Long‐term culture of skin biopsies: maintenance of fibroblast production and competency of reprogramming

All skin biopsies were conducted under the approval of Health Sciences Review Committee #2 of the Institutional Review Board of Vanderbilt University Medical Center, IRB #080369. All research was performed in accordance with relevant guidelines/regulations including those in accordance with the Declaration of Helsinki. Written informed consent/assent was obtained from each subject, and a single 3 mm punch skin biopsy was obtained from the posterior aspect of an arm of each subject. Tissue was dissected in half longitudinally, cut downwards from the exterior surface. Each half was then plated in a single well of a 6‐well culture plate. A sterile glass slide cover was placed on the skin section. 3 mL of DMEM Complete [1× DMEM (Gibco 11995‐065, Grand Island, NY, USA), 10% FBS (Corning 35‐015‐CV, Manassas, VA, USA), 100 U·mL⁻¹ Penicillin/Streptomycin (Gibco 15140‐122), 1× MEM Non‐Essential Amino Acids (Gibco 11140‐050)] was added to each well, completely submerging the biopsy and slide. Plates were incubated at 37 °C and 5% CO₂, gently aspirating and changing media twice a week. After 1–3 weeks, fibroblast‐appearing cells began to migrate out from the skin chunk. When confluent, the still intact skin chunks were moved to a new 6‐well plate, with a fresh sterile glass slide cover to keep the tissue adherent to the culture dish. Fibroblasts from the original well were passaged using 0.1% Trypsin–EDTA and plated on a 10 cm petri dish with DMEM Complete for further growth and expansion before cryopreservation.

For storage, cells were trypsinized and resuspended in fibroblast freezing media [50% DMEM Complete, 40% FBS (Corning 35‐015‐CV), 10% DMSO (Sigma‐Aldrich D8418, Saint Louis, MO, USA)]. The cell solution was then aliquoted into multiple cryotubes for each generation, frozen overnight at −80 °C, and then moved to liquid nitrogen for long‐term storage. Initial fibroblasts grown from the skin explant were denoted as the first generation (Gen1) and subsequent passaging and resulting fibroblast cultures as Gen2 through Gen16. HIKHOM‐1 was harvested at Gen16 after 16 months in culture; HIKHet‐2 was harvested at Gen6 after 6 months in culture; unrelated controls CH1‐CON and CH2‐CON were harvested at Gen1 after 1 month in culture.

We focused our long‐term efforts on a punch biopsy from a male patient with Hikeshi Associated Leukodystrophy (HAL), a devastating autosomal recessive neurological disorder caused by homozygous loss of function mutations in the HIKESHI gene [41, 42, 43]. We also utilized cells obtained from healthy, asymptomatic first‐degree relatives who had heterozygous HIKESHI mutations as controls. We also obtained skin biopsies from two unrelated healthy individuals (CH1‐CON and CH2‐CON). The genotype of all lines was confirmed by DNA sequencing of the HIKESHI gene and further validated by immunostaining for HIKESHI protein.

Skin biopsies were removed from the culture plate and fixed in 4% paraformaldehyde (PFA) for a minimum of 24 h. The biopsies were then transferred into 70% ethanol and submitted to the Vanderbilt Translational Pathology Shared Resource. Fixed tissue was embedded in paraffin and cut into 5 μm thick sections. One set of slides was stained with hematoxylin and eosin (H&E) [44]. The H&E‐stained slides were submitted to the Vanderbilt Digital Histology Shared Resource and imaged with the Leica SCN400 Slide Scanner (Leica Biosystems, Nussloch, Germany) at 40× magnification.

Tissue sections were deparaffinized and rehydrated by immersion in xylene and a series of ethanol dilutions. Antigen retrieval was performed utilizing heated citrate buffer pH 6 [diH₂O, 11.3 mm Trisodium citrate, 0.05% Tween 20]. Slides were incubated with blocking buffer [1× PBS, 5% Normal Donkey Serum (Sigma‐Aldrich D9663), 0.3% Triton X‐100] for 1 h at room temperature. Primary antibodies (Table 1) were diluted in the blocking buffer, and 100 μL was added to each well in a humidified chamber at 4 °C overnight. Secondary antibodies were diluted in blocking buffer and incubated at room temperature for 2 h. Slides were then washed twice with 1× PBS and mounted using Prolong Antifade DAPI mounting medium (Invitrogen P36935↗, Carlsbad, CA, USA) and allowed to dry overnight before imaging on the EVOS FL (Waltham, MA, USA) Auto microscope.

Eight‐well chambered slides (Ibidi 80841, Gräfelfing, Germany) were coated with 0.1 mg·mL⁻¹ Poly‐l‐lysine (Sigma‐Aldrich P4707) for 20 min. Poly‐l‐lysine was aspirated, and slides were washed twice with dH₂O and air dried for 2 h. Fibroblasts were harvested from a confluent T25 flask using 0.25% Trypsin–EDTA solution (Gibco 25200‐056) for 7 min at 37 °C before being spun down at 300 × g for 5 min, and the pellet was resuspended in 1 mL DMEM Complete. A Countess Cell Counter was used to count cells resuspended in DMEM Complete, plating 2000 cells into each well of the slide. Slides were incubated overnight at 37 °C. Media was aspirated, and slides were washed with 1× PBS before proceeding with antibody staining as above.

Fibroblasts were plated on 8‐well chambered slides (Ibidi 80841) at 2500 cells per well. Cells were fixed and stained for expression of Ki67 and imaged on the EVOS FL Auto microscope at 10× magnification. evos Software captured 15 randomized images each from four wells. The captured images were analyzed and the total number of DAPI nuclei and Ki67 positive cells was counted. Cell counts were averaged for the well and each well was denoted as a data point for statistical analysis and graphing. Graphing was performed on graphpad prism.

A proliferation assay was conducted on fibroblasts over a 7‐day period. Fibroblasts were harvested from a confluent T25 flask using 0.25% Trypsin–EDTA solution (Gibco 25200‐056) for 7 min at 37 °C. Fibroblasts were spun down, and the pellet was resuspended in 1 mL DMEM Complete. Cells were counted on the Nexcelom Cellometer Auto T4 Cell Counter and resuspended in DMEM Complete at 10 000 cells·mL⁻¹. 100 μL of cell solution was added to 8 wells per genotype of 8 96‐well plates (1 for each day) for 1000 cells per well. Plates were incubated at 37 °C. A Thermofisher CyQuant Direct Cell Proliferation Assay (C35011↗) was utilized for analysis. For Day 0 to Day 7, the staining solution was prepared per the kit's instructions and 100 μL was added per well. Plates were incubated for 1 h at 37 °C. Plates were then read at 485/520 nm with a Biotek Synergy HT plate reader. Data were reported per previously published methods [45, 46, 47]. Data and statistics were compiled using biotek gen5 and graphpad prism.

Fibroblasts were plated in a 24‐well plate at 100 000 cells per well with 4 wells for each of the 5 lines (HIKHet‐2 Gen1, Gen 6, and HIKHOM‐1 Gen1, Gen 6, Gen 16) [48]. Cells were incubated overnight at 37 °C, and the following day a scratch was made in the middle of the well using a sterile pipette tip. The cells were then imaged every 2 h for 24 h on the evos xl Core microscope at 4× magnification. The scratch area was quantified using the Wound Healing Size Tool PlugIn on imagej [49]. The data were analyzed, and the slope was determined to calculate the t^1/2 gap values across genotypes [50]. The statistics were compiled using graphpad prism.

Fibroblasts were harvested and resuspended as above. RNA was extracted using the Qiagen RNAeasy Kit (Qiagen 74104, Hilden, Germany) and Qiagen RNA‐free DNase kit (Qiagen 79254). Samples were quantified using a NanoDrop and submitted to Vanderbilt Technologies for Advanced Genomics core laboratory.

Paired‐end reads were obtained from Illumina NovaSeq6000 PE150 Sequencing (San Diego, CA, USA). RNA sequence processing was performed with pyrpipe v0.0.5. First sequence adapters from the reads were trimmed with trimgalore v0.6.6 and aligned with star v2.7.7a to the Genome Reference Consortium Human Build 38 patch release 14 (GRCh38.p14). Assembly of alignments was done with StringTie v2.14 and the GRCh38.p14 NCBI RefSeq annotation assembly (GCF_000001405.40). Gene abundances were estimated using stringtie. Gene and transcript read counts were extracted for each sample with prepDE.py provided by Johns Hopkins University Centre for Computational Biology. pydeseq2 v0.3.5 was used to perform single and multiple factor differential expression analysis among Generations 1 and 6 for both cell lines and Generation 16 from HIKHOM‐1. Each paired read was processed independently.

Genes with adjusted P‐values by the Benjamini and Hochberg method under or equal to 0.001 and log fold change greater than 1 or less than −1 were selected for Gene Ontology Enrichment Analysis (GOEA) using goatools (v1.3.2) [51]. All human genes from Gencode annotations release 39 (GRCh38.p13) served as the background population set. Benjamini–Hochberg P‐values were obtained for each Gene Ontology biological process, cellular component, and molecular function pathway. Upregulated and downregulated genes were analyzed separately. Gene associations were taken from the human Gene Ontology 2023 library.

Fibroblasts were harvested and resuspended as above. DNA was extracted using the DNAeasy Kit (Qiagen 69504). Samples were quantified using a NanoDrop (Wilmington, DE, USA). Samples were submitted to CD Genomics for low‐pass whole genome sequencing and analysis. Samples were run through a quality test and then utilized to create a construct library. The generated library was sequenced using the Illumina HiSeq sequencing platform, producing raw data. The raw data were processed to filter out various reads, including low‐quality and redundant reads. The sequencing quality score is determined, and the distribution of base content is calculated from the raw data. The reads were then aligned with a reference genome, and copy number variations (CNV) were determined. Deletions and duplications were identified and annotated with gene names or Ensembl IDs. Chromosome diagrams were generated for each genotype, depicting the CNVs. Statistics were conducted throughout the process for sequencing reads and alignment.

GOEA was performed in the same manner as with gene expression using DNA duplications and deletions from Generation 16 fibroblasts. Both duplications and deletions were considered a change to the gene and used as the study population. Changes already present in Generation 1 were removed from this set. A P‐value of 0.001 or less was used to identify modified pathways.

Fibroblasts were reprogrammed using the CytoTune‐iPS 2.0 Reprogramming Kit (Invitrogen A16517). Fibroblasts were plated at 2 × 10⁵ cells per well of a Matrigel‐coated [Corning #354277, diluted in DMEM/F12 per manufacturer's instructions] 6‐well plate. Cells were fed with DMEM Complete. The following day (Day 0), 1 well was dissociated from the plate and the cell number was counted. The volume of virus required was calculated using the formula:

Viruses hKOS and cMYC had a MOI of 5 and KLF4 a MOI of 3. Virus was added to DMEM complete without Penicillin/Streptomycin, adding 1 mL per well of cells. The following day (D1), the virus solution was carefully removed, and the cells were fed with DMEM Complete, changing the media every day from Day 1 to Day 6.

On Day 7, the cells were disassociated with 0.25% Trypsin and counted. They were plated on Matrigel‐coated 6‐well plates with densities of 5 × 10⁴, 7.5 × 10⁴ and 1 × 10⁵ cells per well. The cells were moved to Essential 8 Medium (Gibco A15169‐01) with Supplement (Gibco A15171‐01). Cells were then grown in E8 Medium, feeding every other day from Day 7 to Day 21. Cells were then moved to mTesR1 media (StemCell Technologies #85857, Vancouver, Canada). Individual colonies were identified and picked by scraping; colonies were transferred to a Matrigel‐coated 48‐well plate and grown on mTeSR1. Colonies were passaged over 5 generations and then fixed with 4% PFA for 30 min at room temperature. Immunofluorescence on colonies was done as above.

Reprogrammed iPSCs from HIKHet2 Gen1 and HIKHOM1 Gen16 were assessed using the trilineage embryoid body assay [52, 53]. iPSCs were grown to confluency, dissociated with ReLeSR (Stem Cell Technologies #100‐0483), and resuspended in mTeSR1 media. Cells were plated on low‐adherence plates for 4 days, feeding daily with a mixture of mTeSR1 and embryoid body media [1× DMEM/F12 + GlutaMAX (Gibco 10565‐018), 20% KnockOut Serum Replacement (Gibco 10828‐028), 100 U·mL⁻¹ Penicillin/Streptomycin (Gibco 15140‐122), 1× MEM Non‐Essential Amino Acids (Gibco 11140‐050), 952.38 μm beta‐mercaptoethanol]. On Day 4, embryoid body clusters were plated on 3 Matrigel‐coated plates (one each for endoderm, mesoderm, and ectoderm). Cells were fed every day with embryoid body media. On Day 15, endoderm and mesoderm plates were fixed with 4% PFA for 20 min at room temperature. On Day 20, the ectoderm plate was fixed. All plates were stained using the immunofluorescence technique listed above.

All analysis and graphs were produced using graphpad prism version 10.5.0 for Windows, GraphPad Software, Boston, MA, USA, www.graphpad.com↗. The conducted tests, number of replicates, and P‐values are listed in the corresponding Figure legend.

We thawed fibroblasts that were obtained from different generations and genotypes and measured proliferation over 7 days (Fig. 3A,B). There were no significant changes in proliferation between HIKHet‐2 Gen1 and HIKHet‐2 Gen6 over 7 days. However, there was a significant decrease in proliferation between HIKHOM‐1 Gen1 and HIKHOM‐1 Gen6 from Day 6 onwards (Fig. 3B). Furthermore, there was a significant decrease in proliferation in HIKHOM‐1 Gen16 compared to both HIKHOM‐1 Gen1 and HIKHOM‐1 Gen6 from Day 4 onwards. It was also noted that HIKHOM‐1 cell lines overall had a higher proliferative rate than HIKHet‐2 cells. This could be attributed to age differences between the donors or possibly due to the homozygous mutation in the HIKESHI gene.

To determine the percentage of proliferating cells, cells were fixed and stained for the proliferation marker Ki67 [59] (Fig. S3A). Cell counts indicated that the percentage of Ki67‐expressing cells decreased over generations (Fig. 3C). There was a downward trend between Gen1 and Gen6 of both HIKHet‐2 and HIKHOM‐1 genotypes; however, this decrease was not significant. There was a significant decrease between HIKHOM‐1 Gen1 and HIKHOM‐1 Gen16, suggesting that long‐term culture (over 15 months) slows but does not halt proliferation. Similar to the proliferative assay, the HIKHOM‐1 genotype displayed a higher percentage of Ki67‐positive cells across all generations when compared to the HIKHet‐2 cells.

To analyze the presence of proliferative cells in the biopsy samples after long‐term culture, the biopsies were ultimately fixed, paraffin‐embedded, and immunostained for proliferative markers. Immunofluorescence staining indicates a lack of Ki67 positive and PCNA positive cells in HIKHOM‐1, HIKHet‐2, and unrelated control skin biopsies (data not shown). The validity of this assay was confirmed utilizing multiple antibodies and positive control tissue. Cells were also fixed and stained for the apoptosis marker, Cleaved Caspase 3 (CC3) [60]. Fibroblasts were negative for CC3, indicating minimal, if any, apoptotic cell death (Fig. S3B). Antibody validity was confirmed utilizing iPSCs as positive control cells (Fig. S3B).

To study the proliferative and repair effect of the fibroblasts, the cells were analyzed using a wound healing or scratch assay [48]. There is a significant increase in the t^1/2 gap of the HIKHOM‐1 Gen16 cells when compared to the other 4 lines (HIKHet‐2 Gen 1, HIKHet‐2 Gen 6, HIKHOM‐1 Gen 1, and HIKHOM‐1 Gen 6), indicating a longer time required to close the gap and lower proliferative capacity (Fig. 3D). There appears to be no significant difference between the t^1/2 gap rates of the other lines (HIKHet‐2 Gen 1, HIKHet‐2 Gen 6, HIKHOM‐1 Gen 1, and HIKHOM‐1 Gen 6), indicating that the lowered proliferation is linked to the longer‐term culture. The gap area data indicate that the HIKHOM‐1 Gen16 was unable to close the gap after 24 h of culture, whereas the other lines (HIKHet‐2 Gen 1, HIKHet‐2 Gen 6, HIKHOM‐1 Gen 1, and HIKHOM‐1 Gen 6) successfully closed the gap within the 24‐h period (Fig. S4).

To determine appreciable differences in gene expression, RNA was obtained from proliferating fibroblast lines from different generations and genotypes. RNA sequencing was conducted, and we observed significant upregulation and downregulation of various genes between HIKHOM‐1 Gen1 and HIKHOM‐1 Gen16 (Fig. 4). Compared to HIKHOM‐1 Gen1, various genes in HIKHOM‐1 Gen16 such as HAPLN1, LHX9, and SDK2 were highly upregulated, and SMO, C3, and EMILIN2 were highly downregulated (Table S1). Gene enrichment analysis comparing HIKHOM‐1 Gen16 to HIKHOM‐1 Gen1 indicated that a variety of pathways were changed. Pathways relating to the extracellular matrix, ion channel activity, and transmembrane transport were upregulated. Pathways relating to the regulation of the immune system, response to stimulus, and cell surface receptor signaling were downregulated (Table S2). Analysis between HIKHOM‐1 Gen1 and HIKHOM‐1 Gen6 as well as HIKHet‐2 Gen1 and HIKHet‐2 Gen6 also displayed changes in expression and gene enrichment, but substantially fewer changes to expression and pathways were observed, implying long‐term culture amplified transcriptional changes (Fig. S1).

As prolonged time in culture could predispose cells to genomic instability and acquisition of mutations, we processed DNA from proliferating fibroblast lines from different generations and genotypes for low‐pass whole genome sequencing [61, 62, 63, 64]. This allowed us to assess chromosomal stability of long‐term culture and to detect any large DNA sequence gains or losses. Chromosomal maps indicate that there were no significant chromosomal gains or losses that developed over generations (Fig. S2).

Further enrichment analysis determined that between HIKHOM‐1 Gen1 and HIKHOM‐1 Gen16, only a single pathway was significantly changed. A deletion in Chromosome 22 affected genes in the APOBEC3 cluster including APOBEC3A, APOBEC3B, APOBEC3C, APOBEC3D, and APOBEC3F (Table S3). This resulted in potential reductions in deaminase activity and negative regulation of viral processes [65, 66, 67, 68]. Despite these genomic changes, the cells displayed no increased susceptibility to contamination. The genomic analysis also indicated that mutations occurred at random and no specific pathway, such as cancer‐related pathways, was targeted.

To determine if prolonged propagation of skin biopsies could produce fibroblasts that retain competency for reprogramming to iPSC, HIKHet‐2 Gen1, and HIKHOM‐1 Gen16 fibroblasts were reprogrammed with Sendai virus expressing Yamanaka factors. By Day 16, colonies of round‐shaped cells were evident as well as surrounding cells maintaining a typical fibroblast morphology. On Day 21 after transduction with Sendai viruses, emerging iPSC‐like colonies were manually picked and replated. These were grown out for several more days before being imaged. The individual colonies were flat and consistent with iPSC morphology (Fig. 5A), consisting of small, round cells with a high nuclear/cytoplasmic ratio with prominent nucleoli [69].

To provide further evidence that HIKHOM‐1 G16 fibroblasts retained competency for reprogramming, individual colonies were immunostained for pluripotency markers Nanog, Oct4, TRA‐1‐60, SSEA3, and SSEA4 [70, 71]. All lines stained positive for all markers (Fig. 5B), indicating pluripotent status and successful reprogramming. HIKHet‐2 Gen1 was also reprogrammed and stained for pluripotency as a control (Fig. 5C).

HIKHet‐2 Gen1 and HIKHOM‐1 Gen16 iPSCs were further validated using a trilineage embryoid body assay (Fig. 5D). Both lineages expressed the endoderm marker GATA4, the mesoderm marker Smooth Muscle Actin, and the ectoderm markers Sox1 and β‐III‐Tubulin, indicating that both cell lines can be successfully differentiated into the three germline lineages [72, 73].

RNA Sequencing data was deposited in Gene Expression Omnibus (GEO) under accession GSE271347↗.