Vol. 52 No. 1/2005 57–70 QUARTERLY Review

Hepatitis C virus (HCV) infection is an important cause of chronic hepatitis, cirrhosis, hepatocellular carcinoma and liver failure worldwide. Chronic hepatitis C virus infection is treated with interferon-a (IFN-alpha), pegylated interferon-alpha (PEG-IFNalpha) alone or in combination with ribavirin; however, a significant fraction of patients either fail to respond or relapse after cessation of therapy. Efforts to identify and develop highly specific and potent HCV inhibitors have intensified recently. Each of the virally encoded replication enzymes has been a focus of studies as well as viral receptors and the host immune system. This review summarizes recent progress in the search for novel anti-HCV agents.

Hepatitis C virus (HCV) infection is a serious disease and can develop into chronic hepatitis, liver cirrhosis or hepatocellular carcinoma.After the identification of the virus in 1989, reliable diagnostic methods have led to the realisation that infection afflicts about 300 million people worldwide (Nishikawa et al., 2003).Interferons (IFNs) are the foundation of antiviral therapy.Pegylated interferons have an improved efficacy compared to standard IFN because of longer half-lives and decreased clearance but current combination treatment with an antiviral nucleoside, ribavirin, is still not sufficiently effective.
These drugs work non-specifically by inhibiting HCV replication and stimulating the immune response, but they have numerous side effects.The efforts in developing new therapeutics are hampered by difficulties with replication of HCV in cell cultures.Animal infection models are limited to chimpanzee (Lanford & Bigger, 2002) or immunodeficient mice carrying engrafted human liver cells (Mercer et al., 2001).The understanding of genomic organisation, availability of the three-dimensional structures of virally encoded enzymes and the recent development of an HCV-replicon system in Huh-7 cells have led to an advance in the development of antivirals, many of which are currently in clinical trials.

GENOMIC ORGANISATION AND POLYPROTEIN PROCESSING
Hepatitis C virus is a positive-sense single-stranded RNA virus with a genome of about 9.6 kb (Takamizawa et al., 1991).It consists of a single open-reading frame (ORF) flanked by 5¢ and 3¢ non-translated regions (NTR) (Tanaka et al., 1995).
The 5¢NTR contains the internal ribosome-entry site (IRES) which mediates the initiation of viral RNA translation.The HCV genomic RNA encodes a polyprotein of 3010-3011 amino acids which undergoes cotranslational and posttranslational proteolytic processing in the cytoplasm or in the endoplasmic reticulum of the infected cell to give rise to four structural and six non-structural (NS) proteins (Hijikata et al., 1991).Figure 1 summarises the information concerning the identification and the function of individual gene products.
The structural proteins consist of the capsid or core protein C, two envelope glycoproteins E1 and E2, and small hydrophobic polypeptide p7.The non-structural proteins are NS2, NS3, NS4A and B, and NS5A and B. NS2 has a zinc-stimulated protease activity that cleaves the NS2/NS3 junction (Love et al., 1996).NS3, a 70 kDa protein, has been subject of intensive study because it is a multifunctional molecule with a trypsin-like serine protease catalytic domain at the N-terminal 181 residues (Tomei et al., 1993), and a nucleoside triphosphatase (NTPase)/RNA helicase domain at the C-terminal two-thirds.NS4A is a small protein (8 kDa) that is an NS3 protease cofactor (Failla et al., 1994).NS4B is a hydrophobic, 27 kDa protein, and it has been suggested that it may function as an anchor to secure part of the HCV replication apparatus to the endoplasmic reticulum (ER) membrane (Lundin et al., 2003).Two cytoplasmic phosphoproteins, p56 and p58 (56 kDa and 58 kDa, respectively) are produced from the NS5A region of the HCV genome (Kaneko et al., 1994;Neddermann et al., 2004).Both proteins are phosphorylated at serine residues in the region between amino acids 2200 and 2250 and in the C-terminal region of NS5A.Although the function of NS5A in viral replication is unknown, it is possible that NS5A might be involved in the resistance of HCV to a-interferon therapy (Tan & Katze, 2001).The NS5B protein of HCV has been overexpressed both in insect cells infected with recombinant baculovirus and in E. coli (Lohmann et al., 1997).The purified enzyme exhibits a highly processive primer-independent RNA-dependent RNA polymerase (RdRp) activity capable of copying in vitro transcribed full-length genomic HCV RNA (Oh et al., 1999).

HCV REPLICATION
The infected cells are predominantly hepatocytes.Detailed information about the mo-lecular mechanism of HCV replication is not available, but it is thought that it may be similar to that of other positive-stranded RNA viruses (De Francesco, 1999).HCV infection begins with attachment, which is mediated by a specific interaction between cell surface molecules on the target cells and the viral envelope proteins.The mechanisms of virus entry remain unknown, as are the receptors for HCV entry are not known at present.The virus is probably engulfed by receptor-mediated endocytosis, and the positive-strand RNA genome is delivered to the cytoplasm.The positive-strand RNA is translated to produce a single large polyprotein which is then cleaved into structural and non-structural proteins important for replication.In order for HCV to replicate, negative-strand RNA must be synthesised using the incoming positive-stranded RNA as a template.The negative-strand replicative intermediate is then used as a template to synthesise positive-strand progeny RNA, which is packed into viral capsids.Because of the high replication rate and a lack of a proof-reading function of NS5B, the HCV genome has high genetic variability (Tan et al., 2002).A schematic model of HCV replication is shown in Fig. 2.

CURRENT THERAPIES
IFN-a is a member of the interferon family of multifunctional proteins involved in antiviral defence, cell growth regulation and immune activation.The efficacy of IFN-a monotherapy was unsatisfactory resulting in a sustained virological response (SVR) in only 14% of patients infected with HCV genotype 1 (HCV-1) (Shepherd et al., 2004).Semi-synthetic protein-polymer conjugates of interferon with polyethylene glycol have been developed (Wang et al., 2002).These conjugates protect the protein from degradation, reduce its immunogenicity, and prolong exposure to the drug by sustained absorption, restricted volume of distribution and sustained high serum concentration.Therapy with pegylated interferon is associated with significantly greater SVR compared to the non-pegylated formulation (SVR 31%) (Lake-Bakaar, 2003).
Ribavirin is a guanosine analog with minimal antiviral activity against HCV.It demonstrates, however, significant clinical synergism when administered in combination with interferon.In the trials that tested pegylated interferon plus ribavirin against non-pegylated interferon plus ribavirin the combined percentage of sustained virological response was 55%.The relative risk (RR) for remaining infected was reduced by 17% for pegylated interferon plus ribavirin compared with non-pegylated interferon plus ribavirin.Response to therapy varied according to the viral genotype.Amantadine blocks entry of influenza A virus into cells.Used in combination with ribavirin and interferon as triple therapy, it may have some benefit compared to dual or monotherapy.Recently it has been shown that amantadine blocks the viral ion channel formed by p7 protein (Griffin et al., 2003).Current treatment with pegylated interferons combined with weight-based ribavirin provides the highest sustained virological response rates.

FUTURE ANTIVIRAL THERAPIES
Several potential antiviral targets have been identified from the knowledge of the structure of the HCV genome.These include (a) the internal ribosome entry site (IRES) which could be blocked by molecules that inhibit the binding to ribosomes; (b) core protein and E1/E2 proteins, which could be attacked by antisense RNA, (c) protease or helicase susceptible to inhibition by specific inhibitors, (d) RNA-dependent RNA polymerase that could by attacked by specific inhibitors.The recently developed subgenomic HCV replicon which allows high-level replication of HCV RNAs in the human hepatoma cell line Huh-7 is widely used to unravel the principles of HCV replication and for drug screening (Lohmann et al., 1999;2001).The first subgenomic HCV replicon consists of 5¢-NTR which directs expression of neomycin phosphotransferase II from its gene (neo) cloned downstream, followed by the encephalomyocarditis virus internal ribosome entry site (IRES) which drives expression of HCV non-structural proteins NS3-NS5B from their gene (Zhu et al., 2003).The replicon-containing cells support only replication of HCV RNA encoding structural proteins, not the entire viral replication.A variety of modified replicons have been produced to improve the efficiency of colony formation, and to include different subtypes and strains of HCV (Huang & Deshpande, 2004).Replication can be measured using a variety of techniques: by counting the number of cells in which replication is established (neo selection experiments) by measuring viral RNA (Northern blotting or RNAase protection assay) or proteins (Western blotting or enzyme-linked immunosorbent assay), or by using reporters such as luciferase expressed from the replicon (Murray et al., 2003).

Inhibitors of RNA-dependent RNA polymerase
Several structural classes of NS5B RdRp inhibitors that are active in the cell-based replicon assay have been identified (Carroll et al., 2003;Gu et al., 2003;Tomei et al., 2004).Inhibitors can be classified into three categories according to their chemical structure: nucleoside analogues, pyrophosphate mimics and non-nucleoside inhibitors (NNIs).Among the nucleoside analogues, 2'-or 3¢-substituted nucleosides act as chain terminators and effectively inhibit replication of HCV subgenomic replicon.Merck Research Laboratories and Isis Pharmaceuticals Inc. described a detailed structure-activity relationship (SAR) study of 2¢-and 3¢-substituted ribonucleosides (Eldrup et al., 2004).However, only a few of the analogues have show activity in cell-based subgenomic replicon assay.Triphosphates of 2¢-C-methyladenosine (2¢-C-MeA, 1; Fig. 3) and 2¢-O-methylcytidine (2¢-O-MeC, 2; Fig. 3) demonstrated similar inhibition of NS5B in vitro (IC 50 2.5 mM and 3.5 mM), but significantly higher concentrations of 1 than 2 were detected in cells (Carroll et al., 2003).This is consistent with the greater potency of 2¢-C-methyladenosine (EC 50 = 0.3 mM) versus 2¢-O-methylcytidine (EC 50 = 21 mM) in cells.Valeant Pharmaceuticals patented adenine analogues, one of them (3, Fig. 3) in pharmacokinetic studies in mice showed a bioavailability of 100% and did not exhibit significant acute toxicity at doses of up to 160 mg/kg (Ni & Wagman, 2004).Non-chain-terminating nucleoside analogues are also feasible, especially those with mutation-inducing abilities.Recent studies suggest that once misincorporated they cause errors in viral replication (Stuyver et al., 2003), as is also suggested for the ribavirin (Maag et al., 2001).The introduction of the amino acid valine at the 3¢-position of ribose gave NM-283 (4, Fig. 3), currently in development by Idenix Pharmaceuticals Inc., which is now into clinical trials (Standring, 2003).NM-283 has low toxicity in vitro (CC 50 > 100 mM) and in vivo (no observed adverse effect level (NOAEL) = 166 mg/kg for monkeys).In early clinical studies oral dosing of NM-283 (200 mg/day for 7 days) in HCV infected patients reduced the viral titer by up to 0.49 log (Ni & Wagman, 2004).
Compounds interfering with the binding of phosphoryl groups of the nucleotide substrates at the active site have been identified.Over 200 compounds including alkyl-, phenyl-, pyrrole-and thiophene-substituted diketoacids were evaluated against HCV NS5B polymerase activity (Altamura et al., 2000).Among them, several phenyldiketoacids and 2-aryl-4,5-dihydroxy-carboxypyrimidines demonstrated low nanomolar IC 50 values (5 and 6, Fig 4 .) in the cell-based HCV subgenomic replicon assay (De Francesco, 2000).The benzo-1,2,4-thiadiazine derivative of hydroxyquinoline 4] were shown to be potent inhibitors of the synthesis of the positive and negative strand RNA by HCV RdRp with IC 50 0.17-0.14mM (7) and, respectively, 0.06-0.04mM (8) (Dhanak et al., 2002).A significant advance in the understanding of the NS5B polymerase was provided by crystallographic studies of the apoenzyme and of complexes with nucleotides or RNA template (Ago et al., 1999;Bressanelli et al., 2002).In all structures reported the fingers, palm, and thumb subdomains characteristic of all known RNA and DNA polymerases may be seen (Fig. 5).The unique feature of HCV NS5B is its encircled active site with overall globular shape instead of the typical U shape found in other polymerases (Lesburg et al., 1999).Among the numerous non-nucleotide compounds documented to have inhibitory activity, (2S)-2-[(2,4-dichloro-benzoyl)-(3-trifluoromethyl-benzyl)-amino]-3-phenyl-propionic acid has been shown to bind on the protein surface in a narrow cleft in the thumb domain (Wang et al., 2003) (9, Fig. 5).

Inhibitors of NS3/4A serine protease
The second enzyme very important for viral replication is the NS3 serine protease which is responsible for processing the HCV polyprotein into smaller functional proteins.NS3 becomes fully activated as a serine protease when bound in a heterodimeric complex with NS4A.X-ray crystallography (Love & Parge, 1996) and nuclear magnetic resonance (NMR) spectroscopy (McCoy et al., 2001) have shown that the structure consists of two domains, both composed of a b-barrel and two a-helices.The catalytic triad comprises histidine 57, aspartate 81 and serine 139 and is located between the two domains (Fig. 6).
Initial investigations revealed that the N-terminal peptide products derived from the sequences of the HCV NS3 protease substrates   are competitive inhibitors of this enzyme, so these peptides became leads for the design of peptidomimetic inhibitors (Llinas-Brunet et al., 2000).Lamarre and his team started with a weak hexapeptide enzyme inhibitor Asp-Asp-Ile-Val-Pro-Cys and K i of 79 mM (Steinkuhler et al., 1998).Next they shortened the size to a tripeptide and finally stabilized the compounds by intramolecular linking to obtain BILN-2061 (10, Fig. 7; Lamarre et al., 2003) patented by Boehringer Ingelheim Corp.In cell-free assays the K i was 0.3 nM, and in the subgenomic replicon model the compound was effective at penetrating cells and displayed an EC 50 value of 4 nM.The 50% cytotoxic concentration (CC 50 ) in Huh-7 cells was acceptable at 33 mM (Lamarre et al., 2002).BILN-2061 was not significantly active against human proteases such as leukocyte elastase and liver cathepsin B. In a phase 1B study of eight genotype 1 infected patients with advanced liver fibrosis, oral treatment with 200 mg BILN-2061 for two days was well tolerated and led to one log decrease in virus titre (Benhamou et al., 2002).Another class of inhibitors was designed by replacing the carboxylic acid moiety of the peptide inhibitors with an a-ketoacid moiety (11, Fig. 7), which improved potency, with the overall K i values between 10 pM and 67 nM (Narjes et al., 2000).
Three new series based on the novel cores, such as the bicyclic aminopyridone derivative (12, Fig. 8) IC 50 0.12 mM (Zhang et al., 2002), the pyrimidinone derivative (13, Fig. 8) IC 50 0.10 mM (Glunz et al., 2003) and the pyra-zinone derivative (14, Fig. 8) IC 50 20 nM (Zhang et al., 2003) were developed by Bristol-Myers Squibb.The successful clinical use of protease and polymerase inhibitors still has to overcome the hurdle of the emergence of resistant HCV variants.Bearing in mind the high mutation rate observed for HCV it is highly likely that viruses expressing a mutant protease or polymerase with decreased sensitivity could emerge during therapy.

Inhibitors of nucleoside triphosphatase (NTPase)/RNA helicase
The C terminal region of NS3 is a member of a large class of helicases that unwind dsRNA in a nucleoside triphosphate (NTP)-dependent fashion in a 3¢ to 5¢ direction and plays a key role in viral replication.Different threedimensional structures of the isolated NS3 helicase domain have been determined (Yao et al., 1997;Cho et al., 1998;Kim et al., 1998).HCV helicase consists of three structural domains separated by clefts, together forming Y-shaped molecule (Fig. 9).
The active site for NTP hydrolysis can be identified as the APTG 207 SGKT sequence, known as the NTP binding motif I, which is involved in binding the g-phosphate of NTP.Motif II DECH 293 is proximal to motif I and is involved in binding of the Mg 2+ -NTP substrate.His 293 is located at the bottom of the interdomain cleft and is essential for the coupling of the NTPase activity to polynucleotide binding.Domain 1 is connected to domain 2 via a flexible linker region corresponding to motif III T 322 ATPP.The second domain contains the highly conserved Q 460 RRGRTGRG-RRG sequence identified as motif VI.Studies of several helicases have evaluated the effects of mutations in this motif; however, a role of this motif has not been clearly defined.Gln 460 lies at the bottom of the cleft and is thought to interact with His 293 of the DECH 293 box found on the opposite side of the cleft on the inner face of domain 1.Arginines -461, 464, and 467 have been proposed by Yao et al., (1997) to be involved in binding single-stranded RNA in the cleft between domains 1 and 2. However, the structure published by Kim et al. (1998), where the helicase is complexed with dU 8 shows that Arg 461 points away from the cleft and is hydrogen bonded to Asp 412 and Asp 427 is not consistent with this interpretation.In the X-ray structure of the nucleic acid-bound HCV helicase (Kim et al., 1998;Fig. 9), the oligonucleotide binds in the orthogonal binding site and contacts relatively few conserved residues.The cleft is large enough to bind ssDNA (or ssRNA) but is too small to bind a double-stranded DNA (or RNA; De la Cruz et al., 1999).
Serebrov and Pyle ( 2004) used a novel combinatorial time-resolved approach to gain in-sight into the mechanism of unwinding of duplex RNA by HCV helicase.They suggest that the helicase has a leading edge and a lagging edge.The leading edge binds to a duplex region of RNA, while the lagging edge binds to the single stranded 3¢ tail.The conformational changes associated with the binding of ATP and its hydrolysis provide the energy for separation of 18 bp of the duplex.
Because the ATPase activity is coupled to strand separation, researchers are pursuing potential specific inhibitors that target the ATPase and helicase activities of the NS3 protein.Non-hydrolysable ATP analogues do not substitute for ATP in the RNA unwinding reaction.It was pointed out that using adenosine-5¢-(3-thio)-triphosphate (ATP-g-S) or b,g-methylene ATP (AMPPCP) only a low level or inhibition of the unwinding activity of the NTPase/helicase of HCV was detected (Gallinari et al., 1998).The adenosine analogue 5¢-O-(4-fluorosulphonylbenzoyl)-adenosine (FSBA) (15, Fig. 10) has been found to react irreversibly with many enzymes which use NTPs as substrates, such as kinases, ATPases, and polymerases (Colman, 1983).This compound might be considered an analogue of ATP or ADP.The structure and size of FSBA is such that it resembles ATP, and its reactive sulphonyl fluoride moiety may occupy a position similar to that of the g-phosphate in NTP.The 5¢-O-(4-fluorosulphonylbenzoyl)-esters of ribavirin and purines were synthesized and tested as inhibitors of the nucleotide triphosphatase/helicase activities of HCV and related Flaviviruses (Bretner et al., 2004a).Whereas the ATPase activity of the West Nile virus enzyme could be completely blocked by all tested compounds at 500 mM, they did not cause any measurable inhibition of the ATPase activity of HCV.When the unwinding activity of the viral enzymes was tested under standard conditions, only weak inhibition was obtained with 5¢-O-(4-fluorosulphonylbenzoyl)-inosine (FSBI, IC 50 ³ 120 mM).Preincubation of the HCV enzyme with 5¢-O-FSB esters increased the inhibitory effect.The influence on the activity of HCV polymerase was also examined, and FSBI exhibited the highest inhibitory activity with IC 50 = 80 mM.
When DCBTR and TBBT were tested in the HCV subgenomic replicon system in Huh-5-2 cells at 11 mM, reduction of viral RNA to 65% was observed (Bretner & De Clercq, unpublished results).Tests in Vero and HeLa Tat cells showed a decrease of cytotoxicity of N-alkyl derivatives as compared to the parent compound -TBBT.A series of ring-expanded heterocycles, nucleoside and nucleotide analogues (RENs) containing the imidazo-[4,5-e][1,3]diazepine ring system have been synthesized and tested as potential inhibitors of HCV and some related Flaviviridae NTPase/helicases (Zhang et al., 2003).Some of them inhibited the viral helicase activity with IC 50 values in the micromolar range.The RENs-5¢-triphosphates did not influence the unwinding reaction, but exerted an inhibitory effect on the ATPase activity of the enzymes (IC 50 0.55 mM).A large series of 2,3,5-trisubstituted-1,2,4-thiadiazol-2-ium salts were claimed to exhibit anti-NS3 helicase activity with low micromolar activity in vitro (Janetka et al., 2000), and the group with the benzimidazole ring was patented (Diana & Bailey, 1997).In vitro analysis of these compounds in the subgenomic replicon system may yield important information as to their ultimate potential.Future studies of NS3 will probably focus on obtaining the co-crystal structure (enzyme-compound) to allow lead optimization and improved specificity.

CONCLUDING REMARKS
The current standard treatment for chronic HCV infection, interferon a and ribavirin, succeeds in only 50% of patients and can cause serious side effects.The development of new antiviral agents with minimum side-effects, designed to inhibit the activities of proteins essential to HCV proliferation, is thus crucial to curbing this viral pandemic.Inhibitors in Phase II clinical development include those that target the HCV NS3 protease, NS5B polymerase and IRES.Some companies are focusing their efforts on developing ribavirin analogues, whereas others try to develop an effective and safe prophylactic vaccine.Correlations between compound potency in replicon assays with enzymatic activity, and ultimately efficacy in man, will significantly accelerate the progression of drug discovery and development, leading to novel treatments for HCV infection.
I am grateful to Professor T. Kulikowski (Institute of Biochemistry and Biophysics, PAS, Warszawa, Poland) for many helpful discussions.

Figure 5 .
Figure 5.A ribbon diagram of HCV NS5B/inhibitor (9) complex structure.The protein is colored according to the secondary structure.Asp 318 in active site coordinating Mg 2+ ions during the polymerization reaction is shown with red and blue surface.Inhibitor is shown as stick model.Figure generated with MDL Sculpt using 1NHU coordinates from PDB.

Figure
Figure 6.A ribbon diagram of HCV NS3-NS4A protein complex structure.The protein is colored according to the secondary structure, and NS4A protein is yellow.Asp 81 in the active site is shown with red and blue surface; a blue boll represents Zn 2+ .Figure generated with MDL Sculpt using 1JXP coordinates from PDB.

Figure
Figure 7.Chemical structures of HCV NS3 serine protease inhibitors.

Figure 9 .
Figure 9.A ribbon diagram of HCV NS3/dU8 complex structure.The protein is colored according to the secondary structure.Asp 290 in the DExH motif is shown with red and blue surface.The sulphate ion, shown as stick model appears to occupy the position of b-phosphate of ADP.The DNA is colored green.Figure generated with MDL Sculpt using 1A1V coordinates from PDB.