Interpreting low-dose Cohort 1 data as functionally active strengthens Phase 1b results and bodes well for Phase 2 data.
The premise of this investment thesis is that a low-dose placebo effect for a gene-therapy asset held by Voyager Therapeutics (NASDAQ:VYGR), a biotech company focused on brain disorders, is instead a true low-dose active effect. This further strengthens the possibility of positive Phase 2 results release, which would cause market appreciation of the stock in 2020.
Investing in early-stage biotechnology companies is both complicated, due to the complex nature of the assets in development (both technical and scientific), and risky, due to the high rate of clinical trial failure. Investing in Parkinson’s Disease ((NYSE:PD) gene-therapy is further discouraged by recent Phase 2 and 3 clinical development failures, due to lack of efficacy, which have set a “bearish” investor sentiment regarding this specific sub-sector of biotechnology assets. Voyager’s slightly less ambitious goal, to first tackle motor symptoms by focusing on genetically “boosting” the dopamine synthesis pathway, seems more achievable: In Phase 1b data, 3 out of 4 patient cohorts show a clear improvement and up to 1-year delay in disease progression. A confounding factor is the possibility that there is a large magnitude placebo effect – observed in the lowest-dose cohort which failed to show increased dopamine synthesis (one of the many biomarkers used).
Contrary to the original interpretation of low-dose Cohort 1 long-term efficiency as a “placebo” effect, I instead believe that motor improvements in Cohort 1 derive from a low-dose active effect, which further strengthens the argument for a positive Phase 2 readout. This is supported by 3 observations: 1) Cohort 1 has a baseline offset that prevents readout for improvement during ON-medication evaluation, 2) Cohort 1 dopamine synthesizing activity was measured at 6-months but improvements in OFF-medication response were registered at 12-months, which means activity might have increased in between and 3) In a similar trial (GT-AADC from PTC therapeutics) half the dosage used in Cohort 1 sufficed to motor improvement.
Since VY-AADC02 the Phase 2 trial design is largely based off the Phase 1b trial early success, I predict that the likely Phase 2 positive outcome will make the stock outperform. Voyager is ahead of competitors in clinical development and will likely be the first company to successfully market a gene-therapy for PD, which albeit not a cure, will durably improve motor symptoms and delay disease progression.
Voyager sails to deliver Parkinson’s Disease first functional gene-therapy controlling dyskinesia and lowering levodopa requirements for up to 3 years
Voyager Therapeutics is a clinical-stage biotechnology company that targets severely debilitating and lethal neurological diseases, such as Parkinson’s and Huntington’s. Voyager’s platform comprises a multi-modal adeno-associated viral (AAV) vector gene therapy which can replenish, deplete or block target disease causing proteins. The company also designs therapy specific injection methods to target key regions of the central nervous system (CNS) to achieve higher penetrance and efficacy of the associated gene therapy. I will focus on VY-AADC02, targeting Parkinson’s disease.
Voyager’s PD early clinical development success, VY-AADC02, relies on replenishing Aromatic l-Amino acid DeCarboxylase (AADC), an enzyme that converts Levodopa into Dopamine, the neurotransmitter that coordinates motor function (see Exhibit 1). Increased AADC activity (up to 2-fold compared to patient baseline, close to healthy subjects levels) was achieved locally and with high coverage (54%) in the putamen region of the brain through a combination of 1) a safe and neuron-specific AAV delivery vector (follow this link for the original publication), 2) a genetic construct carrying the human cDNA of AADC, resulting in robust expression of AADC to high levels, and 3) a short duration surgical frontal brain injection technique into the putamen. This is a one-shot gene-therapy whose functional benefits can extend for a long time period (up to 15 years in non-human primates) and 3 years, but possibly longer in humans, see Exhibit 2).
Exhibit 1 – Mechanism of action of VY-AADC02
The gene therapy expresses AADC in the neurons, allowing the more efficient conversion of L-dopa into dopamine and restoring synaptic function and motor control. This reduces dyskinesia and maintains sensitivity to low-dose Levodopa, the standard-of-care for PD patients. Adapted from Voyager’s August 2019 and Axovant’s June 2019 corporate presentations.
Voyager’s VY-AADC02 efficacy is further supported by Axovant’s Axo-Lenti-PD and PTC Therapeutics GT-AADC preliminary clinical data:
Unlike previous gene-therapy clinical failures using similar approaches (ProSavin, a lentiviral vector approach being one of the most recent examples) VY-AADC02, limited by the carrying capacity of the AAV vector to a single gene (human cDNA of AADC), achieved functional motor improvement with a delay in disease progression over 12 months. The VY-AADC02 single-shot gene therapy is safe and, unlike lenti-viral therapies, does not integrate into the target cell genome therefore having low carcinogenic potential compared to integrating vectors. A direct comparison with competing drugs is presented in Exhibit 2.
One of VY-AADC02 most direct competitors, Axovants’s Axo-Lenti-PD (formerly OxPharma’s OXB-102) uses larger viral particles and expresses 3 genes in the Dopamine synthesis pathway and is a re-engineered ProSavin, where AADC was placed closer to a promoter, possibly to increase expression of this important enzyme. While a large vector and a complex genetic construct might lead to similar complications which ProSavin faced, one of two patients clearly showed improvement in motor function (see Exhibit 6A). The trial design for the clinical advancement of the VY-AADC02 and its direct competitor Axo-Lenti-PD, are similar in terms of patient population and primary and secondary endpoints. However, VY-AADC02 has been tested in a larger patient population, defined its therapeutic dosage and has biomarker data available, placing it at least 12 months ahead in development relative to Axo-Lenti-PD.
PTC Therapeutics is developing GT-AADC (AAV2-hAADC), an adeno-associated virus carrying human cDNA of AADC, very similar to VY-AADC02, to treat hereditary AADC deficiency which causes early motor development complications (Clinical Trial references: NCT01395641, NCT02926066, NCT02852213). The low dose single administration of AAV2-hAADC, which is lower than the cohort 1 dosage in the VY-AADC02 Phase 1b studies (see Exhibit 2), sufficed to achieve functional efficacy. This result suggests that the VY-AADC02 cohort 1 dosage, initially thought to be below therapeutic levels is also partially efficient, as I will discuss in more detail below. GT-AADC was shown to be efficient in 20 out of 26 children (average age 4±2 year old): patients responded well to treatment showing significant and durable (12-60 months) motor development improvement (see Exhibit 4). The efficiency of treatment, for a constant dosage injected in the GT-AADC trial, is inversely correlated with the age and size of the putamen: 2 year-old children respond better than older patients, which suggests that relative dosage and brain development stage influence therapy outcome.
Exhibit 2 – Table comparing three gene therapies that use AADC as a vector.
Note: Find clinical trial information for these assets here: VY-AADC02, Axo-Lenti-PD and GT-AADC. vg – viral genomes (non-integrating), and tu – transducing units (integrating). Although vg and tu cannot be directly compared, it is expected that the number of viral integration events is 2 orders of magnitude lower than the number of vector genomes. Putamen coverage refers to the volume of putamen region of the brain targeted with MRI guided injection. Axo-Lenti-PD and GT-AADC coverage was not available and was estimated based on vector characteristics. Data adapted from publications and corporate presentations (Voyager, Axovant and PTC).
VY-AADC02 shows robust functional motor improvements in pre-clinical and clinical Phase 1b trials, reverting dopamine depletion by targeting its genetic cause
VYGR VY-AADC02 could become the first gene therapy treatment for Parkinson’s disease. The etiology of the disease is complex, and it has numerous genetic and environmental factors, being diagnosed at a later age (60 years old). What is consistent is the depletion of dopamine, which causes mild tremors to loss of motor function (dyskinesia), and the disease also progresses towards cognitive and psychological dysfunction (see Exhibit 3). This can be countered by injecting dopamine’s precursor levodopa, but over time low dosages lose efficiency and patients suffer from secondary effects of higher levodopa intake. Rescuing AADC expression, the enzyme that synthesizes dopamine, is likely to maintain an efficient response to low dose levodopa, controlling PD motor symptoms and preventing side effects of high levodopa medication.
Exhibit 3 - Clinical manifestations of PD before and after VY-AADC02 treatment.
(Dyskinesia refers to severe problems with motor coordination.)
Summary of Pre-Clinical and Clinical results:
- Pre-clinical studies in primates show long-term expression of AADC (up to 15 years) result in motor function improvement (original data).
- Safe putaminal VY-AADC02 injection in humans, with 30-40% coverage of putamen, posterior trajectory injection method has a shorter surgery duration and showed 54% putaminal coverage.
- Increased AADC enzyme activity detected and F-Dopa uptake, detected by PET-imaging which lead to a decrease in Levodopa intake around 40%.
- Disease progression is delayed at 12 months* (from stage 3 to stage 2 category mH&Y) both for regular injection (PD-1101) and posterior trajectory injection (PD-1102).
- 50% reduction in dyskinesia (less 3h of dyskinesia per day) and reduction of on-medication time (7h less per day).
- Decrease of 9 points in the PDQ-39 PD severity test (self-reporting test scores show the equivalent to a 25% increase in quality of life).
*Note: 18-month assessment of 8 patients in 4Q2019, for posterior trajectory injection, positive at 12 months, reported May 2019.)
Scientific, clinical and commercial investment aspects of VY-AADC02 based on 12-month data
I believe VYGR management is developing the right asset, VY-AADC02, to conquer a very profitable market which will overcome the short-duration response and side-effects of levodopa therapy for Parkinson’s disease. VY-AADC02 open-label Ph.1b data interpretation: dosage escalation from cohort 1 to cohort 4 (frontal injection) correlates with motor symptom and quality of life improvement.
VY-AADC02 replenishes AADC, helping to increase the levels of dopamine in the putamen, a brain region responsible for motor control (see Exhibits 4 and 5). The significant putamen coverage due to the AAV vector and injection methods, the high AADC expression of the transgene, and most importantly the alleviation of disease symptoms and improved the quality of life, coupled with a lowering of levodopa medication dosage suggest the drug is efficient.
VYGR phase 2/3 trial design is robust, supported by relevant biomarker data and focused on the positive motor function and quality of life assessment outcomes (RESTORE-1). Restore-1 trial will consist of 42 randomized, double-blind, placebo Surgery controlled patients. Patients diagnosed with PD >4 years showing treatment resistant dyskinesia will be eligible, which should accelerate the recruitment in the >250 medical centers across the US. The primary endpoint will be time without dyskinesia over 12 months, the secondary endpoints are other motor function and quality of life measurements (UPDRS scores, PDQ-39 questionnaire and CGI). The trial will also measure non-motor symptoms and safety. Biomarker data regarding VY-AADC02 putaminal coverage, AADC activity and levodopa dosage changes will be acquired.
Exhibit 4 –Clinical data describing the efficiency of VY-AADC02 (n=15 patients).
A. Hours without dyskinesia
B. Improvement in quality of life (score reduction means improvement)
C. Disease progression (mH&Y stages)
Cohort 2 and 3 have higher dosage of VY-AADC02 which corresponds to higher AADC enzyme activity, as determined by PET studies (adequate biomarker assessment measuring F-Dopamine uptake). Data adapted from Voyager’s corporate presentation.
Key risks and difficulties in interpretation
Although the Phase 1b results of VY-AADC02 show functional improvements and delayed disease onset, it is unclear if the Phase 2 results will support the initial findings. Previous attempts at treating PD using AADC have failed and trials were discontinued (ProSavin), most likely due to low putaminal coverage or enzyme expression. Now it seems that the new posterior injection method, smaller AAV viral particle size compared to lenti-viral vectors, non-integrative expression and a tailored AADC transgene design together achieve both higher putamen coverage and cellular expression, supporting a positive phase 2 outcome. Similar results have been obtained for the same indication by Axo-Lenti-PD in one out of two patients (Exhibit 6). Also, motor development improved in AADC deficient children treated with GT-AADC (nearly identical in design to VY-AADC02, 20 out of 26 patients).
VY-AADC02 cohort 1 is highly responsive to medication (Levodopa) when compared to cohorts 2 and 3, which might account for the observed absence of improvement in the on-state.
A confounding factor is cohort 1 where AADC activity measured at 6 months was much lower than in the other cohorts, correlating with low coverage. Nonetheless, at 12 months, where no measurements of coverage or activity are available, a strong “placebo effect response” appears for the non-active dose cohort 1. I think this result is likely due to an increase in AADC activity, and possibly better putaminal coverage, from month 6 to 12. Because the OFF-medication and time without dyskinesia improvements in cohort 1 was similar in magnitude as higher dosage cohorts 2-4, efficiency of VY-AADC02 was wrongly attributed to being a placebo.
Cohort 1 had an unusually high ON-medication baseline response (7.6 vs 17.0 in symptom severity scale), which prevents a direct comparison with other cohorts. It is surprising to see that in Table 2 of Chadwick et al, 2019 cohort 1 has a much better baseline response to medication on-time (by 9 points) than the two other patient groups. These fluctuations among patient groups can occur when the sample size is low (n=5 patients per cohort) but curiously it was only registered for this parameter, all the other parameters have no statistically significant differences. This optimality in levodopa response is likely the cause of the failure in the improvement of quality of life over time registered in cohort 1 (e.g. if a student obtains a 98% or an A in the school exam it will be hard to improve over time, see Exhibit 5A).
The fact that there was a “ceiling for improvement” due to cohort 1 being already optimal from baseline has launched skepticism and misinterpretations of the data. Nonetheless there was an improvement in the off-medication sate, at 6 months, and an improvement for 1.6h in time without dyskinesia at 12 months, which I believe to be the consequence of a low dose effect: In my opinion, the data actually supports improvement in cohort 1 which is not due to placebo effect but to an effect of a low dosage in the medication off-state (Exhibit 5B).
This is supported by another study in 26 child patients from GT-AADC (PTC therapeutics) which received a similar magnitude of viral dosage (cohort 1= 8.30x1011 vector genomes versus GT-AADC = 2.37x1011 vector genomes), in the same region of the brain. Considering that the adult brain putamen volume is 2x larger than the infantile brain (see Exhibit 2, note that putamen volume range varies widely from 1-8 years of age (Giedd, Snell et al. 1996, Choe, Ortiz-Mantilla et al. 2013)), and that coverage was lower in the GT-AADC trial (lower vector dosage and injection was not MRI guided), the approximate effective loading of viral genomes should be similar, if not twice as high for cohort 1 in VY-AADC, and result in comparable AADC activity levels in both studies. This comparison is nonetheless challenging due to the low numbers of patients tested in both studies but shows significant and longitudinal motor improvement, considering off-state data. The robust reduction in Levodopa dosage from cohort 1 to 4 also suggest that the lower dose has activity. Again, curiously, in cohort 1 a single patient had to receive a supratherapeutic dosage (4500mg/day) which shifted the average reduction towards an increase in Levodopa requirements for cohort 1.
Exhibit 5 –PD motor symptoms comparison in VY-AADC treated patients (n=5 patients/cohort, Phase 1b).
A. Change during on-medication stage. Cohort 1 was already optimal leaving little room for improvement based on therapy (see red outline on A and C).
B. Change during off-medication stage improved significantly, suggesting a low-dose effect, which was expected by comparison with the efficiency of response with 50% relative dosage in the GT-AADC trial data (see Exhibit 6B).
C. Patient cohort parameters. Cohort 1 has an unusually high response to Levodopa compared to cohorts 2 and 3.UPDRS – Unified Parkinson’s Disease Rating Scale (symptom quantification system). Data was adapted from Voyager’s corporate presentation and Chadwick et al, 2019.
Exhibit 6 – Motor function improvements in two other AADC brain gene-therapy trials:
A. Axovant’s Axo-Lenti-PD (n=2 patients for up to 6 months
B. PTC Therapeutics GT-AADC developmental motor scale in children with severe AADC deficiency (n=18 for 12-month study and n=8 patients for 60-month study). UPDRS and PDMS are standard and comparable Parkinson’s motor symptoms score systems. Viral dosage was 50% when compared to VY-AADC02 cohort 1, estimated by lower vector number and lower injection efficiency used in GT-AADC and normalized by brain volume size (children vs adults). Data was adapted from corporate Axovant’s and PTC therapeutics corporate presentations.
Exhibit 7 – Financial modeling of Voyager based on VY-AADC02 lead candidate success:
A. Discounted Cash Flow analysis for USA and EU, based on a 50/50 deal between Voyager and Neurocrine for the US, and 5% royalties in Europe. Conservative assumptions on pricing due to increased regulatory pressure, cost of PD palliative care and patient pool size. Treatment capacity was modeled conservatively as well (maximum 10,000 treatments per year at 100% market penetrance over 10 years).
B. Model assumptions tabulation
C. Risk-adjusted Net Present Valuation (rNPV) for VY-AADC02 asset as being Voyager’s leading asset and the major contributor to value (other assets in development excluded from this analysis for simplicity).
What investors need to reflect upon regarding Voyager and VY-AADC02:
- Voyager’s lead asset, VY-AADC02 has shown remarkable improvements in motor function, robust for over 24 months in 4 cohorts of patients in Phase 1b trials (n=15 patients for PD-1101 and n=8 patients for PD-1102) while Axo-Lenti-PD from Axovant, is convincingly efficient in only one out of two patients, followed only for 6 months.
- PTC Therapeutics data show that GT-AADC, similar to VY-AADC02 in design, delivery and targeting the same brain region, efficiently restores motor development in pediatric AADC deficiency (patients followed up to 60 months).
- Voyager’s posterior trajectory surgery (PD-1102) reduces the standard stereotatic injection operative time (PD-1101, Axo-Lenti-PD and GT-AADC) and complications, increasing coverage of the putamen (see Exhibit 2).
- The quality of life and disease progression measurements for VY-AADC02 show a 1-year delay in PD progression (from stage 3 to stage 2 of mH&Y scale). Quality of life improvement was similarly observed for GT-AADC and for one patient in Axo-Lenti-PD, but changes in PD progression for Axo-Lenti-PD were not measured.
- Axovant’s Axo-Lenti-PD and PTC Therapeutic’s GT-AADC show preliminary data that supports functional success in AADC gene therapy targeting the putamen, comparable across the three studies using PDMS or UPDRS symptom scales. For Axo-Lenti-PD I estimate a developmental time delay of at least 12 months, and GT-AADC targets a different and smaller patient population (AADC deficiency is a rare disease with early in life versus late in life onset).
- While gene-therapies such as VY-AADC02 and Axo-Lenti-PD are not a cure for PD, and are tainted by the skepticism in the PD gene-therapy field (ProSavin), VY-AADC02 attenuates motor symptoms and delays PD progression.
- I believe the wrong interpretation of low-dosage VY-AADC02 cohort 1 efficiency as a “placebo effect” casted doubts on active drug-derived efficacy in higher-dosage cohorts 2 to 4. This resulted in a “bearish” response from investors.
- Backed by the fact that even lower dosages, and possibly lower putaminal coverage, in GT-AADC show efficiency in children, it is likely that VY-AADC02 cohort 1 low dosage effects are active non-placebo. This represents a contrary investment opportunity for the bearish investor sentiment Voyager currently faces.
- I estimate that Voyager stock may be currently under-valued, especially since its leading asset, VY-AADC02, is ahead of competitors in clinical development. Upon successful Phase 2 and 3 trial completion, the stock will outperform: I predict $65/stock in 2022, compared to the current valuation of $17/stock (see Exhibit 7C).
A bullish view on Voyager, and Parkinson's Disease asset VY-AADC02, which will likely proceed to Phase 3
In light of my observations, investors have a bullish opportunity to invest in Voyager. VY-AADC02 is the best asset in clinical development to show efficacy in controlling PD motor symptoms. The preliminary success of similar assets GT-AADC and Axo-Lenti-PD from competitor companies, still in Phase 1 of development, support the viability and efficacy of the approach. Specifically, VY-AADC02 decreases ON-medication (Levodopa) and OFF-medication dyskinesia, but also delays disease progression and extends the lifespan of the standard-of-care, by eliminating the need for a Levodopa cumulative dosage increase during at least 36-months (trial ongoing). Instead of accepting the designation of Cohort 1 as “inactive placebo” or neglecting the contribution of Cohort 1, based on the fact that it is not directly comparable to the other cohorts due to its high baseline Levodopa responsiveness, I found that it actually supports a low-dose effect (see Exhibits 2-6).
I believe VYGR is set to outperform short-term (Phase 2 top-line data release in 2H2020) and might even outperform long-term (pending positive outcomes of Phase 2, Phase 3 and PDUFA). Investors should look-out for long-term Phase 1b data on the frontal injection (Cohort 4) efficacy (I estimate it to be released in November 2019) and topline Phase 2 efficacy data likely to be released in 1H2020. Additionally, FDA approval of VY-AADC02 would set a regulatory precedent favoring the clinical development of gene-therapy assets targeted at neurological disorder.
Disclosure: I am/we are long VYGR. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: Disclaimer: This article was written by Miguel Coelho, Ph.D. member of Beacon Value Partners (BVP), a team of analysts. Mikalai Malinouski, Ph.D. provided critical comments to improve content and clarity. This article contains forward-looking statements and information, including statements regarding the safety, efficacy and potential clinical benefits of VY-AADC02, and the possible Ph.2 data release date in 2020 and the timing of the FDA's review of Voyager's future NDA filing. Acceptance of the NDA filing does not determine the final evaluation of the data submitted in the NDA and is not a guarantee of approval. The author is solely expressing his ideas and opinions and declares no responsibility for any third-party investment decisions regarding the companies listed.