Here, we investigated how LPS-induced inflammation alters the function of drug transporters in microglia, the primary CNS target of HIV, using a clinically relevant concentration (50 nM) of the antiretroviral medication saquinavir as a prototypical probe substrate and the following model systems: a rat microglia cell line, HAPI, and primary cultures of rat and mouse microglia. Furthermore, we examined at a molecular level, what mechanisms may drive the observed changes in saquinavir accumulation and retention by microglia following an inflammatory LPS challenge. As noted in another rat microglia cell line
, accumulation of saquinavir into HAPI microglia cells was rapid, reached a plateau by one hour, and was increased significantly by a potent P-glycoprotein inhibitor PSC833. In this model, an increase in saquinavir accumulation in the presence of PSC833 provides an indirect measure of compound efflux by the transporter. Following both short (6 hour) and long term (24 hour) LPS exposure in microglia, the overall accumulation of saquinavir decreased in a dose-dependent manner, with significant decreases observed at 24 hours at doses greater than 2.5 ng/ml LPS. Using LPS in the presence and absence of the P-glycoprotein inhibitor PSC833, the decrease in saquinavir accumulation was only partially explained by increases in P-glycoprotein function, that is, by increased P-glycoprotein-mediated efflux of compound from the intracellular compartment to the outside of the cell. The remainder of the unaccounted saquinavir transport surprisingly could not be explained by increases in efflux or protein expression of Mrp1, a transporter known to handle saquinavir efficiently. Although less likely, a decrease in potential uptake of saquinavir into the cells via decreased SLC uptake transporter expression/function was also considered. Transcripts of seven well- characterized SLC transporters, some already well known to interact with ARs, were examined in the presence and absence of LPS. With the exception of Slc22a2, none of these transporters (at the transcriptional level) were expressed significantly in HAPI microglia. Furthermore, Slc22a2 transcript levels in HAPI microglia were unchanged following LPS exposure. Therefore, it is unlikely that a change in SLC uptake transporters explains the reduced accumulation of saquinavir following LPS treatment.
While it was clear that LPS exposure decreased accumulation of saquinavir significantly in microglia, at least partially through a P-glycoprotein pathway, protein levels of that transporter were unchanged. One explanation for this may be that transporter function was altered by signaling pathways downstream of TLR4, resulting in for example, transporter dephosphorylation, deglycosylation, or tyrosine nitration. Indeed, activation of NF-κβ in HT29 colon cancer cells decreases transport function of another drug transporter, human MRP3, via tyrosine nitration of the protein
. This suggests that TLR4 signaling regulates microglial P-glycoprotein activity to some extent, which is consistent with the fact that cytokines and NO are produced 6 to 24 hours later in the microglial response to LPS but fail to impact P-glycoprotein function/saquinavir accumulation in the current study. The role of TLR4 in P-glycoprotein regulation is particularly relevant to pharmacotherapy in HIV, as there is increasing evidence that HIV proteins may activate macrophages through a TLR4 dependent pathway. In fact, a recent study shows that HIV1-Vpr induces cytokine production from macrophages through TLR4/MyD88
A second explanation for the discrepancy between altered P-glycoprotein function and expression following LPS treatment is altered trafficking of P-glycoprotein from intracellular stores to the cell surface. To actively efflux compounds, P-glycoprotein must be correctly orientated on the plasma membrane. In polarized cells such as brain capillary endothelium and choroid plexus epithelia, proper routing of intracellular reserves of transporter protein to the plasma membrane on the apical side is achieved through a series of complex molecular signaling events. In brain capillaries, intracellular stores of P-glycoprotein may cycle into and out of the endothelial membranes following exposure to proinflammatory mediators as a short-term adaptive compensation mechanism to cellular stresses
. Mechanisms contributing to trafficking of drug transporter proteins within microglia have not been identified. However, immunohistochemical studies of P-glycoprotein in microglia have localized the protein to both the plasma and nuclear membranes, demonstrating that intracellular compartments for the protein do indeed exist
[11, 43, 44] and might be recruited in response to cellular stress.
The interaction of LPS with microglia at the molecular level and subsequent signaling pathway activation have been well described elsewhere
. At the cell surface level, LPS activation of TLR4, scavenger receptors and NADPH oxidase have all been implicated as initial events that initiate downstream intracellular signaling changes in microglia. Inhibition of the scavenger receptors (by fucoidan) and NADPH oxidase (by DPI) in the present studies did not attenuate the decrease in saquinavir accumulation following LPS challenge, whereas a TLR-4 neutralizing antibody caused partial attenuation. By decreasing TLR4 activity to a large extent using microglia from TLR4 deficient mice, full attenuation of the changes in saquinavir transport in the presence of LPS in primary microglia was seen. This demonstrates that TLR4 signaling at the cell surface is sufficient to initiate a signaling cascade that affects P-glycoprotein downstream.
In microglia, surface engagement of TLR4 by LPS leads to activation of multiple intracellular pathways including those connected to NF-κβ, AP-1, JAK/STAT, and multiple protein kinase pathways. Recent studies by Gibson et al.,
 have shown a role for NF-β in the regulation of P-gp in a mouse microglia cell line, BV-2. Interestingly, in this study, LPS at doses of 1 to 500 ng/ml for 12 hours reduced P-gp expression (mRNA and protein), and function using the fluorescent P-gp probe rhodamine 123. In the present study using primary cultures of mouse microglia, 10 ng/ml LPS decreased saquinavir accumulation significantly at 6 and 24 hours, presumably due to increased saquinavir efflux. The observed decrease in saquinavir accumulation in the mouse cultures was, however, modest compared to primary rat cultures, suggesting potential species differences. Whether species differences in molecular mechanisms or specific substrate handling can explain these discrepancies, remains to be confirmed.
Of all the molecular pathways examined in the present study, only inhibition of NF-κβ and MEK1/2 reversed the changes in saquinavir accumulation in microglia following LPS exposure. Given that several pro-inflammatory factors that are known activators of NF-κβ (for example, TNF-α and IL-1β) were shown to have no effect, these findings support that NF-κβ is necessary, but not sufficient to change saquinavir accumulation. These results are in stark contrast to findings in freshly isolated rat brain capillaries where LPS also initiates activation of TLR4, which downstream is connected to alterations in TNF-α, ET-1, iNOS and PKC activation, and ultimately results in increased P-glycoprotein protein expression and consequently function in the capillaries
. This may not be surprising, as the transporter profile in glial cells is quite different compared to cells of the BBB. Most notably, cultured microglia do not express significant levels of Mrp2
 or mRNA of any of the important SLC uptake transporters (that is, Slco221a2, 1a5, and Slc22a8) expressed at the BBB. Given the redundant nature of the LPS response in microglia (that is, multiple pathways are initiated via multiple cell surface receptors), we cannot rule out the possibility that compensatory pathways mask the effects of inhibition or activation of a single pathway in our cell cultures. Further investigations in vivo using knockdown strategies may be helpful to fully elucidate all the pathways that are involved.
In summary, we have demonstrated that exposing microglial cells to LPS decreases cellular accumulation of one representative antiretroviral medication. The ability of LPS to significantly decrease saquinavir accumulation was consistent between microglia derived from multiple species (rats versus mice), multiple strains within the same species (Wistar versus Fisher rats), and multiple cell preparations (cultured cell line versus primary cells). Using PSC833, a non-immunosuppressive cyclosporine-A analog and potent P-glycoprotein inhibitor, the decrease in saquinavir accumulation in cultured microglia was consistent, in part, with an increase in P-glycoprotein-mediated drug efflux. This increase in transporter activity and its absence in cells from TLR4-deficient mice suggest an important role for TLR4 in microglial P-glycoprotein function and demonstrate its importance for HIV pharmacotherapy. These results confirm that the presence of neuroinflammation within the brain parenchymal compartment can further exacerbate the ability of glial cells to actively extrude antiretroviral agents, and explains in part why treatment of neurologically-based HIV strains remains difficult despite our best efforts.