Penetrance

How was penetrance re-estimated?

Minikel and colleagues compared case data against large population controls: 16,025 prion disease cases, 60,706 population control exomes and 531,575 individuals genotyped by 23andMe.

Missense variants in PRNP previously reported as pathogenic were at least 30 times more common in the population than the prevalence of genetic prion disease would predict. The authors attribute part of that excess to benign variants falsely assigned as pathogenic, while concluding that other variants have genuine effects on susceptibility, conferring lifetime risks that range from under 0.1 percent to about 100 percent.

They also report that truncating variants in PRNP have position-dependent effects, with true loss-of-function alleles found in healthy older individuals, and state that this supports the safety of therapeutically suppressing prion protein expression. That second finding is now load-bearing for the PrP-lowering programmes described at therapeutics research.

The method has a structural limitation worth stating. Population databases are not age-matched to disease onset and contain people who may yet develop disease, so they give an upper bound on allele frequency in the unaffected population rather than a clean denominator. The direction of the bias is toward overestimating how common these alleles are among people who will stay well, which makes the re-scaling conservative rather than exaggerated.

Is a penetrance figure the same as a risk for a person?

No, and the general variant interpretation literature has moved toward case-by-case assessment for this reason. Gudmundsson and colleagues examined clinically relevant ClinVar variants in 807,162 individuals in gnomAD and then assessed 734 predicted loss of function variants in 77 genes associated with severe, early-onset, highly penetrant haploinsufficient disease. They found explanations for the apparent lack of disease in 701 of 734 variants, 95 percent, and conclude that individuals with genuinely unexplained absence of disease are rare, while deep case-by-case assessment is needed to avoid false assignments of disease risk, particularly in unaffected people.

That paper is about haploinsufficiency genes rather than PRNP, where the mechanism is gain of function, so it is cited here for the methodological point only: a published penetrance number is an aggregate over a variant, not a statement about a carrier.

Why does variable age at onset matter so much?

Because it decides whether a trial in healthy carriers can be run. Minikel and colleagues assembled age at onset or death data from 1,094 individuals with high-penetrance PRNP mutations, generated survival and hazard curves, tested for genetic modifiers and estimated statistical power.

Age at onset varies over several decades for the commonest mutations. Neither sex, nor parent’s age at onset, nor codon 129 genotype added explanatory power that could be used to stratify a trial. Their power calculations indicate that a randomised preventive trial with clinical onset as the endpoint would need hundreds or thousands of at-risk individuals, which they describe as posing prohibitive cost and delay and as likely exceeding the number of people available.

Their conclusion is that biomarkers suitable as surrogate endpoints are essential for prevention, with longer trial duration, increased enrolment and historical controls in a post-marketing study as routes to confirming clinical benefit later. This is the argument behind the regulatory question set out at evidence for regulators.

The practical message for a family member is uncomfortable and important: no one can tell a carrier when, or whether, disease will begin, and a parent’s age at onset does not predict a child’s.

What has changed in counselling practice?

Goldman and Vallabh set out the current position. They give 10 to 15 percent of prion disease cases as caused by autosomal dominant gain-of-function PRNP variants, and note that rarity and phenotypic variability complicate diagnosis, often obscuring family history and leaving families unprepared when an index case occurs. Three developments drive their update: a prion-detection assay that has transformed symptomatic diagnosis, systematic characterisation of penetrance, age of onset and illness duration across PRNP variants in a global cohort, and the prospect of genetically targeted therapies in which healthy at-risk individuals may be best positioned to benefit.

They also report that the traditional genotype to phenotype correlation has weakened over time, and suggest that the term genetic prion disease now serves providers better than the historical subtype names. Their stated aim is to expand access to genetic counselling and testing services for people who want them.

Testing is a personal decision and not testing is a valid choice. See predictive testing and the right not to know.

Related: genetics hub, codon 129, the PRNP gene and variants.