The debut of antisense drug discovery in the early 1990s was met with great enthusiasm. In contrast to small molecule drugs, which exert their effects by binding to proteins, antisense drugs bind to RNA, preventing the protein from being synthesized in the first place, or modifying the structure of the protein that is formed. This novel approach created hope for the development of new drugs that targeted biological pathways previously regarded as completely intractable due to the inability to identify small molecules with sufficient binding affinity or selectivity.
In 1991, antisense drug discovery company ISIS Pharmaceuticals (ISIS) went public and within 6 months, achieved a market capitalization of $400 (nearly $700M in 2013 dollars). In the ensuing 21 years, more than 75 antisense compounds entered the clinic in over 125 indications, with ISIS as sponsor or licensor of approximately two-thirds of these. Despite this massive R&D effort, only a single product was approved for marketing by the FDA. Vitravene, an intraocular injection for the treatment of CMV retinitis, received a very narrow label and was discontinued after a few years due to poor sales. In May of 2012, ISIS's share price stood at $9.89/share, corresponding to a 21-year return of -1.1%.
Figure 1. ISIS Share Price history
Since May of 2012, ISIS's share price has more than doubled from $9.89 to $21.17 (as of this writing on June 20). A string of successes that is remarkable from a historical point of view has created tremendous investor optimism regarding future revenue potential. These include:
- The announcement of encouraging biomarker data from Phase I trials of three liver-targeted antisense oligionucleotide (ASO) drugs for the treatment of diabetes in December 2012
- The approval of Kynamro, an ASO targeting apoB-100, for the treatment of homozygous familial hypercholesterolemia in January 2013
- The announcement of positive efficacy data from a Phase I trial of ISIS-SMNRX, an ASO designed to increase levels of the of the missing protein SMN in children with spinal muscular atrophy (SMA) in March 2013
- The announcement of favorable efficacy data from a Phase I trial of ISIS-CRPRX, an ASO designed to reduce C-reactive protein levels as a potential therapy for cardiovascular disease, rheumatoid arthritis, and multiple myeloma in March 2013
- The announcement of several partial responses in patients with treatment refractory non-Hodgkins lymphoma in a Phase I trial of ISIS-STAT2Rx in June 2013
ISIS is on a roll. Is it a blip or a harbinger of yet more good news to come? To address this question we tapped the BioMedTracker Database produced by Sagient Research Systems to better understand the history, gain some insights into its current success, and predict the likelihood of success of its current pipeline. A more complete version of this report is available through BioMedTracker.
Analysis of Trial Success Rates Using the BioMedTracker Database
We began our investigation by using the BioMedTracker database to determine the clinical trial success rates for ASO clinical development programs and compare them to those for the universe of all 12,796 drugs in the database. Our search turned up 93 antisense drugs currently or formerly in development for 161 different indications. We used this list, trial results, and other available data in BioMedTracker to assess the factors associated with success in antisense drug development. We then evaluated ISIS's current pipeline to see if we could identify a trend toward greater incorporation of these success factors. In short, we asked whether ISIS was getting smarter or had simply had a string of good luck.
In order to avoid individually evaluating over 250 clinical trial results to assess them as "successful" or "not successful," for our first pass analysis we defined development programs as successful in Phase II only if the sponsor initiated a Phase III trial. In Phase III success was defined as the filing of an NDA, and at the NDA stage success was defined as FDA approval. Suspension was defined as an explicit announcement by the sponsor that development was discontinued, or by the disappearance of the compound from the company pipeline. All other programs were designated as unresolved, and excluded from the analysis.
Historical Failure Rates are Unusually High in ASO Programs
Compared to all drug development programs, ASO development programs have lower success rates in all stages examined (Figure 1). On average, a Phase II ASO program has a 5% likelihood of approval compared to 15% for development programs for all drugs.
Figure 2. Stage Success Rates and Phase II Likelihood of Approval for Antisense and All Drugs in the BioMedTracker Database
Antisense Drugs Have Trouble Getting Into Cells
One important difference between antisense drugs and traditional small molecule drugs is that antisense drugs are very large and have a large negative charge. While traditional small molecule drugs typically contain fewer than 60 atoms and is uncharged, a typical antisense drug contains 700 atoms and carries a negative charge of -18 to -20. Traditional small molecule drugs can enter cells by passive diffusion across the cell membrane. Conversely, antisense drugs enter by a more complex and much less efficient process called endocytosis. This suggests that an antisense drug may need to be present in the tissues at higher concentrations compared to a traditional small molecule drug to be equally effective.
Antisense Drugs Distribute Unevenly in the Body
Because of their high molecular weight, antisense drugs also tend to distribute unevenly in the body, with a tendency to accumulate in the liver and the proximal tubules of the kidney, and to a lesser extent in the skin, bone marrow, muscle and intestines. Low concentrations are found in other organs. We note that other studies have found somewhat different distributions, but all such studies we are aware of consistently report high levels in the liver, spleen, and proximal tubules of the kidney. Qualitatively similar but less detailed results have been observed in the clinic.
Figure 3. Distribution of Radiolabeled Antisense Drug ISIS-3082 in a Mouse
Based on this data, we decided to evaluate whether the success of antisense programs correlated with the organ being targeted. Our initial analysis focused on the results of Phase II programs only, because the number of examples of Phase III programs is small and because Phase I programs are not consistently reported.
High Success Rates are Found for Antisense Development Programs that Target the Liver
Examination of Phase II trial data from BioMedTracker gave the results shown in Figure 2.
- Two of four Phase II development programs targeting the liver were successful by our criteria of Phase III trial initiation representing success and drug suspension representing failure. Closer examination of the data from these trials showed that one of the two suspended liver-targeted ASOs was suspended for economic reasons unrelated to efficacy. Examination of Phase II development programs for which advancement/suspension decisions haven't been made yet identified one additional compound with compelling evidence of efficacy in Phase II but no additional examples of ASOs that failed to demonstrate efficacy. Overall, efficacy was observed in four of five Phase II development programs targeting the liver and for which trial results were available
- A high success rate was found for antisense programs targeting the blood/bone marrow, but on closer inspection, all four of these programs involved the same drug, so this result is less compelling
- Three out of six programs (50%) targeting the skin, muscle, or GI tract were successful. This result is also in accord with expectations based on preclinical distribution data
- Only one program out of 15 targeting prostate, breast, or other modestly perfused organs was successful. This result is expected based on the modest distribution to these organs seen pre-clinically
- No program targeting the kidney was successful despite a relatively high level of distribution. This may reflect the fact that ASO accumulation is restricted to a specific region of this organ (the proximal tubules)
- ASOs targeting the lung were fairly successful as measured by our advancement/suspension criteria. This result was unexpected based on the data in Figure 1. Review of the lung targeted trials revealed that the six Phase II failures included both asthma and lung cancer trials. The three "successful" trials were all single-arm lung cancer trials. Two of these compounds have since been suspended from development in Phase III due to lack of efficacy. Phase III results for the third are still pending
- Two programs targeting the eye and one targeting the CNS were successful. On inspection of the trial protocols, all three of these programs were found to involve local injection of the ASO directly into the target tissue
While this approach has important limitations (small number of examples, varying company standards for advancing compounds into Phase III), the data seems mostly supportive of our hypothesis that ASO development program success is higher when targeting tissues where ASOs naturally accumulate in high concentrations. Specifically, programs targeting the liver appear to have an above average chance at success.
Figure 4. Phase II Success Rates for Antisense Drug Development by Target Organ
Local Administration of the Antisense Drug is Also Associated with High Success Rates
Following up on the observations above regarding local administration to the eye and CNS, we found that three of five programs (60%) involving local administration of an antisense drug progressed to Phase III, compared to 14 of 68 (21%) of programs involving systemically administered ASO drugs.
Examination of ASO development programs that could not yet be characterized as having been advanced or suspended revealed four additional Phase II programs for which efficacy data had been disclosed. Three of these demonstrated compelling evidence of efficacy (ISIS-SMNRx in spinal muscular dystrophy, Alicaforsen enema in pouchitis, and PF-06473871 in wound healing). A fourth program (ASM8 in asthma) provided only equivocal evidence for efficacy and was judged likely to be suspended. This analysis suggests an even higher success rate for Phase II programs involving locally administered antisense drugs. The success rate calculated using our suspension/Phase III trial initiation criteria is compared to that for systemically administered antisense drugs in Figure 5.
Figure 5. Trial Phase II Success Rates for Systemically and Locally Administered ASOs as Measured by Progression/Suspension Criteria
ISIS's Pipeline is Increasingly Focused on ASOs that are Locally Administered or Which Target the Liver
Our analysis above suggests that antisense programs are most likely to be successful if 1) the drug targets the liver or 2) the ASO is injected directly into the target tissue (local administration).
Figure 6 shows the fraction of ISIS internal and partnered drug development programs targeting the liver or employing local administration from 2003-2013. Over this period, liver targeted and local administered ASOs increased from 12% to 41% of ISIS's clinical development programs.
Figure 6. Proportion of ISIS Internal and Partnered Drug Development Programs Targeting the Liver or Employing Local Administration, 2003-2013
An Overview of ISIS Pipeline Suggests More Successes to Come
Recent clinical successes reported by ISIS are recapped in Table 1. We note that the recent doubling of the company's share price has been largely coupled with clinical successes that appear to be due to the company's increasing focus on therapeutic programs in which local administration or the pursuit of biochemical targets located in the liver helps ensure that sufficient amounts of drug are getting into the cells.
Table 1. Some Recent Clinical Successes Reported By ISIS
Table 2 shows a timeline of upcoming ISIS catalysts taken from the BioMedTracker database. While we defer a detailed discussion of individual drugs to a future article, we note the following.
- Four of the seven drugs in Table 1 are liver-targeted and one is administered by local injection
- All four liver-targeted drugs and the single locally administered drug have demonstrated effects on biomarkers in Phase I, thus providing evidence that they knock down their target mRNAs. This does not by any means guarantee success in Phase II, but these drugs have already surmounted a hurdle that has played a major role in the failure of many previous antisense drugs
- Custirsen and ISIS-EIF4ERx each have upcoming catalysts in both prostate and lung cancer. Custirsen is in Phase III and ISIS-EIF4ERx is in Phase II for these indications
- The preclinical data show modest accumulation of ASOs in prostate tissue. A Phase I trial of custirsen showed 92% knockdown of its RNA target in prostate biopsy samples, but protein knockdown appeared less complete. The Phase II efficacy data for custirsen in prostate cancer was equivocal.
- The preclinical data show modest accumulation of ASOs in lung as well, though the organ has very high blood flow that could partially compensate for this. Historically, Phase II success rates for ASO programs targeting lung are reasonably high. The Phase II data from custirsen's single arm lung cancer trial suggested activity comparable to recently approved drugs.
- Neither target is clinically validated.
- Overall we are concerned with prior failures in prostate trials and ambivalent to mildly positive regarding the two lung cancer trials.
Table 2. Upcoming Catalysts from the BioMedTracker Database
Polar bears don't eat penguins because they live at opposite ends of the earth. Antisense drug developers took many years to learn that ASO drug discovery programs are more likely to be successful if the target RNA is located in tissues in which ASOs naturally accumulate. Among these, the liver stands out as an organ with a large number of therapeutically relevant targets. Efficacy also appears to be associated with local drug administration, which can provide very high drug levels in the target tissue.
The recent successes of ISIS in the clinic appear to be based on the company's increasing focus on targets found in the liver, where antisense drugs naturally accumulate. Additionally, ISIS and its partners are pursuing multiple programs in which high concentrations of the ASO in target tissues is guaranteed by local administration.
Lastly, we caution that this article has focused entirely on the criteria for demonstrating efficacy, but safety is another important criterion for drug approval. Antisense drugs exhibit a number of side effects in clinical trials including injection site reactions, thrombocytopenia, activation of the alternate complement system, and prolongation of the activated partial thromboplastin time, each of which appears related to non-specific binding effects. In each indication, regulatory and commercial success will be determined by the risk/benefit ratio of treatment, as well as by the safety and efficacy of other available treatments.