Near infrared lasers in flow cytometry

Abstract

Technology development in flow cytometry has closely tracked laser technology, the light source that flow cytometers almost exclusively use to excite fluorescent probes. The original flow cytometers from the 1970s and 1980s used large water-cooled lasers to produce only one or two laser lines at a time. Modern cytometers can take advantage of the revolution in solid state laser technology to use almost any laser wavelength ranging from the ultraviolet to the near infrared. Commercial cytometers can now be equipped with many small solid state lasers, providing almost any wavelength needed for cellular analysis.

Flow cytometers are now equipped to analyze 20 or more fluorescent probes simultaneously, requiring multiple laser wavelengths. Instrument developers are now trying to increase this number by designing fluorescent probes that can be excited by laser wavelength at the “edges” of the visible light range, in the near ultraviolet and near-infrared region. A variety of fluorescent probes have been developed that excite with violet and long wavelength ultraviolet light; however, the near-infrared range (660–800 nm) has yet seen only exploitation in flow cytometry. Fortunately, near-infrared laser diodes and other solid state laser technologies appropriate for flow cytometry have been in existence for some time, and can be readily incorporated into flow cytometers to accelerate fluorescent probe development. The near infrared region represents one of the last “frontiers” to maximize the number of fluorescent probes that can be analyzed by flow cytometry. In addition, near infrared fluorescent probes used in biomedical tracking and imaging could also be employed for flow cytometry with the correct laser wavelengths. This review describes the available technology, including lasers, fluorescent probes and detector technology optimal for near infrared signal detection.

1. Introduction.

Flow cytometry has become a critical and nearly ubiquitous technology in the field of cell biology, particularly in the biomedical sciences. Flow cytometers permit the analysis of very large numbers of single cells using lasers to excite cell-associated fluorescent probes and filter and detector combinations to detect the often low levels of fluorescence associated with these markers. The original flow cytometers from over forty years ago could only detect one or two fluorescent markers using a single laser source [1]. Modern cytometers now employ multiple laser sources simultaneous to excite a vast array of fluorescent probes, with the simultaneous detection of fifteen or more extracellular or intracellular proteins becoming routine [2]. The wide variety of laser wavelengths now available from small diode-pumped solid state (DPSS) or direct laser diodes means that virtually any biological fluorescent fluorescent probe with a visible light excitation and emission spectra can be analyzed by flow cytometry, if the correct laser wavelength is available. Commercial cytometer manufacturers have taken full advantage of modern laser technology to provide a wide variety of laser wavelengths for single cell analysis [3,4].

The earliest flow cytometers developed in the 1950s and 1960s required an intense light source both to measure cellular light scattering, and to excite the first fluorescent probes. The first cytometers uses mercury arc lamps, capable of producing much brighter and spectrally distinct light that that available from other lamp sources. However, development of the earliest flow cytometers coincided fortuitously with the development of the first lasers, which also become a practical technology in the 1960s. Most of the earliest cytometers used water-cooled argon and krypton ion lasers as their primary light sources [3,4]. The traditional blue-green 488 nm laser line produced by argon-ion sources was and remains (via newer solid state sources) the primary laser line for most flow cytometric analysis. The 488 nm line excites fluorescein, a low molecular weight fluorochrome that forms the molecular basis for a wide variety of fluorescent cell tags and physiological probes. The cyanobacterial energy exchange protein phycoerythrin (PE) and its tandem dyes, including PE-Texas Red, PE-Cy5 and PE-Cy7, are also excited at 488 nm, and could be distinguished spectrally from fluorescein using bandpass filters, dichroic mirrors and electronic compensation of spectral overlap. These technologies allowed multicolor analysis of multiple cell surface markers. By combining fluorescein, PE and the PE tandem dyes, up to six cell markers could be analyzed simultaneously.

While the blue-green 488 nm line was the primary excitation wavelength used for flow cytometry, the other laser lines produced by argon-ion and krypton-ion lasers were also employed [3,4]. Gas lasers can produce ultraviolet, violet, blue, green, yellow and red laser lines, and all were used to excite specialized fluorochromes in the early days of flow cytometry. The ultraviolet lines produced by argon and krypton lasers could excite dyes like the DNA binding probes DAPI and Hoechst 33342 and 33258, and the green and yellow lines from krypton sources could excite rhodamine and sulforhodamine based fluorescent probes [5]. The long red 641 and 647 nm lines produced by krypton-ion lasers were used to excite newly available red-excited phycobiliproteins like allophycocyanin (APC), and the monomeric cyanin dyes Cy5 and Cy7. These dyes could also be excited with helium-neon (HeNe) lasers emitting at 633 nm, which were air-cooled and easier to use [6,7]. However, the large size, high cost and heavy maintenance requirements of these water-cooled laser sources made cytometry with more than one laser source difficult.

This changed with the advent of laser diodes in the late 1980s. These small, air-cooled and relatively inexpensive lasers allowed the easy incorporation of additional lasers into cytometers already equipped with a primary blue-green source [8,9]. The first laser diodes emitted in the near-infrared and long red range, from approximately 660 nm to above 1200 nm. Laser diodes emitting at approximately 640 nm were introduced into flow cytometers in the early 1990s, as replacements for krypton-ion and HeNe sources for the excitation of APC and Cy5 [9]. Small HeNe lasers continued to be used in flow cytometers for some time, but red laser diodes became the dominant second laser source and remain so today. A red laser diode can excite APC, its tandem dyes (APC-Cy5.5 and APC-Cy7) as well as other low molecular weight dyes like Alexa Fluor 700, allowing up to three fluorescent probes to be added to the probes excited at 488 nm. Adding this second laser source allowed eight to nine fluorescent markers to be analyzed simultaneously.

Violet laser diodes, based on different semiconductor chemistry than red and near infrared diodes, were the next laser type to see wide usage in flow cytometry. These modules excite in the 395 to 415 nm range, just within the visible spectrum [10]. Originally developed in the mid-1990s, they were first incorporated into flow cytometers in the late 1990s by Shapiro and colleagues [11]. The water-cooled krypton-ion lasers used early in flow cytometry could also produce violet lines at 407, 413 and 415 nm, motivating the development of several fluorescent probes that could be excited in this range. Cascade Blue and Pacific Blue were two coumarin-based low molecular weight fluorescent probes that excited in this range and emitted in the blue 450 to 480 nm range. The incorporation of small violet laser diodes into cytometers as a third source revived the use of these dyes, and spurred the development of additional fluorescent probes that could take advantage of this laser wavelength [12,13]. Quantum dots or nanocrystals were developed in a form that could be conjugated to antibodies; a series of quantum dots ranging in emission from 525 to 800 nm (Qdot 525, 585, 605, 655, 705 and 800) were initially used with violet laser excitation [14]. More recently, the Brilliant Violet dyes (Sirigen) have been developed, a group of polymer based fluorescent probes that can be “built” with specific excitation and emission characteristics. The Brilliant Violet series (including BV412, BV510, BV570, BV605, BV650, BV705 and BV787) can be easily conjugated to antibodies, and can be used simultaneously to add seven additional fluorescent parameters to the probes already excited by blue-green and red lasers [15]. A cytometer now equipped with blue-green, red and violet lasers could now excite up to sixteen fluorescent probes simultaneously, a substantial number. Many commercial cytometers are now equipped with these three laser wavelengths.

Other laser wavelengths from modern solid state lasers have also been incorporated into flow cytometry. Green and yellow lasers, including 532, 552 and 561 nm, are now common fixtures on cytometers. These laser wavelengths provide more efficient excitation of phycoerythrin and its tandems, and allow better excitation of red fluorescent proteins like DSred and mCherry, which are poorly excited at 488 nm [16,17,18]. However, these useful wavelengths, while giving improved excitation of many fluorescent probes, have not increased the total number of simultaneous parameters available for flow cytometry. With red and violet laser used to their maximum utility, the search was on to look for additional excitation wavelengths and fluorescent probes that could add to the existing array of simultaneous markers. Ultraviolet (UV) lasers have been used in flow cytometry (albeit on a limited basis) since its inception, being originally generated by argon and krypton sources. They were primarily used to excite DNA binding dyes like DAPI and the Hoechst dyes, and to excite the calcium flux probe indo-1 [5]. UV excited molecules suitable for conjugation to antibodies for cell surface labeling, including the coumarin dyes, were not bright, and had to contend with considerable amounts of autofluorescence generated by UV excitation [4]. The primary UV excited fluorochrome aminomethylcoumarin (AMCA) has seen little use in flow cytometry for this reason. This situation changed with the development of the Brilliant Ultraviolet (BUV) series, similar in principle and structure to the Brilliant Ultraviolet dyes. These probes excite well at 355 nm, and emit at uniform wavelengths from the violet to the red. At the time of this writing, BUV395, BUV496 and BUV737 were available. These probes are not well excited by violet laser excitation, and conversely the Brilliant Violet dyes are not well excited by UV. Their emission wavelengths have been staggered with the BV dyes, permitting even better spectral separation. Like the BV dyes, the BUV dyes will add an additional group of parameters for high-dimensional flow cytometry for cytometers equipped with UV lasers. UV lasers remain expensive fixtures on cytometers, but the recent development of small solid state units will make their incorporation into flow cytometers much more feasible.

At this point, the near UV, violet and most visible wavelengths have been extensively “tapped” as excitation sources for flow cytometry, both for the excitation of novel fluorochromes, and to increase the total number of simultaneous fluorochromes that can be used for high-dimensional cell analysis. The near-infrared (NIR) laser range is the only remaining area of the spectrum that has not been extensively investigated for flow cytometry. NIR laser diodes based on InGaP, AlGaInP and GaAs semiconductor chemistry were some of the earliest laser diodes developed (Figure 1), so laser sources in this range applicable for flow cytometry (ranging from 660 nm to 800 nm) are available. A growing number of fluorescent probes are also available; Cy5.5, Cy7, Cy7.5 and their derivatized counterparts Alexa Fluor 680, 700, 750 and 790 are available in chemical forms appropriate for protein conjugation [19,20]. One reason for this lack of work has been the dynamic wavelength range of photomultiplier tubes (PMTs), the primary detector technology in modern flow cytometers. Sensitivity in bialkali photocathode PMTs starts to decrease above 750 nm, and even more sensitive multialkali PMTs start to lose sensitivity above 800 nm. However, the existing array of fluorescent probes in this range emit at 820 nm or shorter, and sensitivity in this range is still sufficient for cell surface marker detection. The optics of most commercial cytometers are also compatible with shorter wavelength NIR sources (660 to 730 nm), and several manufacturers in fact offer lasers in this range as options. Since the NIR laser range is probably the “final frontier” for adding additional fluorochromes to our already large polychromatic arrays, their use in flow cytometry should be investigated and defined. In this review, we will describe the integration of NIR laser sources into flow cytometers, and investigate their utility in exciting the increasing number of NIR fluorochromes currently available. We will also discuss the use of other detection technologies with improved sensitivity in the NIR range, such as avalanche photodiodes.

Figure 1.

The wavelength range of laser diodes, and their semiconductor chemistry. Inset, a Power Technologies 705 nm laser diode in the IQ10 package. All NIR lasers in this study were of this form factor.

2. Methods.

2.1. NIR lasers.

In this study we used a series of laser direct diodes extending into the NIR range, including (1) a 660 nm module emitting at at a maximum power level of 110 mW, (b) a 685 nm module emitting at 40 mW, (c) a 705 nm module emitting at 40 mW, and a 730 nm module emitting at 40 mW. An example of these single wavelength lasers is shown in Figure 1 (inset). All laser modules were manufactured by Power Technologies, Inc. (Alexander, AK, USA). All lasers were of the Power Technologies IQ (instrument quality) cylindrical configuration with built-in Peltier temperature control, and equipped with adjustable anamorphic optics to modify the beam spot. All beams spots were elliptical with 1 mm W × 0.7 mm H dimensions at 1 meter. For comparison purposes, all lasers were attenuated to 20 mW power output immediately prior to the flow cell of the cytometer.

2.2. Visible red lasers.

Most multilaser cytometers are equipped with a red laser source, either a HeNe or solid state unit. For this study a HeNe 633 nm (JDS Uniphase) emitting at 20 mW was used for comparison to a visible red laser source. In addition, a recent trend in red laser usage has been toward “shorter” red sources, in the 620 to 630 nm range. Since red laser diodes on many commercial cytometers emit in the 640–645 nm range, their wavelengths fall very close to emission range of fluorescent probes such as APC, giving the potential for inadvertent laser light “contamination of the fluorochrome detector and resultant high background noise. Shorter wavelength red lasers (so-called HeNe replacements) can minimize this problem, and are starting to be available in forms compatible with flow cytometers. A “short” red 620 nm fiber laser source (MPB Communications, Quebec, Canada) was also used as a red laser comparison source. As with the NIR sources above, power level of both lasers was maintained and/or attenuated to 20 mW.

2.3. Flow cytometry.

A LSR II flow cytometer (BD Biosciences, San Jose, CA) was used in these studies. The LSR II is a quartz cuvette based cytometer equipped with a blue-green 488 nm primary laser source and Hamamatsu R3896 multialkali photomultiplier tubes with specifications for good long red sensitivity up to 800 nm. The NIR lasers described above were mounted on the instrument in the same position, with the same beam guidance and focusing optics used for each laser source. Laser power was verified immediately prior to the flow cell, and was set to 20 mW for all tested laser sources. The same light collection optics, fiber optics, PMTs, filters and dichroics were used for each NIR laser source. The only exception to this was in experiments where the red HeNe 633 nm laser and a NIR source were used simultaneously, where distinct detector clusters were employed. All bandpass filters were manufactured by Semrock (Rochester, NY), with the specifications given in the Results.

2.4. Fluorochromes and cells.

The following red and NIR excited fluorescent probes were used in this study: Alexa Fluor 647 (AF647), Alexa Fluor 660 (AF660), Alexa Fluor 700 (AF700), Alexa Fluor 750 (AF750) and Alexa Fluor 790 (AF790)[20]. All fluorochromes were produced by Molecular Probes Life Technologies, Thermo-Fisher. The excitation and emission spectra for these probes is shown in Figure 2. For some experiments, an EL4 mouse lymphoma cell line was labeled with biotin-conjugated anti-mouse CD44 (BD Biosciences Pharmingen, Torrey Pines, CA, USA) followed by washing and subsequent labeling with a streptavidin conjugate of the above fluorochromes. The resulting labeled cells were then analyzed by flow cytometry. In some experiments, cells were substituted with antibody capture microspheres, latex beads conjugated with an anti-kappa chain antibody. The biotin conjugate above was incubated with an aliquot of beads, followed by washing and labeling with the streptavidin fluorochrome conjugate. In both cases, the resulting cell and bead suspensions gave well-labeled samples for evaluation of fluorochrome excitation and detection.

Figure 2.

Excitation and emission spectra for Alexa Fluor 647, 660, 700, 750 and 790 nm. Grey line, excitation spectra; black line, emission spectra. Extinction coefficient (EC), quantum yield (QY) and fluorescence lifetime (t) are listed to the right. The quantum yield and extinction coefficient of Alexa Fluopr 790 have not been determined.

2.5. Data analysis.

All data was acquired using FACSDiVa data acquisition software (BD Biosciences) and analyzed using FlowJo version 7.6.5 for PC (FlowJo LLC). Degree of fluorochrome fluorescence was expressed as a staining index (SI), using the median and slope distributions of labeled and background cell or bead populations to calculate a value proportional to the level of fluorochrome fluorescence, as described previously [18]. Compensation was calculated in FlowJo using the package’s automated spillover matrix algorithm.

3. Results.

3.1. Diode laser installation and alignment.

Many commercial cytometers are equipped with a red laser, either a red HeNe emitting at 633 nm (now less common), or a red laser diode emitting between 635 and 645 nm, with an average emission value above 640 nm. The lasers used in this study were direct or GaInP and AlGaInP diodes, emitting at 660 nm, 685 nm, 705 nm and 730 nm (shown in Figure 1). This represented a good typical distribution of wavelengths in this color range. Direct laser diodes are relatively economical compared to diode pumped solid state (DPSS) lasers and are quite suitable for flow cytometry. The 730 nm was the highest single wavelength laser tested; laser diodes also emit at 780 and 785 nm (GaAlAs diodes) but are more complex and expensive.

Alignment of NIR lasers requires an alignment microsphere that can be excited by NIR light. Conventional alignment particles for visible lasers may not work well for this application; even microspheres that can be excited by visible red light may not excite well at NIR wavelengths. Blue Beads from Polysciences, Inc. (Warrington, PA, USA) are often used for red laser alignment and can be used to align the 660, 685 and 705 nm lasers. They are detected in the 650 to 800 nm range. The AccuDrop microspheres used for drop delay estimation on BD Biosciences cell sorters are designed to be excited by a NIR laser, and can work very well for NIR laser alignment. They are similarly detected in the 650 to 800 nm range.

3.2. Laser safety.

The NIR lasers used in this study are Class IIIb laser sources and should be handled and shielded appropriately. Laser light above 685 nm is not easily visible to the human eye, and should be used to particular caution as the beam cannot be directly visualized. Laser eyewear certified for IR use often blocks emission of considerably longer wavelengths than the NIR lasers illustrated here; be certain to check the specifications on any eyewear used to ensure that it protects in the 600 – 730 nm range. An infrared viewing card, coated with an IR sensitive phosphor, can be used to visualize NIR beam paths at lower laser power. An infrared viewer can also be used to visualize NIR beam paths and spots.

3.3. Fluorochromes.

A series of Alexa Fluor fluorescent probes (developed by Molecular Probes Life Technologies, now a Thermo Fisher company) with red and NIR excitation/emission characteristics were evaluated for their ability to be excited by these NIR laser sources. Their spectra are shown in Figure 2. Alexa Fluor 647 (λ_EX = 650 nm, λ_EM = 668 nm) is a red-excited fluorochrome with excitation/emission characteristics similar to the cyanin dye Cy5, and is commonly used in traditional cytometry using a red HeNe or laser diode for excitation. Alexa Fluor 660 (λ_EX = 663 nm, λ_EM = 690 nm) sees much less use in flow cytometry, as it ideally requires a somewhat longer laser source. Alexa Fluor 700 (λ_EX = 702 nm, λ_EM = 723 nm) also benefits from a longer laser source, but is somewhat excited at shorter red wavelengths, allowing it to be paired with APC and APC-Cy7 for three color detection off a red laser. It has similar spectral characteristics to the earlier cyanin dye Cy5.5 Alexa Fluor 750 (λ_EX = 749 nm, λ_EM = 775) and Alexa Fluor 790 (λ_EX = 782 nm, λ_EM = 805 nm) are not commonly used in flow cytometry due to their long emission maxima, although Alexa Fluor 750 and the previously developed cyanin dye Cy7 are often paired with APC as tandem dyes.

EL4 mouse thymoma cells labeled with each of these dyes were analyzed using the four NIR lasers described above, using a wide variety of bandpass filters with both wide and narrow detection windows . They were compared to a traditional red HeNe 633 nm laser source, and a “short” red 620 nm fiber laser. Power levels, optical paths, collection fibers, dichroics, bandpass filters and PMTs were the same for all lasers tested. The results are shown for each probe in Figure 3.

Figure 3.

Fluorescence intensity of Alexa Fluor 647 (a), 660 (b), 700 (c), 750 (d) and 800 nm (e) with varying laser wavelengths and detection filters. Laser wavelength (in nanometers) is plotted on the x-axis, and bandpass filter wavelength and window width in nanometers is plotted on the y-axis. Fluorescence intensity on the z-axis is expressed as a staining index (SI), a relative value proportional to sample fluorescence versus background [18].

3.4. Alexa Fluor 647.

The results for Alexa Fluor 647 are shown in Figure 3a. As expected, both short red 620 nm and HeNe 633 nm lasers provided excellent excitation of this probe. The emission range was in the expected 660 to 710 nm range based on the excitation filters used, but emission did extend well into the 720 to 730 nm range as well, suggesting that Alexa Fluor 647 will require considerable compensation for spectral overlap when used with adjacent fluorochromes. Interestingly, the 660 nm laser also gave considerable excitation in this longer wavelength range.

3.5. Alexa Fluor 660.

Alexa Fluor 660 has seen much less usage in flow cytometry. The results for this probe are shown in Figure 3b. Interestingly, the shorter 620 and 633 nm laser wavelengths also efficiently excited this probe, suggesting that it and Alexa Fluor 647 could be used together. The longer 660 and 685 nm NIR laser diodes also provided good excitation, albeit at longer emission wavelengths.

3.6. Alexa Fluor 700.

Alexa Fluor 700 is known to be somewhat excited at shorter red wavelengths, and is often used in conjunction with allophycocyanin (APC) and the tandem conjugate dye APC-Cy7 as a third immunolabeling fluorochrome for flow cytometry. However, the 685 and 705 nm lasers provided far better excitation than the shorter red sources (Figure 3c). Emission extended well into the 750 nm range and beyond, indicating that it would likely require significant levels of compensation when used with the longer Alexa Fluor probes.

3.7. Alexa Fluor 750.

As expected, Alexa Fluor 750 also required 685, 705 or 730 nm excitation to provide detectable fluorescence; the shorter wavelength sources produced almost no detectable signal (Figure 3d). Emission was well into the 800 nm range for all laser wavelengths.

3.8. Alexa Fluor 790.

This probe spectrally resembles the earlier cyanin dye Cy7.5, and sees little use in flow cytometry, either alone or as a tandem dye acceptor molecule. It has spectral properties to other NIR dyes like indocyanin green, and is used for imaging techniques requiring NIR probes. Like Alexa Fluor 750, it is well-excited by the longer NIR lasers and shows little excitation by the shorter modules. Interestingly the 730 nm module only produced adequate excitation for fluorescence detected beyond the 800 nm point (Figure 3e).

3.9. Initial conclusions.

The objective of these descriptive studies was to predict whether NIR lasers and probes can be added to the existing array of fluorescent probes, thereby increasing the total number of fluorescent markers that can be analyzed on a multi-laser flow cytometer. The above excitation/emission information suggested several possibilities. (1) Alexa Fluor 660 could be paired with Alexa Fluor 647 or another red excited probe (i.e. Cy5, Alexa Fluor 633 or APC) instead of the more usual Alexa Fluor 700. Alexa Fluor 660 is far better excited with a red laser than Alexa Fluor 700 is. (2) A longer wavelength NIR laser would be better suited to excite Alexa Fluor 700, which could be paired with Alexa Fluor 750 or Alexa Fluor 790, excited with the same laser. This could be used in combination with a normal red exciting Alexa Fluor 647 or APC, Alexa Fluor 660 and the APC tandem dye. This arrangement would permit excitation of five fluorochromes with a red and a NIR laser, instead of the usual three using a red laser alone.

3.10. Combining Alexa Fluor 647, 660, 700 and 750.

To test these proposed configurations, a BD LSR II was configured with both 633 and 685 nm lasers, aligned to separate detector clusters. The 685 nm laser was chosen because it gave good excitation of Alexa Fluor 700, 750 and 790, and its emission avoided both the APC/Alexa Fluor 647 and the Alexa Fluor 700 emission ranges.

First, EL4 cells labeled with Alexa Fluor 647 and Alexa Fluor 660 or Alexa Fluor 700 were analyzed using a single red HeNe laser. Although Alexa Fluor 660 has a closer emission to Alexa Fluor 647 than Alexa Fluor 700, the signals were still easily distinguishable after compensation; less than 15% compensation was required to separate Alexa Fluor 647 and 660, compared to less than 2% for Alexa Fluor 700 (Figure 4a and b). The 685 nm laser also allowed the addition of Alexa Fluor 700, with better excitation than the red HeNe (Figure 4c). The level of compensation was high (140%), but still acceptable for a digital system. Alexa Fluor 750 or 790 could then be added as a fourth fluorochrome with excitation by the 685 nm laser (Figure 5a and b), still with acceptable levels of compensation with Alexa Fluor 660 and 700. Spectrally, Alexa Fluor 750 and 790 are very similar; while Alexa Fluor 790 has slightly less spectral overlap into Alexa Fluor 700, the difference with Alexa Fluor 750 was slight (Figure 5b). It should be observed that some signal spreading occurred after compensation, as would be expected from fluorescent probes with close spectral properties. As the required number of simultaneous fluorochromes increases, the use of probe combinations with close emission spectra is required; however, this overlap needs to be carefully assessed when planning a high-dimensional protocol. The strategy of using these problematic probes for cellular markers with clearly positive or negative populations (as opposed to cellular proteins requiring measurement based on mean fluorescence intensity) may need to be employed here.

Figure 4.

a, Simultaneous Alexa Fluor 647 and 660 nm analysis using a single HeNe laser. The filter specifications, laser configuration and compensation matrix are shown. b, Simultaneous Alexa Fluor 647 and 700 nm analysis using a single HeNe laser, also with laser configuration and compensation matrix. c, Simultaneous analysis of Alexa Fluor 647, 660 and 700 using dual 633 and 685 nm lasers, also with filter information, laser configuration and compensation matrix indicated. Cells labeled with each fluorochrome are indicated by the indicated scatterplot colors.

Figure 5.

a, Simultaneous analysis of Alexa Fluor 647,660, 700 and 750, using the indicated filters, laser configuration and compensation matrix. Cells labeled with each fluorochrome are indicated by the indicated scatterplot colors. b, Compensation matrix for simultaneous analysis of Alexa Fluor 647, 660, 700 and 790 (replacing Alexa Fluor 750).

3.11. Addition of Alexa Fluor 700 and 750 or 790 to APC and APC-Cy7 labeling.

While the low molecular weight NIR fluorochromes can be combined, most antibody conjugates are available using the phycobiliprotein APC and its tandem APC-Cy7. The most practical use of these fluorochromes would therefore be as additions to this pair. Alexa Fluor 700 is already commonly used as a third intermediate fluorochrome with APC and APC-Cy7, but as shown in Figure 4 it is not optimally excite by the typical red lasers. In Figure 6, Alexa Fluor 700 is added to two labeling schemes with APC and APC-Cy7, with excitation using either the same HeNe red laser, or with the spatially separated 685 nm module. Both methods work and give acceptable levels of spectral overlap; however, excitation of Alexa Fluor with the NIR laser increases sensitivity. The addition of the 685 nm laser module also allows the addition of either Alexa Fluor 750 or 790. APC-Cy7 and both Alexa Fluor 750 and 790 appeared to be compatible with one another when using spatially separated lasers, despite their similar emission spectra. The spectral overlap and population spreading observed when using spectrally close probes is noted here as well, again requiring careful marker selection when designing high-dimensional panels. Although not tested here due to limitations in the number of PMTs per laser on our test instrument, the addition of Alexa Fluor 660 with excitation by the HeNe red laser should also be possible, albeit with increased spectral overlap.

Figure 6.

a, Simultaneous analysis of APC, Alexa Fluor 700, APC-Cy7 and Alexa Fluor 750, with Alexa Fluor 700 excitation by the red HeNe laser. The filter specifications, laser configuration and compensation matrix are shown. b, Simultaneous analysis of APC, Alexa Fluor 700, APC-Cy7 and Alexa Fluor 750, with Alexa Fluor 700 excitation by the NIR 685 nm laser diode, with filters, laser configuration and compensation matrix also shown. Cells labeled with each fluorochrome are indicated by the indicated scatterplot colors.

4. Discussion

These results indicate that NIR laser diodes can be readily incorporated into a commercial flow cytometer with minimal optical modification, and that these lasers can be used to excite NIR fluorochromes not normally accessible to the more conventional laser sources found on most instruments. Multiple laser spanning the visible spectrum are becoming the norm on flow cytometers, and laser wavelengths that previously saw little use in cytometry (such as the violet and ultraviolet) have been profitably exploited to increase the number of fluorescent probes that can be analyzed simultaneously. The NIR region is probably the “last frontier” of the visible spectrum for subsequent additions to the fluorochrome array for polychromatic flow cytometry.

The fluorochrome combinations tested here are spectrally compatible, although there is overlap and both a requirement for spillover compensation and some observed population spreading. Increasing laser power might overcome some of this overlap by increasing the signal-to-noise of the NIR excited probes. The laser modules used in this study are at roughly the maximum power available for a single mode laser for these wavelengths. Higher power modules are possible but at a considerably higher cost. Diode-pumped solid state NIR lasers are also available in this range at higher power levels, but are also more expensive. Although not specifically tested with the Alexa Fluor NIR probes, increased power levels with the shorter wavelength Alexa Fluor dyes have generally not yielded improved signal-to-noise ratios; these probes saturate at relatively low power levels. However, the effect of higher power levels on these long wavelength probes remains to be tested.

The number of fluorescent probes that can be excited in this region is expanding as well. The Alexa Fluor dyes demonstrated here are only a sampling of the NIR probes available based on the cyanin (Cy dyes). The DyLight 650, 755 and 800 probe series (Pierce Thermo Fisher), the IRDye 650, 680, 700, 750 and 900 series (Li-Cor, Lincoln, NE, USA), the VivoTag 645, 680, 750 and 800 series (Perkin-Elmer)[21], the CF series from Biotium and the TRACY dyes from Sigma-Aldrich are only a sampling of dyes available in this spectral range. Almost all of these dyes were developed for applications other than cytometry, particularly in vivo imaging. The NIR range is preferred for imaging applications where tissue penetration and low background is crucial for sensitivity. However, there is no reason these probes cannot be retasked for cytometric applications. Many of them are available in formats appropriate for immediate labeling (such as straptavidin conjugates) or reactive forms for antibody conjugation. While tissue penetration is generally not an issue for flow cytometry, decreased cellular autofluorescence and improved signal to noise ratios would be a considerable benefit.

In addition to the need for additions to the array of fluorochromes for simultaneous immunolabeling, there are a number of NIR fluorescent physiological probes and tagging reagents that might have utility for flow cytometry. Indocyanin green is a widely used in vivo tagging agent using in hepatic and other clinical studies. With an excitation maxima close to 780 nm and an emission beyond 800 nm, it cannot be detected using conventional cytometers. Dyes like indocyanin green, the NIR heptamethine dye IR-780 iodide [22], IRDye800CW [23], cyanine dye IR783 [24] and other NIR tracking agents typically used for imaging could be accessible to flow cytometry as well. There are also a significant number of recently reported probes with novel structures, again aimed at imaging but with potential for cytometry. Including a series of tetraarylazadipyrromethene NIR fluorescent dyes [25] and the rhodamine-derived dyes Si-pyronine, Si-rhodamine, Te-rhodamine, and Changsha NIR dyes [26]. Development of probes in this spectral region is therefore vigorous and ongoing.

One technical issue with NIR probe detection is the reliance of flow cytometers on PMTs as detectors. The dynamic range of PMTs starts to decline at about 800 nm; modern PMTs are somewhat more sensitive in this range, but remain essentially non-sensitive by 900 nm. This is not a major issue for the fluorochromes discussed here, with emission maxima no greater than 820 nm. However, any attempt to extend the wavelength range of emission detection will encounter this barrier. Fortunately, other semiconductor-based detectors are now available that can overcome this problem; avalanche photodiodes (APDs) are now found in many detection devices, and can efficiently detect light well into the NIR. APDs have been investigated as detectors for long red signal detection in flow cytometers [27], and have been incorporated into commercial cytometer designs, notably the BD Biosciences FACSArray. These detectors are much more robust than PMTs and are far more economical as well.

Abbreviations

AF647

Alexa Fluor 647

AF660

Alexa Fluor 660

AF700

Alexa Fluor 700

AF750

Alexa Fluor 750

AF790

Alexa Fluor 790

APC

allophycocyanin

APD

avalanche photodiode

Cy

cyanin

HeNe

helium-neon

NIR

near infrared

PE

phycoerythrin