Granzyme B PET Imaging for Assessment of Disease Activity in Inflammatory Bowel Disease

Abstract

Developing a noninvasive imaging method to detect immune system activation with a high temporal resolution is key to improving inflammatory bowel disease (IBD) management. In this study, granzyme B (GZMB), typically released from cytotoxic T and natural killer cells, was targeted using PET with ⁶⁸Ga-NOTA-GZP (where GZP is β-Ala-Gly-Gly-Ile-Glu-Phe-Asp-CHO) to detect early intestinal inflammation in murine models of colitis. Methods: Bioinformatic analysis was used to assess the potential of GZMB as a biomarker for detecting IBD and predicting response to treatment. Human active and quiescent Crohn disease and ulcerative colitis tissues were stained for GZMB. We used IL-10^-/- mice treated with dextran sulfate sodium (DSS) as an IBD model, wild-type C57BL/6J mice as a control, and anti-tumor necrosis factor as therapy. We used a murine GZMB-binding peptide conjugated to a NOTA chelator (NOTA-GZP) labeled with ⁶⁸Ga as the PET tracer. PET imaging was conducted at 1, 3, and 4 wk after colitis induction to evaluate temporal changes. Results: Bioinformatic analysis showed that GZMB gene expression is significantly upregulated in human ulcerative colitis and Crohn disease compared with the noninflamed bowel by 2.98-fold and 1.92-fold, respectively; its expression is lower by 2.16-fold in treatment responders than in nonresponders. Immunofluorescence staining of human tissues demonstrated a significantly higher GZMB in patients with active than with quiescent IBD (P = 0.032).⁶⁸Ga-NOTA-GZP PET imaging showed significantly increased bowel uptake in IL-10^-/- mice with DSS-induced colitis compared with vehicle-treated IL-10^-/- mice (SUV_mean, 0.75 vs. 0.24; P < 0.001) and both vehicle- and DSS-treated wild-type mice (SUV_mean, 0.26 and 0.37; P < 0.001). In the IL-10^-/- DSS-induced colitis model, the bowel PET probe uptake decreased in response to treatment with tumor necrosis factor-α (SUV_mean, 0.32; P < 0.001). There was a 4-fold increase in colonic uptake of ⁶⁸Ga-NOTA-GZP in the colitis model compared with the control 1 wk after colitis induction. The uptake gradually decreased to approximately 2-fold by 4 wk after IBD induction; however, the inflamed bowel uptake remained significantly higher than control at all time points (week 4 SUV_mean, 0.23 vs. 0.08; P = 0.001). Conclusion: GZMB is a promising biomarker to detect active IBD and predict response to treatment. This study provides compelling evidence to translate GZMB PET for imaging IBD activity in clinical settings.

Keywords: 68Ga-NOTA-GZP; PET imaging; colitis; granzyme B; inflammatory bowel disease.

Graphical abstract

FIGURE 1.

GZMB as biomarker in IBD for detecting active disease and evaluating treatment response using bioinformatic analysis. (A) Number of significantly altered genes in active UC vs. normal tissues. (B) Upregulated genes that were in common in UC 8 datasets ranked by log₂ fold change. (C) Heat map of up- and downregulated genes in UC and CD tissues compared with normal tissue. (D) Number of altered genes in active CD vs. normal tissues. (E) Upregulated genes that were in common in CD 7 datasets ranked by log₂ fold change. (F) Gene expression of GZMB in patients before and after treatment analyzed by 1-way ANOVA followed by adjusting P value using Benjamini–Hochberg method (data are mean ± SD). (G) Top 15 genes significantly decreased after treatment at different time points (green spots represent top 5 genes, which decreased after treatment, and red represents GZMB). (H) Evaluation of GZMB gene expression in active and inactive UC using t test analysis. FDR = false-discovery rate; log₂FC = log₂ fold change.

FIGURE 2.

(A) Immunofluorescence evaluation of GZMB expression in active IBD. (B) Immunofluorescence evaluation of GZMB expression in quiescent IBD. (C) Quantitative analysis of GZMB to 4′,6-diamidino-2-phenylindole ratio in active IBD vs. quiescent IBD (data are mean ± SEM, 4/group). (D) Hematoxylin and eosin staining of corresponding active IBD samples at indicated magnification. (E) Hematoxylin and eosin staining of corresponding quiescent IBD samples at indicated magnification. Scale bars represent 100 μm. DAPI = 4′,6-diamidino-2-phenylindole.

FIGURE 3.

(A) Coronal representation of ⁶⁸Ga-NOTA-GZP PET imaging patterns alongside CT scans, encompassing both combined PET/CT and standalone PET views in IL10^−/− mice treated with vehicle, DSS, or DSS plus anti-TNF, as well as wild-type animals treated with DSS or vehicle. (B) Representative axial view of ⁶⁸Ga-NOTA-GZP PET/CT uptake pattern in IL10^−/− mice treated with DSS and wild-type mice treated with vehicle. (C) Quantitative analysis of ⁶⁸Ga-NOTA-GZP PET/CT SUV_mean in DSS-treated IL10^−/− mice compared with other groups (7–13/group). (D) ⁶⁸Ga-NOTA-GZP uptake in ex vivo PET/CT imaging experiment of extracted colonic specimens (3/group). (E) Temporal change in ⁶⁸Ga-NOTA-GZP PET/CT uptake during 4 wk after DSS induction (data are mean ± SEM, 8/group). %ID = percentage injected dose; WT = wild-type.

FIGURE 4.

Evaluation of colonic cytokine levels in different groups (M1–M4 represent each mouse’s data; red box represents cytokine levels < 250 ng/mL).