Chemical p38 MAP kinase inhibition constrains tissue inflammation and improves antibiotic activity in Mycobacterium tuberculosis-infected mice

Host-modulating therapies have become an important focus in the development of novel concepts for improved management of tuberculosis (TB). Previous in vitro studies revealed that the p38 MAP kinase signaling pathway coordinates several inflammatory and stress responses in Mycobacterium tuberculosis (Mtb)-infected host cells. Here we extend these findings and show that in vivo treatment of Mtb-infected C57BL/6 mice with doramapimod, a p38 MAP-kinase inhibitor, results in reduced inflammation, granuloma formation and lung pathology. Moreover, doramapimod, together with standard antibiotic treatment, significantly reduced lung and spleen mycobacterial loads compared to antibiotic treatment alone. Our in vivo data suggest the opportunity to repurpose p38 MAPK inhibitors for adjunct host directed therapies. We also provide first data on safety of p38 MAPK inhibition which is of relevance for future application of these substances in inflammatory diseases and concomitant TB.

In continuation of our previous in vitro findings we consequently sought to determine the potential of p38 MAPK inhibition as an adjuvant treatment for TB and to determine the risk for exacerbation of the disease during monotherapy. To this end, we here analyzed the outcome of experimental TB in vivo under p38 MAPK inhibition with and without antibiotic treatment during acute and chronic infection of C57BL/6 mice.

Results
Doramapimod treatment reduces tissue inflammation in the C57BL/6 mouse model of acute Mtb infection. Doramapimod is a well-defined orally bioavailable p38 MAPK inhibitor with confirmed activity in both humans and mice 10 . One day after aerosol infection with Mtb H37Rv we started treatment of mice with 30 mg/kg doramapimod or vehicle. After 28 days of treatment, mice were sacrificed and evaluated for histopathological changes and bacterial load in the lungs. In doramapimod-treated mice, lung inflammatory scores such as peribronchitis and perivasculitis were significantly reduced (Fig. 1a,b) and alveolitis was lowered by trend (Fig. 1a). In addition, doramapimod treatment significantly reduced the number of granumlomas (occupied area of the lung section) in infected mice (Fig. 1a,c). Interestingly, a similar bacterial load was found in lungs of doramapimod-and vehicle-treated mice (Fig. 1d). Together, p38 MAPK inhibition during acute experimental TB limits inflammation in the lungs yet does not modulate mycobacterial loads.
p38 MAPK inhibition impacts Mtb-induced inflammatory immune responses in lungs of chronically infected C57BL/6 mice. To evaluate the impact of p38 MAPK inhibition on the inflammatory immune response, lung infiltration and bacterial loads on an already established disease, we started doramapimod treatment 28 days after aerosol infection of C57BL/6 mice with Mtb. Production of pro-inflammatory cytokines in lungs was determined 14 and 28 days after initiation of treatment using cytometric-Bead-Array of lung homogenates. On day 42 of infection, the concentration of IL-6, TNF, IL-12p40 and IFNγ were significantly reduced in doramapimod-treated animals ( Fig. 2a; white and black bars). Fourteen days later (day 56), the production of IL-1β, TNF, IL-17A and IL-12p40 remained significantly impaired in treated mice ( Fig. 2a; white and black bars). This diminished inflammatory cytokine release in the lungs of doramapimod-treated and Mtb-infected mice was accompanied by a strikingly attenuated pulmonary tissue inflammation and granulomatous response ( Fig. 2b; white and black circles, and Fig. 2c). Neither doramapimod nor antibiotic treatment had a profound effect on IL-10 production in infected lungs ( Supplementary Fig. 1). In contrast to the modulating effect on the inflammatory immune response during chronic Mtb infection, doramapimod treatment alone had no impact on bacterial loads in lungs and spleen ( Fig. 3; white and black circles). In summary, p38 MAPK inhibition during chronic experimental TB impairs pro-inflammatory immune responses in the lungs but does not affect mycobacterial growth.
Adjuvant doramapimod treatment positively affects the inflammatory response and reduces the mycobacterial load. We next analyzed the adjuvant effect of doramapimod treatment on the antiinflammatory and anti-mycobacterial effect of standard antibiotics. To this end, chronically infected C57BL/6 mice received on day 28 of infection doramapimod along with isoniazid and rifampicin and the outcome of experimental TB was compared to mice that received vehicle, doramapimod or antibiotics alone. Overall, treatment with antibiotics resulted in a pronounced decrease of cytokine production in the lungs of Mtb-infected animals ( Fig. 2a; grey bars). On day 42, the release of IL-6, and IL-12p40 was further reduced by addition of the p38 MAPK inhibitor to the antibiotic treatment ( Fig. 2a; grey bars). The initial reduction of cytokine levels in lungs of Mtb-infected mice treated with the combination of doramapimod, isoniazid and rifampicin were accompanied by a further reduction of the granulomatous response on day 56 of infection ( Fig. 2b; grey circles). Strikingly, this decreased tissue pathology was associated with significantly diminished bacterial loads in lungs and spleen of doramapimod, isoniazid and rifampicin-treated C57BL/6 mice (lung: 5.194 ± 0.201; spleen; 3.548 ± 0.213) when compared to animals that received antibiotics (lung: 5.399 ± 0.119; spleen: 3.837 ± 0.296) alone ( Fig. 3; grey circles). Together, p38 MAPK inhibition has an adjuvant impact on the anti-mycobacterial effect of antibiotic treated along with an additive impairment of pro-inflammatory immune responses and tissue pathology.

Discussion
Targeting signaling pathways which mediate hyperinflammation, necrosis and tissue damage represents a promising approach for adjunctive HDT in management of TB 3 . In this study we followed up on in vitro and ex vivo data in which we were able to link p38 MAPK to inflammation and necrotic cell death in Mtb infected host cells 7,11 . We now provide in vivo evidence, that p38 MAPK is a key signaling molecule in Mtb pathogenesis. Bacterial infections are well known to activate p38 MAPK either directly by secreted factors and components of the bacterial cell wall or indirectly through the release of pro-inflammatory cytokines like IL-1 or TNF from activated host cells 11 . Therefore, p38 MAPK plays an important role in coordinating the immune response of the host and is often targeted by pathogens to promote virulence and ensure pathogen survival 8 . Histopathological investigation of human biopsies revealed p38 MAPK phosphorylation in macrophages surrounding granulomas in TB patients, indicating that this kinase is engaged and may be a potential target for HDT in TB 12 . Chemical inhibition of p38 MAPK indeed reduced the inflammatory response and granuloma formation in Mtb infected C57BL/6 mice ( Figs. 1 and 2). Despite significant reduction of cytokines known to be essential for control of the disease in humans and animals, doramapimod treatment had no unfavourable effect on the bacterial load in both acute and chronic infection models of the disease as shown in this work. Similar effects were seen in our ex vivo assays in which doramapimod potently protected infected human macrophages from Mtb induced cell death Scientific RepoRtS | (2020) 10:13629 | https://doi.org/10.1038/s41598-020-70184-x www.nature.com/scientificreports/ without reducing the intracellular bacterial load 7 . The reason for this may be found in the multiple regulatory effects of this kinase, which not only involve cytokine release but also regulation of autophagy and induction of necrotic host cell death 7,13 . However, our findings stand in contrast to the effects seen with chemical inducers of autophagy such as metformin, an approved drug which is also promoted as a host-directed therapeutic in TB. Induction of autophagy using metformin has little effect on host cell survival but impairs intracellular bacterial growth 14 . Whether p38 MAPK positively or negatively affects autophagy in TB requires further investigations. www.nature.com/scientificreports/ Nevertheless, it is well-known that Mtb-driven activation of the p38 MAPK pathway leads to cyclooxygenase II accumulation, prostaglandin E2 and cyclic AMP release and ultimately matrix metalloproteinase-1 secretion which are known to play a pivotal role in Mtb pathogenesis and tissue destruction 12 . Thus, p38 MAPK inhibition in Mtb-infected tissue may generate an immunological microenvironment that is host protective despite the down-regulation of pro-inflammatory cytokines. Interestingly, this environment improves antibiotic activity as shown in our adjunctive treatment experiments with rifampicin and isoniazid (Fig. 3). Since p38 MAPK is a validated target in autoimmune and inflammatory diseases with several completed or ongoing clinical trials 8 , our findings have important implications for future clinical use of these drugs. We provide first in vivo data on safety of monotherapy with p38 MAPK inhibitors and concomitant Mtb infection. It is well known that inhibition of pro-inflammatory cytokines such as TNF in Mtb-infected mice results in exacerbation of the disease, increased inflammation, a higher bacterial burden and impaired survival 15 . In contrast, we observed no adverse effects after treatment with doramapimod and subsequent decrease in TNF levels indicating Colony enumeration assay. Bacterial loads in lungs and spleens were quantified at different time points after infection with Mtb as described earlier 16 . Briefly, organs were removed aseptically, weighed and homogenized. Tenfold serial dilutions of organ homogenates were then plated, incubated at 37 °C and colonies were counted after 21 days.
Histopathology. For histology, one lung lobe per mouse was fixed in 4% formalin-PBS, set in paraffin blocks, and sectioned (2-3 µm). Histopathological preparations were performed using standard protocols for  17 . Histopathological parameters peribronchiolitis, perivasculitis, alveolitis and granuloma formation were semiquantitatively scored as absent, slight, moderate, marked or strong, noted as 0, 1, 2, 3, 4, and 5, respectively. In this score the frequency as well as the severity of the lesions were incorporated. Granuloma formation was scored by estimating the occupied area of the lung section.
Quantification of cytokine production. The concentrations of cytokines in lung homogenates from uninfected and infected mice were determined by a Cytometric-Bead-Array (CBA) (BD Biosciences, Heidelberg, Germany) as described 16 . The quantity of cytokines per lung was calculated based on the ratio of lung to sample weight.
Statistical analysis. The significance of differences between parameters under different therapies was examined with unpaired T tests for Fig. 1 and one-way ANOVA and multiple comparisons for Figs. 2 and 3, where p values < 0.05 were considered significant.