Four to eight-week-old C57BL/6 female mice were obtained from Charles River Laboratories (L’Arbresle, France) and kept in quarantine for at least one week before any experimental procedure. Rag2-deficient (Rag-2^–/–) C57BL/6 mice were bred in our specific pathogen-free rodent facility. Mice were housed in this facility (#D25-056-7) in standard plastic cages with cellulose bedding in environmentally controlled and germ-free rooms at 22°C, and kept under a 12-hour light/dark cycle. Pellet food and sterile tap water (Hydropack; Plexx) were available ad libitum. Experimentation (#02831) was approved by the local ethic committee (#58) and the French Ministry of Higher Education and Research (Ministère de l’Enseignement Supérieur, de la Recherche et l’Innovation) and was conducted in accordance with the European Union Directive 2010/63.

An intestinal mucosal wound was performed as previously described (26). To wound the C57BL/6 mouse colon, a biopsy forceps was inserted into the working channel of the endoscope (superslim 8.4 Fr, 2.8 mm, flexible video ureteroscope [URF V2; Olympus]) and introduced into the colon of a 2% isoflurane-anaesthetized 6 to 8-week-old mice. Under visual control, a small piece of mucosa was carefully removed. SuperMApo or vehicle (X-vivo) was administrated i.p. the day of the injury and 48 hrs later (1 mL per mouse/injection). The process of wound healing was subsequently monitored by daily endoscopies.

Naive CD25⁻CD45RB^highCD4⁺ T cells were sorted by flow cytometry (cell sorter SH800Z, SONY) from magnetically enriched CD4⁺ spleen and lymph node cell suspensions (EasySep Mouse CD4⁺ T Cell Isolation Kit; StemCell Technology) to a >95% purity. According to the described model (27), 5.10e5 naive T cells sorted from C57BL/6 were injected into the tail vein of 6 to 8-week-old Rag-2^–/– C57BL/6 mice to induce colitis. Mice were monitored daily for clinical assessment of colitis and twice a week by video-endoscopy to appreciate colon mucosa. SuperMApo or vehicle (X-vivo) was administrated i.p. at day 10, after T cell transfer, and 48 hrs later (1 mL/mouse/injection). At that time (10 days post-T cell transfer) mice experimented a 6 to 7 MEICS score, out of 15. In additional in vivo experiments, TGF-β, IGF-I and VEGF recombinant growth factors (R&D systems), or PBS, were administrated i.p. at day 10, after T cell transfer, and 48 hrs later (1 mL per mouse/injection; TGF-β: 700 pg/mL, IGF-I: 1050 pg/mL and VEGF 350 pg/mL).

Chronic DSS colitis was induced in 6 to 8-week-old C57BL/6 mice with the addition of 3% of DSS (MP Biomedicals LLC) in drinking water for seven consecutive days. Then, regular water was given for the next 10 days, and this cycle was repeated twice. Control mice were given with tap water only. Mice were monitored daily for clinical assessment of colitis and twice a week by video-endoscopy to appreciate their colon mucosa. SuperMApo or vehicle (X-vivo) were given i.p. the day of the first DSS cycle initiation, and 48 hrs later (1 mL/mouse).

The mice were monitored daily for clinical assessment of colitis according to a total clinical score of 0 to 9 assigned to each animal and composed of the sum of the body weight loss (0 to 5% of weight loss = 0; 5 to 10% = 1; 10 to 15% = 2; 15 to 20% = 3; >20% = 4), the feces appearance (normal = 0; diarrhea = 1; bloody stools = 2), and the general behavior (normal = 0; mild alteration = 1; piloerection and motor activity altered = 2; severe alteration = 3). The mice were monitored twice a week by video-endoscopy under 2% isoflurane-induced anesthesia, without bowel preparation nor starving. Video-endoscopy was performed using a superslim 8.4 Fr (2.8 mm) flexible video ureteroscope (URF V2; Olympus), connected to OTV-S190 and CLV-S190 visera elite monitor system (Olympus), with a narrow band imaging system enhancing visibility of vascular structures. Endoscopic assessment of mucosal inflammation was reported with the murine endoscopic index of clinical severity (MEICS) score (28), composed of five parameters rated from 0 to 3, including: thickening of the colon, changes in the vascular pattern, fibrin deposition, granularity of the mucosal surface, and stool consistency.

The mice were submitted to gavage with fluorescein isothiocynate (FITC)-dextran (MW 4000; Sigma-Aldrich) at 80 mg/kg of body weight, five hrs prior to blood collection through retro orbital puncture. Fluorescence intensity in the serum was measured at 520 nm (Perkin Elmer). FITC-dextran concentrations were determined from a standard curve generated by a serial dilution of FITC-dextran. Quantification of regenerating islet derived protein 3-γ (REG-3γ) concentrations in plasma was determined by ELISA (Cloud clone corporation) according to the manufacturer’s instructions.

Pro-resolving factors from mouse or human pro-resolving macrophages (called SuperMApo) were produced as described previously (14). Briefly, SuperMApo consisted in the supernatant of a 48-hr culture of apoptotic thymocytes (rendered apoptotic by a 35 Gy-irradiation) with thioglycolate-elicited peritoneal macrophages, in X-vivo culture medium to a 5 to 1 ratio, respectively (14). Human SuperMApo consisted in the 48-hr culture supernatant of M-CSF blood monocyte-derived macrophages cultured with apoptotic autologous peripheral blood mononuclear cells (rendered apoptotic by a 35 Gy-irradiation) to a 5 to 1 ratio in MEM medium, respectively (14). Supernatants were collected, centrifuged to eliminate debris, filtered through a 0.22 µm filter and frozen or lyophilized and frozen for conservation. SuperMApo was used directly in culture experiments (see Functional Assays) or injected i.p. for in vivo experiments (1 ml/mouse repeated 48 hrs later, in total 2 ml per mouse). In some in vitro experiments, VEGF (AF493NA), TGF-β (MAB240) and IGF-I (AF791) neutralizing antibodies or control anti-goat IgG (AB-108-C) or anti-mouse IgG1 antibodies (MAB002) (all from R&D systems), were used in addition to SuperMApo. In some in vivo experiments, growth factors were immuno-magnetically depleted from SuperMApo. Briefly, 10 µg/ml of antibodies directed against VEGF and IGF-I (the same clones than in blocking experiments) as well as 50 µg/ml of anti-TGF-β antibody (2G7 clone; provided by Prof. L. Chatenoud) were incubated for one hour with SuperMApo at room temperature. Then, goat anti-mouse IgG (Polysciences Europe GMBH) or rabbit anti-goat IgG (Rockland) coated magnetic beads were incubated separately with a pre-formed antigen (Ag)-antibody (Ab) complex for 20 minutes or one hour respectively, at room temperature. The Ag-Ab-bead complexes were then captured using a magnetic separator. Protein levels of FGF, EGF, TGF-β, IGF-I and VEGF in SuperMApo were determined by ELISA (all from R&D systems), according to the manufacturer’s protocols.

Immediately after sacrificing the mice, tissue samples from the proximal, median and distal parts of the colon were collected and fixed overnight in 5% formalin. Paraffin inclusions were performed on the macroscopic injured area. Then, 4 µm sections were stained with hematoxylin-eosin-saffron (HES). A histological evaluation was performed by an experimented pathologist blinded to the nature of the mice being examined, using a five-degree severity score based on inflammatory cell infiltration, erosions and ulcerations, epithelial hyperplasia and mucin depletion, as previously described (29). Cytofix/Cytoperm buffer (eBioscience) was applied for the detection of nuclear staining. The primary antibodies used for immunostaining of mouse tissue and primary fibroblasts were: rabbit anti-mouse Ki-67 (clone SP6; Abcam), rabbit anti-human α-smooth muscle actin (α-SMA) (polyclonal; Abcam) and Alexa 488-conjugated rabbit anti-vimentin (Clone D21H3; Cell signaling). The secondary antibody was Alexa Fluor 555 goat anti-rabbit IgG (H+L) (Life Technologies). The nuclei were counterstained with DAPI (Sigma-Aldrich). Fluorescence analysis was performed with Olympus IX81 and Fluoview FV1000 software (Olympus). For quantitative analyses, at least four to five representative high-power fields (HPF, 40X) per section were evaluated in a blinded manner.

Total RNA were extracted from tissues or cells using the RNeasy mini kit (Qiagen) and 1 to 2 µg of RNA was reversed-transcribed into cDNA using the high capacity RNA to cDNA kit (Applied Biosystems). Then, cDNA were quantified by real time RT-PCR using the power SYBR green PCR master mix (Applied Biosystems) using Col1a1 (Mm0080 1666_m1), Col3a1 (Mm0125 4476_m1), Fn1 (Mm0125 6744_m1) and GAPDH primers (from Applied Biosystems). Relative mRNA levels were determined using the ΔΔCt method. Values were expressed relative to GAPDH levels.

Fibroblasts were isolated from C57BL/6 mice colon mucosa or from human colon biopsies (30). Colon biopsies were collected according to standard procedures. This study was approved by the Ethics Committee of Besançon (#14/456; date of approval: July 13^th, 2021), and all participants signed an informed consent form. Mucosal sheets were treated with 5 mM of EDTA (Life Technologies) in calcium-free HBSS (Life Technologies) five times at 37°C for 15 min to remove the epithelial layer. The mucosal sheets without epithelial cells were then cut into small pieces and cultured in completed RPMI-1640 medium (Life Technologies) containing 10% of FBS (Life technologies), 1% of penicillin/streptomycin and 10 mg/mL of collagenase D (Roche) for 60 min at 37°C, 5% CO₂. Isolated fibroblasts were cultured in RPMI-1640 medium containing 10% of FBS, 1% of penicillin/streptomycin, 10 mM of sodium pyruvate (Lonza), 1 mM of non-essential amino acids (Lonza) and 0.05 mM of 2 beta-mercaptoethanol (Sigma Aldrich). The cultured fibroblasts were used for experiments after two to four passages. The immortalized MODE-K cell line derived from mouse small IEC was obtained from INSERM U1111 (31). Cells were cultured in completed RPMI-1640 medium at 37°C, 5% CO₂.

Data are expressed as mean ± standard error of the mean (s.e.m.), and were analyzed using adapted statistical tests, as indicated in figure legends, using GraphPad Prism software (5.01 version). A p-value below 0.05 was considered statistically significant.

The pro-resolving functions of the factors released by macrophages after co-culture with apoptotic cells (named SuperMApo) have been described previously (14). These factors were evaluated for their capacity to stop naive T cell transfer-induced colitis. SuperMApo was injected twice (at a 48-hr interval) in mice exhibiting ongoing colitis, as attested by a MEICS score of 6 to 7 out of 15 (determined by live colonic endoscopy). The day after the last injection, colonic lesions stopped progressing in SuperMApo-treated mice, as attested by the MEICS score stagnation, whereas these lesions aggravated in vehicle-treated mice ( Figure 1A ). A thickening and increased granularity of the colon mucosa, a high fibrin deposition as well as thinner and tortuous blood vessels demonstrated lesion worsening between the last day of treatment (day 10) and day 19 in vehicle-treated mice ( Figure 1A ). MEICS score improvement after SuperMApo treatment was associated with reduced weight loss and reduced colitis clinical score ( Figures 1B, C ). Intestinal necropsy 10 days post-treatment showed an increased length of the colon in IBD mice receiving SuperMApo compared to the colon length in vehicle-treated mice. Colons of vehicle-treated mice were shorter, hyperemic and contained fewer feces due to massive diarrhea ( Figures 1A, D ). Altogether, this indicates inflamed colons. Histological examination of the intestines showed a reduced infiltration of inflammatory cells, notably in the rectum, in SuperMApo-treated mice ( Figure 1E ). Overall, these data demonstrate that the injection of pro-resolving factors released by efferocytic macrophages in mice with ongoing IBD could control IBD progression and inflammatory cell infiltration.

We further evaluated the wound healing properties of factors produced by efferocytic macrophages in DSS-induced colitis model. Administered on the day of colitis induction and 48 hrs later, SuperMApo strongly prevented colitis severity as attested by a reduced clinical score (which reached a score of 0 by day 48 following treatment), and a reduced MEICS score ( Figures 2A, B ) throughout the 50 days of the experiments. Body weight loss was not significantly improved ( Figure 2C ). Vehicle-treated mice exhibited worse colitis lesions, such as high fibrin deposition within the colon, a thickened and bleached mucosa with spontaneous bleeding, superficial ulceration and edema ( Figure 2B ). Macroscopic and histologic examinations confirmed mucosa improvement in mice receiving pro-resolving factors ( Figures 2D, E ). Overall, these factors exhibit pro-resolving properties in inflamed intestine mucosa.

Fibroblasts and epithelial cells are the main cells involved in tissue wound healing. Next, we addressed whether these cells are affected by pro-resolving factors released by macrophages after efferocytosis (i.e., SuperMApo treatment). First, primary mouse fibroblasts extracted from colon tissues were cultured with SuperMApo. α-SMA expression was strongly increased by SuperMApo and, at the same time, mRNA levels of extracellular matrix-associated genes coding for fibronectin, type I and III collagens, were decreased ( Figures 3A, B ). SuperMApo-treated mouse fibroblasts demonstrated enhanced proliferative and migratory properties ( Supplemental Figures 1A, B ). This was further confirmed using a scratch assay where these fibroblasts showed enhanced wound healing capacities ( Figure 3C ). In addition, using floating collagen gels, SuperMApo-treated fibroblasts demonstrated increased contraction capacities ( Figure 3D ). Thus, in vitro, primary fibroblasts isolated from colon demonstrated higher healing properties in the presence of pro-resolving factors released by macrophages after efferocytosis.

We then analyzed the effects on IEC using the MODE-K intestinal immortalized epithelial cell line. We observed a similar improvement of healing functions, such as increased proliferative and migratory capacities, as well as increased wound closure properties ( Supplemental Figures 1D–F ). In addition to the gain in healing properties, SuperMApo allowed MODE-K epithelial cells to acquire phagocytic capacities to eliminate apoptotic cells ( Figure 3E ), as previously demonstrated in the lung mucosa (33). These data indicate that pro-resolving factors produced by efferocytic macrophages trigger wound healing properties in fibroblasts and IEC, providing IEC with phagocytic activity as well.

SuperMApo-treated mouse fibroblasts demonstrate all the features of activated fibroblasts but without overexpressing extracellular matrix-associated genes. This suggests that fibroblasts acquired a specific pro-resolving profile. These properties were therefore evaluated on human fibroblasts isolated from colon biopsies (from inflamed lesions and control tissues of patients diagnosed with IBD). In the presence of human SuperMApo (i.e., the secretome of human macrophages that have eliminated human apoptotic cells), the wound healing functions of control fibroblasts (i.e., isolated from control tissue) were strongly increased including higher proliferative properties and higher wound closure capacities ( Figures 3F, G ). Concerning inflamed lesion-derived fibroblasts, proliferative properties and wound closure capacities were also increased and less modestly ( Figures 3F, G ). These data demonstrate that pro-resolving factors released by efferocytic macrophages could restore the healing functions of IBD fibroblasts, and thus support the use of SuperMApo in IBD patients.

The enhanced healing functions induced by pro-resolving factors released by efferocytic macrophages were then addressed in vivo using a biopsy forceps-wounded colonic mucosa model in which inflammation is localized to the injured tissue and not systemic. In this model, i.p. injection of SuperMApo on the day of mucosa injury significantly accelerated the lesion wound healing observed by video endoscopy, compared to the lesions of control vehicle-treated mice ( Figure 4A ). We addressed the integrity of the intestinal barrier after naive T cell transfer-induced colitis by using FITC-Dextran oral administration and its quantification in the mouse serum. Reduced systemic levels of FITC-Dextran were found in colitis mice receiving SuperMApo compared to control mice ( Figure 4B ). This attested for a conserved barrier permeability, further supported by a lower MEICS score ( Figure 4C ). In addition, the levels of the intestinal barrier integrity marker, the regenerating islet-derived protein 3γ (REG3γ) (34, 35), were also found reduced in colitis mice treated with SuperMApo ( Figure 4D ). These results suggest a restored intestinal barrier permeability by administration of pro-resolving factors. This was associated with a marked in vivo proliferation of cells within the crypts and activation of fibroblasts in the colon mucosa of colitis mice receiving SuperMApo, as attested by increased Ki67⁺ cell number per mucosal crypt, and with an increased expression of α-SMA in colonic fibroblasts, respectively ( Figures 4E, F ). Moreover, a slight decrease of colon mRNA expression of extracellular matrix-associated genes Fn1, Col1a1, Col3a1 and Tgf-β was observed in colitis mice receiving SuperMApo treatment. This decrease was significant for Fn1 ( Figure 4G ). Overall, our data demonstrate that treatment of mice experiencing ongoing colitis with pro-resolving factors released by efferocytic macrophages restores intestinal barrier permeability by inducing IEC proliferation and activating local fibroblasts in vivo.

Several pro-resolving factors have been shown to be released by macrophages after efferocytosis (36, 37). Thus, different growth factors, including EGF, FGF, TGF-β, IGF-I and VEGF, were quantified in the SuperMApo supernatant. TGF-β, IGF-I and VEGF were detected in significant higher quantities than in control supernatants, whereas EGF and FGF were not detectable ( Figure 5A ). When blocking antibodies directed against TGF-β, IGF-I and VEGF were used, we observed a non-significant decrease of SuperMApo-induced α-SMA expression ( Supplemental Figure 2A ). The reduction of SuperMApo-induced Col1a1, Col3a1 and Fn1 mRNA expression in primary fibroblasts was not significantly affected by growth factor blockade ( Supplemental Figure 2B ). The proliferative and contractive properties of mouse primary fibroblasts were not affected by antibodies blocking these growth factors ( Supplemental Figures 2C, D ). However, blockade of TGF-β or IGF-I significantly reduced the proliferation of IEC line MODE-K ( Figure 5B ). In addition, SuperMApo-enhanced cell migration of fibroblasts and IEC was reduced by the addition of blocking antibodies whatever the neutralized growth factor (i.e., TGF-β, IGF-I or VEGF) ( Figure 5C ). Thus, while proliferation and migration of IEC induced by SuperMApo are sustained by TGF-β, IGF-I and VEGF growth factors, only migration of fibroblast is associated to these growth factors.

The role of TGF-β, IGF-I and VEGF growth factors was further addressed in vivo using the T cell-transfer colitis model in which all the growth factors tested in vitro were infused together. While colon lesions were not ameliorated – as attested by a similar MEICS score between growth factor-treated and control mice ( Figure 6A ), weight loss and the composite clinical score were slightly improved by the treatment with the three growth factors ( Figure 6B ). We next decided to deplete the SuperMApo supernatant from TGF-β, IGF-I and VEGF using antibody-coupled microbeads before its injection to colitis mice. Depletion of TGF-β, IGF-I and VEGF from SuperMApo inhibited the improvement of the MEICS score induced by SuperMApo ( Figure 6C ). However, the absence of TGF-β, IGF-I and VEGF in the SuperMApo supernatant partially reduced the properties of SuperMApo in terms of weight loss prevention and clinical score reduction ( Figure 6D ). These data demonstrate that TGF-β, IGF-I and VEGF participate in intestinal mucosal healing induced by SuperMApo but are not sufficient to resolve global IBD inflammation.

Altogether, these results highlight pro-resolving factors released by efferocytic macrophages as a promising therapeutic approach to both circumvent inflammation and promote lesion healing in the setting of inflammatory bowel diseases.