- Top of page
- Experimental procedures
Many phase II protective genes contain a cis-acting enhancer region known as the antioxidant response element (ARE). Increased expression of these genes contributes to the protection of cells from oxidative stress. Transgenic reporter mice were created that carry in their genome the core ARE coupled to the human placental alkaline phosphatase (hPAP) reporter gene. Primary cortical cultures derived from these mice were treated with tBHQ resulting in a dose-dependent increase in hPAP activity. Histochemical staining for hPAP activity was observed in both glia and neurons from tBHQ-treated cultures. The tBHQ-mediated increase in hPAP was not affected by the antioxidant glutathione monoethyl ester (GSHEE), whereas the increase in hPAP following DEM treatment was completely blocked by GSHEE. Pre-treatment of cultures with the PI3-kinase inhibitor LY 294002 demonstrated a dose-dependent decrease in tBHQ-induced hPAP activity. In addition, the tBHQ-mediated expression of ARE-driven genes in primary cortical cultures was blocked by LY 294002. Interestingly, basal expression of Nrf2 was also inhibited by LY 294002. We theorize that increased levels of genes controlled by the ARE are important for cellular protection against oxidative stress. These ARE-hPAP transgenic mice will be an important in vivo model for testing our hypothesis.
A cell's ability to protect itself from oxidative stress caused by the toxic accumulation of reactive oxygen species (ROS) and electrophiles, relies, in a large part, on the levels of phase II protective enzymes. The 5′ flanking region of many of these protective genes contain a cis-acting enhancer region known as the antioxidant response element (ARE). These include, but are not limited to, heme oxygenase-1 (HO-1), glutathione-S-transferases (GSTs), gamma-glutamylcysteine ligase (catalytic and regulatory subunits; GCL) and NAD(P)H:quinone oxidoreductase 1 (NQO1; Li and Jaiswal 1992; Mulcahy and Gipp 1995; Prestera et al. 1995; Moinova and Mulcahy 1998). The ARE was first identified in 1990 by Rushmore and Pickett (1990) as a regulatory element in the upstream region of the rat GST-Ya subunit. The core sequence is 5′-GTGACnnnGC-3′. It was later demonstrated that both the basal and inducible activity of NQO1 by planar aromatics, phenolic antioxidants and H2O2, was mediated through the ARE (reviewed in Dinkova-Kostova and Talalay 2000). Furthermore, it has been demonstrated that Nrf2 regulates the gene expression of phase II detoxifying enzymes through binding to the ARE (Venugopal and Jaiswal 1996; Itoh et al. 1999; Moinova and Mulcahy 1999; Wild et al. 1999).
The experiments defining ARE activation are based primarily on studies in hepatoma cell lines. NQO1, however, is located in most tissues, including brain as was seen by immunohistochemical-positive staining in rat glial cells in the forebrain (Schultzberg et al. 1988), Bergmann radial glial cells in the cerebellum (Schultzberg et al. 1988), and glial cells in human brains (Wang et al. 2000). Murphy et al. (1991) showed that pre-treatment of a neuroblastoma cell line with tert-butylhydroquinone (tBHQ), a known activator of the ARE, significantly attenuated glutamate toxicity by 80%. More recently it was demonstrated that pre-treatment of the N18-RE-105 neuroblastoma-retina hybridoma cell line with a different ARE activator, dimethyl fumarate (DMF), protected against dopamine and H2O2 toxicity. This protective effect was attributed to elevated intracellular GSH and increased activities of NQO1 and GST. A stable cell line overexpressing NQO1 and GST, however, was not resistance to dopamine toxicity (Duffy et al. 1998). This suggests that these two enzymes alone were not sufficient in protecting against dopamine toxicity, but that the coordinate induction of genes regulated by the ARE may have provided the mechanisms necessary for cell protection and survival.
In primary cultures of cerebellar glial cells and in a human neuroblastoma cell line, our laboratory has defined the core ARE sequence containing a complete 5′ palindrome and conserved 3′ GC nucleotide pair that is required for maximal activation by tBHQ (Ahlgren-Beckendorf et al. 1999; Moehlenkamp and Johnson 1999). Activation of the ARE in both culture systems was correlated with increased NQO1 activity. We have also shown that activation of the ARE in IMR-32 human neuroblastoma cells treated with tBHQ and diethyl maleate (DEM) was dependent upon the transcription factor, Nrf-2 (Lee et al. 2001a). Interestingly, activation of the ARE by tBHQ was independent of oxidative stress (Lee et al. 2001a). In addition, it was demonstrated that activation of the ARE in these human neuroblastoma cells was through a phosphatidylinositol-3-kinase (PI3-kinase) dependent pathway that lies upstream of Nrf-2 (Lee et al. 2001b).
To further understand how the ARE is controlled in primary cultures and in vivo, we have created two founder lines of transgenic reporter mice. Fifty-one basepairs of the rat NQO1 promoter region containing the ARE coupled to the human placental alkaline phosphatase (hPAP) reporter gene have been inserted into the genome of these mice. The studies presented in this manuscript were aimed at determining the cell-specific patterns of ARE activation in primary cortical cultures derived from the ARE-hPAP transgenic mice and if tBHQ-mediated activation of the ARE was associated with oxidative stress or PI3-kinase.
- Top of page
- Experimental procedures
We have successfully created mice that carry in their genome a stable hPAP reporter gene that is basally and transcriptionally controlled by a segment of the rat NQO1 promoter containing the core ARE sequence. This was demonstrated by the dose-dependent activation of hPAP by tBHQ in primary cortical cultures derived from the ARE-hPAP mice. In primary cortical cultures treated with tBHQ for 72 h, cell-specific activation of hPAP was identified in astrocytes and neurons. Histochemical staining for NQO1 followed in a similar fashion. A selective PI3-kinase inhibitor, LY 294002, blocked the activation of the ARE by tBHQ suggesting this activation was mediated through a PI3-kinase dependent mechanism involving Nrf2.
Recently published data from our lab demonstrated that the activation of the ARE by tBHQ and DEM in IMR-32 neuroblastoma cells was mediated through the transcription factor Nrf2 (Lee et al. 2001a). In addition, activation of the ARE by tBHQ was independent of oxidative stress, while activation of the ARE by DEM was dependent on oxidative stress. This implies that tBHQ and DEM can activate the ARE by different signaling pathways that converge on Nrf2. Similarly, in primary cultures derived from the ARE-hPAP reporter mice, pre-treatment with GSHEE blocked activation of the ARE by DEM but not tBHQ, suggesting the activation of the ARE by tBHQ was independent of oxidative stress, whereas activation of the ARE by DEM was a result of oxidative stress. Furthermore, total cellular GSH levels from primary cultures treated with DEM showed a marked decrease compared to that of vehicle which was reversed by pre-treatment with GSHEE. GSH levels in primary cultures treated with tBHQ were unaffected. However, to truly confirm these observations, continuing experiments in our laboratory are focused on measuring levels of oxidative stress biomarkers such as 8-oxo-7,8-dihydro-2′-deoxyguanosine, a marker of DNA damage, and 4-hydroxynonenal, a lipoxidation product.
We theorize that the activation of the ARE in neurons is also through a Nrf2-dependent mechanism. To test this theory, we have begun to cross breed our ARE-hPAP reporter mice with the Nrf2 knock-out mouse (Chan and Kan 1999) to study the effects of the absence of Nrf-2 on basal and inducible activation of the ARE in primary cortical cultures and in vivo. Work in our lab has also shown that activation of the ARE in neuroblastoma cells is regulated by PI3-kinase (Lee et al. 2001b). Likewise, in primary cortical cultures derived from the ARE-hPAP mice, LY 294002, in a dose-dependent manner, inhibited tBHQ activation of the ARE, although not completely at 72 h. This incomplete inhibition may suggest an alternative ARE-signaling pathway that does not depend on PI-3 kinase, or perhaps that the half-life of tBHQ is longer than that of LY 294002. Thus, as evident at the 24-h time point, LY 294002 would be a more effective inhibitor of PI-3 kinase than at 72 h. RT-PCR of RNA isolated from these cultures treated with tBHQ showed an induction of NQO1 and GCLR gene expression that was blocked by LY 294002. Interestingly, the basal gene expression of Nrf2 was inhibited by LY 294002. This loss of Nrf2 expression could significantly contribute to the inhibitory effects of LY 294002 on ARE-driven genes. It is possible that treatment with LY 294002 does not directly block tBHQ-mediated signaling, but indirectly modulates the level of Nrf2 that is required for activation of the ARE. If PI3-kinase is responsible for maintaining expression of Nrf2, then inhibition of this kinase would result in reduced levels of Nrf2 and loss of inducibility by tBHQ. In either case, it would appear that PI-3 kinase plays a major role in ARE-driven gene expression in primary cortical cultures.
Chan and Kan (1999) demonstrated the importance of the transcription factor Nrf2 when Nrf2 knockout mice fed BHA showed extreme pulmonary toxicity. Gene expression of NQO1, HO1 and CGLR were significantly reduced in the lungs of these animals suggesting these ARE-containing genes, regulated by Nrf2, play a critical role in protection from oxidative stress. Preliminary data using oligonucleotide microarrays and RT-PCR to compare the expression profiles of control neurons from Nrf2–/– and Nrf2+/+ littermates, indicate that NQO1, CGLR, GCLC and HO-1 mRNA levels are significantly reduced in the Nrf2–/– cultures making them more susceptible to oxidative stress (J. M. Lee and J.A. Johnson, unpublished observation). These findings support our hypothesis that the coordinate induction of the ARE-containing genes may confer protection against neuronal cell death.
Activation of the ARE by tBHQ was observed in neurons and astrocytes. Neurons that stained positive were in close proximity to astrocytes. Perhaps the astrocytes in these cultures are secreting obligatory factors required for activation of the ARE in neurons. This observation was supported by the increased number of neurons staining positive for hPAP activity in cultures maintained in the absence of AraC, a mixed culture system of astrocytes and neurons (Fig. 2b), compared with cultures maintained in the presence of AraC which are predominately neurons (Fig. 2d). This is the first report indicating that the ARE and NQO1 can be activated and expressed, respectively, in cultured neurons. The data suggest that although post-mitotic-differentiated neurons in human and rodent brain are consistently negative for NQO1, they may have the capacity to reverse this differentiation imposed repression of the ARE and genes controlled by this enhancer element.
In contrast to the data presented here, transient transfection of ARE reporter gene constructs in rat primary cortical cultures that were treated with DMF for 24 h, showed preferential activation of the ARE in astrocytes, but not in neurons (Murphy et al. 2001). We attribute the inability of neurons to express the ARE in this system to a shorter (24 h vs. 72 h) exposure of the cultures to the ARE activator, DMF. The longer exposure may afford the non-neuronal cell population the time needed to manufacture and secrete factors required by neurons to activate the ARE. In addition, Laxton et al. (2001) demonstrated that in rats given a unilateral focal cerebral infarct, NQO1 activity was significantly increased in astrocytes localized to the cortex and subcortical areas as compared with the non-ischemic side. NQO1 activity in neurons was not detected. In this system, ischemia resulted in an acute oxidative insult and rapid neuronal cell death. It would therefore be unlikely to observe NQO1 activity in neurons that are dying or in the process of apoptosis.
In support of the ARE activation in neurons, Wang et al. (2000) reported that in brains from AD patients, NQO1 activity was observed in hippocampal neurons and was undetectable in the same neuronal population in healthy controls. It was hypothesized that increased NQO1 was a result of the activation of the ARE in response to the reactive oxygen species generated by the accumulation of amyloid-beta protein (Aβ). However, as stated earlier, the stable overexpression of NQO1 in a neuroblastoma cell line was not sufficient to block oxidative stress-induced cell death (Murphy et al. 1991). This would imply that the coordinate activation of ARE-driven genes may be required for neuroprotection. In the AD brain, all the surviving neurons in close proximity to the pathology were NQO1-positive, while only a small fraction were positive for phospho-tau. The question remains as to the effect of increased ARE-driven gene expression on Aβ toxicity. Future experiments with the ARE-hPAP reporter mice therefore will aim to correlate the activation of the ARE in neurons and glia with neuroprotection against Aβ toxicity in primary cortical cultures.
Clearly, screening for ARE-activating compounds could prove to be important in the role of drug discovery. The ARE-hPAP reporter mice have the potential to be used for such purposes to determine the cell-specific activation of the ARE by various drugs. The ARE-hPAP reporter mice could also be used as an extremely sensitive model for the determination of chemicals that cause oxidative stress. Lastly, the identification of the signaling pathway in the activation of the ARE has not fully been elucidated. Studies using primary cultures or tissue from the ARE-hPAP reporter mice in correlation with hPAP activity can only further our knowledge of the ARE signaling pathway.