Understanding intratumoral heterogeneity to defeat cancer
Figure 1 – schematic representation of intratumoural heterogeneity-created with BioRender
Abstract
Despite the advances in research, cancer remains the second cause of death in the world. Intratumoral heterogeneity is the main obstacle in the fight against cancer because it drives to drug resistance and tumour progression. The recently developed single-cell RNA sequencing methods opened new opportunities to address the tumour transcriptional heterogeneity. By using this methodology, Gavish et al.1 have identified new transcriptional metaprograms, which define cell type composition within the tumours and identify pre-cancerous cells, that may represent a new therapeutic target. This study allows for a more comprehensive understanding of several cancer types and may offer novel therapeutic and diagnostic avenues.
Review
Intratumoural heterogeneity (ITH)
The development of tumours is a complex process driven by genetic, epigenetic and microenvironment changes that diversify cellular phenotypes, making cancer treatment challenging2. The term ‘tumour heterogeneity’ is used in diagnostic pathology to describe the differences in appearance both between different tumors (intertumoral heterogeneity) and within the same tumor (intratumoral heterogeneity). With the introduction of molecular pathology, heterogeneity has been studied extensively through transcriptional profiling at the RNA level3. It has been observed that ITH appears to be associated with worse tumour phenotypes and drug resistance. Therefore, understanding ITH is crucial to improve cancer diagnosis, prognosis, and therapy as this is a fundamental determinant of tumour behaviour and response to treatment.
Discussion
In recent years, the development of more sequencing techniques, such as single-cell RNA sequencing (scRNA-seq) in ITH analysis has enabled better identification of expression programs in each cell. Metaprograms (MPs) are defined as sets of genes most shared between malignant and non-malignant cells in different tumours.
Through previous scRNA-seq analysis, 198 tumour cell lines were studied to identify 12 MPs in vitro. Another study analysed scRNA-seq data for 62 primary tumours of various cancer types and found 16 MPs. In this study, to understand the different cellular states in human tumours, data from different sources, such as tumour datasets from human tumours and mouse models, were utilized and then integrated using computational approaches that allowed the elimination of batch effects.
Starting from a huge amount of data, a non-negative matrix factorization (NMF) was used to characterize ITH programs, identifying approximately 5,000 “robust NMF programs” which were subsequently grouped into cluster eliminating doublets and low-quality programs. After clustering, the resulting MPs were further divided into 11 families based on functional similarities. Among the 11 MPs families ultimately identified, cell cycle and stress were found to be the two primary indicators of ITH. The remaining nine hallmarks occur less frequently and are specific to unique contexts (Figure 2).
Figure 2 – Identification of 11 MPs families-created with BioRender and Canva
Analysis of MPs in non-malignant and pre-cancerous cells revealed similarities in the expression of some MPs with malignant tumour cells, suggesting the presence of intrinsic heterogeneity also in the cells of origin. Moreover, the analysis of pre-malignant tissues confirmed an intermediate stage between malignant and non-malignant cells, wherein there is a higher similarity of MP expression to malignant than to non-malignant epithelial cells.
The analysis was performed using the same method used to define MPs in malignant cells. This approach allowed data to be compared between malignant and non-malignant cells. Using previously published cell annotations, they applied the NMF method to define robust normal programs of these cells. They integrated the data obtained with those of the malignant cells already available. This analysis highlighted the possibility of pre-existence of malignant MPs.
Finally, they studied the presence of MPs within the tumor microenvironment (TME) to determine the effect of interactions between different cell types in the TME. They managed to identify positive and negative correlations between various MPs. In particular, they observed that four clusters of MPs show a positive correlation with each other: the cell cycle, interferon, stress, and heat shock proteins (HSPs). This positive correlation between clusters suggests a possible coordinated response of different cell types present in the same TME.
Conclusions
In conclusion, this study is of fundamental importance for understanding tumours at the transcriptional level. It highlighted two forms of heterogeneity within tumours: intertumoral heterogeneity, which reflects cumulative differences between tumours due to genetic and epigenetic variations, and intratumoral heterogeneity, which is more dynamic and linked to events that shape the state of all cells inside a single tumour. This distinction could contribute to the development of more effective therapies.
Thanks to this study, it is also possible to have an updated database with recent information, since all the collected data from this study are now available through the Curated Cancer Cell Atlas (3CA). Furthermore, the software tools provided by the authors allow the quantification and visualization of gene expression patterns, facilitating the understanding of ITH and the design of new therapeutic strategies.
Comparison of overall cellular states between tumours has highlighted the need for targeted therapeutic approaches, given the complexity of tumours. The identification of shared cell subpopulations between tumours offers new prospects for combination therapies targeting coexisting cellular states, while differentiation therapies could direct tumour cells towards more benign or reactive states.
Understanding whether there are more heterogeneous tumours means that they may require differentiated treatment. It may be necessary to use a combination of drugs to target all the different cell populations within the tumour. It is tempting to speculate that also epigenetics can play a very important role; for example, inter- and intra-tumoural DNA methylation patterns may provide new prognostic parameters. In general, the more heterogeneous a tumour is, the more difficult the treatment is and the lower the survival rate of patients may be.
From a practical point of view, this study offers new perspectives for research in the oncologic field as well as in the treatment of cancer patients. This study lays the foundation for improved tumour treatment, potentially overcoming current challenges related to drug resistance and relapse. The final aim is to develop new therapeutic and diagnostic tools for patients.
Developing studies appropriate for precision medicine faces many barriers, including identifying and validating predictive molecular biomarkers, technical limitations of molecular tests, limited research advancement for some cancer types, and time- and cost-related issues. Finding ways to accurately identify and control tumor heterogeneity and translating these achievements into patient benefits are major goals in modern oncology. Precision medicine will not be possible until heterogeneity-related issues are resolved 4.
References
- Gavish, A. et al. Hallmarks of transcriptional intratumour heterogeneity across a thousand tumours. Nature 618, 598–606 (2023).
- Marusyk, A., Almendro, V. & Polyak, K. Intra-tumour heterogeneity: a looking glass for cancer? Nat Rev Cancer 12, 323–334 (2012).
- Stanta, G., Jahn, S. W., Bonin, S. & Hoefler, G. Tumour heterogeneity: principles and practical consequences. Virchows Arch 469, 371–384 (2016).
- Seoane, J. & De Mattos-Arruda, L. The challenge of intratumour heterogeneity in precision medicine. Journal of Internal Medicine 276, 41–51 (2014).