Introduction

Hereditary spastic paraparesis (HSP) represents a group of genetically heterogeneous diseases that mainly involve the spinal portion of upper motor neurons [1]. HSPs show autosomal dominant, autosomal recessive, X-linked, or mitochondrial inheritance, with over 85 genes loci identified, multiple pathogenic mechanisms [2], and an ample array of neurological and extra neurological accompanying clinical features [3, 4]. From a phenotypical point of view, they are classified into pure or complex [5, 6]. Usually, autosomal dominant HSP (AD-HSP) present with pure phenotypes and are more frequent than autosomal recessive HSP (AR-HSP) [7]. SPG18 stands out as one of the less common forms of HSP documented to date, resulting from either monoallelic or biallelic mutations in the ERLIN2 gene. ERLIN2 is a lipid-raft-associated protein situated within the endoplasmic reticulum (ER) featuring an SPFH domain [8], and it forms an ERLIN1/2 complex with the closely related ERLIN1 protein; the function of this complex is to bind RNF170, a ubiquitin ligase (E3), which targets activated inositol 1,4,5-trisphosphate receptors (IP3R). These receptors are subsequently ubiquitinated and degraded [8, 9].

SPG18 is in most cases inherited in an autosomal recessive manner (AR-SPG18) [10,11,12,13,14,15], but few autosomal dominant SPG18 have surfaced (AD-SPG18) [16,17,18,19]. The clinical picture of AR-SPG18 usually displays early onset spastic paraparesis, complicated by intellectual disability, motor and speech development delay, multiple joint contractures, seizures [10,11,12, 14, 17], and in one case progressive juvenile primary lateral sclerosis [13]. AD-SPG18 shows upper motor neuron abnormalities and mild dorsal column abnormalities [16,17,18,19]. MRI alterations (i.e., thin corpus callosum—TCC) have been reported in two families [10, 17]. No other laboratory of imaging characteristics has been found. Furthermore, it has been pointed out how some patients with SPG18, both AR and AD, could show phenoconversion to amyotrophic lateral sclerosis (ALS)-like syndromes [20].

Herein, we report two novel ERLIN2 pathogenic variants found in a series of Italian SPG18 patients. We also review genetic and clinical data from all SPG18 cases described in the existing literature, discussing about genetic variability, phenotypical features, and the importance of zygosity in genetic counselling.

Materials and methods

Patients

This multicentric case series study was performed in accordance with the Declaration of Helsinki statements. Written informed consent and ethical approval (CE Lazio) were obtained. In the past 8 years, in a single laboratory, we tested 944 patients with clinical evidence of HSP without a genetic diagnosis, using a multigene targeted resequencing panel (TRP, n = 710) or exome sequencing (ES, n = 234) or both (n = 114) and investigated the coding exons and flanking introns of the genes known to be associated with HSPs [6, 21, 22]. Five patients (4 men; 1 woman) from four families were identified and recruited from three Italian neurology centers (University of Rome Sapienza, Azienda Ospedaliera San Camillo Forlanini, IRCCS Stella Maris Foundation in Pisa). These patients were enrolled in the study and underwent further investigation and analysis. Family and clinical history were collected. All patients underwent neurological examination, clinical cognitive assessment through Montreal Cognitive Assessment (MoCA) [23], and brain MRI (Table 1).

Table 1 Clinical, demographic, and radiological details of our patients
Full size table

Molecular and database search

DNA extraction was carried out using peripheral blood lymphocytes obtained from the patients and modalities of next-generation sequencing (NGS) analysis for TRP and ES using methodologies already reported [22, 24]. Search for variants of ERLIN2 was done using population databases (dbSNP, 1000genome, EVS) and local databases, and their pathogenicity was assessed according to the American College of Medical Genetics and Genomics (ACMG) guidelines [25]. Literature was reviewed using PubMed and Google Scholar, and findings were collected in Table 2.

Table 2 Details about clinical, demographical, and radiological findings in all SPG18 cases described in the literature so far
Full size table

Results

Cases clinical reports

Clinical, imaging, and laboratory features are summarized in Table 1. Pedigrees are shown in Fig. 1. All families had Italian descent with no reported consanguinity. Three of five individuals had disease onset between 25 and 50 years, while two had juvenile onset (< 10 years). The overall initial manifestation was spastic paraparesis (5/5), with mild dorsal column signs and symptoms (3/5). In one case (BII.1), at the time of the last neurological examination and after 20 years from HSP onset, bulbar and appendicular lower motor neuron signs were identified, suggesting an ALS-like syndrome. Patient AI.2 (mother of AII.1) showed first signs of spastic paraparesis at the age of 50, and her disorder evolved into a rapidly progressive ALS-like syndrome after 27 years (77 years old). She died of ab ingestis pneumonia 12 months later. In one proband, we found maculopathy, while in another one, we observed congenital cataract. No one showed intellectual disability (ID) or extraneurological complications. They all underwent brain MRI: one patient had white matter alterations (WMA). Disease course was slowly progressive (mean 26.4 y at the time of the last examination).

Fig. 1
figure 1

Families’ pedigree. Roman numerals represent the generation. Arabic numerals identify individuals. Arrows indicate the probands. Mutations on the bottom of each pedigree. (Fig. 1 should be placed before the section “Result”)

Full size image

Molecular findings

Gene testing identified five ERLIN2 variants, one of which was a compound heterozygous (Fig. 1). The c.502G > A [p.Val168Met], c.866 T > C [p.Ile289Thr], and c.374A > G [p.Asn125Ser] pathogenic variants have already been reported [19, 20, 26]. The other two variants, i.e., the heterozygous c.615G > C [p.Lys205Asn] (CADD score is 24.1) and c.481C > A [p.Pro161Thr] (CADD score is 27.2—found in compound heterozygosis with c.866 T > C [p.Ile289Thr], previously reported [26]) were absent in our in-house databases as well as in population databases (dbSNP, 1000genome, gnomAD) and were classified initially as “hot” VUS and then revised as likely pathogenic according to the ACMG guidelines. Missense variants were indicated as “probably damaging” and “damaging” by multiple in silico predictors (including PolyPhen-2 and SIFT).

Discussion

We present five previously unreported Italian patients with SPG18. These patients harbored already published disease-associated variants in ERLIN2, and in two cases, we detected novel, likely pathogenic variants. This is the 14th report about ELRIN2-related motor neuron disease (Table 2). Thus far, a total of 80 cases from 23 SPG18 families (14 biallelic, 9 monoallelic-3 cases presented as sporadic) have been described worldwide [10,11,12,13,14,15,16,17,18,19, 26, 27] (Table 2) since the original description by Al-Yahyaee et al. in 2006 [10]. In their work, patients showed complicated AR-HSP (6 cases with intellectual disability and thin corpus callosum, 3 with seizures). ERLIN2 mutation as causative of SPG18 was first mapped in an AR Turkish family with early onset intellectual disability, motor impairment, and multiple joint contractures in 2011 [11]. Meanwhile, a Saudi family with a mutation in the same locus (flanking the D8S532 marker) received a formal diagnosis. Afterwards, other ERLIN2 pathogenic variants were identified in additional AR-SPG18 [13,14,15, 17, 26, 27] families. Data showed a rather uniform phenotype characterized by early onset, complicated and severe spastic paraparesis, usually with a short disease duration. Shortly after, five families with pure HSP harboring monoallelic ERLIN2 missense mutations had been separately described [16,17,18,19]. The clinical manifestation in those families manifests a late onset, progressive and pure spastic paraparesis, with mild dorsal column abnormalities, bearing a strong resemblance to that of other AD-HSP cases. The comparison between AR-SPG18 and AD-SPG18 forms led to the hypothesis that the clinical phenotype of SPG18 may depend on the mode of inheritance, with AD-SPG18 manifesting as a juvenile-adolescent onset pure HSP, whereas AR-SPG18 being a complicated form with earlier onset and more severe course as the likely consequence of a complete loss of function [14]. Presumably, the underlying molecular mechanism in AD-SPG18 is different. Identification of domain-specific mutations in cases from both European and Asian populations makes possible that monoallelic pathogenic variants induce a site-specific dominant negative effect, by disrupting the ERLIN2/ERLIN1 complex, which leads to the more pure and less severe phenotype [28]: functional studies would be required to clarify such issue. To date, only a limited number of HSP loci have been documented to potentially have both AD and AR inheritance patterns, like REEP2/SPG72, ALDH18A1/SPG9, KIF1A/SPG30, and ATL1/SPG3A [29,30,31,32]. This emphasizes the conundrum in categorizing HSPs solely based on their observable characteristics and that the notion of zygosity no longer serves as a constraining factor when determining a molecular diagnosis in HSP. Indeed, a broader range of HSP genes might be associated with different inheritance patterns, thus having implications for the diagnostic success rate and potentially indicate variations in disease-related characteristics. The existence of different inheritance modes per single disease clearly has implications for genetic counselling, ousting the classic division into AD, AR, X-linked, or mitochondrial patterns. In our series, 4 out of 5 patients have a monoallelic pathogenic variant, and they show a pure HSP. Considering our cases, it appears that AD-SPG18 is as prevalent as AR-SPG18 worldwide. This consideration has implication for counselling and prognosis.

Amador et al. [20] described four pedigrees with an ALS form: two families with monoallelic pathogenic variants (c.502G > A [p.Val168Met] and c.926C > T [p.Ala309Val]), one family with the biallelic pathogenic variant c.899A > T [p.Asp300Val], and a sporadic case with monoallelic pathogenic variant c.374A > G [p.Asn125Ser]. Except for one of the AD families (c.926C > T [p.Ala309Val]), which showed an ALS onset and course without HSP signs, the other cases all exhibited pure HSP-ALS phenoconversion after 20–39 years. After developing ALS-like syndrome, four patients died after a rapidly evolving disease in about 12–18 months. Indeed, from our series, we report two cases of pure HSP-ALS evolution. Patient AI.2, mother of AII.1 with monoallelic pathogenic variant c.502G > A [p.Val168Met], and BII.1 with monoallelic pathogenic variant c.615G > C [p.Lys205Asn], after a long duration of slowly progressive pure HSP (27 years in AI.2, 20 years in BII.1), showed a phenoconversion to an ALS-like syndrome. In one case (AI.2), this condition was rapidly progressive, ultimately leading to death in less than 1 year. To date, nine SPG18 patients with HSP-ALS have been described worldwide. Despite the few cases, we can observe that HSP-ALS conversion is more frequent in AD cases than in AR ones (seven versus two—Tables 1 and 2). Also, phenoconversion seems to occur only in phenotypically pure SPG18 cases (both AD and AR).

A last comment deserves the incidence of the c.502G > A variant, occurring in 11 cases worldwide [19, 20], and only found as a monoallelic gene change. We cannot exclude this variant as a hotspot for a peculiar phenotype of SPG18 with ALS-like syndrome, whereas the multiple genetic background of reported patients makes it unlikely a common ancestor. However, further research is needed to elucidate the disease mechanisms of ERLIN2-related disorders, as well as improved genotype–phenotype correlations.

Conclusion

We report a large series of Italian SPG18 patients, confirming the prevalence uniformity of both AD and AR forms, as previously described. Phenoconversion of SPG18 into ALS-like syndrome seems to be more frequent in AD-SPG18. We expand the mutational scenario, adding new c.502G > A [p.Val168Met] AD cases with HSP-ALS phenoconversion, thus pointing out the predictive value of this pathogenic variant. The link between HSP and ALS is well known [33,34,35]. Several HSPs may show ALS-like syndrome during their course, like SPG7 [36], 10 (allelic with Charcot-Marie-Tooth type 2) [37, 38], 11 (allelic with Charcot-Marie-Tooth type 2X) [39, 40], 15 (Kjellin S.) [41], 17 (Silver S.) [42], 20 (Troyer S.) [43]. On the other hand, a few familial ALSs are caused by HSP mutations. The possibility of HSP-ALS phenoconversion has significant clinical implications. Clinicians should be aware of the potential for ALS-like symptoms in individuals with HSP, especially those with mutations in genes that are shared between the two disorders. On the contrary, it is important to bear in mind that, in cases of familial ALS [40], genetic testing should include HSP panels. Recognition of phenoconversion is crucial, as it can inform prognosis, management, counselling, and treatment decisions.