Daniel Dan Liu

Primary human NSPCs thus make a great system for modeling gliomas, allowing for modular introduction of genetic manipulations into isogenic cell types at defined developmental stages. Being human cell lines, they also recapitulate human-specific cell types not found in mice. 10/

To see if our results match up with real patient data, we reanalyzed some published GBM scRNA-seq datasets. Higher EGFR expression in the tumor indeed correlated with more GPC-like tumor cells. This suggests EGFR is not just a marker, but also a driver of the GPC-like state. 9/

The mutation-of-origin had some even more dramatic effects, with CDK4 overexpression resulting in largely neuron-like tumor cells, and EGFR interestingly expanding the GPC-like tumor cells. 8/

The cell-of-origin had some interesting effects on the resulting tumor. NSC-derived tumors had the greatest cellular heterogeneity, while OPC-derived tumors harbored more differentiated oligodendrocyte-like cells. 7/

Our lines engrafted robustly, growing and migrating aggressively. We only transplanted the cells on one side of the brain, but here you can see them invading cross-hemisphere. 6/

So we asked: can we use these primary human NSPCs to model how cell-of-origin and mutation-of-origin drive different glioma subtypes?

Using isogenic NSC, GPC, and OPC neurosphere lines, we created glioma-like lines by introducing defined combinations of oncogenic driver genes. 5/

We previously described how to purify 10 distinct neural stem and progenitor cell types from the developing brain, including tripotent radial glia (NSCs), bipotent glial progenitor cells (GPCs), and unipotent oligodendrocyte progenitor cells (OPCs). 4/
bsky.app/profile/dani...

Even more interesting is that different glioma subtypes seem to embody different stem cell hierarchies: low-grade oligodendrogliomas have a lot of mature oligodendrocyte-like cells, while diffuse midline gliomas mostly contain OPC-like cells(Figure from Suvà& Tirosh, 2020). 3/

Brain tumors known as gliomas are notoriously hard to treat, in part due to how heterogeneous the cancer cells are. In fact, glioma cells seem to mirror a corrupted neural stem cell hierarchy found in normal development. (Figure credit Neftel et al., 2019). 2/

Modeling glioma intratumoral heterogeneity with primary human neural stem and progenitor cells Gliomas are characterized by extensive intratumoral heterogeneity, recapitulating a hierarchy reminiscent of normal neurodevelopment. This facet is difficult to recapitulate in disease models. Here, W...

Our new study, “Modeling glioma intratumoral heterogeneity with primary human neural stem and progenitor cells,” is online now at @stemcellreports.bsky.social ! Spearheaded by our talented undergrad @danielgao970.bsky.social. A thread on the study: 1/

Cortical Development 2022 was the first conference I ever attended, and so it was particularly special to return to Sicily for @corticodevelopment.bsky.social 2025 as a speaker. Featuring a sneak peak on some exciting new findings on neural stem cells in the postnatal human brain (stay tuned!)

Looking forward to meeting this amazing group of neuroscientists at the @utah.edu Rising Stars Symposium!

Weissman Lab representing at the @isscr.org Athens International Symposium on Neural Stem Cells! Glad to share our work on purifying neural stem/progenitor cells from the human brain, and a preview of our work using those cells to model glioma heterogeneity.

Hi Oliver, would love to join this! MD-PhD student working on neural stem cells and human neurodevelopment here.

Our 2024 CSHL conference on human brain development and stem cell models is in full swing, with over 300 participants!

This year’s event features exciting new discoveries and dynamic discussions, driven largely by the energy & insights of trainees

#Assembloids #Organoids #Brain
#CSHL2024Brain

A huge thanks to everyone who made this study possible: my mentor Irv Weissman, the lab's brain team (Rahul, Nobuko, Joy, Anna et al.), as well as @czbiohub.bsky.social and Stanford MSTP for their support! 16/

Cell Press: STAR Protocols STAR Protocols is an open access, peer-reviewed journal from Cell Press. We offer structured, transparent, accessible, and repeatable step-by-step experimental and computational protocols from all are...

We have since published a methods paper detailing the dissociation, staining, and cell sorting protocols needed to replicate this work, which can be easily expanded to incorporate new surface markers. 15/

Purification scheme for neural stem and progenitor cells from the human brain.

And here’s our final purification scheme! My hope is that this study provides a framework for studying the cellular basis of human brain development, including cell-autonomous functions and interactions between cell types. 14/

Contrasting Rodent and Human Neocortical Development (Lui, Hansen, Kriegstein, Cell 2011)

These GPCs mostly lived in the outer subventricular zone, a layer of the cortex that is greatly expanded in primates. These cells may thus play an evolutionary role in the dramatic white matter expansion characteristic of human brains. 13/

Because we had the index sort data, we could immediately figure out how to purify these cells. Through clonal differentiation and transplant assays, we found these cells made oligodendrocytes and astrocytes, but NOT neurons. We thus named it the glial progenitor cell (GPC). 12/

Identification of a cell type with mixed glial markers.

And that’s that! Or so we thought, if not for this mysterious cluster in our scRNA-seq dataset. It expressed both astrocyte and oligodendrocyte markers. I thought it was an artefact, but we indeed found such cells in the human cortex with immunostaining. 11/

Purified neurons are unipotent.

And then the neurons, which were CD24+THY1–. We can sort out early/late excitatory neurons as well as inhibitory neuron precursors, which uniformly give rise to neurons in vitro. 10/

Purified OPCs are unipotent.

Up next, the oligodendrocyte (OL) lineage, which were all THY1 hi. We figured out how to further purify pre-OPCs, OPCs, and mature OLs, and show that these cells are truly unipotent in vivo (they give rise to OLs and nothing else). 9/

Purified radial glia are tripotent.

First off, the neural stem cells (and astrocytes) which were CD24–THY1–. We further figured out how to purify two subsets of radial glia (ventricular and outer radial glia), and showed they give rise to all three neural lineages in vitro and in vivo. 8/

Index sorting allows mapping of transcriptomic cell types onto their cell surface marker profile.

Now, we can map transcriptomically-defined cell types onto the FACS plot. Lo and behold, surface marker expression mapped onto cell types remarkably well! Now that we can purify, we can start doing experiments. 7/

Index sorting option on FACSDiva software.

But scRNA-seq doesn’t do us any good on its own. That is where *index sorting* comes in. While sorting, we record exactly which single cell was deposited into each well for sequencing. This gives a direct map between transcriptome and immunophenotype (surface markers). 6/

FACS purification of human brain cells for scRNA-seq

We started by staining dissociated brain tissue with a variety of different surface markers, which we analyzed with flow cytometry. We sorted single cells for Smart-seq, and captured all expected cell types. 5/

The neural stem cell hierarchy in the developing cerebral cortex.

In brain development, neural stem cells (radial glia) give rise to the neurons, oligodendrocytes, and astrocytes of the mature cerebral cortex. But, purification strategies for all these different cell types are not as thoroughly worked out. That is the gap this project sought to fill. 4/

The hematopoietic stem cell hierarchy

The importance of prospective isolation is particularly salient in the blood system, where over the years, the Weissman lab and many others have developed sophisticated gating strategies for distinct hematopoietic stem/progenitor cell types. 3/

First, the motivation: the study of human stem cells relies on robust prospective isolation methods. To study a cell type, you need to be able to purify it to perform rigorous functional experiments. This is especially the case in human tissues where classical genetic tools are not available. 2/