So gene expression changes after A273 treatment clustered into 17 groups using weighted gene co-expression network analysis. That sounds good. Then to allow treatment effect analysis, an eigengene value was calculated for each patient at baseline and 14 weeks. Makes sense.
Genes were clustered together using WGCNA1
, starting from their
correlation of TPM2 expression values across samples: WGCNA created
17 clusters (size: 13 to 11,190 genes)
• The cluster’s eigengene3 was used as summary of expression level of
the cluster
Placebo patients were compared to 30 mg and 50 mg treated cohorts looking at the differences between 14 week eigengenes to baseline. Two of the 17 clusters showed the most change and these 2 clusters include genes that likely play a role in AD and PD. On a qualitative level this sounds very promising.
But then we get this:
Both identified clusters are up-regulated for treated patients compared to
placebo: Cluster 2 eigengene expression was significantly increased for
patients treated with ANAVEX®2-73 high oral dose compared to placebo
(p = 0.021)
So 34 comparisons (17 clusters/eigengenes changes over 14 weeks comparing placebo to 30 mg and 17 eigengene changes comparing placebo to 50 mg).
Of these 1 was significant at p=0.021 (cluster two comparing 50 mg (high) dose to placebo). I assume 3 trended (cluster 2, 30 mg and cluster 1, 30 mg and 50 mg). How unexpected is one 0.021 p value among 34 tests?
The poster data is not negative and definitely thought provoking but doesn't seem near enough to get a pharmacodynamic/response biomarker qualified. Most importantly, there was no data linking changes in cluster 1 and 2 to the actual benefit. This does not move the needle, IMO, to allow the phase 2b/3 to stand on its own for accelerated approval. If a second study is needed we are talking about a 2025/2026 NDA for AD --- all eyes back to Rett and Feb'ish.
A .40-.50 move on the open sounds about right. I'll have a side of fries with that.
