GLUT-1 Enhances Glycolysis, Oxidative Stress, and Fibroblast Proliferation in Keloid

A keloid is a fibroproliferative skin tumor. Proliferating keloid fibroblasts (KFs) demand active metabolic utilization. The contributing roles of glycolysis and glucose metabolism in keloid fibroproliferation remain unclear. This study aims to determine the regulation of glycolysis and glucose metabolism by glucose transporter-1 (GLUT-1), an essential protein to initiate cellular glucose uptake, in keloids and in KFs. Tissues of keloids and healthy skin were explanted for KFs and normal fibroblasts (NFs), respectively. GLUT-1 expression was measured by immunofluorescence, RT-PCR, and immunoblotting. The oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured with or without WZB117, a GLUT-1 inhibitor. Reactive oxygen species (ROS) were assayed by MitoSOX immunostaining. The result showed that glycolysis (ECAR) was enhanced in KFs, whereas OCR was not. GLUT-1 expression was selectively increased in KFs. Consistently, GLUT-1 expression was increased in keloid tissue. Treatment with WZB117 abolished the enhanced ECAR, including glycolysis and glycolytic capacity, in KFs. ROS levels were increased in KFs compared to those in NFs. GLUT-1 inhibition suppressed not only the ROS levels but also the cell proliferation in KFs. In summary, the GLUT-1-dependent glycolysis and ROS production mediated fibroblast proliferation in keloids. GLUT1 might be a potential target for metabolic reprogramming to treat keloids.


Introduction
A keloid is a proliferative scar that extends beyond the original injury borders with excess fibroblast proliferation and increased collagen production [1]. It appears as a red, indurated, disfiguring tumor and may be accompanied by intractable pruritus, pain, and even contractions. They tend to develop in high tension skin areas after minor traumas Life 2021, 11, 505 3 of 18 controls (n = 6) were recruited (Table 1). A keloid was defined as a skin lesion exhibiting continuous growth beyond the margin 6 months after trauma or surgery. Patients were excluded if they received treatment with intralesional corticosteroid injection or liquid nitrogen therapy within the past 3 months or they had associated invasive cancers. Tissue samples were taken from the center of the keloid with a 3-mm biopsy punch. Control samples of normal skin were obtained from perilesional skin after elective surgical excision for melanocytic nevus or epidermal cysts. The protocol of this study was approved by the Institutional Review Board of Kaohsiung Veterans General Hospital (VGHKS16-CT5-10).

Cell Culture and Treatment
Human fibroblasts were isolated from dermal tissues of control or keloid skin. Dermal tissues were cut to 1-2 mm 3 ; then, fibroblasts were cultured in DMEM (12100046, Gibco, Waltham, MA, USA) medium supplemented with 10% fetal bovine serum (FBS) (A4766801, Gibco, Waltham, MA, USA) and 1% penicillin-streptomycin (P/S) (15140122, Gibco, Waltham, MA, USA), and maintained at 37 • C humidified air with 5% CO 2 . Fibroblasts with the third to sixth passage were used for the experiments. Three strains of KFs and three strains of NFs derived from punched tissue samples were used in the study. To examine the role of GLUT-1 on NFs and KFs, fibroblasts were starved in serum-free DMEM for at least 12 h before treatment using 10 µM WZB117, a GLUT-1 inhibitor (6143, PeproTech, Rocky Hill, NJ, USA) for 48 h.

Seahorse Analysis for OCR and ECAR
A Seahorse XF24 Extracellular Flux Analyzer (Agilent, Santa Clara, CA, USA) was used to continuously monitor the extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) of fibroblasts. An XF Glycolysis Stress Test kit (103020, Agilent, Santa Clara, CA, USA) was used for glycolysis and an XF Mito Stress Test kit (103015-100, Agilent, Santa Clara, CA, USA) was used to test mitochondrial stress. Fibroblasts (1 × 10 4 cells /well) were seeded in an XF24 culture plate in DMEM supplemented with 2% FBS at 5% CO 2 at 37 • C for 18-20 h. Before the experiment, fibroblasts were washed with PBS, replaced with 625 µL DMEM (pH = 7.4) without FBS and sodium bicarbonate, and then incubated in a CO 2 -free incubator at 37 • C. For the ECAR assay, 10 mM glucose, 1 µM oligomycin, and 75 mM 2-deoxy-glucose (2-DG) (a glucose analog) in XF Glycolysis Stress Test kit (103020, Agilent, Santa Clara, CA, USA) were injected to the medium ( Figure 1A). First, glucose was injected to induce glycolysis. Subsequent treatment with oligomycin shut down OXPHOS with a compensatory glycolytic capacity. Next, by inhibiting glycol-Life 2021, 11, 505 4 of 18 ysis with 2-DG, glycolysis reserve was estimated by the glycolysis capacity divided by basal ECAR. For the OCR assay, 1 µM oligomycin (a complex V inhibitor), 1 µM carbonyl cyanide-phospho-(p)-trifluoro-methoxyphenyl-hydrazon (FCCP) (a proton gradient uncoupler), and 0.5 µM rotenone (a complex I inhibitor)/antimycin A (a complex III inhibitor) in an XF Mito Stress Test kit (103015-100, Agilent, Santa Clara, CA, USA) were subsequently injected to the medium ( Figure 1C). First, baseline cellular OCR was measured, from which basal respiration was derived by subtracting non-mitochondrial respiration. Next, oligomycin was added, and the resulting OCR was used to derive ATP-production respiration (by subtracting the oligomycin rate from baseline cellular OCR) and protonleak respiration (by subtracting non-mitochondrial respiration from the oligomycin rate). Further, FCCP was added to collapse the inner membrane gradient, driving the ETC to function to its maximal rate, and maximal respiratory capacity was derived by subtracting non-mitochondrial respiration from the FCCP OCR. Lastly, ETC function was shut down by addition of rotenone and antimycin A, which revealed the non-mitochondrial respiration. The mitochondrial spare capacity was calculated by subtracting basal respiration from maximal respiratory capacity.
Test kit (103020, Agilent, Santa Clara, CA, USA) were injected to the medium (Fi First, glucose was injected to induce glycolysis. Subsequent treatment with ol shut down OXPHOS with a compensatory glycolytic capacity. Next, by inhibitin ysis with 2-DG, glycolysis reserve was estimated by the glycolysis capacity d basal ECAR. For the OCR assay, 1 µ M oligomycin (a complex V inhibitor), 1 µ M cyanide-phospho-(p)-trifluoro-methoxyphenyl-hydrazon (FCCP) (a proton gra coupler), and 0.5 µ M rotenone (a complex I inhibitor)/antimycin A (a complex tor) in an XF Mito Stress Test kit (103015-100, Agilent, Santa Clara, CA, USA) we quently injected to the medium ( Figure 1C). First, baseline cellular OCR was m from which basal respiration was derived by subtracting non-mitochondrial re Next, oligomycin was added, and the resulting OCR was used to derive ATP-p respiration (by subtracting the oligomycin rate from baseline cellular OCR) an leak respiration (by subtracting non-mitochondrial respiration from the oligomy Further, FCCP was added to collapse the inner membrane gradient, driving th function to its maximal rate, and maximal respiratory capacity was derived by su non-mitochondrial respiration from the FCCP OCR. Lastly, ETC function was s by addition of rotenone and antimycin A, which revealed the non-mitochondria tion. The mitochondrial spare capacity was calculated by subtracting basal re from maximal respiratory capacity. The continuous ECARs and were recorded and analyzed using an XF94 extracellular flux analyzer. The scheme of EC shows the section of the time trace corresponding to each module. Subsequently, 10 mM µ M oligomycin, and 75 mM 2-DG were injected to the medium. (B) Representative ECA measured in KFs and NFs (1 × 10 4 cells/well). Data are presented as mean ± SEM. (N = 6, ** p < 0.01, *** p < 0.001.). The representative data from 3 independent experiments are sh Continuous OCRs values were recorded and analyzed. The scheme of continuous of OC the section of the time trace corresponding to each module. Subsequently, 1 µ M oligomy FCCP, and 0.5 µ M rotenone/antimycin A were injected to the medium. (D) Representativ were measured in KFs and NFs (1 × 10 4 cells/well). Data are presented as mean ± SEM (N 0.05, ** p < 0.01, *** p < 0.001.). Shown is the representative data from 3 independent expe

Immunohistochemistry (IHC) for GLUT-1 In Tissue
Paraffin-embedded samples of keloid and normal tissue were cut and mo coated slides. After de-paraffinization in xylene and rehydration with graded a lutions, antigen retrieval was performed in sodium citrate buffer (100 °C, 20 m Novolink Polymer Detection System (RE7150-K, Leica, Wetzlar, Germany) wa

Immunohistochemistry (IHC) for GLUT-1 In Tissue
Paraffin-embedded samples of keloid and normal tissue were cut and mounted on coated slides. After de-paraffinization in xylene and rehydration with graded alcohol solutions, antigen retrieval was performed in sodium citrate buffer (100 • C, 20 min). The Novolink Polymer Detection System (RE7150-K, Leica, Wetzlar, Germany) was carried Life 2021, 11, 505 5 of 18 out according to the manufacturer's introduction. The sections were neutralized with endogenous peroxidase using a Peroxidase Block for 5 min, blocked by incubation with Protein Block (RE7102, Leica, Wetzlar, Germany) for 5 min, then incubated overnight at 4 • C with mouse monoclonal antibody anti-GLUT-1 (1:100 dilution, sc3772282, Santa Cruz, CA, USA). After rinsing by phosphate buffered saline (PBS) (11-223-1M, Biological Industries, Cromwell, CT, USA), the slides were incubated with Novolink Polymer for 30 min. Afterwards, the sections were developed with 3,3 -diaminobenzidine (DAB) working solution, counterstained with hematoxylin for 5 min, dehydrated, and mounted by mounting medium. Finally, 6 view fields were randomly selected on each section and observed under a microscope (Olympus, Tokyo, Japan) to determine the expression.  The total RNA from fibroblasts was extracted using 1 mL TRIzol reagent (15596018, Invitrogen, Waltham, MA, USA). After the addition of 200 µL 1-Bromo-3-chloropropane (B9673, Sigma, St. Louis, MO, USA), RNA was precipitated with 500 µL isopropanol (278475, Sigma, St. Louis, MO, USA) and pelleted by centrifugation at 13,000× rpm at 4 • C for 20 min. The pelleted RNA was washed with 1 mL 75% ethanol, then centrifuged at 13,000× rpm at 4 • C for 10 min. After removal of the supernatant, it was air-dried for 6 min and suspended in nuclease-free water. The RNA concentration was determined by Spectrophotometer (Eppendorf, Hamburg, Germany). For RT-PCR, 1.5 µg samples of total RNA were reverse transcribed into cDNA in 20 µL of reaction mixture containing 200 units of Superscript III enzyme, 0.5 µL random primer, 0.5 µL 25 mM dNTP, 4 µL 5× first strand buffer, 1 µL 0.1 M dithiothreitol (DTT), and 1.8 µL H 2 O in a buffer. The RT procedure was performed according to the protocol suggested by the Superscript III manufacturer (18080085, Invitrogen, Waltham, MA, USA). The resulting cDNA was frozen at −20 • C for mRNA detection. Target genes were amplified by PCR performed with fast SYBR green PCR Master mix (4385612, Thermo Fisher, Waltham, MA, USA). Next, 3 µL cDNA from RT was replicated in PCR reactions in a total volume of 10 µL containing 5 µL fast SYBR Master mix, and 2 µL of the target gene's forward and reverse primers mix at Table 2. RT-PCR primer sequence.

Gene
Sequences (

Mitochondrial Superoxide Detection
The MitoSOX TM Red mitochondrial superoxide indicator kit (36008, Thermo Fisher, Waltham, MA, USA) was applied to measure mitochondrial superoxide in fibroblasts. Fibroblasts (1 × 10 5 cells/well) were seeded overnight in a 12 well-plate then incubated with 1 mL 5 µM MitoSOX reagent solution for 10 min at 37 • C in a 5% CO 2 incubator protected from light. After staining, the cells were replaced with fresh culture medium and observed under a microscope (Olympus, Tokyo, Japan) to determine the superoxide production from mitochondria.

Cell Proliferation Assay
Proliferation was measured by WST-1 Proliferation Assay (11644807001, Roche, Basel, Switzerland). Fibroblasts were seeded at 3 × 10 3 cells/well (100 µL) in 96-well plates in DMEM with 2% FBS overnight. Fibroblasts were treated with WZB117 at indicated concentration in <2% FBS DMEM and determined at indicated time points. The cell medium was replaced with the WST-1 reagent and then incubated at 37 • C for 2 h. Absorbance was monitored with an Epoch TM 2 Microplate Spectrophotometer (BioTek, Winooski, VT, USA) at 450 nm with a reference wavelength of 620 nm.

Statistical Analysis
All analyses were conducted with Prism 6.0 (GraphPad Software). All other values were expressed as either means ± standard error in the mean (SEM) in three independently performed experiments or as a representative value obtained in three independently performed experiments. Group differences were evaluated by Student's t-test or Mann-Whitney U test. A p < 0.05 was considered statistically significant.

Enhanced Glycolysis in KF
To investigate the mitochondrial energy metabolism in keloids, we first analyzed ECARs and OCRs, which are surrogates of glycolysis and oxidative mitochondrial activity, Life 2021, 11, 505 7 of 18 respectively. The ECAR was measured by sequential injections of glucose, oligomycin, and 2-DG with an XF94 Flux analyzer (for details, see Figure 1A in Methods). Our data showed that, in comparison with NFs, KFs had higher ECARs including rates of glycolysis, glycolytic capacity, and glycolysis reserve ( Figure 1B). Next, OCR was measured with an XF94 Flux analyzer after sequential injections of oligomycin, FCCP, and rotenone/actimycin A (for details, see Figure 1C and in Methods). On the other hand, the dynamic components of OCR, including the basal respiratory OCR, proton-leak OCR, ATP production OCR, and maximal respiration OCR were similar among KFs and NFs ( Figure 1D). These results suggest that the KFs showed a tendency towards glycolysis.

Glycolytic Enzymes Were Upregulated in KFs
Since increased ECARs in the KFs was associated with a reciprocal decrease in OCR, we next asked the mechanism by which ECAR (glycolysis) is enhanced in KFs. To address that, we measured the expression of enzymes in glycolytic pathway, including GLUT-1, hexokinase, GPI, PFK, aldolase, PKM2, LDH, and PDK1 by RT-PCR in KFs (Figure 2A). The results showed that GLUT-1, hexokinase, GPI, and aldolase were significantly upregulated at 1.5-fold, 2-fold, 3-fold, and 2-fold in KFs, respectively, as compared with those in NFs, suggesting that the induction of GLUT-1, hexokinase, GPI, and aldolase may contribute to the high glycolytic phenotype in KFs ( Figure 2B).

Enhanced Glycolysis in KF
To investigate the mitochondrial energy metabolism in keloids, we first analyzed ECARs and OCRs, which are surrogates of glycolysis and oxidative mitochondrial activity, respectively. The ECAR was measured by sequential injections of glucose, oligomycin, and 2-DG with an XF94 Flux analyzer (for details, see Figure 1A in Methods). Our data showed that, in comparison with NFs, KFs had higher ECARs including rates of glycolysis, glycolytic capacity, and glycolysis reserve ( Figure 1B). Next, OCR was measured with an XF94 Flux analyzer after sequential injections of oligomycin, FCCP, and rotenone/actimycin A (for details, see Figure 1C and in Methods). On the other hand, the dynamic components of OCR, including the basal respiratory OCR, proton-leak OCR, ATP production OCR, and maximal respiration OCR were similar among KFs and NFs ( Figure 1D). These results suggest that the KFs showed a tendency towards glycolysis.

Glycolytic Enzymes Were Upregulated in KFs
Since increased ECARs in the KFs was associated with a reciprocal decrease in OCR, we next asked the mechanism by which ECAR (glycolysis) is enhanced in KFs. To address that, we measured the expression of enzymes in glycolytic pathway, including GLUT-1, hexokinase, GPI, PFK, aldolase, PKM2, LDH, and PDK1 by RT-PCR in KFs (Figure 2A). The results showed that GLUT-1, hexokinase, GPI, and aldolase were significantly upregulated at 1.5-fold, 2-fold, 3-fold, and 2-fold in KFs, respectively, as compared with those in NFs, suggesting that the induction of GLUT-1, hexokinase, GPI, and aldolase may contribute to the high glycolytic phenotype in KFs ( Figure 2B).

GLUT-1 Expression Was Increased in Both KFs and Hyperfibrotic Regions in Keloid Tissues
Since GLUT-1 is the key and initial enzyme in the glycolysis and it was found to be significantly increased in KFs, we asked whether GLUT-1 expression is increased in keloid tissues. We then measured the GLUT-1 expression in keloids and normal skin tissues by immunohistochemical analysis. The results showed that the expression of GLUT-1 was present in the epidermis from both keloids and control skin ( Figure 3A). Interestingly, in the dermis of the keloids, the expression of GLUT-1 was markedly increased with high collagen deposits as compared with that in the control skin ( Figure 3B). In parallel, the

GLUT-1 Expression Was Increased in Both KFs and Hyperfibrotic Regions in Keloid Tissues
Since GLUT-1 is the key and initial enzyme in the glycolysis and it was found to be significantly increased in KFs, we asked whether GLUT-1 expression is increased in keloid tissues. We then measured the GLUT-1 expression in keloids and normal skin tissues by immunohistochemical analysis. The results showed that the expression of GLUT-1 was present in the epidermis from both keloids and control skin ( Figure 3A). Interestingly, in the dermis of the keloids, the expression of GLUT-1 was markedly increased with high collagen deposits as compared with that in the control skin ( Figure 3B). In parallel, the expression of GLUT-1 in fibroblasts was examined by immunofluorescence staining. The GLUT-1 expression was distributed along the plasma membrane and throughout the cytoplasm of fibroblasts ( Figure 3C). Notably, GLUT-1 expression was nearly 7-fold higher in KFs compared to NFs ( Figure 3D). Interestingly, treatment of KFs with WZB117, a GLUT-1 inhibitor, substantially decreased the elevated level of GLUT-1 in KF ( Figure 3D). Taken together, these results suggest that GLUT-1 may contribute to the development of fibrotic lesions in keloids.

GLUT-1 Expression Is Critical for Glycolysis but Not Mitochondrial OXPHOS in KFs
We noted the increased GLUT-1 expression and increased glycolysis in KFs. We then asked whether GLUT-1 regulates glycolysis in keloids. The ECARs were measured in KFs and NFs with and without treatment with WZB117, a GLUT-1 inhibitor ( Figure 4A). Before treatment, the KFs had significantly increased ECARs in terms of glycolytic ATP output ( Figure 4B), glycolytic capacity ( Figure 4C), and glycolytic reserve ( Figure 4D), which is consistent with previous experiments (Figure 1). After treatment with WZB117, the enhanced glycolysis, glycolytic capacity, and glycolysis reserve were significantly suppressed in KFs ( Figure 4B-D), indicating that GLUT-1 regulates enhanced glycolysis in keloids.
Next, we asked whether GLUT-1 regulated mitochondrial OXPHOS, an indicator of oxidative stress, in KFs. We measured the OCRs in KFs and NFs with or without WZB117 ( Figure 5A). Consistent with our previous experiments (Figure 1), KFs and NFs had comparable OCRs, including basal respiration OCR ( Figure 5B), ATP-production OCR ( Figure 5C), proton-leak OCR ( Figure 5D), and the maximal respiration OCR ( Figure 5E). However, our further studies showed that spare respiratory capacity of KFs was nearly 3-fold higher than that of NFs ( Figure 5F).
After treatment with WZB117 in NFs, the basal respiration OCR (2-fold) and maximal respiration OCR were significantly increased. Although NFs also showed increased ATP production OCR and proton-leak OCR, the increase did not reach statistical significance. Additionally, whereas WZB117 treatment suppressed spare respiration capacity in both NFs and KFs ( Figure 5F), the spare respiratory capacity was higher in KFs despite WZB117 treatment. Taken together, these results suggest that KFs are fueled predominantly by GLUT-1 dependent glycolysis rather than mitochondrial OXPHOS. That is, KFs behave toward the Warburg-like effect to meet the energetic demand of fibroproliferation property, in which GLUT-1 has a critical role.

Blocking GLUT-1 Decreased the Enhanced ROS Formation in KFs
The oxidative state is maintained by a balance between antioxidant and pro-oxidant levels. To determine whether GLUT-1 regulated ROS levels, MitoSox was used as an indicator of ROS in NFs and KFs, when they were treated with or without WZB117 ( Figure 6A-C). We traced the ROS fluorescence signaling at different time points in each group and counted more than 50 cells in each field to quantify the signals. The results showed that, in the baseline, KFs showed a nearly 2-fold increase of ROS levels compared to those in NFs ( Figure 6A). While WZB117 treatment did not alter the ROS levels in NFs after one or 24 h, WZB117 treatment suppressed ROS levels in KFs after 1 or 24 h. In addition, the ROS levels in KFs returned to the baseline ROS levels as in NFs ( Figure 6B,C). These data indicated that GLUT-1 mediated enhanced ROS levels, creating oxidative stress in keloids.

GLUT-1 Is Required for Proliferation of KFs
In our previous experiments ( Figure 5F), spare respiratory capacity, a measure of the ability of the mitochondria to respond to physiological stress, was enhanced in KFs regardless of GLUT-1 inhibition. We were interested in whether cell proliferation of KFs would be regulated by GLUT-1. To evaluate the effect of GLUT-1 on the cell proliferation in keloids, cell viability and proliferation were measured after WZB117 treatment. At baseline, KF exhibited a slight increase in the cell viability compared to NFs in accordance with the increased spare respiratory capacity. WZB117 treatment suppressed cell viability in both KFs and NFs, although the cell viability in KFs was reduced to a greater extent than that in NF ( Figure 7A). To consolidate this finding, further proliferation analysis was performed using Ki-67 labeling ( Figure 7B). WZB117-treated KFs showed a significantly reduced Ki-67 Life 2021, 11, 505 9 of 18 intensity compared with controls ( Figure 7C), suggesting that GLUT-1 regulates fibroblast proliferation, particularly in keloids. expression of GLUT-1 in fibroblasts was examined by immunofluorescence staining. The GLUT-1 expression was distributed along the plasma membrane and throughout the cytoplasm of fibroblasts ( Figure 3C). Notably, GLUT-1 expression was nearly 7-fold higher in KFs compared to NFs ( Figure 3D). Interestingly, treatment of KFs with WZB117, a GLUT-1 inhibitor, substantially decreased the elevated level of GLUT-1 in KF ( Figure 3D). Taken together, these results suggest that GLUT-1 may contribute to the development of fibrotic lesions in keloids.  put ( Figure 4B), glycolytic capacity ( Figure 4C), and glycolytic reserve ( Figure 4D) is consistent with previous experiments (Figure 1). After treatment with WZB117, hanced glycolysis, glycolytic capacity, and glycolysis reserve were significant pressed in KFs ( Figure 4B-D), indicating that GLUT-1 regulates enhanced glyco keloids. Next, we asked whether GLUT-1 regulated mitochondrial OXPHOS, an indi oxidative stress, in KFs. We measured the OCRs in KFs and NFs with or without W ( Figure 5A). Consistent with our previous experiments (Figure 1), KFs and NFs ha parable OCRs, including basal respiration OCR ( Figure 5B), ATP-production OCR 5C), proton-leak OCR ( Figure 5D), and the maximal respiration OCR ( Figure 5E) ever, our further studies showed that spare respiratory capacity of KFs was nearly higher than that of NFs ( Figure 5F).
After treatment with WZB117 in NFs, the basal respiration OCR (2-fold) and m respiration OCR were significantly increased. Although NFs also showed increas production OCR and proton-leak OCR, the increase did not reach statistical signi Additionally, whereas WZB117 treatment suppressed spare respiration capacity NFs and KFs (Figure 5F), the spare respiratory capacity was higher in KFs despite W treatment. Taken together, these results suggest that KFs are fueled predomina GLUT-1 dependent glycolysis rather than mitochondrial OXPHOS. That is, KFs

Blocking GLUT-1 Decreased the Enhanced ROS Formation in KFs
The oxidative state is maintained by a balance between antioxidant and pro-oxidant levels. To determine whether GLUT-1 regulated ROS levels, MitoSox was used as an indicator of ROS in NFs and KFs, when they were treated with or without WZB117 (Figure 6A-C). We traced the ROS fluorescence signaling at different time points in each group and counted more than 50 cells in each field to quantify the signals. The results showed that, in the baseline, KFs showed a nearly 2-fold increase of ROS levels compared to those in NFs ( Figure 6A). While WZB117 treatment did not alter the ROS levels in NFs after one or 24 h, WZB117 treatment suppressed ROS levels in KFs after 1 or 24 h. In addition, the ROS levels in KFs returned to the baseline ROS levels as in NFs ( Figure 6B,C). These data indicated that GLUT-1 mediated enhanced ROS levels, creating oxidative stress in keloids.

GLUT-1 Is Required for Proliferation of KFs
In our previous experiments ( Figure 5F), spare respiratory capacity, a measure of the ability of the mitochondria to respond to physiological stress, was enhanced in KFs regardless of GLUT-1 inhibition. We were interested in whether cell proliferation of KFs would be regulated by GLUT-1. To evaluate the effect of GLUT-1 on the cell proliferation in keloids, cell viability and proliferation were measured after WZB117 treatment. At

Discussion
In this study, KFs revealed an increased GLUT-1-dependent glycolysis rate and spare respiratory capacity, suggesting that KFs behave according to a Warburg-like effect to meet the energy demand of increased cell viability. Moreover, enhanced GLUT-1 activity was shown in keloid tissues and KFs, and this regulates not only glycolytic rate but also the increased ROS levels in keloids. The GLUT-1 inhibitor suppressed cell viability, which is associated with the inhibited glycolytic function and reduced spare respiratory capacity in KFs. These results indicated that increased GLUT-1 expression in keloids may regulate the metabolic reprogramming and increased fibroblast proliferation in keloids (Figure 8).
fe 2021, 11, x FOR PEER REVIEW In this study, KFs revealed an increased GLUT-1-dependent glycolysis r respiratory capacity, suggesting that KFs behave according to a Warburg meet the energy demand of increased cell viability. Moreover, enhanced GL was shown in keloid tissues and KFs, and this regulates not only glycolytic the increased ROS levels in keloids. The GLUT-1 inhibitor suppressed cell via is associated with the inhibited glycolytic function and reduced spare respira in KFs. These results indicated that increased GLUT-1 expression in keloids the metabolic reprogramming and increased fibroblast proliferation in keloi Keloids are characterized by the accumulation of extracellular matrix a broblast proliferation, both of which are energy-intensive processes [44,45]. oxidative phosphorylation, glycolysis meets the energy demand more qui larly in proliferating cells [32]. The adaptation of the cell's energy requirem the high anabolic rate and low catabolic cellular of active fibroblasts [46]. In glycolytic enzymes and glucose transporters are abnormally high in human fibroblasts [47]. In accordance with previous studies, our results demonstr behave toward the Warburg effect with a partial increment of OXPHOS (hi piratory capacity) ( Figure 1B,D). Enhanced glucose influx and increased g zymes may be pathophysiological mechanisms in keloids.
In our experiments, GLUT-1 was expressed in keloids and it mediated of glycolysis in keloids. Changes in the metabolic state and oxidative stress GLUT-1 expression [48]. GLUT-1, responsible for glucose intake, contribute sion, differentiation, and transformation in response to stresses and various cesses [49,50]. For example, many cancer cells overexpress GLUT-1 to fulfil extra glucose intake because GLUT-1 can be activated quickly [51]. By inhi 1, the proliferation of cancer cells was decreased dramatically in the breast a cell lung cancer [52]. In idiopathic pulmonary fibrosis, GLUT-1-dependent critical for parenchymal fibrosis and airway inflammation in a bleomycininjury model [53]. Glycolysis was significantly suppressed after transfection 1 shRNA or treatment with GLUT-1 inhibitor. The expression of collagen I an in lung fibroblasts decreased accordingly [54]. In our study, KFs showed hig Keloids are characterized by the accumulation of extracellular matrix and excess fibroblast proliferation, both of which are energy-intensive processes [44,45]. Compared to oxidative phosphorylation, glycolysis meets the energy demand more quickly, particularly in proliferating cells [32]. The adaptation of the cell's energy requirements mirrors the high anabolic rate and low catabolic cellular of active fibroblasts [46]. In lung fibrosis, glycolytic enzymes and glucose transporters are abnormally high in human and murine fibroblasts [47]. In accordance with previous studies, our results demonstrated that KFs behave toward the Warburg effect with a partial increment of OXPHOS (high spare respiratory capacity) ( Figure 1B,D). Enhanced glucose influx and increased glycolytic enzymes may be pathophysiological mechanisms in keloids.
In our experiments, GLUT-1 was expressed in keloids and it mediated enhancement of glycolysis in keloids. Changes in the metabolic state and oxidative stress can regulate GLUT-1 expression [48]. GLUT-1, responsible for glucose intake, contributes to cell division, differentiation, and transformation in response to stresses and various disease processes [49,50]. For example, many cancer cells overexpress GLUT-1 to fulfil the need for extra glucose intake because GLUT-1 can be activated quickly [51]. By inhibiting GLUT-1, the proliferation of cancer cells was decreased dramatically in the breast and non-small cell lung cancer [52]. In idiopathic pulmonary fibrosis, GLUT-1-dependent glycolysis is critical for parenchymal fibrosis and airway inflammation in a bleomycin-induced lung injury model [53]. Glycolysis was significantly suppressed after transfection with GLUT-1 shRNA or treatment with GLUT-1 inhibitor. The expression of collagen I and fibronectin in lung fibroblasts decreased accordingly [54]. In our study, KFs showed higher basal glycolytic rates, capacity, and glycolysis reserve, which provided higher glycolytic ATP output ( Figure 1B). After exposure to WZB117 treatment, ECAR (including glycolysis, glycolytic capacity, glycolytic reserve) was selectively decreased in KFs but not in NFs (Figure 4). In contrast, after WZB117 treatment, OCR (including basal respiration, maximal respiration, and spare respiratory capacity) was decreased in NFs, not in KFs ( Figure 5). Additionally, KFs showed augmented spare respiratory capacity compared to NFs ( Figure 5F). These results indicated that KFs rely heavily on GLUT-1-mediated glycolysis whereas NFs are dependent on OXPHOS in fibroblast proliferation.
The KFs in our study showed a GLUT-1-mediated induction of ROS. Induction of ROS is implicated in the dysregulation of apoptosis and can damage both DNA and protein in keloid tissues [55]. Increased ROS production in KFs suggests an impaired cellular antioxidant system in keloids [41,42]. The presence of ROS, the byproduct of OXPHOS, indicates the aerobic metabolism works in the process [56]. A vicious circle of ROSstimulated glucose uptake and glucose-stimulated ROS production can be triggered [57]. The increased ROS observed in KFs in this study is consistent with several previous reports [41,42], although the other study showed ROS generation was lower in KFs than in NFs. This discrepancy may result from the complex redox homeostasis mechanisms [36,58]. ROS promote fibroblast proliferation, myofibroblasts differentiation, EMT, and collagen deposition [59][60][61]. The glycolysis might be an adaptive shift to avoid the oxidative stress caused by OXPHOS [36]. Through NADPH generating from PPP and mediating transport of dehydroascorbic acid, glucose may serve as an antioxidant [62]. Evidence from previous works indicates that increased ROS levels stimulate increases in cellular glucose uptake at both slow and fast time scales. In L6 myoblasts, decreased GLUT-1 activity increases ROS levels, which suggests an ROS scavenging role for glucose [63]. In our study, basal ROS levels were increased in KFs. After WZB117 treatment, the ROS levels decreased in accordance with decreased fibroblasts proliferation (Figures 6 and 7). Taken together, these results suggest that the inhibition of glucose influx may regulate the development of oxidative stress and coordinate the glucose metabolism in keloids.
Although the activation of glycolysis is observed in wound healing processes, the expression profiles of glycolytic enzymes are not consistent in different studies and fibrotic models. How GLUT-1 regulates glycolysis or OXPHOS in fibroblasts is seldom discussed in the literature. Vinaik et al. reported that keloids derived from patients with extensive burns showed upregulations of several glycolytic enzymes, including GLUT-1, HK2, PFK1, and PFK2 [64]. That study utilized the tissue and fibroblasts derived from burn-induced keloids, but not those from common keloids, which we were investigating. Moreover, in that study, the blocking molecules for GLU-1 were different in that study and in our study (shikonin and WZB117, respectively). Using the radiation-induced fibrosis in human and mouse models, Zhao et al. reported that the upregulation of glycolysis contributes to increased extracellular matrix deposition, as evidenced by that GLUT-1 inhibition with WZB117-reduced ECM accumulation [65]. However, that study included patients who received full-dose radiotherapy to the neck for cancer treatment, for which the fibrotic skin model was different from this study in patients with common keloids. The intrinsic factors of cancer itself and the exogenous radiation therapy in these patients confounded the experimental findings in that study.
For glycolysis in keloids, Wang et al. showed that in hypoxia-promoted proliferation, GLUT-1 expression was enhanced, as well as migration and collagen synthesis, and autophagy in KF [66]. However, in that study, no blocking measures for GLUT-1 or other glycolytic enzymes were performed. Our study showed, with GLUT-1 blockage by WZB117, that GLUT-1 mediated enhanced glycolysis and decreased ROS formation in KFs (Figures 4 and 6, respectively) while increasing mitochondrial OXPHOs in NFs. Chen et al. investigated the inhibition of glycolysis-regulated KF proliferation through 2-DG, which is different from the inhibitor used in our study (WZB117). Moreover, that study measured the expression of several glycolytic enzymes, including HK2, PKM2, and LDHA, but not GLUT-1 [67]. Taken together, our study is distinct to others with ample experimental evidence showing that GLUT-1 mediated the enhanced glycolysis in keloids.
There are some limitations in this study. First, this study lacks a suitable keloid animal model that may limit the extrapolation of in vitro results to in vivo. Second, there were limited numbers of samples of tissues and tissue-derived fibroblasts (6 from keloids and 6 from healthy skin, 6 pairs). Relatively few keloid patients received surgical intervention due to a high likelihood of recurrence after surgery. Therefore, for research purposes, it is difficult to collect a large number of keloid samples, which is a general limitation of keloid research [2,4,68,69]. For this reason, not all the confounders could be matched, including age, gender, and the body sites, which may affect the intrinsic property of the fibroblasts in the present study. Further large-scale and animal studies may be required.

Conclusions
Our findings showed that human KFs exhibit Warburg-like effects of glycolysis in bioenergy utilization. By mediating glycolysis, GLUT-1 promoted KFs proliferation in concordance with increases in ROS and spare respiratory capacity. Whereas GLUT-1 has been identified as a therapeutic target for cancer [70], our study provides evidence that GLUT-1 is also a potential therapeutic target for keloids and other fibrotic diseases.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: All data generated or analyzed during this study are included in this published article.