The spectrum of LYST mutations in Chediak-Higashi syndrome: a report of novel variants and a comprehensive review of the literature

Abstract

Introduction:

Chediak-Higashi syndrome (CHS) is a rare autosomal recessive disorder characterized by partial oculocutaneous albinism, a bleeding diathesis, immunological dysfunction, and neurological impairment. Bi-allelic loss-of-function variants in LYST cause CHS. LYST encodes the lysosomal trafficking regulator, a highly conserved 429 kDa cytoplasmic protein with an unknown function.

Methods:

To further our understanding of the pathogenesis of CHS, we conducted clinical evaluations on individuals with Chediak-Higashi syndrome enrolled in our natural history study. Utilizing genomic DNA Sanger sequencing, we identified novel pathogenic LYST variants. Additionally, we performed an extensive literature review to curate reported LYST variants and classified these novel and reported variants according to the ACMG/AMP variant interpretation guidelines.

Results:

Our investigation unveiled 11 novel pathogenic LYST variants in eight patients with a clinical diagnosis, substantiated by the presence of pathognomonic giant intracellular granules. From these novel variants together with a comprehensive review of the literature, we compiled a total of 147 variants in LYST, including 61 frameshift variants (41%), 44 nonsense variants (30%), 23 missense variants (16%), 13 splice site variants or small genomic deletions for which the coding effect is unknown (9%), 5 in-frame variants (3%), and 1 start-loss variant (1%). Notably, a genotype-phenotype correlation emerged, whereby individuals harboring at least one missense or in-frame variant generally resulted in milder disease, while those with two nonsense or frameshift variants generally had more severe disease.

Conclusion:

The identification of novel pathogenic LYST variants and improvements in variant classification will provide earlier diagnoses and improved care to individuals with CHS.

INTRODUCTION

Lysosome-related organelles (LROs) are specialized cellular compartments that exhibit a unique morphology, composition, and function. They play crucial roles in diverse physiologic processes such as pigmentation, blood coagulation, immunity, and neurological function. Dysfunctional LROs can arise from Mendelian LRO-related disorders, including Chediak-Higashi syndrome (CHS, MIM 214500). CHS is a rare autosomal recessive disease characterized by partial oculocutaneous albinism, a bleeding diathesis, immunological dysfunction, and neurological impairment [1]. Individuals with CHS are also more likely to develop hemophagocytic lymphohistiocytosis (HLH), a life-threatening hyperinflammatory condition that manifests as persistent fever, cytopenia, hepatosplenomegaly, and multiorgan failure, which can be effectively treated by allogeneic hematopoietic stem cell transplantation (HSCT) [2]. The clinical features of CHS stem from abnormalities in various cell-type specific LROs: abnormal melanosomes in melanocytes lead to defective pigmentation, deficient dense granules in platelets lead to abnormal aggregation and a bleeding diathesis, aberrant azurophilic granules in the neutrophils lead to impaired chemotaxis and reduced bactericidal activity, and atypical lytic granules in natural killer (NK) cells and cytotoxic T cells lead to defective cytotoxicity (Figure 1A). A diagnostic hallmark of CHS is the presence of enlarged intracytoplasmic granules, which can be observed on peripheral blood smear (Figure 1B) and serve as a pathognomonic characteristic [1].

Figure 1.

Clinical findings of Chediak-Higashi syndrome (CHS) and affected lysosome-related organelles (LROs) in CHS. (a) A schematic of the affected LROs and respective cell types in CHS and their associated clinical features including melanosomes in the melanocyte leading to partial oculocutaneous albinism (OCA), dense granules in the platelet leading to a bleeding diathesis, azurophilic granules in the neutrophil and lytic granules in NK and cytotoxic T cells leading to immunological dysfunction. (b) Clinical findings of CHS include enlarged azurophilic granules in a neutrophil on peripheral blood smear (left), altered distribution of melanin in the hair shaft by light microscopy analysis (center), and iris transillumination of the eye by slit-lamp biomicrography (right). Abbreviations: CHS, Chediak-Higashi syndrome; LRO, lysosomal-related organelle; NK cell, natural killer cell; OCA, oculocutaneous albinism.

Significant clinical overlap is observed between CHS and other Mendelian multisystemic LRO-related disorders, including Hermansky-Pudlak syndrome and Griscelli syndrome. Hermansky-Pudlak syndrome is a phenotypic series due to recessive mutations in multiple heterogeneous loci (MIM 203300, MIM 608233, MIM 614072, MIM 614073, MIM 614074, MIM 614075, MIM 614076, MIM 614077, MIM 614171, MIM 617050, MIM 619172) characterized by oculocutaneous albinism, a bleeding diathesis, neutropenia, granulomatous colitis, and pulmonary fibrosis [3 4]. Griscelli syndrome (MIM 214450, MIM 607624, MIM 609227) is an autosomal recessive disorder characterized by partial albinism alone or in combination with neurological or immunological dysfunction. These disorders are caused by loss-of-function variants in genes encoding protein complexes required for LRO biogenesis and transport, respectively [4].

CHS is caused by bi-allelic loss-of-function variants in the LYST gene [5 6]. The LYST gene is located on chromosome 1q42.3 and comprises 53 exons within a 13,466 bp mRNA transcript spanning approximately 222 kb of genomic DNA [5 7] (Figure 2A). LYST encodes the lysosomal trafficking regulator (LYST), a highly conserved 429 kDa cytoplasmic protein with an unknown function [5 8]. LYST contains several functional domains, including the highly conserved pleckstrin homology (PH) and beige and CHS (BEACH) domains as well as several ARM/HEAT and WD40 domains throughout the protein [5] (Figure 2B). Human LYST is expressed in lymphoid tissues such as the thymus, lymph node, and spleen [6]. In the mouse, Lyst mRNA has the highest levels of expression in the brain, lung, liver, kidney, and muscle, with lower levels in the heart, spleen, and testis. The LYST protein has been detected in the brain, lung, and spleen [8]. Studies using cellular and animal models suggest that LYST plays a role in the biogenesis and maturation of lysosomes and LROs, as well as in the vesicular transport of proteins in the endolysosomal system [9–20].

Figure 2.

Novel and previously reported LYST variants in Chediak-Higashi syndrome (CHS). (a) Schematic of LYST mRNA showing the locations of previously reported (white symbols) and novel (red symbols) variants found in this study. The ATG start codon and stop codon are represented in the schematic as arrowheads; coding regions are indicated in grey and untranslated regions are indicated in light grey. Pathogenic start-loss (pentagon), missense (circle), frameshift (square), nonsense (diamond), in-frame (triangle), and splice site (star) variants are present throughout the LYST gene. This schematic is not drawn to scale. Note that p.Y2026* is the predicted amino acid change for c.6077dup and c.6078C>A. (b) Schematic of LYST protein and its functional domains, including the seven WD repeats (red), the pleckstrin homology (PH) domain (dark blue), and the beige and CHS (BEACH) domain (light blue). This schematic is drawn to scale. The missense and in-frame variants represented in (a) are color-coded in red or light blue if they are located in a WD repeat or the BEACH domain, respectively. (c) The percentages of each coding effect are represented as pie charts for novel variants (n = 11), reported variants (n = 136), and all variants (n = 147). Variants for which further studies have been performed at the level of RNA or protein (n = 9) have been categorized according to the coding effect determined by these studies. Of these, 1 frameshift variant and 2 in-frame variants were due to substitutions at a canonical splice site, 2 frameshift variants were due to small deletions of a canonical splice site, and 3 frameshift variants and 1 in-frame variant were due to variants leading to partial or complete exon skipping determined by cDNA Sanger sequencing but for which the causative gDNA change could not be identified. Variants for which the coding effect was unknown include canonical (n = 10) and non-canonical (n = 2) splice site variants as well as a single exon deletion (n = 1). Abbreviations: BEACH, beige and CHS domain; CHS, Chediak-Higashi syndrome; LYST, lysosomal trafficking regulator; PH, pleckstrin homology domain; WD, WD40 domain.

The advent of next-generation sequencing has revolutionized the analysis of genetic data, enabling rapid identification of pathogenic variants for rare disorders like CHS. Although variant interpretation remains a challenge, it serves as a powerful tool for molecular diagnosis when paired with a comprehensive clinical evaluation and clinical acumen. In this study, we describe 11 novel variants in eight individuals with CHS, further expanding the genetic spectrum of LYST variants in CHS. Additionally, we present a comprehensive review of all reported LYST variants and classify them using the ACMG/AMP variant interpretation guidelines. Furthermore, we discuss genotype-phenotype correlation and highlight the tools needed for the molecular diagnosis of these patients.

MATERIALS AND METHODS

Patients

A cohort of individuals with CHS was evaluated clinically under the protocol, NIH clinical protocol 00-HG-0153, “Investigations into Chediak-Higashi Syndrome and Related Disorders” (NCT00005917, clinicaltrials.gov) or NIH clinical protocol 76-HG-0238, “Diagnosis and Treatment of Patients with Inborn Errors of Metabolism or Other Genetic Disorders” (NCT00369421, clinicaltrials.gov) approved by the National Human Genome Research Institute’s Institutional Review Board and provided written informed consent. Clinical evaluation included a patient and family history, physical examination, a comprehensive review of the medical records, a complete blood count, and collection of biospecimens including primary cells. Whole-mount electron microscopy was used to assess platelet dense granules that are often decreased in number in CHS. Peripheral blood smear analysis and microscopic hair examination were used to assess the presence of giant intracytoplasmic granules in leukocytes and pigment clumping in hair shafts, i.e., pathognomonic clinical features of CHS.

Cell culture

Primary dermal fibroblasts from individuals with CHS were cultured from forearm skin biopsies using standard protocols. Fibroblasts were cultured in high glucose DMEM (11965092, Gibco) with 10% fetal bovine serum (10082, Gibco) and 1× penicillin-streptomycin (15140122, Gibco) at 37°C with 5% CO₂. A lymphoblastoid cell line from an individual with CHS (GM02431, Coriell Institute for Medical Research) was cultured in RPMI 1640 medium (61870036, Gibco) with 10% fetal bovine serum (10082, Gibco) and 1× penicillin-streptomycin (15140122, Gibco) at 37°C with 5% CO₂.

Genetic analysis

Genomic DNA (gDNA) was extracted from peripheral blood or cultured primary dermal fibroblasts or lymphoblastoid cell lines using standard procedures. Regions of the LYST (NM_000081.3) coding sequence (exons 3 to 53), including intron-exon boundaries, were amplified by PCR using the primers listed in Supplementary Table 1. The size and specificity of the PCR products were confirmed by agarose gel electrophoresis. Sanger sequencing was performed using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) and separated on the 3130xl Genetic Analyzer (Applied Biosystems). Data were analyzed using Sequencher v5.0 software (Gene Codes Corporation, Ann Arbor, MI).

Literature search

A comprehensive literature search was conducted using PubMed (https://pubmed.ncbi.nlm.nih.gov) to gather reported variants in LYST associated with CHS from peer-reviewed publications. Additional databases were queried, including ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/), the Leiden Open Variation Database (LOVD version 3.0, https://www.lovd.nl/), and the Human Gene Mutation Database (HGMD Professional 2021.4, www.hgmd.cf.ac.uk/ac/index.php). The search in all databases was last performed in June 2022. All LYST variants are described according to the current Human Genome Variation Society (HGVS) nomenclature guidelines based on National Center for Biotechnology Information (NCBI) reference sequence NM_000081.3 (GRCh37/hg19 genome assembly).

The classification of variant pathogenicity was conducted following the standards and guidelines presented by the ACMG/AMP [21]. The strength of the PVS1 criteria for null variants was adjusted according to published recommendations [22]. In addition, several alleles identified by cDNA Sanger sequencing but for which the causative gDNA change could not be identified have been scored using the decision trees for canonical splice sites and deletions as the consequence of the variant had been determined through RNA studies [23]. The strength of the PS4 criteria was adjusted to PS4_Supporting if the variant has been observed in at least two unrelated individuals with CHS. The presence of the pathognomonic enlarged intracytoplasmic granules on peripheral blood smear or clinical diagnosis by experts of CHS was required to include the PP4 evidence for phenotype specificity. ClinVar was not considered a reputable source for the PP5 and BP6 evidence according to published recommendations [24]. Variants with an allele frequency of >1% were not included in the study. Further details are presented in Supplementary Tables 2 and 3.

Reported and novel LYST variants were considered disease-causing variants and presented in Figure 2 and Tables 1–3 if the variant met one of the following conditions considering variant, variant configuration, and associated phenotype features: (1) a rare (PM2) null variant (nonsense, frameshift, canonical splice site, initiation codon, single or multiexon deletion) (PVS1) or a rare (PM2) missense or non-canonical splice site variant with multiple lines of computational evidence supporting a deleterious effect (PP3) confirmed to be in trans with a pathogenic allele (PM3); (2) a rare (PM2) null variant (nonsense, frameshift, canonical splice site, initiation codon, single or multiexon deletion) (PVS1) or a rare (PM2) missense or non-canonical splice site variant with multiple lines of computational evidence supporting a deleterious effect (PP3) associated with an individual confirmed to have the pathognomonic enlarged intracytoplasmic granules and/or clinically diagnosed with CHS by an expert (PP4); or (3) a rare (PM2) null variant (nonsense, frameshift, canonical splice site, initiation codon, single or multiexon deletion) (PVS1) or a rare (PM2) missense or non-canonical splice site variant with multiple lines of computational evidence supporting a deleterious effect (PP3) likely in trans with a pathogenic allele (i.e., the individual has another pathogenic allele, but the variants have not been phased) and associated with an individual presenting with at least two clinical features suggestive of CHS, including partial oculocutaneous albinism, immunological findings such as recurrent infections and hemophagocytic lymphohistiocytosis (HLH), and neurological manifestations such as cognitive impairment, peripheral neuropathy, ataxia, and parkinsonism. With respect to variant features, an exception was made for in-frame variants since there are limited computational tools to determine in silico pathogenicity. A variant was considered disease-causing and included in Figure 2 and Tables 1–3 if the variant was rare (PM2), resulted in a protein length change in a non-repeat region (PM4), and had a CADD phred score ≥20. These conditions are summarized in Supplementary Table 4. Bi-allelic or potentially bi-allelic LYST variants that require further evidence for pathogenicity are presented in Supplementary Table 5. Mono-allelic LYST variants that require further evidence for pathogenicity are presented in Supplementary Table 6. Variants reported without phenotypic information were not included in the study.

TABLE 1.

Disease-causing LYST variants in individuals with Chediak-Higashi syndrome classified as uncertain significance by ACMG/AMP guidelines.

Coding sequence change	Amino acid change	Coding effect	Reference(s)
c.2T>C	p.(Met1?)	Start-loss	[44]
c.172C>G	p.(Leu58Val)	Missense	[51]
c.772T>C	p.(Cys258Arg)	Missense	[43]
c.2570C>G	p.(Ser857Cys)	Missense	[40]
c.2897_2902delinsCAT	p.(Met966_Arg968delinsThrCys)	In-frame	[52]
c.3939G>Aa	p.?	?	[53]
c.4189T>G	p.(Phe1397Val)	Missense	[54]
c.4688G>A	p.(Arg1563His)	Missense	[25]
c.5784+5G>T	p.?	?	[40]
c.5996T>A	p.(Val1999Asp)	Missense	[25]
c.7136T>C	p.(Leu2379Pro)	Missense	[55]
c.?b	p.Gly2718_Gln2786del	In-frame	[23]
c.10424C>T	p.(Ser3475Phe)	Missense	[35]
c.10702-?_10800+?delc	p.?	?	[53]

^aThis exonic variant has multiple lines of computational evidence supporting a deleterious effect on splicing rather than leading to a synonymous variant, so the amino acid change was noted as p.? rather than p.(Gln1313=) and the coding effect was noted as? rather than synonymous.

^bcDNA Sanger sequencing demonstrated that this allele leads to complete skipping of exon 31 (c.8152_8358del) resulting in a frameshift variant, however the causative gDNA variant could not be identified.

^cThis variant was reported as a CNV of exon 48 as c.3568-?_3600+?del, but the coding sequence coordinates of exon 48 are further downstream at c.10702_10800. The reported coordinates reflect the amino acid position of exon 48.

TABLE 3.

Disease-causing LYST variants in individuals with Chediak-Higashi syndrome classified as pathogenic by ACMG/AMP guidelines.

Coding sequence change	Amino acid change	Coding effect	Reference(s)
c.118dup	p.(Ala40Glyfs*24)	Frameshift	[6]
c.148C>T	p.(Arg50*)	Nonsense	[61]
c.338_339delinsAGATCTTTGAGTGGA	p.(Ser113Lysfs*23)	Frameshift	[62]
c.433C>T	p.(Arg145*)	Nonsense	[44]
c.575dup	p.(Leu192Phefs*6)	Frameshift	[46]
c.925C>T	p.(Arg309*)	Nonsense	[63]
c.1044_1045insA	p.(Leu349Thrfs*2)	Frameshift	[64]
c.1168del	p.(Val390Cysfs*23)	Frameshift	[23]
c.1303G>T	p.(Glu435*)	Nonsense	[65]
c.1331del	p.(Leu444Tyrfs*14)	Frameshift	[53]
c.1467del	p.(Glu489Aspfs*78)	Frameshift	[5]
c.1507C>T	p.(Arg503*)	Nonsense	[32]
c.1540C>T	p.(Arg514*)	Nonsense	[42]
c.1897A>T	p.(Lys633*)	Nonsense	[66]
c.1902dup	p.(Ala635Serfs*4)	Frameshift	[67]
c.2016dup	p.(Arg673Glnfs*12)	Frameshift	[23]
c.2063G>A	p.(Trp688*)	Nonsense	[53]
c.2158C>T	p.(Gln720*)	Nonsense	This study
c.2311C>T	p.(Gln771*)	Nonsense	[68]
c.2374_2375del	p.(Asp792Phefs*6)	Frameshift	[40]
c.2413del	p.(Glu805Asnfs*2)	Frameshift	[41]
c.2413G>T	p.(Glu805*)	Nonsense	[23]
c.2454del	p.(Ala819Hisfs*5)	Frameshift	[25]
c.2623del	p.(Tyr875Metfs*24)	Frameshift	[69]
c.2655dup	p.(Arg886Thrfs*5)	Frameshift	[20]
c.2749_2750del	p.(Arg917Glyfs*5)	Frameshift	[68]
c.2962C>T	p.(Arg988*)	Nonsense	[23]
c.3017_3018del	p.(Lys1006Argfs*14)	Frameshift	This study
c.3073_3074del	p.(Asn1025Glnfs*6)	Frameshift	[61]
c.3085C>T	p.(Gln1029*)	Nonsense	[61]
c.3194del	p.(Leu1065*)	Nonsense	[57]
c.3202C>T	p.(Gln1068*)	Nonsense	[53]
c.3310C>T	p.(Arg1104*)	Nonsense	[5]
c.3433del	p.(His1145Ilefs*7)	Frameshift	This study
c.3434dup	p.(His1145Glnfs*9)	Frameshift	[25]
c.3622C>T	p.(Gln1208*)	Nonsense	[70]
c.3726del	p.(Lys1242Asnfs*25)	Frameshift	[71]
c.3938del	p.(Gln1313Argfs*4)	Frameshift	[20]
c.3944dup	p.(Val1316Cysfs*16)	Frameshift	[59]
c.4052C>G	p.(Ser1351*)	Nonsense	[25]
c.4157del	p.(Asp1386Valfs*4)	Frameshift	[34]
c.4159dup	p.(Thr1387Asnfs*35)	Frameshift	[57]
c.4274del	p.(Leu1425Tyrfs*2)	Frameshift	[25]
c.4322_4325del	p.(Glu1441Valfs*12)	Frameshift	[13]
c.4353G>A	p.(Trp1451*)	Nonsense	[13]
c.4498del	p.(Arg1500Glufs*5)	Frameshift	This study
c.4508C>G	p.(Ser1503*)	Nonsense	[40]
c.5006del	p.(Asn1669Ilefs*28)	Frameshift	[46]
c.5034_5038del	p.(Gly1679Thrfs*20)	Frameshift	[72]
c.5061T>A	p.(Tyr1687*)	Nonsense	[25]
c.5317del	p.(Arg1773Aspfs*13)	Frameshift	[39]
c.5491C>T	p.(Gln1831*)	Nonsense	This study
c.5506C>T	p.(Arg1836*)	Nonsense	[47]
c.5519del	p.(Ser1840Tyrfs*2)	Frameshift	[46]
c.5541_5542del	p.(Arg1848Serfs*3)	Frameshift	[59]
c.5601del	p.(Lys1867Asnfs*11)	Frameshift	[20]
c.6077dup	p.(Tyr2026*)	Nonsense	[73]
c.6078C>A	p.(Tyr2026*)	Nonsense	[25]
c.6159_6160del	p.(Met2053Ilefs*31)	Frameshift	[74]
c.6676C>T	p.(Arg2226*)	Nonsense	[62]
c.6694G>T	p.(Gly2232*)	Nonsense	[65]
c.6712C>T	p.(Arg2238*)	Nonsense	[71]
c.7060–1G>A	p.Leu2354Metfs*16	Frameshift	[39 40 75]
c.7060–1G>T	p.?	?	[40]
c.7207C>T	p.(Arg2403*)	Nonsense	[64]
c.7291del	p.(Leu2431Trpfs*3)	Frameshift	[57]
c.7555del	p.(Tyr2519Ilefs*10)	Frameshift	[39]
c.7604dup	p.(Asn2535Lysfs*2)	Frameshift	[76]
c.7645C>T	p.(Gln2549*)	Nonsense	[77]
c.7763del	p.(Lys2588Serfs*20)	Frameshift	This study
c.7786C>T	p.(Arg2596*)	Nonsense	[78]
c.8080C>T	p.(Gln2694*)	Nonsense	[71]
c.8127_8131delinsTTCTGATATGTA	p.(Val2710Serfs*4)	Frameshift	[66]
c.8221C>T	p.(Arg2741*)	Nonsense	[79]
c.?a	p.His2787Serfs*2	Frameshift	[23]
c.8380dup	p.(Tyr2794Leufs*8)	Frameshift	[80]
c.8583G>A	p.(Trp2861*)	Nonsense	[25]
c.8782C>T	p.(Gln2928*)	Nonsense	[81]
c.9044+1G>T	p.?	?	[65 82]
c.9106+1G>T	p.?	?	[78]
c.9107–20_9109del	p.Gly3036Glufs*16	Frameshift	[39 68]
c.?b	p.Ser3055Lysfs*4	Frameshift	[23]
c.9377_9389del	p.(Gly3126Valfs*2)	Frameshift	[83]
c.9453del	p.(Lys3151Asnfs*34)	Frameshift	[84]
c.9560+1G>C	p.?	?	[82]
c.9590del	p.(Tyr3197Leufs*62)	Frameshift	[67]
c.9628–9_9628dup	p.(Tyr3210Phefs*17)	Frameshift	[39]
c.9784+2T>C	p.?	?	[85]
c.9893del	p.(Phe3298Serfs*7)	Frameshift	[42]
c.9926–4_9936del	p.Phe3310Serfs*4	Frameshift	[23]
c.9930del	p.(Phe3310Leufs*36)	Frameshift	[40]
c.10100del	p.(Lys3367Argfs*34)	Frameshift	[55]
c.10120_10127del	p.(Ala3374Cysfs*16)	Frameshift	[53]
c.10127A>G	p.(Asn3376Ser)	Missense	[41]
c.10395del	p.(Gly3466Alafs*2)	Frameshift	[25]
c.10445_10446insCA	p.(Val3483Argfs*33)	Frameshift	[64]
c.10551_10552del	p.(Tyr3517*)	Nonsense	[40]
c.10776C>G	p.(Tyr3592*)	Nonsense	[57]
c.10883dup	p.(Tyr3628*)	Nonsense	[86]
c.11002G>T	p.(Glu3668*)	Nonsense	[70 87]
c.?c	p.Ala3680Glyfs*8	Frameshift	[23]
c.11183del	p.(Asn3728Metfs*4)	Frameshift	[84]
c.11216dup	p.(Leu3739Phefs*14)	Frameshift	[88]
c.11308_11311del	p.(Asn3770Valfs*5)	Frameshift	This study

^acDNA Sanger sequencing demonstrated that this allele leads to partial skipping of exon 32 (first 31 bp) (c.8359_8389del) resulting in a frameshift variant, however the causative gDNA change could not be identified.

^bcDNA Sanger sequencing demonstrated that this allele leads to complete skipping of exons 39 and 40 (c.9163_9560del) resulting in a frameshift variant, but the causative gDNA change could not be identified. Exons 39 and 40 could not be amplified from the gDNA from the patient CHD25 reported in Kuptanon et al. (2023), suggesting a homozygous genomic deletion in this region.

^ccDNA Sanger sequencing demonstrated that this variant leads to complete skipping of exon 51 (c.11039_11195del) resulting in a frameshift variant, however the causative gDNA change could not be identified.

RESULTS

Clinical findings of seven individuals with Chediak-Higashi syndrome

The clinical characteristics of all the patients presented in the study are summarized in Table 4. These patients included seven individuals enrolled in our natural history study and one previously reported individual (CHD10), who was noted to have clinical manifestations of CHS [25]. All seven individuals presented with the giant intracytoplasmic granules in leukocytes, a pathognomonic feature that establishes a clinical diagnosis of CHS [1]. Partial oculocutaneous albinism was present in all seven individuals. Five individuals manifested with a bleeding diathesis and/or absent platelet dense granules, which play a critical role in platelet aggregation. Six individuals had a history of recurrent infections or HLH; among them, four individuals underwent hematopoietic stem cell transplantation (HSCT) as a preventative measure against future HLH episodes. Neurological manifestations were observed in four individuals, including absent deep tendon reflexes (DTRs), ataxia, lower extremity weakness, and peripheral neuropathy, while neurological examinations were normal in three individuals at ages 5 (CHD7), 7 (CHD27), and 11 (CHD29) years. The age at diagnosis ranged from neonatal to 4 years of age, highlighting the early onset of symptoms and the need for prompt clinical recognition.

TABLE 4.

Summary of the clinical characteristics of the individuals with CHS in whom novel LYST variants were identified.

Patient ID	CHD1	CHD7	CHD10b	CHD14	CHD27	CHD28	CHD29	CHD30
LYST variants	c.148C>T (p.Arg50*)a; c.11267+1G>A (p.?)	c.2158C>T (p.Gln720*); c.3017_3018del (p.Lys1006Argfs*14)	c.10022A>G (p.His3341Arg); c.10022A>G (p.His3341Arg)	c.4498del (p.Arg1500Glufs*5); c.7763del (p.Lys2588Serfs*19)	c.3433del (p.His1145Ilefs*7); c.5491C>T (p.Gln1831*)	c.3433del (p.His1145Ilefs*7); c.7951G>T (p.Val2651Phe)a	c.10003C>T (p.Arg3334Cys); c.11308_11311del (p.Asn3770Valfs*5)	c.9457G>C (p.Ala3153Pro); c.9457G>C (p.Ala3153Pro)
Age at diagnosis	2 years	Early childhood (2–5 years)	N/A	Toddler age (13 months to 2 years)	Toddler age (13 months to 2 years)	Infancy (28 days to 12 months)	Early childhood (2–5 years)	Neonatal (birth to 27 days)
Alive	No (age at death 40’s [years])	No (age at death at early childhood [2–5 years])	N/A	Yes (age at last follow up (middle childhood [6–11 years])	Yes (age at last follow up 20’s [years])	Yes (age at last follow up middle childhood [6–11 years])	NA (age at last follow up middle childhood [6–11 years])	Yes (age at last follow up 30’s [years])
Immunological history	Skin infections and pneumonia in infancy	No unusual infections	N/A	Recurrent otitis media	Infected urachal cyst	Severe, nonspecific infection in infancy	No unusual infections	No unusual infections
Inclusions in WBC	Present	Present	N/A	Present	Present	Present	Present	Present
Bleeding history	Easy bruising	Easy bruising	N/A	Not observed	Not observed	Not observed	Not observed	Easy bruising
Platelet dense granules	Absent	Not tested	N/A	Absent	Absent	Not tested prior to transplant	Not tested	Not tested prior to transplant
Partial albinism	Present; gray sheen to hair	Present	Nystagmus	Present; gray sheen to hair	Present; gray sheen to hair	Present; blonde at birth	Present	Present
HLH	No	Yes, age 5	N/A	No	No	No	No	Yes, age 7
Neurologic history	Absent DTRs, ataxia, weakness, shuffling gait at age 18	No abnormalities at age 5	Peripheral neuropathy	Anxiety, absent DTRs at age 11	Absent DTRs, ataxia, LE weakness	ADHD; neurologic exam normal at age 7	Neurologic exam normal at age 11	Absent DTRs, LE weakness, peripheral neuropathy, cerebellar signs, basal ganglia signs
HSCT	No	No	N/A	Yes, age 20 months; second transplant at 2 years	Yes, age 3 years	Yes, age 9 months	No	Yes, age 10 years

Abbreviations: ADHD, attention deficit hyperactivity disorder; DTR, deep tendon reflex; HLH, hemophagocytic lymphohistiocytosis; HSCT, hematopoietic stem cell transplantation; LE, lower extremity; N/A, not available; WBC, white blood cell.

^aPreviously reported LYST variant.

^bCHD10 refers to the individual from which the GM02431 lymphoblastoid cell line (Coriell Institute of Medical Research) was derived and previously reported in Karim et al. (2002).

Molecular genetic findings of seven individuals with Chediak-Higashi syndrome

To establish a molecular diagnosis, we performed targeted Sanger sequencing of LYST. We identified 11 novel disease-causing variants in LYST, expanding the mutational spectrum of CHS (Figure 2, Table 4, and Supplementary Table 7). These novel variants were distributed throughout the gene and no recurring variants were observed (Figure 2A). Notably, all novel missense variants were located in the BEACH domain of LYST (Figure 2A and B). With thorough clinical phenotyping and a confirmed clinical diagnosis of CHS, missense variants accounted for a larger proportion of the novel variants identified (27% of novel variants versus 16% of previously reported variants) (Figure 2C), underscoring the importance of detailed clinical examination and assessment of the enlarged intracytoplasmic granules pathognomonic for CHS.

The spectrum of LYST variants and a review of the literature

A total of 1860 unique variants were obtained from the literature, ClinVar, the Leiden Open Variation Database, and the Human Gene Mutation Database. Among these variants, 1651 lacked phenotypic information and were therefore excluded from the study. Of the 209 remaining variants associated with specific phenotypic information, 136 variants were found to be disease-causing variants while 73 variants required further evidence for pathogenicity. In this study, we report 11 novel disease-causing variants, resulting in a total of 147 unique disease-causing variants (Figure 2 and Tables 1–3). Additional information on the variants, including location, variant type, CADD phred score, gnomAD frequency, and ACMG/AMP classification, are presented in Supplementary Table 7.

The distribution of the 147 disease-causing LYST variants (Figure 2 and Tables 1–3) revealed no mutational hotspots as the variants were distributed throughout the gene and most variants were private (Figure 2A). Many of the disease-causing missense and in-frame variants are in known functional domains, including the WD repeat and BEACH domains (Figure 2A and B). The majority of variants introduce a premature termination codon, including frameshift (41%) and nonsense (30%) variants. Missense (16%), splice site variants or small genomic deletions for which the coding effect is unknown (9%), in-frame (3%), and start-loss (1%) variants accounted for the remaining variant types (Figure 2C). While the ClinVar and LOVD databases have reported many more variants in LYST, most of these variants have been reported in the absence of phenotypic information, which is crucial for variant interpretation, and hence were not included for interpretation.

From previous publications, there are bi-allelic LYST variants that require further evidence for pathogenicity (Supplementary Table 5). These variants mostly consist of missense or non-canonical splice site variants. Few variants were considered rare (PM2) with multiple lines of computational evidence supporting a deleterious effect (PP3), while some variants even had multiple lines of computational evidence supporting a benign effect (BP4). Regarding variant configuration features, some variants were homozygous variants or compound heterozygous variants confirmed to be in trans, but others were compound heterozygous variants that were not confirmed to be in trans or potentially compound heterozygous variants where the second allele was not identified. With respect to phenotype features, none of the variants were associated with an individual presenting with enlarged intracytoplasmic granules on peripheral blood smear (PP4), although some variants were associated with individuals presenting with one or two clinical features of CHS. Some variants were associated with individuals presenting with features not typically associated with CHS. In all cases, the variants did not satisfy any of the conditions summarized in Supplementary Table 4 required to be considered a disease-causing variant.

Mono-allelic LYST variants associated with other disorders (Supplementary Table 6) and common LYST variants associated with pigmentary and immunological traits have also been reported and are further described in the Supplementary Information.

DISCUSSION

CHS is a rare autosomal recessive disorder characterized by partial oculocutaneous albinism, a bleeding diathesis, immunological dysfunction, and neurological impairment [1 26 27]. The pigmentary manifestations can include pigment dilution of the skin, eyes, and hair. Notably, the hair may be silvery or have a metallic sheen, a quality often observed in animal models of CHS (further described in the Supplementary Information). Hyperpigmentation of the irides [28] and skin [29–32], particularly in sun-exposed areas, has been previously reported in some populations. The bleeding diathesis is typically mild and manifests as easy bruising and mucosal bleeding. The immunological dysfunction manifests as frequent and recurrent infections, commonly affecting the skin and upper respiratory tract. Most patients develop neurological complications from the second decade of life that can include cognitive impairment, sensory and motor neuropathies, ataxia, spastic paraplegia, and parkinsonism [32–35]. Recent associations of amyotrophic lateral sclerosis (ALS) [36] and HLH [37] with bi-allelic missense variants in LYST without the presence of the pathognomonic enlarged intracytoplasmic granules require further investigations to determine whether these findings represent expansion of the clinical spectrum of CHS.

Currently, individuals diagnosed with CHS are classified as having a classical (typical) or atypical clinical presentation [1 26]. Classical CHS patients typically have an earlier onset of disease and experience severe hematological and immunological complications, including an increased risk for HLH, which is the most common cause of mortality in individuals with CHS. On the other hand, patients with atypical CHS are likely to be identified later in life because of a milder clinical presentation with the near absence of severe immunological symptoms, including a lower risk for developing HLH. While early bone marrow transplantation complicates the classification of a patient as having a classical or atypical clinical presentation, it effectively eliminates the risk of developing HLH. However, nearly all patients develop the long-term neurologic sequelae of the disease regardless of the clinical presentation or bone marrow transplant status [33 38].

Establishing genotype-phenotype correlation for LYST and CHS has been challenging due to the rarity of the disorder and the predominantly private nature of the variants [39 40]. Nonetheless, several studies have attempted to correlate the clinical phenotypes with molecular genotypes [25 41 42]. These studies have generally concluded that null variants are associated with the more severe classical presentation of CHS, while missense or in-frame variants are associated with the milder atypical form of the disease. In general, our review of the literature of reported CHS cases suggests that harboring at least one missense or in-frame variant leads to milder disease, while harboring two nonsense or frameshift variants leads to more severe disease. However, there are several notable exceptions. One individual with classical CHS, developing HLH at the age of 3 years, had a homozygous missense variant (c.11362G>A, p.Gly3725Arg) located in the WD6 domain [43]. We also follow a patient with classical CHS who developed HLH at the age of 6 years and had a homozygous missense variant (c.9457G>C, p.Ala3153Pro) located in the BEACH domain; her older sibling also presented with classical CHS and succumbed from HLH. In contrast, a remarkable individual with relatively mild disease was diagnosed with CHS at the age of 67 years despite harboring a start-loss (c.2T>C, p.Met?) and nonsense (c.433C>T, p.Arg145*) variant [44]. These illustrative cases suggest that some amino acid residues are critical for LYST structure, function, and/or stability and that there may be genetic or environmental interactions that attenuate the phenotypic effects of start-loss, nonsense, or frameshift variants.

Studies of siblings can also provide insight into genotype-phenotype correlations. While siblings with CHS often present with a similar clinical presentation [34 45 46], significant clinical variability has also been observed. For example, two siblings with a homozygous nonsense variant (c.5506C>T, p.Arg1836*) presented with divergent phenotypes whereby one sibling presented with fair skin and hair and died from HLH at 5 months of age while the other sibling presented with dark skin and hair and developed HLH at 4 years of age [47]. In another example, two siblings with compound heterozygous frameshift variants presented with differing phenotypes; one sibling died from HLH at 18 months of age while the older sibling was still immunologically asymptomatic at 10 years of age [39]. Even more striking is the example of another family in which the proband developed HLH at 2 years of age, but a sister and cousin died before the age of 2 years and three cousins without bone marrow transplants had not developed HLH at 11, 16, and 40 years of age [39]. These clinical findings suggest that there are genetic modifiers of LYST; indeed, there is evidence of genetic modifiers of LYST homologs from animal studies [48].

The function of LYST is largely unknown but is believed that loss-of-function variants in LYST lead to CHS through misfolding, impaired function, and/or reduced stability of LYST. Most disease-causing variants are nonsense or frameshift variants that result in reduced levels LYST. Despite the size of LYST, which poses a challenge for experimental investigations, several studies have illuminated important characteristics of LYST. Structural and biochemical studies have demonstrated that the beige and CHS (BEACH) and pleckstrin homology (PH) domains associate and interact with one another, and functional studies have provided evidence for the requirement of both domains for activity, altogether suggesting that these domains function as a single unit [49]. Variants in either of these domains may disrupt mutual interactions and decrease the functionality of LYST. Yeast two-hybrid studies have identified several LYST-interacting proteins, including proteins involved in vesicular transport and signal transduction [50]; variants in the regions of LYST involved in these interactions would likely be detrimental for certain aspects of LYST protein function. Further functional studies have demonstrated that variants located in different domains of LYST differentially affect NK cell functions, including lytic granule size, polarization, and exocytosis [15].

While the influx of next-generation sequencing data has led to the identification of many LYST variants, variant interpretation remains challenging due to limited evidence for pathogenicity. Here, we classified both novel and previously reported LYST variants associated with Chediak-Higashi syndrome using the ACMG/AMP variant interpretation guidelines (Supplementary Tables 2 and 3). The assessment of the pathognomonic enlarged intracytoplasmic granules on peripheral blood smear is crucial for clinical diagnosis, while sequencing of parents or additional family members is essential for molecular diagnosis to phase variants. Additionally, the evaluation of intronic variants by cDNA Sanger sequencing or RNA-seq can aid in variant interpretation and classification [23]. While clinical acumen alone has played a role in the classification of variant pathogenicity, the development of an assay or other measures to assess the pathogenicity of LYST variants would improve the interpretation of variants overall.

CONCLUSION

We present a comprehensive review of the spectrum of LYST variants in CHS and expand the allele spectrum by identifying 11 novel variants, resulting in a total of 147 variants in LYST associated with CHS. Our analysis revealed a genotype-phenotype correlation where individuals carrying at least one missense or in-frame variant tend to have a milder disease, while those with two protein-truncating experience more severe manifestations. By classifying both the novel and reported LYST variants using the ACMG/AMP variant interpretation guidelines, we provide valuable insights into their pathogenicity and clinical significance. This classification framework enhances our understanding of the genetic basis of CHS and facilitates accurate molecular diagnosis for affected individuals.

The identification and classification of these LYST mutations contributes to an earlier diagnosis and improved care for individuals with CHS. The interpretation of variants of unknown significance identified through next-generation sequencing remains a challenge, necessitating the development of specialized assays and tools to better assess their pathogenicity. Finally, the absence of a treatment for CHS underscores the necessity to further understand the natural history of the disease, the function of LYST, and the biological mechanisms underlying pathogenesis of CHS. These insights are essential for the development of targeted personalized therapies and the design of clinical trials aimed at improving the prognosis and quality of life for individuals affected by CHS.

TABLE 2.

Disease-causing LYST variants in individuals with Chediak-Higashi syndrome classified as likely pathogenic by ACMG/AMP guidelines.

Coding sequence change	Amino acid change	Coding effect	Reference(s)
c.1463C>G	p.(Ser488*)	Nonsense	[56]
c.2015dup	p.(Tyr672*)	Nonsense	[57]
c.4361C>A	p.(Ala1454Asp)	Missense	[25]
c.4862+1G>Aa	p.Val1564_Arg1621del	In-frame	[56 58]
c.5715del	p.(Asn1905Lysfs*3)	Frameshift	[57]
c.5784+1G>T	p.?	?	[57]
c.7951G>T	p.(Val2651Phe)	Missense	[33]
c.7982C>G	p.(Ser2661*)	Nonsense	[59]
c.8281A>T	p.(Arg2761*)	Nonsense	[59]
c.8428G>A	p.(Glu2810Lys)	Missense	[25]
c.8770C>T	p.(Gln2924*)	Nonsense	[57]
c.8802–2A>G	p.?	?	[57]
c.9457G>C	p.(Ala3153Pro)	Missense	This study
c.9476A>G	p.(Asp3159Gly)	Missense	[34]
c.9488A>G	p.(Tyr3163Cys)	Missense	[35]
c.9706C>T	p.(His3236Tyr)	Missense	[56]
c.9827_9832del	p.(Asn3276_Thr3277del)	In-frame	[45]
c.9844_9845del	p.(Ser3282Ilefs*3)	Frameshift	[57]
c.9925G>A	p.(Gly3309Ser)	Missense	[32]
c.9925+1G>Ab	p.Asp3262_Glu3308del	In-frame	[23]
c.10003C>T	p.(Arg3335Cys)	Missense	This study
c.10022A>G	p.(His3341Arg)	Missense	This study
c.10095G>C	p.(Lys3365Asn)	Missense	[55]
c.10222G>A	p.(Gly3408Arg)	Missense	[33]
c.10374+1G>T	p.?	?	[34]
c.10747G>c	p.(Gly3583Arg)	Missense	[60]
c.11173G>A	p.(Gly3725Arg)	Missense	[43]
c.11196–1G>A	p.?	?	[40]
c.11267+1G>A	p.?	?	This study

^aThis variant was first reported by Lasseaux et al. (2018). Serra-Vinardell et al. (2020) demonstrated that this canonical splice site variant leads to complete skipping of exon 14 (c.4689_4862del) resulting in an in-frame deletion by cDNA Sanger sequencing.

^bcDNA Sanger sequencing demonstrated that this variant leads to complete skipping of exon 43 (c.9785_9925del) resulting in an in-frame variant.

^cThis variant was incompletely reported in Antunes et al. (2013). The reported predicted amino acid change was p.(Gly3583Arg), so the variant could be c.10747G>A or c.10747G>C.

KEY MESSAGES.

What is already known on this topic

• Bi-allelic loss-of-function variants in the LYST gene cause Chediak-Higashi syndrome (CHS), but variant interpretation is challenging due to the large size of the gene, the unknown function of LYST, and the lack of robust cellular assays to assess variant pathogenicity.

What this study adds

• Identification of 11 novel LYST variants, a comprehensive review of LYST variants in the literature, and ACMG/AMP classification of all novel and reported LYST variants associated with CHS.

How this study might affect research, practice or policy

• Our identification of novel LYST variants and ACMG/AMP classification of reported variants would likely lead to an earlier molecular diagnosis of individuals with CHS, and earlier considerations for treatment to prevent the life-threatening hyperinflammatory condition hemophagocytic lymphohistiocytosis (HLH).

DATA AVAILABILITY STATEMENT

The novel variants described in this study have been submitted to the ClinVar database (https://www.ncbi.nlm.nih.gov/clinvar/).