**4. Discussion**

The present study confirmed that the STZ-evoked diabetic state aggravates neuronal damage caused by a transient cerebral ischemia and subsequent reperfusion in rats. We found that diabetes enhances the production of O2 − , activates caspase-3, and induces the expression of proinflammatory cytokines (TNF-α and IL-1β) in the brain. These detrimental effects are markedly potentiated by cerebral ischemia and reperfusion, leading to greater infarct growth and aggravation of apoptosis and inflammation. Our results show that chronic supplementation with AA inhibits the apoptotic changes and proinflammatory responses, and attenuates the exacerbation of cerebral injury and neurological deficits in the diabetic state. These beneficial effects of AA could be attributed to its antioxidant and anti-inflammatory properties. We observed that the expression level of SVCT2 increases in response to MCAO/Re in the nondiabetic cortex, which is accompanied by an increase in the total AA (AA + DHA) in the tissues, and that these responses are abolished in the diabetic rats. AA supplementation to the diabetic rats restored these responses to the levels of the nondiabetic rats. Therefore, supplementation with AA may enhance the transport of AA into

neuronal cells, resulting in reinforcement of the antioxidant defense and alleviation of oxidative ischemic injury in the brain of diabetic rats.

Because rats can synthesize their own AA, unlike humans, they seem to hardly suffer from low levels of internal AA [19]. However, there is experimental evidence that STZ-induced type 1 diabetic rats have low levels of AA in the plasma, liver, kidney, and other tissues [23,34,35]. It is conceivable that the enhanced oxidative stress in diabetes consumes the circulating AA and that the resulting diminution in the AA concentration leads to further enhancement of the reactions mediated by free radicals. This study indicates that the diabetic rats have low levels of AA in the plasma, and that chronic oral pretreatment with AA in diabetic rats decreases the cerebral O2 − generation, apoptosis, and infarction induced by MCAO/Re; all of these effects could be a result of improved antioxidant status in the diabetic brain. Indeed, we observed that AA supplementation to the diabetic rats prevents the diminution of the total AA levels in the cortex associated with ischemia-reperfusion. On the other hand, we found that AA supplementation causes insufficient downregulation of proinflammatory cytokines and has no effect on MPO expression in diabetic rats after MCAO/Re. The relative low dose of AA (100 mg/kg) which we used in this study might not be efficacious against the severe inflammatory responses induced by the combination of stroke and diabetes.

Concentrations of AA in plasma and peripheral extracellular fluid after ingestion are strictly controlled by intestinal absorption, tissue transport, and renal excretion. Chen *et al*. [36] shows that orally administrated AA at doses >200 mg/kg is declined to ≈60 μM in plasma and extracellular fluid of femoral muscle in rats. In the brain, concentration of AA is relatively high compared with that in the plasma [24]. Two mechanisms of AA recruitment into the brain are known: the transport of AA through SVCT2, a specific transporter for AA, and the transport of DHA across the BBB through GLUT1 with subsequent immediate reduction in the brain. The gradation of AA concentrations between the plasma and the CNS is thought to be maintained by SVCT2, because AA is mainly present in the reduced form in the plasma. On the other hand, the cerebroprotective effects of AA supplementation remain controversial, whereas beneficial effects of DHA administration for neuronal oxidative stress and inflammation seem to be well established [37–39]. Previous studies do not support acute cerebroprotective effects of treatment with AA under normal conditions because AA is not transported into the brain rapidly, owing to the high affinity and low ability for AA transport by SVCT2. Ahn and colleagues [40] reported that chronic AA supplementation has little or no effect on elevated oxidative stress in STZ-induced diabetic brain, whereas it is efficacious in the liver and kidneys. This phenomenon can be attributed to a lack of SVCT2 expression in endothelial cells of the BBB [41,42]. Recently, the evidence indicating that AA supplementation suppresses diabetes or ischemia-induced oxidative damage in the hippocampus and cortex of experimental animals has been accumulated [43–45]. These regions are not only vulnerable to oxidative stress but abundant in SVCT2 compared to other regions in the brain, namely, there may be some kinds of differences in sensitivity for AA supplementation depending on brain regions. Therefore, it seems possible that previous experiments using whole-brain samples might have not been adequate to detect the cerebroprotective effects of AA supplementation.

The system of DHA transport through GLUT1 in the BBB may not substantially contribute to total brain AA concentrations under normal conditions, because DHA concentrations in the plasma are usually only 0%–2% of AA concentrations [19,46]. Nonetheless, during oxidative stress in the CNS such as stroke and/or treatment with DHA, this route could play a significant role in the recruitment of AA [39,47]. Our study demonstrated that diabetes downregulates the expression of GLUT1 in the cortex, in line with previous studies [21,25]. AA supplementation to the diabetic rats markedly upregulated the basal expression of GLUT1; however, it exerted no significant increase in the total AA levels in the cortex of the rats. Taken together, contribution of the DHA transport system through GLUT1 to the cerebroprotection seems to be limited in the diabetic rats. The upregulation of GLUT1 expression by AA supplementation was observed also in the cortex of nondiabetic sham-operated group, despite that there were little changes in the total AA levels in the plasma of the rats. These results may indicate that orally supplemented AA has no direct enhancing effect on the expression of GLUT1 in the nondiabetic rats. However, AA concentration in the plasma is considered to increase within several hours after oral administration and decline to the basal level 24 h after the administration by excretion of the excess of AA, at which we collected the plasma samples from the experimental animals. Therefore, one possibility is that intermittent increases in AA concentration in the plasma repeated for two weeks could stimulate expression of GLUT1 in the cortex. Future research needs to be done with regard to the detailed mechanism underlying the upregulation of GLUT1 expression by AA supplementation.

SVCT2 has been shown to be crucial for maintaining AA for protection against oxidative stress in the CNS. It has been shown that cultured cells lacking this protein are vulnerable to oxidative stress [18]. In the present study, diabetes itself caused no apparent changes in the expression of SVCT2 in the cortex. On the other hand, after MCAO/Re, the expression of SVCT2 in the penumbral cortex, of which the majority seems to be localized in neurons, is upregulated in the nondiabetic rats. Gess and colleagues [20] have demonstrated that SVCT2 is specifically expressed in brain capillary endothelial cells and transports AA across the BBB during the subacute phase after MCAO/Re in mice. We also observed that expression levels of SVCT2 in capillary endothelial cells are increased after MCAO/Re in the nondiabetic rat cortex. This has been proposed as one of the mechanisms that protect the brain against oxidative injury under pathological conditions such as stroke and neurodegeneration. In contrast, we found that the upregulation of SVCT2 expression was abrogated in the diabetic rats, suggesting that the cerebroprotective mechanism via AA recruitment may be disrupted in the diabetic state. The abnormality of endothelial cell function in STZ-diabetic rats has been indicated to be related to the ROS generation in microvascular walls [17]. Seno and coworkers [48] demonstrated that the inflammatory cytokines TNF-α and IL-1β suppress the transport of AA through SVCT2 in human umbilical vein endothelial cells. Indeed, we detected a robust increase in the expression these cytokines in the brain of diabetic rats after MCAO/Re, which may reverse the upregulation of SVCT2 by MCAO/Re itself. AA supplementation to the diabetic rats upregulated the expression of SVCT2 in the cortex and restored the total AA levels to the nondiabetic levels. The antioxidative properties of AA may protect endothelial functions including the transcriptional regulation of SVCT2 against the exacerbated ischemic oxidative injury and the enhanced proinflammatory responses in the diabetic state, and may improve AA transport through SVCT2 in the BBB. Further experiments using radiolabeled tracers or other rodent models such as SVCT2+/− mice are needed to elucidate these cerebroprotective mechanisms of AA, including its effects on the specific transport of AA into the diabetic brain.
