Targeted Regression of Hepatocellular Carcinoma by Cancer-Specific RNA Replacement through MicroRNA Regulation

Abstract

Hepatocellular carcinoma (HCC) has a high fatality rate and limited therapeutic options with side effects and low efficacy. Here, we proposed a new anti-HCC approach based on cancer-specific post-transcriptional targeting. To this end, trans-splicing ribozymes from Tetrahymena group I intron were developed, which can specifically induce therapeutic gene activity through HCC-specific replacement of telomerase reverse transcriptase (TERT) RNA. To circumvent side effects due to TERT expression in regenerating liver tissue, liver-specific microRNA-regulated ribozymes were constructed by incorporating complementary binding sites for the hepatocyte-selective microRNA-122a (miR-122a), which is down-regulated in HCC. The ribozyme activity in vivo was assessed in mouse models orthotopically implanted with HCC. Systemic administration of adenovirus encoding the developed ribozymes caused efficient anti-cancer effect and the least hepatotoxicity with regulation of ribozyme expression by miR-122a in both xenografted and syngeneic orthotopic murine model of multifocal HCC. Of note, the ribozyme induced local and systemic antitumor immunity, thereby completely suppressing secondary tumor challenge in the syngeneic mouse. The cancer specific trans-splicing ribozyme system, which mediates tissue-specific microRNA-regulated RNA replacement, provides a clinically relevant, safe, and efficient strategy for HCC treatment.

Figure 1. Scheme for the TERT-targeting trans-splicing ribozyme-induced selective expression of therapeutic RNA in cancer cells through microRNA regulation.

Trans-splicing ribozymes with target sites to liver-specific miR-122a at the 3′-UTR of the 3′exon recognize TERT RNA at the targeted uridine residue by selective base-pairing through their internal guide sequence in cancer cells lacking the miRNA. The ribozymes then remove the sequence downstream of the target site and replace it with the 3′ exon exerting anti-cancer activity.

Figure 2. MiR-122a-regulated hTERT-specific cytotoxicity of Ad-PRT-122aT.

(A) Expression vectors encoding hTERT-targeting ribozymes. (B) Schematic diagram of the tetracycline-inducible system for miR-122a expression in HepG2 cells. (C) Selective cytotoxicity of the ribozyme-encoding adenoviral vectors according to miR-122a. Stable HepG2-Tet-on cells expressing miR-122a were infected with each adenoviral vector at various MOI and inoculated with 100 μM GCV (Left panel; with 100 μM tetracycline, right panel; without tetracycline). (D) Hep3B or SKLU-1 cells were infected with each adenovirus at various MOI and inoculated with 100 μM GCV. Cell viability was determined via MTS assay. Results were presented as means ± SD of triplicate experiments. (E) RNA analysis of the adenoviral-infected cells. The Hep3B cells were mock-infected (Mock) or infected with PEPCK-ribozyme adenovirus at 10 MOI. SKLU-1 cells infected with Ad-PRT-122aT were mixed with mock-infected Hep3B cells (mix). TSMs generated in the cells were amplified, yielding a DNA fragment of 187 bp. Human GAPDH RNA was amplified as an internal control. NTC denotes non-template control. Note that cropped gel images are used in this figure and the gels were run under the same experimental conditions (upper panel). RNA level of TSM was measured using qRT-PCR in Hep3B cells infected with each adenovirus and expressed as a percentage of TSM level of Ad-PRT-mut 122aT. Data are mean values ± SD of triplicate experiments (bottom panel). (F) Representative sequences of TSMs generated from Hep3B cells infected with Ad-PRT-122aT or Ad-PRT-mut 122aT in (E) were shown.

Figure 3. Anti-cancer efficacy and regulation of transgene expression of the adenovirus in xenograft model of HCC.

(A) Ten mice with multifocal HCC per group were systemically administered with 1 × 10¹¹ v.p. of each adenoviral vector, followed by GCV treatment. The HCC weights of each mice group were determined and plotted. Average tumor mass was presented with SD. Representative livers with tumor burdens of each mouse group treated with GCV were photographed (bottom panel). (B) Microscopic findings of sliced entire livers from each representative mouse in (A). The sectioned and paraffin-embedded liver tissues were stained with H&E. Microscopically, deep blue-colored nodules indicate HCC, while red-colored tissues are non-tumoral liver cells. (C) Liver enzyme levels in the virus and GCV-treated mice were represented as means ± SD. (D) RNA and genomic DNA patterns of mice injected with adenovirus. Ribozyme, transgene expression (HSVtk, TK), or TSM production was analyzed using RT-PCR. Endogenous 18S RNA was amplified as an internal control. Adenovirus genome (E4) and 18S DNA were analyzed using genomic DNA PCR. NTC denotes non-template control. Note that cropped gel images are used in this figure and the gels were run under the same experimental conditions (upper panel). TSM or TK RNA level and adenovirus genome level was measured using real-time PCR in liver or tumor tissues from mice infected with each adenovirus and expressed as a percentage of the level of Ad-PRT-mut 122aT. Data are average values ± SD (bottom panel). (E) Representative sequences of TSMs generated from tumor in the HCC mouse model infected with Ad-PRT-122aT or Ad-PRT-mut 122aT in (D) were shown.

Figure 4. mTERT-targeted HSVtk RNA expression and cytotoxicity through miR-122a control in cells.

(A) Schematic diagram of adenoviral backbone vectors. (B) Hepa 1–6 or SKLU-1 cells were infected with each adenovirus and inoculated with 100 μM GCV. Cell viability was determined via MTS assay. Results represent the means ± SD of three independent experiments. (C) RNA analysis of the adenovirus-infected cells. Hepa 1–6 cells were infected with each adenovirus at 10 MOI. SKLU-1 cells infected with Ad-mCRT-122aT were mixed with mock-infected Hepa 1–6 cells (mix). TSMs generated in the cells were amplified, yielding a DNA fragment of 177 bp. Mouse GAPDH RNA was amplified as an internal control (upper panel). RNA level of TSM was measured using qRT-PCR in Hepa 1–6 cells infected with each adenovirus and expressed as a percentage of TSM level of Ad-mCRT-mut 122aT. Data are mean values ± SD of triplicate experiments (bottom panel). (D) Representative sequences of TSMs generated from the cells infected with Ad-mCRT-122aT or Ad-mCRT-mut 122aT in (C) were shown. (E) Cytotoxicity of each adenovirus and GCV treatment in Hepa 1–6 and SKLU-1 cells. (F) TSMs of Ad-mPRT-122aT and Ad-mPRT-mut 122aT in Hepa 1-6 cells were detected by southern blotting. Note that cropped gel images are used in this figure and the gels were run under the same experimental conditions.

Figure 5. Cancer-specific regression by miR-122a-regulated ribozyme and GCV in syngeneic orthotopic model of HCC.

(A) Syngeneic orthotopic C57BL mice with multifocal HCC were systemically administered with 1 × 10¹¹ v.p. of each adenoviral vector (n = 7 each), followed by GCV treatment. Liver enzyme levels of the mice were measured and represented as means ± SD. A representative liver with tumor burdens of each group was photographed (bottom panel). (B) The HCC weights of each group in (A) were determined and plotted. Average tumor mass was presented with SD. (C) Ribozyme RNA (HSVtk, TK) and viral genomic DNA (E4) patterns of normal liver tissue (lane N) and HCC (lane T) from syngeneic mice model infected with adenovirus were analyzed using PCR. The 18S RNA and 18S genomic DNA were amplified as internal controls. Note that cropped gel images are used in this figure and the gels were run under the same experimental conditions (upper panel). TK RNA level and virus genomic DNA level was quantified using real-time PCR in liver or tumor tissues from mice infected with each adenovirus and expressed as a percentage of the level of Ad-mPRT-mut 122aT. Data are average values ± SD (bottom panel). (D) Seven mice of the allogenic and athymic BALB/c model with orthotopic and multifocal HCC for each group were systemically administered with 1 × 10¹¹ v.p. of adenoviral vectors and GCV. The HCC weights of each group were determined and plotted. Average tumor mass was presented with SD.

Figure 6. Induction of systemic anti-tumor immunity by treatment with Ad-mPRT-122aT and GCV.

(A) Immune cell infiltration during in vivo assessment of antitumor effects of Ad-mPRT-122aT. The liver tissue of syngeneic HCC mice treated with Ad-mPRT/GCV or control/GCV was stained with each antibody. Empty triangles indicate cancer cells and filled triangles represent immunohistochemically stained cells. Scale bar was shown at bottom left corner. (B) Schematic representation of challenge experiments with parental HCC after adenovirus infection and GCV treatment in vivo. Syngeneic mice (n = 5) were given intrasplenic injection of Hepa 1–6 cells and infected with 1 × 10¹¹ v.p. of virus on day 9. GCV treatment was given from days 10 to 19, and 6 × 10⁶ Hepa 1–6 cells were subcutaneously injected into the flanks of the mice on day 20. Tumors were harvested on day 30. (C) Tumor volumes of subcutaneous HCC mass of each group of mice were determined and plotted. Average tumor volume was presented with SD.