Lipid Hydroperoxides Induce Apoptosis in T Cells Displaying a HIV-associated Glutathione Peroxidase Deficiency*

immunodeficiency (HIV)-infected human T cell to be extremely to peroxide (H202)-induced apoptosis due to a HIV-associ- ated catalase deficiency. Here we report that HTV gene expression additionally renders 8EB cells 10-fold more sensitive than either uninfected cells or HIV-in- fected but nonexpressing 8EBL cells to killing by 15-hy-droperoxyeicosatetraenoic acid (lS-HPETE), as well as several other hydroperoxy fatty acids. Whereas the viability of A3.01 and 8ESL cells was relatively unaffected by exposure to 10 p~ 15-HPETE, similarly treated 8ES cells underwent apoptosis, as demonstrated by morpho- logical changes and the presence of fragmented DNA.

The unique susceptibility of 8EB cells was attributable to their inability to convert 15-HPETE to 1S-hydroxyeicosatetraenoic acid (15-HETE) owing to a marked reduction in glutathione peroxidase activity. Since oxidized lipids have been reported to accumulate in oxidatively stressed, HTV-infected individuals, a HIVassociated glutathione peroxidase deficiency may contribute to the depletion of CD4 T cells that occurs in the acquired immune deficiency syndrome (AIDS).
The depletion of CD4 T cells in acquired immune deficiency syndrome (AIDS1 has been suggested to result from apoptosis (1)(2)(3)(4), but the underlying mechanisms have not been defined. It has been proposed that oxidative stress may be involved (5,6), as various indicators of oxidative stress, including decreased thiol levels, antioxidant deficiencies, and evidence of lipid peroxidation, have consistently been observed in human immunopayment of page charges. This article must therefore be hereby marked * The costs of publication of this article were defrayed in part by the "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Control, Atlanta, GA 30333.  deficiency virus (HIV)-infected individuals (7)(8)(9). Furthermore, studies with acutely and chronically HIV infected T cell lines have demonstrated a correlation between HIV gene expression and decreased levels of the antioxidants thioredoxin, superoxide dismutase, and catalase (10)(11)(12)(13). The resultant compromised ability of HIV-infected T cells to avoid oxidative stress can lead to increased HIV replication (14,15) and a markedly enhanced susceptibility to cell killing by hydrogen peroxide (Hz02) (13). Both effects may contribute to the depletion of CD4 T cells in vivo.
One indirect way in which HzOz can kill cells is via the oxidation of membrane polyunsaturated fatty acids to form fatty acid hydroperoxides (16). Oxidation of arachidonic and linoleic acids, the major polyunsaturated fatty acids of mammalian cell membranes, can lead to the formation of hydroperoxyeicosatetraenoic acids (HPETEs) and hydroperoxyoctadienenoic acid (HPODEs), respectively. By extracting a hydrogen atom from adjacent fatty acyl chains, lipid hydroperoxides can initiate an autocatalytic chain reaction culminating in membrane destruction and cell death (16). Hydroperoxy fatty acids can also be synthesized intracellularly by the action of cellular lipoxygenases. In many cases the formation of HPETEs coincides with an increased intracellular production of reactive oxygen species (ROS) in response to an extracellular stimulus. For example, upon binding to cell surface receptors, tumor necrosis factor-a (TNF) elicits the intracellular formation of both ROS and oxidized products of arachidonic acid (17,181. The fact that cells susceptible to TNF-mediated cytolysis can be protected by limiting arachidonic acid availability (19-21) or oxidation (22)(23)(24) indicates that HPETE may be one of the cytolytic effector molecules the immune system uses to kill tumor cells and foreign pathogens. Given that HIV-infected individuals may display elevated levels of TNF production (25,261, and that such individuals are typically infected with numerous opportunistic pathogens which can enhance the release of ROS from phagocytic cells (27), the formation of lipid hydroperoxides may be of particular importance in AIDS. Furthermore, HIV itself has been reported to evoke the formation of oxidized arachidonic acid metabolites (28,29).
Mammalian cells usually avoid the toxic effects of HPETEs by converting the lipid hydroperoxides to more benign hydroxyeicosatetraenoic acids (HETEs) in a reaction requiring reduced glutathione (GSH), and catalyzed by the antioxidant enzyme, glutathione peroxidase (30): HPETE + 2GSH + HETE + GSSG + HzO. Since HIV-infected T cells have previously been shown to be defective in several cellular antioxidants, most notably GSH (31), it was of interest to determine whether HIVinfection affects T cell glutathione peroxidase activity. Therefore, the chronically HIV-infected 8E5 cell line, which was previously shown to display a HIV-associated catalase deficiency (13), was used to determine the effects of HIV infection on 15-HPETE metabolism. The results described herein demonstrate a direct relationship between HIV gene expression and impaired glutathione peroxidase activity, which increases T cell susceptibility to hydroperoxy fatty acid-induced apoptosis.
VmbilityAssay-A3.01,8E5 and 8E5L cells grown in 10% FBSRPMI were diluted 15 with RPMI to yield -4 x lo5 celldm1 in 2% FBS/RPMI. Triplicate aliquots of 0.2 ml were added to flat-bottomed, 96-well culture plates, followed by the addition of either arachidonic acid, 15-HPETE, or (13,34) was added to all of the culture wells, followed 4 h later by 20 pl of 37% formaldehyde. Cell viabilities were subsequently assayed by measuring the optical density at 492 nm (690 reference filter) using an Anthos model 2001 plate reader. Background values corresponding to wells containing only 2% FBSIRPMI, MTS, and 3.7% formaldehyde were subtracted and the resultant data expressed as the "percent of control," where the values obtained for each untreated cell line were set at 100%. In one series of experiments, the cells were incubated overnight in the presence of 1 m~ buthionine sulfoximine (BSO) to deplete them of GSH (35) prior to exposing them to 15-HPETE.
Apoptosis Assay-DNA fragmentation was assayed as described (13,37). Briefly, 2 x lo6 cells were pelleted and lysed by the addition of ice-cold 20 m~ Tris-HCI (pH 7.4) containing 10 m~ EDTA and 0.2% Triton X-100. Following centrifugation to pellet high molecular weight DNA, the supernatant was treated sequentially with proteinase K and RNase A. The DNA remaining was extracted twice with phenol, once with chlorofodisoamyl alcohol (24:1), and then precipitated overnight. Recovered DNA fragments were separated by electrophoresis through a 1.5% agarose gel and visualized with ethidium bromide. Assay of Cellular Antiodants-Cellular levels of GSH were determined using a GSH-400 colorimetric assay kit obtained from Bioxytech (Paris, France). Triplicate aliquots of cells grown in 10% FBS/RF"I were collected by centrifugation and washed once with phosphate-buffered saline. The cell pellets were treated with 5% sulfosalicylic acid followed by two cycles of freeze-thaw. Supernatants recovered after centrifugation at 10,000 rpm for 15 min were assayed for GSH as per instructions provided with the kit. "Triplicate cell samples ( 5 x lo7 cells each) to be assayed for glutathione reductase activity or glutathione peroxidase activity were suspended in 1 ml of 10 m~ potassium phosphate buffer (pH 7.0) containing 0.2% Triton X-100, followed by one cycle of freeze-thaw (38). After centnfugation (10,000 r p m / l O mid, supernatants were assayed spectrophotometrically for glutathione reductase (39) and glutathione peroxidase (40) activities.

RESULTS AND DISCUSSION
Using a human T cell lineage consisting of A3.01 (uninfected), 8E5 (chronically HIV-infected, constitutive HIV-expressive), and 8E5L (HIV-infected, nonexpressive) cells, we previously demonstrated a n unique susceptibility of the 8E5 cell line to killing by H202 (13). Furthermore, it was shown that HIV gene expression was associated with reduced levels of catalase activity, thereby impeding the cells' ability to degrade H202 (13). The same cell lineage was used to determine whether HIV expression renders T cells susceptible to killing by another peroxide compound, 15-HPETE. Within 4-10 h after exposure to 2.5-10 1.1~ 15-HPETE, 8E5 cells displayed a dose-dependent reduction in cell viability (LDS0 -2 p) such that at a dose of 10 p~ 15-HPETE, ~5 % of the cells remained viable (Fig. 1). By comparison, A3.01 and 8E5L cells were at least 10-fold more  20 p). Similar results were observed with 5-HPETE, 12-HPETE, 13-HPODE, and 13-HPOTE, but equivalent doses of arachidonic acid, 15-HETE, 9(10)-EODE, or 12(13)-EODE either had much less (9,lO-EODE) or no effect on any of the three cell lines (Figs. 1 and 2). Experiments in which the 10% FBSRPMI cell cultures were diluted with a serum-free medium containing selenium (13,371 yielded identical results (data not shown), thus ruling out selenium deficiency (41) as being responsible for the enhanced sensitivity of 8E5 to hydroperoxy fatty acids.
The 15-HPETE-induced killing of 8E5 cells was accompanied by marked changes in cellular morphology, most notably the presence of crescent-shaped and fragmented nuclei (421, which were evident by both phase contrast microscopy (Fig. 3) and fluorescence microscopy of acridine orange-stained cells (not shown). Additionally, gel electrophoresis of DNA from 8E5 cells exposed to 15-HPETE displayed the presence of -200-base pair DNA fragments (Fig. 4), which are definitive of apoptosis (42). Although other investigators have shown that lipid hydroperoxides such as HPETEs are toxic for mammalian cells (34,411, to our knowledge the ability of lipid hydroperoxides to directly induce apoptosis has not previously been reported. Mammalian cells usually detoxify hydroperoxy fatty acids by converting them to hydroxy fatty acids in a reaction catalyzed by glutathione peroxidase and requiring GSH (30,38,41). To confirm the importance of this antioxidant system in the relative resistance ofA3.01 and 8E5L to the cytolytic effects of lipid hydroperoxides, the three cell lines were incubated overnight in the presence of BSO, a n inhibitor of glutathione synthesis (35), followed by exposure to 15-HPETE. As expected, BSO treatment rendered A3.01 and 8E5L as susceptible as 8E5 to 15-HPETE-induced apoptosis (Figs. 1 and 3).
To determine if 8E5's susceptibility to killing by 15-HPETE reflected a n inability to detoxify lipid hydroperoxides, intact A3.01, 8E5, and 8E5L cells were compared for their ability to convert 15-HPETE to 15-HETE (Fig. 5). The detoxifying reduction of 15-HPETE to 15-HETE was efficiently carried out by A3.01 and 8E5L cells, but 8E5 cells were almost completely void of such activity. It should be noted that the 15-HPETE and 15-HETE recovered from the cell supernatant accounted for 935% of the hydroperoxy fatty acid initially added, suggesting that glutathione peroxidase constitutes the major degradative pathway.
The inability of 8E5 cells to convert 15-HPETE to 15-HETE suggested a HIV-associated defect in either the synthesis or utilization of GSH. The three cell lines were therefore compared for their levels of GSH and glutathione peroxidase activity, as well as glutathione reductase activity, the enzymatic step that regenerates GSH from oxidized glutathione (GSSG) (39). As shown in Fig. 6 (a and b ) , A3.01, 8E5, and 8E5L cells displayed equivalent levels of GSH and glutathione reductase activities, making it unlikely that the susceptibility of 8E5 to hydroperoxy fatty acids was due to the cells' inability to synthesize or regenerate GSH. However, the level of glutathione peroxidase activity in 8E5 was significantly less than in A3.01 or 8E5L: 65% lower when assayed using H202 as the peroxide substrate (Fig. &), and 85% lower with 15-HPETE as the peroxide substrate (Fig. 6d). This finding is in agreement with at least one report of decreased glutathione peroxidase activity in HIV-infected individuals (43). Collectively, these findings causally relate HlV gene expression with a glutathione peroxidase deficiency, a lack of fatty acid hydroperoxide detoxification, and enhanced susceptibility to lipid hydroperoxide-induced apoptosis.

12-HPETE
The antioxidant imbalance that coincides with HlV infection may contribute to the depletion of CD4 T cells and development of lymphomas which are characteristic ofAIDS. By shifting the redox balance of T cells to an oxidative equilibrium, HIV can promote its own replication (14, 15, 44), thereby enhancing virus cytopathicity. Furthermore, based on our present and previous studies (13), a HN-associated antioxidant deficiency may predispose HN-infected, oxidatively stressed T cells to apoptosis. Thus, T cells localized to, or passing through, infected lymphoid tissues may be exposed to ROS released by activated phagocytic and inflammatory cells (27,45). Another possibility is that endogenous ROS production in T cells will be increased following activation by antigen or TNF (17, 46). In either case, a resultant formation of lipid hydroperoxides may trigger the "activation-induced death" of HN-infected, antioxidant-deficient T cells (3). Finally, fatty acid hydroperoxides can act as tumor promoters (471, and their persistence in HIVinfected, glutathione peroxidase-deficient cells may play a role in AIDS-related lymphomagenesis. Such considerations provide a rationale for the proposed therapeutic use of antioxidants in AIDS (8) to alleviate oxidant-dependent damage attributable to activated macrophages or other phagocytic cells.
However, since the oxidants released by these cells are also vital in host defense, suppression of oxidant-mediated cytolysis could also suppress oxidant-mediated killing of opportunistic pathogens.