Absence of TNF-α Promotes Anemia of Inflammation and Inflammatory Bowel Disease

Anemia of inflammation (AI) is a hallmark of a multitude of disorders including infections, autoimmune diseases and malignancies. The pathophysiology of AI is currently under investigation. While the hematological aspect of AI and the involvement of cytokines such as IL6, IL1α, IL1 β has been subject to extensive study, the inflammatory and immunological arm has only recently begun to be explored. This involves a complex interplay of cytokine mediators both pro and anti-inflammatory, mediating a broader innate and ultimately adaptive immune response that could potentially be linked to the hematological deficiencies seen under AI. In particular, highly elevated levels of the cytokine TNF a have been noted in inflammatory conditions, including an established acute mouse model of AI induced by a single injection of the heat killed pathogen Brucella abortus (BA). The latter results observed by us prompted us to furtherinvestigate the role of this cytokine in BA induced AI.

To study the above we injected mice lacking TNFα (TNFα^-/-) with BA and studied hematological, inflammatory and immunological changes in our model. Following a single injection of BA, we found that TNFα ^-/- mice develop an irreversible macrocytic, hyperchromic anemia with hemoglobin levels dropping as low as 7 g/dL within two weeks of BA injection and not returning to normal even after eight weeks unlike their WT counterparts. This is further characterized by a complete abolition of Bone Marrow (BM) erythropoiesis and extra medullary erythropoiesis in the spleen (SPL). We also found elevated serum levels of erythropoietin and iron, indicating that the resulting anemia is not caused by a lack of these factors. The anemia was accompanied by a fall in the lifespan of RBCs in peripheral blood and a concurrent increase in production of reactive oxygen species by progenitors and terminally differentiated RBCs. Moreover, our experiments using WT GFP⁺ RBCs in BA treated TNFα^-/- mice indicate that even the healthy GFP⁺ RBCs are cleared faster in the latter animals. Preliminary data, suggest that macrophages derived from BA treated TNFα^-/- express pro-inflammatory markers and that the same cells can consume more RBCs than macrophages from control animals, suggesting increased macrophagocytosis. Subsequently, we looked at changes in serum cytokine levels over time after injection of BA. We found that unlike WT animals injected with BA, TNFα^-/- mice show a sustained production of IFNγ, IL12-p40 and IL1β. We also found an expansion of splenic resident macrophages and T cells which are cell types implicated in the production of the cytokines mentioned previously. Finally, to extend the reach of this model to other disorders we treated TNFα^-/- mice with low doses of dextran sodium sulfate (DSS) and found that only these animals developed inflammatory bowel disease (IBD) characterized by blood in the feces and severe anemia, unlike WT animals treated the same way.

Our observations on the upregulation of IFNγ and IL12 is leading us to study macrophage activation and T helper cell polarization in the BM, SPL and lymph nodes. Moreover, we are also investigating the potential crosstalk between these effectors, cytokines and erythropoiesis looking at oxidative stress, survival and cell differentiation of erythroid progenitors. As a first step, we are investigating the recovery from BA of IFNγ^-/- mice, as TNFα^-/- mice with AI show a significant upregulation of IFNγ. Preliminary complete blood count of BA treated IFNγ ^-/- mice indicates that these animals do recover from an inflammatory insult and the anemia is reversible, unlike the TNFα^-/- mice. These results suggest that absence of IFNγ could potentially rescue the severely impaired phenotype that we found in our TNFα^-/- mice and the recovery in BA treated TNFα/IFNγ double knock out mice.

Given the history of TNFα in enabling inflammatory reactions and promoting cell death, the view of this cytokine in supporting recover under AI as inferred from our results is relatively novel and suggest caution in the use of TNFα inhibitors. Furthermore, our data point to an important role in the crosstalk between TNFα and INFγ and in their effect on erythropoiesis.