Abstract
Perianal fistulas (PAFs) represent a severe complication of Crohn’s disease (CD). Despite the advent of biologic and small-molecule therapeutics for luminal disease, PAFs in CD (CD-PAF) are relatively resistant to treatment, with less than 50% responding to any therapy. We report an injectable, biodegradable, mechanically fragmented nanofiber-hydrogel composite (mfNHC) loaded with adipose-derived stem cells (ADSCs) for the treatment of fistulas in a rat model of CD-PAF. The ADSC-loaded mfNHC results in a higher degree of healing when compared to surgical treatment of fistulas, which is a standard treatment. The volume of fistulas treated with mfNHC is decreased sixfold compared to the surgical treatment control. Molecular studies reveal that utilization of mfNHC reduced local inflammation and improved tissue regeneration. This study demonstrates that ADSC-loaded mfNHC is a promising therapy for CD-PAF, and warrants further studies to advance mfNHC toward clinical translation.
INTRODUCTION
Inflammatory bowel disease (IBD) includes Crohn’s disease (CD) and ulcerative colitis (UC) and is a complex, multifactorial, immune-mediated illness (1, 2). There are approximately 6.9 million IBD cases globally, and this number is rising (1). In 2020, there were approximately 2.5 million IBD cases in the United States, and this number is expected to increase to 3.5 million by 2030 (2). Currently, it is estimated that 780,000 patients in the United States suffer from CD, and 33,000 new cases are diagnosed each year (3). Perianal fistulas (PAFs) occur in approximately 30 to 40% of CD (4–6). PAF in CD (CD-PAF) is estimated to affect more than 15,000 new patients each year (7). CD-PAF can be associated with perianal pain, purulent or feculent drainage, frank fecal incontinence, and superimposed infections. These complications of CD-PAF are associated with reduced quality of life and morbidity. Current clinical treatment strategies aim to promote long-term fistula healing while avoiding incontinence and diverting stomas (8). These approaches, however, are successful in less than 50% of patients (6, 9–20). For example, randomized clinical trials showed that infliximab, a tumor necrosis factor–α (TNF-α) antibody that is one of the most effective treatments for luminal inflammation, induces complete closure in a mere 23% of CD-PAF (18–20). Other medical therapies, including antibiotics or immunomodulators, are even less effective in the long-term treatment of CD-PAF, with as many as 70% of fistulas relapsing upon discontinuation of treatment (15–17). Another approach in the treatment of CD-PAF is surgical. Unfortunately, surgical interventions to address PAFs in CD are marred by frequent fistula recurrence, as well as local complications such as damage to anal sphincter that can lead to fecal incontinence (11, 12).
A newer treatment of CD-PAF consists of allogeneic adipose-derived stem cells (ADSCs) injected into and around the fistula tract (6, 13, 14, 21, 22). In a phase 3 clinical trial, in addition to surgical treatment of PAFs, a single local administration of 120 million ADSCs into the fistula tract induced clinical PAF remission at week 52 in 59.2% of patients versus 41.6% achieved in the standard surgical treatment with placebo [phosphate-buffered saline (PBS)] injection arm. On the basis of these results, the ADSC treatment has recently been approved by the European Medicines Agency for the treatment of complex CD-PAF that do not respond to conventional and/or biologic therapies (13, 14). This treatment has not yet been approved in the United States. Although it has not been studied specifically for fistulas, it has been hypothesized that the beneficial effects of ADSCs at the site of injury are potentially limited by ADSC migration away from the site (23). A potential strategy to improve the efficacy of ADSCs would therefore include a methodology to retain ADSCs within the fistula tract. Another investigational approach to the treatment of PAF in CD is the utilization of bioprosthetic materials as fillers for fistula tracts (9, 24–26). As shown in a previous study (25), the approach, however, has multiple complications, including a 20% rate of plug extrusion and a 15% rate of surgical site infection. In addition, the overall healing rate was only 56%. The long-term healing rate is even lower, due to the fact that the recurrence rate is high (25, 26). In summary, the overall success of all current treatments for CD-PAF, including injection of ADSCs into fistula tract, still leaves more than 40% of CD-PAF without adequate treatment (6).
On the basis of these studies and our clinical experience, we advance that a systemic treatment targeting luminal inflammation (such as biologics or small molecules) needs to be coupled with a local treatment of CD-PAF for maximum impact on fistula tracts. The ideal local fistula treatment should be applied directly in the fistula tract to address local inflammation, as well as the physical defect that allows drainage of stool, pus, and blood. We designed and tested such a tissue repair scaffold to satisfy multiple conditions: (i) ease of application into the fistula tract, (ii) high porosity to allow regenerative host cell migration into the scaffold, (iii) substantial mechanical stiffness to retain its shape and integrity within the fistula tract, (iv) immunomodulatory properties and enhanced angiogenic responses and blood vessel in-growth into the scaffold, and (v) ability to co-deliver ADSCs and retain them within the composite for maximum ADSC effect within the fistula tract. Building on our previous work and aiming to fulfill these design criteria, we have created a hyaluronic acid (HA) hydrogel covalently linked to electrospun poly(ε-caprolactone) (PCL) nanofiber fragments forming an integrated nanofiber-hydrogel composite (NHC). This NHC is mechanically fragmented (mfNHC) and loaded with ADSCs and then tested in a clinically relevant rat model of CD-PAF.
RESULTS
Fabrication of mfNHC with tunable stiffness and pore size
mfNHC was synthesized by conjugating PCL fiber fragments with HA hydrogel as we previously described (27, 28). For ease of administration, we developed an NHC composite that can be injected into the fistula tract. We designed and synthesized mfNHC of several stiffness levels to identify the optimal material stiffness to allow injection into the fistula tract, as well as host cell infiltration into the composite. First, we tuned the composite pore size, because this mechanical feature directly affects the ability of host regenerative cells to migrate into the composite (28). HA was first reacted with glycidyl acrylate to conjugate the acrylate groups to HA to generate an NHC precursor, acrylate-modified HA (HA-Ac). Next, electrospun PCL nanofibers were prepared as described before (28). PCL fibers were then surface-activated with plasma treatment to get carboxylic groups on the surface of PCL fiber, followed by converting the carboxyl group to the thiol-reactive maleimide (MAL) group to generate MAL-functionalized PCL (MAL-PCL) fibers. The functionalized PCL fibers were cryo-milled into fragments with length of 40 to 80 μm. NHC was then fabricated by cross-linking HA-Ac, MAL-PCL fibers, and dithiol poly(ethylene glycol) (PEG-SH) at 37°C for 16 hours (Fig. 1A). MAL-PCL fiber fragments and HA-Ac were conjugated to PEG-SH cross-linked network, forming interfacial covalent bonds to generate an integrated composite structure during the gelation process (Fig. 1A). As NHC mechanical properties and pore size are determined by the cross-link network density, we tested multiple HA concentrations of 5, 10, and 15 mg/ml while keeping MAL-PCL fiber content constant. The measured shear modulus G′ of NHC stiffness at HA concentration of 5, 10, and 15 mg/ml was 100, 250, and 400 Pa (Fig. 1C). NHC was then mechanically fragmented using our previously reported method (29) by passing it through a stainless steel mesh with a uniform pore opening of 0.009 inches to produce mfNHC with an average diameter of 131.3 ± 18.8 μm. As presented in Fig. 1A, the preformed NHC was loaded in two syringes, connected with a union fitted with a stainless steel mesh inside. NHC was sieved by passing through the mesh back and forth twice. The generated mfNHC via this method has an irregular and aspherical shape but with a relatively narrow size distribution (29). The microstructure of mfNHC at different shear storage moduli G′ (100, 250, or 400 Pa) was analyzed by scanning electron microscopy (SEM). Representative images are shown in Fig. 1B. The SEM image displays the PCL fiber–HA hydrogel network with fibrillary microarchitecture, with various pore sizes as a function of the NHC stiffness. The average pore size was 0.12 ± 0.08 mm2 for 100-Pa NHC (NHC-100), 0.06 ± 0.02 mm2 for 250-Pa NHC (NHC-250), and 0.03 ± 0.01 mm2 for 400-Pa NHC (NHC-400; Fig. 1D). As expected, the pore size of NHC decreased with increasing stiffness. As the concentration of HA increased in NHC, the internal microstructure became more compact, and the pore size decreased. NHC-250 and NHC-400 showed more uniform size distribution compared to NHC-100.
