Towards a New Diagnostic to Identify Lung Cancer Patients Who Respond to Immunotherapy by Analyzing Blood-born microRNAs

Timothy Rajakumar, MD, PhD & Bruno Steinkraus, PhD

“It is more important to know what sort of person has a disease than to know what sort of disease a person has”. Hippocrates

Two and a half thousand years after this adage was recorded, medical science is nearing the point at which we can do justice to this insightful guiding philosophy. Nowhere is this better exemplified than in recent advances that have occurred in the personalized approach to oncology. This is particularly vital in the emerging field of immunotherapy, in which a patient’s own immune system is redirected to become a powerful anti-cancer force.

Immunotherapies — such as anti-PD-(L)1 therapy — have demonstrated remarkable efficacy in a subset of patients. Yet, it is difficult to predict exactly which patients will respond to them. An important question remains: Who should be given these powerful drugs?

In a recent study, we have sought to address this dilemma through the systematic profiling of the immune system in a cohort of immunotherapy treated lung cancer patients and have defined a molecular signature that is predictive of the success of immunotherapy. We believe that our approach could pave the way towards a more individualized use of immunotherapies so that they can be given to the patients who will benefit from them most.

Effective for Some, but Not for All

Immunotherapies have been developed based upon the groundbreaking work by Honjo and Allison who found that some tumors evolve to evade the immune system via the hijacking of the cytotoxic T lymphocyte-associated antigen 4 (CTLA4) and/or Programmed Death 1 (PD-1) signaling pathways (Hoos 2016). The therapeutic blockade of such immune checkpoints can be highly effective in patients whose cancers exploit these pathways. That said, these patients have been challenging to identify.

The current gold standard biomarker to select patients for anti-PD-(L)1 therapy is the measurement of tumour PD-L1 expression. The rationale behind this is that patients whose tumours express higher levels of PD-L1 are likely to be reliant on the activity of this immunosuppressive signaling pathway and therefore at increased susceptibility to its inhibition. Whilst this hypothesis is certainly supported by clinical trials that have shown a weak positive correlation between the efficacy of immunotherapies and PD-L1 expression (Mok et al. 2019), this biomarker unfortunately displays a relatively poor predictive value. This is not surprising given the complex dynamics between tumour and host immune system that govern the response to immunotherapy, and the fact that PD-L1 expression only reflects one side of this multifactorial equation. Furthermore, the measurement of PD-L1 expression only reflects a snapshot of a small portion of the tumour, without consideration of the tumour as a heterogeneous whole.

A New Approach Using microRNAs

To address these limitations, we set out to define a more holistic biomarker that broadly represented the host immune phenotype and its potential to mediate a beneficial immunotherapy response.

We collected blood from patients with late-stage lung cancer who were due to be treated with immunotherapy and performed genome wide analyses. These analyses focused on a fascinating class of RNA known as “microRNAs,” which have gained a reputation as master orchestrators of gene expression and have been linked to virtually all physiological processes in health and disease (Bartel 2018). As such, the expression of these miRNAs can be thought of as tiny thermometers reflecting molecular health.

Instead of conducting a liquid biopsy, which would analyze circulating tumor cells, we have chosen a rather unique approach to a blood-based test that collects and measures the molecular profiles of whole blood samples. This enables the collection of clinical material in an extremely simple manner, avoiding the laborious processing associated with the sampling of alternate biological fluids.

This technique also offers a twofold benefit in that often-problematic sample-to-sample variation was minimized, and any findings from this study could be more rapidly developed into tests that may fit into existing clinical workstreams. One challenge of sampling whole blood is the enormous number of non-immune cell (e.g., red blood cells, platelets) compared to immune cells, from which we aimed to infer immunophenotypes and immunotherapy response. This necessitated the development and optimization of a technical sample processing pipeline that enriched the measurement of miRNAs coming from cells of the immune system and is described in more detail in the accompanying paper.

Using our optimized technology, we have measured miRNAs from the blood of 334 well-characterized lung cancer patients and applied machine learning methods to identify an expression signature that can predict the survival of cancer patients due to start immunotherapy. This 5-miRNA signature — the miRisk score — was validated in an independent cohort of patients where it was shown that patients predicted to be at low risk survived for a significantly longer time following treatment than those deemed high risk.

Finally, we sought to understand the biological relevance of the 5 miRNAs that comprise miRisk and how they may be linked to the interaction between immune system and tumour in the context of immunotherapy. For this we created a blood cell miRNA expression atlas by measuring miRNAs in 10 purified dominant cell types from blood to trace the cellular origin of the miRisk miRNAs. This revealed a striking bias towards an origin in cells of a particular branch of the immune system known as myeloid cells. This finding is consistent with a growing body of evidence implicating the role of myeloid immune cells in configuring the tumor microenvironment, where they have been shown to be associated with anti-tumor immunity (Liu et al. 2020)

Furthermore, we have used bioinformatic methods to identify potential targets of regulation by miRisk miRNAs. This analysis revealed several predicted miRisk miRNA targets in the PD-L1 pathway, including in PD-L1 itself. Together, this makes a direct functional link between miRisk miRNAs, relevant cells of the immune system and the mechanism of action of immunotherapy drugs.

Next Steps

In conclusion, our study describes a miRNA signature that can predict the response of lung cancer patients to immunotherapy, measured from a simple blood test. Given its immune-centric origin, this signature may also be of predictive value in other cancers that are treated with immunotherapies, which we seek to explore this in future clinical trials. Our next steps include the validation of the miRisk score in additional lung cancer cohorts and other cancer types where PD-(L)1 inhibition is utilized. Overall, the miRisk score could represent an important contribution to the personalized management of oncological patients.

Bartel DP (2018) Metazoan MicroRNAs. Cell 173:20–51.

Hoos A (2016) Development of immuno-oncology drugs — from CTLA4 to PD1 to the next generations. Nat Rev Drug Discov 15:235–247.

Liu Y, Zugazagoitia J, Ahmed FS, et al (2020) Immune Cell PD-L1 Colocalizes with Macrophages and Is Associated with Outcome in PD-1 Pathway Blockade Therapy. Clin Cancer Res 26:970–977.

Mok TSK, Wu Y-L, Kudaba I, et al (2019) Pembrolizumab versus chemotherapy for previously untreated, PD-L1-expressing, locally advanced or metastatic non-small-cell lung cancer (KEYNOTE-042): a randomised, open-label, controlled, phase 3 trial. Lancet 393:1819–1830.

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in

Subscribe to the Topic

Cancer Biology
Life Sciences > Biological Sciences > Cancer Biology

Related Collections

With collections, you can get published faster and increase your visibility.

AI in precision oncology

This Collection is a partnership between npj Precision Oncology and npj Breast Cancer. It will bring together articles on all facets of AI in cancer research.

Publishing Model: Open Access

Deadline: Apr 19, 2024

Innovations in cancers of the central nervous system

This Collection invites research on tumors involving the central nervous system, including primary glial tumors, meningiomas and metastatic tumors in adults.

Publishing Model: Open Access

Deadline: Mar 01, 2024