You are on page 1of 8

A Phase I Clinical Trial of the Treatment of Crohn’s Fistula by Adipose Mesenchymal Stem Cell Transplantation

Damián García-Olmo, M.D.,1,2,3 Mariano García-Arranz, Ph.D.,2 Dolores Herreros, M.D.,1,2 Isabel Pascual, M.D.,1,2 Concepción Peiro, Ph.D.,4 José Antonio Rodríguez-Montes, M.D.1,3
1 2

Department of General Surgery, “La Paz” University Hospital, Madrid, Spain Cell Therapy Laboratory, “La Paz” University Hospital, Madrid, Spain 3 Department of Surgery, Universidad Autónoma de Madrid, Madrid, Spain 4 Department of Pharmacology and Therapeutics, “La Paz” University Hospital, Madrid, Spain

PURPOSE: The effective management of fistulas in patients with Crohn’s disease presents an extremely challenging problem. Mesenchymal adult stem cells extracted from certain tissues, such as adipose tissue, can differentiate into various cell types. Therefore, we have tried to use such cells to stimulate healing of Crohn’s fistulas. METHODS: We designed a prospective Phase I clinical trial, involving five patients with Crohn’s disease, to test the feasibility and safety of autologous stem cells transplantation in the treatment of fistulas. We also studied the expression of various cell markers and the growth rates of the lipoaspirate-derived cells that were used for transplantation. RESULTS: One patient was excluded because of bacterial contamination of cultured cells. We inoculated nine fistulas in four patients with autologous adipose tissue-derived stem cells at Passage 3 or earlier. Eight inoculated fistulas were followed weekly for at least eight weeks. In six fistulas, the external opening was covered with epithelium at the end of Week 8, and, thus, these fistulas were considered healed (75 percent). In the other two fistulas, there was only incomplete closure of the external opening, with a decrease in output flow (not healed; 25 percent). No adverse effects were observed in any patient at the end of the follow-up period (minimum

follow-up,12 months; maximum follow-up, 30 months; follow-up average, 22 months). CONCLUSIONS: To our knowledge, this is the first report of a clinical trial of cell therapy using autologous stem cells obtained from a lipoaspirate. Our results indicate that our protocol is feasible and safe for the treatment of fistulas in Crohn’s disease. The number of patients included and the uncontrolled nature of Phase I clinical trials do not allow demonstration of the effectiveness of the treatment. However, the results of the present study encourage to perform further studies in Phase II. [Key words: Crohn’s fistula; Stem cell; Clinical trial]

Presented in part at IFATS-2004, Pittsburgh, Pennsylvania, October 3 to 5, 2004. Correspondence to: Damia ´ n Garcı ´a-Olmo, M.D., Servicio de Cirugía General C, Hospital Universitario La Paz, Paseo de la Castellana 261, 28046, Madrid, Spain, e-mail: Dis Colon Rectum 2005; 48: 1416–1423 DOI: 10.1007/s10350-005-0052-6 © The American Society of Colon and Rectal Surgeons Published online: 17 May 2005

he reported incidence of fistula in Crohn’s disease ranges from 17 to 50 percent.1 Management of fistulas in patients with Crohn’s disease continues to present an extremely challenging problem, because many such fistulas do not respond to available treatments.2 Such fistulas and their recurrence are a distressing complication that significantly reduces the quality of life of affected patients. Recent improvements in medical treatment (e.g., treatment with Infliximab®) and expert surgical management have decreased the need for complicated surgery. However, many patients are not cured. Failure of fistulas to heal is probably the result of the suboptimal quality of tissues that have been affected by Crohn’s disease.3,4 Mesenchymal adult stem cells extracted from certain tissues, such as adipose tissue, can differentiate into various types of cell.5 Moreover, lipoaspirates ob-



The raw lipoaspirate was washed extensively with . 7 STEM CELLS IN CROHN’S FISTULA Table 1. the canula was moved through the adipose abdominal-wall compartment for mechanical disruption of the fatty tissue. mental handicap. five in enterocutaneous fistulas (4 different fistulas in 1 patient). or Octreotide concurrently with this procedure.5 20 30 20 15 1417 Implant 1 2 3 NI 4 5 6 7 8 9 Patient 001 002 002 003 002 002 001 004 005 002 Gender F F F M F F F M M F Type of Fistula Rectovaginal Enterocutaneous Enterocutaneous Perineal Rectovaginal Enterocutaneous Rectovaginal Enterocutaneous Perineal Enterocutaneous Middle line Periumbilical Suprapubic Right lower quadrant Right lower quadrant Passage 1 1 1 1 2 2 1 2 2 3 Outcome Healed Healed NA NA Not Healed Healed Healed Not Healed Healed Healed F = female. with signature of the informed-consent form. All enterocutaneous fistulas had low flow. Patients 001 and 002 required two liposuction procedures. Summary of Clinical Results Age (yr) 35 40 40 36 40 41 37 36 32 41 Culture Time (days) 6 9 7 NA 14 8 12 16 31 12 Cell Number (×106) 6. 80 to 100 ml of raw of lipoaspirate were obtained from each patient. and AIDS. <50 ml per day. clinical trial. 35.2–37. We reported previously the successful treatment of one patient with a Crohn’s fistula by autologous stem-cell lipoaspirate transplantation. presence of one or more complex Crohn’s fistulas (enterocutaneous fistula. No. The Ethics Committee was kept informed about the progress of the study throughout the clinical trial. 3 males.1 ± 2. suprasphincteric fistula. and chondrogenic potential. A saline solution and the vasoconstrictor epinephrine were injected into the adipose compartment to minimize blood loss. PATIENTS AND METHODS The Phase I clinical trial was designed to test the feasibility and safety of autologous stem cell transplantation for treatment of Crohn’s fistulas.4 (range.Vol. under local anesthesia and general sedation. We describe a small. During the clinical trial. A hollow blunttipped canula was introduced into the subcutaneous space through a small incision (<0. Isolation of Stem Cells Adipose tissue was obtained by liposuction. Remicade. diagnosis of Crohn’s disease at least five years before the trial. which we designed to ascertain the feasibility and safety of the use of autologous adipose stromal stem cell transplantation for treatment of Crohn’s fistulas. The patients were selected according to the following inclusion criteria: older than aged 18 years. Nine cell implants were performed: three in rectovaginal fistulas. extreme thinness. allergy to local anesthetics. NA = not analyzed. 31. no stem cells survived cryopreservation. tained from human adipose tissue contain a population of stem cells of putative mesodermal lineage with myogenic. previous diagnosis of cancer. NI = no implant. The Clinical Trial and Ethics Committee of La Paz Hospital approved the protocol on April 12. No patient was treated with Total Parenteral Nutrition. 2002. average age. and/or rectovaginal fistula) that had been unrespon- sive to medical treatment and unsuccessfully treated by classic surgery at least twice.7 Crohn’s fistulas provide a model system for wound healing under some of the worst possible conditions.1 9 8 NI 10 3.5 cm in diameter).6 We have tested the effects of transplantation of such cells in the treatment of Crohn’s fistulas. In this way. adipogenic. we studied the expression of various markers of differentiation in the lipoaspirate-derived cells used for transplantation. M = male. With gentle suction.5) years) were enrolled in the study from April 2002 to November 2003. and agreement to participate. and were located in the abdominal wall (Table 1). because after the first liposuction. The exclusion criteria were as follows: failure to meet inclusion criteria. and one in a suprasphincteric perianal fistula. 48. Five patients (Patients 001–005. and a detailed informed-consent form was generated to be signed by patients.

Biotechnology Preparation of Inoculum For clinical use. cells were used after three or fewer passages (Table 1). Chemicon. Medium was replaced every two days. Göttingen. Gibco BRL. Characterization of Cells by Immunofluorescence Staining Cells were plated at low density in DMEM plus 10percent FBS on glass coverslips in 24-well plates. Chemicon. 1/50. cells were incubated at 4°C overnight with primary antibodies against the following cell markers at the indicated dilutions: 1) alphaactin. human synovial fibroblasts also were cultured and their growth rates determined. UK) to remove blood cells. Chandlers Ford.5-percent. Sigma. and the suspension was centrifuged at 110 × g for 5 minutes. and human synovial fibroblasts were used as positive controls for immunostaining with the various antibodies. Then. For staining of ␣-actin. Cells were resuspended in 0. Trypsinization was stopped by addition of DMEM plus FBS. 1/50. 1/50. Cells were then passaged with trypsin-EDTA (Gibco BRL) at a dilution of 1:3. Human aortic smooth muscle cells. we omitted the primary antibodies. 1/200. which contained 10 percent fetal bovine serum (FBS. FL). we determined the numbers of immunopositive cells in different fields and compared them to the numbers of stained nuclei. The extracellular matrix was digested with a solution of Type II collagenase (0. CYMBUS. Glostrup. St. 7) desmin. with replacement of the culture medium every three to four days. 8) cytokeratin. cells were washed with PBS and fixed in acetone for 10 minutes at −20°C. Randomly selected fields were exported to a computer (MacIntosh G3. 1/100. Gibco BRL).. UK. After blocking with a PBS that contained 4-percent goat serum and 0.16 M NH4Cl and allowed to stand for 10 minutes at room temperature (RT) for lysis of erythrocytes. 2-percent. LTD. 1/100. Cells were washed in PBS and the suspension was centrifuged .. 1/100.99). The suspension of cells was centrifuged at 250 × g for 10 minutes. 2) vimentin.6diamidino-2-phenylindole (DAPI). saline. After cells had attached to the substratum (3 hours). the culture medium was replaced by DMEM supplemented with 1-percent antibiotics plus 0. Tampa. At 24-hour intervals. and 9) S-100. by monitoring absorbance at 595 nm. The cells were maintained in culture in the same medium and under the same conditions until they reached approximately 80 percent confluence. Cells were then mounted in Mobiglow (MoBiTec.075 percent. Japan). Hants. Germany) and observed with an epifluorescence microscope Eclipse TE300 (Nikon. In each case. Dako. 6) c-Kit. and local anesthetic. 1/100. Quantitation of Cell Growth Cells were plated in 24-well plates at a concentration of 5 × 103 cells/cm2. Paisley. Denmark. Louis. For transplantation. 4) Factor VIII. human umbilical vein endothelial cells. after nuclear staining with crystal violet. Then the dishes were washed with PBS to remove nonadhering cells and cell fragments. Dako. and cells were resuspended in DMEM plus 10 percent FBS and 1 percent ampicillin/streptomycin mixture (Gibco. 5-percent. For immunofluorescence studies. Dako. Cupertino. Cell Culture Cells were cultured for 24 hours at 37°C in an atmosphere of 5-percent CO2 in air. Apple Computer Ink.1-percent Triton X-100. cells were fixed with 1-percent glutaraldehyde and the number of cells per well was determined. Then. Cell characterization was performed using cells at passages 1 to 9. A standard curve was constructed to establish the relationship between cell number per well and absorbance at 595 nm (r2 = 0. Dako. For negative controls.1418 ´A-OLMO ET AL GARCI Dis Colon Rectum. cells were fixed in 4-percent paraformaldehyde for 10 minutes at RT. Scotland. July 2005 sterile phosphate-buffered saline (PBS. CA) through a Spot1 camera (Diagnostic Instruments Inc. cells were incubated with the appropriate fluorescein isothiocyanate (FITC)-conjugated or tetramethylrhodamine isothiocyanate chloride (TRITC)-conjugated second antibodies (Sigma. MO) for 30 minutes at 37°C to release the cellular fraction. CA. Gibco BRL) in balanced salt solution (5 mg/ ml. Dako. BRL) and then were plated in 100-mm tissueculture dishes at a concentration of 10 to 15 × 103 cells/cm2. Sigma. we used cells between passages 1 and 3. the collagenase was inactivated by addition of an equal volume of Dulbecco’s modified Eagle’s medium (DMEM. Nuclei were counterstained with 4 Ј . 3) CD 90. Gibco BRL). 1/100. The mixture was centrifuged at 250 × g. Tokyo. Cell cultures were trypsinized for three minutes at 37°C. As positive controls for testings of each batch of serum. or 10percent FBS. 1/50) for 45 minutes at RT. 5) CD 34.

Baxter. 7 STEM CELLS IN CROHN’S FISTULA 1419 again at 150 × g for 5 minutes. The blister in the rectal mucous after cells had been injected close to the sutured internal opening. 48. we injected cells into the wall of the track. independent of previous observations. the patient was dismissed and follow-up visits were scheduled at the outpatient clinic. follow-up average. all tracks were scraped. we used a vaginal approach with detachment of the posterior vaginal wall. Madrid. maximum follow-up. Postoperative Care No bandages were applied.0. In case of rectovaginal fistulas. depending on the growth of the cultured cells (Table 1). In certain cases. The rectal mucosa had been damaged by Crohn’s disease and was extremely fragile. Follow-Up Schedule Weekly follow-up was scheduled for eight weeks after surgery. in cases of enterocutaneous fistula. tracks were filled with fibrin glue. When accessory tracks were detected. Presurgical Procedure In the case of enterocutaneous fistulas. an advancement vaginal flap was constructed. In the cases of rectovaginal and perianal fistulas. and then the skin was sutured. in an attempt to improve the obturation of the fistulas’ tracts. 22 months). and evaluated. we injected cells into the rectal mucosa. 1). 12 months. Spain) before combination of the kit’s two components. The number of injected cells ranged from 3 to 30 × 106. sectioned. one year after the first implant (implant #1). No. The specimens were embedded in paraffin. there was a monthly follow-up (minimum follow-up. during the surgical procedure associated with implant #6. One to three days after surgery. Injection of Cells Using a needle. In all cases. Histopathologic Analysis Two histopathologic samples were obtained. . In the case of perianal fistulas. One specimen (Patient 002) was obtained from the area of an enterocutaneous fistula (7 months after implant #2 and 10 days after implant #3). they were filled with fibrin glue. Liquid intake was initiated 12 hours after the procedure and solid food 6 hours after. The other specimen (Patient 001) was obtained from the rectovaginal wall. Cells were resuspended at between 3 and 30 × 106 cells/ml in 1 to 2 ml of Ringer lactate solution and put in a suitable syringe. Termination of the Surgical Procedure In the case of enterocutaneous fistulas. The time from the beginning of preparation of the inoculum to the end of the injection was less than 90 minutes in all cases. In the case of rectovaginal fistulas. 30 months. After eight weeks.Vol. one-half of the cells were resuspended in the thrombin component of a fibrin glue kit (Tissucol® Duo 2. The gap was completely separated and the rectal opening was closed with 3/0 absorbable stitches. as previously described. the main track was cored out and the rectal hole was closed with 3/0 absorbable stitches through the sclerotic mucosa. Figure 1. we observed a fluid-filled blister on the area of the injection after the injection (Fig. close to the sutured internal opening. Patients were considered healed when a total epithelialization of the external opening was evident after eight weeks. stained with hematoxylin and eosin.

Cell numbers ± standard deviations are shown in terms of absorbance at 595 nm. which did not differ significantly from the population-doubling time of human synovial fibroblasts cultured under the same conditions (35. Eight fistulas were considered suitable for retention in the study and followed for at least eight weeks (Table 1). and CD34. Table 1). results from 3 independent experiments).1420 ´A-OLMO ET AL GARCI Dis Colon Rectum. were only detected at Passages 1 through 3 (7 and 12 percent immunopositive cells. which also is found on the surface of endothelial cells. a marker of mesenchymal cytoskeletal cells (Table 2). with 99 percent immunopositive cells from Passage 4 onward (Fig. An enterocutaneous fistula in Patient 002 was eliminated from the study because of emergency abdominal surgery for a new enterovesicular fistula that had resulted in acute sepsis.6 ± 1. t-test. The marker of endothelial cells. 4). RESULTS Five patients were included in the study and seven liposuctions were performed (Table 1). as reported by the patients (25 percent. The laparotomy required the resection of the implant area. Levels of other markers ten-day intervals. 2.6 ± 0. By contrast. The mean populationdoubling time at these concentrations of serum was 37. and 10 percent. ␣-actin. In some cases. as indicated). There also was no direct relationship between the patient’s gender or age and healing. we could not adhere to the minimum eight-week follow-up schedule in this case. a marker of cell proliferation. Viable adipose tissue-derived stromal cells were successfully isolated and cultured from all seven lipoaspirates (LPAs). Two histopathologic samples were obtained seven months (enterocutaneous fistula) and one year (rec- .05. Surgical and implantation procedures were performed without additional technical difficulty in all nine treated fistulas. The fibroblast marker CD90. July 2005 Figure 2. Data are from a representative experiment with triplicate wells. No expression of the neuroectodermal marker S100 or the ectodermal marker keratin was observed in any of the LPA-derived cells at any time. allergic reactions) were observed in any of the cases studied. with maximal proliferation between 5 and 10 percent FBS (Fig.6 hours. not healed. For example.5. with time. Fig.5. increased with time. the expression of c-Kit (CD117). Patient 003 was eliminated from the trial during the implant procedure as a result of the discovery of contamination by gram-positive bacteria of the cultured lipoaspirated cells. cells were cryopreserved and thawed before implantation. The growth rate of adipose tissuederived stem cells (ADSC) depended on the serum concentration. von Willebrand factor (Factor VIII). Human synovial fibroblasts were cultured in the presence of 5 or 10 percent FBS. The bacteria were identified as Oerkovia xanthineolytica. however. At Passage 1. P > 0. There was no direct relationship between the number of cells injected or culture time and success of the procedure. Expression of vimentin was maintained at the same level up to and including Passage 9. which was found in 17 percent of LPA-derived cells at Passage 1 was no longer detectable at Passage 7. In six fistulas. anaphylaxis. the external opening had epithelialized by Week 8 and these fistulas were considered healed (75 percent.4 hours. 3). respectively).g. Nine fistulas from four patients were inoculated with ADSC after three or fewer passages (Table 1). These cells were grown in culture and passaged at seven. with a decrease in output flow. No immediate adverse reactions (e. initially expressed in approximately 80 percent of LPA-derived cells. Therefore. a high percentage (90–95 percent) of ADSC expressed vimentin.. The other two had only incomplete closure of the external opening. 2). was found in 99 percent of cells from Passage 6 (Table 1). Growth curves of lipoaspirate-derived cells at different concentration of fetal bovine serum (FBS) (0.

and we observed a complete healing in six of eight procedures. Cells from Patients 001 were Passage 6 cells subsequent to implant #6. No. Indirect immunofluorescence characterization of adipose tissue-derived stem cells. ± = 6–15%. all liposuction procedures yielded a clinically useful number of cells with characteristics of stem cells. We chose adipose tissue as the source of stem cells because of their capacity for myogenic differentiation5. a longer follow-up and MRI evaluation would be necessary to prove the longevity of this healing. liposuction fat is available in large quantities and can be harvested with minimal adverse effects on the patient.6-diamidino-2-phenylindole (DAPI)-stained nuclei. + = 16–50%. fibroblasts) c-Kit (stem cell marker) Factor VIII (differentiation marker) Alpha-actin (marker of vascular smooth muscle cells) Vimentin (marker of mesoderm origin. B. We followed our patients according to the program scheduled. ++ = 51–85%. Results of the Characterization of Cells by Immunofluorescence Staining Antigen P1 + + + + + +++ + − +/− P3 ± ++ + ± + +++ − − − P4 − ++ ++ ± ± +++ − − − P6 − +++ +++ − − +++ − − − P7 − +++ +++ − − +++ − − − 1421 P9 − +++ +++ − − +++ − − − CD 34 (stem cell marker) CD 90 (stem cell marker. vimentin. The frequency of immunopositive cells is indicated as follows: − = <5%. mesenchymal cells) Desmin (marker of mesoderm origin. Figure 3.9 In our study. Probably. muscle cells) S-100 (marker of neuroectoderm origin) Keratin (marker of ectoderm origin) P = passage number.Vol. tovaginal fistula) after surgery.6 and the fact that fistulas respond well to muscle transplants. 7 STEM CELLS IN CROHN’S FISTULA Table 2. c-Kit. DISCUSSION In a previous report. CD90. No cytologic transformation was detected in a complete series of histopathologic sections. such as those with a myocardial infarct.4 Moreover. and C. we designed the present Phase I clinical trial to evaluate the feasibility and safety of autologous adipose stromal stem cell transplantation for the treatment of unresponsive Crohn fistulas. 48. a cell-mobilization procedure is required that can be dangerous to some patients.7 Thus.3. Blue color indicates 4Ј. Other groups have used bone marrow-derived8 stem cells but. we described the successful cell-based treatment of a young female with a recurrent rectovaginal fistula that had been unresponsive to medical treatment. and +++ = 86–100%. A. in such cases. It is important to note that Crohn’s disease provides the worst conditions for a surgical approach to fistulas because of the fragility of the tissue and the enormous .

The number of patients included and the uncontrolled nature of Phase I clinical trials do not allow demonstration of the effectiveness of the treatment.10 Stem cells might differentiate into connective. stomatherapist nurse (La Paz University Hospital-Madrid-Spain). or scar tissue. Fistula before (A) and eight weeks after (B) injection of cells. however. we are now starting a Phase II clinical trial that will allow us to evaluate adequately the efficacy of this stem cell-based therapy in the treatment of complex fistulas.A. because patients underwent a combined therapy and it might be possible that the stem cells were not responsible for the success achieved. We observed a complete healing in 75 percent of cases using our treatment. CONCLUSIONS To our knowledge. American Gastroenterological Association medical position statement: perianal Crohn’s disease. Department of Pharmacology (U.M.1422 ´A-OLMO ET AL GARCI Dis Colon Rectum. F. our treatment seemed to be effective.. Our study shows that such cells are safe for the treatment of fistulas in Crohn’s disease.11 However. ACKNOWLEDGMENTS The authors thank Dolores Garcia-Olmo. these results should be taken with the pertinent care. In this way. The biologic mechanism that underlies the therapeutic success of ADSC transplantation is unknown. (Albacete General Hospital-Spain) and Paloma De-La-Quintana. problems associated with healing in these patients. July 2005 Figure 4. but we have no way of distinguishing transplanted from local connective-tissue cells. for their useful collaboration during the development of this study. REFERENCES 1. and C. Ph. We saw typical scar tissue in the histopathologically examined fistulas. Alternatively. American Gastroenterological Association Clinical Practice Committee. for providing us with the room and part of the material to perform some of the indirect immunofluorescence assays.). Moreover. However. Nevertheless. Chief of the Research Unit. Gastroenterology 2003. . secretion of growth factors by the stem cells might facilitate wound healing. Our patients were chosen because they had been unresponsive to medical treatment and at least two previous surgical procedures. no ethical conflicts were identified by our ethics committee because the cells were autologous. muscle.125:1503–7. Sánchez-Ferrer. new outbreaks of Crohn’s disease may still produce new fistulas in any given patient that will need to be treated again using the cryopreserved autologous cells from that patient. this is the first clinical trial using autologous stem cells obtained from a lipoaspirate. the results of the present study encourage to perform further studies.D.

Eur J Surg 2000. Lancet 2004. Acta Gastroenterol Belg 2001. Zuk PA. gut or liver cells? N Engl J Med 2002.7:211–28. The vulture and stem cells. Lorenz HP. Gomez-Garc´ ıa L. Inflamm Bowel Dis 2002.22:560–7. D’Hoore A.166:218–22. 7 STEM CELLS IN CROHN’S FISTULA 1423 2. Mizuno H. Hedrick MH. Zhu M. No. Garc´ ıa-Olmo D. 7.8:106–11.64: 223–6. Garc´ ıa-Arranz M. Wexner SD.349:1480–1. Myogenic differentiation by human processed lipoaspirate cells. Zhu M. Can human hematopoietic stem cells become skin. . 4. Mizuno H. et al.346: 770–2.109: 199–209. 10. Levy C. N Engl J Med 2003. et al. Nat Biotechnol 2004. Nogueras JJ. Garc´ ıa-Olmo MA. Garc´ ıa-Olmo D. 8. Management of internal fistulas in Crohn’s disease. Benhaim P. 9. Matsubara H. Cowan CM. Adipose-derived adult stromal cells heal critical-size mouse calvarial defects. Multilineage cells from human adipose tissue: implications for cell-based therapies. 6. Risk of the coronary arteries of intracoronary stem cells infusion G-CSF cytokine therapy. Penninckc F. Advancement flap plasty for the closure of anal and recto-vaginal fistulas in Crohn’s disease. 5.363:746–7. 3. 11.18: 451–4.Vol. et al. Rius J. Tissue Eng 2001. Plast Reconstr Surg 2002. Nessim A. Zuk PA. Filez L. Shi YY. 48. Autologous stem-cell transplantation for treatment of recto-vaginal fistula in perianal Crohn’s disease: a new cell-based therapy. Int J Colorectal Dis 2003. Tremaine WJ. Aalami OO. Gracilis transposition in complicated perianal fistula and unhealed perineal wounds in Crohn’s disease. Abkowitz JL.