Leucine-rich α-2 glycoprotein 1 (LRG1) during inflammatory complications after allogeneic stem cell transplantation and CAR-T cell therapy

Abstract

Background: Previous data indicated that the leucine-rich α-2 glycoprotein 1 (LRG1) pathway contributes to vascular dysfunction during cancer growth. Therapeutic targeting of LRG1 normalized tumor vessel dysfunction and enhanced the efficacy of anti-cancer adoptive T cell therapy. A major clinical problem after allogeneic hematopoietic stem cell transplantation (alloHSCT) and after chimeric antigen receptor (CAR) T-cell therapy is the induction of hyperinflammatory side effects, which are typically associated with severe endothelial dysfunction.

Methods: We investigated LRG1 in preclinical models and in patient samples.

Results: In prospective studies, we found elevated LRG1 serum levels in patients with cytokine release syndrome and immune effector cell-associated neurotoxicity syndrome after CAR-T-cell therapy as well as in patients with acute graft-versus-host disease (aGVHD) after alloHSCT.In preclinical models of aGVHD, we found vasculature-associated LRG1 upregulation as well as LRG1 pathway gene upregulation. The genetic deletion of LRG1 in alloHSCT donors and in alloHSCT recipients led to reduced clinical and histological aGVHD. In line with this, LRG1 deletion led to clinically and histologically reduced disease severity in experimental inflammatory models of colitis (dextran sulfate sodium colitis) and paw edema. LRG1 deletion reduced inflammation-related vascular leakiness, endothelial cell proliferation, and migration.

Conclusions: The current data support the hypothesis that LRG1 is an attractive therapeutic target after alloHSCT and after CAR-T cell therapy for cancer because of its role in dysfunctional tumor vessels as well as in inflammatory complications.

Keywords: Cytokine release syndrome; Graft versus host disease - GVHD; Immune related adverse event - irAE; Immunotherapy; Stem cell.

Figure 1. Serum level of LRG1 in aGVHD and CRS patients. (A) LRG1 serum levels of patients with aGVHD grades II and III versus patients without aGVHD before and after alloHSCT. (B) LRG1 serum levels of patients with CRS scores I–III before and after CAR T-cell therapy. (C) LRG1 serum level of patients without CRS versus patients with CRS grades I–III on day 3 after CAR T cell infusion. (D) LRG1 serum levels of patients with ICANS scores I–III before and after CAR T-cell therapy. (E) LRG1 serum level of patients without ICANS versus patients with CRS grades I–III on day 3 after CAR T cell infusion. Number of patients: 9 patients with aGVHD grades II–III, 9 matched control patients without aGVHD; 27 patients with CRS (all grades) 7 patients without CRS (grade 0); 12 patients with ICANS (all grades), 21 patients without ICANS (grade 0). Error bars indicate mean±SD. Significance was tested with paired ((A, B, D) measurements of serum from the same patients at different time points) or unpaired Student’s t-test (A, C, E). aGVHD, acute graft-versus-host disease; alloHSCT, allogeneic hematopoietic stem cell transplantation; CRS, cytokine release syndrome; ICANS, immune effector cell-associated neurotoxicity syndrome.

Figure 2. Expression data of whole liver tissue and isolated liver sinusoidal endothelial cells at different time points after alloHSCT. (A) Experimental schema of the LP/J → C57BL/6 model, detailed description in the methods section. mRNA (B) and protein (C) expression in the liver and mRNA expression in isolated sinusoidal endothelial cells (D). mRNA expression of members of the TGFβ pathway (E, F), the proangiogenic co-receptor ALK1 (G) and the angiostatic co-receptor ALK5 (H). For qPCR and proteomics data, we used the chemotherapy-based minor mismatch model LP → C57BL/6 and harvested liver tissue from syngeneic and allogeneic transplanted animals at the indicated time points after HSCT. Further, we isolated liver sinusoidal endothelial cells on day +2 and day +15 after HSCT for qPCR analysis. Fold change refers to the relative expression compared with wild-type controls. Error bars indicate mean±SD. Significance was tested with unpaired Student’s t-test. N=2–12 samples per group. alloHSCT, allogeneic hematopoietic stem cell transplantation; BM, bone marrow; BMT, bone marrow transplantation; LSECs, liver sinusoidal endothelial cells.

Figure 3. LRG1 in blood vessels and knockout of LRG1 in aGVHD. (A) Experimental schema of the 129/SV → C57BL/6 model, detailed description in the methods section. (B) Quantification of LRG1 positive area (left) and LRG1/CD31 ratio (right) in the liver on day 15 after HSCT. (C) Representative images of increased LRG1 expression in the liver during aGVHD. (D) Quantification of LRG1 positive area (left) and LRG1/CD31 ratio (right) in the colon on day 15 after HSCT. (E) Representative images of increased LRG1 expression in the colon during aGVHD. For immunohistological staining (B–E) we used the chemotherapy-based minor mismatch model 129 → C57BL/6 and harvested tissue on day 15 after HSCT. Colon and liver sections were stained against LRG1 and CD31 and counterstained with 4’6-diamino-2-phenylindole (DAPI). (B, D) n=4–5 per group. (F) Clinical aGVHD scores of B6 WT and LRG1−/− mice used as alloHSCT recipients on day 8 after transplantation. (G) Experimental schema of the 129/SV → C57BL/6 model, detailed description in the methods section (H) Clinical GVHD scores of mice receiving either B6 WT or LRG1−/− donor cells on day 12 after transplantation. (I) Experimental schema of the B6 WT/LRG1−/− → Balb/C model, detailed description in the methods section (F) n=10 mice per group, mouse model: 129 → B6 WT/LRG1-/-, (G) n=7–8 mice per group, mouse model: LRG1−/− → Balb/c. (J) Quantification of vessel density in the liver. (K) n=4–5 mice per group, mouse model: B6 WT/LRG1−/− → BDF. Error bars indicate mean±SD, significance tested by unpaired Student’s t-test. aGVHD, acute graft-versus-host disease; BMT, bone marrow transplantation; HSCT, hematopoietic stem cell transplantation; WT, wild-type.

Figure 4. Role of LRG1 in DSS Colitis. (A) Experimental schema of the DSS-induced colitis model. LRG1−/− mice and B6 WT littermates were challenged with 2.5% DSS in their drinking water for 8 days. Mice were monitored for DAI score every second day. On day 9 mice were sacrificed, and organs were taken for detailed examinations. (B) Colitis DAI of LRG1−/− mice and WT littermates from day 2 to day 8. (C) Quantitative comparison of colon length and spleen weight of LRG1−/− and B6 WT on day 9 after DSS treatment start. (D) Representative pictures of colon and spleen from LRG1−/− and B6 WT mouse. (E) Histopathological score determined on H&E-stained colon sections of B6 WT and LRG1−/− mice. (F) Representative pictures of H&E staining of colon sections from B6 WT and LRG1−/− mice. Expression of CD11b (G) and CD3 (I) of B6 WT and LRG1−/− mice without colon inflammation and during experimental colitis. (H, J) Representative images of immunological staining against CD11b (H) and CD3 (J) with and without colitis. (K) Histological examination of CD31 expression and ZO1⁺ vessels in B6 WT and LRG1−/− mice during colitis. (L) Representative images of CD31 and ZO1 staining with and without colitis. (M) mRNA expression of the typical pathway gene TGFβ in B6 WT and LRG1−/− mice during colitis. N=12–15 per group (B–E), n=5–9 per group (G), n=7–9 per group (I), n=4 per group (K). Error bars indicate mean±SD, significance tested by unpaired Student’s t-test. DAI, Disease Activity Index; DSS, dextran sulfate sodium; WT, wildtype.

Figure 5. Influence of LRG1 on local inflammation. (A) Experimental schema of the paw edema model of local inflammation. LRG1−/− mice and B6 wild-type (WT) littermates were injected with a 1% carrageenan solution into one footpad and 0.9% saline into the other footpad. Footpad swelling was determined by measuring footpad thickness every hour (indicated by the ruler), using the footpad thickness before injection as a baseline (B) Amount of increase in paw thickness of the footpad injected with carrageenan in WT and KO mice. (C) Extent of the increase in paw thickness excluding the swelling of the control foot injected with NaCl. (D) Sections from footpad biopsies were stained for CD31 and analyzed with ImageJ. (E, F) The additional in vivo Evans blue assay provided information about the vascular integrity on local inflammation. 3 hours after carrageenan injection, mice were intravenously injected with Evans blue, and punches of footpads were taken 30 min later to determine the amount of extravasated Evans blue into the vessel surrounding tissue. n=25–27 per group (B, C), n=5 per group (D), n=11–14 per group (E, F). Error bars indicate mean±SD, significance tested by unpaired Student’s t-test.

Figure 6. Impact of LRG1 on endothelial cell behavior. (A) Freshly isolated liver sinusoidal endothelial cells were stained for the endothelial-specific markers CD31 and VCAM. (B–D) Migration rate of endothelial cells during wound closure with addition of LRG1−/− and B6 WT serum to the growth medium. n=3 runs per group for each assay. Error bars indicate mean±SD, significance tested by unpaired Student’s t-test. MCECs, murine cardiac endothelial cells; WT, wild-type.