Identification of Genes Differentially Expressed Between Midgestational and Postnatal Mouse Skin Through Suppression Subtractive Hybridization

Objective. The aim of this study was to identify differentially expressed genes between the skin of midgestational (embryonic day [ED] 14) and postnatal mice that might be involved in the wound healing process. Materials and Methods. BALB/c mice with dated pregnancy underwent laparotomy and hysterotomy on ED14, and skin from 20 fetuses were harvested. Full-thickness dorsal skin was also harvested from 30 three-month-old postnatal mice. Total ribonucleic acid (RNA) and mRNA were purified side-by-side from the harvested skins. Differentially expressed cDNA fragments were isolated by suppression subtractive hybridization (SSH). The cDNA fragments were subsequently cloned, sequenced, and identified through a Nucelotide Basic Local Alignment Search Tool (BLASTN) search. Results. Twenty differentially expressed transcripts were identified, including 18 uniquely expressed in fetal mice skin and 2 from postnatal mice skin. The known genes identified include RPS29, Nedd5, EndoA, TIEG, and eEF-1α, which may play an important role in scarless wound healing of embryonic mice. Two novel genes uniquely expressed in fetal mouse skin were also identified. Conclusion. Through SSH, 20 differentially expressed genes in the skin of midgestational and postnatal mice were identified. Two novel genes were identified that were uniquely expressed in the midgestational mouse skin. The authors suggest these genes might be involved in the scarless wound healing process.

Early to midgestational mammalian skin heals rapidly from wounds without scar formation, while adult wounds heal with fibrosis and scars. Previous studies of postnatal and fetal (≤ embryonic day [ED]  15) wound healing in mice reveal many differences in inflammatory response.¹ Many genes are differentially expressed between the 2 processes, including cellular mediators,² cytokines,³ growth factors,⁴ extracellular matrix modulators,⁵ stem cell specific genes,⁶ and many others.^7-9 Despite these investigations, the mechanism of fetal wound healing remains largely unknown and the identification of new genes may help in a better understanding of the scarless wound healing process. 

Suppression subtractive hybridization (SSH) coupled with polymerase chain reaction (PCR) is an accurate and comprehensive transcriptome approach that allows the discovery of differentially expressed genes between 2 cell types or tissues.¹⁰ The effectiveness of SSH in enriching differentially expressed genes has been optimized and utilized in many studies.^10,11 

In the current study, to identify novel genes that might play a potential role in scarless wound healing, the authors used SSH to identify differentially expressed genes between the skin of midgestational (ED14) and postnatal mice, corresponding to scarless and scarring phenotype, respectively.

Animals and skin harvest. All animal protocols were reviewed and approved by the Institutional Committee on Laboratory Animals’ Care and Use of Weifang Medical College, Shandong, China. Twenty E14 mice (from Central Animal Laboratory, Weifang Medical College, Shandong Province, China) and 30 three-month-old postnatal mice were sacrificed. For fetuses, a 0.5 cm x 0.5 cm section of skin was excised from the dorsal side with a No. 11 blade. Care was taken to restrict the incision to integument only, avoiding taking the subcutaneous tissue. For postnatal mice, with the mice anesthetized, a 1 cm x 1 cm section of dorsal skin was excised using a No. 11 blade. Care was also taken to avoid subcutaneous tissue. All incisions were excised from similar locations from each fetus or mouse. The skin samples were flash frozen and stored in liquid nitrogen immediately after harvest.

Total ribonucleic acid preparation and suppression subtractive hybridization. Total RNA and Poly mRNA were isolated separately from the 2 samples using an RNeasy Mini Kit (QIAGEN Sciences, Germantown, MD) and Oligotex Direct mRNA Mini Kit (QIAGEN Lake Constance GmbH, Stockach, Germany). Ribonucleic acid was measured using an Ultrospec 2000UV spectrophotometer (Pharmacia Biotech, Uppsala, Sweden) which measured ultraviolet absorbance at wavelengths 260 and 280. A ratio between 1.7 < A260/A280 < 2.0 indicates the RNA has good purity. A ratio lower than 1.7 indicates the sample might be contaminated with protein or phenol, and a ratio above 2.0 indicate that the sample might be polluted by isorhodanic acid. In this case, an A260 (0.371)/A280 (0.708) of 1.9 indicated the integrity of total RNA wasn’t affected. 

The RNA integrity was further verified by electrophoresis on 1% formaldehyde denaturing agarose gel and visualized by ethidium bromide staining. The cDNA was reverse-transcribed using a ReverAid First Strand cDNA Synthesis Kit (MBI, Ontario, Canada); using the following oligo-dT anchor primer: 5’-AAGCAGTGGTAACAACGCAGAGTACT-3’ and 5’-AAGCAGTGGTAACAACGCAGAGTGCGGCCGCGGG-3’ according to the instruction of the PCR-Select cDNA Substraction Kit (Clontech Laboratories, Inc, Mountain View, CA). First strand cDNA fragments were digested with RsaI, which was part of the cDNA substraction kit, and purified by ethanol precipitation. The RsaI-digested cDNA fragments from the 2 types of skin tissues were divided into 2 portions, and each was ligated to a cDNA adaptor as tester cDNA at 16°C overnight using Fast-Link DNA Ligation and Screening Kit (Epicentre Technologies Corp, Madison, WI). The adaptor sequences were: 

1) CTAATACGACTCACTATAGGGCTCGAGCGGCCGCCCGGGCAGGT, and 

2) CTAATACGACTCACTATAGGGCAGCGTGGTCGCGGGCGAGGT. 

Excess driver cDNA from the skin tissies of the postnatal mice was added to each tester cDNA, and the samples were heat-denatured and then allowed to anneal for the first hybridization at 68°C for 8 hours. The 2 samples from the first hybridization were mixed together, denatured driver cDNA was added to further enrich differentially expressed sequences, and the mixture was incubated at 68°C overnight to complete the second hybridization. Differentially expressed cDNA fragments were diluted and were amplified by a 2-step PCR with the following condition: 25s at 94°C, 20s 66°C, 30s 72°C, 27 cycles. The primer was CTAATACGACTCACTATAGGGC. 

Complementary DNA cloning and identification. The PCR products from the above procedures were purified by QIAquick PCR Purification Kit (QIAGEN Lake Constance GmbH, Stockach, Germany). The nested PCR products were cloned into a standard plasmid vector (T-Vector PCR Product Cloning Kit, Shanghai Sangon Biological Engineering Technology & Services Co Ltd,  Shanghai, China). The ligation products were transform into Escherichia coli JM109 (Shanghai Sangon Biological Engineering Technology & Services Co Ltd, Shanghai, China) and positive clones were selected with 5-bromo-4-chloro-3-indoyl-β-D-galactopyranoside (X-gal) on medium containing isopropyl-β-D-thiogalactopyranoside (IPTG) by white-blue selection. White colonies were selected and plasmid DNA was isolated from the bacterial using a QIAGEN Plasmid Mini Kit (QIAGEN Lake Constance GmbH, Stockach, Germany). 

An ABI Prism 7700 sequence detector (PE Biosystems, Foster City, CA) was used to perform DNA sequencing. The sequences were identified using BLASTN and BLASTX (translated cDNA sequences) program. 

A total of 37 clones from the SSH cDNA library were selected for further analysis using bioinformatics tools. By searching for sequence homology in NCBI GenBank database with BLASTN programs, 20 cDNA sequences obtained homologous sequences with significant matches (E-values < 1×10-4) in the SSH library (Table 1). 

All the cDNA fragments were from 300 base pair (bp) to 3 kilobase (kb), and 17 of the 20 pieces of cDNA fragments are between 300 bp to 700 bp. Among the 20 pieces of specific cDNA fragments isolated and identified, 18 pieces were expressed specifically in the fetal skin of BALB/c mice, and the other 2 pieces in the skin of adult BALB/c mice (Table 1). Of the 20 genes identified in the subtracted library, 18 were known genes, including RPs29, eEF-1α, TIEG, RPs15, Nedd5 (SEPT2), CDCrel-1 (SEPT5), and EndoA (Table 1). Two novel genes were identified in this study: F32-5 and F4-4. The authors searched for the highly homologous ESTs of F32-5 and F4-4 by the electronic hybridization walking method and obtained multiplex expressed sequence tags. The Open Reading Frame (ORF) of F32-5 and F4-4 were confirmed by the ORF Finder (Figures 1A, 1B). Specific primers were designed to amplify the 2 genes’ ORF cDNA by PCR.

F32-5 Forward primer: 5’— GCTCTAGACACAATGCACATCAATGTGGA— 3’; Reverse prime: 5’—GCGGTACCTACCTCAAGACACTTTGCAGT— 3’. F4-4 Forward primer: 5’ — GCTCTAGAGATGCCCTCCAGCAAGA— 3’; Reverse primer:  5’—GCGGTACCTCTCACACTGTCTCCATGT— 3’.

In this study, SSH was used to identify candidate genes that showed differential expression between the skin of midgestational (ED14) and postnatal mice. The SSH technique can increase the screening efficiency of low abundance mRNA by 2-step PCR.^10,11 The false positive rate in this study was further decreased by a 2-round subtractive hybridization in SSH.^10,11 

Cutaneous scarless wound healing in mammalian fetuses has been a research focus for the last several decades.^12,13 In mice, cutaneous wounds generated early in development (≤ ED15) heal scarlessly, while wounds generated in postnatal skin form scars.¹⁴ Most previous research focused on the skin wound of murine fetus at ED15 or later in gestation, or in adult specimens.^7,14,15 In this study, the authors hypothesized there are many specially expressed genes involved in the wound healing process that are responsible for the age-related differences in wound healing outcomes between the skin of midgestational and postnatal mice. By comparing the abundance of specific cDNA fragments between the skin of embryonic and postnatal mice, it is possible to uncover new genes that might play a major role in the molecular mechanisms of scarless wound healing.

Twenty predicted genes were obtained, among which 18 genes are from the skin of the ED14 mice and 2 from the skin of 3-month-old mice. The known genes in the subtracted library include RPS29, eEF-1α, TIEG, RPS15, Nedd5 (SEPT2), CDCrel-1 (SEPT5), and EndoA (Table 1). It has been proved that RPS29,¹⁶ Nedd5 (SEPT2), EndoA,¹⁷ and eEF-1α18 are related to the proliferation and apoptosis of fibroblasts and epidermal cells. RPS15¹⁹ has been found to be activated in various tumors, such as insulinomas, esophageal cancers, and colon cancers. A tissue-specific transcription factor, TIEG,²⁰ can regulate the growth and proliferation of a variety of cell types. Some genes of the Septin gene family such as Septin7 and Septin9 and TIEG were found to be activated in tumors. It is not certain whether Nedd5 (SEPT2) is activated in tumors. The Septin gene family plays a very important role in the cell division. Nedd5 (SEPT2)²¹ is a novel cytoskeletal component interacting with actin-based structures. In interphase and postmitotic cells, Nedd5 localizes to fibrous or granular structures depending on the growth state of the cell. The Cdcrel-1 (SEPT5) gene, which was expressed primarily in the brain and thought to be involved in platelet and neuron exocytosis, binds to SEPT8 (KIAA0202).²² The fact that the authors’ research cDNA library consists of many genes known to play a role in the wound healing process confirmed the authors’ method is valid. However, the role of the other 11 genes in wound healing have not been reported and this would be an interesting topic for further study. 

Complementary DNA F32-5 and F4-4, which are uniquely expressed in mouse embryo skin, have a high homology with precollagen genes and keratin genes respectively. Previous studies show the synthesis and secretion of collagen and keratin²³ have an important role in the wound healing process. Another study shows type III collagen is expressed and temporally regulated during prenatal and postnatal rat skin morphogenesis.²⁴ Therefore, it is highly possible that the involvement of F32-5 and F4-4 in the restoration of wounded murine skin results in scarless wound healing.

The authors identified 20 genes that are uniquely expressed either in midgestational (ED14) or postnatal mouse skin using an improved SSH. This study identified a number of known genes involved in the wound healing process and uncovered 2 novel genes that might play important roles in the scarless healing process. Further research is needed to investigate the functions of these genes by transgenic expression in cultured skin fibroblasts and keratinocytes.

The authors would like to thank the Shandong Province Natural Science Foundation of China for its financial support (Grant No. z2000c03).

From the Department of Aesthetic, Plastic, and Burn Surgery, Shandong Provincial Hospital, Shandong University, #324, Jingwu Road, Jinan City, Shandong, PR China; Plastic Surgery Research Center, Weifang Medical College, #288 Shengli East Road Weifang, Weifang City, Shandong Province, PR China; and Department of Aesthetic, Plastic, and Burn Surgery, Yuhuangding Hospital, #20 Yuhuangding East Road Yantai, Yantai City, Shandong Province, PR China=

Address correspondence to:
Ran Huo, PhD
Department of Aesthetic, Plastic, and Burn Surgery
Shandong Provincial Hospital, Shandong University
#324, Jingwu Road, Jinan City, Shandong, 250021 PR China.
wangpeng1765@126.com 

Disclosure: The authors disclose no financial or other conflicts of interest.