Lab-grown organoids reveal how glioblastoma outsmarts treatment Tumor organoid models offer new insight into how the brain and immune system shape cancer behavior.

🌡️ HOTT isn't just about temperature.

Human Organoid Tumor Transplantation (HOTT) is one of the models Dr. Bhaduri's lab (@bhadurilab.bsky.social) developed to understand how interactions with #glioblastoma tissue and immune cells contribute to therapy failure.

#Neuroskyence @dgsomucla.bsky.social

(5/5) Why it matters: In vivo ablation of NVP-associated programs remodels the tumor and significantly extends survival. These studies move us from descriptive atlases to a mechanistic, targetable understanding of tumor organization. Check them out!

(4/5) Next: Meet the NVP (Neurovascular Progenitor) 🧬. This rare population co-expresses neural and perivascular programs. Using patient-derived lineage tracing, we proved NVPs can bridge the gap, giving rise to both neural and mesenchymal fates.

(3/5) These relationships organize into reproducible "Tracks". By understanding these clonal architectures, we can design "hierarchy-informed" combination to disrupt the interactions that drive tumor growth.

(2/5) First up: CellTagging! We integrated DNA barcoding with scRNA-seq in 235k+ cells from 9 patient tumors. We found that GBM growth is sustained by multiple non-redundant progenitors, not just one dominant stem cell. #GBM #CancerResearch

(1/5) How does glioblastoma maintain its immense heterogeneity? 🧠 Two new companion papers from the Bhaduri Lab @UCLA now on @bioRxiv offer a lineage-resolved view of GBM hierarchies. Let’s dive in! 🧵👇

www.biorxiv.org/content/10.6...
www.biorxiv.org/content/10.1...

Special thanks to our amazing collaborators:
Nathanson Lab, Prins Lab, UCLA Neurosurgery, as well as the members of the Bhaduri Lab who have provided advice, edits, and moral support throughout. We look forward to hearing what the community thinks!

💡iHOTT is a human, patient-specific, matched platform to study tumor–immune interactions. It recapitulates key features of tumor-immune interactions, offering a new tool for immunotherapy discovery.

📄Paper: biorxiv.org/content/10.1...
🌍Data: cells.ucsc.edu?ds=ihott

🔗Were there shared antigens across patients? We used GLIPH2 to find out. The short answer: not really. Most T cell clonotypes were private to an individual tumor. This tells us something important: GBM’s immune landscape is highly personalized *and* iHOTT robustly captures that.

To understand the T cell response, we performed TCR sequencing on treated samples. Clonal diversity increased & new clones emerged 🌈. These novel clones were driven by CD4 T-cells and expressed TCF7, a marker for stem-like T cells that has been linked to improved outcomes.

But MOST interestingly ⭐, when we benchmarked iHOTT against patients treated with pembrolizumab as part of a clinical trial, we noted the *exact same* cell type shifts, validating that iHOTT recapitulates patient biology. This was an exciting finding ‼️

We treated iHOTT with pembrolizumab 💊to understand the effects of immunotherapy. Interestingly, we noted expansion of CD4+ T cells and B cells and enrichment pathways reflecting immune activation. The cytokine profile also changed with increases in IL15, IL25, and IL17F.

Our model preserved immune and tumor cell types after co-culture. We also saw immune activation and upregulation of GBM-linked cytokines like IL-6, IL-8, IL-10, and G-CSF, notably seen *only* when tumor cells and PBMCs were cultured together.

Most GBM models miss one of two things:
🛡️ a functional immune system
🧠 a human brain microenvironment

iHOTT has both.
🧫 We co-culture freshly isolated patient tumor cells + matched PBMCs on human cortical organoids. It’s patient-specific, matched, and immunocompetent.

🧵Excited to share our new preprint introducing iHOTT - an autologous tumor-immune co-culture model that captures patient-specific responses in #Glioblastoma

💥Now on
@biorxivpreprint
: biorxiv.org/content/10.1...

Led by Dr. Shivani Baisiwala, Neurosurgery Resident in the lab

A huge thank you to the amazing collaborators and members of the Bhaduri Lab for their contributions to this study! Antoni Martija, @prnano9, @Jalbsoto, Dan Jaklic, @MilJessenya, Rista White, @JacquiMMartn1, Dakshesh Rana, @GeschwindLab

Taken together, our findings suggest that the early entry of thalamocortical afferents in human and nonhuman primate cortex mediates communication with cortical progenitors. This demonstrates that the human thalamus actively sculpts cortical architecture during early development! 🧩

Leveraging the tractability of the organoid model, we knocked out thalamic NRXN1 to show that when disrupted, thalamic axons fail to form stable contacts with cortical outer radial glia, leading to impaired upper layer neurogenesis.

Our key player: NRXN1, a cell adhesion molecule expressed by thalamic neurons, which directly contacts outer radial glia. 🤝 We identify NRXN1-NLGN1 gene expression and protein colocalization along thalamocortical afferents and outer radial glia in the assembloid and human primary cortical tissue.

Cre-dependent anterograde tracing demonstrates that thalamocortical axons make physical connections with cortical outer radial glia - a primate-enriched progenitor cell type central to cortical expansion. 🧠 But what types of cell-cell interactions mediate this contact?

Interestingly, severance ✂️ of the thalamic input midway through neurogenesis actually maintains these increased upper layer neurons. But how is early thalamic input changing the future fate of cortical cell types? Our work indicates communication with progenitors may be the answer…

Utilizing the human thalamocortical assembloid model, we show that thalamic input increases the generation of upper layer cortical neurons compared to cortex-only controls. 🧫 This work shows extrinsic cues like thalamic input can shape intrinsic cortical programs.

Excited to present our new preprint led by @claudianguyen95 uncovering how thalamic input shapes human cortical development! We discover that thalamic axons promote the generation of upper layer cortical neurons through NRXN1-mediated contacts with outer radial glia. www.biorxiv.org/content/10.1...

Congratulations @prnano9.bsky.social!!! We are thrilled that this paper is out and hope everyone enjoys using the datasets and meta-atlas tools!!

We hope this dataset including metabolomics and scRNA-seq data will be useful to others who are interested in metabolism and cortical cell fates! Shoutout to @madelineandrews.bsky.social and all of our collaborators and the @uclastemcell.bsky.social for contributions and input on this project.

This led us to a model that indicates that glycolysis, through PPP, seems to restrict radial glia from generating oRGs, astrocytes, and inhibitory neurons too early in development. This may be why inhibition of glycolysis and PPP led to an increase of these cell types.

To test if PPP was guiding radial glia cell fate, we (painfully) extracted EGFP labeled radial from organoids and knocked down PPP genes. Indeed, same story. More oRGs, astrocytes, and inhibitory neurons paired with a decrease in PPP metabolite abundance.

This was fascinating because PPP also peaks in primary tissue at the same time as glycolysis. Was glycolysis working through the PPP? We tested this by inhibiting PPP with pharmacological inhibition, and also saw an increase in oRGs, astrocytes, and inhibitory neurons.

Organoids love sugar! Using staining and scRNA-seq, we saw an increase in outer radial glia (oRGs), astrocytes, and inhibitory neurons. When we examined metabolite abundance in this experiment many pathways, including the pentose phosphate pathway (PPP) were downregulated.

This was intriguing because glycolysis has been a flash point and somewhat controversial in organoids. Is it up in organoids? And if so, does it impact cell fate? We tested this by lowering glucose levels in organoid culture and measuring how cell types changed.