Effect brain irradiation with X-ray/Proton beam on circulating immune cell in rodents: A longitudinal modeling approach
Thao-Nguyen Pham (1,2), Julie Coupey (1), Jérome Toutain (1), Serge M. Candeias (3), Gaël Simonin (4), Marc Rousseau (4), Omar Touzani (1), Juliette Thariat (2)*, Samuel Valable (1)* (*Authors equally contributed)
(1) Normandie Univ, UNICAEN, CNRS, ISTCT, GIP Cyceron, Caen, France, (2) Laboratoire de physique corpusculaire UMR6534 IN2P3/ENSICAEN, France - Normandie Université, France, (2) Univ. Grenoble Alpes, CEA, CNRS, IRIG-LCBM-UMR5249, 38054, Grenoble, France, (4) CNRS, IPHC, UMR 7178, Strasbourg University, 67200, Strasbourg, France
Objectives
Leucocyte can influence tumor growth and response to radiotherapy, one of the standard therapies in brain tumors. While radiation-induced lymphopenia is being studied extensively, radiation effects on lymphoid and myeloid subtypes was relatively less addressed [1]. This project aims at determining the influence of irradiation parameters both on acute effects and the subsequent recovery of leucocyte subpopulations following brain irradiation in rodents using modeling.
Methods
Leucocyte subpopulations (LS) concentration (T CD4, T CD8, B, NK cell, neutrophil and monocyte) after brain irradiation with either X-ray (XR) or proton (PT) beam in mice were obtained. Mice (n=96) were irradiated for whole-brain or hemispheric irradiation at 1 or 2 Gy/min twice a day on 4 consecutive days until 20 Gy. Blood samples were withdrawn at day 0, 3, 7, 14, 21 and 28 after irradiation. Relative concentration over control group at each time point was calculated.
Relationship between radiation parameters and LS during irradiation were first investigated. Physiology-based parameters were estimated based on the hypothesis that dose to the blood induces leucocyte variations. Then, relationship between the physiology-based parameters and LS levels were then investigated. Finally, network modeling was applied to investigate the relationship between LS under irradiation.
Following XR irradiation, the recovery kinetic of T (CD4, CD8) cell were identified through selecting the appropriate function using Akaike information criterion. For B cell, a mechanistic modeling approach was also applied based on prior knowledge of lymphocyte physiology. Individual-based modeling was then applied in assuming inter-individual heterogenous baseline using nlminxr version 2.0.8 [2]. For myeloid subpopulations and lymphoid subpopulations following PT irradiation, k-means clustering analysis was applied to identify the main kinetic trend.
Interplay between LS was described in a Bayesian network. Model structure learning was based on minimized Bayesian information criterion.
Modeling, simulation and data analysis were applied in RStudio version 4.1.2 [3].
Results
Particle had most impact on radiation-induced lymphopenia in all subpopulations, followed by spatial coverage. Besides, particle also impacted irradiated area size and interactions. Using vertical XR irradiation of mice brain, the chin and neck (which contain supplying arteries) were also included in the irradiated area. Myeloid population was almost insensitive to XR, which is in line with reported knowledge of radioresistance of this population.
T cell kinetic was best described by negative exponential function and was able to recover to their initial level within a month. For B cell, a two-compartment model was selected as optimal model structure. The predicted baseline value was approximate 1, thus B-cell level nearly reaches its initial level after completion of irradiation. NK cell, on the other hand, more than 95% of individuals’ NK cell level could not recover within 1 month after IR.
PT treatment induced little change in lymphocyte kinetic in all subpopulations. Data was grouped into 2 clusters: 95% of individual have a stable lymphoid subpopulations concentration following PT irradiation and only 1 mouse had an increase cell count. Thus, brain irradiation using PT beam do not impact lymphoid subpopulations.
For myeloid subpopulations, the majority of individual have little change in cell count by time.
Following XR irradiation, there was an inter-dependence in homeostatic regulation of T CD4+ and CD8+, which was strong early after irradiation then weaken by time. The interplay of innate immune cell occurred early after XR irradiation then disappeared later, which on the other hand occurred late after PT irradiation.
Conclusion
Joint use of radiation parameters into physiology-based parameters in a structured network model provides new insights on radiation effects on the immune system as well as the difference effect between XR and PT irradiation, including differential response and interactions between lymphoid and myeloid lineages. The conservative effect of PT irradiation compared to XR could be explained by the reduction of total blood flow in irradiated area and the mean dose delivered to the leucocyte exposed to radiation. Predicting leucocyte subpopulations kinetic following brain irradiation is critical for determining the optimizing the use of radiotherapy and immunotherapy combination.
References
[1] Pham, T. N., Coupey, J., Candeias, S. M., Ivanova, V., Valable, S., & Thariat, J. (2023). Beyond lymphopenia, unraveling radiation-induced leucocyte subpopulation kinetics and mechanisms through modeling approaches. Journal of experimental & clinical cancer research : CR, 42(1), 50. https://doi.org/10.1186/s13046-023-02621-4
[2] Fidler M, Xiong Y, Schoemaker R, Wilkins J, Trame M, Hooijmaijers R, Post T, Wang W (2022). nlmixr: Nonlinear Mixed Effects Models in Population Pharmacokinetics and Pharmacodynamics. R package version 2.0.8, https://CRAN.R-project.org/package=nlmixr
[3] RStudio Team (2020). RStudio: Integrated Development for R. RStudio, PBC, Boston, MA URL http://www.rstudio.com/