Delivery Courier? Microglia Drop Neprilysin Off Near Plaques in Mouse Brain

The blood-brain barrier border protects the brain by regulating the flow of molecules, peptides, and cells, but it also keeps out many therapeutics. Now, researchers have harnessed the power of a resident drug courier, i.e., microglia.

  • Researchers genetically engineered human iPSC-microglia to deliver neprilysin when the cells are near amyloid plaques.
  • Both local injection and brain-wide engraftment of neprilysin-secreting microglia reduced Aβ pathology.
  • Only widespread engraftment helped reduce pathologies and improve neuronal density in the subiculum.

Mathew Blurton-Jones and his team at the University of California, Irvine, have engineered human iPSC-derived microglia (iMG) to ferry protein therapeutics into the brain. They CRISPR-engineered the immune cells to express the Aβ-degrading enzyme neprilysin, but only in areas where the cells encounter amyloid plaques. When the researchers injected the couriers into mouse models of Alzheimer’s disease, the cells not only reduced the amount of amyloid, but also improved downstream aspects of Alzheimer’s pathogenesis. “While there are, of course, other approaches to reduce [Aβ] levels, this study demonstrated the powerful potential of iPSC-microglia to provide a novel immune cell therapy for a broad array of neurological diseases,” senior author Blurton-Jones wrote to Alzforum.

Researchers lauded the paper. “This work represents a potential paradigm shift in how we approach the challenge of delivering therapeutics across the BBB,” wrote Bart De Strooper at University of College London (comment below). “Replacing endogenous microglia with genetically modified microglia to deliver therapeutics could be a game-changer in the treatment of brain disorders,” he wrote.

“In my view, this proof-of-concept study brings forward several important and noteworthy findings,” wrote Eva Simoncicova, a Ph.D. student in Marie-Eve Tremblay's lab at the University of Victoria in British Columbia, Canada. “Leveraging microglia—the brain’s primary immune cells, which are typically present at the scene of any pathology—as vehicles for therapeutic delivery introduces novelty into the field of microglial targeting,” wrote Simoncicova.

Several years ago, both Blurton-Jones and De Strooper’s teams came up with ways to wipe out mouse microglia and replace them with human ones in the mouse brain (Apr 2019 news). The former lab introduced a mutation that enabled microglia to resist inhibitors that would normally kill them, and used these engineered cells to repopulate resident microglia (Jan 2023 news). These implanted human microglia can prevent, and even reverse, aspects of neurodegeneration (Jun 2024 news). This encouraged Blurton-Jones and first author Jean Paul Chadarevian, also at UC Irvine, to try engineering these microglia to deliver a therapeutic protein from within the central nervous system.

They chose neprilysin, an enzyme that cleaves the Aβ peptide, among other substrates, and has been studied extensively for its mechanism and as a potential therapeutic (Feb 2001 news). In the past, researchers have tried several approaches to deliver neprilysin, such as viruses, neural stem cells, and non-viral constructs (Spencer et al., 2008; Blurton-Jones et al., 2014; Rofo and Metzendorf et al., 2022). Chadarevian et al. sought a safer, more controlled approach.

When microglia encounter amyloid plaques, they express a particular set of genes. Specifically, in microglia that are close to a plaque, the promoter CD9 turns on this known plaque-responsive signature. But if the cells move just a few microns away, CD9 is completely off. “It’s almost like this binary genetic switch,” said Chadarevian.

Picking a Switch. In brain regions of mice that have been injected with human iPSC-derived microglia, CD9 (y-axis) is expressed in proportion to the amount of amyloid plaque (x-axis) in a certain brain region. [Courtesy of Chadarevian et al., Cell Stem Cell, 2025.]

The scientists used CRISPR to insert the coding sequence for neprilysin under the control of the CD9 promoter in iMG. The hope was that when a plaque activated CD9, the microglia would produce the amyloid-degrading protein. They injected both MITRG mice, an older line that can support human immune cells, and 5x-MITRG transgenics, which are a cross between MITRG and 5xFAD mice, with engineered iMGs that produced either membrane-bound or secreted neprilysin. Both methods reduced amyloid and preserved synapses, albeit to slightly different degrees. While membrane-bound neprilysin reduced more amyloid, the secreted neprilysin better protected synapses, as measured by ELISAs for the presynaptic protein synaptophysin and the postsynaptic protein. “Ultimately, that’s what matters in Alzheimer’s disease,” said Blurton-Jones.

Moving forward with the secreted version, Chadarevian and colleagues injected the engineered microglia into the brains of 5x-MITRG mice when they were 2 months old. In one group, the researchers only injected the genetically engineered cells into specific brain areas and kept the mouse’s own microglia. In another, the researchers replaced the mouse’s original microglia with the neprilysin-producing iMGs that were resistant to the CSF1Ri, gave the mice that inhibitor, and thus ensured that only the grafted microglia would survive and repopulate the mouse brain

Surprisingly, the former, local delivery worked about as well as brain-wide engraftment in certain ways. Both methods reduced Aβ monomers and oligomers when measured biochemically. What’s more, both generated improvements in secondary Alzheimer’s pathologies, including a reduction in astrogliosis, neuroinflammation, and plasma NfL.

However, in the subiculum, a part of the hippocampus that starts amyloid deposition early and becomes packed with plaque in these mice, only widespread engraftment was able to reduce plaque load, dystrophic neurites, and astrogliosis, and preserve neuronal density. The subiculum lies near the injection site but was not directly injected. Therefore, the mice with local engraftment had few human cells in that area. In the replacement group, the microglia migrated and engrafted throughout the subiculum. “In areas where there's a lot of amyloid pathology, it may take more of these cells in order to have an effect,” said Blurton-Jones.

Damage control. As amyloid plaques mature, they damage neuronal axons, which begin to swell. In this figure, halos of dystrophic neurons (yellow) form around plaques (blue) in the subiculum. Axonal damage is reduced the most in brain-wide engraftment (fourth panel) of neprilysin-secreting human microglia (red). [Courtesy of Chadarevian et al., Cell Stem Cell, 2025.]

“I was surprised to see how this approach not only resulted in an amelioration of the amyloid plaque pathology—which was somewhat expected based on the specific biological used—but led to a normalization of several neuroinflammatory, neuronal damage, and astrogliosis markers,” wrote Renzo Mancuso at the University of Antwerp in Belgium (comment below). “This means to me that the strategy used was not only able to clear amyloid plaques but also significantly alter the downstream intricate network of cellular interactions that ultimately underlies Alzheimer’s disease.”

Importantly for safety, the plaque-induced neprilysin production did not lead to measurable cleavage of neprilysin’s other substrates, the scientists report.

Could iMG be engineered to respond to specific pathologies in other diseases in which microglial activation is known to play a part? The scientists used breast cancer brain metastasis and multiple sclerosis to address the question. They injected iMG into mouse models of these diseases and found that the microglia mounted unique transcriptional responses to each pathology. For each disease, the team identified a pathology-specific promoter that they believe could work similarly to the way CD9 responds to amyloid plaques. “The main thing we are trying to highlight is that here we have a platform which can be more broadly applied,” said Chadarevian.

“This is a completely new paradigm for preclinical therapy development. It is highly modular and customizable,” wrote Chris Bennett at the University of Pennsylvania.

Next, the team hopes to publish a paper on a less-invasive route than intracranial infusion that they discovered to get microglia into the brain. They are also exploring whether the engineered microglia could work in lysosomal storage diseases. Regarding the approach of first depleting endogenous microglia and then replacing them with modified iMGs, Mikael Simons of DZNE in Munich suggested that the microglial disease Nasu-Hakola might be a good one to study (see comment below).––Andrea Tamayo

Andrea Tamayo is a freelance writer in Brooklyn, New York.

Comments

    • User Profile Image Bart De Strooper
      UK Dementia Research Institute@UCL and VIB@KuLeuven
    • Posted:

    This is a fantastic, highly important paper. Delivering biologics to the brain for therapeutic applications is notoriously difficult, often requiring intrathecal delivery of gene therapies or complex engineering of blood-brain barrier shuttles. The approach here—although invasive—is conceptually very innovative. Replacing endogenous microglia with genetically modified microglia to deliver therapeutics could be a game-changer in the treatment of brain disorders.

    I’m actively working on microglia transplantation myself so I wasn’t surprised by the concept, but there’s a big difference between expecting a certain outcome and actually doing the hard work to prove that it works. The successful demonstration of this concept is impressive.

    The data in this paper are convincing, and now further work is needed to explore whether this highly invasive but potentially transformative approach is translatable to humans. One major strength is the ability to achieve widespread brain coverage—something that remains a challenge for current gene therapy or biologics modified for BBB crossing.

    The example of using neprilysin is compelling, but it would be valuable to see whether clinically available anti-amyloid antibodies could be delivered using this strategy, since intravenous delivery is currently highly inefficient. A key question is how microglia would respond to secreting such antibodies themselves—something that needs careful investigation.

    Microglia play a central role in the pathogenesis of Alzheimer’s disease. Many genetic risk factors are linked to their activation states. The ability to replace dysfunctional microglia with engineered, "therapeutic" microglia is a promising direction. Beyond that, this work represents a potential paradigm shift in how we approach the challenge of delivering therapeutics across the BBB. Given that only about 0.3 percent of intravenously administered antibodies reach their target in the brain, a microglia-based delivery system could be revolutionary. Of course, more work is needed to assess efficacy, safety, and patient benefit—but this study opens up a whole new avenue worth pursuing.

    View all comments by Bart De Strooper

  2. This proof-of-concept study brings forward several important findings. Leveraging microglia—the brain’s primary immune cells, which are typically present at the scene of any pathology—as vehicles for therapeutic delivery introduces novelty into the field of microglial targeting. While current strategies have been focused primarily on directly modulating or inhibiting microglial activity, this approach takes it a step further by using advanced gene-editing techniques to guide microglia in their response to existing microenvironment. In this context, rather than temporarily alleviating disease symptoms, such as altered microglial activity, or solely reducing the load of pathological protein aggregates, utilizing these long-lived cells could offer the brain a more sustained and longer-term defensive mechanism. A growing body of research emphasizes the importance of heterogeneity in the microglial population and their state identity. The idea to selectively target individual microglial states, rather than the entire population, harnessing their specialized functions, seems like the most promising path forward for therapeutic advancement.

    A surprising observation was the small-scale engraftment of cells and the localized microglia-mediated increase in membrane-bound or secretory neprilysin. This increase not only induced objectively beneficial brain-wide effects across categories of symptoms, ranging from amyloid beta deposits to perineural net alterations, but appears to do so without interfering with some of the other signalling pathways related to neprilysin’s endogenous peptidase activity. While further research is needed to assess the long-term impact of artificially increasing neprilysin (or other therapeutic agents in future), as well as to determine whether other conditional ligands could inadvertently stimulate the respective regulatory promoter region, the applicability of this method across a range of neurological disorders is considerable.

    The strength of this study lies in its well-founded research concept, which directly addresses a critical gap in the field, and its thoughtful and robust experimental design that integrates a diverse range of models, methodologies, and targets. As tau seeding and phosphorylation is considered another major pathological hallmark of AD, it would be relevant if the authors next assessed the indirect impact of the neprilysin treatment on tau tangle formation. As the AD field continues to critically re-evaluate the amyloid hypothesis, which traditionally placed amyloid deposition at the center of the AD pathology, investigating the interplay between these two proteins remains a crucial area of inquiry.

    It is, however, important to highlight that the authors investigated many other relevant emerging AD hallmarks, including drawing connections between neprilysin exposure and amyloid toxicity associated with pre-synaptic axons, as well as its impact on astrocytic activity, perineuronal nets coverage, and dystrophic neurons across various brain regions. Similarly, although sex was not considered as a biological variable in this proof-of-concept study, it is a clinically relevant factor that warrants investigation. AD progression as well as microglia properties are distinct between men and women, underscoring the importance of incorporating sex as a biological variable in future research.

    Compared to the significant public focus on identification of therapeutic targets and development of therapeutic agents per se, the field of drug delivery—which in the end defines whether a drug will or will not have a clinically relevant neurological impact—does not seem to receive the same level of exposure or interest. By harnessing the potential of using a patient’s own cells, which carry the impact of an individual’s lifetime exposures, for therapeutic targeting of the brain immune system responses, this type of study brings the field one step closer to achieving the goal of a more personalized therapy.

    Given the numerous parallels across the aging decline and neurodegenerative disease spectrum of symptoms, such as synaptic and neuron loss, altered cytokine release and glial reactivity, this study could be inspirational for both the public and research community, extending its relevance even beyond the AD field.

    View all comments by Eva Simoncicova

  3. The most significant component of this paper lies in the novelty of using iPSC-derived microglia as a delivery system, and that this approach is able to change the course of Alzheimer’s disease pathology. Also very interesting is that—as cleverly presented by the authors—this system holds promise beyond AD and could become a highly transversal platform.

    I was surprised to see how this approach not only resulted in an amelioration of the amyloid plaque pathology—which was somewhat expected based on the specific biological used—but also led to a normalization of several neuroinflammatory, neuronal damage and astrogliosis markers. This means to me that the strategy used was not only able to clear amyloid plaques but also to significantly alter the downstream intricate network of cellular interactions that ultimately underlies Alzheimer’s disease.

    This work is very elegant. Complementary studies could be carried out to see how this system would react to the addition of other pathological hallmarks, i.e., tau. One other point to consider would be that the authors used a large number of genetic manipulations in both the iPSC lines and the host mice. I am curious to see how they envision translated this to humans.

    This preclinical work will need to go through very thorough safety and efficacy follow-up research before possibly arriving at a clinical setting. Even so, it has the potential to open new avenues that could be complementary to current therapeutic strategies by harnessing the intrinsic power of our own brains to protect ourselves from Alzheimer’s disease.

    View all comments by Renzo Mancuso

  4. This paper takes many important steps forward in demonstrating the great potential of microglia adoptive transfer for the treatment of AD. Successes in fibrosis and autoimmunity highlight the growing application of immune cell therapies outside of oncology.

    Taken together, prior studies extending this concept to CNS degenerative disease (including Yoo et al., 2023; Mishra et al., 2023; Milazzo et al., 2024; Colella et al., 2024), the clinical success of hematopoietic stem cell therapy for human leukodystrophies, and strong evidence that microglia play a central role in AD, all strongly support the development of engineered microglia therapies.

    Here, the authors elegantly employ ex vivo genome editing to therapeutically engineer microglia that, in a pathology-specific manner, produce the amyloid-degrading enzyme neprilysin. They engraft them in the brain by direct intracranial injection, and deeply validate that context-dependent neprilysin production, particularly a secreted form, is highly and broadly pathology modifying. As a bonus, they show that iPSC microglia have distinct transcriptional responses to demyelination, tumor infiltration, and amyloid pathology, nominating many genes that could provide context-dependent payload regulation.

    The paper is groundbreaking for the field because it directly demonstrates how an endogenous anti-amyloid mechanism can be specifically boosted by cell engineering in a long-lived resident cell population to reduce amyloid toxicity. This is a completely new paradigm for preclinical therapy development. It is highly modular and customizable.

    In clarifying the extent to which microglia processing contributes to amyloid pathology, these results also improve our understanding of the pathobiology of AD. One surprising finding is that neprilysin engineering increases amyloid phagocytosis by microglia in vitro. Although speculation, it is exciting to consider that amyloid processing rather than uptake may be the more important therapeutic target. Importantly, in some experiments the authors dually engineer donor iPSC microglia to insert neprilysin into the CD9 locus, and an “inhibitor resistance” mutation in the CSF1R locus. This raises the exciting additional possibility for combinatorial microglial engineering, for example to both increase amyloid uptake and its processing.

    References:

    Colella P, Sayana R, Suarez-Nieto MV, Sarno J, Nyame K, Xiong J, Pimentel Vera LN, Arozqueta Basurto J, Corbo M, Limaye A, Davis KL, Abu-Remaileh M, Gomez-Ospina N. CNS-wide repopulation by hematopoietic-derived microglia-like cells corrects progranulin deficiency in mice. Nat Commun. 2024 Jul 5;15(1):5654. PubMed.

    Milazzo R, Montepeloso A, Kumar R, Ferro F, Cavalca E, Rigoni P, Cabras P, Ciervo Y, Das S, Capotondo A, Pellin D, Peviani M, Biffi A. Therapeutic efficacy of intracerebral hematopoietic stem cell gene therapy in an Alzheimer's disease mouse model. Nat Commun. 2024 Sep 13;15(1):8024. PubMed.

    Mishra P, Silva A, Sharma J, Nguyen J, Pizzo DP, Hinz D, Sahoo D, Cherqui S. Rescue of Alzheimer's disease phenotype in a mouse model by transplantation of wild-type hematopoietic stem and progenitor cells. Cell Rep. 2023 Aug 29;42(8):112956. Epub 2023 Aug 8 PubMed.

    Yoo Y, Neumayer G, Shibuya Y, Mader MM, Wernig M. A cell therapy approach to restore microglial Trem2 function in a mouse model of Alzheimer's disease. Cell Stem Cell. 2023 Oct 5;30(10):1392. PubMed.

    View all comments by Chris Bennett

  5. This very interesting paper describes a potentially important therapeutic strategy for delivering molecules into the brain for long-term expression. If isogenic cells are used for transplantation, this method may even offer advantages over viral delivery. One could envision a scenario in which endogenous microglia are first depleted and then replaced with genetically modified, isogenic iPSC-derived microglia. Diseases that primarily affect microglia, such as Nasu-Hakola disease, would be a compelling starting point for exploring this approach.

    View all comments by Mikael Simons

  6. It is exciting to see that the clinical data generated on Trontinemab seems to hold together with a larger number of patients. The drop in amyloid burden is both fast and deep, which might be important for a clinical response. Also, interesting to see that the ARIA rate continues to be low, which speaks to the way Trontinemab works—lower dose and another route into the brain compared to conventional antibodies.

    Due to the unique mode of action of Trontinemab, with a more rapid and pronounced effect on the amyloid burden, it will be interesting to see what the Phase 3 program will look like. Which dose, frequency, and duration will be used to confirm the clinical benefit and prepare for a potential long-term use in patients with risk of Alzheimer’s disease?

    The clinical data now starting to emerge using active transport across the blood brain barrier mediated by TfR is fueling an overall interest in the pharma industry to apply this approach in areas other than Alzheimer’s disease. We are likely to reach a point soon where active delivery across the blood-brain barrier will be both required and the standard when using large molecules like antibodies to treat brain disorders where the target is behind the blood-brain barrier.

    View all comments by Per-Ola Freskgård

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Chimeric Mice: Can They Model Human Microglial Responses?
  2. Healthy, Drug-Resistant Microglia Reinvigorate Mouse Brain
  3. Microglial Transplants Reverse Age-Related Pathology in Mice
  4. Neprilysin Steps out of the Shadows

Paper Citations

  1. Spencer B, Marr RA, Rockenstein E, Crews L, Adame A, Potkar R, Patrick C, Gage FH, Verma IM, Masliah E. Long-term neprilysin gene transfer is associated with reduced levels of intracellular Abeta and behavioral improvement in APP transgenic mice. BMC Neurosci. 2008;9:109. PubMed.
  2. Blurton-Jones M, Spencer B, Michael S, Castello NA, Agazaryan AA, Davis JL, Müller FJ, Loring JF, Masliah E, LaFerla FM. Correction: Neural stem cells genetically-modified to express neprilysin reduce pathology in Alzheimer transgenic models. Stem Cell Res Ther. 2024 Mar 25;15(1):88. PubMed. Corrected paper.
  3. Rofo F, Metzendorf NG, Saubi C, Suominen L, Godec A, Sehlin D, Syvänen S, Hultqvist G. Blood-brain barrier penetrating neprilysin degrades monomeric amyloid-beta in a mouse model of Alzheimer's disease. Alzheimers Res Ther. 2022 Dec 5;14(1):180. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. Chadarevian JP, Davtyan H, Chadarevian AL, Nguyen J, Capocchi JK, Le L, Escobar A, Chadarevian T, Mansour K, Deynega E, Mgerian M, Tu C, Kiani Shabestari S, Carlen-Jones W, Eskandari-Sedighi G, Hasselmann J, Spitale RC, Blurton-Jones M. Harnessing human iPSC-microglia for CNS-wide delivery of disease-modifying proteins. Cell Stem Cell. 2025 Apr 9; Epub 2025 Apr 9 PubMed.

Annotate

To make an annotation you must Login or Register.