Research

Projects

Optical image-guided photomedicine

Far-red and near infrared light-activated therapies offer unique mechanisms for triggering cancer cell death that are agnostic to classical drug-resistance. Light-activated photochemistry can also be exploited for optically active nanomedicine that facilitates spatiotemporal control of drug release to exploit synergies in multi-drug combination therapies. Our group is presently investigating tumor-targeted, cell-activatable approaches for photodynamic therapy (PDT) to selectively kill drug-resistant and heterogeneous cancer cell phenotypes (see the schematic below for tumor-targeted, activatable photoimmunotherapy, taPIT). Concurrently, we are developing new miniature laser scanning microscopy systems and fluorescent probes to recognize drug-resistant phenotypes and to guide multi-targeted PDT.

This research direction is inspired by Prof. Spring's prior work as a postdoctoral fellow in Tayyaba Hasan's group (Wellman Center for Photomedicine, Massachusetts General Hospital and Harvard Medical School) that resulted in a patent-pending technology enabling selective destruction and quantitative molecular imaging of cancer micrometastases (PNAS 2014; featured in a Science Translational Medicine editorial as "ready for translation") as well as near IR-activated nanomedicine for spatiotemporally-controlled, tumor-focused release of molecular inhibitors (Nature Nanotechnology 2016).

Schematic of taPIT where clinical antibody–photodynamic chromophore conjugates are used to selectively light up cancer cells for fluorescence imaging as well as for tumor-confined destruction upon near-IR irradiation. Adapted from Spring et al. PNAS 2014.

Online monitoring of cancer micrometastases using in vivo fluorescence microendoscopy of of cetuximab–benzoporphyrin derivative (Cet–BPD, no light), a clinical antibody–photodynamic chromophore conjugate. The top row of images demonstrates disease progression in a mouse model of micrometastatic cancer in the absence of taPIT. These mice did not receive phototherapy after administration of Cet–BPD. The bottom row demonstrates significant tumor reduction following near-IR taPIT using Cet–BPD. Adapted from Spring et al. PNAS 2014.

Multiphoton microendoscopy

Multiphoton excitation (MPE) fluorescence imaging enables microscopic resolution at depth in tissue (see images below). Nonlinear excitation occurs through a multiphoton process wherein the linear single-photon excitation (1PE) energy (visible light) normally used to make the quantum leap from the ground state to the singlet excited state system is divided up among multiple photons. These near-IR photons must be spatio-temporally synchronized to induce a multistep excitation event through one or more virtual electronic states. The shift to lower energy photons reduces tissue scattering and absorption effects to enable deeper microscopy than possible using linear excitation wavelengths. Multiphoton imaging also provides intrinsic optical sectioning via the femtoliter-sized focal spot, eliminating the need for the traditional confocal pinhole. Read more on multiphoton microscopy

Linear and nonlinear excitation patterns. Image credit: Chris Xu, Cornell University

Two-photon imaging depth using different wavelengths with the single-photon limit overlaid. Adapted from D Kobat, C Schaffer, C Xu et al. (Optics Express 2009).

We are building a versatile in vivo video-rate hyperspectral multiphoton microendoscopy imaging platform that can multiplex many fluorophores in parallel. Therefore, detailed MPE spectra are desirable for engineering laser excitation wavelengths paired with spectral emission detector channel profiles in order to resolve individual components of a cocktail of targeted molecular probes. One of our goals is to validate our MPE spectra instrumentation by comparison to previously measured MPE spectra and cross-sections (e.g., by Chris Xu and Watt Webb, and Muetze et al.). We will then further characterize these spectra into the near-IR (beyond 1,000 nm) where there are significant knowledge gaps in the literature.

Building on recent advances in fiber-optic multiphoton endoscopy, our group's in vivo video microscopy technology development will focus on malignancies of the brain, pancreas and ovaries, and will be applicable to other disease models as well. These technologies are miniaturized for minimally invasive mouse model imaging with potential for clinical translation for use in humans.

Left: Schematic of a video multiphoton fiber-scanning microendoscope published by Guillaume Ducourthial during his graduate studies in Frédéric Louradour’s group (Université de Limoges, France). Adapted from Ducourthial et al. (Nature Scientific Reports 2015). Right: MEMS (microelectromechanical systems) fabrication of a newly designed low-cost probe.

Femtosecond fiber lasers

Because MPE is a relatively weak effect, ultrafast solid state lasers (~$100k) have been used traditionally to generate the peak powers required to generate sufficient fluorescence signal. Inspired by progress in low-cost (~$10k) and compact ultrafast pulse fiber laser technology invented by Frank Wise and colleagues (Cornell University), our group fabricates custom femtosecond fiber lasers.

An all-normal dispersion (ANDI) femtosecond fiber laser with rigorously verified mode-locking built in our laboratory in collaboration with the Wise laboratory (Cornell University).

GPU-accelerated hyperspectral image analysis

For true online optical biopsy and monitoring of disease, multidimensional image analysis must occur at the same rate as data acquisition, and conventional (i.e., pixel-by-pixel) computation techniques aren't adequate to handle such tasks. Instead of sequential analysis, we implement parallel processing on graphical processing units (GPUs), which reduces computation time to video rate speeds. For example, it takes 6 to 8 hours to analyze one minute (5.7 GB) of hyperspectral video in MATLAB. Our algorithm, however, can process the same data at over 150 fps, thereby eliminating this bottleneck from the data analysis pipeline. Techniques like feature tracking, cell counting, micro-mosaicking and quantitative abundance maps are immediately available to the user with these new video-rate processing algorithms.

Low-cost custom LED-arrays for light-activated therapy

Photodynamic therapy needs a light source to activate the photosensitizers. We are developing an automated, low-cost, and easy-to-use illumination platform featuring high-power LED array as the light source. It illuminates multiple wells in a 24- or 96-well plate uniformly for replicated samples.

Fluorescent probes and biosensors

With the emergence of fluorescence-guided cancer surgery, molecular-targeted fluorescent probes are being translated into the clinic to light up cancer cells during tumor resection. We are developing multicolor probes for molecular imaging to enable visualization of heterogeneous human cancers during surgery and microendoscopy-guided photomedicine in collaboration with the Hasan laboratory.

Activatable clinical antibody–fluorophore–quencher conjugates for multicolor cancer imaging. Optimization of the fluorophore to quencher loading ratio yields a 10 to 20-fold fluorescence activation potential depending on the antibody without compromising its biological activity (e.g., epitope binding). Adapted from G Obaid, BQ Spring, T Hasan et al. (J Biomed Opt 2017).

Cancer biology and physics

Resistance to chemotherapy is an ongoing problem for cancer treatment. Ovarian cancer is a prime example of a malignancy with high rates of recurrence and mortality due to drug-resistance. We are presently developing new cell culture and mouse models of chemoresistant ovarian cancer. Towards this goal, we are using novel cell lines and cell culture techniques developed by Cellaria Bio (Cambridge, MA). These models are patient-derived and cultured in proprietary media such that the heterogeneity of the original tumor is preserved during passaging in culture, whereas traditional methods frequently select for particular tumor cell subpopulations with loss of diversity. Chemotherapy dose cycling with characterization of cell-surface biomarkers are ongoing to investigate unique biomarker profile signatures associated with drug-resistance (and specific cell phenotypes such as stem-like and mesenchymal cells). Scientists at Cellaria Bio are providing critical input, primary cell lines and high-throughput assays to compare the new tumor models to original tumor histopathology and genetics.