A recent article1 has questioned the clinical benefits from chronic hepatitis C treatment with direct-acting antivirals (DAAs). In particular, the analysis of Cochrane review of clinical data (138 randomized clinical trials with a total of 25 232 participants)2 yielded one major conclusion – there was no sufficient evidence to judge that DAAs reduce the long term risks of death and hepatocellular carcinoma associated with chronic hepatitis C1.
This, however, was not surprising because large phase IV trials, often called “large simple trials”, are required to make such conclusions. Post-marketing observational trials are made to identify the discrepancies between the observations made during randomized clinical trial and those made in actual clinical practice. The follow-up in these trials is limited to simple and measurable clinical outcomes like the incidence of death and hepatocellular carcinoma. The results of one such large simple study3, in the French ANRS CO22 Hepather cohort of patients, have very recently been published. During this study, 7 344 patients with chronic hepatitis C were treated with DAAs, whereas 2 551 patients did not receive any treatment until the last follow-up visit (~3 years on average)3. It was found that patients treated with DAAs had reduced risk for hepatocellular carcinoma and mortality3. The overall conclusion of the study was that DAA treatment should be considered in all patients affected with hepatitis C.
One of the limitations of Hepather cohort study3 is short follow-up period and this precludes addressing some of the issues raised by the Cochrane review analysis1. Moreover, it was stated in the Hepather cohort publication that hepatitis C virus (HCV) ribonucleic acid (RNA) assessments were included up to 24 weeks after the end of DAA treatment. The results of these assessments, however, are not included in the publication. This brings us to another important issue highlighted in Cochrane review analysis. The usage of word “cure” — it is known that some patients after clearing the HCV from their blood end up having the same virus years later strongly suggesting that the virus was not completely removed from the body1,4.
The way words are used shapes the way we think. The word “cure” has a powerful meaning, especially to those being affected by a disease. “Cure” typically means restoration of health. When the word “cure” is combined with the term “antiviral therapy”, it is natural to think that as a result of such therapy the virus will be completely removed from a person affected by a disease.
How do pharmaceutical companies that sell DAAs, drugs used to treat hepatitis C, use the word “cure”? They provide a definition – cure means that HCV is not detected in the patient blood 12 weeks after treatment ends, known as sustained virologic response (SVR12).
What does SVR12 mean? It means that the genetic material of HCV is below the detectable limit of a procedure used to detect viral RNA in the blood 12 weeks after completion of DAA therapy. If the SVR12 is not achieved and the virus is still detected in the blood, then it is considered a “relapse”. If HCV RNA is detected in the blood after SVR12, then it is considered “late relapse” or “reinfection”. Thus, according to pharmaceutical companies, the hepatitis C cure means that a person affected by this disease achieved SVR12 and HCV RNA is not detected in his/her blood 12 weeks after DAA therapy discontinuation.
Achieving SVR12, however, does not mean that HCV has been completely removed from a person affected by hepatitis C1,5. Why? …The short answer is — because HCV multiplies primarily in the human liver. The longer answer requires diving into the details of HCV multiplication.
HCV multiplies in human cells called hepatocytes, the primary functional and structural constituents of liver accounting for ~80% or this organ’s volume. HCV multiplication in these liver cells typically leads to a continuous presence of 106 – 107 HCV infectious virus particles (virions) in each milliliter of HCV-infected person’s blood6. Every HCV virion carries a single viral genome in the form of RNA surrounded by a protein-lipid coat. During HCV multiplication, numerous HCV RNA genomes and viral protein-lipid coats are produced. HCV RNA genomes and protein-lipid coats are assembled into infectious virions and released into the bloodstream.
The production of HCV RNA genomes is called viral RNA replication. Viral RNA replication is performed by a replicase, a multi-component viral molecular machine that first utilizes viral RNA genome to synthesize complementary viral RNA, which subsequently serves as a template for the overproduction of viral genomes. HCV replicase consists of at least 5 components represented by viral non-structural proteins: NS3, NS4A, NS4B, NS5A, and NS5B encoded by the viral genome. These components have many additional individual functions. One such function is the formation of hollow spherical compartments derived from internal lipid membranes of the liver cell. These spherical compartments are called replicase complexes, because they contain HCV replicases and viral RNAs, and serve as protective sites for active viral RNA replication.
The “cure” in its true meaning — complete hepatitis C virus removal — can be achieved only when the following is true years after DAA treatment. First, the blood of the person is free of the virus. Second, the liver of the person is free from the virus. Third, the liver cells are free from the HCV replicase complexes. It has been suggested7 that repeat HCV RNA blood testing beyond SVR12 is required to determine whether the patient is cured.
At large, there are two types of HCV replicase complexes: (i) those that were present before DAA treatment; and (ii) those that appeared after starting DAA treatment. It turns out that various DAAs have different effect on the two types of HCV replicase complexes. Some DAAs prevent only the formation of newly formed HCV replicase complexes, whereas others penetrate into and disrupt the activity of both preformed and newly formed replicase complexes. Some DAAs blunt the virus production, whereas others just reduce it. In this solution, we evaluate the mechanism of DAAs action to understand how to maximize the probability of true cure.
As mentioned above, the components of HCV replicase have additional functions and effects. Some replicase components have activities that might preclude the effect of certain DAAs. Some replicase components’ functions have never been impeded (inhibited) to remove (cure) the replicating HCV RNA from hepatocytes in scientific laboratories. Finally, some replicase components have conflicting effects on the antiviral defense system of the cell, termed innate immunity. In this solution, we analyze such effects because of their paramount significance for choosing the best possible treatment option for hepatitis C.
Pharmaceutical companies benchmark their DAA combinations against each other primarily based on the following criteria: (i) the percentage of persons affected by hepatitis C who achieve SVR12 as a result of DAA treatment; (ii) the HCV genotypes that can be treated using DAAs; and (iii) the duration of DAA treatment. It becomes increasingly clear, that such benchmarking is oversimplified and cannot determine the best treatment option.
In the Hepather cohort, approximately 90% of patients with chronic hepatitis C were treated with DAA combination containing HCV polymerase inhibitor, primarily sofosbuvir. In this solution, we bring 6 arguments strongly suggesting that inhibiting HCV polymerase might not be the best possible option to treat hepatitis C.
HCV replicon cells – the major tool for direct-acting antivirals (DAAs) discovery
HCV multiplies in human liver cells called hepatocytes. To search for antiviral drugs, researchers and pharmaceutical companies needed a surrogate system supporting hepatitis C virus RNA replication in human liver cells. Obtaining such surrogate system took more than 10 years after the isolation of the first full-length RNA genome of hepatitis C virus8, previously known for over a decade as non-A, non-B hepatitis (NANBH) infectious agent.
A German group of researchers created a shortened modified version of hepatitis C virus RNA, a subgenomic RNA replicon, capable of replication but incapable of virion formation in cells derived from liver cancer cells9. These Huh-7 cells, derived from a hepatoma tissue of a Japanese patient with hepatocellular carcinoma10, supported replication of subgenomic HCV replicons9.
Subgenomic HCV RNA replicons9 lacked genes encoding the components of viral protein coat and had an additional neomycin phosphotransferase gene that conferred the resistance to an antibiotic called G418, an analogue of neomycin. To get subgenomic HCV replicons inside the Huh-7 cells a short electrical pulse was used, which created temporary pores in the cells’ lipid membranes. Only cells which contained high amounts of subgenomic HCV RNA produced enough neomycin phosphotransferase protein required to inactivate the G418 antibiotic, whereas other cells that either had no replicon or supported low level HCV replication died due to antibiotic action. Thus, G418 antibiotic treatment allowed the selection of cells supporting subgenomic HCV RNA high-level replication.
In particular, after delivering the electric pulse to a mixture containing Huh-7 cells and subgenomic HCV RNA replicons, the Huh-7 cells were seeded onto plastic cell culture dish in a special liquid growth medium. Cells attached to plastic within several hours and covered the surface of the plastic cell culture dish. After 24 hours, the growth medium was replaced with a fresh one containing G418 antibiotic and the same procedure, termed G418 antibiotic selection, was repeated every 2–3 days for a period of 3–5 weeks. During this selection, the majority of cells died, whereas single cells supporting high level replication of HCV replicons grew and formed colonies, visible as small “islands” on the culture dish with clean cell-less plastic surrounding the “islands”. By isolating these colonies and growing them further the German researchers established the first subgenomic HCV replicon cell lines9.
These subgenomic HCV replicon cell lines were the long-sought surrogate system for HCV replication in human liver cells. Using the same approach, subgenomic HCV replicon cell lines for nearly all HCV genotypes were created and used as the major tool for the discovery of all currently available DAAs.
Colony formation assay – using the tool for DAAs discovery
Colony formation assay was the primary experimental procedure extensively and successfully used for the discovery and characterization of almost all current DAAs. During this procedure, HCV replicon cell lines were grown in the presence of increasing amounts of compounds and constant amounts of G418 antibiotic to identify the DAAs we know today.
If increasing amounts of a compound resulted in increased death of HCV replicon cells and this was not due to compound toxicity, then a potential DAA candidate was identified. Subsequently, the effect of the identified compound on the replication of other related and unrelated viruses was evaluated to determine the specificity of the compound’s effect.
During colony formation assay, increasing amounts of DAA are added to HCV replicon cells growing on plastic cell culture dish in the presence of G418 antibiotic. By targeting any of the NS3/4A, NS5A, or NS5B components of the viral replicase, DAA inhibits the replication of subgenomic HCV replicon RNA. This leads to a decrease in the amount of subgenomic HCV replicon RNA present in the replicon cell. Decreased amount of subgenomic HCV replicon RNA translates into reduced amount of neomycin phosphotransferase protein, which inactivates the G418 antibiotic. Inability to inactivate all the G418 molecules results in stopping of the cell’s protein synthesis machinery, which leads to cell death.
The result of colony formation assay seen on the cell culture dish is similar to the result of establishing the subgenomic HCV replicon cell lines described in the previous section. In particular, cell colonies formed from single cells are visible as “islands” on the plastic. These colonies support high-level replication of subgenomic HCV replicon. There are now, however, at least two types of cell colonies on the plastic. First, there are colonies harboring subgenomic HCV replicon identical to the one initially present (parental) in the replicon cells. Second, there are colonies with subgenomic HCV replicons, which acquired mutations conferring the resistance to the DAA used. In any case, the presence of colonies in colony formation assay strongly suggests that the amount of DAA used is not sufficient for reducing the subgenomic HCV replicon level. If raising the amounts of a particular DAA does not result in the death of all replicon cells in colony formation assay, then the efficiency of such DAA is under question.
Taking into account the information provided above, it becomes clear that the colony formation assay does not show the elimination of the HCV RNA from the replicon cells. In other words, colony formation assay does not show the true cure. The death of all replicon cells in colony formation assay, however, is a strong indication that the true cure might be achieved.
Currently available DAA combinations – the primary difference
There are three major classes of DAAs currently used to treat hepatitis C in different combinations and their generic names end in: (i) “-previr” – NS3/4A protease inhibitors; (ii) “-buvir” – NS5B polymerase inhibitors; and (iii) “-asvir” – inhibitors of NS5A, a critical protein for viral replication and infectious virion assembly. The NS5A inhibitors are the most powerful inhibitors of HCV and interfere with multiple stages in replication and virion formation of HCV. Due to this reason, NS5A inhibitors are present in all DAA combinations currently used11. Thus, the primary difference between various DAA combinations is the second component used. This second component is either the HCV protease or polymerase inhibitor. There is also a single triple-DAA combination, which targets all three components of the viral replicase.
How do current most recommended11,12 DAA combinations perform in colony formation assay – the major tool for DAAs discovery and efficiency evaluation? Table 1 (below) gives an overview of the colony formation assay usage for the evaluation of various DAA combinations efficiency.
Colony formation assays allow the researchers to determine the half-maximal (50%) effective concentration (EC50) – the DAA concentration at which 50% inhibition of HCV replication is achieved. At large, there are two different types of EC50 values, which depend on what is assumed under the term “inhibition”. In colony formation assay, the absolute EC50 values are obtained. The absolute EC50 values are calculated based on two numbers: (1) the number of colonies in the absence of DAA (no effect); and (2) zero colonies (100% inhibition) meaning the death of all HCV replicon cells in the colony formation assay. For example, the colony formation assays for glecaprevir (ABT-493) & pibrentasvir (ABT-530) and grazoprevir (MK-5172) & elbasvir (MK-8742) DAA combinations have been performed as described in the previous section and the results were published in original papers13,14.
|Genotypes targeted||Colony formation assay results published?|
|Glecaprevir& pibrentasvir||NS3/4A& NS5A||1–6||Yes13|
|NS3/4A& NS5A||1, 4||Yes14|
|Sofosbuvir& velpatasvir& voxilaprevir||NS5B& NS5A& NS3/4A||1–6||No15,17,18|
Surprisingly, the results of colony formation assays for all HCV NS5B polymerase inhibitor (sofosbuvir)-based DAA combinations, listed in Table 1 (above), have not been published by the respective manufacturer. Why? Because another assay type was used for obtaining the EC50 values.
As mentioned above, the colony formation assay is the surrogate model system that mimics the HCV multiplication in the human liver cells. As in human liver cells infected with HCV, the HCV replicon cells support persistent high-level HCV replication. Colony formation assay typically takes 3–5 weeks.
There is, however, an alternative much faster HCV replication assay which takes only 3–4 days. In this transient assay, subgenomic HCV replicons are introduced into modified Huh-7 cells which support high-level viral replication. Transient replication assays can also be performed with cells harboring subgenomic HCV replicon cells without antibiotic selection. Calculations of EC50 values for sofosbuvir-based DAA combinations were performed using such transient replication assays. During transient replication assays, the HCV replicon cells do not die in DAAs presence – they emit light, when special compound is added. In transient replication assays, the relative EC50 values are obtained because light emission is never zero meaning that cells always contain HCV RNA. Thus, in transient assays, the inhibition of HCV replication is never 100%. The relative EC50 values are calculated based on two numbers: (1) the maximum light amount emitted in the absence of DAA (no effect); and (2) the minimum light emitted in the presence of DAA. For example, relative EC50 values were calculated individually for PSI-797715 (sofosbuvir), GS-588516 (ledipasvir), GS-581617 (velpatasvir), and GS-985718 (voxilaprevir).
Transient replication assay can be viewed as a surrogate for colony forming assay, which is a surrogate for HCV multiplication in human liver cells. Transient subgenomic HCV replication assay provides only an initial (3–4 days) snapshot of DAA effect, whereas colony formation assay demonstrates what is happening during 3–5 weeks of DAA treatment. Moreover, colony formation assay provides the absolute EC50 values and better represents the hepatitis C treatment duration (8–12 weeks). Thus, colony formation assay is, without a doubt, more powerful, more biologically- and clinically-relevant assay than transient replication assay.
Conclusion #1: Due to the fact that colony formation assay is more biologically- and clinically-relevant assay than transient replication assay, the DAAs that were evaluated in the former assay have the advantage over DAAs that were evaluated in the latter assay. Thus, simultaneously targeting NS3/4A and NS5A proteins has the advantage over NS5B & NS5A inhibitor combinations.
The score: Sofosbuvir
0 – 1 NS3/4A DAAs
We have mentioned above that the colony formation assay does not show the elimination of the HCV RNA from the replicon cells – the assay does not show the true cure. In other words, the colony formation assay shows the HCV replicon cell death as a result of G418 antibiotic action due to decreased viral RNA replication. It is important to highlight the surrogate nature of the colony formation assay. In particular, DAA combinations used to treat hepatitis C have no extra antibiotics added and the HCV RNA genomes replicating and producing viruses in the liver have no neomycin phosphotransferase genes. The ability to cure the Huh-7 cells of subgenomic HCV replicon, however, is a strong indication that using the same strategy might result in the same effect in an HCV-infected individual undergoing DAA treatment.
There are two widely used cell lines that have been cured of subgenomic HCV replicons using two different strategies by American19 and German20 groups of researchers. First, the Huh-7.519 cell line has been cured of HCV RNA by continuously growing the subgenomic HCV replicon cells in the presence of interferon-alpha (IFN-α) and the absence of G418 antibiotic. After IFN-α treatment, the HCV RNA could not be detected (detection limit ~10 molecules) in Huh-7.5 cells. Second, the Huh-7-Lunet20 cell line has been cured of HCV replicon by treating the replicon cells with a selective inhibitor in the absence of G418. Both cured cell lines supported increased HCV RNA replication compared to Huh-7 parental cells19,20.
The selective inhibitor used for obtaining the Huh-7-Lunet cells has not been specified. Selectivity, however, means that this inhibitor targeted one of the HCV replicase components. Given the date of the publication20 describing Huh-7-Lunet cells, the potential candidates include only HCV protease inhibitors21 either BILN2061 (culprivir) or VX-950 (telaprevir). Later publications using the same subgenomic HCV replicon cell line showed that the cure can be achieved by using either DEBIO-02522 (alisporivir) or the combination of DEBIO-025 (non-selective inhibitor) and VX-950 (selective NS3/4A protease inhibitor)23. The latter work23 also demonstrated that HCV polymerase inhibitors available at that time could not compete with protease inhibitors as judged by colony formation assay. Thus, it is highly likely that the HCV protease inhibitor was used for the generation of the Huh-7-Lunet cells.
Thus, the true cure can be achieved in the laboratory settings for Huh-7 cells harboring subgenomic HCV replicons and growing on plastic dishes if NS3/4A protease inhibitor is used. The same cannot be verified for NS5B polymerase inhibitors, even though respective claims had been done. In particular, it was claimed that R1479, a NS5B polymerase inhibitor, cures the cells of subgenomic HCV RNA in 15 days (importantly, colony formation assay was not performed)24. The colony formation assay results23, however, did not support this claim. Moreover, there are no available and widely used cured cells that were generated using NS5B polymerase inhibitor treatment either alone or in combination with any other selective or non-selective inhibitor.
Conclusion #2: Demonstrated ability of NS3/4A protease inhibitors to cure the Huh-7 cells of HCV RNA in laboratory setting strongly suggests that DAAs simultaneously targeting NS3/4A and NS5A proteins have the advantage over NS5B & NS5A inhibitor combinations.
The score: Sofosbuvir
0 – 2 NS3/4A DAAs
As defined above, the true “cure” implies the absence of HCV and its replicase complexes from the liver. Translating this into laboratory setting – “cure” is the absence of HCV and its replicase complexes from hepatocytes living on plastic dishes. It is thought that for the assembly of infectious viral particles replicase complexes interact with a specialized lipid storage containers of the cell called lipid droplets25. It is also thought that after the virions have been assembled they are rapidly released by the liver cells26.
If HCV is not released by an infected liver cell, then remaining healthy liver cells not become infected. How do different classes of DAAs perform in infectious virus release experiments? It was demonstrated that NS5A and NS3/4A protease inhibitors were capable of early, near complete (~100%) inhibition of infectious virus release, whereas NS5B polymerase inhibitors acted slowly and achieved only ~80% inhibition26–28.
As mentioned earlier, there are two types of HCV replicase complexes: (i) those that were present before DAA treatment; and (ii) those that appeared after starting DAA treatment. Various DAAs have different effects on the two types of HCV replicase complexes. By following the viral RNA synthesis in HCV-infected cells researchers found that NS5A inhibitors block the formation of newly forming viral replicase complexes, whereas NS3/4A protease inhibitors block the viral RNA synthesis in all replicase complexes26. This type of experiment has not been performed for NS5B polymerase inhibitors (e.g., sofosbuvir).
The activity of HCV replicase complex is measured by the amount of viral RNA produced in an infected cell. Strikingly, NS5B polymerase inhibitors reduce the viral RNA amount by only 80%29. This finding is directly supported by experiments conducted with subgenomic HCV replicon cells, in which ~80% reduction in the amount of replicon RNA is observed when NS5B polymerase inhibitors are added30. A very recent back-to-back comparison of NS3/4A protease (simeprevir) and NS5B polymerase inhibitor (sofosbuvir) effects on viral RNA synthesis showed the superiority of the former inhibitor class, when the same DAA concentrations were used31.
During hepatitis C treatment, HCV RNA level in the blood of an infected individual declines in two phases. During the first phase, a rapid drop in viral load is observed, whereas the second phase results in further slow decline of HCV RNA. Using mathematical modeling of HCV RNA levels measured in patients as a function of time, it was demonstrated that the second phase of viral RNA decline is 2–3-fold faster for NS3/4A protease inhibitors (e.g., telaprevir) when compared to the decline observed for NS5B polymerase inhibitors (e.g., sofosbuvir). Based on this result, it was proposed that the NS3/4A protease and NS5A inhibitor combination can inhibit viral replication, assembly, and release32.
Conclusion #3: Two facts demonstrate directly the advantages of NS3/4A protease inhibitors over NS5B polymerase inhibitors. First, NS5B polymerase inhibitors (including sofosbuvir) are unable to completely block the HCV release by infected cells, whereas NS3/4A protease and NS5A inhibitors are capable of almost immediate and near complete inhibition of virus release. Second, NS5B polymerase inhibitors, when used at therapeutic doses, are outperformed by both NS3/4A and NS5A inhibitors in viral RNA replication experiments. In addition, mathematical modeling of real patient data indicates that NS3/4A protease and NS5A inhibitor combination is optimal for viral replication, assembly, and release inhibition.
The score: Sofosbuvir
0 – 3 NS3/4A DAAs
Resistance to DAAs means that much higher drug concentrations than are being used in standard treatment regimens are required to inhibit HCV replication. DAA resistance can lead to unsuccessful hepatitis C treatment. Often, when talking about HCV DAAs a “high barrier to resistance” of NS5B polymerase inhibitor, sofosbuvir, is highlighted. It is generally thought that the NS3/4A protease and NS5A protein inhibitors have a low barrier to resistance. In particular, it is assumed that the mutations arising in NS3/4A and NS5A due to error-prone HCV replication confer the resistance to DAAs targeting these proteins much easier, than mutations that accumulate in NS5B polymerase. Why? …Because the major mutation33 associated with sofosbuvir resistance results in a 5–20-fold drop of NS5B polymerase activity in a test tube34. Due to this reason, it is assumed that the viruses harboring this mutation are not fit enough to “survive” in infected liver cells.
The mechanism for mutation accumulation during HCV replication is relatively simple. Viral replicases have a property of incorporating wrong RNA building blocks, called nucleoside monophosphates, into a nascent growing RNA strands during replication. As a result, nascent viral RNAs contain errors and differ from the original viral RNA template genetically. Thus, viral replicase complexes of every single infected liver cell produce a mixture (also called a swarm or a cloud) of genetically-different mutant viruses or quasi-species. Sometimes, the errors accumulated during viral replication can lead to mutations in viral proteins. These mutations can either prevent the DAA binding or reduce the effect of DAA by some other mechanism, which leads to resistance. At the same time, the mutations in viral RNA and viral proteins encoded by this RNA can lead to either increased or reduced virus replication capacity or fitness (phenotype difference).
The researchers from Spain and USA produced highly-fit HCV swarms by serially cultivating HCV for a prolonged time on plastic with Huh-7.5 cells35. They found that these highly-fit HCV swarms were resistant to sofosbuvir and at the same time did not have sofosbuvir-specific resistance mutations28. Similar observations were made for NS3/4A protease (telaprevir) and NS5A (daclatasvir) inhibitors36. During chronic infection, HCV replicates for a prolonged time in the liver environment of an infected individual. The highly-fit HCV swarms used in laboratory studies, mentioned above, mimic the HCV swarms present during chronic infection in patients. Thus, both resistance mutations and high HCV fitness define the sensitivity of HCV to antiviral treatment.
Conclusion #4: It is not known whether resistance mutations or high HCV fitness play the key role in defining the sensitivity of HCV to DAA treatment. Consequently, a therapeutic advantage of NS5B polymerase inhibitor, sofosbuvir, which is assumed due to its “high barrier to resistance” might not be clinically relevant.
The score: Sofosbuvir
1 – 4 NS3/4A DAAs
We have mentioned direct-acting antivirals many times throughout this NextGenRnD Solution. There are three major classes of DAAs targeting NS5A protein, NS3/4A protease, and NS5B polymerase. In the previous section, we have also mentioned that mutations accumulating in these viral proteins can preclude the binding of DAAs. It turns out, however, that not all DAAs bind to their target viral proteins and remain bound. In other words, the manner in which HCV replicase proteins are targeted by DAAs is different. This begs a question: How direct are direct-acting antivirals?
NS3/4A and NS5A DAAs bind their respective viral targets and remain bound for a prolonged time or permanently, whereas NS5B DAA, sofosbuvir, binds to the viral polymerase only for a very short time (transiently). In particular, sofosbuvir, in a modified form, is incorporated by the HCV NS5B polymerase into growing nascent viral RNA. Sofosbuvir is called chain terminator, because once it is incorporated into growing viral RNA strand (chain), the viral polymerase cannot extend this RNA any further.
It is generally assumed that sofosbuvir-terminated viral RNA chain cannot be extended further to complete the synthesis of full-length viral RNA. In a real-world scenario, however, sofosbuvir molecule might be removed by HCV NS5B polymerase and viral RNA synthesis might be re-enabled. It turns out that viral polymerase has an additional activity, which can facilitate the excision of chain-terminating inhibitor molecule in the presence of pyrophosphate (PPi)37. Thus, PPi-mediated sofosbuvir excision by the viral polymerase might be responsible for the inability of sofosbuvir to fully (or near completely) inhibit the HCV release and replication (see argument #3 for details). The experiments addressing potential pyrophosphate-mediated sofosbuvir excision by NS5B polymerase have not yet been published.
Pyrophosphate is an abundant molecule in human plasma and it is thought that a substantial amount of PPi is secreted by the liver38. In a study with rats, it was found that ethanol and acetate increased the liver PPi content 2.4- and 13-fold respectively39. Furthermore, in a study with human volunteers, it was found that ethanol consumption triggered very high blood acetate levels40. Thus, alcohol consumption during sofosbuvir-based therapy should be avoided not just because of the stress imposed on the human liver but also due to NS5B polymerase excision activity potentially targeting sofosbuvir incorporated into viral RNA.
In addition, sofosbuvir has a so-called off-target activity. In particular, it can be incorporated into cellular (host) RNAs by human mitochondrial RNA polymerase41,42. This can lead to disruption of synthesis of some essential host RNAs in mitochondria, the energy powerhouses of the cell. Defective host RNA synthesis in mitochondria can lead to decreased energy production. It has been proposed that energy failure is an important factor influencing the development of both acute and chronic hepatic encephalopathy43.
Conclusion #5: Mutations accumulating in the components of viral replicase (NS5A, NS3/4A, or NS5B) can render HCV resistant to respective DAAs. However, in addition to the above-mentioned mechanism, the viral NS5B polymerase has an excision activity that might preclude the effect of sofosbuvir and other chain-terminating inhibitors. Due to this extra unwanted NS5B activity the NS3/4A DAAs have an advantage over chain-terminating NS5B DAAs including sofosbuvir. Finally, because sofosbuvir interacts with HCV NS5B polymerase only for a short time and has an identified off-target activity – it is the least direct of all direct-acting antivirals available for hepatitis C treatment.
The score: Sofosbuvir
1 – 5 NS3/4A DAAs
Virtually all human cells have an innate immune system – the first line of defense against viruses like HCV. The innate immune system is capable of recognizing and counteracting the viral infection. Recognition of viral infection typically triggers induction of type I interferons (IFNs), which promote an antiviral state. In mammals, type I interferons and several other cytokines — molecules triggering antiviral and inflammatory states — provide cues that determine the subsequent reaction of adaptive immune system, which consists largely of two types of lymphocytes: T and B cells44. One of the adaptive immune system’s reactions is the destruction of the infected cell44,45. Thus, activation of the innate immune system in HCV-infected hepatocytes can lead to their elimination by the cells of adaptive immune system.
We have already mentioned that some HCV replicase components have conflicting effects46 on the antiviral defense system of the cell. Paradoxically, these replicase components, however, do not make up the replicase complexes themselves and as a result do not participate in HCV RNA replication. It turns out that during HCV RNA replication, the replicase components are overproduced47 and have additional functions inside the host cell being invaded. In particular, the NS3/4A protease inhibits the type I IFN response46,48,49, whereas NS5B polymerase activates it46,50.
It was demonstrated that producing HCV NS5B polymerase in mouse liver and human hepatocytes triggers significantly increased levels of type I IFNs (IFN-α and IFN-β)51. Later studies demonstrated that genetically modified mice producing a polymerase similar to HCV NS5B had a life-long increased type I IFNs and cytokine production with no health issues (e.g., no liver pathology, normal fertility and longevity, etc.) and were resistant to typically lethal viral infections52,53. Moreover, human cells expressing the same polymerase were resistant to human immunodeficiency virus 1 (HIV-1)52. It appears that the NS5B polymerase might be used against the HCV itself. In fact, it was proposed that maintaining persistent IFN production in HCV infected cells through inhibition of NS3/4A protease might be helpful in prevention of chronic HCV infection51. Thus, blocking the NS3/4A protease with DAAs might indeed help maintain IFN production, whereas inhibiting the NS5B polymerase might preclude it.
There is a big difference between highly toxic IFN-α therapy (used prior to DAAs availability) and the IFNs produced by infected liver cells. The former is systemic, which means that essentially high concentrations of injected IFN are present throughout the body, whereas the latter is local – present in high concentrations only in the vicinity of HCV-infected liver cells. Localized IFNs and cytokines presence typically leads to highly specific and precisely targeted adaptive immune system reaction. Such local activation of adaptive immune system might lead to destruction of HCV-infected cells.
What does the above body of evidence mean? First of all, including the immune system into the equation that evaluates the potential of DAA effects leads to novel perspectives on hepatitis C treatment. It appears that in the context of immune system it is more beneficial to target NS3/4A protease, whereas targeting NS5B polymerase might lead to inability of human cells to detect the viral infection (i.e., inability to elicit type I IFN response). In fact, it was demonstrated that DAA combinations including NS5B inhibitor (dasabuvir or sofosbuvir) suppressed antiviral type I IFN response during the first four54 to twevle55 weeks of treatment period. In addition, as discussed earlier, type I IFNs can cure the replicon cells of HCV RNA on plastic19.
Conclusion #6: NS3/4A protease inhibits the antiviral type I IFN response, whereas NS5B polymerase activates it. Thus, DAAs targeting NS3/4A have an advantage over NS5B DAAs.
The score: Sofosbuvir
1 – 6 NS3/4A DAAs