Long-term Safety and Efficacy of Local Microinjection Combining Autologous Microfat and Adipose-Derived Stromal Vascular Fraction for the Treatment of Refractory Perianal Fistula in Crohn’s Disease

See editorial on page 2128. See editorial on page 2128. Perianal fistulas are frequent and morbid Crohn's disease (CD) complications that critically affect patient quality of life.1Kotze P.G. et al.Gut. 2018; 67: 1181-1194Crossref PubMed Scopus (108) Google Scholar Despite the best available therapies, durable remission in complex perianal fistulas is achieved in only 37% of patients,2Molendijk I. et al.Inflamm Bowel Dis. 2014; 20: 2002-2028Crossref Scopus (99) Google Scholar and recurrence and postoperative anal incontinence remain frequent. Recently, cell therapies using expanded mesenchymal stem cells (MSCs) have shown promising effects.1Kotze P.G. et al.Gut. 2018; 67: 1181-1194Crossref PubMed Scopus (108) Google Scholar The adipose-derived stromal vascular fraction (ADSVF) is an easily accessible source of cells with angiogenic, immunomodulatory, and regenerative properties,3Bora P. et al.Stem Cell Res Ther. 2017; 8: 145Crossref PubMed Scopus (321) Google Scholar but remains unevaluated in this context. We hypothesized that ADSVF combined with the trophic and volumizing effects of microfat grafting4Nguyen P.S. et al.J Plast Reconstr Aesthet Surg. 2012; 65: 1692-1699Abstract Full Text Full Text PDF PubMed Scopus (98) Google Scholar could promote tissue healing. We reported the first clinical trial evaluating the injection of autologous ADSVF and microfat in the management of complex CD perianal fistulas. Patients were aged older than 18 years and diagnosed with complex perianal fistula associated with CD for at least 6 months with controlled luminal disease (Crohn's Disease Activity Index score <220). Fistula(s) were refractory to conventional treatment, defined as failure of at least 1 surgical treatment (including drainage and seton placement) associated with anti-TNF therapy. Exclusion criteria included anal abscess, stenosis, severe proctitis, and body mass index <18 kg/m2. The patients were not excluded based on the number of fistula tracts. Patients were subjected to a first examination under anesthesia with drainage by seton placement, if indicated. The procedure was performed 1 week later and required 2 procedures on the same day under anesthesia.5Philandrianos C. et al.Stem Cell Res Ther. 2018; 9: 4Crossref PubMed Scopus (20) Google Scholar First, lipoaspiration was performed to acquire ADSVF using Celution 800/CRS (Cytori Therapeutics, San Diego, CA). The second surgery involved microfat harvesting with a closed-circuit filtration system (Puregraft, Solana Beach, CA), preparation of fistula (seton removal, curettage of fistula tracts and suture of the internal openings) and the immediate injection of both microfat and ADSVF into the wall of the fistula and surrounding tissues. Patient characteristics and cell therapy procedures are detailed in the Supplementary Material. Patients were monitored for safety at baseline and at 1, 2, 6, 12, 16, and 48 weeks after injection. Efficacy analysis included clinical evaluation of fistula, evaluation of disease activity by Perianal Crohn's Disease Activity Index/Crohn's Disease Activity Index scores, and assessment of quality of life by the Small Inflammatory Bowel Disease Questionnaire. Combined remission was defined as a complete cessation of suppuration of all the external openings confirmed via magnetic resonance imaging assessment (absence of collections >2 cm in 3 axes) at weeks 12 and 48. Clinical response was defined as an evident decrease in the suppuration score. Ten patients with a median age of 36 years (range, 19–63 years) were treated (Supplementary Table 1). Median CD and fistula durations were 8 years (range, 2–38 years) and 4 years (range, 1–15 years) respectively. Patients presented 1 (n = 4), 2 (n = 5), or 3 (n = 1) fistula tracts associated with 1 (n = 4), 2 (n = 3), or more than 2 (n = 3) external openings. Median of 22.8 million (range, 10.9–47.8 million) viable nucleated cells was injected, including 38.4% (range, 22.6%–50.2%) MSCs and 6.4% (range, 3.9%–15.7%) endothelial cells. Median quantity of microfat injected was 10.8 mL (range, 5–17 mL). Three serious adverse events occurred: 2 flares and 1 new fistula tract. They were moderate in intensity (grade 2), but categorized as serious adverse events because of the need for patient hospitalization. Non-serious adverse events were moderate pain at the lipoaspiration site (40%) that resolved in less than one week under simple oral analgesia and a cutaneous reaction secondary to anesthetics. No case of post-treatment incontinence was described. Blood inflammatory parameters (C-reactive protein, fibrinogen, white blood count) were unchanged compared with baseline. Seventy percent of patients had a clinical response at week 12, and 80% at week 48; 20% and 60% of patients achieved combined remission at weeks 12 and 48, respectively. In addition, 13 of 17 treated external openings showed complete re-epithelialization or an absence of drainage at week 48. Two patients showed complete occlusion of fistula tracts on magnetic resonance imaging at week 48 (Figure 1A). A significant improvement in perianal disease severity was observed with a decreased PDAI score (7.3 at baseline, 3.8 at week 12 and 3.4 at week 48; P = .002) and an increased quality of life score (P = .038) (Figure 1B). We provide some of the first evidence that injection of the ADSVF and microfat is a feasible and safe treatment for CD perianal fistula.6Garcia-Olmo D. et al.Int J Colorectal Dis. 2009; 24: 27-30Crossref PubMed Scopus (131) Google Scholar The serious adverse events were typical and possibly related to the natural course of disease, although the impact of the treatment cannot be fully excluded, particularly for the occurrence of the new fistula tract. The fistula improvement is noteworthy, as combined remission was observed in 60% of patients. This result is comparable to that reported for cultured MSCs.7Panes J. et al.Gastroenterology. 2018; 154: 1334-1342Abstract Full Text Full Text PDF PubMed Scopus (300) Google Scholar The surgical procedure may play a key role in this therapeutic approach. However, Panes et al7Panes J. et al.Gastroenterology. 2018; 154: 1334-1342Abstract Full Text Full Text PDF PubMed Scopus (300) Google Scholar reported combined remission in 34% of patients in the "placebo group" (surgical procedure alone), whereas a significantly higher proportion of patients achieved combined remission after MSC injection (50%). These data suggest an additional effect of cells compared with surgical treatment alone. As usually observed in phase 1 clinical trials, the procedure used heterogeneous doses of ADSVF cells recovered from a standardized lipoaspirate volume. Future studies should investigate the relationship between cell doses and healing effects. The proposed procedure constitutes a promising alternative to more complex strategies requiring ex vivo expansion of stem cells and/or scaffolding materials.8Dietz A.B. et al.Gastroenterology. 2017; 153: 59-62Abstract Full Text Full Text PDF PubMed Scopus (134) Google Scholar Preclinical models have documented the superiority of ADSVF over purified stem cells for tissue repair, owing to the heterogeneity of cells in the ADSVF.3Bora P. et al.Stem Cell Res Ther. 2017; 8: 145Crossref PubMed Scopus (321) Google Scholar The ADSVF also improves the efficiency of fat autografts used in soft-tissue defects. Reciprocally, microfat acting as a plug and filler is likely to favor the local persistence of cell activities. Therefore, our results support the healing potential of 2 minimally manipulated autologous products—the ADSVF and microfat—which act in synergy to confer regenerative and volumizing effects. While our findings should not be overinterpreted as a proof of benefit because of the low number of patients and the uncontrolled nature of the trial, they strongly encourage future evaluation of the efficacy of the ADSVF and microfat in placebo-controlled trials. Such trials will shed light on whether this strategy represents an easily accessible cell-based therapy with a favorable benefit–cost profile for patients with refractory fistulizing CD. The authors are grateful to Julie Veran and the technical team of the Assistance Publique Hôpitaux de Marseille cell therapy unit who supervised the manufacture of cell products; to the staff of the hematology laboratory of Assistance Publique Hôpitaux de Marseille (Prof Françoise Dignat-George) for support in quality-control assays; to Dr Jeremy Magalon who contributed to the interpretation of the data and critical revision of the manuscript; to Dr Caroline Prost and Jerôme Soussan who reviewed all patients' magnetic resonance imaging examinations in a blinded fashion; to Elisabeth Jouve who provided methodological assistance in the study design and performed the statistical analysis; and to Dr Pierre-Charles Orsoni and Prof Dominique Casanova for sharing their expertise in the surgical aspects of treatment. We also acknowledge Prof Guy Magalon, who participated in the design of the trial and promotes innovative cell therapy within our university hospital. Author contributions: Study concept and design: JCG and FS. Collection and acquisition of data: JCG and MS. Analysis and interpretation of data: MS and FG. Management of patients' surgical treatment: CV and CP. Drafting of the manuscript: MS and FG. Critical revision of the manuscript for important intellectual content: FS and JCG. Study supervision: FS and JCG. This is an open-label, single-arm prospective phase 1 clinical trial conducted at 1 site from December 2015 to March 2018 (ie, to the end of the 48-week follow-up). The study protocol was approved by the local Ethics Committee and the French Drug Agency (EudraCT number: 201325) and was registered at ClinicalTrials.gov (NCT02520843). Written informed consent was obtained from all patients before enrollment. The study was performed in accordance with the Declaration of Helsinki and current Good Clinical Practices. An independent data monitoring committee ensured the integrity of the trial and safety of participants. Eligible patients were aged older than 18 years and diagnosed with perianal fistula associated with CD for at least 6 months. They presented with a controlled luminal disease (Crohn's Disease Activity Index score <220) and fistulas categorized as complex according to the American Gastroenterological Association definition. Fistulas were refractory to conventional treatment, defined as failure of at least 1 surgical treatment (including drainage and seton placement) associated with anti-TNF therapy (failure of anti-TNF was defined as no response 12 weeks after induction treatment was initiated or after 12 weeks of maintenance treatment with a stable dose). Exclusion criteria included activated severe Crohn's disease (Crohn's Disease Activity Index score >220); no previous treatment for perianal fistulizing Crohn's disease; anal abscess or stenosis or severe proctitis; infectious diseases, including hepatitis B virus, hepatitis C virus, syphilis, human T-lymphotropic retrovirus or human immunodeficiency virus infection; active tuberculosis; signs of septicemia; introduction of new treatment within 3 months before study treatment; steroid treatment or having been treated with steroids within the month before the study; history of malignancy during the past 5 years; contraindication to magnetic resonance imaging and to the anesthetic or surgical procedure; body mass index <18 kg/m2. Pregnant women or women who were breastfeeding were also excluded. Eligible patients were enrolled based on the inclusion and exclusion criteria described and underwent a fistula preparation visit, including examination under anesthesia with drainage by seton placement if clinically indicated at least 1 week before administration of the experimental product. These procedures were performed to ensure homogeneity of the baseline characteristics of the study population. If a seton was placed, it had to be withdrawn immediately before administration of the experimental product. The patients and disease characteristics at baseline are detailed in Supplementary Table 1. The cell therapy procedure required 2 separate operations, including 1 surgical procedure for fat harvesting and another surgical procedure for microfat and SVF injection. A liposuction procedure using the tumescent technique with manual aspiration under local anesthesia of harvesting areas and sedation, if needed, was performed. Harvesting areas were infiltrated with a solution of physiological saline and xylocaine/adrenaline to allow tumescence of fat with minimal blood loss. Harvesting was performed under sterile conditions with a 10-mL syringe in a closed circuit using a Khoury cannula with 12 lateral 2.5 × 1.5 mm openings with a 2-way non-return valve, sterile tubing, and a 250-mL collection bag. Once collected, adipose tissue was immediately processed at the authorized Cell Therapy Unit by standard operating procedures in accordance with current Good Manufacturing Practices. The ADSVF was obtained within 4 hours of lipoaspiration. Collected lipoaspirate was transferred into the tissue collection chamber of the Celution800/CRS system (Cytori Therapeutics), washed to remove free blood and lipid, and then enzymatically digested using Good Manufacturing Practices–grade reagents to release the ADSVF of the lipoaspirate. The ADSVF was then concentrated by short centrifugation and automated wash cycles. At the end of the automated Celution process, the ADSVF was aseptically recovered and resuspended in 5 mL Lactate Ringer's solution. The final cell suspension was transferred into a syringe (10 mL) for injection. The remaining volume was used for sterility testing and biological characterization. Biological characterization of the ADSVF included standardized regulatory quality-control assays required for qualification of the cell-based product. Total viable nucleated cell recovery and viability percentage were determined using a Nucleocounter NC100 (ChemoMetec, Lillerød, Denmark). The various cell subsets of the ADSVF were identified using flow cytometry analysis (Navios instrument; Beckman Coulter, Brea, CA) with a panel of cell surface makers, in accordance with the International Federation for Adipose Therapeutics and Science and the International Society for Cellular Therapy recommendations. The markers CD45, CD34, CD90, CD146, and CD14 were used in combination with DRAQ5 and 4′,6-diamidino-2-phenylindole to exclude debris, red blood cells, and dead cells. The frequency of adipose-derived mesenchymal-like stem cells was estimated using a colony-forming unit–fibroblastic clonogenic assay (Supplementary Table 2). The final product contained 10 million to 60 million viable nucleated cells. Administration of the investigational treatment was performed under general anesthesia. The harvesting of microfat was performed with a 2-mm st'RIM cannula (Thiebaud Medical, Margencel, France) connected to a 10-mL syringe. The harvested fat was injected directly into a closed-circuit PureGraft 50-mL system filtration pocket (PureGraft, San Diego, CA). The PureGraft system allows the purification of adipose tissue by eliminating excess fluid, the lipid phase, blood cells, and fragments through filtration by a membrane in a sterile environment. The quantity of harvested fat was approximately 50 mL, which allowed a final volume of autologous microfat of approximately 15 mL ready to inject. The cleaned microfat was directly recovered by connecting the PureGraft system to 1-mL syringes to allow precise reinjection around the fistula. Before injection, patients first had any setons removed if present and then underwent careful curettage of the fistula tracts. The internal openings were closed with resorbable sutures. The microfat was injected using a 21-gauge cannula from the st'RIM device around the fistula to fill the adjacent tissues and collapse the fistula. The ADSVF solution was then injected using a 25-gauge needle into the wall of the fistula and surrounding tissues. Finally, the external opening was filled by periostial injection of the remaining microfat. Patients were allowed to leave the hospital after 24 hours of monitoring with a 1-week prescription of antibiotics (metronidazole and/or ciprofloxacin).Supplementary Table 1Patients and Disease Characteristics at Baseline (n = 10)CharacteristicsDataMen, n (%)6 (60)Median age, y (range)36 (19–63)Smoker, n (%) Yes3 (30) No3 (30) Past4 (40)BMI, kg/m2, median (range)22 (20–33)CD duration, y, median (range)8 (2–38)CDAI score, median (range)119 (22–421aThe inclusion of patient 9 who had a CDAI of 421 at baseline was considered a deviation from the protocol and was submitted for authorization by the study's sponsor. The high CDAI score of this patient was related to functional disorders due to secondary surgery (ileocolectomy) without luminal disease activity (normal ileocolonoscopy, small bowel magnetic resonance imaging and inflammatory parameters).)SIBDQ score, median (range)41 (25–70)C-reactive protein, mg/L, median (range)5 (2–21)Median fistulas duration, y (range)5 (1–15)Median PDAI score (range)7 (2–11)Number of fistulas tracts, n (%) 14 (40) 25 (50) >21 (10)Type of fistulas, n (%) Intersphincteric1 (10) Trans sphincteric8 (80) Suprasphincteric0 (0) Extrasphincteric1 (10)Fistula external openings, n (%) 14 (40) 23 (30) >23 (30)History of surgery for fistula, n (%)10 (100)Fistulas prior treatment failure, n (%) At least 1 anti-TNF (± IS)10 (100) 2 anti-TNF8 (80) Vedolizumab3 (30) Derivation ileostomia2 (20)CDAI, Crohn's Disease Activity Index; PDAI, Perianal Crohn's Disease Activity Index; SIBDQ, Small Inflammatory Bowel Disease Questionnaire.a The inclusion of patient 9 who had a CDAI of 421 at baseline was considered a deviation from the protocol and was submitted for authorization by the study's sponsor. The high CDAI score of this patient was related to functional disorders due to secondary surgery (ileocolectomy) without luminal disease activity (normal ileocolonoscopy, small bowel magnetic resonance imaging and inflammatory parameters). Open table in a new tab Supplementary Table 2Characteristics of Adipose-Derived Stromal Vascular Fraction (n = 10)CharacteristicsDataVolume of ADSVF injected, mL5Viability, %, median (range)83.9 (75.6–88.5)No. of viable nucleated cells injected, ×106, median (range)22.8 (10.9–47.8)SVF cells subsets, %, median (range) Leukocytes36.0 (18.8–60.3)Macrophages/monocytes14.2 (8.3–24.1)Lymphocytes11.1 (5.9–36.7)Neutrophils5.8 (2.0–12.1) Transitional cells and pericytes16.9 (9.1–27.0) Endothelial progenitor cells6.4 (3.9–15.7) Mesenchymal stem cells38.4 (22.6–50.2)CFU-F, median (range)1.6 (0.1–3.6)CFU, colony-forming units–fibroblast. Open table in a new tab CDAI, Crohn's Disease Activity Index; PDAI, Perianal Crohn's Disease Activity Index; SIBDQ, Small Inflammatory Bowel Disease Questionnaire. CFU, colony-forming units–fibroblast.