Mucosal gene therapy using a pseudotyped lentivirus vector encoding murine interleukin-10 (mIL-10) suppresses the development and relapse of experimental murine colitis

Inflammatory bowel disease (IBD), comprising Crohn’s disease (CD) and ulcerative disease (UC), is thought to result from abnormal interactions between gut associated lymphoid tissue and enteric microflora. This abnormal mucosal immune response is probably facilitated by defects in epithelial barrier function [1]. Interleukin (IL)-10 plays a crucial role in mucosal immunoregulation, inhibiting aspects of both the innate and cell mediated inflammatory response [2]. IL-10 has broad immunoregulatory activity and acts to suppress intestinal inflammation on several levels. Gene-targeted IL-10 knockout mice (IL-10−/−) and IL-10 receptor 2 deficient mice spontaneously develop an enterocolitis by 2–3 months of age with multifocal inflammatory lesions throughout the gastrointestinal tract [3]. In addition, mutations in the gene encoding the IL-10 receptor subunit proteins have been found in IBD patients [4]. IL-10 suppresses the release of many other proinflammatory cytokines and chemokines, including tumor necrosis factor alpha (TNF-α), IL-1, IL-6, and IL-8. Finally there is strong evidence that IL-10 acts to promote the differentiation and augment the activity of regulatory T cells [5].

Administration of systemic IL-10 is sufficient to inhibit inflammation and abrogate experimental colitis models [6, 7]. However, clinical trials in Crohn’s disease have shown that, although daily systemic IL-10 injections are safe and well tolerated, they have minimal therapeutic efficacy [8, 9]. One explanation for these disappointing results is that daily systemic therapy does not deliver adequate mucosal levels of IL-10 that would be needed to control proinflammatory responses associated with active CD. Thus, strategies that result in greater mucosal exposure to IL-10 in the gastrointestinal tract may prove effective in treating IBD.

Several IL-10 delivery systems targeting the gastrointestinal mucosa have been reported; rectal administration using an adenovirus vector [10], oral delivery using non-pathogenic bacteria (Lactococcus lactis) [11], and oral nanoparticles [12]. Recently, we demonstrated the ability of a vesicular stomatitis virus envelope glycoprotein (VSV-G) pseudotyped lentiviral vector (LV) to infect colonic mucosal tissue via the apical surface in vivo, after intraluminal instillation per rectum in a healthy murine model and ex vivo in a human intestinal explant system [13]. In this study, we tested whether efficient local gene transfer can be accomplished using LV expressed murine IL-10 (mIL-10) in a murine colitis model.

In this study, we have demonstrated for the first time that an LV could be used as a novel gene delivery system for the treatment of DSS colitis. Topical gene therapy using a LV encoding mIL-10 was found to suppress not only the development but also the relapse of the colitis. The LV platform is efficient and provides a unique gene transfer system with a number of features that differentiate it from other topical gene therapy approaches that have been evaluated for the treatment and prevention of experimental models of colitis.

Adenovirus vectors [10, 24, 26‐29] and cationic lipids complexed to immunoregulatory genes [30], have reported efficacy in the treatment of murine colitis. However, a major disadvantage of these systems is that the transgene expression is transient since genes are transduced into mucosal epithelial cells that only live for about 2–3 days. Alternatively, an approach using Lactococcus lactis only transiently increased mucosal IL-10 concentration because the bacteria did not colonize the intestine [11, 31]. Ex vivo retroviral gene transfer of CD4+ T cells with IL-10 did not have any therapeutic benefit in acute colitis of immunocompetent mice, though it did prevent induction of colitis in a transfer colitis model using immunodeficient mice [32].

Here we chose to evaluate whether a LV encoding IL-10 was effective in treating or preventing DSS induced murine colitis for three reasons. Firstly, this colitis model induces profound epithelial barrier disruption that might facilitate LV transduction and access to the lamina propria [19]. Secondly, this DSS colitis model shows severe inflammation in the distal colon [19, 20, 22], the area known to have maximal LV transduction. Finally, mucosal CD4+ cells, an important target of LV transduction, also play a critical role in the pathogenesis of IBD colitis [33]. Within the DSS model, increases in mucosal CD4+ T cells occur in a dose dependent fashion such that with 5% DSS induced colitis is associated with significantly more CD4+ T cells compared to 3% DSS colitis [34]. Supporting our hypothesis, LV expressed mIL-10 showed efficacy against acute and relapsing murine DSS colitis.

Notably, we found that topical delivery of a LV encoding IL-10 could suppress not only the development but also prevent relapse of DSS colitis. This is a unique feature of the LV approach and has not been seen in other traditional vector systems. We suggest that the reason for this property might relate to the ability of the LV to transduce both colonic mucosal epithelial and mucosal CD4+ T cells following rectal administration [13]. Mucosal CD4+ T cells play a dominant role in controlling gastrointestinal mucosal inflammation, especially regulatory CD4+ T cells through secretion of anti-inflammatory cytokines such as IL-10 [5, 7]. Interestingly, although not statistically significant, administration of two vector doses showed a trend toward achieving more therapeutic efficacy than a single dose. In this context, it is notable that IL-10 levels were comparable between the single dose and double dose groups, perhaps due to feedback down regulation of endogenous IL-10 in response to vector-derived IL-10. Alternative explanations might include the induction of antibodies or other immunological processes that reduce the efficiency of lentiviral infection and or biological availability of IL-10 in the mucosal tissue. Additional studies are warranted to examine whether higher vector concentrations or frequency of administration might achieve further benefit.

In addition to the barrier disruption caused by DSS colitis, 20% ethanol pre-treatment was also required for optimal LV transduction, perhaps by helping to dissolve or wash away physical barriers obstructing vector access to the cellular surfaces of the mucosal epithelium. While 20% ethanol pre-treatment in itself induces no mucosal inflammation, in the presence of 3% DSS-induced colitis we obtained transduction levels comparable to those observed in our previous studies using 50% ethanol pre-treatment to enhance vector transduction efficiency in the absence of pre-existing mucosal inflammation [13]. Furthermore, similar approaches have been tested both pre-clinically and in clinical trials, e.g., using ethanol or surfactants to improve transduction efficiency of vectors instilled into the bladder, and so this approach has potential for clinical translation in IBD.

There are a number of limitations to our study. Use of additional control groups such as an IL-10 plasmid or lentiviral vector without the IL-10 insert could have provided insight as to the independent contribution of the lentiviral vector and the IL-10 on the subsequent modulation of mucosal infection. In addition we did not measure plasma IL-10 or other immunological parameters that could have provided explanations as to why a second dose of the LV did not appear to provide benefits above or beyond those seen with a single dose of LV.

In conclusion, we have demonstrated that topical gene therapy using a VSV-G LV vector to deliver therapeutic levels of IL-10 could have promise as a novel therapy for IBD colitis. Follow-up studies are needed to determine whether this vector system would be effective in treating other models of murine colitis such as the IL-10 knock out model of murine colitis [3].