Acta Diabetol DOI 10.

1007/s00592-011-0358-x

SHORT COMMUNICATION

Benecial effects of the synthetic antioxidant tert-butyl bisphenol on the hepatic microcirculation in a rat model of diabetes mellitus
Aisling C. McMahon Sarah N. Parry Victoria L. Benson Paul K. Witting David G. Le Couteur

Received: 17 October 2011 / Accepted: 29 November 2011 Springer-Verlag 2011

Abstract Diabetes mellitus is associated with oxidative injury to the vasculature. Here, the link between oxidative stress and ultrastructural changes in the hepatic microcirculation was investigated as well as the effects of a synthetic antioxidant, tert-butyl bisphenol (tBP). The study focused on the impact of experimental diabetes on liver sinusoidal endothelial cell (LSEC) fenestrations, which are pores in the liver endothelium that facilitate substrate transfer between blood and hepatocytes. Adult male rats were rendered diabetic using streptozotocin (60 mg/kg) and administered 12 IU insulin daily. After 8 weeks, animals received either 100 mg/kg tBP or vehicle alone, on 2 consecutive days. Livers were harvested 24 h later under isouorane anaesthesia (5% v/v in O2(g) by inhalation) and xed for scanning electron microscopy to evaluate fenestrations or for immuno-histochemical assessment of nitrotyrosine, a marker of nitrosative stress. Median fenestration diameter increased signicantly following 8 weeks of diabetes (80 nm vs. 70 nm controls; P \ 0.001). LSEC porosity increased by *50% (P \ 0.001). Treatment with tBP reversed these changes completely. Periportal nitrotyrosine staining was increased in diabetic livers, and this was abrogated by tBP, indicating that tBP reduced nitrosative stress in the liver. Early diabetes caused an increase in

fenestration diameter and porosity. This was reversed by acute treatment with tBP, suggesting a link between nitrosative stress and regulation of liver endothelial fenestrations, and indicates that antioxidant therapy may protect the liver microvasculature against the effects of diabetes mellitus. Keywords Diabetes mellitus Endothelial cell Liver sinusoidal endothelial cells Fenestrae Oxidative stress Antioxidant Abbreviations tBP Tert-butyl bisphenol LSEC Liver sinusoidal endothelial cell

Introduction Vascular disease is a major cause of morbidity and mortality in diabetes mellitus, with oxidative stress playing a pivotal role. However, little is known about its effects on the hepatic microvasculature, a complex three-dimensional network of small blood vessels called sinusoids. Unlike capillaries, sinusoids possess a fenestrated endothelium. Fenestrations are complete pores in the liver sinusoidal endothelial cells (LSECs) approximately 30300 nm in diameter. Bidirectional exchange of substrates, including lipoproteins, across the Space of Disse is facilitated through these fenestrations. Changes in LSEC fenestrations have implications for liver function including lipoprotein metabolism [1]. We previously studied the effects of the oxidant tertbutyl hydroperoxide on isolated LSEC morphology and showed a dose-dependent increase in periportal 3-nitrotyrosine staining and LSEC porosity, largely due to increased

A. C. McMahon (&) S. N. Parry V. L. Benson D. G. Le Couteur Centre for Education and Research on Ageing, ANZAC Medical Research Institute, University of Sydney, Concord Repatriation General Hospital, Sydney, Australia e-mail: aisling.mcmahon@sydney.edu.au P. K. Witting Discipline of Pathology, University of Sydney, Sydney, NSW, Australia

123

Acta Diabetol

fenestration size and intracellular gaps [2]. Similarly, hydrogen peroxide (H2O2) markedly increases LSEC porosity by increasing fenestration diameter [3]. Together, these data indicate that enhanced oxidative and nitrosative stress can play a role in modulating liver fenestrations. We hypothesised that enhanced oxidative and nitrosative stress in diabetic livers would be associated with changes in LSEC ultrastructure, and investigated whether such changes could be ameliorated by acute treatment with a synthetic polyphenolic antioxidant. We previously described the potential for the synthetic antioxidant t-bisphenol (3,30 ,5, 50 -tetra-t-butyl-biphenyl-4,40 -diol; tBP) as a neuro-protective agent in a cell culture model of hypoxia-reoxygenation injury [4] and demonstrated that tBP inhibits oxidative stress in vivo [5]. tBP has a low redox potential compared with other phenolic antioxidants [6], and its diphenoquinone oxidation product can be chemically reduced in successive one-electron reduction steps to regenerate the phenolic antioxidant. Recycling of tBP through chemical reduction in the diphenoquinone, combined with its low redox potential, increased bioavailability and lack of obvious toxicity, suggests that tBP may be a suitable candidate to inhibit oxidative and nitrosative stress and potentially ameliorate alterations in LSEC fenestrations in diabetes mellitus.

Methods Male Wistar rats (Animal Resources Centre, Perth, WA) aged 68 weeks were housed in pairs on a 12-h light/dark cycle and acclimatised for 1 week. Experimental procedures were approved by Sydney South West Animal Welfare Committee and adhered to the Australian Code of Practice for the care and use of animals for scientic purposes. Animals were rendered diabetic using streptozotocin (60 mg/kg i.p.; Sigma) in citrate buffer pH 4.0 or received buffer alone (controls). Blood glucose was monitored (CareSens Blood Glucose Monitoring System, iSens inc, Seoul, Korea) and maintained at 1525 mmol/l with daily insulin injections (12 IU/d). After 8 weeks, animals received either 100 mg/kg tBP in Intralipid 20 (2 ml/kg/d commercial fat emulsion, 20% soybean oil, Baxter Healthcare, Deereld, IL, USA) s.c. or Intralipid alone, on 2 consecutive days. Twenty four hours after the nal dose of tBP or vehicle, animals were anaesthetised with isouorane (5% v/v in O2 (g) by inhalation), blood glucose was immediately measured from a tail vein sample and livers were harvested. After removal of one lobe for biochemical assessment and light microscopy, livers were perfusion-xed through the portal vein (34 ml/min, PBS) followed by 10 ml xative

(1% v/v glutaraldehyde, 4% w/v paraformaldehyde, 2 mM CaCl2, 2% w/v sucrose, 0.1 M cacodylate buffer, pH 7.4). The xed liver was then removed, cut into small pieces and left in xative overnight at 4C. The tissue was rinsed in buffer, osmicated, dehydrated and incubated in hexamethyldisilazane. Tissue blocks were mounted on stubs and sputter-coated with platinum prior to scanning electron microscopy (JEOL model 6380, JEOL Ltd, Tokyo, Japan). Ten images from different sinusoids were recorded from 3 to 4 tissue blocks from each liver (mag 920,000). Fenestration diameters were measured using ImageJ (http://rsb.info.nih.gov/ij). For controls, 4129 fenestrations were counted from 50 micrographs, n = 5 livers; diabetics 4125, 31, n = 3; controls treated with tBP 4952, 55, n = 5; and diabetics treated with tBP 2518, 34, n = 3, respectively. Data were analysed using Kruskal Wallis one-way analysis of variance on ranks for nonnormally distributed data with post hoc analysis using Dunns method for multiple comparisons. Samples designated for immuno-histochemistry were xed overnight in 4% w/v paraformaldehyde, then stored in 70% EtOH until processed and parafn-embedded (Leica TP1020, Heidelberg, Germany). Semi-thin (5 lm) sections were stained with haematoxylin and eosin. For detection of nitrotyrosine, sections were treated with 10 mM sodium citrate buffer for antigen retrieval, then incubated with H2O2 to quench endogenous peroxidase activity and blocked with 10% v/v goat serum in PBS. Finally, sections were incubated with rabbit anti-nitrotyrosine, followed by biotinylated goat anti-rabbit antibodies (Sigma, 1 h each) in a humidied chamber. An avidinbiotin complex kit (Extravidin Peroxidase; Sigma) was used in conjugation with diaminobenzidine (DAB, DAKO Cytomation, CA, USA) to visualise positive staining. Sections were counterstained with Harriss haematoxylin and mounted with DPX. Imaging was performed with StereoInvestigator software (MBF Inc, Williston, VT, USA). Where required, samples of unxed liver tissue were homogenised and the lipophilic components extracted into hexane and methanol (5:1 v/v) as described previously [4, 5]. The extract was then dried under vacuum and the residue resuspended in isopropyl alcohol for analysis with reversed-phase liquid chromatography as described in detail elsewhere [4, 5].

Results The presence of tBP in the liver was conrmed by analysis with liquid chromatography (nal concentration 23.0 4.2 pmol/mg protein, n = 2 livers, each sampled in duplicate). Blood glucose levels in controls were not inuenced by tBP (8.9 1.5 vs. 8.6 1.1 mmol/l, respectively).

123

Acta Diabetol Fig. 1 Representative scanning electron microscopy of liver sinusoidal endothelium. a control rats; b diabetic rats; c control rats treated with tBP; d diabetic rats treated with tBP. Diabetes was associated with increased fenestration diameter and porosity, which was reversed by acute treatment with tBP. Changes are representative of n = 10 independent assessments using n = 35 livers per group. Arrows indicate normal fenestrations. e Median and range of sinusoid porosity presented for each group. Porosity was calculated as the % area occupied by fenestrations. Diabetes was associated with a statistically signicant increase in porosity which was reversed by acute treatment with tBP

Control

Diabetic

Control +tBP

Diabetic +tBP

In diabetic animals, blood glucose was elevated (25.0 5.3 mmol/l) but not inuenced by tBP (24.8 1.4 mmol/l). Scanning electron microscopy

animals (80 nm, P \ 0.001; Fig. 2a). Treatment with tBP restored median fenestration diameter in diabetic animals to 70 nm (Fig. 2b). Immuno-histochemistry

There was a signicant increase in porosity between controls (median 6.03%, range 3.419.80%) and diabetics (median 9.21% range 7.2411.79%; P \ 0.001; Fig. 1ad), which was normalised by treatment with tBP (median 6.19%, range 3.898.01%), Fig. 1e. Median fenestration diameter was 70 nm in controls but was greater in diabetic

Nitrotyrosine staining was not evident in control livers. In contrast, periportal staining was markedly enhanced in diabetic livers, consistent with enhanced nitrosative stress being associated with diabetes mellitus (Fig. 2c, d). Diabetic rats treated with tBP had reduced peroxynitrite

123

Acta Diabetol

Discussion It is established that diabetes mellitus is associated with enhanced oxidative and nitrosative stress and that these factors may promote endothelial dysfunction, a key process that occurs early in the development of vascular disease. Mitochondrial overproduction of reactive oxygen species is considered a key mechanism that initiates a cascade of events leading to endothelial injury. Among these pathways is the production of excess nitric oxide which reacts with superoxide radical anion to generate peroxynitrate, a potent oxidant that yields nitrotyrosine adducts in proteins [7, 8]. Accordingly, nitrotyrosine detection is widely used as a surrogate marker for oxidative/nitrosative injury. Here, we found extensive periportal liver staining for nitrotyrosine in diabetes mellitus. A likely explanation is that this zone has the highest oxygen tension, leading to greater production of superoxide radical anion than the hypoxic downstream central region. We previously showed that oxidative stress has marked effects on the hepatic microvasculature. Tert-butyl-hydroperoxide, whether injected directly into the portal vein in vivo or administered to isolated LSECs, caused increased porosity and enhanced gap formation. Porosity increased fourfold at the highest doses in association with increased periportal nitrotyrosine staining [2], while in another study, treatment with H2O2 increased porosity by two-thirds [3]. Fenestration diameter increased with H2O2 treatment, while tert-butyl-hydroperoxide increased the number of gaps without changing fenestration diameter. In this study, we have shown that streptozotocin-induced diabetes yielded increased periportal nitrotyrosine staining associated with increased porosity of *50% and increased fenestration diameter of *10%. This supports the conclusion that endothelial changes in the hepatic microcirculation in diabetes mellitus are linked to oxidative/nitrosative stress. Few reports describe the effects of diabetes mellitus on the hepatic microcirculation. In a preliminary study in rats 5 weeks after streptozotocin-induced diabetes mellitus, we reported increased fenestration diameter (from 69 to 75 nm) and porosity (from 5.6 to 8.2%) [9]. We also reported that long-term diabetes mellitus in baboons is associated with decreased porosity and fenestration diameter, as well as other markers of capillarisation of the LSEC [10]. It is possible that short-term diabetes mellitus induces injury to LSECs characterised by increases in fenestrations and gaps, while long-term disease causes scarring and brosis of the liver endothelium with loss of fenestrations. Although ageing is associated with reduced porosity and fenestration diameter in rats as well as other species, there are no reports of the long-term effects of diabetes mellitus in rats. Diabetes mellitus-associated fenestration abnormalities have important clinical implications because

pp

cl

Fig. 2 a and b The effect of diabetes mellitus and tBP on frequency distribution of fenestration diameter. Diabetes mellitus was associated with increased diameter compared with controls, and this was reversed by tBP. Nitrotyrosine immuno-histochemistry of the liver. c control rats; d diabetic rats; e control rats treated with tBP; and f diabetic rats with treated with tBP. Diabetes mellitus was associated with marked staining of the periportal region (pp) compared with minimal staining in the centrilobular region (cl). Acute administration tBP reduced periportal staining in diabetic animals. Magnication 910

staining compared with untreated diabetic rats (Fig. 2f). Slight positive staining was observed in some but not all control animals following tBP treatment (Fig. 2e).

123

Acta Diabetol

fenestrations are vital for the passage of lipoproteins including chylomicron remnants from the sinusoidal blood into the extracellular space of Disse. Thus, any increase in endothelial porosity in early diabetes mellitus might contribute to fatty liver, while the reduced endothelial porosity in chronic diabetes mellitus might contribute to dyslipidemia, both common features in diabetic patients [1, 9]. Importantly, the effects of diabetes mellitus on LSECs were abrogated by acute treatment with tBP. The data suggest that this restoration is linked to its antioxidant properties, since a diminution in immuno-histochemical markers of oxidative damage occurred concomitantly with the ultrastructural restoration. Previous cell culture studies have shown that tBP has promise as a neuro-protective agent [4] and that it also prevented excessive oxidative damage in kidney epithelial cells in an experimental model of rhabdomyolysis [6]. Previously, long-term dietary supplementation with tBP was shown to reduce atherosclerotic lesions in apoE-/-/LDLR-/- mice [5]. Longer-term studies of tBP are now required to determine whether the benecial effects of tBP on the liver endothelium may be translated into clinical benets related to fatty liver, dyslipidemia and diabetes mellitus. In conclusion, streptozotocin-induced diabetes was associated with increased fenestration diameter and porosity, which was reversed by tBP. This suggests a link between oxidative and nitrosative stress and control of LSEC fenestration and indicates that antioxidant therapy may provide protection for the liver microvasculature against the effects of diabetes mellitus.
Acknowledgments The study was funded by the Diabetes Australia Research Trust, the Ageing and Alzheimers Research Foundation and NHMRC project grants. The authors thank Tharani Sabaretnam for technical assistance with HPLC.

References
1. Le Couteur DG, Fraser R, Cogger VC, McLean AJ (2002) Hepatic pseudocapillarisation and atherosclerosis in ageing. Lancet 359:16121615 2. Cogger VC, Muller M, Fraser R, McLean AJ, Khan J, Le Couteur DG (2004) The effects of oxidative stress on the liver sieve. J Hepatol 41:370376 3. Cogger VC, Mross PE, Hosie MJ, Ansselin AD, McLean AJ, Le Couteur DG (2001) The effect of acute oxidative stress on the ultrastructure of the perfused rat liver. Pharmacol Toxicol 89:306311 4. Duong TT, Antao S, Ellis NA, Myers SJ, Witting PK (2008) Supplementation with a synthetic polyphenol limits oxidative stress and enhances neuronal cell viability in response to hypoxia-re-oxygenation injury. Brain Res 1219:818 5. Witting PK, Pettersson K, Ostlund-Lindqvist AM, Westerlund C, Eriksson AW, Stocker R (1999) Inhibition by a coantioxidant of aortic lipoprotein lipid peroxidation and atherosclerosis in apolipoprotein E and low density lipoprotein receptor gene double knockout mice. Faseb J 13:667675 6. Shanu A, Parry SN, Wood S, Rodas E, Witting PK (2010) The synthetic polyphenol tert-butyl-bisphenol inhibits myoglobininduced dysfunction in cultured kidney epithelial cells. Free Radic Res 44:843853 7. El-Remessy AB, Tawk HE, Matragoon S, Pillai B, Caldwell RB, Caldwell RW (2010) Peroxynitrite mediates diabetesinduced endothelial dysfunction: possible role of Rho kinase activation. Exp Diabetes Res 2010:247861 8. Ren XY, Li YN, Qi JS, Niu T (2008) Peroxynitrite-induced protein nitration contributes to liver mitochondrial damage in diabetic rats. J Diabetes Complications 22:357364 9. Jamieson H, Dobbs BR, Day WA, Rogers GWT, Fraser R (2001) The liver sieve in diabetes: are the ultrastructural changes similar to those seen in alcoholism? In: Wisse E, Knook DL, De Zanger R, Arthur MJP (eds) Cells of the hepatic sinusoid. Kupffer Cell Foundation, Leiden, pp 123124 10. Jamieson HA, Cogger VC, Twigg SM et al (2007) Alterations in liver sinusoidal endothelium in a baboon model of type 1 diabetes. Diabetologia 50:19691976

123