The Clinical, Radiological and Genetic Spectrum of PLA2G6-Associated Neurodegeneration: An Experience From a Tertiary Center

Patients and methods

We performed a retrospective observational study of all patients exhibiting a biallelic pathogenic or likely pathogenic variants in the PLA2G6 gene based on exome sequencing in the probands. The case records were subjected to detailed data extraction comprising details of age at onset (AAO), age at presentation, first symptom at onset, the spectrum of motor symptoms, movement disorders encountered, non-motor clinical features including cognitive, psychiatric, and additional relevant clinical features, genotype, management strategies including levodopa responsiveness and drug-induced dyskinesia. All patients were classified based on the clinical phenotype into INAD, aNAD, dystonia-parkinsonism (DP), dystonia parkinsonism myoclonus (DPM), early-onset parkinsonism (EOP), complex dystonia and complicated hereditary spastic paraparesis (cHSP) phenotype. The study was approved by the institute’s ethics committee. Written informed consent was obtained from patients for the publication of recorded videos.

Genetic analyses

The blood samples of patients were subjected to Genomic DNA extraction using QIAamp DNA Blood Mini Kit (Qiagen Germany, #51104). Subsequently, raw reads were aligned to the human reference genome (GRCh37) based on the BMA-mem algorithm. Polymerase chain reaction (PCR) duplicates were removed using the Picard toolkit (https://broadinstitute.github.io/picard/) [4]. The variants were identified based on the Genome Analysis Toolkit (GATK) framework (Broad Institute, Cambridge, MA, USA). Variants were exposed to base quality score recalibration for filtration, following which annotations would be done in the ANNOVAR platform (http://www.openbioinformatics.org/annovar/) [5]. The variants with a minor allele frequency (MAF) >0.01, suggestive of common occurrence in the population, were not included. A comparison of data with the Exome Aggregation Consortium (ExAC), 1000 Genome project, and gnomAD database (https://gnomad.broadinstitute.org/) was performed. Each of the individual sequence variants were tested using PolyPhen-2, Sorting Intolerant from Tolerant (SIFT) webserver, and MutationTaster [6, 7, 8]. The variants were assigned a pathogenicity coding according to the American College of Medical Genetics and Genomics (ACMG) guidelines into benign, likely pathogenic or pathogenic [9]. Each of these genes was analyzed in the mutation databases of ClinVar for novelty evaluation.

Review of literature

We conducted a systematic search across the accessible medical database of PubMed, Google Scholar, and Scopus, using the medical subject headings (MeSH): “PLA2G6”, “PLAN”, “INAD”, “ANAD”, “atypical NAD”, “Hereditary Spastic Paraparesis”, parkinsonism”, “young onset Parkinson’s disease”, “early onset Parkinson’s disease”, “dystonia”, and “disorder” to identify all the relevant studies. The articles were subjected to title and abstract screening. A shadow search of the reference articles was done to avoid missing key articles. Our data extraction process was based on identifying unique demographic and clinical parameters, classifying the clinical phenotype, and describing the specific genetic variants so identified. Studies not presented in English or lacking patient details were excluded from consideration.

Statistical methods

Categorical variables were depicted in terms of frequencies, whereas continuous variables were designated as median values along with their respective interquartile ranges. The initial analysis involved descriptive statistics to examine the demographic, clinical, and neuroimaging parameters. A narrative description encompassing clinical phenotypes and genotypes and their correlation with neuroimaging parameters was made. Statistical calculations were performed using SPSS version 23.0.

Demographics

In this retrospective analysis, we analyzed 26 patients (22 families) with genetically confirmed PLAN. Eight of these patients were part of a multicentric publication [3]. There was a male predominance (n = 15, 57.69%) distributed across 22 families. The median age at presentation was 22.5 years (IQR: 6.5–29.0), while the median age at onset (AAO) was 13.0 years (IQR: 2.7–20.5) and the patients presented after a median duration of illness of 3.5 months (IQR: 1.5–10.5). Consanguinity was observed in 53.8% of patients (n = 14), while 42.3% (n = 11) had a positive family history (Table 1).

Clinical features

Investigations and management

Neuroimaging features

Brain MRI (available for 25 patients) showed mineralization in various areas, including the caudate in 9/25 patients (34.6%), the putamen in 8/25 patients (30.8%), the GPi in 14/25 patients (53.8%), and the substantia nigra in 14/25 patients (53.8%). Atrophy was noted in different brain structures, of which cerebellum was involved most commonly (23/25;88.5%) followed in order by diffuse cerebral (10/25;38.5%), caudate (4/25;15.4%), frontal-predominant atrophy (3/25;11.5%), optic nerve (2/25;7.7%) and corpus callosum (1/25; 3.8%). Other findings included white matter signal changes in 12 patients (12/25;46.2%) and claval hypertrophy in 10 patients (10/25;38.5%). With regards to the individual subtypes, classic claval hypertrophy was observed in 80% (n = 4/5) of patients of INAD followed by ANAD (1/3;33.3%) and DP (2/5;40%). This claval hypertrophy refers to the hypertrophy of the gracile tubercle formed by the nucleus and fasciculus of the gracilis due to spheroid bodies deposition in sensory nuclei of the medulla [10]. The predominant MRI abnormalities were diffuse involvement of white matter (4/5;80%) in the DP subtype. Mineralization of substantia nigra and putamen was noted in a single patient with an EOP phenotype. The two patients with complex dystonia lacked any signature MRI involvement except for non-specific cerebellar atrophy (Figure 1, Table 1).

Figure 1

Genetic analysis

Exome sequencing (Figure 2, Table 2) revealed biallelic disease-causing variants in the PLA2G6 gene in all patients in homozygous state in 21 (21/26;80.8%) and compound heterozygous state in the remaining 5 patients (5/26;19.2%). Among patients with homozygous variants, missense variants were noted in 19 patients, while the remaining two patients had truncating variants (Stop-gain: 1 and frameshift: 1). The patients with homozygous truncating variants had the severe INAD phenotype. In five patients with compound heterozygous variants, three patients had missense/missense configuration and one patient each had missense/stop-gain and missense/frameshift configuration. The c.2222G>A;p.Arg741Gln variant was the most commonly identified variant in our cohort, which was identified in 12 patients (9 families). This variant has been previously reported and is predominantly seen in Asian subpopulations [1]. The next common variant was a novel c.2405T>C;p.Leu802Pro missense variant, which was observed in two patients. It was identified in homozygous state in one patient with EOP and in compound heterozygous configuration with a previously reported pathogenic stop-gain variant in another patient with cHSP phenotype.

Figure 2

Table 2

Details of disease-causing variants (Transcript ID: NM_003560.4).

FAMILY	PATIENT NO	AGE (YEARS)	GENDER	AAO (YEARS)	FAMILY HISTORY	CONSANGUINITY	FIRST SYMPTOM	CLINICAL PHENOTYPE	EXON	c.DNA	PROTEIN CHANGE	CONSEQUENCE	ACMG CRITERIA	ZYGOSITY	NOVEL
1	1	22	F	19.5	–	III	Slowness of ADL	DP	15	c.2239C>T	p.Arg747Trp	Missense	Pathogenic (PS3PP5PM1PM2PP2PP3)	Hm	–
2	2	34	F	26	+	III	Behavioural changes	DP	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
3	3	1.9	M	0.5	+	No	Developmental regression	INAD	12	c.1723delT	p.Thr575 Argfs*8	Frameshift	Likely pathogenic (PVS1PM2)	Hm	+
4	4	29	F	15	+	No	Walking difficulty and leg posturing	DP	5	c.763C>T	p.Pro255Ser	Missense	VUS-LP (PM2PP2PP4)	Hm	+
5	5	1.9	M	1.4	–	III	Developmental regression	INAD	12	c.1612C>T	p.Arg538Cys	Missense	Pathogenic (PM5PS3PM1PM2PP2PP3PP5)	Hm	–
6	6	29	M	12	+	II	Cognitive decline	DPM	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
6	7	24	F	12	+	II	Cognitive decline	DPM	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
7	8	21	M	20	–	No	Speaking difficulty	cHSP	3	c.292T>C	p.Ser98Pro	Missense	VUS–LP (PM2PP2PP4)	Hm	+
8	9	8	M	2	–	No	Slurring of speech	ANAD	10	c.1547C>T	p.Ala516Val	Missense	Pathogenic (PM1PM2PP2PP3PP5)	CH	–
									10	c.1471C>T	p.Leu491Phe	Missense	Likely pathogenic (PM1PM2PP2PP3)	CH	+
9	10	7	F	4	–	No	Walking difficulty	ANAD	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	CH	–
									13	c.1763T>C	p.Leu588Pro	Missense	Likely pathogenic (PM1PM2PP2PP3)	CH	+
10	11	28	M	25	–	No	Slowness of ADL	EOP	15	c.2197G>C	p.Asp733His	Missense	Likely pathogenic (PM2PP2PP3PP4)	CH	+
									14	c.1897G>A	p.Ala633Thr	Missense	Likely pathogenic (PM1PM2PM5PP2PP3)	CH	+
11	12	2	F	1	–	II	Regression of milestones	INAD	5	c.667C>A	p.Pro223Thr	Missense	Likely pathogenic (PM2PM5PP3PP4)	Hm	+
12	13	26	F	16	+	II	Toe walking	cHSP	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
	14	32	M	20	+	II	Ataxia	cHSP	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
13	15	3.5	F	2	–	III	Regression of milestones	INAD	13	c.1799G>A	p.Arg600Gln	Missense	Pathogenic (PM1PM2PM5PP2PP3PP5)	Hm	–
14	16	5	M	3	–	No	Regression of milestones	ANAD	3	c.379G>A	p.Val127Met	Missense	VUS–LP (PM2PP2PP4)	Hm	+
15	17	29	M	14	–	No	Behavioural issues	DP	6	c.835delA	p.Ile279SerfsTer26	Frameshift	Pathogenic (PVS1PM2)	CH	+
									7	c.991G>T	p.Asp331Tyr	Missense	Pathogenic (PM1PM2PM5PP2PP3PP5)	CH	–
16	18	36	F	10	–	No	Toe walking and running since childhood	DP	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
17	19	23	M	18	+	II	Left leg dragging followed by walking difficulty	cHSP	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
18	20	28	M	24	+	No	Slowness in daily activity	DPM	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
18	21	31	M	28	+	No	Memory impairment	DPM	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
19	22	30	F	28	–	Yes	Slowness of ADL	EOP	17	c.2405T>C	p.Leu802Pro	Missense	VUS-LP (PM2PM3PP2PP4)	Hm	+
20	23	2	M	1.5	–	II	Developmental regression	INAD	14	c.1903C>T	p.Arg635Ter	Stop gain	Pathogenic (PVS1PM2PP5)	Hm	–
21	24	22	F	22	+	II	Posturing and tremulousness of head and right upper limb	cDYT	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
21	25	20	M	11	+	II	Posturing and tremulousness of head and right upper limb	cDYT	15	c.2222G>A	p.Arg741Gln	Missense	Pathogenic (PS3PP5PM1PM2PM5PP2)	Hm	–
22	26	15	M	5	–	No	Walking difficulty, stiff leg	cHSP	17	c.2370T>G	p.Tyr790Ter	Stop gain	Pathogenic (PVS1PS3)	CH	–
									17	c.2405T>C	p.Leu802Pro	Missense	Likely pathogenic (PM2PM3PP2PP4)	CH	+

[i] +: Yes; –: No.

AAO: Age at onset; ACMG: American College of Medical Genetics; ADL: Activity of daily living; ANAD: Atypical neuroaxonal dystrophy; cDNA: complementary Deoxyribonucleic acid; cDYT: Complex dystonia; CH: Compound heterozygous; cHSP: complicated Hereditary spastic paraparesis; DP: Dystonia parkinsonism; DPM: Dystonia parkinsonism myoclonus; EOP: Early onset Parkinsonism; F: Female; Hm: Homozygous; INAD: Infantile Neuroaxonal Dystrophy; M: Male; PLAN: PLA2G6-associated neurodegeneration; VUS-LP: Variant of uncertain significance probably likely pathogenic.

In total 19 unique variants were identified, of which 15 were missense variants, two were stop-gain and the remaining two were frameshift variants (Figure 2). Among these 19 unique variants, 11 variants were novel. Of these 11 novel variants, 9 were missense variants (c.292T>C;p.Ser98Pro, c.379G>A;p.Val127Met, c.667C>A;p.Pro223Thr, c.763C>T;p.Pro255Ser, c.1471C>T;p.Leu491Phe, c.1763T>C;p.Leu588Pro, c.1897G>A,p.Ala633Thr, c.2197G>C;p.Asp733His and c.2405T>C;p.Leu802Pro) and two were frameshift variants (c.835delA;p.Ile279SerfsTer26 and c.1723delT;p.Thr575 ArgfsTer8). Of these, as per ACMG criteria, eight variants could be classified as pathogenic or likely pathogenic, while three variants (c.292T>C;p.Ser98Pro, c.379G>A;p.Val127Met and c.763C>T;p.Pro255Ser) could be only classified as a variant of uncertain significance (VUS) at best. However, as per the recent scoring update to ACMG criteria, these variants could be reclassified as VUS more likely to be pathogenic than benign (VUS-LP) [9]. One novel variant (c.2405C>T;p.Leu802Pro) was seen in two patients, in homozygous state in a patient with EOP and in compound heterozygous trans configuration with a previously reported pathogenic stop-gain variant in a patient with cHSP owing to which it could be classified to likely pathogenic variant.

Case vignettes

Patient-13: A 26-year-old previously healthy female with consanguineous parentage, normal birth and developmental history presented with a complex neurological illness spanning 10 years (Video 1). Initially, she noticed walking difficulty with stiffness in limbs and occasional toe walking. In addition, she had frequent falls and inappropriate smiling. Her elder brother also had walking difficulty for 12 years duration, along with inappropriate smiling. On examination (Video 1), both patients had mild upgaze restriction that improved on vestibulo-ocular maneuverer, pseudobulbar affect, spasticity, and brisk deep tendon reflexes in all four limbs with extensor plantar response bilaterally and cerebellar abnormality. MRI of the proband revealed mineralization of basal ganglia and substantia nigra with cerebellar and cortical atrophy. Exome sequencing revealed a homozygous pathogenic missense variant (c.2222G>A;p.Arg741Gln). The siblings were treated symptomatically with baclofen and physiotherapy.

Patient-7: A 24-year-old female born to consanguineous parentage with normal birth and developmental history presented with a twelve-year history of progressive cognitive impairment, followed by tremulousness in upper limbs, slowness, abnormal posturing and walking difficulty. In addition, the patient had a behavioral abnormality in the form of anger outbursts, crying spells, low mood and anxiety. On examination (Video 2), the proband had parkinsonism, rest tremor, facial predominant myoclonus, pyramidal signs, generalized dystonia and cerebellar involvement. MRI showed diffuse cortical and cerebellar atrophy with basal ganglia mineralization. Exome sequencing revealed a previously reported pathogenic homozygous missense variant (c.2222G>A;p.Arg741Gln). She was treated with levodopa with a good response but developed disabling choreiform dyskinesia. Her elder brother also had similar cognitive, behavioral, and motor symptoms of 12 years duration and, on examination, had parkinsonism, rest tremor, dystonia, pyramidal signs, myoclonus, and cerebellar abnormality. The rest tremor persisted in an outstretched position as well albeit with similar severity, and in the presence of cerebellar signs and distal hand dystonia, it can be a component of rubral tremor rather than parkinsonian rest tremor. Sanger sequencing in the brother confirmed the presence of c.2222G>A;p.Arg741Gln in homozygous state. He also responded to levodopa but with disabling dyskinesia.

Patient-25: A 20-year-old male born to consanguineous parentage with normal birth and developmental history presented with an eight-year history of head and right upper limb posturing and tremulousness, alongside intellectual disability (Video 3). He demonstrated upgaze restriction, ataxic dysarthria, and generalized dystonia, with normal gait and brisk reflexes. MRI also revealed cerebellar atrophy. Exome sequencing revealed a previously reported pathogenic homozygous missense variant (c.2222G>A;p.Arg741Gln). The neck dystonia was associated with axial paroxysms of jerky dystonia commonly seen in patients with myoclonus dystonia (DYT-SGCE). However, there were no significant variant identified in SGCE gene. He did not respond to levodopa therapy and was put on clonazepam and baclofen for dystonia. His elder sibling, 22-year-old girl also had similar symptoms but of only 4 months duration. Examination revealed upgaze restriction, generalized dystonia, appendicular ataxia, and brisk reflexes, with a dystonic gait. MRI showed cerebellar atrophy. Sanger sequencing in the sister confirmed the presence of homozygous c.2222G>A;p.Arg741Gln variant. Her dystonia partially improved with levodopa therapy.

PLA2G6-associated cHSP phenotype

Our study cohort of 5 patients of cHSP phenotype represents the second largest series on this unique phenotype in the medical literature (Table 4). Previously, there were only 4 publications that reported on cHSP phenotype [12, 13, 14, 15]. The largest series of 6 patients by Koh et al. from Japan, had a female preponderance, high prevalence of cognitive decline (5/6), and cerebellar atrophy on neuroimaging (5/6) [13]. At par with the above phenotype, cognitive decline was observed in 4/5 (80%) of our patients, and all except one exhibited cerebellar signs. The unique features identified in our cohort included dysarthria (n = 3, 60%), dystonia (n = 3, 60%, generalized in n = 2, 40%), epilepsy (n = 2, 40%), and gaze restriction (n = 3, 60%).

With regards to neuroimaging, cerebellar atrophy and mineralization in substantia nigra were seen in all patients with available brain MRI (n = 4). There were 3 patients (75%) that were detected to have mineralization involving caudate, putamen, and globus pallidus interna (GPi). Unlike previous reports, white matter signal changes were reported in 75% of patients (n = 3/4), while claval hypertrophy was reported in 2 patients. Interestingly, the most common genetic variant: c.2222G>A identified in 3 patients (60%), was not previously reported with PLA2G6-related complicated cHSP phenotype. The difference in clinical phenotypes and neuroimaging findings as compared to the previous patients could be accounted for by the differences in the genotypic spectrum of the identified variants in the PLA2G6 gene.

PLA2G6-associated phenotypes related to c.2222G>A variant

The c.2222G>A is the most common genetic variant identified in our cohort of PLAN (34.6%). Previously 19 patients have been reported in the literature on this genetic variant (Table 5) [1, 2, 16, 17, 18, 19, 20, 21, 22]. Amongst the patients reported previously, a large proportion of them belong to Asian ancestry, and of these, 12 patients either are from India or have an Indian ancestry [1, 17, 18, 19, 20, 21]. Our study contributes to the largest single-center series of patients identified with c.2222G>A variant. The demographic parameters of this subgroup were at par with previous publications, exhibiting a nearly equal sex ratio, similar AAO (Median: 28 years; IQR: 23.5-33 years), a similar prevalence of family history (77.8% vs 78.9%), and consanguinity (66.7% vs 72.2%). The unique attributes that were recognized include the occurrence of cHSP phenotype (n = 3, 33.3%), DPM (n = 3, 33.3%), and aNAD (n = 1, 11.1%), which were not known previously. While previous studies have suggested a much greater prevalence of DP (n = 16/19, 84.6%), only 2 of our patients (22.2%) displayed such a phenotype. A recent report demonstrated a remarkable response to bilateral subthalamic nuclei DBS in a patient diagnosed with PLAN carrying this specific variant in the homozygous state [21]. At three months follow-up, the patient experienced significant relief from dyskinesia (>90%), improvement in OFF-state (UPDRS-III score 61 to 16), and reduction in levodopa-equivalent daily dose (1050 mg to 275 mg). This emphasizes the crucial importance of identifying this variant in patients clinically suspected of having this condition [1]. However, the long-term sustenance of the benefit needs to be studied. There are two more reports of STN-DBS in patients with PLAN but with different variants. In the report by Wirth et al., the patient carried compound heterozygous variants (c.109C>T;p.Arg37X and c.232G>T;p.Ser774Iso) in PLA2G6 gene and noted around 70% improvement in UPDRS-III OFF state score at one-year follow-up [23]. In the other report by Choi et al. (conference abstract), the authors mention that the siblings carrying compound heterozygous variants (c.359G>A;p.Trp120Ter and c.1742G>A;p.Arg581Gln) in PLA2G6 gene had an excellent response to bilateral STN-DBS [24]. However, no further details are available. In view of only a few case reports, one needs to be cautious with respect to the long-term benefits of STN-DBS in patients with PLAN.