Infection promotes Ser-214 phosphorylation important for generation of cytotoxic tau variants

Abstract

Patients who recover from hospital-acquired pneumonia exhibit a high incidence of end-organ dysfunction following hospital discharge, including cognitive deficits. We have previously demonstrated that pneumonia induces the production and release of cytotoxic oligomeric tau from pulmonary endothelial cells, and these tau oligomers can enter the circulation and may be a cause of long-term morbidities. Endothelial-derived oligomeric tau is hyperphosphorylated during infection. The purpose of these studies was to determine whether Ser-214 phosphorylation of tau is a necessary stimulus for generation of cytotoxic tau variants. The results of these studies demonstrate that Ser-214 phosphorylation is critical for the cytotoxic properties of infection-induced oligomeric tau. In the lung, Ser-214 phosphorylated tau contributes to disruption of the alveolar-capillary barrier, resulting in increased permeability. However, in the brain, both the Ser-214 phosphorylated tau and the mutant Ser-214-Ala tau, which cannot be phosphorylated, disrupted hippocampal long-term potentiation suggesting that inhibition of long-term potentiation was relatively insensitive to the phosphorylation status of Ser-214. Nonetheless, phosphorylation of tau is essential to its cytotoxicity since global dephosphorylation of the infection-induced cytotoxic tau variants rescued long-term potentiation. Collectively, these data demonstrate that multiple forms of oligomeric tau are generated during infectious pneumonia, with different forms of oligomeric tau being responsible for dysfunction of distinct end-organs during pneumonia.

Keywords: Pseudomonas aeruginosa; cytotoxicity; endothelium; long-term potentiation; oligomeric tau; pneumonia.

FIGURE 1

Immunodepletion of Ser-214 tau from supernatants generated following P. aeruginosa infection of rat PMVEC depletes cytotoxic activity of the supernatants. Cytotoxicity was assessed after incubation of naïve PMVECs treated with supernatant collected from PMVECs that had been infected with either ExoY⁺ (B and C) or PA103 strains of P. aeruginosa (D and E). Both untreated supernatant (B and D) and supernatant that was immunodepleted using anti-Ser-214 antibody (C and E) were used. Cells incubated with HBSS alone (A) were used as a negative control. Quantitation of cell killing is shown (F). * denotes p ≤ .05 when comparing B to C; # denotes p ≤ .05 when comparing D to E; N = 4. Bar = 50 μm.

FIGURE 2

Mutation of Ser-214 to Ala decreases the production of cytotoxic oligomeric tau following P. aeruginosa infection. (A) Immunoblot analysis for Ser-214 tau in whole-cell extracts prepared from tau knockout cells (KO) and KO cells that were transfected individually with constructs encoding either ON4R, 1N4R, 2N4R, or big tau (BT). Cell extracts were prepared from both uninfected cells (U) and cells that had been infected for 4 h with ExoY⁺ bacteria (I). M_r in kDa. (B) Immunoblot analysis of supernatant collected from PMVECs expressing either wild-type 1N4R tau (1 N4) or mutated Ser-214 tau (Mut) PMVECs following infection by ExoY⁺ bacteria using T22 anti-oligomeric tau antibody and anti-Ser-214 antibody. M_r in kDa. (C) Immunoblot analysis of nonmutated 1N4R-expressing PMVECs (S214) and mutant 1N4R expressing PMVECs (S214A) for total tau expression using Tau1 antibody (Left panel). Both uninfected (C) and ExoY⁺-infected cells (ExoY) were analyzed. The same samples were also analyzed by immunoblot using antibody specific for Ser-214 tau. Tau in mutant cells could not be phosphorylated on Ser-214 (middle panel). The same samples were analyzed by immunoblot using Actin antibody as a loading control (right panel). M_rs in kDa. (D) Analysis of cytotoxicity of PMVECs expressing 1N4R tau and 1N4R tau mutated at Ser-214. Supernatant was collected from WT PMVECs (b), tau KO PMVECs (c), tau KO PMVECs expressing 1N4R tau (d), and tau KO PMVECs expressing 1N4R tau mutated at Ser-214 (e) following infection by P. aeruginosa strain ExoY⁺. The supernatants were applied to naïve PMVECs and cell killing was quantified (f). Cells treated with HBSS served as a negative control (a). * denotes p ≤ .05 when comparing 1N4R to Mut and # denotes p ≤ .05 when comparing wild type to KO.

FIGURE 3

Pseudomonas aeruginosa infection leads to activation of PKA contributing to production of cytotoxic tau oligomers. (A) Immunoblot analysis of human hippocampal extract using antibody against Ser-214 tau failed to identify Ser-214 phosphorylated tau in the untreated control brain homogenate (C). PKA was immunoprecipitated from PMVEC extract and then incubated with ATP and hippocampal homogenate, and then, the mixture was analyzed by immunoblot analysis using antibody against Ser-214 tau (MV). Incubation of the brain extract with PKA immune-isolated from uninfected PMVECs demonstrated that PMVEC PKA can phosphorylate Ser-214 of tau. (B) PKA isolated from PMVECs that had been infected with either ExoY⁺ or PA103 strains of P. aeruginosa was more effective at phosphorylating brain tau than PKA isolated from uninfected (Un) control PMVECs. PKA was immunoprecipitated from untreated control PMVECs (Un) and PMVECs that were infected with either ExoY⁺ (ExoY) or PA103 (PA103), the immunoprecipitated PKAs were incubated with hippocampal extract and ATP, and Ser-214 phosphorylation was measured by immunoblotting using antibody against Ser-214 tau. Ser-214 phosphorylation by PKA isolated from ExoY⁺- and PA103-infected cells was increased 2.00 ± 0.79-fold and 1.98 ± 0.37-fold, respectively, relative to PKA-isolated untreated cells. Both values represent a significant increase in kinase activity (p ≤ .05, N = 7, ANOVA). (C) Inhibition of PKA using the PKI inhibitor blocked the release of hyperphosphorylated tau oligomers. Untreated control PMVECs (Un) and PMVECs treated with the PKI inhibitor (PKI) were infected with ExoY⁺. Supernatants were collected and then analyzed by immunoblot using T22 anti-oligomeric tau antibody and anti-Ser-214 tau antibody. M_rs in kDa. (D) Inhibition of PKA decreased cytotoxic activity of supernatants generated by ExoY⁺ infection of PMVECs. PMVECs were treated with either HBSS (a), control ExoY⁺ supernatant (b), or supernatant that was collected from PMVECs that were pretreated with PKI prior to infection with ExoY⁺ (c). Quantitation of cell killing is shown (d). * denotes p ≤ .05 when comparing supernatant collected from untreated ExoY⁺ infected cells to supernatant from cells pretreated with PKI prior to infection, N = 3.

FIGURE 4

Pharmacological activation of PKA does not lead to generation of cytotoxic hyperphosphorylated tau. (A) Cultured PMVECs were either treated with rolipram (10 μM) and forskolin (100 μM) for 9 h (b) or incubated with ExoY⁺ bacteria for 6 h (c). Both treatments induced cell retraction and gap formation in the monolayer. An untreated control confluent monolayer of rat PMVECs is also shown (a). Bar = 50 μm. (B) Treatment with rolipram and forskolin leads to increased PKA activity. PKA was immunopurified from lysates generated from untreated control cells (C) and cells treated with rolipram and forskolin (R/F), and then, the immunopurified material was incubated with human brain extract along with ATP. Ser-214 tau levels then were measured by immunoblot using Ser-214 antibody followed by band intensity quantitation using ImageJ software, (n = 6, 1.39 ± 0.16-fold increase in PKA activity in the drug-treated cells, p ≤ .05, ANOVA). (C) Phospho-tau was not secreted from cells that were activated with rolipram and forskolin. Whole-cell lysates (Lys) and cell culture supernatants (Super) were collected from untreated control cells (C) and cells that were treated with rolipram and forskolin (R/F), and then, each sample was analyzed by immunoblot using either anti-Ser-214 antibody or T22 oligomeric tau antibody. Molecular weights in B and C are in kDa. (D) Cytotoxicity was assessed by adding supernatants collected from either ExoY⁺-infected PMVECs (b) or PMVECs that were treated with rolipram and forskolin (c). Control cells were incubated with HBSS alone (a). Cytotoxic activity was not detected in the supernatant from rolipram and forskolin-treated cells. Bar = 50 μm.

FIGURE 5

ExoY⁺ promotes Ser-214 phosphorylation of tau which is important for its lung cytotoxicity. (A) Supernatant was collected from wild type, tau knockout, 1N4R expressing, and Ser-214 mutant tau-expressing cells following ExoY⁺ infection, concentrated to 150 μL and then introduced into the circulation of isolated perfused rat lungs. Filtration coefficient ($K_f$) was recorded at 2 and 4 h postaddition of cytotoxic supernatant. Baseline $K_f$ among groups was 0.15 ± 0.02 g min⁻¹ cm H₂O⁻¹ 100 g⁻¹. * denotes p ≤ .05, using two-way ANOVA with Tukey's multiple comparison test. (B) Representative images of lungs at T = 0 (Beginning) and 4 h after the addition (End) of supernatants collected from ExoY⁺-infected wild type, tau knockout, 1N4R, and 1N4R mutant tau-expressing PMVECs. The yellow arrows show lung damage caused by the phospho-tau-containing supernatant.

FIGURE 6

Endothelial cells release neurotoxic tau that inhibits synaptic plasticity. Summary plot of LTP. Field excitatory postsynaptic potential (fEPSP) slopes were normalized to those before the theta-burst stimulation (TBS; delivered at time 0) and plotted against time (mean ± SEM). Brain slices were untreated control (control) or treated with supernatants collected from wild-type (WT), tau knockout (KO), 1N4R overexpressing (1N4R), or Ser-214 to Ala mutant overexpressing (S214A) endothelial cells exposed to ExoY⁺. The responses from the control and KO treatment were significantly higher than the other groups (F(4,52) = 14.29; * indicates p ≤ .0001; ANOVA). The bottom shows representative averages of five traces obtained from timepoints “a” (pre-TBS; black), “b” (5–8 min post-TBS; blue), and “c” (57–60 min post-TBS; red). X-axis scale bar represents 2 ms and Y-axis scale bar represents 0.1 mV.

FIGURE 7

Phosphorylation of tau is required for inhibition of LTP. (A) PMVECs expressing the 1N4R form of endothelial tau were infected with ExoY⁺ bacteria, and supernatants were collected 6 h later. The untreated supernatant then was directly analyzed by immunoblot (Before) or was pretreated (After) with Lambda protein phosphatase (LPP; “+” = LPP treated sample, “−“= sample treated with reaction buffer alone) prior to being analyzed by immunoblot using either pan-tau antibody (Tau1) or Ser-214-tau antibody. LPP treatment abolished Ser-214 immunoreactivity. Molecular weights are in kDa. (B) Hippocampal LTP was abolished by supernatants collected following endothelial cell infection with ExoY⁺ bacteria. However, LPP treatment of the supernatant inactivated cytotoxicity of tau and rescued LTP to normal levels. The bottom shows representative averages of five traces obtained from timepoints “a” (pre-TBS; black), “b” (5–8 min post-TBS; blue), and “c” (57–60 min post-TBS; red). X-axis bar represents 2 ms and Y-axis bar represents 0.1 mV. * Denotes p ≤ .05 using ANOVA.

FIGURE 8

Ser-214 phosphorylation is required for endothelial tau-induced seeding of neuronal tau. (A) Quantitation of BiFC fluorescence following treatment of HEK293 cells with either HBSS (negative control; fluorescence level of these cells was set to 100%), membrane-permeant cyclic nucleotides (cNMP; positive control), or supernatant collected from wild-type (WT), tau knockout (KO), 1N4R tau-expressing (1N4R), or Ser-214 mutant (S214A)-expressing cells following ExoY⁺ infection (HBSS vs. WT **p ≤ .0043; HBSS vs. KO ns 0.8823; HBSS vs. 1N4R ****p ≤ .0001; HBSS vs. 1N4R-mutant ns 0.9997; HBSS vs. cNMP * 0.0494; N = 9). (B) Representative BiFC images of negative (Opti-MEM and HBSS) and positive (cAMP and cGMP) control cultures of HEK293 cells. (C) Representative BiFC images of HEK293 cells treated with supernatant collected from either untreated (Control) or ExoY⁺-infected (ExoY) wild-type (WT), tau knockout (KO), 1N4R, or Ser-214 mutant (S214A) tau-expressing cells.