Cellular cholesterol modifies flow-mediated gene expression

Abstract

Downregulation of heme oxygenase-1 (HO-1), cyclooxygenase-2 (COX2), and nitric oxide synthase-2 (NOS2) in the kidneys of Dahl rodents causes salt sensitivity, while restoring their expression aids in Na⁺ excretion and blood pressure reduction. Loading cholesterol into collecting duct (CD) cells represses fluid shear stress (FSS)-mediated COX2 activity. Thus, we hypothesized that cholesterol represses flow-responsive genes necessary to effectuate Na⁺ excretion. To this end, CD cells were used to test whether FSS induces these genes and if cholesterol loading represses them. Mice fed either 0% or 1% cholesterol diet were injected with saline, urine volume and electrolytes were measured, and renal gene expression determined. FSS-exposed CD cells demonstrated increases in HO-1 mRNA by 350-fold, COX2 by 25-fold, and NOS2 by 8-fold in sheared cells compared with static cells (P < 0.01). Immunoblot analysis of sheared cells showed increases in HO-1, COX2, and NOS2 protein, whereas conditioned media contained more HO-1 and PGE₂ than static cells. Cholesterol loading repressed the sheared mediated protein abundance of HO-1 and NOS2 as well as HO-1 and PGE₂ concentrations in media. In cholesterol-fed mice, urine volume was less at 6 h after injection of isotonic saline (P < 0.05). Urinary Na⁺ concentration, urinary K⁺ concentration, and osmolality were greater, whereas Na⁺ excretion was less, at the 6-h urine collection time point in cholesterol-fed versus control mice (P < 0.05). Renal cortical and medullary HO-1 (P < 0.05) and NOS2 (P < 0.05) mRNA were repressed in cholesterol-fed compared with control mice. Cholesterol acts to repress flow induced natriuretic gene expression, and this effect, in vivo, may contribute to renal Na⁺ avidity.

Keywords: blood pressure; collecting duct; flow; shear.

Fig. 1.

Fluid shear stress (FSS) induces heme oxygenase-1 (HO1), cyclooxygenase-2 (COX2), and nitric oxide synthase-2 (NOS2) mRNA expression compared with static controls. Inner medullary collecting duct 3 cells were exposed to 0.4 dyn/cm² of FSS (n = 4–6) or maintained under static conditions (n = 3–6) from 2 to 6 h. Total RNA was extracted from shear-exposed and static cells, and PCR was performed using specific gene-specific TaqMan primers. Steady-state mRNA expression of HO-1 (51.8 ± 6.2-, 63.0 ± 2.6-, and 351 ± 35-fold increase, respectively) and COX2 (6.3 ± 0.7-, 4.0 ± 0.7-, and 29.5 ± 2.0-fold increase, respectively) increased at 2, 4, and 6 h in shear-exposed cells versus static control cells (#P < 0.05). NOS2 mRNA abundance increased above static control at 4 and 6 h (2.5 ± 0.2- and 8.5 ± 0.9-fold, respectively; #P < 0.05 vs. the static group).

Fig. 2.

Protein abundance of cyclooxygenase-2 (COX2) and heme oxygenase-1 (HO-1) in inner medullary collecting duct 3 (IMCD3) cells exposed for up to 4 h. A: COX2 and HO-1 protein expression by Western blot analysis in fluid shear stress (FSS)-exposed and static IMCD3 cells. B: densitometry demonstrating no difference in COX2 expression at 2 h; however, at 4 h, shear-exposed cells expressed ~2.5-fold increased COX2 (#P < 0.05) compared with static cells. C: HO-1 densitometry similarly showed no effect of FSS at 2 h; however, at 4 h, there was an ~6-fold (#P < 0.05) increase in protein abundance compared with static cells.

Fig. 3.

Nitric oxide synthase-2 (NOS2) protein expression increases at 2 and 4 h in fluid shear stress (FSS)-exposed cells. A: Western blot analysis of NOS2 expression in cells exposed to 0.4 dyn/cm² of FSS versus static controls. B: NOS2 abundance was greater in shear-exposed cells at 2 h (#P < 0.05) and 4 h (#P < 0.05) versus static controls.

Fig. 4.

Fluid shear stress (FSS) induces cyclooxygenase-2 (COX2), nitric oxide synthase-2 (NOS2), and heme oxygenase-1 (HO-1) abundance at 6 h. A: Western blot analysis of protein lysates generated from inner medullary collecting duct 3 cells exposed to FSS versus static controls to measure immunodetectable COX2, NOS2, and HO-1. B: densitometry illustrating the ~2-fold increase in COX2, ~6-fold increase in NOS2, and ~20-fold increase in HO-1 in FSS-exposed cells (#P < 0.05) compared with static cells.

Fig. 5.

PGE₂ and heme oxygenase-1 (HO-1) expression in the media of fluid shear stress (FSS)-exposed and static cells. A: PGE₂ concentration in the media bathing the cells, normalized to the total amount of protein lysate, was significantly greater at each time point compared with static controls (#P < 0.05). B: HO-1 concentration in the media bathing the cells, normalized to the total amount of protein lysate, was not different at 2 h but was significantly greater at 4 and 6 h compared with static controls (#P < 0.05).

Fig. 6.

Cellular cholesterol content after cholesterol loading and this effect on ATP-binding cassette subfamily A member 1 (ABCA1) gene expression. A: inner medullary collecting duct 3 (IMCD3) cells were loaded with a cholesterol/methyl-β-cyclodextrin mixture for ~36 h, whereas control cells were maintained under serum-free conditions. Cellular cholesterol content was ~2-fold greater in cholesterol-loaded cells (#P < 0.05) versus control cells. B: ABCA1 expression was increased >3-fold in cholesterol-treated cells (#P < 0.05) compared with control cells.

Fig. 7.

Cholesterol-loaded cells express less shear-induced nitric oxide synthase-2 (NOS2) and heme oxygenase-1 (HO-1) than control cells after 4 h of fluid shear stress (FSS). A: Western blot analysis of control and cholesterol-treated inner medullary collecting duct 3 cells after exposure to 4 h of FSS. Steady-state protein abundance of NOS2 and HO-1 was less in cholesterol-treated versus control cells. B: densitometric analysis of control cells (n = 22–24) and cholesterol-loaded cells (n = 22–24) cells confirmed that NOS2 and HO-1 protein expression was reduced in cholesterol-treated cells (*P < 0.05 vs. control cells); however, COX2 was unchanged.

Fig. 8.

Soluble PGE₂ and heme oxygenase-1 (HO-1) are reduced in conditioned media of cholesterol-treated compared with control inner medullary collecting duct 3 cells after 4 h of fluid shear stress. Concentrations of PGE₂ and HO-1 were determined (in pg/ml) and normalized to the quantity of cellular protein lysate (in µg). Data sets were then normalized to the control shear. PGE₂ and HO-1 in conditioned media was only 0.35 ± 0.07 and 0.26 ± 0.03 in cholesterol-loaded cells (#P < 0.05), respectively, compared with control sheared cells.

Fig. 9.

Urine volume as a percentage of injected saline measured sequentially from 2 to 6 h in control and cholesterol-fed mice. Male FVB mice fed either 0% or 1% cholesterol diet for 6 wk were placed into a metabolic cage and injected with isotonic saline at 5% of body weight, and urine was collected sequentially from 2 to 6 h. Urine volume is presented as a percentage of the injected isotonic saline by volume. No difference in excretion of the saline load was noted between 2 and 4 h; however, at 6 h, urine volume as a percentage of injected saline in the cholesterol-fed group was less than 50% of the control group. No difference was noted between total urine volumes excreted over 6 h.

Fig. 10.

Renal cortical (A) and medullary (B) mRNA expression of heme oxygenase-1 (HO-1), cyclooxygenase-2 (COX2), and nitric oxide synthase-2 (NOS2) in 6- to 8-wk control and cholesterol-fed mice. After injection of isotonic saline and urine collection, control and cholesterol-fed mice were euthanized, the renal cortex and medulla was extracted, and total RNA was generated. Steady-state abundances of HO-1 and NOS2 were reduced in the cortex (A) and medulla (B) of cholesterol-fed mice (#P < 0.05) versus control mice, whereas no difference was noted in COX2 expression.

Fig. 11.

Renal immunoblot expression of cyclooxygenase-2 (COX2) did not differ between control and cholesterol-fed mice. Total protein lysates from kidneys of mice fed 0% (n = 8) and 1% (n = 8) cholesterol diet for 6–8 wk were immunoblotted for COX2 and β-actin. Although no statistical difference in expression was noted between control and cholesterol-fed mice, steady-state renal medullary expression of COX2 tended (*P = 0.08) to be less in cholesterol-fed mice.

Fig. 12.

Renal immunoblot expression of nitric oxide synthase-2 (NOS2) was less in cholesterol-fed compared with control mice. Total protein lysates from kidneys of mice fed 0% (n = 8) and 1% (n = 8) cholesterol diet for 6–8 wk were immunoblotted for NOS2 and β-actin. NOS2 tended to be less in the renal cortex (*P = 0.08) of cholesterol-fed versus control mice, whereas NOS2 in the renal medulla of cholesterol-fed mice was significantly less (#P = 0.05) compared with control mice.

Fig. 13.

Renal immunoblot expression of heme oxygenase-1 (HO-1) did not differ between control and cholesterol-fed mice. Total protein lysates from kidneys of mice fed 0% (n = 4) and 1% (n = 4) cholesterol diet for 6–8 wk were immunoblotted for HO-1 and β-actin. Although no statistical difference in expression was noted between control and cholesterol-fed mice, steady-state renal medullary expression of HO-1 tended (*P = 0.06) to be less in cholesterol-fed mice.