Lack of glucagon-like peptide-1 receptor signaling worsens blood stem cell transplant rejection in mice

All in vivo and in vitro mouse experiments utilized age-matched, female mice that are 8 to 16 wk old. Mice were housed in a temperature-controlled room at 22.2 °C on a half day light-dark cycle. Mice were fed a regular diet (PicoLabLaboratory Rodent Diet 5LOD), and tap water was available ad libitum. Mouse experiments were approved by the Institutional Animal Care and Use Committee (IACUC) at Vanderbilt University Medical Center and were conducted according to the guidelines for the Care and Use of Laboratory Animals prepared by the Institute of Laboratory Animal Resources, National Research Council. BALB/c and C3.SW-H2/SnJ mice were purchased from Jackson Laboratories and housed in our colony until the time of experimentation. GLP1RKO and WT controls on the C57BL/6J background were bred and maintained in our colony. The creation and characterization of these mice has been previously described. ® b , ^{251 251}

Bone marrow-derived HSCs were harvested from the tibias and femurs of donor mice (C57BL/6J WT, BALB/c WT, C3.SW-H2/SnJ). Briefly, bones were dissected from mice euthanized with pentobarbital overdose. Bones were cleaned of all soft tissue and cut at the ends with a scalpel. Up to 4 bones were placed in 0.5 ml snap tubes that had been punctured at the bottom with a 19-gauge needle. In brief, 0.5 ml tubes containing bones were closed and placed inside of larger 1.5 ml snap tubes. The combined tubes were centrifuged to collect bone marrow cells. Cells were passed through a 70 μM strainer and erythrocyte lysis was carried out with an ammonium chloride-based lysing reagent (Tonbo). Cell counts were obtained with a hemocytometer and an aliquot of cells diluted in Trypan blue. Bone marrow cells collected from mice of identical strains were pooled to create the HSC graft. All steps after RBC lysis were performed in BMT Media (1640 RPMI w/L-glutamine, RMPI 1640 with 10% fetal bovine serum (FBS), Penicillin/Streptomycin, HEPES, Sodium Pyruvate). b ®

HSCs were isolated and depleted of RBCs as described above. Cells were incubated with Mouse Fc Block (BD™), and subsequently with anti-mouse Thy1.2 (CD90.2) (30H12) in BMT Media (5 μg/mL) for 30 min. Lysis of labeled cells was carried out with Low-Tox-M Rabbit Complement (Cedarlane) diluted 1:120 in BMT Media at 37 °C for 45 min. ® ®

Spleens were harvested from donor mice (C57BL/6J WT, BALB/c WT, C3.SW-H2/SnJ) euthanized with pentobarbital overdose. Spleens were smashed through a 70 μM strainer and erythrocyte lysis was carried out with an ammonium chloride-based lysing reagent (Tonbo). Cell counts were obtained with a hemocytometer and an aliquot of cells diluted in Trypan blue. T cells were isolated from splenocytes with the Pan T Cell Isolation Kit II, Mouse (Miltenyi™) according to the manufacturer protocol. b ®

All mice were conditioned for transplantation (C57BL/6J WT and GLP1RKO) with lethal irradiation through placement in a Cesium-137 irradiator. Lethal irradiation doses were split evenly between two irradiations 3 h apart to limit gastrointestinal toxicity. Mice, all on the C57BL/6J background, received 11 Gy (2 × 5.5 Gy) in a 12-chamber Mouse Pie Cage (Braintree) on a rotating platform. Retro-orbital transfer of BM HSCs and T cells occurred 2 h after the last irradiation. Mice were anesthetized with a formulation of ketamine and xylazine injected intraperitoneally. Grafts were injected into the retro-orbital sinus of mice with a 1 ml TB syringe in 100 µl of serum-free RPMI. All grafts used 5.0 × 10BM HSCs. T cell containing grafts for transplant experiments in,, andused 1.5 ×10T cells. For the T Reg experiment performed in, 5.0 ×10BALB/c BM HSCs + 1.5 ×10Balb/c T cells were given with either no T Regs, 6.0 × 10WT C57BL/6J T Regs, or 6.0 ×10GLP1RKO C57BL/6J T Regs. Mice were kept on heating pads after injections to allow proper recovery. For the duration of an experiment, mice were housed in standard caging, but cages were changed daily to ensure cleanliness and reduce risk of infection. For survival experiments, mice were marked as dead if (1) the mouse died naturally, (2) the mice lost 30% of its body weight and was euthanized, or (3) the mouse was recommended for euthanasia by animal care due to illness (having reached a humane endpoint). ® 6 5 6 5 5 5 Figs. 1 2 3 Fig. S2

Mice were euthanized with pentobarbital overdose immediately prior to organ harvest. All tissues were fixed in 10% formalin. Lungs were insufflated with 10% formalin prior to excision. Livers were excised whole. Duodenum sections of the small intestine were taken (approximated as the proximal 1/3 of the distance from the pyloric sphincter to the ileocecal valve). Large intestines were cut distal to the cecum and at the rectum. Intestinal sections were cut longitudinally and fixed for 1 d in folded filter paper. Intestines were then rolled and held with a 30-guage needle and allowed to fix for an additional day. Samples were embedded in paraffin, cut, and stained with hematoxylin and eosin. Assessment of GVHD pathology was carried out by a board-certified veterinary pathologist who was blinded to the groups.

Serum was collected via cardiac puncture of recently euthanized mice. Blood was collected in snap tubes and allowed to coagulate before centrifugation. Cytokine detection was carried out via Luminex assay using x-map technology via the MagPix system. The assay was carried out by the VUMC Analytical Services Core. The MILLIPLEX MAP Mouse Cytokine/Chemokine Magnetic Bead Panel—Premixed 32 Plex—Immunology Multiplex Assay (Millipore™) was used for analysis of the following cytokines: interferon (IFN)-γ, interleukin (IL)-1β, IL-2, IL-3, IL-6, IL-10, IL-17, and tumor necrosis factor (TNF)-α.

Flow cytometry was performed on cells from the bone marrow and peripheral blood. Isolation of bone marrow cells was identical to the method described in the “Isolation of Bone marrow-derived HSCs” section. Blood was collected via cardiac puncture of recently euthanized mice. Blood was deposited into 0.5 ml EDTA lined purple top vacuum tubes. RBC lysis was carried out, and single cell suspensions were counted for staining. Cells were first stained in a Live/Dead™ Fixable Aqua Dead Cell stain kit to assess viability. Cells were then blocked with BD Fc Block™ for 10 min in FACS buffer (3% FBS in 1× phosphate buffered saline [PBS]). Fluorescent antibodies for surface staining were added for 20 min. Cells were washed and resuspended in the eBiosciences™ FoxP3/Transcription Factor Staining Buffer set overnight at 4 °C. Intracellular staining was carried out for 45 min in manufacturer provided permeabilization buffer. Flow cytometry was performed on a 4-laser CytekAurora. Known quantities of 123count eBeads™ were added to samples to assess total cell quantities accurately. Analysis of flow cytometry data was performed on FlowJo software (10.8.1). ®

Mice were injected intraperitoneally with 500 μg of anti-mouse Thy1.2 (CD90.2) (30H12) depleting antibody resuspended in 100 μl of sterile PBS. Injections occurred 4 d before irradiation and transplant.

Naïve CD4T Cells were isolated from spleens of WT and GLP1RKO mice on the C57BL/6J background with the Easy Sep™ Naïve CD4+ T Cell Isolation Kit (STEMCELL™). Cells were plated on a FalconNon-Tissue Culture Treated 24-well plates previously coated with Ultra-LEAF α-CD3ε (BioLegend) at a density of 500,000 cells per well with 10 μg/ml of recombinant TGF-β (Peprotech) and 100 IU/ml of recombinant human IL-2 (NIH). Cells were cultured for 4 d before harvest for Flow Cytometry analysis and adoptive transfer with BM HSCs + splenic T cell grafts. + ® ®

All statistical analysis was carried out in GraphPad Prism 10 software. The significance of differences in survival represented in Kaplan-Meier curves were assessed with a Mantel-Cox test (Logrank). Significance for weight curves was assessed with a 2-way repeated measures analysis of variance (ANOVA). Tukey’s multiple comparisons test was used for comparison between groups. Cytokine concentration and GVHD score data were assessed with a 2-way ANOVA with Tukey’s multiple comparisons test. Significance in flow cytometry data was assessed with an un-paired-test. t

To assess the role of GLP1R signaling in mediating GF, we performed a set of HSCTs into WT and GLP1RKO mice on the C57BL/6J background. These mice received HSCs from C57BL/6J (Syngeneic), C3.SW-H2/SnJ congenic mice (allogeneic, minor antigen-mismatch) or BALB/c (allogeneic, MHC mismatch). For each transplant, a small dose of splenic T cells from the HSC donors was included with the graft to mimic the inclusion of donor T cells in human HSCT procedures. A T cell dose of 1.5 ×10was low enough to not cause GVHD and to not overcome the residual T cells and NK cells in the host that might reject the HSC graft. Additionally, we performed the HSCTs within 2 h after the last irradiation session of conditioning, thereby decreasing the likelihood of complete host T cell and NK cell clearance upon contact with the graft. After transplant, mice were monitored for weight loss and survival (). In this model, WT and GLP1RKO mice receiving syngeneic transplants fully recovered their weight after HSCT and experienced no mortality out to day 25. There was no significant difference in either weight loss () or survival (). An initial dip in weight loss was observed in all groups, which is likely attributed to illness related to the Csirradiation. Similarly, there was no significant difference in weight loss () or survival () between WT and GLP1RKO mice receiving minor antigen-mismatched grafts. Both groups recovered their weight after 60 d, and there was no significant mortality. In contrast, GLP1RKO mice experienced significantly increased weight loss () and mortality () compared to WT mice after MHC-mismatched HSCT. While 80% of WT mice survived out to 28 d, all GLP1RKO mice died within 24 d post-transplant. b 5 137 Fig. 1A Fig. 1B Fig. 1C Fig. 1D Fig. 1E Fig. 1F Fig. 1G

Given the survival difference that existed between WT and GLP1RKO mice, we hypothesized that GLP1RKO mice exhibited worsened GF after MHC-mismatched HSCT compared to WT controls. To test this hypothesis, we repeated our transplant model frombut harvested all mice at day 7 post-transplant to assess engraftment in both the peripheral blood and bone marrow compartments. For WT and GLP1RKO mice receiving syngeneic HSCT, we compared engraftment by enumerating live CD45positive cells in the peripheral blood and bone marrow. Due to the strains available, we were unable to differentiate between recipient and donor cells for syngeneic and minor antigen-mismatch transplants. Instead, total CD45+ cell counts were used to assess graft survival. Given that mice receiving syngeneic and minor antigen-mismatched grants survived in our model, we hypothesized that we would not see a decrease in engraftment in these two conditions. As shown in, there was no significant difference in engraftment between WT and GLP1RKO mice receiving syngeneic transplants. Similarly, WT and GLP1RKO mice receiving minor antigen-mismatched HSC grafts showed no difference in engraftment reflected in the peripheral blood, but to our surprise, GLP1RKO mice had more CD45cells in the bone marrow compared to WT (). We were able to differentiate between donor and recipient cells for mice receiving MHC-mismatched HSC grafts, and so we assessed donor chimerism and total live donor CD45cells in WT and GLP1RKO mice receiving allogeneic HSCT from BALB/c mice (). GLP1RKO mice had significantly decreased donor chimerism and total numbers of donor CD45cells in the peripheral blood and bone marrow compared to WT controls (). These results validated our hypothesis that GLP1RKO mice dying in the setting of MHC-mismatched allogeneic transplant were doing so because of GF. Fig. 1 Fig. 2A Fig. 2B Fig. 2C Fig. 2D + + + +

Our HSCT model included a low dose of donor T cells, and so we aimed to confirm that GVHD was not contributing to the deaths of these mice. We investigated GVHD by repeating our allogeneic HSCT model and harvested tissues on day 10 and 14 post-transplant, examining liver, lung, small intestine, and large intestine (). We selected these time points because GLP1RKO mice had begun to die by day 7, and so an underlying pathology hypothetically responsible for this death would be observable at day 10 and day 14. A board-certified veterinary pathologist blinded to the groups evaluated all tissues and found no histologic evidence of GVHD in any tissue for any group. These negative histologic findings for the GLP1RKO mice at timepoints when they were dying supported our hypothesis that GVHD was not contributing to systemic illness and mortality in this model. To obtain further evidence against the presence of GVHD, we obtained serum samples at day 7 and evaluated concentrations of a host of cytokines that are typically altered in the setting of GVHD. In line with our histology findings, GLP1RKO allogeneic mice did not show an increase in IFN-γ, IL-1β, IL-2, IL-6, or TNF-α compared to WT allogeneic mice (). To rule out morbidity and mortality from other hematologic complications, we performed complete blood counts (CBCs) with differentials on peripheral blood samples from transplanted mice on day 7 (). We found no significant difference in hemoglobin concentration, hematocrit, or platelet counts. However, GLP1RKO mice had significantly fewer total white blood cells and neutrophils and a nearly significant trend towards fewer lymphocytes than WT mice. These findings of leukopenia on CBC support our assertion that GLP1RKO mice are experiencing increased GF in the setting of MHC-mismatch HSCT compared to WT mice. Fig. 3A Fig. 3B Fig. 3C

To definitively rule out GVHD as the cause of increased mortality in GLP1RKO mice receiving MHC-mismatched HCST compared to WT, we repeated our transplant model without donor T cells. We ensured the absence of donor T cells in the HSC graft through α-CD90.2 antibody-dependent complement-mediated lysis as previously described,and we added no additional splenic T cells. WT and GLP1RKO mice receiving syngeneic T cell-depleted (TCD) HSC grafts appropriately recovered their weight, while WT and GLP1RKO allogeneic TCD groups did not. GLP1RKO mice receiving TCD HSCT experienced an increase in weight loss compared to WT counterparts that trended toward significance (), and they had statistically significantly reduced survival compared to WT after transplant (). The persistence of decreased survival in GLP1RKO mice in HSCT models without donor T cells strongly suggests that GVHD was not responsible for the differences in mortality observed in our earlier experiments. ²⁵¹ Fig. 3D Fig. 3E

Next, we aimed to determine if GF was mediated by recipient lymphocytes rejecting MHC-mismatched allogeneic grafts in the GLP1RKO mice. To assess this, we devised an antibody depletion model that would clear recipient mice of CD90cells, particularly T cells, before irradiation. We used an intra-peritoneal injection of α-CD90.2 antibody to effectively deplete circulating and splenic T cells within 4 d (), and we verified that these antibodies persist long enough to deplete donor CD90.2T cells present in the HSC graft (). With these recipient T cells depleted, we performed our transplants after lethal irradiation with syngeneic HSCs (C57BL/6J) and MHC-mismatched allogeneic HSCs (BALB/c) without additional T cells. Our use of a CD90.2 depleting antibody precluded the use of C57BL/6J or BALB/c (CD90.2) donors for T cells, as circulating antibody would deplete these as well (). We hypothesized that CD90cell depletion in the recipient mice would abrogate the GF observed in GLP1RKO mice receiving MHC-mismatched HSC grafts. After HSCT, both WT and GLP1RKO syngeneic mice recovered weight by day 18 and did not experience mortality, aligning with our previous transplant models (). Both WT and GLP1RKO mice transplanted with BALB/c HSCs also recovered weight and did not experience mortality (). This result suggested that depletion of recipient CD90cells prevented the rejection of the MHC mismatched allografts in GLP1RKO mice. We confirmed proper engraftment and a lack of rejection in these mice; no differences were observed in donor chimerism or total donor CD45cells at day 20 post-transplant in the bone marrow between WT and GLP1RKO (). + + + + + Fig. S1A and B Fig. S1C Fig. 4A Fig. 4B, -C Fig. 4B, C Fig. 4D, E

It has been reported in the literature that GLP1R signaling influences regulatory T cell (T Reg) homeostasis.We hypothesized that the difference in host T cell-mediated rejection of BM HSCs in GLP1RKO mice might be due to a deficiency in T Reg suppressive function. If this hypothesis were true, supplementation of WT T Regs in a GLP1RKO recipient would more effectively reverse graft rejection than supplementation with GLP1RKO T Regs. To test this hypothesis, we polarized T Regs from naïve CD4+ splenic T cells from WT and GLP1RKO C57BL/6J mice (). WT and GLP1RKO T cells achieved near identical T Reg polarization, assessed by CD25/FOXP3 double positivity on flow cytometry (). Host MHC-matched WT or GLP1RKO T Regs (6.0 × 10) were added to MHC-mismatched HSC + splenic T cell grafts () given to lethally irradiated GLP1RKO mice. Surprisingly, there was no benefit in prevention of weight loss or survival with the addition of either genotype of T Regs (). This result does not support the notion that GLP1R signaling-dependent differences in T Reg function are responsible for different degrees of allogeneic HSC graft rejection. ^{251 251} Fig. S2A Fig. S2B Fig. S2C Fig. S2D and E , 5