Introduction

Cerebral malaria (CM), the most lethal complication of Plasmodium infection, leads to death in 15-20 % of the cases, mostly children under five in sub-saharan Africa. CM is often difficult to recognize in Plasmodium-infected populations because of concomitant diseases. Moreover, CM is frequently associated with multiple organ failure in adults, which may worsen neurological signs. A widely-used experimental model of pediatric cerebral malaria (ECM) is obtained by infecting susceptible mice with Plasmodium Berghei ANKA (PbA)¹. Recently, it was suggested that other organs than brain could be affected by PbA in ECMl². A better knowledge of this model is required to improve our understanding of ECM physiopathology in a context of multiple organ failure. In this pilot study, we investigated liver structure and metabolism during ECM by using an approach combining anatomical MRI, relaxometry and proton spectroscopy.

Methods

C57BL/6J mice were inoculated with parasitized red blood cells and monitored daily for the clinical/neurological signs of ECM¹. Healthy controls and infected animals were imaged from d5 to d8 after PbA injection. The MR protocol was performed under isoflurane anesthesia at 11.75 T (vertical Bruker AVANCE 500 WB MR system). Transverse MR images (FOV 24×24 mm², slice thickness 0.5 mm) were acquired using prospective respiratory gating (PC-SAM, Small Animal Instruments Inc). Anatomical imaging of the liver (matrix 240x240, in-plane resolution 100µm) was performed using a 2D spin-echo sequence (repetition time TR≥448 ms; echo time TE = 14 ms) with 20 contiguous slices. T₂ maps (TR ≥ 9 s; 12 equally spaced echoes at TE = 7.5 to 120 ms; matrix 64×64, 2 accumulations) were acquired in a single slice. T₂-maps were computed assuming mono-exponential decay³ using ImageJ. One large ROI was delineated, avoiding dorso-ventral phase-encoded flow artifacts on a slice devoid of respiratory motion artifact, and T₂-histograms were analyzed. A ROI in the spinal cord was used as internal reference. A Kruskal-Wallis test followed by Dunn’s multiple comparison was applied on the average T₂-values of the ROIs at each time point. ¹H-MRS of the liver was performed at d7 only using the PRESS sequence (TE= 16.51ms, TR= 5000ms, voxel size= 27mm³) after shimming with FASTMAP. Both water-saturated (NS=32) and non-saturated spectra (NS=8) were acquired (saturation with VAPOR). Spectra were processed under CSIAPO using AMARES⁴. Several parameters related to lipid composition were derived from lipid resonances⁵ including the saturated lipid component (SL), the total of unsaturated lipids (ULtot), the fraction of unsaturated lipids (fUL), the fraction of saturated lipids (fSL), the fraction of polyunsaturated lipids (fPUL) and the fraction of monounsaturated lipids (fMUL). Mann-Whitney test was used to compare control and infected animals (significance set to P<0.05).

Results

All infected mice developed ECM within 8 days following PbA injection and showed the usual clinical and neurological signs of the disease¹. ECM was associated with body weight loss (ca 11% at d7 after PbA injection). Anatomical MRI did not reveal any overt hepatic lesion, but a reduction in subcutaneous fat was visible at the end of the follow-up concomitant with body weight loss. Metabolic results (control mice: n=4, ECM mice: n=4) showed a significant increase in fPUL (Figure 1) in ECM mice as well as a trend for a decrease in water signal intensity at d7 (Figure 2). Liver T₂ values increased significantly with disease severity up to d6 or d7 and decreased thereafter (Figure 3), while T₂-values in spinal cord did not change by more than 8%. On average, liver T₂ was increased by 31% at d6 and by 22% at d7, compared to controls (Figure 4).

Discussion/Conclusion

The T₂ increase might be due to changes in interstitial water, protein content, lipid composition and/or lipid mass. Anomalies in water content and lipid composition were detected by ¹H-MRS at the latest stage of ECM. The increase in fPUL reflects an increase in lipids with short T₂ ⁵, however still superior to the highest T₂ value measured with relaxometry in this study (16 ms). Changes in T₂ values measured by relaxometry might be partly explained by loss of free water as suggested by the spectroscopic results. A longitudinal ¹H-MRS study on a larger cohort is required to better understand the origin of T₂ variation and the extent of hepatic lipid remodeling in ECM. Such study could help determine whether the alteration of these parameters is a side-effect of acute clinical degradation (dehydration and body weight loss at the late stage of ECM), or a direct consequence of liver dysfunction caused by PbA.

Acknowledgements

France Life Imaging (FLI).