Microglia are known as the brain’s sometimes-overzealous clean-up crew, but what happens to people if these immune cells are missing from the get-go? Two studies published March 29 in the American Journal of Human Genetics describe nine people who carry mutations in both copies of the CSF-1R gene, a receptor that is vital for development and maintenance of microglia. In the most severe cases, key brain structures were not there, such as the corpus callosum that connects the two hemispheres. Bones were overly dense and malformed. While it is uncertain what the findings portend about the role of microglia in late-onset neurodegenerative disease, they make clear that the cells are indispensable components of brain development and function.

  • Researchers describe nine cases of bi-allelic CSF-1R mutations.
  • Severe cases had no microglia and profound abnormalities in brain and bones.
  • Symptom onset varied from prenatal to early adulthood.

Microglial progenitors derive from the embryonic yolk sac and migrate into the budding brain early in development. There, their renewal and expansion are driven by colony stimulating factor 1. Some animal studies suggest that the cells are pivotal in supporting myelination of the developing brain, assembling brain circuits, and pruning synapses (Squarzoni et al., 2014; Wlodarczyk et al., 2017Hammond et al., 2018). 

People who carry one nonfunctional copy of CSF-1R develop a condition known as adult-onset leukoencephalopathy with axonal spheroids and pigmented glia (ALSP). This neurodegenerative disease degrades white matter and leads to dementia and motor impairment (Rademakers et al., 2011). In some cases, the mutated copy of CSF-1R may exert a dominant-negative effect—corralling the healthy copy and stifling signaling by more than half (Pridans et al., 2013). 

What would happen in people who had two mutant copies of CSF-1R? Researchers led by Tjakko van Ham of Erasmus University in the Netherlands and James Bennett of the University of Washington in Seattle described two such cases, an infant and a 24-year-old man. The baby came into Bennett’s care when he was born at Seattle Children’s Hospital. His parents had traveled from Alaska to have their child there after abnormalities were detected in utero. He was born with multiple congenital brain abnormalities, including complete absence of a corpus callosum, enlarged ventricles, and a cerebellar defect called a Dandy-Walker malformation, in which parts of the cerebellum fail to develop. He had low blood calcium and dense bones, a condition the researchers diagnosed as osteopetrosis. During his short life, he had trouble breathing, was unable to eat properly, and had epilepsy. He died from a bacterial infection at 10 months of age.

Upon autopsy, the researchers found no microglia in the brain, although they did spot some cells expressing Iba1, a microglial marker that can also be expressed on circulating myeloid cells, and CD68, which marks phagocytes. Genetic testing revealed that the infant carried two copies of CSF-1R with a point mutation that disrupted splicing, leading to a protein product with a nonfunctional kinase domain. The parents, who were cousins, each carried a single copy of the variant. Neither had yet reported symptoms of ALSP, but were only 40 years old at last contact.

Missing Microglia. Immunocytochemistry revealed Iba1+ microglia throughout the brain of a control, with ramified processes (G and H). An infant with homozygous mutations in CSF-1R only had scant Iba1+ cells near blood vessels (I), which were rounded (J and K). [Courtesy of Oosterhof et al., American Journal of Human Genetics, 2019.]

The 24-year-old patient is still alive. At age 12, he started having seizures and regressed developmentally. Now he is unable to walk, talk, or feed himself. An MRI revealed abnormalities similar to the infant’s, including a partial loss of the corpus callosum, enlarged ventricles, atrophy of the cerebellar vermis, and white-matter loss. However, he had neither hypocalcemia nor dense bones. Two of his siblings had died at ages 4 and 21 with similar symptoms. He carries a homozygous mutation that was also predicted to affect function of the kinase domain of CSF-1R.

The researchers speculated that the brain abnormalities in both cases were caused by a lack of microglia, and that the skeletal problems in the infant were likely the result of a loss of osteoclasts. These tissue-resident macrophages break down extra bone tissue, and also depend on CSF-1R signaling.

The Dutch researchers had previously generated CSF-1R-deficient zebrafish, and reported that the minnows had no microglia, dense bones, and died prematurely (Oosterhof et al., 2018). A proteomic analysis found the fish lacked Cux1a, a homolog of human CUX1. This neuronal transcription factor is required for the projection of axons that make up the corpus callosum. The researchers went back to the postmortem samples from their first case, and found about half as many CUX1+ neurons as in an age-matched control. They proposed that microglia likely play an instrumental role in the formation of the corpus callosum by supporting CUX1+ neurons.

The second paper, by researchers led by Shiro Ikegawa at the RIKEN Center for Integrative Medical Sciences in Tokyo, described seven people with bi-allelic CSF-1R mutations among three families. The most severe case was a 5-year-old Brazilian boy, who had similar brain deformities as those found by Bennett and colleagues. He had severe bone abnormalities, which the researchers diagnosed as dysosteosclerosis, a condition distinct from osteopetrosis, Ikegawa emphasized. He noted that the diagnosis of osteopetrosis in the infant in Bennett’s paper was likely incorrect. The other cases—one from a Japanese family and five from an extended family of Chaldeans, who are Iraqi Christians, in the U.S.—had varying degrees of brain abnormalities, dysosteosclerosis, and symptom onsets and severities. While many had seizures, intellectual disabilities, and motor problems, such as rigidity or trouble walking, one completed college before troubles with coordination and speech started to emerge. The carriers in the Brazilian and Japanese families were compound heterozygotes, meaning that they had a different mutation in each copy of the CSF-1R, while the Chaldean carriers were homozygous. The researchers reported that the mutations nixed kinase activity of the receptor by varying degrees.

Authors of both studies proposed that the severity of disease across cases correlated with the extent to which CSF-1R function was lost. They placed the disease phenotypes along the same spectrum as those in heterozygous carriers with ALSP. Though people with ALSP have a later onset and do not have skeletal problems, their disease profoundly affects white matter, including the corpus callosum, van Ham pointed out. He thinks that heterozygous carriers of CSF-1R mutations may start life with enough microglia for the brain to develop, but cell numbers gradually wane over time as CSF-1R levels are too low to support their renewal. At that point, scarcity of trophic factors made by microglia, such as insulin-like growth factor-1, may lead white-matter structures to crumble, he added. In homozygous carriers, these structures may never develop in the first place, leading to more severe developmental abnormalities, he said.

Florent Ginhoux of the Agency for Science, Technology, and Research in Singapore commented that these cases provide invaluable insight into human biology. He, too, noted that the wide spectrum of clinical and pathological manifestations caused by the mutations likely reflect their severity. He said work is needed to quantify the loss of CSF-1R function caused by each mutation, and understand how that correlates with the severity of the disease.

Kim Green of the University of California, Irvine, had a different take on the data. He agrees that the extent of the abnormalities, particularly in the infant with homozygous CSF-1R mutations, illustrates that microglia play a critical role in brain development. However, he thinks people with residual CSF-1R function, such as heterozygous mutation carriers, or homozygous carriers with less severe mutations, have a fundamentally different disorder caused by malfunction of mutated microglia. He noted that mice expressing a single copy of CSF-1R have more microglia than wild-type mice, a curious finding he is investigating (Chitu et al., 2015). He suspects that people with residual CSF-1R likely have a disorder caused by the actions of those dysregulated cells, rather than their loss. Van Ham believes the zebrafish model, in which loss of CSF-1R correlates with loss of microglia, may more accurately reflect the human situation. Van Ham added that people with ALSP share several of the same pathological characteristics, such as axonal spheroids, as observed in the infant with homozygous mutation. They also have fewer microglia in some regions of the brain.

What can these cases tell scientists about the role of microglia in late-onset neurodegenerative disease? To van Ham, the findings highlight a pivotal function for microglia in supporting axonal growth and maintaining proper wiring in the brain. He speculated that in AD, microglia could become chronically distracted by plaques, leading to a loss of physiological functions, and perhaps the erosion of white-matter tracts the cells usually support. On the flip side, Green found that using a CSF-1R inhibitor to deplete microglia from adult mice did not cause cognitive problems, and even assuaged neuroinflammation and synaptic damage in mouse models of amyloidosis (Apr 2014 news; Spangenberg et al., 2016). Researchers have also reported that nixing the cells with this inhibitor prevented Aβ plaque formation in 5xFAD transgenic mice, while another study reported that genetic ablation of the cells in adult APP/PS1 and APP23 mice appeared to have no effect on plaque development (Mar 2018 newsOct 2009 news). However, van Ham said it is important to consider all physiological functions of microglia before toying with reducing their numbers.—Jessica Shugart


  1. These two papers on CSF1R deficiency in humans are obviously rather interesting in the context of our own work. CSF1R mutations have been associated with the autosomal-dominant disease, hereditary diffuse leukoencephalopathy with axonal spheroids (HDLS, or ALSP). Our earlier work showed that the coding mutations in this disease lead to an altered protein that lacks signalling functions and is a dominant-negative. It is dominant in two respects. The mutant protein forms inactive heterodimers with the normal gene product. So in an HDLS patient, only 25 percent of dimeric receptors are functional. Further, the nonfunctional receptors are still expressed on the cell surface, where they compete for ligand. We have made a mouse model of the human disease and confirmed that there is a dominant effect on macrophage biology and also that a homozygote is lethal (unpublished). Both these studies support the conclusion that individuals with a genuine heterozygous loss of function do not develop disease. The inheritance is recessive. By extension, haploinsufficiency is not the explanation for dominant inheritance of HDLS. We would suggest that in the very small number of cases where haploinsufficiency has been inferred there is some other form of dominant interference so that there is much more than 50 percent loss of CSF1R function.

    Interestingly Guo et al. also show that the mutants they have studied may not produce a 100 percent loss of CSF1R protein in their patients. So, they should not be compared directly to full knockouts in animal models. There are two misconceptions in both manuscripts. Firstly, there is no compelling evidence for functional expression of CSF1R outside the myeloid lineages. The small number of published studies of CSF1R expression in epithelia and neurons were disproved by later studies. So, the phenotypes observed in patients and animals are due to deficiencies in macrophage lineage cells. Secondly, the CSF1R mutation in mice is considerably less penetrant on the original mixed genetic background than in the inbred C57Bl6 genetic background. The heterozygous mutation in outbred mice has no effect (again arguing against haploinsufficiency) despite the fact that it is not dosage-compensated.

    The bone phenotypes observed by Guo et al. (largely due to the loss of osteoclasts) and brain phenotypes are similar to those observed in both receptor (CSF1R) and ligand (CSF1) mutations in both mice and rats. We recently published a knockout of CSF1R in rats (Pridans et al., 2018). These rats are born normally, and develop a postnatal growth retardation and osteopetrosis. They lack microglia entirely, and also share with the human patients a thinning of the corpus callosum and enlarged lateral ventricles. However, there is no loss of Cux1 mRNA in any brain region. By contrast to mutant inbred mice, many CSF1R knockout rats survive well into adulthood. The phenotype is variable and more penetrant in males. Since our published study, we have found that the impact of the CSF1R knockout in rats is also influenced by genetic background, being more severe on an inbred genetic background.

    There are also quite likely to be epistatic interactions in humans. For example, it is worth noting that  a variation at the Csf1 (ligand) locus associates with Paget’s disease in humans. We have also worked on novel CSF1R mutations in mice. We have recently generated a hypomorphic mutation associated with an enhancer deletion that entirely lacks microglia throughout development but has no brain phenotype. The mice are not osteopetrotic and grow normally. 

    The impacts of the CSF1R mutations in both Guo et al. and Oosterhof et al. are attributed to the loss of microglia. Our work suggests that dysregulation of peripheral macrophages is an essential contributor to brain pathology in CSF1R mutations and that microglia may actually be partly redundant.


    . Pleiotropic Impacts of Macrophage and Microglial Deficiency on Development in Rats with Targeted Mutation of the Csf1r Locus. J Immunol. 2018 Nov 1;201(9):2683-2699. Epub 2018 Sep 24 PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Mutations in Microglia Protein Cause Rare Dementia
  2. Microglial Magic: Drug Wipes Them Out, New Set Appears
  3. Wiping Out Microglia Prevents Neuritic Plaques

Research Models Citations

  1. 5xFAD (B6SJL)
  2. APPswe/PSEN1dE9 (line 85)

Paper Citations

  1. . Microglia modulate wiring of the embryonic forebrain. Cell Rep. 2014 Sep 11;8(5):1271-9. Epub 2014 Aug 21 PubMed.
  2. . Microglia and the Brain: Complementary Partners in Development and Disease. Annu Rev Cell Dev Biol. 2018 Oct 6;34:523-544. Epub 2018 Aug 8 PubMed.
  3. . CSF1R mutations in hereditary diffuse leukoencephalopathy with spheroids are loss of function. Sci Rep. 2013 Oct 22;3:3013. PubMed.
  4. . Colony-Stimulating Factor 1 Receptor (CSF1R) Regulates Microglia Density and Distribution, but Not Microglia Differentiation In Vivo. Cell Rep. 2018 Jul 31;24(5):1203-1217.e6. PubMed.
  5. . Phenotypic characterization of a Csf1r haploinsufficient mouse model of adult-onset leukodystrophy with axonal spheroids and pigmented glia (ALSP). Neurobiol Dis. 2015 Feb;74:219-28. Epub 2014 Dec 9 PubMed.
  6. . Eliminating microglia in Alzheimer's mice prevents neuronal loss without modulating amyloid-β pathology. Brain. 2016 Apr;139(Pt 4):1265-81. Epub 2016 Feb 26 PubMed.
  7. . Formation and maintenance of Alzheimer's disease beta-amyloid plaques in the absence of microglia. Nat Neurosci. 2009 Nov;12(11):1361-3. PubMed.

Other Citations

  1. APP23

External Citations

  1. Wlodarczyk et al., 2017

Further Reading


  1. . Dysosteosclerosis is also caused by TNFRSF11A mutation. J Hum Genet. 2018 Jun;63(6):769-774. Epub 2018 Mar 22 PubMed.

Primary Papers

  1. . Homozygous Mutations in CSF1R Cause a Pediatric-Onset Leukoencephalopathy and Can Result in Congenital Absence of Microglia. Am J Hum Genet. 2019 Mar 29; PubMed.
  2. . Bi-allelic CSF1R Mutations Cause Skeletal Dysplasia of Dysosteosclerosis-Pyle Disease Spectrum and Degenerative Encephalopathy with Brain Malformation. Am J Hum Genet. 2019 May 2;104(5):925-935. Epub 2019 Apr 11 PubMed.