Diphtheria toxin induced but not CSF1R inhibitor mediated microglia ablation model leads to the loss of CSF/ventricular spaces in vivo that is independent of cytokine upregulation

Background Two recently developed novel rodent models have been reported to ablate microglia, either by genetically targeting microglia (via Cx3cr1-creER: iDTR + Dtx) or through pharmacologically targeting the CSF1R receptor with its inhibitor (PLX5622). Both models have been widely used in recent years to define essential functions of microglia and have led to high impact studies that have moved the field forward. Methods Using either Cx3cr1-iDTR mice in combination with Dtx or via the PLX5622 diet to pharmacologically ablate microglia, we compared the two models via MRI and histology to study the general anatomy of the brain and the CSF/ventricular systems. Additionally, we analyzed the cytokine profile in both microglia ablation models. Results We discovered that the genetic ablation (Cx3cr1-iDTR + Dtx), but not the pharmacological microglia ablation (PLX5622), displays a surprisingly rapid pathological condition in the brain represented by loss of CSF/ventricles without brain parenchymal swelling. This phenotype was observed both in MRI and histological analysis. To our surprise, we discovered that the iDTR allele alone leads to the loss of CSF/ventricles phenotype following diphtheria toxin (Dtx) treatment independent of cre expression. To examine the underlying mechanism for the loss of CSF in the Cx3cr1-iDTR ablation and iDTR models, we additionally investigated the cytokine profile in the Cx3cr1-iDTR + Dtx, iDTR + Dtx and the PLX models. We found increases of multiple cytokines in the Cx3cr1-iDTR + Dtx but not in the pharmacological ablation model nor the iDTR + Dtx mouse brains at the time of CSF loss (3 days after the first Dtx injection). This result suggests that the upregulation of cytokines is not the cause of the loss of CSF, which is supported by our data indicating that brain parenchyma swelling, or edema are not observed in the Cx3cr1-iDTR + Dtx microglia ablation model. Additionally, pharmacological inhibition of the KC/CXCR2 pathway (the most upregulated cytokine in the Cx3cr1-iDTR + Dtx model) did not resolve the CSF/ventricular loss phenotype in the genetic microglia ablation model. Instead, both the Cx3cr1-iDTR + Dtx ablation and iDTR + Dtx models showed increased activated IBA1 + cells in the choroid plexus (CP), suggesting that CP-related pathology might be the contributing factor for the observed CSF/ventricular shrinkage phenotype. Conclusions Our data, for the first time, reveal a robust and global CSF/ventricular space shrinkage pathology in the Cx3cr1-iDTR genetic ablation model caused by iDTR allele, but not in the PLX5622 ablation model, and suggest that this pathology is not due to brain edema formation but to CP related pathology. Given the wide utilization of the iDTR allele and the Cx3cr1-iDTR model, it is crucial to fully characterize this pathology to understand the underlying causal mechanisms. Specifically, caution is needed when utilizing this model to interpret subtle neurologic functional changes that are thought to be mediated by microglia but could, instead, be due to CSF/ventricular loss in the genetic ablation model.