The following antibodies were purchased from Biolegend or BD Bioscience: B220-Pacific Blue (RA3-6B2), CD3e-FITC & APC-Cy7 (145-2C11), CD4-APC (GK.1.5), CD8α-BUV395 & PE (53-6.7), CD11b-PerCP Cy5.5 (M1/70), CD11c-BUV786 (N418), CD25-APC-Cy7 (3C7), CTLA-4-PE (UC10-4F10-11), Gr-1-PE-Cy7 (RB6-8C5), ICOS-BUV510 (C398.4a), Ly6C-BUV510 (HK1.4), Ly6G-APC (1A8), MHC II (1-A^b)-FITC (KH74), MHC II (1-A^d)-FITC (39-10-8), and PD1-BUV786 (29F.1A12). αGalCer/mCD1d tetramers were provided by NIH tetramer facility.

All animal experiments were conducted per The University of Texas MD Anderson Cancer Center’s Institutional Animal Care and Use Committee guidelines. Six to 8- week old female, C57BL/6 and BALB/c, mice were purchased from Jackson Laboratories (Bar Harbor, ME).

BRNPs were prepared as described previously (12) and lyophilized for long term storage. Briefly, polyethylene glycol-modified bilirubin (PEG-BR) was dissolved in chloroform and dried under nitrogen gas, where it was further dried to generate a film layer. To formulate the nanoparticle, PEG-BR based film was hydrated with phosphate buffered saline (PBS) where it was sonicated for 10 minutes to produce BRNPs. For longer term use, BRNPs were lyophilized and stored at -30°C. Lyophilized BRNPs were resuspended in PBS at a concentration of 5 mg/ml prior to use and stored at 4°C and used within 7 days.

Indocyanine green (ICG), a near-infrared fluorophore, was loaded in the BRNPs for in vivo near-infrared imaging of organs from injected mice using Xenogen IVIS Lumina in vivo imaging system. To yield ICG-BRNPs, the film layer of PEG-BR was hydrated using 1 mg/mL aqueous solution of ICG followed by sonication for 15 minutes in the dark (14).

Ten to 15- week old female BALB/c mice were irradiated at a myeloablative dose of 8 Gy using ¹³⁷Cs irradiator on day 0. The next day, mice intravenously received a total of 5 x 10⁶ bone marrow cells with or without 5 x 10⁶ splenocytes from age-matched donor female C57BL/6 mice, and with 5 daily intravenous doses of 10 mg/kg BRNP or equivalent volume of vehicle from day 0 to day 4. Subsequently mice were assessed three times weekly for survival and clinical GVHD.

Clinical GVHD was scored as previously established by Cooke et al., 1996 (17). Briefly, recipient mice were scored from 0-2 for Weight (0: <10% loss, 1: ≥10% <25%, 2: ≥25%), Fur (0: normal, 1: mild to moderate, 2: severe ruffling), Posture (0: normal, 1: kyphosis at rest, 2: kyphosis impairing movement), Activity (0: normal, 1: stationary 50% of the time, 2: stationary unless stimulated), and Skin (0: normal, 1: scaling paws or tails, 2: lesions). Moribund mice were sacrificed according to IACUC guidelines.

Recipient mice were sacrificed on day 8 after the transplantation. Organs (spleen, liver, lung, skin, small intestine) and serum was procured, and part of organs were fixed with 10% buffered formalin. Fixed organs were embedded in paraffin and sectioned in 6µm thickness, followed by tissue staining with hematoxylin and eosin. GVHD associated damages were evaluated by pathologist in a blind manner. Pathologic GVHD score was defined as follows: Lung: the presence of perivascular and peribronchial inflammation (ratio) and degree of interstitial pneumonia (1:<25%, 2: <50%, 3: <75%, 4:≥75); Liver: the presence of portal and central vein inflammation (ratio) and number of necrotic foci per view in 10x object, small intestine; number of single cell necrosis per view in 40x object; skin: presence of lymphocyte infiltrates in hair follicles (ratio).

Single cell suspension was prepared from liver, spleen, and lung, and stained with antibodies specific for various surface antigens such as CD3ε, CD4, CD8α, CTLA-4, PD-1, ICOS, B220, iNK TCR, and Fixable Viability Stain 620 in PBS for 30 min at 4°C. The cells were then washed with 1X PBS, fixed in 2% paraformaldehyde, and stored at 4°C until acquisition. For intracellular staining of transcription factor – FoxP3 and Helios, cells were fixed and permeabilized using BD Cytofix/Cytoperm buffer (Beckton Dickinson, NJ) for 20 min after surface antigen staining, followed by incubation with antibodies specific for transcription factors in BD Perm/Wash buffer for 45 min. After wash, the samples were acquired using LSRFortessa™ X-20 Cell Analyzer (Beckton Dickinson, NJ), and FlowJo version 10.3 (Tree Star, OR) was used for analysis.

First, blood from each individual mouse was collected by retro-orbital vein puncture. The blood from each mouse was placed in tubes to be centrifuged at 600xg for 5 minutes. Serum was collected and stored at -80°C until further analysis. Sera were diluted 1:4 in assay diluent from BD™ Cytometric Bead Array (CBA, Becton-Dickinson, NJ) to determine the serum cytokine concentrations of Th1/Th2/Th17 type cytokines (IL-2, IL-4, IL-6, IFN-γ, TNF-α, IL-17A, and IL-10). The sera were incubated with mixed capture beads and subsequently washed according to the manufacturer’s instructions. Samples were acquired using LSR Fortessa Cell Analyzer (Beckton, Dickinson and Company, Franklin Lakes, NJ), and FlowJo version 10.3 (Tree Star, Ashland, OR) was used for analysis.

On day 0, 10- to 15- week old BALB/c mice received a myeloablative dose of 8 Gy TBI. The following day, mice were intravenously infused with 5 x 10⁶ bone marrow cells with or without 5 x 10⁶ splenocytes isolated from C57BL/6 mice. Additional BALB/c mice received 8 Gy TBI on day 4. On Day 5, mice were intravenously infused with BRNP-ICGs at a single dose of 35 mg/kg per mice and sacrificed 5 hours after infusion. The major organs (spleen, kidney, heart, lungs, liver, small intestine, and large intestine) were harvested to measure fluorescence intensities using a Xenogen IVIS Lumina in vivo imaging system (PerkinElmer, MA) with an ICG filter channel and an exposure time of 5 seconds.

All statistical analyses were performed using GraphPad Prism Software version 8.00 (La Jolla, California) for Windows. Data sets were analyzed using either student t-test or Mann-Whitney U tests and p values for comparisons between groups were determined. For survival analysis, Log-rank (Mantel-Cox) survival test was used. P value less than 0.05 was considered to be statistically significant.

Previous studies have demonstrated that BRNPs selectively accumulate at the area of inflammation elicited by various tissue injury from ischemia, allergen, dextran sodium sulfate and as well as autoimmune processes (10–15). As acute GVHD cascade involves systemic and local inflammation caused by conditioning regimen and infiltrating donor immune cells, we sought to investigate whether BRNPs preferentially accumulate in the area of initial tissue injury – especially in the liver and gut from conditioning regimen of radiation commonly used in allogeneic stem cell transplantation. In bone marrow transplantation (BMT), alloreactive memory T cells or primed naïve T cells may start proliferating from 3-5 days after transplantation, further magnifying the inflammation from initial tissue injury, leading to detectable pathologic findings of GVHD as early as day 7. Therefore, we sought to demonstrate the biodistribution of BRNPs at day 6 after conditioning regimen with total body irradiation (TBI) at a myeloablative dose followed by MHC mismatched donor cell transplantation ( Figure 1 ).

First, there was a physiologic accumulation of ICG-BRNPs in colon through the course. ICG-BRNPs were detectable in the small intestine of recipient mice, 6 days after TBI, and significantly increased in mice after major MHC mismatched BMT. ICG-BRNP accumulation was observed in additional GVHD-target organ, i.e. liver of the mice only after BMT, and was not detectable in the non-GHVD target organs, that is spleen, kidney, and heart of both the groups ( Figure 1B ). These results support that BRNP preferentially accumulates at GVHD-target organs after the BMT.

Previous in vivo pharmacokinetics study demonstrated that half-life of BRNPs is 9.3 hours with a peak level of bilirubin in serum being 100 mg/dl immediately after BRNPs at 50 mg/kg was intravenously infused (11). As serum bilirubin can increase in setting of liver injury from conditioning regimen and GVHD post transplantation (1–3), we chose BRNPs given to recipient mice intravenously daily at a dose of 10 mg/kg from day 0 to day 4 to allow near normalization of serum bilirubin level by day 7 or 8. Moreover, single or multiple infusions of BRNP at a dose of 10 mg/kg has been shown to effectively prevent hepatic ischemic perfusion injury (10), and pancreatic islet xenotransplantation (16), supporting our dose and dosing strategy.

After myeloablative irradiation, BALB/c mice (MHC d-haplotype: H-2^d, IA/E^d) received 5x10⁶ bone marrow cells isolated from C56BL/6 on day 1 to reconstitute recipient immune system, or 5x10⁶ splenocytes in addition to 5x10⁶ bone marrow cells to induce graft versus host disease. Portion of these mice were infused with BRNP at a dose of 10 mg/kg or vehicle control daily from day 0 to day 4 ( Figure 2A ). Mice were monitored for clinical GVHD (weight, fur, activity, posture, and skin) and survival for 90 days. We observe transient weight loss and corresponding increase in GVHD score around day 7, which was related to radiation treatment ( Figure 2B ). As expected, recipient mice with BMT only did not develop clinical GVHD and survived through day 90 from BMT ( Figure 2C ), but mice infused with a mixture of splenocytes and bone marrow cells started to develop progressive clinical signs of GVHD around day 14-such as ruffled fur, hunched back, decreased activity, and weight loss compared to BMT control (p=0.0007) with significant GVHD-related mortality (HR 5.898, p=0.0134). Treatment with BRNP prevented recipient mice from developing clinical GVHD and resulted in improved overall survival compared to vehicle treatment group (HR 0.2055, p=0.0112), demonstrating that targeting early inflammation cascade with BRNP is feasible and has therapeutic benefits in BMT ( Figure 2D ).

The initial tissue injury may start from conditioning chemotherapy or radiation in transplantation, but subsequent damage to organ mediated by alloantigen specific donor T cells is detrimental. In murine BMT model, GVHD related tissue damage can be detectable as early as day 7 post transplantation. To investigate whether BRNPs rescue recipient mice from GVHD-related tissue damage, we performed histopathological analyses on GvHD target organs such as the liver, lung, gastrointestinal tract (GI), and skin on recipient mice at 8 days after major MHC-mismatched BMT ( Figure 3A ).

We observed a significantly lower inflammatory immune cell infiltration in hair follicles of skin (p=0.0232), less inflammation in the portal tracks and central veins of liver (p=0.0373, p=0.0042, respectively), less single cell necrosis in small intestine (p=0.0029), and a trend of decrease in peribronchiolar and perivascular inflammation in lung (p=0.1007, p=0.0524, respectively) from BRNP treated mice when compared to the mice treated with vehicle ( Figure 3B ). Of note, histopathologic finding of tissue injury from recipient mice treated with BRNP did not have significant differences from recipient mice with BMT only, demonstrating that BRNP treatment peri transplantation could prevent histopathologic GVHD of target organs.

Inflammatory cytokines such as IL1-β, TNF-α, IL-6, IFN-γ, etc., are first produced by innate immune cells in response to tissue damage caused by conditioning regimen, this inflammatory responses is further augmented once alloantigen specific T cells are activated and expanded in the area of damaged tissue (8). Previous studies indicate higher serum levels of these pro-inflammatory cytokines are associated with higher organ damage causing more severe onset of acute GVHD (18). We also observed significantly higher levels of inflammatory cytokines such as IFN-γ and TNF-α ( Figure 4 ) in sera from mice with GVHD than those from mice without GVHD as early as 8 days post transplantation, confirming systemic inflammatory responses occurred during GVHD. Another proinflammatory cytokine, IL-6, was detectable but there was no difference between mice with and without GVHD, likely due to the time of evaluation (data not shown). The sera level of IFN-γ was significantly reduced in mice treated with BRNPs compared to mice without BRNPs (p=0.0164) but still elevated compared to mice with BMT only (p=0.0479). In contrast, there was no difference in the serum of TNF-α in mice treated with BRNPs compared to untreated mice with GVHD, and still significantly elevated compared to mice with BMT only (p=0.0039). Our results indicate that BRNP treatment peri-transplantation reduced the sera level of IFN-γ, a potent proinflammatory cytokines produced by lymphoid cells (19).

GVHD is a disease of dysregulated donor T cells where naïve T cells are inadvertently primed to recognize alloantigens released during initial tissue injury, and subsequently mediate further tissue damage in transplant recipients. As we observed that BRNP treatment significantly improved clinical GVHD and GVHD-related mortality by the reduction of local tissue damage and systemic inflammatory cytokines, we investigated whether BRNP treatment prevents the early recruitment of donor derived T cells along with various inflammatory or regulatory immune cells such as neutrophils (CD11b⁺GR1⁺), invariant Natural Killer T cells (iNK T, αGalCer/CD1d tetramer⁺) or regulatory T cells (Treg, CD4⁺FoxP3⁺Helios⁺) 8 days after major MHC mismatched BMT among mice that received BMT only for normal hematopoiesis, bone marrow plus splenocytes for inducing GVHD, and bone marrow plus splenocytes with BRNP treatment ( Supplementary Figure 1 ). First, we did not see significant differences in frequencies of T cells or consistent differences of major T cells type – CD4⁺, CD8α⁺, CD4^-CD8α^- T cells systemically in spleen or locally in lung or liver, between mice developing GVHD with or without BRNP treatment ( Figure 5A ). However, there was a significantly delayed reconstitution of another innate regulatory immune cells, iNK T cells, in mice developing GVHD, and BRNP treatment increased the frequency of iNK T cells, in spleen and lung and reduced in liver ( Supplementary Figure 2 ). Interestingly, CD8α⁺ iNK T cells, an effector/inflammatory subset of iNK T cells, were significantly increased in spleen, and locally in liver in mice developing GVHD (p = 0.0952, p = 0.0952, respectively), with a reciprocal decrease in CD4⁺ iNK T cells, regulatory subset of iNK T cells (p = 0.0952, p = 0.0952, respectively), compared to mice with BMT only. BRNP treatment peri-transplantation significantly blunted the expansion of CD8α⁺ iNK T cells in spleen and liver and let to restoring CD4⁺ iNK T subsets ( Figure 5B ). Of note, mice developing GVHD had significantly lower frequencies of conventional Tregs systemically in spleen and locally in liver and lung compared to mice with normal hematopoiesis, and there was a trend towards an increase in splenic Treg frequencies with BRNP treatment ( Figure 5C ). Lastly, there was an expansion of neutrophils in the spleen, locally in liver and lung of mice with GVHD compared to those of mice with normal hematopoiesis. Again, BRNP treatment reduced infiltration of neutrophils compared to mice with no treatment ( Figure 5D ).

Although we did not see consistent alteration in major T cell subsets associated with GVHD, we observed significant upregulation in Inducible T cell Costimulator (ICOS) in both CD4⁺ and CD8α⁺ T cell subsets in mice developing GVHD compared to those with normal hematopoiesis, and the expression of ICOS was reduced after BRNP treatment ( Figure 6 ). Likewise, immune checkpoint inhibitors, - PD-1 but not CLTA-4 were highly upregulated in T cells from mice developing GVHD compared to those with normal hematopoiesis but BRNP treatment reduced PD-1 expression only in splenic T cells ( Supplementary Figure 2 ).

Altogether, we demonstrated that inflammatory cells such as neutrophils or CD8α⁺ iNK T cells were significantly infiltrated, locally in mice with GVHD, and BRNP treatment blunted expansion of neutrophils and CD8α⁺ iNK T cells. Although there was no significant alteration of major conventional T cell subsets, there was significant upregulation of ICOS and PD-1 in T cells during GVHD, and again the expression of these proteins was reduced with BRNP treatment, more in splenic T cells than liver or lung residential T cells. Our findings suggest that BRNP treatment ameliorates GVHD-related tissue damage via dampening early activation of conventional T cells and infiltration of inflammatory innate immune cells.