HIV-1 Tat promotes astrocytic release of CCL2 through MMP/PAR-1 signaling

Abstract

The HIV-1 protein Tat is continually released by HIV-infected cells despite effective combination antiretroviral therapies (cART). Tat promotes neurotoxicity through enhanced expression of proinflammatory molecules from resident and infiltrating immune cells. These molecules include matrix metalloproteinases (MMPs), which are pathologically elevated in HIV, and are known to drive central nervous system (CNS) injury in varied disease settings. A subset of MMPs can activate G-protein coupled protease-activated receptor 1 (PAR-1), a receptor that is highly expressed on astrocytes. Although PAR-1 expression is increased in HIV-associated neurocognitive disorder (HAND), its role in HAND pathogenesis remains understudied. Herein, we explored Tat's ability to induce expression of the PAR-1 agonists MMP-3 and MMP-13. We also investigated MMP/PAR-1-mediated release of CCL2, a chemokine that drives CNS entry of HIV infected monocytes and remains a significant correlate of cognitive dysfunction in the era of cART. Tat exposure significantly increased the expression of MMP-3 and MMP-13. These PAR-1 agonists both stimulated the release of astrocytic CCL2, and both genetic knock-out and pharmacological inhibition of PAR-1 reduced CCL2 release. Moreover, in HIV-infected post-mortem brain tissue, within-sample analyses revealed a correlation between levels of PAR-1-activating MMPs, PAR-1, and CCL2. Collectively, these findings identify MMP/PAR-1 signaling to be involved in the release of CCL2, which may underlie Tat-induced neuroinflammation.

Keywords: CCL2; HIV; Tat; matrix metalloproteinase; protease activated receptor 1.

Figure 1.

Tat-induced CCL2 release from astrocytes is significantly reduced by MMP inhibition. A. Primary mouse cortical astrocytes were treated for 24-hours under serum-free conditions. Treatment with 100 ng/mL of recombinant Tat promoted CCL2 release as detected by ELISA. Unpaired t test, p < 0.0178, N = 3 biological replicates (3 separate cultures with 4 - 6 replicates per condition in each culture). B. Cells were treated the same as the previous experiment with or without the broad-spectrum MMP-inhibitor GM6001 (10 µM). GM6001 significantly attenuated Tat-induced CCL2 release. One-way ANOVA followed by Newman-Keuls multiple comparisons test, p < 0.05, N = 2 biological replicates (2 separate cultures with 3 replicates per condition in each culture). * p < 0.05

Figure 2.

Tat induces MMP-3 and MMP-13 expression in astrocyte cultures. A. Astrocytes were treated as described in the previous experiments. Tat exposure led to an increase in MMP-3 protein levels as detected by ELISA. Unpaired t test, p = 0.0016, N = 3 biological replicates (3 separate cultures with 3 - 6 replicates per condition in each culture). B. MMP-13 mRNA was increased by Tat when compared to control. Unpaired t test, p = 0.0005, N = 3 biological replicates (3 separate cultures with 6 replicates per condition in each culture); mRNA fold expression normalized to controls with statistical analysis done using the fold change values. C. A specific MMP-13 activity assay was used to detect potentially active enzyme species in the conditioned media of control and Tat-stimulated astrocyte cultures. Unpaired t test, p = 0.1709, N = 2 biological replicates (2 separate cultures with 6 replicates per condition in each culture). ** p < 0.01, *** p < 0.001

Figure 3.

PAR-1 activating MMPs induce CCL2 release from astrocytes. A. Astrocytes were treated for 24-hours with pro-MMP-13, catalytic MMP-13 (cMMP-13) and IL-1β. Catalytic MMP-13 was comparable to IL-1β in stimulating CCL2 release, while pro-MMP-13 was similar to control. Experiment conducted two times with separate cultures; single experiment shown with 3 technical replicates per condition. Control levels were non-detectable (N.D.); B. Catalytic MMP-13 also stimulated release of the pro-inflammatory cytokine IL-6. Unpaired t test, p = 0.0098, N = 3 biological replicates (3 separate cultures with 4 replicates per condition in each culture). C. cMMP-3 was also found to potently stimulate CCL2 release. Unpaired t test, p < 0.0001, N = 4 biological replicates (4 separate cultures with 3 - 4 replicates per condition). ** p < 0.01, **** p < 0.0001

Figure 4.

Genetic KO or blockade of PAR-1 reduces CCL2 release following MMP exposure. A. Primary cortical astrocytes were derived from WT or PAR-1 KO mice and treated with IL-1β. WT and PAR-1 KO astrocytes did not differ in their ability to release elevated CCL2 in response to IL-1β stimulation. Unpaired t test, p = 0.8686, N = 3 biological replicates (3 separate cultures with 3 - 4 replicates per condition in each culture). B. Astrocytes from both genotypes were treated with 20 nM of cMMP-13. cMMP-13 treated PAR-1-KO astrocytes displayed a reduction in CCL2 release when compared to WT astrocytes. Unpaired t test, p = 0.0001, N = 3 biological replicates (3 separate cultures with 3 - 4 replicates per condition in each culture). C. Pharmacological inhibition of PAR-1, using 2 µM vorapaxar, resulted in a significant reduction in MMP-induced CCL2 release. Unpaired t test, p = 0.0135, N = 3 biological replicates (3 separate cultures with 4 replicates per condition in each culture). * p < 0.05, *** p = 0.0001

Figure 5.

PAR-1-activating MMPs are detected in HIV+ post-mortem brain tissue and MMP levels correlate with CCL2. A. Representative Western blots shown for MMP-13 and PAR-1. In addition to the pro- and cleaved forms of MMP-13, a prominent band corresponding to the active catalytic domain of MMP-13 (~25 kDa) was readily detected. The PAR-1 antibody revealed bands corresponding to glycosylated and unglycosylated forms (66 and 47 kDa, respectively). Data were normalized to GAPDH. B. Representative Western blots shown for CCL2 and MMP-3. The CCL2 antibody revealed multiple bands corresponding to glycosylated forms of the protein. A single band was detected for MMP-3 (~54 kDa). Data were normalized to GAPDH. Each vertical lane across the representative blots corresponds to the same human sample. C. Within individual samples, protein levels were positively correlated. The catalytic domain of MMP-13 was significantly correlated with PAR-1 (Pearson correlation; r = 0.80, r² = 0.64, p = 0.003, n = 11), and with CCL2 (Pearson correlation; r = 0.66, r² = 0.43, p = 0.028, n = 11). MMP-3 was significantly correlated with CCL2 (Pearson correlation; r = 0.92, r² = 0.86, p < 0.0001, n = 11). Best fit line and 95% confidence band generated using linear regression.

Figure 6.

Hypothetical model illustrating Tat-induced CCL2 release via MMP/PAR-1 signaling. Recruitment of monocytes via CCL2 signaling may potentiate neuroinflammation that is characteristic of the HIV-infected brain. These pathological signaling pathways may underlie symptoms associated with HAND.