Uncovering the
hidden errors
that drive
neurodegeneration

We study both rare mutations that run in families and common variants that modify disease risk.

Loss-of-function mutations in progranulin (GRN) are a common cause of inherited FTD, and common variants in TMEM106B strongly influence who develops the disease. How these two genes connect was unknown. We found that they converge on a single molecule: GRN loss leads to buildup of C-terminal fragments (CTFs) of TMEM106B in lysosomes, and the protective haplotype of TMEM106B keeps CTF levels low (Zeng et al., 2026 (opens in new tab)). CTF levels may be a disease driver we can measure and target.

In most FTD and ALS, the RNA-binding protein TDP-43 leaves the nucleus and forms aggregates in the cytoplasm. We ask which errors from its loss matter for neurons and how they propagate to other cells.

In the nucleus, TDP-43 controls how thousands of RNAs are processed. We found that TDP-43 loss changes where many RNAs end (alternative polyadenylation, APA), impacting expression of disease-relevant genes such as VPS35 (Zeng et al., 2025 (opens in new tab); Maheswari Jawahar, Zeng et al., 2026 (opens in new tab)). We also found that nonsense-mediated decay (NMD) hides thousands of cryptic splicing errors caused by TDP-43 loss (Zeng, Sianto et al., 2025 (opens in new tab)). Bringing these hidden errors to light points to new disease drivers and drug targets.

If we know the mistake, can we fix it?

Each mechanism we define points to a way to fix it. We are exploring ways to promote degradation of TMEM106B CTFs, and designing antisense oligonucleotides (ASOs), short synthetic strands that bind RNA, to correct cryptic splicing and APA changes, with the goal of repairing defects in the cells that need it.