A mutation in *CCDC91*, Homo sapiens coiled-coil domain containing 91 protein, cause autosomal-dominant acrokeratoelastoidosis

Acrokeratoelastoidosis (AKE) (OMIM:101850) is an autosomal dominant hereditary skin disorder characterized by bilateral multiple hyperkeratotic papules primarily distributed on the lateral portion of the palmoplantar regions, first described by Oswaldo Costa in 1953. Lesions typically appear in childhood or early adulthood and gradually increase in size and number, becoming grouped or plaque-like. The primary etiology of AKE has remained elusive; both sporadic and familial autosomal dominant cases have been reported. Proposed risk factors for sporadic cases include light exposure and trauma. Earlier genetic mapping in a large pedigree suggested possible linkage to chromosome 2, and AKE was previously classified as type III punctate palmoplantar keratoderma (PPKP3). However, genes implicated in other palmoplantar keratodermas (e.g., AAGAB in PPKP1) have not been found to be involved in AKE, and the pathogenic gene(s) remained unidentified. In this study, the authors report identification of a splicing mutation in CCDC91 (OMIM:617366) in a Chinese family with AKE, and functional analyses suggesting that CCDC91 influences vesicle transport in the Golgi apparatus, thereby affecting elastin synthesis and secretion.

Subjects and clinical characteristics: A three-generation Chinese AKE family (22 living individuals) was recruited. Diagnoses were made by dermatologists based on clinical and histological features. The proband (III-7) had an 18-year history of multiple keratotic papules on hands; lesions became swollen and whitish after 5 minutes of immersion. Skin biopsy showed marked hyperkeratosis, thickened granular layer, acanthosis, and sparse superficial dermal perivascular inflammation; VVG staining revealed thick, curved, coarse, fragmented, and reduced elastic fibers in the reticular dermis. DNA isolation: Peripheral blood genomic DNA was extracted using QIAamp DNA Blood Mini Kit; quality assessed by NanoDrop2000. Genotyping and linkage analysis: Genome-wide SNP genotyping using Illumina Human 660W-Quad_v1 BeadChips (655,214 probes). Multipoint parametric linkage analysis was performed with MERLIN v1.1.2; 6,886 markers analyzed. Whole-exome sequencing (WES): Libraries prepared from 1 µg DNA (proband and three relatives), captured with Roche NimbleGen SeqCap EZ Human Exome v3.0 (64 Mb), sequenced on Illumina HiSeq 2000. Reads aligned to hg19 using BWA; variants filtered against dbSNP138, 1000 Genomes, and HapMap. Sanger sequencing validation: Candidate variants from NGS were validated by Sanger sequencing (primers in Table S1) using 3730xl sequencer; analyses with DNASTAR. RT-PCR: PBMCs isolated by Ficoll gradient; RNA extracted with TRIzol; cDNA synthesized (PPrimeScript II kit). Primers spanning the splice site (Table S1) used for PCR and sequencing to assess splicing. Cell culture: Human skin fibroblasts (HSF) and HEK293T cells cultured in high-glucose DMEM with 10% FBS and penicillin-streptomycin at 37°C, 5% CO2. CCDC91 knockdown in HSF: Four shRNAs targeting CCDC91 (NM_018318.5) were designed, cloned into lentiviral vectors, packaged in 293T cells, and used to infect HSF with puromycin selection (3 µg/ml). Knockdown efficiency assessed by qPCR (SYBR Premix Ex Taq II) and Western blot; GAPDH as control. CRISPR/Cas9 knockout in HEK293T: Two gRNAs targeting exon 11 (validated via CRISPR design tool) cloned into PX459 (pSpCas9-2A-Puro). HEK293T transfected with Lipofectamine 2000, selected with puromycin (3 µg/ml). CCDC91 cDNA sequenced to verify exon 11 deletion. Immunofluorescence: Cells fixed in 4% formaldehyde, blocked with 1% BSA. Primary antibodies: CCDC91 and GM130 (Golgi marker). Appropriate Alexa Fluor secondary antibodies used; nuclei stained with DAPI; imaged on Zeiss LSM880. In si-CCDC91 HSF, tropoelastin (TE), human aortic alpha elastin, and fibrillin-1 were immunostained to assess elastic fiber components. Transmission electron microscopy (TEM): Cells fixed in glutaraldehyde, post-fixed in OsO4, dehydrated, embedded in Spurr resin, sectioned (EM UC6), stained with uranyl acetate and lead citrate, imaged on JEM-1230 at 120 kV. ELISA: Culture media from fibroblasts after 10 days were collected; insoluble elastin quantified with Human Elastin SimpleStep ELISA kit (ab239433). Statistical and bioinformatics analysis: Data expressed as mean ± SD; one-way ANOVA used with p<0.05 significance. CCDC91 3D structure modeled using FR-t5-M and I-TASSER; side chains optimized with CIS-RR; models visualized with PyMOL. PROVEAN used to evaluate deletion (L309_Q367del).

• Linkage analysis in a three-generation AKE family mapped a susceptibility region on chromosome 12 between rs7296765 and rs10784618 (46.93–56.50 cM), with a maximal LOD score of 3.55.
• WES identified a heterozygous splice-site mutation in CCDC91 (NM_018318: exon 11: c.1101+1G>A) in affected individuals; absent in unaffected relatives.
• RT-PCR from patient PBMCs showed alternative splicing: controls had a single 606 bp cDNA fragment, while patients had 606 bp and 429 bp fragments. Sanger sequencing of RT-PCR products confirmed exon 11 skipping due to c.1101+1G>A, causing an in-frame deletion of 59 amino acids (L309–Q367del) in CCDC91.
• Structural and conservation analyses: The deleted region is highly conserved among mammals. Modeling indicated loss of a stable motif (two alpha-helices opposing antiparallel beta-sheets); mutant protein exhibited a narrower, more concave structure relative to wild-type. PROVEAN predicted the deletion as deleterious (score −115.041).
• Golgi apparatus abnormalities upon CCDC91 loss: In si-CCDC91 HSF, immunofluorescence showed CCDC91 co-localization with GM130 and a dispersed, vesicular Golgi morphology. TEM revealed swollen, fragmented Golgi cisternae and increased lysosomes and autophagosomes compared to controls. CRISPR-mediated exon 11 knockout in HEK293T similarly produced significant Golgi swelling.
• Effects on elastic fiber components: In si-CCDC91 HSF, globular tropoelastin deposits with increased intracellular immunoreactivity were observed, and extracellular insoluble elastin staining was reduced. Fibrillin-1 microfibril assembly appeared unchanged.
• Quantitative decrease in elastin: ELISA of culture media demonstrated significantly decreased insoluble elastin in si-CCDC91 fibroblasts versus controls (p<0.05).
• According to the abstract, lysyl oxidase activity showed no significant changes, supporting a defect primarily in elastin transport rather than crosslinking.
• Overall, findings implicate CCDC91 in Golgi-mediated vesicular transport of tropoelastin; its disruption leads to Golgi swelling, intracellular TE accumulation, and reduced extracellular elastin deposition, explaining elastic fiber abnormalities in AKE.

This study addresses the long-standing question of the genetic basis of acrokeratoelastoidosis (AKE) by mapping and identifying a pathogenic splice-site mutation in CCDC91 in a multigenerational family. The mutation (c.1101+1G>A) leads to exon 11 skipping and deletion of 59 amino acids in a highly conserved region of CCDC91, predicted to perturb its structure and function. Functional assays in fibroblasts and HEK293T cells demonstrate that reduced or altered CCDC91 disrupts Golgi morphology, increases lysosomal and autophagosomal compartments, and causes tropoelastin accumulation within the Golgi, with concomitant reduction in extracellular insoluble elastin. Fibrillin-1 assembly and lysyl oxidase activity are not significantly affected, indicating a selective disturbance in elastin trafficking rather than in microfibril scaffold formation or crosslinking. These data collectively support a mechanistic model in which CCDC91, a Golgi-associated coiled-coil protein reported to interact with GGA adaptors, facilitates vesicular transport from the Golgi to downstream compartments. Its dysfunction compromises tropoelastin export and extracellular elastin deposition, leading to fragmented and diminished dermal elastic fibers characteristic of AKE. The clinical features observed (bilateral, symmetric keratotic papules; whitening after water exposure) are concordant with elastin abnormalities rather than with CFTR-related aquagenic palmoplantar keratoderma, consistent with the absence of CFTR variants. The lack of CCDC91 variants in two sporadic AKE cases suggests genetic heterogeneity and potentially distinct mechanisms between familial and sporadic forms. The findings broaden AKE pathogenesis to include Golgi vesicular transport defects and identify CCDC91 as a key regulator of elastic fiber homeostasis in skin and potentially other tissues.

The study identifies a heterozygous splice-site mutation in CCDC91 (c.1101+1G>A) as a causative alteration for autosomal-dominant AKE in a Chinese family. The mutation induces exon 11 skipping, leading to a deleterious in-frame deletion (L309–Q367del). Functional analyses demonstrate that CCDC91 loss disrupts Golgi structure and tropoelastin trafficking, resulting in intracellular TE accumulation and reduced extracellular insoluble elastin without affecting fibrillin-1 microfibrils or lysyl oxidase activity. These findings establish CCDC91 as a critical regulator of elastin transport and expand the pathogenic mechanisms underlying AKE. Future research should include broader genetic screening in additional familial and sporadic AKE cohorts, in vivo modeling to confirm skin-specific mechanisms, and detailed dissection of CCDC91 interactions with Golgi adaptor proteins (e.g., GGA1-3) to define the transport pathway for tropoelastin.

• The genetic discovery is based on a single multigenerational family; generalizability requires validation in independent cohorts.
• Functional studies were performed in vitro using HSF and HEK293T cells; in vivo confirmation in patient-derived tissues or animal models is needed.
• Two sporadic AKE cases did not harbor CCDC91 variants, suggesting genetic heterogeneity and indicating that CCDC91 may not account for all AKE forms.
• While fibrillin-1 and elastin were examined, the study did not comprehensively assess all components and regulators of elastic fiber assembly beyond those reported, leaving other pathways potentially involved.