Content uploaded by Chengcheng Deng
Author content
All content in this area was uploaded by Chengcheng Deng on Jan 29, 2023
Content may be subject to copyright.
Br J Dermatol 2023; 188 : 84–93
https://doi.org/10.1093/bjd/ljac025
Advance access publication date: 5 November 20 22 Translational Research
Accepted: 16 September 2022
© The Author (s) 2022. Published by Ox ford University Press on behalf of British A ssociation of Dermatologists. All rights reserved. For permissions,
please e-mail: journals.permissions@oup.com
Risk single-nucleotide polymorphism-mediated
enhancer–promoter interaction drives keloids through
long noncoding RNA down expressed in keloids
Cheng-Cheng Deng ,1 Li-Xue Zhang,1 Xue-Yan Xu,1 Ding-Heng Zhu,1 Qing Cheng,1
ShufengMa,2,3,4 Zhili Rong 1, 2,3 and Bin Yang 1
1Dermatology Hospital, Southern Medical University, Guangzhou, China
2Cancer Research Institute, School of Basic Medical Sciences, Southern Medical University, Guangzhou, China
3State Key Laboratory of Organ Failure Research, National Clinical Research Center of Kidney Disease, Key Laborator y of Organ Failure
Research (Ministry of Education), Southern Medical University, Guangzhou, China
4Department of Nephrology, Shenzhen Hospital, Southern Medical University, Shenzhen, China
Correspondence: Bin Yang. Email: yangbin1@smu.edu.cn
Abstract
Background Keloids represent one extreme of aberrant dermal wound healing and are characterized by fibroblast hyperproliferation and
excessive deposition of extracellular matrix. Genetics is a major factor for predisposition to keloids and genome-wide association study has
identified a single-nucleotide polymorphism (SNP) rs873549 at 1q41 as a susceptibility locus. The SNP rs87354 9, and the SNPs in strong link-
age disequilibrium (LD) with rs873549, may be involved in keloid development. However, the functional significance of these SNPs in keloid
pathogenesis remains elusive.
Objectives To investigate the function and mechanism of SNP rs87354 9 and the SNPs in strong LD with rs8735 49 in keloids.
Methods SNPs in strong LD with rs873549 were analysed using Haploview. The expression levels of the genes near the susceptibility locus
were analysed using quantitative real-time polymerase chain reaction. The interaction between rs1348270-containing enhancer and the long
noncoding RNA down expressed in keloids (DEIK) (formerly RP11-400N13.1) promoter in fibroblasts was investigated using chromosome
conformation capture. The enhancer activit y of the rs1348270 locus was evaluated using luciferase reporter assay. Knockdown experiments
were used to explore the function of DEIK in keloids. RNA-Seq was performed to investigate the mechanism by which DEIK regulates the
expression of collagens POSTN and COMP.
Results rs1348270, an enhancer-located SNP in strong LD with rs8735 49, mediated looping with the promoter of DEIK. The risk variant was
associated with decreased enhancer–promoter interaction and DEIK down- expression in keloids. Mechanistically, downregulation of DEIK
increased the expression of collagens POSTN and COMP through upregulating BMP2. Furthermore, correlation analysis revealed that DEIK
expression was inversely correlated with BMP2, POSTN and COMP expression in both keloid and normal fibroblasts.
Conclusions Our findings suggest that the risk variant rs1348270 is located in an enhancer and is associated with the downregulation of
DEIK in keloids, and that downregulation of DEIK increases the expression of collagens POSTN and COMP through BMP2 in keloid fibroblasts.
These findings will help to provide a more thorough understanding of the role played by genetic factors in keloid development and may lead
to new strategies for screening and therapy in keloid-susceptible populations.
Linked Article: Hamzehlou et al. Br J Dermatol 2023; 188: 7.
What is already known about this topic?
• Genetic factors are considered to play impor tant roles in keloid formation.
• Genome-wide association study has identified the single-nucleotide polymorphism (SNP) rs87354 9 at 1q41 as a susceptibility locus
for keloids.
• Collagens and skeletal system development, ossification, and osteoblast differentiation -associated proteins, such as P OSTN and
COMP, are enriched in keloid fibroblasts compared with normal scar fibroblasts.
What does this s tudy add?
• rs1348270, an SNP in strong linkage disequilibrium with tag SNP rs873549, is located in an enhancer and is associated with the
downregulation of the long noncoding RNA down expressed in keloids (DEIK).
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
85
Risk SNP drives keloids through lncRNA DEIK, Deng et al.
Keloids represent benign fibroproliferative tumours and are
characterized by fibroblast proliferation and excessive dep-
osition of extracellular matrix.1–3 Keloids continue to grow
over time and rarely regress spontaneously. Keloids can be a
significant burden for the affected individual in terms of cos-
metic harm, pruritus, pain and limited joint movement.1,2,4
Numerous factors have been identified as the cause of
keloid development, such as genetic predisposition, stress,
endocrine factors, hyperactive inflammation and immuno-
logical conditions.1,2 However, the mechanistic details of
keloid pathogenesis remain to be fully characterized.
Genetic factors are considered to play an important role
in keloid formation.1,2 Research has suggested that there is
a strong genetic contribution to keloid formation. Familial
clustering is common in keloid scarring, and the incidence of
keloids in individuals with African ancestry has been reported
to be up to 15 times higher than in white populations.1,2
Agenome-wide association study (GWAS) in a Japanese
population has identified three susceptibility loci (rs873549
at 1q41, rs940187 and rs1511412 at 3q22.3, rs8032158 at
15p21.3) for keloids.5 Among these loci, only the associ-
ation of rs8032158 with keloids has been studied further
in a function study.6 The high frequency of risk allele C in
rs8032158 in keloids is associated with a selectively higher
expression of transcript variant 3 of NEDD4 in the activation
of the nuclear factor-κB pathway.6 However, the most sig-
nificant association with keloids was observed at rs873549
[combined P = 5· 89 × 10−23, odds ratio (OR) = 1·77] on chro-
mosome 1 in the Japanese GWAS study, and the associa-
tion of rs8735 49 with keloids was confirmed in a study of
the Chinese Han population.5,7 These findings suggest that
rs8735 49 may play an important role in keloid formation, but
until now the functional significance of rs873549 in keloid
formation has not been studied.
GWASs have uncovered thousands of genetic variants
that influence risk for complex human traits and diseases.
However, functional studies to delineate the causal genetic
variants and biological mechanisms underlying the observed
statistical associations with disease risk have been lim-
ited.8–10 Several factors have made it difficult to bridge the
gap between the statistical associations linking locus and a
functional understanding of the biology underlying disease
risk. Firstly, there are often many coinherited single-nucleo-
tide polymorphisms (SNPs) in strong linkage disequilibrium
(LD) with the most significant disease-associated SNP, com-
prising a haplotype. Within the haplotype, genetic variants in
strong LD often have statistically indistinguishable associa-
tions with disease risk. Further investigations are required
to determine which of the linked SNPs are functional.
Secondly, > 90% of disease-associated SNPs are located
in nonprotein-coding regions of the genome, and most of
these SNPs are located far from the nearest known gene.
This observation increases the difficulty in investigating the
function of these SNPs.8–10 One clue arises from the obser-
vation that these SNPs are enriched in predicted transcrip-
tional regulatory regions, such as enhancers, insulators and
so on.8–10 This suggests that these SNPs may affect disease
risk by altering the genetic regulation of one or more target
genes. Further experimental approaches are required to help
elucidate the function of these SNPs in disease.
Some studies have suggested that gene expression in
human keloids is altered from dermal to chondrocytic and
osteogenic lineage.11–13 Our single-cell RNA sequencing
study also suggested that skeletal system development,
ossification and osteoblast differentiation-associated pro-
teins, such as POSTN and COMP, were enriched in keloid
fibroblasts compared with normal scar fibroblasts.14 Is has
been suggested that POSTN and COMP play an important
role in the overdeposition of keloid fibroblast extracellu-
lar matrix.15,16 However, the reason for the increase in the
expression of these chondrocyte- and osteocyte-associated
genes in keloids is still unknown.
In this study, we aim to investigate the function of rs873549
and its strong LD SNPs in keloids. Our results indicate that
rs1348270, an SNP in strong LD with tag SNP rs873549,
is located in an enhancer and associated with the down-
regulation of a long noncoding (lnc)RNA RP11-400N13.1
(ENST00000416510.1), which we have termed lncRNA down
expressed in keloids (DEIK), in keloid fibroblasts. Further stud-
ies reveal that downregulation of DEIK increases the expres-
sion of collagens and chondrocyte- and osteocyte-associated
genes POSTN and COMP through promoting BMP2 expres-
sion. These findings will help to improve our understanding of
the role that genetic factors play in keloid development, and
provide potential targets for keloid therapies.
Materials and methods
Haplotype analysis
Haplotype analysis was carried out using Haploview to
determine whether the keloid susceptibility locus rs873549
was in LD with any possible candidate SNPs within a 10-kb
region centred on rs873549. LD was set at r2 > 0·9 and
SNPs were grouped based on LD.
• Downregulation of DEIK promotes keloid development by increasing the expression of collagens and chondrocyte- and osteocyte -
associated genes POSTN and COMP in keloid fibroblasts.
• Downregulation of DEIK promotes POSTN and COMP expression by activating BMP2.
What is the translational message?
• The G allele at rs1348270 is a risk allele for keloids and can be used for screening in a keloid- susceptible population.
• Restoring DEIK expression in keloid fibroblasts may be an effective treatment for patients with keloids.
• Targeting the BMP2 pathway may serve as a treatment strategy for keloids.
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
86 Risk SNP drives keloids through lncRNA DEIK, Deng et al.
Normal and keloid fibroblast culture
This study was approved by the Medical and Ethics
Committees of Dermatology Hospital, Southern Medical
University, and each patient signed an informed consent
form before enrolling in this study. Keloid tissues were har-
vested during plastic surgery from patients confirmed to
have clinical evidence of keloids (Table S1; see Supporting
Information). Normal skin tissues were obtained from vol-
unteers as control participants (Table S1). Fibroblast cul-
tures were established as previously described (File S1; see
Supporting Information).17 The cell strains were maintained
and stored in a liquid nitrogen container, and cells from pas-
sages 3 through 7 were used in this study.
Chromosome conformation capture (3C) assays
The 3C assays were performed as previously described
with some modifications.18 In brief, 5 × 106 primary fibro-
blasts were fixed with 1% formaldehyde and lysed with
cell lysis buffer. Nuclear extracts were digested with XhoI
(FD0694, Thermo Fisher, Waltham, MA, USA) at 37 °C over-
night. Digestion efficiencies of different sites were 93–96%
as determined by quantitative real-time polymerase chain
reaction (qRT-PCR) in which yields of several amplicons that
span XhoI sites were compared with yields of neighbouring
amplicons that were not disrupted by digestion. Digested
chromatin was diluted in ligation buffer and T4 DNA ligase
and incubated at 16 °C overnight. DNA fragments were
extracted using phenol-chloroform and subjected to Taqman
qRT-PCR amplification using the primers and probes listed
in Table S2 (see Supporting Information). Four bacterial arti-
ficial chromosome (BAC) clones (RP11-356K6, RP11-265G5,
RP11-14B15, RP11-355C20) containing human genomic
region 1q41 used for the 3C assay were purchased from
BACPAC Genomics (Emeryville, CA, USA) and used for 3C
control template preparation. The ligation product from XhoI-
digested BAC DNA was used to verify primer efficiency and
also used to normalize 3C interaction frequency.
Statistical analysis
All experiments were performed in triplicate and repeated
at least three times. Statistical analyses were performed
using SPSS software (version 19.0) (IBM, Armonk, NY,
USA). Data were provided as mean ± SD. A two-tailed
unpaired Student’s t-test or the Mann–Whitney U-test were
employed to compare the values between subgroups for
quantitative data, and P < 0·05 was considered statistically
significant.
Results
rs1348270 is located in an enhancer and may
regulate the expression of nearby genes in
fibroblasts
As there are often SNPs in strong LD with the tag SNP, we
first analysed a 10-kb region centred on keloid susceptibility
locus rs873549 by Haploview and discovered 33 SNPs in
strong LD (r2 > 0·90) with rs873549 in this locus (Figure1a).
As all of these SNPs were located in nonprotein-coding
regions of the genome, we hypothesized that these SNPs
may affect keloid risk by altering nearby gene expression.
Thus, we next explored whether these SNPs were located
in cis-acting elements by using the Homo sapiens fibroblast
of dermis DNase I hypersensitive sites and histone ChIP-
Seq data from the ENCODE dataset. We found that among
these SNPs, only rs11118946 and rs1348270 were located
in cis-acting elements; rs11118946 was located in an insu-
lator (marked by CTCF peak) and rs1348270 was located in
an enhancer (marked by H3K4me1, H3K4me2 and H3K27ac
peaks) (Figure 1b). Most of the genes within 500 kb of the
two SNPs are lncRNAs (Figure 1c).
To determine whether the cis-acting elements where
the two SNPs were located can regulate the expression
of nearby genes, we used iKA-CRISPR human embr yonic
stem cells constructed in our lab.19 In these cells, when trans-
fected with a guide RNA containing a 20- bp target sequence
(20-bp gRNA), the doxycycline (Dox)-induced Cas9 protein
was able to cleave genomic DNA. Firstly, we designed three
gRNAs to target the rs11118946-containing insulator and
rs1348270-containing enhancer (Figure 1d) ( Table S3; see
Supporting Information). The tracking of indels by decompo-
sition method showed that all the gRNAs we designed could
cut the targets effectively (Figure 1e). We then derived fibro-
blasts from the cells and transfected the three gRNAs target-
ing the rs11118946 -containing insulator or the three gRNAs
targeting the rs1348270-containing enhancer. Western blot
analysis confirmed the induction of Cas9 expression in the
fibroblasts following Dox treatment (Figure 1f ). The fibro-
blasts were treated with Dox for 2 days or left untreated.
They were then harvested and the expression levels of
the genes near rs11118946 and rs1348270 were analysed.
Except for the five genes shown in Figure1g , we could not
detect the expression of other genes in the fibroblasts using
qRT-PCR (Figure 1c)(Table S4; see Supporting Information).
In the rs1348270 gRNA group, the expression of DEIK (RP11-
400N13.1), LINC01705 and DUSP10 changed significantly in
the fibroblasts when treated with Dox (Figure 1g). Ho wever,
we could not detect any significant gene-expression
changes in the rs11118946 group following treatment with
Dox (Figure 1g). Taken together, these results indicate that
rs1348270 is located in an enhancer and may regulate the
expression of DEIK, LINC01705 and DUSP10 in fibroblasts.
The keloid risk-associated G allele at rs1348270
resulted in decreased expression of DEIK in
fibroblasts
The GWAS study identified rs873549 at 1q41 as a susceptibil-
ity locus for keloids, and the C allele was the risk allele.5,7 The
G allele at rs1348270 has complete linkage with the C allele
at rs8735 49; therefore the G allele at rs1348270 is also a risk
allele for keloids. We examined the genotypes of rs1348270
in the primary normal and keloid fibroblasts and confirmed
that risk allele G was much more frequent in keloid fibroblasts
than in normal fibroblasts (Figure 2a). Next, we explored
the association between rs1348270 genotypes and DEIK,
LINC01705 and DUSP10 expression in normal and keloid
fibroblasts. The results showed that the expression of DEIK
downregulated significantly in fibroblasts with the GG gen-
otype at rs1348270, but there was no significant difference
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
87
Risk SNP drives keloids through lncRNA DEIK, Deng et al.
Figure 1 rs1348270 is located in an enhancer and may regulate the expression of nearby genes in fibroblasts. (a) Haploview analysis of a 10 -kb
region centred on keloid susceptibility locus rs873 549. The red arrow indicates rs873549 location. (b) rs11118946 and rs1348 270 were found to be
located in cis-acting elements by analysing the Homo sapiens fibroblast of dermis DNase I hypersensitive sites (DHS) and histone ChIP -Seq data
from the ENCODE dataset. (c) Genes in the 1000 -kb region centred on rs11118946 and rs1348270. (d) Illustration of gR NAs targeting rs11118946-
containing insulator and rs1348 270-containing enhancer. (e) Tracking of indels by decomposition metho d indicated that all the gRN As were able to
cut the targets effectively. (f) Western blot analysis of Cas9-p300 expression with or without dox ycycline (D ox) treatment in iKA -CRI SPR human
embryonic stem cell- derived fibroblasts. (g) The fibroblasts were treated with Dox for 2 days or left untreated and were then harvested. The
expression levels of the genes near rs11118946 and rs1348270 were analysed using quantitative real-time polymerase chain reaction. Error bars
represent SD (n = 3 ). **P < 0·01. GADPH, Glyceraldehyde 3-phosphate dehydrogenase.
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
88 Risk SNP drives keloids through lncRNA DEIK, Deng et al.
in the expression of LINC01705 and DUSP10 among the
three genotypes at rs1348270 (Figure 2b). Importantly, the
expression of DEIK was downregulated significantly in keloid
fibroblasts compared with normal fibroblasts, but there was
no significant difference in the expression of LINC01705 and
DUSP10 in keloid and normal fibroblasts (Figure 2c). These
results suggest that the G allele at rs1348270 may downreg-
ulate the expression of DEIK in keloid fibroblasts.
The rs1348270-containing enhancer loops to the
DEIK promoter
The results above suggest that rs1348270 is located in
an enhancer and can regulate the expression of DEIK. To
investigate whether rs1348270 -containing enhancer inter-
acts with the DEIK promoter in fibroblasts, we performed
chromosome conformation capture (3C) using digestion by
XhoI enzyme. The 3C technique assesses whether a can-
didate fragment (e.g. the DEIK promoter) physically inter-
acts with a region of interest (e.g. rs1348270-containing
enhancer).18 Of all 21 XhoI sites in this region, we observed
that the rs1348270 locus interacted with the DEIK promoter
(Figure3a). rs1348270 comprises an A nonrisk allele and a G
risk allele. The interaction with the DEIK promoter was spe -
cific to the A nonrisk allele in fibroblasts with AG genotype
(Figure 3b). We further evaluated the enhancer activity of
the rs1348270 locus, for both the A and G alleles, using lucif-
erase reporter assay. Although the reporter vector contain-
ing the risk G allele of rs1348270 produced luciferase activity
similar to that of the control vector, the reporter vector con-
taining the A nonrisk allele demonstrated significantly higher
luciferase activity (Figure 3c). Taken together, these findings
indicate that the rs1348270 -containing enhancer loops to
the DEIK promoter in fibroblasts with an A genotype.
Downregulation of DEIK in normal fibroblasts
increases the expression of collagens and
chondrocyte- and osteocyte-associated genes
To explore the function of DEIK in keloid formation and devel-
opment, two DEIK-targeted small interfering (si)RNAs were
designed to reduce DEIK expression in the normal fibroblasts
NF5 and NF11 (both AA genotype) (Table S5; see Supporting
Information). The efficiency of DEIK knockdown was con-
firmed by real-time PCR (Figure 4a). We first explored whether
cell proliferation changed in DEIK knockdown fibroblasts. Cell
counting assay indicated that DEIK knockdown had no signif-
icant impact on fibroblast proliferation (Figure4b). Collagen I
and III are important extracellular proteins in keloid formation.
Therefore, we subsequently explored the expression of col-
lagen I and III in DEIK knockdown fibroblasts. We found that
DEIK knockdown increased the expression of collagen I and
III in fibroblasts (Figure 4c, d).
Our study, along with studies carried out by other
researchers, suggests that chondrocyte- and osteocyte-
associated genes were enriched in keloids compared
with normal skin.11–14 These genes, such as POSTN and
COMP, have been suggested to promote the expression
Figure 2 The keloid risk-associated G allele at rs1348270 resulted in decreased expression of the long noncoding RNA down expressed in keloids
(DEIK) in fibroblasts. (a) Frequency of rs1348270 genotypes in the primar y normal and keloid fibroblasts. (b) The association between rs1348 270
genotypes and DEIK , LINC01705 and DUSP10 expression in normal and keloid fibroblasts was analysed using quantitative real -time pol ymerase
chain reaction (qRT-PCR). *P < 0·05. (c) qRT-PC R analysed the expression of DEIK, LINC01705 and DUSP10 in both keloid and normal fibroblasts.
Error bars represent SD (n = 3). **P < 0· 01.
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
89
Risk SNP drives keloids through lncRNA DEIK, Deng et al.
of collagens in keloid.15,16 To test whether DEIK expres-
sion is associated with chondrocyte- and osteocyte-asso-
ciated gene expression, we next explored the expression
of POSTN and COMP in DEIK knockdown fibroblasts. The
results showed that the expression of POSTN and COMP
was significantly increased in DEIK knockdown fibroblasts
(Figure 4e, f). Taken together, these results suggest that
down-expression of DEIK increased the expression of col-
lagens and chondrocyte- and osteocyte-associated genes
in fibroblasts.
DEIK regulates chondrocyte- and osteocyte-
associated gene expression through BMP2
To examine the mechanism by which DEIK regulates
POSTN and COMP expression, we performed RNA-Seq
to compare the gene expression in DEIK knockdown fibro-
blasts and control fibroblasts. The top 10 increased and
decreased genes in DEIK knockdown fibroblasts are shown
in Figure5a. Of these, BMP2 was consistently increased
in all the DEIK knockdown fibroblasts (Figure 5a). Enzyme -
linked immunosorbent assay, qRT-PCR and Western blot
analysis validated that BMP2 expression increased in DEIK
knockdown fibroblasts (Figure 5b–d).
BMP2 has been shown to play an important role in chon-
drocyte- and osteocyte-associated gene expression.20,21
To test whether DEIK regulates POSTN and COMP
expression through BMP2, we next treated the DEIK knock-
down fibroblasts using a BMP2 inhibitor, LDN193189. The
results showed that LDN193189 treatment significantly
decreased the expression of POSTN and COMP in the DEIK
knockdown fibroblasts (Figure 5e), suggesting that DEIK
regulates POSTN and COMP expression through BMP2.
Correlation between DEIK and chondrocyte- and
osteocyte-associated gene expression
Finally, we examined the clinical relevance between DEIK
and chondrocyte- and osteocyte-associated genes. We col-
lected the RNA of normal skin and keloid primary fibroblasts
and performed qRT-PCR to detect the expression of DEIK,
BMP-2, POSTN and COMP. Correlation analysis revealed
that DEIK expression was inversely correlated with BMP2,
POSTN and COMP expression (Figure 6a–c). These results
suggest that DEIK expression was inversely correlated with
chondrocyte- and osteocyte-associated gene expression in
primary fibroblasts from normal skin and keloids.
Discussion
Although it has been more than 10 years since the suscep-
tibility loci for keloids have been identified, the molecular
mechanisms underlying the causal actions and biological
Figure 3 The rs1348 270-containing enhancer loops to the down expressed in keloids (DEIK) promoter. (a) Quantification of a 3C assay of the DEIK
genomic region. The data represent relative frequencies of interaction between the anchor region near the DEIK transcription start site and XhoI
digestion sites (circles). All X hoI digestion sites in this region were tested. Relative enrichment of interaction was determined by 3C –quantitative
real-time polymerase chain reaction and normalized to interaction with a bacterial artificial chromosome control librar y. (b) Sanger sequencing of
3C input DN A and 3C amplicons from the DEIK–rs1348270 locus. The blue box highlights risk single-nucleotide protein rs1348270. R represents
nucleotides A and G. (c) Luciferase assay using plasmids containing the rs1348270 locus with either the risk or nonrisk allele of rs1348270. The
pGL3 -promoter plasmid (empty) was used as a baseline control. Luciferase signal was normalized to Renilla signal. Error bars represent SD (n = 3).
**P < 0· 01.
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
90 Risk SNP drives keloids through lncRNA DEIK, Deng et al.
Figure 4 Downregulation of down expressed in keloids (DEIK) increases the expression of collagens and chondrocyte - and osteocy te- associated
genes. (a) Normal fibroblasts (NFs) N F5 and NF11 (both are A A genotype) were transfected with scrambled small interfering (si)RN A or DEIK-
targeted siRNAs (siDEIK) for 48 h, and DEIK mRNA levels were detec ted using quantitative real-time polymerase chain reaction (qRT-PCR). (b) Cell
proliferation was analysed using cell counting assay in DEIK knockdown NFs. (c– d) The mRN A and protein levels of c ollagen I and III were detected
using real-time PCR (c) and Western blot (d) in DEIK knockdown NFs. *P < 0· 05, **P < 0· 01 and ***P < 0·001. (e, f) The mRNA and protein levels of
chondrocyte- and osteocyte -associated genes POSTN and COMP were detected using real-time P CR (e) and Western blot (f) in DEIK knockdown
NFs. Error bars represent SD (n = 3 ). *P < 0·05 , **P < 0·01 and ***P < 0· 00 1.
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
91
Risk SNP drives keloids through lncRNA DEIK, Deng et al.
Figure 5 Downregulation of down expressed in keloids (DEIK) regulates chondrocyte - and osteocy te-associated gene expression through BMP 2.
(a) The top 10 upregulated and downregulated genes in D EIK knockdown cells compared with control cells. (b–d) The mRNA and protein levels of
BMP2 were detected using real -time polymerase chain reaction (PCR) (b), enzyme-linked immunosorbent assay (c) and Western blot (d) in DEIK
knockdown N Fs. **P < 0·01 an d ***P < 0·0 01. (e) The NFs were treated with LDN193189 or dimethyl sulfoxide, and the mRN A levels of POS TN and
COMP were detected using real -time PCR. Error bars denote SD (n = 3). *P < 0· 05, **P < 0· 01 and ***P < 0 ·0 01.
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
92 Risk SNP drives keloids through lncRNA DEIK, Deng et al.
effects of keloid risk-associated SNPs are largely unknown.
In this study, we illustrated the function of rs1348270, an
SNP in complete linkage with keloid risk-associated SNP
rs8735 49, in keloid fibroblasts. Our results suggest that
rs1348270 is located in an enhancer and is associated with
the downregulation of an lncRNA DEIK in keloids. Further
mechanism studies reveal that downregulation of DEIK
increases the expression of chondrocytic and osteogenic
lineage genes through BMP2. These findings are important
in helping us to understand the role of genetic factors in
keloids and could help to identify potential targets for keloid
therapies.
The most significant association with keloids was
observed at rs873549 on chromosome 1 in both Japanese
and Chinese GWAS studies.5,7 However, the mechanism
by which this SNP affects keloid formation and devel-
opment is still unknown. In this study, we found that it
was not rs8735 49, but rather its complete linkage SNP
rs1348270, that was the causal variant of keloids. To our
knowledge, there has been no previous study on the asso-
ciation between rs1348270 and any disease or trait, and our
study is the first to connect this SNP with a fibrotic disease,
namely keloids, and to describe the mechanism by which
this SNP affects keloid development.
GWASs have identified thousands of SNPs associated
with predisposition to disease. Over 90% of these SNPs
are located outside of the exons of protein-coding genes.8–10
These noncoding SNPs are enriched in regulatory regions,
such as enhancers or insulators.8–10 Efforts to interpret the
functional consequences of noncoding SNPs have mainly
focused on the regulation of protein-coding genes, and
lncRNA regulation was studied in only a few cases. We
found that the keloid risk-associated SNP rs1348270 is
located in an enhancer (Figure 1b), and chromosome con-
formation capture experiments suggested that the gen-
otype of rs1348270 determined the interaction between
the rs1348270-containing enhancer and the DEIK promoter
(Figure 3a, b). Thus, we identified a new SNP that could
affect the function of enhancer in keloid.
Although microarray and RNA-Seq studies have identified
several lncRNAs that are increased or decreased in keloids,
and some of these lncRNA functions have previously been
studied,22–25 the understanding of the function of lncRNAs
in keloid formation is still limited. In this study, we identified
an lncRNA, DEIK, which was decreased in keloid fibroblasts
(Figure 2c). Functional studies suggested that DEIK down-
regulation increased the expression of collagen and chon-
drocytic and osteogenic lineage-associated genes POSTN
and COMP in keloid fibroblasts (Figure 4e, f). Importantly,
we connected the expression of DEIK with SNP rs1348270
in keloids, and thus revealed the mechanism by which DEIK
was downregulated in patients with keloids. These findings
will help in understanding the role of lncRNAs in keloid for-
mation and development, and may also provide potential
targets for keloid therapies.
Our study, along with the research of others, suggests
that the expression of chondrocytic and osteogenic lin-
eage genes, such as POSTN and COMP, is increased in
keloid fibroblasts.11–14 It has been suggested that POSTN
and COMP enhance extracellular matrix overdeposition
in keloids.15,16 However, the mechanism by which the
expression of these genes is increased in keloids is still
unknown. In this study, we provide a genetic explana-
tion for this phenomenon. Keloid risk-associated SNP
rs1348270 is associated with the expression of DEIK, and
DEIK was found to regulate the expression of POSTN and
COMP through BMP2 (Figure 5e). BMP2 has been shown
to be a strong inducer of bone and cartilage formation and
can stimulate the expression of chondrocytic and osteo-
genic lineage genes. Considering that there are already a
variety of small-molecule inhibitors for BMP2, our find-
ings may be used to develop new treatment strategies
for keloids.
Some studies have suggested that BMP2 could promote
the formation of scars in skin wound healing,26 but the func-
tion of BMP2 and its regulation in wound-induced fibrotic
skin diseases, such as keloids, is still unclear. Our study
found that lncRNA DEIK was a regulator of BMP2 in keloid
fibroblasts. It has been demonstrated that lncRNAs can reg-
ulate gene expression through the following mechanisms:
(i) recruiting transcription factors or chromatin-modifying
complexes to their DNA targets, (ii) by forming heterogene -
ous nuclear ribonucleoprotein complexes, (iii) by acting as
decoys to sequester RNA-binding proteins and microRNAs,
or (iv) by directly interacting with RNA and DNA by base
pairing.27,28 The mechanism by which DEIK regulates the
expression of BMP2 remains to be explored.
In summary, our study suggests that rs1348270 is located
in an enhancer and is associated with the downregulation
of the lncRNA DEIK in keloids, and downregulation of DEIK
Figure 6 Correlation between down expressed in keloids (DEIK) and chondrocyte - and osteocyte- associated gene expression. (a– c) Correlation
analysis bet ween DEIK expression and BMP2 expression (a), POST N expression (b) and COMP expression (c) in normal and keloid fibroblasts.
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
93
Risk SNP drives keloids through lncRNA DEIK, Deng et al.
increases the expression of chondrocyte- and osteocyte-as-
sociated gene expression through regulating BMP2. These
findings provide valuable insights into the pathogenesis of
keloids and could help to identify novel targets for keloid
treatment.
Acknowledgments
We acknowledge the National Natural Science Foundation
of China and the Science and Technology Foundation of
Guangzhou for funding the study. We also thank all the patients
who contributed their tissues for primary fibroblast culture.
Funding sources
This work was supported by grants from the National Natural
Science Foundation of China (81903189 and 82073418)
and the Science and Technology Foundation of Guangzhou
(202102020025).
Conflicts of interest
The authors declare they have no conflicts of interest.
C.-C.D., L.-X.Z. and X.-Y.X. contributed equally to this work.
Z.R. and B.Y. are cosenior authors.
Data availability
RNA-Seq data have been deposited in the Gene Expression
Omnibus database under accession codes ‘GSE202293’. All
other relevant data supporting the key findings of this study are
available within the article and its Supporting Information files
or from the corresponding author upon reasonable request.
Ethics statement
This study was approved by the Medical and Ethics
Committees of Dermatology Hospital, Southern Medical
University, and each patient signed an informed consent
form before enrolling in this study.
References
1 Andrews JP, Marttala J, Macarak E et al. Keloids: the paradigm
of skin fibrosis – pathomechanisms and treatment. Matrix Biol
2016; 51:37–46.
2 Limandjaja GC, Niessen FB, Scheper RJ, Gibbs S. The keloid dis-
order: heterogeneity, histopathology, mechanisms and models.
Front Cell Dev Biol 2020; 8:360.
3 Macarak EJ, Wermuth PJ, Rosenbloom J, Uitto J. Keloid dis-
order: fibroblast differentiation and gene expression profile in
fibrotic skin diseases. Exp Dermatol 202 1; 30:132– 4 5.
4 Naik PP. Novel targets and therapies for keloid. Clin Exp Dermatol
2022; 47: 507–15.
5 Nakashima M, Chung S, Takahashi A et al. A genome-wide asso-
ciation study identifies four susceptibility loci for keloid in the
Japanese population. Nat Genet 2010 ; 42:76 8 –71.
6 Fujita M, Yamamoto Y, Jiang JJ et al. NEDD4 is involved in inflam -
mation development during keloid formation. J Invest Dermatol
2019; 139:333–41.
7 Zhu F, Wu B, Li P et al. Association study confirmed susceptibility
loci with keloid in the Chinese Han population. P LOS ONE 2013 ;
8:e 6 2 377.
8 Gallagher MD, Chen- Plotkin AS. The Post- GWAS era: from asso-
ciation to function. Am J Hum Genet 2018; 102:717 –30 .
9 Sud A, Kinnersley B, Houlston RS. Genome-wide association
studies of cancer: current insights and future perspectives. Nat
Rev Cancer 2017; 17:6 92–704 .
10 Visscher PM, Wray NR, Zhang Q et al. 10 years of GWAS dis-
covery: biology, function, and translation. Am J Hum Genet 2017;
101:5–22.
11 Fuentes-Duculan J, Bonifacio K M, Suárez- Fariñas M et al.
Aberrant connective tissue differentiation towards cartilage
and bone underlies human keloids in African Americans. Exp
Dermatol 2017; 26:721 –7.
12 Lin P, Peng R, Zhao M et al. RNA-sequencing reveals plantar
keloid has potential bone/cartilage characteristics. Exp Dermatol
20 21; 30:997–8.
13 Naitoh M, Kubota H, Ikeda M et al. Gene expression in human
keloids is altered from dermal to chondrocytic and osteogenic
lineage. Genes Cells 2005; 10:1081– 91.
14 Deng CC, Hu YF, Zhu DH et al. Single-cell RN A-seq reveals fibro-
blast heterogeneity and increased mesenchymal fibroblasts in
human fibrotic skin diseases. Nat Commun 2 0 21; 12:3709.
15 Inui S, Shono F, Nakajima T et al. Identification and characteriza-
tion of cartilage oligomeric matrix protein as a novel pathogenic
factor in keloids. Am J Pathol 2011; 17 9:19 51– 60.
16 Zhang Z, Nie F, Kang C et al. Increased periostin expression
affects the proliferation, collagen synthesis, migration and inva-
sion of keloid fibroblasts under hypoxic conditions. Int J Mol Med
2014; 34:2 53 – 61.
17 Deng CC, Zhu DH, Chen YJ et al. TRAF4 promotes fibroblast pro-
liferation in keloids by destabilizing p53 via interacting with the
deubiquitinase USP10. J Invest Dermatol 2019; 139:1925–35.e5.
18 Hagège H, Klous P, Braem C et al. Quantitative analysis of chro-
mosome conformation capture assays (3C-qPCR). Nat Protoc
2007; 2:17 22 – 33.
19 Ma S, Lv J, Sun J et al. iKA-CRISPR hESCs for inducible and
multiplex orthogonal gene knockout and activation. FEBS Let t
2018; 592:2 2 3 8 – 47.
20 Salazar VS, Gamer LW, Rosen V. BMP signalling in skeletal devel-
opment, disease and repair. Nat Rev Endocrinol 2016; 12:203 –21.
21 Halloran D, Durbano HW, Nohe A . Bone morphogenetic protein -2
in development and bone homeostasis. J Dev Biol 2020; 8:19.
22 Liang X, Ma L, Long X, Wang X et al. LncRNA expression profiles
and validation in keloid and normal skin tissue. Int J Oncol 2 015;
47:182 9 –38.
23 Duan X, Wu Y, Zhang Z, Lu Z. Identification and analysis of dys-
regulated lncRNA and associated ceRNA in the pathogenesis of
keloid. Ann Transl Med 2020; 8:222.
24 Zhao X , Jie X, Gao YK et al. Long non-coding RNA CACNA1G-AS1
promotes proliferation and invasion and inhibits apoptosis by reg-
ulating expression of miR-20 5 in human keloid fibroblasts. Biosci
Rep 2020; 40:BSR20192839.
25 Zhang J, Liu CY, Wan Y et al. Long non-coding RNA H19 pro-
motes the proliferation of fibroblasts in keloid scarring. O nco l Lett
2016; 12:2835 –9.
26 Stelnicki EJ, Longaker M T, Holmes D et al. Bone morphogenetic
protein-2 induces scar formation and skin maturation in the sec-
ond trimester fetus. Plast Reconstr Surg 199 8; 101:12 –19 .
27 Wilusz JE, Sunwoo H, Spector DL. Long noncoding RNAs:
functional surprises from the RNA world. Genes Dev 2009;
23:149 4 –50 4.
28 Engreitz JM, Ollikainen N, Guttman M. Long non-coding RNA s:
spatial amplifiers that control nuclear structure and gene expres-
sion. Nat Rev Mol Cell Biol 2016; 17:75 6 –70.
Supporting Information
Additional Supporting Information may be found in the
online version of this article at the publisher’s website.
Downloaded from https://academic.oup.com/bjd/article/188/1/84/6806087 by guest on 29 January 2023
