RNA sequence analysis of dermal papilla cells’ regeneration in 3D culture

Abstract It is well known that dermal papilla cells (DPCs) are crucial for hair follicle growth and regeneration. However, dermal papilla cells in 2D culture could lose their ability of regeneration after several passage intervals. As opposed to DPCs in 2D culture, the DPCs in 3D culture could passage extensively. However, the molecular mechanisms of DPCs’ regeneration in 3D culture remain unclear. Accordingly, gene sequencing is recommended for the investigation of hair regeneration between 2D and 3D culture, the three groups were established including DPCs in passage 2 in 2D culture, DPCs in passage 8 in 2D culture and DPCs in passage 8 in 3D culture. The differentially expressed genes (DEGs) were identified using the Venn diagram of these three groups, which included 1642 known and 359 novel genes, respectively. A total of 1642 known genes were used for Gene Ontology (GO), Kyoto Gene, Genomic Encyclopedia (KEGG) pathway enrichment and protein‐protein interaction (PPI) analyses, respectively. The functions and pathways of DEGs were enriched in biological regulation, signal transduction and immune system, etc. The key module and the top 10 hub genes (IL1B, CXCL12, HGF, EGFR, APP, CCL2, PTGS2, MMP9, NGF and SPP1) were also identified using the Cytoscape application. Furthermore, the qRT‐PCR results of the three groups validated that the hub genes were crucial for hair growth. In conclusion, the ten identified hub genes and related pathways in the current study can be used to understand the molecular mechanism of hair growth, and those provided a possibility for hair regeneration.

studies reported that dermal papilla is responsible for hair generation, its exact mechanism is not clear. 3,4 Cell culture technology is vital for the study of diseases and cancer. Currently, the three-dimensional (3D) culture technology is rapidly growing in the life science sector. As compared to 2D culture, cells in 3D culture technology are closer to the complex in vivo conditions. 5 A significant advantage of 3D culture over 2D culture conditions is the decrease in the gap between cell models in vitro and in vivo. Besides, the features of 3D culture cells are closer to the complex in vivo condition. 6 Some researchers found that 3D cell cultures could induce pluripotent stem cells (iPSC) to generate personalized tissue models. 7 As the dermal papilla cells (DPCs) are unique mesenchymal cells, we previously found out that DPCs could also induce human hair growth in mice by transplanting intact DPCs and epithelial cells. 4,8 However, such a process requires large numbers of human DPCs as they lose the ability of hair inductivity in 2D in vitro culture. Therefore, most researchers reported that the 3D-spheroid culture of DPCs could restore DPCs' inductive characteristics; however, the science behind such observations is unknown. 9 RNA-Seq is an excellent approach to study transcriptome profiling using deep-sequencing technologies. 10 The transcriptome includes all sets of transcripts in a cell, which is crucial for further analysis of its molecular constituents and gene functionality. The objective of transcriptomics is to categorize all species of the transcript, which includes mRNAs, LncRNAs, CircRNAs and MicroRNAs. 11 RNA-Seq provides a new approach to analyse eukaryotic transcriptomes using a novel, high-throughput and cheaper DNA sequencing methods than traditional methods such as microarray technology 12 and Sanger sequencing technology. 13 Therefore, in this study, we applied the novel RNA-Seq technology to investigate the effect of 3D culture on reversing the ageing process of DPCs. Gene Ontology (GO), Kyoto Gene and Genomic Encyclopedia (KEGG) pathway enrichment analyses, as well as protein-protein interaction (PPI) network analysis, were conducted on genes to identify the molecular processes of DPCs' development and progression in 3D culture as compared to 2D culture.

| Human dermal papilla cells in 2D culture
Intact human dermal papillae were isolated from hair follicles in occipital scalp during hair transplantation surgery of patients. We removed the matrix and exposed the dermal papilla from the follicles by fine needles. Papillae were isolated by cutting through their stalk. Next, 6-8 papillae were cultured in 35-mm dishes containing 20% foetal calf serum (FBS) in Dulbecco's modified eagle medium (DMEM) with 1% three antibiotics medium (penicillin, streptomycin and fungizone).
Papillae were adhered to the bottom of the culture dish by fine needles. We observed that cells could migrate from the papillae, and the intact papillae eventually collapsed after 7 days. When dermal papilla cell density approached 80% confluence, the cells were passaged at 1:2 using trypsin containing 0.5% ethylene diamine tetraacetic acid (EDTA). These cells were passaged twice and tagged as Passage 2 in 2D culture (P2-2D group). The dermal papilla cells were passaged 8 times, and therefore, the cells were named Passage 8 in 2D culture (P8-2D group). The dermal papillae cells were from four donors, so each group contained four repeated samples.

| Human dermal papilla cells in 3D culture
GravtityPlus hanging-drop plates were used for 3D culture. The plates consisted of the following components: (a) The bottom plate with a reservoir. (b) GravityPLUS Plate (raster plate) with 12x8-well strips. (c) Lid.

| Total RNA extraction and mRNA library construction
P2-2D group, P8-2D group and P8-3D group' RNA were extracted with TRIzol. 14 Agilent 2100 Bioanalyzer was used to detect the quality of RNA. The cDNA libraries were synthesized and amplified with the supplemented oligo primers. Next, they were loaded onto the flow cell channels of a BGISEQ-500 platform for sequencing at the Beijing Genomics Institute (BGI), Shenzhen, China.

| Identification of differentially expressed genes (DEGs)
The most common limitation of sequence data is the presence of false positives. The presence of false positives can be accounted for by adjusting the P-value (Q-value) as well as calculating the Benjamini and Hochberg false discovery rate, which also simplifies the identification of statistically significant genes. A logarithmic fold change value (logFC)> 1 and an adjusted P-value <.05 were used to choose the differentially expressed genes (DEGs) among three groups (P2-2D group, P8-2D group and P8-3D group). The overlap of two datasets (between P8-2D group and P2-2D group, between P8-2D group and P8-3D group) was identified by drawing Venn diagrams.

| GO and KEGG enrichment analysis of the DEGs
The Database for Annotation, Visualization and Integrated Discovery (DAVID; version 6.8; http://david.ncifc rf.gov) 15 is an online database that analyses bioinformatic data with online tools and offers an exhaustive result of the functions of genes or proteins.

Kyoto Encyclopedia of Genes and Genomes (KEGG) is used to com-
prehend the advanced features of a biological system using molecular datasets at the scale at which it is generated by high-throughput experimental techniques. KEGG (KEGG, https://www.kegg.jp/) is a database that helps users assign related molecular process, diseases and pathways of genes by high-throughput technology. 16 Gene Ontology (GO) is an essential web-based tool for annotating genes and unifying biological processes, cellular components and molecular function of target genes. 17 The biological function of the DEGs was analysed using the online database, DAVID. A Q-value (Adjusted P-value) of <.05 was considered to indicate statistical significance. in Cytoscape to analyse the PPI network, which can identify densely connected and bipartite network modules of target genes. In this study, the Cytoscape application was used to analyse the PPI network of DEGs; we also applied the MCODE plugin to identify the most crucial modules. The standards that were applied were as follows: degree cut-off = 2, node k-score = 2, score cut-off = 0.2, Max depth = 100 and MCODE scores >5.

| Hub genes selection and analysis
CytoHubba is an app plugin in Cytoscape that can obtain hub genes.
After the construction of the gene network, the top 10 with degree ≥10 were identified as hub genes. [20][21][22] The term 'degree' is defined as a number in the Cytoscape setting, which represents the relevance of one gene with other genes.

| Quantitative RT-PCR
The total RNAs were extracted using the RNAeasy kit (Takara

| Cellular morphology of three groups
When DPCs were passaged twice in the 2D culture, the cells formed a star-shaped morphology that tends to stick together ( Figure 1A).
In contrast, the shape of each cell became longer and less concentrated after the DPCs were passaged eight times than those DPCs that were passaged twice in 2D the culture, respectively ( Figure 1B). novel genes, as shown in the Venn diagram (Figure 2A).

| Classification and enrichment analysis of DEGs
Gene

| Construction of the PPI network and module analysis
The most significant module with 65 nodes and 991 edges was constructed in Figure 2C by using the Cytoscape MCODE application.

| Selection and analysis of Hub genes
Top ten genes with degrees ≥10 were recognized as hub genes.

F I G U R E 4 Kyoto encyclopedia of genes and genomes (KEGG) analysis of
DEGs. (A) Blue and red columns mean the cellular process and environmental information processing, respectively. The green column indicates the processing of genetic information processing, whereas the brown column represents human diseases. The grey and pink column indicates the metabolism and organismal system, respectively. (B) The size of the bubble means gene number; the colour depth means Q-value, and the rich ratio means the gene number/the total gene number in the y-axis item

P2-2D, P8-2D and P8-3D' FPKM (Fragments per Kilobase Million)
value of the hub genes are shown in Table 1. The changing trend of hub genes among the three groups is shown in Figure 5. We found out that the hub genes' functionality partially returned to P2-2D groups' expression level (except for CXCL12 and HGF) for DPCs that were cultured in 3D conditions. Figure 6 showed that the mRNA expression levels of hub genes (top 10) were similar to the RNA-seq results, that is, the expression levels of hub genes tends to return to the level as P2-2D groups after DPCs were passaged 8 times in 3D culture. has become a standard tool to study the molecular mechanism of diseases and species. 35,36 To the best of our knowledge, we are the first to explore the mRNA expression profiles of DPCs between 2D

| D ISCUSS I ON
and 3D culture using high-throughput sequencing. We identified the top ten hub genes (IL1B, CXCL12 According to our studies, when DPCs were passaged eight times, the cell morphology of DPCs could transform from star shape to spindle shape; and likewise, the ability of cells to form clusters or colonies decrease. Following the passage of DPCs (8 times) in 3D culture, the cells came together to form a multicellular spheroids, which are similar to the structure of hair follicles' dermal papilla in vivo. Hang drop 3D culture technology induces DPCs to form a spheroid using gravity, which could simulate in vivo conditions. Therefore, we could use hang drop 3D culture technology to restore the cellular function of DPCs, partially. However, the related molecular mechanism is still controversial. Therefore, we divided DPCs into three groups: P2-2D group, P8-2D group and P8-3D group. Each group included four repeated samples to avoid the influence between different individuals.
DPCs were offered by four donors, and patients with systemic disease were excluded from this study. The experimental results were verified in different individuals to increase credibility of founding.
We acquired the genes (DEGs between P8-2D group and P2-2D group) that had changed after DPCs were passaged multiple times.
Moreover, we also obtained the DEGs between P8-2D group and P8-3D group consisting of genes that had changed when DPCs were cultured in 3D conditions. Next, we made the Venn diagram of two datasets (P8-2D vs. P2-2D, P8-2D vs. P8-3D) to identify 2001 genes including 1642 known genes and 359 novel genes that regulate hair barely any report that SSP1, CCL2 and APP hub genes could regulate hair growth. The hub genes could be potential biomarkers for the diagnosis and treatment of hair loss. In the future, we attempt to control the expression of hub genes in vitro and in vivo to observe the changes in the morphology of DPCs. Therefore, further research is necessary to verify our findings, which are a limitation to our current study.
In conclusion, the present study identified 2001 DEGs, 10 hub genes and related signalling pathways that were associated with functional recovery of dermal papilla cells in 3D culture. The genes were identified as mechanism of DPCs' regeneration in 3D condition and potential biomarkers to control hair growth.

ACK N OWLED G EM ENTS
The authors thank Professor Jinmin Zhao from Guangxi Key Laboratory of Regenerative Medicine for support of laboratory equipment.

CO N FLI C T O F I NTE R E S T S
There are no conflicts to declare.