From an ASH1L finding to evidence for care

There is compelling evidence that haploinsufficiency is the main mechanism of disease, and for several variants (e.g. intra- or whole-gene deletions, nonsense, frameshift, and splice-site variants) is clear that lead to nonsense-mediated decay (NMD). However, it is also acknowledged that some of them, because of different reasons, may escape NMD, cause de deletion or insertion of in-frame segments of amino acids, etc. Therefore, it is convenient to perform a detailed analysis of the variant and then to perform gene expression analysis and/or western blot of the resulting protein. On the other hand (and linked with a below question), it is possible that ASH1L may be associated with more than one condition, depending on the mechanism of disease. To date, it is clear that monoallelic loss-of-function leads to the current syndrome, but considering emerging evidence for similar genes (e.g. KMT2D, DOT1L), it may be associated with other disorders if the variants cause either a gain-of-function or a dominant negative effect, which are mainly cause by missense or in-frame variants. To confirm these hypotheses, two types of analyses are required at least: 1) to compare phenotypes from patients with those different types of variants through a genotype-phenotype correlation, and 2) to create a DNA methylation episignature. Additional research may also shed light on those hypotheses, but at least these two approaches will lead to clearer situations. The episignature will also solve those cases with uncertain findings in this gene.

To solve those questions, it is necessary to have a detailed analysis of the genomic context. For example, in our 2018 work we only selected patients with minimum compromising of other genes other than ASH1L to be completely sure of its contribution to this novel disorder. For larger deletions, it is likely that other genes may also contribute to the phenotype and a detailed correlation among all patients with deletions may lead to delineate the minimum phenotype and to identify the other contributing genes, especially considering that there are several metrics that ease their identification (pLI, pHaplo, etc). Similarly for duplications, as it is predicted that duplications are also deleterious in humans according to the pTriplo score of 0.82. For rearrangements that have their breakpoints within the gene, although it is very likely they will produce loss-of-function, it is relevant to see if they also affect a topologically-associated domain, which may affect the expression of other relevant genes. Knockdown of ASH1L or overexpression of this gene in zebrafish, for example, may shed light in confirming these findings and have a correlation with what has been seen in humans.

In my experience with that method, it is crucial to find a distinguishable episignature first, and for that, it is necessary to include patients with variants that undoubtedly lead to haploinsufficiency. Then, missense variants should be tested to see if they "behave" like loss-of-function variants, or they cluster differently, which may lead to the identification of the other potential disorders (see aforementioned hypothesis). CNV and splice-site variants require a more detailed genetic analysis first in order to predict their potential effect and therefore, to predict its episignature.

At the moment, episignatures only have a diagnostic utility, but I can predict they could have a "management" utility as well. For the KMT2C episignature work (in which I was also part of), we see that patients with a "weaker" episignature have also a milder phenotype, but the problem was that the sample size could not confirm or discard that hypothesis. Also, I think episignatures may be useful as a "biomarker": these enzyme-encoding genes have counteracting enzymes, which can be inhibited to rebalance histone methylation (e.g. ASH1L methylation is counteracted by KDM2A and KDM2B). There are several compounds that are being tested in this way, especially for cancer, so if we see some episignature that changes with a further therapy, then we can say it is potentially useful in an objective way. Again, this is only an hypothesis.

Difficult to say now, first of all it is necessary to see what happens in big cohorts so the publication of the 61 compiled patients is a must. The frequencies of different manifestations will shed light on more focused management, and subsequent surveillance will shed light on the natural history. But at the moment, very few patients have been published so I cannot have a stronger opinion. However and considering the broad effect of this gene on human body, it is likely that multiple evaluations are required for all patients.

All of them are necessary, and although a pharmacological treatment is mostly considered as the "final answer", I think current non-pharmacological management is crucial as it has been demonstrated that cause long-term benefits for neurodevelopmental disorders (e.g. see MECP2 work done in mice exposed to stimulating environments, which is been replicated in patients with MECP2-pathies). In this sense, to have a ASH1L-neuropsychological profile will be very useful as it may lead to a more focused/personalized management by therapists, including patients with different ages.

Difficult questions as they require to solve trans-countries difficulties: each country and even every institution have their own acts, policies and regulations on data sharing, biobanks, etc. However, I do think that the only way to solve those problems is through collaboration. For many disorders, there are patient organisations in many countries, and a coordinated work amongst them is required to solve common, but difficult questions. If all ASH1L organisations have a common plan, then every ASH1L patient organisation in every country can work on adapting that plan to their specific contexts and countries´ requirements, caring that a minimum core of objectives are kept intact. Finally and considering that I am researcher from the Global South, it is necessary a fair work and genuinely collaborating work, where everyone is a peer and no asymmetries are allowed, except for specific situations when required.