An Interdisciplinary Collaboration between Engineering and Medicine.

Mark Borden, PhD, Asst. Professor of Chemical Engineering Jessica Kandel, MD, Assoc. Professor of Surgery Darrell Yamashiro, MD/PhD, Assoc. Professor of Pediatrics

An ultrasound molecluar imaging scan showing the extent of αVβ3 integrin (a marker for angiogenesis) exrpession in the tumor microenvironment. Left: the image shows microbubble signals in green. Right: the plot shows a signal intensity change after a destruction pulse to deliniate signal from targeted vs. free microbubbles.

Mark Borden, PhD

Dr. Borden is an Assistant Professor of Chemical Engineering at Columbia University. His research entails the design and engineering of novel microbubble constructs for ultrasound molecular imaging and targeted drug delivery. Dr. Borden is leading the implementation of microbubble technology to assess tumor vasculature via ultrasound.

Jessica Kandel, MD

Dr. Kandel is an Associate Professor of Surgery. Her research specializes in the regulation of angiogenesis in pediatric solid tumors, including Wilms Tumor and neuro-blastoma. Dr. Kandel is leading the development of animal models and ex vivo techniques and translation of results to clinical practice.

Shashank Sirsi

Darrell Yamashiro, MD/PhD

Dr. Yamashiro is an Associate Professor of Pediatrics and Pathology and Cell Biology (in Surgery). He is the co-director (with Dr. Kandel) of the Pediatric Tumor Biology Laboratory, which focuses on the role of angiogenesis in promoting the growth and metastasis of pediatric solid tumors.

Shashank Sirsi, PhD

Dr. Sirsi is a Postdoctoral Research Scientist working with Dr. Borden. He is the key liaison between the laboratories and is responsible for the execution of microbubble development and ultrasound imaging and therapy experiments.

Combined Molecular/Ultrasound Analysis of Tumor Response to VEGF

Funding: NIH/NCI R21 CA 139173

PI: Borden and Kandel

Our ultimate goal is to develop an innovative ultrasound technique to monitor and guide anti-angiogenic therapy for children with clinically aggressive Wilms tumors (WT). We are developing this technology, which includes molecularly targeted microbubble contrast agents and sophisticated imaging methods, while exploring tumor vascular changes during initial and chronic blockade of vascular endothelial growth factor (VEGF). We hypothesize that molecular changes in WT angiogenesis can be correlated with vascular changes revealed by noninvasive ultrasound.

Children with unfavorable histology Wilms tumor (WT) and metastatic disease continue to experience high mortality rates. These patients urgently require new therapies. Drs. Kandel and Yamashiro have recently reported Phase I data indicating excellent tolerance of the anti-vascular endothelial growth factor (VEGF) antibody bevacizumab (BV) in refractory pediatric tumors. Because this therapy has been validated in adult cancers, it may provide an attractive option for patients with aggressive WT; however, methods of assessing tumor response clinically are lacking. This is a particularly critical issue for pediatric cancer patients, in whom long-term tumor control is the goal. In previous studies, Drs. Kandel and Yamashiro reported that experimental WT were initially strikingly suppressed by VEGF inhibitors. Yet consistent with clinical observations that adults treated with BV virtually all progress, they found that even highly responsive xenografts resumed growth if treatment was sustained. The mechanism of resistance to VEGF blockade is poorly understood, and clinical endpoints of resistance remain undefined. Emerging data suggest that tumors subjected to VEGF inhibition exhibit features of ischemic injury, including induction of damage response pathways and vessel remodeling. Further, distinct changes in gene expression, vascular assembly, and perfusion occur both acutely and chronically. For example, Drs. Kandel and Yamashiro have previously reported that VEGF inhibition can cause striking loss of branching vasculature and ischemia by 24 hours, whereas long-term blockade results in vessel remodeling, recovery of flow, and tumor progression. Key molecular markers of the response to vessel injury include members of gene families that are essential to angiogenesis, including integrins (alphaVbeta3), VEGF receptors (VEGFR-1 and -2), and Notch family members (Jagged-1), and mediators of the response to hypoxia (such as COX-2).

High frequency ultrasound is an emerging technology that can provide rapid and longitudinal assessment of the anatomic, functional, and physiological response of WT vasculature to VEGF inhibitors. The excellent sensitivity of newly available commercial scanners to sonographic contrast agents (microbubbles) echo-signatures facilitates visualization of vessel architecture, quantification of blood flow, and molecular imaging of endothelial biomarkers in solid tumors. Yet this technology is still in its infancy, and further development of long-circulating and targeted microbubbles is critical for realizing its full potential as a means of evaluating dynamic changes in vessel structure and function. In particular, it is essential to develop a platform suitable for clinical point-of-care use. In these studies, we will investigate vascular remodeling during VEGF blockade using high frequency ultrasound, in the specific context of experimental WT, and using novel microbubble tools and ultrasound imaging techniques. Our goal in these studies is to relate acute and chronic molecular changes in WT angiogenesis with highly quantitative and sensitive architectural and flow characteristics and vascular biomarker expression patterns revealed by ultrasound.

Xenogen image of mouse tumor transfected with a bioluminescent reporter gene.

Novel microbubble-based gene delivery vehicles targeting solid tumors

Funding: St. Baldrick’s Foundation

PI: Yamashiro

Novel microbubble-based gene delivery vehicles will be engineered to provide an efficient and selective means of transfecting target tumors with DNA. The research will not only result in advanced gene delivery technology, but will also better characterize the underlying mechanisms of ultrasound-microbubble gene delivery. Together, these studies may move this concept and technology rapidly toward clinical treatment of children with cancer.

Previous workers have used microbubbles and ultrasound to deliver genetic material, in an effort to bypass the need for viral vectors. By exploiting the ability of ultrasound to precisely control microbubble destruction, researchers have demonstrated highly specific tissue targeting of proteins and plasmids to the heart (Shohet et al., 2000; Mayer and Bekeredjian, 2008), tumors (Nie et al., 2008), and other tissues (Unger et al., 2004; Chen et al., 2006). Although promising, the current state of this novel gene delivery method has yet to achieve efficient transfection within a clinically practical microbubble concentration. Microbubbles bearing genetic material, e.g. by charge-coupling of DNA to the cationic surface, can be introduced into cells by transient sonoporation. DNA-loaded microbubbles are injected intravenously, and high pressure, low-frequency ultrasound is then applied to the region of interest to destroy microbubbles as they pass through the local microcirculation, producing a site-specific concentration of the gene. Ultrasound-induced oscillation of the gas core of the microbubble creates pores in endothelium and surrounding cell membranes, through which genetic material may pass to enter cytoplasm. Larger molecules, such as DNA plasmids, may also enter cells via an endocytic pathway. In addition, microbubble-based agents can be observed by ultrasound, demonstrating localization and concentration of injected materials.

However, a challenge is posed by the finite surface area of current microbubble vehicles, which limits loading capacity (nucleic acids are insoluble in the gas phase and thus cannot be encapsulated). For example, current loading capacity of lipid-coated microbubbles is approximately 80 μm² for a 5-μm diameter microbubble. Considering a “hit-and-stick” adsorption model, the surface density is approximately 0.0001 pg/μm² for a 10 kbp DNA plasmid, resulting in an estimated maximum loading density of ~0.01 pg/microbubble. Thus, a method is needed to condense DNA and increase the total available surface area of the microbubble. One technique we have previously used to increase loading capacity is by layer-by-layer (LbL) assembly of a polyelectrolyte multilayer (PEM) composed of DNA and a biocompatible polycation (Borden et al., 2007). Unfortunately, although this led to excellent loading capacity, the bioavailability was poor, leading to insufficient transfection efficiency. In the current research, we are developing an alternative strategy to improve DNA packaging for enhanced release from the microbubble shell and subsequent intracellular trafficking.

Copyright Columbia University