Guerrero-Cázares, Hugo

Abstract The purpose of this study was to identify glioblastoma (GBM) proteins that contribute to the hijacking of the tumor microenvironment through extracellular vesicles (EVs). GBM is the most aggressive primary brain tumor in adults. It presents significant challenges due to its universal recurrence and limited survival rates, exacerbated when its located proximal to the lateral ventricles (LV). The cellular subpopulation of brain tumor-initiating cells (BTICs) plays a pivotal role in tumor initiation and invasiveness, interacting with the microenvironment, particularly the cellular components in the subventricular zone (SVZ) in the LV. We employed an integrated multi-omic strategy combining cell-specific proteomics of EV cargo and whole-cell transcriptomics of EV-treated neural progenitor cells (NPCs) to define mechanisms by which BTICs remodel the tumor microenvironment. Using the TurboID cell-specific protein labeling system, we identified a previously unrecognized repertoire of BTIC-derived proteins selectively packaged into EVs. Among these, DPYSL2 (CRMP2) emerged as a prominent EV cargo as it is implicated in cancer cell migration and invasion. Functionally, exposure of human NPCs to BTIC-derived EVs significantly increased NPC proliferation and migratory capacity in vitro, findings that were recapitulated in vivo following intracranial and intranasal delivery of GBM-derived EVs. To define the molecular basis of these phenotypic changes, we performed RNA sequencing of NPCs following EV treatment. Bioinformatic integration, including ontogenic and pathway enrichment analyses, revealed robust activation of oncogenic signaling pathways, notably JAK/STAT, mTOR, and WNT, accompanied by upregulation of downstream effector genes. Importantly, integrative proteomic-transcriptomic analyses identified DPYSL2 as a mechanistic link between EV cargo and NPC transcriptional reprogramming. Prior studies demonstrate that DPYSL2 can interact with JAK kinases, promoting STAT3 activation and transcription of genes such as Vimentin (VIM) and CALCA, consistent with the transcriptomic changes observed in EV-treated NPCs. Overall, we have identified DPYSL2 as a novel regulator of tumor microenvironment hijacking mediated by extracellular vesicles. This multi-omic approach is the first use of cell-specific proteomics to study extracellular vesicles and the first time pairing this method to identify upstream regulators of transcriptomic changes in neural progenitor cells. Collectively, these findings establish a direct molecular connection between BTIC-derived EV proteins and functional reprogramming of neural progenitor cells, providing mechanistic insight into how glioblastoma exploits developmental programs to promote tumor progression. Citation Format: Marissa Russo, Emily Norton, Weiwei Wang, Shreyartha Mukherjee, Daniel Wickland, TuKiet Lam, Hugo Guerrero-Cazares. DPYSL2 mediates extracellular vesicle-driven reprogramming of the glioblastoma tumor microenvironment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB228.

Abstract Glioblastoma (GBM) is the most common and aggressive primary brain tumor in adults with a poor prognosis of 15 months despite aggressive therapies. GBM malignancy is often driven by a high degree of inter- and intra-tumoral heterogeneity that is dynamic through time and regulated by interactions between tumor cells and their neural microenvironment. For instance, GBM tumors located near the lateral ventricles (LV) of the brain grow more aggressively and result in significantly worse patient survival, compared to tumors that grow far from the LV. This suggests that interactions between GBM and cells in the LV microenvironment may regulate tumor malignancy. Understanding the mechanisms underlying GBM malignancy near the LV is essential to determine if regulation of tumor biology is driven by the presence of healthy neural cells, and to ultimately develop therapeutic approaches that target GBM-LV interactions. Here, we tested the hypothesis that region-specific cellular identities and activation of unique transcriptional programs regulate GBM malignancy near the LV. Paired primary biopsies were collected from 14 GBM patients. Image-guided biopsies proximal and distal to the brain’s LV (LV-proximal GBM and LV-distal GBM, respectively) were processed for bulk RNA sequencing, single-nucleus RNA sequencing, and in situ spatial transcriptomic (Xenium) analyses. We found significant inter-patient and intra-tumoral transcriptional heterogeneity in GBM samples collected proximal and distal to the LV. After accounting for tumoral and patient heterogeneity, we found that proximal GBM tumors exhibit an increased immune-related gene signature, compared with tumors distal to the LV. Single-nucleus RNA sequencing and in situ spatial transcriptomic analyses showed no significant differences in the proportion of cellular states or identities between proximal and distal GBM samples. However, within each cellular identity, significant differences in gene expression were identified between proximal and distal GBM, suggesting that distinct cell-specific transcriptional pathways characterize regional differences in GBM biology. Our results suggest that GBM tumors near the LV have increased interactions with immune cells that lead to the enrichment of immune-related transcriptional pathways, compared to tumors growing far from the LV. This work will inform the development of therapies that target GBM-immune cell interactions to decrease tumor malignancy and improve individual patient outcomes. Citation Format: Maria F. Gonzalez-Aponte, Jhan C. Salazar, Aura Figueroa-Gonzalez, Emily S. Norton, Paola Suarez-Meade, Hugo Guerrero-Cazares. Integrating RNA sequencing and spatial transcriptomic analyses reveals drivers of glioblastoma malignancy near the brains’ lateral ventricles [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Brain Cancer; 2026 Mar 23-25; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2026;86(6_Suppl):Abstract nr B035.

Abstract Cancer remains one of the most complex and devastating diseases, in part because tumors are not static entities but dynamic ecosystems that evolve in real time. Over the decades, researchers have proposed two major frameworks toexplain tumor development: 1) the clonal evolution model, which emphasizes the stepwise accumulation of mutations insingle cells, and 2) the cancer stem cell model, which posits a hierarchical tumor structure driven by a small population of stem-like cells capable of self-renewal and tumor initiation. In brain tumors, as in other cancer types, the presence of cancer stem cells has been demonstrated; however, their evolutionary role in glioblastoma, the most aggressive primary brain tumor, remains poorly understood. A key unanswered question is how cancer stem cells in the brain evolve intomalignant populations and how evolutionary processes shape this transition. Since Darwin, evolutionary biologists have sought to understand how traits evolve across lineages and have developed several tools to address questions regarding how lineages, species and populations have evolved through time. In recent years there has been an increase in the use of evolutionary statistical methods to study cancer evolution. These approaches remain niche in a few tumor types (e.g., blood and lung cancer), but the rapid growth of sequencing technologies and large-scale datasets has created new opportunities to address evolutionary questions across cancer types that have been understudied (e.g., brain tumor). Researchers are now using whole-genome, exome, single-cell, spatial, and multi-omic data to reconstruct tumor histories, test adaptive hypotheses, and in few cases, quantify rates and modes of evolution within tumors and between tumor types. However, there is not yet a consolidated framework for which evolutionary analyses are most appropriate to test cancer evolution, nor accessible guidance on how to implement these approaches in oncological datasets. We synthesize evolutionary statistical methods currently used in brain cancer research and propose additional approaches for future studies. We identify more than 30 evolutionary tools developed to study trait evolution, diversification, and lineage dynamics in biological systems, of which only one has been applied to brain cancer. We argue that adapting these methods offers a powerful opportunity to investigate cancer stem cell dynamics and tumor progression in glioblastoma and other understudied cancers. By bridging evolutionary biology and cancer genomics, this work provides a roadmap for expanding the evolutionary toolkit used to study tumor evolution and for generating testable hypotheses about malignant transformation. Citation Format: Jhan C. Salazar Salazar, Sean W. McHugh, Maria F. Gonzalez-Aponte, Hugo Guerrero-Cazares. From healthy brain to cancer cells: Understanding the origins of glioblastoma through the lens of Darwinian evolution [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Brain Cancer; 2026 Mar 23-25; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2026;86(6_Suppl):Abstract nr B017.

Cancer stem cells (CSCs) are increasingly recognized as central drivers of tumorigenesis, therapeutic resistance, and recurrence across diverse malignancies. This review synthesizes our current understanding of CSC biology across CNS tumors, with a focus on glioblastoma, where stem-like cells are sustained by specialized and overlapping tumor microenvironmental niches. Perivascular, hypoxic, invasive, immunosuppressive, and extracellular matrix-associated niches cooperatively enforce stemness, metabolic adaptability, immune evasion, and phenotypic plasticity, enabling CSC persistence despite maximal surgical resection and standard-of-care therapy. Notably, CSCs extend beyond radiographically defined tumor margins and populate peritumoral regions, providing a biological basis for near-universal recurrence. Advances in multiparametric imaging, stem cell-based ex vivo and in vivo models, and single-cell and spatial profiling have refined insight into CSC heterogeneity, niche dependence, and treatment resistance. Together, these findings reframe therapeutic strategies, highlighting the need for function-preserving maximal resection and multimodal therapies that target both CSC-intrinsic pathways and their supportive microenvironments.

Bioreducible nanoparticles were engineered for safe and effective systemic siRNA delivery, including crossing the blood–brain barrier to achieve in vivo gene silencing in an orthotopic glioblastoma mouse model.

OBJECTIVEChordoma is a slow-growing, locally aggressive cancer that is minimally responsive to conventional chemotherapy and radiotherapy and has high local recurrence rates after resection. Currently, there are no rodent models of spinal chordoma. In the present study, the authors sought to develop and characterize an orthotopic model of human chordoma in an immunocompromised rat.METHODSThirty-four immunocompromised rats were randomly allocated to 4 study groups; 22 of the 34 rats were engrafted in the lumbar spine with human chordoma. The groups were as follows: UCH1 tumor–engrafted (n = 11), JHC7 tumor–engrafted (n = 11), sham surgery (n = 6), and intact control (n = 6) rats. Neurological impairment of rats due to tumor growth was evaluated using open field and locomotion gait analysis; pain response was evaluated using mechanical or thermal paw stimulation. Cone beam CT (CBCT), MRI, and nanoScan PET/CT were performed to evaluate bony changes due to tumor growth. On Day 550, rats were killed and spines were processed for H & E–based histological examination and immunohistochemistry for brachyury, S100β, and cytokeratin.RESULTSThe spine tumors displayed typical chordoma morphology, that is, physaliferous cells filled with vacuolated cytoplasm of mucoid matrix. Brachyury immunoreactivity was confirmed by immunostaining, in which samples from tumor-engrafted rats showed a strong nuclear signal. Sclerotic lesions in the vertebral body of rats in the UCH1 and JHC7 groups were observed on CBCT. Tumor growth was confirmed using contrast-enhanced MRI. In UCH1 rats, large tumors were observed growing from the vertebral body. JHC7 chordoma–engrafted rats showed smaller tumors confined to the bone periphery compared with UCH1 chordoma–engrafted rats. Locomotion analysis showed a disruption in the normal gait pattern, with an increase in the step length and duration of the gait in tumor-engrafted rats. The distance traveled and the speed of rats in the open field test was significantly reduced in the UCH1 and JHC7 tumor–engrafted rats compared with controls. Nociceptive response to a mechanical stimulus showed a significant (p < 0.001) increase in the paw withdrawal threshold (mechanical hypalgesia). In contrast, the paw withdrawal response to a thermal stimulus decreased significantly (p < 0.05) in tumor-engrafted rats.CONCLUSIONSThe authors developed an orthotopic human chordoma model in rats. Rats were followed for 550 days using imaging techniques, including MRI, CBCT, and nanoScan PET/CT, to evaluate lesion progression and bony integrity. Nociceptive evaluations and locomotion analysis were performed during follow-up. This model reproduces cardinal signs, such as locomotor and sensory deficits, similar to those observed clinically in human patients. To the authors’ knowledge, this is the first spine rodent model of human chordoma. Its use and further study will be essential for pathophysiology research and the development of new therapeutic strategies.