Oral Interleukin-10 Alleviates Polyposis via Neutralization of Pathogenic T-Regulatory Cells

IL10 is critical for maintenance of immune homeostasis, particularly at the gastrointestinal interface (1). In both humans and mice, IL10 deficiency is linked to the development of a wide range of immune-mediated pathologies (2, 3), including cancer (4).

Preclinical studies in murine models of inflammatory disease have demonstrated therapeutic utility for recombinant IL10 (5, 6). In clinical trials, however, intravenous administration of IL10 to patients with inflammatory bowel disease (IBD) resulted in dose-dependent paradoxical effects with only nominal therapeutic benefit (7). Because of the lack of methods for administering IL10 in a tissue-specific manner, the clinical potential of IL10 as a biologic therapeutic remains undetermined.

Previous studies have revealed that bioerodible sustained-release microparticles prepared via phase inversion nanoencapsulation (PIN) can be used for oral delivery of biologic molecules to the intestine (8). When applied to models of intestinal inflammation and dysplasia, this strategy affords the ability to target physiologic quantities of protein to the disease microenvironment with minimal escape into circulation. Although the utility of PIN for delivery of nucleic acids and peptides to the gut mucosa has been established, its effectiveness in targeting bioactive protein therapeutics to gut-associated lymphoid tissue and mesenteric lymph nodes (MLN) remains undefined.

On the basis of the previous findings that adoptive transfer of CD4⁺CD25⁺ T cells suppresses intestinal disease in the APC^min/+ mouse model of spontaneous gastrointestinal polyposis in an IL10-dependent manner (9), we sought to determine whether IL10-encapsulated PIN microparticles could directly modify the intestinal inflammatory landscape and alter the natural history of this genetically driven preneoplastic disease. The results provide the first evidence for the therapeutic potential of an oral particulate cytokine formulation, in this case IL10, to ameliorate both local and systemic disease. Furthermore, our findings identify pathogenic T-regulatory cells (pgTreg) as a primary therapeutic target of IL10 in the dysplastic intestine.

C57BL/6 (B6), B6.SJL-Ptprc^aPepc^b/BoyJ (CD45.1), APC^min/+, DEREG, and RAG1^−/− mice were purchased from Jackson Laboratory. BALB/c mice were purchased from Taconic. APC^min/+-DEREG mice were created in house. All experiments were conducted in accordance with guidelines set forth by the Institutional Animal Care and Use Committees at the University at Buffalo (Buffalo, NY) and Wadsworth Center (Albany, NY).

IL10-loaded PIN particles were prepared using a modified PIN process (10). Two formulations were produced: (i) control (no cytokine) and (ii) recombinant murine IL10 (Peprotech, Inc., >98% purity, <1 EU/μg) with a loading of 0.5 μg cytokine/mg of particles. Control or IL10 microparticles were provided by oral gavage (1 mg particles in 0.3 mL sterile water) three times per week for 4 weeks starting at 10 weeks of age. For experiments utilizing FITC-BSA microparticles, single 30 mg bolus doses were given.

The following antibodies were used for immunofluorescence and IHC: CD45R/B220 (RA3-6B2, BD Pharmingen), CD11c (N418, eBioscience), Fc receptor block; anti-CD16/CD32 (2.4G2, BD Pharmingen), CD3 (F7.2.38, Agilent Technologies), and von Willebrand factor (vWF; A0082, Dako). The following antibodies were used for flow cytometry: CD45.1 (A20, eBioscience), CD4 (RM4-5, eBioscience), IL17A (TC11-18H10, BD Pharmingen), FoxP3 (FJK-16s, eBioscience), RoRγt (Q31-378, BD Pharmingen), CD44 (IM7, eBioscience), IFNγ (XMG1.2, BD Pharmingen), CD8α (53-6.7, eBioscience), and γδ-TCR (eBioGL3, eBioscience). Anti-mCD4 (GK1.5, BioXCell) was used for in vivo CD4⁺ T-cell depletion.

Tissues were embedded in OCT compound (Sakura Finetek) and snap frozen in liquid nitrogen. Of note, 20-μm serial cryosections were immunostained as indicated previously (11). Scanning laser confocal images were collected using a Leica SP5 ABOS microscope and processed using Fiji Software (http://fiji.sc/).

Formalin-fixed intestines were opened longitudinally and polyp burdens were quantified using a dissecting microscope.

Spleens and MLN were processed into single cell suspensions. Intestines were digested and fractionated into lamina propria mononuclear cells (LPMC) and intraepithelial lymphocytes, as described previously (12). For experiments requiring detection of intracellular antigens (FoxP3, RoRγt, IL17A, and IFNγ), cell suspensions were cultured for 5 hours in the presence of Golgistop (5 μL/mL; BD), phorbol myristate acetate (50 ng/mL; Sigma), and ionomycin (1 μg/mL Sigma). Cells were then permeabilized and fixed using an intracellular staining kit (eBioscience) overnight at 4°C.

Paraffin-embedded tissue sections (5 μm) were stained with hematoxylin and eosin. For IHC, slides were subjected to antigen retrieval before application of the indicated primary antibodies. Biotinylated secondary antibodies (anti-rat IgG, BD Pharmingen), HRP-SA conjugate (Invitrogen), and diaminobenzidine chromogen (Dako) were used for visualization. Images were taken at ×400 magnification.

For intestines, histologic scores were assigned in a blinded fashion using criteria established in our laboratory (see Supplementary Data). For spleens, histologic scores were assigned using modified criteria from a previously published report (see Supplementary Data; ref. 13).

Red blood cell (RBC), hemoglobin (HgB), and hematocrit (HCT) levels were determined using a Vetscan HM5 automated analyzer (Abaxis Veterinary Diagnostics).

Mortality criteria for APC^min/+ mice were developed and applied to survival studies. A combined scoring system, taking into account the degree of weight loss and the severity of change in activity was selected (see Supplementary Data).

RNA was isolated from nonpolyp (for WT and APC^min/+ mice) and polyp sections (for APC^min/+ mice) of terminal ileum using the RNeasy Mini Kit (Qiagen) and subjected to gene expression profiling. Briefly, RNA was processed into cRNA using the Affymetrix Genechip 3′ IVT Expression Protocol and applied to Affymetrix Mouse Genome 430A 2.0 arrays. Interference from bacterial RNA was selected against using a probe-set enriched in oligos extending into 3′-poly-A tails. Details with respect to sample processing and preparation can be accessed from the Gene Expression Omnibus (NIH) under accession number GSE49970.

DEREG or APC^min/+-DEREG mice received 1 μg of Diphtheria toxin (DT, Sigma) dissolved in 100 μL of Dulbecco PBS by intraperitoneal injection on days 0, 1, and 14.

MLN CD45.2⁺CD4⁺CD25⁺ cells and total lymph node CD45.1⁺CD4⁺CD25⁻ cells were isolated by magnetic selection from either APC^min/+ or BoyJ mice, respectively. A total of 1 × 10⁶ CD45.1⁺CD4⁺CD25⁻ cells were CFSE labeled using the Vybrant CFDA SE cell tracer kit and adoptively transferred with 4 × 10⁵ MLN CD45.2⁺CD4⁺CD25⁺ as indicated, into 8-week-old female RAG1^−/− recipients.

A total of 1 × 10⁶ magnetically selected MLN CD4⁺CD25⁺ T cells from 14-week-old APC^min/+ mice were adoptively transferred into sex-matched 10 week old APC^min/+ recipients. Four days before transfer, recipient mice were transiently depleted of their CD4⁺ cells using a single 300 μg i.p. dose of anti-CD4 antibody.

Steady-state mRNA levels were detected with SYBR Green PCR Master Mix (Applied Biosystems) using the Mx3000p qPCR system (Agilent Technologies). Results were normalized to the expression of β-actin. The expression level was scaled using the 2^−ΔΔCt method with the average levels obtained for wild-type intestines set arbitrarily to 1. Primer sequences utilized were: β-actin forward 5′-TCACCCACACTGGCCCATCTACGA-3′, reverse 5′-TGGTGAAGCTGTAGCCACGCT-3′; IL1β forward 5′- GCCCATCCTCTGTGACTCAT-3′, reverse 5′-AGGCCACAGGTATTTTGTCG-3′; IL6 forward 5′-CCATCCAGTTGCCTTCTTGG-3′, reverse 5′-TTTCTGCAAGTGCATCATCG-3′; IL17A forward 5′-CTGAGCTTCCCAGATCACAGAG-3′, reverse 5′-CGCAAAAGTGAGCTCCAGAAAG-3′; GM-CSF forward 5′-CACGTTGAATGAAGAGGTAGAAG-3′, reverse 5′-CATGTTCAAGGCGCCCTTGAG-3′; TNFα forward 5′-GAACTGGCAGAAGAGGCACT-3′, reverse 5′-AGGGTCTGGGCCATAGAACT-3′.

Statistical calculations were performed using Student t test in pairwise comparisons of groups. Log-rank tests were utilized for survival studies. P values of 0.05 or less were considered statistically significant.

Oral PIN microparticles loaded with FITC-BSA were taken up and retained in the Peyer's patches and MLNs of both APC^min/+ and wild-type mice (Fig. 1A and Supplementary Fig. S1A and S1B). Microparticles were not detected in the colon, liver, or spleen at any time point after feeding (Supplementary Fig. S1C). Moreover, when IL10-loaded microparticles were delivered in bolus doses, IL10 could not be detected in serum (data not shown). These results suggest that uptake of microparticles and release of IL10 were effectively localized to intestinal immune structures.

We directly tested whether oral administration of IL10 microparticles could suppress established polyposis in APC^min/+ mice. Therapy resulted in a more than 2-fold reduction in polyp burden (Fig. 1B) and a decreased severity of disease pathology in the intestine (Supplementary Fig. S2A and S2B). In APC^min/+ mice, intestinal disease is accompanied by predictable systemic abnormalities, namely anemia, splenomegaly, and weight loss (14). To determine whether amelioration of local disease would result in improvement of systemic symptoms, we monitored experimental animals over the course of therapy. In all cases, treatment with IL10 resulted in significant benefit (Fig. 1C–E and Supplementary Fig. S2C). Relief of systemic symptoms was likely secondary to the suppression of chronic intestinal inflammation, as such activity has been shown to cause anemia (15) and to promote the establishment of extramedullary hematopoiesis (16). Accordingly, analysis of APC^min/+ mouse spleens revealed significant megakaryocytosis, which was abrogated following treatment (Fig. 1D). Chronic treatment resulted in a 15% extension of both median and maximal lifespan, suggesting that IL10 could suppress but not arrest disease (Fig. 1F and Supplementary Fig. S3A).

We sought to define the potential mechanisms underlying the therapeutic activity of oral IL10. Because IL17 has recently been implicated as an essential driver of polyp growth in APC^min/+ mice (17), and IL10 has been shown to suppress IL17 production by T cells (18), we hypothesized that the observed therapeutic effects of IL10 microparticles may partially be attributed to modification of T-cell activity. APC^min/+ nonpolyp and polyp areas demonstrated increasing levels of T-cell activation, with a progressive shift from a balanced T-helper (Th) 1/2/17 profile to an IL1β/IL6/IL23-enriched Th17-promoting profile during early-stage disease (Fig. 2A). Importantly, quantification of select Th17-related inducer/effector cytokines confirmed that IL10 microparticles were highly effective at suppressing the expression of these genes in mice with established disease (Fig. 2B). In contrast, treatment did not result in significant changes in IFNγ or IL4 levels (data not shown). Taken together, these results demonstrate that IL10 microparticle therapy strongly antagonizes IL17-mediated inflammation.

T cells are thought to be the primary producers of IL17 during chronic intestinal inflammation (19). Accordingly, we examined the IL17 expression profile in various types of intestinal T cells from our APC^min/+ mice. Although intraepithelial γδ-T cells did not produce significant levels of IL17 during disease (Supplementary Fig. S3B), there was an increase in both the total number of CD4⁺ T cells and in their ability to produce IL17 (Fig. 2C). IL10 treatment did not impact the absolute numbers of CD4⁺ T cells, but did act to dramatically reduce the percentage of IL17-producing cells within this population (Fig. 2C). Because both FoxP3⁻RORγt⁺IL17⁺ Th17 cells and FoxP3⁺RoRγt⁺IL17⁺ pgTregs are known to be a major contributor to polyposis in this model (19–21), we examined the impact of IL10 on IL17 expression by these two subsets. IL10 treatment resulted in a dramatic reduction in IL17 production by pgTregs, with only a modest decline in IL17 production by Th17 cells (Fig. 2C). Collectively, these results suggest that of the IL17-producing T-cell subtypes present during disease, pgTregs are uniquely sensitive to modulation by IL10.

Next we examined whether modulation of pgTregs was the primary mechanism underlying IL10 microparticle therapeutic effect. We crossed APC^min/+ mice with the DEREG murine model of inducible Treg depletion and developed a subtotal Treg depletion protocol appropriate for our desire to assess the contribution of the IL10–pgTreg axis to our model system. Subtotal Treg depletion resulted in a more than 70% reduction in FoxP3⁺ cells over 28 days, yet avoided the catastrophic myelo- and lympho-proliferative disorder typically seen in mice undergoing total Treg ablation (Supplementary Fig. S4 and S5A; ref. 22). Subtotal Treg depletion reduced polyp burden, anemia, and splenic pathology (Fig. 3A and Supplementary Fig. S5B and S5C), supporting the notion that pgTregs are the major contributors to both local and systemic disease. Administration of IL10 microparticles further enhanced depletion-induced therapeutic benefits, suggesting that IL10 was acting, in part, through neutralization of residual pgTregs. Consistent with this hypothesis, analysis of MLNs revealed that subtotal Treg depletion reduced the prevalence of pgTregs, whereas IL10 both diminished the pgTreg presence and promoted an increase in conventional Tregs; MLN Th17 cell numbers, however, remained unaffected (Fig. 3B and Supplementary Fig. S5D and S5E). Analysis of the LPMCs revealed a similar pattern, that is, both subtotal Treg depletion and IL10 therapy diminished the proportion of IL17-positive cells, but the most significant reduction was obtained in the combination treatment group (Fig. 3C). Whether the superior effect that was obtained in this group was due to the appearance of a qualitatively unique Treg population is currently under investigation.

To determine whether the dual activity of IL10 on pgTregs and Tregs resulted in altered functional capacity within the total Treg population, we tested the ability of posttherapy MLN CD4⁺CD25⁺ T cells to suppress naïve T-cell proliferation and activation. In comparison with cells isolated from control-treated APC^min/+ mice, the cells from IL10-treated APC^min/+ mice were significantly more effective at suppressing these processes both in vivo and in vitro (Fig. 4A and Supplementary Fig. S6C).

The above results demonstrated that the suppressive capacity of APC^min/+ MLN CD4⁺CD25⁺ T cells was restored by IL10 microparticle therapy. We predicted that functional rescue of APC^min/+ intestinal Treg population via reversal of the pgTreg/Treg ratio was ultimately responsible for disease amelioration. To confirm this notion, we adoptively transferred MLN CD4⁺CD25⁺ T cells from control or IL10-treated APC^min/+ mice into 10-week-old APC^min/+ recipients. IL10-conditioned CD4⁺CD25⁺ T cells reduced polyposis, corrected anemia, and decreased the prevalence of Th17 cells in recipient MLNs (Fig. 4B and Supplementary Fig. S7A–S7C). Splenic disease, however, was not alleviated after transfer (Supplementary Fig. S7D and S7E). Whether the inability to restore full therapeutic effect was due to the limitations of the experimental approach, that is, accumulation/persistence of transferred cells in the intestine, or to the absence of Treg-independent effects of IL10 microparticle therapy, that is, impacts on the innate immune system, was not determined.

Our findings demonstrate that tissue-specific delivery of a particulate cytokine formulation can change the natural history of a highly penetrant genetic condition. We propose that tissue-localized IL10 serves as a molecular rheostat, specifically modulating pgTreg function during intestinal inflammatory and dysplastic disease. The overall concepts tested here are expected to extend to other inflammation-driven diseases, especially those manifesting at mucosal surfaces, including IBD, asthma, and genital tract infection.

N.K. Egilmez is the vice-president of and has ownership interest (including patents) in TherapyX, Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A.Y. Chung, K. Khazaie, N.K. Egilmez

Development of methodology: A.Y. Chung, Q. Li, C. LeVea, K. Khazaie

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.Y. Chung, S.J. Blair, M. De Jesus, K.L. Dennis, C. LeVea, L.P. Virtuoso, N.G. Battaglia, S. Furtado, E. Mathiowitz, N.J. Mantis

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.Y. Chung, Q. Li, S.J. Blair, K.L. Dennis, J. Yao, Y. Sun, E. Mathiowitz, N.K. Egilmez

Writing, review, and/or revision of the manuscript: A.Y. Chung, C. LeVea, Y. Sun, N.J. Mantis, K. Khazaie, N.K. Egilmez

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.Y. Chung, T.F. Conway, K. Khazaie, N.K. Egilmez

Study supervision: N.K. Egilmez

The work was supported by the NIH (AI092133 to N.K. Egilmez; HD061916 to N.J. Mantis; and CA160436 to K. Khazaie) and the Howard Hughes Medical Institute (M. De Jesus).