Two strikes and you’re out: Gene-gene mutation interactions in HCM

Francis Crick: “While Ockham's razor is a useful tool in the physical sciences, it can be a very dangerous implement in biology. It is thus very rash to use simplicity and elegance as a guide in biological research”

Gregor Mendel is credited with first understanding the fundamental characteristics of genetic inheritance through his studies of trans-generational expression of phenotypes in pea plants (Pisum sativum). As every middle schooler knows, through careful cross-breeding he discerned dominant and recessive modes of inheritance for, but not blending of, parental traits. He rightly concluded that inheritance is achieved through vertical transmission of trait-encoding factors that we now call genes.

An important underlying assumption to Mendel’s mechanistic view of inheritance is that one inheritance unit (he called them “factors” ¹; we call them genes) encodes one trait. This relationship holds for the seven pea phenotypes he catalogued, but is the exception rather than the rule in mammalian biology. For example, stature is an inherited trait, but over 50 different genes are thought to contribute to human height ². Nevertheless, heritable diseases are commonly referred to as “Mendelian” when they are attributed to mutations in a single gene. The public database OMIM (Online Mendelian Inheritance in Man; http://www.omim.org) lists >4,000 Mendelian disorders with a known molecular basis, plus >1,700 others with an unknown molecular basis. While some of these diseases, such as cystic fibrosis (autosomal recessive mutation of CFTR gene), Huntington’s disease (autosomal dominant mutation of HTT gene), and hypertrophic cardiomyopathy (HCM; autosomal dominant mutations of the MYH7 ³ and other sarcomeric protein genes), are caused by mutations in one gene, even for so-called Mendelian diseases the contribution of additional genetic modifiers has yet to be well modeled. In this issue of Circulation Research, Blankenberg and colleagues show how the combinatorial effects of two different HCM-associated mutations can be critical determinants of disease phenotype ⁴.

Humans have an unusual potential to develop genetic diseases because we suffer from an abnormally high prevalence of deleterious mutations. As a species we have accumulated damaging mutations much faster than we have cleared them ^5–7; here’s why: The rate of non-synonymous DNA mutations in protein coding genes is constant, ~1×10⁵/ gene/generation ⁶. Fitness loss for a mutated protein (i.e. the odds that a non-synonymous mutation will be damaging) is also constant, ~1% ⁶. Damaging mutations tend to be cleared from the population through evolutionary suppression over many generations. Thus, in a stable long-term population negative selection restrains the prevalence of damaging DNA mutations. However, the human population has been anything but stable over a relatively short period of evolutionary time. In 5,000 BC there were ~5 million Homo sapiens on the planet. The advent of agriculture and transition from nomadic life to urban civilizations prompted an exponential increase in population, resulting in ~500 million humans by the middle ages, a 100-fold increase. The industrial revolution and modern technologies stimulated a second exponential increase in population, from 1 billion in 1800 to over 7 billion now. Thus, in just a few hundred generations human population has increased by three orders of magnitude (doubling just within the authors’ lifetimes!). This represents a tremendous number of new genomes to accumulate damaging mutations with insufficient time to clear them. Accordingly, 73% of all protein coding variants and 86% of all deleterious SNPs are only 5,000–10,000 yrs old ⁷. Mathematically, every person currently alive is carrying one or more severely damaging DNA mutations, and every protein coding gene is represented by someone carrying a dysfunctional variant, most of which are rare or private.

Mutations in MYH7 and MYBPC3, which encode β myosin heavy chain and myosin binding protein C, respectively, account for approximately half of inherited HCM. Clinical genetic testing for HCM has become more commonplace and revealed marked inter-individual variability in disease, i.e. the same gene mutation is present in individuals with a spectrum of HCM, even within the same family. Differences in disease phenotype (i.e. symmetric vs asymmetric hypertrophy and the occurrence of arrhythmias), variable disease severity, and different patterns of onset all suggest a role for modifiers that may be genetic or environmental.

Blankenberg et al ⁴ examined three MYH7 mutations, V606M, R453C, and R719W, mapping to unique regions of the β-myosin heavy chain (MHC) head domain. V606M has been characterized as a “mild” HCM-associated mutation in humans associated with later onset disease and less severe hypertrophy ⁸. V606 lies in the 50 KDa portion of the myosin S1 head, where it contributes to the actin-binding site ⁹. R453C has been associated with more severe human HCM and is located within the γ-phosphate sensing domain of the myosin ATPase ¹⁰. When engineered in vitro into human β myosin, the R453C mutation has reduced actin-activated ATPase and directs slower in vitro sliding of actin filaments ¹¹. The final mutation, R719W, is also associated with more severe human HCM, is located within the myosin converter domain, and has been linked to increased elastic distortion of individual myosin heads ¹².

To better evaluate their individual and combined phenotypes on identical genetic backgrounds, Blankenberg introduced these three human MYH7 HCM mutations into the respective positions of the mouse Myh6 gene (encoding αMHC, the major myosin heavy chain protein found in the murine heart). Somewhat recapitulating the human genotype-phenotype spectrum, R453C caused hypertrophy by 26 weeks of age in mice, whereas V606M had no significant effect on left ventricular hypertrophy at that same time. Even when homozygous, the V606M mutation was insufficient to produce cardiac hypertrophy. However, when the V606M mutant mice were crossed with the R453C mutant mice, the dual heterozygous mutant progeny mice developed more hypertrophy than either of the parent mutant strains, demonstrating additive phenotypic effects of these two mutations.

Functionally, these allelic combinations offer the potential for fresh molecular insight. The V606M mutation with its ability to modify actin binding is insufficient to produce HCM, even when homozygous, i.e. in the absence of any normal MHC. However, altered actin binding plus reduced ATPase evoked significantly more hypertrophy, suggesting a disease model wherein the cardiac sarcomeres contain a mixture of myosin heads, some with different actin binding and others with reduced ATPase. These mixed mutations may or may not reflect a situation where each allele produces half the total MHC protein, as V606M and R719W may not be expressed at equal levels to the normal allele ¹⁰. It is also possible that some mutations may alter mRNA stability and splicing, provoking reduced expression of the mutant protein relative to normal MHC, thereby leading to less hypertrophy.

This work demonstrates that a seemingly innocuous secondary mutation may, in some cases, evoke more severe HCM. As caveat, an inbred murine genome and heart differs markedly from the human HCM condition. Nonetheless the findings suggest that multiple mutations are likely to elicit a more severe phenotype. To translate this finding to the human condition requires estimating the frequency of MYH7 genetic variation in the population at large, including DNA variations not yet linked to HCM phenotypes. Where it has been studied, the frequency of potentially pathogenic MYH7 mutations in the general population is much higher than would be predicted from the prevalence of HCM ^{13, 14}. HCM occurs in roughly 1 in 500 individuals, but DNA variants thought to predispose to HCM are present at 5–10 fold higher frequency. Taken together, these population-based estimates and the current findings in mice show how individual DNA variants that contribute little or nothing to cardiac phenotypes in otherwise normal individuals may lead to more severe disease in the context of a more pathogenic HCM mutation. While these analyses focused on the MYH7 gene, the “second genetic hit principle” is certainly not restricted to this gene; functional significance of other sarcomere gene DNA variants is likely to be revealed in the context of a pathogenic MYH7 HCM mutation. For this reason, Blankenberg and colleagues advocate more broad-based genetic screening to help identify those at highest risk for faster progression.

A markedly increased potential for gene-gene interactions is one of the consequences of a human genetic landscape replete with mutations that are both rare and predicted to be damaging. If a given mutation is “causal”, other functionally significant DNA variants can almost certainly act as phenotypic modifiers. Depending upon genetic context, DNA variants may therefore be categorized as primary or “driver”, associated, or co-morbid mutations, and may exhibit variable expression or reduced penetrance. This genetic reality not only provokes inter-individual variability in phenotype, but blurs the distinction between monogenic, bigenic, and polygenic or complex inheritance. Indeed, it is likely that the genetic component of phenotype can only be explained by the cumulative functional consequences of all coding and non-coding DNA sequence variants in the same gene, in other genes of the same pathway, and finally in all genes that impact the affected organ system (Figure). And, this approach still fails to incorporate epigenetic and environmental influences. Serious attempts at this type of systems-level genetic integration will need to be both comprehensive and unbiased. Accordingly, the initial discovery of a genetic disease might be accomplished using conventional genetic linkage of the suspected primary mutation in a relatively few affected subjects, but to fully understand how the spectrum of personal genotypes help determine individual phenotype will require identifying all genomic DNA variants in large cohorts of subjects presenting with the disease phenotype, regardless of primary mutation.

Figure 1. Schematic depiction of variable genetic and non-genetic influences on individual phenotype.

Left, different forms of secondary mutations. Right, role of secondary mutations in context of other modifying influences.