CELL CULTURE

Three types of MLO-Y4 cells Cx43+/+ (presenting an empty vector as a negative control to be able to evaluate the effects of non-silencing) and Cx43-/- (cells deficient in connexin 43, transfected with an RNAi) were used, which were kindly provided by Dr. L. I. Plotkin, and Cx43+/+ cells in which α2 integrin was silenced by RNAi. These cells were seeded at a concentration of 24,000 cells/cm² and cultured with α-Modified Eagle’s Medium (α-MEM) (Gibco, ThermoFisher Scientific, ES) supplemented with 2.5 % calf serum (Calf Serum; CS), 2.5 % fetal bovine serum (Fetal Bovine Serum; FBS), 1 % L-glutamine, 1 % penicillin/streptomycin and Puromycin from Streptomyces alboniger (Sigma Aldrich, ES) at a concentration of 10 μg/ml.

All the surfaces on which these cells were seeded had to be previously collagenized with type I collagen at 0.01 % acetic acid13. Cells were kept at 37 °C and 5 % CO₂.

SILENCING OF α2 INTEGRIN

MLO-Y4 Cx43+/+ cells were transfected with three different α2 silencers (RNAi) (5 nM) (ThermoFisher Scientific, ES), targeting the α2 integrin sequence, using lipofectamine RNAiMax (Life Technologies, Carlsbad, CA, USA). The siRNAs were added in serum-free medium for 24 h. The scrambled sequence (SCR) (RNAi control, Santa Cruz Technology, TX, USA) was used as a negative control to evaluate the non-targeted effects of silencing (RNAi off-targeted).

PCR

RNA extraction was performed with TRIZOL^® (Ambion, FosterCity, CA, USA). For RNA reverse transcription (RTPCR, reverse transcriptase polymerase chain reaction) the kit (Applied Byosistems, Grand Island, NY, USA) and the thermocycler (Eppendorf, Hamburg, Germany) were used.

To analyze the expression of the α2 integrin (Fw 5´CCATGATGGGTCGAAGCTGA3´; Rv 5´CTTCGTCGGCCACATTGAAA3´) SYBR Green (Sybr promix ex Taq, Takara, Otsu, Japan) was used. The level of α2 integrin expression was analyzed using β-actin as control gene (Fw 5´GAACCCTAAGGCCAACCGTG3´; Rv 5´ACCAGAGGCATACAGGGACAG3´). Triplicates of all conditions (Cx43+/+ and Cx43-/-) were performed. The expression change of the genes was calculated based on the value of 2^-ΔΔCt.

IMMUNOFLUORESCENCE

Multiwell plates (Falcon, ES) were seeded at 30,000 cells/well. Cells were grown until they reached 80 % confluency and serum-free medium was added for 24 h to induce primary cilium formation. Cells were then fixed with 4 % paraformaldehyde (PFA) and permeabilized with 0.5 % Triton X-100. Next, the blocking solution composed of 10 % bovine serum albumin (BSA), supplemented with 5 % goat serum, was added for 1 h. Subsequently, the following primary antibodies were kept overnight at 4 °C under agitation: polyclonal anti-Cx43 produced in rabbit (Sigma, ST. Louis, MO, USA) (1:1000 dilution in 10 % BSA and 5 % goat serum); and mouse monoclonal acetylated anti-α tubulin (Sigma) (1:1000 dilution in 10 % BSA and 5 % goat serum), to observe the primary cilium. Secondary antibodies were then added: Alexa fluor^® 488 goat anti-mouse primary cilia (Invitrogen Molecular probes, Thermo Fisher ScientificTM, ES) (1:1000 dilution in 10 % BSA and 5 % goat serum), to Cx43 Alexa fluor^® 568 anti-rabbit IgG (Live technologies, Thermo Scientific™, ES) (1:1000 dilution in 10 % BSA and 5 % goat serum). After 1 h of incubation, 4’-6-diamidino2-phenylindole (DAPI) (1:10,000 dilution) was added. The nuclei, the primary cilium and the Cx43 were visualized with the epifluorescence microscope (Leica CTR 6000). Triplicates were imaged and a total of 100 cells from each condition (Cx43+/+ SCR and α2 and Cx43-/-) were analyzed. The fusion of the individual images of the primary cilium, Cx43 and cell nuclei into one was performed with the ImageJ program, which allows processing digital images, is capable of calculating the area and the statistics of the value of the pixel selected by the user and to measure distances.

STIMULATION BY PTH1R AGONISTS (PTH AND PTHRP)

To perform stimulation by PTHrP and PTH, cells were plated at a density of 25,000 cells/cm² and both ligands were added at a concentration of 10-7 molar (M), for 10 min.

CELL DEATH AND ADHESION ASSAY

Cell viability was determined by trypan blue exclusion, a method that stains dead cells blue, allowing the percentage of live and dead cells to be calculated with respect to the total using a Neubauer camera and a bright field optical microscope (Leica DM5500B). to do the counting.

After trypsinizing the cells and 30 minutes after reseeding, non-adherent cells were counted with the Neubauer chamber. Next, images of different fields of the cells adhered to the Petri dish were taken, using an inverted bright field microscope (Leica DM5500B), to calculate the percentages of cells adhered to the total. A total of 9 fields of each condition were analyzed in triplicate (Cx43+/+ SCR and α2 and Cx43-/-).

WESTERN BLOT ANALYSIS

Protein extraction was performed for each condition in duplicate (Cx43+/+ SCR and α2 and Cx43-/-), using RIPA Buffer (150 mM NaCl, 1.0 % IGEPAL^® CA-630, 0.1 % SDS (sodium dodecyl sulfate), 50 mM Tris, pH 8, Sigma-Aldrich, St Louis, MO, USA), protease inhibitor (PI) (1:100 dilution, Calbiochem^®, ES) and phosphatase inhibitor (IF) (1:100 dilution, Calbiochem^®, IT IS). Proteins were quantified by the bicinchoninic acid (BCA) assay (Pierce™ BCA Protein Assay Kit, Thermo Fisher Scientific, ES). To carry out the reading, the Variouskan Flash plate reader (Thermo Fisher Scientific, ES) was used and three readings were made at 562 nm using the SkanIt Software 2.4.3 RE program.

Protein extracts were separated using a polyacrylamide gel. Proteins were subsequently transferred to a nitrocellulose membrane (Bio-Rad, Hercules, CA, USA). The blocking was carried out with 5 % milk powder in Tris saline buffer with Tween20 (TTBS) at 0.05 %, for 1 h under stirring at room temperature. Next, the following primary antibodies were incubated for 24 hours at 4 °C and under agitation: anti-phospho-p44/42 MAPK (Erk 1/2) (Cell Signaling, Beverly, MA, USA), anti-p44/42 MAPK (Erk 1/2), and anti-α-tubulin (Sigma Aldrich, ES). Subsequently, secondary antibodies were added. Chemiluminescence development was carried out with Clarity^TM Western ECL substrate (Bio-Rad). The intensity of the bands was quantified by densitometry, using Dnr Bio Imaging System MF ChemiBIS3.2 and the programs Gelcapture and QuantityOneTM (Bio-Rad).

STATISTICAL ANALYSIS

The confidence limit established in all the statistical tests was 95 %. Therefore, results with a p value (p) p < 0.05 are considered statistically significant.

For the comparison of means ± standard deviation (SD), the GraphPad Prism 8 program was used. To compare means of more than two groups, the ANOVA test of a single factor and the ANOVA Welch test were used. Equality of variances was not considered. For multiple analyses, Dunnett’s test, Tukey’s test and the non-parametric Kruskal Wallis test were used. Being t the statistic that analyzes if the measures of the two conditions are equal or not and gl that would be the degrees of freedom that indicate the number of values that can be assigned arbitrarily.

INTEGRIN GENE EXPRESSION MODULATED BY CONNEXIN 43 IN OSTEOCYTIC CELLS

The RNA expression of integrins α2, α6, β1, β3 and β6 and Cx43 was analyzed by means of quantitative PCR, to study the possible relationship in the expression of these two families of proteins.

Figure 1 shows how the α2 integrin significantly decreases its expression level in cells of the Cx43-/- line compared to Cx43+/+ cells (t = 13.93, gl = 4, p = 0.0002). On the contrary, the α6, β1 and β3 integrins significantly increase their expression in the Cx43-/- cell line compared to Cx43+/+ (t = 3.646, df = 4, p = 0.0219; t = 5.501, g = 4, p = 0.0053, t = 26.18, gl = 4, p < 0.0001, respectively). In the case of β6, no significant differences were observed between the two cell lines (t = 0.99, df = 4, p = 0.378).

Figure 1. RNA expression of integrins α2, α6, β1, β3 and β6 in Cx43 +/+ and Cx43-/- cells. The results are expressed as mean ± SD of an experiment performing triplicates of each experimental condition *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. 

These results indicate that the expression of connexin 43 conditions the expression pattern of various integrins in osteocytic cells.

CONNEXIN 43 AND INTEGRIN α2 REGULATE THE LENGTH OF THE PRIMARY CILIUM IN OSTEOCYTIC CELLS

Immunofluorescence was performed to observe the possible relationship and interaction between the primary cilium, Cx43 and α2 integrin, three well-known mechanosensors of osteocytes, and to determine if the development and length of the primary cilium could depend on the presence of Cx43 and α2 integrin (Fig. 2).

Figure 2. Immunofluorescence of MLO-Y4 Cx43+/+ (SCR and RNA α2) and Cx43 -/- cells. The cell nucleus was visualized with DAPI (blue), the primary cilium with the anti-α-acetylated tubulin antibody (green) and the Cx43 with the anti-Cx43 antibody (red). Images were captured with a confocal fluorescence microscope (40X). Scale bar = 50 μm. 

MLO-Y4 Cx43+/+, MLO-Y4 Cx43+/+ cells silenced with α2 integrin (RNAi α2), and MLO-Y4 Cx43-/- cells were compared. In the results shown in figure 3, it was observed that all cell lines are capable of developing primary cilium and that this organelle originates on the cell surface. In addition, it was observed that Cx43 presents a different distribution in Cx43+/+ and α2 RNAi cells, since in cells in which the α2 integrin was not silenced, the presence of Cx43 predominates in the cell membrane, an expected location, since it is where it forms hemichannels and UCs. While in the conditions in which the cells do not present α2 integrin, Cx43 is distributed in the cell cytoplasm and not so focused on the plasma membrane. Immunofluorescence images also showed that Cx43 and the primary cilium do not co-localize.

Figure 3. Comparison of primary cilium length (μm) between lines Cx43+/+ (SCR and RNAi α2) and Cx43-/-. A. The results show the mean value ± SD of the length of the primary cilium in mean μm from the images made at different fields. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. B. Representative images in which the nucleus can be visualized by DAPI (blue) and the primary cilium by anti-α-acetylated tubulin antibody (green). The images were captured with a confocal fluorescence microscope (40X). Scale bar = 50 μm. 

In order to quantitatively compare the development of the primary cilium and its length, images of each cell type (Cx43+/+ SCR, Cx43+/+ RNAi α2 and Cx43-/-) were analyzed. These images were taken from different fields using fluorescence microscopy (40X).

Statistical analyzes were carried out comparing the means ± SD of the length in μm of all the cilia analyzed in the cells Cx43+/+ SCR (2.37), Cx43+/+ RNAi α2 (2.08) and Cx43-/- (1.52). The results obtained indicate that the length of the primary cilium in cells in which α2 integrin was silenced and in cells deficient in Cx43 was significantly less than in scrambled cells. The p-value was: p = 0.0017 for the comparison of SCR vs RNAi α2, p < 0.0001 SCR vs Cx43-/- and p < 0.0001 RNAi α2 vs Cx43-/- (Fig. 3).

Based on these observations, we can suggest that connexin 43 and integrin α2 are involved in primary ciliary elongation in osteocytic cells.

REGULATION OF ERK KINASE 1/2 PHOSPHORYLATION BY CONNEXIN 43 IN OSTEOCYTIC CELLS

To characterize the response of MLO-Y4 cells (Cx43+/+ and Cx43-/-) to stimulation by agonists, the PTH1R, PTH and PTHrP (1-37) receptor ligands were added to the culture medium at a concentration of 10^-7 M, for 10 min. Immediately afterwards, protein extraction was performed to analyze P-ERK expression by Western blot. The purpose of this experiment was to determine if Cx43 deficiency influenced the activation of the PTH1R receptor after stimulating it with their respective ligands (PTH and PTHrP) and to analyze the intracellular signaling pathways that this stimulus activates and the cellular response that it triggers, specifically the ERK 1/2 phosphorylation.

The results obtained show that in Cx43+/+ cells, stimulation with PTH and PTHrP increases ERK 1/2 phosphorylation (mean ± SD of the duplicates made) with respect to cells that were not stimulated with any ligand (CE). In the case of Cx43-/- cells, PTH stimulation decreased ERK phosphorylation relative to Cx43+/+ cells that were stimulated with PTH, and the agonist peptide PTHrP did not stimulate Cx43-/- cells. No significant variations were obtained in the levels of P-ERK between the different experimental conditions (Fig. 4).

Figure 4. Analysis of P-ERK after stimulating MLO-Y4 Cx43+/+ and Cx43-/- cells with PTH and PTHrP. In Cx43+/+ cells, stimulation with PTH and PTHrP increases ERK 1/2 phosphorylation vs. cells that were not stimulated with any ligand (CE) (2.200 ± 0.696 vs. 0.4240 ± 0.150 and 1.11 ± 0.554 vs. to 0.4240 ± 0.150, respectively). In Cx43-/- cells, PTH stimulation decreased ERK phosphorylation relative to Cx43+/+ cells that were stimulated with PTH (0.865 ± 0.003 vs. 2.20 ± 0.696). Total ERK and tubulin values were used to normalize P-ERK values. Results are expressed as mean ± SD of a duplicate experiment of each experimental condition vs. CE. 

Our observations suggest that connexin 43 regulates ligand-activated PTH1R-dependent phosphorylation of ERK 1/2.

CX43 EXPRESSION REGULATES VIABILITY AND ADHESION OF OSTEOCYTIC CELLS

A cell death assay was performed with Cx43+/+ SCR, RNAi α2 and Cx43-/- cells in order to determine whether the lack of expression of integrin α2 and Cx43 would modify the signaling pathways involved in the cell response, causing its apoptosis. Figure 5A shows how the number of live Cx43-/- cells decreases significantly with respect to Cx43+/+ SCR cells (p = 0.0035). In the case of the analysis of dead cells, we see that their number increases significantly in Cx43-/- cells compared to Cx43+/+ SCR (p = 0.0074).

Figure 5. Analysis of cell death (A) and cell adhesion (B) in Cx43+/+ (SCR and α2 ARNi) and Cx43-/- cells. The results show the mean ± SD values of live and dead cell counts from a single experiment with six replicates for each condition, and the counts of adhered and non-adhered cells on the plate. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. 

As for the Cx43+/+ RNAi α2 cell line, as there were no significant differences in the number of live cells (although there was a slight tendency towards a decrease in their number) and it presented a very similar number of dead cells with respect to cells. Cx43+/+ SCR, it is suggested that the silencing of the α2 integrin influences the proliferation of osteocytes and not their death.

On the other hand, an adhesion assay was also performed on Cx43+/+ SCR, RNAi α2 and Cx43-/- cells. Figure 5B shows that the number of Cx43-/- cells adhered to the collagen plaque decreases significantly with respect to Cx43+/+ SCR and α2 RNAi cells (p < 0.001 and p = 0.0039, respectively). In the case of counting non-adherent cells, their number increased significantly both in Cx43+/+ RNAi α2 cells and in Cx43-/- compared to Cx43+/+ SCR (p = 0.0146 and p = 0.0134, respectively).

Cell adhesion was quantified using brightfield images, which also allowed the morphology of the different cell types to be analyzed, and it was observed that the Cx43+/+ SCR cells had a more elongated shape and a greater number of dendritic extensions, which would potentially allow them to have a higher adhesion capacity. In contrast, Cx43+/+ RNAi α2 and Cx43-/- cells showed a more rounded structure, probably due to the absence of both integrins and connexins, both transmembrane proteins that favor cell adhesion (⁴).

These results are related to those obtained in the cell death assay, since we observed a greater number of dead cells in the Cx43-/-, due to the lack of Cx43, adhesion to the culture surface would be difficult, which could cause cells to detach and die due to lack of substrate binding.