Gapmer antisense oligonucleotides cleave target RNA effectively in vivo, and is considered as promising therapeutics. Especially, gapmers modified with locked nucleic acid (LNA) shows potent knockdown activity; however, they also cause hepatotoxic side effects. For developing safe and effective gapmer drugs, a deeper understanding of the mechanisms of hepatotoxicity is required. Here, we investigated the cause of hepatotoxicity derived from LNA-modified gapmers. Chemical modification of gapmer’s gap region completely suppressed both knockdown activity and hepatotoxicity, indicating that the root cause of hepatotoxicity is related to intracellular gapmer activity. Gene silencing of hepatic ribonuclease H1 (RNaseH1), which catalyses gapmer-mediated RNA knockdown, strongly supressed hepatotoxic effects. Small interfering RNA (siRNA)-mediated knockdown of a target mRNA did not result in any hepatotoxic effects, while the gapmer targeting the same position on mRNA as does the siRNA showed acute toxicity. Microarray analysis revealed that several pre-mRNAs containing a sequence similar to the gapmer target were also knocked down. These results suggest that hepatotoxicity of LNA gapmer is caused by RNAseH1 activity, presumably because of off-target cleavage of RNAs inside nuclei.

No abstract

Triantennary N-acetyl galactosamine (GalNAc, GN3) and lipophilic ligands such as cholesterol and a-tocopherol conjugations dramatically improve the distribution and efficacy of secondgeneration antisense oligonucleotides (ASOs) in the whole liver. To characterize ligands for delivery to liver cells based on pharmacokinetics and efficacy, we used a locked nucleic acid gapmer of ASO targeting apolipoprotein B as a model compound and evaluated the amount of ASO and apolipoprotein B mRNA in the whole liver, hepatocytes, and nonparenchymal (NP) cells as well as plasma total cholesterol after administration of ASO conjugated with these ligands to mice. Compared with unconjugated ASO, GN3 conjugation increased the amount (7-fold) and efficacy (more than 10-fold) of ASO in hepatocytes only and showed higher efficacy than the increased rate of the amount of ASO. On the other hand, lipophilic ligand conjugations led to increased delivery (3-to 5-fold) and efficacy (5-fold) of ASO to both hepatocytes and NP cells. GN3 and lipophilic ligand conjugations increased the area under the curve of ASOs and the pharmacodynamic duration but did not change the half-life in hepatocytes and NP cells compared with unconjugated ASO. In the liver, the phosphodiester bond between ASO and these ligands was promptly cleaved to liberate unconjugated ASO. These ligand conjugations reduced plasma total cholesterol compared with unconjugated ASO, although these ASOs were well tolerated with no elevation in plasma transaminases. These findings could facilitate ligand selection tailored to liver cells expressed in disease-related genes and could contribute to the discovery and development of RNA interference-based therapy.

γ-Glutamyl transpeptidase (GGT) catalyzing the cleavage of γ-glutamyl bond of glutathione and its S-conjugates is involved in a number of physiological and pathological processes through glutathione homeostasis. Defining its Cys-Gly binding site is extremely important not only in defining the physiological function of GGT, but also in designing specific and effective inhibitors for pharmaceutical purposes. Here we report the synthesis and evaluation of a series of glutathione-analogous peptidyl phosphorus esters as mechanism-based inhibitors of human and Escherichia coli GGTs to probe the structural and stereochemical preferences in the Cys-Gly binding site. Both enzymes were inhibited strongly and irreversibly by the peptidyl phosphorus esters with a good leaving group (phenoxide). Human GGT was highly selective for l-aliphatic amino acid such as l-2-aminobutyrate (l-Cys mimic) at the Cys binding site, whereas E. coli GGT significantly preferred l-Phe mimic at this site. The C-terminal Gly and a l-amino acid analogue at the Cys binding site were necessary for inhibition, suggesting that human GGT was highly selective for glutathione (γ-Glu-l-Cys-Gly), whereas E. coli GGT are not selective for glutathione, but still retained the dipeptide (l-AA-Gly) binding site. The diastereoisomers with respect to the chiral phosphorus were separated. Both GGTs were inactivated by only one of the stereoisomers with the same stereochemistry at phosphorus. The strict recognition of phosphorus stereochemistry gave insights into the stereochemical course of the catalyzed reaction. Ion-spray mass analysis of the inhibited E. coli GGT confirmed the formation of a 1:1 covalent adduct with the catalytic subunit (small subunit) with concomitant loss of phenoxide, leaving the peptidyl moiety that presumably occupies the Cys-Gly binding site. The peptidyl phosphonate inhibitors are highly useful as a ligand for X-ray structural analysis of GGT for defining hitherto unidentified Cys-Gly binding site to design specific inhibitors.