High diagnostic value of plasma Niemann‑Pick type C biomarkers in adults with selected neurological and/or psychiatric disorders
Abstract
Late-onset Niemann-Pick type C (NP-C) is a rare, underdiagnosed lysosomal disease with neurological manifestations. A specific treatment, miglustat, can stabilize the disease if given early. Recently, three plasma screening biomarkers (PSBs) were developed [cholestane3β,5α,6βtriol (C-triol), 7-ketocholesterol (7-KC), and lysosphingomyelin-509 (LSM-509)], allowing a simpler and quite robust screening of patients suitable for genetic testing. The objective of our study was to evaluate practical utility and feasibility of large-scale PSB screening for NP-C in selected adult patients. Patients were prospectively enrolled if they showed, starting from 12 years of age, at least one of the three initial neuro-psychiatric manifestations described in NP-C: (1) gait disorder (cerebellar and/or dystonic); (2) cognitive decline with frontal lobe syndrome; (3) atypical psychosis. PSBs were measured in plasma of all patients and, if positive (LSM-509 and/or C-triol + 7-KC elevated), sequencing of NPC1 and NPC2 genes was performed. A total of 251 patients [136 males, 115 females; median age 42.1 (range 12.2–85.6) years] were screened. Six patients had positive PSBs. Two were confirmed to have NP-C (0.8% diagnostic yield, both with all three PSBs highly increased, especially LSM-509). False-positive rate was 1.2%, which was identical if only considering LSM-509. By contrast, false-positive rates were 8.1% and 5.7% for 7-KC and C-triol, respectively. We showed that selecting patients with neurologic and/or psychiatric symptoms consistent with NP-C for large-scale PSB screening is a simple and valid strategy to identify new adult NP-C patients, and would probably lead to earlier diagnosis and treatment administra- tion if widely applied.
Keywords : Niemann-pick type C · Biomarkers · Lysosphingomyelin-509 · Screening
Introduction
Niemann-Pick type C (NP-C; OMIM #257220) is a genetic lipid storage disease with autosomal recessive inheritance, resulting from a malfunction of cholesterol trafficking within the late endosome/lysosome (LE/LY) [1]. NP-C has an extremely wide clinical spectrum, substantially varying with the age of onset. In the late-onset form of the disease (i.e., starting from the second decade of life), neurologi- cal symptoms at onset are gait disorder (cerebellar and/or dystonic), and/or cognitive decline, and/or psychosis [2].
Other manifestations can occur and may help for achieving the correct diagnosis but none is constant or specific (pre- ceding hearing loss; vertical saccadic gaze palsy (VSGP); splenomegaly and/or hepatomegaly; dysarthria, dysphagia, and myoclonus later on). Miglustat is the only available treatment specific for NP-C, and it was reported to stabilize disease progression in a significant subset of late-onset NP-C patients, especially when treated early in the course of the disease [3].
While a complex procedure needing a skin biopsy (filipin test) was formerly used to screen patients for NP-C, three plasma screening biomarkers (PSBs) have been recently introduced in clinical practice: cholestane-3β,5α,6β-triol (C-triol) and 7-Ketocholesterol (7-KC) (both choles- terol oxides, or oxysterols) and lysosphingomyelin-509 (LSM-509). These PSBs demonstrated good diagnostic per- formances (sensitivity 87.5–100% and specificity 91–100%) [4–6]. Therefore, they could be easily used to screen a larger number of adult patients with a clinical picture consistent, even if not necessarily specific, with late-onset NP-C, in order to improve detection of unrecognized cases and achieve earlier diagnosis. However, this promising diagnos- tic strategy needs evaluation.
In this study, we prospectively screened for NP-C, in a real-life setting and through PSB measurement, an adoles- cent/adult population having clearly defined clinical pheno- types, i.e., at least one of the three initial cardinal manifesta- tions (CM) described in late-onset NP-C: (1) gait disorder related to cerebellar ataxia and/or dystonia; (2) cognitive decline with prominent frontal lobe syndrome; and (3) atypi- cal psychosis [1].
Methods
This was a multicenter prospective study conducted between August 2016 and November 2018 (28 months) in France. Referring physicians were asked to fill a standardized form to collect clinical information from the enrolled patients, who gave their written consent. Patients included pre- sented one or more CM of unclear etiology, starting after the age of 12: (1) gait disorder (cerebellar and/or dystonic),
(2) cognitive decline (if associated with frontal lobe syn- drome), and (3) atypical psychosis (i.e., psychosis associ- ated with at least one of intellectual disability; prominent visual hallucinations; catatonia; resistance to at least three neuroleptics; onset before the age of 15). Patients were cat- egorized by investigators (D.M. and Y.N.) as “simple” (one CM) or “complex” (≥two CMs), and as “low”, “medium”, or “high suspicion” for NP-C according to the NP-C suspi- cion index (NP-C SI) [7]. Blood samples of the patients were drawn at the time of recruitment as part of standard care and sent to our Metabolic Biochemistry Department (eMethods). We set PSB cut-offs on the basis of the results we had in our cohort of late-onset NP-C patients (n = 22) (Fig. 1). We considered patients PSB + if they had high levels of LSM- 509 (i.e., > 94 nmol/L) and/or if both C-triol and 7-KC were found increased (i.e., > 32 and 76 nmol/L, respectively). Next-Generation Sequencing (NGS) was performed in all PSB + patients, analyzing coding exons and exon–intron junctions of NPC1 and NPC2 genes, through bioinformat- ics pipeline allowing detection of single-nucleotide variation (SNV) and copy number variation (CNV).
Results
A total of 288 patients were initially enrolled. Thirty-seven patients were excluded (lack of inclusion criteria or insuffi- cient clinical data), resulting in a final cohort of 251 patients [136 males, 115 females; median age 42.1 (range 12.2–85.6) years] (Fig. 2). Figure 3 illustrates the distribution of the three cardinal manifestations (CM) among the included patients: 149 had one CM (59.4%,), 96 had two (38.2%), and 6 patients had all three CM (2.4%). A total of 104 patients had a gait disorder (41.4%), 123 had cognitive decline (49%), and 132 patients had a psychotic condition (52.6%). Six patients (2.4%) were PSB + : five had all three biomark- ers increased, one had C-triol and LSM-509 increased but normal 7-KC. Five out of the six PSB + patients had NGS (one was lost to follow-up): two had two heterozygous muta- tions in NPC1 (patients NP-C#1 and NP-C#2, see details in Table 1), whereas no mutation was found for three patients. One of these three negative patients was lately reported to harbor bi-allelic mutations in SACS gene, responsible for Autosomal Recessive Spastic Ataxia of Charlevoix-Sague- nay (ARSACS).
The two confirmed NP-C patients (both “complex”) had the three biomarkers highly elevated, especially LSM-509, this latter reaching the values of 16.9 (patient NP-C#1) and 7.8-fold (patient NP-C#2) the upper limit of normal range (Fig. 4). Overall false-positive cases were 1.2% (3/248). If we only consider LSM-509, the rate of false-positive cases was identical, whereas it was 8.1% and 5.7% for 7-KC and C-triol, respectively. If we examine only the group of “com- plex” patients (n = 102), the diagnostic yield increased from 0.8% (2/251) to 2% (2/102), with no false-positive cases identified (Fig. 2). Considering only “high NPC-SI” patients, diagnostic yield was 2.3% (2/87), with three false-positive cases.
During the same 28-month period, 534 blood samples from patients ≥ 12 years old were sent for NP-C screening in our laboratory, outside our prospective screening study. Regarding these patients, limited or no clinical data were available. Among them, three were diagnosed with NP-C, all having the three PSBs found abnormally high. Their phe- notype, retrospectively collected via the same questionnaire used in the screening study, is also described in Table 1. All three had at least one CM.
Discussion
Our study demonstrates that large-scale PSB screening in selected adult patients with defined neurologic and psy- chiatric can be a simple, efficient (0.8% diagnostic yield),accurate (1.2% false-positive cases), and probably cost- effective approach to identify late-onset NP-C. We found no evidence that measuring C-triol and 7-KC in addition to LSM-509 could improve accuracy, but combined oxysterol measurement in our study had the same diagnostic yield and false-positive rate than LSM-509 alone, and therefore possibly useful if LSM-509 measurement is not available. Targeting “complex” patients (i.e., presenting at least two of the CMs used in this study) increased diagnostic yield up to 2%, a very high percentage for an ultra-rare disease like NP-C. However, this approach would probably delay the diagnosis, since more than 40% of late-onset patients only have one CM at onset [1]. Therefore, in a context of a single unexplained CM, to avoid patients over-screening, we suggest that presence of additional NP-C signs (mainly hearing loss, VSGP, and visceromegaly) should definitely prompt to screen for NP-C. Another approach to improve the screening procedure would be to find, within a single CM, better semiological clues to select patients for NP-C screening, such as (1) a mixed gait disorder where cerebellar ataxia is associated with dystonia; (2) a prominent impair- ment of frontal cognitive functions in cognitive decline, with a relative sparing of instrumental functions, at least initially. NPC-SI, currently used to identify subjects at risk for NP-C, despite its accuracy, is not widely known among neu- rologists or psychiatrists, is time demanding, and requires items that are not often sought (as visceromegaly) or whose assessment is often challenging (e.g., cataplexy or VSGP) for clinicians with no expertise in NP-C. By contrast, the selection strategy for screening described here is very simple to use in clinical practice, and was shown to be as relevant as NPC-SI. We therefore suggest to routinely select adult patients for NP-C screening as we did in our study, instead of complexifying the diagnostic process using NPC-SI tool. In medical literature, several screening studies for NP-C have been conducted, all with some limitations: (1) the use of gene sequencing as a screening test, which is expensive, and sometimes difficult to interpret due to NP-C genetic complexity [8, 9]; (2) the use of a single biomarker (C-triol [10–12] or 7-KC [13] or LSM-509 [14]); (3) the lack of defined inclusion criteria for tested patients [10–12, 14], which precludes to identify a target population for screen- ing. Our study used the three biomarkers mostly reported in the literature, which are accurate and cost-effective, especially if compared to genetic tests, and allowed a com- parison between them. We specifically targeted adolescent/ adult patients, who are known to have a clinical phenotype [2] and a biomarkers profile [4, 6, 15, 16] clearly different from children. Lastly, we used defined inclusion criteria, based on a descriptive study conducted on a population of adult patients with NP-C [1]: this has a major importance because NP-C is a highly heterogeneous disease with dif- ferent phenotypes, changing with the age of onset.
As limitations, we did not perform gene sequencing for NP-C in all the patients included in this study, thus possi- bly missing some cases. However, several studies, as well as our cohort of adult patients diagnosed with NP-C, have previously shown LSM-509 to bear a sensitivity of 100% in excluding NP-C [5, 14, 17]. In addition, we specifi- cally ask clinicians to perform genetic sequencing if they retained nevertheless a high suspicion of NP-C despite the negativity of PSBs, but no further patient was tested. Also, among PSB + patients with negative NP-C genetic testing (n = 3), we did not perform intronic studies. However, one was finally reported to have two mutations in SACS gene, responsible for the Autosomal Recessive Spastic Ataxia of Charlevoix-Saguenay (ARSACS), and regarding the two others, the possibility of two undetected deep intronic muta- tions is unlikely, notably in the absence of consanguinity, as those mutations are very rare (< 1% of disease-causing mutations in NPC1 reported in HGMD Pro database, and no deep intronic pathogenic mutation reported in NPC2 so far). In conclusion, we showed that selecting patients with neurologic and/or psychiatric symptoms consistent with NP-C for large-scale PSB screening is a simple and valid strategy to identify new adult NP-C patients, and would probably lead to earlier diagnosis and treatment administration if widely applied.