Ytterbium-ion femtosecond laser reduces damage during biomedical imaging
Over the past decade, nonlinear microscopy techniques have become powerful tools for biomedical imaging on the cellular level. The nonlinear responses used for imaging, such as two-photon-excited fluorescence, fluorescence lifetime, and second-and third-harmonic generation, are efficiently induced only at the focus of a laser beam, inherently providing spatial registration of pump and probe like that in confocal microscopy. Three-dimensional images with high axial and depth resolution can be generated by recording a stack of images at different focal planes, providing noninvasive optical sectioning of a sample. To date, femtosecond pulses with wavelengths around 800nm, produced by titanium (Ti):sapphire lasers, remain the most popular source of excitation for these techniques because they provide the high peak powers that are needed to drive the nonlinear processes.
Observation of dynamic processes in living cells, however, often requires continuous imaging for 5–30min. Yet we have found that imaging with a Ti:sapphire laser in live cells and organelles such as cardiomyocytes (heart-muscle cells),1 yeast cells,2 and plant chloroplasts cannot be extended to more than a few minutes.3 Indeed, beyond that point the samples begin to degrade, for instance, bleaching or undergoing structural changes.
The predominant sources of degradation are heat generated by two-photon absorption, accumulation of long-lived triplet states, and photogenerated reactive oxygen species. Two-photon absorption of indigenous molecules is relatively strong around the 780–900nm region,4 but its effects can be greatly reduced by choosing lasers operating at longer wavelengths and at lower pulse repetition rates.
In this context, ytterbium (Yb)-ion-based femtosecond lasers5 operating around 1μm are particularly suitable, and can be used with high-resolution fluorescence-lifetime imaging and second-and third-harmonic generation microscopy. In addition to reduced two-photon absorption and autofluorescence, longer excitation wavelengths have less scattering and higher penetration depths in biological tissue. Furthermore, the second and third harmonics fall into the visible range, rather than UV as for 800nm excitation, resulting in higher throughput and simpler detection.
At still longer wavelengths, like those in the 1200–1500nm range produced by chromium (Cr):forsterite and erbium:fiber lasers, increased absorption by water limits the usable laser power for biological imaging. It is also worth noting that direct diode pumping of Yb-ion-doped laser gain media5 considerably reduces the cost of a system in comparison with Ti:sapphire and Cr:forsterite lasers, which are usually pumped by expensive solid-state or fiber lasers. Therefore, Yb-ion-based femtosecond solid-state lasers are ideally suited for biomedical imaging applications.
We recently developed a diode-pumped low-repetition-rate Yb:potassium badolinium tungstate (Yb:KGW) laser (see Figure 1) for application in nonlinear multimodal microscopy.6 When pumped with a 15W fiber-coupled laser diode at 980nm, the laser delivered up to 0.85W of average power with ∼200fs-duration pulses at a repetition rate of 14.6MHz, corresponding to a pulse energy of 60nJ and a peak power of ∼300kW.
The laser was passively mode locked by a semiconductor saturable absorber (SESAM) and used two Brewster prisms made of SF10 glass for dispersion compensation. The pulse repetition rate was reduced by incorporating two identical intracavity telescopes. Each telescope had a 1:1 imaging ratio and was constructed from a concave mirror with a radius of curvature of 2m and a plane folding mirror. Operation at low repetition rates not only helps reduce photo damage to the samples but is also beneficial for applications involving fluorescence lifetime measurements. In this case longer time between consecutive pulses (67ns vs. the usual 10ns) allows accurate measurements of long-lived fluorophores (up to ∼20ns).7
As an example, we investigated in vivo the structure of muscle cells from larvae of the fruit fly, Drosophila melanogaster, using this laser and a home-built nonlinear laser scanning microscope3,6,8 capable of simultaneous detection of two-photon fluorescence, second- and third-harmonic signals. The strong second-harmonic signal originating from the actin-myosin protein microcrystalline structure allowed visualization of the anisotropic bands of basic units of muscle contraction called sarcomeres (see Figure 2).
Continuous imaging of specimens up to several hours also enabled detailed studies of muscle contraction dynamics.8 No structural or functional alterations were observed during imaging, confirming that excitation at 1μm is highly appropriate for in vivo biological imaging. In contrast, when a Ti:sapphire laser was used for imaging cardiomyocytes under comparable conditions, rapid sample degradation was observed.1 We attribute this drastic reduction in useful imaging time to processes following efficient two-photon excitation of intrinsic cellular pyridine nucleotides, flavin adenine dinucleotides, and lipoamide dehydrogenase.1,4
In another experiment, this laser allowed prolonged fluorescence-lifetime imaging of isolated chloroplasts by significantly reducing fluorescence bleaching compared with 800nm excitation.3 The induced fluorescence quenching, and the decrease in second- and third-harmonic signals, was much stronger when imaged with a Ti:sapphire laser. The difference in the light-induced changes most probably reflects the excitation of chlorophylls by the 800nm excitation, producing detrimental reactive oxygen species via a triplet state. Two-photon excitation of carotenoids at 1μm, however, does not lead to this process, because their triplet state has a lower energy.9
The ultrashort pulse laser we developed represents significant advantages over previous excitation sources in biomedical imaging applications. The laser was successfully used for detailed, noninvasive structural and functional investigations of different biological samples. It has great potential for application in other areas, for example, optical DNA sensing technology.7
This work was supported by the Canada Foundation of Innovation, the Ontario Innovation Trust, and the Natural Sciences and Engineering Research Council of Canada.