Single-cell proteomics defines the cellular heterogeneity of localized prostate cancer
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (PREreview)
Abstract
Article activity feed
-
-
This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/4608736.
Prostate cancer can be assessed using clinicopathological information which takes into account factors such as patient age, PSA levels and also Gleason score. This enables clinicians to determine the severity and progress of the tumour, as well as to determine the most efficient therapeutic agents to treat the disease. Whilst most patients can be treated successfully to reduce tumour burden and to prevent tumour spread, some patients may progress to a more advanced disease state or develop metastatic disease, resulting in worse clinical outcomes. De Vaargas Roditi et al. suggest that the basis for this discrepancy in patient outcomes lies in the heterogeneous nature of prostate tumours …
This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/4608736.
Prostate cancer can be assessed using clinicopathological information which takes into account factors such as patient age, PSA levels and also Gleason score. This enables clinicians to determine the severity and progress of the tumour, as well as to determine the most efficient therapeutic agents to treat the disease. Whilst most patients can be treated successfully to reduce tumour burden and to prevent tumour spread, some patients may progress to a more advanced disease state or develop metastatic disease, resulting in worse clinical outcomes. De Vaargas Roditi et al. suggest that the basis for this discrepancy in patient outcomes lies in the heterogeneous nature of prostate tumours which is not accurately identified using current clinicopathological techniques. This current work therefore aims to identify biomarkers that are adequately able to represent the inherent heterogeneity of prostate tumours, and work in parallel with the current Gleason score grading. Such biomarkers may be able to accurately predict patients which will go on to develop more advanced and/or metastatic disease through more tailored and frequent testing and monitoring.
Using mass cytometry analysis conducted on single cells taken from prostatectomies, the authors hypothesise that using this technique combined with a novel method of unsupervised cell clustering, they would be able to identify multiple common tumour cell populations in a highly sensitive and time efficient manner. This method also has the unique ability to identify rare/less common cell populations which are harder to detect using traditional sequencing methods such as single cell RNA sequencing. This is a very powerful technique as the number of cells analysed from the clinical samples is very large, and therefore allows a large sampling from the cohort upon which to conduct further analysis. The ability of Franken to be able to discern clusters of cells based on a gradient of protein expression was particularly interesting as it increased sensitivity, and it allowed the authors to identify many more cell clusters based on the panel of antibodies which they used. Compared to other methods, such as single cell sequencing which may be restricted in read depth due to cost and cell number following dissociation, mass cytometry paired with the Franken pipeline represents a very user friendly and cost-effective tool for researchers to use. There is a lot of potential for this to have impact not only in prostate cancer, but also in many different areas of research.
This study provides a comprehensive overview of cell clusters which are present in clinical prostate cancer cases across a variety of tumours ranging from Grade II to Grade V. There is a great and pressing need for reliable and common biomarkers in prostate cancer to aid the identification of patients who are at risk of developing advanced stage disease or metastatic spread, and this research identifies multiple cell populations which may be of use to identify such markers.
The authors remark that they find similar clusters of cells in the ABPT and tumour regions, which was interesting. This could have been due to limitations induced by the macroscopic sampling technique, but a more pronounced difference would have been the expected result, particularly between the tumour samples and the ABPT. It may be interesting to see whether a comparison between truly benign prostate tissue (e.g. from a healthy patient) would display the same cell clusters or not? It may also be beneficial to have age-matched controls to better discern the differences between healthy and diseased tissue. Additionally, have the authors looked into changes between ethnicity or age of patients and how that relates to the identified cell clusters?
It was also surprising to see that there was only one stromal cell population outside of the multiple immune cell populations, given the large number of stromal cells which are found within both healthy tissue and prostate tumours. Could the authors comment what these cells could be and whether the finding of only one stromal population fit with their expectations for the data?
The data displayed in Figure 2 (a-d) is very clear, and the use of UMAPs to present the data makes it obvious how all of the different cell populations are represented equally across all of the patient samples. This is a great visual way to see that there don't seem to be inter-patient/batch variabilities, which again highlights what a powerful tool Franken is as this is a great benefit to researchers wanting to study large datasets or clinical cohorts. A very minor point about the Figures would be to note that in Figure 2, the page was cropped a bit short at the bottom of the page so panel e and f are not completely visible.
Within the manuscript the authors state that the panel of proteins they used does not cover the entire proteome, and that therefore there could be cell type present which do not fit into the clustering (represented by population NE01. Do the authors have any suggestions for what these cell types would be? Or if they were to expand the number of proteins that they used in the mass cytometry experiments, do they believe that many more cell types would be determined? Or do the authors feel like this panel sufficiently covers the cell populations which are present within human tumours? To validate this data, and to expand on the findings, it would be interesting to carry out a multi-channel fluorescent staining (such as Codex or Hyperion) in parallel on these clinical samples to see spatially how the identified cell clusters are located within the tumour and adjacent benign tissue regions. I would be curious to see if there is any correlation between the distinct areas of the tumours (as defined by the Gleason score or particular pathology) and the cell populations which are present in them. This also may aid the identification of biomarkers and to associate particular cell populations/clusters with more vs. less advanced disease states. It may also be helpful to better distinguish between the ABPT and RPT tissue samples, and to identify areas of tissue contamination.
For future studies, it would be very interesting to see how the cell populations change when looking at more advanced disease, such as CRPC or NEPC and how the particular populations such as the basal or luminal cells change over time and during AR positivity vs. negativity.
-
