Skip to main content
European Commission logo print header

The Role of Granzymes A, B and M in Sepsis

Final Report Summary - GRANZYMES IN SEPSIS (The Role of Granzymes A, B and M in Sepsis)


Aim of the study:

The inflammatory process is necessary for adequate containment and elimination of pathogens. However, when this response is inadequately regulated, systemic manifestations can result in pathophysiologic changes, such as observed in sepsis. Granzymes (gzms) are serine-proteases produced mostly by cytotoxic lymphocytes. Studies on gzms have been mainly focused on their cytotoxic role. However, growing evidence points to a role in inflammation and the immune response.
Based on previous findings we hypothesized that gzms play an essential role in the host response to bacterial sepsis. To test this we proposed the following objectives:
● To determine which cells produce specific gzms during bacterial sepsis through a combined clinical (Intensive Care Unit patients with and without sepsis, and controls) and experimental (sepsis mouse models) approach.
● To determine which cells produce specific gzms during infection by an intracellular bacterium through a clinical approach (patients with pulmonary tuberculosis and controls).
● To determine the role of gzm A and gzm B in the host response during bacterial sepsis through studies in sepsis mouse models with granzyme-deficient mice.

Work performed:

To determine the cells that produce gzms during bacterial sepsis in humans, blood from 32 healthy controls, 23 patients with sepsis and 11 patients with systemic inflammatory response syndrome (SIRS) was analyzed by flow cytometry. The percentage of different lymphocyte populations (cytotoxic lymphocytes, NK (natural killer) T cells, γδ T cells and NK cells) expressing gzm A, B, M and K was measured, as well as the median fluorescence intensity (MFI) of each gzm in positive cells.
To determine the cells that produce gzms during pulmonary tuberculosis in humans, blood from 12 healthy controls and 18 patients was also analyzed by flow cytometry, measuring the percentage of the same lymphocyte populations expressing gzm A, B and K, and the MFI. In addition, extracellular levels of gzm A and B were measured in plasma from 34 healthy controls and 58 patients by ELISA.
The expression of gzm A and B was analyzed similarly by flow cytometry in mouse models of the two most common causes of sepsis in wild-type (WT) mice. For that, mice were analyzed before infection and at early, intermediate and late time-points after infection with Klebsiella pneumoniae (pneumonia) or E. coli (peritonitis). Additionally, mRNA levels of gzm were measured by qRT-PCR in lung homogenate (pneumonia) or peritoneal lavage fluid (PLF, peritonitis).
To determine the role of gzm A and B in the host response during bacterial sepsis through mouse models, mice deficient for gzm A, gzm B or gzm A+B were compared with WT mice with regard to key host responses during pneumonia, peritonitis or primary sepsis. For the pneumonia model mice were intranasally infected by K. pneumoniae (gram-negative) or S. pneumoniae (gram-positive), for peritonitis mice were intraperitoneally inoculated with E. coli, and for primary sepsis mice were intravenously infected with K. pneumoniae. Mice were sacrificed at early, intermediate and late time-points of the disease and blood and organs (lung, spleen and liver), and PLF in peritonitis, were collected for the following measurements: bacterial loads (quantitative cultures in lung, spleen and liver homogenates, blood and PLF); histopathology of the lungs and liver, and kidney in peritonitis (semi-quantitative scores to obtain insight in the extent of injury and inflammation); influx of neutrophils to the primary site of infection (leukocyte counts and differentials in PLF, myeloperoxidase levels in lung homogenates and semi-quantitative Ly6 staining of lung tissue slides); cytokine (TNFα, IL-1β, IL-6, IL-10, IL-12, IFNγ) and chemokine (MCP-1, MIP-2, KC) levels in lung homogenates, plasma and PLF; and clinical chemistry to obtain insight into the extent of distant organ injury (plasma levels of ASAT, ALAT and LDH, and also creatinin and urea in peritonitis). For the pneumonia model with K. pneumoniae a survival study was also done.
In addition, wt mice were depleted of NK and NKT cells, infected intranasally with K. pneumoniae and sacrificed at intermediate and late time-points. Blood and organs were harvested and bacterial loads, histopathology, cytokines and chemokines, and clinical chemistry were measured.

Main results and potential impact:

The percentage of cells expressing gzms was similar in patients with sepsis and SIRS, but lower than in healthy controls. Similar results were found for MFI, although gzm B presented not significantly higher levels than in controls. This led us to think that the difference with regard to controls is caused by the inflammatory process, not by the bacterial infection per se. The same results were observed in the group of sepsis caused by pneumonia. However, the patients with tuberculosis showed higher percentage and MFI, as well as extracellular levels of gzms, than controls. This differs from the results in sepsis and SIRS, and seems to be in line with results observed during viral infections. The increase observed in percentage and MFI of gzm A, but decrease of gzm B MFI, at a late time-point of infection with K. pneumoniae in wt mice is in consonance with the results in septic patients. In addition, the pneumonia model showed increased pathology scores in gzm B and AB KO mice and increased pro-inflammatory cytokine levels in gzm AB KO mice, suggesting that the presence of gzm B is attenuating lung inflammation, strengthened by gzm A. In line with those results, the primary sepsis model by K. pneumoniae showed higher inflammation and necrosis in liver of gzm B KO mice, as well as higher levels of IFNγ.
The peritonitis model in wt mice showed lower percentage of NK cells positive for gzms in PLF compared to the high levels present in blood and spleen. Gzms KO mice, mainly gzm B, showed higher renal damage measured by histopathology and by creatinin and urea levels in plasma. However, higher bacterial loads were found at late time-point in gzm B and AB KO mice, which can be contributing to some of these results.
Overall our results point to a protective role for gzms in host defense against severe infection. This project provides information on the regulation of gzms production during sepsis/severe bacterial infection, which sheds light on the role of gzms in inflammation and can contribute to the development of future innovative treatments and methods of disease management. This is relevant since sepsis is a major public health problem that is becoming more common and whose attributed mortality rates are rising, being not only a problem at human level, but upsetting the costs of healthcare services. Thereby, any progress in this field is in the spotlight worldwide.

Researcher: M. Isabel García-Laorden (
Scientist in charge: Prof. Tom van der Poll (
Center for Experimental and Molecular Medicine. Academish Medish Centrum. Amsterdam, Netherlands.