Critical Role of Macrophage FcγR Signaling and Reactive Oxygen Species in Alloantibody-Mediated Hepatocyte Rejection

Abstract

Humoral alloimmunity negatively impacts both short- and long-term cell and solid organ transplant survival. We previously reported that alloantibody-mediated rejection of transplanted hepatocytes is critically dependent on host macrophages. However, the effector mechanism(s) of macrophage-mediated injury to allogeneic liver parenchymal cells is not known. We hypothesized that macrophage-mediated destruction of allogeneic hepatocytes occurs by cell-cell interactions requiring FcγRs. To examine this, alloantibody-dependent hepatocyte rejection in CD8-depleted wild-type (WT) and Fcγ-chain knockout (KO; lacking all functional FcγR) transplant recipients was evaluated. Alloantibody-mediated hepatocellular allograft rejection was abrogated in recipients lacking FcγR compared with WT recipients. We also investigated anti-FcγRI mAb, anti-FcγRIII mAb, and inhibitors of intracellular signaling (to block phagocytosis, cytokines, and reactive oxygen species [ROS]) in an in vitro alloantibody-dependent, macrophage-mediated hepatocytoxicity assay. Results showed that in vitro alloantibody-dependent, macrophage-mediated hepatocytotoxicity was critically dependent on FcγRs and ROS. The adoptive transfer of WT macrophages into CD8-depleted FcγR-deficient recipients was sufficient to induce alloantibody-mediated rejection, whereas adoptive transfer of macrophages from Fcγ-chain KO mice or ROS-deficient (p47 KO) macrophages was not. These results provide the first evidence, to our knowledge, that alloantibody-dependent hepatocellular allograft rejection is mediated by host macrophages through FcγR signaling and ROS cytotoxic effector mechanisms. These results support the investigation of novel immunotherapeutic strategies targeting macrophages, FcγRs, and/or downstream molecules, including ROS, to inhibit humoral immune damage of transplanted hepatocytes and perhaps other cell and solid organ transplants.

Figure 1.. In vitro macrophage-mediated hepatocytotoxicity is contact-dependent.

FVB/N hepatocytes were isolated and cultured in vitro. Allogeneic hepatocytes were incubated with alloserum or control serum from naïve mice. RAW 264.7 macrophages were activated by pretreatment with IFN-γ (2.5 ng/mL; 18 hours). Following pretreatment, macrophages (Mac; 1.5×10⁶ cells) were washed and co-incubated with hepatocytes (1.5×10⁵ cells). A) In positive control co-cultures, activated macrophages and allogeneic hepatocytes were co-cultured in the same well. When activated macrophages and viable hepatocytes were co-cultured with alloantibody for 8 hours, significant cytotoxicity was observed [57.3±8.0%, p<0.0005 compared to hepatocytes cultured alone (8.2±0.5%), with serum from naïve mice (7.5±0.7%), or with alloserum (15.1±1.3%), as denoted by “*”; n=4 for all conditions] as reflected by LDH release in the culture supernatant. B) In other co-cultures, macrophages were separated from viable hepatocytes by a transwell membrane. To activate macrophages in the transwell group, cells were co-cultured with formalin-fixed alloantibody-incubated hepatocytes (ffHc) in the transwell and viable hepatocytes were in the bottom well. Following 8 hours of co-culture, supernatant was analyzed for LDH release. In the absence of activated macrophage cell contact with viable allogeneic hepatocytes (transwell), minimal cytotoxicity was detected in co-cultures (last bar: activated macrophages= 11.4±1.6%; p=ns compared to hepatocytes cultured alone and p=.0004 compared to positive control culture in (A); n=4 for all conditions). Data for co-cultures in (A) and (B) are representative from triplicate experiments. C) To confirm activation of macrophages by formalin-fixed alloantibody-incubated hepatocytes, macrophages were lysed at 15, 30, and 60 minutes after co-culture. Macrophages were tested for phosphorylated-Erk, SerAkt, and NFκB by western blot. Beta actin was used as a loading control.

Figure 2.. In vitro macrophage-mediated hepatocytotoxicity is FcγRI- and FcγRIII-dependent.

Macrophage:hepatocyte co-cultures consisted of FVB/N hepatocytes and RAW 264.7 macrophages activated by pretreatment with IFN-γ (2.5 ng/mL; 18 hours). Prior to co-culture RAW macrophages were incubated with FcγRI and/or FcγRIII blocking antibody (2 μg per million macrophages) or control IgG. The macrophages were washed with PBS and then added to the hepatocyte co-culture for an 8 hour incubation (1.5×10⁶ cells macrophages and 1.5×10⁵ hepatocytes). Control IgG-treated macrophages mediated hepatocytotoxicity against alloantibody incubated target hepatocytes (49.8±2.7%; n=24). Anti-FcγRI mAb (24.3±4.3%; n=11) and anti-FcγRIII mAb (24.7±4.0%; n=9) treatment significantly blocked macrophage-mediated hepatocytotoxicity (p<0.0001 for both, as denoted by “*”). Combination treatment with both anti-FcγRI and anti-FcγRIII mAbs completely inhibited macrophage-mediated hepatocytotoxicity (3.2±1.0%; n=12, p<0.0001, as denoted by “**”). Data is combined from duplicate experiments.

Figure 3.. Alloantibody-mediated hepatocellular allograft rejection is FcγR-dependent.

C57BL/6 (wild-type; WT) and Fcγ chain KO mice (H-2^b) were transplanted with allogeneic FVB/N hepatocytes (H-2^q) on day 0. A cohort of recipients was CD8-depleted (days −3, −1, and weekly post transplant until day +49) to suppress cell-mediated rejection. A) Allogeneic hepatocyte rejection occurred rapidly in both WT (n=5; MST=day 10) and Fcγ chain KO (n=6; MST=day 14) recipients. B) Alloantibody-mediated hepatocyte rejection in CD8-depleted WT mice also occurred rapidly (n=5; MST= day 14). In contrast alloantibody-mediated rejection was not observed in CD8-depleted Fcγ chain KO recipients (n=4) by the end of the study period (day 50). C) Alloantibody titer was measured on day 14 post transplant. WT (titer=90±5) and Fcγ chain KO (titer=75±10) recipients both produced similar alloantibody titers (control serum from naïve mice represented by the dashed line). CD8-depleted WT mice produced markedly increased alloantibody (titer=1800±89) compared to WT recipients (p<0.0001, as denoted by “*”). CD8-depleted Fcγ chain KO recipients also exhibited a significantly increased alloantibody titer (titer=1062±157) compared to Fcγ chain KO recipients (titer=75±10; p<0.0001, as denoted by “**”). Titers in CD8-depleted WT and Fcγ chain KO recipients were comparable. Data for all is combined from duplicate experiments.

Figure 4.. In vitro macrophage-mediated hepatocytotoxicity is reactive oxygen species-dependent.

All macrophage:hepatocyte co-cultures consisted of FVB/N hepatocytes, activated macrophages and alloserum. RAW 264.7 macrophages and BMM were activated by pretreatment with IFN-γ (2.5 ng/mL; 18 hours). A) During co-culture of hepatocytes and RAW macrophages, cultures were treated with DMSO control, UO126, LY294002, or BAY11–7058. While treatment with LY294002 (58.3±2.3%; n=12, 20 μM; p=ns) and BAY11–7085 (51.9±2.5%; n=12, 5 μM, p=ns) did not inhibit macrophage-mediated hepatocytotoxicity compared to positive control (58.9±2.1%; DMSO control, n=20), treatment with UO126 (28.4±2.0%; n=14, 5 μM; p<0.0001, as denoted by “*”) significantly inhibited macrophage-mediated cytotoxicity of alloantibody-incubated hepatocytes. Data is combined from triplicate experiments. B) In a separate cohort, co-cultures were incubated with DMSO (vehicle control), anti-TNF-α mAb, cytochalasin D, superoxide dismutase (SOD), or apocynin. Both superoxide dismutase (200 U/mL= 69.8±2.7%; 500 U/mL= 35.4±1.6%; 1,000 U/mL=28.7±3.0%; n=4 for all, p<0.0005 for 500 and 1,000 U/mL, as denoted by “**”) and apocynin (0.25 mM=16.9±4.0%; 0.5 mM=9.6±3.0%; 1.0 mM=0±0%; n=5 for all, p<0.0001 for all, as denoted by “***”) significantly inhibited macrophage-mediated hepatocytotoxicity (apocynin abrogated cytotoxicity at 1.0 mM) as compared to DMSO-treated co-cultures (0.1%=62.6±3.3%, 0.5%= 66.9±3.8%, 1.0%= 65.5±2.9%; n=6 for all). RAW cells incubated with anti-TNF-α mAb at all doses tested (5 μg/mL=61.8±5.5%; 10 μg/mL=66.6±1.1%; 20 μg/mL=64.3±1.9%; n=4 for all) had no effect on macrophage-mediated hepatocytotoxicity. Cytochalasin D (inhibitor of actin polymerization and FcγR co-localization) significantly impaired macrophage-mediated hepatocytotoxcity (1 μg/mL=37.3±1.1%; 5 μg/mL=44.2±1.8%; 10 μg/mL=39.3±2.8%; n=4 for all, p<0.009 for all doses, as denoted by “*”). Data is combined from duplicate experiments. C) Bone marrow macrophages (BMM) were isolated from WT and p47 KO mice and co-cultured with allogeneic hepatocytes and alloserum. Macrophage-mediated hepatocytotoxicity was significantly impaired in co-cultures containing p47-deficient BMM (8.3±3.8%; n=6) which cannot produce ROS compared to WT BMM (38.6±6.1%; n=6, p=0.0009, as denoted by “*”). Data is combined from duplicate experiments.

Figure 5.. Macrophage-mediated alloantibody-dependent hepatocyte rejection is mediated by FcγR and ROS.

Fcγ chain KO (H-2^b) mice were transplanted with allogeneic FVB/N hepatocytes (H-2^q) on day 0. All recipients were CD8-depleted (days −3, −1, and weekly post transplant until day +35) to suppress cell-mediated rejection. On the same day of transplant, recipients were adoptively transferred (AT) with bone marrow macrophages from WT, Fcγ chain KO, or p47-deficient mice. A) Representative data shows that adoptively transferred bone marrow macrophages cells were >98% macrophages as determined by dual expression of F4/80 and CD11b by flow cytometry. B) Adoptive transfer of WT bone marrow macrophages precipitated rapid rejection (MST= 17; n=4) while adoptive transfer of Fcγ chain KO (n=6) and p47-deficient (n=5) bone marrow macrophages did not (MST > 45 days, p<0.0001 for both).