TAZ inhibits glucocorticoid receptor and coordinates hepatic glucose homeostasis in normal physiological states

Essential revisions:

1. The effects of TAZ overexpression were much more impressive than its under-expression when looking at PTT. In fact if it weren't for the almost non-existent error bars for each BG measurement on the PTT I would almost doubt there is much of a significant effect of TAZ KO. How do the authors explain this? Is this because the mice were so fasted that hepatic TAZ levels are already so low that further reduction in its expression has little effect? This raises the issue of how physiological the level of overexpression of TAZ was.

During a PTT, the intraperitoneally injected pyruvate bolus elicits a glycemic excursion that reflects hepatic gluconeogenesis. Unlike a glucose tolerance test, in which injected glucose induces an increase in the blood glucose concentration up to several fold relative to the basal glucose concentration, we observed a small rise of the blood glucose concentration in the PTT (from ~50–75 mg/dl at the basal time point to ~100 mg/dl at 30 min after pyruvate injection), and thus much smaller variations within each group. The discrepancy of glucose readings at the basal time point among experiments reflects the length of fasting and when the fast started relative to the dark cycle. To demonstrate the small variation, we have plotted our PTT data using SEM (Author response image 1 A, C, and E) and SD (Author response image 1 D, and F).

Author response image 1

Download asset Open asset

We concluded that lowering hepatic TAZ expression via virus-mediated knockdown or genetic KO increases glucose production based on the following evidence. (i) Author response tables 1-3 contain our raw data of the PTT in mice infected with AdshTAZ or AdshCon ( Author response table 1) and L-TAZ KO and control floxed mice (Author response table 1; Author response table 2). The p values at each time point were determined by a two-way ANOVA after Sidak’s multiple comparison correction. (ii) Two-way ANOVA analyses showed significant differences between AdshTAZ versus AdshCon (p < 0.001) and L-TAZ KO versus floxed controls for both female and male mice (p < 0.001). (iii) Unpaired two-tail Student’s t-test analyses showed that the area under the curve significantly differed between AdshTAZ versus AdshCon (p < 0.001, Figure 2G) and L-TAZ KO versus floxed controls for both female and male mice (p < 0.001) (Figures 2L–M). We have included these statistical results in the figure legends to clarify our results and to support our conclusion (Author response table 3) .

In Figure 3, if I am to interpret this correctly, the level of TAZ in total liver was 3-fold higher in overexpressing mice than controls whereas in the pericentral regions it was expressed at comparable levels to endogenous. Does this mean that there is much TAZ expression in other parts of the liver where TAZ would normally not be found? This needs to be addressed in the manuscript.

We thank the reviewer for pointing out the distribution of overexpressed TAZ in the mouse liver. Figure 1E, Figure 1F, and Figure 1—Figure supplement 2 show that endogenous TAZ protein is primarily expressed in pericentral hepatocytes. However, Figure 3—Figure supplement 1 shows that TAZ was overexpressed not only in pericentral hepatocytes, but also in periportal hepatocytes (which did not endogenously express TAZ). This has been clarified in the revised manuscript.

2. It would be important to measure TAZ protein concentrations in liver of diabetic mouse models i.e hyperglycemic (genetic or nutritional models of diabetes and/or insulin resistance). Does TAZ affect the binding of GR to gluconeogenic gene promoters under hyperglycemic/diabetic conditions?

3. Would the overexpression of TAZ prevent the hyperglycemia characteristic of db/db mice for example? It would help determine the potential role of TAZ in pathophysiology.

We agree with the reviewer that it is important to understand how TAZ is regulated in the insulin resistant pathologic state and its role in that state. To address this, we measured expression of hepatic TAZ in the following insulin resistant mouse models: HFD-fed mice, leptin-deficient (ob/ob) mice, and liver-specific insulin receptor substrate 1 and 2 double KO mice (L-DKO, a model of hepatic insulin resistance and hyperglycemia). Hepatic TAZ protein expression was reduced in L-DKO mice (Author response image 2 A) but elevated in ob/ob mice (Author response image 2B) . In addition, HFD feeding increased the hepatic TAZ protein level (Figure 8D) and L-TAZ KO in HFD-fed mice increased the glucose concentration after mice were challenged with pyruvate (Figure 8E). Moreover, Figures 8A–C show that overexpression of hepatic TAZ reduced blood glucose concentration, glucose production, and mRNA expression of hepatic gluconeogenic genes in hepatic insulin resistant LDKO mice. Taken together these data suggest that hepatic TAZ regulates glucose production in insulin resistant hyperglycemic mice. However, our data also indicate that regulation of hepatic TAZ is complex in HFD-fed insulin resistant states because HFD feeding increased the hepatic TAZ protein level in fasted mice, but to a lesser extent in fed mice (Figure 8D).

Author response image 2

Download asset Open asset

4. Why do the authors not consider – and discuss – the possibility of regulation of glycogenolysis by TAZ-GR? The reduction in liver mass with TAZ knockdown seems more consistent with promoting glycogenolysis (as glycogen will take up more space/account for more liver mass than gluconeogenic precursors) vs gluconeogenesis.

We thank the reviewer for pointing this out. We have included data (Figure 4 ̶ Figure supplement 1) and discussed the regulation of glycogen metabolism by TAZ-GR given that GR is a key regulator of glycogen homeostasis.

5. The interpretation of the ITT is questionable: the authors state that there was no difference in insulin sensitivity, but if we calculate the plasma glucose concentrations during the ITT based on the time zero plasma glucose concentrations in the PTT, we would expect plasma glucose in the ITT to drop to ~60 mg/dl in the floxed mice and ~70 mg/dl in the L-TAZ KO animals. Given this degree of hypoglycemia, not only insulin sensitivity, but also hypoglycemia counterregulation (which involves glucocorticoids!) would modulate plasma glucose. Please discuss.

We thank the reviewer for pointing this out. We have added a new figure (Figure 2—Figure supplement 5C) plotting raw blood glucose concentrations. Original Figure 2—figure supplement 5B plotted glucose concentrations normalized to that at time zero as percentages. New Figure 5C shows that the glucose concentration in L-TAZ KO mice was increased at all time points throughout the ITT; the curve of L-TAZ KO mice paralleled that of control floxed mice. This suggests that the reduced glucose concentrations are primarily due to low basal glucose levels, not alterations in insulin sensitivity. Therefore, we concluded that L-TAZ KO does not significantly alter the insulin sensitivity of mice.

In addition, TAZ KO or overexpression did not significantly affect the plasma insulin or glucagon level (Figure 2—Figure supplement 3A and Figure 3—Figure supplement 2F) or activation of key components/targets of the insulin and glucagon signaling pathways in hepatocytes in vitro and in vivo (Figure 2—Figure supplement 2, Figure 3—Figure supplement 2F, and Figure 4—Figure supplement 3). Therefore, we conclude that regulation of glucose production by TAZ is unlikely due to direct regulation of insulin sensitivity or insulin/glucose signaling pathways. In addition, TAZ overexpression reduced blood glucose concentration, glucose production, and mRNA expression of gluconeogenic genes in L-DKO mice (Figures 8A–C), suggesting that TAZ inhibits gluconeogenesis independent of insulin signaling pathway. However, we agree with the reviewer that although our data show that TAZ inhibits the induction of gluconeogenic gene expression by a GR agonist, dexamethasone, the changes in glucose concentrations induced by hepatic TAZ overexpression or KO may lead to hormonal or non-hormonal compensatory regulation in the liver or extra-hepatic tissues.