In the present study, we tried to explore the association of phenotype-genotype in NPD comprehensively and lay the foundation for understanding the mechanisms of this rare disease. With the strict quality-controlled literature search, we collected 144 cases with comprehensive pathophysiological characteristics of NPD. We also found a connection between some variants and the phenotypes (NPD type A or B). Type A is correlated with more severe mutations, while patients with the non-neuro-related type NPB normally have mutations in SMPD1 with a mild effect. In addition, following the model, we have found a threshold of 4.45% and that it can be taken into account to discriminate the majority of cases with NPA from clinical phenotypes less severe of NPD (intermediate and NPB). Furthermore, we also explored the expression landscape of the SMPD1 in different cell types of fetal development and adult tissues, which offered us the opportunities to better understand pathogenic mechanisms underlying NPA and NPB at a single cell type level. At the same time, the difference in SMPD1 expression levels on different cell types provides an important resource for the precise diagnosis of the disease in clinical application.
The residual ASM activity has been regarded as one of the clinical features to distinguish NPA from NPB in Chinese people (13), while some bibliographic data that supports ASM residual activity threshold is not definitive for discriminating between type A and B (23–25). Normally, < 5% of effective residual ASM activity in situ is observed in NPA, whereas 5–20% is detected in NPB (8, 26–28). However, in the literature we have mined, > 5% of the cases were still diagnosed as NPA (Ding et al., 2016; Ceron-Rodriguez et al., 2019; Al-Eitan et al., 2020). Several intermediate cases indicated that the residual activities of the ASM enzyme were broad (23, 29). Therefore, the ASM activity is not always associated with the so-called well-defined subtypes. Our model indicated that 4.45% of ASM activity could be the threshold to distinguish NPA from other subtypes. However, Hu, Maegawa (13) suggested that the cutoff value for differentiating the two clinical forms was 1.685 nmol/17 h/mg protein (approximately 12.2% to the reference, 13.7 nmol/17 h/mg protein as the reference value in Chinese patients only, and all ASM activities were measured within single laboratory using the same method. Although the SMPD1 gene sequencing appears to be a golden standard for NPD diagnosis, the method might be less available in the less-developed regions. ASM activity is still a standard for deficiency diagnosis (30). The threshold proposed in the present study was derived from multinational samples (i.e., leukocytes, skin fibroblasts, or dried blood spots) and different ethnic groups (31) with different measurement approaches to ASM activity, which further indicates that ASM activity is a common feature used to differentiate NPA from other subtypes for counselling, prognostication, and the interpretation.
SMPD1 gene mutations have been reported in many countries and ethnic groups, and its mutation prevalence varies from one ethnic group to another (9, 16–18, 20, 32). In this study, we observed frequency differences in the same sites of SMPD1 protein between the East Asians in gnomAD and Han Chinese in the "Huabiao" project (Fig. 3C). In the gnomAD, the most common variant in East Asians is p.Lys189GlnfsTer4 (c.564dup), while p.Glu508Lys (c.1522G > A) is the most common one in Han Chinese. Besides, the high-frequency mutation sites of the SMPD1 gene highly vary in different populations reported in the literature, such as Ashkenazi Jews, Italians, Spanish, Turks, Chinese, and Dutch. Moreover, p.F333SfsX52, p.L304P, and p.R498L are the most common SMPD1 gene mutations among Ashkenazi Jews, which morally cause NPA (33, 34). This racial difference in the mutation frequency may also contribute to a massive difference in the prevalence of NPA and NPB in various races and cause different phenotypes. Globally, the most common mutation is p.Arg610del, which has been associated with NPB (18, 35). Similarly, p.Arg610del is also dominant in our collected cases. In contrast, the most common mutations among Chinese patients are p.Arg3AlafsX76 and p.H284SfsX7 (18).
Furthermore, with the two databases, ClinVar and ANNOVAR, and a deep learning algorithm to improve the reliability of the EVE model, we predict unreported 21 variants that could be pathogenic (36), which can provide new information to interpret the related variants in SMPD1 gene testing for NPD. The comparison of different databases shows the frequency of variant sites of the SMPD1 gene in the Chinese Han group from Huabiao to the east Asian in the gnomAD. It is believed that these high-risk mutations might lead to spontaneous abortion as the SMPD1 gene expression is high in the CNS system during development. These novel variants aggregate in the domains of the protein. Fourteen variants (66.67%) are involved in sphingolipid metabolism reactions in the Calcineurin-like phosphodiesterase domain (from 255th to 462nd amino acid). Sphingomyelin was included in this type of phosphodiesterase superfamily, demonstrating that the amino acid changes due to variants would impact the function of the ASM and finally lead to the severe phenotype - NPA or the milder one - NPB. Metallo-dependent phosphatase-like domain (202nd to 497th amino acid) found 80.95% of the variants. This domain is associated with metabolite damage-control (37). Hence, if the variants occur in these domains, it is highly likely to lead to LSD, even NPD (38).
Studies that comprehensively expounded pathways related to NPD are barely found. In the present results, within the data above, we infer that the following scenario could be the mechanisms underlying NPD. The expression profiles of SMPD1 in cells and tissues in healthy people help explain the complex symptoms of NPD. We can further connect the clinical phenotypes to the mutation pattern based on the SMPD1 expression profiles in fetal and adult tissues. ASM is an enzyme essential for neurodevelopment. Normally, the mutations in the catalytic domain of SMPD1 have severe pathogenic effects because the lost catalytic function of the enzyme can significantly decrease ASM activity, which causes the accumulation of sphingomyelin and other sphingolipids that are toxic at elevated and nonphysiological levels. The clinical manifestations include rapid progressive psychomotor deterioration, liver and spleen enlargement, respiratory disease, jaundice, and death within three years (13, 17, 39). Liver and spleen enlargement could be a compensation mechanism of the body to sustain ASM activities. Considering that SMPD1 is universally expressed in many different cell types and tissues, it is expected that the dysfunction of SMPD1 should have a significant impact on many tissues, indicating that symptoms of NPD should present in the whole body without abundant specificity. Consistently, the phenotypes reported in NPD are all consistent with short stature, osteoporosis, sea-blue histiocytosis, microcytic anaemia, and bone-marrow foam cells. The expression profile of SMPD1 in various cell types during development and various adult tissues can help us comprehensively decipher the potentially affected cell types and tissues of SMPD1 mutation, which might have been ignored clinically. In addition, according to its expression profile, SMPD1 is highly expressed in the heart, pancreas, eyes, and kidneys during fetal development, which indicates that the dysfunction of the SMPD1 gene should have possibly influenced these organs and the related phenotypes such as renal involvement in NPD is rarely reported (40, 41). Therefore, in clinical application, clinicians are recommended to conduct comprehensive examination during diagnosing patients with potential NPD, paying attention to the pathological abnormalities of these organs, such as the heart, pancreas, and kidneys, during fetal development hepatosplenomegaly, splenomegaly, and neurological abnormalities.
We further deduced that potential pathogeny at a gene level could correlate with the types of mutations. Severe mutations resulting from deletion or insertion and stop gain led to the premature termination of the synthesis of the polypeptide chain of the SMPD1 gene, or mutated polypeptide chains produce enzymes without biological activity or barely active domains. In patients with NPB, a single missense mutation only changes an amino acid, resulting in defective ASM with partial catalytic activity. Therefore, the ASM activity of NPB is higher than that of NPA, which explains the perspective that patients with NPA/NPB have the same pathogenic mutated genes, but the clinical manifestations are quite different. The pathogenic mutations of the SMPD1 gene are primarily found in compound heterozygotes; the phenotype-genotype association study is particularly complicated. Thereupon, the expression profiles of the SMPD1 in different cell types of fetal development and adult tissues would be a tool for understanding the pathogenic mechanisms underlying NPA and NPB. The period from 2 weeks post-conception to early childhood is crucial for developing the brain and other CNS organs (42). Large amounts of sphingomyelin are needed to be converted to develop non-CNS cells (42). In the present study, the expression pattern of SMPD1 (Fig. 4A) in various cell types of fetal demonstrated that SMPD1 dysfunction should significantly affect the functions and development of the circulatory system (heart and kidneys) and nervous system, which are essential for the survival of the fetal and infants. The placenta, vital to support fetal growth, also presents a high SMPD1 gene expression. The individual clinical symptoms strongly correlate with the severity of SMPD1 mutations, as the mutations would result in the functional decrease or even loss of the ASM activity in those cell types. Once those mutations cause decreased or forfeit ability of ASM results in the ASM substrates and sphingomyelin accumulation, which would negatively affect individual fetuses. Finally, the excessive amount of accumulated sphingomyelin might lead to NPD phenotypes at an early age (namely, Type A) or the NPB (the late-onset phenotype). Among 53 adult tissues, the expression of SMPD1 is relatively high in the liver and the spleen (liver ranked 13, and the spleen ranked 20, Fig. 4C), which also suggests high levels of sphingomyelin in both tissues. Individuals with low ASM activities might not be able to convert sphingomyelin timely; thus, patients with NPA and NPB are featured with progressive hepatosplenomegaly and other organ dysfunction (3, 17, 43).
Our study is the first to comprehensively elucidate the effects of SMPD1 mutation on cell types and at the tissue level, which provides new insights into the genotype-phenotype association and can help in the precise diagnosis of NPD. Admittedly, our study has certain limitations; the number of cases included in this study is relatively small, which could influence the AUC results; more cases should improve the model's performance. In this study, we fail to comprehensively detect the relationship between phenotypes and genotypes because of incomplete phenotype data from some reported cases. In addition, we found the area or race specificity to the frequency of the variants, but it should be noted that some mutations with population or area-specific prevalence could also result from the bias of case study and collection. However, we compared the frequency of the variants collected to the public databases, gnomAD and the results are consistent. For example, p.Arg610del are the most domain variant in both the documented NPD patients (gnomAD databases) and the cases we collected.