FEATURE

Organic growth

Stephen Mounsey takes a look at some of the applications for biophotonics, both in the laboratory, and elsewhere, and looks at the ways in which photonic technologies are enabling the life sciences

Most of the features printed in Electro Optics over the course of the last year have touched on the challenges of the recession. Biophotonics, however, is an application area that has weathered the storm particularly well – in part, because of good levels of public sector funding. Many applications, old and new, look set to come up swinging as the markets recover.

Optics has been a vital tool of biological research since Robert Hooke first used a microscope to describe the cellular structure of cork in 1665. The abilities of a conventional microscope are limited in the power of what they can resolve by the diffraction limit – a fundamental limit related to the wavelength of the light used. Confocal microscopy bypasses some of the limits of the conventional (wide field) approach by constricting the view to a single focal plane within the sample. In order to make the technique versatile, modern microscopes offer a number of different illuminating wavelengths, usually provided by several different lasers. The wavelength used has important consequences for the depth of the focal plane, as well as illuminating different samples in different ways.

Fluorescence imaging takes the importance of the illuminating wavelength a step further by introducing fluorescent marker compounds to the a biological sample. The dyes used (called markers of labels) will bind to parts of a biological sample, and different markers have varying degrees of specificity as to which parts of the sample they’ll bind with. All of these markers, however, contain a fluorophore – a section of the dye molecule that fluoresces when excited by a specific wavelength of light. It is this fluorescence that forms the spectacular images captured by confocal fluorescence microscopy. In order to build up a multi-coloured micrograph, with different colours representing different dyes bound within the sample, several wavelengths of light must be used.

John Clowes, director of business development at UK-based fibre laser manufacturer Fianium, describes the difficulties associated with providing several wavelengths of light: ‘Traditionally, a large confocal microscope might have up to eight individual lasers. These could be laser diodes, frequency doubled Nd:YAG lasers, frequency-tripled lasers in UV wavelengths, 488nm gas lasers, argon ion lasers, or other complex sources. Nowadays, the [manufacturers] are generally moving towards laser diodes, but in the past they used to use a combination of gas lasers, YAG lasers, and laser diodes. A lot of people are investing time and effort into developing new wavelength [laser] diodes, and increasing their power. For example, there are now 488nm diodes available, and these replace one of the argon ion laser lines that the lasers used to use.’ This is important, because a solid-state laser diode is cheaper and easier to integrate than a gas laser, or a DPSS laser.

Horses for courses

While high-end microscopes use lasers to achieve the specific wavelengths required, monochromatic light sources across a selection of wavelengths can be obtained through the use of more simple, non-lasing LEDs. These are substantially cheaper than lasers. Paul Knight, divisional manager at UK distributor Pacer International, explains the importance of having the right wavelength when it comes to these LEDs: ‘With an LED, as with a laser, you can select the wavelength you want to use with minimal filtering, so that you get the most efficient solution. In the real world, in the human world if you like, you can mix and match wavelengths to fool the eye into believing that it’s seeing white, or a brighter green, or a duller red, but in the biological world, you can’t fake it. We’re talking much more about developing the LED structure to give a more specific wavelength than we would have to in other applications,’ he says – adding that, although an LED cannot offer the performance of a laser in terms of wavelength specificity and stability, they are still suitable for some applications. ‘Some of the LEDs are pretty narrow bandwidth now, and it’s a question of what advantages a laser would offer [for an application] – it’s a case of horses for courses. If all you need is some fluorescence excitation, then an LED will have sufficient light, in a given wavelength band, to do the job,’ he says. Knight notes, however, that LEDs suffer a limitation in terms of wavelength drift over continuous periods of operation, although he also states that this can be controlled to an extent by using a sensor to provide feedback to the LED’s control electronics.

Fianium offers a different solution to the same problem. Fianium’s supercontinuum systems use ultrafast pulses generated by its picosecond fibre laser, coupled into a photonic crystal fibre (PCF), to create a supercontinuum laser – a broad-bandwidth, essentially white laser. Clowes explains why this is preferable to other approaches: ‘However a confocal microscope works, you still need up to eight individual lasers, and you need to beam-combine each of those, and have the control electronics for each of them. But, with a supercontinuum, you can effectively have all eight at once, if you’ve got the right filters,’ he says. Additionally, the filtered supercontinuum source may be able to excite fluorescent markers more effectively than other lasers: ‘Because diodes or other lasers are typically developed to emit at a single given wavelength, researchers may find that a particular dye has peak excitation at, say 470nm, but they have to use a 488nm laser to excite it,’ he says, noting that diode and DPSS lasers are commonly available at 488nm – initially developed as replacements for the aging argon ion technology. ‘This approach is not optimal but, with a supercontinuum laser, the user can go straight to the optimum wavelength every time,’ says Clowes. Filters are used to remove parts of the spectrum that are not required.

Detecting

Microscopes of the past used the human eye as their de facto detector but, in many modern applications, such as fluorescence imaging, the light intensities involved are many orders of magnitude too small to be visible to the naked eye; detectors are used. Jim Owens, sales manager at Hamamatsu Photonics UK, describes their requirements: ‘Imaging cameras for life science research need to be high-quality, low-noise, and fast – and, typically, that’s been fairly expensive,’ he says, adding that the current standard in such applications is the costly electron multiplying CCD (EMCCD). Hamamatsu is introducing a new line of CMOS-based cameras, branded as scientific CMOS (SCMOS), which aim to be a cost-effective alternative to EMCCD detectors. ‘It’s the kind of technology that’s in a camera phone, but with much higher quality and much better signal-to-noise ratio. So, this is really something that’s going to be appropriate to life science imaging. It’s the same kind of technology as CMOS, but it’s really not the same thing,’ he says, adding that, when compared to a standard CMOS, the SCMOS is built to more rigorous manufacturing tolerances with associated electronics designed to minimise noise throughout. ‘If you’re taking a normal photo with your camera, you really don’t care about these things, but in these [life sciences] applications, users are sometimes trying to pick out individual photons. This year, we’re really going to see these [SCMOS] cameras move from the industrial area, where they’ve already penetrated, into applications such as high-grade microscopes. They’re going to be cheaper than EMCCDs, they’re going to be able to run fast while giving brilliant images, and they’ll do for most of the standard CCD imaging that people have done until now in the life sciences,’ says Owens, although he points out that there will still remain some applications for which EMCCD detectors are best-suited; EMCCDs will still be useful for applications requiring the highest sensitivities.

Everything in sequence

According to Owens, another significant advancement with important consequences for biophotonics is the technique of ‘time delay and integration’, or TDI. TDI is a method of reading the data off a CCD at a rate which is in sequence with some other external speed (such as the scanning speed, for example). ‘In TDI, the user either needs a very short exposure time, e.g. when things are moving very quickly, such as in a conveyer belt. We’re offering TDI now for high-throughput scanning, with single model sensitivity. This is a key enabling imaging technology, which is being integrated into second- and third-generation solutions for gene sequencing,’ he says, adding that Pacific Biosciences and Illumina are two companies active in this field.

At around a micrometre in length, mitochondria are difficult for optical microscopes to resolve. Here, a fluorescent marker compound has bound to certain sites on the mitochondrion, allowing its shape to be seen.

The science of genetics has seen much technological advancement over the past 20 years, initially as a result of the human genome project’s drive to sequence an entire copy of a complete genetic code. High-speed techniques now allow a human genome to be sequenced in the space of a couple of weeks. These techniques are based on laser excitation of fluorescent markers, each of which binds to a specific nucleotide in the sequence. There remains, however, a drive to reduce the cost of sequencing a complete genome, as Matthias Schulze, segment marketing director at Coherent, explains: ‘Improving speed and lowering the effective cost are key drivers in DNA sequencing, particularly in the push to implement the technology at the clinical level. Specifically, the “holy grail” target is a total cost of $1,000 to sequence a person’s entire genome, at which point widespread clinical adoption should take place. The present cost is approaching $20,000, so we are definitely getting much closer.’ Schulze adds that sequencing techniques are moving to higher-power lasers at more varied wavelengths in order to improve the speed. ‘A key metric is the length of DNA that can be sequenced in a single run, that is, before cross-talk errors build up. Originally, a single 488nm ion laser was used, but now three lasers are typically used in order to produce four well-separated and easily distinguishable fluorescent signals,’ he says, adding that the preferred laser wavelengths are 488nm, 532nm with the third being longer than 630nm.

According to Schulze, other laser requirements for faster, cheaper sequencing are low-noise, excellent mode quality for tight focusing in the interaction zone, and output power at 1W or more in order to produce a high signal-to-noise fluorescence signal. Furthermore, the lasers are often ‘buried’ within an OEM’s instrument, and so they must be reliable over their lifetime. Coherent offers DPSS and optically-pumped semiconductor lasers (OPSLs) to meet the requirements of 488nm and 532nm lasers, but the red (>630nm) wavelength had, until recently, only been obtainable by the use of diode lasers – which the company says are not reliable enough for integration into OEM systems. In response to this market demand, Coherent added a 1W, 639nm laser to its Genesis family of OPSLs, which the company is already shipping to life scientists.

An exclusive screening

In the life sciences, specificity is often key; specific markers may only fluoresce when bound to a particular protein, and only when excited by a particular wavelength of light. Markers are not the only way to achieve this specificity, however, as Chiraz Frydman, a product specialist at Horiba Scientific, explains: ‘Surface plasmon resonance, or SPR, allows users to follow a molecular interaction. This is a label-free technique, with no fluorescence, no radioactivity, and no colourimetry. We work with biochips,’ says Frydman. Biochips in this application are single substrates, typically pieces of silicon or glass the size of a microscope slide, onto which arrays of many microscopic test sites are added. ‘The biochips are functionalised by adding some chemicals to the surface, and after that we will add one molecule – one ligand [a specific biologically active molecule] – to the surface, and this ligand will be used to catch other molecules from the sample solution that is being analysed,’ she says. The ligands used can be highly specific to a particular chemical, or even to a particular kind of cell or protein. When a sample solution (blood, for example) is added to the biochip, the ligand’s target compounds, if present, will interact with it. Frydman explains that the interaction is observable on the surface of the chip as a change in its reflectivity.

By adding several hundred compounds to a single biochip and by monitoring the whole chip with a high sensitivity CCD array, the technique is capable of screening a sample solution for hundreds of compounds in one go. Frydman states that while other manufacturers have developed SPR for screening applications, no other developer has used a CCD array to offer such parallelism when screening.

The ultimate outcome of biochip technology could be a cheap, disposable test on every GP’s desk, leveraging photonics technology to offer an instant diagnosis for every ailment. Although this is a long way off, the pace of development demonstrated over the past two decades certainly suggests that anything is possible.