Evidence for dispensability of protein kinase R in host control of tuberculosis

Abstract

Genetic deficiency of protein kinase R (PKR) in mice was reported to enhance macrophage activation in vitro in response to interferon-γ (IFNγ) and to reduce the burden of Mycobacterium tuberculosis (Mtb) in vivo (Wu et al. PloS One. 2012 7:e30512). Consistent with this, treatment of wild-type (WT) macrophages in vitro with a novel PKR inhibitor (Bryk et al., Bioorg. Med. Chem. Lett. 2011 21:4108-4114) also enhanced IFN-γ-dependent macrophage activation (Wu et al. PloS One. 2012 7:e30512). Here we show that co-treatment with IFN-γ and a new PKR inhibitor identified herein to be highly but not completely selective likewise induced macrophages to produce more reactive nitrogen intermediates (RNI) and tumor necrosis factor alpha (TNF-α) and less interleukin 10 (IL-10) than seen with IFN-γ alone. Unexpectedly, however, this new PKR inhibitor had a comparable effect on PKR-deficient macrophages. Retrospective investigation revealed that the PKR-deficient mice in (Wu et al. PloS One. 2012 7:e30512) had not been backcrossed. On comparing genetically matched PKR-deficient and WT mice, we saw no impact of PKR deficiency on macrophage activation in vitro or during the course of Mtb infection in vivo. In addition, although 129S1/SvImJ macrophage responses to IFN-γ were greater than those of C57BL/6J macrophages, PKR was not required to mediate the IFN-γ-dependent production of IL-10, RNI or TNF-α in either strain. Together the data cast doubt on PKR as a potential therapeutic target for tuberculosis.

Keywords: IFN-γ; Macrophage; PKR; TNF-α; Tuberculosis.

Figure 1.. Compound 85 inhibits PKR in vitro and enhances macrophage activation with minimal toxicity.

A) Representative concentration-response curve for inhibition of human PKR by C85 performed with pure recombinant human PKR and 2-fold dilutions of C85. IC₅₀s were determined by nonlinear regression analysis fitting to a formula Y=Bottom + (Top-Bottom)/(1+10^((LogIC₅₀-X) *HillSlope)). Mean value for IC₅₀ from three independent experiments was determined to be 0.42+/−0.14 μM at final 10 μM ATP. Experiment was performed at variable ATP of 10 μM (red circles, 2.5 μM final in assay), 20 μM (black circles, 5 μM final in assay) or 40 μM (blue circles, 10 μM final in assay). B) Representative images of BMMs treated with indicated concentrations of C85 for 24 hrs at magnification 20X. The scale bar indicates 100 μ. The experiment was performed three times. C) Dose-response curve of C85 and its impact on IL-10 reduction, nitrite production and TNFα production by WT BMMs primed with IFNγ for 24 hrs. then treated with C85 for 48 hrs. NT = No treatment. Results in (C) are means ± SD pooled from two independent experiments with each condition performed in triplicate. p-value < 0.05 = *; < 0.01 = **; < 0.005 = ***; < 0.001 = **** by two-way ANOVA – Tukey’s multiple comparisons test.

Figure 2.. PKR does not play a role in the effect of C85 on activation of macrophages from C57BL/6J background.

A) C85 reduces IL-10 and increases nitrite and TNFα in response to IFNγ in WT (B6) macrophages. B) C85 reduces IL-10 and increases nitrite and TNFα in response to IFNγ in PKR-regulatory domain KO (B6) macrophages. C) C85 reduces IL-10 and increases nitrite and TNFα in response to IFNγ in PKR-kinase domain KO (B6) macrophages. NT = No treatment. V = Vehicle control. Results in (A - C) are means ± SD pooled from two independent experiments with each condition performed in triplicate. p-value < 0.05 = *; < 0.01 = **; < 0.005 = ***; < 0.001 = **** by two-way ANOVA – Tukey’s multiple comparisons test.

Figure 3.. Macrophages from 129S1/SvImJ mice show enhanced activation in response to IFNγ.

A) IFNγ does not induce the production of IL-10 but does induce the production of nitrite and TNFα in WT (129S1/SvImJ) macrophages. B) IFNγ does not induce the production of IL-10 but does induce the production of nitrite and TNFα in PKR regulatory domain KO (129S1/SvImJ) macrophages. NT = No treatment. V = Vehicle control. Results in (A) and (B) are means ± SD pooled from two independent experiments with each condition performed in triplicate. p-value < 0.05 = *; ns = non-significant by Student’s t-test.

Figure 4.. WT and PKR-deficient mice on the C57BL/6J background show similar Mtb burden.

A) CFU burden in the lung; B) CFU burden in the spleen. Numbers of mice (n) = 3 for day 1; n = 4 for day 14; n = 5 for days 29, 57 and 120. Data are means ± SEM performed once with the number of mice specified above.