Challenging the concept of immunothrombosis

Raghavendra Palankar and Andreas Greinacher

In this issue of Blood, Beristain-Covarrubias et al elegantly reveal that infection-driven thrombosis induced by Salmonella Typhimurium (STm) in mice is monocyte dependent and follows organ-specific kinetics.1

(A) Onset of immunothrombosis in mice infected with STm shows up early in spleen (24 hours) and quickly subsides, whereas in liver, it emerges at a much later time point (day 7). (B) These thrombi are rich in platelet aggregates (CD41), monocytes (Ly6G+ and F4/80+), and neutrophils (myeloperoxidase [MPO] and citrullinated histone H3 [H3Cit]) and resemble the typical immunothrombosis event. Surprisingly, these thrombi contain very few bacteria, even in the presence of high bacterial burden in spleen and liver, thus raising questions on the notion that immunothrombosis is primarily a protective response against systemic bacterial infection. On the contrary, (C) in vitro, significantly high number of bacteria were captured upon platelet aggregation induced by STm.

Thrombi first occur in the spleen where they are dissolved within 1 to 2 days. Thrombi are found in the liver several days later. This is surprising because the bacterial load is comparable in both organs. Because these thrombi contain only a few bacteria despite a high bacterial burden in the organs themselves, the authors conclude that immunothrombosis per se may not facilitate bacterial containment during systemic infection (see figure). Potentially, platelets are more relevant for the capture of STm, because the bacteria were found in platelet aggregates induced by STm in vitro in platelet-rich plasma.

Sepsis, a systemic infection by pathogenic bacteria, elicits inflammation and can trigger (uncontrolled) thrombosis referred to as immunothrombosis (or thromboinflammation).2 These processes may rapidly manifest as multiple organ failure with life-threatening complications.3 During systemic bacterial infection, monocytes and neutrophils as well as platelets respond to pathogen-associated molecular patterns or damage-associated molecular patterns.4 In addition, platelets are directly activated by secreted bacterial products or bind indirectly through binding of plasma proteins to bacterial surface proteins, which results in rapid activation. Thus, platelets, neutrophils, and monocytes cooperatively initiate and amplify blood coagulation. Current opinion suggests that the thrombi triggered by bacterial infections provide protection against invading pathogens.5 Mechanistically, this is thought to be achieved by (1) limiting bacterial dissemination by their containment within thrombi, (2) forming protective barriers of fibrin network that prevent bacterial movement in and out of the blood vessels, thereby (3) coordinating concentrated antimicrobial cellular immune responses in and around the thrombi where bacteria are localized.6 However, there are still major gaps in our understanding of the complex interaction networks between host cells and different pathogens. The study by Beristain-Covarrubias et al adds another layer of complexity, which shows that these interactions between hemostasis and pathogens lead to immunothrombosis that occurs with different kinetics and dynamics at the organ level.

Previously, a mouse model of systemic STm infection revealed that TLR4-specific and interferon-γ–dependent inflammation drives upregulation of podoplanin on macrophages derived from circulating monocytes in the liver. Lesions in the vessel wall expose podoplanin to the C-type lectin-like receptor-2 (CLEC-2) on platelets thereby initiating thrombus formation.7 The thrombi in the liver appeared a week after the infection began. However, despite the spleen’s having a bacterial burden similar to that in the liver, histologic investigation of the spleen showed no apparent thrombi.

These intriguing observations prompted Beristain-Covarrubias et al to take a second look at both spleen and liver, and they showed that thrombi also occur in the spleen but at a much earlier time point. Interestingly, the thrombi captured very few bacteria in vivo. Conversely, platelets that aggregated in vitro showed close association with STm. These findings challenge the concept that immunothrombosis provides a fortified cellular and fibrillary framework to capture and contain bacteria in circulation. In support of this argument, there is some evidence to suggest that, in humans, Salmonella Typhi is able to activate both the coagulation system and fibrinolytic pathways.8 This may explain why the authors observed few or negligible numbers of STm in the thrombi in vivo and why the thrombi dissolve within days in the spleen. The reason for the delayed thrombus formation in the liver might be the production of natural anticoagulant proteins in high concentrations within the liver, such as antithrombin, protein C, and protein S, which prevent formation of thrombi until these proteins are depleted as a result of bacteria-induced liver cell damage. This would nicely link the recent concept of shock liver (ischemic hepatitis)–induced venous thrombosis of the microcirculation with sepsis-associated microthrombosis9 (ie, the risk of microthrombosis might be primarily related to the concomitant impairment of antithrombin and the protein C pathway because of liver cell injury).

Additional studies are currently underway to identify organ-specific markers of infection in model animals that will help address the role of the tissue-specific microenvironment in bacterial host defense.10 Further insight into host defense mechanisms governing immunothrombosis may help establish potentially useful organ-specific targets for therapeutic interventions in sepsis.


  • Conflict-of-interest disclosure: The authors declare no competing financial interests.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.