Adenylyl cyclase 3 regulates osteocyte mechanotransduction and primary cilium

Abstract

Osteocytes are accepted as the primary mechanosensing cell in bone, but how they translate mechanical signals into biochemical signals remains unclear. Adenylyl cyclases (AC) are enzymes that catalyze the production of second messenger cyclic adenosine monophosphate (cAMP). Osteocytes display a biphasic, cAMP response to fluid shear with an initial decrease in cAMP concentrations and then an increased concentration after sustained mechanical stimulation. To date, AC6, a calcium-inhibited AC, is the primary isoform studied in bone. Since osteocytes are calcium-responsive mechanosensors, we asked if a calcium-stimulated isoform contributes to mechanotransduction. Using a transcriptomic dataset of MLO-Y4 osteocyte-like cells from the NIH Gene Expression Omnibus, we identified AC3 as the only calcium-stimulated isoform expressed. We show that inhibiting AC3 in MLO-Y4 cells results in decreased cAMP-signaling with fluid shear and increased osteogenic response to fluid flow (measured as Ptgs2 expression) of longer durations, but not shorter. AC3 likely contributes to osteocyte mechanotransduction through a signaling axis involving the primary cilium and GSK3β. We demonstrate that AC3 localizes to the primary cilium, as well as throughout the cytosol and that fluid-flow regulation of primary cilia length is altered with an AC3 knockdown. Regulation of GSK3β is downstream of the primary cilium and cAMP signaling, and with western blots we found that GSK3β inhibition by phosphorylation is increased after fluid shear in AC3 knockdown groups. Our data show that AC3 contributes to osteocyte mechanotransduction and warrants further investigation to pave the way to identifying new therapeutic targets to treat bone disease like osteoporosis.

Keywords: Adenylyl cyclase 3; Gsk3beta; Mechanotransduction; Osteocyte; Primary cilium; cAMP.

Figure 1:

AC3 inhibition results in increased osteogenic response to fluid shear. (A) AC3 knockdown resulted in a significant decrease in gene expression (N=7 experimental repeats with an average knockdown created from 3–4 control and knockdown samples per experiment). (B) In static samples, Ptgs2 gene expression is unchanged. (C) Dynamic fluid-flow resulted in an increased expression of Ptgs2 in the AC3 KD group relative to controls only after 60 minutes of flow (C). Statistical analysis included a one-sample t-test (A), unpaired t-test (B), and 1-way ANOVA followed by Bonferroni's multiple comparison test (C). +++ p < 0.001. *** p < 0.001 compared to control of same flow duration. N = 6–11 from three experimental repeats for 2, 15, and 30 minutes of flow, and N = 16–18 from four experimental repeats for 60 minutes of flow.

Figure 2:

cAMP levels are reduced with an AC3 knockdown. (A) Static cAMP levels, while diminished in AC3 knockdown samples, were not statistically different from control. (B) After fluid flow application, static normalized flow samples were lower in AC3 knockdown samples, though only significantly reduced after 60 minutes of fluid flow. N=6 samples. Data analysis included a t-test (A) and 1-way ANOVA followed by Bonferroni's multiple comparison. ** p < 0.01 between control and AC3 KD. + p < 0.05 and ++ p < 0.01 for one-sample t-test of flow/static normalized cAMP levels compared to one (i.e. no change with flow).

Figure 3:

AC3 localizes to the primary cilium and contributes to flow-induce ciliary regulation. (A) Immunostaining revealed that AC3 localizes to the primary cilium in MLO-Y4 cells and murine long-bone osteocytes. NucBlue was used to stain the nucleus. Anti-acetylated alpha-tubulin marked the primary cilium and an anti-AC3 antibody was used to detect AC3 localization. In vitro images are confocal maximum projections and in vivo images are maximum projections from deconvolved confocal microscopy data. Scalebar = 5 μm. (B) Static primary cilia length is similar between control and AC3 knockdown groups. (C) Primary cilia shorten with flow in knockdown cells, while there is no significant change in control. N = 6. **p< 0.01 between experimental groups. + p < 0.05 for one-sample t-test of flow/static normalized length to one (i.e. no change with flow).

Figure 4:

Loss of AC3 alters GSK3β regulation (A) There was no difference in the total content of Creb, but there was greater variance in the AC3 KD group. (B) Flow caused a significant increase in Creb phosphorylation in the control group (+++ p < 0.001), but not with a knockdown (p = 0.066). (D) There was no difference in the total Gsk3β content in static samples, but with flow the amount of GSK3β increased in the knockdown group (* p < 0.05 from control flow. + p < 0.05 from AC3 KD static). (E) Flow-induced fold-change in GSK3β-phosphorylation was not affected by flow application, but was significantly different with AC3 KD (* p < 0.05). One-way ANOVA followed by Bonferroni Multiple comparison tests were performed. One-sample two-sided t-test were performed on static-normalized flow data with a theoretical mean of one. Data (N = 5–6) are from three independent experiments with one western blot shown (C,F). Full blots are shown in supplemental material.