Meningeal-tumor interactions define distinct modes of leptomeningeal colonization in medulloblastoma

doi:10.21203/rs.3.rs-8042018/v1

Download PDF

Research Article

Meningeal-tumor interactions define distinct modes of leptomeningeal colonization in medulloblastoma

https://doi.org/10.21203/rs.3.rs-8042018/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Medulloblastoma, the most common malignant brain tumor in children, shows a pronounced tendency to spread to the leptomeninges. Leptomeningeal metastasis accounts for nearly all medulloblastoma-related deaths, yet the cellular and molecular mechanisms driving this process remain poorly understood. Progress has been hindered by limited access to patient samples, the fragile anatomy of the leptomeninges, and the lack of robust preclinical models.

Here, we developed an in vitro model that captures key features of the early leptomeningeal niche. We demonstrate that human meningeal cells promote medulloblastoma survival and proliferation under nutrient-deprived conditions both in vitro and in vivo. Using this system, we uncovered mechanisms that govern distinct patterns of leptomeningeal colonization in vivo. This physiologically informed and experimentally validated model offers a tractable platform for elucidating the molecular basis of leptomeningeal colonization and for identifying therapeutic vulnerabilities that can prevent medulloblastoma dissemination.

Medulloblastoma

leptomeningeal metastasis

adhesion

co-culture models

Nearly all deaths from medulloblastoma, the most common malignant brain tumor of childhood, are due to leptomeningeal metastasis [1–5]. The biological mechanisms that enable tumor cells to survive and expand within the unique leptomeningeal compartment remain poorly defined[6]. Consequently, therapeutic strategies capable of eradicating disseminated tumor cells from the leptomeninges are lacking, particularly for patients who relapse after standard-of-care therapy[7].

The leptomeninges, comprising the pia mater, subarachnoid space (filled with cerebrospinal fluid - CSF), and arachnoid mater, were once viewed primarily as structural coverings to the brain and spinal cord but are now recognized as dynamic microenvironments that actively participate in central nervous system health and disease, including cancer[8–11]. Within the CSF, medulloblastoma cells exist in at least two states: free-floating within the fluid or adherent to the leptomeningeal surface of the brain and spinal cord. Adherent lesions may form diffuse, laminar coatings or discrete nodules along the leptomeninges, yet the mechanisms governing these distinct metastatic patterns remain unknown[12, 13]. This knowledge gap underscores the need to understand how interactions between tumor cells and leptomeningeal stromal elements influence colonization, survival, and progression.

Advances in leptomeningeal metastasis biology have been hindered by several experimental and clinical barriers[14]. Surgical access to leptomeningeal lesions is limited due to their diffuse nature and lack of therapeutic benefit from resection, precluding comprehensive molecular analysis of matched primary and metastatic samples. In the few available cases, metastatic lesions exhibit marked biological divergence from their primary tumors, underscoring the need to study the metastatic niche directly[15]. Moreover, while mouse models have been invaluable for identifying molecular drivers of medulloblastoma initiation, endpoints are reached due to the rapid growth of cerebellar tumors that cannot be surgically resected, not leptomeningeal metastasis[16]. Lastly, conventional in vitro models fail to replicate the cellular composition and nutrient-deprived microenvironment of the CSF, limiting their translational relevance.

Organotypic co-culture systems have proven valuable for modeling tumor–host interactions that occur during metastatic dissemination to lung and bone marrow in other solid tumor contexts[17–20]. Leptomeningeal stromal-tumor co-cultures have been applied in studies of adult cancers, but have not been well characterized for medulloblastoma; moreover, previous in vitro findings have not been validated within the in vivo leptomeningeal niche[21, 22]. To enable reductionist studies of leptomeningeal biology in medulloblastoma, we developed and validated an in vitro model that recapitulates key features of the leptomeningeal niche. By co-culturing established and patient-proximal medulloblastoma cell lines with primary human leptomeningeal fibroblasts, we demonstrate that meningeal cells promote tumor cell survival and proliferation under nutrient-deprived conditions. We further identify medulloblastoma models that reproducibly form laminar or nodular growth patterns in vitro and in vivo, revealing a cell-biological mechanism that explains this morphological divergence. This physiologically informed, experimentally validated model enables the dissection of the molecular determinants of leptomeningeal colonization and the identification of therapeutic vulnerabilities to prevent medulloblastoma dissemination.

Cell Culture

Cell lines utilized in this manuscript were routinely authenticated with short tandem repeat profiling and Mycoplasma testing (Laboratory Corporation of America). Cells used for experiments that were grown in suspension were routinely replaced within two months after being revived from liquid nitrogen storage. Primary cells used for experiments were used within 10 passages after being revived from liquid nitrogen storage. Medulloblastoma cell lines (and their molecular sub-group) used in this study include: D425 (Group 3), MED-411FH (Group 3), MED-2112FH (Group 3), and CHLA-01-MED (Group 4). D425 Med (D425) cells (RRID: CVCL_1275) were derived from a primary medulloblastoma tumor resection at Duke University Medical Center and obtained from the laboratory of Dr. Jeffery Leonard[23, 24]. Med-411FH (MED411) cells and Med-2112FH (MED2112) cells are cell lines derived from early passage patient-derived xenografts developed by the laboratory of Dr. Jim Olson and distributed by the Brain Tumor Resource Lab at Seattle Children’s Hospital[25]. CHLA-01-MED was obtained from American Type Culture Collection as described previously[26]. Human meningeal cells (HMC) (ScienCell Research Laboratories, Cat. #1400) are primary leptomeningeal cells derived from human donors. We have used > 5 distinct lots with similar results. D425 and HMC cells were cultured in DMEM (Corning, 10-013-CV) supplemented with 10% FBS (Atlanta Biologicals, S11150H). MED411, MED2112, and CHLA-01-MED cells were cultured in NeuroCult NS-A Basal Medium (Human) (Stemcell technologies, Cat. #05750) supplemented with NeuroCult NS-A Proliferation Kit (Human) (Stemcell technologies, Cat. #05751), human epidermal growth factor (20 ng/mL, ThermoFisher, AF-100-15-500UG) and human fibroblast growth factor-basic (20ng/mL, ThermoFisher, 100-18B-50UG).

HMC survival optimization

HMC were grown to confluence in 24-well plates. Media was exchanged for either basal growth media, serial dilutions of growth media with artificial CSF (ACSF, Tocris, Cat. No. 3525), ACSF + 60mg/dL glucose, or DMEM/F12 (Gibco, Cat. #21041025). After 72 hours, cell viability was assessed using alamarBlue (ThermoFisher, A50100), and cells were photographed under light-phase microscopy.

Medulloblastoma-leptomeningeal fibroblast co-culture

Human medulloblastoma cells and meningeal cells were co-cultured to generate an in vitro model that captures key interactions between tumor cells and the surrounding stroma of leptomeningeal metastasis within a reductionist setting. HMCs were seeded onto 0.4 µm transwell filters or 1.5# glass coverslips (Electron Microscopy Sciences, 72204-04) in 6- or 24-well dishes and cultured with DMEM supplemented with 10% FBS. After HMCs formed a confluent monolayer (48–72 hours), the base media was removed and replaced with serum-free, growth-factor free DMEM/F12 containing medulloblastoma cells. For transwell experiments, DMEM/F12 was placed in the top and bottom wells.

Cell-free extracellular matrix

HMC were grown to confluence on 1.5# coverslips for 72 hours. Cell-free HMC-derived matrix was generated as previously described [27]. Medulloblastoma cells were plated onto coated coverslips for 48–72 hours in their base media before processing for immunofluorescence.

Animal Studies

All animal procedures were approved by the Institutional Animal Care and Use Committee of the Abigail Wexner Research Institute at Nationwide Children’s Hospital (IACUC protocol AR15-00022). CB17-SCID mice [CB17SC scid−/−; RRID: IMSR_TAC:cb17sc] were used for all in vivo experiments.

Subcutaneous Tumor Model: Mice were anesthetized and injected subcutaneously in the flank with 5 × 10⁵ medulloblastoma cells, with or without 1 × 10⁶ HMCs. Tumor growth was monitored biweekly via caliper measurements. Mice were euthanized, and tumors were harvested upon a total tumor volume of 2000 mm³.

Spontaneous Leptomeningeal Metastasis Model: Mice were anesthetized with isoflurane and placed in a stereotaxic apparatus (Kopf Instruments, Model 940). Intracerebellar injection of 8 × 10⁴ medulloblastoma cells in 3 µL of Neurocult media was performed at 2mm lateral, 2mm posterior, and 2.5 mm deep relative to lambda. Post-operative care included analgesia and daily monitoring. At neurological endpoint (rapid weight loss and an enhanced body condition score of less than 9, and/or neurological clinical signs including head tilts, seizures, and/or log rolling), mice were euthanized and processed for histology.

Experimental Leptomeningeal Metastasis Model: Medulloblastoma cells (2.5 × 10⁴ in 3 µL Neurocult media) were injected into the cisterna magna under aseptic conditions. Post-operative care and endpoint monitoring mirrored the spontaneous model.

Histology

Mice were euthanized at defined endpoints, sternotomy was performed, and mice were perfused with 20 ml of 0.9% saline injection (BD PosiFlush, cat #306546) followed by 30 ml of neutral buffered formalin (NBF) (Sigma-Aldrich, HT5012-1CS) through the left ventricle. Heads and spinal columns were cleaned of connective tissue and placed in NBF at 4°C for 48 hours. Brains were dissected from skull and spines were decalcified in 10% EDTA (Elabscience, E-IR-R112) for 10 days at room temperature, and then processed for paraffin embedding (FFPE). For standard histological observation, FFPE sections were counterstained with hematoxylin and eosin via conventional methods.

Immunohistochemistry and Immunofluorescence

Unstained 4µm sections were deparaffinized with xylene and rehydrated through ethanol series. Heat-mediated antigen retrieval was performed with Tris-EDTA buffer (TE; pH 9) with 0.2% v/v Tween 20 (Fisher Bioreagents, BP337-100).

Cells growing on #1.5 glass coverslips or 0.4µm filters were fixed in NBF for 10 minutes at room temperature. Following three rinses in 1x Phosphate Buffered Saline (PBS) (Corning, 21-040-CM), coverslip or histological sections were then incubated for at least 1 hour in blocking solution consisting of PBS, 0.2% triton100 (v/v) (Sigma, X-100-100ml), and 3% bovine serum albumin (w/v) (Sigma, A-7888) at room temperature. Coverslips or histological sections were then incubated overnight in the following primary antibodies (diluted in blocking solution): hamster anti-Podoplanin (Developmental Studies Hybridoma Bank(DSHB), 8.1.1), rabbit anti-Fibronectin (Abcam, ab2413), rabbit anti-Collagen IV (Abcam, ab6586), rabbit anti-Synaptophysin (Abcam, ab32127), rabbit anti-GFAP (Abcam, ab68428), rabbit anti-CD31 (Abcam, ab182981), rabbit anti-pan laminin (Novus Biologicals, NB300-144), rabbit anti- Cleaved Caspase 3 (Cell Signaling Technologies-CST, 9664), rabbit anti- phospho-Histone Ser10 (CST, 53348), goat anti-CD45 (R&D Systems, AF114), goat anti-OTX2 (R&D Systems, AF1979), sheep anti-S100A6 (R&D Systems, AF4584), or mouse anti-Vimentin (Abnova, SRL33). Following three rinses in PBS, sections or coverslips were incubated with appropriate AlexaFluor-labeled secondary antibodies (Invitrogen), DAPI (Invitrogen, D1306) and, where indicated, Phalloidin AlexaFluor 488, 568 or 647 (Invitrogen), or Wheat Germ Agglutinin (WGA) AlexaFluor 647 (Invitrogen, W11261) diluted in blocking solution for 1 hour at room temperature. Following PBS rinses, coverslips and tissue sections were mounted in Fluoromount G (Invitrogen, 00495802).

Microscopy and image analysis

Confocal microscopy images were obtained using a Crest-X-Light V3 spinning disk confocal on a Nikon Ti2-E inverted microscope, equipped with either a Hamamatsu ORCA Fusion camera or a Zeiss 800 laser scanning microscope, and featuring 4X, 20X, and 40X air objectives. Raw ND2 or czi files were imported into ImageJ for post-image processing and quantitation. All manipulation of images (brightness/contrast) were done uniformly to images. Mitotic (pH3+) and apoptotic (CC3+) were quantified by creating binary images with a manual threshold, then subjecting the binary images to the Analyze Particles function (size > 10µm²). Index was generated by dividing the cells (pH3 + or CC3+) by the total number of medulloblastoma cells (OTX2+). Spreading of medulloblastoma cells in co-culture was quantified by generating a binary image of OTX2 image and processed using the Skeletonize function. Skeletonized images were then quantified with Analyze Skeleton function. Percent spreading on HMC-derived ECM was quantified manually using the Cell Counter Plugin. H/E images were obtained with Aperio FL ScanScope digital slide scanner with 20X objective, and metastasis length was quantified using Aperio ImageScope Software (V12.4.3).

Statistical analysis

All statistical analyses for in vitro assays and animal studies were performed using GraphPad Prism v10. The number of biological replicates (independent experiments or animals) is specified in the figure legends, along with the appropriate statistical tests and P values. Data are presented as the mean with error bars that represent the standard error of the mean.

Pial fibroblast-medulloblastoma adhesion within the early metastatic niche

To guide the design of physiologically relevant in vitro models, we first generated models of the early leptomeningeal metastatic niche in vivo. Among medulloblastoma subgroups, Group 3 and Group 4 tumors most frequently disseminate to the leptomeninges[3]. Because Group 4 models are not widely available, we orthotopically injected three human Group 3 medulloblastoma cell lines (D425, MED411, and MED2112) into the cerebella of immunocompromised mice to assess spontaneous metastasis (Fig. 1A). At humane endpoint, brains and spines were examined histologically for evidence of dissemination. Although all models formed robust primary tumors with comparable survival (Fig. 1B), small leptomeningeal lesions were detected only in D425 and MED411 (Fig. 1C). These findings highlight the potential limitations of orthotopic xenografts for studying leptomeningeal metastasis, as rapid primary tumor growth restricts the development of marked leptomeningeal disease.

We next defined the stromal architecture of the leptomeningeal niche in D425 spontaneous metastasis lesions. The cellular aspects of the leptomeninges include pial and arachnoid fibroblasts, penetrating blood vessels, a subpial basement membrane, tissue-resident and recruited immune cells, and the glial limitans [8, 9, 28–30]. Multiparameter immunohistochemistry identified pial and arachnoid fibroblasts (Podoplanin), endothelial cells (CD31), basement membrane (laminin), glial limitans (GFAP), and immune cells (CD45). Tumor cells, identified by OTX2 expression (not shown for clarity), were consistently juxtaposed to pial fibroblasts, suggesting a preferential tumor–fibroblast interaction during early dissemination (Fig. 1D-G). Based on these observations and recently published data[31], we selected leptomeningeal fibroblasts as the primary stromal component for developing an in vitro co-culture model of the early leptomeningeal niche.

Human leptomeningeal fibroblasts support medulloblastoma survival and growth

Leptomeningeal fibroblasts can be isolated from human donors and cultured in vitro [32]. To enhance the physiological relevance by mimicking the nutrient-deprived state of the CSF, we first defined the minimal media conditions capable of supporting primary human leptomeningeal fibroblasts (HMC). Standard HMC media (DMEM + 10% FBS) was diluted with increasing concentrations of artificial CSF (aCSF) containing physiologic glucose or replaced with serum- and growth factor–free DMEM/F12 (Supplemental Fig. 1A). While HMCs failed to survive in aCSF, confluent monolayers were maintained in DMEM/F12 (Supplemental Fig. 1B–C). Under these conditions, HMCs retained expression of pial markers including fibronectin, collagen IV, laminin, S100A6, and vimentin (Supplemental Fig. 1D). Thus, HMCs can be maintained under serum- and growth factor–free conditions that approximate the nutrient limitations of CSF.

Supplemental Fig. 1

Optimization of HMC growth under serum-free and growth-factor-free conditions. A) Schematic depicting the dilution conditions followed to minimize growth factor and serum conditions while maximizing HMC proliferation and survival. B) Quantification of cellular viability of HMCs under different nutrient conditions. The plot represents results for n = 3 experiments. C) Representative phase contrast images of HMCs under the different nutrient conditions. D) Representative immunofluorescence images of HMCs demonstrating that they retain expression of common pial markers under serum- and growth-factor-free conditions. Scale bar = 50µm.

Using these optimized conditions, we next tested whether HMCs promote medulloblastoma cell survival and proliferation in short-term co-culture (Fig. 2A). Under nutrient-deprived conditions, D425 and MED411 cells exhibited reduced proliferation and increased apoptosis, as shown by phospho-H3 (S10) and cleaved caspase-3 (CC3) staining (Fig. 2B–E). Co-culture with HMCs restored proliferation and reduced apoptosis to near steady-state levels, indicating that leptomeningeal fibroblasts provide trophic support under nutrient-limited conditions.

To assess whether HMCs enhance tumorigenicity in vivo, we performed dual flank injections in immunocompromised mice, injecting tumor cells alone on one side and tumor + HMCs on the opposite side (Fig. 2F). Co-injection with HMCs increased tumor establishment (D425: 55.5% tumor take rate in single injection vs 88.9% tumor take rate in co-injection, MED411: 66.6% tumor take rate in single injection vs 100% tumor take rate in co-injection) and accelerated growth in both D425 and MED411 models (Fig. 2G). Notably, HMC injection alone did not result in tumorigenesis, suggest the difference is not attributable to expansion of meningeal cells. Together, these data demonstrate that leptomeningeal fibroblasts actively promote medulloblastoma survival and proliferation both in vitro and in vivo.

Distinct laminar and nodular morphologies reflect divergent modes of leptomeningeal colonization

Medulloblastoma metastatic lesions adopt diffuse laminar or distinct nodular phenotypes in patients, though the mechanisms underlying these morphologies have not been elucidated due to lack of preclinical models. Having validated short-term co-culture feasibility, we next examined whether extended co-culture recapitulates the morphologic diversity of leptomeningeal metastases. A panel of medulloblastoma neurosphere lines—including Group 3 models (D425, MED411, MED2112) and one Group 4 model (CHLA-01-MED)—was co-cultured with HMCs under nutrient-deprived conditions for 7 days. Two reproducible and morphologically distinct growth patterns emerged: a laminar phenotype (D425, MED411) in which tumor cells disseminated from neurospheres and spread across the fibroblast monolayer, and a nodular phenotype (MED2112, CHLA-01-MED) characterized by compact, spherical aggregates with limited dispersion (Fig. 3A). To generate a quantitative metric of laminar versus nodular growth, OTX2 + tumor cell masks were converted to skeletonized representations, enabling measurement of branch number and extent as indicators of surface spreading (Fig. 3B-C). Laminar models (D425 and MED411) demonstrated significantly higher branching than nodular models (MED2112 and CHLA-01-MED), indicative of a more diffuse growth pattern in vitro.

We next sought to determine whether in vitro behaviors mirrored in vivo dissemination patterns. Because leptomeningeal disease progression is prevented by rapid primary tumor growth in spontaneous models (Fig. 1), we hypothesized that direct injection of tumor cells into CSF would allow for leptomeningeal tumor progression and evaluation of growth patterns between different models. Previous works have seeded medulloblastoma cells into CSF through injection of tumor cells into lateral ventricle or cisterna magna[31, 33, 34]. Medulloblastoma cells are thought to enter CSF through either direct invasion into 4th ventricle or following hematogenous dissemination[35]. Because the CSF within the 4th ventricle circulates directly into the cisterna magna, we chose intracisternal injections to inoculate D425, MED411, or MED2112 cells directly into the CSF. The resulting metastatic lesions recapitulated the in vitro patterns: laminar models (D425 and MED411) exhibited extensive surface coating of the intracranial and spinal meninges, whereas the nodular model MED2112 formed discrete, localized deposits (Fig. 3D). We quantified lesion length as a surrogate measure of laminar versus nodular growth and found that laminar lesions extended significantly farther along the pial surface than nodular lesions (Fig. 3E). We acknowledge the limitations of using quantitative metrics such as length to capture categorical morphologic features, mirroring the challenges faced clinically in assessing leptomeningeal disease[36, 37]. Collectively, the combination of in vitro and in vivo findings provide preclinical evidence for biologically distinct modes of leptomeningeal colonization.

Laminar lesions disrupt the fibroblast layer and adhere to the subpial basement membrane

We performed long-term live-cell imaging of HMC–tumor co-cultures to directly visualize tumor–fibroblast interactions. Both laminar and nodular medulloblastoma cells adhered rapidly to fibroblasts within hours of plating (Supplemental Videos 1–4). However, laminar cells progressively displaced the fibroblast monolayer and spread along the underlying substrate, whereas nodular cells remained compact and adherent to pial cells. Fixed co-cultures imaged after 5 days confirmed these distinct morphologies at high resolution (Fig. 4A–E).

In vivo immunohistochemistry of D425, MED411, and MED2112 leptomeningeal lesions corroborated these findings: laminar tumor cells penetrated the pial fibroblast layer to adhere directly to the laminin-positive subpial basement membrane. We did not observe invasion through the basement membrane or disruption of the glial limitans, consistent with clinical observations that parenchymal extension of disseminated disease is rare in medulloblastoma. In contrast, nodular lesions remained juxtaposed to intact pial fibroblasts (Fig. 4F–H). Thus, laminar and nodular lesions represent distinct modes of tumor–meningeal interaction, differing in their degree of fibroblast displacement and engagement with the subpial basement membrane.

Adhesion to meningeal extracellular matrix distinguishes laminar from nodular cells

We next sought to determine how medulloblastoma cells interact with extracellular matrix (ECM) derived from leptomeningeal fibroblasts given the differential patterns of spread observed in previous experiments. To directly quantify medulloblastoma behavior on pial ECM, we generated cell-free HMC ECM by culturing HMC on coverslips and performing a decellularization method described previously[27] (Fig. 5A–C). Decellularized ECM retains both the structure and biological properties of leptomeningeal-deposited ECM, even in the absence of leptomeningeal fibroblasts (noted by the absence of nuclear DAPI staining). D425, MED411, MED2112, and CHLA-01-MED were plated on decellularized HMC-derived ECM for 24 hours. Laminar cells (D425 and MED411) adopted an elongated morphology with filopodia-like projections characteristic of spreading and migration. In contrast, nodular cells (MED2112 and CHLA-01-MED) retained cortical actin organization and failed to spread on HMC-derived ECM (Fig. 5D-F). Lastly, nodular cells exhibit significantly increased apoptosis when cultured on HMC-derived extracellular matrix, indicating reduced fitness and impaired adaptation to ECM-dependent growth conditions (Fig. 5G).

Leptomeningeal metastasis remains the principal cause of treatment failure and death in children with medulloblastoma, yet its underlying biology has remained elusive. Here, we developed an in vitro model that faithfully mirrors the morphologic diversity of leptomeningeal metastasis observed in vivo. Using this system, we identify two distinct modes of leptomeningeal colonization—laminar and nodular—that arise from differential tumor–stroma adhesion mechanisms. Our findings build upon previous iterations of in vitro meningeal and ECM medulloblastoma models by providing in vivo validation[22, 38, 39]. Future studies will define the molecular mechanisms that distinguish laminar and nodular phenotypes and determine whether cellular plasticity exists between these states. Integrating in vitro co-culture systems with both spontaneous and experimental leptomeningeal metastasis models establishes a versatile platform for delineating the stage-specific roles of metastatic mediators; whether they govern invasion and dissemination from the primary site, colonization of the leptomeninges, or subsequent progression within the leptomeningeal niche [34, 40–42]. Additionally, understanding the pathways that regulate these phenotypic transitions may have translational significance, given differential therapeutic sensitivities between clinical phenotypes [12, 43]. The model described here provides a tractable platform to investigate therapeutic response as a function of leptomeningeal colonization architecture.

The leptomeningeal niche as a trophic microenvironment

A recent landmark paper from the Taylor group identified medulloblastoma-induced changes to the leptomeningeal environment during colonization and established the importance of tumor-host interactions in leptomeningeal metastasis[31]. Our findings agree with those observations by providing direct evidence that human meningeal cells provide trophic support to medulloblastoma cells. Co-culture of medulloblastoma cells with HMC rescued tumor cell proliferation and suppressed apoptosis under nutrient-deprived conditions, demonstrating that fibroblast-derived cues compensate for the metabolic limitations of CSF. Similar trophic effects have been described in lung and bone marrow metastases, suggesting that stromal support is a conserved feature of the metastatic niche across organ systems[44]. By defining leptomeningeal fibroblasts as key mediators of this process, our study establishes a cellular foundation for dissecting paracrine and adhesion-dependent interactions that sustain leptomeningeal metastasis.

Modeling implications and future directions

Traditional xenograft models are limited in their ability to capture the temporal or spatial features of leptomeningeal dissemination due to rapid intracerebellar tumor growth and limited sampling of metastatic lesions. The co-culture system described here complements these in vivo models by offering a reductionist platform for mechanistic dissection and therapeutic testing. Integration of single-cell and spatial transcriptomic profiling will enable mapping of tumor–stroma signaling networks and identification of context-specific dependencies. In parallel, the model provides an experimental framework for evaluating compounds that disrupt adhesion-dependent survival. Future studies will delineate the mechanisms that distinguish cell–cell from cell–matrix adhesion in nodular and laminar metastases, as disrupting these interactions may represent context-specific therapeutic vulnerabilities within the leptomeningeal niche.

While our model captures key aspects of the leptomeningeal niche, it lacks other microenvironmental components, such as immune or vascular cells, that may modulate colonization dynamics and progression[31, 35, 45, 46]. Incorporating these elements in future iterations will further enhance physiologic relevance. The models used in this study were predominantly Group 3 medulloblastoma lines with MYC amplification, reflecting both their aggressive and metastatic phenotype and their availability [47]. Future work will extend these studies to WNT-, SHH-activated, and non-MYC–amplified Group 3/4 tumors, ideally using patient-derived xenografts or primary human specimens, to capture the full molecular spectrum of medulloblastoma leptomeningeal metastasis. Ultimately, coupling these advanced models with clinical and genomic data from patient-derived metastasis samples will enable a more complete understanding of metastatic evolution in medulloblastoma.

Our study delineates modes by which medulloblastoma colonizes the leptomeninges. We identify differential interactions with fibroblast-derived extracellular matrix as a key determinant in influencing metastatic phenotype. This framework establishes the leptomeningeal niche as both a driver of metastatic persistence and a therapeutic vulnerability, offering new avenues for targeting the most lethal stage of pediatric medulloblastoma.

Ethics approval and consent to participate

This study used pre-established and patient-derived human cell lines that are commercially available (Brain Tumor Resource Lab, ATCC, Sigma-Aldrich and ScienCell). According to institutional and national guidelines, research involving such cell lines does not require ethics committee approval because the material is anonymized and information identifying the donor is not accessible.

Consent for publication

De-identified MR images were obtained after approval from Nationwide Children’s Hospital Institutional Review Board. Consent was not deemed to be necessary.

Competing interests

LJG is supported by Pelotonia on a Pelotonia Graduate Scholar Fellowship.

Funding

This work was funded by CancerFree KIDS (Wolf Pack Research Grant, JBR), St. Baldrick’s Foundation Scholar Award (JBR), Alex’s Lemonade Stand Foundation Young Investigator Award (JBR), Hyundai Hope On Wheels, Bridge2K (JBR), and Nationwide Children’s Hospital (JBR). LJG is supported by Pelotonia on a Pelotonia Graduate Scholar fellowship. RDR is supported by Alex’s Lemonade Stand Crazy 8 Award, NIH R01CA260178, NIH CTSA Grant UL1TR002733, CancerFree KIDS Foundation, and Nationwide Children’s Director Strategic Development Fund.

Acknowledgments

We are indebted to Dr. Robert Wechsler-Reya, Elizabeth Thompson, and Daphne Koubourli for their technical expertise in in vivo modeling. We are grateful to members of the laboratory of Dr. Ryan Roberts for valuable discussions and critical reading of this manuscript. We also acknowledge the Histopathology and Microscopy Core facilities at Nationwide Children’s Hospital Abigail Wexner Research Institute for their excellent technical expertise. The authors acknowledge the use of ChatGPT5, solely for text editing and clarity. All schematics in this manuscript were generated using BioRender.

Authors’ information (optional)

Not applicable.

Author Contribution

LJG- design of work, acquisition, analysis, and interpretation of data, draft of manuscriptLR- design of work, acquisition, analysis, and interpretation of data, reviewed manuscriptAG-acquisition, analysis, and interpretation of data, reviewed manuscriptAA-acquisition of data, reviewed manuscriptJW-acquisition of data, reviewed manuscript JL-acquisition of data, critical materials, review and revision of manuscriptRR-interpretation of data, review and revision of manuscript JR-conception and design of work, acquisition, analysis, and interpretation of data, review and revision of manuscript

Acknowledgement

We are indebted to Dr. Robert Wechsler-Reya, Elizabeth Thompson, and Daphne Koubourli for their technical expertise in in vivo modeling. We are grateful to members of the laboratory of Dr. Ryan Roberts for valuable discussions and critical reading of this manuscript. We also acknowledge the Histopathology and Microscopy Core facilities at Nationwide Children’s Hospital Abigail Wexner Research Institute for their excellent technical expertise. The authors acknowledge the use of ChatGPT5, solely for text editing and clarity. All schematics in this manuscript were generated using BioRender.

Data Availability

All data supporting the findings of this study are available within the paper and its Supplementary Information.

Hill RM et al (2020) Time, pattern, and outcome of medulloblastoma relapse and their association with tumour biology at diagnosis and therapy: a multicentre cohort study. Lancet Child Adolesc Health 4(12):865–874
Michalski JM et al (2021) Children's Oncology Group Phase III Trial of Reduced-Dose and Reduced-Volume Radiotherapy With Chemotherapy for Newly Diagnosed Average-Risk Medulloblastoma. J Clin Oncol 39(24):2685–2697
Ramaswamy V et al (2013) Recurrence patterns across medulloblastoma subgroups: an integrated clinical and molecular analysis. Lancet Oncol 14(12):1200–1207
Robinson GW et al (2018) Risk-adapted therapy for young children with medulloblastoma (SJYC07): therapeutic and molecular outcomes from a multicentre, phase 2 trial. Lancet Oncol 19(6):768–784
Sabel M et al (2016) Relapse patterns and outcome after relapse in standard risk medulloblastoma: a report from the HIT-SIOP-PNET4 study. J Neurooncol 129(3):515–524
Fults DW, Taylor MD, Garzia L (2019) Leptomeningeal dissemination: a sinister pattern of medulloblastoma growth. J Neurosurg Pediatr 23(5):613–621
Levy AS et al (2021) Temozolomide with irinotecan versus temozolomide, irinotecan plus bevacizumab for recurrent medulloblastoma of childhood: Report of a COG randomized Phase II screening trial. Pediatr Blood Cancer 68(8):e29031
Derk J et al (2021) Living on the Edge of the CNS: Meninges Cell Diversity in Health and Disease. Front Cell Neurosci 15:703944
Freret ME, Boire A (2024) The anatomic basis of leptomeningeal metastasis. J Exp Med, 221(4)
Weller RO et al (2018) The meninges as barriers and facilitators for the movement of fluid, cells and pathogens related to the rodent and human CNS. Acta Neuropathol 135(3):363–385
Remsik J, Boire A (2024) The path to leptomeningeal metastasis. Nat Rev Cancer 24(7):448–460
Dufour C et al (2012) Metastatic Medulloblastoma in Childhood: Chang's Classification Revisited. Int J Surg Oncol, 2012: p. 245385
Zapotocky M et al (2018) Differential patterns of metastatic dissemination across medulloblastoma subgroups. J Neurosurg Pediatr 21(2):145–152
Wilcox JA et al (2024) Leptomeningeal metastases from solid tumors: A Society for Neuro-Oncology and American Society of Clinical Oncology consensus review on clinical management and future directions. Neuro Oncol 26(10):1781–1804
Morrissy AS et al (2016) Divergent clonal selection dominates medulloblastoma at recurrence. Nature 529(7586):351–357
Turunen JR, Kumthekar P, Ahmed AU (2025) Beyond the brain: exploring the impact of animal models of leptomeningeal disease from solid tumors. Acta Neuropathol Commun 13(1):103
Carlson P et al (2019) Targeting the perivascular niche sensitizes disseminated tumour cells to chemotherapy. Nat Cell Biol 21(2):238–250
Ghajar CM et al (2013) The perivascular niche regulates breast tumour dormancy. Nat Cell Biol 15(7):807–817
Montagner M et al (2020) Crosstalk with lung epithelial cells regulates Sfrp2-mediated latency in breast cancer dissemination. Nat Cell Biol 22(3):289–296
Reinecke JB et al (2025) Metastasis-Initiating Osteosarcoma Subpopulations Establish Paracrine Interactions with Lung and Tumor Cells to Create a Metastatic Niche. Cancer Res
Alhaddad H et al (2024) Spatial transcriptomics analysis identifies a tumor-promoting function of the meningeal stroma in melanoma leptomeningeal disease. Cell Rep Med 5(6):101606
Svalina MN et al (2016) IGF1R as a Key Target in High Risk, Metastatic Medulloblastoma. Sci Rep 6:27012
Bigner SH et al (1990) Amplification of the c-myc gene in human medulloblastoma cell lines and xenografts. Cancer Res 50(8):2347–2350
Studebaker AW et al (2017) Oncolytic Herpes Virus rRp450 Shows Efficacy in Orthotopic Xenograft Group 3/4 Medulloblastomas and Atypical Teratoid/Rhabdoid Tumors. Mol Ther Oncolytics 6:22–30
Brabetz S et al (2018) A biobank of patient-derived pediatric brain tumor models. Nat Med 24(11):1752–1761
Xu J et al (2015) Disseminated Medulloblastoma in a Child with Germline BRCA2 6174delT Mutation and without Fanconi Anemia. Front Oncol 5:191
Hellewell AL, Rosini S, Adams JC A Rapid, Scalable Method for the Isolation, Functional Study, and Analysis of Cell-derived Extracellular Matrix. J Vis Exp, 2017(119).
DeSisto J et al (2020) Single-Cell Transcriptomic Analyses of the Developing Meninges Reveal Meningeal Fibroblast Diversity and Function. Dev Cell 54(1):43–59e4
Pietila R et al (2023) Molecular anatomy of adult mouse leptomeninges. Neuron 111(23):3745–3764e7
Rua R, McGavern DB (2018) Advances in Meningeal Immunity. Trends Mol Med 24(6):542–559
Abeysundara N et al (2025) Metastatic medulloblastoma remodels the local leptomeningeal microenvironment to promote further metastatic colonization and growth. Nat Cell Biol 27(5):863–874
Remsik J et al (2021) Characterization, isolation, and in vitro culture of leptomeningeal fibroblasts. J Neuroimmunol 361:577727
Choi SA et al (2016) In vivo bioluminescence imaging for leptomeningeal dissemination of medulloblastoma in mouse models. BMC Cancer 16(1):723
Martirosian V et al (2021) Medulloblastoma uses GABA transaminase to survive in the cerebrospinal fluid microenvironment and promote leptomeningeal dissemination. Cell Rep 35(13):109302
Garzia L et al (2018) A Hematogenous Route for Medulloblastoma Leptomeningeal Metastases. Cell 172(5):1050–1062e14
Chamberlain M et al (2017) Leptomeningeal metastases: a RANO proposal for response criteria. Neuro Oncol 19(4):484–492
Peng J et al (2020) Evaluation of RAPNO criteria in medulloblastoma and other leptomeningeal seeding tumors using MRI and clinical data. Neuro Oncol 22(10):1536–1544
Fiorilli P et al (2008) Integrins mediate adhesion of medulloblastoma cells to tenascin and activate pathways associated with survival and proliferation. Lab Invest 88(11):1143–1156
Linke F et al (2021) 3D hydrogels reveal medulloblastoma subgroup differences and identify extracellular matrix subtypes that predict patient outcome. J Pathol 253(3):326–338
Grausam KB et al (2017) ATOH1 Promotes Leptomeningeal Dissemination and Metastasis of Sonic Hedgehog Subgroup Medulloblastomas. Cancer Res 77(14):3766–3777
Li M, Deng Y, Zhang W (2021) Molecular Determinants of Medulloblastoma Metastasis and Leptomeningeal Dissemination. Mol Cancer Res 19(5):743–752
Sanghrajka RM et al (2023) KMT2D suppresses Sonic hedgehog-driven medulloblastoma progression and metastasis. iScience 26(10):107831
Le Rhun E et al (2025) Point/Counterpoint: Intrathecal therapy for patients with leptomeningeal metastases from solid tumors. Neuro Oncol 27(9):2225–2231
Monteran L, Zait Y, Erez N (2024) It's all about the base: stromal cells are central orchestrators of metastasis. Trends Cancer 10(3):208–229
Nor C et al (2025) Therapeutic radiation drives leptomeningeal dissemination of medulloblastoma through an innate immune process. Dev Cell
Remsik J et al (2025) Interferon-gamma orchestrates leptomeningeal anti-tumour response. Nature 643(8073):1087–1096
Ivanov DP et al (2016) In vitro models of medulloblastoma: Choosing the right tool for the job. J Biotechnol 236:10–25

No competing interests reported.

Download PDF

Editorial decision: Revision requested
12 Dec, 2025
Reviews received at journal
11 Dec, 2025
Reviews received at journal
11 Dec, 2025
Reviewers agreed at journal
19 Nov, 2025
Reviewers agreed at journal
17 Nov, 2025
Reviewers invited by journal
17 Nov, 2025
Editor assigned by journal
15 Nov, 2025
Submission checks completed at journal
11 Nov, 2025
First submitted to journal
05 Nov, 2025

You are reading this latest preprint version

Meningeal-tumor interactions define distinct modes of leptomeningeal colonization in medulloblastoma

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Cell Culture

HMC survival optimization

Medulloblastoma-leptomeningeal fibroblast co-culture

Cell-free extracellular matrix

Animal Studies

Histology

Immunohistochemistry and Immunofluorescence

Microscopy and image analysis

Statistical analysis

Results

Pial fibroblast-medulloblastoma adhesion within the early metastatic niche

Human leptomeningeal fibroblasts support medulloblastoma survival and growth

Distinct laminar and nodular morphologies reflect divergent modes of leptomeningeal colonization

Laminar lesions disrupt the fibroblast layer and adhere to the subpial basement membrane

Adhesion to meningeal extracellular matrix distinguishes laminar from nodular cells

Discussion

The leptomeningeal niche as a trophic microenvironment

Modeling implications and future directions

Conclusions

Declarations

Consent for publication

Funding

Author Contribution

Acknowledgement

Data Availability

References

Additional Declarations

Supplementary Files

Status:

Version 1