Cis-4-[18F]fluoro-L-proline PET/CT molecular imaging quantifying liver collagenogenesis: No existing fibrotic deposition in experimental advanced-stage alcoholic liver fibrosis.

Background and purpose: Heavy alcohol drinking-induced alcoholic fatty liver, steatohepatitis, and early-stage alcoholic liver fibrosis may progress to advanced-stage alcoholic liver fibrosis (AALF)/cirrhosis. The lack of non-invasive imaging techniques for the diagnosising collagenogenesis in activated hepatic stellate cells (HSCs) can lead to incurable liver fibrosis at the early reversible stage. Proline has been known as the most abundant amino acid of collagen type 1 synthesized by activated HSC with the transportation of proline transporter. cis-4-[¹⁸F]fluoro-L-proline ([¹⁸F]proline) was reported as a useful tool to quantify collagenogenesis in experimental alcoholic steatohepatitis. This study aims to use [¹⁸F]proline micro PET as non-invasive imaging to quantify liver collagenogenesis in HSC of experimental AALF.

Methods: AALF model was set up by a modified Lieber-DeCarli liquid ethanol diet for 12 weeks along with intraperitoneal injection (IP) of CCl ₄ (0.5 ml/kg) between the 5th and 12th weeks. Controls were fed an isocaloric liquid diet and IP. PBS. In vitro [³H]proline uptake by HSCs isolated from livers was quantified using a liquid scintillation counter. Collagen type 1 production in HSCs culture medium was assayed by ELISA. Ex vivo liver collagen type 1 and proline transporter protein were compared between AALF rats (n = 8) and mice (n = 8). [³H]Proline uptake specificity in ex vivo liver tissues was tested using unlabeled proline and transporter inhibitor benztropine at different doses. Liver H&E, trichrome stain, and blood biochemistry were tested in rats and mice. In vivo, at varying times after instillation, dynamic and static [¹⁸F]proline micro PET/CT were done to quantify tracer uptake in AALF mice (n = 3). Correlation among liver collagen, liver SUVmax, normalized liver-to-brain ratio, normalized liver-to-thigh ratio, and fluoro-proline-induced collagen levels in ex vivo liver tissues were analyzed.

Results: In vitro HSCs study showed significant higher [³H]proline uptake (23007.9 ± 5089.2 vs. 1075.4 ± 119.3 CPM/mg, p < 0.001) in HSCs isolated from AALF rats than controls and so was collagen type 1 production (24.3 ± 5.8 vs. 3.0 ± 0.62 mg/ml, p < 0.001) in HSCs culture medium. Highly positive correlation between [³H]proline uptake and collagen type 1 by HSCs of AALF rats was found (r value = 0.92, p < 0.01). Ex vivo liver tissue study showed no significant difference in collagen type 1 levels between AALF rats (14.83 ± 5.35 mg/g) and AALF mice (12.91 ± 3.62 mg/g, p > 0.05), so was proline transporter expression between AALF rats (7.76 ± 1.92-fold) and AALF mice (6.80 ± 0.97-fold). Unlabeled fluoro-proline induced generation of liver tissue collagen type 1 and [³H]proline uptake were specifically blocked by transporter inhibitor. In vivo [¹⁸F]proline micro PET/CT imaging showed higher SUVmax in liver (4.90 ± 0.91 vs. 1.63 ± 0.38, p < 0.01), higher normalized liver/brain ratio (12.54 ± 0.72 vs. 2.33 ± 0.41, p < 0.01), and higher normalized liver/thigh ratio (6.03 ± 0.78 vs. 1.09 ± 0.09, p < 0.01) in AALF mice than controls, which are all positively correlated with fluoro-proline-induced levels of collagen in liver tissue (r value ≥ 0.93, p < 0.01) in AALF mice, but not correlated with existing liver collagen. Liver histology showed increased collagen in the liver of AALF mice. Blood serum ALT and AST levels were remarkably higher in AALF mice than in controls, but there is no significant difference in blood fibrotic parameters HA, A2M, TGFβ1, and MMP1.

Conclusions: [¹⁸F]proline micro PET/CT might be useful to visualize collagenogenesis in activated HSC of experimental AALF but fails to quantify existing liver collagen in AALF mice. [¹⁸F]proline has the potential sensitivity to assess the activity and severity of liver fibrosis.