brugada.net

The science

The measurement, and the question nobody has answered.

The starting number is solid. What it means inside a human heart is not. That gap is the honest centre of this project, and it gates every route on the site.

Where the variant sits

The sodium channel is a large protein threaded through the surface of a heart cell. Before the part that spans the membrane begins, a stretch of it hangs inside the cell, called the N-terminal domain, meaning the first section of the protein rather than the part embedded in the membrane. Position 104 is in that stretch, and it sits against a negatively charged neighbour at position 84.

Replacing arginine with glutamine removes a positive charge from a position that is buried inside the folded protein rather than exposed at its surface, and that is where the structural argument starts. It comes from an experimentally determined structure rather than a computational model. RCSB 8VYJ, cryo-electron microscopy

A caution about this section. Five separate analyses in this project have measured how buried that position is, and they disagree with each other depending on the method and on whether the domain is measured alone or inside the whole channel. No single percentage is quoted here for that reason. Two candidate explanations for why the position matters, local strain and exposed water-repelling surface, were each tested and eliminated when a variant known to be harmless scored just as badly.

The open question

A person with this variant has one working copy of the gene and one broken copy. There are two very different things the broken copy could be doing, and they lead to different conclusions about treatment.

  • It simply sits out. The broken copy never reaches the cell surface, the working copy carries on normally, and the total is about half of normal.
  • It interferes. The broken copy actively disrupts the working copy, dragging the total below half.

The measured value sits below half, which is what makes this a real question rather than a theoretical one.

Corrected since: the figure that was here has been withdrawn. It drew the comparison as 50 percent expected against 34.1 percent measured. Both numbers were rescaled by dividing the laboratory measurement by two, which assumes the two copies add perfectly. The same study measured the two-copy case directly and got 218.4 percent of one copy rather than 200, so the correct pair is 45.8 against 31.3. The figure is held back until it can be redrawn rather than shown with the superseded pair. The sample-size panel it carried was computed against the old pair and has not been recalculated.
31.3%

of the current a normal two-copy heart would produce, rescaled from the laboratory measurement. Simple failure of one copy would give 45.8 percent, so about 15 percentage points are unexplained.

Unresolved

That gap is the entire question, and the standard assay cannot resolve it even in principle. The two competing explanations are algebraically degenerate, meaning they predict the same current. Measuring harder with the same method cannot separate them, which is a property of the measurement itself rather than a shortcoming of the studies that used it.

What would settle itAn assay that resolves each copy separately rather than measuring only the total. The design is on the experiments page, and its internal control is stated in advance.

This matters beyond one variant. If the broken copy interferes with the working one, the target is not this variant but the whole class of changes in this gene that suppress their healthy partner, which is a fifth to a third of Brugada patients who have an identified gene. If it simply sits out, that class does not exist and several routes on this site lose their rationale.

Why a single number is not the whole story

The 68.3 percent figure comes from a cell that is not a heart cell, engineered to make this protein. Those cells lack the partner proteins and the alternative forms of the protein that a real heart muscle cell has. The authors of the closest methodological work say so themselves. Whether the interference operates in a human heart cell is the thing that has not been tested.

There is also direct evidence that the two systems can disagree. For one variant, the same measurement gave a 33 to 46 percent current reduction in heart cells grown from a patient against 50 percent in an engineered line, with additional differences that the engineered system did not show at all.