|Home | About | Journals | Submit | Contact Us | Français|
TIM-3 is constitutively expressed on subsets of macrophages and dendritic cells. Its expression on other cells of the innate immune system and its role in fetomaternal tolerance has not yet been explored. Here we investigate the role of TIM-3 expressing innate immune cells in the regulation of tolerance at the fetomaternal interface (FMI) using an allogeneic mouse model of pregnancy. Blockade of TIM-3 results in accumulation of inflammatory granulocytes and macrophages at the utero-placental interface and up regulation of pro-inflammatory cytokines. Furthermore, TIM-3 blockade inhibits the phagocytic potential of uterine macrophages resulting in a build up of apoptotic bodies at the utero-placental interface that elicits a local immune response. In response to inflammatory cytokines, Ly-6ChiGneg M-MDSCs (monocytic myeloid derived suppressor cells) expressing iNOS and arginase 1 are induced. However, these suppressive cells fail to down-regulate the inflammatory cascade induced by inflammatory granulocytes (Ly-6Cint Ghi) and apoptotic cells; the increased production of IFNγ and TNFα by inflammatory granulocytes leads to abrogation of tolerance at the fetomaternal interface and fetal rejection. These data highlight the interplay between cells of the innate immune system at the FMI and their influence on successful pregnancy in mice.
Since Medawar’s first hypothesis on the mechanism of avoidance of immune attack by the semi-allogeneic fetus, substantial research in reproductive and transplant immunology has addressed this paradigm. Successful pregnancy requires that the maternal immune system does not attack the fetus that has fetal histocompatibility antigens inherited from the father. A deleterious immune attack is avoided by orchestration of cellular, hormonal and enzymatic factors. In recent years, it has become apparent that a Th2 cytokine profile is crucial to maintain successful pregnancy (1-3).
Pregnancy induced regulatory T cells (4) and the negative co-stimulatory molecule PD-L1 have been shown to be important in fetal acceptance in murine pregnancy (5, 6). Further, innate immune cells are also critical for initiating and coordinating an immune response against paternal antigens (7). Decidual macrophages and dendritic cells (DC) have also been shown to have a pivotal role in establishing a tolerogenic microenvironment at the fetomaternal interface (8, 9).
In this study we explore whether the molecule TIM-3 plays a role in inducing fetomaternal tolerance. TIM-3 was first described as a molecule specifically expressed on the surface of IFNγ producing Th1 and Tc1 cells (10). TIM-3 is a pattern recognition receptor specialized for recognition of phosphatidylserine exposed on apoptotic cells (11). Another ligand for TIM-3 is galectin-9 (12). Galectin-9 is an S-type lectin ubiquitously expressed in cells and on certain epitheliums (13). It binds to TIM-3 expressed on activated (IFNγ producing) Th1 and Tc1 but not to Th2 cells (10) to terminate T cell response by induction of apoptotic signals (12, 14). A role for TIM-3 has also been described in T cell exhaustion of virus-infected CD8 cells (15-21).
Besides being expressed on activated T cells, TIM-3 is constitutively expressed on cells of the innate immune system in both mice and humans (10, 22, 23). TIM-3 expressed on dendritic cells and on subsets of macrophages mediates phagocytosis of apoptotic cells and cross-presentation of antigens (23), and synergizes with Toll Like Receptors (TLRs) to enhance inflammatory responses (22). Transgenic overexpression of TIM-3 on T cells results in an increase in the population of CD11b+Ly6Ghi granulocytic myeloid derived suppressor cells (G-MDSC) in mice (24). The role of TIM-3 in innate immune cells is likely complex as many cell types are involved in the regulation of the innate immune response by various mechanisms.
The role of TIM-3 has also been studied in allograft tolerance (14, 25). For example, TIM-3 deficient mice are reported to be refractory to tolerance induction by donor specific transfusions (DST) or treatment by CTLA4-Ig or anti-CD40L (CD 154) antibody (14). However, little is known whether TIM-3 plays a role in regulating the immune system at the fetomaternal interface.
In this study, we use MHC mismatched pregnancy to explore the role of TIM-3 on uterine myeloid cells in inducing or maintaining fetomaternal tolerance. We find that TIM-3 is expressed on monocytes and granulocytes infiltrating the uterus as well as on different subsets of uterine macrophages and dendritic cells (DC). Treatment of pregnant mice with a TIM-3 blocking antibody resulted in the failure of uterine macrophages to clear apoptotic and dying cells. The resulting inflammation activates two subsets of myeloid cells. The first subset, the suppressor M-MDSC up-regulate the expression of iNOS and arginase 1; the second subset, the inflammatory CD11b+Ly6Cint Ghi cells increase their production of pro-inflammatory cytokines such as IFNγ and TNFα. M-MDSC are unable to counterbalance the over powering inflammation created by CD11b+Ly6Cint Ghi cells, and the resultant effect is increased fetal resorption and death.
CBA/CaJ and C57BL/6J mice were purchased from The Jackson Laboratory and maintained in our animal facility according to institutional and National Institute of Health guidelines. 7-8 week old females were mated to either C57BL/6 or CBA/CaJ males to induce pregnancy and were inspected every morning for vaginal plugs. The day of visualization of a plug was designated as day 0.5 of pregnancy. In some experiments, pregnant mice were monitored until parturition and the number of live pups recorded to assess the effect of anti-TIM-3-antibody treatment on litter size. In other instances, mice were sacrificed before the termination of pregnancy and the percentage of resorbed embryos was calculated (resorbed embryos/total embryos ×100).
Pregnant females received 4 injections of RMT3-23 (anti-TIM-3) antibody intraperitoneally at doses of 500 μg, 500 μg, 250 μg, 250 μg at days 6.5,8.5,10.5 and 12.5, respectively.
Uteri from pregnant mice were dissected free from the mesometrium and removed by cuts at the ovaries and cervix. These uteri were carefully dissected with scissors to remove fetal and placental tissue, washed twice in ice-cold PBS (Phosphate Buffered Saline) or HBSS (Hank’s Balanced Salt Solution) and shredded carefully. Minced uteri were enzymatically digested for 20 minutes at 37°C in HBSS/Ca/Mg containing 200 U/ml hyaluronidase, 1mg/ml collagenase type IV, 0.2 mg/ml DNase and 1 mg/ml bovine serum albumin/fraction V as previously described with some modifications (26). Cells were then washed in HBSS free of enzymes and incubated in the same buffer for another 15 minutes at 37°C prior to filtration through a 70μm cell strainer. The draining para aortic and iliac lymph nodes were not subjected to enzymatic digestion and cells were carefully minced and filtered through a 70μM cell strainer.
CBA/CaJ females were mated with C57Bl/6 males and uteri were harvested on gestational day 7.5 and 10.5. Uteri were immersion-fixed in 10% buffered formalin and embedded in paraffin. After deparaffinization, uterine sections were stained using ApopTag® Fluorescein Direct In Situ Apoptosis Detection Kit (Millipore, Billerica, MA) according to the manufacturer’s instructions. Nuclei were counterstained with DAPI (ProLong® Gold Antifade Reagent, Invitrogen Life Technologies, Grand Island, NY). Images were captured on Zeiss LSM 510 Meta confocal microscope and the number of TUNEL positive cells was quantified as TUNEL positive cells/ mm2 of tissue.
Total RNA was extracted using TRIzol reagent (Invitrogen Life Technologies, Grand Island, NY). Complementary DNA was synthesized using SuperScript III first-strand synthesis system for RT-PCR (Invitrogen Life Technologies). mRNA expression of iNOS, arginase 1, IFNγ, TNFα and GAPDH were examined using TaqMan gene expression assays. Triplicate samples were examined in each condition. A comparative threshold cycle (CT) value was normalized for each sample using the formula: ΔCT = CT (gene of interest)− CT (GAPDH), and the relative expression was then calculated using the formula 2−ΔCT.
For FACS analysis, cells were isolated from uterus of mice at different time points of gestation and pretreated with Fc blocking monoclonal antibody 24G.2 for 10 minutes at 4°C and stained with flurochrome labeled antibodies against CD45 (30-F11), CD11c (N418), CD11b (M1/70), F4/80 (BM8), MHCII (Ia/Ie), TIM-3 (B8.2C12), Ly-6C (HK1.4), Ly-6G (1A8), CD3 (17A2), CD4 (GK1.5), CD8 (53-6.7) purchased from Biolegend (San Diego CA), Invitrogen, eBioscience (San Diego, CA) or BD biosciences (San Jose, CA). The fluorochromes used were FITC, PE, PERCP.Cy5.5, APC, PE-Cy7, APC-Cy7, Alexa fluor 700, PE-texas red and brilliant violet. Samples were washed with FACS buffer, acquired on a 5 Laser 12 Color BD LSRII FACS (BD biosciences, San Jose, CA) and analyzed with Flowjo software (Tree star, Ashland, OR). For sorting experiments, cells were isolated from uterine tissue as described earlier, stained with antibodies against CD45 APC (30-F11), CD11b PE-Cy7 (M1/70), TIM-3 PE (B8.2C12), Ly-6C FITC (HK1.4) and Ly-6G Percp-Cy5.5 (1A8). Cells were sorted on a 5 lasers FACS Aria (BD biosciences) or Moflo cell sorter (Beckman-Coulter, Brea CA) into TRIzol reagent (Invitrogen Life Technologies).
For apoptosis experiments, cells were stained using the Annexin V apoptosis detection kit from BD pharmingen (San Jose, CA), acquired on a 2 lasers 4 colors FACS calibur (BD biosciences, San Jose, CA) and analyzed with Flowjo software (Tree star, Ashland, OR).
Immunofluorescence antigen labeling for TIM-3 was performed on the paraffin embedded mouse uterus samples. Paraffin-embedded sections were prepared and treated with Trilogy (Cell Marque, CA) for antigen retrieval. Sections were then incubated with 1mg/ml sodium borohdyride (ICN chemicals) for 5 minutes at room temperature. After three washes with TBS, the sections were incubated with 5% normal donkey serum (Jackson ImmunoResearch Lab Inc, West Grove, PA) for one hour at room temperature. The sections were then incubated with Goat anti-TIM-3 (1:200, Santa Cruz, Santa Cruz, CA) or Rabbit anti-Mac2 (1:300, CedarLane, Burlington, NC) antibody overnight at 4°C.
Slides were washed three times and Dylight 649 conjugated donkey anti-goat secondary antibody or Dylight 549 donkey anti-rabbit secondary antibody (1:200, Jackson Immuno Research Lab).
Slides were then washed three times with Tris Buffered Saline (TBS) and mounted with Prolong Gold anti-fade mounting media containing DAPI (Invitrogen). Confocal images were taken using a Zeiss LSM510 Meta confocal system and Zeiss LSM510 image acquisition software (20x/0.8 Plan-Apochromat objective and a 40x/1.3 Oil Plan-Apochromat objective).
Apoptotic cells were prepared by incubating Tsras2 trophoblast cell line from FVB mice with 1%H2O2 for 10 minutes at 37°C. The percent of apoptosis was assessed using the annexin V apoptosis detection kit (BD biosciences). For measuring in vitro activation of macrophages by apoptotic cells, 1 million RAW264.7 cells were pretreated with Fc block followed by RMT3-23 antibody or control and co-cultured with 0.1×106 apoptotic trophoblast cells for 24h at 37°C. After 24h, cells were washed twice with PBS and RNA was isolated according to the protocol described earlier.
RAW264.7 macrophage cell line or uterine F4/80+ macrophages flow sorted from gestational day 10.5 uterus of CBA/CaJ females (mated with C57Bl/6 males) were plated on a 48-well or a 96-well plate at a density of 5×104 cells/well for 4 hours at 37°C and then washed twice to remove non-adherent cells. The adherent cells were then incubated with RMT3-23 antibody (30μg/ml) for 30 minutes prior to culture with FITC labeled beads (Cayman Chemicals, Ann Arbor, MI) for 2h at 37°C. The cells were washed 3 times with PBS to remove unbound beads and fluorescence intensity was assessed using an inverted microscope or by flow cytometry on a FACS Calibur according to the manufacturer’s instructions.
Tsras2 cell line was a kind gift from Adrian Erlebacher (NYU Langhone Medical Center, NY, NY). Cells were maintained in DMEM: F12 supplemented with 20%FCS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 10 mM HEPES, 1 μM sodium pyruvate and 100 μM 2-mercaptoethanol.
RAW264.7-cell line (ATCC) was maintained in complete DMEM with L-glutamine supplemented with 10%FCS, 100 U/mL penicillin, 100 μg/mL streptomycin and 10 mM HEPES.
Unpaired two-tailed t test was used to analyze the statistical significance between two groups. Kruskal-Wallis followed by Dunns post hoc test was used to analyze the statistical significance among multiple groups. P < 0.05 was considered statistically significant.
The mouse model of allogeneic pregnancy involving mating of CBA/CaJ females with C57Bl/6 males (27, 28) was used to study the changes taking place in the composition of cells infiltrating the fetomaternal interface on gestational days 7.5, 10.5 and 12.5. (See S1 and S2 for gating strategy). Analysis of DC subtypes showed that CD11bhi and CD11bneg subtypes do not vary between day 7.5 and 12.5. In contrast CD11blo cells decreased significantly in number in late pregnancy (day 12.5) (Figure 1a). Analysis of macrophages in the uterus and draining para aortic and iliac lymph nodes of pregnant mice showed two subtypes: mature F4/80+MHCIIhi cells and immature F4/80+MHCIIlo cells (Figure 1b). Immature macrophages reside predominantly in the uterine mucosa. The percentage of these uterine F4/80+MHCIIlo cells increases significantly from 0.8 percent in the peri-implantation period (day 7.5) to 7.8 percent at late gestation (day 12.5).
Inflammatory monocytes/Myeloid derived suppressor cells are phenotypically characterized as CD11b+Gr1+cells. Our results show that these cells are further divided into three subsets at the uteroplacental interface based on staining by specific Gr1 monoclonal antibodies (Ly6C and Ly6G) (Figure 1c). Granulocytic myeloid derived suppressor cells (G-MDSC) have a CD11b+Ly6CnegGhi phenotype and constitute 0.5 to 3% of the CD11b+Gr1+ population. Monocytic MDSCs (M-MDSC) have a CD11b+ly6ChiGneg phenotype and constitute 20-30% of the population. Granulocytes make up 2-4% of CD11b+Gr1+ cells and have a CD11b+Ly6CintGhi phenotype. Resident monocytes, which constitute the rest of the population, are Ly-6C and G negative (data not shown). CD11b+Ly6CintGhi cells remain unchanged throughout pregnancy compared to G-MDSC and M-MDSC, which peak at mid (day 10.5) and late gestation (day 12.5), respectively. Table 1 summarizes the mean percentage of myeloid cells on gestational days 7.5, 10.5 and 12.5 in the uterus.
To address whether TIM-3 is involved in promoting immune cell homeostasis at the fetomaternal interface, TIM-3 expression was analyzed on gestational day 10.5 in uterine sections obtained from allogeneically mated mice. TIM-3 expression was observed in the uterus as determined by immunofluorescence staining (Figure 2).
In addition to the above described TIM-3 expression at the uteroplacental interface by histological methods, we assessed TIM-3 expression on myeloid cells present in the uterus and draining lymph nodes by flow cytometry on gestational days 7.5, 10.5 and 12.5. It was observed that in the draining LN, TIM-3 expression is significantly up regulated at mid gestation (day 10.5) in all three DC subsets namely CD11c+11bhi, CD11c+blo and CD11c+11bneg (data not shown). TIM-3 expression was significantly up regulated at midgestation (day 10.5) in uterine CD11c+11bhi DC as well. Conversely, uterine CD11blo and CD11bneg DCs expressed relatively constant levels of TIM-3 at day 10.5 gestation (Figure 3a).
Similar to DCs, macrophages are known to constitutively express TIM-3 (23, 29, 30). We also evaluated whether TIM-3 expression on mature and immature uterine macrophage populations and on cells residing in the draining para aortic and iliac draining lymph nodes play a role in modulating allo-immune pregnancy. The draining LNs contained very few macrophages (data not shown). The gestational uterus contained a higher number of macrophages compared to the regional draining LNs and the majority of cells were immature, expressing low levels of MHC class II. TIM-3 expression remained constant in both subsets of uterine macrophages during early stages of pregnancy but immature cells were observed to down regulate TIM-3 expression at day 12.5 (Figure 3b).
Different subsets of myeloid progenitor and suppressor cells were observed to express TIM-3 in a unique manner in pregnancy. TIM-3 expression in the uterus was negative in the three subsets of myeloid derived cells at gestational day 7.5. Granulocytic MDSC (G-MDSC) and granulocytes up regulated TIM-3 at mid-gestation compared to monocytic MDSC (M-MDSC) which showed peak expression at day 12.5 (Figure 3c). Thus, pregnancy in mice results in variable expression of TIM-3 on the surface of the different cells residing locally in the uterus and in the regional lymph nodes.
To investigate the role of TIM-3 in fetomaternal tolerance, we studied the effect of TIM-3 blockade on syngeneic and allogeneic pregnancy. Syngeneic pregnancy did not show any difference in litter size after treatment (Figure 4a). Pregnant CBA/CaJ females (mated with C57Bl/6 males) challenged with TIM-3 blocking antibody (RMT3-23) according to the protocol described earlier were observed to be more susceptible to fetal loss manifested by a reduction in litter size (Figure 4b), and increased rate of resorption at midgestation compared to untreated pregnant controls (Figure 4c). This data shows that TIM-3 plays a protective role during successful pregnancy in vivo.
TIM-3 blockade was observed to result in infiltration of macrophages and inflammatory Ly-6Cint Ghi cells at the utero-placental interface at mid-gestation in allogeneic (Figure 5a,b and c) but not in syngeneic pregnancy (Figure 5d and e). Granulocytes are phenotypically characterized by co-expression of Ly-6C and Ly-6G (Ly-6Cint Ghi cells). Uterine Ly-6Cint Ghi cells were also observed to express TIM-3. We isolated uterine TIM-3 positive and TIM-3 negative Ly-6Cint Ghi cells from both treated and untreated allogeneically impregnated mice. We observed that TIM-3+ and TIM-3- Ly-6Cint Ghi cells up regulate TNFα mRNA following TIM-3 blockade. TIM-3+ Ly-6Cint Ghi cells also up-regulated IFNγ mRNA (Figure 6a). In addition, macrophages (RAW 264.7) co-cultured in vitro with apoptotic trophoblast cells (Tsras 2 cells) in the presence of TIM-3 blocking antibody were also observed to up regulate mRNA expression of TNFα (Figure 7c).
On the other hand, ex-vivo isolated suppressor cells such as CD11b+Ly6ChiGneg up regulated the expression of the immunoregulatory molecules iNOS and arginase1 following treatment with RMT3-23 (Figure 6b).
Our results suggest that the above-mentioned suppressor cell population fails to counterbalance the potent pro-inflammatory effect of CD11b+Ly6CintGhi cells in vivo thereby leading to pregnancy failure.
To investigate whether TIM-3 plays a role in phagocytosis and clearance of apoptotic cells from the utero-placental interface, TIM-3 expression on the surface of uterine F4/80+ macrophage population (Figure 3b) as well as on RAW264.7 macrophages was first confirmed (Figure 7a). Next, uterine F4/80+ macrophages were isolated from gestational day 10.5 uteri of CBA/CaJ mice and their phagocytic properties were analyzed in the presence of RMT3-23 antibody. Fc receptors of uterine macrophages were blocked to rule out the possibility that binding of RMT3-23 to FcR might affect phagocytosis. Macrophages were then pre-treated with RMT3-23 or PBS and incubated ex vivo with FITC-labeled beads. It was observed that both treated and untreated uterine macrophages are capable of internalizing the particles as shown in Figure 7b. Interestingly, RMT3-23 treated cells were shown to engulf significantly fewer particles (Figure 7b). In another experiment, RAW264.7 cells were pre-treated with RMT3-23 antibody and incubated with FITC-labeled beads for 2h before quantitative analysis by flow cytometry. Similarly RMT3-23 treated RAW264.7 macrophages engulfed significantly fewer particles (Figure 7a). Thus, TIM-3 blockade using RMT3-23 treatment significantly abrogates phagocytosis at the utero-placental interface. As described in the previous section, macrophages co-cultured with apoptotic trophoblast cells in vitro in the presence of anti-TIM-3 antibody up regulated mRNA expression of TNFα (Figure 7c).
To evaluate whether TIM-3 blockade in vivo has any effect on apoptosis at the utero-placental interface, pregnant CBA/CaJ females were injected with RMT3-23 at days 6.5 and 8.5 of pregnancy. Uteri were harvested at day 10.5 and apoptotic cells were detected by TUNEL method on uterine sections; staining for Annexin V/7AAD was performed on uterine cell preparation. Significantly increased number of TUNEL+ cells were observed in uterine sections from RMT3-23 treated group as compared to that from untreated control group (Figure 8a,b). In addition, Annexin V staining showed increased apoptosis of uterine CD4+ population but not CD8+ cells (Figure 8c).
This result is consistent with the possibility that an increased number of cells undergo apoptosis upon blockade of TIM-3. Alternatively the observation of increased numbers of apoptotic cells could be a result of defective clearance of these cells by phagocytic cells. Apoptotic and dying cells have been shown to send a “danger” signal that can regulate the immune system and enhance inflammation (31-33).
The fetomaternal interface is a complex environment since multiple mechanisms operate to protect the fetus against attack by the maternal immune system. These include complement system (34, 35), catabolism of tryptophan by IDO (27), regulation by galectin-1 (36), regulatory T cells (4, 37-39) and T cell apoptosis (40). Previously we described a role for the inhibitory co-stimulatory molecule PDL1 in maintaining tolerance at the fetomaternal interface (5, 6, 41).
In this study, we have examined the role of TIM-3 molecule and contributions of macrophage and myeloid cells to fetomaternal tolerance; areas that have not been explored till date. TIM-3 is expressed on the surface of terminally differentiated T cells, dendritic cells and macrophages. On the surface of T cells, TIM-3 binds to galectin-9 and inhibits Th1 response by triggering T cell death (12, 42). On antigen presenting cells (APCs), TIM-3 synergizes with Toll like receptor (TLR) and increases the secretion of pro-inflammatory cytokines (22). TIM-3 also functions as a phosphatidylserine receptor and is involved in the clearance of apoptotic cells and antigen cross presentation (23, 43). Our study demonstrates the involvement of TIM-3 in promoting tolerance at the fetomaternal interface.
We observed that TIM-3 is expressed on the surface of innate immune cells, viz. uterine macrophages and IMC/myeloid derived suppressor cells in the CBA/CaJ (x C57Bl/6) model of allogeneic pregnancy. Our study shows an important role for TIM-3 in modulating the function of these two cell types. This is a significant finding given that innate immune cells constitute more than 80% of the cells found in the decidua/uterus during pregnancy (26, 44).
During pregnancy, apoptosis is particularly seen in trophoblast and uterine NK cells (45, 46). Apoptosis of these cells is increased in threatened pregnancy situations such as pre-eclampsia and intra uterine growth restriction (IUGR) (47, 48). Trophoblast cells express paternal antigens. Therefore, mechanisms must exist to clear apoptotic trophoblast cells so that the cells themselves, or the paternal antigens released by these cells do not cause tissue damage and fetal rejection (49). Macrophages have been shown to play a key role in mediating the resolution of inflammatory conditions by phagocytosing and clearing apoptotic cells (50). Our data shows that following TIM-3 blockade, apoptotic cells accumulate at the uteroplacental interface because of a deficiency in the phagocytic capacity of uterine macrophages leading to inflammatory events and subsequent fetal rejection (Figure 4, ,6a,6a, ,77 and and88).
Our data does not support the hypothesis that reduced phagocytosis of apoptotic cells by macrophages results in reduced presentation of antigens and hence reduced alloimmune responses. Rather, our alloimmune response data obtained by stimulating splenocytes from pregnant females with irradiated stimulators from allogeneic males showed no differences in Th1/Th2 or Th17 alloimmune responses following TIM-3 blockade (data not shown). Also, we did not see a difference in the size of DC population subtypes (CD11c+11bhi, CD11c+blo, and CD11c+11bneg) following anti-TIM-3 treatment (data not shown). Therefore, it is unlikely that phagocytosis and cross presentation of cell-associated antigens with MHC class I molecules is mediated by DCs (31, 51-53). Further studies are needed to determine if cross presentation of antigens by DCs is affected despite unchanged size of population subtypes following anti-TIM-3 treatment, and if fetal rejection/acceptance is influenced as a result.
Further, we also observed that the decrease in litter size and increase in fetal resorption correlated with an increase in the population of F4/80+ macrophage locally in the uterus. Under normal conditions, F4/80+ uterine macrophages, a major APC (antigen presenting cell) (8, 26) with the ability to clear apoptotic cells, remain more or less constant during various stages of pregnancy (49, 50). An increase in their number following TIM-3 blockade suggests that these macrophages may play a role in causing resorption. In an earlier study, administration of anti-TIM-3 mAb was shown to enhance disease severity in experimental autoimmune encephalitis model with a concurrent increase in infiltration of activated macrophages at the site of injury (10).
TIM-3 expression was also observed in myeloid derived suppressor cells (Figure 3). Based on the predominant expression of either Ly6C or Ly6G, three populations CD11b+Ly6CnegGhi granulocytic MDSC (G-MDSC), CD11b+Ly6ChiGneg monocytic MDSCs (M-MDSC) and CD11b+Ly6CintGhi cells can be distinguished. Cells with CD11b+Ly6CintGhi phenotype have no suppressive activity (54). G-MDSCs and M-MDSC are capable of suppressing T cells (55) and are shown to modulate immune response in various disease models (54, 56-60). All three subtypes were observed to express TIM-3 in this study (Figure 3). CD11b+Ly6ChiGneg M-MDSC and CD11b+Ly6CintGhi inflammatory granulocytes (Figure 1) predominated at the uteroplacental interface. Results obtained here suggest that both M-MDSCs and CD11b+Ly6CintGhi cells become activated as a consequence of inflammatory changes occurring in the uterus. Activated M-MDSCs up regulate iNOS and arginase 1 (Figure 6b), and CD11b+Ly6CintGhi granulocytes up regulate the expression of TNFα and IFNγ in response to treatment or to inflammation (Figure 6a). Resorption observed here could be a result of the failure of M-MDSC cells to counterbalance the effect of inflammatory CD11b+Ly6CintGhi granulocytes. TIM-3 blockade has been shown to result in pro-inflammatory cytokine release in a model of liver ischemia (61). Also, in a TIM-3 transgenic mouse model, over-expression of TIM-3 on T cells resulted in an expansion of CD11b+Gr1+F4/80 low granulocytic MDSCs and suppression of T cell responses while other cellular compartments remained unchanged between transgenic mice and littermate control (24).
We conclude that TIM-3 blockade results in a decrease in phagocytic properties of uterine macrophages leading to a failure to clear apoptotic and dead cells from the uterus. This leads to accumulation of apoptotic cells at the utero-placental interface causing activation of inflammatory granulocytes. The result is increased local inflammation and fetal rejection. Future studies will address the relative contributions of various innate immune cells in TIM-3 modulated fetomaternal tolerance and interactions between them.
We thank X. Li and V. Kuchroo for critical reading of the manuscript and for helpful discussions. We also thank Lay-Hong Ming for her help with confocal microscopy.
This work is supported by research grants from the National Institutes of Health (RO1 AI 84756-01A2 and R21AI076794 to I.G).
The authors declare that there are no conflicts of interest.