|Home | About | Journals | Submit | Contact Us | Français|
Recessive dystrophic epidermolysis bullosa is an incurable, often fatal mucocutaneous blistering disease caused by mutations in COL7A1, the gene encoding type VII collagen (C7). On the basis of preclinical data showing biochemical correction and prolonged survival in col7−/− mice, we hypothesized that allogeneic marrow contains stem cells capable of ameliorating the manifestations of recessive dystrophic epidermolysis bullosa in humans.
Between October 2007 and August 2009, we treated seven children who had recessive dystrophic epidermolysis bullosa with immunomyeloablative chemotherapy and allogeneic stem-cell transplantation. We assessed C7 expression by means of immunofluorescence staining and used transmission electron microscopy to visualize anchoring fibrils. We measured chimerism by means of competitive polymerase-chain-reaction assay, and documented blister formation and wound healing with the use of digital photography.
One patient died of cardiomyopathy before transplantation. Of the remaining six patients, one had severe regimen-related cutaneous toxicity, with all having improved wound healing and a reduction in blister formation between 30 and 130 days after transplantation. We observed increased C7 deposition at the dermal–epidermal junction in five of the six recipients, albeit without normalization of anchoring fibrils. Five recipients were alive 130 to 799 days after transplantation; one died at 183 days as a consequence of graft rejection and infection. The six recipients had substantial proportions of donor cells in the skin, and none had detectable anti-C7 antibodies.
Increased C7 deposition and a sustained presence of donor cells were found in the skin of children with recessive dystrophic epidermolysis bullosa after allogeneic bone marrow transplantation. Further studies are needed to assess the long-term risks and benefits of such therapy in patients with this disorder. (Funded by the National Institutes of Health; ClinicalTrials.gov number, NCT00478244.)
Loss of Skin Integrity That Leads to trauma-induced blisters and erosions is a defining feature of epidermolysis bullosa, a heterogeneous group of more than 20 inherited blistering diseases with highly variable clinical severity.1 One of the most severe forms is recessive dystrophic epidermolysis bullosa, caused by loss-of-function mutations in the collagen type VII (C7) gene (COL7A1).2-6 These mutations result in severely diminished expression of C7, a collagen localized at the dermal–epidermal junction. C7 is the major component of the anchoring fibrils that tether the epidermal basement membrane to the dermal matrix. In the absence of normal C7 expression, these fibrils do not form properly, and epidermal–dermal adherence is lost beneath the lamina densa of the basement membrane.
From birth on, children with recessive dystrophic epidermolysis bullosa have painful erosions and blisters on mucosal membranes and skin, often resulting in esophageal strictures, mutilating scars, local and systemic infections, joint contractures, fusion of fingers and toes, and aggressive squamous-cell carcinomas.7-9 For patients with the Hallopeau–Siemens variant of recessive dystrophic epidermolysis bullosa, severe mucocutaneous disease is apparent at birth and the median survival is approximately 30 years, whereas persons with other types of recessive dystrophic epidermolysis bullosa have a median survival of 55 to 65 years.10
To date, the care of patients with recessive dystrophic epidermolysis bullosa has been palliative and restricted to the treatment of individual wounds.11,12 To assess the potential effectiveness of bone marrow stem-cell transplantation for the systemic correction of recessive dystrophic epidermolysis bullosa, we evaluated this approach in a mouse model13 and observed that a stem-cell–enriched fraction of bone marrow prevented postnatal death in a proportion of mice with recessive dystrophic epidermolysis bullosa. Fifteen percent of the treated mice survived for longer than 80 days after transplantation; these mice had expression of wild-type C7 in skin and mucosa, formation of new anchoring fibrils, and resistance to blistering.14 We then conducted a phase 1–phase 2 clinical trial of bone marrow transplantation for the treatment of severe recessive dystrophic epidermolysis bullosa.
Between October 2007 and August 2009, we enrolled seven children with recessive dystrophic epidermolysis bullosa in a trial of immunomyeloablative chemotherapy and transplantation of allogeneic stem cells. Follow-up data are reported through January 8, 2010.
To optimize engraftment, we used a modified conditioning regimen for immunomyeloablation that consisted of busulfan (0.8 mg per kilogram of body weight per dose for children who weighed 12 kg or more and 1.1 mg per kilogram for those who weighed less than 12 kg, delivered intravenously every 6 hours on days 9 to 6 before transplantation, with dose modifications based on pharmacokinetics, targeting 1000 μmol per minute for the area under the curve), fludarabine (25 mg per square meter of body-surface area per day, given intravenously on days 5 to 3 before transplantation), and cyclophosphamide (50 mg per kilogram per day, given intravenously on days 5 to 2 before transplantation). (The treatment schema is shown in Fig. 1 in the Supplementary Appendix, available with the full text of this article at NEJM.org.) Immunoprophylaxis against graft-versus-host disease consisted of cyclosporine (targeting trough levels of 200 to 400 μg per liter on day 3 before transplantation to day 100 after transplantation, with the dose tapered by 10% each week thereafter) and mycophenolate mofetil (15 mg per kilogram on day 3 before transplantation to day 45 after transplantation).
On the day of transplantation, five patients received unfiltered marrow stem cells from an HLA-identical sibling; one of the five (Patient 1) concomitantly received umbilical-cord blood from the same sibling donor. Patient 3 received umbilical-cord blood from an unrelated donor (HLA-5/6, with B antigen mismatched). Owing to severe chemotherapy-related toxicity, one patient (Patient 2) did not receive the scheduled marrow infusion. Except for intensive skin care and placement of a gastrostomy tube to minimize vomiting, supportive care was standard for the recipients of an allogeneic bone marrow transplant.15 The study protocol was approved by the University of Minnesota institutional review board. Parents of all the study participants provided written informed consent, and Patient 7 provided assent.
To quantify and determine the location of C7, skin-biopsy specimens were obtained at perilesional sites on the extremities, approximately 1 mm away from an erosion, and were stained with anti-C7 antibodies. Six areas were selected at random locations along the dermal–epidermal junction to quantify the C7 fluorescent-signal intensity (see the Methods section in the Supplementary Appendix).
Skin samples were also fixed and processed for transmission electron microscopy, as described previously.14 Qualitative and quantitative assessments, made at specific time points after treatment, focused on the development of anchoring fibrils beneath the lamina densa. Previously defined criteria for ultrastructural evaluation were applied16-18; specifically, the insertion of a fibrillar structure into the dermal surface of the lamina densa and central cross-banding, a fan-shaped appearance, or both were considered morphologic hallmarks of anchoring fibrils. Electron-microscopical images were reviewed independently by three dermatopathologists without knowledge of clinical outcome data.
We carried out fluorescence in situ hybridization (FISH) with the use of established methods (see the Supplementary Appendix) and determined the extent of blood and skin chimerism by means of competitive polymerase-chain-reaction analysis of polymorphic variable-number tandem-repeat regions.15 We used a quantitative enzyme-linked immunosorbent assay to test for anti-C7 antibodies in blood samples.19 Images were scanned with the Epson Perfection V700 Photo Scanner, and densitometry was performed with the use of ImageJ software.
Six children with recessive dystrophic epidermolysis bullosa between the ages of 15 months and 14.5 years completed treatment and could be evaluated. Characteristics of the patients and grafts and the outcomes of transplantation are summarized in Table 1. Pretransplantation evaluations showed extensive cutaneous disease in each of the six patients and severe mucosal disease in four of the six, who required esophageal dilation and gastrostomy-tube placement for nutritional support. Five patients had severe mitten deformities, and four used wheelchairs. In addition, two had clinically significant renal impairment (glomerular filtration rate, 26.01 and 79.49 ml per minute per 1.73 m2 of body-surface area in Patient 2 and Patient 4, respectively) and four had severe iron-deficiency anemia.
Despite a 50% reduction in the dose of fludarabine because of a reduced glomerular filtration rate, Patient 2 died before the bone marrow infusion could be carried out as a result of hemorrhagic cardiomyopathy that was probably due to cyclophosphamide cardiotoxicity. Among the six patients who underwent transplantation, neutrophil recovery (≥5×108 cells per liter) occurred on days 11, 24, 16, 12, 14, and 18, in Patients 1, 3, 4, 5, 6, and 7, respectively, and platelet recovery (≥5×1010 platelets per liter) occurred on days 43, 33, 46, 37, and 41 in Patients 1, 4, 5, 6, and 7, respectively, but was not achieved in Patient 3. Regimen-related toxicity included grade 3 mucositis (in six patients); transient, clinically significant hyperbilirubinemia (in four patients), with a peak total bilirubin level of 6.9 to 19 mg per deciliter (118 to 325 μmol per liter); renal insufficiency requiring 3 days of hemodialysis for fluid management (in two patients), with peak creatinine levels of 1.66 mg per deciliter (146.7 μmol per liter) in Patient 4 and 1.58 mg per deciliter (139.7 μmol per liter) in Patient 5; and opportunistic infections (candidemia in Patients 4 and 7 and viral reactivations in Patient 3 [adenovirus, Epstein–Barr virus, BK virus, and human herpesvirus type 6] and Patient 4 [BK virus]). No patient had acute or chronic graft-versus-host disease. Five patients were alive 799, 387, 268, 196, and 130 days after bone marrow transplantation, with Patients 1, 4, 5, and 6 having been taken off all immunosuppressive therapy and Patient 7 tolerating the tapering of the cyclosporine dose. Patient 3, who required a second transplant because of graft rejection, had hematopoietic recovery (on day 22 after the second infusion of umbilical-cord blood) with complete marrow chimerism; however, subsequent infections resulted in his death on day 183.
Mucocutaneous blistering was not adversely affected early after transplantation except in Patient 7, in whom grade 4 cutaneous toxicity developed on day 17, presumably related to the chemotherapy regimen. All the patients had increased wound healing and decreased mucocutaneous blistering within the first 100 days after transplantation (Fig. 1). Patients 4, 5, and 7 had marked clinical improvement as early as day 30. The percentage of body-surface area affected was reduced in Patients 1, 3, and 6 according to parents' reports and clinical observation, with more objective evidence provided by documented reductions in bandage use (assessed on the basis of the weight of precut dressings) in Patient 5 at 6 months and in Patient 4 at 1 year, as compared with bandage use before transplantation (Table 1). The parents of each patient reported substantial improvements in the rate of wound healing and marked reductions in oral mucosal lesions. We tested the skin of the two most recently enrolled patients (Patients 6 and 7) for resistance to blister formation on exposure to negative pressure. In both patients, the time to blister formation gradually increased after transplantation and exceeded that observed before transplantation (Fig. 2 in the Supplementary Appendix).
Skin-biopsy specimens were evaluated for the presence of C7 (Fig. 1). In Patient 1, skin immunofluorescence microscopy at baseline revealed faint linear and interrupted C7 immunolabeling at the dermal–epidermal junction; in contrast, by day 200 there was bright linear C7 immunoreactivity at the dermal–epidermal junction, which was even more intense by day 365 after transplantation. Skin-biopsy specimens from Patients 3, 4, 5, and 7 also revealed increases in C7 immunoreactivity at the dermal–epidermal junction after transplantation. In addition, we found that cultured fibroblasts from skin-biopsy specimens obtained after transplantation produced C7, as detected on Western blotting (data not shown).
Skin-biopsy specimens from all the patients except Patient 1 were evaluated with the use of six anti-C7 antibodies, which could potentially recognize different epitopes on wild-type C7. Testing with at least one anti-C7 antibody revealed a progressive increase in C7 expression as measured by increased fluorescence intensity over time in five of the six patients (Fig. 2); in Patient 6, C7 could not be detected at any point in time. C7 was either absent or barely detectable with at least one anti-C7 antibody before treatment in Patients 3 and 4; however, expression of C7 became readily apparent after transplantation and increased substantially over time.
Electron micrographs obtained at baseline revealed a complete absence of mature anchoring fibrils in all the patients. After transplantation, electron micrographs in all but one patient revealed only scanty, wispy structures under the lamina densa, as shown in Patient 7 (Fig. 3). Although this fibrillar material could represent very rudimentary anchoring fibrils or fragmented elastic microfibrils, more precise interpretation or reproducible quantification was not possible. In Patient 4, we observed occasional fibrillar structures below the lamina densa that were consistent with rudimentary anchoring fibrils (Fig. 3), but these structures were thin and generally lacked central cross-banding. After transplantation, all patients had an increase in the number of such structures, but none had the morphologic hallmarks of normal anchoring fibrils. Other ultrastructural changes included thickening of the lamina densa in Patients 4 and 5 (Table 1 and Fig. 3).
Because anti-C7 antibodies could interfere with the benefit of transplantation, we assayed levels of anti-C7 antibodies in serum. Before transplantation, we detected low antibody titers in Patients 1, 3, and 7; no anti-C7 antibodies have been detected since transplantation.
After transplantation, all patients (including Patient 3, who required a second umbilical-cord blood transplant) had complete replacement of the marrow and blood by donor lymphohematopoietic cells. Moreover, all patients had evidence of donor cells in the skin, with a median of 20% (range, 0 to 93) being of donor origin. The percentage of donor cells varied between time points and with the biopsy site (Table 1). Patient 5 had a sex-mismatched donor. With the use of FISH to identify X and Y chromosomes (male donor), anti-CD45 and anti-CD31 fluorochromes were used to determine the identity of the donor cells in the skin-biopsy specimens. We observed extravascular localization of CD45+ (hematopoietic) donor cells and CD45−CD31− (nonhematopoietic, nonendothelial) donor cells in the papillary dermis and epidermis in this patient (Fig. 3 in the Supplementary Appendix).
We found that the infusion of allogeneic bone marrow can partially correct the C7 deficiency and improve skin and mucosal integrity in patients with recessive dystrophic epidermolysis bullosa. The observation that partial C7 correction is associated with a clinical benefit is consistent with previous work showing that a C7 level as low as 10% of the wild-type level of expression is associated with stability of the skin.20 The rates of recovery and ultimate outcomes varied among the surviving patients. Patients 4 and 5 had rapid and substantial clinical improvement relative to baseline, whereas Patient 1 had slow improvement, with only a modest overall benefit, and Patient 7 had a recurrence of blistering after an early period (days 30 to 60) of almost no blistering.
Unexpectedly, we detected substantial proportions of donor cells in the skin and mucosa after treatment; these proportions varied over time and with the location of the biopsy site. Many of the donor cells were located some distance away from blood vessels, and many were hematopoietic in origin (CD45+ cells). Although more work needs to be done to determine the identity of the CD45−CD31− donor cells, we favor the possibility that these healthy donor cells residing in the skin secrete C7 and that the secreted C7 is subsequently incorporated into the lamina densa at the dermal–epidermal junction.
The effect of new C7 on the morphologic features of anchoring fibrils, however, is not clear. Because all the patients except Patient 6 had mutant C7 expression at baseline (with Patients 4, 5, and 7 having higher levels of expression than Patients 1 and 3), it is possible that small amounts of newly formed C7 assembled with wild-type C7 early after transplantation to form a trimer and thus improved mucocutaneous integrity. This hypothesis also predicts that the C7 gene dose (i.e., the zygosity status of the donor) could influence the rate of improvement and is consistent with the blunted overall clinical outcome in Patient 1, who had a heterozygous donor with only one normal C7 gene. An initial scaffold of rudimentary anchoring fibrils in the host might be necessary for optimal deposition of wild-type C7 from the allogeneic cells, but it is clear that a clinical response does not depend on the presence of increased numbers of anchoring fibrils. Our data are limited to assessments of skin-biopsy specimens obtained within the first year after transplantation. Although rare anchoring-fibril–like structures were seen in Patient 4, the development of anchoring fibrils that fulfill the established ultrastructural criteria may require a longer interval after bone marrow infusion. Together, our findings suggest that the infusion of bone marrow from a healthy donor can ameliorate recessive dystrophic epidermolysis bullosa in humans, as has previously been shown in the mouse model of the disorder. Substantial efforts are under way to understand the physiology of the apparent clinical response after bone marrow transplantation and to identify the stem-cell population responsible for this effect.
Despite the apparent beneficial effect of bone marrow stem cells in patients with recessive dystrophic epidermolysis bullosa, allogeneic bone marrow transplantation remains a high-risk procedure typically reserved for patients with imminently life-threatening disease, such as those with high-risk lymphohematopoietic malignant disease.21 Before this clinical trial, it was not known whether patients with preexisting mucocutaneous disease could tolerate the conventional conditioning regimens used as preparation for allogeneic bone marrow transplantation. Although grade 3 mucositis occurred in all the patients in our study, as is typical for patients without preexisting mucocutaneous disease, only one patient had severe cutaneous toxicity. Notably, no patient had uncontrolled cellulitis, despite pretransplantation bacterial or fungal skin colonization. Clearly, the risks of bone marrow transplantation need to be weighed against the risks and benefits of alternative innovative therapies now being explored (e.g., injections of C7 protein or ex vivo grafting of allogeneic or gene-corrected autologous cells).17,22-35 Although these measures may prove to ameliorate morbidity, recessive dystrophic epidermolysis bullosa is a systemic disease, and the benefits of these other interventions have thus far been brief or limited to small areas of skin. Therefore, new approaches, such as the one we describe here, are being explored. Already, we and others are considering modifications to enhance safety, such as the coinfusion of mesenchymal stromal cells or the use of reduced-dose conditioning before bone marrow transplantation, recognizing that the latter approach may increase the risk of opportunistic infection as a result of delayed immune reconstitution brought about by the additional immunosuppression.
In summary, our data suggest that the infusion of allogeneic bone marrow stem cells can lead to increased C7 deposition and reduced blistering in patients with severe recessive dystrophic epidermolysis bullosa. However, the unique skin and mucosal membrane defects of this disease pose a particular challenge to any bone marrow transplantation program. Clearly, much remains to be learned regarding the mechanism of the apparent functional correction as well as the long-term risks and benefits of this therapeutic approach, including the risk of squamous-cell carcinoma, which may occur after chemotherapy or as a result of incomplete correction of the underlying disease. Despite the potential benefits of marrow transplantation, it is a high-risk therapeutic approach that could shorten the expected survival of patients with recessive dystrophic epidermolysis bullosa, particularly those with less severe clinical manifestations. Nevertheless, the results to date support the study of stem-cell–based approaches for patients who have the severest forms of epidermolysis bullosa.
Supported in part by grants from the National Institutes of Health (5R01-HL049997, to Dr. Blazar), from the Ministry of Health, Labor, and Welfare of Japan and the Ministry of Education, Culture, Sports, Science, and Technology of Japan (to Dr. Ishida-Yamamoto), and from the University of Minnesota Academic Health Center, the Epidermolysis Bullosa (Liao Family) Research Fund, the Sarah Rose Mooreland EB Fund, and the Children's Cancer Research Fund.
We thank Dr. Angela M. Christiano for her helpful review of the original treatment protocol; Ron McElmurry, Brandon Peacock, Lily Xia, Amanda Kobs, Cindy Eide, Trevor Keyler, Christopher Lees, Carolyn Mann-Sandbakken, Meghan Munger, Michael Gitler, Dr. Ana Junqueira, Teresa Kivisto, and Pat Fidler for their technical and clinical help; Drs. Mei Chen and David Woodley for the anti-C7 antibody and for testing of anti-C7 antibodies in patients' blood before and after transplantation; and Noriko Takashita and Stomi Igawa for preparing the electron-microscopical images for the study.
Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.