Applying the lotus-effect to liver cancer research

Predicting liver cancer prior to its development to improve early detection and survival.

Project Timeline:
March 2022 – September 2025 (Expected Completion).

Project Summary:
Hepatocellular carcinoma (HCC), the most common type of liver cancer, accounts for approximately 10% of cancer- related deaths worldwide. In Australia, HCC is among the fastest-growing causes of cancer mortality. It often arises from chronic liver conditions such as viral hepatitis, alcohol-related liver disease, and metabolic dysfunction-associated fatty liver disease (MAFLD). Unfortunately, current methods do not reliably predict which patients will develop HCC, apart from using abdominal ultrasounds in cirrhotic patients. This limitation results in delayed diagnoses and poor survival outcomes.

My research aims to identify early signs of liver cancer before existing methods can detect them. Our research team has recently discovered a specific cell type in the liver at high risk of becoming cancerous, termed disease-associated hepatocytes (daHeps). I am working to develop techniques to detect these daHeps in individuals with liver disease, potentially enabling earlier detection and significantly improving patient outcomes.

Research Progress

Using long-read Oxford Nanopore sequencing, my current work is characterising the whole genome methylation landscape of human pre-cancerous daHeps. I have been analysing these complex DNA sequencing data to identify molecular markers that distinguish precancerous cells from healthy ones. So far, I have successfully sequenced and interpreted DNA profiles from several liver samples.

The next phase of my research involves detecting these markers in blood samples. Since blood contains DNA fragments from dying liver cells, my goal is to develop a simple blood test that can identify and quantify these precancerous cells (daHeps) early.

In addition to DNA sequencing, I am utilizing spatial transcriptomics (GeoMx Digital Spatial Profiler) to explore the spatial organization of daHeps. Using this technique, I have mapped the locations of daHeps in the liver tissues of patients with cirrhosis and liver cancer. My findings indicate that daHeps are concentrated in areas of fibrotic damage and immune activity, suggesting that these cells play a crucial role in the progression of liver disease. By charting their distribution within the liver, I aim to better understand their function and enhance their use as early biomarkers of liver cancer. This could lead to substantial advancements in the management of chronic liver disease, facilitating earlier diagnoses and more personalized treatment plans, ultimately improving patient outcomes.

Project title

Supercharging natural killer (NK) immune cells to eliminate leukaemia.

Summary

Therapies that harness the patient’s own immune system to fight cancer (known as immunotherapies) have shown remarkable success for patients with advanced leukaemia. However, these therapies are expensive, time-consuming, and not all patients are able to donate their own cells. This means there is an urgent need for an alternative approach. Our solution is to use the cancer killing immune cell, the natural killer (NK) cell, which can be collected from healthy donors, stored for later use, and given to multiple patients as soon as treatment is required. In this project we will develop new methods to enhance the activity of NK cells against leukaemia using immune-boosting molecules known as interferons

Aim 1: We will also use cutting-edge genetic technologies to understand why some NK cells are better at killing leukaemia than others

Aim 2: and we will then use this information to “supercharge” NK cells to further enhance and refine their cancer killing abilities

Aim 3: Our longterm goal is to create a highly effective NK cell therapy to ensure all patients can receive this life-saving treatment.

Project update – JSeptember 2024

Aim 1 – Completed December 2022
Aim 2.
In this aim we have used several cutting-edge sequencing technologies to identify new targets which can be used to improve NK cell activity against leukaemia. We have performed several major sequencing experiments on healthy donor NK cells with “high” and “low” abilities to kill leukaemia to identify the major genes and pathways that control NK cell function. From our sequencing analyses we have identified a target gene that we believe drives improved NK cell responses against leukaemia (investigated in Aim 3). We have also performed a more in-depth analysis at the single cell level to further explore the genes and pathways that control different populations of NK cells. From these experiments, we have found a panel of target genes that we are currently exploring in the lab. We hope to have some exciting results from these experiments soon.
Aim 3
Results from Aim 2 have allowed us to identify a target gene that we believe is a critical regulator of NK cell function. Using drugs that either activate or block this target, we have shown that several important aspects of NK cell biology can be controlled including proliferation, metabolism, responses to immune molecules, and anti-cancer function. Notably, we found that blocking this target significantly reduces the ability of NK cells to kill leukaemia in the lab even in the presence of known NK cell activators. We are currently writing a paper about this project which we hope to submit for publication by the end of the year.

We continue to be excited by the results from this project and believe that our new findings will significantly impact the field of NK cell therapy. We hope these findings can be used to create more effective and safer cell therapies that will improve the outcomes for patients with leukaemia.

Project Title

Using artificial intelligence combined with medical image data to improve glioblastoma patient outcomes.

Period Covered by this Report

30 September 2023 – 30 September 2024.

Project Summary

Glioma is a cancer of the brain and glioblastoma is the most aggressive form of glioma. Glioblastoma is difficult to treat and has poor survival rates with treatment involving surgery to remove as much tumour as possible, followed by chemotherapy at the same time (and after) radiotherapy. Doctors use medical images to inform treatment decisions, such as magnetic resonance imaging (MRI). Doctors visually inspect these images and make a judgement on whether current treatment is effective or not. These assessments inform on predicted patient survival and guide further treatment decisions. However, the digital data available within medical images can be harnessed to unlock further information about a patients prognosis. This process is called radiomics, which is often used to enhance predictive modelling. However, the methodology in radiomics studies can be highly variable, which complicates clinical feasibility; therefore, I have pivoted my work to explore how to best apply radiomics and artificial intelligence (AI) in a way that will increase the likelihood of clinical impact, described below under “radiomics modelling”. Specifically, radiomics is important for positron emission tomography (PET) when combined with a radioactive tracer in glioma imaging called O-(2[18F]-fluorethyl)-L-tyrosine (FET). Preparation for radiomics modelling will be conducted in parallel with credentialing analysis from the ongoing Australian, multicentre FET PET in Glioblastoma (FIG) trial. I am also investigating the use of an automated artificial intelligence pipeline applied to a relatively new MRI sequence called amide proton transfer (APT), which has also shown promising predictive power for patient survival.

Research Progress

RADIOMICS MODELLING
To best understand the key steps for clinical translation of radiomics models I needed to utilize available tools for evaluating radiomics publications. We found this was crucial before conducting radiomics analysis on either FET PET or APT MRI data. One prominent tool in the literature is called the Radiomics Quality Score (RQS), which consists of 16 criteria that measure clinical feasibility and is used by researchers when reviewing published radiomics research. However, a comprehensive metaanalysis (grouping results from multiple publications) of this tool across different cancer types had not been conducted.

Therefore, I systematically evaluated 130 reviews which applied the RQS in their reporting. A total of 3258 quality scores were extracted from these reviews and the subsequent analysis provided an update on the progress of radiomics research and identified that progress has occurred with newer publications, although this progress has been slow.

A manuscript was drafted of this systematic review and meta-analysis, which is currently under review in European Radiology. This work was also accepted as a poster for the European Association of Nuclear Medicine Congress 2024, which I will be attending. Additionally, this paper allowed me to identify what the general pitfalls and shortcomings are in radiomics research, which was then used as evidence in a set of guidelines; up-to date best-practices to conduct robust and realistic radiomics modelling moving forward. These guidelines have been turned into a draft manuscript, which is currently in the latter stages of revisions by co-authors.

FET PET
In my previous report, the Nuclear Medicine Physician credentialing analysis was accepted for publication in the European Journal of Nuclear Medicine and Molecular Imaging (DOI: 10.1007/s00259-023-06371-5). Since then, the Radiation Oncology credentialing analysis was accepted for publication in Physics and Imaging in Radiation Oncology (DOI: 10.1016/j.phro.2024.100568). Since I am yet to obtain a sufficiently sized dataset, with patient outcomes, from the ongoing FIG trial to conduct radiomics modelling, I focused on developing an automatic FET PET segmentation network.

However, rather than developing this on the limited data available, a senior researcher from the Institute of Neuroscience and Medicine in Juelich, Germany, reached out to collaborate to validate a segmentation network that they had already developed on a much larger dataset. Therefore, I conducted a collaborative study which looked at how well this segmentation network performed on a set of FET PET images of glioblastoma patients from our local institution. Although the network performed very well, it demonstrated a bias towards underestimating tumour volumes and sometimes segmented regions of the brain that weren’t tumour. This provided important insight, as this indicated that model fine-tuning or manual adjustment will be necessary if this network is to be deployed nationally. A manuscript has been drafted with planned submission to the Journal of Nuclear Medicine.

APT MRI
Steady progress has been made for the APT MRI arm of the PhD. Upon receiving approval for our retrospective study protocol, 73 glioblastoma patients from Sir Charles Gairdner Hospital met our inclusion criteria and their respective imaging, demographic, treatment, and outcomes data have been retrieved. The imaging data has since been used to validate an artificial intelligence based, open-source, automatic brain tumour segmentation toolkit, conducted with direct consultation from an experienced neuroradiologist.

A plugin for the medical imaging computing platform “3D Slicer” was also created for user-friendly execution of this toolkit and has been made available to other researchers. The work conducted so far has been accepted for oral presentation at the Engineering and Physical Sciences in Medicine conference (2023). Future steps are to apply these automatically generated segmentations to APT MRI and extract informative data from the tumours of each of the patients, which will be used in combination with survival data to develop a prognostic model.

Project Title

Boosting immune cells to find and kill pancreatic cancer cells.

Project Summary

Pancreatic cancer (PC) is one of the deadliest cancers that has a 5-year survival rate of only 11%. PC is usually detected very late, due to the cancer cells’ rapid growth and early metastatic capacity. PC is also very hard to treat and the therapies currently used don’t often work. With the number of people diagnosed with PC on the rise, we urgently need to improve our ways of detecting and treating this fatal disease.

My research focusses on understanding how PC cells transform the body’s immune cells by sending tiny ‘bubbles’, called exosomes. Exosomes are naturally released by all cells and carry many different molecules (proteins, lipids, genetic information and other substances). Cells use exosomes as a way of communicating by delivering molecules from one cell to another.

Research Progress

In my PhD, I am examining how PC cells use exosomes to transport biologically active lipids to immune cells, and if this contributes to disease progression. Our recent study revealed that PC cells, but not normal pancreatic cells, release exosomes containing an important lipid called sphingosine-1-phosphate (S1P). S1P is involved in many cell processes, and is also a key player in controlling the movement and function of immune cells.

The first part of my project has focussed on understanding whether S1P is involved in shaping the immune responses induced by PC-derived exosomes. I am currently looking at if these exosomes prevent macrophages (a type of immune cell) from polarising into anti-tumour phenotypes, which are involved in recognising and killing tumour cells. To study this, I have treated macrophages in different resting states (pro- or anti-tumour) with exosomes isolated from PC cell lines. These macrophages have then been analysed using different techniques to assess protein and DNA expression, to determine how the exosomes are affecting macrophage phenotype and function. In addition to this, I have treated the cells with inhibitors specific to the S1P receptors to see if blocking S1P prevents the effect of the exosomes. Interestingly, my data shows that when two of these receptors are blocked together, the effect of the exosomes is significantly decreased and the anti-tumour activity is rescued. In the next part of my project, we will validate if exosomes from PC patients contain abnormal amounts of S1P. If confirmed in the clinic, this will provide a novel marker for detecting PC.

Outcome / Impact

In whole, this project will add valuable insight into how exosomes promote PC development and progression. Exosomes are also an attractive source for disease-specific markers as they appear very early in disease and are found in all body fluids (blood, urine, saliva, tumour ascites), meaning samples can be collected early and non-invasively. Our data can provide novel, non-invasive markers to aid in detecting PC during its early stages. These results may also offer new ways to predict or track how different patients will respond to treatment, guiding the use of personalised therapies. As a result, this may improve the poor outcomes seen in PC by increasing patient survival and quality of life. Indeed, we believe this project can aid in the fight against pancreatic cancer.

Project title

Exploring the role of deubiquitinating enzymes to MAPK pathway inhibitor resistance in BRAF mutant melanoma.

Project timeline

May 2024 – November 2027 (Expected Completion).

Summary

Cutaneous melanoma is the third most diagnosed cancer in Australia, with rising morbidity and mortality rates observed worldwide. Despite advancements in melanoma therapies, the disease’s heterogeneity underscores the need for personalized treatment approaches. The hyperactivation of the MAPK pathway, particularly through the BRAFV600E mutation, is a hallmark of melanoma, prompting the development of targeted therapies. However, the efficacy of these targeted therapies is often hindered by primary or acquired resistance.

Recent research has highlighted deubiquitinating enzymes (DUBs) as critical regulators of melanoma progression and therapeutic resistance. DUBs control protein stability by removing ubiquitin modifications, influencing key processes like cell survival, proliferation, and metastasis in melanoma progression. DUB inhibitors have shown potential in preclinical studies, promoting apoptosis and reducing tumour growth. However, developing selective and potent DUB inhibitors remains a challenge due to the complexity of DUB functions and potential off-target effects. Further research is required to refine these inhibitors, explore combination therapies, and fully understand DUB-related mechanisms of resistance in melanoma. This project aims to investigate the role of DUBs, particularly USP9X, in melanoma progression and resistance to MAPK-targeted therapies. By advancing our understanding of these molecular pathways, we hope to contribute to the development of more effective, personalized treatment strategies for melanoma.

Research Progress

Our current research focuses on evaluating the efficacy of a novel USP9X inhibitor in treating various metastatic melanoma cell lines harbouring the BRAFV600E mutation. We are using a cell viability assays to assess the sensitivity of different cell lines to this inhibitor. Thus far, 3 out of 5 cell lines have shown sensitivity to the treatment, while the remaining two have exhibited resistance. To expand our dataset and improve the robustness of our findings, we are exploring a potential collaboration with Edith Cowan University (ECU) to gain access to additional melanoma cell lines.

In parallel, we are conducting western blot analyses to identify specific biomarkers that may be modulated by the USP9X inhibitor in the sensitive cell lines. This will help us better understand the molecular mechanisms underlying the observed responses.

As we approach Milestone 1 (due at the end of October), we are also preparing for the upcoming candidacy review. Looking ahead, our next steps will involve testing the USP9X inhibitor in combination with a targeted BRAF inhibitor to explore potential therapeutic synergy between the two treatments. If successful, this approach has the potential to transform the treatment landscape of melanoma, leading to improved patient outcomes and a significant enhancement in their quality of life.