Selenium Compounds Affect Differently the Cytoplasmic Thiol/Disulfide State in Dermic Fibroblasts and Improve Cell Migration by Interacting with the Extracellular Matrix

Abstract

Deficient wound healing is frequently observed in patients diagnosed with diabetes, a clinical complication that compromises mobility and leads to limb amputation, decreasing patient autonomy and family lifestyle. Fibroblasts are crucial for secreting the extracellular matrix (ECM) to pave the wound site for endothelial and keratinocyte regeneration. The biosynthetic pathways involved in collagen production and crosslinking are intimately related to fibroblast redox homeostasis. In this study, two sets of human dermic fibroblasts were cultured in normal (5 mM) and high (25 mM)-glucose conditions in the presence of 1 µM selenium, as sodium selenite (inorganic) and the two selenium amino acids (organic), Se-cysteine and Se-methionine, for ten days. We investigated the ultrastructural changes in the secreted ECM induced by these conditions using scanning electron microscopy (SEM). In addition, we evaluated the redox impact of these three compounds by measuring the basal state and real-time responses of the thiol-based HyPer biosensor expressed in the cytoplasm of these fibroblasts. Our results indicate that selenium compound supplementation pushed the redox equilibrium towards a more oxidative tone in both sets of fibroblasts, and this effect was independent of the type of selenium. The kinetic analysis of biosensor responses allowed us to identify Se-cysteine as the only compound that simultaneously improved the sensitivity to oxidative stimuli and augmented the disulfide bond reduction rate in high-glucose-cultured fibroblasts. The redox response profiles showed no clear association with the ultrastructural changes observed in matrix fibers secreted by selenium-treated fibroblasts. However, we found that selenium supplementation improved the ECM secreted by high-glucose-cultured fibroblasts according to endothelial migration assessed with a wound healing assay. Direct application of sodium selenite and Se-cysteine on purified collagen fibers subjected to glycation also improved cellular migration, suggesting that these selenium compounds avoid the undesired effect of glycation.

Keywords: HyPer biosensor; cellular migration; extracellular matrix; human dermic fibroblast; selenium cysteine; selenium methionine; sodium selenite.

Figure 1

Comparison of redox responses of HyPer-expressing human fibroblasts cultured at 5 and 25 mM glucose. In (A), the time courses of 37 single-cell recordings of HyPer obtained from cells maintained at 5 mM glucose (NG) are shown with grey lines and the average of this group is depicted with a black line. A white box and dotted line indicate the moment and period of exposure to 500 µM H₂O₂. (B) is similar to A, but here, 23 single-cell recordings come from to fibroblasts maintained at 25 mM glucose (HG). In both cases, imaging data were collected from four independent experiments. (C) compares baseline values of the biosensor obtained from NG fibroblasts (empty bar) and HG fibroblasts (filled bar). Basal values were collected from 5–6 random fields from each coverslip, a procedure repeated in four independent experiments. This procedure rendered 129 and 133 cells overall for NG and HG, respectively. The data correspond to the average ± SE, and the asterisk represents the significant difference between the groups according to Student’s t-test. In (D), HyPer signal recovery was obtained right after H₂O₂ removal, when the biosensor is oxidized and presents the maximal signal. Data correspond to the average ± SE of 37 cells for NG (empty circles) and 23 for HG (filled circles). These data sets present the average ± SE of 24 recordings for NG-PX-12 (empty triangles) and 23 cells for HG-PX-12 (filled triangles). (E) compares the recovery observed in the four groups from (D) with data corresponding to average ± SE. The asterisk indicates significant differences between the groups according to one-way ANOVA with Bonferroni post hoc tests.

Figure 2

Redox alterations induced by the treatment with selenium compounds in human fibroblasts. (A) Comparison of baseline values obtained from HyPer-expressing fibroblasts maintained at 5 (empty bars) and 25 mM glucose (filled bars) and treated for ten days with sodium selenite (SS, 1 µM), selenocysteine (SeCys, 1 µM), or selenomethionine (SeMet, 1 µM). For this, baseline values from 136 to 170 cells were collected from four independent experiments. The data correspond to the average ± SE, and the asterisks represent significant differences between the treated groups versus non-treated according to one-way ANOVA with post hoc Bonferroni tests. (B) the time course of HyPer signal increase in NG fibroblasts upon hydrogen peroxide exposure. The data correspond to the average ± SE of 37 cells for NG (green), 29 cells for NG-SS (empty squares), 23 cells for NG-SeCys (filled stars), and 37 cells for NG-SeMet (filled triangles) conditions. The inset graph shows the rate constants (b) obtained from data fitting to the function $HyPer ratio=\left(Maximal-baseline\right)\ast e^{bt}$. The asterisks indicate significant differences between the treated and non-treated groups according to Kruskal–Wallis tests with Dunn post hoc tests. (C) Identical to B, but here time courses of HyPer signal increase were obtained from 23 HG fibroblasts (red), 23 cells for HG-SS (1 µM) (empty squares), 28 cells for HG-SeCys (1 µM) (filled stars), and 33 cells for HG-SeMet (1 µM) (filled triangles) conditions. (D) the time courses of HyPer signal recovery obtained from NG fibroblasts (green) and subjected to treatments with SS (empty square), SeCys (filled stars), and SeMet (filled triangles). The data are expressed as the percentage of the maximal signal obtained from the oxidized biosensor and correspond to the average ± SE. The inset graph shows the rate constants (b) obtained from fitting data from each single-cell recording to an exponential decay function. (E) similar to (D), with the difference that here the data correspond to HG fibroblasts (red), with the same symbols for the treatments. The asterisk indicates a significant difference within the NG and HG groups according to Kruskal–Wallis tests with Dunn post hoc tests. (F) the comparison of the PX-12 effects on the recovery (1 µM, gray bars) with those that are untreated. In the left panel are the NG fibroblasts treated or not with selenium compounds. In the right panel, we have the percentage of recovery obtained from HG fibroblasts. The data correspond to average ± SE and the asterisks indicate significant differences within the groups according to one-way ANOVA with Bonferroni post hoc tests.

Figure 3

Effect of selenium compound treatments on the abundance of extracellular collagen secreted by CCD1068Sk cells. In (A), the panel of images shows the staining with Picrus Sirius Red on the surface that CCD1068Sk cells left after decellularization. The black bar in the lower right corner of each image represents 250 µm. NG and HG fibroblasts were treated with 1 µM sodium selenite (SS), selenocysteine (SeCys), or selenomethionine (SeMet). In (B), the area covered by the staining with Picrus Sirius Red was quantified by following a sequential step that included background subtraction/RGB stack/threshold in the ImageJ software. Experiments were conducted in triplicate and repeated three times. The data are expressed as average ± SE. The asterisks indicate significant differences within the NG (empty bars) and HG (filled bars) groups with different selenium compounds according to Kruskal–Wallis tests with post hoc Dunn tests.

Figure 4

Effect of high glucose and selenium compounds on thickness, numbers of branches, and crosslinking of fibers secreted by dermal fibroblasts. CCD1068Sk cells were maintained under the collagen synthesis protocol for 10 days in normal glucose (NG, 5 mM) and high glucose (HG, 25 mM) in the presence of 1 µM selenium compounds: sodium selenite (SS), selenocysteine (SeCys), and selenomethionine (SeMet). Subsequently, the samples were decellularized and fixed. (A) a set of representative images taken with a scanning electron microscope at 50,000×. The black bar in the lower right corner of each image represents 2 µm. (B) refers to a comparison of thickness, (C) shows the number of branches, and (D) indicates crosslinking nodes of at least 50 fibers per condition [NG, empty bars; HG: filled bars]. The number of branches was obtained after skeletonizing the images and only branches associated with an intersection were considered. At least 9000 intersections were used for this analysis. The data represent the average ± SE from two to three independent trials. The asterisks indicate significant differences within NG or HG groups according to Kruskal–Wallis tests with post hoc Dunn tests.

Figure 5

Sodium selenite and SeCys but not SeMet restore the quality of ECM produced by HG fibroblasts and glycated collagen fibrils for endothelial migration. (A) shows a panel of representative images of wound assays performed on TIME cells taken 8 h after growing in ECM produced by NG and HG fibroblasts in the presence of 1 µM sodium selenite (SS), selenocysteine (SeCys), or selenomethionine (SeMet). For scale, the black bar at the corner of the image represents 500 µm. In (B), the quantification of the gap area filled by TIME cells (p6-10). The wound closure percentage is the average ± SE of 18 observations from three independent experiments. The hash indicates significant differences in the treatments compared to the NG group according to Kruskal–Wallis tests with post hoc Dunn tests. In (C), similar to (A), TIME cells were seeded on purified collagen type I treated or not with methylglyoxal (MGO) and selenium compounds. (D) shows the quantification of the gap area filled by TIME cells (p6-10). The data represent the average ± SE from 18 observations from three independent experiments. The hash indicates significant differences in the treatments compared to the native group according to Kruskal–Wallis tests with post hoc Dunn tests.

Figure 5

Sodium selenite and SeCys but not SeMet restore the quality of ECM produced by HG fibroblasts and glycated collagen fibrils for endothelial migration. (A) shows a panel of representative images of wound assays performed on TIME cells taken 8 h after growing in ECM produced by NG and HG fibroblasts in the presence of 1 µM sodium selenite (SS), selenocysteine (SeCys), or selenomethionine (SeMet). For scale, the black bar at the corner of the image represents 500 µm. In (B), the quantification of the gap area filled by TIME cells (p6-10). The wound closure percentage is the average ± SE of 18 observations from three independent experiments. The hash indicates significant differences in the treatments compared to the NG group according to Kruskal–Wallis tests with post hoc Dunn tests. In (C), similar to (A), TIME cells were seeded on purified collagen type I treated or not with methylglyoxal (MGO) and selenium compounds. (D) shows the quantification of the gap area filled by TIME cells (p6-10). The data represent the average ± SE from 18 observations from three independent experiments. The hash indicates significant differences in the treatments compared to the native group according to Kruskal–Wallis tests with post hoc Dunn tests.