Probiodrug — Update 31 March 2016

The key message during Probiodrug’s FY15 results call was that the Phase IIa trial is on track although the safety/tolerability data (primary endpoints) are delayed to end 2016 (previously mid-2016). Exploratory efficacy results (secondary endpoints) should be announced three to four months later. The delay is due to several challenges, including high competition in enrolling treatment-naïve patients. To help counter these challenges, Probiodrug has taken measures including increasing the number of trial sites, but the company emphasized that the quality of the process remains a priority.

While AD accounts for half of all dementia patients with c 5.2m people affected by AD in Western Europe alone, this is misleading in terms of the patient pool accessible to a specific trial. The first issue is that recently the focus has shifted to clinically established, early stage AD, which is believed to be more susceptible to new treatments. Prodromal stage AD or minimal cognitive impairment (MCI) is also a desirable population for trials, but poses challenges because of screening hurdles as there are no reliable tests to identify those who will convert to clinical AD. As a result, the majority of the ongoing late stage trials try to recruit either early stage AD or those patients who have just been diagnosed. Besides competition access to treatment-naïve patents makes enrolment even more challenging.

The other major issue is retention, ie how many of the enrolled patients actually complete the study, as this affects the intention-to-treat analysis of the trial results. Patients with dementia pose particular problems regarding their retention on study trials, which can be partially solved by study partners – a closely related supervising person, eg spouse or adult children). One overview study published in 2010 analysed 29 clinical trials that investigated patients ranging from MCI to severe AD and found that the percentage of patients who complete the trial varied significantly from 46% to 95%.1 Good retention begins with the enrolment of “quality” patients. As a result, if the trial enrolment is delayed, relaxing the screening criteria in AD patients might seem lucrative, but could be perilous in terms of retention and the quality of the final data (although it might be a good strategy in other indications). Increasing the number of trial sites could potentially be a better strategy in this case. The trade off, however, is higher costs in the short term.

In general, a desirable AD patient profile for the majority of ongoing AD trials is recently clinically validated and early AD, treatment-naïve people with as few co-morbidities as possible. According to World Alzheimer Report 2015 (Alzheimer’s disease International), there were an estimated c 2.6m people newly diagnosed with dementia in North America and Western Europe, of which AD accounts for 60-80% (alzheimer.org) or around 1.8m. Mild AD is estimated to account for around 27% (World Alzheimer Report 2014), which reduces the target population to c 490k. The final size of the accessible patient pool is difficult to estimate, but it could be multiple times less than the figure above, which is due to the fact that most AD patients are older than 75 years of age and are therefore likely to be affected by co-morbidities or are not AD treatment naïve. In addition, geographical distance further limits access to these patients as clinical trials tend to compete for the leading neurology centres.

There are currently 96 active Phase I/II or later stage clinical trials in the US and Europe’s top five countries focused on prodromal or clinical stage AD. In total, these trials are seeking to recruit an impressive 31,818 patients (clinicaltrials.gov). Exhibit 1 shows a breakdown according to stage.

R&D spend in 2015 increased to €10.2m (vs €8.0m in 2014) and was largely in line with our expectations of €10.7m. The increase was mainly a result of initiating and conducting the SAPHIR study, which is fully underway. G&A expenses of €3.3m stayed flat year-on-year and were somewhat below our estimate of €3.6m. Accordingly, the net loss increased to €13.5m in 2015 (vs €11.4m in 2014). Probiodrug indicated that the net loss for 2016 may be larger than that incurred in 2015 due to additional costs related to SAPHIR. We believe this may be primarily related to R&D spend, which we adjusted to €11.1m expected in 2016 compared to our previous estimate of €9.8m. Our G&A estimate for 2016 was adjusted only slightly downwards. We continue to expect a licensing deal in 2017 and PQ912-related R&D spend will be borne by a partner.

Probiodrug reported cash of €21.4m at end 2015, which was boosted by the gross proceeds of €13.5m from a private placement in November. Our post-FY15 model suggests that this should be sufficient to fund operations until around mid-2017. According to the latest Probiodrug’s enrolment projections, the first data from SAPHIR may be reported by the end of 2016, while the final results should be released three to four months later. As per our current estimates, existing funding is sufficient to reach the readout. However, it is uncertain whether the company will be required to repay tax provisions of €2.6m; if it does, the company estimates the cash reach is the beginning of Q117. In either case, the readout will be imminent at that point and, in our view, the future operations of Probiodrug will depend on the data rather than on the short-term liquidity situation.

We continue to expect that if the trial results are positive, a licensing deal in 2017 is likely given the high profile of Probiodrug’s R&D programme among the larger players in the AD field. Although our valuation includes risk-adjusted milestones from a partner for PQ912 that could be triggered by licensing and the start of Phase III in 2017 (more details in our initiation report), our financial forecasts do not include any such income. Hence, our forecasts include €5.5m of illustrative financing included as long-term debt on the balance sheet in 2017. This is based on the cash need that we estimate is required to run the business through to the end of 2017.

Our Probiodrug valuation is increased to €242.3m (from €234m) or €32.6/share due to rolling our model forwards in time, which more than offset the increase in R&D costs. We previously assumed the deal in 2017 and still maintain this assumption given the six-month delay (which increased R&D costs). Our other assumptions for PQ912 are unchanged. While no value is assigned to the preclinical pipeline given its earlier stage of development, it represents potential upside to our current valuation that could be added as the projects progress into clinical stage. The breakdown of our rNPV valuation, which uses a discount rate of 12.5%, is shown in Exhibit 2.

Our partnering assumptions are also maintained and include a 15% royalty on global sales. We keep our assumptions of peak sales of c €6.2bn in 2028, assuming launch in 2022, for the treatment of mild AD only. Our peak sales are based on the assumption that PQ912 proves to be disease-modifying and is therefore able to command a premium price. We assign a 20% probability of success, based on the risks associated with AD drug development.