Microscopy is an essential tool for clinical diagnosis and for understanding the pathogenesis of cancer. By examining alterations in nuclei architectures, a hallmark of cancer cells, pathologists identify and assess tumor progression. Inside the nucleus, genomic DNA folds into compact chromatin structures that regulate DNA transcription and genomic stability, defining cancer development. However, how chromatin structures change over the tumorigenesis process is beyond the detection limits of conventional microscopes.
Yang Liu, an optical bioengineer at the University of Pittsburgh, exploits the power of light and develops advanced imaging technologies and quantitative analysis tools to capture clearer pictures of subcellular structures. Liu and her team specialize in stochastic optical reconstruction microscopy (STORM), a fluorescence imaging technique that visualizes nanoscale molecular organization. Aiming to understand epigenetic regulations in cancer cells, they use STORM to look at chromatin organization directly in patient-derived tissues across different stages of cancer. By improving the imaging capability of STORM, the team has resolved previously invisible chromatin structures down to a resolution of 20–30 nm, gaining valuable information on malignant transformation and opening new opportunities for detecting cancer at earlier stages.
What interested you in using microscopy to study cancer epigenetics?
Pathologists rely on conventional light microscopes to examine distinct morphological features of cell nuclei, such as sizes and shapes, to identify tumors. However, a clinical dilemma is that precancerous lesions that are morphologically similar may have different levels of risk for developing malignant cancers. It is also challenging to detect cancers at earlier stages when cells still appear normal.
During my PhD training at Northwestern University, I worked on developing biomedical imaging technologies. I wanted to understand epigenetic changes in cancer at the molecular level using advanced microscopes. So, I looked into chromatin organization to identify features that can differentiate high risk from low risk precancerous lesions and allow the early detection of cancer. Although my training was not in chromatin biology, I was motivated to apply cutting-edge imaging technologies to visualize epigenetic changes and eventually improve risk stratification for cancer patients.
What is STORM?
STORM is a type of single molecule localization microscopy; it is a 2014 Nobel Prize-winning technique. The resolution of conventional microscopy is limited by diffraction, which means that it is hard to distinguish individual molecules because their fluorescently-labeled images overlap. Xiaowei Zhuang at Harvard University and Eric Betzig at the University of California, Berkeley developed the approach of activating only a small subset of fluorescent molecules while turning off the rest. Imagine them like stars blinking. At any given frame, we detect a random set of blinking molecules. By repeating the process, we can accumulate tens of thousands of frames to reconstruct the full image in super resolution. STORM combines a traditional wide field fluorescence microscope, imaging buffers that trigger the blinking events, and an image acquisition and localization system. Among various types of super resolution microscopy techniques such as structured illumination microscopy and stimulated emission depletion microscopy, STORM is one of the simplest and most cost effective approaches, so it has great potential in clinical settings.
What is the biggest challenge of studying chromatin structure with STORM?
Another challenge comes with the slow data acquisition speed. When STORM was first developed in 2015, scientists mostly used it to study monolayer cell cultures or bacteria. But for paraffin-embedded tissues, we need to analyze hundreds of cells to achieve statistical significance, which means increasing the imaging throughput.
To overcome these challenges, we developed an algorithm that removes the background from the original image. This significantly improves the image quality and resolution, reduces artifacts, and increases reproducibility. We are also establishing a high throughput and high resolution system to enable high content phenotyping and study heterogeneous biological systems, such as the tumor microenvironment.
What chromatin structural changes have you observed during cancer progression?
We used mouse and human tissues to study the entire carcinogenesis process and found chromatin to be open and fragmented throughout cancer progression. We observed an increased degree of disruption in chromatin compaction at different stages, including during early carcinogenesis when cells still appear normal, as precursor lesions develop that grow into small tumors, and in malignant invasive cancers. To our surprise, although the opening of chromatin structures continues throughout the carcinogenesis process, the biggest changes occur very early, even before the tumor forms. By looking at various tumor types, such as colon, prostate, and lung, we also realized that this pattern of chromatin changes is universal across a wide range of cancers.
What do these changes tell us about cancer?
For many years, people believed that chromatin was more condensed in cancer cells because cancer cells undergo rapid replication and division. But by looking at the expression of histone proteins and heterochromatin marks, scientists in epigenetics research have found evidence suggesting open chromatin structures in cancer cells. In our study, we indeed observed open and disrupted chromatin structures in tumor tissues.
We also learned from many basic biological studies that compacted chromatin protects the cell from environmental insults. Once the protection is broken, the cell transitions from a normal phenotype to a state prone to carcinogenesis more easily. This is in accordance with our observation that the most remarkable chromatin fragmentation occurs at an early stage of tumorigenesis. Chromatin decompaction is an early event that leads to active transcription and facilitates malignant transformation.
How can super resolution imaging of epigenomic signatures help scientists detect cancer earlier?
The current clinical practice for colon cancer diagnosis relies on routine colonoscopy screening that identifies precursor lesions and colon polyps for a large number of patients. Patients need to do the screening annually, but only a small percentage of them eventually develop colon cancer, which makes this process inefficient.
We hope to analyze the chromatin structures of existing clinical samples with super resolution imaging to identify those that might progress into malignant tumors. We are working with a large database of well documented patient samples to correlate their chromatin features with cancer development. In this way, we will be able to stratify the cancer risks of patients and monitor patients with high risks more closely while examining patients with lower risks less frequently. This approach has the potential to improve clinical practice by providing more guided surveillance.
We are applying this technique to lung cancer. We collaborate with researchers at the Hillman Cancer Center who collected many sputa samples from patients. Hopefully, by analyzing cells from these samples and looking at their chromatin signatures, we can detect lung cancer at an early stage.
What has been the most rewarding part of your research?
I am passionate about solving problems and excited about new discoveries and technologies that allow us to do what we have not been able to do in the past. Doing research is a long journey with lots of failures and frustrations, but the discoveries made in the end are worth all the effort. I encourage my students to focus less on the final outcome, and to look at the process and solve one problem at a time. We’re driven by the curiosity of exploring new things, and accomplishing high-impact work that can eventually help patients is truly rewarding.
This interview has been condensed and edited for clarity.