Effective gene transfer with sustained gene expression is an important adjunct to the study of intestinal inflammation and future therapy in inflammatory bowel disease. Recombinant adeno-associated virus (AAV) vectors are ideal for gene transfer and long-term transgene expression. The purpose of our study was to identify optimal AAV pseudotypes for transduction of the epithelium in the small intestine and colon, which could be used for studies in experimental colitis. The tropism and transduction efficiencies of AAV pseudotypes 1–10 were examined in murine small intestine and colon 8 wk after administration by real-time PCR and immunohistochemistry. The clinical and histopathological effects of IL-10-mediated intestinal transduction delivered by AAVrh10 were examined in the murine IL-10−/− enterocolitis model. Serum IL-10 levels and IL-10 expression were followed by ELISA and real-time PCR, respectively. AAV pseudotypes 4, 7, 8, 9, and 10 demonstrated optimal intestinal transduction. Transgene expression was sustained 8 wk after administration and was frequently observed in enteroendocrine cells. Long-term IL-10 gene expression and serum IL-10 levels were observed following AAV transduction in an IL-10−/− model of enterocolitis. Animals treated with AAVrh10-IL-10 had lower disease activity index scores, higher colon weight-to-length ratios, and lower microscopic inflammation scores. This study identifies novel AAV pseudotypes with small intestine and colon tropism and sustained transgene expression capable of modulating mucosal inflammation in a murine model of enterocolitis.
- small intestine
gene targeting to the gastrointestinal tract is vital for the study of intestinal disorders and for the development of therapeutics. The challenges of deciphering the mechanisms involved in the pathogenesis of inflammatory bowel disease (IBD) and the difficulties of treatment warrant the study of intestinal gene targeting and endogenous protein expression. Compartmentalized immune responses and the evolution of small interfering RNA (siRNA) and microRNA in the study of mucosal inflammation would greatly benefit from sustained transgene expression in the intestines (7, 21, 24, 47, 52). This concept is limited by the technique of gene transfer to the intestines in which many viral vector and nonviral strategies have been studied with limited success (3, 6, 8, 11, 22, 26, 40, 43, 44, 46, 48). Three key concepts should be followed when selecting a viral gene delivery system: long-term expression, lack of toxicity, and many effective serotypes. Recombinant adeno-associated viral (AAV) vectors, a single-stranded DNA virus, fulfill these criteria and have increased AAV's potential as a delivery vehicle for gene transfer applications(9). Several studies in small and large animals have demonstrated their long-lasting and therapeutic effects (19, 34, 45). Further work must be done to investigate AAV gene transfer to the intestines.
We previously reported the transduction efficiencies of AAV serotypes 1, 2, and 5 in the small intestine (SI) and colon of mice in vivo were limited (36). Recently, a biodistribution analysis of neonatal mice systemically injected with AAV reported intestinal transduction, but did not differentiate between small bowel and colon (20). However, neonatal mice are a challenging model to use in the study of intestinal inflammation. The development of self-complementary AAV (scAAV) vectors to bypass the limiting aspects of second-strand synthesis (30) will help to identify serotypes with gut tropism. The rationale of scAAV vector is to increase the biological efficiency of the vector by rapidly forming transcriptionally competent double-stranded DNA. In addition, the discovery of new AAV serotypes (14, 15) will lead to vectors with improved intestinal tropism.
The goal of this study was to identify a serotype that can achieve high transduction efficiency in SI and colonic tissues and is capable of modulating mucosal inflammation in vivo in an experimental model of IBD. We examine tropism and transduction efficiencies of AAV serotypes (pseudotypes) 1–9 and rh10 using a scAAV-encoding enhanced green fluorescence protein (GFP) in an AAV2 genome in a mouse model. Three different routes of vector delivery are examined because of limitations in the techniques of administration (36, 42). An optimal vector and route of delivery, in this case AAVrh10 (AAV10) via superior mesenteric artery (SMA) injection, was identified and then used to examine the clinical and histopathological effects of intestinal IL-10 expression in the IL-10−/− model of enterocolitis. The IL-10−/− model was used as a clear model in which to examine biology of murine IL-10 (mIL-10) delivered by AAV10 transduction.
MATERIALS AND METHODS
The Animal Care Services Facility and the Institutional Animal Care and Use Committee at the University of Florida, Gainesville, FL, approved all procedures performed on mice. Healthy 6- to 8-wk-old male BALB/cJ mice (Jackson Laboratory, Bar Harbor, ME) were used for the AAV serotype intestinal tropism experiments. Eight- to 10-wk-old male homozygous IL-10 gene deficient mice on a C57BL/6 background (Jackson Laboratory) were used for the IL-10 gene transfer experiments. All mice were housed in specific pathogen-free facilities, maintained throughout on a 12-h dark-light cycle, and provided food pellets and water ad libitum. Mice were transferred to conventional housing conditions following surgical procedures.
Plasmids and AAV Vectors
The Hpa-Trs-SK plasmid encoding enhanced GFP with a cytomegalovirus (CMV) immediate early promoter and one mutated terminal repeat was a kind gift from Doug McCarty (29, 30). The scAAV encoding GFP used in the intestinal tropism experiments were packaged similar to methods previously described (51). Briefly, vectors were prepared by transfection of HEK 293 cells with pHpa-trsSK and appropriate helper plasmid by using liposomes to generate capsids pseudotyped 1 through 10. All vectors were purified on an iodixanol density gradient. Vector genome titers were determined by dot-blot hybridization. The pCBIL-10 plasmid was generated by using mIL-10 cDNA cloned into the p43.2 plasmid with AAV2 inverted terminal repeats and a fused CMV immediate early enhanced chicken β-actin promoter (16). pCBIL-10 was then used to generate recombinant single-stranded AAV2/10-IL-10 (AAV10-IL-10) by the methods described above.
Three different methods of vector administration were utilized to assess intestinal tropism: namely small bowel gavage, SMA injection, and enema. For each administration technique group, experiments were performed in triplicate with each pseudotype and control for a total of 33 mice per group.
Small bowel gavage.
Mice were fasted overnight without bedding, and had access to water containing 20 mM N-acetyl-l-cysteine (NAC) (Sigma-Aldrich, St. Louis, MO). NAC was used as a mucolytic agent for the intestinal mucosal surface (38). To further assist with intestinal crypt mucus extrusion, pilocarpine, 30 mg/kg (Sigma-Aldrich), was given by intraperitoneal (IP) injection 45 min prior to vector administration (38). Mice were anesthetized with inhaled isoflurane and laparotomy was performed. A silk tie was temporarily placed around the third portion of the duodenum. The proximal small bowel was lavaged with 500 μl of 20 mM NAC via a tuberculin syringe. After 5 min 5 × 1010 physical particles of scAAV-GFP in 1 ml of 1× phosphate-buffered saline (PBS) was injected into the proximal bowel. After another 5 min the proximal bowel tie was removed, the incision was closed, and the mouse was recovered with appropriate analgesics.
These injections were performed as previously described (37). Briefly, 5 × 1010 physical particles of scAAV-GFP in 200 μl of PBS were injected into the SMA via a 250-μl gas-tight syringe (Hamilton; Reno, NV).
Mice were fasted overnight, pretreated with NAC and pilocarpine as described above, and anesthetized with inhaled isoflurane. The colon was washed with intrarectal injection of 300 μl of 20 mM NAC using a stainless steel 1-in. straight round-tip needle (Roboz; Gaithersburg, MD) and allowed to drain without sedation for 30 min. Mice were reanesthetized and 5 × 1010 physical particles of scAAV-GFP in 600 μl of PBS were given by enema. Appropriate enema fluid volumes were determined through preliminary experiments injecting methylene blue (Fisher Scientific; Pittsburgh, PA) intrarectally and visualization of the fluid reaching the cecum.
Histology of scAAV-GFP Tissues
Eight weeks after scAAV-GFP administration, mice were anesthetized with isoflurane and blood was drained through inferior vena cava puncture followed by normal saline infusion. Mice were killed by cardiac puncture. Organs were removed by sterile biodistribution protocol with sterile instruments used between each organ. The SI was removed, opened longitudinally, washed of debris, and sectioned into thirds (proximal, middle, and distal). Each third was sectioned into three equal longitudinal portions. One section was snap frozen and stored at −80°C for future DNA extraction. The second section was snap frozen in RNAlater (Ambion; Austin, TX) and stored at −80°C for future RNA extraction. The third longitudinal section was rolled on a sterile cotton tip applicator stick and fixed in 4% paraformaldehyde (Sigma-Aldrich), cryoprotected in 18% sucrose solution for at least 4 h, and embedded and frozen in optimal cutting temperature (OCT) compound. Tissues were further processed by theMolecular Pathology Core at the University of Florida (Gainesville, FL) and the Central Microscopy Research Facility at the University of Iowa (Iowa City, IA). Frozen sections (4- or 10-μm thickness) were mounted in Vectashield with 4,6-diamidino-2-phenylindole (DAPI; Vector Laboratories; Burlingame, CA) for direct GFP visualization.
Smooth muscle actin.
OCT slides were processed as described above. Tissues were blocked for nonspecific staining with 10% normal goat serum. The primary antibody (Abcam) was incubated on tissues at 1:500 for 1 h at room temperature. The slides were rinsed with Tris-buffered saline (TBS) buffer before incubation with goat anti-rabbit AF 555 (Molecular Probes/Invitrogen) secondary antibody at 1:500 for 45 min at room temperature. Nuclear counterstaining was performed by use of Vectashield with DAPI (Vector Laboratories, H-1200).
OCT slides were processed as described above followed by trypsin digestion for antigen retrieval. Staining proceeded using the M.O.M. Kit (Vector Laboratories) by slightly modifying the manufacturer's suggestions. Mouse IgG was blocked with the M.O.M. IgG blocking solution provided in the kit and prepared as directed. Prior to application of the primary antibody, the tissues were equilibrated with the M.O.M. diluent provided in the kit. The primary antibody (Millipore; Billerica, MA) was diluted to 1:50 in the same diluent and incubated on tissues overnight at 4°C. Tissues were brought to room temperature and rinsed with TBS buffer. Goat anti-mouse AF 594 secondary antibody (Molecular Probes/Invitrogen) was applied to tissues at 1:1,000 for 45 min at room temperature and rinsed. Nuclear counterstaining was performed by use of Vectashield with DAPI (Vector Laboratories, H-1,200).
Frozen sections were brought to room temperature in TBS buffer then permeabilized with Triton X-100 0.1% in PBS 1× pH 7.4. Tissues were rinsed then blocked in 10% normal goat sera (Sigma, G-9023) in PBS 1×. Primary antibody rabbit anti-chromogranin A (anti-CGA) (Novus Biologicals, NB120-15160, 0.2 mg/ml) was diluted 1:50 in blocking solution and incubated on tissues overnight at 4°C. Tissues were brought to room temperature and rinsed with PBS buffer. Secondary antibody goat anti-rabbit-Alexa Fluor 647 conjugated (Invitrogen, A21246, 2 mg/ml) was diluted 1:500 in blocking solution, and incubated on tissues for 1 h at room temperature, and rinsed. Nuclear counterstaining was performed by use of Vectashield with DAPI (Vector Laboratories, H-1200).
The same protocol was used as for CGA staining with primary antibody rabbit anti-GFP (Invitrogen, A11122, 2 mg/ml): 1/200 diluted in blocking solution and secondary antibody goat anti-rabbit-Alexa Fluor 568 conjugated (Invitrogen, A21069, 2 mg/ml).
Tissues were examined under a Zeiss 710 confocal microscope (Zeiss; Jena, Germany) (with the following excitations: direct GFP 488 nm; anti-smooth muscle 555 nm; anti-vimentin 594 nm; anti-CGA 561 nm; anti-GFP 633 nm; nuclei 405 nm) by utilizing a sequential method of acquisition for optimal signal separation. Postacquisition images were processed with ImageJ (NIH).
Genomic DNA (gDNA) was isolated from entire-length colon sections or entire-length distal SI sections, prepared as described above, by using a DNeasy kit (Qiagen; Valencia, CA) according to manufacturer's instructions. Total RNA was isolated from analogous colon and SI sections by using RNeasy mini kit (Qiagen) according to manufacturer's instructions. Sterile PCR experiments were conducted by the Powell Gene Therapy Center Toxicology Core at the University of Florida (Gainesville, FL). Presence of viral genome DNA was detected by real-time PCR analysis using total gDNA. Primers and probe used against the CMV promoter region of scAAV-GFP were forward 5′-CGTCTACGGTGGGAGGTCTATATTAA-3′, reverse 5′-GGATCGGTCCCGGTGTCT-3′, probe 6FAM-ACGCCATCCACGCTGTTTTGACCT-TAMRA. Presence of transgene product, GFP mRNA, was detected by reverse transcriptase PCR of total RNA isolated from tissues, followed by real-time PCR analysis. Primers and probe used against the GFP gene of scAAV-GFP were forward 5′-GTGGCGGATCTTGAAGTTCAC-3′, reverse 5′-CTACAACAGCCACAACGTCTATATCA-3′, probe 6FAM-ATGCCGTTCTTCTGCTTGTCGGCC-TAMRA. Presence of the transgene product mIL-10 was detected by reverse transcriptase PCR of total RNA isolated from tissues as described above, followed by real-time PCR analysis. Primers and probe used against the mIL-10 gene of AAV10-IL-10 were forward 5′- CAGCCG GGAAGACAATAACTG-3′, reverse 5′-CGCAGCTCTAGGAGCATGTG-3′, probe 6FAM-ACCCACTTCCCAGTCGGCCAGAG-TAMRA. Quantitative real-time PCR was performed by use of an ABI 7900HT system (Applied Biosystems; Foster City, CA). For vector genome analysis 1 μg of extracted gDNA was used in TaqMan Universal Master Mix (ABI) under the following conditions: 40 cycles of 94°C for 40 s, 37°C for 2 min, 55°C for 4 min, and 68°C for 30 s. For mRNA transcript analysis 100 ng of extracted RNA was used in One-Step RT PCR Master Mix (ABI) for the reverse transcriptase reaction followed in the same program by cDNA amplification using the following conditions: RT: 48°C for 30 min; TaqMan Activation: 95°C for 10 min; PCR: 40 cycles of 95°C for 15 s, 60°C for 1 min. DNA and RNA samples were assayed in triplicate and data was analyzed with ABI 7900HT Prism sequence detector software (ABI). The third replicate was spiked with DNA at a ratio of 100 copies/μg of gDNA. If at least 40 copies of the spike-in DNA were detected, the DNA sample was considered acceptable for reporting vector DNA copies. When the copy number of the vector DNA found in that sample was greater than 100 copies/μg, the sample was considered positive and the measured copy number/μg was reported. If fewer than 100 copies/μg were present, the sample was considered negative. If less than 1 μg of gDNA was analyzed to avoid PCR inhibitors, the spike-in copy number was reduced proportionally to maintain the 100 copies/μg DNA ratio.
Treatment of IL-10−/− Colitis
1 × 1011 physical particles of AAV10-IL-10 or control (AAV10-GFP) were administered to mice via SMA injection as described above or via portal vein injection. Portal vein injections were performed as previously described (28). Briefly, a midline incision was made and the abdominal skin and muscles were retracted back. The intestines were gently moved to the left side of the abdominal cavity, covered with a gauze sponge, and irrigated with warm saline. The portal vein was then exposed. Using a 29G tuberculin syringe, we delivered 1 × 1011 physical particles in 200 μl of PBS into the portal vein. The delivery was confirmed with visualization of the liver blanching during injection. Once hemostasis was confirmed, the intestines were bathed with warm saline and the abdominal muscles and skin were sutured. A total of eight animals were divided into the two experimental groups: the AAV10-IL-10 treatment group (n = 5) and AAV10-GFP control group (n = 3). Mice were allowed to recuperate for 2 wk postoperatively in conventional housing.
Synchronization of colitis.
Some mice developed loose stools during the recovery period. Colitis was then further triggered and synchronized with nonsteroidal anti-inflammatory administration (2). Piroxicam (Sigma-Aldrich) was pressed into standard rodent chow pellets at a dose of 200 ppm (Harlan Teklad; Madison, WI) and provided to all mice ad libitum for 2 wk. The diet was then replaced with standard rodent chow ad libitum for the remainder of the experiment.
Clinical evaluation of IL-10−/− colitis.
Colitis activity was monitored daily. Weight, stool consistency, and stool blood were measured every 3–5 days for each mouse for a total of 60 days. Individual indexes and disease activity index (DAI) scores were determined and calculated by a system previously described (33). Index scores were determined as follows: change in weight [0 (<1%), 1 (1–5%), 2 (5–10%), 3 (10–20%), 4 (>20%)]; stool consistency (0 well formed pellets, 2 pasty pellets, 4 liquid stool); and stool blood (0 negative, 2 occult positive or 4 visible bleeding). Weight change was calculated as the percent difference between the original weight and the new weight. Occult blood in mouse fecal pellets was detected by Hemoccult test kit (Beckman Coulter; Brea, CA). DAI score at each time point was determined by the average of the scores for each animal. Blood was obtained every 2 wk through tail vein bleeding for IL-10 level analyses.
Histopathological assessment of IL-10−/− colitis.
At the end of the experiment, mice were euthanized and drained of blood through inferior vena cava phlebotomy while anesthetized. Serum was separated in Statspin vials (Fisher Scientific) and stored at −80°C until further use. Small intestine, colon, and spleen were removed by sterile technique. The spleens were weighed. The intact colons were weighed and measured for a weight-to-length ratio. Small intestine and colon were opened longitudinally and washed of debris. Each colon was sectioned into three equally sized longitudinal sections along the entire length of the colon. Two sections were snap frozen in RNAlater (Ambion) and stored at −80°C for future processing; and the third section was fixed and processed for histological evaluation. The entire length of the colon was examined for histopathology. The SI was sectioned into thirds. The distal third was further sectioned into three equal longitudinal lengths and stored as described above. Tissues were fixed with 10% neutral buffered formalin and routinely processed, and 6-μm sections were stained with hematoxylin and eosin for light microscopy analysis. Slides were coded and samples were examined by a blinded gastrointestinal pathologist (L. Dixon). Inflammation was scored similar to systems previously described (1, 35). Briefly, tissues were scored for degree of mononuclear cell infiltration in the lamina propria (0–3), presence of neutrophils (0–4), submucosal to transmural involvement (0–3), number of crypt abscesses (0–1), goblet cell loss (0–2), mucosal erosion/ulcerations (0–1), and abnormal crypt architecture including epithelial hyperplasia (0–3) (maximum score 17) across two affected fields of each tissue analyzed. Severity of the inflammatory changes in the tissues was based on the sum of the scores reported for each parameter. Total scores were averaged for the two different regions and then averaged among replicates.
Cytokine assessment of IL-10−/− colitis.
Mouse serum was analyzed for the presence of mIL-10 as previously described (17) by using a bead-based technology cytokine kit (Millipore) according to manufacturer's instructions. IL-10 gene expression was quantified in full-length sections of distal SI and colon by real-time PCR as described above.
All values are presented as means ± SE. Data were analyzed on GraphPad Prism software (San Diego, CA) by analysis of variance (ANOVA) with Bonferroni posttest and Student's t-test. Pearson correlation was used to assess the linear relationships between clinical scores and IL-10 levels. Statistical significance was set at P < 0.05.
AAV Intestinal Tropism and Transduction Efficiency
AAV vector pseudotypes 1 through 10 were evaluated to determine optimal SI and colon transgene expression. Three different administration techniques were used to identify optimal delivery of vector to the intestines: 1) small bowel gavage, 2) enema, and 3) SMA injection. Small bowel gavage and enema did not yield significant SI or colonic transduction. PCR for presence of vector genome in the SI of gavage-treated mice and the colon of enema-treated mice were negative (Fig. 1). The copy numbers reported with gavage of AAV8 and AAV10 were not consistently positive to suggest efficient transduction.
SMA injection provided the best delivery method with significant SI and colon transduction. AAV pseudotypes 4, 7, 8, 9, and 10 demonstrated efficient tissue tropism in the SI and colon (Fig. 2, A and C). These same vectors also provided the best transduction efficiencies in both the SI and colon (Fig. 2, B and D). Interestingly, despite a lower level of infection, AAV1-mediated transduction efficiencies in the colon similar to the other optimal vectors. The localized gastrointestinal biodistribution of all the pseudotypes revealed contamination of the vectors outside of the SI and colon. Examination of the stomach revealed similar tropism and transduction efficiency patterns seen in the SI and colon (Fig. 3, A and B). Multiple other pseudotypes demonstrated liver tropism and transduction efficiency (Fig. 3, C and D).
The presence of vector genome and transgene expression in the SI and colon 8 wk after AAV administration would suggest the transduction of intestinal progenitor cells and/or nondividing cells [e.g., endothelial, nerve, and mesenchymal cells (myoepithelial cells, smooth muscle cells, fibroblasts and myofibroblasts)] in the intestine. Localization of transduced cells was evaluated by examination of tissues for GFP positive cells using confocal microscopy. Green fluorescent cells were found scattered in the epithelium of the SI (Fig. 4) and the colon (Fig. 5). Transduction appeared to be patchy with certain areas expressing more GFP than other. Most of the green fluorescent cells appeared to be epithelial cells with only a few rare cells with morphology suggestive of myoepithelial cells, endothelial cells, and serosal cells. Immunofluorescence did not increase the sensitivity of identifying more GFP-positive cells (Fig. 6). It was technically difficult to quantify the primary location of all cell types with transgene expression; however, qualitative data suggest more positive cells in the proximal colon. To further examine cell type tropism, tissues were also stained for smooth muscle actin and vimentin to identify pericryptal myoepithelial cells/muscle cells and myofibroblasts/fibroblasts, respectively (Fig. 7, A–D). Colocalization with GFP-expressing cells was not seen with either cell marker in AAV8-, 9-, or 10-treated tissues. Tissues were also stained for chromogranin A to identify enteroendocrine cells (EECs). Colocalization of green (GFP) and red (CGA) was seen in rare EECs in AAV10-treated SI and was not conclusive in colon (Fig. 7, E and F).
Clinical Evaluation of Experimental Colitis After AAV-IL-10 Gene Delivery
One of the optimal AAV pseudotypes was selected to examine its potential to express an immunomodulatory protein in the SI and colon in experimental colitis. To determine whether intestinal expression of transgene is an effective means of modulating mucosal inflammation we chose a clean model to examine gene delivery effects. We analyzed the effects of mIL-10 gene delivery in the IL-10−/− murine model of enterocolitis. AAV10 was selected as a vector of choice given the significant intestinal transduction efficiencies observed. AAV10 encoding mIL-10 was administered to the intestine of IL-10−/− mice via SMA injection. Given our findings of significant liver transduction following SMA delivery of vector, we also added a second arm to our experiment by administering the same AAV10-IL-10 vector to another group of IL-10−/− mice via portal vein injection. Experimental control groups were generated by administering AAV10 encoding GFP to IL-10−/− mice via SMA injection or portal vein injection. Two weeks after vector delivery, enterocolitis was synchronized with piroxicam as described in materials and methods.
The clinical course of IL-10−/− colitis was followed by monitoring animal weight changes, stool consistency, and presence of blood in stool. The DAI score was determined for each experimental group by using the three mentioned factors. Following piroxicam synchronization, the DAI was significantly decreased in both AAV10-IL-10 treatments groups compared with their respective controls (Fig. 8). Only mild differences were noted in clinical activity between SMA treated and portal treated with more time points of lower DAI scores in the SMA-treated animals.
Histopathological Evaluation of Experimental Colitis After AAV-IL-10 Gene Delivery
Gross pathology of colitis was assessed by examination of colon weight-to-length ratio 30 days after piroxicam synchronization. A heavier and/or shortened colon corresponds with worse colitis. The SMA treatment group had lower ratios or less structural effects of inflammation than the control group (Fig. 9). The SMA delivered AAV10-IL-10 group had less gross inflammation than the portal delivered AAV10-IL-10 group; however, the difference was not significant. No significant differences in colon weight-to-length ratios were seen between portal-delivered AAV10-IL-10 and its control.
Microscopic histological examination was performed by a gastrointestinal pathologist blinded to treatment groups and analysis revealed moderate to severe inflammation of the colonic mucosa and submucosa (Fig. 10). Features included epithelial hyperplasia with chronic inflammatory cell infiltration, patchy neutrophil infiltration with crypt abscesses, areas of transmural inflammation, and erosions. Inflammation was mild in the distal ileum. The inflammatory features were scored as outlined in materials and methods. Microscopic inflammation scores were significantly less in the both SMA and portal AAV10-IL-10 treatment groups compared with their respective controls (Fig. 11). Inflammation was not significantly different in the distal ileum among all groups. A trend was noted toward less proximal colon inflammation in the SMA AAV10-IL-10 treatment group compared with the portal AAV10-IL-10 treatment group.
Overall, transgene derived IL-10 treatment by all administration routes showed significantly improved clinical response and less inflammation compared with control groups.
Long-Term IL-10 Transduction Efficiency
Serum IL-10 was measured in all mice throughout the experiment at different time points by serial tail vein blood draws. Serum IL-10 levels peaked 2 wk after AAV10-IL-10 administration with high levels in both SMA and portal delivery groups (Fig. 12). IL-10 levels decreased but persisted even after 8 wk. IL-10 levels in the portal delivery group showed more variability in IL-10 levels compared with the SMA delivery group. Mean IL-10 levels in AAV10-GFP treatment groups were below detection limits.
Tissue transduction efficiency was also determined by real-time PCR for IL-10 genome from RNA extracted from SI, colon, and liver in each treatment group (Fig. 13). IL-10 copy number was greater in the colon (entire length section) compared with the distal SI (entire length section) in both AAV10-IL-10 treatment groups. IL-10 copy numbers were similar between SMA and portal AAV10-IL-10 administration groups, which corresponds to the similar serum IL-10 levels. Liver IL-10 copy numbers were greatest and correspond to the findings from the initial scAAV-GFP serotype tropism experiments.
We have identified several new AAV serotypes that provide intestinal tropism with sustained transgene expression in mice. We also demonstrate amelioration of colitis in the IL-10-null model of IBD, following AAV-mediated delivery of IL-10 via SMA injection, intended to direct gene therapy to the intestine, compared with portal vein delivery. Taken together, these data suggest that distinct AAV serotypes provide improved gut transduction and can allow the study of the physiology of intestinal inflammation.
The directed vascular route of vector delivery is again superior to luminal routes of administration (36, 41). Attempts to bypass the gastric contents or modify the mucus layer of the intestines did not improve the infectious capability of AAV. Certainly, multiple other factors that are part of the intestinal mucosal barrier could also be limiting vector infection (12). In addition, the capsid characteristics of AAV vectors thus far favor basolateral binding over apical binding (10). Thus, until the development of modified capsids that can bind to the apical or luminal aspect of the intestinal epithelium, the vascular route is the preferred mode of delivery for AAV vectors.
These experiments did not examine the IP injection technique because this was inferior compared with the SMA injection method in our previous work to examine AAV gut transduction, albeit with different vectors (36). The SMA injection method was chosen to facilitate identification of a gut tropic vector using a targeted delivery method. However, we do not suggest that the IP injection technique could not be used for newly identified AAV vectors. As more small bowel and colon tropic vectors are developed, future studies should examine both delivery techniques.
AAV serotypes 4, 7, 8, 9, and 10 provided the most efficient gut transduction. Not unexpectedly, we observed substantial liver transduction with all serotypes except 4. We have previously demonstrated the ability of particles and vectors to escape from the mesenteric arterial circulation and enter the portal circulation with deposition in the liver (37). The preferred tropism and improved transduction efficiencies of AAV serotypes 8 and 9 have also been demonstrated in other organ systems and cells types (4, 23, 39, 50). Serotypes 4 and 7 had not yet been investigated in the SI and colon. The amino acid structure of serotype 4 capsid is significantly divergent from the other serotypes and, interestingly, was the only serotype to show less transduction in the liver. Previous studies have also demonstrated transduction efficiencies with serotype 7 in nonintestinal tissues that are similar to serotypes 8 and 9 (13, 25, 50). We also demonstrated that AAV10 provided effective long-term intestinal transduction. AAV10 or AAVrh10 identified from nonhuman primates has not been extensively examined in different organs or in vivo (5, 14, 18, 23). A recent study investigating AAV transduction efficiencies in neonatal mice confirms our AAV10 findings (20).
Localization of cellular transgene expression appeared to be in epithelial cells with only a few colocalizing specifically to EECs. However, lower transgene expression levels may not have been detected and hence may underestimate additional cell types containing AAV genome. Conversely, it could also be that the villous/gland epithelial cells that differentiate into secretory type cells tend to accumulate GFP, or that EECs, which turn over slower than other cell lineages, allow for prolonged visualization of transduced cells. AAV can also transduce fibroblasts and smooth muscle cells, but these cells were not identified by IHC. Future studies for cellular localization should also utilize gene tagging for improved IHC sensitivity and specificity.
AAV10 may in fact have some inherent properties that support its use in intestinal transduction (31); thus it was selected to test in an experimental model of colitis. Delivery of mIL-10 to IL-10−/− experimental colitis was selected as a clean model in which to examine IL-10 transgene levels. Given the observed liver transduction with SMA injection we also added a portal injection treatment arm to the experiment. Certainly, liver contamination limits the ability to study the intestinal specific effects of localized transgene expression since it may mimic systemic administration.
Significant differences in clinical disease activity and histopathology were seen in mIL-10-transferred animals compared with GFP-transferred animals. Single-stranded AAV10-IL-10 also provided long-term gene expression as demonstrated by IL-10 serum levels and SI and colon gene expression levels. A tighter control of serum IL-10 levels was observed in the SMA injection group compared with liver-directed delivery. Serum IL-10 levels decreased over time, which corresponds to observations of transgene expression made by other groups (27, 49).
Comparison of intestinal-transduced and liver-transduced animals showed similar clinical and histopathological responses. This lack of difference could be due to the liver expression seen after SMA injection and the intestinal expression seen after portal injection. However, SMA-delivered AAV10-IL-10 resulted in more time points at which the DAI score was significantly improved compared with control, better colon weight-to-length ratios, and slightly better histology inflammation scores in the proximal colon. More mice had higher IL-10 copy numbers in the colon in the SMA group than the portal group, which might explain the slightly better clinical response. Taken together these data suggest there might be a better clinical response following SMA injection-mediated delivery of IL-10. To fully understand the physiological differences between intestinal transgene expression and liver transgene expression; intestine specific capsids or intestinal specific transgene expression must be developed.
Our study was limited by duration in that it only examined 8 wk after vector administration, thus it is unknown whether IL-10 serum levels or IL-10 genome would have continued to decline over time or whether there would be a difference between the two delivery techniques over time. Conversely, our study is also limited in that we did not examine AAV-mediated gene transfer early after vector administration, but it has been well established that AAV-mediated transgene expression is much higher early on and declines over time (13, 19, 20). Indeed, our serum IL-10 time course data corroborate these observations.
Postmortem examinations also revealed adverse effects of sustained IL-10 gene expression. Spleen weights were significantly greater in the IL-10 treatment groups compared with controls (SMA: 454.5 ± 124.2 vs. 152.8 ± 29.29 mg; portal: 361.4 ± 119.8 vs. 127.8 ± 2.503 mg). Higher serum IL-10 levels correlated with increased spleen size in the portal treatment group (r = 0.9761; P = 0.004) and the SMA treatment group (r = 0.9578; P = 0.010). Others have also reported the adverse effects of sustained IL-10 transgene expression (32).
In summary, we report the intestinal tropism and transduction efficiencies of AAV serotypes 1 through 10 using three different vector delivery techniques. We also demonstrate that AAV-mediated transduction with mIL-10 can effectively modulate mucosal inflammation and the clinical features of experimental colitis. The SMA injection technique cannot eliminate hepatic transduction and thus can result in transgene expression that could mimic systemic administration. Future work is needed to develop SI- and colon-specific targeting of AAV and vector genomes that promote tissue specific expression.
Research support was provided by the Crohn's and Colitis Foundation of America, Career Development Award to S. Polyak.
No conflicts of interest, financial or otherwise, are declared by the author(s).
We thank Chantal Allamargot and the Central Microscopy Research Facility (University of Iowa, Iowa City, IA) for assistance in immunofluorescence studies and confocal microscopy. We also thank Dr. Timothy Wang (Columbia University, New York, NY) for kindly sharing his GFP immunofluorescence protocol.
- Copyright © 2012 the American Physiological Society