Hyperglycemia attenuates fibroblast contractility via suppression of TβRII receptor modulated α-smooth muscle actin expression

Abstract Nonhealing wounds are a common complication in patients suffering from diabetes with hyperglycemia being the most deteriorating factor for this serious pathological condition. Despite the great body of data, the molecular mechanisms by which high glucose affects cellular physiology are still poorly defined. Here we used primary human foreskin fibroblasts cultured in normo- and hyperglycemic conditions to study the mechanisms leading to altered cell contractility. Our results demonstrated that 25 mmol/L glucose effectively reduced fibroblasts ability to contract fibrin gels, and this physiological change was accompanied by a decrease in alpha-smooth muscle actin expression and the percentage of spontaneously differentiated myofibroblasts in the population of high glucose-treated fibroblasts. These changes were a result of hyperglycemia-induced attenuation of TGF-β1 signaling, involving specific suppression of TGF-β receptor type II but not type I expression. Decreased production of the receptor abolished the ability of exogenously added TGF-β1 to induce Smad2/3 phosphorylation in the presence of high glucose concentrations. Our results identify TGF-β receptor type II as hyperglycemia expression-sensitive receptor and add further aspect to the complex way in which high glucose can affect the wound healing process.


Introduction
Skin wound healing is a dynamic process of repair and replacement of damaged tissue consisting of four main and overlapping processes: hemeostasis; inflammation, proliferation and maturation. The proliferation stage is characterized by angiogenesis, collagen deposition, epithelialization and wound contraction [1,2]. During this stage, new connective tissue, known as granulation tissue is formed. It is composed essentially of extracellular matrix, variable amounts of inflammatory cells, small vessels and fibroblastic cells that after activation become myofibroblasts [3,4]. These specific wound fibroblasts have particular characteristics distinguishing them from normal dermal fibroblasts. They express α-smooth muscle actin (α-SMA), have enhanced contractility when compared to normal fibroblasts and synthesize excessive amounts of extracellular matrix proteins [5,6]. The most powerful stimulator of fibroblast to myofibroblast conversion is transforming growth factor-β (TGF-β) [7][8][9]. TGF-β is released in the wound predominantly by inflammatory cells such as monocytes and macrophages. This cytokine, together with the alternatively spliced extra domain A containing (EDA) form of fibronectin [10] and the increased stiffness in the wound environment [11,12] govern myofibroblast differentiation. The myofibroblast formation appears to be a two-step process involving an intermediate proto-myofibroblast stage characterized with the presence of stress fibers, containing only β-and γ-cytoplasmic actins [12,13]. De novo expression of α-SMA -an actin isoform typical for smooth muscle cells -and its incorporation into existing stress fibers marks the appearance of differentiated myofibroblasts [12].
Studies on the transcriptional activation of the α-SMA gene demonstrate a complex regulation dependent on the cellular type. In myofibroblasts and smooth muscle cells, TGF-β can stimulate α-SMA promoter acting through a response element present in both cell types [14,15]. The signal from TGF-β is mediated by a heteromeric complex of type I and type II transmembrane receptors. The activated receptor complex recruits and phosphorylates Smad2 and Smad3, which, when activated, bind to Smad4 and translocate into the nucleus (ten Dijke, 2006). There, the complex activates transcription of TGF-β-responsive genes in a Smad-binding element (SBE)-dependent or independent manner [16]. A recent study implicates ligand-activated peroxisome proliferator-activated receptor (PPAR) δ in the upregulation of α-SMA in myofibroblast transdifferentiation of human dermal fibroblasts [17]. The effect of PPAR δ is a result of direct binding to a direct repeat-1 (DR1) site in the α-SMA promoter, and stimulation of TGF-β expression, leading to recruitment of Smad3 to a SBE in another region of the promoter.
Myofibroblasts expressing α-SMA have the capacity to generate strong contractile force necessary for shrinkage of the wound granulation tissue [18] and for the overall healing process [19,20]. Deviation from this normal process leads to pathologies like altered wound healing in Diabetes mellitus. The wound healing process in such patients is worsened due to hyperglycemic conditions that lead to major chronic complications, such as diabetic foot ulcers [21]. In vitro studies also demonstrate deteriorating effect of hyperglycemia on aspects of cellular physiology that are important for wound healing involving cell contractility [22], migration [23] and expression of integrin subunits α5 and αv [24]. The proliferation and sensitivity toward growth factors are also affected [25] and these in vitro studies are confirmed by analysis of diabetic wound fibroblasts, which also demonstrate lower responsiveness to EGF, IGF-I, bFGF and PDGF-AB [26] and TGF-β1 [27]. Despite the accumulation of a wealth of data, the exact mechanisms of glucose action on these processes are not fully elucidated. Since myofibroblast contractile phenotype is tightly related to the expression of α-SMA (see, for reviews, [12,28]), we hypothesized that high glucose concentrations would inhibit the expression of α-SMA, thus leading to delayed myofiroblast differentiation and attenuated cell contractility. The purpose of the present study was to elucidate the effects of physiological (5.5 mmol/L) and hyperglycemic (25 mmol/L) glucose concentrations on the expression of α-SMA in primary human skin fibroblasts. Here we demonstrate that elevated glucose levels attenuate the transcription of TGF-β type II receptor (TβRII) leading to a decrease in TβRII protein levels and suppressed activation of Smad2/3. The inhibited TGF-β signaling pathway results in decreased synthesis of α-SMA and reduced cellular contractility.

Cell cultures
Primary human foreskin fibroblasts (HFF) were a gift from S. Yamada (NIDCR, NIH) and were used at passages 9-15. HFF were cultured at 37 °C, 95% humidity and 5% CO 2 in Dulbecco's modified Eagle's medium (DMEM) containing either 1000 mg/L (L0060) or 4500 mg/L (L0104) glucose purchased from Biowest (Nuaillé, France). Culture medium was supplemented with 10% fetal bovine serum (S1520) and a mixture of antibiotics -penicillin (100 U/mL) and streptomycin (100 μg/mL) (L0010) -all from Biowest. HFF cells were propagated in low glucose medium, and the treatment with high glucose concentrations was performed for at least 48 h before testing. In some experiments, a control for the effect of osmotic pressure was used by incubating the cells in low glucose medium supplemented with 20 mmol/L mannitol.
For TGF-β1 treatments cells were cultured for 24 h in medium without serum and then incubated with the same medium containing different concentrations of the cytokine for additional 24 h.

Fibrin gel contraction assay
Fibrin gel contraction assay was done essentially as described by [29] except that 2 × 10 5 cells were plated on top of the fibrin gels and the incubation at 37 °C was extended to 48 h. Since variations in the amount of fibrinogen and thrombin used in the preparation of gels affect cell behavior, we chose a final fibrinogen concentration of 5 mg/mL and 1 U/mL thrombin, which has been shown to support normal cell morphology and proliferation [30]. The medium used for gel preparation and incubation contained high and low glucose concentrations or mannitol as indicated in the previous section. Fixed and stained gels were photographed with camera-equipped Nikon SM Z155 stereomicroscope (Nikon Instruments Inc., Melville, NY). To evaluate the contraction induced by fibroblasts, the diameter of four perpendicular axes of each gel was measured using Image J open source software.
Immunostaining was performed with the indicated antibodies at room temperature for 1 h, followed by appropriate secondary FITC-conjugated antibodies. The position of cell nuclei was visualized by staining with PureBlu™ Hoechst 33342 Nuclear Staining Dye (BioRad Laboratories, CA). Stained samples were mounted in Fluoromount™ aqueous mounting medium (F4680, Merck), Immunofluorescence microscopy images were obtained using an Axiovert 200M microscope (Carl Zeiss) equipped with an Axiocam MR3 camera (Carl Zeiss).

Total RNA extraction and PCR analysis
mRNA from HFF cells was isolated using RNeasy mini kit (ID: 74104, qIAGEN, Stockach, Germany) following the manufacturer's instructions. First strand cDNA was synthesized using RevertAid H Minus First Strand cDNA Synthesis Kit (K1631, Thermo Scientific) according to the manufacturer's instructions. Polymerase chain reaction (PCR) analysis was carried out using Red Taq Master Mix (A180303, Genaxxon Bioscience, Ulm, Germany) on a T100™ Thermal Cycler PCR (Bio-Rad, Hercules, CA). Reaction products were separated on 3% agarose gels and imaged; the results were quantified using the ImageJ software. All samples were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression. Oligonucleotide primers used for all PCR reactions are shown below:

Data analysis
All data were evaluated using GraphPad InStat Program (GraphPad, San Diego, CA). Data were expressed as mean values with standard deviations (± S.D.) and analyzed with a one-way analysis of variance (ANOVA) followed with Tukey-Kramer Multiple Comparisons Test. Differences were considered statistically significant at the level of p < 0.05.

Elevated glucose levels suppress fibroblast contractility
To evaluate the effect of hyperglycemia on the ability of fibroblasts to apply contractile forces on the surrounding extracellular matrix we used fibrin gels. Fibrin is a physiological substance playing a major role in wound healing where it acts as the first scaffold encountered by fibroblasts entering the wound site. Cells used for these experiments were grown in different glucose concentrations for at least two days before the experiment. We cultured HFF cells in low glucose DMEM, containing 1 g/L glucose (5.5 mmol/L) and in high glucose DMEM, containing 4.5 g/L glucose (25 mmol/L) to mimic normal and hyperglycemic conditions, respectively. Since mannitol does not enter the cells, we used media containing 20 mmol/L of this sugar alcohol as a control for the possible effects of increased osmotic pressure. Two days post-seeding, the gels populated with cells grown under low glucose conditions and in the presence of mannitol reduced their average diameter with about 50% (49.5 ± 0.8 and 50.1 ± 3.9, respectively) ( Figure 1 LG and M). Unlike this, high glucose concentrations had relaxation effect on the cells, leading to lower gel contraction. The average gel diameter under these conditions was about 70% (67.4 ± 2.1) of the diameter of the control gels without cells (Figure 1, HG).

Hyperglycemia decreases the expression of α-SMA
Since myofibroblasts contractility is tightly related to α-SMA expression, we studied the distribution of this marker for fibroblast activation in cultures grown under normal and hyperglycemic conditions. Immunofluorescence studies with anti-α-SMA antibodies showed variable staining intensity and numbers of α-SMA positive myofibroblasts, depending on the glucose concentration present in growth medium ( Figure 2). The reaction varied from brightly stained and clearly visible fibers in some cells, through a pail and more diffuse α-SMA signal to complete absence of reaction in the majority of cells (Figure 2(a)). Nevertheless, the percentage of α-SMA-expressing fibroblasts in high glucose medium (18.4 ± 5.5) was reduced almost twice when compared with cells grown in low glucose (36.4 ± 2.5) (Figure 2(b)). These results were confirmed by Western blotting experiments, demonstrating similar and statistically significant two-fold reduction in α-SMA expression by fibroblasts cultured in the presence of high glucose (Figure 3, + Serum). Comparable to the results for fibrin gel contraction, mannitol did not affect the amount of α-SMA production.
The expression of α-SMA is governed by a complex of factors with TGF-β playing a major role. Because the serum used in culture medium contains physiological levels of TGF-β, we repeated the immunoblotting experiments in the absence of serum to assess the possibility that high glucose concentrations would affect the TGF-β signaling cascade. The results confirmed that under serum-free conditions fibroblasts expressed similar amounts α-SMA regardless of the glucose concentration in the medium (Figure 3, -Serum).

High glucose concentrations interfere with TGF-β signaling by suppressing the expression of TβRII
Serum is a complex fluid containing hormones, growth factors and a number of other biologically active compounds that may affect α-SMA expression. To confirm  by direct experiments the possibility that high glucose imposes its action by interfering with TGF-β signaling, we treated 24-hour serum-starved cultures, grown in low or high glucose, with increasing amounts of exogenously added TGF-β1 (Figure 4).
Testing the expression of α-SMA, we found a dose-dependent and statistically significant increase when HFFs in low glucose were stimulated with 1 ng/ mL (202.5 ± 12.1%) and 2 ng/mL (177.1 ± 12.6) TGF-β1 (Figure 4(a), LG). Contrasting to these results, cells in hyperglycemic conditions did not respond to TGF-β1 stimulation and the amount of α-SMA was similar to the amount detected in non-treated cells independently of the amount of the cytokine used (Figure 4(a), HG).
The main signal transducers of the TGF-β signaling cascade are Smad family of proteins with receptor-regulated Smad2/3 being activated through phosphorylation following TGF-β1 stimulus. To test the activity of the TGF-β pathway, we assayed Smad2/3 phosphorylation in the same samples used for α-SMA evaluation. Again, cells grown in low glucose responded to stimulation with the strongest increase in Smad2/3 phosphorylation after stimulation with 1 ng/mL TGF-β1 (315.8 ± 30.9) followed by fading in activation when 2 ng/mL (242.0 ± 38.0) or 5 ng/mL (141.7 ± 36.9) were hFF treated as in "a." Reaction with anti-tubulin antibody (tubulin) was used as an internal control for loading. Right panel represents change in pSmad2/3 expression calculated by normalizing signals from pSmad2/3 to the corresponding signals from tubulin (tubulin) and presenting them as percentage from pSmad2/3 expression in hFF without tgF-β treatment. the data represent mean values ± S.D. of triplicate assays. *p < 0.01, **p < 0.001. used (Figure 4(b), LG). Correspondingly to the results for α-SMA expression, high glucose blocked the response to TGF-β1 and the detected level of Smad2/3 phosphorylation was almost equal in all samples including control, non-treated cells (Figure 4(b), HG).
TGF-β1 signals through a complex cell-surface receptor consisting of 2 type II and 2 type I transmembrane serine/threonine kinases. Binding of type II receptors (TβRIIs) to the ligand results in association and phosphorylation of type I receptor kinases (TβRIs), leading to activation of their kinase activity. Activated TβRIs then phosphorylate cytoplasmic Smad2/3 at the C-terminus. The established deficiency of Smad2/3 phosphorylation under hyperglycemic conditions prompted us to test the status of TGF-β receptors. First, we tested the transcription of both TGF-β receptors in HFFs cultured in low and high glucose concentrations. While the amount of TβRI mRNA was not affected by different growth conditions, the message for TβRII was reduced by 80% when high glucose was present in the growth medium ( Figure 5(a)). This result was confirmed also at the protein level. Immunoblotting experiments with antibodies against TβRII demonstrated a significant decrease in the amount of protein (45.2 ± 3.7%) of this type of TGF-β receptor ( Figure 5(b)).

Discussion
Fibroblasts activation is an essential process for normal wound healing. The resulting myofibroblasts are involved in strengthening of the wound extracellular matrix by producing large amounts of collagen and other extracellular matrix proteins, and subsequent wound contraction, which is necessary for the healing process [5, 28;Tay et al., 2021]. The excessive contractile force generated by myofibroblasts is connected to their specific ability to express and incorporate the smooth-muscle specific α-SMA into the actin stress fibers [12,13,19,31]. Impaired fibroblasts activation, which occurs in the elderly or in various diseases like diabetes, leads to pathological conditions such as nonhealing wounds. In the case of diabetes, hyperglycemia is emerging as an important factor not only leading to the induction of pro-inflammatory molecules [32] and to downregulation of collagen [33] but also affecting strongly cell physiology.
With the present work, we demonstrated that high glucose concentrations could down-regulate the expression of the myofibroblast marker α-SMA, and this was manifested physiologically by a statistically significant reduction in the contractility of human foreskin fibroblasts (Figure 1). Similar reduction in collagen gel contraction is reported for skin fibroblasts [34], implying that the simplified in vitro system involving fibroblasts and media with different glucose content can be used to study the mechanisms by which hyperglycemia imposes its negative action on cell contractility.
Actin stress fibers are regarded as cell contractile machinery [35][36][37] that are responsible for generation Figure 5. effect of hyperglycemia on the expression of tgF-β receptors. (A) pcR analysis of tβRi and tβRii expressed by hFF cells cultured for 48 h in medium containing 5.5 mmol/l glucose (lg) or 25 mmol/l glucose (hg). the percent change in the expression of each tgF-β receptor in high glucose is presented relative to the expression of the same receptor in low glucose after normalization to gapDh, which was used as an internal control. (B) Western blot analysis of tgF-β type ii receptor (tβRii) expression by fibroblasts cultured as in "a" but for 72 h. Reaction with anti-glyceraldehyde 3-phosphate dehydrogenase (gapDh) antibody was used as a control for sample loading. the graph represents the change in tβRii protein expression calculated by normalizing the signals from tβRii to the corresponding signals from gapDh and presenting them as percentage of tβRii expression in hFF cultured in low glucose. graphs shown represent pooled data from three independent experiments. error bars represent mean values ± S.D. *p < 0.01, **p < 0.001. of isometric tension [38]. Enrichment of these cytoskeletal elements with α-SMA in myofibroblasts is dependent on substrate stiffness and focal adhesion maturation [11] and leads to generation and transmission of strong mechanical forces to the extracellular matrix [28]. Although the expression of α-SMA during wound healing is tightly controlled by complex environmental cues, spontaneous fibroblast-myofibroblast transdifferentiation can occur in in vitro cultures depending on substrate rigidity [39]. Our results confirmed this finding and demonstrated an average of 36% α-SMA-positive HFFs when cells were grown on the rigid plastic surface of the culture dish in the presence of 5.5 mmol/L glucose (Figure 2(b)). In agreement with the results from the fibrin gel contraction assays, elevation of the glucose content to 25 mmol/L caused a two-fold reduction in the number of myofibroblasts, implying that hyperglycemia may have directly or indirectly affected the activity of the α-SMA gene. Testing the expression of α-SMA in the absence of serum revealed similar levels, unrelatedly to the glucose concentration in the growth media ( Figure 3, -Serum). In the presence of low and high glucose as well as in media with mannitol, serum-starved human fibroblasts produced comparable amounts of α-SMA, suggesting that hyperglycemia did not impose its effect directly on the α-SMA gene. In favor of indirect influence, different levels of expression were detected only in the presence of serum, indicating that high glucose most probably interfered with a signaling pathway activated by factor(s) present in the serum.
As the α-SMA gene is a well-established target of the TGF-β pathway [7][8][9]14], we focused our subsequent experiments on this cascade. Addition of serum, which contains TGF-β1 (Figure 3, +Serum), or exogenous TGF-β1 in serum-free conditions ( Figure  4(a)) restored the inhibitory effect of hyperglycemia on α-SMA and allowed us to identify the TGF-β signaling cascade as a glucose sensitive pathway. These results were verified further by probing the activation status of TGF-β effectors Smad2/3. The stimulation of these receptor-regulated Smads (R-Smads) is a result of phosphorylation of the two Ser residues in the distal C-terminal Ser-Xaa-Ser motif [40]. By using phosphospecific antibody recognizing the phosphorylation status of these residues (Ser 423/425) we established the same pattern of activation corresponding to the pattern of α-SMA expression after TGF-β1 stimulation (Figure 4(b)) with maximal effect at 1 ng/mL TGF-β1 in low glucose and marked inability for stimulation in the presence of high glucose concentrations.
The phosphorylation level of Smad2/3 is a net result of the activity of kinases and phosphatases that recognize these proteins as substrates. While Smad2/3-specific phosphatases are still under elucidation [40], the process of C-terminal serine phosphorylation is well studied and is attributed to the activity of TGF-β receptor type I kinase. After dimer TGF-β1 ligand binds to two TβRII, due to its higher affinity for this type of receptor, a TβRI dimer is recruited and a functional heterotetrameric receptor complex is assembled [41]. Within the complex, the constitutively active TβRII phosphorylates TβRI in the regulatory GS domain, located above the kinase domain [42]. The phosphorylation of GS domain, increases binding of R-Smads and promotes their phosphorylation [43]. Since both receptor types play important role in the transmission of the TGF-β1 signal, we tested the effect of high glucose concentrations on the status of each of the receptors. While the transcription of TβRI was not influenced, the TβRII message was reduced by 80% in cells grown in high glucose concentrations ( Figure 5(a)) and an almost similar decrease was detected at the protein level ( Figure 5(b)).
Hyperglycemic conditions have been shown to attenuate the expression of different cellular receptors including vitamin D and estrogen receptor β [44], insulin receptor [45] and hepatic scavenger receptor class B type I (SR-BI) [46]. In the last study, a molecular mechanism for glucose action is proposed connecting p38 MAPK and Sp1 to the glucose suppression of SR-BI promoter activity. The TβRII promoter lacks a consensus TATA box, and together with other TATA-less promoters like SR-BI, depends strongly on the regulation by Sp1 transcription factor [47]. Indeed several Sp1-biding sites have been shown to play an important role in TβRII activation [48] but whether they may be involved in the transmission of the repressive effect of glucose, like in SR-BI, remains to be elucidated.
Despite the unclarified mechanism of glucose action on TβRII the obtained results allowed us to outline a route of glucose action on cell contractility. It starts with suppression of TβRII expression, leading to attenuation of TGF-β1 signaling through reduction of Smad2/3 phosphorylation, decreased expression of α-SMA and suppressed ability of primary human fibroblasts to transdifferentiate to myofibroblasts and exert strong contractile forces. Decreased TβRII expression has been reported for some pathological conditions like venous ulcers [49] and cancer [50] but to our knowledge this is the first report demonstrating the suppressive effect of hyperglycemia on TFG-β type II receptor. Our results add additional information on the mechanisms by which hyperglycemia may interfere with the normal processes of wound healing in diabetic patients. They should be considered when new therapies for diabetic wound are designed, especially when treatments with exogenous growth factors are envisaged.

Authors' contributions
AM and RP conceptualized and designed the study.
AE, GG and PK developed methodology. AE, GG, BA and SP acquired data and performed analysis.
AM and RP wrote and revised the manuscript. RP acquired funding.

Data availability
All data that support the findings reported in this study are available from the corresponding author upon reasonable request.

Disclosure statement
No potential conflict of interest was reported by the authors.