Molecular Pathology of Laminopathies

Abstract

The nuclear envelope is composed of the nuclear membranes, nuclear lamina, and nuclear pore complexes. Laminopathies are diseases caused by mutations in genes encoding protein components of the lamina and these other nuclear envelope substructures. Mutations in the single gene encoding lamin A and C, which are expressed in most differentiated somatic cells, cause diseases affecting striated muscle, adipose tissue, peripheral nerve, and multiple systems with features of accelerated aging. Mutations in genes encoding other nuclear envelope proteins also cause an array of diseases that selectively affect different tissues or organs. In some instances, the molecular and cellular consequences of laminopathy-causing mutations are known. However, even when these are understood, mechanisms explaining specific tissue or organ pathology remain enigmatic. Current mechanistic hypotheses focus on how alterations in the nuclear envelope may affect gene expression, including via the regulation of signaling pathways, or cellular mechanics, including responses to mechanical stress.

INTRODUCTION

In 1994, Daniela Toniolo and colleagues (1) reported that mutations in a novel gene caused X-linked Emery-Dreifuss muscular dystrophy. Two years later, the encoded transmembrane protein named emerin after Alan Emery was localized to the nuclear envelope, presumably the inner nuclear membrane (2, 3). In 1999, Ketty Schwartz and colleagues (4) reported that mutations in LMNA encoding lamin A and lamin C (lamin A/C), peripheral proteins of the inner nuclear membrane, cause the phenotypically identical autosomal dominant Emery-Dreifuss muscular dystrophy. Soon after, mutations in LMNA were reported to cause Dunnigan-type familial partial lipodystrophy (5-7), mandibuloacral dysplasia type A (8), an axonal peripheral neuropathy (9), and Hutchinson-Gilford progeria syndrome (HGPS) (10, 11). These initial studies led to the identification of a group of diseases caused by mutations in genes encoding proteins of the nuclear envelope, which are now often referred to as nuclear envelopathies or, more often, laminopathies.

In the past two decades, additional laminopathies have been described. Investigators have also provided better descriptions of the genetics, clinical phenotypes, and pathology. The development of cellular and animal models has led to a better understanding of nuclear envelope protein functions and how their dysfunction may relate to tissue-specific pathology. However, deciphering mechanisms of how these disorders affect specific tissues, especially when the mutant genes and encoded proteins are widely expressed, has remained elusive. This review covers up-to-date aspects of the molecular pathology of the laminopathies, summarizing what we know and pointing out gaps in our understanding of this group of diseases.

THE NUCLEAR ENVELOPE

The nuclear envelope separates the nucleoplasm from the cytoplasm in eukaryotic cells. It consists of the nuclear membranes, nuclear pore complexes (NPCs), and nuclear lamina (Figure 1). The nuclear membranes are separated into three morphologically distinct domains: inner, outer, and pore. The outer nuclear membrane is contiguous with the endoplasmic reticulum and similarly contains ribosomes on its outer surface. It is separated from the inner nuclear membrane by the perinuclear space, a continuation of the endoplasmic reticulum lumen. The nuclear lamina, a meshwork on intermediate filament proteins called lamins, which form noncanonical 3.5-nm-diameter filaments, lines the nucleoplasmic side of the inner nuclear membrane (12). The pore membranes connect the inner and outer nuclear membranes where NPCs are located. The pore complexes are megadalton structures that mediate active and passive transport of substances between the nucleus and the cytoplasm (13).

Figure 1.

Schematic diagram of the nuclear envelope showing the nuclear membranes, nuclear lamina, and a nuclear pore complex. Ribosomes are on the rough endoplasmic reticulum and continuous outer nuclear membrane. The nuclear pore complexes are associated with the pore membranes, and lamina and chromatin are associated with the inner nuclear membrane.

A unique collection of transmembrane proteins concentrate in the inner nuclear membrane, along with a few in the pore and outer nuclear membrane of interphase cells (Figure 2). An initial subtractive proteomics analysis study of extracts of rodent livers identified approximately 80 integral inner nuclear membrane proteins (14). After synthesis on membrane-bound ribosomes, transmembrane proteins most likely concentrate in the inner nuclear membrane as a result of retention by binding to lamins or chromatin after lateral diffusion through the interconnected endoplasmic reticulum and nuclear membranes (15). The pore membrane contains a small number of transmembrane nucleoporins that anchor the pore complexes (13). The outer nuclear membrane generally shares transmembrane proteins with the endoplasmic reticulum but contains uniquely localized KASH domain proteins, termed nesprins in mammals. The KASH domains of nesprins interact within the perinuclear space with the luminal domains of integral inner nuclear membrane SUN proteins, forming the linker of nucleoskeleton and cytoskeleton (LINC) complex (16, 17). SUN proteins bind to lamins. The cytoplasmic domains of different nesprins interact directly or indirectly with actin, microtubule, and intermediate filaments. These connections provide a nucleocytoskeletal network mediating proper positioning of the nucleus within the cell and the transduction of force from the cytoplasm to the nucleus (18).

Figure 2.

Schematic diagram showing select transmembrane proteins of the inner, pore, and outer membranes of the nuclear envelope. Representative transmembrane proteins that concentrate in the inner nuclear membrane are SUN1/2, emerin, LAP1, NET25, LAP2 β, MAN1, and LBR. Gp210 is a representative integral protein of the pore membrane. Nesprins concentrate in the outer nuclear membrane by binding within the perinuclear space to the luminal domains of SUN proteins and also bind to cytoskeletal filaments.

Most nuclear envelope proteins appear to be expressed ubiquitously in most different cell types. For example, lamin A/C is present in most terminally differentiated somatic cells (19). However, there are differences in the nuclear envelope proteome of different cells and tissues (20). This topic has been relatively poorly studied but is likely important in understanding the generation of different cellular mechanics, cell cycle regulation, signaling, gene expression, and genome organization that may underlie the tissue-specific nature of most laminopathies.

LAMINOPATHIES

Mutations in LMNA Cause Four Major Types of Pathology

LMNA on chromosome 1q21.2-q21.3 encodes the A-type lamins lamin A, which is posttranslationally processed from a precursor protein, prelamin A, and lamin C (21, 22). Prelamin A and lamin C are identical for the first 566 amino acids. As a result of alternative splicing of RNA in the region encoded by exon 10, lamin C has six unique carboxyl-terminal amino acids and prelamin A 98.

Prelamin A contains a cysteine-aliphatic-aliphatic-any amino acid (CAAX) motif at its carboxyl terminus that triggers a series of processing reactions to generate lamin A (23). First, protein farnesyltransferase catalyzes the addition of a farnesyl moiety to the cysteine. Second, either Ras converting CAAX endopeptidase 1 or zinc metallopeptidase, STE24 homolog (ZMPSTE24) catalyzes the endoproteolytic cleavage of the -AAX. Third, isoprenylcysteine carboxyl methyltransferase catalyzes methylation of the farnesylcysteine. Finally, ZMPSTE24 recognizes the farnesylated protein and catalyzes an endoproteolytic cleavage leading to the removal of the last 15 amino acids, including the farnesylcysteine α-methyl ester.

Mutations in LMNA cause four major types of pathology involving (a) multiple systems with features of accelerated aging (progeroid disorders), (b) striated muscle, (c) adipose tissue, and (d) peripheral nerve (Table 1). Within each major pathology type, there are specific disorders that were described on the basis of clinical phenotypes before the genetics were elucidated, with overlap between them. However, while the major pathology types may share some features, they are clearly distinct.

Table 1.

Mutations in LMNA cause four groups of diseases affecting different tissues

Type of pathology	Specific disorders
Progeroid disorders
Prelamin A processing-deficient laminopathies	Hutchinson-Gilford progeria syndrome
	Other progeroid disorders
Prelamin A processing-proficient laminopathies	Mandibuloacral dysplasia type A
	Other progeroid disorders
Striated muscle disease
Dilated cardiomyopathy with variable skeletal muscle involvement	Emery-Dreifuss muscular dystrophy
	Limb-girdle muscular dystrophy 1B
	Dilated cardiomyopathy with minimal/variable skeletal muscle involvement
	Congenital muscular dystrophy
Adipose tissue disease
Lipodystrophy	Dunnigan-type familial partial lipodystrophy
	Atypical lipodystrophy syndromes
Peripheral neuropathy
Charcot-Marie-Tooth disease type 2	Charcot-Marie-Tooth disease type 2B1

Progeroid Disorders

The multisystem diseases with features of accelerated aging can be divided into processing-deficient and processing-proficient progeroid laminopathies (24). In the first group, mutations in LMNA lead to defective prelamin A processing and accumulation of farnesylated variants. Similarly, mutations in ZMPSTE24 can lead to defective prelamin A processing, prelamin A accumulation, and progeroid syndromes. In the processing-proficient disorders, LMNA mutations generate amino acid substitutions in lamin A/C, but prelamin A is normally processed.

The most common albeit still ultrarare progeroid disorder caused by defective prelamin A processing is HGPS. Children with HGPS have, among other symptoms, growth impairment, sclerotic skin, micrognathia, decreased subcutaneous fat, alopecia, prominent cutaneous vasculature, fingertip tufting, and decreased bone density; they generally die in their teens from complications of occlusive cardiovascular or cerebrovascular disease (25). A dominant de novo c.1824C>T (Gly608Gly) mutation activates a cryptic splice leading to an in-frame deletion of 150 base pairs in pre-mRNA encoding prelamin A (10, 11). As a result, an internally truncated prelamin A variant, called progerin, accumulates. Progerin lacks 50 amino acids, including the second ZMPSTE24 recognition site, and hence retains a farnesylated, carboxymethylated cysteine at its carboxyl terminus (Figure 3).

Figure 3.

Progerin is a farnesylated prelamin A variant that accumulates in Hutchinson-Gilford progeria syndrome (HGPS). The cysteine-isoleucine-serine-methionine (CISM) [cysteine-aliphatic-aliphatic-any amino acid (CAAX) motif] in prelamin A triggers three sequential reactions: (①) Protein farnesyltransferase catalyzes the addition of a farnesyl moiety to the cysteine (C), (②) CAAX endopeptidase 1 or ZMPSTE24 catalyzes the endoproteolytic cleavage of the -ISM, and (③) isoprenylcysteine carboxyl methyltransferase catalyzes methylation of the farnesylcysteine. Normally, ZMPSTE24 (wild type) then recognizes farnesylated prelamin A and catalyzes an endoproteolytic cleavage (scissors), leading to removal of the last 15 amino acids, including the farnesylcysteine α-methyl ester, to generate prelamin A. In HGPS, an LMNA mutation activates an RNA cryptic splice, leading to expression of an internally truncated prelamin A variant, called progerin, that lacks 50 amino acids (red dashed line). This deletion includes the second ZMPSTE24 recognition site, and hence it is cleaved and retains a farnesylated, carboxymethylated cysteine at its carboxyl terminus.

Other LMNA mutations can also lead to accumulation of progerin and cause phenotypes more or less severe than HGPS roughly in proportion to the amount of progerin expressed (24). An LMNA mutation leading to expression of a farnesylated prelamin A variant with an internal deletion of only 35 amino acids causes a progeroid disorder less severe than HGPS (24). One study also reported a patient with a de novo LMNA c.1940C>T transversion that resulted in a Leu to Arg amino acid substitution at residue 647, abolishing the ZMPSTE24 cleavage site; she accumulated full-length prelamin A with a single amino acid change and suffers from a relatively mild progeroid disorder (26).

Progeroid disorders also arise from loss-of-function mutations in ZMPSTE24. Severity of disease correlates with residual activity of the prelamin A processing protease (27). Total loss of function causes the neonatal lethal progeroid disorder restrictive dermopathy. Homozygous or compound heterozygous partial loss-of-function mutations cause mandibuloacral dysplasia type B or clinically similar progeroid disorders. In these patients, full-length farnesylated prelamin A accumulates to varying extents. The most prominent clinical features in these patients are hypoplasia of the mandible and clavicles, acro-osteolysis, and lipodystrophy. The patient with an LMNA point mutation that abolished the ZMPSTE24 recognition site had very similar clinical features (26).

Considerable evidence indicates that farnesylated prelamin A and its farnesylated variants are responsible for pathology in HGPS and the other processing-deficient progeroid laminopathies. Cultured cells from patients with these disorders and genetically modified mouse models have abnormal nuclear morphology with blebbing of the nuclear envelope (28-34). Treatment of these cultured cells, as well as transfected cells expressing progerin, with protein farnesyltransferase inhibitors or a statin plus an aminobisphosphonate that reduce protein prenylation restores normal nuclear shape (31-36). More significantly, Fong et al. (37) originally showed that treatment of Zmpste24 null mice with a protein farnesyltransferase inhibitor reversed profound progeroid phenotypes and prolonged survival. Since then, several other studies have confirmed the beneficial effects of blocking protein farnesylation in mouse models of HGPS and in mice with ZMPSTE24 deficiency (38-40). Human clinical trials have also shown beneficial effects of the farnesyltransferase inhibitor lonafarnib in children with HGPS (41, 42). In 2020, the United States Food and Drug Administration approved lonafarnib for the treatment of HGPS and other processing-deficient progeroid laminopathies (42).

While the accumulation of prelamin A or variants, not loss of lamin A function, is clearly responsible for the molecular pathology in processing-deficient progeroid laminopathies, the downstream mechanistic defects these farnesylated proteins induce are less well understood. Fibroblasts from patients with HGPS have decreased viability and increased apoptosis under repetitive mechanical strain, suggesting that accumulation of progerin may lead to the death of certain cells exposed to stress (43). Unprocessed prelamin A and progerin also perturb DNA damage repair responses, resulting in genomic instability (44). Progerin induces DNA replication fork stalling and nuclease-mediated fork degradation, causing replication stress and genomic instability, which are accompanied by upregulation of the cGAS/STING pathway and activation of an interferon-like innate immune response (45). Progerin and farnesylated prelamin A may also bind to and directly modulate the functions of proteins involved in DNA repair or other nuclear processes. For example, in induced pluripotent stem cells from patients with HGPS, progerin binds to DNA-dependent protein kinase catalytic subunit, which is known to be involved in DNA repair and various aging-related cellular events (46). Despite intensive investigation, exactly how and to what extent these or other pathogenic mechanisms may contribute to cellular dysfunction in processing-deficient progeroid laminopathies remain to be firmly established.

Point mutations in LMNA that do not cause farnesylated prelamin A or variant accumulation can also cause progeroid disorders. Mandibuloacral dysplasia type A is caused by a homozygous arginine-to-histidine amino acid substitution at residue 527 (8). These patients have similar symptoms to those with ZMPSTE24 partial loss-of-function mutations. Dominant LMNA mutations have also been described in patients diagnosed with atypical Werner syndrome (47). Several other homozygous, heterozygous, and compound heterozygous LMNA mutations leading to amino acid substitutions have further been linked to progeroid disorders (24). The dominant amino acid substitutions tend to cluster in structured regions of lamin A/C, including the rod domains conserved among intermediate filament proteins and an immunoglobulin-like fold domain in the tail domain. The heterozygous amino acid substitutions mostly cluster in a portion of the immunoglobulin-like fold domain and disrupt the interaction of lamin A/C with barrier-to-autointegration factor (48). A homozygous point mutation in the gene encoding barrier-to-autointegration factor also causes a progeroid disorder, and the resulting amino acid substitution disrupts binding to lamin A/C (49). Barrier-to-autointegration factor was originally discovered as a host protein that prevents a DNA copy of a retroviral genome from integrating into itself. It was subsequently found to be involved in processes that protect genome integrity such as nuclear envelope reformation at the end of mitosis, repair of ruptured envelopes, and the DNA damage response (50). Disruption of barrier-to-autointegration factor binding to lamin A may alter its normal function. Hence, as in processing-deficient laminopathies, genomic instability may be a downstream cellular defect in processing-proficient laminopathies.

Striated Muscle Disease

The first identified pathogenic mutations in LMNA segregated in affected family members with autosomal dominant Emery-Dreifuss muscular dystrophy (4). The clinical diagnosis of Emery-Dreifuss muscular dystrophy is based on distinctive features of early joint contractures, humeroperoneal wasting, and weakness and cardiomyopathy with early conduction defects (51). Subsequent studies showed that LMNA mutations cause dilated cardiomyopathy with other muscle groups affected, such as limb-girdle muscular dystrophy, or even minimal to no skeletal muscle pathology (52-56). In most of these disorders, the age of onset is usually in later childhood or early adulthood, but LMNA mutations also sometimes cause congenital muscular dystrophy that presents in the first year of life with the eventual development of cardiomyopathy (57).

Dominant LMNA mutations that cause striated muscle disease lead to single amino acid substitutions, small deletions, RNA splicing defects, or haploinsufficiency. Extremely rare heterozygous mutations have been described (55, 58). Striated muscle disease-causing LMNA mutations likely lead to loss of some aspect of lamin A/C function, as demonstrated in Lmna null mice (59). Fibroblasts from these mice, as well as from patients with LMNA mutations and striated muscle disease, have altered nuclear morphology (59, 60). Mouse fibroblasts lacking lamin A/C have increased nuclear deformation, defective mechanotransduction, and impaired viability under mechanical strain (61). These fibroblasts as well as transfected cells expressing lamin A variants that cause muscle disease also have defective movement of the nucleus when polarizing for migration (62). Cryoelectron tomography analysis of mouse fibroblasts homozygous for an Lmna point mutation that causes striated muscle disease shows apparently unaltered organization of the lamin filaments but increased nuclear surface area, reduced heterochromatin, and increased lamin B1 and B2 expression (63).

Loss of lamin A/C leads to mislocalization of the integral inner nuclear membrane protein emerin to the bulk endoplasmic reticulum (59). Expression of some muscle disease–associated point mutant lamin A variants also leads to a partial mislocalization of emerin (64, 65). Emerin and lamin A/C directly interact (66). Mutations in the gene encoding emerin cause phenotypically near-identical disease to the striated muscle disease caused by LMNA mutations (1). Both lamin A/C and emerin interact with lamina-associated polypeptide (LAP) 1, another transmembrane protein of the inner nuclear membrane (67). Mutations in the gene encoding LAP1 also cause muscular dystrophy and cardiomyopathy in humans, and deletion of the protein from mouse skeletal and cardiac muscle, respectively, causes these phenotypes (67-69). These findings suggest that a complex of lamin A/C, emerin, and LAP1 has a critical role in striated muscle maintenance.

Adipose Tissue Disease

In 1974, Dunnigan and colleagues (70) described a dominantly inherited form of partial lipodystrophy with loss of subcutaneous fat from the limbs and trunk at around the onset of puberty. About 25 years later, three groups reported mutations in LMNA in individuals with so-called Dunnigan-type familial partial lipodystrophy (5-7). Subsequent to the peripheral fat loss, patients develop insulin resistance, diabetes mellitus, hypertriglyceridemia, and associated complications such as hepatic steatosis and steatohepatitis (5-7, 71, 72). LMNA mutations can also cause atypical lipodystrophic syndromes different from the Dunnigan-type (73).

About 90% of the mutations causing Dunnigan-type patient lipodystrophy are located in LMNA exon 8. These mutations create amino acid substitutions that lead to a diminution of the conserved positively charged character of a solvent-exposed surface in the lamin A/C immunoglobulin-like fold domain (74, 75). In contrast, amino acid substitutions in the same region that cause striated muscle disease disrupt the overall structure of the immunoglobulin-like fold domain (Figure 4). This suggests that LMNA mutations causing partial lipodystrophy destroy a positively charged interaction site for a lamin A/C binding partner important in adipocyte function. This portion of lamin A/C binds to the transcription factor sterol response element binding protein 1, and some data suggest that its transcriptional activity is altered by a lipodystrophy-causing lamin A variant (76, 77). Overexpression of lamin A in preadipocytes inhibits lipid accumulation, triglyceride synthesis, and expression of adipogenic markers, whereas embryonic fibroblasts lacking lamin A/C accumulate more intracellular lipid (78).

Figure 4.

Amino acid substitutions that cause Dunnigan-type familial partial lipodystrophy diminish the positivity of a surface of the lamin A/C immunoglobulin-like fold domain, while those in the same region causing striated muscle disease disrupt overall structure. (a) Localization of the amino acid substitutions (red) within the immunoglobulin-like fold domain causing striated muscle disease. (b) Three amino acid substitutions (green) causing Dunnigan-type familial partial lipodystrophy at R482, G465, and K486 that do not affect overall fold domain structure but diminish the positive charge of a solvent-exposed surface. In the disease-causing variants, glycine (neutral) at residue 465 is replaced by an aspartic acid (negative); arginine (positive) at residue 482 is replaced by a glutamine, tryptophan, or leucine (all neutral); or lysine (positive) at residue 486 is replaced by an asparagine (neutral). Figure adapted with permission from Reference 75.

Peripheral Neuropathy

An LMNA mutation generating an arginine-to-cysteine amino acid substitution at position 298 in the rod domain of lamin A/C causes autosomal recessive peripheral neuropathy (9). Individuals with this Charcot-Marie-Tooth disease type 2B1 suffer from an axonal neuropathy with variability in the age of onset and the course of the disease (79). Sciatic nerves of Lmna null mice have a reduction of axon density, axonal enlargement, and nonmyelinated axons similar to phenotypes of human peripheral axonal neuropathies (9). However, mice homozygous for the corresponding human point mutation affecting residue 298 do not develop a detectable peripheral neuropathy phenotype (80).

Mutations in Genes Encoding B-Type Lamins

B-type lamins are encoded by two independent genes in humans, LMNB1 and LMNB2. Lamin B1 and lamin B2 are expressed in nearly every cell type from early stages of development. Deficiency in lamin B1 or lamin B2 in mice leads to defects in neuronal migration and layering within the cerebral cortex and cerebellum, indicating the crucial roles of these lamins in the developing brain (81, 82). Consistent with these phenotypes in knockout mice, de novo mutations in LMNB1 have been identified in seven individuals with primary microcephaly (83). Another study reported that mutations in LMNB1 and LMNB2 cause syndromic microcephaly (84). Padiath et al. (85) reported that duplications of LMNB1 cause autosomal dominant leukodystrophy, an adult-onset demyelinating disorder. Overexpression of lamin B1 in BAC transgenic mice results in aberrant myelin formation, axonal degeneration, demyelination, and cognitive and motor defects (86). B-type lamins, although widely expressed, appear to have special functions in the central nervous system.

LAMINOPATHIES INVOLVING ENVELOPE PROTEINS OTHER THAN LAMINS OR ZMPSTE24

Mutations in genes encoding nuclear envelope proteins other than lamins or the prelamin A processing enzyme ZMPSTE24 also cause rare inherited diseases (Table 2). Several of these mimic the cardiac and skeletal muscle diseases caused by LMNA mutations. Others affect bone or multiple organ systems.

Table 2.

Genes encoding nuclear envelope proteins other than lamins linked to human diseases

Nuclear envelope compartment	Gene	Protein	Affected tissues	Disease
Inner nuclear transmembrane	EMD	Emerin	Striated muscle	X-linked Emery-Dreifuss muscular dystrophy and related myopathies/cardiomyopathy
	LEMD3	MAN1	Bone Skin	Osteopoikilosis Nonsporadic melorheostosis Buschke-Ollendorff syndrome
	LBR	LBR	Neutrophils Bone Multisystem	Pelger-Huet anomaly Greenberg skeletal dysplasia
	TMEM43	LUMA	Striated muscle	Arrhythmogenic right ventricular dysplasia
	TOR1AIP1	LAP1	Striated muscle Multisystem	Cardiomyopathy/muscular dystrophy Multisystemic abnormalities with progressive neurological degeneration and early death
	LEMD2	NET25 (LEM2)	Multisystem	Progeria-like facial phenotypes with other developmental abnormalities
	SUN1/SUN2	SUN1/SUN2	Striated muscle	Muscular dystrophy/cardiomyopathy
Outer nuclear transmembrane	SYNE1	Nesprin-1	Central nervous system Striated muscle	Autosomal recessive spinocerebellar ataxia 8 Arthrogryposis multiplex congenita Muscular dystrophy/cardiomyopathy
	SYNE2	Nesprin-2	Striated muscle	Cardiomyopathy/muscular dystrophy
	SYNE4	Nesprin-4	Inner ear	High-frequency hearing loss
Perinuclear space	TOR1A	TorsinA	Central nervous system	DYT1 dystonia

Emerin

As discussed above, mutations in EMD encoding emerin, which interacts with lamin A/C and LAP1, cause X-linked Emery-Dreifuss muscular dystrophy (1). Virtually all these mutations cause a loss of emerin expression (2, 3, 87, 88). The clinical symptoms of X-linked Emery-Dreifuss muscular dystrophy are almost identical to those of the autosomal form caused by LMNA mutations. As with LMNA mutations that cause striated muscle disease, dilated cardiomyopathy with skeletal muscle involvement different than classical Emery-Dreifuss can occur with loss of emerin (89, 90). Deletion of emerin, deletion of lamin A/C, and cardiomyopathy-causing LMNA mutations lead to some of the same cell signaling defects, such as ERK1/2 activation (91-93).

Two independent laboratories generated emerin knockout mouse lines. Unlike human patients, these emerin-deficient mice display minimal to no muscular dystrophy or cardiomyopathy phenotypes (94, 95). However, there is a significant excess of LAP1 relative to emerin in mouse skeletal muscle compared with that of human. In mice, loss of LAP1 from skeletal muscle causes pathology, and combined deletion of emerin significantly exacerbates it (67). This suggests that LAP1, which binds to emerin, may compensate for its depletion from mouse striated muscle.

MAN1

MAN1 was originally identified by autoantibodies from a patient with a collagen vascular disease as a nuclear envelope protein that cofractionated with nuclear lamins (96, 97). It has two transmembrane segments with nucleoplasmic amino- and carboxyl-terminal domains. The amino-terminal region contains a LAP2-emerin-MAN1 (LEM) domain, a globular module for approximately 40 amino acids common to several inner nuclear membrane proteins. A genome-wide linkage analysis identified heterozygous loss-of-function mutations in LEMD3 encoding MAN1 in families with osteopoikilosis, nonsporadic melorheostosis, and Buschke-Ollendorff syndrome, disorders of excessive bone growth sometimes with skin abnormalities (98). Experiments in cultured cells and knockout mice clearly showed that MAN1 antagonizes bone morphogenic protein and transforming growth factor-β signaling by binding to Smad2 and Smad3 (98-102). Subsequent research determined the structural basis of MAN1 binding to Smad2, Smad3, and their inactivating phosphatase PPM1A (103). Hence, MAN1 provides an inner nuclear membrane-localized scaffold for inhibiting Smad2/Smad3-mediated signaling. This inhibitory mechanism is apparently most prominent in bone and to some extent skin cells, given the phenotypes of patients with loss-of-function mutations.

LBR

LBR is a polytopic integral protein of the inner nuclear membrane that interacts with B-type lamins and heterochromatin proteins (104-106). It is also homologous to sterol reductases of the endoplasmic reticulum (107). Hoffmann et al. (108) identified that heterozygous mutations in the LBR gene cause the benign Pelger-Huet anomaly, characterized by hypolobulation and altered chromatin structure of neutrophil nuclei. Subsequently, Waterham et al. (109) identified a homozygous LBR truncation mutation in patients with Greenberg dysplasia, a perinatal lethal syndrome that affects bone and other organ systems. A later study of three fetuses with Greenberg dysplasia identified mutations in LBR that resulted in the loss of sterol reductase activity (110). Tsai et al. (111) reported that disease-causing LBR mutations perturb LBR’s ability to engage in cholesterol synthesis, with some mutations possibly leading to defective binding of a cofactor necessary for enzymatic activity and others causing degradation. Ichthyosis mice carry a spontaneous mutation in Lbr and have multiple abnormalities, including alopecia, syndactyly, hydrocephalus, and neutrophil nuclear morphology similar to human Pelger-Huet anomaly (112). Mice with an Lbr gene trap mutation are phenotypically similar to ichthyosis mice (113).

LUMA

LUMA is an integral protein of the inner nuclear membrane that contains four transmembrane segments and is expressed in all or most cell types (114). A positional cloning study of 15 families with arrhythmogenic right ventricular dysplasia identified heterozygosity for a serine-to-leucine missense mutation at codon 358 in TMEM43 encoding LUMA (115). Germline Tmem43 null mice and knock-in mice with the pathogenic serine-to-leucine amino acid substitution in LUMA have normal cardiac function (116). However, transgenic mice overexpressing a human pathogenic LUMA variant in cardiomyocytes die at young ages and recapitulate aspects of the human disease, including cardiomyocyte death and severe fibrofatty replacement (117).

LAP1

LAP1, encoded by TOR1AIP1, was originally identified in rat liver extracts as an integral membrane protein of the inner nuclear membrane associated with the nuclear lamina (118). Further investigation showed that it binds to lamin A/C and lamin B1 (119). There are at least two isoforms expressed from TOR1AIP1 in humans (120). LAP1 has a nucleoplasmic amino-terminal domain, a single transmembrane segment, and a carboxyl-terminal domain in the perinuclear space (121). Within the perinuclear space, LAP1 interacts with and is necessary to activate the AAA+ ATPase torsinA (122-124). In addition to binding to nuclear lamins, LAP1 interacts with emerin in the nucleoplasm (67). In mice, it is essential for postnatal skeletal muscle development and maintenance and proper cardiac function (67, 68, 125). Recessive mutations in TOR1AIP1 that disrupt the LAP1B isoform cause cardiomyopathy and muscular dystrophy (69). TOR1AIP1 mutations leading to a combined loss of both LAP1 isoforms causes multisystem disease with severe progressive neurological impairment, bilateral cataracts, growth retardation, and early lethality (126).

NET25

NET25 (also called LEM2) was identified by a subtractive proteomic study and is encoded by the LEMD2 gene (14). NET25 is structurally related to MAN1 and contains two transmembrane segments and a LEM domain in its amino-terminal region (127). It recruits ESCRT to repair ruptured interphase nuclei and to promote nuclear envelope reformation in mitosis (128, 129). NET25 is also required for proper myogenesis (130). Two collaborating research groups have identified a de novo missense LEMD2 mutation in two individuals with progeroid facial phenotypes and neurological anomalies (131).

SUN Proteins

SUN proteins are integral inner nuclear membrane protein components of the LINC complex. Sequence variants in SUN1 and SUN2 have been reported in patients with muscular dystrophy and cardiomyopathy; however, segregation with disease in the affected families has not been demonstrated (132).

Nesprins

Nesprins are integral proteins localized to the outer nuclear membrane (there are also some small inner nuclear membrane–localized isoforms) that along with SUN proteins form the core LINC complex. SYNE1 encodes nesprin-1, which has several isoforms that interact with actin and microtubules. Deletion mutations in SYNE1 cause a recessively inherited cerebellar ataxia (133). A homozygous splice site mutation in SYNE1 causes recessive arthrogryposis multiplex congenita, a disorder characterized by congenital joint contractures and reduced fetal movements (134).

Studies from genetically modified mice implicate nesprin-1 and nesprin-2 function in cardiac and striated muscle function (135-137). Mutations in SYNE1 and SYNE2 encoding these human proteins have been associated with cardiomyopathy and muscular dystrophy; however, segregation within affected family members has not been demonstrated for SYNE1 (138-140). Horn et al. (141) identified a homozygous truncating mutation in SYNE4 encoding nesprin-4, an outer nuclear membrane LINC complex protein expressed in the hair cells of the inner ear. They further showed that Syne4 null mice, as well as mice lacking SUN1, have progressive hearing loss. In these mice, cochlear outer hair cells are formed but degenerate as hearing matures, while the inner hair cells remain intact. This result suggests that nucleocytoplasmic connections mediated by the LINC complex are essential for the viability of the outer hair cells.

TorsinA

A dominantly inherited in-frame TOR1A mutation leading to a single glutamic acid deletion in torsinA causes DYT1 dystonia (142). Wild-type torsinA is mainly localized throughout the endoplasmic reticulum, but the dystonia-causing variant is preferentially localized in the perinuclear space (143). TorsinA is an AAA+ ATPase that is inactive unless it binds to LAP1 in the perinuclear space or to LULL1 in the bulk endoplasmic reticulum (122, 123). Although the dystonia-causing glutamic acid deletion is not located in the active site of torsinA, its deletion compromises binding to LAP1/LULL1, leading to significantly diminished ATP hydrolysis activity (122, 144).

Transgenic mice overexpressing the dystonia-causing torsinA variant in neurons exhibit abnormal involuntary movement defects and have perinuclear inclusion bodies that contain ubiquitin, lamin A, and torsinA (145). Germline deletion of Tor1a in mice causes perinatal lethality associated with abnormal nuclear membrane morphology of neurons (146). Various conditional deletions of Tor1a in subsets of neurons cause dystonic movements in mice, along with selective neurodegeneration; these results demonstrate a cell-autonomous function of torsinA in neurons (147, 148). Lack of torsinA function appears to lead to neuronal defects during development, as deleting it in mouse embryos causes dystonia but deleting it from adult mice leads to no abnormalities. Conversely, restoring torsinA in juvenile DYT1 mice rescues motor defects, but there is no benefit from torsinA repletion in adult mice (149). Depletion of torsinA or its activator LAP1 from hepatocytes surprisingly causes reduced hepatic very-low-density lipoprotein secretion and steatosis (150). However, no genome-wide association studies so far have linked genes encoding these proteins to lipid metabolism defects or fatty liver disease in humans.

Nuclear Pore Complex Proteins

NPCs are macromolecular structures composed of multiple copies of approximately 30 distinct proteins, most of which are called nucleoporins (151). NPCs mediate the passive and active transport between the nucleus and the cytoplasm, but growing evidence indicates that the NPCs have transport-independent roles including cell differentiation, cell cycle progression, gene expression, and epigenetic regulation (152). Many of these roles appear to be cell type specific, and pore complex composition may vary between cells, as mutations in genes encoding constituent proteins cause diseases involving specific organs (Table 3). A detailed discussion of these disorders is beyond the scope of this review on laminopathies; however, they have been reviewed elsewhere (153, 154).

Table 3.

Genes encoding nuclear pore complex proteins linked to human diseases

Gene	Protein	Affected tissues	Disease
AAAS	Aladin	Autonomic nervous system	Triple-A syndrome
NUP155	Nup155	Heart	Atrial fibrillation
NUP62	Nup62	Central nervous system	Infantile bilateral striatal necrosis
GLE1	GLE1	Motor neurons	Lethal congenital contracture syndrome-1
RanBP2/NUP358	RanBP2 (Nup358)	Central nervous system	Acute necrotizing encephalopathy
NUP107	Nup107	Kidney	Galloway-Mowat syndrome
NUP107, NUP93, NUP205, NUP85, NUP133, NUP160	NUP107, NUP93, NUP205, NUP85, NUP133, NUP160	Kidney	Steroid-resistant nephrotic syndrome

IN SEARCH OF PATHOGENIC MECHANISMS

As we have already discussed, the molecular and cellular consequences of disease-causing mutations in several genes encoding nuclear envelope proteins are understood. For example, in HGPS, there is an accumulation of a farnesylated prelamin A that is likely responsible for abnormal cellular function. Pathogenic lamin A/C variants in Dunnigan-type partial lipodystrophy have an alteration in the surface charge of part of the proteins. However, the mechanisms of how alterations in the nuclear envelope proteins, most of which are expressed in multiple cell types, lead to pathology affecting specific tissues or organ systems remain for the most part enigmatic. In only a few cases, such as mutations in SYNE4 causing hearing loss, can cell type–specific expression of the gene explain the organ pathology (141).

The field has to a large extent focused on two general hypotheses to attempt to explain the tissue-selective nature of most laminopathies. One hypothesis is that the nuclear envelope regulates cell-specific transcription and that alterations in its structure lead to pathogenic changes in gene expression. The second hypothesis is that defects in the nuclear envelope make cells susceptible to damage by mechanical stress; this premise has often been invoked to explain the fact that striated muscle is often affected in laminopathies. This so-called mechanical stress hypothesis has gained traction as research into the role of the LINC complex in cellular force transduction has expanded. Some investigators have combined these two hypotheses, suggesting that increased sensitivity of cells with nuclear envelope defects to mechanical stress leads to increased activation of stress-responsive signaling pathways.

The most clear-cut example of altered gene expression as a pathogenic mechanism resulting from mutations in a gene encoding a nuclear envelope protein is the case of MAN1. As discussed above, MAN1 functions as an inner nuclear membrane scaffold to deactivate Smad2 and Smad3. Loss of this inhibitory protein leads to increased transforming growth factor-β signaling (98-103). Transforming growth factor-β is a potent stimulant of bone formation (155). Osteopoikilosis, Buschke-Ollendorff syndrome, and nonsporadic melorheostosis caused by mutations in LEMD3 encoding MAN1 are characterized by increased bone density, a phenotype consistent with enhanced expression and repression of genes regulated by transforming growth factor-β. However, it is not entirely clear why MAN1 plays a significant role in inhibiting transforming growth factor-β signaling only in bone and in some cases skin, as other tissues are not affected by mutations leading to its partial loss of function.

For nearly four decades, investigators have hypothesized that the nuclear envelope functions in the maintenance and the alterations of the three-dimensional structure of the genome during development, differentiation, and the cell cycle (156). There are numerous reports of altered chromatin organization and gene expression driven by disease-causing alterations in nuclear lamins and other nuclear envelope proteins. However, there are no robust examples of how alterations in the nuclear envelope in a laminopathy lead to chromatin structural changes that directly influence the expression of any single gene or group of genes involved in pathogenesis. Nonetheless, the hypothesis that disease-associated defects in the nuclear envelope alter chromatin in a way that directly leads to pathogenic gene expression changes remains prevalent in the field.

Since early reports that cells lacking lamin A/C have altered mechanical properties (61), many studies have focused on the hypothesis that stress-induced cellular damage or altered responses to mechanical strain underlie pathology in laminopathies. One study has demonstrated that in HGPS, progerin expression combined with mechanical stress promotes arterial smooth muscle cell death that is reduced by disruption of the LINC complex (157). Similarly, skeletal muscle cells with disease-causing alterations in lamin A/C have reduced nuclear stability and nuclear envelope damage, both of which are reduced by disruption of the LINC complex (158). Migrating fibroblasts, neurons, and myonuclei with alterations in lamins also have increased nuclear envelope rupture, DNA damage, and cell death (158-160). Increased susceptibility of cardiomyocytes to continuous mechanical strain may also explain the increased activation of stress-induced signaling pathways such as ERK1/2 and AKT/mTOR in hearts of mice with cardiomyopathy-causing Lmna mutations (91, 161). Abnormal activation of these pathways has detrimental effects on heart structure and function, whereas blocking these pathways has beneficial effects (161, 162).

Space limitations for this review prohibit us from citing all of the publications on testing the mechanical stress and gene expression hypotheses. Despite all of the research addressing these two hypotheses, neither one in and of itself is entirely satisfying. Alterations in both cell stress responses and gene expression may occur simultaneously to cause pathology. Other hypotheses must be proposed and tested as well. Mutations in different laminopathy-associated genes, or even different mutations in the same gene such as LMNA, clearly have different consequences in different cell types. Hence, no single hypothesis will likely explain all of the diverse laminopathies.

A major limitation of much of the research to date has been that data obtained in cell culture models, often fibroblasts from affected patients or transfected cells, cannot always be readily applicable to what is observed in affected tissues and organs. Nonetheless, studies in model cell systems are essential, as it is difficult to assess the effects of mechanical stress or certain other insults on intact organs. Measured gene expression alterations in animal or patient tissues also may not reflect the direct effect of a laminopathy-causing gene mutation but rather may indicate secondary consequences, such as inflammation or fibrosis. Therefore, future research more closely combining cell culture and in vivo approaches will help move the field forward. Furthermore, physicians who care for patients with laminopathies and pathologists who can appreciate the underlying tissue and organ dysfunction must interact more with basic scientists to assure that phenomena observed in cultured cells and even small animal models are relevant to what occurs in affected humans. To accomplish this interaction, disciplinary barriers will need to be broken down, and collaborations in which recognition and grant funding are shared will need to increase. Broader interdisciplinary research may be the only approach to elucidate the mechanisms underlying the broad range of diseases caused by mutations in genes encoding proteins of a fascinating structure common to virtually all eukaryotic cells.