Probiodrug unveils encouraging results of phase 2a SAPHIR study in AD patients

The SAPHIR study is the first clinical trial to investigate the Glutaminylcyclase (QC) inhibitor PQ912 in patients with early AD over a treatment period of 12 weeks.

The highest dose of 800mg bid PQ912 used in the Phase 1 multiple dose study and showing a very high target occupancy was compared to placebo to identify early efficacy and safety signals to optimally plan future long-term dose ranging studies.

The primary objective of the SAPHIR study was to investigate the safety and the tolerability of PQ912. Secondary objectives were to assess early effects of PQ912 on the exploratory endpoints NTB, EEG, cerebro-spinal fluid (CSF) biomarkers related to the QC-inhibition (mechanism of action) and to the AD pathology.

Methodology: The SAPHIR study is a double-blind, placebo controlled randomised study carried out in 7 European countries at 21 study sites in treatment naïve patients with early AD disease. The statistical analysis of the endpoints was based on two-sided tests with a significance level of alpha 0.05 and not corrected for multiplicity. Results presented in this press release are for the Intention to treat (ITT) population based on a ‘complete case’ analysis.

Results: A total of 120 patients were randomised in the SAPHIR study, 60 to the placebo arm and 60 to PQ912 arm. Treatment arms were well balanced with respect to age, gender, disease severity and APOE4 status. The mean MMSE (Mini-Mental State Examination) score at baseline was 25.5 (min-max 21-30).

Safety and tolerability results (primary objective): There were no statistically significant differences of PQ912 vs placebo between the number of patients experiencing an adverse event (PQ912 n=49, placebo n=45) or the number of patients with a serious adverse event (PQ912 n=8; placebo n=5).

Patients in the treatment arm did show a significantly higher discontinuation rate due to SAE or grade 3 adverse events compared to patients in the placebo arm (PQ912 n=6; placebo n=0, p=0.027) and the total number of patients non-adherent to randomised treatment for any reason was higher in the treatment arm (PQ912 n=26; placebo n= 2; p<0.01).

Skin and gastrointestinal organ system related adverse events were observed in a higher frequency in the PQ912 arm compared to placebo and occurred in the majority in the first half of the treatment period. Dose reductions prescribed by the investigator were identical in the treatment and the placebo arm (both n=5).

Results of the secondary exploratory endpoints:

Molecular biomarkers in the CSF: CSF analyses showed a highly significant QC inhibition (p=0.001), corresponding to a calculated target occupancy of 92% (median), which was achieved in patients with last drug intake within 24 hours before CSF sampling. A decrease in pGlu-Abeta oligomers in the CSF was observed in the treatment arm whereas pGlu-Abeta oligomers increased in the placebo arm.

This directional change demonstrates, together with the significant QC-enzyme inhibition, a strong and robust target engagement.

There was a strong trend for reduction in the level of neurogranin, a marker of synaptic dysfunction in the ITT population in the treatment arm compared to placebo (p=0.1), which became significant if 3 patients starting prohibited concomitant medication during the study were excluded (p=0.046, a 5% absolute reduction of baseline in neurogranin observed in the treatment arm).

There was also a strong trend in the mean reduction of YKL 40, a biomarker of inflammation, in the PQ912 arm compared to placebo (p= 0.07, 5% absolute reduction of baseline–level in the treatment arm).

EEG: The analysis of the EEG power spectra showed a significant reduction of theta power in the PQ912 arm compared to placebo (p=0.002). Slow wave theta activity is reported to increase with the onset and progression of AD. Further analysis of functional connectivity and EEG network parameters is pending.

Neuropsychological test battery (NTB): Patients in the placebo arm showed overall a stable performance with no or marginal change between baseline and week 12. Performance on the ‘One Card Back Test’, an assessment of working memory showed a statistically significant effect in favour of PQ912 (p=0.05, Cohen’s d = 0.24) while the ‘Detection Test’ an assessment of attention, also showed a meaningful improvement under PQ912 (Cohen’s d=0.20) although not sufficient to reach statistical significance.  Performance on the five other cognitive assessments, as well as on their aggregate scores, were not influenced by treatment with PQ912 for 12 weeks (Cohen`s d < 0.2).

Additional analysis comprising all endpoints, further CSF biomarker and subpopulations will continue during the next several months and the full results of the SAPHIR study are intended to be reported at scientific congresses and published in scientific journals.

Philip Scheltens, Director of the Alzheimer Center VU University Medical Center Amsterdam, and Principal Investigator of the SAPHIR study commented: "The population studied was very representative of patients with early AD. The primary objective of the SAPHIR study was safety and tolerance. Although differences between treatment arms were observed we are confident that the drug is safe and well tolerated in the AD population.

"We set out to detect small signals on the various sensitive secondary exploratory outcome measures in a relatively short time frame and were happy to see very strong target engagement, significant effects on a working memory task and EEG theta power and encouraging results in the right direction on synaptic and inflammatory CSF markers. These results point to a direct effect on pGlu-Abeta with beneficial effects on synaptic function, even in such a short treatment period."

Inge Lues, Chief Development Officer at Probiodrug commented: "We are very positively encouraged by the outcome of the SAPHIR trial and look forward to the results of further analyses. The results observed support our hypothesis of pGlu-Abeta being synaptotoxic.

"The SAPHIR study was designed to guide the future development of PQ912 in AD. In this first in-patient study, we used a high dose achieving about 90% QC occupancy to meet two goals: to get a first picture of frequency and types of safety and tolerability events and at the same time to test for early signals of efficacy in a relatively short treatment period of 12 weeks."

The results will guide the design of future studies of PQ912 and support the use of lower doses of PQ912 still reaching relevant target occupancy of the QC enzyme, as well as the option to introduce a titration phase for the 800 mg bid dose. We are encouraged by having observed early indicators of response to PQ912 in the CSF, the EEG and the NTB.

The results we have today in our hand are tremendously useful to deliver a tailored future development program.