Decreased ZO1 expression causes loss of time-dependent tight junction function in the liver of ob/ob mice

Diabetes patients are at a high risk of developing complications related to angiopathy and disruption of the signal transduction system. The liver is one of the multiple organs damaged during diabetes. Few studies have evaluated the morphological effects of adhesion factors in diabetic liver. The influence of diurnal variation has been observed in the expression and functioning of adhesion molecules to maintain tissue homeostasis associated with nutrient uptake. The present study demonstrated that the rhythm-influenced functioning of tight junction was impaired in the liver of ob/ob mice. The tight junctions of hepatocytes were loosened during the dark period in control mice compared to those in ob/ob mice, where the hepatocyte gaps remained open throughout the day. The time-dependent expression of zonula occludens 1 (ZO1, encoded by Tjp1 gene) in the liver plays a vital role in the functioning of the tight junction. The time-dependent expression of ZO1 was nullified and its expression was attenuated in the liver of ob/ob mice. ZO1 expression was inhibited at the mRNA and protein levels. The expression rhythm of ZO1 was found to be regulated by heat shock factor (HSF)1/2, the expression of which was reduced in the liver of ob/ob mice. The DNA-binding ability of HSF1/2 was decreased in the liver of ob/ob mice compared to that in control mice. These findings suggest the involvement of impaired expression and functioning of adhesion factors in diabetic liver complications.


Introduction
Obesity associated with binge eating leads to the development of metabolic diseases, such as diabetes and dyslipidemia. Diabetes and obesity are strongly linked to the onset and progression of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) [1]. Although NAFLD and NASH progress to fatty liver, there exists an underlying relationship between the alleviation of diabetes and NAFLD that could be exploited to establish liver-focused treatments [2]. Since diabetes patients are at a high risk of developing hepatitis and liver cancer [3], it is necessary to include a treatment protocol for NAFLD alongside with diabetes in these patients. Though therapies that focus on normalizing the abnormal metabolic system are in practice, satisfactory therapeutic outcomes have not been obtained. At present, research to analyze the morphological development of the liver is underway; however, several aspects remain to be elucidated.
Regardless of obesity, diabetes patients show high incidence of three major complications, namely, retinopathy, kidney disease, and peripheral neuropathic pain. Angiopathy is the major cause of these complications resulting from a disrupted capillary system [4], as seen from the analysis of tight junctions that act as barrier mechanisms. The expression of tight junction-forming factors was signi cantly reduced in the retina of diabetic mice [5]. Similarly, decreased expression of zonula occludens-1 (ZO1) in the small intestine of diabetic mice is associated with in ammation and endotoxin invasion [6]. The binding disorders of hepatic tight junction are also one of the risk factors for causing liver disease.
Functionally, tight junctions in the liver prevent the leakage of bile from the bile duct and facilitate the uptake of nutrients [7]. However, the relationship between the alteration of liver-associated tight junctions and the associated pathological conditions remains unclear. In the living system, factors linked to the defense and maintenance of homeostasis are controlled by the biological clock system. In type 2 diabetes, the expression amplitude of the clock gene, which is the core of the biological clock system, is attenuated [8]. Conversely, the onset of type 2 diabetes may cause an irregular rhythm in the membrane protein activity [9]. Therefore, elucidating the pathophysiology of diabetes warrants biochemical analysis based on the biological clock system. Since alteration in the expression of tight junction-related factors was dependent on the time difference [10], it was assumed that disrupted physiological time-dependency in diabetes would result in the abnormal expression of liver tight junction-related factors, leading to the pathological manifestation of hepatitis. This study aimed to evaluate and explore the mechanism associated with the transformation of altered liver tight junction function in diabetes using ob/ob mice.

Animal experiments
Six-week-old male C57BL/6J Ham Slc-ob/ob mice and age/sex-matched C57BL/6J Ham Slc-+/+ mice (control mice) were purchased from the Japan SLC Inc. (Shizuoka, Japan). Mice were housed in a lightcontrolled room at a temperature of 24 ± 1 °C and 60 ± 10 % humidity, with food and water available ad libitum. In the light/dark cycle, zeitgeber time (ZT) with ZT0 and ZT12 were de ned as time of lights on and off, respectively. During the dark period, a dim red light was used to aid in animal treatment. Vascular permeability was assessed by intravenous (i.v) administration of Evans blue dye (30 mg/kg) at ZT0 and ZT12, according to previous reports [11]. After 1 h post-injection, the mice were euthanized and thoroughly perfused with 0.9 % saline to clear the circulation of any residual dye. The perfused liver was collected after xation with 4 % paraformaldehyde. Frozen liver sections were processed for uorescent microscopic evaluation and photographed with a BZ-9000 microscope. For the assessment of Evans blue in ltration, 100 mg of the liver was soaked in 500 µL of 4 % paraformaldehyde, incubated at 55 °C for a day (to extract Evans blue), and centrifuged; the absorbance of the supernatant was measured at 610 nm using an In nite® 200 PRO plate reader (Tecan Group Ltd., Männedorf, CH).
RNA isolation and quantitative real-time polymerase chain reaction (qRT-PCR) Carlsbad, CA, USA). All experiments that used kits were performed as per the manufacturer's instructions. The data were normalized to 18S ribosomal RNA gene (Rn18s) used as the internal control. Nucleotide sequences of speci c primers of genes used in the study are listed in Table 1. Hepatic membrane proteins were extracted from the mice liver using the FractionPREP™ cell fractionation kit (K270; Biovision, Mountain View, CA, USA) as per the manufacturer's instructions. Denatured samples containing 20 µg of each protein fraction were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto a polyvinylidene di uoride membrane. Separated proteins were stained with Coomassie Brilliant Blue (CBB) as a control for equal loading of the membrane fraction. The membranes were incubated with primary antibodies: anti-ZO1 (1:1000; ab15602; Abcam, Cambridge, UK). Speci c antigen-antibody complexes were visualized using HRP-conjugated anti-rabbit IgG (1:10000; sc-2032; Santa Cruz Biotechnology, Santa Cruz, CA) and ImmunoStar LD (Wako Chemicals, Osaka, Japan). Visualized images were scanned by a BIO-RAD ChemiDoc™ Touch Imaging System (Bio-Rad, Hercules, CA, USA).

Transcriptional activity assay
Mouse heat shock factor 1 (HSF1) and HSF2 expression vectors were purchased from OriGene (HSF1: MR208087, HSF2: MR208286, Guangzhou, China). Mouse tight junction protein 1 (mTjp1)-reporter vectors containing the mTjp1 promoter region spanning from -650 to +50, -248 to +50, and -50 to +50 (relative to the transcription start site, +1) were constructed using the VectorBuilder services (VectorBuilder Inc, Chicago, IL, USA). Hepa1-6 cells were seeded at a density of 2 × 10 4 cells/well in 24well culture plates. Cells were transfected with 50 ng of the green uorescent protein (GFP) reporter construct and 500 ng (total) of the respective expression vector using Lipofectamine ® 3000 Reagent (Thermo Fisher Scienti c). The pCMV-6 empty vector was added to obtain a constant nal DNA concentration in all transfections. After 24 h post-transfection, the uorescence intensity was analyzed using an In nite ® 200 PRO plate reader (Tecan Group Ltd.). The ratio of GFP intensity to protein concentration in each sample served as a measure of normalization.

Chromatin immunoprecipitation (ChIP) analysis
ChIP assay was performed with reference to past papers [12]. Cross-linked chromatin from liver was sonicated on ice and nuclear fractions were obtained by centrifugation at 10,000 × g for 5 min.
Supernatants were incubated with the following antibodies: anti-HSF1 (1:200; ab16502; Abcam), anti-HSF2 (1:200; ab32360; Abcam), or rabbit anti-IgG (1:200; sc66931; Santa Cruz Biotechnology). DNA was puri ed using the DNA puri cation kit (Promega, Madison, WI, USA) as per the manufacturer's instructions and ampli ed by PCR for the surrounding HSF response element (HSE) in the 5′ -anking region of the mouse Tjp1 gene. Primer sequences used for ampli cation were as follows: forward, 5′-AATGGTATGGCATAGGAGTGG -3′; and reverse, 5′-TTACGCTTGACAAAGAGGAAG-3′. TB green premix Ex Taq™ (Takara Bio) with the StepOnePlus™ Real-Time PCR System (Life Technologies) was used to quantify the products. All data were normalized to the PCR products of input DNA. The quantitative reliability of PCR was evaluated by kinetic analysis of the ampli ed products to ensure that signals were derived exclusively from the exponential phase of ampli cation. ChIP was performed either in the absence of antibodies or in the presence rabbit IgG as a negative control.

Statistical analyses
All data are expressed as mean ± standard error of the mean (SEM). Statistical analyses were performed using the GraphPad Prism software (ver. 8; GraphPad Software, San Diego, CA, USA). Differences among the groups were analyzed by two-way ANOVA, followed by Tukey's post-hoc and Sidak's post-hoc tests. P < 0.05 was considered statistically signi cant. Although no statistical methods were used to predetermine the sample size, the sample sizes used in the present study are similar to those reported in previous studies [6,9,13]. The experiments were not randomized.

Attenuation of time dependency of tight junctions in the mouse diabetic liver
To examine possible alterations in the diurnal variation of tight junctions in the diabetic liver, vascular leakage was analyzed using Evans blue in wild-type and ob/ob mice. Fluorescent microscopic analysis revealed dark period-dependent permeation of Evans blue in the hepatic parenchymal cells of wild-type and time independent increase of Evans blue permeation in the liver of ob/ob mice (Fig. 1a, b). These results indicate attenuated time dependency and tight junction functions in the diabetic liver, in vivo.
In uence of diabetes on the expression of tight junction-related genes in the mouse liver Tight junctions are adhesion sites on the cell membrane that are densely packed with multiple interacting proteins. Although the attenuation of the expression of tight junction factor, ZO1, has been deciphered in ob/ob mice, the time-dependent changes remain to be elucidated. To identify the attenuated genes associated with diurnal variation in diabetes, the mRNA expression levels of signi cant tight junctionrelated genes were evaluated. Accordingly, the expression levels of Tjp1, Tjp2, occludin (Ocln), claudin 1 (Cldn1), Cldn3, and Cldn5 were measured. The expression of Tjp1 mRNA was signi cantly decreased in the light periods in ob/ob mice (Fig. 2a-f). Since the time dependency of ZO1 expression was attenuated in ob/ob mice, we speculated that ZO1 might play a central role in the time dependency of tight junctions in diabetes.
The expression of Tjp1 mRNA exhibited signi cant diurnal time-dependency in the liver of wild-type mice maintained under light/dark cycle conditions (Fig. 2a). The plasmalemmal expression of ZO1 indicated signi cant time dependence similar to that of the mRNA expression in the liver of wild-type mice (Fig. 3a,  b). The expression of Tjp1 mRNA and ZO1 protein was continually decreased throughout the day in the liver of ob/ob mice compared to that of the wild-type mice (Fig. 2a, 3b). These results indicate that decreased of ZO1 expression causes a decrease in hepatic tight junction function in ob/ob mice.

Transcriptional regulation of Tjp1 by HSFs
Sequence analysis of the promoter region of Tjp1 genes revealed a highly conserved HSE located between -468 and -448 base pairs (bp) upstream from the transcription start site (relative to the transcription start site, +1) (Fig. 4a). Further, the repression of Tjp1 transcription by HSFs was explored. The activity of Tjp1 reporters was diminished by the elimination of their promoter sequences (Fig. 4b).
To con rm whether the transcriptional ability of HSFs affects the time-dependent expression of ZO1, the nuclear expression level of HSF1/2 was measured in the liver of ob/ob mice. The expression of Hsf1/2 mRNA indicated signi cant time dependence in the liver of wild-type mice (Fig. 4c). In contrast, the expression of H Hsf1/2 mRNA was continually decreased throughout the day in the liver of ob/ob mice (Fig. 4c). ChIP analysis demonstrated signi cantly enhanced binding of HSF1/2 to the promoter region of the gene containing the HSEs at dark phase compared to that observed at light phase in the liver of wildtype mice (Fig. 4d). In contrast, the amount of HSF1/2 binding to the Tjp1 promoter in the liver of ob/ob mice was decreased in both light and dark phases (Fig. 4d). These results indicate that the decrease in HSF1/2 expression during diabetes mellitus eliminates the time-dependent expression uctuation of ZO1 and disrupts the tight junction mechanism of the liver.

Discussion
Morphological wear or physiological effects, including angiopathy and modulation of insulin signals, are associated with the complications linked with tissue damage in diabetes [4,15]. In this study, we found that the formation of tight junctions was diminished because of the decreased of ZO1 expression in the liver of diabetic mice. The results indicate a loss of function of the defense mechanism in healthy hepatocytes due to attenuated diurnal variability in diabetic liver.
Vascular permeability associated with angiopathy is evaluated using Evans blue because of its high binding a nity to albumin and retention in the blood. The perisinusoidal space or space of Disse in the liver is a region around the sinusoid where plasma collects, and hence with the highest retention of Evans blue compared to all other tissues [16]. The space of Disse is liver-speci c and serves as a site that expresses several transporters and facilitates the uptake of nutrients, such as fatty and amino acids [17].
The expansion of the space of Disse causes damage to the hepatic morphology and leads to the development of liver damage [18]. Since vascular permeability was enhanced, both in the resting and active period in diabetic mice, it was considered as an indication for the occurrence of possible liver damage.
Tight junctions in the liver are essential for maintaining the physiological function and hence, abnormal functioning of the same is associated with liver diseases [19]. Few studies have de ned tight junctionrelated factors that are altered in liver diseases. The expression of ZO1, a protein that forms tight junctions, is downregulated in liver diseases, such as NASH and hepatic cancer [20,21]. The expression of ZO1 was decreased in ob/ob mice, suggesting the possibility of NASH-related liver damage in diabetes, as proven clinically.
Reduction of HSF1 and HSF2 expression leads to liver injury. In HSF1 knockout mice, increased cytokine production and decreased clearance of reactive oxygen species, exacerbates the progression of hepatitis [25]. HSF2 positively regulates Psmb5 expression constituting 20S core proteasome complex; which suggests that Decreased proteasome activity is associated with liver fat accumulation and liver disease [26,27]. Attenuated expression of HSFs increased the signals involved in promoting hepatitis in ob/ob mice. HSF1 drives the cascade of heat shock proteins which affects the expression of genes with downstream physiological functions. HSF1 binds to clock genes and controls the circadian clock mechanism in living organisms [22]. HSF1 functions as resetting the biological clock mechanism affected by changes in temperature and UV stimulation [23]. The amplitude of the clock gene expression reduces in ob/ob mice [24], but its association with HSF1 remains unclear. The expression of the clock gene may be related to the decrease in HSF1 expression (Fig. 4c-d), which causes liver injury in ob/ob mice.
Complications associated with systemic organs have been reported in people with diabetes. Several studies have focused on the signaling mechanism associated with the suggested risk of development of hepatitis. However, morphological evaluation and differential expression pro les of the liver during the course of diabetes have not been elucidated. The present study characterizes the morphological changes induced by diabetes in the liver and elucidates a novel mechanism involving cell-binding disorders in diabetes.   signi cantly different between the two groups; ##, P < 0.01; signi cantly different from wild-type mice at the corresponding time point (two-way ANOVA with Tukey post hoc test).