2023 CYP PhD project Clark2023-02-23T13:02:42+00:00

MRI-based tracking of senescent cell population dynamics in brain tumours to monitor and optimise senolytics-based therapies

Primary supervisor: Chris Clark, UCL

Secondary supervisor: Faraz Mardakheh, Queen Mary University of London, Jamie Dean, UCL

Project

Background

Cellular senescence is a key biological process regulating the growth and treatment response of many brain tumours. Some brain tumours, such as pilocytic astrocytoma, exhibit high levels of cellular senescence before treatment. Others, including diffuse midline glioma, initially have smaller numbers of senescent cells, but widespread senescence is induced in response to standard-of-care therapy. As senescence has been shown to cause treatment resistance and relapse, there is great interest in pharmacologically targeting these cells with ‘senolytic’ drugs. Preclinical research has shown that the timing of administration of senolytics drastically affects their efficacy [1]. Monitoring senescent cell dynamics in patients would enable the timing of administration of senolytics to be personalised to maximise their efficacy. However, longitudinally monitoring senescent cell population dynamics in brain tumour patients is not currently possible. Therefore, the development of a tool capable of non-invasively tracking the fraction of senescent cells in tumours would represent a substantial advance. The supervisory team has recently performed a successful proof-of-concept demonstrating that the relative size of different cell populations within gliomas can be inferred from routinely acquired MRI scans using machine learning (unpublished). This project builds upon that work to develop a machine learning-based computational tool to infer the fraction of senescent cells within brain tumours from MRI.

Hypothesis

We hypothesise that the fraction of senescent cells within tumours can be inferred from MRI using machine learning models developed from matched histological, molecular and radiological data.

Aim 1: Characterise senescence in brain tumours: The student will perform a detailed characterisation of senescence in brain tumour samples using histology and molecular data from large publicly available datasets, including the Pediatric Brain Tumor Atlas [2], and Great Ormond Street Hospital. Senescent cells exhibit a distinctive morphology which allow them to be accurately identified on histological imaging using deep learning [3]. RNA-sequencing and DNA-methylation signatures of senescent cells have been developed. Combining these signatures with matrix factorisation enable the fraction of senescent cells to be inferred from bulk sequencing measurements. Multi-region data will be used to determine whether cellular senescence is associated with different environments within tumours.

Aim 2: Develop machine learning models to senescent cell fraction from MRI data: The student will train and validate machine learning models to predict the fraction of senescent cells using matched histology, RNA-sequencing, DNA methylation and MRI data from large publicly available datasets, including the Pediatric Brain Tumor Atlas [2], and Great Ormond Street Hospital. The histology and molecular data will be used to define the fraction of senescent cells in a sample. Machine learning models will be trained to predict the senescent cell fraction from quantitative features extracted from MRI measurements informed by Aim 1.

Aim 3: Design novel MRI sequences to enhance the accurate detection of senescent cells: The student will develop a diffusion MRI sequence optimised to enhance the detection of senescent cells. The distinctive morphology of senescent cells creates differences in water diffusion compared with other cells. The shape and cellular dimensions of senescent cells, characterised in Aim 1, will be integrated into mathematical models of the diffusion signal [4] thereby allowing the detection of specific cellular characteristics. The ability of these new sequences to detect senescent cells will be preclinically and clinically evaluated in the future.

Candidate background

This project would suit a candidate with a background in engineering and physical sciences, experience of programming and computational biology is desirable.

Potential Research Placements

  1. Juan-Pedro Martinez-Barbera, UCL Institute of Child Health
  2. Kish Mankad, Great Ormond Street Hosptital
  3. Faraz Mardakheh, Barts Cancer Institute/ Queen Mary University of London

References

  1. Rahman, M., et al. (2022) Selective vulnerability of senescent glioblastoma cells to BCL-xL inhibition. Mol Cancer Res 20:938-948.
  2. Shapiro, J.A., et al. (2022). OpenPBTA: An open pediatric brain tumor atlas. bioRxiv [link]
  3. Heckenbach, I., et al. (2022). Nuclear morphology is a deep learning biomarker of cellular senescence. Nature Aging 2:742-755.
  4. Gyori, N.G., Clark, C.A., Alexander, D.C., Kaden, E. (2021). On the potential for mapping apparent neural soma density via a clinically viable diffusion MRI protocol. Neuroimage 239:118303.
available PhD projects
apply now