Nanoparticles that generate light through a mechanism known as second harmonic generation have been used to image live tissue. The particles overcome many problems associated with fluorescent probes for bioimaging.
Biologists love optical microscopy. Open any cell-biology journal and you will be struck by a collection of colourful microscopy images worthy of a Seurat retrospective, with fluorescent staining patterns that help to unravel the complexities of cell signalling and development, neuroscience and cancer. But while microscopes have grown steadily more sophisticated over the years, so much so that single-molecule bioimaging is no longer uncommon, many of the fluorescent probes used for microscopy have remained unchanged for decades or longer. These traditional organic 'fluorophores' can present problems for biologists' increasingly demanding imaging experiments — for example, limits on probe brightness and stability often curb the types of data that can be acquired.
Recently, a new and unexpected class of probe has emerged that overcomes many of the shortcomings of traditional fluorophores: inorganic nanoparticles. These nanoparticles are based on elements far beyond carbon in the periodic table, and have optical properties entirely unlike those of their organic predecessors. Reporting in Proceedings of the National Academy of Sciences, Pantazis et al.1 describe nanoparticles that convert near-infrared (nIR) light to visible light, a property that may overcome several persistent problems in bioimaging. What's more, the authors illustrate the utility of their nanoparticles for live-tissue imaging.
When fluorescent molecules absorb light, they enter an excited state and, after a certain lifetime, can return to the ground state by emitting lower-energy light that is observed as fluorescence (Fig. 1a). The energy difference between absorbed and emitted light is known as the Stokes shift and ensures, for example, that fluorescein (the fluorophore that is the traditional workhorse of bioimaging) and green fluorescent protein (GFP) emit green light after excitation with blue light. There are, however, two key drawbacks to using Stokes-shifted probes for cellular imaging. First, cells have many Stokes-shifted fluorescent molecules of their own, leading to unwanted background fluorescence that can be particularly problematic in sensitive experiments such as single-molecule imaging. Second, the intense light typically used by microscopes to excite fluorescent molecules can damage cells, especially light at the ultraviolet (UV) and blue end of the visible spectrum. Photodamage may also occur indirectly, if the excited state of a fluorescent molecule reacts with a nearby molecule instead of emitting fluorescence. Adding insult to injury, the fluorophore is often 'photobleached' in this process, rendering it dark and useless.
One way in which microscopists have circumvented these problems is by building two-photon microscopes, which use pulsed nIR lasers powerful enough to deliver two photons to a probe nearly simultaneously. The additive energy of the photons excites the fluorophore and leads to anti-Stokes emission, in which light is emitted at higher energies (shorter wavelengths) than the excitation source (Fig. 1b). Not only is the nIR light that is used to stimulate anti-Stokes emission less damaging to cells than UV or visible light, but it also scatters less and penetrates farther into tissue, making it useful for whole-animal imaging experiments. A drawback of two-photon excitation is that it is extremely inefficient, so that ultra-high peak powers of pulsed lasers are essential for achieving the large photon flux needed to produce observable signals.
In their work, Pantazis et al.1 use a technique called second harmonic generation (SHG; Fig. 1c) to produce an optical signal for bioimaging. In SHG, two photons interact with certain asymmetric materials — such as non-centrosymmetric crystals, which lack a central point of symmetry — to produce a single photon with twice the energy (half the wavelength) of the incident photons. Because SHG is based on scattering, rather than fluorescence, there is no excited state and therefore no photobleaching or cytotoxic by-products. SHG has been used to image repeating asymmetric cellular structures such as cytoskeletal proteins, as well as other biological systems to which certain SHG-active organic dyes have been added2,3. Previous work has also shown that SHG works well when nanoparticles made of noble metals or metal oxides are used as probes4,5.
Pantazis et al.1 now extend SHG to whole-animal imaging. They started by surveying a series of nanocrystals for SHG activity in vitro, and found that 30-nanometre barium titanate (BaTiO3) crystals produced the strongest signal, exhibiting none of the undesirable on/off blinking behaviour of comparable fluorescent probes. When the authors injected BaTiO3 nanocrystals into live zebrafish embryos, the resulting SHG images showed minimal background signal and little apparent loss of brightness over time, consistent with a lack of photobleaching or chemical breakdown of the probe. This suggests that these particles could be useful in tracking cell lineages during embryonic development, as the stable particles are passed from cell to cell. It will be interesting to see for just how long the particles can be tracked in living systems.
Like two-photon imaging, SHG requires high-power pulsed nIR lasers, but another new class of anti-Stokes-emitting nanocrystal may obviate this need. These crystals, known as phosphorescent upconverting nanoparticles (Fig. 1d), also display no observable photobleaching or blinking6,7, and have been imaged in live mice8. A key distinction is that upconverting nanoparticles are visible with about 1,000-fold weaker illumination intensity than is needed for SHG of the BaTiO3 nanoparticles, such that even a modest continuous-wave laser is sufficient to make single particles visible7. But it might be possible to reduce the photon flux required for the activity of SHG-active particles by changing the surface chemistry of BaTiO3 nanoparticles, or by using other nanoparticle shapes or compositions.
Will inorganic nanoparticles have as large an impact on bioimaging as fluorescent proteins such as GFP have had? The answer mainly depends on how readily nanoscientists can adapt them for easy integration into cells. Quantum dots — luminescent semiconducting nanoparticles that have exceptional stability and brightness — have been used in bioimaging for more than a decade. But their use has been hampered in part by difficulties in attaching them to biological molecules and by their large size. Pantazis and colleagues' BaTiO3 nanoparticles1 are even larger, more than twice the diameter of a ribosome.
Recent advances in the synthesis of quantum dots9 and upconverting nanoparticles10 have shrunk them to antibody-sized structures that should be more palatable for biologists. Similar reductions in the size of SHG-active nanoparticles, as well as improvements in surface passivation (which makes the particles biocompatible) and in methods for attaching them to biological targets, could dramatically expand their reach. Cell biologists may not completely abandon fluorescence in favour of phosphorescence or light scattering by inorganic nanoparticles, but the absence of blinking, background noise and photobleaching may be too alluring for some to resist.