Tolerogenic effect of mouse fibroblasts on dendritic cells

Summary

There is controversy about the immunomodulatory effect of fibroblasts on dendritic cells (DCs). To clarify this issue, in this study, we have evaluated different features of fibroblast‐primed DCs including their ability to express co‐inhibitory and co‐stimulatory molecules, pro‐inflammatory and anti‐inflammatory cytokines and their ability to induce T‐cell proliferation. We also examined migratory capacity of DCs to lymphatic tissues and present fibroblast‐derived antigens after encountering fibroblasts. The results of our in vitro study showed that both co‐inhibitory (programmed death ligand 1 and ligand 2 and B7H4) and co‐stimulatory (CD86) molecules were up‐regulated when DCs were co‐cultured with fibroblasts. In an animal model, we showed that intra‐ peritoneal injection (IP) of both syngeneic and allogeneic fibroblasts significantly increased both total DC count and expression level of co‐inhibitory and co‐stimulatory molecules on DCs. Priming of DCs with syngeneic and allogeneic fibroblasts reduced the proliferation of CD4⁺ and CD8⁺ T cells. Even activation of fibroblast‐ primed DCs failed to restore their ability to induce T‐cell proliferation. Likewise, priming of DCs with fibroblasts blocked the ability of ovalbumin‐pulsed DCs to induce proliferation of ovalbumin‐specific CD4⁺ T cells. Compared with non‐activated DCs, fibroblast‐primed DCs had significantly higher expression levels of interleukin‐10 and indoleamine 2, 3 dioxygenase. Fibroblast‐primed DCs had a significantly reduced interleukin‐12 expression level compared with that of activated DCs. After priming with fibroblasts, DCs were able to migrate to lymphatic tissues and present fibroblast‐derived antigens (ovalbumin). In conclusion, after priming with fibroblasts, DCs gain tolerogenic features. This finding suggests the potential role of fibroblasts in the maintenance of immune tolerance.

Introduction

Dendritic cells (DCs) are the most important antigen‐presenting cells in the body and are responsible for the activation of T cells against foreign antigens as well as regulating the immune response against self‐molecules. It is well known that cytokines secreted from other types of cells affect the fate of DCs and, ultimately, the subtypes of induced T cells.1 Skin is a barrier to protect the body from foreign antigens, but they can enter when this barrier is breached. Langerhans cells and other DCs in the skin engulf these antigens and present them to the naive T cells located within the skin draining lymph nodes.2 On the other hand, dermal fibroblasts are the major cells in proximity to DCs during their migration toward lymph nodes. There is a controversy in the literature about the effect of fibroblasts on DCs. Some studies support a stimulatory effect of fibroblasts on DCs. A series of studies by Simon's group shows that fibroblasts can induce maturation of DCs,3 promote migration of DCs4 and support expansion of interleukin‐17 (IL‐17) ‐producing T cells.5 On the other hand, there are some clues to suggest immunoregulatory effects of fibroblasts on DCs. For instance, it has been shown that fibroblasts can reduce the production of IL12‐p70 upon activation of DCs.6 Also, it has been shown that cell–cell contact with fibroblasts is required for the suppression of MHC class II and CD40 on DCs, whereas soluble prostaglandin E₂ (PGE₂) released by fibroblasts is responsible for the suppression of production of IL‐12 and TNF‐α by DCs.7

So far, we know that the effect of other types of stromal cells on DCs is to some extent tolerogenic. For instance, mesenchymal stem cells (MSCs) are known for their tolerogenic effects on DCs. It has been shown that MSCs can impair the ability of ovalbumin‐pulsed DCs in vivo to induce ovalbumin‐specific T‐cell proliferation. However, this effect was mostly attributed to decreased migration of ovalbumin‐pulsed DCs injected subcutaneously to the mouse recipients of MSCs.8 Also, it has been shown that haematopoietic stem‐cell‐derived DCs cultured in vitro with MSCs are defective in stimulating alloreactive T cells and this effect is attributed to activation of Notch pathway in DCs.9 MSC‐treated mature DCs gain an immature phenotype and that cannot be reversed by lipopolysaccharide activation.10 The MSCs from sources other than bone marrow, including umbilical cord blood, can have a similar impact on DCs.11 Further, renal fibroblasts induce increased expression of co‐inhibitory molecules, B7H1 and B7DC, on DCs, in addition to decreased expression of IL‐12. Also, renal fibroblast‐conditioned DCs are less potent in stimulating T‐cell proliferation.12

As reviewed above, there is no consensus about the effect of fibroblasts, which reside in almost all tissues, on DCs. Understanding more about this interaction might shed light on the physiology of cell–cell interactions in the skin. Besides, it can help us to design cell therapy approaches to treat skin‐related diseases. Here, we therefore hypothesize that conditioning of DCs with dermal fibroblasts can convert the immature DCs to tolerogenic ones. Tolerogenic DCs are a subset of DCs that are responsible for regulating the immune responses.13 The main features of these cells include high expression levels of co‐inhibitory molecules, IL‐10 and indoleamine 2,3‐dioxygenase (IDO) and low expression of IL‐12, the ability to induce anergy in effector T cells, expansion of regulatory T cells, and migratory capacity to T‐cell areas in secondary lymphoid tissues.14 In this study, we tried to test these features in DCs upon encountering the fibroblasts.

Materials and methods

Mice

Eight‐week‐old female C3H/Hej (C3H) and C57BL/6 (B6) mice were purchased from Jackson Laboratories (Bar Harbor, ME) and kept in Blusson Spinal Cord Injury Center Animal Care Facility. OT‐II mice and Thy1.1 congenic C57BL/6 mice purchased from Jackson Laboratories (Sacramento, CA) were bred and maintained as OT‐II × Thy1.1 homozygous mice at the Child & Family Research Institute Animal Care Facility. The OT‐II mice have CD4 T cells that harbour a T‐cell receptor that recognizes ovalbumin. Mice expressing a membrane‐bound chicken ovalbumin gene under the direction of the chicken β‐actin promoter (Act‐mOVA) were purchased from Jackson Laboratories. Care and maintenance of all animals were in accordance with the principles of laboratory animal care and the guidelines of the institutional Animal Policy and Welfare Committee.

Isolation and culture of dendritic cells and fibroblasts

To culture DCs, we followed the method established by Lutz et al.15 After removing the surrounding tissues from femur bones in sterile conditions, both ends of the bones were cut and the marrow was flushed out. Two million cells were plated in non‐adherent 60 × 15‐mm dishes (VWR, Edmonton, Canada) in RPMI‐1640 medium containing 10% fetal bovine serum (FBS; Hyclone, Logan, UT), 1% penicillin/streptomycin (Invitrogen, Carlsbad. CA) and granulocyte–macrophage colony‐stimulating factor (GM‐CSF; eBioscience, San Diego, CA). At days 3, 6 and 8, half of the medium in each dish was taken and centrifuged at 500 g for 5 min. The supernatant was discarded and the pellet was suspended in fresh medium containing GM‐CSF and added to the same dishes. At day 10, naive DCs are ready for characterization or co‐culturing studies. DCs characterized by expression of both CD11c and MHC II molecules comprise > 80% of cells in suspension.

To culture fibroblasts, after removing the hairs with hair clippers and sterilization of back skin with povidone iodine, a 2 × 2‐cm piece of full thickness skin was taken from the euthanized mouse. Washed several times in PBS containing 2% penicillin/streptomycin, the hypodermis fat tissue was removed with a sterile blade in sterile condition. The skin was cut into small pieces and each piece was put on the 150 × 25‐mm culture dish (Corning Inc., Corning, NY) with the dermis facing down. Small drops of FBS were placed on each piece of skin and the dish was kept in a 37° incubator supplemented with 5% CO₂ for 4 hr. Afterwards, Dulbecco's modified Eagle's medium (Hyclone) containing 10% FBS and 1% penicillin/streptomycin was added very slowly to the dish; care was taken to avoid detachment of the pieces from dish surface. After 1–2 weeks, the fibroblasts migrate out of the skin and adhere to the dish surface; the cells can be sub‐cultured. At passages 5–9, the fibroblasts were used for co‐culturing studies.16

Conditioning of dendritic cells

After 10 days of culturing the DCs from bone marrow cells with the mentioned protocol, the non‐adherent cells in the dishes were collected for further conditioning in different groups. For conditioning, 2 × 10⁶ DCs were cultured in 6 ml of RPMI‐1640 medium containing 10% FBS, 1% penicillin/streptomycin and GM‐CSF in non‐adherent dishes for another 48 hr. To activate the DCs, 200 ng/ml of the Toll‐like receptor 9 agonist CpG oligodeoxynucleotide 1826 (5′‐TCCATGACGTTCCTGACGTT‐3′) (custom synthesized by Sigma‐Aldrich, Oakville, ON, Canada), was added to the culture medium.17 For co‐culturing with fibroblasts, 2 × 10⁵ fibroblasts were seeded in non‐adherent 6‐cm dishes 4 days before adding DCs.

B6 cell lysate was added to C3H DCs to enable them to present B6‐derived antigens to B6‐pre‐sensitized C3H T cells.18 To prepare the cell lysate, a pellet of 10⁷ B6 bone marrow cells resuspended in 100 μl PBS was frozen in −80° and thawed four times, followed by passing through a 70‐μm cell strainer (Fisher Scientific, Pittsburgh, PA) to remove the big chunks. From this solution, 10 μl was added to each dish of C3H DCs. In experiments where presentation of chicken albumin on B6 DCs was needed, 1 mg/ml of ovalbumin protein (Sigma) was added to each dish of B6 DCs for 24 hr and then washed out.19

Co‐culturing of DCs/lymphocytes and T‐cell proliferation assay

Non‐adherent cells from each dish of conditioned DCs were collected and treated with 50 μg/ml of mitomycin C (Sigma) for 30 min. Then, 5 × 10⁴ DCs suspended in 100 μl media were seeded in round‐bottom 96‐well plates (Corning Inc.).

Skin draining lymph node cells from B6‐pre‐sensitized C3H and OT‐II mice were used as responder cells. Twenty million lymphocytes were stained with 2·5 μg/ml of carboxyfluorescein succinimidyl ester (CFSE) for 10 min. After washing the cells, 2 × 10⁵ lymphocytes in 100 μl of medium was added to DCs. RPMI‐1640 medium containing 10% FBS, 1% penicillin/streptomycin and 30 U/ml of IL‐2 (Roche, Mannheim, Germany) was used for co‐culturing of DCs and lymphocytes and was changed every other day.20 After 6 days, the cells were stained with CD4, CD8 and 7‐aminoactinomycin D antibodies and read with FACS to assess T‐cell proliferation.

Flow cytometry and kynurenine assay

All flow cytometry analyses were performed with a BD Accuri C6 Flow Cytometer (BD Biosciences, San Jose, CA). All the antibodies were bought from eBioscience. For characterization of C3H DCs, they were stained with FITC‐conjugated anti‐mouse CD11c (1 : 200) and peridinin chlorophyll protein‐eFluor 710‐conjugated anti‐mouse MHC II (I‐AK) (1 : 200) antibodies. B6 DCs were stained with FITC‐conjugated anti‐mouse CD11c (1 : 200) and allophycocyanin (APC) ‐conjugated anti‐mouse MHC II (I‐Ab) (1 : 200) antibodies. Co‐stimulatory and co‐inhibitory antibodies that were used to characterize the DCs included phycoerythrin (PE) ‐conjugated anti‐mouse CD86 (1 : 200), PE‐conjugated anti‐mouse CD274 (PD‐L1 or B7H1) (1 : 200), PE‐conjugated anti‐mouse CD273 (PD‐L2 or B7DC) (1 : 200) and PE‐conjugated anti‐mouse B7‐H4 (1 : 200).

For T‐cell proliferation assays, 6 days after co‐culturing with DCs, lymphocytes were stained with PE‐conjugated anti‐mouse CD4 (1 : 1000), APC‐conjugated anti‐mouse CD8 (1 : 300) antibodies and 7‐aminoactinomycin D (1 : 75) and read with FACS.

To track the green fluorescent protein‐positive (GFP⁺) DCs, 3 days after injection of 10⁷ C3H fibroblasts into the peritoneal cavity of B6 GFP⁺ mice, the peritoneal lavage cells were re‐injected into the peritoneal cavity of B6 non‐GFP mice. Three and 10 days after injection, the recipient mice were killed, their peritoneal lavage and mesenteric lymph node (MLN) cells were stained with APC‐conjugated anti‐mouse CD11c (1 : 200) antibody and read with FACS to track the GFP⁺ DCs.

To check the presentation of ovalbumin on B6 DCs after in vitro co‐culturing with B6 and Act‐mOVA fibroblast or after intraperitoneal injection of B6 and Act‐mOVA to B6 recipients, co‐cultured DCs and peritoneal lavage and MLN cells of the recipient mice were stained with APC‐conjugated anti‐mouse CD11c (1 : 200) and PE‐conjugated MHC I/OVA (H‐2Kb/SIINFEKL) (1 : 200) and read with FACS.

After treating B6 and C3H DCs in different groups for 48 hr, the supernatants were collected and then the concentration of mouse IL‐12 p70 and IL‐10 was determined by BD Cytometric Bead Array kit (CBA) (BD Biosciences) according to the manufacturer's instructions using a C6 Accuri flow cytometer (BD Biosciences). Also, kynurenine (by‐product of IDO enzyme activity) level was measured in the supernatants as previously described.21 Briefly, proteins in the conditioned medium were precipitated by trichloroacetic acid. After centrifugation, 0·5 ml of supernatant was incubated with an equal volume of Ehrlich's reagent at room temperature for 10 min. The reaction mixture was measured spectrophotometrically at 490 nm. The concentration of kynurenine in the conditioned medium was calculated according to a standard curve of defined kynurenine concentration.

Quantitative PCR

After treating B6 and C3H DCs in different groups for 48 hr, total RNA was isolated and purified using an RNeasy kit (Qiagen, Valencia, CA) following the manufacturer's instructions. The concentration and purity of the extracted RNA were checked using a Nano Drop 2000 Spectrophotometer (ThermoScientific, Waltham, MA USA). The absorbance ratio at 260/280 nm was checked as a measure of purity of RNA. After DNase I treatment (Invitrogen), cDNA was synthesized from 1 mg of total RNA using a Superscript II First Strand cDNA Synthesis kit (Invitrogen). Quantitative PCR was performed on an Applied Biosystems 7500 PCR machine using the following PCR cycling conditions: 95° for 5 min, 40 cycles at 95° for 15 seconds and 60° for 1 min. U6 snRNA expression was measured as the house keeping gene (F: 5′‐ctcgcttcggcagcaca‐3′, R: 5′‐aacgcttcacgaatttgcgt‐3′). The following primers were used for quantitative PCR: mouse IL‐10 (F: 5′‐cagccgggaagacaataacg‐3′, R: 5′‐ccgcagctctaggagcatg‐3′), mouse IL‐12p35 (F: 5′‐gtgattctgaagtgctgcgt‐3′, R: 5′‐ctttgatgatgaccctgtgc‐3′), Mouse IDO (F: 5′‐aagggcttcttcctcgtctc‐3′, R: 5′‐aaaaacgtgtctgggtccac‐3′) (Invitrogen). CYBR Green dye (Roche) was used for detection of PCR products and relative quantity of expression of each gene was measured by ΔΔC_T method. Non‐activated B6 DC sample was assigned as the reference sample and fold change of all the samples was compared with this one.

Florescence microscopy

To confirm the existence of GFP⁺ DCs in MLNs of recipient non‐GFP B6 mice, the lymph nodes were frozen in cryomatrix (ThermoScientific). Using the cryostat (ThermoScientific), 5‐μm sections were obtained. The slides were fixed for 15 min in 1 : 1 acetone/ethanol followed by 15 min in PBS. The slides were visualized using a Zeiss fluorescent microscope. Pictures were taken through axiovision 4 software by an AxioCam camera (Carl Zeiss, Oberkochen, Germany).

Statistical analysis

Graphs are generated using graphpad prism software version 5·04 (GraphPad, San Diego, CA). Data are shown as mean ± SD of three or more observations. Statistical analysis was performed using SPSS Statistics software version 22 (SPSS Inc., Chigaco, IL). To perform analysis where two groups were compared, a paired Student's t‐test was applied. To compare more than two groups, one‐way analysis of variance with post‐hoc evaluation of Tukey was used. A P‐value < 0·05 was considered statistically significant.

Results

Co‐inhibitory and co‐stimulatory molecules are increased on DCs upon co‐culturing with dermal primary fibroblasts

Bone‐marrow‐derived DCs (BMDCs) from B6 and C3H mice were cultured for 48 hr with or without fibroblasts from B6 mice. Afterwards, non‐adherent cells were collected and DCs were checked for the expression of co‐inhibitory and co‐stimulatory molecules. B6 DCs, which were cultured with syngeneic fibroblasts, had a significantly higher expression of co‐inhibitory molecules, programmed death ligand 1 (PD‐L1) and ligand 2 (PD‐L2). Also, C3H DCs that were cultured with allogeneic fibroblasts had elevated expression of the co‐inhibitory molecule PD‐L1. In both groups, co‐stimulatory molecule, CD86, was also significantly increased (Fig. 1).

Figure 1.

Evaluation of co‐inhibitory and co‐stimulatory molecules on dendritic cells (DCs) upon co‐culturing with primary dermal fibroblasts. B6 and C3H bone marrow‐derived (BM) DCs were cultured in the absence (control) or presence (B6‐fib) of B6 fibroblasts for 48 hr. Representative flow cytometry plots are shown on the top row of this figure. After gating the cells based on the size (forward scatter; FSC) and granularity (side scatter; SSC), cells having CD11c and MHC II were identified as DCs and the mean fluorescence intensity (MFI) of co‐inhibitory and co‐stimulatory molecules were checked on DCs. Lower row graphs summarize the MFI of co‐inhibitory molecules, programmed death protein ligand 1 (PD‐L1) and ligand 2 (PD‐L2) and B7H4, and also the co‐stimulatory molecule, CD86, on B6 (left) and C3H DCs (right). Data are presented as the mean ± SD of at least three separate experiments. P < 0·05 was considered significant.

To validate this finding in vivo, we injected 10⁶ B6 fibroblasts in Dulbecco's modified Eagle's medium plus 10% FBS as the vehicle (Control) into the peritoneal cavity of B6 and C3H mice. After 10 days, the peritoneal lavage cells were collected and checked for the expression of co‐inhibitory and co‐stimulatory molecules on DCs. The DCs were characterized as CD11c⁺ MHC II⁺ cells. The result showed that both syngeneic and allogeneic fibroblasts caused a significant increase in total DC count within the peritoneal cavity compared with that of control. The mean florescence intensity (MFI) of co‐inhibitory molecule, PD‐L1 significantly increased on DCs in both syngeneic and allogeneic fibroblast groups. However, the levels of PD‐L2 and B7H4 were only increased in the allogeneic fibroblast‐treated group. Although the MFIs of CD80 in the allogeneic fibroblast group and of CD40 and CD86 in both allogeneic and syngeneic fibroblast groups were higher than in the control group, this difference was not statistically significant. Only the MFI of CD40 was significantly increased on DCs in the syngeneic fibroblast treated group (Fig. 2).

Figure 2.

Detection of co‐inhibitory and co‐stimulatory molecules on dendritic cells (DCs) isolated from peritoneal cavity of the mice that received intraperitoneal injection of fibroblasts. Ten days after intraperitoneal injection of B6 fibroblasts or control vehicle to C3H and B6 mice, the peritoneal lavage cells were analysed by flow cytometry. (a) After gating for size (forward scatter; FSC) and granularity (side scatter; SSC), CD11c⁺ MHC II ⁺ cells were gated and defined as DCs. The mean fluorescence intensity (MFI) of co‐inhibitory and co‐stimulatory molecules was measured on DCs. The next panels show the results of statistical analysis on data collected from all repeats of the experiments. (b) Total peritoneal lavage (PL) DCs were counted in different groups using counting beads. (c–h) The MFI of co‐inhibitory molecules, programmed death protein ligand 1 (PD‐L1) and ligand 2 (PD‐L2) and B7H4, and also the co‐stimulatory molecules, CD80, CD86 and CD40 on PL DCs are shown, respectively. Data are presented as the mean ± SD of at least three separate experiments. P < 0·05 was considered significant.

Direct and indirect pathways of antigen presentation in DCs were impaired upon priming with fibroblasts

To investigate whether contact with fibroblasts changes the ability of DCs to induce T‐cell proliferation, a series of experiments was conducted. After in vitro culturing of B6 and C3H BMDCs, they were divided into four groups. B6 DCs were cultured alone, treated with CpG, co‐cultured with B6 fibroblasts, or co‐cultured with fibroblast and CpG. C3H DCs were treated in the same manner. Moreover, B6 bone marrow cell lysate was also added to all groups of C3H DCs to enable them to present B6‐derived antigens. After 48 hr of incubation, DCs were further treated with mitomycin‐C and used as stimulator cells. Lymphocytes of skin draining lymph nodes of B6‐pre‐sensitized C3H mice (received B6 skin transplant 1 month before) were stained with CFSE and used as responder cells. After 6 days of mixed co‐culture, the cells were analysed by flow cytometry to check their proliferative status. As shown in Fig. 3(a), only CpG‐activated DCs were able to induce T‐cell proliferation in CD4 and CD8 T cells. Priming of DCs from both B6 and C3H mice with B6 fibroblasts resulted in reduced proliferation of CD4 and CD8 T cells. Even simultaneous activation with CpG could not restore the ability of DCs to induce T‐cell proliferation. As B6 DCs can directly activate allogeneic C3H lymphocytes, we considered it a direct pathway of antigen presentation. However, C3H DCs present B6‐derived antigen to syngeneic B6‐pre‐sensitized C3H lymphocytes in an indirect and/or semi‐direct pathway. This finding shows that both direct and indirect pathways of T‐cell activation are impaired upon priming of DCs with fibroblasts. However, it cannot be excluded that presentation of B6‐derived antigens on C3H DCs has happened through a semi‐direct pathway rather than an indirect one.

Figure 3.

T‐cell proliferation in response to fibroblast‐conditioned dendritic cells (DCs). (a) Lymphocytes from B6‐pre‐sensitized C3H lymph nodes were stained with CFSE as responders and cultured for 6 days with mitomycin‐C‐treated B6 and C3H bone marrow‐derived (BM) ‐DCs as stimulators. T‐cell proliferation was measured in the following DC treatment groups: (i) non‐activated DCs (DC alone, black lines in histograms), (ii) CpG‐activated DCs (DC + CpG, blue lines), (iii) B6‐fibroblast‐primed DCs (DC + fib, red lines), and (iv) CpG‐activated and B6‐fibroblast‐primed DCs (DC + fib + CpG, green lines). After gating for forward scatter–side scatter (not shown) and live cells (7AAD ⁻ population), CD4 and CD8 cells were identified and proliferation of these T‐cell subsets was compared in the four above‐mentioned groups of B6 and C3H DCs. Graphs summarize the results of a T‐cell proliferation assay from three independent experiments for both B6 and C3H stimulators. Proliferation of CD4 and CD8 T cells are shown separately. (b) Ovalbumin‐specific lymphocytes from OT‐II mice were stained with CFSE and cultured for 6 days with mitomycin‐C‐treated B6 BM‐DCs from five different pre‐conditioned sets: (i) non‐pulsed non‐activated DCs (black), (ii) ovalbumin‐pulsed non‐activated DCs (brown), (iii) ovalbumin‐pulsed CpG‐activated DCs (blue), (iv) ovalbumin‐pulsed B6‐fibroblast‐primed DCs (red), and (v) ovalbumin‐pulsed, B6‐fibroblast‐primed and CpG‐activated DCs (green). After gating the live CD4 T cells (CD4⁺ 7AAD ⁻), the CFSE dilution was checked as a measure of T‐cell proliferation. T cells in both ovalbumin‐pulsed DC groups either with (blue) or without (brown) CpG activation proliferated extensively.

Ovalbumin‐pulsed activated DCs fail to induce proliferation of OTII T cells upon fibroblast priming

To further confirm the previous finding, we examined whether fibroblast priming of DCs can arrest the ability of DCs to induce proliferation of antigen‐specific T cells. Similar to the previous experiment, B6 DCs were cultured in vitro and treated in five different groups: (i) non pulsed non‐activated, (ii) ovalbumin‐pulsed non‐activated, (iii) ovalbumin‐pulsed CpG‐activated, (iv) ovalbumin‐pulsed and B6‐fibroblast‐primed, and (v) ovalbumin‐pulsed, B6‐fibroblast‐primed, and CpG‐activated. Lymphocytes of OTII mice, which contain ovalbumin‐specific CD4 T cells, were used as responder cells. After CFSE staining and co‐culturing with DCs from the above‐mentioned groups, proliferation of responder cells was checked using flow cytometry. Similar to the previous experiment, priming with fibroblasts inhibited the ability of ovalbumin‐pulsed DCs to induce the proliferation of ovalbumin‐specific CD4 T cells (Fig. 3b). Although the DCs were pulsed with ovalbumin, they could induce proliferation of OTII cells both in the presence and in the absence of CpG activation. However, after being co‐cultured with fibroblasts, they lost their ability to induce OTII cell proliferation. Similarly, the ability of DCs to induce cell proliferation was not restored even with CpG activation.

Elevated expression of anti‐inflammatory cytokines and reduced expression of pro‐inflammatory cytokine in fibroblast‐primed DCs

To further characterize fibroblast‐conditioned DCs, their expressions of pro‐ and anti‐inflammatory cytokines were checked using real‐time quantitative PCR. B6 and C3H BMDCs were divided into four groups, including non‐activated, CpG‐activated, B6‐fibroblast‐primed, and CpG‐activated plus B6‐fibroblast‐primed. As shown in Fig. 4, compared with non‐activated DCs, both B6 and C3H fibroblast‐primed DCs had significantly higher mRNA and secreted protein levels of IL‐10. The IDO mRNA level was only significantly elevated in fibroblast‐primed B6 DCs, whereas kynurenine (a by‐product of IDO enzyme activity) level was significantly increased in the fibroblast‐primed C3H DC group. Upon activation, both B6 and C3H DCs had significantly higher amounts of the pro‐inflammatory cytokine IL‐12. However, priming with fibroblasts significantly reduced the expression of IL‐12 even upon CpG activation (Fig. 4c).

Figure 4.

Cytokine expression in dendritic cells (DCs) cultured in the presence or absence of fibroblasts. B6 and C3H bone marrow‐derived (BM) ‐DCs were cultured in four different groups for 48 hr: (i) non‐activated DCs (DC), (ii) CpG‐activated DCs (activated DC), (iii) B6‐fibroblast‐treated DCs (DC + fib), and (iv) CpG‐activated and B6‐fibroblast‐primed DCs (activated DC + fib). (a) The mRNA expression levels of interleukin‐12 (IL‐12), IL‐10 and indoleamine 2,3‐dioxygenase (IDO) in different groups of treated DCs are shown in three graphs from top to bottom (n = 3). The relative quantity (RQ) of each cytokine is calculated by normalizing to the amount for the B6 non‐activated DC group using the ΔΔC_T method. U6 was used as the internal control. (b) Supernatants of DC culture dishes in different groups were collected after 48 hr of conditioning and secreted levels of IL‐12P70 and IL‐10 were measured by the BD Cytometric Bead Array kit. The level of kynurenine (by‐product of IDO enzyme activity) was measured by spectrophotometric kynurenine assay. Graphs are shown respectively from top to bottom. Data are presented as the mean ± SD of at least three separate experiments. P < 0·05 was considered significant.

DCs are able to migrate to lymphatic tissues upon priming with fibroblasts

To test the ability of DCs to migrate to the lymphatic tissues after being primed with fibroblasts, we designed the following experiment. Ten million C3H fibroblasts were injected into the peritoneal cavity of GFP B6 mice. After 2 days, the peritoneal cavity was washed and the peritoneal lavage cells were injected into the peritoneal cavity of non‐GFP B6 mice. After 3 and 10 days, peritoneal lavage and MLN cells of these mice were checked by flow cytometric analysis to track the GFP⁺ DCs. Further, fluorescence microscopy was used to locate GFP⁺ cells within the frozen sections of the recipients' MLNs. As presented in Fig. 5, GFP⁺ CD11c⁺ DCs were present in peritoneal lavage and also MLNs of recipient me 3 days after injection, but not after 10 days. This finding confirms that after priming with fibroblasts, DCs can migrate to lymphatic tissues.

Figure 5.

Evaluating the ability of dendritic cells (DCs) to migrate to the lymph nodes after priming with fibroblasts. To test the migratory capacity of DCs after priming with fibroblasts, C3H fibroblasts were injected into the peritoneal cavity of the B6 GFP mice. After 2 days, the peritoneal lavage cells were injected into the peritoneal cavity of non‐GFP B6 mice. Three and 10 days later, the GFP ⁺ cells were tracked in the peritoneal cavity and mesenteric lymph nodes (MLNs) of recipient mice. (a) 3 days (upper plots) and 10 days (lower plots) after intraperitoneal injection of GFP ⁺ cells, peritoneal lavage (left plots) and MLN cells (right panels) were stained for CD11c and tested with flow cytometry. (b) The migration of GFP ⁺ cells was further confirmed by detecting them in MLNs using fluorescence microscopy.

DCs are able to present fibroblast‐derived antigens

To investigate the ability of DCs to present fibroblast‐derived antigens, B6 or Act‐mOVA fibroblasts were co‐cultured in vitro with B6 DCs. After 48 hr of co‐culturing, DCs were checked for presentation of ovalbumin on their MHC I. A percentage of Act‐mOVA fibroblast‐primed DCs stained positive with the antibody that detects MHC I/ovalbumin (Fig. 6a). Likewise, to see whether DCs can present fibroblast‐derived antigens in draining lymph nodes in vivo, 10⁶ B6 or Act‐mOVA fibroblasts were injected into the peritoneal cavity of B6 mice. After 3 days, MLN cells were checked with flow cytometry to detect CD11c⁺ DCs that present MHC I/ovalbumin. Positive DCs were found in the Act‐ovalbumin fibroblast‐ treated group but were not present in the B6‐fibroblast‐treated group (Fig. 6b).

Figure 6.

Evaluating the capacity of dendritic cells (DCs) to present fibroblast‐derived antigens in the context of MHC I. (a) B6 DCs were cultured in the presence of B6 fibroblasts (left plot) or β‐actin–ovalbumin (Act‐OVA) fibroblasts (right plot). CD11c⁺ DCs that present ovalbumin on MHC I were detected with flow cytometry. (b) Similarly, after injection of B6 fibroblasts (left plot) or Act‐OVA fibroblasts (right plot) into the peritoneal cavity of B6 mice, CD11c⁺ DCs that present ovalbumin on MHC I were tracked in mesenteric lymph nodes using flow cytometry.

Discussion

In this study, we address the potential role of fibroblasts in the tolerogenicity of DCs. We hypothesized that priming with fibroblasts has a tolerogenic impact on DCs. To test this hypothesis, we tested different characteristics of tolerogenic DCs in BMDCs after being primed with fibroblasts. These features include the expression of co‐inhibitory molecules and anti‐inflammatory cytokines, the ability to migrate to lymphatic organs and present antigens, and induction of anergy in T cells.

The major signals needed for the activation of T cells include (but are not limited to) the interaction between the T‐cell receptor and MHC–peptide complex and the interaction of co‐stimulatory molecules on antigen‐presenting cells with their receptors on T cells.22 Although co‐stimulation helps the activation of DCs upon introduction of any danger signal, co‐inhibitory molecules help to maintain tolerance to self‐antigens.23 PD‐L1 and PD‐L2 are among the most important co‐inhibitory molecules that both bind to PD‐1 on T cells, which in turn starts an inhibitory signalling pathway in T cells. Expression of PD‐L2 is restricted to DCs, macrophages and B1 cells,24 whereas PD‐L1 is expressed on a wider variety of cells.23 B7H4 is another co‐inhibitory molecule of the B7 family that is expressed on antigen‐presenting cells and its receptor is still unknown.25 Our results showed that these co‐inhibitory molecules, as well as the co‐stimulatory molecule CD86, are up‐regulated on DCs co‐cultured with fibroblasts. We also confirmed this in an in vivo system by injecting the fibroblasts into the peritoneal cavity of recipient mice and checking these molecules on retrieved peritoneal lavage DCs. We found an influx of CD11c⁺ MHC II⁺ DCs into the peritoneal cavity upon injection of either syngeneic or allogeneic fibroblasts. Both syngeneic and allogeneic fibroblast groups up‐regulated co‐inhibitory and co‐stimulatory molecules on DCs. Although co‐stimulatory molecules were up‐regulated on DCs upon priming with fibroblasts, we believe the ratio of co‐inhibitory to co‐stimulatory molecules was in favour of the immunomodulatory state for DCs. Indeed, tolerogenic DCs comprise a spectrum of DCs with different maturation status, which includes immature DCs and other type of DCs with higher levels of co‐stimulation.26 Therefore, the expression of co‐stimulatory molecules on DCs does not per se prove their stimulatory nature.

As one of the main functions of DCs is to stimulate the proliferation of naive T cells, we have asked whether conditioning of DCs with fibroblasts could attenuate this function of DCs. Our results showed that upon priming with either syngeneic or allogeneic fibroblasts, DCs lose their capacity to stimulate the proliferation of CD4 and CD8 T cells in both direct and indirect (or semi‐direct) pathways. Even activation of DCs with a TLR ligand, CpG, could not retrieve this ability of DCs. This finding is consistent with those previously reported for stromal cells like mesenchymal stem cells9 and renal fibroblasts.12 One possible explanation for the reduced stimulatory ability of fibroblast‐primed DCs could be the increased expression of anti‐inflammatory cytokines and reduced expression of pro‐inflammatory cytokines. We showed that the expression of IL‐10 and IDO was increased in fibroblast‐primed DCs compared with that shown in immature DCs. In general, IL‐10 suppresses T‐cell responses, especially CD4 effector and memory cells.27 IDO, a tryptophan‐degrading enzyme, also inhibits T‐cell proliferation,28, 29 through local depletion of tryptophan30 and production of T‐cell cytotoxic by‐products.31 Moreover, IL‐12p70 expression was significantly lower in fibroblast‐primed DCs compared with activated DCs. Autocrine IL‐12 signalling is needed for DCs to produce interferon‐γ and so its absence affects the stimulatory status of DCs.32

Another feature of functional DCs including tolerogenic ones, is their ability to migrate to lymphatic organs and present antigens. We used GFP⁺ DCs to show the ability of fibroblast‐primed DCs to migrate from the peritoneal cavity to MLNs. The fact that GFP⁺ DCs were present on day 3 but not on day 10 post‐injection in the peritoneal cavity suggests that either all of them have migrated to other places before day 10 or they have undergone apoptosis and cleared. We also showed that upon priming with fibroblasts, host DCs migrated from the peritoneal cavity to MLNs and presented fibroblast‐derived antigens. We used Act‐mOVA mouse fibroblasts to prime peritoneal cavity DCs. These transgenic mice express chicken ovalbumin under the control of β‐actin promoter on the surface of all body cells.33 As this antigen does not exist in B6 recipient mice, its presentation on DCs proves that they have been able to receive the fibroblast antigen and migrate to MLNs to present it to naive T cells. As there is no commercially available antibody to detect ovalbumin on MHC II of B6 mice, we were only able to evaluate its presentation on MHC I. Presentation of ovalbumin on MHC I can either show that the fibroblast cell membrane is being fused to the DC membrane or that DCs have engulfed the fibroblast‐derived antigens and cross‐presented them on their MHC I.

In conclusion, our results show that conditioning of DCs with fibroblasts triggers them to express higher levels of co‐inhibitory molecules and anti‐inflammatory cytokines and lower levels of pro‐inflammatory cytokines. Fibroblast conditioning arrests the ability of DCs to induce T‐cell proliferation in both direct and indirect pathways. These DCs can migrate to the regional lymph nodes and present fibroblast‐derived antigens. Together, these data are suggestive of the tolerogenic nature of DCs after being conditioned with fibroblasts. This finding sheds light on the role of skin fibroblasts in the maintenance of self‐tolerance and regulation of immune responses in the skin. Moreover, it might be used to design cell therapy approaches to control the immune responses in dermal autoimmune diseases or even in the setting of transplantation.

Authors contribution

M.K‐M designed and performed the experiments, analysed the data and contributed to writing and editing the manuscript, R.B.J contributed to designing the experiments, performing the in vivo and flowcytometry experiments and editing the manuscript, S.S‐E and MR.P contributed to in vivo experiments and imaging studies, J.L contributed to flowcytometry studies and breeding the mice and editing the manuscript, M.P.M contributed to Real Time PCR, R.T.K contributed to culturing and transducing the fibroblasts, J.D contributed in designing the experiments and editing the manuscript, A.G supervised the project, designed the experiments and contributed in writing and editing the manuscript.

Disclosures

The authors declare they have no conflict of interests.