Statistics from Altmetric.com
First-generation protease inhibitors: where are we now?
The first-generation protease inhibitors (PIs) of hepatitis C virus (HCV), telaprevir and boceprevir, are the harbingers of important advances in the treatment of chronic HCV infection. Improved response rates have been observed in both previously untreated and previously treated patients who have failed pegylated interferon (PEG-IFN) and ribavirin (RBV) treatment.1–5 Sustained virological response (SVR) rates of >70% have been reported. Despite the side effects of the agents and their continued necessity for a PEG-IFN and RBV backbone, the high rates of early viral response translate into higher rates of cure than with PEG-IFN and RBV alone in genotype 1 infection. Their efficacy in the treatment of genotype 2–6 has not been fully tested.6 Activity in in vitro cell culture systems showed similar efficacy of PIs against genotypes 2a, 5a and 6a and comparatively low but varying efficacy against genotype 3a isolates.7 Replication-competent intergenotype chimaeras inserted into the JFH clone of HCV have shown susceptibility to telaprevir, with genotypes 1 and 2 being most susceptible and genotypes 4 and 5 most resistant.8
Previously untreated (naïve) patients without severe fibrosis respond better to the first-generation PIs than those with advanced fibrosis or cirrhosis. Although response rates in patients with cirrhosis are improved compared with PEG-IFN and RBV treatment, they remain suboptimal. Patients with a prior null response (defined as a decline in HCV RNA of <2 log10 after 12 weeks of treatment) and cirrhosis are less likely to respond to retreatment with telaprevir or boceprevir.9 ,10 Approximately 15% of prior null responders and cirrhosis respond to first-generation PIs, and only 35% of those with a prior partial response and cirrhosis respond to retreatment. Inherited IL28b haplotypes continue to influence response rates.11 ,12 Response rates are slightly lower in subtype 1a compared with subtype 1b.
These advances have brought higher rates of cure but more complexity to the treatment of hepatitis C—a paradox of progress.13 Thus, despite the technological and medical advances, the first-generation PIs create more problems rather than making treatment more straightforward.14 A complex process of decision making is required to assess the indications for treatment of naïve and previously treated patients, and for patients with mild disease versus those with cirrhosis or advanced cirrhosis. For example, do all genotype 1 patients likely to respond to PEG-IFN treatment—that is, those with early stages of fibrosis and a favourable IL28b CC haplotype—need triple therapy, considering the increase in costs due to first-generation PIs, the side effects and the risk of resistance? Or can we avoid exposure to PIs in some (table 1)?
Cost-effective studies have examined the selective administration of PEG-IFN RBV versus the triple combination of telaprevir or boceprevir together with PEG-IFN RBV. Five competing strategies were compared in a cost-effective analysis.15 Boceprevir-RVR and telaprevir-IL28B strategies were the most effective and cost-effective of the evaluated strategies and were dominant—that is, dual therapy was a cost saving first-line therapy in patients with IL28B CC genotype or in those who achieved a rapid virological response (RVR). This and other studies provide evidence that a selective allocation system requires consideration. However, there are caveats: pricing affects the cost savings and the reimbursement setting is important. The American Association for the Study of Liver Diseases 2011 update practice guidelines do not recommend selective allocation of patients to triple treatment with first-generation HCV PIs. Similarly, the National Institute for Health and Clinical Excellence (NICE) in the UK and the German authorities do not recommend selective allocation.16–18 Other countries do. Some cost savings could be made. However, fewer than one in five genotype 1 naïve patients have a RVR and the cost savings made are not large. Patients with a IL28b CC haplotype who do become undetectable after 4 weeks of dual therapy can elect for treatment with the dual regimen alone, but will require a full 48 weeks of IFN in most cases; abbreviated treatment with PEG-IFN alone only suffices for those with a low viral load. Conversely, the addition of a PI offers treatment-naïve patients without cirrhosis and responsive to PEG-IFN with a RVR the likelihood of shorter courses of treatment. This strategy adds a layer of complexity and necessitates IL28b testing or a ‘lead in’ with PEG-IFN and RBV for all patients. Determining an appropriate RVR brings another layer of intricacy to decision making based on response criteria. The assays used differ in their analytical sensitivity, particularly their lower limits of detectability and quantitation of HCV RNA. Assays used in the pivotal phase III trials may show slightly different performances from assays provided by commercial laboratories. Thus, it is unclear at this stage how to manage patients with a week 4 HCV RNA result of ‘detectable but not quantifiable’. Do these patients always need PIs? Is response-guided therapy possible in these patients? Finally, the exact day of timed testing at week 4 has to be considered if results are delayed.
What have we learnt from the clinical experience with first-generation PIs?
Telaprevir and boceprevir are generally recognised as an advance in treatment of chronic hepatitis C. The efficacy of these agents has been substantiated; shortening of treatment to 24 weeks is a major therapeutic advance. The high positive predictive value of >90% in patients with RVR is an important motivator. However, clinical experience in patients with cirrhosis has been tempered by the potential for serious adverse events in patients with severe liver disease. The side effects of treatment require management and intensive monitoring in some. Rashes (of varying grades of severity and duration) have been reported in 55% of patients treated with telaprevir. Most drug-related dermatitis has been mild to moderate in intensity. Therein lies a difficulty. The risk of severe drug-related reactions requires careful longitudinal evaluation as telaprevir-induced rashes may progress somewhat unpredictably to a more severe drug-related rash. Approximately 5% of patients experience severe rash during treatment with telaprevir in combination with PEG-IFN and RBV. Drug rash with eosinophilia and systemic symptoms occurred in 0.4% and Stevens–Johnson syndrome in 0.1% in clinical trials.19 The evaluation of the course and outcome of the rash poses a difficulty as even moderately severe rashes may be transient and improve. However, a severe progressive drug-induced dermatitis requires stopping first telaprevir and subsequently PEG-IFN and RBV if improvement is not rapidly observed. Nausea, anorectal symptoms and dysgeusia are unpleasant symptoms that affect many patients.
The management of anaemia, however, has proved to be most problematic. Anaemia occurs with both telaprevir and boceprevir and is a class effect of the first-generation PIs.20 The effect on haemoglobin is unfortunate, adding to the haemolytic anaemia induced by RBV and the bone marrow suppressive effect of PEG-IFN. Haemoglobin concentrations of <10 g/l have been reported in one-third of patients and <8.5 g/l in approximately 15%. The fall in haemoglobin requires RBV dose reductions or erythropoietin use; previously, RBV dose reductions were shown to be detrimental during PEG-IFN and RBV. Recent data suggest that the addition of a direct acting viral PI reduces (but most certainly does not eliminate) the need for RBV.21 RBV dose reductions resulted in the same SVR as erythropoietin use in a randomised controlled trial (perhaps testimony to the inadequate effect of erythropoietin).22 The timing of RBV dose reductions in patients less responsive to PEG-IFN could still be important, and secondary measures have been required.23 Approximately 2% of patients received a blood transfusion in clinical trials, although the number was higher (15%) in patients with cirrhosis in the French expanded access programme.24
The requisite backbone of PEG-IFN means that other problems associated with PEG-IFN and RBV use—including depression, psychiatric symptoms, worsening of liver function and severe infections—still complicate a regimen including PEG-IFN and render a large group intolerant of PEG-IFN ineligible for treatment.
What are the implications of resistance to first-generation PIs?
Treatment failure with telaprevir or boceprevir is associated with the potential development of viral drug resistance that could limit future treatment options. PIs target the N-terminal serine protease domain of the NS3 protein. Common resistance mutations to boceprevir and telaprevir have been described;25–28 several mutations located in the catalytic site of the protease confer cross-resistance to multiple NS3 PIs.29 It has been noted that the frequency and pattern of first-generation PI mutations are different between HCV subtypes 1a and 1b.
What threat does resistance pose?30 Unlike HIV, HCV does not replicate via a DNA intermediate and the HCV genome does not integrate into the human genome; as a result, there is no stable genetic reservoir with HCV infection. Thus, once drug selection pressure is discontinued, the advantage is lost and resistant variants are again outgrown by more fit wild-type virus. After discontinuation of telaprevir or boceprevir, levels of wild-type variants increase relatively rapidly and may completely replace resistant variants in the quasispecies.27 ,31–33 The result is the theoretical possibility of restoration of drug susceptibility. However, the speed of disappearance of the resistant variants from the dominant quasispecies varies by subtype. In an analysis of phase III telaprevir studies, Bartels and colleagues showed that V36M/A/L, T54A/S, R155K/T and A156S/T mutations emerged frequently in patients failing treatment. Among patients with subtype 1a HCV, the median time to loss of resistant variants V36M and R155K was approximately 6 and 10 months, respectively, whereas loss of the combination of both mutations required approximately 13 months.34 ,35
Other ongoing studies suggest that, among patients with emergent resistance during telaprevir-based therapy, 85% did not have detectable resistant variants at a median of 29 months following treatment failure.36 Similar data exist after boceprevir resistance.37
To date there are very few data on the outcomes of retreatment of patients who have failed boceprevir or telaprevir. Preliminary data suggest that patients may respond to PI retreatment after short-term exposure.38 ,39 However, on balance it is unlikely that, despite the apparent reversion of resistance with time, retreatment with first-generation PIs in an inappropriate regimen will be successful. The efficacy of TMC435 or MK5172 plus PEG-IFN RBV to treat first generation PI resistance is unproven. More clinical data are required to confirm that retreatment with next generation PIs together with other drug classes or with agents from the original class but within a new regimen will be successful.
Successful retreatment following failure with boceprevir and telaprevir will rely on future regimens and the development of agents that lack cross-resistance across classes.29 However, in some patients, particularly those with cirrhosis and genotype 1a HCV infection, reversion to wild type may be a somewhat lengthy process and disease progression may occur during the interval. The importance of the specific retreatment regimen cannot be overemphasised. In addition, most studies have used population sequencing methods to study reversion to wild type, a relatively insensitive method. For promising second-generation PIs such as MK5172, it will need to be established that the drug concentrations achieved in vivo are sufficient to control the replication of Resistance associated variants (RAVs) without toxicity. Thus, the clinical relevance of RAVs identified years after short-term exposure to boceprevir or telaprevir remains unclear. For now, the adherence to stopping rules to ensure prompt discontinuation of futile therapy and thereby limit the likelihood of developing compensatory mutations that may prolong the durability of RAVs is important.
What about IFN-sparing regimens?
Current treatments will have a limited impact on disease if the stated aim is to treat very large numbers of infected persons to reduce the morbidity from liver disease caused by HCV. For several reasons there is a concrete expectation that first-generation PIs will be displaced in the not too distant future by a next generation of PIs with fewer severe side effects (in particular, without anaemia) and more convenient daily or twice daily dosing. TMC435 and BI291335 (both PIs) are examples; phase III trials are nearing completion so licensing could be achieved by 2013/2014.40–42 What then? Are regimens of direct acting antivirals (DAAs)—for example, a second-generation PI or an NS5b polymerase inhibitor or an NS5a inhibitor with PEG-IFN and RBV—the future of hepatitis C therapy?
We have limited but interesting information. The bar has been set high by recent data obtained in IFN-sparing regimens. For example, the ATOMIC study in patients with naïve genotype 1 has shown that a regimen of a nucleotide polymerase inhibitor (NPI) GS-7977 (sofosbuvir) together with PEG-IFN and RBV was highly effective (>90% SVR) in naïve patients with genotype 1 and 3 treated for only 12 weeks.43 It is likely that several other trials under evaluation with second-generation DAAs may yield similar results (with the same constraints of the IFN-RBV backbone). Such a regimen may yet appeal to persons not intolerant of or not ineligible to IFN and may be more affordable, given that the patent for PEG-IFNα will expire shortly and RBV is inexpensive.
However, retreatment with a second wave PI (eg, TMC435) is not optimal in prior null responders with cirrhosis because of the poorly understood poor responses to IFN in patients with cirrhosis.44 Paradoxically, therefore, quadruple PEG-IFN-containing regimens—that is, PEG-IFN together with RBV and two DAAs—are being tested. Preliminary evidence has suggested the efficacy of quadruple IFN/DAA regimens which may prove more effective than telaprevir or boceprevir and PEG-IFN with RBV.45 However, safety has to be demonstrated in patients with cirrhosis. Thus, quadruple regimens present another paradox in that a quadruple regimen could be required for difficult-to-treat patients with cirrhosis but will carry a greater risk. There is the possibility that PEG-IFNs with fewer haematological side effects (eg, PEG-IFNλ) could be useful in this context, but this too requires proof.46 ,47 This fact, and perhaps the unwelcome need for quadruple PEG-IFN regimens, may mean that PEG-IFN has a long goodbye.
In theory a combination of drugs that do not share cross-resistance (in the replicon model) should achieve synergistic or additive effects without resistance. Combination oral therapy has been tested and the initial errors have probably been made in the teething stages. We have learnt that combination oral therapy does not necessarily protect against dual resistance as observed, for example, with the combination of VX222, a non-nucleoside polymerase inhibitor (NNPI) and teleprevir in the ZENITH study. Nonetheless, IFN-free treatment would seem to hold the key to the future. Proof-of-concept studies have established that IFN-free cures are indeed possible. This was observed with daclatasvir (an NS5a inhibitor) and asunaprevir (a PI), but only for prior null responders with subtype 1b.45 Daclatasvir and asunaprevir have not proved to be effective for subtype 1a and have not been sufficiently tested in patients with cirrhosis. The regimen also failed more frequently in a population of so-called IFN-intolerant patients with subtype 1b, and appeared to be dependent on adherence (and the pharmacokinetics) in this population in whom SVR rates declined from 91% to 64%.48
Naïve patients with genotypes 1 and 3 treated with daclatasvir and the NPI sofosbuvir for 24 weeks resulted in striking SVR rates without major toxicity. However, these data should be tempered by the small number of patients and also, disappointingly, by the cessation of industry cross-collaboration.49 A 12-week trial of ABT-450 (a ritonavir-boosted PI) and ABT-333 (a NNPI) plus RBV in patients with G1 HCV and previous non-responders achieved SVR12 in 93–95% of treatment-naïve patients but in only 47% of previous non-responders infected with HCV genotype 1.22 Responses were lower in those with IL28b TT haplotype, emphasising the effect of intrinsic IFN responsiveness on the ability to respond to IFN-free regimens reported thus far.
All-oral regimens in development: problems going forward
By contrast, IFN-free regimens have proved worrisome in preliminary data in genotype 1 null responders treated for a relatively short period of 12 weeks with sofosbuvir plus RBV. Relapse occurred in 90% of patients although resistance during treatment was not detected. Ritonavir-boosted PIs within all-oral regimens have so far proved less effective for prior partial responders and null responders than in naïve patients.22 A late relapse more than 24 weeks after the end of treatment in a naïve IL28b CC positive naïve genotype 1 patient treated with ABT-450/r and ABT-072 (a NS5a inhibitor) and RBV has provoked much discussion.50 How much threat do late relapses pose to patients and to development programmes?
Nonetheless, there are several realistic all-oral regimens and the therapeutic landscape is undoubtedly forever changed for the better. A number of potent inhibitors of HCV are being assessed, with a high barrier to resistance and pan-genotypic effects. Some oral regimens are entering late phase II and even phase III assessments. We will need to know whether drug interactions between the DAAs occur. We do not yet know whether RBV remains important, or the appropriate dose of RBV in these regimens. The role of RBV remains a work in progress. Haemolytic anaemia without PEG-IFN is less problematic; other potent DAAs may offer an advantage over RBV in a regimen but the costs will escalate. Whether null responders to PEG-IFN and RBV are null responders to both agents is a possibility—for example, as observed with sofosbuvir and RBV—and perhaps such patients will require a regimen of multiple DAAs without RBV. Several relevant studies examining oral regimens with or without RBV arms will answer these questions. Other regimens are relying on ritonavir boosting to achieve the desired effect.
A list of current regimens undergoing assessment can never be complete, but includes sofosbuvir+RBV in G1,2,3 naïve and intolerant subjects and prior treatment failures; daclatasvir and asunaprevir in IFN-intolerant G1b non-responders; GS5885 (an NS5a inhibitor) and GS9451 (a PI) and GS9190 (tegobuvir, NNPI) in G1 naïve and IFN-intolerant patients; sofosbuvir and TMC435±RBV in G1 null; BI201335 (a PI) and BI207127 (a NNPI)±RBV; daclatasvir and asunaprevir and BMS791325 (NNPI) in 1a and 1b naïve patients; daclatasvir and TMC435±RBV; VX222 (NNPI) and telaprevir+RBV in G1 naïve patients; ABT333 (NNPI) and/or ABT072 (an NS5a inhibitor) and ABT450 (a PI).
Thus, at present it is difficult to make comparisons of the efficacy of PEG-IFN-free versus PEG-IFN-sparing regimens, begging the question of whether IFN-sparing regimens (with IFNs with less haematological toxicity such as IFNλ) will still appeal, given the high response rates and perhaps lower cost. Such regimens may still be applicable in some regions and in some cohorts whereas IFN-free regimens will have far greater appeal in others, including IFN-intolerant, IFN-ineligible or IFN-unwilling patients. We require more data in patients with cirrhosis, decompensated cirrhosis, post-transplant recurrent hepatitis C, HIV and HCV co-infected patients and patients with renal impairment or post-renal transplant patients. We have not yet scratched the surface with resistance. Surely the next group of patients requiring attention will be those resistant to first-generation PIs; will IFN-free regimens be effective?
Why do null responders fail to respond to some oral regimens?
Null responders are supposedly defined by poor responsiveness to PEG-IFN; ‘null’ in theory pertains to a null response to IFN due to a lack of IFN sensitivity and an HCV-induced impairment of the innate immune response. How can this pertain to patients who fail to respond to an IFN-free regimen? The explanation is uncertain as defining and treating a prior null response remains a puzzling conundrum. IL28b single nucleotide polymorphisms have been shown to be an important determinant of the response in patients with subtype 1a treated with a PI and a NNPI, but not in those with subtype 1b.51
Thus, a genetically determined and viral influenced innate response may yet apply to DAAs, pointing to a role for the innate immune response in viral clearance with these agents. The explanation for failure remains uncertain at present but prior null response could also reflect viral perturbations induced by prior treatment, and even possibly RBV resistance. IL28b polymorphisms affect the second phase viral decline (ie, clearance of HCV from cells). Functional studies have shown that HCV may disrupt the innate immune response by abrogating IFN signalling.52
This effect may be nullified in patients treated with potent regimens given for a sufficient time to effect clearance of HCV from the liver and perhaps extrahepatic sites. We may find that there will be an inverse correlation between the influence of the innate response and the potency of DAA agents. Other factors such as treatment duration, pharmacokinetics and pharmacodynamics are critically important questions to answer. Thus, treatment for too short a period may lead to insufficient clearance from the liver or possibly other sites. Is RBV resistance a possibility? There is little opportunity for on-treatment intensification of a failing regimen as these patients relapse after treatment. It will be important to devise appropriate regimens for different patients that might differ between subtypes and naïve versus prior non-responders. It has proved difficult to develop an experimental model to test the abrogation of the immune response.
Who should be treated now and who might best wait for treatment?
Given this broad therapeutic landscape, what is the most relevant and important group to treat now? Many factors determine this paradigm. It is in part determined by the particular reimbursement clauses that pertain in different regions and countries. In the UK, reimbursement is being provided via the NICE guidelines. All naïve and experienced genotype 1 patients are potential candidates for treatment. Should patients at either end of the disease spectrum who have either mild disease not needing immediate treatment or have advanced or severe hepatic fibrosis start treatment with a first-generation PI now or wait for the possibility of improved treatments or even possibly a PEG-IFN-free regimen?
Naïve patients who want treatment and who agree that this is an opportune time for them to be treated should be offered treatment. This contract requires that patients do not have contraindications to PEG-IFN treatment and are motivated to receive a triple combination IFN-containing regimen. Patients should understand the inability to accurately predict disease progression, likelihood of response, the regimens, the side effects and the risks and benefits of treatment, and should be cognisant of treatments that are coming down the line or are given this information. Physicians could consider it rational to give treatment to those with progressive disease but to avoid unnecessary or potentially detrimental treatment. The regimen is not free from the disadvantages of IFN. In giving advice, physicians should clarify the key determinants of response, the likelihood of response (in naïve and prior non-responders), the need for treatment, side effects and safety. Patients with mild disease do not necessarily require immediate treatment, but treatment should not be denied to patients with early stages of disease as those with minimal necroinflammatory changes and minimal fibrosis are candidates for treatment and have a higher likelihood of response.
Patients with compensated cirrhosis are candidates for treatment; IFNα is difficult to apply in decompensated cirrhosis and may precipitate deterioration. Compensated cirrhosis without portal hypertension differs markedly from a later stage of cirrhosis characterised by evidence of portal hypertension and thrombocytopenia or cirrhosis with hepatic decompensation.53 ,54 Physicians must be able to judge the stage of disease and to manage the poor tolerability of triple combination therapy and potential complications of prolonged IFN treatment in patients with advanced cirrhosis.
Psychiatric or other comorbidities may be worsened with IFN treatment so these conditions should be stabilised.55 For prior non-responders and patients with cirrhosis, it may be necessary to take into account the nature of the prior response, the IL28b subtype and HCV subtype if deferring treatment for those unlikely to respond is considered.
For experienced patients, the data to indicate improved SVR rates in patients with first-generation PIs should provide the motivation. However, response rates are suboptimal in prior null responders with cirrhosis and are <10% for previous null responders with cirrhosis and HCV genotype 1a infection. Ultimately, what drives a patient averse to IFN to be treated with telaprevir or boceprevir and PEG-IFN and RBV? What motivates a patient already treated with IFN to want to do it again with a first-generation PI regimen? What role does the physician have in influencing these choices? There are no easy answers. The uniqueness of the HCV treatment decision process has recently been discussed. Deferring HCV therapy pushes the doctor–patient relationship past easily defined boundaries and creates the need for an informed deferral process.56 We have reached an era of personalised medicine in hepatitis C. The importance of specialised nurses has increased in order to facilitate and support treatment, and physician experience is required.
What are some of the key remaining questions?
Protocols for impending investigational trials will have to take into account several levels of uncertainty. Perhaps the most pressing issue is treatment duration in differing categories of patients. Will all oral treatment strategies narrow down to a few regimens initially and then even to a ‘one size fits all’ regimen? This seems unlikely in the shorter term. However, there is not yet clarity as to the degree of segmentation into IFN-sparing regimens versus IFN-free regimens or specific IFN treatment regimens for patients with differing subtypes and host IL28b genotypes. The uncertainty continues to pose developmental difficulties for devising appropriate trial designs.
The licensing of several new DAAs has to date required classical development pathways with a backbone of PEG-IFN and RBV, but the protocols being evaluated may not necessarily correspond with how the DAAs will be used. Cross-collaboration at a relatively early stage should be encouraged to develop optimal regimens with a low risk of failure. After that it is not clear whether ‘light touch’ regulation and reimbursement will allow an appropriate combination of agents that may not have been fully evaluated in phase III studies. What will be the payer's attitude to this approach? Thus, there is still a degree of ‘zig-zagging’ between development plans for IFN-sparing and IFN-free regimens until more clarity emerges. However, the ideal solution cannot be proved until further key data gaps have been closed.
How can more widespread treatment be applied? If the stated aim is to treat millions of patients to reduce the burden of disease, how will this be afforded? Ground-breaking antiviral treatment will have to be priced at a cost that society will bear and that will not stifle innovation. Drug pricing currently shoehorns the cost of drugs to fit cost-effective thresholds, but budget reductions to cope with declining funding may lower the currently acceptable thresholds. Higher prices will reduce the effectiveness of penetration of treatment into key cohorts of infected persons and will not provide the incentive to implement wider programmes of ascertainment and treatment. High costs may negate community treatment algorithms and the involvement of general practitioners in treating ‘easy to treat patients’, thus affecting screening policies. Finally, high prices will impair the opportunity to treat millions of patients in resource-constrained regions of the world for decades.
The first-generation PIs have imposed new demands on monitoring, resistance and safety testing. Although they have brought higher response rates than with PEG-IFN and RBV, they have not simplified treatment and do not overcome the several barriers to diagnosis, care and treatment of hepatitis C at this critical juncture.57 It is to be hoped that future regimens will both simplify and expand treatment for chronic hepatitis C.
Contributors The manuscript has been written by both authors. No professional or additional help has been sought or obtained.
Competing interests GD: Vertex, Abbott, Boehringer Ingelheim, GlaxoSmithKline, Novartis, Pharmasett, Pfizer, Roche/Genentech, Merck, Tibotec, Astek, Achillion, Janssen: Advisory Board; Bristol Myers, Gilead Sciences, Human Genome Sciences, Presidio: Advisory board/Safety Monitoring Board; Zymogenetics: Safety Monitoring Board.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.