Abstract
Diabetic wound is a significant clinical challenge, and stem cell
therapy has shown great potential. This study explores the role of
mesenchymal stem cell (MSC) spheroids (Sp‐MSCs) in healing diabetic
wounds and the use of autologous plasma‐rich platelet fibrin (PRF) as a
scaffold for Sp‐MSCs. Through activation of the coagulation system, PRF
offers a protective fibrin shield for Sp‐MSCs to promote the rapid
recovery migration and proliferation of MSCs while maintaining the
activity of Sp‐MSCs in an inflammatory overload environment by
activating the related genes of Integrin‐β1‐vascular endothelial growth
factor (VEGF), and Wnt/β‐catenin pathways. The inclusion of Sp‐MSCs
accelerates the gelation of PRF and results in improved mechanical
strength. Additionally, PRF enhances the repair function of Sp‐MSCs,
creating a favorable microenvironment for angiogenesis. In the wound
model of diabetic mice, the combination of PRF with Sp‐MSCs accelerates
wound healing. Results show that this combination significantly
promotes wound repair and regulates the immune microenvironment. The
study suggests that PRF is a promising bio‐derived scaffold for stem
cell applications in diabetic wounds, offering new directions for stem
cell therapy and biomimetic scaffold material development.
Keywords: cell spheroid, diabetic wound, mesenchymal stem cell,
platelet‐rich fiber, ROS protection
__________________________________________________________________
Autologous plasma‐derived platelet‐rich fibrin (PRF) is prepared as a
protective shield for mesenchymal stem cell spheroids (Sp‐MSCs). PRF
forms a fibrin shield to protect Sp‐MSCs from the oxidative stress
environment. The nutrients in PRF, particularly the α‐granules, can
enhance the repair function of Sp‐MSCs. This opens up new possibilities
for the clinical application of stem cells to treat diabetic wounds.
graphic file with name ADVS-12-2413430-g008.jpg
1. Introduction
Diabetes mellitus represents a burgeoning global health challenge,
affecting an estimated 9.8% of the worldwide population.^[ [44]^1 ,
[45]^2 ^] A prevalent complication of this metabolic disorder is the
development of diabetic wounds, which affect approximately 15% of
individuals with diabetes and often lead to the formation of chronic,
non‐healing ulcers.^[ [46]^3 ^] The prognosis for these ulcers is grim,
with a long‐term mortality rate reaching 30%, and for those undergoing
amputation, the rate soars to over 70%.^[ [47]^4 ^] Traditional
therapeutic strategies, including self‐monitoring of blood glucose,
wound debridement, flap transplantation, and revascularization, are
often hindered by the complex microenvironment of diabetic wound beds,
thereby limiting their efficacy in fostering wound healing.^[ [48]^4 ^]
Evidence suggests that the protracted healing of diabetic ulcers is
intricately linked to a paucity of mesenchymal stem cells (MSCs) at the
wound site.^[ [49]^5 ^] This deficiency culminates in a dearth of
essential growth factors and nutrients, which in turn amplifies the
detrimental effects of inflammation and reactive oxygen species on the
wound healing process.^[ [50]^6 ^]
Consequently, mesenchymal stem cells (MSCs), endowed with the
capacities for self‐renewal and multilineage differentiation, emerge as
a potent therapeutic approach for the management of diabetic wounds.^[
[51]^7 ^] These cells foster tissue regeneration and cellular survival
through their differentiation into dermal components, secretion of
cytokines, and the release of extracellular vesicles, all of which
contribute to the modulation of the inflammatory milieu.^[ [52]^8 ^]
MSCs derived from human embryonic stem cells have illustrated their
efficacy across a spectrum of disease models, attributable to their
stable characteristics and amenability to large‐scale expansion.^[
[53]^9 , [54]^10 , [55]^11 ^] Moreover, MSC spheroids (Sp‐MSCs)
demonstrate augmented capabilities in replicating in vivo conditions,
tissue reparative processes, and immune modulation, owing to their
three‐dimensional architecture and intensified intercellular
interactions. This renders Sp‐MSCs a superior cell therapeutic strategy
in the realm of regenerative medicine and tissue engineering.^[ [56]^8
, [57]^12 ^] Additionally, this approach is not only cost‐effective but
also enhances the preservation, transportation, and the expedited
restorative and healing attributes of the cells post‐transplantation.^[
[58]^12 , [59]^13 ^]
In the context of diabetic wounds, the pathological microenvironment is
characterized by elevated levels of inflammatory mediators and reactive
oxygen species (ROS) that arise from hypoxia‐ischemia.^[ [60]^14 ^]
Extensive research has demonstrated that the survival rate of engrafted
stem cells is strikingly low, primarily due to the deleterious impact
of inflammation and ROS.^[ [61]^15 , [62]^16 , [63]^17 , [64]^18 ^]
These factors significantly compromise the functionality and viability
of the transplanted mesenchymal stem cells (MSCs), thereby diminishing
the therapeutic potency of stem cell transplantation.^[ [65]^18 ,
[66]^19 , [67]^20 ^] It is, therefore, imperative to establish a
protective shield and nutrient reservoir for MSCs during the repair of
diabetic wounds. This sanctuary is essential for shielding the cells
from inflammation and ROS‐induced injury and for ensuring a continuous
supply of nutrients and growth factors necessary for wound healing.
Consequently, there is a pressing need for the development of a novel
scaffold material capable of orchestrating and sustaining diverse
bio‐microenvironments throughout the healing cascade, thereby
augmenting the efficacy of stem cell therapies. Recently, a variety of
bioactive carriers have been engineered and deployed in stem cell
therapy. These carriers protect MSCs from the intricate and often
hostile microenvironment present at disease sites, thereby empowering
the cells to effectively execute their reparative functions in tissue
regeneration.^[ [68]^8 , [69]^18 ^] Hydrogel materials have emerged as
efficient vehicles for the delivery of MSCs; moreover, they have been
shown to modulate the proliferative activity and stemness of these
cells.^[ [70]^21 ^] Hence, there is a critical need to develop a stem
cell carrier that could emulate the in vivo environment and closely
follow the biological properties of the native tissue. Such a carrier
would play a pivotal role in safeguarding and enhancing the recovery
migration and in preserving the active functionality of Sp‐MSCs, which
are crucial for effective regenerative medicine strategies.^[ [71]^22
^] Platelet‐Rich Fibrin (PRF), a bioactive matrix derived from the
lower layer of plasma following differential centrifugation of whole
blood and subsequent thrombin activation.^[ [72]^23 ^] As a source of
autologous plasma gel, PRF is widely used in the repair of various
tissue injuries.^[ [73]^24 ^] Compared to synthetic materials, PRF is
prepared on autologous tissue, and has no immune rejection reaction.
Meanwhile, the fibrin network in PRF is closer to the tissue, and
provides a good scaffold structure for tissue regeneration and cell
adhesion at the site of wounds.^[ [74]^25 ^] The abundant growth
factors contained in PRF play a positive role in inflammation
regulation and promoting tissue regeneration which play important roles
in accelerating and high‐quality tissue healing.^[ [75]^26 ^]
Therefore, PRF has the potential to become a scaffold carrier for
delivering stem cells for wound healing.
In the context of diabetic wound healing, the hostile microenvironment,
replete with inflammatory cytokines and reactive oxygen species (ROS)
stemming from hypoxia‐ischemia, poses a significant challenge for stem
cell survival. To counteract this, the present study introduces We have
developed a novel approach by coating Spheroid Mesenchymal Stem Cells
(Sp‐MSCs) with PRF gel, thereby creating a protective shield around
Sp‐MSCs. This PRF‐coated Sp‐MSC construct is anticipated to withstand
the ROS‐mediated assault in the inflammatory milieu characteristic of
diabetic wounds, while also providing a biomimetic scaffold that
facilitates Sp‐MSCs colonization and adherence within the wound bed.
This strategy is designed to preserve the viability and enhance the
reparative functions of Sp‐MSCs.
2. Result
2.1. The Fabrication and Characterization of Sp‐MSCs
Here, we use the cell spheres forming plate to prepare mesenchymal stem
cell spheroids (Sp‐MSCs) (Figure [76]1A,a1–a4). To detect Sp‐MSCs
activity, we re‐plate Sp‐MSCs onto the culture plate. MSCs in Sp‐MSCs
recover and crawl out of the spheroids (Figure [77]1A,a5,a6). By
controlling the cell seeding density, Sp‐MSCs expected to contain
different cells (1000/5000/15 000) could be prepared. The results
showed that 1000/5000 cells/spheroid had a higher spheroid formation
rate than 15 000 cells/spheroid (Figure [78]1B,b1). And 5000
cells/spheroid had a lower free cell rate than the other groups
(Figure [79]1B,b2). By measuring the diameter of Sp‐MSCs, we found that
5000 cells/spheroid had the most stable spherical diameter (318.57 ±
10.99 µm). The diameter of 1000 cells/spheroid was smaller (207.08 ±
20.82 µm), and more free cells can be seen in wells. The 15 000
cells/spheroid had a large number of free cells and the Sp‐MSCs had an
uneven diameter (356.07 ± 137.85 µm) (Figure [80]1B,b3; Figure [81]S1A,
Supporting Information). Overall, 5000 cells/spheroid exhibited the
most stable spherical form and the highest cell spheroidization rate.
Therefore, we chose 5000 cells/spheroid for our subsequent experiments.
The live/dead staining revealed that the Sp‐MSCs were stored in MSC
culture medium for 3 days at 25 °C; the vast majority of cells
survived, with only a tiny percentage of cells dying in the middle
(Figure [82]1D; Figure [83]S1B, Supporting Information). MSC surface
antibodies characterized were evaluated using immunofluorescence and
flow cytometry to verify the effect of Sp‐MSCs formation on stem cell
stemness. MSCs markers CD73/CD90 remain highly positive in the sphere
state of cells (Figure [84]1C; Figure [85]S1C, Supporting Information).
Flow cytometry results revealed that compared to MSC, the expression
levels of stem cell surface antibodies CD44, CD73, CD90, and CD105, and
did not show significant changes after Sp‐MSCs dissociated into
individual MSC (Figure [86]1F). After the formation of Sp‐MSCs, stem
cells aggregated into clusters with stable structures and are not
easily loose. Sp‐MSCs slice staining and cytoskeleton staining showed
that the stem cells in Sp‐MSCs were tightly bound (Figure [87]1E).
Sp‐MSCs were stored in an MSC culture medium at 25 °C, and the diameter
of Sp‐MSCs gradually decreased with the extension of storage days
(Figure [88]S1D, Supporting Information). After 7 days of storage, the
diameter of Sp‐MSCs decreased to around 200 µm, and the difference of
Sp‐MSCs’ diameter increased (Figure [89]S1E, Supporting Information).
Sp‐MSCs can still recover and crawl out after 7 days of storage (Figure
[90]S1F, Supporting Information).
Figure 1.
Figure 1
[91]Open in a new tab
The fabrication and characterization of Sp‐MSCs. A) The process of cell
spheroids forming plate method for Sp‐MSCs. a1–a4) Add MSCs into cell
spheroids forming plate wells, spheroids forming, Sp‐MSCs collection.
a5,a6) Image of Sp‐MSCs recovery and crawling out. (Scale bar =
200 µm). B) 5000 cells/spheroid had the most stable structure and the
highest sphere formation rate. b1) sphere formation rate of different
sizes Sp‐MSCs (n = 3), b2) Free cells rate of different sizes Sp‐MSCs
(n = 3), b3) The diameters of different sizes Sp‐MSCs (n = 8),
(*p < 0.05, **p < 0.01). C) Stemness unchanged in Sp‐MSCs. Stem cell
markers (CD90, green; CD73, red; DAPI, blue) of the cells in the
Sp‐MSCs (Scale bar = 100 µm). D)The microscopy images of live/dead
stained Sp‐MSCs after 3 days in vitro culture showed no significant
cell death in Sp‐MSCs (Scale bar = 100 µm). E) H&E staining, Masson
staining, Phalloidin staining of Sp‐MSCs (Scale bar = 100 µm). F) Flow
cytometry of CD44, CD73, CD90, and CD105 of MSCs before forming
Sp‐MSCs, and Sp‐MSCs dissociate into individual cells.
2.2. PRF Combined with Sp‐MSCs, Accelerated Coagulation, and Enhanced
Strength
PRF was prepared by differential centrifugation method. Sp‐MSCs were
then combined directly with PRF (Figure [92]2A,B; Figure [93]S2A,
Supporting Information). After coating Sp‐MSCs with PRF, PRF
crosslinked with Sp‐MSCs (Figure [94]2C; Figure [95]S2B, Supporting
Information), and the gelation time was significantly shortened
(Figure [96]2D,E). The addition of Sp‐MSCs activated the exogenous
coagulation system, which made PRF coagulate into gel faster. Early
application of PRF helps maintain active factors in PRF, and the
shorter gelation time makes clinical application more convenient.
Viscosity measurements revealed that PRF exhibited
temperature‐sensitive properties. The addition of Sp‐MSCs shortened the
gelation time of PRF and increased the viscosity of PRF while
maintaining its temperature‐sensitive properties. Increasing
temperature from 4.0 to 37.0 °C, the viscosity of PRF gel increased
from 347.366 to 4194.887 η mPas^−1, while adding Sp‐MSCs into PRF, the
viscosity increased from 16 810.05 to 786 936.4 η mPas^−1
(Figure [97]2G). The stimulation of the coagulation system promotes
adhesion. High adherence promotes Sp‐MSC colonization in the wound site
and prevents Sp‐MSCs loss due to local administration. PRF and
PRF+Sp‐MSCs both exhibited hydrogel‐like behavior with elastic modulus
(Gʹ) consistently surpassing the loss modulus (Gʹʹ). And the addition
of Sp‐MSCs significantly enhanced the mechanical strength of PRF
(Figure [98]2G). The result of scanning electron microscopy (SEM)
showed that PRF contained a large number of fibrin components that
crosslinked into a network, and Sp‐MSCs adhered to fibers
(Figure [99]2F). After coating Sp‐MSCs, the fibers in PRF aggregated
into bundles. And the fiber bundle diameter thickened from 1.15 ±
1.02 µm to 5.46 ± 2.43 µm by measuring the diameter of PRF fibrin
through SEM images (Figure [100]3J), which revealed that the addition
of Sp‐MSCs enhanced mechanical strength of PRF. The H&E and Masson
staining results showed that Sp‐MSCs were disseminated in PRF fibers.
And due to the activation of PRF coagulation by Sp‐MSCs, fibrin in PRF
aggregated around Sp‐MSCs, forming a fibrin protective shield.
Activated coagulation enriched fibrin around Sp‐MSCs without disrupting
the structure of Sp‐MSCs (Figure [101]2H,I; Figure [102]S2B, Supporting
Information). While there was no significant change in the porosity or
collagen density of PRF fibers after coating Sp‐MSCs
(Figure [103]2K,L). PRF contained various growth factors, including
epidermal growth factor (EGF) was 3.58 ± 0.79 ng mL^−1,
platelet‐derived growth factor (PDGF) was 5.19 ± 0.33 ng mL^−1,
vascular endothelial growth factor (VEGF) was 1.43 ± 0.19 ng mL^−1, and
transforming growth factor β (TGF‐β) was 6.84 ± 3.08 ng mL^−1 (Figure
[104]S2C, Supporting Information).
Figure 2.
Figure 2
[105]Open in a new tab
PRF combined with Sp‐MSCs enhanced PRF strength. A) Diagram of PRF
preparation by differential centrifugation method and combined with
Sp‐MSCs. B) Image of PRF gel. C) Cross‐linking between PRF and Sp‐MSCs.
D) PRF combined with Sp‐MSCs, accelerated the coagulation and gelation
of PRF. E) Quantitative analysis of gelation time of PRF and
PRF+Sp‐MSCs (n = 3, **p < 0.01). F) Scanning electron microscope (SEM)
of PRF and PRF+Sp‐MSCs. G) The addition of Sp‐MSCs enhanced the
strength of PRF (Scale bar = 10 µm). G) The temperature‐viscosity curve
and rheological properties of PRF and PRF+Sp‐MSCs. H,I) H&E staining,
Masson staining revealed PRF‐coated Sp‐MSCs (Scale bar = 200 µm). J,K)
Quantitative analysis based on SEM image of fiber diameter and porosity
rate of fibers of PRF and PRF+Sp‐MSCs (n = 3, **p < 0.01, ns:
p > 0.05). L) Quantitative analysis based on Masson staining collagen
proportion of PRF and PRF + Sp‐MSCs (n = 3, ns: p > 0.05).
Figure 3.
Figure 3
[106]Open in a new tab
The protective efficacy of PRF for the encapsulated Sp‐MSCs. A)
Live/Dead staining showing the cell viability of exposed Sp‐MSCs and
encapsulated Sp‐MSCs in HA/PRF at 2 days during in vitro culture (Scale
bar = 100 µm). B) Live/Dead staining showing the cell viability of
exposed Sp‐MSCs and encapsulated Sp‐MSCs in HA/PRF at 2 days during in
vitro culture after treatment with 500 µm H[2]O[2] (Scale bar =
100 µm). C) Quantitative analysis based on live/dead staining of the
cell viability (n = 3, *p < 0.05, **p < 0.01, ***p < 0.001,
****p < 0.0001). D,E) The gene expression of Bax and the ratio of
Bax/Bcl2 (n = 3, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001,
ns: p > 0.05). F,G) Representative fluorescent images of total
intracellular ROS (stained by DCFH‐DA) in the exposed Sp‐MSCs and
encapsulated Sp‐MSCs in HA/PRF at 2 and 4 days during in vitro culture
after treatment with 500 µm H[2]O[2] (Scale bar = 100 µm). H)
Quantitative analysis based on DCFH‐DA staining of 2 and 4 days ROS
level (n = 3, **p < 0.01, ***p < 0.001, ****p < 0.0001). I,J) PRF
promoted proliferation of MSCs in 2D culture conditions, measured by
CCK‐8 assay (n = 3, **p < 0.01, ***p < 0.001). K,L) PRF promoted
Sp‐MSCs recovery and crawling out (n = 3, **p < 0.01, ****p < 0.0001).
2.3. PRF‐Shelter‐Protected Sp‐MSCs in Oxidative Stress Environment
Transplanted allogenic stem cells for diabetic wound healing are
frequently subjected to oxidative stress and excessive inflammatory
factors, which could cause irreparable damage to the transplanted cells
and endanger therapeutic efficacy.^[ [107]^17 ^]
Therefore, to verify the protective properties of PRF. We first
cultured Sp‐MSCs in medium containing severe oxidative stress (500µM
H[2]O[2])^[ [108]^27 , [109]^28 ^] to simulate the ROS microenvironment
in vitro. After 2 days and 4 days, the viability of Sp‐MSCs was
assessed using the live/dead assay. The oxidative stress level was
assessed using DCFH‐DA to determine the intracellular ROS levels of
Sp‐MSCs. To compare the protective effects of various scaffolding
materials. We selected hyaluronic acid (HA) as a typical scaffold
carrier. And compared the cellular activity and intracellular ROS
levels of Sp‐MSCs exposed and loaded in HA/PRF. Sp‐MSCs coated with PRF
had better activity and less ROS. Under general storage conditions,
Sp‐MSCs coated with HA exhibited more dead cells in the spheroids,
while exposed Sp‐MSCs and Sp‐MSCs coated with PRF showed fewer dead
cells. Less nutrients in HA lead to significant death of MSCs in
spheroid (Figure [110]3A,C). At the same time, the expression of
apoptotic gene Bax was significantly increased in Sp‐MSCs coated with
HA, and the Bax/Bcl2 ratio was also significantly increased compared to
the control group and Sp‐MSCs coated with PRF (Figure [111]3D). Under
oxidative stress conditions, exposed Sp‐MSCs caused massive dead cells
in spheroids. Sp‐MSCs coated with gels showed no significant increase
in dead cells. HA and PRF protected Sp‐MSCs from oxidative stress
environmental interference. While Sp‐MSCs coated with HA still
exhibited cell death in the spheroids (Figure [112]3B,C). And exposed
Sp‐MSCs showed higher expression levels of Bax and Bax/Bcl2 ratio than
the other group (Figure [113]3E). And after 4 days of storage, the
shape of Sp‐MSCs coated with HA was difficult to maintain, the sphere
was broken, MSCs dispersed in HA, and a large number of MSCs apoptosis
occurred. The exposed Sp‐MSCs appeared a small amount of dead cells.
PRF reduced the apoptosis of Sp‐MSCs while maintaining their shape
(Figure [114]S3A,C, Supporting Information). While under oxidative
stress conditions, PRF continued its excellent protective ability
(Figure [115]S3B,C, Supporting Information). In addition, DCF was
quantitatively analyzed for the assessment of intracellular ROS levels.
The results showed that in the oxidative stress environment, exposed
Sp‐MSCs consistently maintained high levels of ROS. And ROS levels
decreased when coated with HA and PRF. Although HA can provide a
physical barrier to assist Sp‐MSCs in resisting adverse environments.
The lack of nutrients in HA leads to broken sphere and extensive cell
apoptosis of Sp‐MSCs. Meanwhile, Sp‐MSCs coated with PRF showed the
lowest levels of ROS in spheroids (Figure [116]3F–H). The above results
indicated that PRF shielded Sp‐MSCs by enriching fibrin around Sp‐MSCs,
protect from adverse conditions while providing nutrients for
maintaining activity of Sp‐MSCs had the lowest cell death rate and ROS
levels in spheroids (Figure [117]3A–H; Figure [118]S3A–C, Supporting
Information).
Before re‐plating Sp‐MSCs onto the culture plate, the culture plate was
wrapped with HA/PRF. Sp‐MSCs were re‐plated onto culture dishes and
then covered with HA/PRF. After 24 h, the image showed that PRF
significantly promoted the recovery migration of Sp‐MSCs, accelerating
the speed of MSCs crawling out and migrating from the spheroids, but HA
had no significant promoting effect (Figure [119]3K,L). The fibrin
scaffold in PRF provided an effective scaffold for Sp‐MSCs recovery and
climbing. Growth factors in PRF also promoted MSC migration, and the
mechanical strength of the PRF gel was closer to the tissue. These
variables allowed Sp‐MSCs to remain active and recover more quickly in
PRF. The CCK‐8 analysis showed that PRF significantly promoted MSC
viability (Figure [120]3I,J). Under 2D culture conditions, HA had no
significant effect on individual MSC activity, proving that HA itself
had no significant toxicity in 2D cell culture. Under 3D conditions,
HA‐coated Sp‐MSCs, reducing the nutrients obtained inside Sp‐MSCs. The
above results indicated that PRF could protect Sp‐MSCs from oxidative
stress conditions, while maintaining Sp‐MSCs activity and promoting MSC
proliferation.
2.4. PRF‐Regulated Sp‐MSCs Secretion Activity and Foster Microenvironment for
Angiogenesis
Diabetic wounds are difficult to heal due to the reduction of new blood
vessels in the wound site, which causes slow local tissue
regeneration.^[ [121]^29 ^] Therefore, the formation of an
angiogenesis‐promoting microenvironment in the wound site is critical
for diabetic wound healing.^[ [122]^30 ^] To verify the effect of PRF
combined with Sp‐MSCs on angiogenesis, Sp‐MSCs were replated on PRF.
And the condition culture medium (CM) of Sp‐MSCs was collected and used
to culture human umbilical vein endothelial cells (HUVECs). Then
assessed proliferation, migration, and tube formation ability of HUVECs
(Figure [123]4A). The CCK‐8 assay showed that PRF+Sp‐MSCs significantly
promoted HUVEC proliferation activity. Compared to the control group,
PRF group and Sp‐MSCs group also had a promoting effect
(Figure [124]4D). The migration ability of HUVECs was assessed using
the scratching assay. PRF+Sp‐MSCs group showed a higher wound healing
rate than the other groups. PRF group and Sp‐MSCs group were higher
than control group (Figure [125]4B,E). In vitro tube formation capacity
was evaluated and the PRF+Sp‐MSCs group showed a higher number of
branches and the longest total branch length. The PRF group promoted
angiogenesis more effectively than the Sp‐MSCs group due to the rich
VEGF content in PRF (Figure [126]4C,F,G). These results collectively
demonstrated the capacity of PRF combined with Sp‐MSCs to promote
HUVECs migration and angiogenesis, creating an angiogenesis‐promoting
microenvironment. By measuring the growth factors secreted by Sp‐MSCs
after recovery. We found that compared to Sp‐MSCs, Sp‐MSCs treated with
PRF showed a certain enhancement in their ability to secrete EGF, VEGF,
and TGF‐β (Figure [127]S3D, Supporting Information). PRF accelerated
the recovery of Sp‐MSCs, enhanced their activity, and promoted their
secretion function. Together, they formed a microenvironment that was
suitable for local tissue repair of wounds, especially vascular
regeneration.
Figure 4.
Figure 4
[128]Open in a new tab
Effect of PRF combined with Sp‐MSCs on angiogenesis in vitro. A)
Schematic diagram of exploring the effect of PRF treated conditioned
medium (CM) on proliferation, migration, and tube formation of HUVECs.
B) PRF combined with Sp‐MSCs enhanced HUVECs migration (Scale bar =
200 µm). C) PRF combined with Sp‐MSCs created a microenvironment that
promoted angiogenesis (Scale bar = 200 µm). D) PRF combined with
Sp‐MSCs enhanced HUVECs proliferation (n = 3). E) Healing rate of
HUVECs in the scratch assay (n = 3, *p < 0.05, **p < 0.01,
***p < 0.001). F,G) Quantitative analysis of tube node and length of
tube formation results (n = 3, **p < 0.01, ***p < 0.001,
****p < 0.0001).
2.5. Exploration of the Transcriptomic Profile of PRF‐Coated Sp‐MSCs
The data presented above indicated PRF's protective efficacy in
shielding Sp‐MSCs from oxidative stress while also promoting
angiogenesis. Thus, to exploit the underlying mechanism of how PRF
exerts protective functions and enhances the repair functions of
Sp‐MSCs. RNA‐seq was used to analyze the transcriptomic changes of the
Sp‐MSCs with and without PRF.
Firstly, we investigated the transcriptomic profile of Sp‐MSCs
uncovered versus combined with PRF to explore the effects of PRF on
intracellular responses of MSCs. Generally, there were 3625
downregulated and 1506 upregulated all differentially expressed genes
(DEGs) between PRF‐coated Sp‐MSCs vs Sp‐MSCs (Figure [129]5B). Further,
Gene ontology (GO) enrichment analysis was performed for the
downregulated and upregulated DEGs, respectively. Kyoto Encyclopedia of
Genes and Genomes (KEGG) enrichment analysis was enriched in protein
synthesis and secretion, epithelial and neural regeneration, cell
cycle, and energy metabolism. After Sp‐MSCs bind to PRF, increased
protein secretion function and enhanced cellular activity of MSCs
(Figure [130]5C). KEGG enrichment analysis further exhibited
significant signaling pathways and indicated that the addition of PRF
enhances the biological activity and secretion function of MSCs. We
validated the secretion‐related genes VEGF and TGF‐β, as well as the
proliferation‐related gene Ki67. PRF‐coated Sp‐MSCs showed enhanced
secretion and proliferation activity. Integrins are a family of cell
adhesion receptors.^[ [131]^31 ^] Integrin‐β1 plays an important role
in MSC migration, paracrine secretion, and angiogenesis regulation.^[
[132]^32 , [133]^33 ^] PRF‐coated Sp‐MSCs showed a significant increase
in integrin‐β1 and VEGF expression (Figure [134]5F). Therefore,
activation of the Integrity pathway may play a key role in the enhanced
repair function of PRF‐coated Sp‐MSCs.
Figure 5.
Figure 5
[135]Open in a new tab
Exploration of transcriptome profile patterns and functional enrichment
analysis. A) The heatmap transcriptomic profiling of Sp‐MSCs,
PRF‐coated Sp‐MSCs, and Sp‐MSCs (H[2]O[2]), PRF‐coated Sp‐MSCs
(H[2]O[2]), (H[2]O[2] mean treatment with 500 µm H[2]O[2]). B) Volcano
map of differentially expressed genes of Sp‐MSCs and PRF‐coated
Sp‐MSCs. C) KEGG pathway enrichment analysis of differentially
expressed genes might take place between PRF‐coated Sp‐MSCs vs Sp‐MSCs.
D) Volcano map of differentially expressed genes of Sp‐MSCs and
PRF‐coated Sp‐MSCs treatment with 500 µm H[2]O[2]. E) KEGG pathway
enrichment analysis of differentially expressed genes might take place
between PRF‐coated Sp‐MSCs vs Sp‐MSCs treatment with 500 µm H[2]O[2].
F) The expression level of VEGF, TGF‐β, and Integrin‐β1 of Sp‐MSCs and
PRF‐coated Sp‐MSCs. G) The expression level of Ki67 and Bax of Sp‐MSCs,
PRF‐coated Sp‐MSCs and Sp‐MSCs(H[2]O[2]), PRF‐coated Sp‐MSCs(H[2]O[2]).
H) The expression level of β‐catenin, FZD2, PORCN, and Axin of Sp‐MSCs,
PRF‐coated Sp‐MSCs, and Sp‐MSCs(H[2]O[2]), PRF‐coated
Sp‐MSCs(H[2]O[2]). (n = 3, *p < 0.05, **p < 0.01, ***p < 0.001,
****p < 0.0001, ns: p > 0.05).
To further explore the protective effect of PRF under an inflammatory
microenvironment, we cultured PRF‐coated Sp‐MSCs vs exposed Sp‐MSCs in
a ROS‐rich environment in vitro and analyzed the change of the
transcriptome in MSCs. It was found that there were 4001 downregulated
and 1325 upregulated DEGs between the two groups (Figure [136]5D). KEGG
enrichment analysis was enriched in cell cycle and cell
division‐related functions (Figure [137]5E), indicated that PRF
protected Sp‐MSCs in an oxidative stress environment and maintained MSC
proliferation activity. In the oxidative stress environment, PRF‐coated
Sp‐MSCs maintained high expression levels of Ki67 while reducing the
expression of apoptosis‐related gene Bax (Figure [138]5G). According to
KEGG enrichment results (Figure [139]5C,E) PRF‐coated will activate
Wnt‐related pathway in Sp‐MSCs. Wnt/β‐catenin signaling plays a crucial
role in the migration of MSCs, apoptosis inhibition.^[ [140]^34 ,
[141]^35 ^] Meanwhile, activation of Wnt/β‐catenin pathway promotes the
proliferation of MSCs under both 2D and 3D conditions.^[ [142]^36 ^]
PRF‐coated Sp‐MSCs showed a significant increase of Wnt/β‐catenin
pathway‐related genes β‐catenin, Frizzled class receptor (FZD2) and
Porcupine homolog (PORCN), simultaneously reducing the expression of
Axis inhibitor (Axin), a negative regulator of Wnt/β‐catenin pathway
(Figure [143]5H). Therefore, PRF not only uses a fibrin network to coat
Sp‐MSCs to protect from adverse environment, simultaneously activated
Wnt/β‐catenin pathway, enhances the survival ability of Sp‐MSCs in
adverse environments, and reduced MSCs apoptosis.
2.6. PRF Combined with Sp‐MSCs Promoted Diabetic Wound Healing
In this study, we used a gene mutation mouse model (db/db mouse) to
establish a diabetic wound model. Subsequently, the diabetic mice were
allocated into four groups (Control/Sp‐MSCs/PRF/PRF+Sp‐MSCs) (Figure
[144]6A). To vividly show the process of wound healing, photos were
taken on days 1, 5, 9, 12, and 15 to record the changes in the wound
area (Figure [145]6B). Statistical results showed that among all groups
(Figure [146]6C), the PRF+Sp‐MSCs group had the highest wound healing
rate (97.7±0.38%) on day 15. The PRF group and Sp‐MSCs group had lower
healing rates (83.5% ± 5.18% and 83.4% ± 4.28%), while the healing rate
of the control group was only 69.6% ± 1.68% (Figure [147]6D).
Figure 6.
Figure 6
[148]Open in a new tab
Effects of PRF combined with Sp‐MSCs in promoting diabetic wound
healing in db/db mouse. A) Schematic diagram of establishing a diabetic
wound and the timeline of animal experiments to test the therapeutic
effect of PRF combined with Sp‐MSCs. B) Representative images of the
wound at days 1, 5, 9, 12, and 15 and the diagrams of time‐evolved
wound areas (Scale bar = 2 mm). C) Statistical analysis line chart of
the wound area. D) Statistical analysis of wound healing rate (n = 3,
*p < 0.05, ***p < 0.001). E) The engraftment of DiR‐labeled Sp‐MSCs
detected by ex vivo imaging at different time points
post‐transplantation. F) Quantification of the fluorescent signal
intensity of the transplanted cells at different time points (n = 3,
*p < 0.05, ns: p>0.05). G,H) H&E staining of wound tissues at day 15,
and quantification of the epidermis repair area of wounds at day 15
(Scale bar = 100/500 µm) (n = 3, *p < 0.05). I,J) Masson staining of
wound tissues at day 15, and quantification of the epidermis repair
area of wounds at day 15 (Scale bar = 100/500 µm) (n = 3, **p < 0.01,
***p < 0.001). K,L) Sirius red staining of wound tissues at day 15, and
quantification of type III collagen percentage of wounds at day 15
(Scale bar = 50/500 µm) (n = 3, *p < 0.05, **p < 0.01, ****p < 0.0001).
To better investigate the therapeutic effect of PRF combined with
Sp‐MSCs, transplanted Sp‐MSCs were labeled with a far‐red plasma
membrane fluorescent probe and then seen applying ex vivo imaging
analysis. Notably, Sp‐MSCs coated with PRF exhibited a higher cell
viability. The fluorescent signals from the PRF‐coated Sp‐MSCs were
higher than Sp‐MSCs. Indicating the in vivo cytoprotective effect of
PRF, which protects Sp‐MSCs from the complex immune microenvironment
(Figure [149]6E,F).
To explore the effect of PRF combined with Sp‐MSCs on the immune
regulation of the inflammation phase and regeneration of the
proliferation phase during diabetic wound healing. H&E staining were
applied to skin specimens on day 5 to observe the overall condition of
wound healing. PRF+Sp‐MSCs group had migrated epithelial tissue around
the wound site. While there was no regenerated epithelium found at the
wound site of the other three groups (Figure [150]S4A, Supporting
Information). It indicated that PRF combined with Sp‐MSCs enabled
diabetic wounds to enter the proliferation phase earlier. In the early
stage of diabetic wound healing, the reconstruction of vascular
regeneration microenvironment plays a vital role in accelerating wound
healing.^[ [151]^37 ^] Further, immunofluorescence staining of CD34,
α‐smooth muscle actin (αSMA), and VEGF showed PRF combined with Sp‐MSCs
created a positive microenvironment for angiogenesis for diabetic wound
healing (Figure [152]S4B,D, Supporting Information). More VEGF and
vascular regeneration cells will make revascularization and faster
tissue repair during the proliferation and remodeling phase. Meanwhile,
PRF combined with Sp‐MSCs jointly regulated the inflammation phase.
PRF+Sp‐MSCs reduced neutrophil recruitment and the expression of
pro‐inflammatory cytokines like tumor necrosis factor‐alpha (TNF‐α),
and promoted the macrophage polarization towards M2 phenotype (Figure
[153]S4C,E, Supporting Information). PRF combined with Sp‐MSCs
effectively regulates the complex immune microenvironment of diabetic
wounds, and shortens the inflammatory phase.
Then we performed histopathological analysis on day 15. H&E and Masson
staining were applied to skin specimens to observe the collagen fibers
and appendage formation in the wound. The newly generated epithelial
tissue in the PRF+Sp‐MSCs group had a larger repair area
(Figure [154]6G,H) and a higher proportion of new collagen coverage
(Figure [155]6I,J). In the newly formed collagen of the skin. Type III
collagen is more expressed in the dermis of newly created skin
collagen, providing the skin with flexibility. And type I collagen is
densely arranged and has a relatively hard texture.^[ [156]^38 ^]
Sirius red staining was utilized to investigate type I and III collagen
on the border of the wound, and the result showed that the PRF+Sp‐MSCs
group showed a larger proportion of type III collagen. The newborn
collagen of the control group was mainly type I. The PRF+Sp‐MSCs group
exhibited more collagen regeneration and a collagen ratio closer to
normal skin (Figure [157]6K,L).
2.7. PRF and Sp‐MSCs Regulated the Inflammatory Environment of Diabetic
Wounds and Promoted Tissue Repair
Tissue immunofluorescence staining was performed to investigate
regeneration and immunological regulation in the wound site of day 15.
Immunofluorescence staining of CD34 and αSMA showed that PRF+Sp‐MSCs
group and Sp‐MSCs group promoted angiogenesis in vivo (Figure
[158]7A,C). Growth‐associated protein‐43 (GAP43) staining was used to
observe nerve tissue regeneration, and both PRF+Sp‐MSCs group and
Sp‐MSCs group showed more nerve tissue in the wound site
(Figure [159]7A,C).
Figure 7.
Figure 7
[160]Open in a new tab
The effect of PRF combined with Sp‐MSCs on tissue regeneration and
immune regulation of diabetic wound healing. A) Immunofluorescence of
αSMA and CD34 immunostaining showed accumulation neovascularization,
and GAP43 immunostaining showed neuroregeneration at the wound on day
15 (Scale bar = 100 µm). B) Immunofluorescence of TNF‐α (white arrow)
immunostaining showed accumulation of pro‐inflammatory factors, CD206
immunostaining showed accumulation of M2 macrophages, and H3Cit
immunostaining showed neutrophil recruitment at the wound on day 15
(Scale bar = 100 µm). C,D) Quantitative analysis based on
immunostaining of αSMA, CD34, GAP43, CD206, TNF‐α, H3Cit (n = 3,
*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns: p > 0.05).
Immunofluorescence results indicated that PRF+Sp‐MSCs group and Sp‐MSCs
group had significantly higher rates of M2 type macrophages (CD206) and
reduced inflammatory factors such as TNFα in the wound site
(Figure [161]7B,D). While the immune regulatory ability of the PRF
group was better than that of the control group. But fell short of
PRF+Sp‐MSCs group and Sp‐MSCs group. It was indicated that Sp‐MSCs
played a role in regulating immunity and accelerating the
transformation of wounds into anti‐inflammatory phenotypes.
PRF+Sp‐MSCs, PRF, and Sp‐MSCs group reduced neutrophil recruitment than
the control group. (Figure [162]7B,D). These results collectively
demonstrated that PRF combined with Sp‐MSCs could regulate the immune
environment in the wound site while also promoting tissue repair and
regeneration.
3. Discussion
Diabetic wounds are among the most severe complications of chronic
diabetes, and stem cell therapy has emerged as a promising treatment
approach.^[ [163]^39 , [164]^40 , [165]^41 ^] However, precisely
delivering mesenchymal stem cells (MSCs) to the wound site and
maintaining their activity remain challenging.^[ [166]^42 ^] The local
microenvironment in diabetic wounds is characterized by the aggregation
of inflammatory cells, an increase in pro‐inflammatory factors, and a
deficiency in blood supply and nourishment. It is difficult for
exogenous therapeutic stem cell drugs to remain active in the diabetic
microenvironment, which significantly reduces their reparative effects.
To address this issue, our study employed PRF gel, derived from
autologous plasma, to coat Sp‐MSCs for the treatment of diabetic
wounds.
3.1. Sp‐MSCs Accelerated PRF Coagulation and Enhanced Mechanical Strength of
PRF
Sp‐MSCs were prepared using an optimized method involving the formation
of spheroids containing 5000 MSCs each, and we have previously
confirmed their excellent performance in vitro and in vivo.^[ [167]^12
^] In this study, we verified these findings (Figure [168]2).
Then we prepared PRF gel from autologous plasma by differential
centrifugation and coated Sp‐MSCs with PRF (Figure [169]2A). PRF
contains various bioactive compounds from platelets consisting of
chemokines, growth factors, and cytokines (Figure [170]S2C, Supporting
Information). These could activate angiogenesis, alter extracellular
matrix (ECM) remodeling, and promote cell recruitment and
proliferation,^[ [171]^26 ^] as well as act against bacterium
infections.^[ [172]^43 ^] Furthermore, PRF is entirely derived from the
autologous plasma of patients, has high biocompatibility, and produces
no immune rejection reaction. We coated Sp‐MSCs with PRF
(Figure [173]2). The addition of Sp‐MSCs shortened the gelation time of
PRF and enhanced the adhesion and mechanical strength of PRF
(Figure [174]2D,E). Improved the shortcomings of prolonged coagulation
time and lower PRF mechanical strength (Figure [175]2G). It is
conducive to accurate delivery of Sp‐MSCs at the wound site. Upon
interaction with PRF, Sp‐MSCs trigger the coagulation cascade of PRF,
subsequently augmenting its mechanical strength. (Figure [176]2D–L).
3.2. PRF Shielding Sp‐MSCs Against ROS and Noxious Inflammatory Factors
In the treatment of diabetic wounds, Sp‐MSCs necessitate a carrier
capable of precisely targeting the wound site, safeguarding the cells
from the detrimental ROS environment, and preserving their viability.
PRF serves as a scaffold for Sp‐MSCs, more importantly, with its
intricate network of cross‐linked fibrin offering a robust shield for
their growth (Figure [177]2H,I). Fibrin in PRF exhibits a significant
capacity to counteract ROS (Figure [178]3B–H). Moreover, PRF is
enriched with growth factors and plasma components that are essential
for sustaining and enhancing the functional activity of Sp‐MSCs
(Figures [179]S2C and [180]S3A, Supporting Information). The
application of PRF to diabetic wounds has been observed to modulate the
immunological microenvironment, thereby creating a protective shield
for Sp‐MSCs in adverse environment.^[ [181]^43 , [182]^44 ^]
Furthermore, PRF has demonstrated the ability to expedite the migratory
recovery and extrusion of Sp‐MSCs (Figure [183]3K,L), thereby enabling
a swift initiation of tissue reparative processes. Sp‐MSCs coated with
PRF have been shown to withstand the intricacies of the diabetic wound
microenvironment, protect the cells from ROS, and consequently
augmenting their survival duration and rate within the wound
(Figure [184]6E,F). Transcriptomic analysis has revealed that, PRF can
activate related genes of Integrin‐β1‐VEGF pathway of MSCs, thereby
promoting MSCs migration, proliferation, and secretion, enhancing their
tissue repair function (Figure [185]5B,C,F). At the same time, under
oxidative stress conditions, PRF can activate the related genes to
Wnt/β‐catenin pathway to protect MSCs activity and reduce apoptosis
under adverse conditions (Figure [186]5B–E,G,H). In the complex
microenvironment of diabetic wounds, PRF provides physical protection
for Sp‐MSCs, mitigating the impact of adverse environment, and
concurrently orchestrates a transition towards an anti‐inflammatory
state by immune regulation (Figure [187]7B; Figure [188]S4C, Supporting
Information).^[ [189]^43 , [190]^44 ^] PRF serves as an adept shield
for Sp‐MSCs, safeguarding them against ROS and noxious inflammatory
factors.
3.3. PRF Enhanced Repair Functions of Sp‐MSCs
PRF serves as a protective shield and nurturing environment for
Sp‐MSCs. The growth factors and plasma components within PRF,
particularly the alpha‐granules, create a conducive nutritional milieu
for Sp‐MSCs.^[ [191]^45 ^] Sp‐MSCs' recovery migration was accelerated
(Figure [192]3K,L), and their proliferation activity was increased
(Figure [193]3I,J). We believe this is closely related to the presence
of a multitude of reparative factors in PRF, For instance,
platelet‐derived growth factors A and B (PDGF‐A and B), transforming
growth factor β (TGF‐β), and vascular endothelial growth factor (VEGF)
(Figure [194]S2C, Supporting Information).^[ [195]^46 , [196]^47 ^] The
growth factors released from PRF could effectively enhance Sp‐MSCs
repair function. PRF could not only accelerate the recovery and
migration of Sp‐MSCs and enhance the proliferative activity of MSCs,
but also enhance MSC paracrine function ability to repair and
regenerate tissues (Figure [197]4). The reduction of
neovascularization, which leads to insufficient local blood supply is
the main factor of difficult wound healing in diabetes.^[ [198]^48 ^]
The growth factors released from PRF could effectively promote vascular
regeneration and tissue repair (Figure [199]S3D, Supporting
Information).^[ [200]^49 ^] Meanwhile, PRF enhanced the metabolic
activity and secretion function of Sp‐MSCs. It has been found that the
combination of PRF with Sp‐MSCs could establish a microenvironment that
fosters vascular regeneration and facilitates tissue repair.^[ [201]^50
, [202]^51 , [203]^52 , [204]^53 ^] Sp‐MSCs coated with PRF gel
enhanced the proliferation and migration ability of surrounding cells
through secretion (Figure [205]4B,D,E). PRF‐coated Sp‐MSCs also formed
a favorable microenvironment for angiogenesis (Figure [206]4C,F,G). It
has been indicated that PRF protects Sp‐MSCs from adverse environmental
assaults, internally fortifying their reparative functions, which
include the enhancement of recovery migration, proliferation, and
secretion of Sp‐MSCs (Figures [207]3 and [208]4; Figure [209]S3,
Supporting Information).
In vivo, Sp‐MSCs encapsulated with PRF significantly promoted wound
healing in diabetes. This synergistic combination has been shown to
foster the regeneration of skin, blood vessels, and nerves within
diabetic wounds (Figures [210]6 and [211]7; Figure [212]S4, Supporting
Information). More importantly, Sp‐MSCs encapsulated with PRF improved
the inflammatory environment of diabetic wounds. Macrophages are innate
immune cells with important roles in immunoregulation and tissue
remodeling.^[ [213]^54 , [214]^55 ^] Sp‐MSCs encapsulated with PRF
transformed macrophages into M2 phenotype in the wound site and reduced
inflammatory factors. At the same time, the aggregation and dysfunction
of neutrophils in the wound site also indicate that diabetic wounds are
prone to infection and immune dysfunction.^[ [215]^56 , [216]^57 ^] PRF
and Sp‐MSCs jointly regulated the immune microenvironment in the wound,
reducing the release of inflammatory factors and the recruitment of
neutrophils (Figure [217]7B; Figure [218]S4C, Supporting Information).
In summary, PRF acts as a biocompatible shield that not only protects
Sp‐MSCs but also amplifies their therapeutic potential, making it a
promising strategy for enhancing tissue repair and regeneration in
diabetic wounds.
3.4. Limitations of PRF and the Emerging Potential of Biomimetic PRF
Alternatives
PRF) is recognized for its superior biocompatibility and significant
regenerative potential; however, the autologous plasma from which it is
derived imposes certain constraints. Initially, the mechanical
properties of PRF necessitate further strengthening.^[ [219]^58 ^]
Additionally, while PRF encompasses an array of growth factors, the
ideal release kinetics to facilitate the healing of diabetic wounds
remain undefined.^[ [220]^59 ^] Moreover, the inability to mass‐produce
PRF, due to the constraints of blood collection and centrifugation
processes, limits its widespread application in wound care.^[ [221]^23
, [222]^60 ^]
The characteristics of PRF, such as platelet concentration, fibrin
structure, and growth factor content, exhibit natural variability,
which may be influenced by factors including individual differences,
blood collection methods, and centrifugation speed.^[ [223]^24 ,
[224]^61 ^] Individual differences, such as age, gender, and health
status, may affect the platelet concentration and growth factor content
of PRF.^[ [225]^62 , [226]^63 ^] Blood collection methods, including
the site, volume, and timing of blood collection, may also lead to
differences in PRF characteristics.^[ [227]^64 , [228]^65 ^] Moreover,
centrifugation speed, a key factor in PRF preparation, can result in
different platelet concentrations and fibrin structures in PRF,^[
[229]^66 ^] thereby affecting its biological properties. In this study,
we found that PRF can effectively protect Sp‐MSCs from the harmful
effects of ROS in diabetic wounds, thereby exerting a more sustained
and beneficial effect. These findings highlights the importance of PRF
characteristic variability for its long‐term clinical implications,
including the effectiveness of tissue repair and regeneration, the
modulation of inflammatory responses, and the development of
personalized treatment strategies. Future research should focus on
elucidating the mechanisms underlying PRF characteristic variability
and on optimizing preparation methods and treatment protocols to
maximize the therapeutic potential of PRF in tissue repair and
regeneration.
Recently, there has been a surge in investigative focus on the
regenerative capabilities of PRF and other plasma‐derived products.
Platelets, known for their distinctive α‐granules, exhibit high
regenerative potential.^[ [230]^67 ^] These α‐granules are repositories
for a variety of growth factors, chemokines, and antimicrobial
proteins, and the unique nature of these granules confers a pivotal
role to platelets in tissue regeneration.^[ [231]^68 ^] Ongoing
research into the constituent factors and underlying molecular
mechanisms is progressively elucidating their roles in the enhancement
of regenerative processes, coagulation, and other related functions.^[
[232]^69 , [233]^70 , [234]^71 ^] Concurrently, the development of
innovative hemostatic agents, such as biomimetic platelets or
platelet‐mimetic particles,^[ [235]^72 , [236]^73 , [237]^74 ^] is
charting new courses for the design and engineering of biomaterials.
The advent of biomimetic α‐granules or synthetic, platelet‐analogous
biomaterials may surmount the challenge of mass production faced by
plasma‐derived biomaterials. These advanced materials promise
consistent active factor content within the gel matrix, independent of
inter‐individual variability, and allow for the utilization of diverse
scaffold materials tailored to specific application scenarios.
Autologous plasma gel emerges as an accessible, cost‐effective, and
biosafe bioactive material, closely resembling native tissue in
composition and mechanical properties. The encapsulation of Sp‐MSCs
within PRF not only bolsters mechanical attributes but also affords
these cells external protection, internal scaffold adhesion, and
nutritional support. The utilization of biomaterial gels derived from
autologous plasma paves novel avenues and strategies for the future
design of regenerative medicine materials.
4. Conclusion
In summary, our study has put forth a strategy for the precise delivery
of Sp‐MSCs using PRF. Our findings further reveal that the
incorporation of Sp‐MSCs can accelerate the gelation process of PRF and
enhance its mechanical strength. This PRF delivery approach not only
provides fibrin protection for Sp‐MSCs against the detrimental effects
of Reactive Oxygen Species (ROS), but also enhances their functional
capabilities, including promoting recovery migration and proliferation.
Moreover, this strategy stimulates the secretory activity of Sp‐MSCs,
collectively creating a microenvironment conducive to cell
proliferation and angiogenesis. Ultimately, the synergistic effect of
PRF and Sp‐MSCs modulates the immune microenvironment, fostering
epithelial, neural, and vascular regeneration in diabetic wounds. This
combination maintains the activity of Sp‐MSCs and enhances their
reparative efficiency in diabetic wound healing, thereby accelerating
the healing process. Therefore, our results indicate that PRF serves as
an effective carrier and shield for the delivery of Sp‐MSCs in the
treatment of diabetic wounds, and the development of smart gel
materials that mimic PRF is an extremely promising direction for future
research.
5. Experimental Section
E‐MSCs Culture and Spheroidal Formation
E‐MSCs were cultured on gelatin (STEMCELL Technologies #07903) in MSC
culture medium prepared with αMEM (Gibco #12571) and 10% HPL (Compass
Biomedical #PLS‐6), 1% GlutaMAX (Gibco #35050), and 1% NEAA (Gibco
#11140). And E‐MSCs were generated as described previously.^[ [238]^10
, [239]^75 , [240]^76 ^] The cells were cultured at 37 °C in the
presence of 5% CO[2] with a medium change every 3 days.^[ [241]^10 ,
[242]^12 ^]
MSCs were digested and cell suspension was prepared. Anti‐adhesion
(STEMCELL Technologies #0 7010) treatment Agreewell800 (STEMCELL
Technologies #34 811) was used. Cell suspension was added to each well,
after centrifugation (100 × g, 3 min), cultured for 24 h to form
Sp‐MSCs. Then Sp‐MSCs were observed and collected in 15 mL centrifuge
tubes, MSC culture medium was added, and stored at room temperature (25
°C).
Cell suspensions of different concentrations were added to form cell
spheroids in Agreewell800 (STEMCELL Technologies #34 811) of varying
sizes.^[ [243]^12 ^] The number of Sp‐MSCs in the sphere‐forming plate
was counted under an inverted microscope. And photos were taken to
measure the diameter of Sp‐MSCs. After collecting Sp‐MSCs, Sp‐MSCs were
left to settle, the supernatant of the culture medium was taken, and
the number of free cells was counted by cell counting.
[MATH:
Sphereforma
mi>tionrate=Sp−<
mi>MSCsforme
mi>dinperwe
mi>llTota
lnumbe
mi>rofmicro
mi>poresperwe
mi>ll :MATH]
(1)
[MATH: Freecells
mi>rate=The<
mspace
width="0.33em">amoun
mi>toffreecells
mi>Totalamoun
mi>tofadded
mi>cells
mi> :MATH]
(2)
Live/Dead Cell Assay
Sp‐MSCs were stored in MSC complete culture medium and incubated at 25
°C for 3 days. Live/dead assay (Yeasen biotech, 40747ES76) was used to
detect cell viability in Sp‐MSCs, and confocal microscopy (Zeiss,
Germany) was used.^[ [244]^77 ^]
Immunofluorescence
Sp‐MSCs were stained with CD73 (BD Pharmingen mouse anti‐human CD73,
561 254), CD90 (BD Pharmingen mouse anti‐human CD90, 555 595), and DAPI
(Antgene, ANT063). A confocal microscope was used for scanning.^[
[245]^12 ^]
Scanning Electron Microscopy (SEM)
Sample was fixed in 2.5% glutaraldehyde fixative, dewatering, and
drying. Specimens were attached to metallic stubs using carbon stickers
and sputter‐coated with gold for 30 s. Images were taken using a
scanning electron microscope (Hitachi, SU8100).^[ [246]^78 ^]
Flow Cytometry
MSC surface markers were detected using BD Stemflow Human MSC Analysis
Kit (BD Biosciences). Sp‐MSCs were dissociated into individual cells.
Cell suspensions were stained via flow cytometry for MSC‐positive
markers (CD44, CD73, CD90, and CD105) and were detected using
corresponding antibodies on BD Accuri 6 flow cytometer (BD
Biosciences).^[ [247]^12 ^]
Preparation of PRF
The preparation methods of PRF have been summarized in relevant studies
and differential centrifugation method was selected to prepare PRF.^[
[248]^23 , [249]^44 , [250]^79 , [251]^80 ^] Mouse (C57BL, female,
6–8w) blood from orbital vein was collected into a vacuum anticoagulant
blood collection tube. And immediately placed in a centrifuge
(Eppendorf Centrifuge 5810R, Germany) and centrifuged at 700 × g, 4 °C
for 8 min, and then 1200 × g, 4 °C for 5 min. Finally, a syringe was
used to extract the lower 1/2layer solution of the plasma layer. 0U
thrombin (Biosharp 9002‐04‐4) per milliliter was added, placed at 25
°C, and left to form platelet‐rich fiber (PRF) gel. PRF was immediately
used for experimentation or storage at −80 °C environment.
Gelation Time of PRF and PRF Combined with Sp‐MSCs
After preparation of PRF, 400 Sp‐MSCs were added into 1 mL PRF
immediately and stayed at 25 °C. The state of PRF until PRF coagulation
was observed and coagulation time was recorded.
Viscosity Measurement
A rheogoniometer (Haake Mars40, Germany) was used to test the viscosity
of PRF/PRF+Sp‐MSCs. PRF and PRF+Sp‐MSCs were prepared for 2 mL each
sample following the aforementioned method. PRF/PRF+Sp‐MSCs gelation
was obtained and put in a rheogoniometer. Control temperature rose from
4.0 to 37.0 °C, heating rate: 1.0 °C min^−1 and the viscosity was
measured.
Growth Factor Determination
Enzyme‐linked immunosorbent assay (ELISA) was used to measure the
content of growth factors in PRF. The fiber block tissue was ground in
PRF and mixed it with the plasma tissue in PRF. It was diluted to a
certain multiple and ELISA kits were used to detect PDGF, EGF, VEGF,
and TGF‐β in it, then analyzed according to the ELISA kit (Jianglaibio,
JL18341, JL20082, JL45919, JL10101).
Stem Cells Proliferation Activity
1000 MSCs were inoculated on 96‐well plates with 100 µL complete
medium, HA group added 10% HA and PRF group added 10% PRF, and
incubated with CCK‐8 reagent for 2h. Optical density (OD) was measured
at 450 nm, and all tests were repeated three times.
PRF Protects Sp‐MSCs from Oxidative Stress
Hyaluronic acid (HA) (Hyprojoint, Bloomage Biotech) was used as a
regular scaffold carrier. After collection Sp‐MSCs, 200 Sp‐MSCs were
coated with 0.5 mL HA/PRF, and the control group was not coated with
gel. Store Sp‐MSCs at 25 °C in MSC culture medium. Oxidative stress
group added 500 µm H[2]O[2].^[ [252]^27 , [253]^28 ^] After 2 and 4
days, live/dead and DCFH‐DA (MX4802, MKbio, Shanghai) staining and
photographed under a confocal microscope for recording.
Sp‐MSCs Migration Capability
Before Sp‐MSCs were inoculated on the cell culture plate, HA/PRF gel
was spread on the culture plate, and the control group was added with
the same amount of blank medium control. After 24 h, photographed under
an inverted microscope and the area of Sp‐MSCs was calculated that had
crawled out of the spheroids.
Sp‐MSCs Secretion Assay
Sp‐MSCs were collected and cultured into 24‐well plate at a density of
25 sp cm^−2. 1 mL MSC culture medium was added, and PRF group added 10%
PRF. After 24 h, the MSC culture medium was replaced by 1 mL α‐MEM with
10% FBS. Then the α‐MEM was collected. The concentration of EGF, VEGF,
and TGF‐β was measured by ELISA.
In Vitro Vascular Formation Ability Experiment
Sp‐MSCs were collected and cultured into six‐well plate at a density of
25 sp cm^−2. And covered with 400 µL PRF. After co‐culturing PRF with
Sp‐MSCs for 48 h, the conditioned culture medium (CM) was collected.^[
[254]^77 ^]
HUVECs Proliferation Activity
1000 HUVECs were inoculated on 96‐well plates and added 10% each
group's CM, and cell viability was measured using CCK‐8 reagent.
HUVECs Migration Assessment
Six‐well plates were inoculated with 6 × 10^5 HUVECs per well, and
after the cells grew to full size, they were scratched with a 200 µL
pipette tip using a straightedge aid, and incubated with different CM.
0 and 12 h were photographed under the inverted fluorescence
microscope.^[ [255]^81 ^]
HUVECs Tube Formation Assessment
60 µL matrix gel was added to 96‐well plates and incubated at 37 °C for
30 min. Then 5000 HUVECs were inoculated on the surface and 100 µL
medium supplemented with 20 µL of different CM was added to each well.
4 h of incubation at 37 °C was used to take pictures using the inverted
fluorescence microscope. Evaluating the ability to promote angiogenesis
by measuring the length of blood vessel formation and the number of
nodes in each group.^[ [256]^81 , [257]^82 ^]
Transcriptomics Analysis
Oxidative stress modeling treatment was the same as before. After a
culture of 2 days, RNA was collected from each group. Output Illumina
platform. Gene expression levels were normalized by calculating TPM
(Transcripts Per Kilobase of exon model per Million mapped reads).
DESeq2 was used to perform the differentially expressed genes (DEGs)
analysis.^[ [258]^83 ^] DEGs were used for subsequent Kyoto
Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were
performed using the Bio conductor packages edgeR.^[ [259]^84 ^]
Hierarchical cluster analysis was performed using clusterProfiler and
heatmap generation was performed using R4.2.1.^[ [260]^85 ^]
RT‐qPCR
Total RNA was extracted from tissues and cells using Trizol
(Invitrogen, USA) and reverse transcribed to cDNA by using a Reverse
Transcription Kit (TOYOBO, OSAKA, Japan). The RNA transcript levels
were analyzed using a QuantStudio3 real‐time PCR system (Thermo Fisher,
USA) and normalized to GAPDH. Primers used in RT‐qPCR are listed in
Table [261]S1 (Supporting Information).^[ [262]^39 , [263]^40 ,
[264]^41 , [265]^42 , [266]^86 , [267]^87 , [268]^88 , [269]^89 ^]
Animal Experiments
All experimental animals used in this study were provided by the
Department of Laboratory Animals of the Central South University
(Changsha, China) and maintained under specific pathogen‐free
conditions. The Research Ethics Committee of Central South University
approved all experimental protocols in this study (csu‐2023‐0196).
Construction of a Diabetic Wound Model: All db/db mice (female, 6–8w)
were subjected to a period of fasting that lasted for 4 h.^[ [270]^90
^] After anesthesia, a circular skin wound (diameter = 10 mm) was built
on the dorsum of the mouse using a surgical scissor, then covering
silver ion antibacterial gel. Subsequently, cover the wound site with
PRF+Sp‐MSCs (100 µL PRF dispersed in 100µL PRF)/PRF (100 µL
PRF)/Sp‐MSCs (200Sp‐MSCs dispersed in 100 µL HA), and leave the control
group untreated. Ultimately, observations and photography of the
wound‐healing process took place on days 1, 5, 9, 12, and 15.^[
[271]^91 , [272]^92 ^]
Ex Vivo Imaging Analysis: To visualize cells, MSCs were labeled with
DiR iodide (1,1′‐Dio ctadecyl‐3,3,3′,3′‐Tetramethylindotricarbocyanine
iodide) (Bergolin Biotechnology, Dalian, China) for 30 min at 37 °C,
then forming Sp‐MSCs. After transplantation of Sp‐MSCs, the wound area
was obtained for imaging by an AniView100X multimode imaging system
(Biolight Biotechnology, Co., Ltd., China). Measure and analyze the
fluorescence intensity of the wound to reflect the survival status of
Sp‐MSCs in the wound.
Wound Healing Mechanism Study: db/db mice were euthanized at
predetermined time points, and dissected to obtain regenerated dorsal
skin, and placed in a 4% paraformaldehyde solution for fixation. The
fixed tissue sections were stained with hematoxylin and eosin (H&E),
Masson, Sirius red staining, CD34, αSMA, GAP43, VEGF, CD206, TNF‐α,
H3Cit.
Image Analysis and Rendering
ImageJ (National Institutes of Health, America) was used for image
correlation measurement and quantitative analysis.
Statistical Analysis
Standard statistical analysis was performed using GraphPad Prizm 9.5.0.
The statistical details in the graphics are presented as the mean ±
standard deviation. The Student's t‐test was used for the determination
of statistical significance between the two groups (n = 2). The ANOVA
test was used for the determination of statistical significance between
multiple groups (n ≥ 3). And the result was deemed statistically
significant at p < 0.05, (where “*” signifies p < 0.05, “**” signifies
p < 0.01, “***” signifies p < 0.001, “****” signifies p < 0.0001, and
“ns” signifies p > 0.05).
Conflict of Interest
The authors declare no conflict of interest.
Author Contributions
J.Z., W.X., Y.X., D.S., and D.W. conducted experiments on in vitro and
in vivo investigation. J.Z., W.X., and D.W. were responsible for animal
experiments. J.Z. and D.W. analyzed the data. J.Z., D.W., and S.L. were
responsible for guiding experiments. J.Z. and D.W. wrote the
manuscript. All authors read and approved the final manuscript.
Supporting information
Supporting Information
[273]ADVS-12-2413430-s001.docx^ (3.6MB, docx)
Acknowledgements