CCl4 -induced liver fibrosis negatively affects LV but not AAV vector transduction.
To induce liver fibrosis, we administered WT mice with CCl4 for 6 weeks. During CCl4 treatment, we measured serum levels of alanine transaminase (ALT) and aspartate transaminase (AST) and observed an increase in both liver enzymes over time (ALT > 3,000 U/L; AST > 2,000 U/L), thus indicating the occurrence of liver damage (Supplementary Fig. 1A, B). In CCl4-treated mice, we observed moderate peri-central fibrosis with incomplete and complete fibrous septa, occasionally bridging between central veins, and mild inflammatory cell infiltrate in the liver capsule (Fig. 1A, Table 1). No major alterations were observed in untreated controls (Table 1). One week after the last CCl4 injection, we administered LV (vesicular stomatitis virus glycoprotein, VSV.G-pseudotyped) or AAV vectors (serotype 8) to both CCl4-treated and untreated control mice (n = 8 mice per group; Fig. 1B). These vectors express either a secreted reporter transgene (human FIX, hFIX) or a fluorescent reporter transgene (mCherry) and are referred to hereafter as LV.hFIX, LV.mCherry, AAV.hFIX and AAV.mCherry (Fig. 1B). The transgenes are expressed under the control of previously described hepatocyte-specific expression cassettes9,21. We selected LV and AAV vector doses expected to achieve 5–20% of liver transduction, based on previous experience9,22. We longitudinally monitored hFIX amounts in the plasma of treated mice and collected livers at the end of the experiment to determine mCherry expression and vector copies per diploid genome (vector genome copy number, VCN) 4 weeks after vector administration (Fig. 1B). In LV-treated animals, we observed between 3- and 4-fold lower hFIX output in CCl4-treated mice compared to controls (Fig. 1C), with a statistically significant difference, paralleled by a reduction of LV VCN in the liver (Supplementary Fig. 1C). We also observed a significantly 2-fold lower mCherry-positive liver area in CCl4-treated mice compared to controls (Fig. 1D, Supplementary Fig. 1D), without detectable difference in LV VCN in the liver (Supplementary Fig. 1E). In AAV-treated animals, we analyzed separately males and females since the transduction efficiency is known to be different between the two sexes23. We did not detect differences in circulating hFIX (Fig. 1E), mCherry-positive liver area (Fig. 1F, Supplementary 1F) or AAV VCN in the liver when comparing mice treated or not with CCl4 (Supplementary Fig. 1G, H). These data show that CCl4-induced peri-central fibrosis caused a reduction in the efficiency of gene transfer into hepatocytes by LV but not AAV vectors.
DDC-induced biliary fibrosis negatively affects both LV and AAV vector transduction.
We next fed WT mice with a DDC-supplemented diet for 5 weeks, to induce biliary liver fibrosis. During administration of the DDC diet, we observed the expected elevation of serum ALT and AST (> 1,000 U/L), indicating liver damage (Supplementary Fig. 2A, B). In DDC-fed mice, we detected cholestasis, marked peri-portal fibrosis, and complete septa bridging between portal areas, forming pseudo-nodules, sinusoidal lining, and mild to moderate peri-ductular inflammatory cell infiltrate (Fig. 2A, Table 2). We did not detect alterations in untreated control mice (Table 2). One week after the end of DDC treatment, we administered LV.hFIX, LV.mCherry, AAV.hFIX, or AAV.mCherry to DDC-treated and untreated control mice (n = 8 mice per group; Fig. 2B). We longitudinally monitored hFIX amounts in the plasma of treated mice and collected livers 4 weeks after vector administration to determine mCherry expression and VCN (Fig. 2B). To reduce the number of experimental animals, we concomitantly delivered AAV.mCherry to CCl4 or DDC-treated mice, using the same control group for both experiments (as shown in Fig.s 1F and Fig. 2F). In LV-treated animals, we observed a remarkable 10-fold reduction in hFIX output at the last time of analysis (Fig. 2C) and a statistically significant 4.5-fold reduction in the mCherry-positive liver area, when comparing DDC-treated mice with controls (Fig. 2D, Supplementary Fig. 2C). We detected similar or lower LV VCN in the liver of DDC-fed mice than in controls (Supplementary Fig. 2D, E). Concerning AAV-vector transduction, we also observed lower circulating hFIX (3.7-fold, Fig. 2E) and mCherry-positive liver area (1.2-fold; Fig. 2F, Supplementary Fig. 2F) in DDC-pretreated mice compared to controls. To compare fibrotic and non-fibrotic mice, independently of the sex, we normalized each value with the mean of the respective control group (Fig. 2G, H) and showed statistically significant differences. In both AAV.hFIX- and AAV.mCherry-injected mice, we confirmed a reduction of liver transduction, as shown by lower VCN in DDC-fed mice of both sexes (Supplementary Fig. 2G-J). These data indicate that the presence of DDC-induced peri-portal fibrosis considerably reduced the efficiency of hepatocyte gene transfer by both LV and AAV vectors.
Milder fibrosis, induced by CCl 4 or DDC, reduces LV and AAV vector transduction efficiency.
To investigate LV- and AAV-vector-mediated gene transfer in the presence of milder fibrosis, we shortened the time of exposure to CCl4 or DDC with the aim to reduce the extent of tissue damage. We administered WT mice with the toxic compounds for 3 weeks. During treatments, we confirmed the induction of liver damage, by showing serum ALT and AST elevation (> 1,500 U/L for DDC and > 3,000 U/L for CCl4; Supplementary Fig. 3A, B). We then analyzed the state of liver fibrosis at the end of the toxic treatments by histopathology. We observed mild peri-central fibrosis in the CCl4 model and moderate peri-portal fibrosis induced by the DDC diet (Fig. 3A, B and Tables 1, 2), displaying similar inflammatory cell infiltrate as described above. One week after the last CCl4 administration or interruption of the DDC diet, we injected LV.mCherry into CCl4-treated, DDC-treated, or control animals (n = 8 mice per group; Fig. 3C). On the other hand, we administered AAV.mCherry only to mice treated with DDC (n = 8 mice per group; Fig. 3C), since we did not observe a reduction of AAV-vector transduction in the presence of CCl4-induced liver fibrosis (see Fig. 1E, F). We observed significant 1.9-fold and 3.2-fold reduction of mCherry-positive liver area in CCl4 + LV and DDC + LV treated mice, respectively (Fig. 3D, Supplementary Fig. 3C) paralleled by a reduction of liver VCN (Supplementary Fig. 3D). Concerning AAV-treated mice, we found a significant 2-fold lower mCherry-positive liver area in DDC-pretreated mice compared to controls of both sexes (Fig. 3E, F, Supplementary Fig. 3E). Consistently with the reduced transgene expression, we found a 2-fold reduction in AAV VCN in the liver, as well (Supplementary Fig. 3F, G). These data show that the milder hepatic damage induced by shorter CCl4 or DDC treatments negatively influenced both LV and AAV vector transduction, as the more severe damage induced in the experimental settings mentioned above (see Fig. 1, 2).
Liver fibrosis due to Abcb11 deficiency slightly reduces LV and AAV vector transduction
We next assessed the impact of liver fibrosis on viral vector transduction in genetic models. We first characterized the state of the liver of 6-month-old Abcb11−/− mice or WT age-matched controls by histopathology. In Abcb11−/− we observed ductular proliferation, mild peri-portal fibrosis with incomplete septa between portal regions, and minimal inflammatory cell infiltrate (Fig. 4A, B, Table 3). None of these alterations were present in control mice (Table 3). We administered Abcb11−/− or WT control mice with LV.mCherry (n = 11 mice per group) or AAV.mCherry (n = 5 mice per group) and collected livers 2 weeks post vectors injection. In LV treated mice, we observed a 1.9-fold decrease in the mCherry-positive liver area in Abcb11−/− compared to controls (Fig. 4C, Supplementary Fig. 4A), paralleled by a similar LV VCN in the liver among the two groups (Supplementary Fig. 4B). In the livers of mice treated with AAV.mCherry, we detected a 1.3-fold reduction in mCherry-positive tissue area in Abcb11−/− compared to WT mice (Fig. 4D, Supplementary Fig. 4C), accompanied by a decrease in AAV VCN in the liver (Supplementary Fig. 4D). These data highlight that the mild peri-portal fibrosis present in Abcb11−/− mice of this age has only a minor negative impact on both LV and AAV-vector mediated gene transfer.
LV and AAV vector transduction is not significantly impaired in Agl−/− mice
We next took advantage of Agl−/− mice, a model of GSDIII previously shown to present with a minimal fibrotic state when adults19. We performed histopathology analysis to better characterize the liver fibrosis of 9-month-old Agl−/− mice (n = 3) or WT age-matched controls (n = 2). We detected a minimal peri-portal fibrosis with mostly incomplete bridging septa and pseudo-nodules, while we did not detect alterations in control mice (Fig. 5A). Of note, the fibrosis in this model appears to be less prominent than the one present in Abcb11−/− mice or in the chemical models shown above (Fig. 1–3). We injected 9-month-old Agl−/− or age-matched control mice (n = 5 per group) with LV.mCherry or AAV.mCherry and we collected livers 2 weeks post vectors administration. In both LV and AAV treated animals, we did not observe significant differences in mCherry-positive liver area (Fig. 5B, C, Supplementary A, C) or liver VCN (Supplementary Fig. 5B, D). These data suggest that the minimal peri-portal liver fibrosis present in Agl−/− mice does not substantially impact gene transfer to hepatocytes.