characterize SCID-related cutaneous disease observed prior to BMT and to compare the eruption to conventional GVHD. Medical records of 51 patients with SCID treated between 1982 and 1999 were reviewed. Ten of 51 (20%) had rash and evidence of maternal engraftment prior to BMT (study group). Eleven of 51 (22%) had no rash or evidence of engraftment pre-bmt but developed GVHD following transplant (control group). Skin biopsies were available for review for 8/10 of the study group and for 8/11 of the control group. Cutaneous findings consisted of a scaling, erythematous maculopapular eruption spread widely over the trunk and extremities, with near-erythroderma in some patients. Microscopically, biopsies from the study group differed significantly from controls. Key differences included parakeratosis (P 0.01), psoriasiform hyperplasia (P 0.04) and spongiosis (P 0.04). The dermatopathologic findings of transplacental GVHD differ from the pattern of post-transplant GVHD. A psoriasiform-lichenoid-spongiotic pattern with necrotic keratinocytes should trigger consideration of SCID and maternal engraftment in the dermatopathologic evaluation of eruptions of infancy. (2001) 28, Keywords: severe combined immunodeficiency disease (SCID); bone marrow transplant (BMT); graft-versus-host disease (GVHD); maternal engraftment Severe combined immunodeficiency diseases (SCIDs) are a heterogeneous group of genetic disorders characterized by deficiency of both T and B cell functions. The mode of Correspondence: Dr IJ Frieden, Departments of Dermatology and Pediatrics, University of California, San Francisco, 1701 Divisadero Street #340, San Francisco, CA 94115, USA Received 22 January 2001; accepted 18 May 2001 inheritance can be autosomal recessive or X-linked. These disorders are rare, with a frequency ranging from one in to one in births. 1 Clinical signs and symptoms commonly manifest by 6 months of age and include failure to thrive, oral thrush, candidal diaper rash, recurrent respiratory infections, cutaneous graft-versus-host disease (GVHD) and oral and genital ulcerations. The nonspecific clinical findings and the significant degree of variability among patients often complicate and delay the diagnosis. The profound lack in immune function leads to infections usually fatal in infancy unless the immune system can be reconstituted. 1 Currently, hematopoietic stem cell transplantation from a variety of sources including bone marrow, peripheral or umbilical cord blood is the primary modality of treatment of SCID, 2 although the early results of gene therapy are encouraging. 3 The major cutaneous manifestations of SCID are skin and mucosal infections, eczematous rashes and GVHD. Patients with SCID are frequently engrafted with transplacentally-derived maternal T lymphocytes. Although maternal T cells in SCID rarely induce lethal, systemic, multiorgan GVHD, cutaneous manifestations of GVHD can be striking. Maternal engraftment may also be asymptomatic. 4 It has been postulated that maternal T cells can prevent engraftment of bone marrow transplants from nonmaternal donors, in particular when HLA-mismatched (haplocompatible) T cell-depleted transplants are used. Because engraftment is incomplete, in most instances maternal cells fail to protect the host against opportunistic infections. 5 In contrast to the well-characterized histopathologic features of GVHD due to BMT, those due to transplacental maternal engraftment are less well studied. Some reports have identified features indistinguishable from post-transplant GVHD, but the reported microscopic features have not been consistent. 6 8 We evaluated the clinicopathologic characteristics of cutaneous GVHD in a group of SCID patients with evidence of maternal engraftment prior to bone marrow transplant. We also sought to determine if the manifestations of GVHD in these patients were indeed distinct from the well characterized forms of GVHD associated with either bone marrow transplantation, the transplantation of other organs or the transfusion of nonirradiated blood. Patients with SCID but no evidence of

2 228 SCID and maternal engraftment maternal engraftment pre-bmt who developed a rash after BMT were used as a control group. Methods Patients The medical records of 51 patients with a diagnosis of SCID who received a BMT at the University of California, San Francisco, Pediatric Bone Marrow Transplant Division between 1982 and 1999 were reviewed. Ten of these patients were found to have evidence of maternal engraftment and rash prior to BMT (group A). An additional 11 did not have evidence of maternal engraftment or a rash pre-bmt but developed rash post BMT (group B).Two patients were found to have evidence of maternal engraftment pre-bmt without rash and thus were excluded from the study. Four patients did not have evidence of maternal engraftment pre-bmt but had a rash pre-bmt. These patients did not have histopathology and were thus also excluded. Twenty-four others, who had neither engraftment nor rash before or after BMT, were also excluded from the study. Data acquisition Patients of white, hispanic, native American, Middle Eastern and Asian background were represented in this study. After 1989, a diagnosis of maternal engraftment was made based on variable number of tandem repeat (VNTR) analysis, also referred to as microsatellite analysis and/or in situ hybridization. VNTR probes give strong hybridization signals allowing for earlier detection of chimerism as well as detection of small numbers of cells. Prior to 1989, methods for detecting maternal engraftment included testing for blood group antigens, polymorphic blood proteins and karyotyping. However, these tests are often not sufficiently sensitive to detect engraftment Maternal engraftment was diagnosed on clinical grounds, which in addition to skin eruption, included failure to thrive, hepatosplenomegaly and a lack of any other explanation for these findings in the presence of the diagnosis of SCID. In addition, histopathology with features of GVHD was also used as supporting evidence for cases prior to the late 1980s (3/10 in group A). At diagnosis, the patients in group A had virtual absence of T and B cell numbers and function. The majority of patients in group A (70%) were T-B- with the remaining 30% being T B+. Two of the T B+ phenotypes were X- linked by familial inheritance pattern. In group B, 55% were T B and 45% were T B+. Two of the T B+ patients in group B were adenosine deaminase (ADA) deficient. Following medical record review, all available skin pathology was reviewed for both groups. Pathology was available for 8/10 in group A and 8/11 in group B. The slides were read by a dermatopathologist who graded all subjects according to standard histologic criteria. The pathologist was aware that all patients had a diagnosis of SCID and had received BMT. He was also aware of which group of patients had a diagnosis of maternal engraftment prior to BMT but was blinded to the prior skin biopsy diagnosis. Chi-square statistical analysis was performed on the histopathological parameters. Results Representative cutaneous lesions are described in Tables 1 and Table 2 and illustrated in Figure 1. The severity ranged from fine erythematous maculopapular lesions involving the trunk and extremities to total erythroderma with alopecia. In addition to rash, virtually all patients had oral thrush and nonspecific symptoms of fever, cough, diarrhea and failure to thrive. Treatment regimens for cutaneous lesions ranged from topical creams and steroids to systemic immunosuppression. Conditioning protocols ranged from none to combinations of systemic immunosuppressives. Seven of the eight patients in group A had rashes that were unresponsive to multiple therapies but resolved after BMT. One patient had a rash that resolved with topical ketoconazole therapy prior to BMT, but did not have a skin biopsy performed. The histopathological results are shown in Table 3 and illustrated in Figures 2 and 3. Of the eight patients in group A who had skin biopsies, only three had a previous histopathologic diagnosis of GVHD. Two had been diagnosed as subacute spongiotic dermatitis, two had vacuolar interface dermatitis and one had a diagnosis of nonspecific dermatitis. Group B was utilized as a control group and of the eight patients with skin biopsies, one had a prior histopathologic diagnosis of GVHD, one had interstitial dermatitis, one had interface dermatitis, one had vascular ectasia and four had superficial perivascular dermatitis. The parameters that showed statistical significance were psoriasiform hyperplasia (P 0.04), parakeratosis (P 0.01), spongiosis (P 0.04), necrotic keratinocytes (P 0.04) and lichenoid infiltrate (P 0.04). A perifollicular infiltrate was observed in 38% of patients in group A and in none of group B, which is borderline insignificant (P = 0.055). The density of lymphocytes and other inflammatory cells did not show statistical significance between group A and group B. However, the very small number of patients in these two groups may have resulted in a lack of statistical power to detect some of the differences that could exist between them. Discussion Our study indicates that maternal engraftment due to the transplacental transfer of T cells is indeed a common phenomenon in patients with a diagnosis of SCID. Eighty-three percent of those engrafted develop a skin eruption as a manifestation of maternal engraftment that is clinically similar to that seen in classic GVHD. However, the cutaneous histopathology in maternally engrafted patients is distinct from the patients with SCID who developed a rash post BMT (group B), and also distinct from classical cutaneous GVHD after transplant for other diseases. Moreover, the definitive diagnosis of maternal fetal GVHD can be problematic because the clinical and histological features are difficult to distinguish from other con-

3 SCID and maternal engraftment Table 1 Clinical findings in transplacental maternal engraftment patients (group A) 229 Patient Age rash Age BMT Total Total CD3 CD3 % Morphology Conditioning Treatment and course onset (m) (y) lymphocyte (10 9 /l) count (10 9 /l) , Diffuse erythematous scaly None Improved on steroids and i.v. 0.6, rash with oozing and Ancef prior to first BMT. 0.8 crusting beginning on chest Resolved post BMT and back, subsequently spreading to scalp, neck, arms and legs with severe involvement on palm, soles and buttocks Generalized eczematous rash CY120 Responded to steroids and and erythroderma with thick ATG80 Velban prior to BMT but scaling in scalp and around 700 cgy recurred when stopped. mouth and nose Initially resolved with CsA and immunosuppression post BMT. Later continued to have chronic eczematous eruption intermittently Diffusely xerotic skin, scaly None Unresponsive to creams and erythematous plaques in ointments given by mother flexural areas, eyebrows, prior to admission. Resolved scalp and diaper area post BMT with systemic and topical corticosteroids Widespread flat-topped CY200 Resolved post BMT with lichenified, papules over ATG80 topical antifungal and topical trunk and extremities with corticosteroids minimal scale. Significant scaling with moderate erythema on scalp and central face , Infiltrated lichenoid plaques Imuran No response to 5% lactic acid 1.2, on extremities and fine Prednisone and anionic base topical 1.8 atrophic scaly papules on ATG80 treatments. Minimal midtrunk. Complete loss of inflammation and scaliness with hair, including scalp, regrowth of scalp hair, eyebrows, eyelashes eyebrows and eyelashes post BMT Fine erythematous ATG80 Response to CsA and maculopapular rash immunosuppression post BMT involving the face, body, extremities, including soles. Most prominent on cheeks Fine erythematous CY200 Responded to ATG and maculopapular rash over ATG80 steroids post BMT trunk and extremities Erythematous, scaly macular None Resolved after 2 months rash on face with raised Nizoral cream and shampoo; sharply demarcated lesions prior to BMT on neck. Erythematous rash around eyes and mouth 9 0 0, Diffuse, desquamating Imuran Resolved with total body 0.2 erythematous rash consistent Prednisone irradiation and with toxic epidermal ATG80 immunosuppression post BMT necrolysis Generalized scaling BU14 Resolved with erythematous rash CY200 immunosuppression post BMT ATG80 ATG80 = anti-thymocyte globulin 80 mg/kg; ATG90 = anti-thymocyte globulin 90 mg/kg; BU14 = busulphan 14 mg/kg; BU16 = busulphan 16 mg/kg; CY120 = cyclophosphamide 120 mg/kg; CY200 = cyclophosphamide 200 mg/kg; NA = not available. Normal ranges: Total lymphocyte count ( /l), total CD3 count ( /l), CD3% (53 84%).

4 SCID and maternal engraftment 230 Table 2 Clinical findings in post BMT patients (group B) Patient Age rash Age BMT Total Total CD3 CD3% Morphology Conditioning Treatment and course onset (m) (y) lymphocyte (10 9 /l) count (10 9 /l) Maculopapular erythematous None Resolved on steroids rash beginning on chest and spreading to face and extremities Diffuse erythematous None Resolved on steroids papular eruption on face, trunk, and extremities Erythematous maculopapular BU16 Self resolved rash on trunk CY200 ATG NA 4 Erythematous maculopapular CY120 Self resolved after mezlocillin rash on neck, axilla, groin ATG90 discontinued and trunk 700 cgy Diffuse erythematous ATG80 Self resolved macular rash. Pruritic papular rash on lower extremities NA 5 Morbilliform eruption CY120 Resolved with topical steroids involving neck, face, behind ATG90 and oral Benadryl ears, scalp and back 700 cgy NA 15 Generalized erythematous ATG80 Resolved on systemic steroids eruption NA 61 Scaly rash over distolateral CY200 Self-resolved aspect of lower extremities. ATG80 Red-purple macular rash over volar surface of forearms Papular hypopigmented rash ATG80 Self resolved on scrotum and erythematous rash on face, trunk and extremities Generalized erythematous CY200 Resolved with systemic macular rash ATG80 immunosuppression Maculopapular rash on None Resolved with increased extensor surfaces of upper cyclosporine dose arms and face ATG80 = anti-thymocyte globulin 80 mg/kg; ATG90 = anti-thymocyte globulin 90 mg/kg; BU14 = busulphan 14 mg/kg; BU16 = busulphan 16 mg/kg; CY120 = cyclophosphamide 120 mg/kg; CY200 = cyclophosphamide 200 mg/kg; NA = not available. Normal ranges: Total lymphocyte count ( /l), total CD3 count ( /l), CD3% (53 84%). ditions that can be seen in SCID patients, such as drug eruptions, exanthems or eczema. In addition, unlike most patients with SCID who have profound lymphopenia, the peripheral blood lymphocyte count may be normal or even elevated in transplacentally engrafted infants because of circulating maternal cells. Maternal engraftment secondary to transplacental transfer of T lymphocytes was observed in 24% (12/51) of all patients with SCID in this series, comparable to the previously reported figure of 25% by Thompson et al. 12 It is possible that a greater percentage was engrafted and that the relative insensitivity of earlier methods precluded accurate detection. Although not utilized in our study, current technology involving locus-specific probes has enabled the retrospective testing of patients for engraftment post BMT. 10 The skin eruptions of GVHD due to maternal engraftment are clinically similar to those seen in patients with GVHD after BMT. These classic lesions have been well characterized and involve an acute phase with morbilliform erythema, papular dermatitis and diffuse erythroderma. The face, neck, palms and soles are usually the initial sites, prior to a generalized eruption. In severe cases, diffuse bullae or toxic epidermal necrolysis is observed. Rashes in our maternally engrafted patients spanned this morphological array. Cutaneous findings ranged from transient morbilliform to total erythroderma with alopecia. Typically, these rashes were refractory to topical treatment prior to BMT and resolved with systemic immunosuppression shortly after BMT. The histopathology of GVHD due to maternal engraftment differs significantly from that due to BMT. As

5 Figure 1 (Patient 9) Generalized rash characterized by scaling and erythema. Table 3 Histopathology chi-square results Parameter Group A Group B P value (%) (%) Psoriasiform hyperplasia * Parakeratosis * Spongiosis * Necrotic keratinocytes * Lichenoid infiltrate * Follicular infiltrate Intradermal lymphocytes Eosinophils Histiocytes *Reaching statistical significance. noted, GVHD secondary to maternal engraftment is characterized by psoriasiform hyperplasia with parakeratosis and variable spongiosis, in contrast to the vacuolar interface pattern observed in conventional GVHD, yet necrotic keratinocytes (suggesting an element of keratinocytotoxicity) are identifiable in both diagnostic groups. 13 The consistent findings of psoriasiform hyperplasia and prominent spongiosis in particular are distinctive and can easily cause diag- SCID and maternal engraftment nostic confusion, since these are seen only rarely in classic GVHD. Lerner et al 14 proposed the following grading scheme of GVHD: grade I, characterized by focal or diffuse vacuolar degeneration (of epidermal basal cells); grade II, characterized by vacuolar degeneration with dyskeratosis (keratinocyte necrosis); grade III, characterized by grade II changes plus subepithelial clefting; and grade IV, characterized by confluent epithelial necrosis. Spongiosis is not typically a prominent feature of GVHD due to BMT but, when present, is usually found in patients with grades III/IV disease, not grades I and II. 15 Our findings of a statistically significant difference in the presence of spongiosis between group A and group B patients supports the conclusion that GVHD due to maternal engraftment has features which are distinct from that due to BMT. Graft-versus-host disease due to maternal engraftment should always be suspected in a neonate with erythroderma. The differential diagnosis of neonatal erythroderma is large including immunodeficiency disorders, infections, viral exanthems, drug eruptions, congenital psoriasis and several forms of ichthyosis. 16 Microscopically, viral exanthems and certain drug eruptions can be difficult to distinguish from acute GVHD. Viral exanthems are characterized by a vacuolar interface pattern, but it is often subtle and necrotic keratinocytes are few compared to GVHD. Drug eruptions, especially drug-induced erythema multiforme (EM) is even more difficult as there are no well-defined histologic criteria to distinguish it from acute GVHD. Involvement of follicular epithelium is more common in GVHD but can be seen in EM. Fortunately, the clinical appearance of GVHD and EM are easily distinguished. Rarely, GVHD in neonates may be confused with ichthyosiform erythroderma, seborrheic dermatitis, atopic dermatitis, psoriasis, pityriasis rubra pilaris, so-called Leiner s disease, Langerhans cell histiocytosis, nutritional deficiencies of zinc, biotin, and essential fatty acids, and even widespread congenital syphilis. However, clinicopathologic correlation is often sufficient to distinguish GVHD from these entities. The differentiation of GVHD caused by maternal engraftment and Omenn s syndrome deserves special consideration. Omenn s syndrome is an autosomal recessive form of combined immunodeficiency with a gene defect that has been mapped to chromosome Typical features include a widespread skin eruption, alopecia, lymphocytosis, hepatosplenomegaly, lymphadenopathy, recurrent infections and failure to thrive. These features can also be seen in some cases of maternally induced GVHD. However, in Omenn s syndrome the degree of eosinophilia, lymphadenopathy and hepatosplenomegaly are usually more pronounced. 16 Some reported cases of Omenn s syndrome have had microscopic features which strongly resemble those of our maternally engrafted patients. 7 In addition, many either did not mention or did not have documentation of maternal engraftment using current, more sensitive methods of detection. Thus, some patients reported as having Omenn s syndrome may actually have had maternally-induced GVHD. Three of our patients had many features of Omenn s syndrome. Patient 10 presented with a total body erythematous rash, fever, hepatomegaly and eosinophilia at 1 month of 231

6 SCID and maternal engraftment 232 Figure 2 A low magnification view of the eruption associated with maternal engraftment shows slight psoriasiform epidermal hyperplasia, focal spongiosis and parakeratosis, and a fairly dense papillary dermal infiltrate. This lichenoid-psoriasiform pattern contrasts distinctly to the pattern of conventional GVHD, in which the epidermis usually displays vacuolar alteration of the lower most epidermis in association with a sparse superficial dermal infiltrate. Figure 3 Higher magnification reveals single necrotic keratinocytes with closely opposed lymphocytes, similar to the pattern of satellite cell necrosis in conventional GVHD. age. He was diagnosed with SCID in 1983, and laboratory studies available at this time may not have been sensitive enough to detect maternal engraftment. Patient 7 presented at age 5 months with a fine erythematous maculopapular rash over his trunk and extremities, lymphadenopathy, hepatomegaly and eosinophilia, but was found to be maternally engrafted by VNTR analysis. Patient 2 presented at 4 months with a generalized eczematous rash accompanied by lymphadenoapthy, hepatosplenomegaly, and eosinophilia and was also found to be maternally engrafted by VNTR analysis. In spite of conclusive evidence of maternal engraftment in patients 2 and 7 and high clinical suspicion in patient 10, coexistent Omenn s syndrome is not necessarily excluded since none had specific gene testing for Omenn s syndrome. The explanation for the differences in cutaneous GVHD in group A and classical GVHD is probably multifactorial. One difference may be the limited T cell receptor diversity of transplacentally acquired maternal T cells in SCID. Thus, Knobloch et al 5 demonstrated that maternal T cells express a profoundly reduced heterogeneity in their TCR repertoire by using a panel of mab specific for TCR V region epitopes. This lack of heterogeneity may correlate with the distinct histopathological findings in those patients receiving maternal T cells transplacentally, as opposed to via BMT. Similarly, T cell functions may be limited, corre-

7 sponding to the number of maternal cells that cross the placenta. Unlike the better documented and more abundant presence of fetal lymphocytes in the circulation of the mother, it has been suggested that only trace amounts of maternal cells enter the fetus. Thus, the distinct features of cutaneous GVHD in patients with maternal engraftment pre-bmt may be due to functional characteristics of the T cells or a functionally limited repertoire of T cells when these patients are engrafted transplacentally. The early diagnosis of SCID is critically important since early BMT can provide a definitive cure and avoid severe morbidity or mortality from the immunodeficiency itself. The clinician should consider the diagnosis of SCID when confronted with a widespread cutaneous eruption and associated pathology as described above. This study highlights the diagnostic value of skin histopathology in infants with widespread rashes along with other symptoms such as thrush, diarrhea and failure to thrive. Recognizing the clinicopathologic findings described in this study may aid in earlier diagnosis of SCID with maternally induced GVHD, which in turn can lead to earlier treatment. Acknowledgements Mort Cowan, MD would like to acknowledge the NIH RO1 AI28339 and March of Dimes FY grants. References 1 Buckley R, Schiff R, Schiff S et al. Human severe combined immunodeficiency: genetic, phenotypic and functional diversity in one hundred eight infants. J Pediatr 1997; 130: Dror Y, Gallagher R, Wara DW et al. Immune reconstitution in severe combined immunodeficiency disease after lectintreated, T-cell depleted haplocompatible bone marrow transplantation. Blood 1993; 8: Cavazzana-Calvo M, Hacein-Bey S, de Saint Basile G et al. Gene therapy of human severe combined immunodeficiency (SCID)-X1 disease. Science 2000; 288: Geha R, Reinherz E. Identification of circulating maternal T and B lymphocytes in uncomplicated severe combined immunodeficiency by HLA typing of subpopulations of T cells separated by the fluorescence-activated cell sorter and of SCID and maternal engraftment Epstein Barr virus-derived B cell lines. J Immunol 1983; 130: Knobloch C, Goldman S, Friedrich W. Limited T cell receptor diversity of transplacentally acquired maternal T cells in severe combined immunodeficiency. J Immunol 1991; 146: Alain G, Carrier C, Beaumier L et al. In utero acute graftversus-host disease in a neonate with severe combined immunodeficiency. J Am Acad Dermatol 1993; 29: Appleton A, Curtis A, Wilkes J, Cant A. Differentiation of materno fetal GVHD from Omenn s syndrome in pre-bmt patients with severe combined immunodeficiency. Bone Marow Transplant 1994; 14: Farrell A, Scerri L, Stevens A, Millard L. Acute graft-versushost disease with unusual cutaneous intracellular vacuolation in an infant with severe combined immunodeficiency. Pediatr Dermol 1995; 12: Gatti R, Nakamura Y, Nussmeier M et al. Informativeness of VNTR genetic markers for detecting chimerism after bone marrow transplantation. Dis Markers 1989; 7: Katz F, Malcom S, Strobel S et al. The use of locus-specific minisatellite probes to check for engraftment following allogeneic bone marrow transplantation for severe combined immunodeficiency disease. Bone Marrow Transplant 1990; 5: Tachinami T, Koizumi S, Yachie A et al. Immune status in two brothers with Omenn s syndrome: no discernible chimerism on FACS analysis using a monoclonal antibody specific for a maternally restricted HLA antigen. J Pediatr Hematol Oncol 1990; 12: Thompson LF, O Connor RD, Bastian JF. Phenotype and function of engrafted maternal T cells in patients with severe immunodeficiency. J Immunol 1984; 133: Parkman R, Rappeport J, Rosen F. Human graft versus host disease. J Invest Dermatol 1980; 74: Lerner K, Kao G, Storb R et al. Histopathology of graft-vshost reaction (GvHR) in human recipients of marrow from HL-A-matched sibling donors. Transplant Proc 1974; 4: Massi D, Franchi A, Pimpinella N et al. A reappraisal of the histopathologic criteria for the diagnosis of cutaneous allogenic acute graft-vs-host disease. Am J Clin Pathol 1999; 112: Hoeger P, Harper J. Neonatal erythroderma: differential diagnosis and management of the red baby. Arch Dis Child 1998; 79: Villa A, Santagata S, Bozzi F et al. partial V(D)J recombination activity leads to Omenn syndrome. Cell 1998; 93: