Conditional knockout of SND1 in hepatocytes
To investigate the effects of hepatocyte-specific deletion of SND1 on liver insulin resistance and acute liver failure, we first constructed SND1 liver conditional knockout (LKO) mice. As shown in Fig. 1a, we first built the SND1 Flox/Flox mice with the loxP allele flanked at the exon 3 of SND1 gene and purchased the wt/wt-albumin-Cre+ mice, which includes the wild type allele and the specific gene sequence for the expression of albumin-induced Cre enzyme. By crossing the two mice, we obtained the SND1 Flox/wt-albumin-Cre+ heterozygous mice with a probability of 1/2. Then, we performed a second round of mating, using the two SND1 Flox/wt-albumin-Cre+ heterozygous mice, to generate the SND1 Flox/Flox-albumin-Cre+ homozygous mice (SND1 LKO mice) with a probability of 3/16. SND1 Flox/Flox littermates were used as wild type controls.
To verify the successful construction of SND1 LKO mice, DNA extracted from mouse primary hepatocytes was used for detecting the existence of SND1 LoxP sites. As shown in Fig. 1b, we performed the genotyping assay using the mixture of three primers (F1 within intron 2 of SND1, F2 within exon 3, and R within intron 3). For the wild-type control mice without LoxP site, SND1 WT band of 376 bp was detected by the PCR assay using the F2 plus R primers (Fig. 1b, lane 1). Because the SND1 Flox/Flox mice contain the LoxP site, the SND1 Flox band of 504 bp was produced by the F2 + R primers (Fig. 1b, lane 2). The existence of albumin-Cre in the SND1 Flox/Flox-albumin-Cre+ homozygous mice (#1 and #2) led to the resection of the sequence between the two LoxP sites and the observation of SND1 LKO band of 279 bp through F1 + R primers (Fig. 1b, lane 3 and 4).Internal control gene was detected in all the above mice (Fig. 1b-c).
Next, we extracted the tissues of liver, spleen, pancreas, and kidney from the SND1 wild type (WT) (#1, #2, #3) and LKO (#4, #5, #6) mice, respectively (Fig. 1d). SMMC-7721 cell lines of SND1 WT and knockout (KO) were used as the sample controls. The lysates were subjected to SDS-PAGE and then immuno-blotted using anti-SND1 antibody, or anti-β-actin antibody. The data in Fig. 1c showed that SND1 protein was completely deleted in the SMMC-7721 cell lines of SND1 KO, but not SND1 WT. However, we only detected the decreased expression of SND1 in the liver tissue of SND1 LKO (#4, #5, #6) mice, when compared with the SND1 WT (#1, #2, #3) mice. There is no difference of SND1 expression between SND1 WT and SND1 LKO mice in the spleen, pancreas, and kidney tissues (Fig. 1d). Considering the albumin protein was specifically expressed in hepatocytes, we further extracted primary hepatocytes from the liver tissue of the mice, and found that SND1 protein was knocked out in the primary hepatocytes of SND1 LKO mice (Fig. 1e). Figure 1f showed the gross morphology of SND1 WT and SND1 LKO mice. These suggested that the mice with the hepatocyte-specific deletion of SND1 were successfully constructed.
Gross morphology and weight analysis in the absence of hepatic SND1
A total of five SND1 WT and five SND1 LKO mice were randomly selected and fed with chow diet (CD). We monitored the bodyweights of WT or LKO mice weekly and did not find a significant difference (Fig. 2a). After 24 w of CD, we measured the weight of liver tissue, and calculate the ratio value of liver weight/ body weight. There is no statistical difference between the SND1 WT and SND1 LKO mice (Fig. 2b). We further measured the weight of white adipose tissue (WAT), and calculate the ratio value of WAT weight/ body weight. Similar negative results were obtained in Fig. 2c.
Next, we further investigated the effect of SND1 liver conditional knockout on the weight of mice with a high fat diet (HFD). We did not observe the significant difference of body weight and liver weight between SND1 WT HFD and SND1 LKO HFD mice (Fig. 3a-d). Nevertheless, the absence of hepatic SND1 can lead to a decreased WAT weight (Fig. 3e, * P < 0.05) or the ratio value of WAT weight/body weight (Fig. 3f, ** P < 0.01). These suggested that SND1 liver conditional knockout can affect the weight of white adipose tissue in mice under the condition of a high-fat diet, but not the gross morphology, body weight, and liver weight of mice.
The glucose homeostasis analysis in the absence of hepatic SND1
We analyzed the glucose homeostasis of SND1 WT and SND1 KO mice with a chow diet. A fasting/refeeding assay was performed. After the fasting treatment of mice with 4w CD for 16 h, we restored the chow diet and measured the blood glucose levels at 0 h, 0.5 h, 1 h, 2 h, 4 h, and 6 h, respectively. As shown in Fig. 4a, we did not observe the statistical difference in blood glucose change between SND1 WT CD and LKO CD mice.
Next, we performed a glucose tolerance test (GTT). After the fasting treatment for 16 h, we injected intraperitoneally 1.5 g/kg glucose solution in the SDN1 WT or LKO mice with 4 w chow diet, and measured the blood glucose levels at 0 min, 15 min, 30 min, 60 min, 90 min, and 120 min, respectively. We observed similar negative results (Fig. 4b). We further performed the insulin tolerance test (ITT) at the 16 w of chow in mice. The 0.75 U/kg insulin was injected intraperitoneally in the mice, and the blood glucose levels at 0 min, 15 min, 30 min, 45 min, 60 min, and 90 min were measured. As shown in Fig. 4c, compared with the SND1 WT CD, no increased AUC value was observed in the SND1 LKO CD. In addition, we performed an acute insulin response assay. At 24 w of CD in mice, SND1 WT and LKO mice were fasted overnight, and anesthetized. After the injection of the insulin solution, the phosphorylation level of Akt protein in the liver tissue was analyzed by western blotting assay. The data of Fig. 4d indicated an increased phosphorylation level of Akt at the time point of 5 min, compared with 0 min, in the SND1 WT mice (#1, #2) and SND1 LKO mice (#3, #4). Nevertheless, no statistical difference of p-Akt/ total Akt ratio was detected between SND1 WT and SND1 LKO mice at the time point of 0 min or 5 min (Fig. 4d).
We also performed the fasting/refeeding assay (Fig. 5a), glucose tolerance test (Fig. 5b-d), insulin tolerance test (Fig. 5e) and acute insulin response assay (Fig. 5f) in the SND1 WT and SND1 LKO mice with a high-fat diet, respectively. In the acute insulin response assay, the treatment of the insulin solution did not increase the phosphorylation level of Akt protein (Fig. 5f), suggesting that high-fat diet results in the occurrence of insulin resistance in SND1 WT and SND1 LKO mice. We did not observe the positive results (Fig. 5). Taken together, SND1 liver conditional knockout fails to influence the glucose homeostasis of mice under the condition of a chow diet or high-fat diet.
Cholesterol level and hepatic steatosis analysis in the absence of hepatic SND1
Very recently, we found that SND1 can regulate cholesterol metabolism in mice through promoting the activity of sterol-regulatory element-binding protein 2 (SREBP2) protein during the induction of a high fat diet [14]. In hepatocellular carcinoma (HCC), the up-regulation of SND1 expression also can influence the cellular cholesterol distribution and homeostasis [23]. Thus, we attempted to measure the level of cholesterol in SND1 WT and LKO mice. As shown in Fig. 6a-c, an increased level of serum total cholesterol, liver free cholesterol, but not liver total cholesterol, was detected in the HFD mice, compared with CD mice. However, we did not observe the statistical difference of the cholesterol level between SND1 WT and SND1 LKO mice with chow or high-fat diets. Moreover, the Hematoxylin-Eosin (H-E) staining of liver sections (Fig. 6d) indicated a significant increase in fat vacuoles in the HFD mice, compared with CD mice, but not a remarkable difference between SND1 WT and SND1 LKO mice. These suggested that hepatocyte-specific deletion of SND1 failed to influence the cholesterol level and hepatic steatosis of mice.
Hepatic failure analysis in the absence of hepatic SND1
First, we analyzed the effect of SND1 hepatocyte-specific deletion in the serum levels of ALT or AST. As shown in Fig. 7a, there is an increased level of ALT in both WT and LKO mice (* P < 0.05) after 6 h of LPS/D-GalN stimulation, compared with the normal saline (NS) control group. Nevertheless, we did not detect the significant difference between WT and LKO mice (Fig. 7a). We also observed that LPS/D-GalN stimulation led to an increase AST level in the WT mice (* P < 0.05), but not LKO mice (Fig. 7b). Furthermore, we performed the quantitative RT-PCR assay to detect the expression level of inflammatory cytokines, including IL-6, IL-1β, and TNF-α. As shown in Fig. 7c, the LPS/D-GalN treatment can induce an enhanced level of IL-6 expression in both WT (** P < 0.01) and LKO (*P < 0.05) mice. However, no difference exists in the presence and absence of hepatic SND1 (Fig. 7c). The similar results were observed for the detection of IL-1β (Fig. 7d, ***P < 0.001 for WT, ** P < 0.01 for LKO) and TNF-α (Fig. 7e, ** P < 0.01 for WT; * P < 0.05 for LKO).
In addition, we performed a Hematoxylin-Eosin (H-E) staining assay to detect the influence of LPS/D-GalN in the histomorphology of liver tissue and investigated whether the deletion of SND1 can affect the role of LPS/D-GalN in the liver tissue damage. As shown in Fig. 7f, LPS/D-GalN treatment results in a significant liver damages (e.g., liver tissue hemorrhage, inflammatory cell infiltration) in LKO mice, compared with the WT mice. Nevertheless, we did not observe the remarkable difference regarding the liver damage extent between WT and LKO mice. These suggested that hepatocyte-specific deletion of SND1 failed to influence the hepatic failure process of mice remarkably.