Gamma entrainment using audiovisual stimuli alleviates chemobrain pathology and cognitive impairment induced by chemotherapy in mice

Abstract

Patients with cancer undergoing chemotherapy frequently experience a neurological condition known as chemotherapy-related cognitive impairment or “chemobrain,” which can persist for the remainder of their lives. Despite the growing prevalence of chemobrain, both its underlying mechanisms and treatment strategies remain poorly understood. Recent findings suggest that chemobrain shares several characteristics with neurodegenerative diseases, including chronic neuroinflammation, DNA damage, and synaptic loss. We investigated whether a noninvasive sensory stimulation treatment we term gamma entrainment using sensory stimuli (GENUS), which has been shown to alleviate aberrant immune and synaptic pathologies in mouse models of neurodegeneration, could also mitigate chemobrain phenotypes in mice administered a chemotherapeutic drug. When administered concurrently with the chemotherapeutic agent cisplatin, GENUS alleviated cisplatin-induced brain pathology, promoted oligodendrocyte survival, and improved cognitive function in a mouse model of chemobrain. These effects persisted for up to 105 days after GENUS treatment, suggesting the potential for long-lasting benefits. However, when administered to mice 90 days after chemotherapy, GENUS treatment only provided limited benefits indicating that it was most effective when used to prevent the progression of chemobrain pathology. Furthermore, we demonstrated that the effects of GENUS in mice were not limited to cisplatin-induced chemobrain, but also extended to methotrexate-induced chemobrain. Collectively, these findings suggest that GENUS may represent a versatile approach for treating chemobrain induced by different chemotherapy agents.

Teaser

Administering a 40-Hz sensory stimulation during chemotherapy attenuates chemobrain pathology in mouse models.

Editor’s summary

Patients with cancer undergoing chemotherapy often experience cognitive impairment or the so- called “chemobrain” that can persist long after chemotherapy has ceased. In two mouse models of chemobrain, Kim et al. now show that a noninvasive 40-Hz audiovisual stimulation designed to entrain gamma neural activity in the brain (GENUS) can ameliorate cognitive impairment and chemobrain pathology in mice. GENUS stimulation was most effective when given during chemotherapy but was less effective if given once chemobrain pathology had progressed. GENUS stimulation should be investigated further as a potential strategy to prevent chemobrain in patients with cancer. —Orla Smith

INTRODUCTION

Chemotherapy has improved survival rates and life expectancy for patients with cancer, but it often comes with considerable side effects. One such side effect is chemotherapy-related cognitive impairment or “chemobrain,” which is characterized by short-term memory loss, attention deficits, and chronic fatigue (1). More complex cognitive functions, such as executive functioning and multitasking, may be affected for months or even years after treatment ends, affecting quality of life (1). Although the condition was recognized more than 40 years ago, our understanding of the molecular and cellular mechanisms underlying chemobrain, as well as effective treatment strategies, remains limited.

Chemotherapeutic agents can cause DNA damage and epigenetic modifications (2–4) in various brain cells, leading to altered gene expression or even cell death. In addition, chronic up- regulation of proinflammatory signals outside the brain, either from peripheral immune responses to cancer cells or secondary effects of cancer treatments (4), can lead to activation of glial cells within the brain. Recent studies suggest that chemobrain may be caused by chronic activation of glial cells such as microglia and astrocytes, resulting in dysregulation of brain network formation (5).

Pharmacological treatments for chemobrain have primarily focused on common pathologies such as white matter damage and neurodegeneration (6). Multiple studies have detected demyelination due to loss of oligodendrocytes in chemobrain (4, 5, 7–9), highlighting the vulnerability of these cells and their precursors to chemotherapeutic agents. Recent research in chemobrain animal models showed that a persistent proinflammatory state of microglia led to incomplete maturation of oligodendrocyte precursor cells (OPCs) into oligodendrocytes (5) and disrupted neuron-oligodendrocyte communication (7). Although restoring the myelin sheath or preventing myelin loss due to reactive glia successfully rescued cognitive impairment in chemobrain animal models (5, 7–9), there are currently no effective therapeutic options for humans (6).

Gamma entrainment using sensory stimuli (GENUS) is a noninvasive sensory stimulation technique that uses light flickering or sound pulsating at 40 Hz to induce 40-Hz neural activity in the brain. This gamma range (30 to 90 Hz) neural activity is thought to relate to cognitive functions such as attention and memory, which are altered in cognitive disorders like Alzheimer’s disease (AD) (10). GENUS treatment has been shown to attenuate AD pathologies in several mouse models of AD (11–15). GENUS also altered the activation state of glial cells such as microglia and astrocytes in both AD mouse models and healthy mice (11–14, 16) and provided neuroprotective effects in an ischemic brain mouse model (17). These animal studies suggest that GENUS holds potential as a noninvasive therapy for treating neurodegenerative disorders. Early clinical trials show GENUS to be safe and tolerable in both cognitively normal individuals and patients with mild AD (18–20). Given the neurodegenerative nature of chemobrain, we decided to test whether GENUS could ameliorate chemobrain pathology and rescue cognitive impairment in two different mouse models of chemobrain induced by administering cisplatin or methotrexate.

RESULTS

GENUS induces 40-Hz neural activity in a cisplatin-induced mouse chemobrain model

We first examined whether GENUS could induce 40-Hz neural activity in a widely used cisplatin- based mouse model of chemobrain. This model was created using an intraperitoneal cisplatin injection schedule that mirrors clinical chemotherapy practices (8, 9). Given the association between cisplatin treatment and sensory impairments such as hearing loss and vision damage (21, 22), we sought to confirm that cisplatin-treated animals could still exhibit 40-Hz neural activity in response to auditory and visual stimuli. For this purpose, animals were implanted with three electroencephalogram (EEG) screws, evenly distributed between the visual cortex and the prefrontal cortex, regions crucial for executive function (23) that are often compromised in patients with chemobrain. After a 2-week recovery period post-surgery, the animals were administered two 5-day courses of cisplatin (2.3 mg/kg) injections, with a 5-day break in between, to induce chemobrain. On day 21, EEG recordings of the local field potential (LFP) revealed power spectral density (PSD) peaks at 40 Hz across all three monitored mouse cortex areas during GENUS stimulation (red versus black lines in Fig. 1A) and only during the period of GENUS stimulation (Fig. 1B). These findings confirm that GENUS induced 40-Hz neural activity in mice after cisplatin treatment.

Fig. 1. GENUS treatment ameliorates chemobrain pathology in mice.

(A) EEG power density was measured in the frontal, somatosensory, and visual cortices of cisplatin-treated mice (average of n = 4 animals from one experiment). (B) Representative spectrogram from the mouse visual cortex of one animal before, during, and after the 40-Hz GENUS stimulation. (C) The experimental design is shown. (D) Images show representative mouse whole-brain slices stained with Hoechst after cisplatin treatment and 40-Hz GENUS stimulation compared to control mice (NS, no stimulation; S, stimulation). Scale bar, 1 mm. (E) Mean ventricle size per mouse group in (D) (PBS S group, n = 15; other three groups, n = 16 animals per group from two replicate experiments). (F) Staining for γH2AX, NeuN, and Olig2 markers with Hoechst counterstain of mouse hippocampal CA1 area brain slices after treatment of mice with cisplatin or PBS with or without the 40-Hz GENUS stimulation is shown. Scale bar, 25 μm. (G) Mean number of γH2AX⁺ cells per mouse group in (F) (PBS NS, n = 8; PBS S, n = 6; cisplatin NS, n = 7; cisplatin S, n = 5 animals from two replicate experiments). (H) Staining for CD68, Iba1, and GFAP markers and Hoechst staining of mouse hippocampal CA1 area brain slices after treatment of mice with cisplatin or PBS with or without 40-Hz GENUS stimulation is shown. Scale bar, 25 μm. (I) Mean number of Iba1⁺ cells per mouse group in (H) (PBS NS, n = 8; PBS S, n = 7; cisplatin NS, n = 8; cisplatin S, n = 9 animals from two replicate experiments). (J) Mean volume of Iba1 CD68 staining per mouse group in (H) (PBS NS, n = 8; PBS S, n = 7; cisplatin NS, n = 8; cisplatin S, n = 9 animals from two replicate experiments). (K) Mean number of GFAP⁺ cells per mouse group in (H) (PBS NS, n = 8; PBS S, n = 7; cisplatin NS, n = 8; cisplatin S, n = 9 animals from two replicate experiments). (L) Representative transmission electron microscopy images of mouse corpus callosum after treatment with cisplatin or PBS with or without the 40-Hz GENUS stimulation. Scale bar, 1 μm. (M) Box plot (mean ⁺ 5 to 95%) of g ratio of axons per mouse group in (L) (n = 60 axons from four animals per group from one experiment). (N) Mean number of myelinated axons per mouse group in (L) (n = 4 animals per group from one experiment). Data are presented as means ± SD; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Two-way ANOVA and Tukey’s multiple-comparison test.

GENUS ameliorates neurodegenerative pathology in a chemobrain mouse model

Previous studies have demonstrated that GENUS treatment can modify neurodegenerative disease pathologies in AD mouse models (12–14). To investigate whether GENUS exerts a similar influence in a chemobrain mouse model, we administered combined visual and auditory 40-Hz stimuli for 1 hour daily, during and after the cisplatin administration period (Fig. 1C). We monitored body weight as a general health marker. Animals treated with cisplatin exhibited marked body weight loss (P < 0.0001), whereas control animals treated with PBS did not. However, GENUS did not significantly affect body weight in either group (fig. S1).

We first found that cisplatin-treated animals that had not received GENUS (cisplatin no stimulation) had significant expansion of the brain lateral ventricles compared to phosphate- buffered saline (PBS)-treated control animals with (PBS stimulation) or without (PBS no stimulation) GENUS (Fig. 1, D and E; PBS no stimulation versus cisplatin no stimulation P < 0.001; PBS stimulation versus cisplatin no stimulation, P < 0.01). Cisplatin-treated animals with stimulation showed similar brain ventricle sizes to PBS no stimulation control mice. Similar effects were seen for changes in white matter volume above the lateral ventricles in these two groups. The corpus callosum in the cisplatin-treated mice with no stimulation was significantly thinner compared to PBS no stimulation animals (fig. S2, A and B; P < 0.05; PBS no stimulation versus cisplatin with stimulation P = 0.9666).

We subsequently evaluated neuronal damage in our chemobrain mouse model. We focused on the hippocampal CA1 and prefrontal cortex areas given their well-documented importance in numerous relevant cognitive functions and their known susceptibility to chemotherapy (24–26). Cisplatin treatment significantly increased the number of cells expressing the DNA-damage marker γH2AX⁺ compared to PBS treatment; GENUS stimulation limited this increase (Fig. 1, F and G; PBS no stimulation versus cisplatin no stimulation, P < 0.0001; cisplatin no stimulation versus cisplatin stimulation P < 0.01). Furthermore, we observed a significant reduction in the number of NeuN⁺ cells in both the CA1 and prefrontal cortex areas after cisplatin treatment (fig. S2, C to E; PBS no stimulation versus cisplatin no stimulation, CA1 P < 0.01; prefrontal cortex, P < 0.01). GENUS stimulation rescued this loss of NeuN⁺ cells in the CA1 (fig. S2D; cisplatin no stimulation versus cisplatin stimulation, P < 0.01), but not in the prefrontal cortex (fig. S2E; P = 0.9991).

Because of their mitotic nature, stem cell populations are susceptible to DNA damage during chemotherapy. Reduced neurogenesis in the hippocampus has been proposed as a contributing factor to cognitive impairment in chemobrain (25). When we used Sox2 as a marker for neural stem cells (NSC) or neural progenitor cells (NPC), we observed a higher number of NSC/NPC with DNA damage in the dentate gyrus of cisplatin-treated animals without GENUS compared to PBS controls (fig. S2, F and G; PBS no stimulation versus cisplatin no stimulation, P < 0.01), as well as a lower number of NSC/NPC (fig. S2, H and I; P < 0.01). GENUS treatment rescued both of these alterations (fig. S2, G and I; cisplatin no stimulation versus cisplatin stimulation, γH2AX⁺ Sox2⁺ cells P < 0.01; Sox2⁺ cells P < 0.01). GENUS stimulation also increased the proliferating cell population in mouse brain as indicated by an increase in the Ki67 marker in both phosphate- buffered saline (PBS)–treated and cisplatin-treated groups (fig. S2, H and J; effect of GENUS, P < 0.05). The number of Ki67⁺ Sox2⁺ cells in the PBS group was higher, although the difference compared to the other groups was not significant (fig. S2, H and K). In contrast, GENUS stimulation significantly increased the number of immature neurons in the brains of both PBS- treated and cisplatin-treated mice (fig. S2, H and L; PBS no stimulation versus PBS stimulation P < 0.05; cisplatin no stimulation versus cisplatin stimulation, P < 0.01).

Cisplatin treatment in mouse models causes neurite and synaptic loss (26). We investigated whether GENUS could also protect neurons from losing synaptic connections after cisplatin treatment (26). Cisplatin treatment reduced the excitatory synaptic marker vGlut1 in both the prefrontal cortex (fig. S3, A and C; effect of cisplatin, P < 0.01) and CA1 (fig. S3, D and F; P < 0.01), while the inhibitory synaptic marker vesicular GABA transporter (VGAT) remained unaffected (fig. S3, A to E). Within the prefrontal cortex, GENUS treatment increased the staining intensities of both VGAT (PBS no stimulation versus PBS stimulation P < 0.05) and vGlut1 (PBS no stimulation versus PBS stimulation P < 0.01) markers (fig. S3, A and B). However, only VGAT staining intensity increased in the cisplatin- administered mouse groups (fig. S3, A and B; cisplatin no stimulation versus cisplatin stimulation P < 0.05).

Chronic inflammatory glial cell activation has been proposed as an underlying cause of chemobrain in methotrexate-treated mice (5, 7). To determine whether cisplatin treatment also resulted in glial activation in mouse brain, we evaluated the numbers of microglia and astrocytes in the hippocampal CA1 region along with the microglial activation marker CD68 and the reactive astrocyte marker C3 (Fig. 1H and fig. S4A). Compared to PBS-treated animals without stimulation, cisplatin-treated animals without stimulation exhibited increased numbers of Iba1⁺ microglia (Fig. 1I; P < 0.001) and glial fibrillary acidic protein (GFAP)⁺ astrocytes (Fig. 1K and fig. S4B; P < 0.0001), as well as increased microglial CD68 expression (Fig. 1J; P < 0.05) and astrocytic C3 expression (fig. S4C; P < 0.01). In contrast, cisplatin-treated animals with GENUS stimulation did not show a significant increase in any of these measures of glial activation (Fig. 1, H to K, and fig. S4, A to C). In the prefrontal cortex, we observed similar microglial behavior as in the CA1 region (fig. S4, D to F), but we did not detect increased GFAP or GFAP and C3 coexpressing cells after cisplatin treatment (fig. S4, G to I). When we stained for microglia in the corpus callosum, we observed similar effects to those seen in the hippocampal CA1 area, specifically that GENUS treatment mitigated the increased number of Iba1⁺ microglia (fig. S4, J and K) and CD68 expression (fig. S4, J and L) induced by cisplatin.

GENUS protects the mouse brain from cisplatin-induced demyelination

Demyelination of neuronal axons is one of the best characterized pathologies in cisplatin-induced chemobrain mouse models (8, 9). Consequently, we investigated whether GENUS treatment could influence the integrity of the myelin sheath in our chemobrain mouse model induced by cisplatin. Using transmission electron microscopy images of the mouse corpus callosum (Fig. 1L), we assessed the myelin sheath thickness. Our findings revealed that, compared to PBS-treated groups, cisplatin-treated animals without GENUS stimulation exhibited a significantly higher g ratio (the g ratio was calculated as the diameter of the neuronal axons divided by the diameter of myelin around the neuronal axons; a higher value indicates a thinner myelin sheath.) (Fig. 1, L and M; PBS no stimulation versus cisplatin no stimulation P < 0.0001; PBS stimulation versus cisplatin no stimulation P < 0.0001) and fewer myelinated axons (Fig. 1, L and N; PBS no stimulation versus cisplatin no stimulation, P < 0.0001; PBS stimulation versus cisplatin no stimulation, P < 0.001). Cisplatin- treated animals with GENUS stimulation demonstrated g ratios (Fig. 1M) and numbers of myelinated axons (Fig. 1N) comparable to those of animals in the PBS control groups. In the prefrontal cortex, staining for myelin basic protein (MBP) showed a significantly lower area positive for MBP in cisplatin-treated animals without GENUS stimulation (fig. S5, A and B; PBS no stimulation versus cisplatin no stimulation P < 0.05; PBS stimulation versus cisplatin no stimulation P < 0.05). However, this reduction was not observed in cisplatin-treated animals with GENUS stimulation (fig. S5, A and B; PBS no stimulation versus cisplatin stimulation, P = 0.9949). We further confirmed that MBP staining represented axonal myelin by costaining for neurofilament. We did not observe any difference in the colocalization of MBP and neurofilament staining across all four mouse groups (fig. S5, C and D).

GENUS promotes gene expression associated with cisplatin resistance

To understand better how GENUS-induced gene expression changes in different brain cells helped to alleviate chemobrain symptoms, we performed single-cell RNA sequencing (scRNA-seq) on hippocampal samples from animals treated with PBS or cisplatin with or without GENUS stimulation. After 21 days of GENUS stimulation, we collected mouse hippocampi and prepared cDNA libraries for scRNA-seq. After rigorous quality control measures, a dataset comprising 19,834 single-cell transcriptomes was curated. These transcriptomes were categorized into 12 distinct clusters after normalization (Fig. 2A). We were able to map these clusters to major cell types based on marker gene expression (fig. S6A). We observed no significant difference in cell- type abundance across animals treated with cisplatin or PBS, regardless of the presence or absence of GENUS stimulation (fig. S6, B and C).

Fig. 2. scRNA-seq reveals an altered transcriptome in oligodendrocytes in mice with chemobrain.

(A) Uniform manifold approximation and projection (UMAP) indicates sequenced single cells from mouse hippocampus (after treatment of mice with cisplatin or PBS with or without the 40-Hz GENUS stimulation) clustered into different cell types (n = 3 animals per group from two replicate experiments). (B) Number of differentially expressed genes (DEGs) across different cell types between mouse groups: cisplatin treatment versus PBS with no stimulation (Cis NS versus PBS NS) and cisplatin treatment with or without GENUS stimulation (Cis S versus Cis NS). (C) Volcano plot shows DEG in oligodendrocytes from mice treated with cisplatin with or without GENUS stimulation (Cis S versus Cis NS). (D) Pathways up-regulated and down-regulated according to GO terms in oligodendrocytes from mice treated with cisplatin with or without GENUS stimulation (Cis S versus Cis NS) are shown. (E) Up-regulated and down-regulated DEGs according to the GO terms “Positive regulation of apoptotic process” and “Negative regulation of apoptotic process” in oligodendrocytes from mice treated with cisplatin or PBS with no stimulation and cisplatin with or without GENUS stimulation (Cis NS versus PBS NS and Cis S versus Cis NS) are shown. (F) Representative images of Klf9 and Hoechst staining of the corpus callosum brain area from mice treated with cisplatin or PBS with (S) or without (NS) GENUS stimulation are shown. Scale bar, 20 μm. (G) Mean signal intensity of Klf9 staining per mouse group in (F) (PBS NS, n = 8; PBS S, n = 7; Cis NS, n = 10; Cis S, n = 9 animals from two replicate experiments). Data are presented as mean ± SD; *P < 0.05, ***P < 0.001, ****P < 0.0001. Two-way ANOVA and Tukey’s multiple-comparison tests.

We first looked at the total number of differentially expressed genes (DEGs) across cell types to determine which cell type was most affected by cisplatin treatment and by GENUS. We found that cisplatin treatment (without GENUS) caused the highest number of DEGs in oligodendrocytes compared to PBS-treated control mice (Fig. 2B and fig. S7A). This confirmed prior reports that oligodendrocytes and oligodendrocyte progenitor cells are particularly sensitive to chemotherapeutic toxicity (2, 5). GENUS treatment had little effect on any cell type in PBS-treated animals (fig. S7A), but in cisplatin-treated animals, GENUS induced the highest number of DEGs again in oligodendrocytes (Fig. 2B and fig. S7A). To explore the effect of GENUS treatment on oligodendrocyte gene expression after cisplatin treatment in more detail, we compared specific DEGs in oligodendrocytes in the hippocampus of cisplatin-induced chemobrain mice with or without GENUS. We found that genes known to provide protective effects against or to be involved in repair of cisplatin-induced damage (Cdkn1a, Ddit4, and Mt1) as well as genes known to be related to anti-inflammatory pathways and cell survival in the context of cisplatin treatment (Sgk1, Sgk3, Hif3a, Pla2g3, and Nfkbia) were up-regulated in cisplatin-treated mice with GENUS stimulation compared to no stimulation. Conversely, genes related to oligodendrocyte apoptosis due to oxidative stress (Jun, Egr1, Fos, JunB, FosB, and Dusp1) were down-regulated in cisplatin-treated mice with GENUS stimulation compared to no stimulation (Fig. 2C). We also found that genes related to the Gene Ontology (GO) terms “ATP (adenosine triphosphate) biosynthetic process,” “ATP metabolic process,” “oxidative phosphorylation,” and “aerobic respiration,” which are all functions associated with mitochondrial function, were up- regulated in cisplatin-treated mice with GENUS stimulation compared to no stimulation (Fig. 2D). This finding indicates that GENUS promotes mitochondrial function in oligodendrocytes, which is important for myelin sheath formation and maintenance, both of which are metabolically demanding processes for oligodendrocytes (27). Conversely, genes related to GO terms “response to chemical” and “regulation of apoptotic process” were down-regulated (Fig. 2D), indicating that GENUS treatment protected oligodendrocytes from cisplatin cytotoxicity. Specifically, we found that GENUS treatment promoted expression of genes involved in negative regulation of apoptotic processes (GO:0043069) but did not have much effect on expression of genes involved in positive regulation of apoptotic processes (GO:0043065) (Fig. 2E and fig. S7B).

Given that our histological data showed that GENUS treatment altered microglia and astrocyte responses to cisplatin treatment, we next analyzed DEGs in these cell types for cisplatin-treated mice with GENUS compared to no stimulation. In hippocampal microglia from cisplatin-treated mice with GENUS stimulation, we found that the cisplatin resistance gene Ddit4 and genes involved in protein translation (Rpl10-ps3, Rps27rt, Rpl23a, and Pabpc1) were up-regulated, whereas genes involved in inflammatory signaling (Atf3, Fos, Ccr5, Klf6, and Gstp1) were down- regulated (fig. S8A). We also found that GO terms related to protein translation such as “organonitrogen compound biosynthetic process,” “cellular macromolecule biosynthetic process,” “cellular amide metabolic process,” and “cellular nitrogen compound metabolic process” were up- regulated in these microglia, whereas “cellular response to chemical stimulus” was down-regulated (fig. S8C). Increased gene expression related to protein translation has been observed in phagocytic microglia in an AD mouse model (28), and phagocytic activity of microglia contributes to tissue homeostasis (29, 30) and reestablishment of myelin sheaths in demyelinating diseases by clearing the myelin debris (31, 32).

In our analysis of DEGs in hippocampal astrocytes from mice administered cisplatin and treated with GENUS, we observed an up-regulation of genes involved in the repair of cisplatin- induced oxidative damage (Mt1 and Mt2). In contrast, genes associated with astrocyte recruitment to injury sites and glial scar formation (Aqp4, Slc6a11, Eno1, and Gpr17l1) were down-regulated (fig. S8B). We also found that GO terms related to locomotion of the cell and subcellular reorganization of molecules such as “transport,” “localization,” “cell adhesion,” “organic substance transport,” and “regulation of cellular component organization” were down-regulated in these astrocytes (fig. S8D), consistent with our earlier immunohistochemical finding that GENUS treatment ameliorated astrogliosis in response to cisplatin toxicity.

To validate our DEG analysis results in oligodendrocytes, we selected two genes, Klf9 and Cdkn1a, that showed differential up-regulation in hippocampal oligodendrocytes from cisplatin- treated mice with GENUS compared to no stimulation. We analyzed their expression in the mouse brain corpus callosum. For this, we used an antibody against the Klf9 protein (Fig. 2F) and a fluorescent in situ hybridization (FISH) probe for the Cdkn1a RNA transcript (fig. S8E). We chose Klf9 for validation given that it mediates oligodendrocyte differentiation and is known to be a key factor in the recovery of the myelin sheath after demyelination (33). Cdkn1a was selected for validation because it plays an important role in oligodendrocyte maturation in addition to its well- known involvement in cell cycle regulation. In addition, a previously published single-cell transcriptomic analysis (34) reported Cdkn1a as one of the top DEGs associated with oligodendrocytes surviving a demyelination event. For both genes, we confirmed the increased expression with GENUS that was indicated by the scRNA-seq results (Fig. 2G and fig. S8F; cisplatin no stimulation versus cisplatin stimulation, Klf9, P < 0.0001; Cdkn1a, P < 0.001). Overall, our scRNA-seq results showed that GENUS treatment increased cisplatin-resistance gene expression and reduced inflammatory gene expression in multiple cell types.

GENUS promotes oligodendrocyte survival after cisplatin treatment

On the basis of the scRNA-seq results, we hypothesized that GENUS treatment might ameliorate demyelination through an effect on the mouse brain oligodendrocyte population. By staining for the marker proteins Olig2 and CC1 to identify oligodendrocytes, we observed that GENUS treatment decreased oligodendrocyte loss in the corpus callosum. Compared to the PBS control groups, the loss was reduced from 65% in the cisplatin-treated mice without GENUS to 35% in the cisplatin-treated mice with GENUS stimulation (Fig. 3, A and B; cisplatin no stimulation versus cisplatin stimulation P < 0.05). This finding indicates that GENUS reduced the damaging effect of cisplatin on mouse brain oligodendrocytes.

Fig. 3. GENUS stimulation promotes oligodendrocyte survival after cisplatin treatment in mice.

(A) Representative corpus callosum brain slices stained for CC1 and Olig2 markers with Hoechst counterstain from mice treated with cisplatin or PBS with (S) or without (NS) GENUS stimulation are shown. Scale bar, 20 μm. (B) Mean number of oligodendrocytes (CC1⁺ Olig2⁺) per square millimeter of corpus callosum in each mouse group in (A) (PBSNS, n = 8; PBS S, n = 8; PBS S, n = 8; PBS S, n = 7 animals from two replicate experiments). (C) Representative corpus callosum brain slices stained for platelet-derived growth factor receptor α (PDGFRα) and Olig2 markers with Hoechst counterstain from mice treated with cisplatin or PBS with (S) or without (NS) GENUS are shown. Scale bar, 20 μm. (D) Mean number of oligodendrocyte progenitor cells (PDGFRα⁺ Olig2⁺) per square millimeter of corpus callosum in each mouse group in (C) (n = 4 animals from one experiment). (E) Design of the experiment is shown. (F) Representative corpus callosum brain slices stained for Edu, CC1, and Olig2 markers with Hoechst counterstain from mice treated with cisplatin or PBS with (S) or without (NS) GENUS are shown. Scale bar, 50 μm. (G) Mean number of newly formed oligodendrocytes (CC1⁺ Olig2⁺ EdU⁺) per square millimeter of corpus callosum in each mouse group in (F) (PBS NS, n = 9; PBS S, n = 9; cisplatin NS, n = 5; cisplatin S, n = 9 animals from two replicate experiments). ROI, region of interest. (H) Representative corpus callosum brain slices stained for the Olig2 marker and TUNEL with Hoechst counterstain from mice treated with cisplatin or PBS with (S) or without (NS) GENUS are shown. Scale bar, 25 μm. (I) Mean number of apoptotic TUNEL⁺ Olig2⁺ cells per square millimeter of corpus callosum in each mouse group in (H) (PBS NS, n = 5; PBS S, n = 7; cisplatin NS, n = 10; cisplatin S, n = 7 animals from two replicate experiments). Data are presented as means ± SD; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Two-way ANOVA and Tukey’s multiple-comparison test.

Next, we tested whether GENUS reduced oligodendrocyte loss by promoting oligodendrocyte progenitor cell differentiation to replenish oligodendrocyte numbers. In the methotrexate-induced chemobrain mouse model, it has been reported that oligodendrocyte progenitor cells do not fully mature into oligodendrocytes, leading to an increase in oligodendrocyte progenitor cell numbers (5). We reasoned that if cisplatin treatment also induced an oligodendrocyte progenitor cell maturation arrest and GENUS increased the number of oligodendrocytes by releasing this arrest, we would see more oligodendrocyte progenitor cells (Olig2⁺ Pdgfrα⁺) in cisplatin-treated mice without GENUS compared to with GENUS stimulation. However, we found a depletion of oligodendrocyte progenitor cells after cisplatin treatment to similar numbers with or without GENUS (Fig. 3, C and D; effect of cisplatin, P < 0.01; effect of GENUS, P = 0.8059). This finding indicates that GENUS did not replenish oligodendrocytes by promoting their differentiation from existing oligodendrocyte progenitor cells. Alternatively, GENUS could have increased new oligodendrocyte progenitor cell formation and accelerated their maturation into oligodendrocytes, resulting in similar oligodendrocyte progenitor cell numbers but increased oligodendrocyte numbers. To test this hypothesis, we injected 5-ethynyl-2-deoxyuridine (EdU) into the animals during the last 3 days of the 21-day GENUS stimulation period to label oligodendrocytes originating from cells dividing during that time period (Fig. 3E). We observed a similar depletion of EdU⁺ oligodendrocytes after cisplatin treatment with or without GENUS (Fig. 3, F and G; effect of cisplatin P < 0.01; effect of GENUS P = 0.7995). Collectively, these findings argue against an effect of GENUS on the replenishment of oligodendrocytes from newly generated precursor cells.

Last, to determine whether GENUS enhanced the survival rate of oligodendrocytes during cisplatin treatment, we stained apoptotic cells using terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL). GENUS stimulation significantly reduced the number of total apoptotic cells (fig. S8, G and H; cisplatin no stimulation versus cisplatin stimulation P < 0.05) and Olig2⁺ apoptotic cells (Fig. 3, H and I; cisplatin no stimulation versus cisplatin stimulation P < 0.01) seen in the corpus callosum brain area after cisplatin treatment. This finding supported an effect of GENUS on the ability of oligodendrocytes to survive cisplatin treatment and confirmed the findings of the scRNA-seq analysis.

GENUS restores cognitive function in a cisplatin-induced chemobrain mouse model

Next, we investigated whether these molecular alterations observed after GENUS treatment improved cognitive function in a cisplatin-induced chemobrain mouse model. An open field test was performed before cognitive testing to see whether cisplatin or GENUS had any effect on locomotion and anxiety in these animals (Fig. 4, A to C). GENUS did not have any effect on animal locomotion, but cisplatin treatment significantly reduced the distance animals moved during the 10-min open field test (Fig. 4B; effect of GENUS P = 0.0792; effect of cisplatin, P < 0.0001). This reduced mobility could indicate peripheral neuropathy, a commonly documented side effect in patients receiving cisplatin (35). Conversely, our analysis revealed a pronounced effect of GENUS on the duration animals spent in the central zone in the open field test (Fig. 4C; effect of GENUS P < 0.05). This suggested a potential anxiolytic effect of GENUS, as evidenced by reduced anxiety in the open field test in both PBS- and cisplatin-administered mouse groups after GENUS treatment.

Fig. 4. GENUS treatment improves cognitive function in mice with chemobrain.

(A) The four mouse groups were subjected to the open field test. (B) The mean value of total distance moved during a 10-min period in the open field test (PBS NS, n = 8; PBS S, n = 8; cisplatin NS, n = 7; cisplatin S, n = 7 animals from two experiments) is shown. (C) The mean value of time spent in the center area during the open field test (PBS NS, n = 8; PBS S, n = 8; cisplatin NS, n = 7; cisplatin S, n = 7 animals from two experiments) is shown. (D) The four mouse groups were subjected to the novel object recognition test. (E) The mean value of the discrimination index for distinguishing between the novel and familiar object is shown (PBS NS, n = 8; PBS S, n = 7; cisplatin NS, n = 8; cisplatin S, n = 8 animals from two experiments). One animal from the PBS stimulation group was excluded as it did not spend more than 10 s exploring both objects. (F) The four mouse groups were subjected to the puzzle box test with different obstacle settings (easy, intermediate, and hard). Animals had to go through the obstacles blocking the door to gain entrance to the preferred dark chamber. (G to I) Average time to reach the dark chamber for the four mouse groups during the puzzle box test is shown (PBS NS, n = 8; PBS S, n = 8; cisplatin NS, n = 7; cisplatin S, n = 7 animals from two experiments). The average time to reach the dark chamber over three trials is shown for the easy obstacle (G), the intermediate obstacle (H), and the hard obstacle (I). Data are presented as means ± SD; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Two- way ANOVA and Tukey’s multiple-comparison test. Images were created with Biorender.com.

Short-term memory impairment and attention deficits rank among the most prevalent symptoms experienced by cancer patients with chemobrain (1). To evaluate these symptoms in our chemobrain mouse model, we used a modified version of the novel object recognition test, which was tailored to be more sensitive to attention and short-term memory deficits (Fig. 4D) (5). We observed that the cisplatin-treated mice without GENUS showed no preference between a novel and familiar object, whereas the PBS-treated mice clearly favored the novel object (Fig. 4D; PBS no stimulation versus cisplatin no stimulation P < 0.05). The cisplatin- treated mice receiving GENUS also demonstrated a marked preference for the novel object, exhibiting a discrimination ratio comparable to the PBS control groups (Fig. 4E; PBS no stimulation versus cisplatin stimulation, P = 0.8414). This underscores the potential of GENUS treatment for mitigating cisplatin-induced short-term memory deficits in this experimental design.

Impaired executive function is another key symptom in cancer patients with chemobrain (1). In rodents, executive function can be assessed with the puzzle box test. This test challenges animals with obstacles of varying levels of difficulty to enter their preferred dark chamber (36). Previous studies have shown that in the cisplatin-induced chemobrain mouse model, animals show impaired performance in the puzzle box test (8, 9). To investigate whether GENUS could improve cisplatin- induced impairment in executive function, we used three obstacles of increasing difficulty: an open corridor (easy), a corridor buried under sawdust (intermediate), and a door plugged with tissue (hard) (Fig. 4F). Without GENUS, cisplatin-treated animals required a longer time to enter the dark box compared to PBS control animals (Fig. 4, G to I; PBS no stimulation versus cisplatin no stimulation, easy P < 0.01; intermediate P < 0.05; hard P < 0.05). GENUS treatment rescued the ability of cisplatin-treated animals to solve the intermediate and hard tasks (Fig. 4, H and I; PBS no stimulation versus cisplatin stimulation, intermediate task P = 0.3660; hard task P = 0.9974) but did not increase their performance in the easy task (Fig. 4G; P < 0.05), possibly because their impaired locomotion became rate limiting.

GENUS treatment during chemotherapy rescues chronic chemobrain pathology in mice

Patients with chemobrain often grapple with the condition for an extended period, which frequently persists for more than 6 months after the cessation of chemotherapy. Therefore, we determined whether GENUS treatment during chemotherapy could prevent development of chemobrain once chemotherapy and GENUS had ended. We designed experimental mouse groups with an extended timeline spanning up to 133 days after initial cisplatin administration (table S1 and Figs. 5 and 6).

Fig. 5. GENUS treatment during chemotherapy prevents progression of chemobrain pathology in mice.

(A) The experimental design for administering GENUS with cisplatin treatment and analysis in three behavioral tests between days 126 and 133 since the start of cisplatin treatment is shown. (B) The mean value of total distance moved during a 10-min period in the open field test is shown (PBS NS, n = 9; PBS S, n = 9; cisplatin NS, n = 15; cisplatin S, n = 15 animals from one experiment). (C) The mean value of time spent in the center area during the open field test is shown (PBS NS, n = 9; PBS S, n = 9; cisplatin NS, n = 15; cisplatin S, n = 15 animals from one experiment). (D) The four mouse groups were subjected to the novel object recognition test and the mean value of the discrimination index for distinguishing between the novel and familiar object was calculated (PBS NS, n = 10; PBS S, n = 10; cisplatin NS, n = 15; cisplatin S, n = 15 animals from one experiment). (E to G) The four mouse groups were subjected to the puzzle box test with different obstacle settings (easy, intermediate, and hard) (PBS NS, n = 9; PBS S, n = 9; cisplatin NS, n = 15; cisplatin S, n = 15 animals from one experiment). The average time to reach the dark chamber over three trials is shown for the easy obstacle (E), the intermediate obstacle (F), and the hard obstacle (G). (H) Representative images of mouse prefrontal cortex brain slices stained for MBP and NF-09 with 4′,6-diamidino-2-phenylindole (DAPI) counterstain. Scale bar, 50 μm. (I) Mean prefrontal cortex area staining positive for MBP per mouse group in (H) (PBS NS, n = 16; PBS S, n = 12; cisplatin NS, n = 12; cisplatin S, n = 14 animals from two replicate experiments). (J) Representative images of corpus callosum brain slices stained for CC1, PDGFRα, and Olig2 with DAPI counterstain are shown. Scale bar, 20 μm. (K and L) The number of CC1⁺ Olig2⁺ cells (K) and the number of PDGFRα⁺ Olig2⁺ cells (L) per mouse group in (J) (PBS NS, n = 12; PBS S, n = 10; cisplatin NS, n = 13; cisplatin S, n = 12 animals from two replicate experiments) are shown. (M) Representative images of hippocampal CA1 brain slices stained for CD68 and Iba1 with DAPI counterstain are shown. Scale bar, 20 μm. (N and O) The mean number of the Iba1⁺ cells (N) and mean volume of Iba1⁺ CD68⁺ area (O) in the hippocampal CA1 region per mouse group in (M) (PBS NS, n = 7; PBS S, n = 6; cisplatin NS, n = 7; cisplatin S, n = 7 animals from one experiment) is shown. (P) Representative images of hippocampal CA1 brain slices stained for C3 and GFAP with DAPI counterstain are shown. Scale bar, 20 μm. (Q) The mean hippocampal area staining positive for GFAP per mouse group in (P) (n = 10 animals per group from one experiment) is shown. (R) Mean number of C3⁺ GFAP⁺ cells per mouse group in (P) (n = 10 animals per group from one experiment). Data are presented as means ± SD; *P < 0.05, **P < 0.01, and ***P < 0.001. Two-way ANOVA and Tukey’s multiple-comparison test.

Fig. 6. GENUS treatment after chemotherapy has limited effect on chemobrain pathology.

(A) The experimental design to test whether GENUS treatment starting at day 105 after initiation of cisplatin chemotherapy could ameliorate chemobrain pathology in mice is shown. (B) The mean value of total distance moved during a 10-min period in the open field test for the four mouse groups is shown (n = 8 animals per group from one experiment). (C) The mean value of time spent in the center area during the open field test for the four mouse groups is shown (n = 8 animals per group from one experiment). (D) The four mouse groups were subjected to the novel object recognition test and the mean value of the discrimination index for distinguishing between the novel and familiar object was calculated (n = 8 animals per group from one experiment). (E to G) The four mouse groups were subjected to the puzzle box test with different obstacle settings (easy, intermediate, and hard) (n = 8 animals per group from one experiment). The average time to reach the dark chamber over three trials is shown for the easy obstacle (E), the intermediate obstacle (F), and the hard obstacle (G). (H) Representative images of prefrontal cortex brain slices stained for MBP and NF-09 with DAPI counterstain are shown. Scale bar, 50 μm. (I) Mean prefrontal cortex area staining positive for MBP in the four mouse groups (PBS NS, n = 10; PBS S, n = 10; cisplatin NS, n = 10; cisplatin S, n = 11 animals from two experiments). (J) Representative images of corpus callosum brain slices stained for CC1, PDGFRα, and Olig2 with DAPI counterstain are shown. Scale bar, 20 μm. (K and L) The number of CC1⁺ Olig2⁺ cells (K) and the number of PDGFRα⁺ Olig2⁺ cells (L) in each mouse group in (J) (PBS NS, n = 8; PBS S, n = 8; cisplatin NS, n = 10; cisplatin S, n = 11 animals from two experiments). (M) Representative images of hippocampal CA1 brain slices stained for CD68 and Iba1 with DAPI counterstain are shown. Scale bar, 20 μm. (N and O) The mean number of the Iba1⁺ cells (N) and mean volume of Iba1⁺ CD68⁺ area (O) in the hippocampal CA1 region per mouse group in (M) (PBS NS, n = 8; PBS S, n = 8; cisplatin NS, n = 9; cisplatin S, n = 10 animals from two experiments) is shown. (P) Representative images of hippocampal CA1 brain slices stained for C3 and GFAP with DAPI counterstain are shown. Scale bar, 20 μm. (Q) The mean hippocampal area staining positive for GFAP per mouse group in (P) (PBS NS, n = 8; PBS S, n = 8; cisplatin NS, n = 8; cisplatin S, n = 6 animals from one experiment) is shown. (R) Mean number of C3⁺ GFAP⁺ cells per mouse group in (P) (PBS NS, n = 8; PBS S, n = 8; cisplatin NS, n = 8; cisplatin S, n = 6 animals from one experiment). Data are presented as means ± SD; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Two-way ANOVA and Tukey’s multiple-comparison test.

We concurrently gave cisplatin and GENUS treatment to mice as in previous cohorts but tested the animals 126 days after the first cisplatin injection (Fig. 5A). At this time point, we did not observe any effect of cisplatin or GENUS treatment on locomotion (Fig. 5B) or anxiety (Fig. 5C) in the open field test. However, cisplatin-treated animals without GENUS were still unable to discriminate novel and familiar objects in the novel object recognition test (Fig. 5D; PBS no stimulation versus cisplatin no stimulation, P < 0.05). Cisplatin-treated animals performed comparably to the PBS control group in easy (Fig. 5E) and intermediate (Fig. 5F) puzzle box test obstacles but required more time to overcome the hard obstacle (Fig. 5G; effect of cisplatin P < 0.05), demonstrating prolonged cognitive impairment after cisplatin treatment. Cognitive deficits in both the novel object recognition test and puzzle box test in the cisplatin-treated mouse groups were rescued by GENUS treatment, indicating a long-lasting effect of GENUS (Fig. 5, D and G; cisplatin no stimulation versus cisplatin stimulation, novel object recognition P < 0.05; puzzle box test, hard P < 0.05).

Next, we investigated chemobrain pathology. In the cisplatin-treated mouse group without GENUS, demyelination (Fig. 5, H and I; PBS no stimulation versus cisplatin no stimulation P < 0.001) and a reduction in mature oligodendrocytes (Fig. 5, J and K; P < 0.05) were observed. In addition, an increase in microglia number (Fig. 5, M and N, and fig. S9, B and C; PBS no stimulation versus cisplatin no stimulation, CA1 P < 0.001; corpus callosum P < 0.01) and CD68 expression (Fig. 5, M and O, and fig. S9, B and D; CA1 P < 0.05, corpus callosum P < 0.0001) and an increase in astrocyte GFAP expression and activation state in CA1 (Fig. 5, P to R; GFAP⁺ cells P < 0.001, GFAP⁺ C3⁺ cells P < 0.05) persisted at 133 days after the start of cisplatin treatment, showing the prolonged effects of chemobrain. In the cisplatin-treated group with GENUS, myelination (Fig. 5, H and I), oligodendrocyte numbers (Fig. 5, J and K), and microglia numbers (Fig. 5, M and N, and fig. S9, B and C) were normalized. However, there was no effect of GENUS treatment on microglial activation state (Fig. 5, M and O, and fig. S9, B and D), and astrocyte GFAP expression and activation state remained elevated (Fig. 5, P to R). We found no effect of cisplatin or GENUS treatment on the number of oligodendrocyte progenitor cells (Fig. 5, J and L). Last, we stained for VGAT and vGlut1, markers of inhibitory and excitatory synapses, respectively. In both prefrontal cortex (fig. S9, E to G) and CA1 (fig. S9, H to J), cisplatin treatment increased VGAT staining intensity (fig. S9, F and I; effect of cisplatin, prefrontal cortex P < 0.01 and CA1 P < 0.0001) but reduced vGlut1 staining intensity (fig. S9, G and J; effect of cisplatin, prefrontal cortex P < 0.05 and CA1 P < 0.01). Notably, GENUS treatment did not exhibit any discernible effect on these changes. Thus, GENUS treatment during cisplatin chemotherapy could prevent some types of cognitive impairment and reduce some chemobrain pathologies, but certain pathologies, like glial activation and synaptic loss, reemerged after the cessation of GENUS treatment (table S1).

GENUS treatment after chemotherapy yields limited effects on chemobrain pathology

To address whether GENUS could be of benefit after chemotherapy, we first treated a cohort of mice with cisplatin and then gave GENUS treatment from day 105 after cisplatin initiation (Fig. 6A). In these mice, we did not observe any effect of cisplatin or GENUS treatment on locomotion (Fig. 6B) or anxiety (Fig. 6C) during the open field test. Post-chemotherapy GENUS treatment did not rescue cognitive functions as measured by the novel object recognition test (Fig. 6D; effect of cisplatin *P < 0.05; effect of GENUS P = 0.7235). Cisplatin-treated animals performed comparably to PBS controls with easy (Fig. 6E) and intermediate (Fig. 6F) obstacles in the puzzle box task but required more time to overcome the hard puzzle box obstacle (Fig. 6G; effect of cisplatin *P < 0.05). Unlike the novel object recognition test, GENUS treatment did rescue the ability of cisplatin-treated animals to solve the hard puzzle box task (Fig. 6G; cisplatin no stimulation versus cisplatin stimulation P < 0.01).

Next, we investigated chemobrain pathology. GENUS treatment 3 months after cisplatin chemotherapy appeared to ameliorate the demyelination pathology observed after cisplatin treatment, but the difference was not significant between the cisplatin-treated groups that did or did not receive GENUS (Fig. 6, H and I). GENUS treatment after chemotherapy was also unable to rescue oligodendrocyte loss in the corpus callosum (Fig. 6, J and K), decrease the elevated microglia number in the CA1 (Fig. 6, M and N), or reduce astrogliosis in the CA1 (Fig. 6, P to R). However, GENUS treatment after chemotherapy reduced the Iba1⁺ cell count (fig. S10, B and C; PBS no stimulation versus cisplatin no stimulation P < 0.05; PBS no stimulation versus cisplatin stimulation P = 0.9692) and CD68 staining (fig. S10, B and D; cisplatin no stimulation versus cisplatin stimulation, P < 0.01) in the corpus callosum of cisplatin-treated animals, as well as CD68 staining in the CA1 (Fig. 6, M and O; P < 0.0001). We found no effect from cisplatin or GENUS treatment on the number of oligodendrocyte progenitor cells (Fig. 6, J and L). When we stained for synaptic markers in the prefrontal cortex (fig. S10E), we observed a notable increase in VGAT staining intensity (fig. S10F; effect of cisplatin P < 0.01) and a decrease in vGlut1 staining intensity in response to cisplatin treatment compared to the PBS group (fig. S10G; effect of cisplatin P < 0.001). Whereas GENUS treatment did not influence the cisplatin-treated animals, it did elevate both VGAT (fig. S10F; PBS no stimulation versus PBS stimulation, P < 0.05) and vGlut1 (fig. S10G; PBS no stimulation versus PBS stimulation, P < 0.05) staining intensity in PBS-treated animals. In the CA1 region, GENUS treatment appeared to normalize the VGAT staining intensity in cisplatin-treated animals (fig. S10, H and I; PBS no stimulation versus cisplatin no stimulation P < 0.05; PBS no stimulation versus cisplatin stimulation P > 0.9999). However, there was no discernible effect of GENUS on the diminished vGlut1 staining intensity resulting from cisplatin treatment (fig. S10, H and J).

Effects of GENUS in the methotrexate-induced chemobrain mouse model

Next, we tested whether the effect of GENUS was specific to cisplatin-induced chemobrain in mice or also extended to chemobrain caused by a different chemotherapeutic agent. We selected methotrexate because it has a different mechanism of action than cisplatin. Cisplatin creates cross- links with the DNA, damaging and preventing the synthesis of DNA. Methotrexate inhibits the enzyme dihydrofolate reductase (DHFR) and thereby disrupts purine synthesis (37). Methotrexate was administered in a schedule analogous to therapeutic timelines, and GENUS was given from the first day animals received methotrexate (100 mg/kg). Methotrexate-treated animals showed slower body weight gain during drug administration, but GENUS had no effect on their body weight during or after the treatment period (fig. S11A). We first tested whether the methotrexate- induced chemobrain mouse model would show 40-Hz entrainment with sensory stimulation. EEG recordings showed a peak response in power spectral density at 40 Hz across all three monitored cortex areas during GENUS stimulation (Fig. 7A) and only during the period of GENUS stimulation (Fig. 7B). These findings confirmed that GENUS induced a 40-Hz neural activity in mouse brain after methotrexate treatment (Fig. 7, A and B).

Fig. 7. GENUS treatment ameliorates chemobrain pathology in the methotrexate- induced chemobrain mouse model.

(A) EEG power density was measured in the frontal, somatosensory, and visual cortices of methotrexate-treated mice (n = 4 animals per group from one experiment). (B) Representative spectrogram from the mouse visual cortex of one animal before, during, and after the 40-Hz GENUS stimulation. (C) Experimental design for GENUS treatment of mice initiated at the start of methotrexate (MTX) treatment. (D) Transmission electron microscopy images of mouse corpus callosum after treating mice with methotrexate or PBS with or without GENUS stimulation. Scale bar, 1 μm. (E) Box plot (mean ⁺ 5 to 95%) of g ratio is shown (n = 30 axons from three animals from one experiment). (F) Representative images of corpus callosum brain slices stained for CC1 and Olig2 with Hoechst counterstain are shown. Scale bar, 50 μm. (G) The number of CC1⁺ Olig2⁺ cells in the mouse groups in (F) (PBS NS, n = 12; PBS S, n = 11; MTX NS, n = 10; MTX S, n = 10 animals from two experiments) is shown. (H) Representative images of hippocampal CA1 brain slices stained for CD68 and Iba1 with Hoechst counterstain are shown. Scale bar, 20 μm. (I and J) The mean volume of the Iba1⁺ area (PBS NS, n = 12; PBS S, n = 12; MTX NS, n = 15; MTX S, n = 14 animals from three experiments) (I) and of the Iba1⁺ CD68⁺ area (PBS NS, n = 7; PBS S, n = 6; MTX NS, n = 7; MTX S, n = 7 animals from two experiments) (J) per cubic millimeter of hippocampal CA1 per mouse group in (H) is shown. (K) The four mouse groups were subjected to the novel object recognition test and the mean value of the discrimination index for distinguishing between the novel and familiar object was calculated (PBS NS, n = 8; PBS S, n = 8; MTX NS, n = 7; MTX S, n = 7 animals from two experiments). (L to N) The four mouse groups were subjected to the puzzle box test with different obstacle settings (easy, intermediate, and hard) (PBS NS, n = 8; PBS S, n = 8; MTX NS, n = 12; MTX S, n = 12 animals from two experiments). The average time to reach the dark chamber over three trials is shown for the easy obstacle (L), the intermediate obstacle (M), and the hard obstacle (N). Data are presented as means ± SD; *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Two-way ANOVA and Tukey’s multiple-comparison test.

We then investigated the response of methotrexate-induced chemobrain animals to GENUS treatment (Fig. 7C) based on histological features. Similar to the findings in the cisplatin-induced chemobrain mouse model, demyelination after methotrexate treatment was rescued by GENUS (Fig. 7, D and E; PBS no stimulation versus methotrexate no stimulation P < 0.0001; methotrexate no stimulation versus methotrexate stimulation P < 0.05). The number of oligodendrocytes was decreased in the methotrexate-treated mice without GENUS compared to the GENUS-treated animals (Fig. 7, F and G; PBS no stimulation versus methotrexate no stimulation P < 0.05; methotrexate no stimulation versus methotrexate stimulation P < 0.01). Iba1 and CD68 expression was increased in the corpus callosum of methotrexate-treated mice without GENUS but decreased with GENUS treatment (Fig. 7, H to J; PBS no stimulation versus methotrexate no stimulation, Iba1⁺ P < 0.0001; Iba1⁺ CD68⁺ P < 0.0001; methotrexate no stimulation versus methotrexate stimulation, Iba1⁺ P < 0.0001; Iba1⁺ CD68⁺ P < 0.05).

Next, we tested whether GENUS could also rescue cognitive impairment in methotrexate- treated animals. Unlike cisplatin, methotrexate did not alter the total distance moved in the open field test, indicating no effect of methotrexate on animal locomotion (fig. S11B). However, the anxiolytic effect of GENUS seen in cisplatin-treated animals was also observed in the methotrexate-induced chemobrain mice, as indicated by an increase in time spent in the center of the arena (fig. S11C; effect of GENUS P < 0.05). In the novel object recognition test assessing short-term memory and attention, methotrexate-treated animals without GENUS showed no preference between novel and familiar objects. However, methotrexate-treated animals with GENUS spent significantly more time exploring the novel object, displaying a discrimination ratio similar to that of the PBS control groups (Fig. 7K; PBS no stimulation versus methotrexate no stimulation P < 0.01; PBS no stimulation versus methotrexate stimulation P > 0.9999). Methotrexate-treated mice with GENUS outperformed their untreated counterparts in the intermediate and hard puzzle box obstacle tasks (Fig. 7, L to N; PBS no stimulation versus methotrexate no stimulation, intermediate P < 0.01; hard P < 0.001; PBS no stimulation versus methotrexate stimulation, intermediate P = 0.3025; hard P = 0. 3476).

Last, we tested which brain pathologies related to methotrexate-based chemobrain could be detected in mouse groups treated or not treated with GENUS 133 days after the first of methotrexate injection, GENUS was administered for 21 days starting from the first day of methotrexate or PBS injection (figs. S12A and S13A). Methotrexate treatment reduced the number of mature oligodendrocytes (fig. S12, B and C; effect of methotrexate P < 0.0001) and increased the number of oligodendrocyte progenitor cells (fig. S12, B and D; effect of methotrexate P < 0.001). In contrast to our findings in the cisplatin-based chemobrain mouse model, we did not observe any effect of GENUS treatment on the number of oligodendrocytes or oligodendrocyte progenitor cells (fig. S12, C and D; two-way ANOVA, effect of GENUS, number of oligodendrocytes P = 0.2986; number of oligodendrocyte progenitor cells P = 0.3282). When we looked at microglial activation in the CA1 area, we found that methotrexate increased both the number of Iba1⁺ cells (fig. S12, E and F; effect of methotrexate P < 0.05) and CD68 expression (fig. S12, E and G; effect of methotrexate P < 0.0001) compared to PBS groups. GENUS treatment did not have any effect on the number of Iba1⁺ cells (fig. S12F; effect of GENUS P = 0.9786) but did decrease CD68 expression (fig. S12G; methotrexate no stimulation versus methotrexate stimulation P < 0.05). The corpus callosum area showed a similar trend with increased Iba1 (fig. S12, H and I; effect of methotrexate, P < 0.001) and CD68 expression (fig. S12, H and J; PBS no stimulation versus methotrexate no stimulation P < 0.001) after methotrexate treatment, and a reduction in CD68 expression after GENUS treatment (fig. S12J; effect of GENUS P < 0.05). Astrocytes showed a similar response in the corpus callosum to methotrexate and GENUS treatment. The area positive for GFAP staining increased in both methotrexate groups (with and without GENUS) compared to PBS-treated groups (fig. S12, K and L; effect of methotrexate P < 0.001), with GENUS treatment reducing the number of C3⁺ GFAP⁺ cells (fig. S12, K and M; methotrexate no stimulation versus methotrexate stimulation P < 0.05). When we measured synaptic markers in this cohort, methotrexate treatment caused a reduction of VGAT and vGlut1 staining intensity in prefrontal cortex (fig. S13, B to D; PBS no stimulation versus methotrexate no stimulation, VGAT P < 0.0001; vGlut1 P < 0.001) and CA1 areas (fig. S13, E to G; PBS no stimulation versus methotrexate no stimulation, VGAT P < 0.0001; vGlut1 P < 0.0001), and GENUS treatment rescued VGAT staining intensity in both regions (fig. S13, C and F; PBS no stimulation versus methotrexate no stimulation, prefrontal cortex P < 0.0001; CA1 P < 0.001).

Last, we examined whether GENUS treatment after chemotherapy could mitigate chemobrain pathology in the methotrexate-induced chemobrain model (figs. S14A and S15A). We found no effect of GENUS treatment on the number of oligodendrocytes or oligodendrocyte progenitor cells (fig. S14, B to D). The number of microglia in the CA1 area (fig. S14, E and F) was comparable between methotrexate-treated mice with or without GENUS, whereas GENUS treatment reduced CD68 expression (fig. S14, E and G, methotrexate no stimulation versus methotrexate stimulation P < 0.01). In the corpus callosum, we found that methotrexate treatment increased both Iba1 expression (effect of methotrexate P < 0.001) and CD68 expression (P < 0.05), but there was no effect from GENUS treatment (fig. S14, H to J). We also did not find any effect of GENUS treatment on astrogliosis and astrocytic C3 expression (fig. S14, K to M). When we measured synaptic markers, we did not find any effect of GENUS on the prefrontal cortex (fig. S15, B to D). However, we found that only methotrexate-treated animals without GENUS had lower VGAT (PBS no stimulation versus methotrexate no stimulation P < 0.05) and vGlut1 (P < 0.001) staining intensity in the CA1 area (fig. S15, E to G).

DISCUSSION

A correctly formed and maintained myelin sheath has an increasingly recognized role in cognitive function (38). Studies in animal models of chemobrain have shown that restoration of oligodendrocyte populations and myelin sheaths can ameliorate cognitive impairment (5, 7–9). Our analyses suggest that GENUS treatment rescued demyelination in mice with chemobrain and promoted oligodendrocyte survival. Some reports from other demyelinating disease animal models suggest that surviving oligodendrocytes do not contribute as much myelin as new oligodendrocytes (39). However, our transmission electron microscopy results show that the surviving oligodendrocytes in the animals with chemobrain given GENUS treatment formed functional myelin sheaths. The observation that GENUS treatment rescued cognitive impairment in mice with chemobrain also supports the notion that surviving oligodendrocytes are able to maintain a functional myelin sheath.

For our transcriptional analyses, we opted for scRNA-seq over single-nucleus RNA-seq (snRNA-seq) to capture cytosolic transcripts and to enrich for glial cell types that are underrepresented in snRNA-seq data. Unfortunately, this approach resulted in a trade-off in terms of adequately representing neurons, preventing us from drawing meaningful conclusions about the neuronal transcriptional profile. However, our histological analysis demonstrated that GENUS treatment could protect against neuronal damage and enhance neurogenesis.

Whereas there have been a few successful attempts to rescue cognitive impairment in chemobrain mouse models (5, 7–9, 40, 41), breakthroughs in treating human patients have been elusive, primarily due to the invasive nature of current treatments. GENUS represents a noninvasive method that has already been demonstrated to be safe in human patients (18–20), potentially accelerating clinical trials. We observed a beneficial effect of GENUS in both cisplatin- and methotrexate-induced chemobrain mouse models during the early phase of pathology. However, careful consideration will need to be given to the selection of candidate patients and the timing of GENUS treatment. Our results suggest that GENUS would most benefit patients when administered during chemotherapy to prevent the progression of the pathology, whereas the benefit would be limited for patients who have already undergone chemotherapy but still exhibit chemobrain symptoms. This preventative approach would be well suited for patients with cancer about to undergo chemotherapy, as clinicians could schedule GENUS treatment simultaneously with chemotherapy.

Persistent long-term pathology after concurrently administered GENUS differed between the cisplatin-based and methotrexate-based chemobrain mouse models. In the methotrexate-based model, the effect of GENUS on oligodendrocytes and microglia seemed to diminish over time, while excitatory and inhibitory synaptic markers were rescued with GENUS treatment. The varying responsiveness to GENUS treatment between the two chemobrain mouse models might stem from the distinct molecular mechanisms that lead to similar pathological symptoms. Cisplatin can diffuse through the cellular membrane and form cisplatin-DNA cross-links that inhibit cellular growth and division (42). On the other hand, methotrexate is taken up by cells through reduced folate carrier 1 (RFC1) and inhibits enzymes such as DHFR and thus purine biosynthesis, impairing DNA replication and repair (43). Methotrexate competes with folate for binding sites and causes folate deficiency, which results in a microglial inflammatory state (44). This could have resulted in the return of chemobrain symptoms after the cessation of GENUS treatment in our study.

We demonstrate the potential of GENUS as a treatment for chemobrain, but there are limitations to our study. We have demonstrated that 21 days of GENUS stimulation can ameliorate chemobrain pathology in our mouse models. However, it remains an open question regarding the optimal treatment duration. Prolonging GENUS stimulation beyond 21 days might enhance the mitigation of chemobrain pathology and prevent its return at day 133 in the chemobrain mouse models. It has already been shown in a clinical trial with human patients with AD that long-term treatment with GENUS for more than 3 months is safe and well-tolerated (20). Therefore, detailed studies on the timeline for GENUS treatment of chemobrain are needed. For our two animal models, we only used adult female mice for our cisplatin-induced chemobrain model and young males for the methotrexate-induced chemobrain model as these chemobrain models have been well- documented in the literature (5, 8). A more detailed study is required to understand whether there are any sex- or age-dependent responses to GENUS treatment. In addition, we did not test whether there were any frequency-specific responses or potential side effects of GENUS treatment in this study.

Prior research has highlighted the therapeutic potential of sensory stimulation or transcranial alternating current stimulation at 40 Hz in mitigating AD pathology in both animal models and human patients (12–15, 18, 20, 45). These benefits were not observed at frequencies of 8 or 80 Hz. Furthermore, sensory stimulation centered around 40 Hz, but not at 10 or 80 Hz, has been shown to alleviate symptoms of other neurological conditions, such as ischemic brain injury (17). This suggests that the 40-Hz frequency might offer therapeutic advantages for a range of neurological disorders. Consistent with these findings, our study revealed that 40-Hz audio and visual stimulation successfully reduced chemobrain pathology in mice. However, given that we have not tested other frequencies in our chemobrain mouse models, the specificity of the observed effects to the 40-Hz frequency still needs to be tested.

MATERIALS AND METHODS

Study design

This study aimed to validate in two mouse models whether GENUS stimulation (auditory and visual stimuli at 40 Hz) could be an effective treatment for chemobrain induced by chemotherapy. For the initial approach to validate whether GENUS stimulation could ameliorate chemobrain pathology and cognitive impairment, animals received cisplatin chemotherapy with GENUS treatment for 21 days. Then, analyses in long-term mouse cohorts were performed to validate the chronic effect of GENUS treatment. These mouse cohorts received GENUS treatment with chemotherapy and cognition and pathology were analyzed 133 days after initiation of chemotherapy. To test whether GENUS treatment could be of benefit after chemotherapy had ended, we treated mice with cisplatin or methotrexate and then gave GENUS treatment for 21 days starting at day 105 after chemotherapy initiation and then analyzed cognition and chemobrain pathology in the mice. Experimenters were blinded when analyzing animal behavior or histological data. All animals were handled and euthanized in accordance with the Massachusetts Institute of Technology Committee on Animal Care. Power analysis for sample size was not performed before the study. All animals in the experiments were included in the data unless specified or the animal was euthanized/deceased before the end point of the cohort because of their poor condition after chemotherapy.

Chemobrain mouse models

Animals were purchased from the Jackson Laboratory and were randomly assigned to each experimental group. Nine-week-old C57/BL6J female mice received intraperitoneal injections of either cisplatin (2.3 mg/kg) or volume-matched PBS for five consecutive days, had a 5-day resting period, and then received another 5 days of cisplatin (2.3 mg/kg) or volume-matched PBS. Postnatal day 21 C57/BL6J male mice received intraperitoneal injection of either methotrexate (100 mg/kg) or volume-matched PBS on postnatal days 21, 28, and 35.

GENUS stimulation

A light-emitting diode (LED) strip of 6000K light temperature was used to generate the visual stimulus. Visual stimulus was a 40-Hz LED flicker with 50% duty cycle (12.5 ms light on, 12.5 ms light off) and 800-lux brightness at the center of the stimulation cage. Auditory stimulus was a 1- ms-long 10-kHz tone played every 25 ms. The stimulation group received 1 hour of 40-Hz audio and visual sensory stimulation immediately after the first injection of chemotherapy (or 105 days after the first injection depending on the experimental design); they then received 1 hour of GENUS stimulation daily until the day before they were euthanized. Animals in the group without stimulation were placed in a room with dim light and ambient sound for 1 hour immediately after the first chemotherapy injection (or 105 days after the first injection depending on the experimental design); the animals stayed in a room with dim light and ambient sound for 1 hour every day until the day before they were euthanized. Each animal was single-caged in a cage covered with a black plastic bag on every side but one side for the duration of the 1-hour stimulation or no stimulation period and then were group housed with their littermates.

Statistical analysis

Data are presented as means ± SD; level of significance was set at *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Two-way ANOVA was used for statistical analysis of the effect of chemotherapy drug and GENUS treatment. Tukey’s multiple-comparison test was used to compare each group mean when possible. Three-way ANOVA was used for statistical analysis of the effect of chemotherapy drug, GENUS treatment, and time. Sample number, replication method, and statistical analysis used for each figure are listed in table S2. All samples or animals were included in the statistical analysis unless otherwise specified.

Data and materials availability:

All data associated with this study are present in the paper or the Supplementary Materials. scRNA-seq data are deposited at NCBI Gene Expression Omnibus (GEO) under the accession number GSE216146.