2D and 3D multiplexed subcellular profiling of nuclear instability in human cancer

Background

Pathologists have been using abnormal nuclei to distinguish cancer cells from healthy tissue for over a century. Nuclear atypia, which includes micronuclei, abnormal nuclear shape and size, and chromosomal bridges, are some of the most visually striking changes that occur during tumour development. Despite these long-standing observations in the field, how nuclear ruptures occur and how they contribute to cancer progression remains mysterious. This is primarily due to the lack of crosstalk between the histopathological slides used in diagnostic labs and data produced by research labs, which is typically derived from live-cell imaging and low-plex immunofluorescence imaging.

To overcome this gap, the authors of this preprint combined conventional microscopy, multiplex imaging techniques such as CyCIF (cyclic immunofluorescence), and 3D electron microscopy. Using this toolkit, the authors dissected the nuclear states present in glioblastoma, the most common and aggressive form of brain cancer in adults, and uncovered previously unknown spatial patterns in nuclear rupture and immune activation.

Key Findings

Above: An overview of the sample types, experimental techniques, and imaging analyses conducted by the authors.

Primary and micronuclear ruptures are common in human cancer and are associated with distinct protein expression patterns

The authors first sought to determine the frequency of nuclear rupture events across different types of cancer. To do this, they stained BAF (barrier-to-autointegration-factor, a marker of nuclear envelope rupture) in over 100 samples representing nine different cancer types and classified ruptures as primary nuclear (PN) or micronuclear (MN). They found that PN ruptures were particularly frequent in glioblastoma, which they confirmed with high resolution techniques such as 3D expansion microscopy and Focused Ion Beam Scanning Electron Microscopy (FIB-SEM). Using FIB-SEM, they found extensive chromatin infiltration into the cytoplasm, often weaving between endoplasmic reticulum tubules and mitochondria.

Above: Quantification of the frequency of PN and MN BAF+ ruptures across different types of cancer. Glioblastoma displays a particularly high frequency of PN ruptures.

Above: Three-dimensional reconstruction of FIB-SEM micrographs showing nuclear envelope ruptures and chromatin weaving through cytoplasmic organelles.

Primary nuclear rupture and lamin expression are correlated with glioblastoma tumor cell states

The authors sought to uncover gene expression patterns associated with nuclear instability in glioblastoma. Using data from The Cancer Genome Atlas (TCGA) Pan-Cancer Study¹, they found that glioblastoma expresses lamin A/C at a much lower level compared to other cancers. Furthermore, The Cancer Dependency Map (DepMap), a dataset containing RNAi and CRISPR knockout data of multiple cancer types, revealed that glioblastoma cells are especially dependent on lamin A/C expression².

To further dissect the expression pattern of nuclear lamins in glioblastoma samples from patients, the authors performed CyCIF using 28 different antibodies against lamins as well as proteins involved in proliferation, nuclear envelope rupture, DNA damage, and cell identity. Interestingly, they found that tumour cells with low expression of lamin A/C were more likely to have a PN rupture, hinting that reduced lamin A/C makes tumour cells more vulnerable to instability.

Multiplexed 3D confocal imaging of nuclear instability and cell state in glioblastoma

Given that glioblastoma is a notoriously heterogeneous disease with four different molecular groups (oligodendrocyte precursor like, neural progenitor-like, astrocyte-like, and mesenchymal-like)³, the authors wondered whether these states differed with respect to nuclear atypia. Using whole slide tissue specimens, which contain tumours as well as the surrounding healthy tissue, the authors found distinct spatial patterns in the distribution of glioblastoma cell states. Mesenchymal-like cells were enriched in peri-necrotic zones, astrocyte-like cells were enriched in the tumour core, while oligodendrocyte-like and neural progenitor-like cells were typically found in the invasive border.

The authors then combined single cell RNA-seq data with multiplex CyCIF to profile the nuclei of different glioblastoma cell states in tissue microarrays from GBM patients. This combination of techniques was necessary given the lack of a defined transcriptional signature for acute nuclear rupture. They found that neural progenitor-like cells were the most proliferative, had the lowest expression of lamin A/C, and the highest rate of nuclear ruptures compared to the other groups. These observations were further confirmed using 20 micron-thick tumour sections, which allowed the authors to make 3D reconstructions of whole cells and clearly visualize rare phenomena such as chromosomal bridging between nuclei.

Above: 3D reconstruction of a 20 micron-thick glioblastoma tissue section highlighting the formation of chromosomal bridges between nuclei and associated DNA damage and nuclear rupture markers.

Primary and micronuclear ruptures are associated with cGAS-STING and interferon signaling.

While immunotherapies have been of great use in targeting different tumour types, it has offered limited clinical benefit for patients with glioblastoma. However, PN and MN ruptures have been demonstrated to activate the pro-inflammatory cGAS/STING pathway in some contexts^4,5. With that in mind, the authors performed CyCIF on glioblastoma tumour microarrays using antibodies against immune, cGAS/STING, DNA damage, and interferon response markers. In tumour cells with BAF+ or cGAS+ nuclear ruptures, the authors found nuclear translocation of inflammatory transcriptional activators such as IRF3 as well as expression of interferon-stimulated genes such as IFITM1/2/3. These protein expression patterns suggest that nuclear atypia in glioblastoma may promote downstream inflammatory signalling.

 Why I chose this preprint

I found this preprint particularly interesting due to my own research on nuclear envelope integrity in medulloblastoma, one of the most common malignant pediatric brain cancers. Medulloblastoma also has a high degree of transcriptional and spatial heterogeneity, and it would be very interesting to investigate whether different cell populations within tumours have different degrees of nuclear instability. I was impressed by the authors’ combination of multiple imaging techniques to capture events that are often missed through fixed cell or tissue imaging using thin sections, such as chromosomal bridging. I believe the findings described in this preprint will inspire other researchers who are interested in nuclear instability to use more comprehensive imaging tools to ensure they aren’t missing rare phenomena.

 Questions for the authors

1. You observed that lamin A/C expression appeared to be lower in neural progenitor-like cells, and that these cell types were enriched at the infiltrating border in glioblastoma. Do you think that having a more deformable nucleus could provide some sort of advantage to invading cells?
2. Glioblastoma is typically thought to be an immune “cold” tumour. Do you believe that in the future, clinicians will be able to integrate nuclear atypia into a predictive measure of responsiveness to immunotherapy?

 References

1. Liu, J. et al. An Integrated TCGA Pan-Cancer Clinical Data Resource to Drive High-Quality Survival Outcome Analytics. Cell 173, 400-416.e11 (2018).
2. Tsherniak, A. et al. Defining a Cancer Dependency Map. Cell 170, 564-576.e16 (2017).
3. Neftel, C. et al. An Integrative Model of Cellular States, Plasticity, and Genetics for Glioblastoma. Cell 178, 835-849.e21 (2019).
4. Denais, C. M. et al. Nuclear envelope rupture and repair during cancer cell migration. Science 352, 353– 358 (2016).
5. Nader, G. P. de F. et al. Compromised nuclear envelope integrity drives TREX1-dependent DNA damage and tumor cell invasion. Cell 184, 5230-5246.e22 (2021).

 

Tags: cancer, glioblastoma, imaging, nuclear, nucleus

doi: https://doi.org/10.1242/prelights.36273

Read preprint (No Ratings Yet)