Research

Introduction

“It is very easy to to answer many of these fundamental biological questions; you just look at the thing!”  -Richard Feymann

In biology, seeing is believing. Optical microscopy is one of the most powerful discovery tools in modern biology and medicine. As chemists, we think seeing is not enough: we need measurements. Optical spectroscopy, which is widely used in analytical chemistry and physical chemistry to identify and quantify chemical species, provides the basis of contrasts for optical microscopy. Ironically, to date, it is quite challenging to perform optical spectroscopy measurements in biological tissue due to the lack of well defined optical beam path as a result of light scattering in tissue. The information obtained from optical microscopy are thus either qualitative or semi-quantitative by limiting sample thickness to <10 um. Therefore there is tremendous needs for further development of microscopy techniques to allow quantitative measurements of molecular and functional processes in biology.

Quantitative Nonlinear Optical Microscopy

Nonlinear optical microscopy (more commonly referred to as multiphoton microscopy in biology) partly address the tissue scattering problem by confining excitation of the targeted molecules to a tiny voxel (~0.1um3). The dominant contrast mechanism is fluorescence, but there is a surge of interests in developing optical imaging techniques that are based on absorption and Raman. Many nonlinear optical processes can be used for this purpose based on the pump-probe approach (Figure 1). We aim to develop spectroscopic imaging tools that can precisely measure the concentration of molecular species at high spatial and temporal resolution.

Quantitative Nonlinear Optical Microscopy

Figure 1: Various nonlinear optical processes can be used as optical contrasts in microscopic imaging with the pump-probe approach.

Hyperspectral stimulated Raman scattering  microscopy

 

Stimulated Raman scattering (SRS)  microscopy is a powerful technique for imaging the spatial distribution of a targeted molecular species that has a well-defined Raman peak which is separate from all other species. However, this is rarely the case in a complex biology system. To better quantify species with overlapping Raman spectra, it is imperative to obtain spectroscopic information. We achieve this by linearly sweeping the probed Raman frequency and synchronizing it to laser scanning. We used an approach called spectral focusing in which the Raman frequency is sequentially tuned by changing the time delay between two linearly chirped femtosecond lasers (see Figure 2). Combining spectroscopic imaging with chemometric methods, we demonstrated that closely related molecular species can be separately quantified. Our goal is to further improve the spectral resolution and spectral coverage to allow better characterization of biological systems.

Raman scattering microscopy

Figure 2: left: spectral focusing tunes the Raman frequency by linearly varying the time delay between the pump pulse and the Stokes pulse; middle: hyperspectral SRS imaging of fixed HeLa cells; right: titration curve of mixture of cholesterol and oleic acid in CDCl3 measured by hyperspectral SRS imaging.

Imaging metabolic turnover of lipids

 

Lipids are important building blocks of cell membranes. They are also essential in maintaining the energy balance of cells by serving as an energy reserve. Thus their turnover (synthesis and degradation) strongly depends on the metabolic needs of cells, allowing us to use lipid imaging to monitor the metabolic states of cells under both normal and disease conditions. We have already demonstrated that different lipid species can be distinguished by hyperspectral stimulated Raman scattering microscopy (Figure 3). We can also trace the metabolic fate of a specific type of lipids using deuterium tracing, a common technique used in Mass spectrometry. We will apply these techniques to understand what happens when cells are under nutritional overload.

Lipids

Figure 3: a,b: hyperspectral SRS imaging of cultured macrophage cells and hepatic cells. c, d: average SRS spectra of lipid droplets in macrophage cell and hepatic cell, respectively. Spectra of pure chemical species cholesterol oleate and trioleate are shown for comparison. e. SRS images of C.elegans fed with deuterated oleic acid in the C-D and C-H stretching region and their ratio. f. Comparing uptake rate of palmitic acid and oleic acid in worms. g. SRS spectra of lipids stored in the droplets are almost identical to their precursors.

Imaging metabolic dynamics of small molecule drugs

 

Small molecule drugs typically do not fluorescence. It is challenging to track their dynamics without resorting to fluorescent tagging, which itself is problematic because it will significantly affect the behavior of small drug molecules. Fortunately drug molecules often have strong and characteristic Raman vibrations that we can capitalize on. With hyperspectral stimulated Raman scattering microscopy, we have the potential to quantify the concentration of drug molecules in subcellular organelles or compartments and follow their intracellular dynamics. One recent example is the direct quantification of lysosomal trapping of tyrosine kinase inhibitor drugs (Figure 4). Importantly, we found that through pH mediated drug-drug interaction, we can improve the drug efficacy of Gleevec, a billion-dollar drug for treating chronic myeloid leukemia.

Figure 4: a. Cartoon of a cell. b. Raman spectra of two chronic myeloid leukemia drugs Gleevec (Imatinib) and Tasigna (Nilotinib) and their chemical structures. c. Simultaneous SRS imaging of Gleevec and two-photon fluorescence imaging of lysotracker revealed that drugs were accumulated in lysosomes. d,e. Time dependent accumulation of Imatinib and Nilotinib. f. Proposed lysosomal trapping model for weakly basic drugs. g. pH-mediated Imatinib-chloroquine interaction removes imatinib from lysosomes as chloroquine concentration increases.

Imaging cell growth and cell dynamics

Tissue is not a homogeneous bag of cells. In particular, cell heterogeneity plays a critical role in cancer development. Thus it is important to understand this heterogeneity at the cellular level. We are particularly interested in understanding heterogeneous cell growth and cell death in tumor tissue, and how they impact the tissue microenvironment. To do this, first we need to build a multimodal optical imaging platform that is capable of quantitatively measuring a number of molecular and functional parameters. We will then integrate information obtained with different modalities and dissect the interrelationship of those variables.

Early cancer diagnosis

 

Cancer at the early stage, when treatment has the highest chance of a complete cure, often eludes the detection with traditional X-ray, ultrasound or MRI imaging. Microscopic examination of H&E stained tissue biopsies has long been the gold standard for cancer diagnosis. However, the undesirability of tissue removal, the long wait time, and a high chance of missing disease tissue necessitate the development of optical “virtual biopsy” technology. We aim to develop a sensitive and accurate cancer diagnosis tool based on nonlinear optical microscopy.