PNPLA6 disorders: what’s in a name?

Abstract

Background:

Variants in the Patatin like phospholipase domain containing 6 (PNPLA6) gene cause a broad spectrum of neurological disorders characterized by gait disturbance, visual impairment, anterior hypopituitarism, and hair anomalies. This review examines the clinical, cellular, and biochemical features found across the five PNPLA6-related diseases, with a focus on future questions to be addressed.

Materials and Methods:

A literature review was performed on published clinical reports on patients with PNPLA6 variants. Additionally, in vitro and in vivo models used to study the encoded protein, Neuropathy Target Esterase (NTE), are summarized to lend mechanistic perspective to human diseases.

Results:

Biallelic pathogenic PNPLA6 variants cause five systemic neurological disorders: spastic paraplegia type 39, Gordon-Holmes, Boucher-Neuhäuser, Laurence-Moon, and Oliver-McFarlane syndromes. PNPLA6 encodes NTE, an enzyme involved in maintaining phospholipid homeostasis and trafficking in the nervous system. Retinal disease presents with a unique chorioretinal dystrophy that is phenotypically similar to choroideremia and Leber congenital amaurosis. Animal and cellular models support a loss-of-function mechanism.

Conclusions:

Clinicians should be aware of choroideremia-like ocular presentation in patients who also experience growth defects, motor dysfunction, and/or hair anomalies. Although NTE biochemistry is well-characterized, further research on the relationship between genotype and the presence or absence of retinopathy should be explored to improve diagnosis and prognosis.

Introduction

Patatin like phospholipase domain containing 6 (PNPLA6) is a protein classified in a nine membered serine hydrolase patatin-like phospholipase family that encodes neuropathy target esterase (NTE). The gene is located on human chromosome 19p13.2 and contains 37 exons. To date, there are at least five splice variants, where the canonical sequence (NM_001166111.2, NP_001159583.1) contains 1375 amino acids, and has a molecular weight of 151 kDa (1, 2). PNPLA6 contains three protein domains: a single pass N-terminal transmembrane domain, three cyclic nucleotide binding (CNB) domains, and a patatin-like catalytic domain (NEST) located near the C-terminus. In particular, the name “patatin” is derived from the major glycoprotein found in potato tubers, which acts as a lipid acyl hydrolase (3). The patatin-like domain is evolutionary conserved across multiple species, ranging from E. Coli to humans (2).

NTE is a membrane protein located on the cytoplasmic side of ER, where it plays a role in phospholipid homeostasis, membrane trafficking, and axonal integrity in cells (4–6). It is classified as a serine hydrolase that functions as a phospholipase B, where it catalyzes the deacylation of glycerophospholipids by evicting two fatty acids from the glycerol backbone (4). NTE has also been shown to have strong lysophospholipase activity, where it can catalyze the eviction of one fatty acid at the sn-1 position of lysophospholipid species ten times faster than known small mammalian cytosolic lysophospholipases (4, 7). NTE plays a substantial role in phosphatidylcholine (PC) homeostasis, where it is involved in the CDP-choline (Kennedy) pathway (8).

PNPLA6 is ubiquitously expressed in adult human tissue (1). In the developing human embryo, PNPLA6 is expressed throughout the central nervous system (CNS) and eye. Stained embryonic tissue taken from the first trimester (Carnegie stages 19 and 23) showed PNPLA6 expression in the neural retina, retinal pigment epithelium (RPE), choroid, anterior and posterior pituitary, cerebellum, and the ventricular zones (1). In postnatal mice, PNPLA6 is also widely expressed throughout the brain and eye (1, 9). In contrast, adult mice have restricted expression that localizes to the cell bodies of Purkinje cells in the cerebellum, as well as the hippocampus and choroid plexus (9, 10). Immunostaining of both developing and adult mouse retina showed expression of NTE in horizontal, amacrine, and photoreceptor cells, with adult retina showing strong expression specifically in cones (11). Another animal model used to study PNPLA6 expression is Drosophila, where the orthologue of PNPLA6, swiss-cheese (sws), is expressed in all stages of development in neurons and glia at lower levels compared to adults (6). Expression of NTE can also be seen around photoreceptor nuclei and along the entire photoreceptor in adult flies (11). Overall, expression of PNPLA6 is consistent through diverse species and is highly present in the central nervous system.

Discovery of PNPLA6/NTE

In 1930, the United States experienced an epidemic colloquially called “Jake Leg” or “Jake Walk” that affected thousands of victims in the mid- and southwestern United States to experience paralysis in the distal muscles of the lower and upper extremities (12). It was determined that the cause of the paralysis was due to a contaminated alcoholic beverage with a Jamaican ginger extract that contained tri-ortho-cresyl phosphate (TOCP). Subsequent smaller incidents in other countries precipitated research into the cause of TOCP related paralysis (13). Finally, research in the 1970’s by M.K. Johnson and colleagues found that the cause of this paralysis, termed Organophosphate induced-delayed neurotoxicity (OPIDN), was caused by the inhibition of a serine hydrolase abundant in the chicken brain (14–17). This enzyme, which was sensitive to mipafox (Neuropathic OP compound) but resistant to paraoxon (Non-neuropathic OP compound), was termed Neuropathy target esterase.

PNPLA6 clinical presentation

PNPLA6 was cloned as the human gene encoding NTE in 1998 (18). In 2008, PNPLA6 was discovered as the gene that encoded NTE (19). Subsequently, several clinical and research articles have been published highlighting individuals or families that are afflicted with PNPLA6 variants. As of May 2023, 95 individuals have been published with biallelic PNPLA6 variants. Patients with biallelic PNPLA6 variants show varying clinical characteristics that mainly affect tissues related to the nervous system. These characteristics include gait disturbance, visual impairment, anterior hypopituitarism, and hair anomalies. The section below summarizes clinical characteristics in the context of PNPLA6-opathies.

Gait disturbances encompass one or a combination of the following clinical manifestations: ataxia, spasticity, and peripheral neuropathy. In particular, the cerebellum, the area of the brain that controls coordination and balance, undergoes atrophy/ataxia in up to 90% of affected patients published thus far, making it a good marker for identifying PNPLA6 associated gait disturbances (19–21). Other markers of ataxia can also include dysmetria, dysarthria, nystagmus, and gait or postural ataxia. Patients with PNPLA6 variants may also experience forms of spasticity, which typically impairs the upper motor neurons, with paresis occurring in the lower extremities. Other markers of spasticity can also include Babinski signs or hyperreflexia. Lastly, sensory disturbances can arise through peripheral neuropathies that usually involve degeneration of the sensory or motor axons. PNPLA6 individuals typically lose either sensation or motor coordination in their distal leg and hand muscles, where symptoms can become progressively worse over time. The progression of the gait disturbance can present as early as 1 year of age (22–24) to as late as 55 years of age (25).

Visual impairment deficits in patients with PNPLA6 variants are usually caused by a unique chorioretinal dystrophy that presents as abnormal pigment clumping under fundus imaging (Figure 1). This phenomenon causes atrophy of the choroidal blood vessels that leads to the loss of integrity of the choroidal layer, as well as the retinal pigment epithelium (RPE) (1). ERGs, retinal imaging, and other visual field tests can confirm the presence of chorioretinal dystrophy found in individuals with PNPLA6 variants. The progression of visual impairment can present as early as 1 year of age (20, 24) to as late as 64 years of age (26).

Figure 1. Clinical symptoms associated with OMCS.

A-B) Fundus imaging of patient diagnosed with OMCS. Fundus shows chorioretinal atrophy with signs of variable pigment findings. C-D) Fundus autofluorescence images of A,B. E-F) Characteristic long eyelashes that is a one of the hallmark signatures of patients with OMCS.

Anterior hypopituitarism can also be seen in PNPLA6 disorders. Through hormonal testing, patients usually experience a decrease in secretion of one or more hormones such as Follicle Stimulating hormone (FSH) or Luteinizing hormone (LH). Anterior hypopituitarism manifesting at an early age can result in developmental delay, short stature, and cognitive impairment. Later in adolescence, lack of growth hormone causes delayed puberty and lack of secondary sexual characteristics (micropenis, primary amenorrhea), also known as hypogonadotropic hypogonadism. The progression of anterior hypopituitarism can present at birth (1, 23) or as late as 25 years of age (26–30).

Lastly, a small subset of patient’s experience hair anomalies, which include long eyelashes (trichomegaly), bushy eyebrows, premature graying, or scalp alopecia (1, 21). These symptoms can usually be easily seen by routine physical examination of the patient. The progression of hair anomalies can present at birth (1) or as late as 18 years of age (31). Overall, patients with biallelic PNPLA6 variants may exhibit one or more of the above common clinical aberrations, with variable ages of onset for each affected tissue.

Individuals with biallelic PNPLA6 variants have been typically diagnosed into five distinct neurological disorders: Spastic paraplegia type 39 (SPG39), Gordon-Holmes (GDHS), Boucher-Neuhäuser (BNHS), Laurence-Moon (LNMS), and Oliver-McFarlane syndromes (OMCS). These neurological disorders will encompass at least one or more of the listed clinical characteristics described previously (Figure 2). Table 1 summarizes the symptoms presented in each syndrome and their relative age of onsets.

Figure 2. Summary of syndromes associated with PNPLA6 variants.

Overview of the prevalent clinical manifestations found in patients with biallelic PNPLA6 variants.

Table 1.

Phenotypic presentation in PNPLA6 related syndromes.

Phenotypic presentation	PNPLA6 Disorder
	SPG39	GDHS	BNHS	LMNS	OMCS
Cerebellar Ataxia	+	+	±	+	±
Peripheral Neuropathy	±	±	±	±	±
Hypogonadotropic Hypogonadism		+	+	+	+
Chorioretinal dystrophy			+	+	+
Hair anomalies					+

Adapted from Synofzik et al. (21). “+”: present; “±”: variable presentation; unfilled cells represent absent phenotype.

Individuals diagnosed with SPG39 exhibit forms of peripheral neuropathy, spastic paraplegia, and cerebellar ataxia. Patients who have done neurological imaging and testing usually present with cerebellar atrophy (19, 32). Patients diagnosed with SPG39 can experience some or all the previously noted symptoms. Interestingly, the onset of gait disturbances in patients with SPG39 can precipitate in early childhood to as late as adulthood (Data shown in chapter 2). The first symptom that presents is usually some form of CNS phenotype, such as ataxia or spasticity. For most individuals, symptoms become progressively worse over time. Several articles have noted that the progression of gait disturbances seen in patients with SPG39 does not severely affect their motor ability, with some patients recorded to date having the ability to walk with either none or some semblance of aid well into adulthood (19, 32). Since individuals with SPG39 only experience gait disturbances, it is considered the mildest or least severe form of the PNPLA6-opathies.

Individuals diagnosed with GDHS and BNHS exhibit forms of gait disturbance, visual impairment, and anterior hypopituitarism. GDHS is typically classified with cerebellar ataxia, hypogonadotropic hypogonadism, and to a lesser extent peripheral neuropathy (33). BNHS is also characterized by the same indications as GDHS, but with the addition of chorioretinal dystrophy leading to vision impairment (20, 34). However, not all individuals with GDHS and BNHS experience forms of gait disturbance (28, 35, 36). This could be due to either CNS or PNS symptoms not presenting yet, since patients diagnosed with GDHS or BNHS are typically diagnosed early on due to other symptoms (delayed puberty or vision impairment). The presentation of delayed puberty is attributed to hypogonadotropic hypogonadism, which occurs due to the lack of LH and FSH released by the anterior pituitary. The lack of LH and FSH results in the lack of secondary sexual characteristics that lead to primary amenorrhea, small penis and testes, or absence of pubic hair/breast development (20, 21). BNHS presents with chorioretinal dystrophy that is usually diagnosed with visual testing such as OCT or fundoscopy, where affected individuals show a thinning of the retina, loss of retinal architecture, and elimination of choriocapillaris and choroidal vessels (1, 20). Individuals with GDHS and BNHS usually have their first symptom present in childhood/adolescence, and onset of further symptoms can manifest as late as late adulthood (data in chapter 2). With the addition of hypogonadotropic hypogonadism and visual impairment compared to patients with SPG39, individuals with GDHS and BNHS are considered the moderate severe form of the PNPLA6-opathies.

Lastly, individuals diagnosed with LNMS or OMCS exhibit similar clinical manifestations as BNHS individuals, except patients with OMCS experience hair anomalies such as trichomegaly or alopecia (1, 11). The clinical difference between individuals with LNMS and BNHS is the first onset of symptoms in LNMS presents at birth or in early childhood. Patients with OMCS also have their first symptoms present at birth or early childhood, with visual impairment or hypopituitarism being the most common first symptom to be discovered. Hair anomalies have not been reported in individuals with LNMS thus far (1). Interestingly, the presence of gait disturbances is more variable among patients with LNMS and OMCS compared to patients with SPG39, GDHS, and BNHS. Nevertheless, since individuals with LNMS and OMCS experience an earlier onset of symptoms compared to previously mentioned syndromes, they both are considered the most severe form of the PNPLA6-opathies.

Genotype-Phenotype-NTE Activity correlations

To date, attempts to find a genotype-phenotype correlation in patients with PNPLA6 variants have been limited thus far. One of the first articles linking individuals with OMCS with biallelic PNPLA6 variants observed that a majority of the patients were compounds heterozygous, and had variants located in both the NEST and CNB domains (11). Synofzik and colleagues showed that in their cohort of patients diagnosed with BNHS, several variants were not restricted to the NEST or CNB domains but can be seen across the entirety of the gene (20). A meta-analysis from Wu et al. demonstrated that having at least one variant in the NEST domain was highly correlated with having chorioretinal dystrophy. Additionally, patients with frameshift, splice, or indel variants were also more likely to have chorioretinal dystrophy (35). Another meta-analysis showed that chorioretinal dystrophy was the most prevalent clinical manifestation presenting at early childhood and hypogonadotropic hypogonadism the most prevalent symptom presenting at adolescence across all patients with PNPLA6 variants (25).

Studies examining the relationship between genotype and activity of NTE have shown promise in describing the basic mechanism of disease due to PNPLA6 variants.. Expressing and purifying only the NEST domain in E. coli cells with missense variants found exclusively in patients with SPG39 showed significantly decreased activity compared to WT NEST (37). This seminal experiment was the first to prove that variants located near or within the enzymatic domain of NTE could alter the catalytic activity of the protein. The same authors also looked at the NTE activity levels in control, unaffected, and affected patient fibroblasts with the same missense variants, and the results showed relatively similar activity levels compared to NEST domain only mutant constructs (38). Adding on to the work by Hein and colleagues, Hufnagel et al. revealed a striking relationship between the NTE activity levels in patient fibroblasts of individuals with SPG39 and OMCS. In this work, the NTE activity of a patient with SPG39 was roughly 75%, whereas the NTE activity of two affected patients with OMCS was roughly 25%. These results indicated the importance of NTE activity in disease severity, where NTE activity and onset of symptoms is inversely correlated (Table 2).

Table 2.

Summary of the genotype-NTE activity relationship thus far (1, 37, 38).

	Genotype	% Activity (relative to Ctrl)
Fibroblast	M1060V/M1060V	73.1
	R938H/R1031Qfs*38	64.3
	R1099Q/G1176S	24.1
NEST domain	M1060V	71.6
	R938H	63.9

Molecular and Cellular biology of PNPLA6/NTE

PNPLA6 (NG_013374.1) contains at least five major transcripts and four major protein isoforms known to date thus far (Table 3) (2). Each isoform includes five domains: an N-terminal transmembrane domain (TMD), three cyclic nucleotide binding domains (CNB), and the patatin like catalytic (NEST) domain.

Table 3.

Summary of PNPLA6 transcripts (2).

Domains	Protein isoform A NP_001159583.1	Protein isoform B NP_006693.3, NP_001159585.1	Protein isoform C NP_001159584.1	Protein isoform D NP_001159586.1
TMD	60–80	12–32	12–32	51–71
CNB1	195–316	147–268	147–268	186–307
CNB2	512–622	464–574	474–548	503–613
CNB3	635–743	587–695	561–669	626–734
NEST	964–1269	916–1221	889–1194	954–1269
Weight (kDa)	151	146	143	150

The differences between these isoforms are the location of the transmembrane domain, and the truncation of the CNB2 domain in isoform C, which is 36 aa shorter than the other isoforms. The N-terminal TMD is predicted to form a 20aa single pass α-helix spanning through the ER membrane (39). The TMD is required for proper ER membrane association and topology, but not necessary for proper catalytic activity of the protein (40). Thus far, there has been little evidence of the role of the CNB domains, although Richardson and colleagues noted that molecular docking simulations found that cAMP and cGMP can bind to these domains favorably (41). Previous research has shown a direct relationship between the levels of cAMP and levels of mRNA and protein in HeLa cells, but not the activity of NTE (42). Additionally, the area where the CNB domains are located contribute to the degradation of the protein via the ubiquitin–proteasome pathway (UPP) (43). The NEST domain contains the catalytic portion of the protein, allowing it to hydrolyze phospholipids by evicting 1 or 2 free fatty acids (4). Target substrates that can be catalyzed by the protein include phosphatidylcholine (PC), phosphatidylglycerol (PG), phosphatidylethanolamine (PE), and phosphatidic acid (PA) (4). NTE can catalyze the hydrolysis of lysophosphatidylcholine (LPC) 10 times faster than other known recombinant lysophospholipases (4, 44). This is further evident in studies of PNPLA6 in vitro, where overexpression of recombinant NTE in Neuro-2a cells causes a greater increase in glycerophosphocholine levels (GPC, product of LPC catalysis) compared to the relative decrease in LPC levels (45). Knockdown of PNPLA6 in HeLa cells displayed a 2.5-fold increase in LPC levels, showing that PNPLA6 enzymatic activity is necessary and sufficient for proper LPC homeostasis. In contrast, the overall levels of PC and other major phospholipids were not perturbed when NTE was either overexpressed (45) or knocked down (46), although their turnover was markedly decreased when PNPLA6 was knocked down (46).

The subcellular localization of the protein was first discovered in 1979, where fractionation experiments in hen brain showed that NTE was isolated in the microsomal fraction, confirming the protein was membrane bound (47). Later studies showed that transfecting fluorescently labeled NTE was present in the endoplasmic reticulum in COS-7 cells (40, 48). Interestingly, transfecting constructs with or without different domains showed that the regulatory domain (location of CNB domains) was able to prevent the protein from localizing to lipid droplets (48). In contrast, transfected constructs harboring no mutations or human missense variants in the PNPLA6 gene did not affect the subcellular localization of the protein to the ER (1).

Activity of NTE can be modulated by the presence of organophosphorus (OP) compounds (discussed in next section) and variants in the gene (discussed in previous section). Interestingly, a yeast two-hybrid screen on a human fetal brain cDNA library showed that a subunit of G proteins, Gβ2, interacts with the C-terminal domain of the protein. Knockdown of Gβ2 does not change the levels of NTE in the cell, but reduces the activity of the protein to ~50%, revealing a potential regulatory mechanism of NTE activity through G-protein signaling (49).

Expression, Purification, and Structure of human NTE

The structure of the full length human NTE protein has yet to be elucidated. Since NTE is a membrane protein, expression, purification, and 3D imaging have been difficult to obtain thus far. Wijeyesakere and colleagues looked at the predicted tertiary structure of human NTE bioinformatically by analyzing the primary structure of and the structure of patatin isoform-17 (pat17, PDB: 1OXW), which shares 30% sequence similarity to the NEST domain of NTE. The study proposed that the NTE catalytic domain contains a catalytic dyad at Ser1014 and Asp1134 that is needed for hydrolysis of phospholipids (39). Based on hydropathy plots, the catalytic domain is predicted to have a hydrophobic facing active site that is associated with the membrane. Additionally, the CNB domains located in the middle of the protein would interact with the cytoplasm due to its hydrophilic nature. A follow up study looking at pat17 (potato lipase) and its interactions with aging and non-aging OP compounds indeed showed that aging does not cause a global conformational change to pat17, implying a similar affect could happen in the NEST domain of NTE (50). Recent advances in protein modeling have also been unable to produce a confident model of the full-length human protein (51).

Expression and purification of the catalytic domain of NTE have proved to be successful, where constructs were expressed in E. coli. Purification of the domain was done via detergent extraction in CHAPS, and subsequent dialysis to incorporate into dioleoylphosphatidylcholine (DOPC) liposomes (52). Specific NTE activity of the catalytic domain of WT and patient specific mutant constructs was able to be elucidated thereafter (37).

PNPLA6/NTE Animal Models

Chicken and overview of OP compounds (1960’s)

Chickens were the first animal model used to study NTE and its interaction with organophosphorus (OP) inhibitors, since other animal models, like rats and mice, were resistant to effects from OP inhibitors (53). Synthetic OP compounds are present in a wide variety of common everyday products such as pesticides and insecticides (41). OP compounds are also used as toxic nerve agents for chemical weapons, with the most notable being (RS)-propan-2-yl methylphosphonofluoridate (sarin), which has been used in numerous modern-day conflicts and terrorism (54). In relation to its effects on NTE, OP compounds can be categorized into “neuropathic” or “non-neuropathic” OP compounds based on their ability to induce OPIDN, a form of paralysis that causes degeneration of distal axons in the CNS and PNS (41, 55–58). M.K. Johnson was the first to show this in chickens, where he observed that neuropathic compounds (such as Mipafox) were able to form “aged” NTE, whereas non-neuropathic OP compounds (such as Paraoxon) were not able to form “aged” NTE (59). “Aging” is defined by a loss of an R-group and a net gain of a negatively charged organophosphyl group covalently bonded to a serine in the active site of an enzyme. This reaction can be seen in NTE found in chicken brain homogenates, where exposure to neuropathic OP compounds causes an inability for the enzyme to catalyze its substrates (60, 61). For OPIDN to occur, 70% of NTE activity in the nervous system of chickens was inhibited via “aging” by a neuropathic OP compound (62). Intriguingly, exposure to non-neuropathic OP compounds prior to exposure to neuropathic OP compounds was able to protect chickens against OPIDN (63). These seminal experiments in chickens not only contributed to the discovery of NTE but also revealed NTE as a target protein for OPIDN that can be protected by pre-exposure to non-neuropathic OP compounds.

Drosophila (1990’s)

The model organism Drosophila has also been extensively used to study NTE function and expression in the nervous system. In Drosophila, the PNPLA6 gene is referred to as sws. Deletion of sws in flies produces an age-dependent degeneration, with the progressive formation of vacuoles in the cortex, glial hyperwrapping, and neuronal apoptosis starting within 3–4 days of adulthood (64). Transgenic pan-neuronal expression of patient specific PNPLA6 mutations from each disorder spectrum in Drosophila sws mutants showed rescue of the locomotion defect in young flies (7d), but only a partial rescue in older flies (14d) (65). Missense variants in the patatin-like phospholipase domain (R1099Q, OMCS; A1029T, GDHS) were able to reduce the size of vacuoles in 14d old sws flies, indicating that partial phospholipase activity can ameliorate the neurodegeneration observed in Drosophila.

Interestingly, sws mutant flies exhibit a motor neuropathy phenotype, but not an eye phenotype, such as loss of photoreceptors. Eye specific deletion of sws in Drosophila via GMR-GAL4 again showed no signs of photoreceptor degeneration after 7d. However, 30d flies were able to show marked signs of degeneration via accumulation of condensed cell bodies or loss of photoreceptors via electron microscopy (11). Interestingly, there was no significant difference in the electroretinograms (ERG) of WT vs. sws mutants at both young (1–3d) and old (30d) flies, implying that photoreceptor degeneration in Drosophila does not affect phototransduction, but does impact photoreceptor viability.

The sws gene also contains phospholipase activity. Mühlig-Versen and colleagues showed that 8–10d sws mutant flies accumulated ~20% more PC, but not PE (6). These results were recapitulated in Drosophila with patient specific mutations across the protein, where no significant difference was seen in PC levels in sws mutants vs. patient specific PNPLA6 mutations. Interestingly, transgenic expression of human PNPLA6 in Drosophila also showed no changes in PC levels compared to sws mutants but showed significantly decreased LPC levels compared sws mutants. This observation was not seen in patient specific PNPLA6 mutations, where all mutants tested had similar LPC levels compared to sws mutants (65). This data implies that activity of human NTE is important for the control of LPC levels and not PC levels in the Drosophila brain.

In response to lipid composition disequilibrium of PC and LPC, previous studies have showed that in Drosophila sws mutants, the activity of the ER stress response increases, and overexpression of proteins involved in ER stress response significantly reduce LPC levels, but not PC levels, relative to the mutant (66). Treatment with Tauroursodeoxycholic Acid (TUDCA), a bile salt that reduces ER stress and is a known inhibitor of apoptosis, partially rescued the neurodegenerative phenotype in sws mutants. These results altogether show again that elevated LPC levels more so than PC levels may play a role in promoting ER stress, which in turn promotes apoptosis and subsequent neuronal cell death, causing locomotion defects.

Mouse (2000’s)

Mice have been used extensively to study the function of Pnpla6/NTE in development and phospholipid homeostasis. In development, Pnpla6 KO mice in C57/BL6 background produced embryonic lethal pups that died around embryonic day 8 (E8) of embryogenesis (9). Developmentally, NTE null embryos displayed a normal phenotype and form definitive germ layers, suggesting that NTE function does not play a major role in early development. NTE deficient mice also produced proper formation of the placental layers. However, the extraembryonic yolk sac of NTE null mice was unable to form a vascular network, resulting in a collapse of the mesodermal and endodermal layers. This subsequent loss of nutrients caused growth retardation and an inability to fuse the cranial neural plate. In addition, apoptosis was present in all three germ layers, and ex vivo cultures of NTE null embryonic cells were able to give rise to different cell types and morphologies. These results suggest that defects in extraembryonic development are the cause of cell death, and placental formation is not essential for embryo survival at E8–8.5, showing that improper vascularization and not placental defects are most likely the major cause of embryonic death in NTE null embryos.

To evaluate the role of NTE in the adult nervous system, a conditional KO (cKO) of the gene in neuronal tissue was generated by breeding an NTE^fl/fl mouse strain with a Nestin-cre (Nes-cre) mouse strain that expressed cre recombinase after E11 in the CNS and PNS (67). The resulting cross produced Nes-cre:NTE^fl/fl embryos with normal mendelian ratios, as well as normal brain structure after E14. Later in adulthood (4–10w), NTE cKO mice exhibited hippocampal vacuolization and a loss of neurons in the thalamus and cerebellum (Purkinje cells). Notably, there was extensive vacuolization in the neuronal cytoplasm, as well as the presence of abnormal reticular aggregates that formed tubulovesicular bodies in the cell body of the neuron. Read and colleagues showed that NTE cKO mice at E18 had axonal lesions that could be characterized as swollen with accumulation of vesicular and cytoskeletal elements, and degenerative with dark dense-bodies and translucent intra-axonal vacuoles (5). These lesions were in the distal parts of the longest sensory spinal axons in the lateral and ventral motor tracts, emanating from neurons in the brain cortex. In support of this, the rate of neuronal secretion of reelin and APP, proteins that are involved in the constitutive secretory pathway, decreased by ~20% in NTE cKO mice. In terms of phospholipid homeostasis, NTE cKO mice see a sustained ~20% increase in PC levels in the brain hemispheres after 1 month. These results show that NTE plays a role in membrane trafficking for the maintenance of mature axons, and that similar phenotypes arise in immature and mature axons when NTE is deficient.

Motor behavioral assessments of adult NTE cKO mice at 4w showed no significant differences in rotarod testing compared to littermates (67). By 4–5 months, cKO mice not only showed a reduction of time on the rotarod (67), but they also showed decreased ability to walk properly on a beam (beam walking test) and in an open field (open-field movement test) (5). Furthermore, clinical signs of hindlimb dysfunction that were present at 4 months progressively worsened up to 14 months, indicating that although spongiform encephalopathy is present early in development (as early as 4w), signs of abnormal motor behavioral do not start until well into adulthood. Additionally, NTE cKO mice were significantly lighter compared to their littermate controls, implying that brain specific deletion of NTE can recapitulate not only the motor neuropathy phenotype but also the anterior hypopituitary phenotype seen in human subjects (67).

Zebrafish (2010’s)

With the recent advances of zebrafish as a model to study human development and disease, several papers have looked at the effect of NTE inhibition and dysfunction in zebrafish. In development, knockdown of pnpla6 lead to distinct morphological changes and lack of neuronal development within 72 hpf (68). Markers of the BMP pathway were upregulated in pnpla6 morpholino morphants and treatment with dorsomorphin rescued the morphological “curly tail” defect, implying a potential role of pnpla6 in neuronal development via the BMP signaling pathway (68). Interestingly, morphant embryos were able to rescue the “curly tail” phenotype by co-injection of pnpla6 morpholino and full-length protein, but not with mRNA that contained human missense variants found in the affected population (1).

Experiments to test the efficacy of zebrafish as a model for OPIDN showed that adult zebrafish was not a good model for OPIDN, even though NTE activity was inhibited 70%. Inhibition of NTE did not lead to any neurological abnormalities and only a mild increase in LPC level after exposure (69).

Conclusions

For the past 70 years, research into PNPLA6 and NTE has evolved from basic biochemical analysis and discovery of OPIDN, to the advancement of next generation sequencing techniques to uncover PNPLA6 as a disease-causing gene for several rare neurological disorders in humans. These studies established the basic understanding of where the protein is expressed and located in the cell, what substrates it can catalyze and how the protein can be inhibited, and how either knockout or knockdown of the gene in vitro and in vivo can cause disruptions in phospholipid homeostasis and proper tissue function. Preliminary evidence has shown that disease onset and NTE activity are inversely correlated, where disease onset becomes more severe as NTE activity decreases (1).

As more individuals with PNPLA6 variants are discovered, the triumvirate relationship between patient genotype, patient phenotype, and activity of NTE still remains undefined. Furthermore, how do different levels of NTE activity precipitate dysfunction in affected tissue, such as the brain and eye? While research on PNPLA6’s role in the brain has been studied, its role regarding the unique chorioretinal dystrophy that is present in affected individuals has yet to be answered. Future work modeling PNPLA6 variants in animal models such as mouse will be important to uncover the cellular and molecular mechanisms that underpin PNPLA6 induced retinopathy. With the advancement of stem cell technologies, it will also be useful to study the lipid and cellular biology in patient derived retinal pigment epithelium to evaluate the lipid species most involved in PNPLA6-associated retinopathy.

Understanding the basic biology that underpins the PNPLA6 disease spectrum will ultimately establish the building blocks for treatment for affected patients. Since PNPLA6 encodes for an enzyme and loss of activity is hypothesized to lead to disease, potential therapies such as AAV directed gene therapy could provide a method for ameliorating the vision loss that occurs in affected patients. Enzyme replacement therapy can also be a potential therapeutic option. Therefore, research into large scale expression and purification of the membrane-associated enzyme will be important for establishing this potential therapy in the future.