Major caveats discovered in the genetic cell ablation model used to study microglia functions

Microglia originate from yolk sac progenitors and migrate to the brain during development to become the innate immune cells of the central nervous system (CNS). In recent years, it has been shown that these cells are essential for proper brain development and maintenance of the homeostasis of the brain under physiological conditions [1]. Many of microglia’s critical roles such as maintaining neuronal function, synaptic plasticity, and regulating astrocyte phenotypes can be attributed to their functional states characterized by distinct morphologies and gene expression profiles [2]. In addition to the homeostatic role of microglia in the CNS, studies show that microglia can exacerbate neuronal dysfunction under pathological conditions such as Alzheimer’s disease (AD) or Parkinson’s disease (PD) [3]. The complexity of microglia’s dynamic role in the brain is not completely understood; however, the recent development of microglia ablation models provides powerful new tools to investigate the role of these cells under both physiological and pathological conditions. In this commentary, we will review what is known about microglia ablation and discuss both potential benefits of applying this technology in pathological brains, as well as potential challenges of using these strategies to “reset” a dysfunctional CNS milieu.

Ablation models of glial cells have been in development since the late 1960s [4] in an attempt to elucidate the function of these cells in the mammalian brain. Early ablation models lacked specificity and targeted multiple cell types, including neurons. These models included chemicals like ethidium bromide (EtBr), X-irradiation, and other toxic substances. Advances in biotechnology have allowed the development of strategies to specifically ablate a cell type of interest with little impact to surrounding cells or tissue. However, because microglia share many similar cellular markers with peripheral macrophages and monocytes, developing ablation techniques that target only microglia without affecting peripheral immune cells has been challenging. Fortunately, significant advancements have been made in recent years in the development of microglia-specific ablation models.

Current strategies of microglia ablation fall into one of two categories: genetic targeting or pharmacological inhibitors. In the early 2000s, transgenic expression of the inducible diphtheria toxin receptor was developed as a novel approach for cell type-specific ablation that countervailed previous ablation models [5]. Genetic approaches of ablating microglia using Cre-inducible diphtheria toxin receptor (DTR) or diphtheria toxin subunit A (DTA), an inhibitor of essential cellular protein synthesis, are generally driven by the promoter for fractalkine receptor CX3CR1. The CX3CR1 receptor is expressed both in microglia and in peripheral macrophages, which presents a potentially confounding variable for models of microglia ablation based on this promoter. To avoid this confound, waiting 4–6 weeks after TAM-induced DTR expression before administering diphtheria toxin (Dtx) allows the DTR + peripheral macrophages to replenish from CX3CR1(−) progenitors rendering them DTR negative [6]. Due to the turnover rate of peripheral macrophages, this model can be utilized for both microglia and peripheral macrophages or microglia-specific ablation, efficiently depleting the cell population within 2–3 days of Dtx treatment. It was not until a decade later that this paramount genetic model would face significant competition with a pivotal pharmacological ablation model.

The first generation of pharmacological inhibitor, PLX3397, inhibits colony-stimulating factor-1 receptor (CSF1R), whose signaling is required for continued survival of microglia. Administration of PLX3397 via diet achieves 90% ablation of microglia within 3 weeks with minimal effect to peripheral macrophages. PLX3397 has several undesirable characteristics including long treatment time, low penetrance across the blood–brain barrier, and off-target inhibition of CSF1R paralogs. These characteristics prompted the development of PLX5622, which displays high specificity for CSF1R and greater brain penetrance [7]. Unlike its predecessor, PLX5622 can effectively ablate microglia within 3–7 days, which is more comparable to the rate of genetic ablation models.

Both genetic and pharmacological strategies have their advantages and disadvantages. The DTR genetic ablation model is fast-acting in its ablation of microglia, better than the PLX5622 model. Apart from the longer time needed for ablation, PLX5622 is easier to implement into experiments as it eliminates the need for breeding of transgenic mice. Furthermore, its minimal effect on peripheral macrophages eliminates the need of a 4–6-week recovery period to allow macrophages to replenish from progenitors. Despite their differences, both models have been widely used in many seminal studies regarding the role of microglia in a variety of neurological functions and processes.

It was not until recently that Bedolla et al. showed that a global ventricular volume/CSF loss was induced with the use of the CX3CR1-iDTR model [8]. In this study, we detailed how the CX3CR1-iDTR and iDTR mice treated with diphtheria toxin (Dtx) but not the PLX5622-treated mice result in a loss of ventricular spaces, astrocyte activation, and upregulation of key cytokines in the brain. Using 3D MRI measures, we show that the loss of ventricular space was not due to parenchyma swelling or altered brain water content, nor was it able to be reversed by repertaxin (an inhibitor of the KC/Gro cytokine receptor CXCR2, which can reduce edema in a stroke model) treatment. Owing to the rigor of this study, we can now distinguish key differences in the use of genetic ablation models in comparison to pharmacological ablation models. Since any mouse strain that carries the iDTR allele will exhibit this shrinkage of CSF phenotype after Dtx treatment, previous studies using this allele aiming to ablate any specific cell types will likely have this confounding phenotype. Given the prevalence of this ablation model across the literature, this finding has a significant impact on data interpretation of past studies and warrants caution against using this genetic model to achieve specific cell-type ablation such as distinct neuronal types or glial cells.

Nonetheless, genetic and pharmacological ablation models have provided us with the tools to expand our field of knowledge on microglia and realize that they are more than just “glue” in the brain according to the nineteenth-century neuroscientists. Precise mechanisms remain controversial for these immune cells’ role in diseases such as dementia and Parkinson’s or their prevalence in stroke outcomes, necessitating more rigorous ablation models. The recent findings regarding the iDTR model demonstrate the need for critical examination of existing ablation models. While this may seem like a step backwards, identifying caveats in current ablation models allows us to continually improve animal models to address these important outstanding questions in the field.