Targeting the Complement Alternative Pathway Permits Graft Versus Leukemia Activity while Preventing Graft Versus Host Disease

Allogeneic hematopoietic cell transplantation (allo-HCT) is a potentially curative option for a variety of benign and malignant hematopoietic disorders (1, 2). The therapeutic benefit of allo-HCT is attributable to the graft-versus leukemia effect (GVL). However, GVHD development remains a significant cause of transplant-related mortality and morbidity (3). As T cells in the donor graft mediate the GVL effect and GVHD concurrently, treatment of GVHD without compromising GVL activity is a major challenge in the field. Current immunosuppressive medications used to prevent or treat GVHD often impair the T-cell–mediated GVL effect and increases the chance of relapse from primary malignancy.

Complement activation is involved in the pathogenesis of GVHD (4–7). Complement proteins secreted from antigen presenting cells (APC), such as dendritic cells (DC), regulate the expansion of naïve and primed effector T cells after allo-HCT (6, 8, 9). These findings suggest that local complement activation may be important for GVHD pathogenicity, as circulating complement in the serum does not correlate with GVHD. With regard to potential complement effector mechanisms, our previous data demonstrated that systemic blockade of complement anaphylatoxin molecules (C3a and C5a) severely impaired GVL activity (5). However, no attempt has been made to target complement activation in target organs to control GVHD and tumor relapse after allo-HCT.

Complement activation is essential for priming naïve T cells and their subsequent differentiation into Th1/Th17 in both the peripheral blood and lymphoid organs (10). Absence of intracellular C3 or C3aR/C5aR signaling on T cells or APCs induces the generation of Tr-1 (11, 12) and regulatory T cells (Tregs; refs. 13–15). Blockade of CD46, C3aR, or C5aR ameliorates T-cell alloreactivity in xenograft GVHD models (11, 14, 15). However, C3^−/− T cells have a comparable capacity to induce GVHD as compared with wild-type (WT) in murine GVHD models (16). Targeting C3 in transplant recipients showed a modest effect on GVHD severity (17). The aforementioned evidence indicates that specific inhibition of local complement signaling is required to control GVHD.

The complement system can be activated via three pathways: the classical (CP), lectin (LP), or alternative (AP; refs. 13, 18). The AP can both spontaneously activate, as well as amplify the CP and LP, and can potentially contribute up to 80% of total complement activation (19). Thus, the AP is important for inflicting maximum injury during the pathogenesis of autoimmune diseases (19–22). Among them, inflammatory bowel disease shares certain pathologic similarities with intestinal GVHD (23); a potent catalyst for systemic GVHD development. This evidence provides rationale for blocking the AP in GVHD target organs to effectively control GVHD.

In this study, using patient samples and murine models of allo-HCT, we demonstrate that complement activation specifically through the AP in injured target organs plays an essential role in GVHD after allo-HCT. To inhibit complement activation in GVHD target organs, we used a complement inhibitor comprised of an AP-specific inhibitory domain (fH) linked to complement receptor 2 (CR2) that selectively binds the membrane bound complement activation product C3d. CR2-fH effectively reduced GVHD while preserving the GVL effect after allo-HCT in multiple murine and human-to-mouse models of GVHD/GVL. Importantly, a humanized version of CR2-fH is currently available and was evaluated in a clinical trial (NCT01335165). Hence, this study has substantial translational potential for patients undergoing allo-HCT.

Tissue samples were collected from patients undergoing allo-HCT. Colon samples were collected from patients for diagnostic purposes, and the tissue blocks used in this study were selected from a sample repository. All the patients' specimens used in this study were deidentified. This study was approved by the Medical University of South Carolina (MUSC, Charleston, SC) Institutional Review Board and was conducted in accordance with International Ethical Guidelines for Biomedical Research Involving Human Subjects, Good Clinical Practice guidelines, the Declaration of Helsinki, and local laws.

fB^−/− and C1q/MBL^−/− mice on C57BL/6 background were bred and housed at the MUSC. FVB (H-2^q, CD45.1), C57BL/6 (H-2^b, CD45.2), B6 Ly5.1 (H-2^b, CD45.1), B6D2F1 (H-2^b/d, CD45.2), C3H-SW (H-2^b, CD45.1), and BALB/c (H-2^d) mice were purchased from the NCI or Charles River Laboratories. Animals were maintained in pathogen-free facilities in the American Association for Laboratory Animal Care–accredited Animal Resource Center at MUSC. All animal procedures were approved by the Institutional Animal Care and Use Committee of MUSC. CR2-fH was generated in Tomlinson's laboratory at MUSC.

As described previously (5), recipient mice were lethally irradiated at 700 cGy for BALB/c and 1,000–1,200 cGy (two split doses, 3 hour interval) for B6 recipient mice using an X-RAD 320 Irradiator (Precision X-Ray). Within 24 hours of irradiation, Balb/c or B6 recipient mice were transplanted with 5.0 × 10⁶ T-cell–depleted bone marrow cells (TCD-BM) from B6 or FVB donors with or without T cells (0.5–1 × 10⁶/mouse). Recipient survival was monitored throughout the experiment. Body weight loss was monitored twice per week and clinical signs of GVHD were monitored once per week and include posture, skin damage, hair loss, ruffled fur, diarrhea, and decreased activity (5, 24). For GVL experiments, tumor cells were injected intravenously on the same day of transplantation. In cases of luciferase transduced acute myeloid leukemia C1498 (2,000 cells/mouse), tumor burden was estimated with bioluminescent imaging (BLI) using Xenogen-IVIS 200 In Vivo Imaging System (Perkin-Elmer). MLL-AF9-GFP⁺ tumors were identified via the percentages of GFP⁺ cells in peripheral blood using flow cytometry. For human-to-mouse xenograft models, sublethally irradiated NSG-A2⁽⁺⁾ mice were transplanted with HLA-A2⁽⁻⁾peripheral blood mononuclear cells (PBMC; 15 × 10⁶). Recipient survival and GVHD severity demonstrated by recipient weight loss were monitored throughout experiment.

As described previously (5, 25), the following antibodies were used for cell-surface staining: anti–CD4 (clone RM4-5, BD Biosciences), anti–CD8 (clone 53-6.7, BD Biosciences), anti–H-2^q (KH114, BioLegend), anti–CXCR3-biotin (CXCR3-173, eBioscience), and anti–CCR6-AF647 (BioLegend, clone 29-2L17), anti-FasL (MFL3, BD Biosciences), anti-PD-1 (MFL3, eBioscience), anti-NK1.1 (PK136, eBioscience), anti-CD44 (IM7, BioLegend), anti-CD62L (MEL-14, eBioscience), anti-TCRβ (H57-597 BD Biosciences), and anti-CD11b (M1/70, eBioscience). Biotinylated antibodies were detected using APCcy7 (BD Biosciences, catalog no. 554063) or PEcy7 (BD Biosciences, catalog no. 557598) conjugated to streptavidin. To measure intracellular cytokines, cells were stimulated for 4–5 hours at 37°C with PMA (100 ng/mL, Sigma-Aldrich) and ionomycin (100 ng/mL; Calbiochem, EMD) in the presence of GolgiStop (BD Biosciences). Fix and permeabilization were performed using Cytofix/Cytoperm Plus (BD Biosciences), followed by staining with the appropriate antibodies including anti–IFNγ (clone XMG1.2, eBioscience), anti–IL-17 (clone TC11-18H10.1, BioLegend), anti–IL-4 (clone 11B11, BD Biosciences), anti–IL-5 (clone TRFK5, eBioscience), anti–FOXP3, (clone FJK-16s, eBioscience), anti-Ki67 (16A8, BioLegend), and anti–pS6-AF467 (Cell Signaling Technology, clone D57.2.2E). Live/dead Yellow Cell Staining Kit (catalog no. L-34968) and CFSE (catalog no. C1157) were purchased from Invitrogen. Apoptosis was measured by Annexin V Kit (BD Biosciences). XenoLight CF750 Labeling Kit was procured from Caliper. Data were analyzed with FlowJo Software (Tree Star). Blood was collected from recipients 14 days after bone marrow transplantation (BMT), and serum cytokines quantified using a Cytometric Bead Assay Kit (BD Biosciences, catalog no. 560485; ref. 25).

DCs were generated from the BM of 8- to 12-week-old mice as described previously (5). BM cells were flushed from the femurs and tibias with RPMI containing 1% FCS, 100 U/mL of penicillin/streptomycin, and 2 mmol/L l-glutamine. The single-cell suspension was then filtered through a nylon mesh strainer (70 mm; BD Biosciences), washing twice with the same medium. BM cells (10 × 10⁶/petri dish) were differentiated in the presence of GM-CSF (20 ng/mL, PeproTech) in complete culture medium (RPMI containing 10% FBS, 100 U/mL of penicillin/streptomycin, 2 mmol/L l-glutamine, and 50 mmol/L b-mercaptoethanol) for 6 days. Half of the medium was replaced with an equal volume of GM-CSF, containing culture medium on the day 3. Immature DCs were stimulated with lipopolysaccharide (LPS; 25 ng/mL, Sigma Aldrich) for 12 hours.

A similar approach was used to stain 10-μm sections as outlined previously (26). Primary antibodies used for these experiments were against murine antigens; anti-C3d (Goat Ab, R&D Systems), anti-CD3e, and anti-fB (Abcam). Directly conjugated antibody used was anti-C3d-FITC (MP Biomedicals). Secondary antibodies used were: Donkey anti-Goat-Alexa Fluor 488nm and 555nm, Donkey anti-rabbit-Alexa Fluor 488nm, 555nm, and 647nm, Donkey anti-rat Alexa Fluor 488nm and 555nm, and DAPI (Vector Labs) was used as nuclear stain. Super-resolution imaging was performed using a Zeiss LSM 880 Confocal Microscope (Zeiss). Uniform field sizes of 240 × 240 × 40 μm dimensions were imaged with a 40× water objective. High-resolution microglial fields were imaged with a 63× oil objective. Individual tracks were used for each channel with sequential imaging to avoid channel bleeding. 3D-rendering was performed using ZEN-Black Software (Zeiss).

Data were analyzed using Prism GraphPad (version 7). As described previously (5, 25), comparisons between two groups were calculated using a two-tailed Student t test. Clinical scores and body weight loss were compared using a nonparametric Mann–Whitney U test. The log-rank (Mantel–Cox) test was utilized to analyze survival data. A P value less than 0.05 was considered significant.

To examine the importance of complement activation in GVHD target organs post allo-HCT, we initially measured complement C3d deposition in the skin, lung, liver, small intestine, and large intestine. We found that C3d deposition was evident in injured target organs of allo-HCT recipients 7 days post-BMT compared with recipients that did not receive T cells (Fig. 1A). We next examined complement gene expression in the intestines of transplanted recipients using NanoString technology. Expression of genes encoding the central complement protein (C3) and AP proteins (fD, fB, and fP), were significantly upregulated in GVHD recipients compared with those without GVHD (Fig. 1B). In contrast, there was no difference in the expression of genes encoding CP or LP proteins (Fig. 1B). We also observed increased expression of genes encoding inflammatory proteins (IFNγ, IL6, and TNFα), cytolytic activity (GrzmB), and cell death (fas) in the intestines of mice with GVHD (Fig. 1C). These results implicate the AP as the central complement regulator in GVHD target organs. To determine whether the AP specifically played an augmented role in GVHD, we next performed allo-HCT using recipient mice deficient for AP (fB^−/−) or CP/LP (C1q/MBL^−/−). Immunologic phenotypes were similar across different mouse strains (Supplementary Fig. S1). We found that ablating the AP caused reductions in gene expression of molecules associated with inflammation, cytolysis, and cell death to levels comparable with recipients without GVHD (Fig. 1D). We observed that GVHD severity was drastically reduced in fb^−/− recipients, as indicated by improved survival and reduced GVHD clinical signs (Fig. 1E and F). Interestingly, the absence of CP/LP had no significant effect on GVHD severity in MHC-mismatched (FVB→B6) models of allo-HCT, supporting the notion that the AP played a more important role than the CP/LP in GVHD (Fig. 1E). To confirm this, we tested an MHC-matched (C3H.SW→B6) model, in which we observed that ablation of CP/LP resulted in accelerated GVHD compared with WT recipients (Fig. 1F).

In the absence of the AP in recipients (FVB-B6 model), complement deposition was markedly decreased in GVHD target organs including liver, lung, and large intestine (Fig. 1G). We also observed that serum proinflammatory cytokines, including TNFα and IFNγ were diminished in fB^−/− recipients after allo-HCT (Fig. 1H). These results indicate a crucial role for the AP, but not the CP/LP, in GVHD development after allo-HCT. As intracellular complement activation can affect T-cell alloreactivity (27), we investigated whether complement generated from donor T cells contributes to GVHD pathogenicity after allo-HCT. We therefore transplanted T cells from WT, fB^−/−, or C1q/MBL^−/− donors into allogeneic recipients. No difference was observed in GVHD severity (Supplementary Fig. S2), suggesting a negligible role for intracellular complement activation in T-cell pathogenicity in the induction of GVHD.

Complement activation via the AP can occur in both hematopoietic and nonhematopoietic recipient cells. Hence, we investigated whether the source of complement generation is important for GVHD development after allo-HCT. To this end, we generated three types of BM chimeric mice (BM chimeras), where the AP was absent in hematopoietic, nonhematopoietic, or present in both compartments (Supplementary Fig. S3A and S3B). Using BM chimeras as recipients, we observed that severe GVHD was evident in chimeras where fB was present in the recipient hematopoietic system [FVB→(WT B6→ fb^−/−)], while significantly reduced GVHD was observed in the chimeras where fB was absent in the recipient hematopoietic system [FVB→(fb^−/−→WT B6); Supplementary Fig. S3A and S3B]. Thus, complement activation through the AP in the recipient hematopoietic, but not nonhematopoietic, compartment plays an important role in the induction of GVHD.

Given that recipient-derived DCs are the most important hematopoietic cells in terms of generating complement to activate donor T cells in target-injured organs (9), we evaluated the contribution of AP-associated complement in recipient DCs to GVHD development post-HCT. BM-derived DCs were differentiated from WT or fB^−/− BM and matured via LPS stimulation. Upon cotransfer of fB^−/− DCs with donor grafts, GVHD-related mortality and morbidity was significantly reduced (Supplementary Fig. S3C and S3D). These data are indicative of a critical role for complement activation via AP in DCs during GVHD development.

Donor T-cell migration and activation in target organs is required for GVHD development (9). We observed a decreased number of donor lymphocytes in the intestine of fB^−/− recipients (Fig. 2A). The expression levels of the intestinal-homing chemokine CCR9 was consistently lower in lymphocytes isolated from fB^−/− recipients compared with WT controls (Fig. 2B). CD103⁺CD8⁺ T cells were previously reported to be pathogenic in the intestine (28). Hence, we examined this population. A lower percentage of donor-derived CD103⁺CD8⁺ cells were found in the intestines of fB^−/− recipients (Fig. 2). Host AP deficiency reduced the frequency (Fig. 3D) and activation of pathogenic CD103⁺ DCs (29), as demonstrated by lower CD80 expression (Fig. 2E) and increased activation as well as frequency of the protective CD8⁺CD11c⁺ DC population (Fig. 2F). Local complement activation results in generation of effector molecules, C3a and C5a, which can differentiate naïve T cells into Th1/Tc1 subsets and reduce Treg differentiation (27). Indeed, deficiency of recipient AP decreased Th1/Tc1 differentiation and increased Treg differentiation, which is demonstrated by reduced percentages of IFNγ-secreting CD4⁺ and CD8⁺ (Fig. 3A and B) and an increased percentage of foxp3⁺ CD4⁺-H2^q+ T cells (Fig. 3C) in the liver of transplanted recipients. Furthermore, Th2 differentiation was significantly reduced in the lungs of fB^−/− recipients indicated by a reduction in percentage of IL4/5-secreting H2^q+ T cells (Fig. 3D). Notably, thymic function was preserved in fB^−/− recipients, indicated by an increased number of CD4⁺CD8⁺ thymocytes in fB^−/− recipients compared with WT recipients (Fig. 3E). On the other hand, we observed similar immunologic characteristics between C1q/MBL and WT cohorts (Fig. 3A and B), indicating a vital role for the AP activation in driving GVHD development after allo-HCT. Interestingly, in lymphoid organs of fB^−/− recipients, we did not observe a significant difference in Th1/Tc1 differentiation (Supplementary Fig. S4A and S4B). While the percentage of splenic DC (Supplementary Fig. S4C) was significantly greater in fB^−/− compared with WT recipient; the numbers of splenic DC were similar (Supplementary Fig. S4G). There was a smaller percentage of CD103⁺DCs in the spleen of fB^−/− compared with WT recipients (Supplementary Fig. S4D). However, the number of these splenic DCs were comparable in different recipient types (Supplementary Fig. S4H). Both the number (Supplementary Fig. S4E) and percentage (Supplementary Fig. S4I) of splenic CD11b⁺DCs were remarkably increased in fB^−/− compared with WT recipients. DCs also expressed a comparable level of costimulator receptor CD80 in both fB^−/− and WT recipients (Supplementary Fig. S4F and S4J). Taken together, absence of the AP drives an anti-inflammatory phenotype of DCs and T cells selectively in target organs, but not in lymphoid organs, after allo-HCT.

Our results thus far indicate that activation of the AP is critical for GVHD pathogenicity. As such, we hypothesized that specific and site-targeted inhibition of the AP in GVHD target organs would effectively control GVHD. We utilized a fusion construct comprised of a complement inhibitory domain (fH) attached to a complement receptor 2 (CR2) fragment to selectively bind C3d, a complement activation product that is deposited in GVHD target organs (Fig. 1A). In initial studies, we labeled the inhibitor with a fluorophore to perform a visual “mapping” of its distribution among target organs in recipient animals (Fig. 4A). Balb/c recipients were transplanted with T cells from luciferase-transduced B6 donors. Migration of donor T cells to the intestine was observed in allogeneic (GVHD), but not syngeneic (without GVHD) recipients 7 days post allo-HCT (Fig. 4A). High fluorescence signal (CR2-fH) was detected in the colon (Fig. 4B), yet low signal was observed in lymphoid (spleen) and non-GVHD organs (heart) of GVHD recipients (Fig. 4B). The localization of CR2-fH in conjunction with donor lymphocytes to target organs suggests that CR2-fH preferentially affects complement generation from immune cells in GVHD target organs of transplanted recipients.

We next evaluated the effect of CR2-fH treatment on the GVL effect after allo-HCT. Given that CR2-fH has a tissue half-life of 2–3 days (30), we treated recipients (0.5 mg/mouse) with CR2-fH or vehicle (PBS) every other day for 14 days post allo-HCT. We observed that CR2-fH treatment effectively reduced GVHD development (Fig. 4C–E). Furthermore, there was a significant reduction in number of recipients treated with CR2-fH that were observed with a leukemia relapse compared with PBS-treated controls (Fig. 4E). These results demonstrate that blocking the AP prevents GVHD while preserving GVL activity.

To substantiate our observation, we examined the effect of CR2-fH on GVHD and GVL activity using a more aggressive and clinically relevant acute lymphoblastic leukemia, MLL-AF9. We observed that CR2-fH treatment dramatically improved recipient survival and decreased GVHD severity (Fig. 5A and B). In addition, on contrary to tumor relapse observed in the recipients of BM and MLL-AF9, the percentages of MLL-AF9 was markedly reduced in the recipients of BM and MLL-AF9 plus T cells regardless of CR2-fH treatment (Fig. 5C). Together, these evidences indicate that CR2-fH treatment reduces GVHD severity while preserving GVL activity. Donor CD8 T cells play a dominant role in mediating the GVL effect via CTL activity (31). In this context, we explored the mechanisms by which CR2-fH treatment preserved GVL activity. At 30 days post allo-HCT, the expression of key molecules driving CTL activity, granzyme B, and perforin, were preserved with CR2-fH treatment (Fig. 5D), which correlates with preserved GVL activity upon CR2-fH treatment.

To extend our findings toward clinical application, we analyzed complement deposition in the colon of patients either with GVHD or without GVHD based on pathologic diagnosis (Supplementary Table S1). We observed that the deposition of complement C3d was higher in biopsies of patients with GVHD (Fig. 6A), and that increased C3d deposition was positively correlated with the number of CD3⁺ lymphocytes (Fig. 6A and B). We also found a trend indicating higher expression of fB in the intestines of patients with GVHD (Fig. 6A and B). Together, these findings suggest that complement activation through the AP likely contributes to GVHD development in patients after HCT.

To further enhance translational potential, we examined the effect of CR2-fH in a human-to-mouse xenograft model. Sublethally irradiated NSG-A2⁽⁺⁾ mice were transplanted with HLA-A2⁽⁻⁾ PBMCs and treated with vehicle or CR2-fH. We found that CR2-fH treatment sufficiently prevented GVHD progression in transplanted recipients. No recipient mortality was observed in transplanted recipients treated with CR2-fH (Fig. 6C). In addition, treated recipients had improved body weight maintenance and donor engraftment was not affected (Fig. 6D).

Complement can be activated via three different pathways (19), but the role of each individual pathway in GVHD has not been extensively studied. Here, we show that specifically inhibiting the AP of complement activation reduces local inflammation within target organs. Furthermore, these data represent a novel approach to treat GVHD while preserving the GVL effect. While a predominant role for the AP in the progression of various autoimmune diseases has been reported, we show for the first time that complement activation through the AP, specifically in target organs, also drives GVHD development after allo-HCT (Fig. 1E). It is not clear how the AP is activated, but intestinal damage caused by total body irradiation conditioning induces translocation of LPS in the gut lumen and exacerbates GVHD severity (32). We found that the AP drives complement activation and GVHD via its activity in target organs (Figs. 1D–G and 6A), but that effects on lymphoid organs were arbitrary. Indeed, AP deficiency decreased the capacity for alloantigen presentation of DCs and reduced the generation of a pathogenic phenotype for DCs in target organs.

Although the role of DC-derived complement has been reported previously (6), it is not clear which pathway is the primary contributor to the complement deposition in GVHD target organs; in particular intestinal GVHD. In this study, we demonstrated that complement activation through the AP is critical for the complement generated in injured target organs (Fig. 1G). Specifically, we have shown that complement effector proteins derived from host DCs play a dominate role in promoting donor T-cell allogeneic responses and GVHD pathogenicity (Supplementary Fig S3C and S3D). Our previous study indicates that complement regulates recipient DCs via modulating their survival during GVHD (5). This study provides evidence that complement activation via the AP is important for DC's ability to generate effector molecules. These deposited complement products are able to regulate the antigen presentation capacity of DCs, which is required for GVHD development.

Complement activation has been observed in nonhematopoietic cells, such as fibroblasts (33), endothelial (34), and epithelial cells (35). Despite their capacity to prime T cells, host hematopoietic APCs (DCs, B cells, and macrophages) are required for maximum activation of donor T cells and subsequent GVHD induction after allo-HCT (9). Therefore, this study suggests an important role for host DCs in amplifying complement activation and deposition in GVHD target/injured tissues.

Complement activation results in differentiation of naïve T cells into Th1/Tc1 through the generation of the anaphylatoxins C3a and C5a (36), and the opsonin C3d (37). We found that, in the absence of the AP, donor T cells skew from Th1/Tc1 differentiation toward Treg generation in GVHD target organs (Fig. 3). Given Th1 cells are pathogenic and Tregs are suppressive in GVHD, T-cell differentiation is an important factor that contributes to effect of the AP on GVHD development (38). Because stimulating DCs by complement-activating products (C3d, C3a, and C5a) creates Th1/Th17 priming cytokines (39, 40), the effect of AP-generated complement products on DCs may precede the modulation of donor T-cell immunity in GVHD recipients.

T-cell migration to GVHD target organs is critical for GVHD development (41). As complement activation generates the chemotactic molecule C5a, which is important for lymphocyte migration to target organs (42), reduced levels of complement activation may lead to fewer infiltrating donor lymphocytes in target organs of AP-deficient recipients (Fig. 2A). In addition, the decreased expression of target organ homing chemokines, such as CCR9 (43), on donor T cells (Fig. 2B) could likely contribute to less T-cell accumulation in the intestine of recipients with GVHD.

The intestine is a primary site of organ damage in GVHD, and intestinal injury also plays a role in amplifying systemic GVHD development (38). In this study, this association is linked with GVHD and complement activation in the intestines in both mice and humans (Fig. 6A and B). Our findings are consistent with a previous study, in which C3 deposition was seen in the skin of patients with acute GVHD (37). Importantly, we provide compelling evidence that local complement activation, specifically in injured target organs, plays an important role in organ damage during GVHD development (Fig. 4A). Furthermore, our study shows that complement effector molecules involved in the pathogenesis of GVHD are generated mainly through the AP in recipients with intestinal GVHD (Fig. 4A). In contrast, we did not observe a significant role for the CP/LP in GVHD development. Consistently, immunologic characteristics of GVHD-driven immune cells in target organs were similar between WT and C1q/MBL^−/− recipient mice (Figs. 2 and 3). Presumably this is due to the fact that complement activation through the CP/LP is dependent on the AP to propagate maximal damage in GVHD target organs (21).

Together, these data suggest that inhibition of the AP in GVHD target organs may provide a therapeutic option for controlling GVHD. Indeed, we found that pharmacologically targeting the AP during complement activation in GVHD target organs effectively protected against GVHD while preserving GVL activity after allo-HCT (Fig. 4C–E). We reasoned that complement inhibition did not affect the differentiation and expansion of donor T cells in lymphoid organs, which may permit these T cells to maintain their role in mediating the GVL effect after allo-HCT (44). In line with this notion, the immunologic phenotype of DCs and T cells were largely unaffected by AP deficiency in the lymphoid organs of transplanted recipients (Supplementary Fig. S4). Accordingly, the CTL activity of donor CD8⁺ T cells (45), an important factor responsible for GVL activity after allo-HCT, is maintained during CR2-fH treatment.

In conclusion, we investigated the role of complement in GVHD/GVL responses after allo-HCT. Specifically, we demonstrate that complement activation, specifically through the host-derived AP activation, is critical for GVHD pathogenesis. Furthermore, specific and localized inhibition of the AP using a site-targeted inhibitor provided protection from GVHD while maintaining GVL activity after allo-HCT. Given that a humanized version of CR2-fH, termed TT30, has been in clinical trial for patients with paroxysmal nocturnal hemoglobinuria (46), these findings have high translational potential for patients undergoing allo-HCT.

J.C. Varela is an employee/paid consultant for and holds ownership interest (including patents) in NexImmune. S. Tomlinson has a patent owned by the Medical University of South Carolina, licensed to Alexion Pharmaceuticals, on targeted complement inhibitors. No potential conflicts of interest were disclosed by the other authors.

Conception and design: H. Nguyen, A. Alawieh, C. Atkinson, J.C. Varela, X.-Z. Yu

Development of methodology: H. Nguyen, S. Tomlinson

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Nguyen, A. Alawieh, D. Bastian, S. Kuril, M. Dai, A. Daenthanasanmak, M. Zhang, S. Iamsawat, S.D. Schutt, M.M. Sleiman, A. Shetty, S. Sun, J.C. Varela

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Nguyen, A. Alawieh, S. Kuril, S.D. Schutt, Y. Wu, C. Atkinson, S. Tomlinson, X.-Z. Yu

Writing, review, and/or revision of the manuscript: H. Nguyen, A. Alawieh, D. Bastian, M.M. Sleiman, A. Shetty, C. Atkinson, J.C. Varela, S. Tomlinson, X.-Z. Yu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): H. Nguyen, Y. Wu, S. Tomlinson, X.-Z. Yu

Study supervision: H. Nguyen, X.-Z. Yu

This work was supported by NIH grants R01s HL140953 and R01HL137373 to X.-Z. Yu, University of Central Florida (UCF) Start-up Grant no. 2540-0715 to H. Nguyen. Institutional resources at the Medical University of South Carolina were supported by NIH support C06 RR15455 and P30 CA138313 grants (to Hollings Cancer Center). We thank cell and molecular imaging, flow cytometry, and pathology cores at the Medical University of South Carolina (MUSC) for their valuable services. We also thank Dr. Angie Duong at MUSC for her coordination in obtaining patient biopsies.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.