Fig. 1. IFN-γ enhances neuronal injury induced by HIV-1 proteins gp120 or Tat in vitro and in vivo. Primary cultured neuronal cells were treated with gp120 (250 ng/ml), Tat (250 ng/ml), IFN-γ alone or gp120 (250 ng/ml), Tat (250 ng/ml) in combination with IFN-γ (100 U/ml; IFN-γ/gp120/or IFN-γ/Tat) for 12 h. Cell cultured media were collected for LDH assay (A) and cell lysates were prepared for neuronal injury examination by Western blot analysis (B). Data are presented as the mean ± SD of LDH release and Western blot band density ratio of Bcl-xL to Bax (n = 3). One-way ANOVA followed by post hoc comparison revealed significant differences between gp120 or Tat and HIV-1 gp120 or Tat plus IFN-γ (**P < 0.001) for both LDH release and band density ratio of Bcl-xL to Bax. (C) Mouse coronal, frozen brain sections were stained with NeuN and NeuN/DAPI. Marked neuronal damage was observed in the gp120 plus IFN-γ condition compared to controls. Similar result were also observed in the Tat plus IFN-γ condition (data not shown). (D) Bcl-xL and Bax protein levels in mouse brain homogenates were analyzed by Western blot. Data are presented as the mean ± SD of Western blot band density ratio of Bcl-xL to Bax (n = 8; 4 male/4 female). One-way ANOVA followed by post hoc comparison revealed significant differences between gp120 or Tat compared to gp120 or Tat plus IFN-γ for band density ratio of Bcl-xL to Bax (**P < 0.001).

Fig. 2. JAK/STAT1 signaling is critically involved in the IFN-γ mediated enhancement of HIV-1 gp120 and Tat-induced neuronal damage. Primary cultured neuronal cells were cotreated with IFN-γ (100 U/ml) and gp120 or Tat at 250 ng/ml in the presence of JAK inhibitor (50 nM) for 12 h. Cell cultured media were collected for LDH assay (A) and cell lysates were prepared for neuronal injury examination by Western blot analysis (B). Data are presented as mean ± SD of LDH release and Western blot band density ratio of Bcl-xL to Bax (n = 3). One-way ANOVA followed by post hoc comparison revealed significant differences between IFN-γ/gp120 or IFN-γ/Tat compared to JAK inhibitor/IFN-γ/gp120 or JAK inhibitor/IFN-γ/Tat (**P < 0.001). Primary neuronal cells derived from STAT1-deficient mice were treated with gp120 or Tat at 250 ng/ml in the presence or absence of IFN-γ (100 U/ml) for 12 h. Cell cultured media and cell lysates from these cells were subjected to LDH assay (C) and Western blot analysis (D). Data are presented as the mean ± SD of LDH release and Western blot band density ratio of Bcl-xL to Bax (n = 5). One-way ANOVA followed by post hoc comparison revealed significant differences between STAT1-deficient neurons compared to wild-type neurons following treatment with IFN-γ/gp120 or IFN-γ/Tat for both LDH release and the band density ratio of Bcl-xL to Bax (**P < 0.001).

Fig. 3. EGCG inhibits IFN-γ-induced JAK/STAT1 phosphorylation; protecting neurons from injury induced by IFN-γ/gp120 or IFN-γ/Tat in vitro. JAK1 and STAT1 protein phosphorylations were examined by Western blot (A, C). Cell lysates were prepared from primary cultured neurons treated with IFN-γ (100 U/ml) for various time points as indicated (A, C). Cell lysates were prepared from primary cultured neurons cotreated with IFN-γ (100 U/ml) and EGCG at different doses as indicated for 60 min (B, D). To examine the putative role of EGCG in opposing neuronal injury induced by IFN-γ/gp120 or IFN-γ/Tat, primary neurons were cotreated with gp120 or Tat at 500 ng/ml in the presence of IFN-γ (100 U/ml) and EGCG (20 μM) for 12 h. Cell cultured supernatants were collected for LDH assay (E) and cell lysates were prepared for Bcl-xL/Bax Western blot analysis (F). Data are presented as the mean ± SD of LDH release and Western blot band density ratio of Bcl-xL to Bax (n = 3). One-way ANOVA followed by post hoc comparison revealed significant differences between IFN-γ/gp120 or IFN-γ/Tat compared to EGCG/IFN-γ/gp120 or EGCG/IFN-γ/Tat for both LDH release and band density ratio of Bcl-xL to Bax (**P < 0.001).

Fig. 4. Mice i.p. injected with EGCG demonstrate significant reductions in neuronal injury after i.c.v. injection of IFN-γ/gp120, IFN-γ/Tat or IFN-γ/gp120/Tat. (A) Coronal, frozen mouse brain sections were stained with NeuN (top panels) and NeuN/DAPI (lower panels) and analyzed for neuron injury/loss. A marked reduction of neuronal damage was observed when EGCG was added to either IFN-γ/gp120 or IFN-γ/gp120/Tat. Similar effects of EGCG were also observed in IFN-γ/Tat condition (data not shown). (B) Bcl-xL and Bax protein levels in mouse brain homogenates were analyzed by Western blot. Data are presented as mean ± SD of Western blot band density ratio of Bcl-xL to Bax (n = 8; 4 female/4 male). One-way ANOVA followed by post hoc comparison revealed significant differences in the band density ratio of Bcl-xL to Bax observed between gp120/IFN-γ or gp120/Tat/IFN-γ compared to gp120/IFN-γ/EGCG or gp120/Tat/IFN-γ/EGCG conditions, respectively (**P < 0.001).