Back

STARR-CRAAVT: A platform to design cell type-specific regulatory elements for next-generation gene therapy

Becker, R.; Choudhury, P.; Oti, M.; Fouani, Y.; Strobel, B.; Ketterer, S.; Rumpel, S.; Park, J.; Kreuz, S.; Maier, U.; Michelfelder, S.; Schön, C.

2025-06-07 genomics
10.1101/2025.06.04.657008 bioRxiv
Show abstract

Precise control of transgene expression through novel enhancer-promoter combinations is a promising strategy for advancing gene therapy mediated by adeno-associated virus (AAV). We present STARR-CRAAVT, a novel STARR-Seq-derived platform to screen for enhancers in the AAV context. Using HepG2 and HaCaT cells as screening models, we applied STARR-CRAAVT for the identification of cell type-specific enhancers. We integrated epigenetic datasets into an in-silico library of putative HepG2 enhancers and captured corresponding fragments from genomic DNA. The fragments were processed into AAV libraries and applied to HepG2 and HaCaT cells. STARR-CRAAVT analysis revealed a selective activity of the libraries confirming the HepG2-directed in-silico design and the specificity of single enhancer-promoter combinations could be validated using luciferase reporter assays. In addition, we scrutinized the impact of key experimental parameters on enhancer identification and found that the used promoter type had significant influence on the ability of candidates to act as enhancers. Furthermore, switching the location of identified enhancers in reporter assays revealed that the level of enhancer activity is highly dependent on the position in the AAV genome. Taken together, our study yields novel insights into enhancer function and demonstrates that STARR-CRAAVT can be employed to identify cell type-specific enhancers, highlighting promoter preference and enhancer positioning as key considerations for enhancer screening campaigns.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.