Analysis of 3D datasets from optically cleared glioblastoma tissues can inform on drug delivery strategies

Research output: Book/ReportPh.D. thesis

Abstract

Glioblastoma (GBM) is the most malignant and lethal type of primary brain tumor. Current aggressive multi-modal treatment consisting of radical resection in combination with radio- and chemotherapy provides median survival time of only around 15 months. Noticeably, compared to a number of other tumor types, GBM has shown an exceptionally high resistance towards multiple novel treatment strategies tested over last decades. Such situation not only calls for utilizing the most recent crossdisciplinary technological advances in order to understand GBM biology and drug targeting better, but also suggests that certain methodological aspects in GBM research may need closer attention and reconsideration.
Among the most promising directions in fighting GBM are improvement in drug delivery through utilization of nanocarriers, and inhibition of immune suppressing activity of the tumor. Success of both strategies is dependent on the delivery of therapeutics across the blood tumor barrier (BTB). Therefore, precise and reliable methods for assessment of therapeutic compound extravasation and distribution in GBM extravascular space (EVS) are of high importance. The first part of the thesis is focused on investigating the methodological limitations implied by the transcardial perfusion deficiency in GBM vasculature, leading to intravascular retention of the injected compound and an exaggeration of the extravasation measurements by a number of widely-used techniques. It is complemented by the study of the artifacts arising from the sectioning of non-perfused tumor vessels limiting the usage of conventional microscopy for the extravasation analysis in GBM. These methodological limitations are further addressed through implementing optical tissue clearing, which circumvents tissue sectioning, allowing to keep the tumor vasculature intact. In order to utilize the full potential of the technique, semiautomatic machine learning-based image analysis workflows were developed allowing for detailed quantitative characterization of compound extravasation, vascular geometry and features of the endothelial binding of therapeutic compound in large 3D datasets.
The second part of the thesis is devoted to the application of the developed methodology for detailed characterization of the changes in GBM targeting with anti-programmed cell death-ligand 1(PD-L1) antibody, GBM angioarchitecture and necrosis formation as a result of radio- and chemotherapy. It is known that GBM microenvironment does not favor extravasation of antibodies due to low and heterogenic permeability of the BTB, as well as high intratumoral pressure hampering the distribution of antibodies in the tumor EVS. However, tumor microenvironment can be remodeled by chemo- and radiotherapy widely used in combination with anti-PD-L1antibody in the current clinical trials. At the same time, the exact effects of such combinations on the GBM targeting by the antibody remain obscure. Developed methodology enabled detection and detailed characterization of an increase in GBM extravascular space targeting by the antibody, as well as changes in its extravasation pattern as a result of radio- and chemotherapy. These changes were accompanied by the remodeling of the vasculature tree through a substantial increase in the volume fraction of large vascular segments. The thesis also reports relations between multiple studied tumor parameters providing an extensive picture of treatment effects on the tumor in context of drug delivery.
As additional methodological developments, the thesis also presents two other tools for the angioarchitecture analysis. The first workflow allows for vasculature analysis in the aged brain addressing the challenge of high lipofuscin autofluorescence levels; the second allows to study endothelial binding of the injected compound by separating the parts of the vasculature bed based on the pattern of their labelling.
We believe that methodological developments and findings reported in the thesis can improve our understanding of the drug delivery against the background of regional tissue pathology, and advance the development of GBM therapy.
Original languageEnglish
PublisherDTU Health Technology
Number of pages208
Publication statusPublished - 2020

Fingerprint Dive into the research topics of 'Analysis of 3D datasets from optically cleared glioblastoma tissues can inform on drug delivery strategies'. Together they form a unique fingerprint.

Cite this