Cheng has built an interdisciplinary research program that develops new imaging tools and nanotechnology to advance basic understanding, diagnosis, and treatment of diseases highly related to human health. A highlight of our nonlinear optical imaging research can be found on the following website: http://www.nibib.nih.gov/HealthEdu/eAdvances/28Dec07. The two major research directions are summarized below.

1. In Vivo Nonlinear Optical (NLO) Imaging to Advance Mechanistic Understanding and Early Diagnosis of Diseases.

Since the invention of first optical microscope in 17^th century, biological science has been largely advanced via a reductionist approach. However, knowledge from individual cells cultured in a dish may not represent their behavior in vivo. Understanding the interactions between cells, extracellular matrix, and stromal molecules in a tissue environment is an emerging frontier of biology. Tissues are usually studied by histology and electron microscopy, in which fixation and staining inhibit real-time monitoring of cellular processes.  NLO microscopy opened up a new window for imaging live tissues with inherent 3D submicron resolution, relatively large penetration depth, and real-time data acquisition. Major NLO imaging modalities include two-photon excited fluorescence (TPEF), second harmonic generation (SHG) or sum-frequency generation (SFG), and coherent anti-Stokes Raman scattering (CARS). Our lab pioneers the development and biomedical applications of CARS microscopy that permits high-speed vibrational imaging (for reviews, see). In a CARS process, the interaction of a pump field E_p(ω_p) and a Stokes field E_s(ω_s) with a sample generates an anti-Stokes field E_as at frequency 2ω_p-ω_s. A unique advantage of CARS is that TPEF and SFG/SHG modalities can be naturally incorporated into the same platform, enabling real-time visualization of various components in a tissue. The principle of multimodal NLO imaging is illustrated in Fig. 1.

Figure 1. (A) Energy diagram of CARS, SFG, and TPEF. (B) Emission spectra of SFG from collagen fibrils (blue), TPEF from nuclear label hoeschst (green), and CARS from intracellular lipid droplets (red). (C) Multimodal NLO imaging of spinal cord white matter with CARS signal from myelin sheath (red), SFG from astrocyte process (green), and TPEF signal from fluorescently labeled nuclei (magenta).

A major focus of our current research is to understand the role of lipids in health and diseases. Lipids serve as structural components of membranes, energy storage, and hormones. Our recent study of single endosome trafficking revealed important roles of cholesterol membrane trafficking in cells. At in vivo level, lipids constitute the major components of white matter in the central nervous system. With obesity and metabolic syndrome reaching an epidemic proportion, the role of lipids in public health is further emphasized. Despite the significance, visualization of lipids has been a challenging topic because most lipid molecules have no specific markers. CARS microscopy provides an ideal tool to overcome this difficulty. The coherent addition of the CARS fields from the high-density C-H vibration in lipid membranes or lipid bodies contributes to a large signal that can be used to monitor lipid organization, metabolism, and storage without the need for labeling. Our lab has pushed the detection sensitivity of laser-scanning CARS microscopy to 20,000 lipid molecules in a single bilayer inside a focal area of 0.25 mm². We have utilized such capability to quantify the partition of deuterated phosphorlipid in co-existing domains and to image domains based on lipid packing density. With multimodal NLO microscopy, our lab is dedicated to understanding the initiation and progression of multiple sclerosis and breast cancer, as summarized below.

Imaging molecular mechanisms of demyelination

Myelin sheath is an extended plasma membrane wrapping around an axon and is crucial for high-speed impulse conduction. Demyelination, the loss of normal myelin sheath, accounts for long-term neurologic disability. Our lab has demonstrated CARS imaging of intact myelin sheath in ex vivo spinal tissues and in sciatic nerve of live animals. The myelin membranes contain about 70% lipid by weight, and the high-density CH₂ groups produce a large CARS signal. With 3D submicron resolution, CARS microscopy is able to image detailed myelin structure around a Node of Ranvier at a speed of a frame per second. The capability of probing intact myelin in the live tissue environment allows examination of enzymatic and cellular activities involved in demyelination.

We have used the resonant CARS signal from symmetric CH₂ stretch vibration (2840 cm^-1) to characterize lysolecithin-induced myelin degradation. Lysolecithin is most widely used in experimental demyelination. The breakdown of myelin membrane was characterized by decrease of CARS intensity and loss of dependence excitation polarization (Fig. 2). By combining CARS imaging with compound action potential recording, we have identified a Ca²⁺ dependent pathway which involves the activation of cPLA₂ and calpain for lysolecithin-induced myelin degradation. With the demonstrated capabilities, we have further evaluated the effect of glutamate excitotoxicity on myelin. Elevated glutamate level was found in cerebral spinal fluid of multiple sclerosis patients.  Our study shows that exposure of white matter to glutamate leads to paranodal myelin splitting and retraction. This observation shows that paranodal myelin degradation is a possible onset of demyelination in multiple sclerosis.

Towards the goal of understanding initiation and progression of multiple sclerosis, we have started CARS imaging of experimental autoimmune encephalomyelitis (EAE) lesions in SJL/J mice via collaboration with Professor Stephen Miller in Northwest University Medical School. EAE is the most frequently studied animal model of multiple sclerosis. In this model, rodents that are immunized with myelin-derived peptides develop a paralytic disease with similarities to the human condition. Our ex vivo CARS images of spinal cord white matter extracted from EAE mice are able to show lesions as small as 10 mM. Such capability is unparalleled by other existing imaging technologies and is being used to understand the correlation between demyelination and the clinical relapsing-remitting pattern of multiple sclerosis. Moreover, our group is developing in vivo CARS imaging of spinal cord white matter on an upright microscope with the aid of a  microprobe lens to minimize trauma damage. This technique enables us to follow the progression of white matter injury or disease in the same animal, so that causalities of different events including T cell infiltration, microglial cell activation, and demyelination can be clarified.

At the in vitro level, our group is developing a demyelination and remyelination model using cultured brain slices. We use lysolecithin to inflict myelin damage. Progression of demyelination and remyelination are monitored by CARS microscopy over one week period, and functional recording is carried out in Riyi Shi’s lab in School of Veterinary Medicine. The model will be used to screen the putative role of 4-aminopyromide (a potassium channel inhibitor used for treatment of multiple sclerosis) in promotion of remyelination and the role of tumor necrosis factor a (an inflammatory agent) in inhibition of remyelination.

The above myelin study is supported by NIBIB R01 grant 007243 from 2007 to 2010 (Cheng PI).

Delineating the role of dietary fat and obesity in cancer development

Dietary fat and obesity are major epi-genetic factors for tumorigenesis and tumor progression. It is generally accepted that diet and obesity account for 30% of preventable causes of cancer. For breast cancer, epidemiology showed that obesity increases the risk of developing breast cancer in postmenopausal women by 50%. Nonetheless, the relationship between diet and cancer incidence remains controversial due to the conflicting experimental results from clinical studies. Currently, a consensus is lacking on the benefits of a certain type of fatty acid or nutritional ingredient to cancer prevention. A contributing factor to the controversy is the lack of mechanistic understanding on how dietary fat or adiposity affects cancer development.

It is established that adipose tissue constitutes an active endocrine and metabolic organ that has far-reaching effects on the physiology of other tissues. Lipolysis of adipose tissue provides circulating free fatty acids (FFA) as a fuel for bodily needs. In upper-body obesity, excess lipolysis is commonly seen, resulting in higher levels of plasma FFA concentration. In addition, adipose tissue is known to secret a variety of bioactive peptides known as adipokines. We hypothesize that high-fat diet and obesity affect the function of epithelial cell and tumor cell membrane, modulate tumor stromal composition, and induce chemotaxis to tumor cells via increased level of circulating free fatty acid and adipokines, consequently promoting tumor initiation and metastasis.

With the long term goal of understanding the role of stroma in tumor development, we have started using NLO imaging to evaluate the impacts of obesity on mammary gland and mammary tumor stroma. Supported by a NCI R03 grant from 2007 to 09, our lab has combined multimodal NLO microscopy with a diet-induced-obesity rat model provided by Dr. Ignacio Camarillo at Purdue University. An advantage of our rat mammary tumor model over a mouse model lies in its progression from hormone-dependent to independent state, an event very common in human breast cancer. By imaging significant components of mammary tumor stroma including adipocytes, collagen fibrils, blood vessels, and macrophages, we showed that obesity significantly increased the density of collagen fibrils in mammary tumor stroma as well as the percentage of carcinoma in animals. This study provided the first visual evidence that supports the relation between obesity and breast cancer risk. By using intravital flow cytometry to enumerate circulating tumor cells in BALB/c mice bearing GFP- and RFP-labeled metastatic M109 cells, our latest work has shown that high-fat diet expedites tumor intravasation (Fig. 3A) without increasing body weight (Fig. 3B) or tumor growth rate (Fig. 3C). Our detailed co-culture study indicates that excess NEFA (Fig. 3D) is a culprit for tumor metastasis. These data demonstrate that NLO imaging and intravital flow cytometry can be effectively used to evaluate the impact of obesity on cancer formation and progression.

     Because it is highly likely that multiple molecules and multiple pathways are simultaneously involved in modulating various aspects of cancer development, we are taking a top-down (from in vivo to in vitro) approach to explore the molecular mechanisms linking high-fat diet / obesity and cancer risk. We first combine multimodal multiphoton microscopy with a diet-induced-obesity breast cancer rat model to determine the impact of high-fat diet and obesity on mammary duct development and tumor stromal composition (e.g. collagen fibrils, macrophages) at 3D sub-micron spatial resolution. We then combine intravital fluorescence flow cytometry with a C57BL/6J mouse model of tumor metastasis to quantify the influence of high-fat diet and obesity on tumor extravasation into bloodstreams. Moreover, we develop intratival imaging with a dorsal skin fold chamber to evaluate the impact of obesity on tumor stroma development. Based on the in vivo observations, we pursue a deeper understanding by investigating the direct impact of adipose tissue on tumor cell membrane property, tumor motility, and gene activities using a tumor cell – adipose tissue co-culture system. A panel of dietary fatty acids and adipokines released from the adipose tissue will then be separately tested to identify the most effective components that affect tumor progression. 

    Our group has also applied CARS microscopy to monitor physiological and pathological processes in arteries. In collaboration with Michael Sturek in IU School of Medicine who developed the Ossabaw swine model of atherosclerosis, we have demonstrated multimodality NLO imaging of arterial walls and atherosclerotic lesions. Based on vibrational signals arising from CH₂-rich membrane, endothelial cells and smooth muscle cells of the arterial walls can be visualized with CARS. In particular, foam cells rich in lipid droplets and lipid bodies in the extracellular matrix of an atheroma are readily observable with CARS microscopy. SHG imaging of collagen fibrils and TPEF imaging of elastin are employed simultaneously. These capabilities are being utilized to investigate the role of adventitia inflammation in atherosclerosis by ex vivo and in vivo multimodal imaging of foam cells, collagen matrix, and vasa vasorum in ApoE deficient mice.

Intravital Flow Detection of Circulating Tumor Cells (CTCs) to Advance Cancer Diagnosis.

CTCs represent a potential alternative to invasive biopsies of tumor tissue for detection, characterization and monitoring of non-haematologic cancers. Although various in vitro assays have been developed, these methods are not sufficient for detection of rare CTCs in early stage cancer due to the limited sample volume. However, in vivo detection of single CTCs in the whole blood is far beyond the sensitivity of current medical imaging technologies such as MRI.

Our lab has recently demonstrated the proof-of-concept of intravital flow cytometry that noninvasively counts rare CTCs in vivo as they flow through the peripheral vasculature. The method involved i.v. injection of a tumor-specific fluorescent ligand followed by multiphoton fluorescence imaging of superficial blood vessels to quantitate the flowing CTCs on a laser-scanning microscope (see Fig. 4A). Folate dye conjugates were chosen because many human carcinomas over-express the folate receptor and because the folate conjugate can be quickly removed from circulation if uncaptured by CTCs (Fig. 4B). Our intravital flow cytometry studies in mice with metastatic tumors demonstrate that CTCs can be quantitated weeks before metastatic disease is detected by other means such as histology (Fig. 4C).

Via collaboration with Philip Low’s group (providing tumor-targeting ligands) in Department of Chemistry, Debbie Knapp’s group (providing dogs with bladder cancers) in School of Veterinary Medicine, and Dr. Wael Harb in Horizon Oncology Center, we are developing a catheter-based fiber-optic probe for intravascular detection of CTCs in large animals and eventually in human patients. Meanwhile, our lab has demonstrated microfluidic CARS flow cytometry for in vitro quantitation of lipid-rich objects (i.e., adipocytes). Development of in vivo CARS flow cytometry is under progress in our lab with the goal of achieving label-free detection of CTCs based on the CARS signal from intracellular lipid bodies.   

2. Nanophotonics and Nanomedicine.

Nanostructures at the scale of 1 to 100 nm often exhibit unique features that bulk materials do not possess. We identify those unique properties applicable to detection and treatment of diseases and injuries. Moreover, we employ intravital optical imaging with 3D submicron resolution to monitor stability and pharmacokinetics of nanoscale drug carriers during circulation and to probe the interaction of these carriers with immune cells, endothelial cells, and circulating tumor cells. Two projects along this direction are summarized below.

Bioconjugated gold nanorods as a NLO imaging and photothermolysis agent.

Gold nanorods produce strong optical responses as a result of plasmon resonance. Being different from nanoparticles, gold nanorods have a transverse and a longitudinal mode, with resonance wavelengths in the visible and near IR regions, respectively. Our lab demonstrated a bright two-photon luminescence (TPL) from CTAB-coated nanorods prepared by a seed growth method. The TPL was found to arise from enhanced two-photon absorption by the longitudinal plasmon resonance. Compared to quantum dots, gold nanorods are less susceptible to photobleaching and are chemically inert under physiological conditions. These properties make gold nanorods an appealing imaging agent for in vivo applications. We are investigating second harmonic generation, third harmonic generation, and coherent anti-Stokes scattering signals from nanorods and their applications to bioimaging.

A unique advantage of gold nanorods is the ease of surface modification. Bio-conjugation allowed us to control the cellular uptake efficiency of gold nanorods and to target tumor cells over-expressing folate receptors. Recently we have shown that nanorods conjugated with arginine-rich peptides can be effectively internalized by activated macrophages in live animals. By exchanging the CTAB on the nanorod surface with polyethylene glycol (PEG) of various molecular weights, we have achieved a circulation half-life time of 5.0 h. Delivery of homing peptide conjugated and PEG-coated nanorods to macrophages in atherosclerotic plaques in ApoE-/- mice is under investigation.

Importantly, gold nanorods are also efficient converters of light energy into heat. We have applied folate-conjugated nanorods to optical hyperthermia of malignant KB cells. Moreover, we have revealed that gold nanorods mediate acute membrane blebbing and subsequent cell death through compromising the membrane integrity and influx of extracellular Ca²⁺. These studies pave the way to in vivo applications of nanorods for photothermolysis of cancer cells and macrophages in inflammatory diseases.

Polymeric micelles for repair of both primary and second damage in spinal cord injury

Polymeric micelles are spherical assemblies of di-block copolymers, containing a hydrophilic PEG shell and a hydrophobic inner core. These micelles have unique properties such as biocompatibility and long blood circulation, and have been widely used as carriers of water insoluble drugs. By recognizing the high-density coronal PEG in a polymeric micelle, we realized the potential of using micelles to fuse membrane breaches created in spinal cord injury. As a proof of principle we have prepared and applied monomethoxy PEG-poly(D,L-lactide acid) (mPEG-PDLLA) micelles to compression-injured ventral white matter isolated from adult guinea pigs. With this approach we could achieve a rapid, significant restoration of impulse conduction and a simultaneous reduction of Ca²⁺ entry into axons in injured spinal cords. With a seed grant from Showalter Foundation, the in vivo therapeutic potential of polymeric micelles is under evaluation using a rat model. 

To evaluate the potential of using the micelles to deliver anti-inflammatory drug to suppress the secondary injury, we have revisited the conventional wisdom that polymeric micelles systemically carry drug molecules until they are taken up into cells followed by intracellular release. We employed Forster resonant energy transfer (FRET) imaging to monitor in real time the release of hydrophobic molecules from circulating poly(ethylene glycol)-poly(_D,L-Lactic acid) micelles. A lipophilic FRET pair, DiIC18 and DiOC18, was physically entrapped into the micelle core by mimicking the loading of hydrophobic drugs. By monitoring the hetero- and homo-FRET efficiency, release of core-loaded probes to model membrane and cell membrane was demonstrated, indicating a membrane mediated pathway for cellular uptake of hydrophobic molecules pre-loaded in polymeric micelles. Furthermore, intravital imaging of blood stream showed that FRET efficiency significantly reduced within 15 min after intravenous injection of FRET micelles into BALB/c mice, implying quick escape of the probes from circulating micelles. These studies have prompted us to make shell-crosslinked micelles to ensure systemic delivery of loaded drug to the injured site.