Site-specific hyperphosphorylation of pRb in HIV-induced neurotoxicity

Abstract

HIV-Associated Neurocognitive Disorder (HAND) remains a serious complication of HIV infection, despite combined Anti-Retroviral Therapy (cART). Neuronal dysfunction and death are attributed to soluble factors released from activated and/or HIV-infected macrophages. Most of these factors affect the cell cycle machinery, determining cellular outcomes even in the absence of cell division. One of the earliest events in cell cycle activation is hyperphosphorylation of the retinoblastoma protein, pRb (ppRb). We and others have previously shown increased ppRb expression in the CNS of patients with HIV encephalitis (HIVE) and in neurons in an in vitro model of HIV-induced neurodegeneration. However, trophic factors also lead to an increase in neuronal ppRb with an absence of cell death, suggesting that, depending on the stimulus, hyperphosphorylation of pRb can have different outcomes on neuronal fate. pRb has multiple serines and threonines targeted for phosphorylation by distinct kinases, and we hypothesized that different stimuli may target separate sites for phosphorylation. Thus, to determine whether pRb is differentially phosphorylated in response to different stimuli and whether any of these sites is preferentially phosphorylated in association with HIV-induced neurotoxicity, we treated primary rat mixed cortical cultures with trophic factors, BDNF or RANTES, or with the neurotoxic factor, N-methyl-d-aspartate (NMDA), or with supernatants containing factors secreted by HIV-infected monocyte-derived macrophages (HIV-MDM), our in vitro model of HIV-induced neurodegeneration. We found that, while BDNF and RANTES phosphorylated serine807/811 and serine608 in vitro, treatment with HIV-MDM did not, even though these trophic factors are components of HIV-MDM. Rather, HIV-MDM targets a specific phosphorylation site, serine795, of pRb for phosphorylation in vitro and this ppRb isoform is also increased in HIV-infected brains in vivo. Further, overexpression of a nonphosphorylatable pRb (ppRb S795A) attenuated HIV-MDM-induced neurotoxicity. These findings indicate that HIV-infection in the brain is associated with site-specific hyperphosphorylation of pRb at serine795, which is not induced by other tested stimuli, and that this phosphorylation contributes to neuronal death in this disease, demonstrating that specific pRb sites are differentially targeted and may have diverse impacts on the viability of post-mitotic neurons.

Figure 1. HIV-infected monocyte-derived macrophage supernatants cause neuronal death in vitro.

A) Reverse transcriptase activity of supernatants collected from HIV-infected human monocyte-derived macrophages (HIV-MDM) from three donors used in this study is shown. B) Loss of MAP2 as measured by MAP2 cell-based ELISA (black bars) and PI exclusion assay via hand counting (gray bars) of primary rat cortical cultures 21 DIV treated for 20 hours with HIV-MDM from three donors at a dilution of 1:20, as compared with untreated cultures or mock-infected MDM supernatants. C) Mock or HIV-MDM were diluted into the media of primary rat cortical neuroglial cultures 21 DIV for 20 hours and neuronal viability was measured by MAP2 ELISA (black bars) and PI exclusion assay (gray bars). (Student’s t-test, * p< 0.01)

Figure 2. Site-specific phosphorylation of pRb phosphorylation is dependent on the stimulus.

A) Western immunoblotting of one set of extracts for expression of the six phospho-pRb forms included in this study is shown. HIV-MDM treatment resulted in phosphorylation of pRb at serine795, while RANTES and BDNF lead to increased pRb phosphorylation on serine608 and serine807/811. Fold changes of ppRb serine608 (B), ppRb serine807/81 (C), ppRb serine795 (D), ppRb serine780 (E) band intensities, as compared with untreated are shown. (F) pRb phosphorylation at serine795 in cultures treated with HIV-MDM is not dependent on NMDA receptor activation. Western immunoblotting of one set of lysates obtained from cultures pre-incubated with 10 μM MK801, 30 minutes before the addition of HIV-MDM or Mock for 16 hours is shown. Fold changes in band intensities over that of the untreated lane are shown. Loading controls were used for normalization (n=3, One-way ANOVA, post-hoc Newman-Keuls, * p<0.05 vs. untreated and vs. Mock).

Figure 3. HIV-MDM treatment leads to increased ppRb serine795 in neuronal nuclei, whereas ppRb serine807/811 accumulates in the neuronal cytoplasm in response to BDNF and RANTES treatment in vitro.

21 DIV primary rat cortical cultures on coverslips were treated with RANTES, BDNF, NMDA, Mock or HIV-MDM for 1, 2 or 4 hours, followed by immunostaining for ppRb serine795 (green, A) or ppRb serine807/811 (green, B), MAP2 (red, A and B), and nuclei (DAPI, blue, A and B). Images were captured by laser confocal microscopy (n=3). Magnification 600X, scale bar = 20 μm.

Figure 4. Cyclin dependent kinase 5 (CDK5) preferentially phosphorylates pRb at serine608 and serine807/811 in response to BDNF.

21 DIV primary rat neuroglial cultures were exposed to indicated treatments for 4 and 16 hours in the presence or absence of a 30 min pre-incubation with the CDK5 inhibitor, Roscovitine (10 μM), or general CDK inhibitor, Olomoucine (50 μM). Whole cell lysates were used for immunoblotting with indicated antibodies. Inhibition of CDK5 activity by Roscovitine does not abolish ppRb serine795 formation in cultures treated with HIV-MDM (A), whereas Roscovitine blocks ppRb serine608 and serine807/811 formation by BDNF (C). Olomoucine effectively blocks all ppRb isoforms shown in cultures in response to indicated treatments (B) and (C). D) Nontoxic Roscovitine and Olomoucine doses were determined by treatment of cultures with the indicated concentrations for 20 hours.

Figure 5. Overexpression of a pRb mutant nonphosphorylatable at serine795 provides protection in neurons treated with HIV-MDM.

A GFP-expressing lentiviral expression system was utilized for overexpression of w.t., ppRb S795A, and ppRb S795E in differentiated neurons in vitro. A) Overexpression of specific mutants was assessed by immunoblotting using antibodies against total pRb and ppRb S795 in rat neurons 3 days post-infection with lentiviral expression vectors. B) Attenuation of HIV-MDM induced MAP2 loss in neurons overexpressing ppRbS795A, as compared with those uninfected, or with those overexpressing w.t. pRb or ppRbS795E. (* p< 0.05, One way ANOVA).

Figure 6. ppRb serine795 levels are increased in HIV-infected brain tissue.

A) and B) Formalin-fixed, paraffin embedded tissue slides from the cases used for immunoblotting were triple-labeled for ppRb serine795 (green), MAP2 (red), and DAPI (blue). Slides were visualized by triple-label immunofluorescent laser confocal microscopy (TLCM) and images taken were quantified for ppRb expression by MetaMorph imaging software. A) Representative composite images of 3 cases from HIV(−) and HIV(+) groups show increased ppRb serine795 immunoreactivity in neuronal nuclei of HIV(+) cases. Scale bar = 20 μm. B) Quantification shows an increase in ppRb serine795 immunoreactivity is statistically significant in HIV(+) group, as compared with HIV(−) group (Student’s t-test, *p<0.05). C) ppRb serine795 increase is statistically significant in the less severely compromised group (non-HAD) compared with the more severely compromised group (HAD). The changes in both non-HAD and HAD groups are statistically significant when compared with the neurocognitively normal group (Student’s t-test, * p<0.05). D) ppRb 807/811 was also examined via TLCM and then quantified using Metamorph. Quantification is show here. E) Whole cell tissue lysates from HIV (+) (n=14) and HIV(−) (n=4) cases (table 3) were blotted for all six ppRb forms included in this study. Western blots for ppRb serine795 and ppRb serine807/serine811 are shown. A band for coomassie blue is included to control for equal loading and protein degradation. E and F) ppRb serine795 levels are increased in HIV(+) cases, as compared with HIV(−) group. G) There is a trend towards significance in the increase in ppRb serine795 levels in the HAD group compared with the non-HAD group. E and H) We did not detect an increase in expression of ppRb serine807/811.