High Speed Multi-photon PLIM

Comments
Add One

Phosphorescence lifetime imaging microscopy (PLIM) allows substances or tissues with different phosphorescence lifetimes to be identified with high spatial resolution. PLIM hasn’t found many practical applications so far, but it could be useful as a way of measuring oxygen concentration in tissues. Depth resolved images can be obtained using multi-photon excitation, a technique which ensures that all the signal comes from the focal plane. Unfortunately, relatively long phosphorescent lifetimes make the point-by-point scanning used in multi-photon microscopy very time consuming. Attempts to improve the frame rate using parallel excitation can result in cross-talk between pixels and blurring of the image. Now, a group from Cornell University has devised a way to acquire parallel excitation PLIM images which are free from cross-talk.

PLIM isn’t as well-known as its cousin, fluorescence lifetime imaging microscopy (FLIM), but the principles are very similar. If a fluorescent molecule is excited with a short pulse of light, it continues to fluoresce for some time afterwards. The intensity of the fluorescent emission decays exponentially and so, theoretically at least, never reaches zero. What we call the ‘lifetime’ is the time constant of this decay – the time taken for the intensity to drop to 1/e or 37% of its maximum. This lifetime is often characteristic of different molecules, even those which have similar excitation and emission wavelengths.

There are a few different ways of measuring the lifetime, although they all generally fall into one of two categories. In time domain FLIM, the sample is illuminated with a short excitation pulse, and the emission intensity is then measured as a function of time. Things are complicated slightly because the excitation pulse can’t be made infinitely short, so what we measure as the lifetime isn’t the true fluorescence lifetime, but this can be corrected for.

The alternative method, frequency domain FLIM, involves modulating the excitation light at a frequency comparable to the fluorescence lifetime. The emission will also be modulated, but with a slight delay. In other words, the finite fluorescent lifetime results in a phase shift between the modulation of the laser source and the collected emission. The size of this phase shift can be used to calculate the lifetime.

Phosphorescence

Phosphorescence differs from fluorescence by the order of magnitude of the time constant. Fluorescence lifetimes can be as short as picoseconds, while phosphorescence lifetimes range from microseconds to hours. So for FLIM we need to use high speed detectors and electronics (or invent clever ways to get around this). For PLIM we have the opposite problem, the long time-constant actually becomes a limit on how fast the images can be assembled. For practical applications, phosphorescent materials with lifetimes at the lower end of the scale are used, but the integration time per pixel is still considerable.

The problem becomes worse if we want to do depth-sectioned imaging. A conventional microscope collects light from as deep into the sample as the light will penetrate, but only a thin layer lying at the focal plane of the microscope objective will be in focus. This results in a large blurred background signal super-imposed on the image, making it difficult to see faint features. So optical depth sectioning – rejecting all the light which isn’t from the focal plane – is pretty much essential for looking at thick, scattering samples.

There are a few ways of implementing depth sectioning in a microscope, with confocal microscopy being the most common. For FLIM/PLIM imaging, two-photon microscopy is an attractive alternative technique. The idea is to illuminate the sample with light at twice the normal excitation wavelength. Because these photons have half the energy needed to excite the fluorophore, two must arrive almost simultaneously for excitation to occur. This means that the probability of excitation depends on the square of the intensity of light, and becomes very unlikely away from the focal plane region where intensity is highest. So almost all of the emission comes from the single plane, and depth sectioning is achieved.  The technique can be extended to three or more photons and so it often known by the more general name of ‘multi-photon microscopy’.

The probability of exciting a fluorophore by the two-photon method tends to be quite small, so high light intensities are needed. This requires a pulsed laser, and a scanning system to sequentially excite and image individual points on the sample. This point sampling also improves the lateral resolution because its avoids the scattering and cross-talk problems of wide-field illumination. While point-by-point scanning is fine for FLIM, it’s not ideal for PLIM because the frame rate is then limited by having to dwell at each point long enough to collect the phosphorescent emission. The idea of parallel excitation has been developed by several groups (see this paper1 for example), but there are problems using these methods for thick, highly scattering samples.

Frequency Multiplexing

The new paper by Howard et al.2 describes a method for performing parallel excitation PLIM while avoiding cross-talk.  Their method involves illuminating an entire line along the sample simultaneously. On its own this would lead to significant cross-talk along the line. To prevent this they ensured that the excitation intensity reaching each point was modulated at a slightly different frequency. This meant than the resulting emission was also modulated at different frequencies depending on its origin along the line.

The signal for the whole line was collected by a single pixel photodetector. The geometry of the line was then reconstructed by decomposing the signal into its frequency components. This neatly avoided the problem of cross-talk, because the original location of the emission was encoded in the modulation frequency. The modulation could also be used for frequency domain PLIM; measuring the phase shift for each frequency allowed the lifetime to be calculated.

The frequency modulation along the line was generated using a custom-made, photo-mask based spatial light modulator. A galvo mirror scanned the line over the reflective phase mask, resulting in sinusoidally varying intensities of different frequencies along the line. This method only produces an image of a line, so to assemble a complete frame they scanned the line across the sample in the other direction using a galvo mirror.

Howard et al. demonstrated their system by imaging mouse brain vasculature in vivo.  They were able to differentiate Ru(dpp3)-nanomicelle dye, which has a phosphorescence lifetime of about 2.5 microseconds, from fluorescent DsRed. They showed images from up to 100 microns in depth, which they claim is the deepest reported PLIM imaging.

More significant was the frame rate they achieved. While integration times of at least 20 s, and up to 5 minutes for the low noise images, might seem quite long, this is a dramatic improvement on previous work (see this paper3, for example). It’s arguably still a fair way short of the kind of ‘real-time’ imaging that would be needed for widespread in vivo use, but the frame-rate was limited by noise and not by having to dwell at each pixel long enough to collect the phosphorescence. So, in principle, integration times could be reduced by increasing the excitation intensity, or finding dyes with a larger two-photon cross-section. It would only take a further improvement by an order of magnitude or so to make two-photon PLIM a viable in vivo technique.

References

  1. H. Choi, D. Tzeranis, J. Cha, P. Clémenceau, S. de Jong, L. van Geest, J. Moon, I. Yannas, and P. So, “3D-resolved fluorescence and phosphorescence lifetime imaging using temporal focusing wide-field two-photon excitation,” Opt. Express 20, 26219-26235 (2012).
  2. Howard, Scott S., et al. “Frequency-multiplexed in vivo multiphoton phosphorescence lifetime microscopy.” Nature Photonics (2012).
  3. Lecoq, Jérôme, et al. “Simultaneous two-photon imaging of oxygen and blood flow in deep cerebral vessels.” Nature medicine 17.7 (2011): 893-898.

Leave a Comment

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>