"We knew that CIC exists in two forms, CIC-Long (CIC-L) and CIC-Short (CIC-S)," Lee said. "Both ATXN1 and ATXN1L bind to the same section on both CICs. CIC-L and CIC-S differ at one end of the protein, suggesting that the two forms play distinct biological roles."

The body also produces a cousin of ATXN1 called ataxin-1-like (ATXN1L). Losing each protein in mice produces very different outcomes:

• Losing ATXN1 does not cause ataxia. Instead, it produces learning and memory deficits and increases amyloid beta production — Alzheimer's-like features affecting the cortex and hippocampus.
• Losing ATXN1L causes lung defects, death of newborns shortly after birth, and hydrocephalus, an abnormal buildup of cerebrospinal fluid in the brain's cavities.

What did the mice reveal?

When the team engineered mice to lack one CIC form at a time, the differences were striking:

• Mice without CIC-S often died early, with severe developmental problems especially in the lungs. Some also accumulated fluid in the brain, grew poorly and eventually died.
• Mice without CIC-L survived but showed behavioral abnormalities, learning and memory difficulties, movement deficits and hyperactivity.

"These findings showed that the two forms of CIC are not interchangeable, each has its own essential function," Lee said.

The symptom overlap was telling: mice lacking CIC-S resembled mice lacking ATXN1L, and mice lacking CIC-L resembled mice lacking ATXN1. Closer examination confirmed specific pairings — CIC-L preferentially binds ATXN1, while CIC-S preferentially binds ATXN1L.

Why does the cerebellum take the hit?

Protein levels vary by brain region and developmental stage. The cerebellum carries the highest baseline levels of CIC in the central nervous system. When overactive mutant ATXN1 arrives, it finds a dense pool of its preferred CIC-L partners there, hyper-stabilizing the complex and disrupting gene regulation specifically within cerebellar Purkinje networks. Other tissues, with lower CIC-L ratios, are spared.

"This means that certain areas are more vulnerable if a specific protein partnership is disrupted," Lee said.

During lung development, ATXN1L and CIC-S are both at high levels, which mirrors the lung failure seen when that pairing is broken.

"Our findings show that subtle differences in the relative abundance of CIC and ATXN1 forms at the protein level, together with the differential complexes they assemble, can result in highly specialized functions and dictate regional vulnerability," Zoghbi said.

What could this mean for treatment?

The findings come from mouse models, so human applications remain preliminary. Still, the study suggests a practical strategy: instead of shutting down a disease protein everywhere and risking side effects in healthy tissue, future therapies could target the specific protein partnership driving disease in a given region — such as the ATXN1/CIC-L complex in the cerebellum.

"Our study illuminates an improved understanding of neurological disease and offers new possibilities to comprehend and treat these conditions more effectively," Zoghbi said.

Esmeralda Villavicencio Gonzalez, Elias M. Rivera, Mark A. Durham, Ronald Richman, Elizabeth H.-Y. Chu, Kailey Xia, Hu Chen, Zhandong Liu, Surabi Veeraragavan and Binoy Shivanna, all at Baylor College of Medicine and/or Texas Children's Hospital, also contributed to the work.