An Explanation for the Disconnect between Z-AAT Editing Efficiency and Alpha-1 Antitrypsin Output

The commonly stated goal of alpha-1 antitrypsin editing approaches in addressing AATD is to achieve 50% editing of the mutant Z-allele.  This is based on human genetics showing that MZ heterozygotes have a much reduced risk of developing liver and lung disease.  

It therefore baffled me that Beam Therapeutics reported a year ago serum AAT levels of only ~12.4uM at day 28 despite having achieved 75-85% ZàM conversion with a 60mg dose of BEAM-302.  This compares to ~15-19uM concentrations seen in MZ heterozygotes, that is people who have just 50% of their AAT alleles in the M state.  

Back of the napkin math would have predicted ~22-30uM AAT levels with 75% M alleles if there was a corresponding relationship between M allele and AAT output.

Beam Therapeutics now reports improved serum AAT levels

This week, Beam Therapeutics reported that one additional year into the clinical trial, mean serum AAT levels in the 60mg cohort have increased to 16.1uM.  This includes an additional 3 patients over the 2025 data cut who had higher baseline AAT levels compared to the initial cohort of n=3.  Curiously, the company did not show how serum AAT changed over time, and now I think they did so on purpose because they did not want to steal the thunder from a future conference presentation.

Z-AAT relief effect following genetic correction takes time

ZZ homozygotes carry a high risk of developing liver fibrosis and ultimately liver failure in their lifetime because of the accumulation of Z-AAT aggregates in hepatocytes and resultant cell stress.  Consequently, less AAT gets out into circulation.  Arrowhead Pharmaceuticals and Takeda are addressing the liver disease aspect of AATD by Z-AAT RNAi knockdown with fazirsiran and are now running an unexpectedly long 4 year study after initial findings indicated slower than expected clearance of AAT aggregates and subsequent amelioration of fibrosis.

It therefore is plausible that the reason why serum AAT level undershoot following BEAM-302 treatment compared to genetic expectations is due to some lag in normalizing AAT output in subjects where Z-AAT had caused stress to the liver for decades of life already.  The increase from the mean 12.4uM reported last year to now 16.1uM could be a reflection of hepatocytes clearing the aggregates to relief endoplasmic reticulum stress and grease the AAT export machinery.  

Alternatively, you would have to explain why the inferred mean for the last 3 subjects comes out at 19.8uM versus the 12.4uM for the first 3 subjects.  Of course, you could try and explain some of it with the differences in baseline AAT output (which actually could also reflect Z-AAT-related liver stress), or time-dependent AAT increases as a result of a competitive growth advantage of corrected hepatocytes, or differences in the way serum AAT was measured last year and this year (turbidimetry vs LC-MS), but none of them explain both the serum concentration undershoot versus genotype expectation and the apparent increase over time.

Be prepared to be pleasantly surprised as more data from the BEAM-302 program emerges.  Also be wary of out-of-context data from investigator-initiated trials of obscure genome editing compounds like the YOLT-202 copycat.