III-05 Salma Bahnasawy

A pharmacometric model on cytokine changes upon live Escherichia coli exposure in a porcine sepsis model

Salma M. Bahnasawy (1), Paul Skorup (2), Katja Hanslin (3), Miklós Lipcsey (4), Lena E. Friberg (1), Elisabet I. Nielsen (1)

(1) Department of Pharmacy, Uppsala University, Uppsala, Sweden (2) Section of Infectious Diseases, Department of Medical Sciences, Uppsala University, Uppsala, Sweden (3) Anesthesiology and Intensive Care, Department of Surgical Sciences, Uppsala University, Uppsala, Sweden (4) Hedenstierna laboratory, Anesthesiology & Intensive Care, Department of Surgical Sciences, Uppsala University, Uppsala, Sweden

Objectives: Effective sepsis management is challenged by sepsis heterogeneity and the lack of clear understanding of the time course of its progression. Describing the kinetics of different cytokines involved as biomarkers of sepsis progression could help to optimise the medical intervention in septic patients. Thorsted et al.1 developed a model that quantifies the relationship between endotoxin (ETX) exposure, a Gram-negative bacteria outer membrane component, and the changes of two cytokines (TNF-α, IL-6). However, there is a lack of knowledge on how the host response could differ upon exposure to intact live bacteria. We aimed is to develop a model describing the cytokine changes triggered by exposure to intact live Escherichia coli. The specific aims were to; i) expand the previous model to describe the kinetics of E. coli in vivo, ii) link the bacterial kinetics to the previously quantified ETX-cytokine relationships and evaluate the need to recalibrate these relationships, iii) evaluate the model by applying it to external literature data, iv) use the model to explore predicted cytokine changes at different bacterial exposure scenarios.

Methods: The data came from three previously published in vivo studies using a porcine sepsis model.2–4 The piglets received a three-hour constant infusion of live E. coli with a total dose of 5*10^8 CFU, and samples of blood bacterial count, ETX, TNF-α, and IL-6. Models were fit to the different dependent variables sequentially. Thorsted’s model was extended to describe the E. coli-ETX relationship, then applied to the cytokines data while fixing the relevant parameters to test the need to update the model structure. Upon getting acceptable cytokines predictions, the parameters were allowed to be re-estimated by using informed priors. The model was applied to extracted literature data5–7 on a similar porcine sepsis model where the animals received increasing infusions of live E. coli with a total dose of 5*10^8 CFU. Model simulations were performed to explore the sensitivity of the host response to the bacterial exposure profile at the same bacterial burden. The total bacterial burden was fixed to  CFU, received as a continuous infusion, with different infusion scenarios; i) 3-h infusion with hourly dose up-titration (two scenarios; 5%, 15%, then 80% , and 20%, 30%, then 50% of total dose), ii) Constant infusion (two scenarios; 3-h, 6-h), iii) 3-h infusion with hourly dose down-titration (two scenarios; 50%, 30%, then 20%, and 80%, 15%, then 5% of total dose).

Results: The analysis included 30 animals with 645 observations and the model consisted of 11 compartments. In the proposed model, ETX is released upon bacterial elimination and triggers cytokine changes. The blood bacterial count was described by a one-compartment model with linear elimination (Cl=152 L/h, V=7.19 L). A scaling factor was estimated to quantify the ETX release per bacterial CFU (0.000079 EU/CFU). The changes in TNF-α concentration were driven by ETX levels using a sigmoidal model. The was increasing over time as a function of ETX levels to accommodate for ETX tolerance development. The ETX levels along with changes in TNF-α concentration stimulate IL-6 production. Transit compartment model was implemented to capture the delays in tolerance development and IL-6 changes. The model successfully described the cytokine changes where ETX was the main driver of these changes. There was no need to modify the ETX-cytokines model structure. Overlay plots of model predictions and external data showed a good description of the data. Exposure to a low initial bacterial burden (i.e. 5%, 15%, then 80% scenario) triggered the lowest cytokine changes. Inversely, the highest cytokine changes were observed with exposure to the highest initial bacterial burden (80%, 15%, then 5% scenarios).

Conclusions: The model predicted the cytokine changes triggered by exposure to intact live E. coli. The ETX released by bacteria was the main driver of cytokine changes indicating the lack of significant difference in cytokine changes when triggered by E. coli or ETX exposure. The model had the flexibility to capture the changes in cytokines under different bacterial exposure scenarios for the external data. The cytokine changes were found to vary based on the bacterial exposure scenario not only on the total bacterial burden. A next step would be to expand the model by including antibiotic treatment and study the host-pathogen-drug interaction which could be valuable for optimising antimicrobial therapy in sepsis.

References:

  1. Thorsted, A. et al. A non-linear mixed effect model for innate immune response: In vivo kinetics of endotoxin and its induction of the cytokines tumor necrosis factor alpha and interleukin-6. PLOS ONE 14, e0211981 (2019).
  2. Skorup, P. et al. Dynamics of Endotoxin, Inflammatory Variables, and Organ Dysfunction After Treatment With Antibiotics in an Escherichia coli Porcine Intensive Care Sepsis Model. Critical Care Medicine 46, e634 (2018).
  3. Hanslin, K. et al. The impact of the systemic inflammatory response on hepatic bacterial elimination in experimental abdominal sepsis. Intensive Care Med Exp 7, (2019).
  4. Skorup, P., Maudsdotter, L., Lipcsey, M., Larsson, A. & Sjölin, J. Mode of bacterial killing affects the inflammatory response and associated organ dysfunctions in a porcine E. coli intensive care sepsis model. Critical Care 24, 646 (2020).
  5. Thorgersen, E. B. et al. CD14 inhibition efficiently attenuates early inflammatory and hemostatic responses in Escherichia coli sepsis in pigs. The FASEB Journal 24, 712–722 (2010).
  6. Barratt-Due, A. et al. Polyvalent immunoglobulin significantly attenuated the formation of IL-1β in Escherichia coli-induced sepsis in pigs. Immunobiology 218, 683–689 (2013).
  7. Barratt-Due, A. et al. Combined Inhibition of Complement (C5) and CD14 Markedly Attenuates Inflammation, Thrombogenicity, and Hemodynamic Changes in Porcine Sepsis. The Journal of Immunology 191, 819–827 (2013).

Reference: PAGE 30 (2022) Abstr 9967 [www.page-meeting.org/?abstract=9967]

Poster: Drug/Disease Modelling - Infection