Transcriptomic profiles of human HaCaT cells in response to angelica polysaccharide

Abstract

The aim of this study was to explore the anti-psoriatic effect and potential mechanism of Angelica polysaccharide (APS) on an in vitro HaCaT cell model. MTS assay was performed to determine whether APS has the ability to inhibit the proliferation of HaCaT cells. RNA-sequencing (RNA-seq) was performed to investigate the underlying mechanism of APS. Quantitative real-time PCR (qRT-PCR) was used to verify the accuracy of RNA-seq data. Our MTS assay results demonstrated that APS time- and concentration-dependently inhibits the proliferation of HaCaT cells. The anti-proliferation property of APS suggests that APS may have anti-psoriatic effect. In the RNA-seq part, comparison between the CK group (i.e., Control group) and ASP groups revealed dramatic differences [468 differentially expressed genes (DEGs) for CK group vs. ASP50 group; 563 DEGs for CK group vs. ASP100 group; 532 DEGs for CK group vs. ASP200 group]. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrich analysis performed on all DEGs failed to find any significant enriched GO terms or KEGG pathways to explain the anti-proliferative effect of APS. All DEGs were then classified into 20 expression profiles by trend analysis. Interestingly, cell proliferation-related GO terms were mostly dispersed in the profile 2 and 17. DEGs enriched in these terms were then analyzed. After literature retrieval, DEGs such as SERPINE1, SMAD6, CTGF, and TGF-β were suspected to closely relevant to the anti-proliferation effect of APS. qRT-PCR results showed similar expression trend with RNA-seq data for 8 out of 10 genes, indicating our sequence data are reliable.

Electronic supplementary material

The online version of this article (10.1007/s13205-019-1828-z) contains supplementary material, which is available to authorized users.

Introduction

Psoriasis is a chronic inflammatory skin disease of unknown etiology, characterized by keratinocyte hyper-proliferation, presence of inflammatory cell infiltrate and neovascularization. Psoriasis affects about 2% of the general population and imposes substantial negative effects on patients’ quality of life because of its annoying symptoms such as pruritus, pain, redness, scaling, flaking and bleeding (Qu et al. 2014). Currently, there are many treatment options for psoriasis (Kushwaha 2017); however, none of them offers the opportunity to cure the disease. Conventional western medications, such as methotrexate, ciclosporin, and retinoids, have been widely used for treatment of psoriasis; unfortunately, they are limited for long-term use due to their potential toxic and side effects (Li et al. 2017; Keseroglu and Gönül 2014). Therefore, looking for safe and effective treatment methods for psoriasis is a critical public concern worldwide.

Traditional Chinese medicine (TCM) formulas have been widely used for treatment of psoriasis in China (Zhang et al. 2014; Farahnik et al. 2017); however, owing to their inherent characteristics, such as complex ingredients and unclear mechanism, the clinical use of TCM was questioned by lots of people. On the contrary, active components from TCM usually have less composition than formulas, making them more acceptable to users. Angelica polysaccharide (APS) is one of the most major active components isolated from the root of Angelica sinensis. Recent evidences suggested that APS may have therapeutic potential for psoriasis. For instance, Jing et al. demonstrated APS can increase apoptosis in psoriasis-like lesions of guinea pigs (Jing and Sheng 2006). Wang et al. reported that APS can alleviate LPS-induced HaCaT cell inflammatory injury (Wang et al. 2019).

The immortalized human keratinocyte cell line HaCaT was commonly used as an in vitro model to study psoriasis (Liang et al. 2017). In the present study, we used different concentrations of APS to detect the effect of APS on the proliferation of HaCaT cells; after then, RNA-sequencing (RNA-seq) was performed to further investigate the underlying mechanism of APS.

Materials and methods

Cell proliferation assay

HaCaT cells, purchased from ATCC (Teddington, UK), were cultured in minimum Eagle’s medium (Invitrogen, Carlsbad, CA, USA) containing 10% fetal bovine serum and 1% penicillin–streptomycin at 37 °C in a 5% CO₂ humidified environment. APS (yellowish amorphous powder, purity ≥ 95%), purchased from Ci Yuan Biotechnology Co., Ltd (Xi’an, Shanxi, China) and dissolved in phosphate buffered saline (PBS), was added to HaCaT cells at final concentrations of 50 mg/L (ASP50 group), 100 mg/L (ASP100 group) and 200 mg/L (ASP200 group), respectively. Same volume of PBS was added in the Control group (CK group). 12 h, 24 h and 36 h after stimulation, MTS assay [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt] was performed according to the manufacturer’s instructions to determine the effect of APS on the proliferation of HaCaT cells. The optical density (OD) value for each sample, determined in triplicate, was recorded at 490 nm using the Multiscan MK3 spectrophotometric microplate reader (Thermo Fisher Scientific, Waltham, MA, USA).

Sample preparation and RNA-seq

HaCaT cells were cultured as the above-mentioned method and stimulated with PBS and APS solutions (50 mg/L, 100 mg/L and 200 mg/L) for 36 h, respectively. After then, cells were collected and cryopreserved for further studies including RNA-seq and quantitative real-time PCR (qRT-PCR). RNA extraction, library construction and sequencing were performed by the Guangzhou Gene Denovo Co. (Guangzhou, China) as previously described (Huang et al. 2017).

RNA-seq data analysis

RNA-seq data analysis was performed as previously described (Huang et al. 2017). Briefly, raw reads were filtered using the FASTX Toolkit to remove reads that containing adaptors, reads that containing more than 10% of ambiguous nucleotides, and reads with low-quality that containing > 50% of bases with Q ≤ 20. Furthermore, clean reads were mapped to the ribosomal RNA (rRNA) database to remove rRNA mapped reads using the Bowtie2 tool. The remaining clean reads were then mapped to the reference genome by the TopHat2 tool. The reconstruction of transcripts was carried out with the Cufflinks software, together with the TopHat2 tool. The gene expression level was normalized using the fragments per kilobase of transcript per million mapped reads (FPKM) method. Genes with fold-change ≥ 2 and a false discovery rate (FDR) of < 0.05 were identified as significant differentially expressed genes (DEGs) using the edgeR package. The Short Time-series Expression Miner (STEM, version 1.2.2b) software was used to cluster DEGs. Profiles with P ≤ 0.05 were considered significantly enriched. All DEGs and DEGs in the significant trend were then subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis.

Quantitative real-time PCR

To further confirm the results of RNA-seq, ten genes were selected for qRT-PCR analysis. Primers used for all tested genes are listed in Table 1. qRT-PCR was performed using the ABI Step One Plus Detection system (Applied Biosystems, Waltham, MA, USA) with the following condition: denaturation at 95 °C for 90 s, followed by 40 cycles of 95 °C for 5 s, 60 °C for 15 s and 72 °C for 20 s. GAPDH was used as the reference gene. Relative gene expression level was calculated using the comparative 2^−△△Ct method.

Table 1.

Primers used for quantitative real-time PCR

Genes	Sequence
HIST1H2BD
Forward	CGGCATCTCTTCCAAGGCAA
Reverse	AGAGCCTTTGGGATTAGGTGTA
HIST1H2BC
Forward	CCCGACACTGGCATCTCTT
Reverse	TGGAGCTGGTGTACTTGGTGA
CD74
Forward	GACAGTCACCTCCCAGAACC
Reverse	CCATACTTGGTGGCATTCTG
HIST1H2AE
Forward	GTGCTGGAATATCTGACGGC
Reverse	AGTAATCATTTCACTTGCCCTTG
HIST1H3H
Forward	CTACCAGAAGTCCACCGAGC
Reverse	ATGATAGTCACCCGCTTGGC
CSF2
Forward	TAAAGTTCTCTGGAGGATGTGG
Reverse	TACTCAGGTTCAGGAGACGC
HSPA1A
Forward	GGGCCTTTCCAAGATTGCTG
Reverse	AGAAATAGTCGTAAGATGGCAGTA
PLAT
Forward	CCTGTCAAAAGTTGCAGCGAG
Reverse	CAGCACTTCCCAGCAAATCC
IGFBP3
Forward	TTTGTGACTTAGGCGGCTGT
Reverse	TCCAATAGTCCCCAAGCAGTA
NUPR1
Forward	GGAAAGGTCGCACCAAGA
Reverse	GCCTCATCTCCAGCTCTGTC
GAPDH
Forward	CGGGAAACTGTGGCGTGA
Reverse	GTGCTCAGTGTAGCCCAGGAT

HIST1H2BD histone cluster 1 H2B family member D, HIST1H2BC histone cluster 1 H2B family member C, CD74 cluster of differentiation 74, HIST1H2AE histone cluster 1 H2A family member E, HIST1H3H histone cluster 1 H3 family member H, CSF2 colony stimulating factor 2, HSPA1A heat shock protein family A (Hsp70) member 1A, PLAT plasminogen activator, tissue type, IGFBP3 insulin like growth factor binding protein 3, NUPR1 nuclear protein 1, transcriptional regulator, GAPDH glyceraldehyde-3-phosphate dehydrogenase

Statistical analysis

MTS and qRT-PCR results are presented as the mean ± SD. Groups were compared with one-way analysis of variance (One-way ANOVA) using the SPSS 19.0 software (SPSS Inc., Chicago, IL, USA). Differences were considered to be significant when P value < 0.05.

Results

APS time- and concentration-dependently inhibits the proliferation of HaCaT cells

Figure 1 shows the MTS results of different concentrations of APS on cultured HaCaT cells. From the results we can see that, APS exhibited significant anti-proliferative effect on HaCaT cells and this effect was time- and concentration-dependent. The inhibitory effect was most obvious at 36 h after stimulation; therefore, in the present study, this time point was chosen for the following experiments to explore the underlying mechanism of APS.

Fig. 1.

APS time- and concentration-dependently inhibits the proliferation of HaCaT cells. Cell proliferation was determined by MTS assay. Data are presented as mean ± SD. **P < 0.01, compared with the CK group; ^△△P < 0.01, compared with the ASP50 group; ^★★P < 0.01, compared with the ASP100 group

RNA-Seq data evaluation

The samples in the CK group and three APS treated groups were named CK-1, CK-2, CK-3, APS50-1, APS50-2, APS50-3, APS100-1, APS100-2, APS100-3, APS200-1, APS200-2, APS200-3, respectively. The major sequencing information for the 12 samples is summarized in Table 2. After filtering the raw reads, over 50 million clean reads were obtained per sample. Over 89% of clean reads were mapped to the reference genome. More than 14,000 genes containing over 800 new genes were obtained for each sample. Results from principal component analysis (PCA) and Pearson’s correlation coefficient analysis indicated that our RNA-Seq data were suitable for further analysis (Figure not shown).

Table 2.

Major sequencing information for each sample

Group	Sample	Clean reads	Mapped ratio (%)	All genes	New genes	Known genes
CK	CK-1	56,457,592	89.0	14,879	815	14,064
	CK-2	50,463,680	89.2	14,816	809	14,007
	CK-3	57,959,838	89.2	14,914	825	14,089
APS50	APS50-1	60,694,838	90.3	14,672	820	13,852
	APS50-2	53,102,402	90.3	14,615	801	13,814
	APS50-3	57,190,644	90.8	14,622	818	13,804
APS100	APS100-1	56,230,804	90.0	14,695	816	13,879
	APS100-2	56,218,244	90.3	14,649	813	13,836
	APS100-3	66,590,600	90.2	14,767	821	13,946
APS200	APS200-1	53,346,416	90.4	14,618	809	13,809
	APS200-2	54,468,214	90.3	14,632	817	13,815
	APS200-3	54,298,784	90.5	14,701	807	13,894

CK, Control; ASP50, 50 mg/L ASP; ASP100, 100 mg/L ASP; ASP200, 200 mg/L ASP

Differentially expressed genes in all test groups

To compare the differential expression levels of genes between the CK group and all APS groups, up and down-regulated gene numbers were calculated (Fig. 2a). A total of 468 DEGs, including 52 up-regulated and 416 down-regulated genes, were identified in the comparison between the CK group and the APS50 group. Similarly, 563 DEGs, including 68 up-regulated and 495 down-regulated genes, were identified in the comparison between the CK group and the APS100 group. There were 532 DEGs, of which 49 were up-regulated and 483 were down-regulated, in the comparison between the CK group and the APS200 group. All DEGs are summarized in the supplemental Table S1. On the basis of the gene expression, a venn diagram was constructed and a total of 891 DEGs were identified in the three comparisons (Fig. 2b).

Fig. 2.

Analysis of DEGs between test groups. a DEGs’ distribution between groups; b Venn diagram constructed with DEGs generated from the comparisons between the CK group and ASP groups

GO and KEGG pathway enrichment analysis of all differentially expressed genes

GO and KEGG pathway enrichment analysis were performed for all DEGs. In the GO enrichment analysis, DEGs were assigned into three categories: cellular component, molecular function, and biological process. The GO terms “cellular process”, “single-organism process” and “metabolic process” were most frequently enriched in biological process; “binding”, “catalytic activity” and “molecular transducer activity” represented the major proportion of the molecular function categories; while “cell”, “cell part” and “organelle” were most enriched in the cellular component. Among these subcategories, thirty five terms were significantly enriched, including 28 for biological process, two for molecular function and five for cellular component. Interestingly, some significantly enriched GO terms in the biological process were related to immune response, such as “mucosal immune response” (GO:0002385), “organ or tissue-specific immune response” (GO:0002251) and “immune response” (GO:0006955). Other GO terms, such as “regulation of cytokine production” (GO:0001817) and “negative regulation of cytokine production” (GO:0001818), were related to the process of cytokine production. Genes enriched in these terms may contribute to the immunomodulatory effect of APS (Supplemental Table S2).

To further investigate the function of these DEGs, KEGG enrichment analysis was performed, and pathways with Q value ≤ 0.05 were considered significantly enriched. Overall, 281 DEGs were assigned to KEGG pathways and 7 KEGG pathways were significantly enriched, including “Systemic lupus erythematosus” (ko05322), “Alcoholism” (ko05034), “Antigen processing and presentation” (ko04612), “Staphylococcus aureus infection” (ko05150), “Steroid hormone biosynthesis” (ko00140), “Inflammatory bowel disease (IBD)” (ko05321) and “Phagosome” (ko04145).

Trend analysis and gene expression clustering

Trend analysis and clustering of the 891 DEGs was performed using the STEM software, and 20 profiles were obtained, including seven significant expression profiles (profiles 2, 1, 0, 17, 4, 3, and 7) (Fig. 3).

Fig. 3.

Trend profiles generated by STEM based on gene numbers. Colored profiles represent significant enrichment profiles with Q value ≤ 0.05. *ENSG00000070748 was treated as half gene and clustered in both profile two and five for it has same correlation coefficient with the two profiles

Profile 2 had 386 DEGs that were rapidly down-regulated after APS stimulation and maintained at a constant level in the three APS groups. GO enrichment data showed that 27 terms were significantly enriched, including 23 for biological process and four for cellular component (Supplemental Table S3).

Profile 17 included 49 DEGs that were rapidly up-regulated after APS treatment and maintained at a constant level in the three APS groups, which was opposite to the Profile 2. In this profile, 113 terms were significantly enriched, including 109 for biological process, 2 for cellular component and 2 for molecular function (Supplemental Table S3). Specially, some significantly enriched GO terms in this profile were related to cell growth and death, such as “positive regulation of programmed cell death” (GO:0043068), “positive regulation of cell death” (GO:0010942), “regulation of programmed cell death” (GO:0043067), “regulation of cell death” (GO:0010941), “cell proliferation” (GO:0008283), “positive regulation of cell proliferation” (GO:0008284), “regulation of cell proliferation” (GO:0042127) and “negative regulation of cell proliferation” (GO:0008285).

Profile 1 included 109 DEGs that were constantly decreased in the APS 50 and APS 100 groups (vs. the CK group) and slightly increased in the APS 200 group (vs. the APS 100 group). Profile 0 had 89 DEGs that showed decreased expression as the APS increased. 46 and 32 DEGs were included in profile 4 and 7, respectively. GO enrichment data showed that no term was significantly enriched for these four profiles.

In profile 3, 36 DEGs were included; only one term, namely, flavonoid metabolic process, was significantly enriched in this profile.

To better know the function of DEGs, KEGG enrichment analysis was performed in all seven significant profiles. Our results showed that seven pathways were significantly enriched, including five for profile 2 and two for profile 7. Five significantly enriched pathways in profile 2 included “Systemic lupus erythematosus” (ko05322), “Alcoholism” (ko05034), “Staphylococcus aureus infection” (ko05150), “Antigen processing and presentation” (ko04612) and “Inflammatiory bowel disease (IBD)” (ko05321). “Cytokine-cytokine receptor interaction” (ko04060) and “Intestinal immune network for IgA production” (ko04672) were significantly enriched in profile 7. No significantly enriched pathway was found in profile 1, 0, 17, 4 and 3.

qRT-PCR validation of DEGs

Ten genes from RNA-seq were randomly chosen to further be verified by qRT-PCR method. qRT-PCR results (Fig. 4a, b) and the FPKM data from RNA-seq (Fig. 4c, d) are shown in Fig. 4. From the figure we can see that, except for HIST1H3H and IGFBP3, other genes revealed similar expression trend in the qRT-PCR and RNA-seq results, indicating our sequence data are reliable.

Fig. 4.

Comparison of the qRT-PCR results and FPKM values from RNA-Seq. a, b qRT-PCR results for 10 DEGs; c, d FPKM values for 10 DEGs

Discussion

Traditional Chinese medicine (TCM) is a complementary and alternative modality for treating psoriasis. TCM formulas, such as “Wen-qing-yin”, “Jia-wei-xiao-yao wan”, “Dang-gui-yin-zi” and “Xue-fu-zhu-yu-tang”, are commonly used in clinical treatment of psoriasis, especially by Chinese people (Weng et al. 2016). Angelica sinensis is one of the most commonly used drug in these formulas. APS is one of important active components isolated from Angelica sinensis. In a previous study, Jing et al. found that APS have a therapeutic effect on the psoriasis-like lesions in a guinea pigs model induced by propranolol ointment (Jing and Sheng 2006). In another study performed by Jing et al. authors found APS can decrease the expression of proliferating cell nuclear antigen (PCNA), one important protein related with cell proliferation and differentiation, in the psoriasis-like lesions (Jing et al. 2008). Our previous study reported that APS can inhibit proliferation of HaCaT cells by delaying cell cycle progression from G0/G1 phase to S phase (Chunshui et al. 2010). Based on these evidences, we suspect that APS may be useful for psoriasis treatment.

It is well known that psoriasis is a disease of epidermal hyperplasia resulted from the abnormal differentiation and hyper-proliferation of keratinocytes; therefore, many therapies had focused on the restrain of the hyper-proliferation of keratinocytes (Li et al. 2012; Tse et al. 2006). Natural products with anti-proliferative effects on HaCaT cells are potentially useful in the treatment of psoriasis (Zhang et al. 2017; Huang et al. 2019). In the present study, we confirmed that APS can time- and concentration-dependently inhibits the proliferation of HaCaT cells, which confirmed that APS have anti-psoriatic potential. To better know the underlying mechanism of this action, GO and KEGG enrichment analysis were performed on all DEGs; however, the results cannot plainly used to explain the anti-proliferative effect of APS. After then, trend analysis was performed on all 891 DEGs. Notablely, cell proliferation-related GO terms were mostly dispersed in the profile 2 and 17, including “regulation of MAPK cascade” (GO:0043408), “positive regulation of MAPK cascade” (GO:0043410), “cell proliferation” (GO:0008283), “positive regulation of cell proliferation” (GO:0008284), “regulation of cell proliferation” (GO:0042127) and “negative regulation of cell proliferation” (GO:0008285). Genes enriched in these terms are summarized in the supplemental table S4. After reviewing published literatures, we found that DEGs such as NOTCH3 (Ota et al. 2014), SERPINE1 (Man et al. 2015), GLI1 (Endo et al. 2006), SMAD6 (Yu et al. 2009), TGFB2 (Jiang et al. 2017), CYR61 (Wu et al. 2017) and FOSL1 (Sobolev et al. 2011) were previously reported in psoriasis studies. Among them, SERPINE1, SMAD6 and CTGF attracted our most concerns. SERPINE1 (also called PAI-1), an important EMT marker, was found significantly increased in the psoriatic epidermal keratinocytes (Man et al. 2015). In another study performed by Nielsen et al. authors found that the plasma level of PAI-1 was significantly elevated in psoriasis patients and significantly decreased during treatment (Nielsen et al. 2002). These evidences indicated that SERPINE1 may play a role in the pathogenesis of psoriasis. The Smad family includes eight different SMAD, namely SMAD1 to SMAD8. SMAD6 and 7, two inhibitory SMADs, represent a negative feedback loop inhibiting the TGF β-SMAD signaling pathway (Di Fusco et al. 2017). Considering that TGF-β is an important cytokine that negatively regulates keratinocyte proliferation and SMAD7 was proved positively regulates keratinocyte proliferation in psoriasis, we speculate that SMAD6 may has similar function with SMAD7, which making it a potential target for psoriasis treatment. Just like SERPINE1 and SMAD7, CTGF is a target gene of TGF-β (Saito et al. 2013). In a study performed by Hayakawa et al. authors used imiquimod to induce a murine psoriatic dermatitis model and found that increase of CTGF showed a tendency to suppress the psoriatic dermatitis through inhibition of suprabasal cells proliferation and macrophage infiltration in the skin; on the contrary, CTGF inhibition using an anti-CTGF antibody slightly worsened the dermatitis (Hayakawa et al. 2018). In the present study, SERPINE1 and SMAD6 were significantly decreased after APS treatment; while, CTGF was significantly increased after APS treatment. These genes, together with TGF-β, which was increased after APS treatment in our RNA-seq data, may closely relevant to the anti-proliferative effect of APS. However, this study has limitations. The most important one is lack of research depth. In the near future, cell and animal experiments are planned to take place to further confirm the underlying mechanism of APS in treating psoriasis.

Conclusions

The present study confirmed that APS can time- and concentration-dependently inhibits the proliferation of HaCaT cells in vitro; the anti-proliferation property of APS suggests that APS may have anti-psoriatic effect. SERPINE1, SMAD6, CTGF, and TGF-β may closely relevant to the anti-proliferation effect of APS.

Accession numbers

RNA-seq data of the present study has been deposited into the SRA database under accession number of PRJNA552662.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Compliance with ethical standards

Conflict of interest

The author declares that they have no competing interests.