Magnetic resonance imaging to assess the brain response to fasting in glioblastoma-bearing rats as a model of cancer anorexia

Abstract

Background: Global energy balance is a vital process tightly regulated by the brain that frequently becomes dysregulated during the development of cancer. Glioblastoma (GBM) is one of the most investigated malignancies, but its appetite-related disorders, like anorexia/cachexia symptoms, remain poorly understood.

Methods: We performed manganese enhanced magnetic resonance imaging (MEMRI) and subsequent diffusion tensor imaging (DTI), in adult male GBM-bearing (n = 13) or control Wistar rats (n = 12). A generalized linear model approach was used to assess the effects of fasting in different brain regions involved in the regulation of the global energy metabolism: cortex, hippocampus, hypothalamus and thalamus. The regions were selected on the contralateral side in tumor-bearing animals, and on the left hemisphere in control rats. An additional DTI-only experiment was completed in two additional GBM (n = 5) or healthy cohorts (n = 6) to assess the effects of manganese infusion on diffusion measurements.

Results: MEMRI results showed lower T₁ values in the cortex (p-value < 0.001) and thalamus (p-value < 0.05) of the fed ad libitum GBM animals, as compared to the control cohort, consistent with increased Mn²⁺ accumulation. No MEMRI-detectable differences were reported between fed or fasting rats, either in control or in the GBM group. In the MnCl₂-infused cohorts, DTI studies showed no mean diffusivity (MD) variations from the fed to the fasted state in any animal cohort. However, the DTI-only set of acquisitions yielded remarkably decreased MD values after fasting only in the healthy control rats (p-value < 0.001), and in all regions, but thalamus, of GBM compared to control animals in the fed state (p-value < 0.01). Fractional anisotropy (FA) decreased in tumor-bearing rats due to the infiltrate nature of the tumor, which was detected in both diffusion sets, with (p-value < 0.01) and without Mn²⁺ administration (p-value < 0.001).

Conclusions: Our results revealed that an altered physiological brain response to fasting occurred in hunger related regions in GBM animals, detectable with DTI, but not with MEMRI acquisitions. Furthermore, the present results showed that Mn²⁺ induces neurotoxic inflammation, which interferes with diffusion MRI to detect appetite-induced responses through MD changes.

Keywords: Cancer anorexia; Diffusion tensor imaging; Fasting paradigm; Glioblastoma; Manganese enhanced MRI.

Fig. 1

Decrease in body weight and blood glucose level in control and GBM rats under fed or fasted conditions. A Body weight changes after 16 h of food deprivation. B Blood glucose levels before the MRI session in fed and fasted conditions. In each box-plot, the central mark indicates the median, the cross mark the mean, and the bottom and top edges refer to the 25-th and 75-th percentiles, respectively. The upper and lower limits of the box extend to the most extreme data points not considered outliers (* p-value < 0.05)

Fig. 2

MEMRI of control and GBM rats, 24 h after Mn²⁺ administration, under fed or fasted conditions. A Anatomical location of the slice selected, including the perimeters of the investigated ROIs in the left contralateral hemisphere: cortex (red), hippocampus (green), thalamus (blue) and hypothalamus (orange). B Representative T₁ maps of a control and a tumor-bearing animal, 24 h after Mn.²⁺ administration, under fed and fasted conditions. The ROIs (cortex, hippocampus, hypothalamus and thalamus) are outlined in the maps. C-F Boxplots of T₁ values from the different ROIs investigated, in control and GBM rats, under fed and fasted conditions. Box plots are represented as indicated in Fig. 1. Outliers are plotted individually using the '•' symbol. (* p-value < 0.05, *** p-value < 0.001)

Fig. 3

Mean diffusivity and fractional anisotropy in the brain of control and GBM rats under fed or fasted conditions 24 h after Mn²⁺ administration. A, B Representative parametric maps of mean diffusivity (MD) and fractional anisotropy (FA) of a control and a tumor-bearing rat, under fed and fasted conditions, respectively. The ROIs (cortex, hippocampus, hypothalamus and thalamus) are outlined in the maps. C, D Boxplots of the MD and FA values, respectively, from the different regions investigated (cortex, hippocampus, hypothalamus and thalamus) in control cohort and tumor-bearing rats under fed and fasted conditions. Box plots are represented as indicated in Fig. 1. Outliers are plotted individually using the '•' symbol. (* p-value < 0.05, ** p-value < 0.01)

Fig. 4

Mean diffusivity and fractional anisotropy in the brain of control and GBM rats under fed or fasted conditions without Mn.²⁺ administration. A, B Representative parametric maps of mean diffusivity (MD) and fractional anisotropy (FA) of a control and a tumor-bearing rat, under fed and fasted conditions, respectively. The ROIs (cortex, hippocampus, hypothalamus and thalamus) are outlined in the maps. C, D Boxplots of the MD and FA values, respectively, from the different regions investigated (cortex, hippocampus, hypothalamus and thalamus) in control cohort and tumor-bearing rats under fed and fasted conditions. Box plots are represented as indicated in Fig. 1. Outliers are plotted individually using the '•' symbol. (*/# p-value < 0.05, ** p-value < 0.01, ***/### p-value < 0.001)