Abstract
Introduction: The role of Adipose-derived mesenchymal stem cells
(AD-MSCs) in skin wound healing remains to be fully characterized. This
study aims to evaluate the regenerative potential of autologous AD-MSCs
in a non-healing porcine wound model, in addition to elucidate key
miRNA-mediated epigenetic regulations that underlie the regenerative
potential of AD-MSCs in wounds.
Methods: The regenerative potential of autologous AD-MSCs was evaluated
in porcine model using histopathology and spatial frequency domain
imaging. Then, the correlations between miRNAs and proteins of AD-MSCs
were evaluated using an integration analysis in primary human AD-MSCs
in comparison to primary human keratinocytes. Transfection study of
AD-MSCs was conducted to validate the bioinformatics data.
Results: Autologous porcine AD-MSCs improved wound epithelialization
and skin properties in comparison to control wounds. We identified 26
proteins upregulated in human AD-MSCs, including growth and angiogenic
factors, chemokines and inflammatory cytokines. Pathway enrichment
analysis highlighted cell signalling-associated pathways and
immunomodulatory pathways. miRNA-target modelling revealed regulations
related to genes encoding for 16 upregulated proteins. miR-155-5p was
predicted to regulate Fibroblast growth factor 2 and 7, C-C motif
chemokine ligand 2 and Vascular cell adhesion molecule 1. Transfecting
human AD-MSCs cell line with anti-miR-155 showed transient gene
silencing of the four proteins at 24 h post-transfection.
Discussion: This study proposes a positive miR-155-mediated gene
regulation of key factors involved in wound healing. The study
represents a promising approach for miRNA-based and cell-free
regenerative treatment for difficult-to-heal wounds. The therapeutic
potential of miR-155 and its identified targets should be further
explored in-vivo.
Keywords: adipose-derived mesenchymal stem cells, miRNA, miR-155-5p,
proteome, porcine wound model, wound healing, fibroblast growth factors
1 Introduction
Adipose-derived mesenchymal stem cells (AD-MSCs) are known for their
plasticity and low immunogenicity, with adequate yield isolated from
fat biopsies and lipoaspirate ([44]Gimble and Guilak, 2003; [45]Kern et
al., 2006; [46]Bailey et al., 2010). Upon appropriate stimulation,
AD-MSCs have the potential to differentiate into a wide variety of cell
types, including osteoblasts, chondrocytes, myocytes, adipocytes,
fibroblasts, smooth muscle, endothelial and epithelial cells
([47]Gimble et al., 2007; [48]Hassan et al., 2014; [49]Azari et al.,
2022). When applied in-vivo, AD-MSCs can enhance tissue regeneration
through their differentiation into skin cells or secretion of paracrine
factors. The latter can initiate the healing process via recruiting
circulating stem cells, as well as other local cells to the wound
microenvironment. AD-MSCs have shown potential to differentiate into
various skin cell types including fibroblasts and keratinocytes. In
addition, AD-MSCs can promote wound healing by releasing growth factors
and cytokines, which promote epithelial migration, neovascularization
and collagen synthesis ([50]Hassan et al., 2014). AD-MSCs modulate the
immune response by suppressing T-cell proliferation and inducing the
production of regulatory T-cells, which play a critical role in
supressing the inflammatory response and promotion of tissue
regeneration ([51]Lin et al., 2013; [52]Ong et al., 2017; [53]Weiss and
Dahlke, 2019). These features rendered AD-MSCs a viable alternative to
epidermal cells in skin regenerative applications, particularly in
cases of extensive skin loss as the donor sites for keratinocyte
isolation are scarce ([54]Gimble et al., 2007; [55]Cherubino et al.,
2011; [56]Kokai et al., 2014; [57]Kirby et al., 2015; [58]Foubert et
al., 2016; [59]Bertozzi et al., 2017).
In this study, we investigated the effect of AD-MSCs in a porcine full
thickness skin wound model and the healing was evaluated using
histopathology and spatial frequency domain imaging (SFDI). SFDI is an
imaging technique that projects a sequence of sinusoidal patterns of
illumination with varying spatial frequencies onto tissue and utilizes
a camera to detect diffusely remitted light. By measuring both the
reduction of remitted light intensity and contrast from these patterns,
models of light transport can calculate the amount of absorption and
light scattering present within the tissue. Another aspect of interest
in SFDI is the penetration depth of light, which depends on the spatial
frequency (f[x]) of the sinusoidal patterns ([60]O’Sullivan et al.,
2012). Quantifying these absorption and scattering properties over
multiple wavelengths permits physiologic interpretations of these
optical signals ([61]O’Sullivan et al., 2012; [62]Zannettino et al.,
2008; [63]Yafi et al., 2011; [64]Saager et al., 2013; [65]Murphy et
al., 2020; [66]Lee et al., 2020; [67]Ponticorvo et al., 2020). This
technique has been previously applied to burn wound assessment
([68]Ponticorvo et al., 2019; [69]Ponticorvo et al., 2020), chronic
wounds ([70]Lee et al., 2020; [71]Murphy et al., 2020), and several
other clinical applications ([72]Yafi et al., 2011; [73]Saager et al.,
2013). To our knowledge, this is the first application of this imaging
approach to assess cell-based therapeutics in a surgical wound model.
The general understanding regarding AD-MSCs’ mechanism of wound repair,
underlying pathways and interplay with other cell types is currently
insufficient. At the wound site, AD-MSCs interact with surrounding
tissue, which is crucial to achieve skin repair. microRNA (miRNA),
fragments of short non-coding RNA, can be considered as one of the most
important mediators in cell-to-cell communication. miRNA can
disseminate through the extracellular fluid to act as signaling
molecules or through direct exchange of exosomes between adjacent
cells. miRNA can also shuttle as a cargo in the exosomes, locally and
through the systemic circulation. The ultimate effect is the alteration
of gene expression and protein production in the recipient cell
([74]Alghfeli et al., 2022). miRNA are involved in the
post-transcriptional regulation of protein-coding genes by interfering
with messenger RNA transcript, leading to its degradation, or -at
least-repression of protein production ([75]Collins, 2011). miRNA can
influence diverse biological processes including cell growth,
development, metabolism, migration, proliferation, differentiation and
apoptosis ([76]Kim and Sung, 2017; [77]Omar et al., 2019).
Additionally, miRNA have been shown to regulate various aspects of
wound healing including cell proliferation, migration, collagen
biosynthesis, and angiogenesis ([78]Wang et al., 2012; [79]Banerjee and
Sen, 2013; [80]Veith et al., 2019). On the other hand, proteins
released by AD-MSCs act as stimulants for cell growth, tissue
granulation, increased macrophage recruitment and improved
neovascularization. These proteins include growth factors, cytokines
and chemokines, such as vascular endothelial growth factor (VEGF),
insulin-like growth factor (IGF), and transforming growth factor-beta
(TGF-β), among others ([81]El-Serafi et al., 2017; [82]Ong et al.,
2017; [83]Shahin et al., 2020; [84]Azari et al., 2022).
AD-MSCs treated wounds showed better properties of healing that
mimicked natural healing. We performed an integrated analysis of our
previously published miRNome data in primary human AD-MSCs, as well as
to relevant protein content, in comparison to normal keratinocytes as
the primary effector in physiological cutaneous wound edge healing
([85]Shahin et al., 2023). The findings were validated through gene
expression study and via anti-miR-driven experimental targeting. The
aim of the study was to identify the main factors involved in skin
wound repair produced by AD-MSCs through a miRNA-based approach.
2 Materials and methods
2.1 Experimental animal procedures
All animal procedures were carried out according to the animal research
ethical approval (No. 1961 DNR 5.2.18-1627/15) in compliance with the
guidelines mandated by the regional animal ethics review board and the
Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines
to ensure humane treatment of research animals. Two female pigs (Sus
scrofa domesticus) weighing between 50 and 60 kg were obtained from a
licensed breeder and included in this study. Animals were housed in
enclosures measuring 2 × 2.5 m^2, with a light/dark cycle of 12 h/12 h
and an ambient temperature of 18°C–20°C. The animals had free access to
water and hay and were fed twice daily. Surgical procedures were
performed under general anaesthesia, induced by intramuscular injection
of 10 μg/kg dexmedetomidin (Dexdomitor; OrionPharma, Danderyd, Sweden)
and 3 μg/kg of tiletamin and zolazepam (Zoletil; Virbac, Kolding,
Denmark). Animals were intubated with an endotracheal tube connected to
an automatic ventilator. General anaesthesia and analgesia were
maintained with intravenous infusion of 3–7.5 mg/kg pentobarbital
sodium (Pentobarbitalnatrium vet.; APL, Kungens Kurva, Sweden) in
combination with 0.5−0.75 μg/kg fentanyl (Leptanal, Janssen, Solna,
Sweden). Vital parameters were monitored by pulse oximetry, capnography
and rectal thermometer. Signs of postoperative pain were treated with
fentanyl patches, 25–50 μg/h for 72 h (Fentanyl Orion; Orion Pharma,
Danderyd, Sweden).
2.2 Porcine wound model
For cell isolation, two full-thickness skin biopsies, measuring
3–5 cm^2, were collected from the neck region of each animal. The
wounds were closed with sutures with prolene 2–0 (Ethicon Inc.;
Somerville, New Jersey, United States) and covered with sterile gauze.
Porcine AD-MSCs (pAD-MSCs) were extracted from the subcutaneous adipose
layer after mechanical separation from the dermis layer. Briefly, 20 mL
of adipose tissue were cut into 0.5–1 cm^2 slices and incubated with
Collagenase I (1 mg/mL) (Gibco, Life Technologies, United Kingdom) at a
3:1 volume ratio on a tube rotator at 37°C for 90 min. The reaction was
stopped by adding Dulbecco’s Modified Eagle Medium (Gibco, Billings,
MT, United States) with 10% fetal bovine serum (Life Technologies, São
Paulo, Brazil). The cells were centrifuged and washed, at least twice,
with phosphate buffered saline (PBS; Life Technologies, Grand Island,
NY, United States) then suspended in 2.5% human serum albumin (HSA)
(Alburex, CSL Behring GmbH, Marburg, Germany). The following day the
animals (n = 2) were anesthetized and sixteen full-thickness excisional
wounds with a radius of 1.5 cm to resemble non-healing wounds were
created paravertebrally on the dorsum of each animal. The experimental
group of pAD-MSCs and a negative control were randomly assigned to a
wound on each pig. Wounds were covered with occlusive dressings
(Tegaderm™, 3M™, MN, United States). Then 1 × 10^6 pAD-MSCs cell
suspension was injected through the dressing into the wound cavity. The
injection site was sealed with another layer of the dressing. The
control wounds were left untreated to delineate physiological healing.
An absorbent foam dressing (Mepilex XT, Mölnycke, Sweden) was added
over the Tegaderm and finally the whole trunk was wrapped with an
elastic dressing (Elastic bandage, Hansbo sport, Västra Frölunda,
Sweden). At 0, 7 and 14-day time points, planimetric measurements
(wound diameters) were taken for each wound. The progress of wound
healing was clinically evaluated by a plastic surgeon at day 7 and 14.
At the 2-week mark, the animals were euthanized by intravenous
injection of 400 mg/kg pentobarbital sodium (Pentobarbitalnatrium vet.;
APL, Kungens Kurva, Sweden).
2.3 Wound evaluation using spatial frequency domain imaging
A wide-field spectral imaging system was employed to non-invasively
quantify the changes within the wound areas at 0, 7, and 14-day
time-points. The specific imaging system is a compact, low-cost,
custom-built imager, developed at Linkoping University ([86]Belcastro
et al., 2020). The system is capable of imaging tissue in a field of
view of approximately 5 cm width, over five spectral bands (458, 520,
536, 556, 626 nm). 11 spatial frequencies were acquired, ranging from 0
to 0.5 mm^−1, in steps of 0.05 mm^−1. Patterns with lower f[x]
penetrate deeper in the tissue, while higher f[x] is shallower. Making
use of this property, the dataset of 11 f[x] was subdivided into
smaller subsets containing 4 frequencies each. These subsets contain
information about different volumes, which allow comparative
measurements from deeper and more superficial tissue ([87]O’Sullivan et
al., 2012) and make qualitative interpretations on the state of new
tissue growth and the underlying supporting structures. Since the data
is in image format, a rectangular area was drawn over the wound sites
as a region of interest. The mean and standard deviation of the
parameters area were calculated at each pixel, to account for the
spatial heterogeneity of the wound, for at least 200 pixels/wound. The
mean scattering coefficient (µ’s) was then fit to the scattering
parameters A and B. The dependance of µ’s to the wavelength of light
(λ) can be modelled with a power law as
[MATH:
μs′λ=A<
/mi>λ−B
:MATH]
, where the two scattering parameters amplitude (A) and slope (B) are
related respectively to the concentration of scattering particles and
their average dimension ([88]Mourant et al., 1998). Higher parameter A
reflects greater number of scattering particles while higher B
parameter indicates smaller size of these particles. Objects that
scatter light in tissue, range from “large” nuclei to “small” collagen
fibrils and other extracellular matrix (ECM) components, so the B
parameter can be used as a rough estimator of the most common
structures in the tissue. In the context of wound healing, increases in
both A and B parameters represent new cell growth
(re-epithelialization) and collagen restructuring/formation. In this
study, the SFDI data in two spatial frequency ranges (f1 and f2) were
processed. These distinct ranges provide insight into the relative
depths at which these changes in tissue structure (scattering) and
function (absorption) occur. Data from f1 represents deeper tissue
volumes, ranging from tissue surface to approximately 0.3–0.5 mm deep,
and f2 is restricted to the more superficial tissue (surface to
∼0.25–0.35 mm).
2.4 Tissue preparation and histological examination
Biopsies were obtained at day 14. The wounds were excised using a
scalpel with an approximately 3 mm margin. Biopsies were fixed in 4%
buffered formaldehyde, dehydrated by immersion in a series of
ethanol-xylene and embedded in paraffin. Tissue sections were cut into
4–5 µm thickness, deparaffinized in xylene, rehydrated and stained with
hematoxylin and eosin (H&E) as well as Masson’s trichrome stain (Sigma
Aldrich, United States). The slides were evaluated by a specialist skin
pathologist blinded to the study design and images were captured using
the Imageview software version X64 (Olympus Corp., Japan). The
epidermal thickness was measured using the analysis function of Adobe
Photoshop 2023 with measurement scale customized to the scale bar. At
least, twenty readings were obtained for each group.
2.5 Analysis of miRNA microarray and protein array data and pathway
enrichment
The miRNA and protein screening data in Keratinocytes and AD-MSCs
([89]Shahin et al., 2023), was analysed to identify the upregulated
proteins in AD-MSCs. Our reported dataset “Differentially expressed
DEmiRNA-mRNA Interaction Network analysis” in AD-MSCs versus
keratinocytes was utilized to identify the enriched pathways in
AD-MSCs. From this analysis, 555 unique target genes related to 58
upregulated miRNA in AD-MSCs were used as input list#1 in the Reactome
database (v83, [90]https://reactome.org/) ([91]Yu and He, 2016) for
pathway enrichment analysis. Similarly, genes encoding upregulated
proteins in AD-MSCs were also investigated independently for pathway
enrichment as input list#2. The following filters were applied, 1)
species: Homo sapiens, 2) statistical significance: p < 0.05. The
resulting lists of pathways from each analysis were compared to
highlight the top matching enriched pathways.
2.6 Predictive miRNA-mediated gene regulations (integrated analysis)
The two-tailed Fisher’s exact test (2 × 2 contingency table) was used
to test the null-hypothesis of no association between DEmiRNA and
differentially expressed proteins. Integrated analysis was performed to
extract predictive miRNA-mediated gene regulations in AD-MSCs. To
maintain a reliable source for miRNA target genes, the Affymetrix
GeneChip^® miRNA 4.0 array annotation (v. HG38) containing the
experimentally validated targets gene symbols corresponding to miRNA
using miRecords for validated targets. All possible gene names
corresponding to the genes encoding for the experimentally identified
upregulated proteins in AD-MSCs were crossed matched with the annotated
gene targets for all differentially expressed miRNA. The matching
DEmiRNA with target genes encoding for any of the upregulated proteins
in AD-MSCs were identified and categorized according to the following 2
assumed regulations:
[MATH: miR
NA → Protein−
coding gene → ProteinORmiRN<
mi>A ─┤ Protein−coding gene → Protein :MATH]
2.7 Transfection with miR-155-5p inhibitor
ASC52telo, hTERT immortalized human adipose-derived mesenchymal stem
cells (ATCC, United States) were used as a model for AD-MSCs in the
transfection study. Lipofectamine RNAiMax (Thermo Fisher Scientific,
Waltham, MA, United States) was used as a transfection reagent,
according to the manufacturer’s instructions. The cells were
transfected with mirVana^© miRNA inhibitor (Anti-miR-155-5p) miRBase
accession #MIMAT0000646 (Ambion, life technologies, TX, United States)
at final oligonucleotide concentration of 30 nM or mock transfection
control (Lipfectamine RNAiMAX + PBS), for 24 h. After the first 24 h,
transfection was stopped by changing media. ASC52telo cells were
harvested for RNA extraction at 24 and 72 h to evaluate the expression
of the downstream genes.
2.8 Gene expression analysis
Quantitative real-time PCR was performed to detect the expression of
the downstream miRNA-gene targets at 24 and 72 h post-transfection (n =
3). Total cellular RNA was extracted using RNeasy mini kit (Qiagen,
Germany). RNA was reverse transcribed into cDNA using Maxima First
Strand cDNA Synthesis Kit (Thermo Fisher Scientific, United States)
following the elimination of double-stranded DNA as recommended by the
manufacturer. Gene expression was determined by the PowerUp^© SYBR
green master mix (Applied biosystem, United States). Sequences for the
oligonucleotide primers of target genes are listed in
([92]Supplementary Table S1). Real-time qPCR was carried out using
Applied Biosystems™ 7500 Real-Time PCR System (Applied biosystem,
United States). The relative fold change in mRNA expression was
calculated using the 2^−ΔΔCT method and normalized to the endogenous
housekeeping gene Glyceraldehyde 3-phosphate Dehydrogenase (GAPDH).
2.9 Statistical analysis
For qPCR and protein arrays, data were analysed using the Data Analysis
ToolPak (Microsoft^® Excel, Microsoft^® Office 365) and the graphs were
created using GraphPad Prism Version 9 (GraphPad Software Inc., San
Diego, CA). Statistical significance was evaluated using Student’s
t-test and statistical significance was determined if p-value < 0.05.
Bar charts showed the mathematical mean and the standard error of mean.
To estimate the upregulated miRNA from the microarray analysis, p-value
< 0.05 was considered as the level of significance. To identify the
enriched pathways in the analysis, Benjamin-Hochberg (BH) corrected
p-value < 0.05 was applied.
3 Results
3.1 Porcine wound healing revealed improved epithelialization and accelerated
wound closure with autologous pAD-MSCs compared to control
Planimetric measurements showed reduction of the wound surface area at
day 14 compared to those of day 0 with 84.2% healing efficiency in
pAD-MSCs treated wounds in respect to their corresponding initial wound
size. At the meantime, negative control showed only 72.1% healing
efficiency. Similar difference could not be shown at day 7.
Histological evaluation of the wound biopsies at day 14 showed that the
epidermis of the negative control wounds showed a variation in
thickness with area of ulcerated wound. Non-healing control wounds
exhibited acanthotic/hyperkeratotic epidermis at the periphery.
Moreover, the dermis showed signs of purpura, blood stasis in addition
to moderate-to-severe inflammation. In contrast, the treatment group
showed uniform epithelization across the whole wound area, in the form
of wide and thick epidermal ridges with mean (SEM) thickness of 168
(14) µm compared to that of the negative control wounds 81.2 (5.1) µm.
Some parts of the epidermis of the treatment group appeared flattened
in the middle, with sporadic acanthosis or hyperkeratosis. The dermis
of the treatment group showed mild inflammation and fibrosis with
better production and arrangement of the collagen in comparison to the
control ([93]Figure 1).
FIGURE 1.
[94]FIGURE 1
[95]Open in a new tab
Evaluation of porcine excisional full thickness wounds at day 14.
Images (A–D) represents the central region of the corresponding wound.
(A) H&E staining of the cell-free control wounds exhibiting
acanthotic/hyperkeratotic epidermis at the periphery, signs of purpura,
blood stasis in addition to moderate-to-severe inflammatory reaction.
(B) H&E staining of the autologous pAD-MSCs treated wounds showing
uniform epithelization across the whole wound area with pronounced
epidermal ridges, the dermis showing mild inflammation. (C) Masson’s
trichrome staining of negative control showing no to little collagen
disposition in negative control wounds. (D) Masson’s trichrome staining
of the autologous AD-MSCs treated wounds showing better collagen
production and arrangement in comparison to the control. (E) Bar chart
showing the reduction in wound surface area at day 14 compared to day
0, while day 7 showing no difference. Values presented as mean
percentages of D7/D0 and D14/D0 with standard error of mean (SEM). (F)
Bar chart showing the difference in epidermal thickness, the mean with
SEM of–at least-twenty read-outs were obtained for each group. Scale
bars = 200 um. Statistical significance was evaluated using Student’s
t-test for unequal variance *p < 0.05, ****p < 0.0001.
SFDI data was acquired on three wounds (two negative controls and one
treated with pAD-MSC) at three different time points: day 0, day 7 and
day 14. It was not possible to obtain SFDI data from the other pAD-MSC
treated wound, according to the animal welfare policy. The second pig
was under anaesthesia and the procedure needed more time, which could
not be granted. Thus, data collection was not possible for some of the
wounds. The investigators were blinded to the treatment received by
each wound. The spatial frequency ranges of the datasets are f[1] =
[0.05–0.2] mm^−1 and f[2] = [0.1–0.25] mm^−1. The raw data was
processed and the optical parameters µ[a] and µ’[s] were obtained. In
([96]Figure 2), the progression of the scattering parameters in time
for the three wounds can be observed. For pAD-MSCs treated wound, an
increasing trend in both the A and B parameter was observed, which is
consistent with the formation of new tissue (new epithelial cells and
ECM). In the control wounds we see a similar trend, but on a smaller
scale for both A and B parameters. The difference in scattering
parameters at day 7 between pAD-MSCs and control is statistically
significant (p < 0.05), while their values are not significantly
different at day 14 (p > 0.05). This suggests a slower cell growth and
tissue restructuration in comparison to the pAD-MSCs treated wound.
Furthermore, the progression of the absorption coefficient in pAD-MSCs
treated wound showed an initial increase in µ[a] at day 7, followed by
a decrease at day 14. The higher value of µ[a] at f[1] compared to f[2]
also suggests that the increase in haemoglobin content is in deeper
tissue, which could be an indication of increased microcirculation. In
both control wounds, there are no significant changes in absorption at
day 7, which suggest weak or absent tissue response. We can however see
a larger increase in µ[a] at day 14, which could be related to an
inflammatory reaction, as observed in the histopathological study.
FIGURE 2.
[97]FIGURE 2
[98]Open in a new tab
Plots of the scattering amplitude (A), scattering slope (B) and
absorption coefficient µ[a] (C) measured at three time points on the
porcine model on a wound treated with pAD-MSC (plain color), and
negative control wounds (patterned color). Data from two datasets with
different spatial frequencies are reported: f[1] = [0.05–0.2] mm^−1
(blue) and f[2] = [0.1–0.25] mm^−1 (red). Low spatial frequencies have
higher penetration depth, so the data from f[1] comes from deeper
tissue compared to f[2]. Statistical significance was evaluated using
Student’s t-test for unequal variance between pAD-MSCs and the control
case *p < 0.05.
3.2 Analysis of miRNA microarray and protein array data reveals
differentially upregulated targets in AD-MSCs
Of the total 378 miRNA differentially regulated in primary AD-MSCs, 264
miRNA (69.84%) were upregulated while 114 miRNA (30.16%) were
downregulated. On the other hand, proteome profiler arrays showed that
26 of the 105 soluble proteins blotted on the array membrane were
significantly upregulated in AD-MSCs. Of those, the most upregulated
were Fibroblast growth factor-7 (FGF-7), also known as (a.k.a)
Keratinocyte growth factor (KGF) [log[2] mean difference (MD) = 5.7,
p-value = 0.006], CD31 a.k.a Platelet and endothelial cell adhesion
molecule 1 (PECAM-1; log[2]MD = 5.2, p-value = 0.04), Endoglin
(log[2]MD = 5.1, p-value = 0.0009), Chitinase 3 like 1 (log[2]MD =
5.03, p-value = 0.001) and Fibroblast growth factor basic (FGF-basic;
log[2]MD = 4.7, p-value = 9.9E-06) ([99]Figure 3; [100]Supplementary
Table S2).
FIGURE 3.
[101]FIGURE 3
[102]Open in a new tab
(A) Heatmap diagram showing the signal intensities of the detected
proteins in the proteome profiler array. Heatmap visualization was
created using shinyHeatmaply. The colour bar represents individual
signal intensity of each sample blue (low intensity) to red (high
intensity). (B) Bar chart detailing the 26 significantly upregulated
proteins in AD-MSCs. Data represented as log[2] mean difference in
expression between AD-MSCs and keratinocytes and the standard error of
mean. Statistical significance was evaluated using Student’s t-test for
unequal variance of 3 independent donor samples (n = 3). *p < 0.05, **p
< 0.01, ***p < 0.001, ****p < 0.0001.
3.3 miRNA-mRNA interactome and enrichment analysis of the miRNA-regulated
targets in AD-MSCs reveal cell signalling-associated pathways
The fisher’s exact test (p-value = 0.008, two-tailed) supported the
association between the 378 DEmiRNA and 28 DEproteins
([103]Supplementary Table S3). Then the integrated analysis carried out
on the 26 upregulated proteins in AD-MSCs and their corresponding
protein-coding genes revealed that only 16 (of 26) matched with the
known gene targets of 54 DEmiRNA ([104]Figure 4A; [105]Supplementary
Table S3). A total of 95 and 92 significantly enriched pathways were
observed resulting from the input gene lists 1 and 2, respectively
([106]Supplementary Tables S4, S5) ([107]Supplementary Figures S1A, B).
Additionally, comparing the above results, 9 significant pathways were
observed, including MAPK family signalling cascades, Transcriptional
regulation of pluripotent stem cells, Signalling by Receptor Tyrosine
Kinases, Signal transduction by Interleukins ([108]Supplementary Table
S6; [109]Figure 4B; [110]Supplementary Figure S1C).
FIGURE 4.
[111]FIGURE 4
[112]Open in a new tab
(A) miRNA-mRNA interaction network in AD-MSCs illustrating the
predicted regulations between 54 DEmiRNAs and the genes encoding for 16
upregulated proteins. (B) The dot plot representing the pathway
enrichment result using the GSEA algorithm. The size of the circle
shows the number of genes in the pathway. The color scale represents
blue (low FDR) to red (high FDR). The enrichment ratio calculates the
enrichment for each pathway. The dot plot visualizing the matching 9
significantly enriched pathways by the input gene sets representing
target genes of the upregulated miRNAs and genes encoding for
upregulated proteins.
3.4 Silencing of miR-155-5p validated the predicted miRNA-mediated gene
regulation
To validate the predicted miRNA-mRNA interactome model, the positive
regulation between miR-155-5p and 4 protein coding genes, which are
FGF2, FGF7, CCL2 and VCAM1 was tested. These genes correspond to the
upregulated proteins: FGF basic log[2] fold change (FC) = 4.7, p-value
= 9.9E-06), FGF-7 (log[2]FC = 5.7, p-value = 0.006), Monocyte
chemotactic protein-1, encoded by CCL2, (MCP-1; log[2]FC = 3.1, p-value
= 0.003) and Vascular cell adhesion molecule-1 (VCAM-1; log[2]FC = 3.6,
p-value = 0.02) ([113]Figure 5A). To test this hypothesis, a gene
silencing experiment was conducted using miR-155-5p inhibitor. The
temporal downstream effect on the predicted target genes was evaluated
at 24 h ([114]Figure 5B) and 72 h ([115]Figure 5C) post-transfection.
Gene expression analysis revealed downregulation at 24 h of FGF7
(log[2]FC = −1.4, p-value = 0.006), CCL2 (log[2]FC = −2.6, p-value =
1.7E-05) and VCAM1 at (log[2]FC = −1.5, p-value = 0.01), with a similar
trend for FGF2 (log[2]FC = −0.4, p-value = 0.08) when compared to the
mock transfection control without reaching statistical significance
([116]Figure 5B). The inhibitory effect of miR-155-5p appeared to be
only transient as the majority of the genes mitigated at 24 h, reverted
to their baseline expression levels after 72 h from terminating the
transfection and alleviating the gene-silencing effect FGF2 (log[2]FC =
0.3, p-value = 0.07), FGF7 (log[2]FC = −0.1, p-value = 0.1), VCAM1
(log[2]FC = 0.4, p-value = 0.2). However, CCL2 appeared to have been
upregulated following the transient downregulation as it showed a surge
of expression at log[2]FC = 1.2, p-value = 0.01) ([117]Figure 5C).
FIGURE 5.
[118]FIGURE 5
[119]Open in a new tab
(A) Schematic represtentation of the predictive miR-155-mediated gene
regulation with FGF2, FGF7, CCL2, and VCAM1 pretaining to wound repair.
(B) Gene expression analysis, 24 h post-transfection with
Anti-miR-155-5p showing gene silencing of FGF2, FGF7, CCL2, and VCAM1.
(C) Gene expression analysis, showing gene silencing effect on FGF2,
FGF7, CCL2, and VCAM1 alleviated 72 h post-transfection. Values in bar
chart represent log2 fold change compared to mock transfection.
Statistical significance was evaluated using Student’s t-test for
unequal variance of 3 independent experiments (n = 3). *p < 0.05, **p <
0.01 and ****p < 0.0001.
4 Discussion
The role of AD-MSCs in wound healing is attracting more attention due
to their feasibility and efficacy. This study verified the regenerative
capacity of AD-MSCs auto-transplants in porcine full-thickness
excisional wounds. At day 14, histological and morphometric analyses
showed AD-MSC expedited wound closure compared to naturally healed
wounds. Similar findings were reported when the cells were applied to a
burn wound as subdermal injection or topical spray ([120]Foubert et
al., 2016). Our findings were verified by histopathology, as well as by
SFDI. The latter can be considered as a non-invasive technique, which
can be directly correlated to biological markers normally associated
with wound healing. The absorption coefficient (µ[a]) gives an
indication of the concentration of chemical species in tissue that are
able to absorb photons. Molecules that absorb light within the spectral
range of the imaging system used in this investigation are haemoglobin
and melanin ([121]Jacques, 2013). The latter was excluded from the
analysis as the epidermis was surgically removed during the wound
creation step. Thus, µ[a] at ∼530 nm was selected to represent the
spatially resolved total haemoglobin present within the wound area
([122]Saager et al., 2018). This parameter gives an approximate
indication of the quantity of blood present in the tissue (e.g., from
micro-circulation or haemoglobin degradation products). In stem-cells
treated wound, the scattering parameters suggested a larger degree of
cell repopulation, while the absorption (from haemoglobin) showed a
faster reactivity of the tissue in time, when compared to negative
control. An important contribution of these optical measurements is the
introduction of additional time points during the healing process. This
complements the data obtained from the biopsy, which can only be
obtained at the end of the study, allowing a more comprehensive
assessment in time. However, a low number of wounds were obtained from
the current porcine wound model, which represents a limitation of this
study. Future studies should include more animals or more wounds per
animal, to further confirm our findings.
The next step of the study was to investigate the molecular signalling
cascades associated with AD-MSCs’ effect in wound healing. This study
analysed the previously identified miRNA expression profile in AD-MSCs
([123]Shahin et al., 2023); integrated with the identified proteome
signature of AD-MSCs. This integration enabled us to generate a
predicted positive regulation between miR-155 and genes encoding for
FGF2, FGF7, CCL2, and VCAM1 in human AD-MSCs. The inhibition of miR-155
elicited gene silencing for FGF2, FGF7, CCL2, and VCAM1 for 24 h,
validating the predicted miR-155-mediated gene regulation. The positive
regulation role of miRNA on target genes can have multiple mechanisms,
including and not limited to, 1) miRNA-mediated post-transcriptional
upregulation 2) translation upregulation or 3) competition with
repressive proteins preventing them from binding to their target sites,
which increases mRNA stability and promotes the expression of target
protein. Nevertheless, miRNA-mediated regulation of gene expression is
a reversible action ([124]Valinezhad Orang et al., 2014). Thus, a
single miRNA can exert both positive and negative gene regulations, and
similarly a single gene could respond in both modulations depending on
system-specific conditions.
miR-155 has been recognized as a regulator and respondent to multiple
inflammatory mediators involved in the cellular immune response against
pathogens including interleukins, tumor necrosis factor and interferons
([125]Mahesh and Biswas, 2019; [126]Rodriguez et al., 2007;
[127]O’Connell et al., 2007). Silencing of miR-155-5p in atopic
dermatitis mouse model reduced both the thickening of the epidermis and
the expression of T helper type 2 (Th2) immune response-associated
cytokines, thymic stromal lymphopoietin and interleukin-33, attenuating
the overall allergic inflammation ([128]Wang et al., 2019). miR-155 is
among the first miRNA to be associated with the ability of inflammatory
cells to recognize invading pathogens particularly through Toll-like
receptors (TLRs). Thus, miR-155 can have an important role in wound
healing research, particularly in chronic infected wounds
([129]Martinez-Nunez et al., 2009). Furthermore, miR-155 is associated
with acceleration of epithelial cell migration through promoting
TGF-β-induced epithelial-to-mesenchymal transition and tight junction
dissolution in normal murine mammary gland epithelial cells ([130]Kong
et al., 2008). The role of miR-155 in wound healing was elucidated in a
full-thickness excision wounds in rats and shown to have accelerated
cutaneous wound closure independent of wound contraction. The authors
provided in-vitro evidence that the regenerative effect of
overexpressing miR-155 was attributed to enhanced keratinocytes
migration via the activation extracellular proteins, particularly
matrix metalloproteinase-2 (MMP-2) ([131]Yang et al., 2017). In
mesenchymal stem cells isolated from aged bone marrow donors, miR-155
was found to be considerably higher than that of young donors MSCs.
Additionally, upregulation of miR-155 in young MSCs led to increased
cellular senescence signified by increased signal of
senescence-associated β-galactosidase (SA-β-gal). This proposes a
positive correlation of miR-155 expression with cellular senescence of
MSCs through regulating the Cab39/AMPK signalling pathway. Aged MSCs
can help to improve the cardiac function when miR-155 was inhibited in
the cells prior to the injection of the infarcted myocardium in murine
model. The effect was evident by enhanced angiogenesis and cell
survival in addition to reduced infract size as well as cardiomyocyte
apoptosis ([132]Hong et al., 2020). Beside cellular senescence, there
is scarcity in the literature describing the role of miR-155 in
AD-MSCs. In addition, miR-155 is released as a main constituent of the
exosomal miRNA cargo in adipose tissue macrophages (ATMs) which can be
transferred to and modulate insulin sensitivity in other cell types
such as adipocytes, myotubes or hepatocytes ([133]Ying et al., 2017).
Furthermore, miR-155 was detected in the extracellular vesicles
isolated from AD-MSCs ([134]Constantin et al., 2022). In intervertebral
disc degeneration (IDD), AD-MSCs-derived exosomal miR-155-5p showed to
inhibit pyroptosis and promoted autophagy and ECM synthesis in nucleus
pulposus cells, in-vitro, through targeting TGF beta receptor 2
(TGFβR2). In another model, exosomes were isolated form AD-MSCs
overexpressing miR-155 and injected in IDD rats which was associated
with improved symptoms of IDD in relation to enhanced autophagy and
reduced pyroptosis ([135]Chen et al., 2023). The histopathological
evaluation of our porcine wounds showed less inflammation in the wounds
receiving pAD-MSCs. The potential of miR-155 to mitigate the
inflammatory response and elicit a positive regenerative effect in the
early phases of wound repair constitute an interesting premise to
explore in the future.
The proteome screening showed upregulation of FGF2, FGF7 and FGF19 in
AD-MSCs. These growth factors play a crucial role in various
regenerative processes, including cell proliferation, differentiation
and migration, as well as mediating angiogenesis ([136]Maddaluno et
al., 2017). According to experimental target verification, both FGF2
and FGF7 are positively regulated by miR-155-5p in AD-MSCs. The
positive regulation role of miR-155 was previously reported in B cells
and associated with specific antibodies production, as well as enhanced
calcium accumulation in response to anti-IgM antibody. The suggested
mechanism was through the inhibition of SHIP1, a tyrosine kinase
negative regulator. In addition, miR-155 has been shown to enhance
TNF-α and IL-6 in macrophages ([137]Mashima, 2015). FGF2 was among the
most abundant angiogenesis inducers detected in mature human adipose
tissue extracts as well in AD-MSCs-released exosomes ([138]Sarkanen et
al., 2012; [139]Kim et al., 2017). FGF2 is also recognized as a key
supplement in AD-MSCs culture medium as FGF2 enhances proliferation and
maintain the cell stemness. FGF2 depletion in serial AD-MSCs
cultivation has been shown to induce autophagy and senescence while
suppressing stemness genes ([140]Ma et al., 2019; [141]Cheng et al.,
2020). In the same context, FGF7 was also upregulated in AD-MSCs and
was shown to be a direct target for miR-155-5p. FGF7 is generally
recognized as keratinocyte growth factor and was found to be abundant
among the soluble paracrine factors detected in adipose tissue, as it
is normally produced by cells of mesenchymal origin ([142]Gabrielsson
et al., 2002). [143]Ceccarelli et al. (2018) demonstrated that human
AD-MSCs showed elevated expression of FGF7 during early phases of
adipogenic differentiation. In keratinocytes, FGF7 is a major regulator
that promotes Tumor Necrosis Factor Alpha (TNF-α) in human
keratinocytes through the FGFR2–AKT–NF–κB signalling axis. FGF7 targets
epithelial cells and exerts an important role in modulating cellular
processes such as proliferation and migration, in addition to
vasculogenesis, and regeneration of various organs ([144]Ceccarelli et
al., 2018). Primarily secreted by mesenchymal cells, FGF7 exerts its
effect through paracrine signalling on epidermal keratinocytes but is
not necessarily secreted by them. FGF7 binds to the KGFR2IIIb receptor
found exclusively on keratinocytes, and promotes the migration,
proliferation and differentiation of epidermal keratinocytes triggering
the inflammatory cascade during wound healing ([145]Belleudi et al.,
2010; [146]Pastar et al., 2014).
Both FGF2 and FGF7 are crucial factors for AD-MSCs to exhibit their
regenerative properties in wound healing, enabling epithelial tissues
to restore structure and function. Both have been reported to be
implicated in the early and late stages of wound healing including
formation of granulation tissue, re-epithelialization as well as the
remodelling phase of wound healing ([147]Werner et al., 1994;
[148]Demidova-Rice et al., 2012; [149]Prudovsky, 2021). FGF2 has
important roles in many cells including dermal fibroblasts,
keratinocytes, endothelial cells, and melanocytes. FGF2 stimulates
angiogenesis and enhances the production of ECM proteins, which are
essential for tissue repair in the formation of granulation tissue. The
mitogenic and angiogenic properties of FGF2 enable its role in tissue
remodelling and neovascularization ([150]Lee et al., 2021). FGF2 is
known to induce neoangiogenesis during tissue repair, tissue
engineering and wound healing through stimulating the differentiation
of AD-MSCs into endothelial cells ([151]Mazini et al., 2020), or
through enhancing the ingrowth of blood vessels by directly
incorporating FGF2 in bioengineered scaffolds ([152]Shahin et al.,
2020). At the wound site, the remodelling effect of FGF2 takes place as
a result of efficient inhibition of fibroblast terminal differentiation
to myofibroblast, which is a key mediator when it is activated in
keloids and hypertrophic scars ([153]Akita et al., 2013).
Another significantly upregulated protein in AD-MSCs and validated
target for miR-155-5p, is the proinflammatory cytokine Chemokine
(CC-motif) ligand 2 (CCL2) a.k.a Monocyte chemotactic protein-1 (MCP1).
In immune response, CCL2 promotes inflammation by binding to its
receptors C-C chemokine receptor type 2 (CCR2) and subsequently
activating monocytes recruitment as well as mediating neutrophil and
macrophage infiltration ([154]Khan et al., 2013; [155]Singh et al.,
2021). A recent study reported that addition of CCL2 directly promoted
adipogenesis by enhancing lipid accumulation in AD-MSCs as well as
accelerated angiogenesis by stimulating tube formation of Human
Umbilical Vein Endothelial Cells (HUVECs) in cell culture. The authors
argue that the latter effect is believed to be attributed to CCL2
upregulating the expression of key vascularization genes VEGF and
VEGFR2 as well as MMPs through the PI3K‐AKTpathway ([156]Zhu et al.,
2022). Conversely, from a wound healing perspective, CCL2 plays a major
role in its earlier phases as it modulates T cell differentiation
toward the Th2 subset ([157]Dipietro et al., 2001). It was among the
earliest and most upregulated genes during the acute inflammatory phase
of murine excisional wounds screened at 6–12 h post wounding, with an
order of ∼70 fold-change ([158]Roy et al., 2008). Recent evidence
highlighted the importance of Nuclear factor erythroid 2 (Nrf2)
together with CCL2 in the healing of epidermal defects in diabetic
mice. The study unveiled an immunogenic signalling network centred on
epidermal keratinocyte-macrophage crosstalk, via Nrf2/CCl2/EGF
signalling axis. Nrf2 promotes CCL2 expression to mediate macrophage
trafficking and direct macrophage production of epidermal growth factor
(EGF), which in turn activates the epidermal progenitors triggering an
early regenerative response after injury ([159]Villarreal-Ponce et al.,
2020). In another study, CCL2 showed discernible effect in treating
early-stage diabetic wounds as it stimulated the healing by
accelerating immune cell infiltration and restoring the macrophage
response. Inflammation subsided earlier in CCL2-treated wounds,
alleviating the otherwise persistent, hyper-inflammatory mode
characteristic of diabetic wounds in 10 days ([160]Wood et al., 2014).
VCAM1 is an endothelial adhesion molecule to constitute the 4th arm of
the proposed miR-155 mediated gene regulation in AD-MSCs. Otherwise
known as CD106 is a cell adhesion factor known to play a role in
regulating stem cell trafficking by modulating the homing or
mobilization of stem cells. Unlike in bone marrow-derived stem cells,
the pattern of AD-MSCs expressing VCAM1 is preserved ([161]Madonna et
al., 2009). Monocytes and lymphocytes can adhere to VCAM1 allowing them
to tag along, penetrating the endothelial surface and inducing a
circulating inflammatory response ([162]Yi et al., 2020). VCAM1
expression was found to be overexpressed at the wound edge at 48–96 h
post injury in murine excisional wounds ([163]Roy et al., 2008). The
pro-angiogenic effect of VCAM1 signalling was evident in bone defect
repair when treated with human serum-derived exosomes as VCAM1
positively regulates vascular endothelial growth factor (VEGF)
([164]Xiang et al., 2023). In our experimental target analysis, the
reported gene silencing effect of Anti-miR-155 on FGF2, FGF7, CCL2 and
VCAM1 reverted to the basal level at 72 h as the method used for
introducing miRNA to the cells was based on cationic-lipid
transfection. For obtaining stable transfection, DNA integration of
miR-155-5p loaded plasmid could have been used. However, transient
transfection was sufficient to prove the effect of miR-155-5p on the
proposed targets.
The rest of the upregulated panel in AD-MSCs can be grouped into
cytokines, growth factors, and chemokines, which can be involved in
inflammatory and vasculogenic processes. Both are closely intertwined
processes in the context of wound healing. Inflammatory cells often
release angiogenic factors, thereby exerting mitogenic and migratory
effects on the endothelium ([165]Yi et al., 2020). The list of
upregulated inflammatory mediators include: Pentraxin-related protein
(PTX3), Complement 5a anaphylatoxin chemotactic receptor 1 (C5AR1),
C-X-C motif chemokine ligand 5 (CXCL5), CD40 ligand (CD40LG), C-X-C
Motif Chemokine Ligand 12 (CXCL12), Hepatitis A Virus Cellular Receptor
2 (TIM-3/HAVCR2) and Transferrin receptor protein 1 (TFRC). The list of
upregulated proangiogenic factors and vasculogenic cytokines beside
FGFs and VCAM1 include: Angiogenin (ANG), platelet and endothelial cell
adhesion molecule 1 (PECAM1/CD31) and Endoglin (ENG/CD105). These
findings are in agreement with the well-known immunomodulatory role of
AD-MSCs in-vivo ([166]Mazini et al., 2020).
Furthermore, our pathway enrichment analysis highlighted the signal
transduction, inflammatory and immune response pathways. As
inflammatory cytokines, interleukins (IL)-1, 6, 8, and 17 are among the
initial factors to be produced in response to skin wounds in order to
participate in the inflammatory phase of wound healing ([167]Xiao et
al., 2020). IL-8, among other cytokines, attracts inflammatory
polymorphonuclear cells (PMNs) in large numbers within the first
24–48 h at the wound site. PMNs in turn act as a major producer of
proinflammatory cytokines, including IL-1α, IL-1β, IL-6, and TNF-α.
2–3 days post injury, monocytes are recruited to the wound bed and
remain there for weeks. Monocytes transform into macrophages, and
secrete IL-1α, IL-1β, IL-6, and TNF-α to perpetuate the inflammatory
state. This inflammatory phase persists for the first 4 days in normal
wound healing and is crucial to initiate downstream repair and mediate
angiogenesis ([168]Kanji and Das, 2017). Although the cytokines
produced by AD-MSCs are involved in the inflammatory signalling
cascades, the established immunomodulatory role of exogenous AD-MSCs in
wound repair is anti-inflammatory in nature. AD-MSCs can downregulate
the immune response through 3 main ways 1) secreting immunosuppressive
factors such as IL-10 and transforming growth factor-beta (TGF-β) which
can promote maturation of suppressor T cells, 2) reducing dendritic
cell maturation, and 3) inhibiting the proliferation of natural killer
(NK) cells ([169]Kokai et al., 2014; [170]Zhou et al., 2022). This
immunosuppressive feature promotes AD-MSCs as suitable candidates for
allogenic cell transplantations. AD-MSCs attenuate wound inflammation
by inhibiting the acute immune response of the host and inducing
healing, even in chronic inflammatory state. MAPK family signalling
cascade is another signalling pathway upregulated by AD-MSCs according
to our pathway enrichment analysis. P38/MAPK promotes ECM production
via positively regulating collagen synthesis ([171]Du et al., 2013). It
would be interesting to verify if the involvement of AD-MSCs in wound
repair is essential through differentiation into keratinocytes, or that
AD-MSCs secretions are sufficient to enhance the natural repair
mechanisms.
miRNA-based therapeutics are emerging as promising strategy with the
potential to improve wound healing. Efforts being devised in
preclinical studies in order to overcome the challenges associated with
miRNA therapies exemplified in guiding sufficient dosing and developing
efficient delivery methods that maintain miRNA stability in the wound
environment ([172]Banerjee and Sen, 2015; [173]Veith et al., 2019). The
synthesis of the double stranded miRNA mimics enhanced their stability
as well as their potential for clinical use. To avoid degradation and
enhance targeting, miRNA can be conjugated with lipid particles or
loaded into specific viruses or nanoparticles ([174]Ho et al., 2022).
Skin wounds have the advantage of being accessible. The strategy of
applying miRNA in an ointment, as a cell-free therapy, can be feasible
for patients, decrease the incidence of non-target effects and allow to
identify local side effects ([175]Inoue et al., 2020). Furthermore,
pre-treatment of stem cells with epigenetic modulators can enhance
their response to the differentiation conditions ([176]El-Serafi and
Hayat, 2012). It would be interesting to investigate the effect of
transient transfection with miR-155 on stem cell differentiation
efficiency into keratinocyte-like cells. These cells can be
transplanted at the wound as miR-155-primed cell-based therapy.
5 Conclusion
This study provides in-vivo evidence that corroborates the restorative
effect of AD-MSCs in full-thickness excisional wounds. AD-MSCs
accelerated wound closure and achieved satisfactory epithelization in
14 days compared to naturally healed wounds as shown with histological,
morphometric and SDFI evaluation. The latter can be considered as a
non-invasive technique that provides read-outs indicative of tissue
repair during the healing process. Future studies should validate
possible correlations between SFDI readings, ideally with a higher
number of wounds. Our computational analysis of the differentially
expressed miRNAs and proteins in human AD-MSCs and keratinocytes
predicted that miR-155 may potentiate the immunomodulatory effect of
AD-MSCs by positively regulating the key proteins FGF2, FGF7, CCL2, and
VCAM1. Furthermore, the predicted regulation was validated
experimentally through transfection with miR-155 inhibitor and
confirmed a positive regulation between miR-155 and the identified four
factors. Each of these factors carries out key functions at different
events within the wound healing process including vascularization,
inflammation, proliferation, and remodelling.
As a follow up study, the therapeutic potential of miR-155 or AD-MSCs
overexpressing miR-155 should be investigated in-vivo, particularly in
difficult-to-heal wound models. However, two critical aspects about
miR-155 must be taken into consideration: firstly, to elucidate its
immunomodulatory effect in order to eliminate the risk of exacerbating
the inflammatory state at the wound site. Secondly, the underlying
mechanisms implicating miR-155 expression with cellular senescence
reported in the context of aged mesenchymal stem cells. Future work
should explore the potential feedforward/back loops affecting how
miR-155 regulates the expression of FGF2, FGF7, CCL2, and VCAM1 and how
these proteins interact with one another, in addition to the potential
reciprocal effects of these proteins on miR-155 expression. Lastly, the
intermediary proteins potentially underlying the positive regulation
between miR-155 and the identified proteins have not been studied in
this report and constitute a limitation of this study.
Acknowledgments