Robert H. Pierce 1, Xiaohu Gao 2
1Program in Immunology, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N, Seattle, WA 98109, USA, 2Department of Bioengineering, University of Washington, 3720 15th Ave NE, WA 98195, USA
Material Matters, 2019, 14.2
A detailed understanding of the dynamics of intercellular communication and intracellular signaling networks is highly sought-after by researchers involved in cell biology, pathology, clinical diagnostics, and drug discovery. Untangling and managing the complex interactions underlying cancer, neurological diseases, and immune system disorders are of particular interest because of their potential to unlock cures. Investigation of these biological processes requires affordable analytical tools with high-resolution, high-sensitivity, and the capability to perform multiplexed analysis. Fluorescent probes, such as organic fluorophores, simultaneously offer many of these features, and are at the core of many fundamental discoveries. At the same time, with more and more genomic, transcriptomic, and proteomic work identifying distinct biological processes, the level of complexity is increasing beyond the capability of conventional fluorophores, hampering the development of diagnostic kits and new therapeutics targeting key molecular mechanisms of pathogenesis.
Quantum dots (QDs), emerged as a result of the explosion in nanotechnology research conducted during the past three decades, offer a promising solution for researchers in need of more capable bioassay and bioimaging technologies. In terms of optical properties, QDs are superior in almost every aspect to organic dyes (Table 1). Their narrow and sizetunable light emission makes them ideal for multiplexed imaging. The Gaussian-shaped emission peaks with a full-width-at-half-maximum (FWHM) of approximately 20 nm to allow 5 to 10 colors to fit into the visible spectrum alone, free of spectral overlap (or at least resolvable by spectral imaging).1 The narrow emission of QDs is complemented by efficient light absorption over a broad spectral range (hundreds of nanometers), enabling simultaneous excitation of multiple colors by a single light source. This feature not only reduces the cost of imaging instrumentation (e.g., filter design) and simplifies data analysis (e.g., fluorescence intensity compensation), but also improves imaging sensitivity, as it can create a large Stokes shift away from the biological sample autofluorescence. QD fluorescence can also be differentiated from background autofluorescence using a gating approach that takes advantage of the long excited-state lifetime of QDs, which is approximately an order of magnitude longer than those of organic dyes. Although in theory the long lifetime could reduce the photon production rate (and consequently probe brightness), most biological imaging applications operate under absorption-limited conditions rather than lifetime-limited conditions, which require a very high excitation photon flux. Therefore, QDs with much higher molar extinction coefficients are found to be brighter probes than organic dyes. For organic dyes and QDs with similar quantum yields, individual QDs are generally 1–2 orders of magnitude brighter than dye molecules. In addition, QDs are significantly more resistant to photobleaching, well-suited for long-term imaging, and tracking applications.
Despite the remarkable optical properties, the synthesis and engineering of QDs have taken a long time to develop to the point where QDs became practical to use in biological imaging and bioassays. Since the initial discovery and establishment of quantum mechanical models for QDs in the early 1980s,2-4 highquality QDs only became commercially available after nearly 20 years of development.5 The long development time period for QDs can broadly be divided in three major milestones: QD core synthesis, QD shell synthesis, and QD surface functionalization.
In 1993, Bawendi and co-workers6 first reported a new route to synthesize monodisperse QD core particles. Deviating from the then-popular aqueous solution-based reactions, this approach instead used high temperature, organometallic reactants, and organic solvents. These conditions enabled rapid crystal nucleation and created a precise control between the competing processes of Ostwald ripening and molecular addition during nanocrystal growth.7 As a result, this approach not only produced highly crystalline, uniform II-VI QDs of specific sizes, but also became the basic framework for the synthesis of other monodisperse nanoparticles, including III-V semiconductor, magnetic, metallic, and perovskite nanoparticles.8-12
However, QDs synthesized using Bawendi’s method often suffered from quantum yields of less than 10%, severely limiting production of bright imaging probes. Due to the small particle size, a large percentage of the atoms were on the QD surface. These surface atoms created crystal defects that trapped charge carriers, thus preventing exciton recombination and fluorescence emission. A simple yet transformational solution was first reported by Hines and Guyot-Sionnest in 1996,13 in which a thin shell was grown on top of the core nanoparticles. The shell was comprised of another semiconductor with a similar crystal lattice and higher energy bandgap. It effectively passivated the core particle surface and confined excitons to the core, thus promoting recombination and fluorescence emission. The initial core-shell QDs reached quantum yields of 30-50%; this was later improved to 80-90% through the use of more complex shells.14
The third technical milestone required to make QDs useful for biomedical applications was surface functionalization. Highquality QDs synthesized in organic solvents have a layer of hydrophobic surface ligands, which are incompatible with biological systems. Initial attempts to make QDs water-soluble were based on ligand exchange and silica coating, but both methods displaced the hydrophobic surface ligands15,16 that were important for QDs’ optical properties. In addition, watersoluble QDs prepared using these methods did not form stable colloids. Freshly synthesized QDs could be used in lab research for proof-of-concept studies but were unsuited for commercial applications. A number of labs explored amphiphilic polymers to make QDs water-soluble while retaining the hydrophobic ligands on their surface.17-20 These polymers solubilize hydrophobic QDs much like the way detergent molecules wrap around tiny oil droplets. For example, Bruchez and coworkers at Quantum Dot Corporation developed a polymer with multiple repeats of hydrocarbon chains and polar head groups (e.g., carboxylic acids).20 The hydrocarbon chains interdigitated into the hydrophobic surface ligands on the QD surface, whereas the polar head groups made the QDs water-soluble (Figure 1). This successful strategy soon became the base structure for most commercially available QD bioconjugates.
Figure 1.Schematic drawing of a common QD probe featuring a core particle capped by a layer of hydrophobic ligands, an amphiphilic encapsulating polymer layer for solubility, a polyethylene glycol (PEG) layer for non-specific binding reduction, and a targeting ligand for biomolecular recognition.
The bioapplications of QDs can also be categorized into three general groups: sensing and detection, in vitro labeling and imaging, and in vivo imaging. For sensing and detection, an optical readout is used in both homogeneous assays (solutionbased detection) and heterogeneous assays (assays on solid supports). The advantage of using QDs lies in their ability to improve the detection sensitivity by virtue of their unique optical properties including brightness and a large Stokes shift. One of the earliest QD biosensors based on fluorescence resonance energy transfer (FRET) was developed by Medintz, Mattoussi and Mauro.21 In this system, recombinant maltose-binding protein (MBP) with a C-terminal histag self-assembled onto the QD surface through chelation. A β-cyclodextrin-fluorescence quencher conjugate bound in the MBP saccharide-binding site causes initial quenching of the QD fluorescence, whereas the presence of maltose capable of displacing the<br> β-cyclodextrinquencher conjugate restores QD fluorescence for detection. A disadvantage of using QDs in FRET assays is their physical size, as well as the surface coating materials, especially for highly stable amphiphilic polymer-coated QDs. These physical barriers increase the distance between energy donors and acceptors, thus reducing FRET efficiency. Fortunately, the reduced efficiency can be overcome using the tunable emission and large surface area of QDs, which allow optimization of spectral overlaps between donors and acceptors and immobilization of multiple energy acceptors per QD, respectively.22 Furthermore, combining QD-based FRET probes with microfluidic devices can enable separation-free, single molecule-level (<50 copies) detection of target DNA sequences. This is possible because the confinement of multiple oligonucleotide probes on the QD surface captures and concentrates multiple target molecules on the same QD.23
The second popular application of QDs is fluorescent labeling and imaging. QDs can be conjugated to targeting ligands including small molecules, antibodies, peptides, and oligonucleotides. These conjugates can then be used to label target molecules immobilized in paper strips, membranes, biochips, gels, or cells. Molecular recognition is achieved by specific interaction between a target molecule and its ligand (e.g. antigen and antibody, and complementary DNA strands), and the target location and abundance are revealed by fluorescence. For example, immunohistochemistry (IHC) has been a workhorse for biological research and clinical diagnosis for more than 50 years. Conventional organic dye-labeled antibodies (or antibody labeling kits) are broadly available, but they suffer from a number of limitations. First, only 2–3 colors of organic dyes can be used in parallel, due to spectral overlap. Second, it is difficult to accurately quantify fluorescence signals because dye molecules quickly photobleach. Third, biological samples such as cells and tissues often have high background signals (autofluorescence) that interfere or even obscure spatially specific labeling. QDs conjugated to primary antibodies address all these problems. Single QD-antibody conjugates are sufficiently small to diffuse into properly fixed cells. Indeed, a number of research groups have demonstrated routine multicolor imaging of cell-surface, cytoplasmic, and nuclear antigens.1,20,24-26
Despite these advantages, QDs are not yet the primary choice for fluorescence labeling largely for two reasons. First, the multiplexing capabilities of QDs only improve staining from 2–3 colors to 5–10, still far below the level necessary for comprehensive molecular profiling. Second, although QD-antibody conjugation protocols have been well-optimized and developed,27 they are labor-intensive, making it costly to produce a large library of conjugates.
Recently, we developed a multicolor, multicycle, molecular profiling (M3P) technology that addressed both problems.28,29 To expand the multiplexing capability, we combined 5–10 color QD-antibody conjugates in a single cocktail, and incubated the mixture with cells or tissue sections to demonstrate parallel multiplexed staining (Figure 2). Following fluorescence microscopy, the stains were removed and the sample was regenerated for another round of multicolor staining. We have developed protocols for complete destaining of the cells with no signal carryover and without affecting cell morphology or biomarker antigenicity, thus allowing the next full cycle of IHC staining to identify a different subset of biomarkers. With each staining cycle, 10 biomarkers can be analyzed, using 10 spectrally distinct QDs. Performing IHC staining for 10 sequential cycles generates 10 subsets of data for the specimen, yielding an overall molecular profile consisting of 100 distinct biomarkers.
Figure 2.M3P technology enabling simple custom probe preparation and highly multiplexed single cell imaging. Key steps involve A) a universal QDadaptor protein platform for single-step purification-free assembly of QD-antibody, B) mixing of multicolor QD probes into a single cocktail, C) parallel multiplexed staining, D) multicolor imaging, and E) destaining for another staining cycle. Bottom panel shows a representative fluorescence image of 5-target staining in Hela cells. Images adapted with permission from reference 28, copyright 2013 Nature Publishing Group.
IHC staining to identify a different subset of biomarkers. With each staining cycle, 10 biomarkers can be analyzed, using 10 spectrally distinct QDs. Performing IHC staining for 10 sequential cycles generates 10 subsets of data for the specimen, yielding an overall molecular profile consisting of 100 distinct biomarkers.
To reduce the labor required to make custom QD-antibody bioconjugates, we replaced the low-yield covalent QD-antibody conjugation with a non-covalent self-assembly protocol joining a universal QD-protein A platform with a variety of intact primary antibodies. The method requires no chemical reactions or purifications for end users, and enables quick and easy preparation of custom QD-antibody panels at extremely small scale (Figure 2A). The non-covalent interaction between the adaptor protein and primary antibodies is sufficiently stable during staining, and does not cross react. Therefore, the M3P technology enables extensive molecular characterization of cells within their native microenvironment in just a few cycles, making QDs practically useful for highly multiplexed IHC.
The third category of QD bioapplications is in vivo imaging. Dubertret et al.17 reported the first use of QDs in a living
organism, a frog embryo. The chemical- and photo-stability of QDs allowed cell lineage tracking and comparative embryology studies for up to four days, without abnormalities in embryo development. For in vivo targeting, Akerman et al.30 showed that QDs guided by peptides concentrated at tumor blood vessels using ex vivo histology sections. Gao and coworkers first demonstrated non-invasive tumor imaging in mice using QD-antibody conjugates.18 Hyperspectral imaging was used to help delineate QD fluorescence from background because mouse skin autofluorescence was still high even when red QDs (relatively long wavelength with large Stokes shift) were used to make the targeted imaging probe. For improved light penetration depth and reduced autofluorescence, Kim et al.31 prepared another type of QD in which the core valence and conduction band edges were both lower or higher than those of the shell. This arrangement effectively red-shifted QD emission to the near infrared (NIR) region, at the cost of exciton recombination rates. These NIR QDs were injected in mice and pigs intradermally, where they quickly drained to nearby lymph nodes, allowing image-guided surgery.
The spectral rainbow enabled by QD technology and the science behind it are both beautiful and fascinating. Ever since QDs were first used in cell staining, they promised a quantum leap for bioassay and bioimaging. Realization of that promise, however, has been delayed by a number of technical hurdles. As a result, QDs struggled to become a robust tool for real biological discoveries over the past 20 years. The vast majority of QD bioapplications to date involve technology developments or proof-of-concept use in model systems. With most of the technical issues addressed recently, identification of killer applications for QDs (significant biological problems that conventional dyes are inadequate for), will be a focus in the coming years. Mutlicolor, quantitative biomarker imaging such as immunofluorescence and fluorescence in situ hybridization in cells and tissues remains an exciting area of research and development. For example, in immunotherapy research, highly multiplexed molecular profiling of the cell genome, transcriptome, and proteome within a native microenvironment is expected to reveal new secrets of the complex and dynamic human immune system. Understanding the molecular, cellular, and organismal interactions requires high-throughput and high-content molecular analysis tool kits. QDs can aid the development of immunotherapies that one day may transform the lives of people affected by various immune diseases.
For in vivo applications, QD probes are expected to have an immediate impact on small animal imaging. In preclinical studies of drug discovery, for example, QDs offer significant advantages over current imaging modalities (e.g., MRI and PET), simultaneously achieving high sensitivity, high resolution, and low cost. In addition, QDs are an ideal prototype material for nanoparticle engineering. A great deal of information, including the effects of particle size, shape, charge, surface coating, and targeting ligand, can be learned using QDs, and then applied to the design of other types of nanomaterials.
The prospects for the in vivo use in human diagnostics and theranostics remains unclear. Any application of QDs in humans must provide benefits that significantly outweigh the risks, and this remains to be shown. One fundamental limitation of optical imaging is that light does not penetrate deeply through tissues; QD toxicity is also a concern. The light penetration hurdle could potentially be overcome by advanced endoscopy technologies.
In summary, recent advances in both chemistry and colloidal science have made QDs a robust and readily available imaging tool. Following their commercial success in display technologies, widespread application of QDs in life science are not far behind. The unique optical properties of QDs complement those of conventional organic dyes and fluorescent proteins and will enable new discoveries in many unexplored or underexplored areas in biology. The future for QDs is bright.