Optical molecular imaging of tumor cells

One of the most promising biological targets for cancer therapy is the epidermal growth-factor (EGF) receptor (EGFR), a transmembrane glycoprotein that controls multiple biological phenomena, including proliferation, growth of new blood vessels from preexisting veins (angiogenesis), tissue invasion, and finally the spread of disease from one organ to another (metastasis).¹ Targeted treatments that selectively inhibit this receptor have found widespread clinical application, but not all patients respond positively. There are few reliable methods to predict response to therapy a priori based on pretreatment tissue-sample tests. Instead, noninvasive imaging of receptor expression and/or activity over time might have greater potential for treatment decisions based on early molecular response assessment.

Optical imaging using fluorescent techniques represents a promising approach since it is inexpensive and quick, involves no exposure to ionizing radiation, and provides spatiotemporal resolution on the basis of relatively small data sets compared with other conventional imaging methods.² Despite encouraging early results using organic fluorophores with emission peaks at near-IR wavelengths, their irreversible photobleaching properties significantly limit repetitive and quantitative imaging. On the other hand, near-IR-emitting quantum dots (QDs)^{3, 4} have advantages that prompted us to use EGF-conjugated quantum dots (EGF-QDs) (see Figure 1) in imaging EGFR expression of human colorectal-cancer tissues transplanted in mice.

Figure 1. Chemical structure of an epidermal growth-factor (EGF)-conjugated quantum-dot (QD) nanoprobe.

EGF was chosen as the preferred molecule for docking to EGFR. First, the EGF binding affinity for its receptor generally exceeds that of synthetic antibodies. Second, in contrast to monoclonal antibodies or antibody fragments, the smaller peptide size permits increased penetration of solid tumors and more rapid clearance. Finally, the polyvalence effect of multiple peptides connected to a single probe permits stronger binding to the receptor.⁵ In vitro EGF-QD nanoprobes evaluated in cells with varying levels of EGFR expression demonstrated excellent binding affinity to cell-surface EGFR. In vivo binding affinity was evaluated through serial imaging of HCT116 (human colorectal tumor 116) colorectal-cancer tissue transplanted in mice after intravenous injection of EGF-QDs (10pmol equivalent of QDs) or unconjugated QDs (see Figure 2). The images were spectrally unmixed and then remixed to extract the QD fluorescence from the tumors.

Figure 2. Representative in vivo images. White circles indicate the tumor site.

Analysis of the image data showed three distinct phases of fluorescence-signal enhancement, i.e., influx, clearance, and accumulation. The initial phases of influx (~3min post-injection) and rapid clearance are caused by the increased vascular volume and permeability of blood vessels in the tumor tissue. At ~1h post-injection, fluorescence signals from both groups reached near-baseline values in an apparent dynamic equilibrium between the vascular and extracellular (and perivascular) space (see Figure 3). Subsequently, between 1 and 6h post-injection, a steady flow of EGF-QD nanoprobes accumulated at a rate of 0.24±0.10h⁻¹, followed by a return to near-baseline levels after 24h. On the other hand, unconjugated QDs exhibited a slow exponential tumor-signal decrease with no progressive accumulation. Pretreatment with the anti-EGFR antibody Cetuximab^® (1.05nmol) further decreased the EGF-QD-mediated tumor-fluorescence signal, confirming the selective binding affinity of EGF-QD nanoprobes. The temporal behavior of the accumulation of EGF-QD nanoprobes in EGFR-expressing tumors suggests that these probes could enable repetitive imaging of EGFR expression during and after therapeutic interventions. In addition, our use of a moderately EGFR-expressing cell line and a probe that also recognizes mouse EGFR in adjacent tissues mimics a clinical scenario where we need to discriminate between tumors and surrounding normal tissue.

Figure 3. Tumor-to-background ratio of (a) QD nanoparticles and (b) EGF-QD nanoprobes.

In vivo observations were validated by ex vivo fluorescence imaging, tissue-homogenate fluorescence, and confocal microscopy of organs harvested 4 and 24h after injection. Liver and spleen demonstrated maximum uptake of EGF-QDs and QDs on ex vivo images. Tumors from EGF-QD-injected animals showed higher fluorescence intensities compared to QDs. Fluorescence from homogenized tissue showed a significant difference between QDs and EGF-QDs at 4h but not at 24h. Confocal microscopy of frozen tissues revealed homogeneously distributed bright fluorescence from EGF-QD nanoprobes and nonhomogeneous, patchy, and relatively weak unconjugated QD fluorescence in tumors at 4h. Negligible fluorescence was observed from tumors at 24h (see Figure 4). Immunofluorescence staining revealed colocalized EGFR- and QD-fluorescence signals in the functional part of tumors (parenchyma) in the EGF-QD group but not in the unconjugated QD group. Instead, the QD group exhibited colocalization of QD fluorescence with patchy areas of CD31 (a human ‘cluster-of-differentiation’ cell-adhesion molecule) vascular staining, suggesting some separation of unbound QDs within leaky blood vessels in the tumor.

Figure 4. Confocal-microscopy images showing the distribution of EGF-QD nanoprobes and unconjugated QDs in different organs. The scale bar represents a length of 50μm.

Optical molecular imaging in combination with nanobiotechnology is emerging as a powerful tool for studying the temporal and spatial dynamics of specific biomolecules and their interactions in real time. We have demonstrated that a peptide-conjugated QD imaging probe can distinguish a tumor overexpressing a certain ubiquitous receptor from adjacent normal tissues that also express the receptor, albeit at a lower level.^5,6 The excellent specific binding affinity, low concentration of EGF-QD nanoprobe needed, and the favorable kinetic parameters for imaging suggest that this probe will permit quantifiable and repetitive imaging of EGFR in tumors to guide treatment decisions based on early molecular response assessments. We will next explore the applicability of similar nanoprobes for early detection of cancers (primary, recurrent, or metastatic), image-guided receptor-targeted sentinel-node tissue tests (biopsies), individualized therapies based on treatment monitoring that provides early indications of positive responses, customized dose-optimization of targeted therapy, and accelerated drug screening and development.

This research was partially funded through grants from the Hitachi Corporation, Japan.