Molecular Analysis of the ABCA4 Gene Mutations in Patients with Stargardt Disease Using Human Hair Follicles

ABCA4 gene mutations are the cause of a spectrum of ABCA4 retinopathies, and the most common juvenile macular degeneration is called Stargardt disease. ABCA4 has previously been observed almost exclusively in the retina. Therefore, studying the functional consequences of ABCA4 variants has required advanced molecular analysis techniques. The aim of the present study was to evaluate whether human hair follicles may be used for molecular analysis of the ABCA4 gene splice-site variants in patients with ABCA4 retinopathies. We assessed ABCA4 expression in hair follicles and skin at mRNA and protein levels by means of real-time PCR and Western blot analyses, respectively. We performed cDNA sequencing to reveal the presence of full-length ABCA4 transcripts and analyzed ABCA4 transcripts from three patients with Stargardt disease carrying different splice-site ABCA4 variants: c.5312+1G>A, c.5312+2T>G and c.5836-3C>A. cDNA analysis revealed that c.5312+1G>A, c.5312+2T>G variants led to the skipping of exon 37, and the c.5836-3C>A variant resulted in the insertion of 30 nucleotides into the transcript. Our results strongly argue for the use of hair follicles as a model for the molecular analysis of the pathogenicity of ABCA4 variants in patients with ABCA4 retinopathies.


Introduction
The ABCA4 gene (OMIM 601691; GenBank:NG_009073.1) encodes an ATP-binding cassette transporter which translocates retinoid intermediates of the visual cycle across the disc membranes of photoreceptor cells [1]. Mutations of the ABCA4 gene are the cause of more than 95% of cases of Stargardt disease (STGD)-the most common form of inherited juvenile macular degeneration-as well as other monogenic retinal diseases, the so-called ABCA4 retinopathies [2].

ABCA4 Gene is Expressed in Human Hair Follicles and Skin
The ABCA4 gene was found to be expressed in human eyebrow hair follicles (n = 8) and skin explants (n = 8) ( Figure 1A). Expression of the ABCA4 gene in hair follicles was over fifty-fold higher than in the skin (p-value < 0.0001 by Mann-Whitney U test). Some, but relatively very few, ABCA4 expressions were also detected in cultured human primary keratinocytes (n = 8), dermal fibroblasts (n = 8) and melanocytes (n = 8) ( Figure 1B). There were no statistically significant differences in ABCA4 mRNA expression between keratinocytes, fibroblasts and melanocytes (p-value = 0.1562 by one-way ANOVA). mRNA expression between keratinocytes, fibroblasts and melanocytes (p-value = 0.1562 by one-way ANOVA). Figure 1. Expression of the ABCA4 mRNA in hair follicles, skin lysates, as well as keratinocytes, dermal fibroblasts and melanocytes as measured by qRT-PCR for TaqMan probe Hs00979586_m1 (ABCA4 exons [35][36]. (A) Expression of the ABCA4 gene in the isolated hair follicles was significantly higher than in total skin lysates. (B) Relatively very low expression of the ABCA4 gene was detected in fibroblasts, keratinocytes and melanocytes. * * * * p < 0.0001 as calculated by Mann-Whitney U test.
Additional evaluation of the particular ABCA4 exons in hair follicles and total skin showed that expression of exons 3-4 was lower as compared to the other exons (Table 1).

Hair Follicles and Skin Cells Express the Full-Length ABCA4 Transcript and Protein
The presence of all ABCA4 exons was confirmed in the analyzed hair follicle samples (n = 4), cultured keratinocytes (n = 4), fibroblasts (n = 4) and melanocytes (n = 4) by means of Sanger sequencing. Western blot analysis of hair follicles, skin protein extracts and cultured skin cells confirmed the presence of the main ABCA4 immunoreactive band, corresponding to a molecular mass of 250 kDa. In one skin sample, expression of the ABCA4 gene was not detected and this sample was used as a negative control. The result of the Western blot analysis of representative hair follicles and skin samples is presented in Figure 2. Expression of the ABCA4 mRNA in hair follicles, skin lysates, as well as keratinocytes, dermal fibroblasts and melanocytes as measured by qRT-PCR for TaqMan probe Hs00979586_m1 (ABCA4 exons [35][36]. (A) Expression of the ABCA4 gene in the isolated hair follicles was significantly higher than in total skin lysates. (B) Relatively very low expression of the ABCA4 gene was detected in fibroblasts, keratinocytes and melanocytes. * * * * p < 0.0001 as calculated by Mann-Whitney U test.
Additional evaluation of the particular ABCA4 exons in hair follicles and total skin showed that expression of exons 3-4 was lower as compared to the other exons (Table 1).

Hair Follicles and Skin Cells Express the Full-Length ABCA4 Transcript and Protein
The presence of all ABCA4 exons was confirmed in the analyzed hair follicle samples (n = 4), cultured keratinocytes (n = 4), fibroblasts (n = 4) and melanocytes (n = 4) by means of Sanger sequencing. Western blot analysis of hair follicles, skin protein extracts and cultured skin cells confirmed the presence of the main ABCA4 immunoreactive band, corresponding to a molecular mass of 250 kDa. In one skin sample, expression of the ABCA4 gene was not detected and this sample was used as a negative control. The result of the Western blot analysis of representative hair follicles and skin samples is presented in Figure 2.

Analysis of the ABCA4 Splice-Site Variants in Hair Follicles
Considering that the expression of the ABCA4 gene in the skin samples was significantly lower than in the hair follicles, we used the latter for further experiments aimed at assessing the sequence of ABCA4 transcripts in patients' samples.
In order to evaluate the usefulness of eyebrow hair follicles as a tool for the molecular analysis of ABCA4 variants, we investigated the consequences of the three ABCA4 splice-site variants c.5312+1G>A, c.5312+2T>G and c.5836-3C>A, previously detected in patients with Stargardt dystrophy [7]. In those patients, ABCA4 gene mutations were found in a compound heterozygous state with missense (NM_000350. 2 [7]. Direct Sanger sequencing revealed that variants c.5312+1G>A and c.5312+2T>G result in the skipping of exon 37 ( Figure 3A,B), whereas mutation c.5836-3C>A leads to the insertion of thirty nucleotides of the intron adjacent to exon 42 ( Figure 3C).

Analysis of the ABCA4 Splice-Site Variants in Hair Follicles
Considering that the expression of the ABCA4 gene in the skin samples was significantly lower than in the hair follicles, we used the latter for further experiments aimed at assessing the sequence of ABCA4 transcripts in patients' samples.
In order to evaluate the usefulness of eyebrow hair follicles as a tool for the molecular analysis of ABCA4 variants, we investigated the consequences of the three ABCA4 splice-site variants c.5312+1G>A, c.5312+2T>G and c.5836-3C>A, previously detected in patients with Stargardt dystrophy [7]. In those patients, ABCA4 gene mutations were found in a compound heterozygous state with missense (NM_000350.

Analysis of the ABCA4 Splice-Site Variants in Hair Follicles
Considering that the expression of the ABCA4 gene in the skin samples was significantly lower than in the hair follicles, we used the latter for further experiments aimed at assessing the sequence of ABCA4 transcripts in patients' samples.
In order to evaluate the usefulness of eyebrow hair follicles as a tool for the molecular analysis of ABCA4 variants, we investigated the consequences of the three ABCA4 splice-site variants c.5312+1G>A, c.5312+2T>G and c.5836-3C>A, previously detected in patients with Stargardt dystrophy [7]. In those patients, ABCA4 gene mutations were found in a compound heterozygous state with missense (NM_000350.

Analysis of the ABCA4 Gene Expression in Hair Follicles of Patients with Stargardt Disease
In order to analyze the overall expression of the ABCA4 gene expression in patients with Stargardt disease we performed qRT-PCR. The results show that ABCA4 mRNA is also expressed in the hair follicles obtained from patients with Stargardt disease, however, this expression was significantly lower than in controls (p = 0.0001 by Mann-Whitney U test) ( Figure 4).

Discussion
Here, we report that full-length ABCA4 transcript is highly expressed in human eyebrow hair follicles and may be thus considered useful for the molecular analysis of splice-site ABCA4 mutation consequences in patients with ABCA4 retinopathies. The applicability of the hair follicle model was confirmed by the analysis of three different splice-site ABCA4 variants from patients with Stargardt disease.
The ABCA4 gene was previously thought to be expressed at very low levels in non-ocular tissues. Although a previous high-throughput mRNA analysis of ABC transporters failed to detect ABCA4 expression in human total skin [39], nevertheless, we detected low levels of ABCA4 expression in human skin explants. On the other hand, a relatively high ABCA4 gene expression was observed in human hair follicles. This is in agreement with the observation made by Haslam et al., who also reported a high expression level of the ABCA4 gene throughout the epithelium of hair follicles [37].

Analysis of the ABCA4 Gene Expression in Hair Follicles of Patients with Stargardt Disease
In order to analyze the overall expression of the ABCA4 gene expression in patients with Stargardt disease we performed qRT-PCR. The results show that ABCA4 mRNA is also expressed in the hair follicles obtained from patients with Stargardt disease, however, this expression was significantly lower than in controls (p = 0.0001 by Mann-Whitney U test) ( Figure 4).

Analysis of the ABCA4 Gene Expression in Hair Follicles of Patients with Stargardt Disease
In order to analyze the overall expression of the ABCA4 gene expression in patients with Stargardt disease we performed qRT-PCR. The results show that ABCA4 mRNA is also expressed in the hair follicles obtained from patients with Stargardt disease, however, this expression was significantly lower than in controls (p = 0.0001 by Mann-Whitney U test) (Figure 4).

Discussion
Here, we report that full-length ABCA4 transcript is highly expressed in human eyebrow hair follicles and may be thus considered useful for the molecular analysis of splice-site ABCA4 mutation consequences in patients with ABCA4 retinopathies. The applicability of the hair follicle model was confirmed by the analysis of three different splice-site ABCA4 variants from patients with Stargardt disease.
The ABCA4 gene was previously thought to be expressed at very low levels in non-ocular tissues. Although a previous high-throughput mRNA analysis of ABC transporters failed to detect ABCA4 expression in human total skin [39], nevertheless, we detected low levels of ABCA4 expression in human skin explants. On the other hand, a relatively high ABCA4 gene expression was observed in human hair follicles. This is in agreement with the observation made by Haslam et al., who also reported a high expression level of the ABCA4 gene throughout the epithelium of hair follicles [37].

Discussion
Here, we report that full-length ABCA4 transcript is highly expressed in human eyebrow hair follicles and may be thus considered useful for the molecular analysis of splice-site ABCA4 mutation consequences in patients with ABCA4 retinopathies. The applicability of the hair follicle model was confirmed by the analysis of three different splice-site ABCA4 variants from patients with Stargardt disease.
The ABCA4 gene was previously thought to be expressed at very low levels in non-ocular tissues. Although a previous high-throughput mRNA analysis of ABC transporters failed to detect ABCA4 expression in human total skin [39], nevertheless, we detected low levels of ABCA4 expression in human skin explants. On the other hand, a relatively high ABCA4 gene expression was observed in human hair follicles. This is in agreement with the observation made by Haslam et al., who also reported a high expression level of the ABCA4 gene throughout the epithelium of hair follicles [37].
Interestingly, we found that mRNA expression of the ABCA4 exons 3-4 was considerably lower as compared to other exons. This may implicate presence of alternate splicing in normal hair follicles, though one cannot exclude the possibility of an insufficient qPCR reaction for this particular probe. Nevertheless, we confirmed the presence of all ABCA4 exons of transcripts isolated from human hair follicles, cultured human keratinocytes, fibroblasts and melanocytes by means of direct Sanger sequencing. Presence of the ABCA4 full-length protein was also confirmed by the Western blot method.
In the hair follicles obtained from healthy controls, a high variability of the ABCA4 gene expression was observed. This was probably caused by individual differences of the ABCA4 gene expression in the population. However, it may be also associated with an unknown function of the ABCA4 in the hair follicles. It is also tempting to speculate that the slightly decreased expression of the ABCA4 gene in the control group (though significantly greater than that observed in the patients) may be related to existence of the pathogenic ABCA4 variants, as has been demonstrated in a canine model of Stargardt disease [40]. This, however, requires further analysis of ABCA4 gene expression in the probands' parents.
Interestingly, ABCA4 mRNA expression in keratinocytes was very low compared to hair follicles, yet ABCA4 protein was detected at a relatively high level by the Western blot method. This difference may be due to variability between samples of different origins. However, this phenomenon may be also explained by the posttranscriptional regulation of the ABCA4 transporter in the skin, as has been observed for other ABCA transporters in tissues in which protein expression was much higher that would be expected based on their level of mRNA expression [41].
Variants potentially affecting RNA splicing can be directly tested with the use of qRT-PCR on accessible cells (e.g., skin cells, lymphoblasts), providing that they constitutively express a sufficient level of the gene of interest. To assess whether hair follicles are a suitable source of material for the study of the consequences of ABCA4 mutations, we analyzed hair follicle-derived mRNA obtained from patients with Stargardt disease carrying representative c.5312+1G>A, c.5312+2T>G and c.5836-3C>A ABCA4 variants [7]. Our findings are in concordance with the results of Schulz et al. who also observed skipping of the 37th exon caused by the c.5312+1G>A mutation using in vitro splicing assay [42]. Furthermore, our results concerning c.5836-3C>A mutation were also consistent with the observations of Sangermano et al. by means of midigene assays [24]. It would also be of great interest to analyze other ABCA4 gene splice variants, especially those located at the 5 region of the ABCA4 gene. Unfortunately, due to lack of appropriate clinical material from patients, analysis of such variants could not be performed in the present study and awaits further investigation.
Analysis of the overall ABCA4 gene expression in patients with Stargardt disease showed a high reduction as compared to healthy controls. This is in agreement with results obtained from different cell types, which also showed a reduction of normally spliced full-length ABCA4 mRNA in fibroblasts of patients with Stargardt disease [23]. It needs to be further elucidated whether the expression level of the ABCA4 gene in patients with Stargardt disease correlates with disease severity.
Eyebrow hair follicles are regularly plucked by millions of women worldwide. Unlike a skin biopsy, no anesthetic is needed and no skin injuries are inflicted upon the patient. In addition, material can be obtained directly from patients and their families and sent to the laboratory without the need of patient arrival. About 10-15 eyebrow hair follicles yield an amount of RNA sufficient to conduct a preliminary molecular assessment of ABCA4 variant pathogenicity and ABCA4 gene expression analysis. This method appears to be simple and inexpensive, especially as it does not require cell culture or the highly advanced generation of mini-, midigenes or photoreceptor precursor cells (PPCs).
It should be stressed, however, that not all molecular and functional analyses can be performed on non-cultured cells from hair follicles. It is impossible to block nonsense-mediated decay in non-cultured hair follicle cells, therefore ABCA4 splice-site variant consequences can be missed in heterozygous patients. Also, reverse transcription reactions are difficult to conduct on large genes where 5 of the gene may not be efficiently retrotranscribed, as compared with 3 of the gene. For this purpose, the construction of mini-or midigenes is more favorable, even though mini-genes may result in the creation of splicing artefacts [13]. In order to improve the accuracy of these assays, interesting results regarding functional analysis of the ABCA4 splice-site variants were obtained with photoreceptor precursor cells generated from skin fibroblasts [24,27,32,33]. Still, the invasiveness of skin biopsies may impede these analyses. Hair follicles may be non-invasively obtained from patients multiple times and may also be a source of cultured epithelial stem cells [43]. Whether consequences of the ABCA4 splice-site variants in cultured hair follicle cells would lead to similar outcomes as those observed in PPCs needs further investigation.
In conclusion, analysis of mRNA from hair follicles appears to be a reliable method to assess the biological consequences of functional changes of the ABCA4 gene in patients with Stargardt disease and possibly other ABCA4 retinopathies. Furthermore, this method is non-invasive for the patients and is relatively simple and inexpensive as compared to other methods based on highly advanced molecular techniques.

Culture of Primary Keratinocytes, Fibroblasts and Melanocytes
Keratinocyte cultures were established as described previously with minor modifications [44]. Briefly, the epidermis was separated from the dermis with 2.5 U/mL of dispase by an overnight digestion at 4 • C. Keratinocytes and melanocytes were released from the epidermis after 5 min of cleavage in 0.25% trypsin at 37 • C with subsequent pipetting. The cells were cultured in Medium 254 (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with Human Melanocyte Growth Supplement-2, PMA-free (Thermo Fisher Scientific, Waltham, MA, USA). Following one-week culture, melanocytes were separated from keratinocytes with a short (up to 2 min) trypsinization. Keratinocytes were subsequently cultured in Keratinocyte Growth Medium (PromoCell, Heidelberg, Germany) supplemented with Keratinocyte Supplement Mix (PromoCell, Heidelberg, Germany). Dermal fibroblasts were released by 0.6 U/mL of Collagenase (Serva Electrophoresis GMbH, Heidelberg, Germany) digestion for 1 h at 37 • C with subsequent pipetting. Then they were cultured in DMEM (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% Fetal Bovine Serum (EURx, Gdansk, Poland). Cells in passages 1-2 were used in subsequent experiments.

RNA Isolation and Quantitative mRNA Measurements
Skin explants cut slightly below the basal layer of the epidermis with shallow fragments of the dermis and eyebrow hair follicles were submerged in the RNAlater Storage Solution (Thermo Fisher Scientific, Waltham, MA, USA) and stored at 4 • C prior to subsequent isolation of total RNA and protein within 6 h. Total RNA from skin explants, skin cells and hair follicles was extracted with NucleoSpin ® TriPrep (with rDNase digestion) (Macherey-Nagel, Dürren, Germany) and transcribed into cDNA with RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA). For quantitative PCR, 20 ng of cDNA per reaction was used. All real-time PCR reactions were performed on the ABI 7500 Fast Real-time PCR system (Applied Biosystems, Inc., Thermo Fisher Scientific, Waltham, MA, USA). Expression levels of the ABCA4 gene in hair follicles, skin cells and skin explants from healthy donors and hair follicles from the patients were analyzed with Hs00979586_m1 TaqMan ® probe (Thermo Fisher Scientific, Waltham, MA, USA). TBP gene (TATA-box binding protein, Hs00427620_m1) was used as a reference gene for further normalizations. All samples were run in triplicates. The relative level of mRNA expression was calculated by the ∆Ct or ∆∆Ct method.
qRT-PCR evaluation of particular ABCA4 exon mRNA expression in hair follicles and skin cells was performed with the use of TaqMan ® probes presented in Table 2. PCR reactions were performed as described above and specific exon expression was presented as number of cycle threshold (Ct).  [45] based on the NM_000350.3 reference sequence. The primers were designed to encompass an area of the five to six adjacent exons, therefore the presence of the full-length ABCA4 transcript was analyzed with 11 sets of primers, listed in Table 3.

Protein Extract Isolation and Western Blot Analysis
Proteins were extracted from the hair follicles and skin explants, as well as cultured keratinocytes, dermal fibroblasts and melanocytes, with NucleoSpin ® TriPrep (Macherey-Nagel), as described in detail by the manufacturer. Lysates were pelleted and resuspended in a buffer containing 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS and 50 mM Tris (pH 8.0). Protein concentration was evaluated with Pierce BCA Protein Assay Kit (Thermo Fisher Scientific Inc.) and 20 µg protein samples were separated by SDS-10% PAGE, transferred onto a nitrocellulose membrane (Thermo Fisher Scientific Inc.), blocked with Pierce Protein-Free Blocking Buffer (Thermo Fisher Scientific Inc.) and incubated at 4 • C overnight with 1:250 rabbit anti-ABCA4 (cat. no. PA5-87983, Thermo Fisher Scientific Inc.) or 1:500 mouse anti-β-actin (cat. no. A1978, Sigma-Aldrich, St. Louis, MO, USA). Secondary HRP-conjugated goat anti-rabbit IgG antibodies (cat. no. GTX213110-01, GeneTex) or rabbit anti-mouse IgG antibodies (cat. no. GTX213111-01, GeneTex, CA, USA) were used and visualized with SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Inc.). As an ABCA4 positive control, 20 µg of Y79 human retinoblastoma cell lysate (sc-2240, Santa Cruz Biotechnology) was used. A skin sample in which ABCA4 mRNA expression was not detected by qRT-PCR served as a negative control. Broad range BlueEasy Prestained Protein Marker (Nippon Genetics Co., Dürren, Germany) was used as a protein standard.

Statistical Analysis
Statistical analyses and graphs were generated with GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA). The nonparametric Mann Whitney U test and the one-way ANOVA were used to compare ABCA4 expression levels in different studied groups. p-value < 0.05 was considered significant.

Conflicts of Interest:
The authors declare no conflict of interests regarding this paper.