In this issue of the JCI, Mottis and colleagues take these observations further to demonstrate that tetracyclines also induce disease tolerance (Figure 1) to influenza virus (IFV) infections (17). Using cellular and germ-free mouse models, where tetracyclines decrease oxidative phosphorylation complex activity and ATP concentrations, the authors demonstrated that the prototypical four-ringed tetracycline doxycycline caused a mild mitochondrial stress response that included both type I IFN signaling and an activating transcription factor 4–mediated (ATF4-mediated) integrated stress response (ISR). Doxycycline caused distinct transcriptional responses in the kidney and liver. While the kidney responded with a transcriptional signature pointing to the activation of the ATF4/ISR pathway, including characteristic translation inhibition, the liver induced a type I IFN response with increased expression of ISGs. Using bone marrow–derived macrophages (BMDMs), the authors identified the release of mtDNA from mitochondria, following its perturbation by doxycycline, as the likely trigger for the initiation of the type I IFN response.

Tetracycline derivatives with minimal antimicrobial activity have increased capacity to induce an adaptive mitochondrial stress response and enhance disease tolerance. (A) Doxycycline, a prototypical tetracycline antibiotic, blocks bacterial and mitochondrial translation, inducing mild proteotoxic mitochondrial stress, which initiates mitochondrial stress responses. (B) 9-tert-Butyl doxycycline (9-TB), a doxycycline derivative with a substitution at the C9 position, has minimal antimicrobial activity but shows substantially greater capacity to induce the UPRmt and mitochondrial stress response (MSR) when compared with parental doxycycline. (C) Mottis et al. (17) showed that both parental doxycycline and 9-TB improved survival of mice in a model of lethal influenza virus infection, by reducing tissue damage but without affecting viral titers. This finding demonstrates that in addition to the antimicrobial properties of tetracyclines (known as resistance), the effect of this class of antibiotics on the host mitochondria triggers disease tolerance mechanisms in viral infections through activation of MSRs.

To avoid the antibacterial effects of tetracyclines on the host microbiome, Mottis et al. (17) identified several derivatives with minimal antimicrobial activity. In particular, a derivative with a substitution at the C9 position, 9-tert-butyl doxycycline (9-TB), retained, and in fact substantially superseded, the effects of the parental doxycycline on induction of mitochondrial unfolded protein response (UPRmt). 9-TB was also much more potent than doxycycline at inducing a mitochondrial stress response (as measured by the capacity to affect mitonuclear protein imbalance) and inducing ISGs in BMDMs. Doxycycline and 9-TB were effective at increasing survival in a lethal IFV infection model when given preventively. They did not affect the viral titers, a fact that points to their capacity to induce disease tolerance, not resistance mechanisms. Critically, while doxycycline affected the gut microbiome as expected, decreasing its bacterial species diversity, 9-TB did not affect the microbiome composition. Moreover, 9-TB was also capable of decreasing the severity of infection and delaying mortality when administered therapeutically (Figure 1). The authors further showed that disease tolerance to IFV infection correlated with the induction of genes associated with lung epithelia and cilia function. In addition, 9-TB (to a greater extent than doxycycline) downregulated genes with roles in inflammatory and immune responses in the lung, liver, and kidney, possibly limiting tissue damage resulting from an excessive inflammatory response to infection. These findings agree with our demonstration that RAbos impair T cell effector function and ameliorate autoimmunity by blocking mitochondrial protein synthesis (18) because T cells may often cause collateral tissue damage.

Going forward, many exciting questions remain. One is how core cellular function perturbations leading to resistance or disease tolerance are sensed. In the case of tetracyclines it is tempting to speculate that inhibiting mitochondrial protein synthesis perturbs the ETC and decreases ATP concentration, potentially leading to the initiation of UPRmt. This possibility is based on the recent finding that ATP is a strong hydrotrope with the ability to prevent the formation of, and dissolve already formed, protein aggregates (19). Alternatively, tetracycline-induced inhibition of mitochondrial protein synthesis may cause an altered stoichiometry of the ETC complex components that are encoded by nuclei and mitochondria, constituting a signal that is sensed and transduced by unknown factors. A second category of questions will emerge from the systematic investigation of the types of infections that may benefit from the disease tolerance–inducing properties of tetracyclines. Different groups of pathogens impose specific types of tissue damage and are antagonized by appropriate nonoverlapping immune effector responses. Each one of these effector mechanisms comes with its own specific immunopathology and requires unique disease tolerance processes to resolve each pathogen-specific type of tissue damage. Of course, tetracyclines are just the tip of the iceberg. Many other classes of immunomodulatory antibiotics cause their own types of perturbations to cellular processes and organelles. The mechanistic study of their effects is likely to reveal fundamental biological insights into the regulation of organismal homeostasis by stress responses. This knowledge may allow us to harness antibiotic effects for therapeutic strategies against infection and other conditions that progress with inflammation and substantial tissue damage and loss of function, including autoimmune, neurodegenerative, and cardiovascular diseases.