Commercialization and Future Developments in Bionanotechnology
MARCEL P. BRUCHEZ
Carnegie Mellon University
Perfection in nanotechnology has long been achieved by biological systems. An enzyme represents a nearly perfect robot, stamping out molecular patterns from unique templates designed to execute individual tasks with nearly perfect efficiency. Evolution has driven these efficient designs to enable life forms to thrive in harsh environments. Evolutionary improvements have developed over a period of at least 3.5 billion years, with impressive results. The fact that we are here at all is a testament to the power and vast potential of nanotechnology.
Recently, we have made the first blunt-fingered attempts to extend the capabilities of biological systems by harnessing innovations in materials chemistry and electronics coupled with biologically defined specificity for both magnetic and fluorescent probes. In these cases, we have succeeded in introducing relatively limited new functionalities to existing biological systems. But we have barely tapped the potential of engineering of these systems, and from here on, our efforts will undoubtedly expand dramatically.
At the present time, we are guided by empirical observations and not by a detailed understanding of the interactions of biological systems with the materials and devices we are preparing. Thus, not only are we blunt-fingered, but we are also nearly blind. Before we can realize substantial commercial rewards and benefits in health and medicine, we will have to expand our efforts dramatically to develop new characterization methods and basic specifications and predictors of biological performance.
DEFINITION OF BIONANOTECHNOLOGY
At the present time, there is no consensus definition of bionanotechnology. To take advantage of the enthusiasm of funding agencies, a number of old (and important) areas, such as colloid science, molecular biology, and implantable materials surface science, have been relabeled “nanotechnology.” In fact, all of these fields, coupled with biological systems, should be included in bionanotechnology. In general, the idea of bionanotechnology is the engineering of interfaces between molecules or materials and biological systems. Looking ahead, the key areas for commercialization will be bringing engineered systems into biological contact and biological function.
The version of bionanotechnology popularized in the media has been largely oversold. The general idea, which was popular 20 years ago as the “magic-bullet” theory of biotechnology and has been adopted as the bionanotechnology target, can be described as the “dump truck” model of technology. In this conception, the technology components consist of a targeting moiety, either biological or nanotechnological, and one or more cargoes, which are envisioned as small machines capable of specific destructive or corrective action.
In reality, designing targeting molecules that are selective for diseased tissues and capable of delivering cargoes larger than a typical antibody has proven extraordinarily difficult, and molecular targeting of nanoscale devices greater than 5 nm outside the vascular space may prove to be prohibitively difficult. However, with no guiding principles for the effective biological direction of nonbiological molecules, this is still an open question.
In this paper, I describe three recent examples of commercialized bionanotechnology, beginning with the one that is the best characterized system. The three are antibody-directed enzyme prodrug therapy (ADEPT), superparamagnetic iron oxide particles for enhancing contrast on magnetic resonance images (MRIs), and quantum-dot technology for biological detection. Each of these shows the potential power and some of the challenges of integrating technologies at the molecular level.
ANTIBODY-DIRECTED ENZYME PRODRUG THERAPY
Perhaps the most salient and relevant example of a bionanotechnology currently being commercialized is the ADEPT method being investigated by Genencor and Seattle Genetics (Figure 1). In this method, an antibody-enzyme fusion is first prepared and isolated. This molecule can be designed with precise chemical (biological) composition, precise linkage geometry, and complete definition and characterization using standard molecular-biology techniques and biochemical methods. The antibody, linked to the enzyme, can be used to target the particular antibody-enzyme conjugate to the site of interest. In this way, a small antibody fragment is used to target a molecular machine (an enzyme) to a par-
ticular site of interest, and the machine is then used to generate a specific molecule at that site. In the clinic, a prodrug (a drug molecule modified to an inactive state that can be converted to an active state in situ) is administered to the patient. After the antibody-enzyme construct reaches its target site, the prodrug is administered and is converted by the enzyme to an active state. The local concentration of the active drug can be driven very high, even though the overall concentration remains very low. Thus, the therapy is both safer and more effective than a high dose of the toxic compound.
ADEPT is a highly characterized, highly effective example of bionanotechnology in action. However, even after 15 years of active research, these targeted prodrug strategies are still in the research or early clinical trial stage and not in general practice. This is a reflection of the complexities of developing biospecific performing technologies, which is likely to be a general problem for the development of nanotechnologies with high clinical impact.
SUPERPARAMAGNETIC IRON OXIDE PARTICLES
A second, more recognizable example of bionanotechnology in clinical use is Ferridex and Combidex superparamagnetic particles, marketed by Advanced Magnetics, which are being commercialized for enhancing MRI signals (Figure 2). The particles are modified with dextran (a polymerized sugar molecule) to create a biocompatible coating, which dramatically reduces nonspecific interactions in the body and increases the contrast of the instrument wherever the particles are present. When administered intravenously, they can easily be measured in a standard clinical MRI instrument. These materials are currently approved for imaging cancers of the liver and spleen.
Recently, the Combidex agent was rejected by the Food and Drug Administration (FDA) because of safety concerns and a lack of efficacy data. Questions regarding safety had arisen when at least one patient died in a clinical trial investigating the use of the Combidex agent for sentinel-node detection (finding
near-nodes that are most likely to contain cancerous cells), which is critical to the grading, staging, and appropriate treatment of cancers. The FDA did recommend, however, that with further appropriately designed trials, the compound may be approvable for specific indications.
I have been extensively involved in work on the third example—the use of fluorescent quantum dots for biological detection in research and, ultimately, clinical applications (Figure 3). Semiconductor nanocrystals (i.e., quantum dots), specifically designed to have intense monochromatic emission spectra, are coupled to biological targeting molecules, such as antibodies and nucleic acids. The conjugates can then be used to detect the presence of particular analytes in biological samples. Although these particles dramatically increase experimental information and sensitivity, the clinical community has been slow to adopt them because of subtle protocol differences between these materials and the typical fluorescent dyes and enzymatic methods used in detection. Many of the protocol differences are thought to arise from distinct size differences between typical probes and nanotechnology-based probes. Such idiosyncrasies are likely to be ubiquitous in nanotechnology-enabled product commercialization.
The technology for the use of quantum dots in biology was initially published in September 1998 in two simultaneous papers in Science (Bruchez et al., 1998; Chan and Nie, 1998). Although these articles generated significant enthu-
siasm in the scientific community, biologically useful products were not launched until November 2002. In the meantime, Quantum Dot Corporation was accused of hoarding the technology, stalling progress, and many other things.
In fact, the reasons for the delay were hardly nefarious—we have no rational framework for “optimizing” these materials. Therefore, although we were working very hard to make products that could be used successfully by the average biologist, every time we made an improvement to any aspect of the system, the entire process had to be revalidated. This empirical approach to product development resulted in a very long development time.
This delay was in addition to unavoidable delays in process development. A nanoparticle designed for a particular application is a complex multilayer structure, shown schematically in Figure 3. Scale-up of the initial chemistries used to make these nanoparticles (as published in Science) was exceptionally dangerous; procedures involved pyrophoric precursors, flammable solvents, and rapid additions and releases of explosive gases. To develop safe, scalable procedures, our scientists had to develop innovative techniques in all aspects of nanoparticle chemistry. Again, every innovation had to be validated through to the utility of
the final material, making the development cycle exceptionally onerous. Just imagine if, to validate a change in one hose of a car, you had to build an entire car with only that change included.
The lack of specifiability of our modules was a key challenge to commercialization. Specification will require detailed basic investigations of the properties and chemistry of nanoparticle materials in biological systems. In addition, we will have to establish analytical tools and quantitative descriptors to detail the distribution of properties present in a population of nanoparticles. This is categorically different from specification for organic molecules and proteins, in which properties can be effectively described by an average. In nanomaterials, performance properties may be dominated by a relatively small population of particles, so averaging cannot always be used.
THE CHALLENGE OF CHARACTERIZATION
Dramatically different tools are necessary for characterization of the three examples I have described. ADEPT, a fully biological system, can be characterized structurally, chemically, and on the basis of activity to ensure that each component of the system is capable of acting independently and that this behavior is preserved as the system components are brought together. Nevertheless, for reasons related to biological complexity, the use of ADEPT in the clinic has not yet proven beneficial. This problem gives some indication of the challenges ahead for nanotechnology solutions.
The second example, Combidex technology, is a homogeneous-particle technology covered with a natural material, dextran, that minimizes the complexity of the system. In this case, the particle size and shape (which can be characterized by electron microscopy) dictate its magnetic properties. The interaction of dextran on the surface dictates the in vivo behavior of these materials. Although the components can be characterized in great detail, the interaction of the dextran with the surface (e.g., the number of surface iron atoms that are actually covered) may be crucial to the fate of these materials in clinical use, an obstacle that was not predicted.
The interaction of molecules with surfaces in complex environments represents a critical area for analytical development. At the moment, many studies are carried out by x-ray photoelectron spectroscopy (XPS), a vacuum technique that does not show many solution interactions in the normal biological environment. Micro-rheology techniques might be valuable in addressing this issue.
The final example, quantum dots, an entirely engineered material, presents many characterization challenges. First, the particle itself is a complex structure, and the best available tools for characterizing these materials are capital intensive and often inaccessible. Essentially, methods such as energy-dependent XPS require a synchrotron source. Other methods, such as Z-contrast scanning transmission-electron microscopy, require unique instrumentation that is available
only in a few laboratories in the world. In addition, these tools are suited either for measuring average properties or measuring single-particle properties, but not both. Bridging the gap to a statistical method that shows single-particle properties in a large population of particles would allow for discrimination of population properties from single-particle properties.
Moving out in the structure, the surface is coated with ligands. Thus, surface interactions will cause the same problems as have arisen for Combidex—routine tools do not give a detailed molecular picture of interactions at the surface. The problem is further complicated as particles are modified for biological applications, for instance by coupling polyethylene glycol molecules to the surface.
The characterization of chemistry on the surface of these particles has not matured to the level of typical organic chemical reactions. In fact, much of the characterization is still inferential (i.e., we analyze what does not react with the particle to determine what does react with it). The tools we have today neither discriminate between adsorbed and reacted materials nor determine whether the chemistry is homogeneous or heterogeneous from particle to particle. These distinctions will be critical for the development of nanoparticle tools with biomedical applications.
Where, then, will bionanotechnology take us? As the examples I have described show, advances have progressed from ADEPT, a completely characterized system with a defined molecular structure (still encountering difficulty in clinical acceptance), to a system in which components are well characterized (Combidex), to quantum dots, a system we still cannot fundamentally characterize. Chemists have tools like mass spectrometry and nuclear magnetic resonance spectroscopy to guide them. Engineers have testing and measurement systems for validating systems as small as a few hundred nanometers. In the middle range, however, nanoengineers (or nanochemists) still do not have the fundamental tools to determine how well they have done their jobs or in what direction they should look for improvements.
Devices are synthesized on molar (~1026) scales, but the characterization tools designed for molecules do not work effectively for bionanotechnology systems. Clearly, the device characterization methods (typically single “device” characterization on enough devices to ensure a reliable measurement of production-run statistics) are inappropriate, especially when a dose is 1013 devices and a minor population component can dominate the bad effects (for instance pore-clogging).
Thus, we have an acute and growing need for specifiability in the design of bionanotechnology tools. To achieve engineerable systems, a concerted effort must be made to conduct a basic scientific investigation of the impact of materials properties on the biological behavior of bionanotechnology systems, com-
bined with a physical scientific investigation of new methods to characterize the detailed physical and population properties of nanometer-scale materials and components. Specifiability will make predictability, falsifiability, and rapid progress in commercial bionanotechnology feasible.
Alderson, R.F., B.E. Toki, M. Roberge, W. Geng, J. Basler, R. Chin, et al. 2006. Characterization of a CC49-based single-chain fragment-beta-lactamase fusion protein for antibody-directed enzyme prodrug therapy (ADEPT). Bioconjugate Chemistry 17(2): 410–418.
Bagshawe, K.D., S.K. Sharma, and R.H.J. Begent. 2004. Antibody-directed enzyme prodrug therapy (ADEPT) for cancer. Expert Opinion on Biological Therapy 4(11): 1777–1789.
Wu, A.M., and P.D. Senter. 2005. Arming antibodies: prospects and challenges for immunoconjugates. Nature Biotechnology 23(9): 1137–1146.
Harisinghani, M.G., and R. Weissleder. 2004. Sensitive, noninvasive detection of lymph node metastases. PLOS Medicine 1(3): 202–209.
Harisinghani, M.G., J. Barentsz, P.F. Hahn, W.M. Deserno, S. Tabatabaei, C.H. van de Kaa, et al. 2003. Noninvasive detection of clinically occult lymph-node metastases in prostate cancer. New England Journal of Medicine 348(25): 2491–2495.
Alivisatos, A.P. 2004. The use of nanocrystals in biological detection. Nature Biotechnology 22(1): 47–52.
Bruchez, M.P., P. Gin, M. Morrone, S. Weiss, and A.P. Alivasotos. 1998. Semiconductor nanocrystals as fluorescent biological labels. Science 281(5385): 2013–2016.
Chan, W.C.W., and S. Nie. 1998. Quantum dot bioconjugates for ultrasensitive nonisotopic detection. Science 281(5385): 2016–2018.
Michalet, X., F.F. Pinaud, L.A. Bentolila, J.M. Tsay, S. Doose, J.J. Li, et al. 2005. Quantum dots for live cells, in vivo imaging, and diagnostics. Science 307(5709): 538–544.