Molecular Pathogenesis
The mechanism by which pathogenic variants in either EPM2A or NHLRC1 result in Lafora disease and the exact role of the Lafora bodies in the pathogenesis of LD have been the subject of intensive research efforts over the past few years.
Pathology in LD consists of the progressive formation of polyglucosans (insoluble glucose polysaccharides that precipitate and aggregate into concretized masses called Lafora bodies), resulting in neurodegeneration. Lafora bodies form in neuronal perikarya and in neuronal short processes (mostly dendrites). Lafora bodies in the neuronal processes are much smaller but they massively outnumber Lafora bodies in the perikarya. Extraneurally, Lafora bodies also form in heart, liver, and skeletal muscle, but cause no symptoms in these organs [Turnbull et al 2011].
A normal glycogen molecule contains up to 55,000 glucose units, yet remains soluble because its glucose chains are short (13 units), each chain is a branch of another, and the whole molecule is a sphere, the surface of which is composed of the hydrophilic ends of chains [Graham et al 2010]. This unique organization allows mammalian cells to store large amounts of carbohydrate energy in a soluble, rapidly accessible form. Without branching, glucose polymers 13 units or longer are poorly soluble and tend to precipitate and crystallize [Hejazi et al 2008]. Polyglucosans are malformed glycogen molecules. They have very long chains, insufficient branches, and a resultant lack of spherical organization. They are more similar to plant amylopectin or starch than to glycogen, and like these plant carbohydrates they are insoluble and they precipitate and accumulate [Minassian 2001].
Glycogen is synthesized through coordinated actions of glycogen synthase and glycogen branching enzyme – the former responsible for chain elongation, the latter for chain branching. Glycogen is digested by glycogen phosphorylase and glycogen debranching enzyme.
The current view on LD pathogenesis suggests that LD is predominantly caused by an impairment in chain-length regulation affecting only a small proportion of the cellular glycogen. The principal function of laforin (gene product of EPM2A) relevant to LD is mediated through malin (gene product of NHLRC1) and directed to preventing glycogen molecules with hyperextended chains. In the absence of either protein, some glycogen molecules at a time precipitate and gradually over time aggregate and amass into Lafora bodies, which, reaching a certain threshold profusion (at age ~14 years in humans), initiate and then drive the progressive myoclonus epilepsy [Nitschke et al 2017; Sullivan et al 2017].
EPM2A
Gene structure.
EPM2A has four exons spanning 130 kb; they are alternatively spliced to form two major EPM2A transcripts [Minassian et al 1998, Serratosa et al 1999, Ganesh et al 2000, Gómez-Garre et al 2000]. NM_005670.3 represents the longer transcript and encodes the longer laforin isoform (a) of 331 amino acids. For a detailed summary of gene, transcript, and protein isoform information, see Table A, Gene and Normal gene product.
Pathogenic variants. To date, more than 70 different pathogenic variants in EPM2A have been reported in more than 100 families [Turnbull et al 2016]. An overview of the different pathogenic variants can be found in the Human Gene Mutation Database or the Lafora Progressive Myoclonus Epilepsy Mutation and Polymorphism Database [Ianzano et al 2005].
Pathogenic variants in EPM2A are scattered all along the coding regions of the gene, but also accumulate in discrete spots of high recurrence. The most common pathogenic variant is the nonsense c.721C>T variant, the so-called "Spanish" pathogenic variant, which accounts for approximately 17% of EPM2A-mediated LD. Its high prevalence is the result of both a founder effect and recurrent events [Gómez-Garre et al 2000, Ganesh et al 2002b]. Large deletions make up 10%-15% of EPM2A pathogenic variants, and the pathogenic c.512G>A variant accounts for approximately 8%.
Table 3.
Pathogenic EPM2A Variants Discussed in This GeneReview
View in own window
| DNA Nucleotide Change | Predicted Protein Change | Reference Sequences |
|---|
| c.512G>A | p.Arg171His |
NM_005670.3
NP_005661.1
|
| c.721C>T | p.Arg241Ter |
Variants listed in the table have been provided by the authors. GeneReviews staff have not independently verified the classification of variants.
GeneReviews follows the standard naming conventions of the Human Genome Variation Society (varnomen.hgvs.org). See Quick Reference for an explanation of nomenclature.
Normal gene product.
EPM2A is known to encode two distinct proteins by differential splicing; a phosphatase active cytoplasmic isoform (a) (laforin, NP_005661.1) and a phosphatase inactive nuclear isoform (b) (NP_001018051.1). Both isoforms of the laforin protein have unique C termini [Ganesh et al 2002c, Ianzano et al 2004]. The carboxyl terminal of isoform (b) targets laforin to the nucleus, a feature that is not shared by longer laforin isoform (a). Ianzano et al [2004] demonstrated that disturbances in the physiologic functions of laforin isoform (a) underlie the pathogenesis of LD, and isoform (b) cannot functionally substitute for laforin isoform (a). The common segment of the laforin isoforms consists of a carbohydrate-binding module and a dual-specificity protein phosphatase domain [Ganesh et al 2000].
Dubey et al identified three additional EPM2A splice variants with potential to code for five distinct proteins in alternate reading frames. The novel isoforms, when ectopically expressed in cell lines, show distinct subcellular localization and interact with and serve as substrates of malin (protein product of NHLRC1). Alternative splicing could possibly be one of the mechanisms by which EPM2A regulates the cellular functions of the proteins it codes for [Dubey et al 2012].
Laforin contains an N-terminal carbohydrate-binding domain (CBD), encoded mainly by exon 1, and a dual-specificity phosphatase domain (DSPD) spanning exons 3 and 4 [Minassian et al 2000b, Ganesh et al 2002b].
Laforin is conserved in all vertebrates; while it has been lost in the vast majority of lower organisms, it is an ancient protein that is conserved in a subset of protists and invertebrates that have undergone slower rates of molecular evolution and/or metabolize a carbohydrate similar to Lafora bodies. The laforin protein holds a unique place in evolutionary biology and has yielded insights into glucan metabolism and the molecular etiology of Lafora disease [Gentry & Pace 2009].
Abnormal gene product. Nonsense variants, insertions, and deletions in EPM2A are predicted to be functionally "null" and to have lost phosphatase activity. Missense variants in EPM2A also result in a lack of phosphatase activity in vitro, resulting in a "null" effect [Fernández-Sánchez et al 2003, Ganesh et al 2006]. Loss of phosphatase activity is not restricted to pathogenic variants located in the DSPD; it has also been observed for pathogenic variants affecting the CBD of EPM2A [Wang et al 2002, Fernández-Sánchez et al 2003]. It is likely that the missense variants affect proper folding of the laforin protein, as illustrated by transfection experiments overexpressing missense mutants, which resulted in ubiquitin-positive cytoplasmic aggregates, suggesting that they were folding mutants destined for degradation [Ganesh et al 2000, Ganesh et al 2002a]. Missense variants also affect the subcellular localization of laforin [Ganesh et al 2002a, Mittal et al 2007] and disrupt the interaction of laforin with R5 and malin (protein product of NHLRC1) proteins that interact with laforin in vivo [Fernández-Sánchez et al 2003, Gentry et al 2005]. It is evident that not all aspects of the protein function have been tested for each missense variant, and that sensitive assays for checking the effect of pathogenic variants on protein function are yet to be developed [Singh & Ganesh 2009].
NHLRC1 (EPM2B)
Gene structure.
NHLRC1 is a single-exon gene spanning 1,188 base pairs that has all of the proposed features of the consensus sequence of a eukaryotic translational initiation site at its 5' end and two putative polyadenylation signals at its 3' end. Northern blot analysis indicates the presence of NHLRC1 as two transcripts of 1.5 kb and 2.4 kb in all tissues examined, including specific subregions of the brain [Chan et al 2003b]. For a detailed summary of gene and protein information, see Table A, Gene.
Pathogenic variants. To date, more than 70 pathogenic variants have been reported in more than 125 families. The majority are missense variants, although insertions, deletions, and nonsense variants have also been found. A heterozygous deletion of the entire NHLRC1 gene has been reported in an Italian and a Serbian family [Lohi et al 2007]. An overview of pathogenic alleles in NHLRC1 is available in the Human Gene Mutation Database or the Lafora Progressive Myoclonus Epilepsy Mutation and Polymorphism Database [Ianzano et al 2005].
The c.468_469delAG
pathogenic variant, involving the removal of two bases in the
coding region, accounts for approximately 8% of
NHLRC1 pathogenic variants, and is by far the most common
deletion. It has been identified in 14 individuals belonging to the same genetic isolate of tribal Oman. All shared a common haplotype, suggesting a
founder effect [
Turnbull et al 2008].
Note: Whereas c.205C>G
pathogenic variant is common in affected persons of Italian and Spanish heritage, both the c.205C>G and c.468_469delAG pathogenic variants have been identified in different ethnic groups, suggesting a recurrent mutation event; these two sites represent hot spots for
NHLRC1 pathogenic variants [
Ganesh et al 2006].
The c.436G>A variant has been reported in individuals with mild LD. The mean age at onset in individuals with this variant reported in the literature is 18 years; this variant is the only one reported with such late onset age [
Ferlazzo et al 2014].
Table 4.
NHLRC1 Pathogenic Variants Discussed in This GeneReview
View in own window
| DNA Nucleotide Change | Predicted Protein Change | Reference Sequences |
|---|
| c.76T>A 1 | p.Cys26Ser |
NM_198586.2
NP_940988.2
|
| c.205C>G 1 | p.Pro69Ala |
| c.436G>A | p.Asp146Asn |
| c.593T>A | p.Ile198Asn |
| c.468_469delAG 1 | p.Gly158ArgfsTer17 |
Variants listed in the table have been provided by the authors. GeneReviews staff have not independently verified the classification of variants.
GeneReviews follows the standard naming conventions of the Human Genome Variation Society (varnomen.hgvs.org). See Quick Reference for an explanation of nomenclature.
- 1.
Normal gene product.
NHLRC1 encodes E3 ubiquitin-protein ligase NHLRC1 (also known as malin), a 395-amino acid protein. Malin contains a zinc finger of the RING type and six NHL-repeat protein-protein interaction domains [Chan et al 2003b].
Abnormal gene product. Nearly all pathogenic variants in NHLRC1 are predicted to result in the loss of function of malin [Chan et al 2003b, Gómez-Abad et al 2005, Singh et al 2005]. For more information, see Turnbull et al [2016].
