Restriction of iron availability to pathogens: mechanisms and contribution to the host antimicrobial defense – RESTRICTIRON

To better understand the role of activin B in sepsis and determine whether activin B is necessary to limit the proliferation of extracellular pathogens, we challenged Inhbb-/- mice (deficient for activin B) with LPS or E. coli, and assessed the bacterial burden and the expression of pro or anti-inflammatory cytokines and of molecules involved in iron homeostasis. We also assessed the level of phosphorylation of the Smad1/5/8 effectors after infection. We made the same analyses on Il6-/- mice.
To determine whether elevated serum activin B is a marker of bacterial persistence in the host, we infected mice with different bacteria: E. coli (Gram – extracellular pathogen), Salmonella enterica typhimurium (Gram – intracellular pathogen), Staphylococcus aureus (Gram + extracellular pathogen), and quantified their hepatic Inhbb expression compared to non-infected mice. We tested different commercial anti-activin B antibodies and two polyclonal rabbit anti-mouse activin B antibodies we produced. We also established a collaboration to generate monoclonal antibodies using our Inhbb-/- mice for immunization.
To better understand the importance of ferroportin mRNA downregulation in the host antimicrobial defense, we first determined when, in the course of an infection, ferroportin mRNA starts to decrease, and for how long. To explore the molecular mechanisms leading to the mRNA downregulation of ferroportin, we examined the possibility that HIF-2a is involved in its mRNA downregulation and compared the level of Slc40a1 repression between Irp1-/- and WT mice challenged with LPS. We also explored the role of Il6 on this repression.
To test whether administration of activin B or early downregulation of ferroportin mRNA can reduce bacteria proliferation in wild-type and in iron-loaded hosts, we first determined how deleterious for the host is hypersideremia in our models and infected WT and severely iron-loaded mice (Bmp6-/-) with wild-type or siderophore-mutants E. coli strains.

We challenged Inhbb-/- mice (deficient for activin B) with LPS or E. coli. Deletion of activin B did not alter the bacterial burden of the mice nor the expression of pro or anti-inflammatory cytokines, or of molecules involved in iron homeostasis. It did not affect the hepcidin expression or the level of Smad1/5/8 effectors phosphorylation, 2 and 4 hours after infection. Experiments carried on Il6-/- mice showed Il6 may be important for early induction of hepcidin, but its effect is taken over by the increase in Smad1/5/8 signaling which is sustained up to 28 hours. Activin B seems dispensable for hepcidin induction
We have infected mice with different bacteria: E. coli, Salmonella enterica typhimurium and Staphylococcus aureus. Inhbb expression is induced in the liver of WT mice infected with each of these pathogens. Inhbb expression remains high in the mice that are going to die from sepsis (E. coli) but normalizes in those that are going to survive (S. aureus).
We established a collaboration to generate monoclonal antibodies using our Inhbb-/- mice for immunization. 12 clones have been obtained and our collaborators are working to obtain an ELISA assay in humans.
We studied the kinetics of ferroportin (Slc40a1) mRNA downregulation after infection with E. coli, Staphylococcus aureus, and Salmonella enterica typhimurium. Slc40a1 mRNA remains repressed in E. coli infection until the mice die. The repression is transient in Staphylococcus infection where the mice recover and absent in the liver of Salmonella-infected mice. Our data suggest that the hypoxia factor Hif2a and the inflammatory cytokine Il6 both play a role in Slc40a1 mRNA repression.
We infected WT and iron-loaded mice (Bmp6-/-) with wild-type or E. coli strains with mutant siderophores. Iron-loaded Bmp6-/- mice are more sensitive to E. coli virulence (wild-type strain) than WT mice. The virulence of mutant E. coli strain is increased in Bmp6-/- mice that are iron over-loaded, compared with WT mice.

We have shown that an increase in activin B mRNA expression occurs in the course of infection with different pathogens. Inhbb seems to stay high in mice with bad prognosis, so it would be important to measure activin B in the serum of these mice and confirm its predictive value. All the commercial anti-activin B antibodies we tested were unfortunately not specific (similar band in WT and Inhbb-/- westerns). We have thus established a collaboration with Pr. O. Ritvos and Dr. G. Savjani to generate monoclonal antibodies using our Inhbb-/- mice for immunization. 12 clones have been obtained so far and will be used to design an effective and specific ELISA assay in humans. We will also test these antibodies in western and immunohistochemistry assays in the mouse.
We observed an induction of hepcidin even in Inhbb-/- mice (deficient for activin B). Interestingly, there is a concomitant activation of Smad1/5/8 signaling, but the identity of the ligand that activates this signaling pathway in the absence of activing B remains unknown. Our in vitro studies suggest a possible role of TGFb. To test this hypothesis in vivo we plan to inhibit TGFb type I receptor Alk5. Immunohistochemistry is being currently performed to investigate which cells (hepatocytes, macrophages…) express P-Stat3 in E. coli-infected Il6-/- mice and whether this is relevant to hepcidin production.
Our data suggested a probable role of Hif2a and of Il6 on the repression of ferroportin in the course of E. coli infection. We will analyze further the exact molecular mechanisms leading to this repression.
We infected WT and iron-loaded mice (Bmp6-/-) with wild-type or mutant siderophores E. coli strains. Interestingly, the iron-loaded Bmp6-/- mice are more sensitive to E. coli virulence (wild-type strain) than WT mice. Moreover, the attenuated virulence of mutant siderophores E. coli strain is increased in Bmp6-/- mice compared with WT mice. This confirms that hypersideremia is deleterious for the host.

Gineste A, Martin P, Oswald E, Coppin H, Roth MP. Evidence for IL-6/STAT3-independent induction of lipocalin-2 in the liver of mice infected with Escherichia coli. Hepatology 2015 Apr 10. doi: 10.1002/hep.27845. [Epub ahead of print]
Gineste A, Besson-Fournier C, Latour C, Martin P, Oswald E, Coppin H, Roth MP. Interleukine 6 is dispensable for hepcidin induction by Escherichia coli and Staphylococcus aureus. European Iron Club, Verona, September 2014 (first price for oral presentation)

Eukaryotic cells and most prokaryotic organisms require iron in order to maintain essential biological functions, including oxygen transport, energy metabolism and DNA synthesis. Not surprisingly, intense competition for iron between host and pathogen occurs during the course of infectious diseases, and the availability of this metal can have a significant impact on both pathogen virulence and host anti-microbial defenses. Several recent reviews have discussed the mechanisms used by microbes to acquire iron during infection of higher organisms, and the importance of this process in the expression of virulence characteristics. On the other side, the mammalian immune system manipulates iron levels in order to limit the multiplication of pathogen organisms. Although our group has recently demonstrated the importance of activin B in this process, the exact iron-sequestering host mechanisms are still imperfectly known. Our first objectives are to determine the importance for the host antimicrobial defense of inducing activin B and hepcidin on one side, and of repressing ferroportin mRNA on the other, and the contribution of these two mechanisms to the restriction of iron not only in the serum but also in the phagosomes. We will use Escherichia coli sepsis and Salmonella infection of mice as experimental models. New tools to monitor bacterial persistence will be developed and we expect to identify new therapeutic targets that we will modulate to reduce bacterial proliferation in infected hosts.

Sepsis occurs in more than 75,000 patients in France annually and is responsible for 30,000 deaths. It is characterized by an initial “cytokine storm” with high plasma levels of proinflammatory cytokines and chemokines, clinical signs of fever, tachycardia, and tachypnea. Improved clinical management has led to survival of the majority of patients in this early period. However, those who survive early sepsis often develop nocosomial infections with organisms not typically pathogenic in immunocompetent hosts because early hyperinflammatory state evolves to a subsequent hypoinflammatory state with significant immunosuppression. The reasons for these cellular changes are still obscure. In another component of the project, we will investigate whether manipulating host iron levels or modulating the relative expression of activins A and B help avoiding immunosuppression in sepsis and its deleterious consequences for the host.

The two partners involved in this project have complementary expertise to investigate host-pathogen interactions, particularly in the regulation of systemic iron metabolism and Salmonella infection (partner 1) and in cellular microbiology and molecular biology of pathogenic Escherichia coli (partner 2). They have been collaborating for the last year on the project and this allowed them to gather the preliminary data needed to construct this grant proposal. Their joined efforts will be essential to decrypt and improve the host strategies to restrict iron availability to both extracellular and intracellular pathogens.