Rather than fixing dysfunctional microglia, why not replace them with shiny new cells? In a collaborative effort to inch microglia transplant closer to reality, researchers led by Chris Bennett, University of Pennsylvania, Philadelphia, and Mathew Blurton-Jones, University of California, Irvine, created mouse macrophages and human microglia, respectively, that resist colony stimulating factor 1 receptor inhibitors. CSF1R activity is essential for macrophage and microglial viability. In the December 30 Journal of Experimental Medicine, they reported that these mutant cells repopulated 75 to 99 percent of mouse brains cleared of native microglia. The inhibitor-resistant cells retained normal gene expression and function.

  • Human microglia and mouse macrophages were engineered to resist CSF1R inhibitors.
  • Gene expression and behavior did not change.
  • The mutant microglia repopulated the entire adult mouse brain.
  • They could be used as “living drugs” to clear toxic proteins.

Researchers applauded the work. “This study sets the new gold standard for studies of microglial replacement, enabling human-relevant studies of microglia biology,” wrote Diego Gomez-Nicola, University of Southampton, U.K. David Gate of Northwestern University in Evanston, Illinois, called this a landmark study. “It has clinical implications for treating neurodegenerative disease, lysosomal storage disorders, and other brain abnormalities,” he wrote (comments below).

In mice, sustained treatment with CSF1R inhibitors clears the resident microglia, a prerequisite for donor microglia to take hold. To facilitate this repopulation, Bennett and Blurton-Jones independently considered transplanting microglia that resist death by CSF1R inhibition, allowing the cells to thrive as the endogenous glia fizzle. After discovering they were tackling the same problem using different approaches, the scientists began collaborating.

Co-first author Sonia Lombroso and colleagues in Bennett’s lab introduced point mutations that would weaken or block the binding of the inhibitor PLX3397 to the ATP site on the kinase but not disrupt its kinase activity. Gomez-Nicola called this strategy brilliant. One CSF1R mutant—a glycine-to-alanine swap at residue 795—enabled cultured mouse macrophages to survive despite high inhibitor concentrations. Bennett used conditionally immortalized macrophages because they are easily modified by viral transduction and readily multiply in culture. What’s more, the mutant kinase phosphorylated its downstream target ERK as well as did wild-type kinase. Using molecular modeling, Blurton-Jones had also identified, in silico, G795A as the optimal mutation to prevent inhibitor binding without altering ATP binding needed for CSF1R kinase activity. “That a tiny side chain methyl group was all it took to disrupt inhibitor binding surprised me,” he told Alzforum.

For their part, co-first author Jean Paul Chadarevian and colleagues in Blurton-Jones’ lab used CRISPR to create human induced pluripotent stem cells with the G795A variant, then differentiated them into microglia. Again, the inhibitor spared these cells.

Would the mutation somehow alter the cells? RNA-Seq indicated almost identical transcriptomes in mutant and wild-type cells. While gene expression in the latter changed dramatically after PLX3397, mutant microglia were unfazed.

The mutant microglia behaved normally in vivo, too. Chadarevian and colleagues transplanted human G795A or wild-type microglia into the brains of newborn mouse pups. After two months, the scientists injected LPS to induce inflammation, then isolated the human microglia. The mutant microglia responded just like wild-type cells by downregulating the homeostatic marker P2RY12, upregulating the inflammatory marker CD45, and retracting cellular processes. The authors concluded that G795A does not disrupt microglial response to inflammation in vivo.

“It’s exciting that the mutation didn't have a fundamental impact on microglial biology, at least in what has been tested thus far,” Joseph Lewcock at Denali Therapeutics, South San Francisco, told Alzforum.

Taking Hold. Over three weeks, G975A macrophages (green) spread from the hippocampus and cortex to inhabit 71 percent of the brain, including 90 percent of the cortex. [Courtesy of Chadarevian et al., Journal of Experimental Medicine, 2022.]

How well might the G795A mouse macrophages and human microglia take hold in the mouse brain? Lombroso and colleagues injected the mutant macrophages into the brains of neonatal pups, then fed them PLX3397 for two weeks to kill off their native microglia. One week after stopping the inhibitor, the mutant macrophages occupied 75 percent of the brain at densities similar to endogenous microglia.

The same was true in adult mice. The researchers fed 2- to 3-month-old animals PLX3397 for two weeks before transplanting macrophages into the hippocampus and cortex, then continued the inhibitor for two more weeks. One week later, three-quarters of the brain held mutant macrophages (see image above). While macrophages and microglia come from different developmental lineages, the former can substitute for the latter in certain circumstances, but they are not a perfect replacement.

For that, Chadarevian and colleagues turned to the iPSC-derived microglia. They transplanted human microglia into the hippocampi and cortices of 2-month-old mice, then, one month later, fed the animals PLX3397 chow for 60 days. While the inhibitor killed most wild-type human microglia within 10 days, G795A cells survived and spread throughout the brain, accounting for all the observable microglia after 60 days (see image below).

Only Mutants. While wild-type human microglia perished in mice after two months of CSF1R inhibitor treatment (left), G795A microglia expanded to fill the entire brain (green, right). [Courtesy of Chadarevian et al., Journal of Experimental Medicine, 2022.]

G795A microglia even persisted 30 days after stopping PLX3997. To the authors, this suggests that microglia replacement might be feasible without continuous dosing. Blurton-Jones told Alzforum that microglial grafts populated the brains of 8-month-old mice, too, indicating that the strategy works in older animals.

“This research is exciting because it provides a platform for very efficient microglia replacement,” wrote Renzo Mancuso of the University of Antwerp, Belgium. “It can be applied to adult mice, which has been very challenging,” he added (comment below). Lewcock thinks that mice populated with homogenous human microglia are a great model for human cells, avoiding the cross-species differences and cell-to-cell variability inherent in endogenous mouse microglia. Oleg Butovsky, Brigham and Women's Hospital, Boston, agreed. “Using G795A microglia is an easier, faster, and cleaner way to generate mice with human microglia than previous models, such as FIRE mice,” he told Alzforum (Jun 2022 news).

Might microglial transplants be possible in people? “I think that safe and efficient engraftment of microglia in people will be feasible for clinical investigation in the next few years,” wrote Christopher Glass, University of California, San Diego (comment below). Taking a more cautious stance, Lewcock noted that multiple technical and biological challenges must be solved before microglia transplants are ready for people. On the safety front, Bennett and Blurton-Jones said that the doses of PLX3397 used in the mouse transplants are known to be safe in people as the inhibitor, also called pexidartinib, is FDA-approved to treat benign tumors. Blurton-Jones co-founded NovoGlia, Inc., a California-based biotech developing ways to engraft iPSC-derived microglia into people.

Butovsky and Gate suggested trying microglial transplants first in rare diseases involving dysfunctional microglia, such as lysosomal storage disorders. Along these lines, Bennett and Blurton-Jones are currently exploring how G795A microglia behave in animal models of such diseases and a rare leukoencephalopathy called ALSP, which is caused by a mutation in CSF1R.

Ultimately, the authors believe that transplant would be useful for any disease with dysfunctional microglia, including Alzheimer’s. Still, many questions remain, chief among them how genetically modified microglia will react in a disease context, Mancuso noted.

Other researchers were skeptical that transplanting healthy microglia would work since replacing the cells does not fix the underlying conditions that tainted the microglia and may eventually corrupt the new healthy cells. Bennett and Blurton-Jones agree that this is possible. Their plan, however, is to use the engineered microglia as “living drugs,” modifying them to target and engulf amyloid plaques or neurofibrillary tangles. “We want to empower microglial donor cells to act on the brain’s environment, not just respond to it,” Bennett said.—Chelsea Weidman Burke

Comments

  1. The idea behind the approach described in this paper is to circumvent known technical issues with existing microglial replacement techniques, including the impact of CSF1R inhibitors on human microglia grafts. In short, one has to first deplete mouse microglia for transplanted microglia to effectively colonize the niche, and this is achieved by using sustained exposure to high concentrations of CSF1R inhibitors. However, these inhibitors naturally affect the human microglia as well, so achieving complete depletion of mouse microglia without affecting human microglia is a complicated business.

    The authors had a brilliant idea, in my opinion: mutating the specific residues in CSF1R that regulate binding of kinase inhibitors, without affecting other areas of CSF1R that bind to its ligands. Indeed, they found such a residue, and its mutation confers resistance to CSF1R inhibitors without affecting normal CSF1R function, an aspect the authors thoroughly addressed in this paper. When this mutated residue is engineered into human iPSC-derived microglia, it enables long-term exposure to specific CSF1R inhibitors (PLX3397, PLX5622) required to deplete mouse microglia, generating an almost empty niche for the human microglia to colonize. The results evidence how superior this approach is to other methods available in the literature, including those from these same groups, enabling proper, full replacement of mouse microglia by their human counterparts. This study sets a new gold standard for studies of microglial replacement, allowing for human-relevant studies of the biology of microglia to be completed.

    The approach is likely sufficient to meet the majority of preclinical experiments one can think of, as long as these are amenable to treatment with PLX3397 or PLX5622. Going forward, it would be useful to ascertain if this residue also confers resistance to other CSF1R inhibitors already in the clinical space, in order to better understand the applicability of this approach to new clinical interventions. Other inhibitors may, or may not, need this particular residue to bind CSF1R, or may need others that can also be mutated by a similar strategy. Opening up the spectrum of compounds that can be coupled to this approach will enable more clinically-relevant experiments to take place, since the clinical space is rich in CSF1R inhibitors other than PLX drugs. On the flip side, identifying residues that only work for a range of inhibitors, but not others, has an advantage: If necessary, the transplanted human microglia could be depleted using those inhibitors that are unaffected by the specific mutated residue, opening up other clinical possibilities and improving safety.

  2. This is a landmark study by Chadarevian and Lombroso of the Blurton-Jones and Bennett labs, respectively. The two groups independently conceived the idea of mutating CSFR1 to be resistant to CSFR1 inhibitors. This straightforward and elegant strategy allows for the engraftment of CSFR1-resistant cells and the subsequent removal of endogenous microglia from the brain. The authors speculate that these transplantable cells—amenable to CRISPR gene editing—may represent a new class of “living drugs.” This is an exciting study that has clinical implications for treating neurodegenerative disease, lysosomal storage disorders, and other brain abnormalities. As the authors mention, future studies ought to finely map the consequences of the CSF1R mutation on microglial signaling, longevity, and function.

    The study is also a fine example of collaborative success in academia. Rather than competing against one another, these two groups joined forces to improve their work for the betterment of science. This is a highly commendable effort!

  3. This is a very interesting paper. I agree that given the pivotal role of microglia in neurodegenerative disorders, strategies that aim to correct microglial functional defects are very relevant. Among those, microglial cell therapy is potentially very exciting and worth exploring.

    This strategy would certainly require 1) knowing how to manipulate microglial function in a way that we boost beneficial functional properties and 2) the technical ability to genetically modify and replace human microglia in the brain.

    This paper is very exciting because it provides a platform for very efficient replacement, and, particularly important, it can be applied to adult mice, which has been very challenging so far. This approach, combined with other modifications of microglia function, provides an exciting avenue to explore microglial cell replacement therapies.

    There are, of course, a couple of notes of caution. How might altering the complex CSF1R phosphorylation pattern change microglial function beyond what is explored here in this paper? And how will genetically modified microglia react in a disease context? These are questions that will need to be addressed. Given that microglia have a particular developmental origin that impacts their physiology and response to disease in adulthood, I wonder if this strategy could be improved by transplanting yolk sac macrophage precursors instead of hematopoietic precursor cells.

    Overall, I think this is a very nice piece of work and very relevant, and congrats to all the authors involved.

  4. There are a number of neurological diseases in which replacement of microglia could have significant therapeutic benefit. There are three major challenges to achieving this goal.

    One is to engraft a cell that can fully recapitulate the phenotype of resident microglia. This cannot currently be accomplished with hematopoietic stem cells derived from the bone marrow, but several laboratories have shown that engrafting an iPSC-derived hematopoietic progenitor cell into a humanized immunodeficient mouse host leads to differentiation of cells that are very similar to resident human microglia. So that challenge is probably mostly solved.

    The second major challenge is to efficiently and safely replace the endogenous cells with engrafted cells. This is the problem this paper addresses, generating progenitor cells that are resistant to CSF1 receptor inhibitors. The paper makes a convincing case that the authors have been able to create a Csf1 receptor that is resistant to widely used inhibitors and that this enables efficient replacement of endogenous cells with engrafted cells. They go on to engineer the mutant receptor into human iPSCs to enable efficient engraftment of these cells into the mouse brain in the presence of the Csf1r inhibitor. This result supports translational relevance. 

    The third major challenge is common to most current cell-replacement strategies, which is to be able to maintain functional cells that are not genetically identical to the host and thus subject to immunologic rejection. This, in principle, can be managed by established methods for immunosuppression and by generating banks of donor iPSCs for optimal matching to recipients. The one additional caveat here is that cells have to be engineered to express the resistant CSF1R, so there is the potential for additional mutations to occur during this process. Overall, the development of these engineered cell lines is an important step toward clinical application.

    The bottom line is that I think the safe and efficient engraftment of microglia in people will be feasible for clinical investigation within the next few years. 

  5. In this paper, Chadarevian et al. have shown that an intriguing, single-point mutation in human CSF1R, G795A, resists inhibition by PLX3397. They have demonstrated successful transplantation of microglia expressing CSF1R with the G795A mutation both in a neonatal and adult model, without disruption of general microglial function and states.

    This discovery is very exciting. It opens up the possibility of developing research and clinical strategies for microglial replacement based on pharmacological depletion of endogenous microglia in the CNS without irradiation or genetic manipulations. One clear advantage of utilizing mutant hCSF1R G795A microglia is the extension of PLX treatment after transplantation. This allows further suppression of the endogenous microglial repopulation that competes with the niche of transplanted microglia for proliferation. The success of multiple strategies applied in this study brings out new possibilities to replace myeloid cells that are in vitro engineered to possess different types of functions that might benefit different diseases. Such benefits would not be limited in the CNS microglial replacement as shown in this paper. For example, this strategy may boost the replacement efficiency of published chimeric antigen receptor macrophages for tumor therapy.

    With the current structural understanding of PLX inhibitors, one fascinating fact is that the G795A mutation probably creates a new hydrophobic surface, bridging with V647 to occupy the docking pocket of PLX3397/5622, and thus reducing inhibitor binding affinities. Such an occupancy shift could also explain equally successful resistance of the G795C mutation to inhibitors. In contrast, the failure of G795V may largely result from different interactions with V647 versus those with the G795A and C mutations, which could disrupt the entire folding of hCSF1R. One potential future direction would be to examine other point mutations at either the V647 site or the G795 site. For example, G795L may “kick-off” the PLX compound further away, prevent tight docking, and, therefore, maintain the resistance of hCSF1R blockage at even higher doses.

  6. With increasing recognition of their role in neurodegenerative diseases, microglia replacement therapy has been a hot topic in recent years. Previous literature reported that lack of endogenous microglia is a prerequisite for efficient engraftment of exogenous microglia/macrophages into the adult brain (Xu et al., 2020; Cronk et al., 2018). The elimination of endogenous microglia requires either pharmacological inhibitors targeting microglial CSF1R signaling, which would suppress the expansion of engrafted microglia as well, or genetic approaches that are not applicable in clinical conditions.

    In this manuscript, Chadarevian, Lombroso, and colleagues engineered human CSF1R to be resistant to inhibitors while preserving its normal signaling capacity. By engrafting human-induced pluripotent stem cell–derived microglia that express the inhibitor-resistant CSF1R (G795A-iMG), they found that the transplanted iMG not only survived long-term PLX3397 treatment, but migrated from the injection sites to other brain regions over time, occupying the vacant space devoid of the endogenous microglia and eventually taking over the entire brain, representing over 99 percent of total observable microglia. The engrafted microglia appear to be transcriptionally and functionally normal. This approach, with striking microglial engrafting efficiency, provides great insights to microglia replacement therapy development.

    However, caution should be taken when considering microglial replacement therapy for Alzheimer’s disease or other neurodegenerative diseases, particularly using this inhibitor-resistant microglia approach. There are a few points to consider:

    1. Under neurodegenerative conditions, the brain stays in a chronic inflammatory state triggered by pathological proteins and injured neurons. Even if sick microglia are replaced with fresh, functionally normal microglia, the newly engrafted cells could quickly be converted to an activated or sick status. In the context of tauopathy, activated microglia drive tau pathogenesis and neurodegeneration (Shi et al., 2019), and a large number of newly activated microglia may not be beneficial, and may even exacerbate the disease.
    1. The inhibitor-resistant microglia cannot be killed by CSF1R inhibitor. Once they occupy the brain, if they become activated and start to do detrimental things, they could not be readily controlled, therefore posing an additional risk.
    1. Although the brain is relatively immune-tolerant, it’s not cut off from peripheral immune communication. What would be the fate of the engrafted exogenous microglia in the long run, and how the brain would be affected if transplant rejection eventually occurs, are still not clear.
    1. What would be the best way to deliver the transplantation in a clinical setting? Bone marrow transplantation involves hazardous procedures such as irradiation or chemotherapy, and the infiltrated peripheral myeloid cells are transcriptionally and functionally distinct from microglia, so it may not be the best choice. Human ipsc–derived microglia better resemble endogenous microglia, but transplantation of ihMG in preclinical models is carried out through intracranial injection. How to translate this to a relatively noninvasive approach clinically, still needs to be explored.

    Despite the questions that remain to be addressed, microglial replacement is still a potentially promising therapeutic approach to treat neurodegenerative diseases, particularly if transplanted microglia are engineered to shed toxic function and retain beneficial ones. In that case, the highly effective microglial engrafting strategy illustrated in this paper could potentially be of great help.

    References:

    . Efficient Strategies for Microglia Replacement in the Central Nervous System. Cell Rep. 2020 Nov 24;33(8):108443. PubMed.

    . Peripherally derived macrophages can engraft the brain independent of irradiation and maintain an identity distinct from microglia. J Exp Med. 2018 Jun 4;215(6):1627-1647. Epub 2018 Apr 11 PubMed.

    . Microglia drive APOE-dependent neurodegeneration in a tauopathy mouse model. J Exp Med. 2019 Nov 4;216(11):2546-2561. Epub 2019 Oct 10 PubMed.

  7. This is an exciting development for the field. Already this work has generated much enthusiasm as seen by the lively commentary above, and our colleagues have delineated both optimistic next steps and points of caution for translation. Like many of us, I am curious to see what this approach teaches us in the context of the diseased brain. Indeed, the notion of engineering “living drugs” is tantalizing, but the therapeutic performance of grafted microglia in disease models will be the ultimate test. No doubt such studies are well on their way. Nevertheless, I’d like to raise a few points for consideration with respect to the preclinical work in animals, which have not been discussed above.

    In the context of Aβ amyloidosis mouse models, it has been shown that chronic treatment with PLX-3397 (275mg/kg) or PLX5622 (1,200 mg/kg) dosed after the onset of plaque deposition leads to incomplete depletion of microglia (Dagher et al., 2015; Spangenberg et al., 2016). These are dosing regimens that deplete >95 percent of microglia in wild-type mice. However, in transgenic APP mice, the plaque-associated microglia appear to be resistant, suggesting that other signals maintain their survival in the presence of chronic CSF1R inhibition. It will be interesting to see if the 600mg/kg doses are sufficient to kill off the plaque-associated microglia to allow for complete replacement as described in this new paper.

    Another major consideration for this paradigm is that drug efficacy, or microglia depletion efficiency, is observed to be sex dependent. In a new paper from our group (Johnson et al., 2023), we demonstrate that PLX3397 drug exposure is about 2x higher in the brain and blood plasma of treated male mice as compared to female mice—this was true for both wild-type and tauopathy mice. Like in Aβ mice, we too observed incomplete depletion (~60 percent decrease in microglia number) using PLX3397 (275 mg/kg) formulated in AIN-76A chow from research diets. These results are consistent with findings from the Holtzman lab using a different tauopathy mouse model. Notably, the Holtzman lab increased the dosing regimen to 400mg/kg to ensure 100 percent microglial ablation in their model. Thus, in the context of the cell replacement paradigm reported by the Blurton-Jones and Bennett labs, such tauopathy mouse models might already be amenable to their approach as described, but sex-dependent phenotypes should be closely monitored.

    Aside from depletion efficiency, another consideration is potential adverse effects in the CNS caused by chronic and high dosing of PLX3397. In our new work, we characterized the PLX-resistant microglia in male mice, which showed morphological and gene-expression signatures consistent with an inflammatory phenotype, which is consistent with findings described by Yang Shi in the Holtzman lab. In contrast, we observed that the PLX-resistant microglia in female mice exhibited phenotypes more like wild-type mice (Johnson et al., 2023). Importantly, our paper demonstrated that PLX-treated male mice also exhibited signs of neuronal excitotoxicity potentially caused by PLX-resistant microglia in a chronic inflammatory state, which we argued precluded the therapeutic benefits of reducing tau pathology. Only the PLX-treated female tauopathy mice demonstrated extended survival and rescue of aberrant behavior (Johnson et al., 2023). 

    Here, the authors describe that ultimately only PLX-3397 was used because of inconsistencies in the quality of PLX-5622 compound from commercial vendors. This is understandable, but future work should consider establishing pure sources of PLX-5622 for this replacement paradigm. PLX-5622 is known to be much more selective for CSF1R over other related kinases in comparison to PLX3397. In a recent study it was shown that after seven days of treatment with PLX3397 (Liu et al., 2019), there was a considerable reduction of oligodendrocyte precursor cells (OPCs), but not with PLX5622 (1200 mg/kg). At 21 days of treatment, the reduction of OPCs was exacerbated—even modest reductions were observed with PLX5622. While no overt changes to mature oligodendrocyte number or myelination were observed in these conditions, it does raise concerns that using much higher doses (600 mg/kg) for long periods might exacerbate this OPC reduction phenotype and muddy interpretations of biological outcomes from studies using this replacement paradigm.

    None of the points raised here are insurmountable. I am still highly motivated by the potential of this experimental paradigm for basic science and preclinical studies. I think this replacement paradigm provides the research community a simpler approach to study human microglia (with a human microglial proteome) in the intact brain, which is vital for myriad investigations on the brain’s most dynamic cell whose biological status is so dependent on its microenvironment.

    References:

    . Colony-stimulating factor 1 receptor inhibition prevents microglial plaque association and improves cognition in 3xTg-AD mice. J Neuroinflammation. 2015 Aug 1;12:139. PubMed.

    . 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.

    . CSF1R inhibitors induce a sex-specific resilient microglial phenotype and functional rescue in a tauopathy mouse model. Nat Commun. 2023 Jan 9;14(1):118. PubMed.

    . Concentration-dependent effects of CSF1R inhibitors on oligodendrocyte progenitor cells ex vivo and in vivo. Exp Neurol. 2019 Aug;318:32-41. Epub 2019 Apr 25 PubMed.

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Sans Microglia, Mice Develop CAA and Die Young

Paper Citations

  1. . CSF1R-Related Adult-Onset Leukoencephalopathy with Axonal Spheroids and Pigmented Glia. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993 PubMed

External Citations

  1. pexidartinib

Further Reading

Primary Papers

  1. . Engineering an inhibitor-resistant human CSF1R variant for microglia replacement. J Exp Med. 2023 Mar 6;220(3) Epub 2022 Dec 30 PubMed.