Volume 135, Issue 3, Pages 871-881.e5 (September 2008) Myeloid-Derived Suppressor Cells in Inflammatory Bowel Disease: A New Immunoregulatory Pathway  Lydia A. Haile, Reinhard von Wasielewski, Jaba Gamrekelashvili, Christine Krüger, Oliver Bachmann, Astrid M. Westendorf, Jan Buer, Roland Liblau, Michael P. Manns, Firouzeh Korangy, Tim F. Greten  Gastroenterology  Volume 135, Issue 3, Pages 871-881.e5 (September 2008) DOI: 10.1053/j.gastro.2008.06.032 Copyright © 2008 AGA Institute Terms and Conditions

Figure 1 Repetitive transfer of splenocytes from CL4-TCR transgenic mice significantly reduces intestinal inflammation. (A) Splenocytes from CL4-TCR transgenic mice were adoptively transferred once (1 × 107 cells on day 27) or 3 times (5 × 105 cells on day 0, 1 × 107 cells on days 12 and 27) into VILLIN-HA mice. Five days after the last transfer, all mice were killed for further analysis. Representative data from one of more than 8 independent experiments are shown. (B) Macroscopic examination of the intestine from naive VILLIN-HA mice and mice after one transfer and 3 transfers of splenocytes. Representative data from one of more than 8 independent experiments are shown. (C) H&E staining of the paraffin-embedded tissues from jejunum (top) and colon (bottom) of mice after one transfer and 3 transfers of splenocytes from CL4-TCR transgenic mice (100- and 400-fold magnification). Data from one of at least 8 independent experiments are shown. (D) Histologic score of tissue sections from intestine after one transfer and 3 transfers. *P < .05. Data represent results from 3 independent experiments with 2 or 3 mice in each experiment group. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Figure 2 HA-specific IFN-γ response is reduced upon repetitive transfer of antigen-specific splenocytes into VILLIN-HA mice. (A) MLN were isolated from VILLIN-HA mice 5 days after one cell transfer or 3 cell transfers and stimulated in vitro with the HA512–520 peptide. Intracellular IFN-γ responses were analyzed by fluorescence-activated cell sorting. Data are representative of 2 independent experiments with 2 mice per group. (B) In vivo proliferation of HA-specific transgenic CD8+ T cells in spleen (SPL) and MLN of VILLIN-HA mice. Carboxyfluorescein succinimidyl ester–labeled splenocytes from CL4-TCR transgenic mice were adoptively transferred to naive mice and mice after 3 cell transfers. Data are derived from 2 independent experiments with 2 mice per group. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Figure 3 CD11b+Gr-1+ MDSCs accumulate following repetitive transfer of antigen-specific CD8+ T cells. (A) Representative fluorescence-activated cell sorter analysis of CD11b+Gr-1+cells in the spleen (SPL) and MLN of VILLIN-HA mice after one cell transfer and 3 cell transfers. Data shown are representative of 3 independent experiments. (B) MDSCs accumulate in the intestine after 3 transfers of HA-specific splenocytes into VILLIN-HA mice. CD11b+Gr-1+ cells from intestine were identified by fluorescence-activated cell sorting. (C and D) MDSCs from naive mice and mice after 3 cell transfers were further characterized using additional surface markers. The panels are gated on CD11b+Gr-1+ cells. Filled histograms show isotype controls. Data are from one of 2 independent experiments. (E) MDSCs from VILLIN-HA mice after 3 transfers have a suppressive function on antigen-specific CD8+ T cells in vitro. MDSCs from VILLIN-HA mice after 3 transfers were incubated with HA512–520 peptide stimulated transgenic splenocytes from CL4-TCR mice. Data from one of 2 independent experiments with similar results are shown. *P < .05 and **P < .01 as determined by Student t test. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Figure 4 DSS treatment does not lead to an increase in the frequency of splenic CD11b+Gr-1+ cells. (A) Colitis was induced in VILLIN-HA mice by the addition of DSS (5%, 7%, 7%) to the drinking water on days 0, 15, and 27, respectively, for 5 days. Mice were killed on day 33. (B) SCID mice were injected with 5 × 105 CD4+CD45RBhigh cells. Mice were killed when they developed wasting disease and diarrhea. The frequency of MDSCs in spleen (SPL) and MLN of mice was determined. (C and D) VILLIN-HA mice were treated with DSS in drinking water or adoptively transferred with antigen-specific splenocytes. Intestinal biopsy specimens were put into culture in complete media and supernatants were collected after 48 hours of culture of the colon, small intestine (SI), and MLN as well as serum and checked by enzyme-linked immunosorbent assay for (C) IFN-γ and (D) GM-CSF. The cytokine levels were normalized to weight of small intestine or colon. Data are pooled from 2 independent experiments with 2 mice per group (*P < .05). Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Figure 5 MDSCs from VILLIN-HA mice after 3 cell transfers express NOS2 and ARG1. (A) Messenger RNA expression of NOS2 and ARG1 was analyzed in MDSCs from naive mice and VILLIN-HA mice after repetitive transfer of CL4-TCR splenocytes. Data are representative of 2 independent experiments. (B) CD11b+Gr-1+ MDSCs isolated from VILLIN-HA mice after 3 cell transfers were incubated together with CL4-TCR splenocytes in the presence of 1 μg/mL HA peptide. NO production was determined after 48 hours from cell supernatants. No NO production was detected in the absence of stimulated T cells. Data are derived from duplicate cultures (*P < .05). (C) Analysis of ARG activity in MDSCs isolated from naive mice or from VILLIN-HA mice after 3 transfers. Data are pooled from 4 independent experiments. (D) NO production (top) and proliferation (bottom) in coculture of MDSCs with HA-pulsed CL4-TCR splenocytes in Transwell plates. Data are derived from triplicate cultures (*P < .05). (E) Coincubation of MDSCs with CL4-TCR splenocytes in the presence of an NO synthase inhibitor (NG-monomethyl-l-arginine [l-NMMA]) and an ARG inhibitor (NG-hydroxy-l-arginine [l-NOHA]). Data are derived from triplicate cultures (*P < .05). (F) HA peptide–stimulated CL4-TCR splenocytes were cultured alone or with CD11b+Gr-1+ cells. CD8+ cells were stained with annexin V and 7-aminoactinomycin D. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Figure 6 Cotransfer of MDSCs with HA-specific CD8+ T cells ameliorates intestinal inflammation. (A) MDSCs were isolated from VILLIN-HA mice after 3 transfers and cotransferred with HA-specific CD8+ T cells in a 4:1 ratio. Mice were monitored for weight loss. Data from 2 independent experiments with 4 mice is shown (*P < .05). (B) H&E sections from ileum (top) and colon (bottom) from VILLIN-HA mice after transfer of CL4-TCR cells (left) or CL4-TCR cells plus MDSCs (right). (C) Histologic score of tissue sections from intestine after single transfer of CL4-TCR splenocytes and in combination with MDSCs (*P < .05). Data from 2 independent experiments with 4 mice are shown. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Figure 7 Patients with IBD demonstrate higher frequency of MDSCs in peripheral blood. (A) Frequency of CD14+ cells in peripheral blood from patients with ulcerative colitis, patients with Crohn's disease, and healthy controls. (B) MDSCs were detected by CD14/HLA-DR staining as described. *P < .05 as shown by Student t test. (C) CD14+ cells from patients with IBD and healthy donors demonstrate ARG activity. ARG activity was determined from isolated CD14+ cells as described. Data shown are from healthy donors (n = 3) and patients with IBD (n = 3). (D) Purified MDSCs from patients with IBD suppress proliferation and cytokine secretion by autologous-stimulated PBMCs cells in a dose-dependent matter. The CD14+ population was sorted from PBMCs of patients with IBD or healthy donors and tested with autologous PBMCs in a suppression assay. Data are derived from 3 patients and 2 healthy donors. Proliferation was determined in duplicate cultures by 3H incorporation. **P < .01 and ***P < .001 versus control as determined by Student t test. (E) Supernatants from the proliferation assay were removed after 48 hours and measured for IFN-γ production by enzyme-linked immunosorbent assay. Data are derived from patients with IBD (n = 3) and healthy donors (n = 2). *P < .05, **P < .01, and ***P < .001 versus control as determined by Student t test. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Supplementary Figure 1 Adoptive transfer of purified CD8+ T cells from Cl-4 TCR mice induce an increase in frequency of CD11b+ Gr-1+ cells in the spleen. CD8+ T cells from CL-4 TCR transgenic mice were adoptively transferred three times (2x105 cells on day 0, 1x106 cells on day 12 and 27) into VILLIN-HA mice. Five days after last transfer, all mice were sacrifeced and splenocytes were stained for Gr-1 and CD11 b and analyzed with flow cytometry. * P < .05 versus naive control. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Supplementary Figure 2 The percentage of CD4+Foxp3+ regulatory T cells is increased in mesenteric lymph nodes but not in spleen after one transfer, however, three transfers did not further increase the frequency of CD4+ cells. CD4+Foxp3+ Cells were analyzed by flow cytometry in the spleen and MLN from either naive mice or after 1 and three transfers of CL-4 TCR spenocytes into VILLIN-HA mice. Two mice are analyzed individually in two independent experiments. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Supplementary Figure 3 No significant differences were found for IL-6, IL-10 and IL-12 after DSS treatment and CD8+ T cell transfer. (A) Colitis is induced in VILLIN-HA mice by the addition of DSS in drinking water on days 0,15, 27. Mice were treated with DSS in drinking water or adoptively transferred with CL-4 TCR splenoytes. Intestine was put into culture in complete media and supernatants were collected from 48 hours culture of the colon, small intestine and MLN as well, as serum and checked by ELISA for IL-6, IL-10, IL-23 (B, C). The cytokine levels were normalized to weight of small intestine or colon. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions

Supplementary Figure 4 Depletion of regulatory T cells does not reverse the attenuated intestinal inflammation observed after third transfer. Mice were depleted of regulatory T cells starting at 5 days before the third transfer and antigen-specific splenocytes from CL-4. TCR transgenic mice were transferred into the VILLIN-HA mice on day 27. (A) Mice were monitored for signs of colitis and the body weight was monitored. (B) Five days after cell transfer all mice were sacrificed and the frequency of splenic MDSCs was determined. Gastroenterology 2008 135, 871-881.e5DOI: (10.1053/j.gastro.2008.06.032) Copyright © 2008 AGA Institute Terms and Conditions