Molecular imaging of liver inflammation using an anti-VCAM-1 nanobody

Abstract

To date, a biopsy is mandatory to evaluate parenchymal inflammation in the liver. Here, we evaluated whether molecular imaging of vascular cell adhesion molecule-1 (VCAM-1) could be used as an alternative non-invasive tool to detect liver inflammation in the setting of chronic liver disease. To do so, we radiolabeled anti-VCAM-1 nanobody (^99mTc-cAbVCAM1-5) and used single-photon emission computed tomography (SPECT) to quantify liver uptake in preclinical models of non-alcoholic fatty liver disease (NAFLD) with various degree of liver inflammation: wild-type mice fed a normal or high-fat diet (HFD), FOZ fed a HFD and C57BL6/J fed a choline-deficient or -supplemented HFD. ^99mTc-cAbVCAM1-5 uptake strongly correlates with liver histological inflammatory score and with molecular inflammatory markers. The diagnostic power to detect any degree of liver inflammation is excellent (AUROC 0.85-0.99). These data build the rationale to investigate ^99mTc-cAbVCAM1-5 imaging to detect liver inflammation in patients with NAFLD, a largely unmet medical need.

Fig. 1. Specificity of ^99mTc-cAbVCAM1-5 SPECT-CT imaging.

a Body weight (n = 10/group/time point, except V-MCD group (n = 9); two-way ANOVA), b liver weight (n = 10/group, except V-MCD group (n = 9); one-way ANOVA), c plasma ALT activity (STD-fed (n = 10) and MCD-fed (n = 9) mice; unpaired Student’s t-test), d liver MCP-1 mRNA (V-STD (n = 4) and V-MCD (n = 5) group; unpaired Student’s t-test) and e liver VCAM-1 mRNA (V-STD (n = 4) and V-MCD (n = 5) group; unpaired Student’s t-test) in STD-fed mice injected with ^99mTc-cAbVCAM1-5 (V-STD), MCD-fed mice injected with control nanobody (C-MCD) and MCD-fed mice injected with ^99mTc-cAbVCAM1-5 (V-MCD). f Representative SPECT-CT liver imaging with ^99mTc-cAbVCAM1-5 or ^99mTc-Control at baseline, 4 and 8 weeks; transversal sections and g SPECT quantification in SUV (C-MCD & V-STD mice (n = 10), and V-MCD mice (n = 9); two-way ANOVA). All data are represented as mean ± SD and significant p values (corrected for multiple testing using Tukey’s post hoc when more than two groups were compared) are given above the bars. Source data are provided as a Source Data file.

Fig. 2. Liver uptake of ^99mTc-cAbVCAM1-5 upon steatohepatitis reversal.

a Representative Hematoxylin & Eosin (H&E) and Oil red-O (ORO) liver staining (x200) in all groups. Liver mRNA expression of b MCP-1, c F4/80, and d VCAM-1 in the three groups of mice (one-way ANOVA, n = 6 for V-STD and n = 5 for V-MCD & V-Reverse groups, corrected for multiple comparisons using Tukey’s post hoc). e Representative SPECT-CT liver imaging with ^99mTc-cAbVCAM1-5 at 1 and 4 weeks, transversal sections. f SPECT quantification in SUV (two-way ANOVA, n = 9 for V-STD & V-MCD mice and n = 8 for V-Reverse, corrected for multiple comparisons using Šídák’s test). All data are represented as mean ± SD and significant p values are given above the bars. Source data are provided as a Source Data file.

Fig. 3. Characteristics of the NAFLD models.

a Body weight, b liver weight, and histological c steatosis score, d inflammatory score, e ballooning score, and f total NAS score in WT mice fed a normal diet (Ctl, n = 8) or a high fat diet (WH, n = 6), in FOZ mice fed a high fat diet (FH, n = 6) and in C57BL6/J mice fed a fat-rich choline supplemented (CSH, n = 6) or deficient (CDH, n = 6) diet (one-way ANOVA and Welch’s t-test). g Representative liver histology in all groups (thick arrow = inflammatory loci, thin arrow = ballooning). All data are represented as mean ± SD and significant p values (corrected for multiple testing using Tukey’s post-hoc when more than 2 groups were compared) are given above the bars. Source data are provided as a Source Data file.

Fig. 4. VCAM1 expression correlates with inflammatory markers.

a F4/80, b MCP-1, and c VCAM-1 mRNA fold change (one-way ANOVA and Student’s t-test, n = 7 for Ctl and n = 6 for other groups). Spearman’s correlations between VCAM-1 mRNA level (fold change compared to housekeeping gene RPL19) and d F4/80 fold change, e histological inflammatory score, and f CLS, respectively. g VCAM-1 protein content per mg of tissue (one-way ANOVA and Student’s t-test, n = 8 for Ctl and n = 6 for other groups). h Total ng of VCAM-1 in the liver (one-way ANOVA and Student’s t-test, n = 8 for Ctl and n = 6 for other groups). All data are represented as mean ± SD and significant p values (corrected for multiple testing using Tukey’s post-hoc when more than 2 groups were compared) are given above the bars. Source data are provided as a Source Data file.

Fig. 5. ^99mTc-cAbVCAM1-5 liver uptake correlates with VCAM-1 expression.

SUV Liver measured in all groups a in vivo by SPECT or b ex vivo (one-way ANOVA and Student’s t-test, n = 8 for Ctl and n = 6 for other groups). c Representative SPECT transversal images. SUV_Lean Liver d in vivo and e ex vivo, and total liver uptake (%ID) f in vivo and g ex vivo of ^99mTc-cAbVCAM1-5 evaluated with SPECT (one-way ANOVA and Student’s t-test, n = 8 for Ctl and n = 6 for other groups). All data are represented as mean ± SD and significant p values (corrected for multiple testing using Tukey’s post-hoc when more than 2 groups were compared) are given above the bars. Source data are provided as a Source Data file.

Fig. 6. ^99mTc-cAbVCAM1-5 uptake robustly reflects on liver inflammation in a fatty liver.

ROC curves for detection of liver inflammation a in the entire cohort (Normal + NAFLD) using SUV_LeanLiver_in-vivo values, or b using %ID Liver_in-vivo values, and c only in animals with liver steatosis (NAFLD only) using SUV_LeanLiver_in-vivo values, or d using %ID Liver_in-vivo values. AUC: area under the curve calculated using Wilson/Brown method, 95% confidence interval (CI) and p values are provided on each panel. Source data are provided as a Source Data file.