Can new optical techniques for in vivo imaging and flow cytometry of the microcirculation benefit sickle cell disease research?

Abstract

Intravital microscopy is a valuable tool for research into sickle cell disease with studies being carried out on transgenic mice and human volunteers. The method has helped to develop an explanation for sickle crises based on cell adhesion to the vascular endothelium followed by logjamming of rigid sickle cells and has stimulated much research into new treatments. In recent years there have been numerous new optical techniques developed for imaging the microcirculation and understanding the circulation of cells within the body, many of which have been further developed into in vivo flow cytometry techniques. This brief review highlights some of the progress made to date in the understanding of sickle cell disease using intravital microscopy. New techniques for imaging the microcirculation and their potential uses in understanding sickle cell disease are discussed. © 2011 International Society for Advancement of Cytometry

Sickle cell disease (SCD) is the most common genetic condition worldwide and it is estimated that 300,000 babies are born with the condition every year and that 5% of the global population (∼340 million) carry the genetic trait. Life expectancy is shortened, with studies reporting an average life expectancy of 42 and 48 years for males and females, respectively (1). SCD is caused by the inheritance of the sickle β globin gene (β^S) and is the homozygous state (SS) when this gene is inherited from both parents. The condition is characterized by blockages of the microcirculation, which cause painful episodes (crises).

Early work focused on the gelation of sickled red blood cells in vitro and is summarized by Eaton and Hofrichter (2). This also guided research into new treatments and although gelation plays a significant role in SCD it is not the only factor. Cell adhesion, inflammation, change in vessel diameter, and coagulation all play roles and are better understood using in vivo models. Intravital microscopy (IVM) has helped to propose an explanation for sickle crises based on cell adhesion to the vascular endothelium followed by logjamming of rigid sickle cells (3) and has stimulated much research into new antiadhesion therapies (4).

IVM is a valuable tool for research into SCD with studies being carried out on transgenic mice and human volunteers. However, in recent years there have been numerous and significant advances in the development of techniques for imaging the microcirculation and understanding the circulation of cells within the body. Many of these have been further developed to provide the capability for in vivo flow cytometry. These have been widely applied in fields, such as cancer research where it is important to monitor circulating white blood cells but they have not yet been adopted by the SCD research community.

The main aim of this brief review article is to highlight some of the new techniques developed for in vivo imaging and cytometry of the microcirculation and suggest how they could benefit SCD research. Some of the progress made to date in the understanding of SCD using IVM in both transgenic mice and human volunteers are highlighted in the following section. New optical techniques for imaging the microcirculation are briefly reviewed in the optical techniques for imaging the microcirculation section. Finally, their potential uses in understanding SCD are discussed.

In Vivo Blood Flow Measurements of SCD

The most widely used tool by researchers in the field of SCD is IVM. Although other optical techniques, such as laser Doppler flowmetry, have been applied to investigate the bulk properties of the microcirculation, the focus in this section is on IVM as it provides the capability to image at the cellular level and has the potential for in vivo flow cytometry. The first section summarises some key papers in the use of IVM in animal models, the second section discusses a smaller body of work relating to measurements on human volunteers.

Animal Studies

IVM has been a crucial tool for researchers in understanding SCD and in the development of new treatments. Barabino et al. (5), Kaul (6), and Hebbel (7) provide useful perspectives on current understanding and future challenges in SCD. Much of this work has focused on adhesion of sickle red blood cells to vascular endothelial cells as an explanation for the occurrence of vaso-occlusive crises. This was first demonstrated in early in vitro studies (8, 9) using static assays where it was shown that there was abnormal adhesion of sickle red blood cells to cultured human endothelium. Transgenic mice expressing human α and β^S globins have proved to be a more realistic model for investigating microvascular blood flow as they exhibit many properties similar to human microvasculature (10). As noted by Kalambur et al (10), transgenic mice exhibit the following useful properties; (i) red blood cells change shape when they become deoxygenated, (ii) adhesion of red blood cells, (iii) rolling and adhesion of white blood cells, (iv) stasis in post capillary venules, and (v) an inflammatory response similar to human SCD. These have all been shown to contribute to sickle crises.

Following early ex vivo work (11), an important in vivo imaging study in the transgenic mouse (3) proposed a two stage model where vaso-occlusion could be caused by adhesion of deformable sickled cells in post capillary venules followed by selective trapping of dense sickle cells. IVM has also been used to demonstrate that sickle cells can adhere to leukocytes (12, 13, 14).

Oxygen plays an important role in the vaso-occlusion process. For example, under hyperoxia transgenic sickle mice show an increase in red blood cell velocity and a decrease in arteriolar diameter, whereas in a control there is also a reduction in arteriolar diameter but a reduction in red blood cell velocity (15, 16). Vascular inflammation also has a role in vaso-occlusive crises. Kalambur et al (10) investigated the effects of hypoxia followed by reoxygenation on the flow and mean venule diameter. It was shown that the mean venule diameter decreased by 9% in the sickle mouse but remained constant in the normal mouse. The mean flow rate and wall shear rate decreased by 55% in the sickle mouse but not in the normal mouse. In addition the white blood cell count (measured by counting fluorescently labelled leukocytes) was higher at baseline for sickle mice than normal. After hypoxia and subsequent reoxygenation the white cell count increased by a factor of 3 in sickle but was unaffected in normal mice.

This fundamental understanding, in part enabled by IVM, has stimulated much research into the adhesion process and treatments that can reduce adhesion of sickle red blood cells to both endothelial cells and leukocytes. For example, IVM has been used to investigate the role of endothelial cell P-selectin in microvascular flow (17, 18). Sickle mice have increased expression of endothelial P-selectin and the velocity of red blood cells is slower, therefore, blocking P-selectin may help prevent crises. Recent work has demonstrated that a novel pan-selectin antagonist can be used to reverse vascular occlusions (19). Vascular occlusion has also been observed to reverse when intravenous immunoglobulin is given as this inhibits interactions between red and white blood cells (20). An (ex vivo) IVM study showed that antioxidants could reduce adhesion of sickle red blood cells (21).

Human Studies

The earliest study in human volunteers with SCD imaged moving red blood cells in 25 SCD patients using a ×10 microscope objective on the nailfold (22). When comparing the nailfold network topography, no significant difference in the patterns of the capillaries was observed between sickle and normal volunteers. However, in a resting state there was more intermittent flow for SCD patients. It has subsequently been shown that intermittent blood flow is variable in SCD patients (23, 24) and can also be observed in healthy volunteers (25). During a crisis, fewer static blood cells were observed and it was suggested that, as the capillaries are larger in the nailfold (typically 9 μm compared with 5 μm in the heart), cells may be sequestered in other organs during a crisis.

A series of papers by Cheung et al.(26–30) demonstrated a system for imaging conjunctival blood vessels in SCD patients. This builds on a number of largely qualitative early studies in this area (24, 31–33). The system was shown to be capable of qualitatively imaging and assessing properties, such as vessel damage and tortuosity and quantitatively measuring red blood cell velocity, total length of vessels in a region and average vessel diameter (26). A comparison of data from 18 SCD patients showed that it was possible to differentiate between patients with severe SCD, nonsevere SCD complications and during a crisis based on measurement of vessel diameter, vascularity and flow. In steady state, all patients exhibited decreased vascularity and red cell velocity compared with a non-SCD control set. During crisis there was a further decrease in vascularity and flow. In a recent study (28), a severity index of vasculopathology was used to compare 14 paediatric and 8 adult SCD patients and significant differences were observed mainly due to a lower number of abnormal shaped vessels and lower vessel tortuosity in the paediatric group. This cross sectional study suggests that it may be possible to monitor the severity of vasculopathology and provide an aid to treatment. However, longitudinal studies are necessary. Measurement of microcirculation flow using IVM has been shown to correlate with measurement of flow in the middle cerebral artery using transcranial Doppler ultrasound (27). The technique has also been shown to be capable of monitoring the effect of drugs on vaso-occlusion (29).

The reduction of blood flow of SCD patients observed in (26) is in contrast with the measurements of Lipowsky et al (22). To address this, a recent study obtained sublingual images (34) of 14 SCD patients during crisis using side stream dark field illumination (35) and showed that flow in the microcirculation is not disturbed between patients in crisis and healthy controls (patients acted as their own controls, a period of 4 weeks after the crisis). In addition, no difference in microcirculatory flow was observed between healthy volunteers and SCD patients. Lipowsky noted that the capillaries of the nailfold were larger than those in other organs, and therefore, this may not be an ideal site from which to infer systemic microcirculation. It may be that under the tongue is also not an appropriate site for imaging SCD, however, van Beers et al (34) point out that sublingual microcirculation is considered to be representative of systemic microcirculation in other conditions, such as sepsis.

This contradiction is yet to be resolved and more research is clearly required in the understanding of circulating sickle cells. In the following section new optical techniques that have potential to make a useful contribution to both animal and human SCD studies are described.

Optical Techniques for Imaging the Microcirculation

As demonstrated in the previous section, conventional IVM has been a valuable tool in SCD research. However, there are limitations with the approach such as being restricted to monitoring thin specimens in two dimensions. This section provides an overview (summarized in Table 1) of new techniques that could make a valuable contribution to SCD.

Table 1. Overview of optical techniques for imaging and in vivo flow cytometry of the microcirculation

Technique	Measurement	Capabilitya	Comments	References
CARS: coherent anti-stokes Raman scattering, (D)OCT: (Doppler) optical coherence tomography, MPM: multiphoton microscopy, SHG: second harmonic generation, FLIM: fluorescence lifetime imaging microscopy, OMAG: optical microangiography, PAM: photoacoustic microscopy, PT: photothermal, LASCA: laser speckle contrast analysis, HSI: hyperspectral imaging, RBC: red blood cell. a Resolution and penetration depth depend upon numerical aperture, wavelength and amount of scattering present. Typical values are presented
MPM	Emission of a single photon due to N photon absorption	3D imaging of endogenous fluorophores, e.g., NADH, tryptophanLateral resolution 0.5 μmAxial resolution 1 μmPenetration depth up to 1 mm	Low photobleaching and phototoxicityIn vivo flow cytometry possibleIn vivo tracking of leukocytesClinical systems available	36–40
SHG	Emission of a single photon due to N-photon anisotropic scattering	No exogenous fluorophoresLateral resolution 0.5 μmAxial resolution 1 μmPenetration depth up to 1 mm	Sensitivity to anisotropic tissue e.g. collagen	36
FLIM	Lifetime of fluorophore	Can separate multiple fluorophores within a sampleResolution dependent on imaging configuration, cellular level - 100s μm	Easily used in conjunction with other techniques, e.g., MPMSensitive to cellular environment e.g. viscosity, pH which play a role in sickling process	41
Raman	Inelastic scattering from molecular vibrations (molecule in ground state)	Can probe molecular structures Resolution, cellular level – 100 s μmLabel free	Sensitivity to in vivo cell hydration, gelation and capillary oxygen saturation should be useful in SCD Weak signals	42–45
CARS	Inelastic scattering from molecular vibrations (molecule in excited state)	Can probe molecular structures Resolution, cellular level Label freeLateral resolution 0.5 μmAxial resolution 1 μmPenetration depth 100μm	Much stronger signals than RamanLow photobleaching and phototoxicitySensitivity to lipid structures e.g., lipid bilayer around cellsProbe water within cells	46–48
OCT, DOCT, OMAG	Backscattered light that maintains coherence	Sensitive to scattering of light within coherence gate 3D imagingLateral resolution 10 μmAxial resolution 3-10 μm	3D imaging of vascular network3D flow imaging in microvessels	49–51
PAM	Acoustic emission after absorption of pulsed light	Contrast due to absorption by haemoglobinLateral resolution 1 μmAxial resolution depends on ultrasound pulse timing	In vivo flow cytometryImaging of single RBCsOxygen saturation of microvessels	52–58
PT	Infrared emission after absorption of pulsed light	Contrast due to absorption by haemoglobinResolution poorer than PAM lateral resolution approximately scales with depth, i.e., ∼100 μm at depth 100 μm	In vivo flow cytometryContrast enhanced by gold nanoparticlesPoorer resolution for blood vessel imaging than PAM	59–61, 57
Full field flow imaging (LASCA, Doppler)	Fluctuations in speckle pattern at detector due to motion	Contrast due to scattering of light by moving RBCsLateral resolution ∼10 s μm for superficial tissue	Limited to 2D imagingCan image flow in microcirculation in real timeSimple instrumentationImage interaction of interconnected flows in different microvessels	62–64
HSI	Absorption spectra of chromophores	Contrast due to oxy- and deoxyhaemoglobin absorption Lateral resolution ∼10 s μm for superficial tissue	Limited to 2D imagingSimple instrumentationOxygen saturation imaging	65, 66

IVM has benefited from the rapid development of nonlinear microscopy techniques. Useful reviews (36, 67) of IVM have highlighted many relevant papers in this area. Multiphoton microscopy involves the absorption of two (or more) photons by the specimen and subsequent emission of a single photon at a shorter wavelength. The process occurs in a small volume that affords high resolution imaging, reduced photo-toxicity, and photo-bleaching. Excitation at longer (usually near-infrared) wavelengths allows a higher penetration depth than conventional IVM. Second and third harmonic generation (SHG, THG) involves scattering of two or more incident photons and recombination into single photon without energy loss. The signals are produced predominantly in the forward scattering direction, however, there is usually sufficient detection because of backscattering of the light by tissue.

Multiphoton IVM has been used to track fluorescently labelled white blood cells in vivo (68). There have been several multiphoton studies using endogenous fluorophores, particularly NADH, NAD(P)H, riboflavins, and tryptophan (37). Li et al (40) have demonstrated imaging of leukocytes through the epidermis and dermis of mouse skin and have investigated the interaction of the leukocytes with the dermal vascular endothelium. This was achieved by exciting endogenous tryptophan using two-photon excitation at 590 nm. Two single frame excerpts from movies obtained from (38) are shown in Figure 1. It was shown that inflammation enhances leukocyte rolling, adhesion and tissue infiltration, properties which are significant in SCD (13, 14, 15). This could be achieved for shallower penetration depths using confocal microscopy (69). The penetration depth and use of endogenous fluorophores may eventually lead to monitoring leukocyte traffic in humans (38). Although nonlinear microscopy is often considered to be a complex tool for the research laboratory, portable commercial systems for clinical use are now available (39).

Figure 1. Single frame excerpts obtained from movies of leukocyte trafficking in skin vasculature (38) (a) Rolling leukocyte in normal BALB/c mouse skin (b) slow rolling and arrest of leukocyte in inflamed skin. The field of view is 200 μm. Reprinted from C. Li et al., ‘Imaging leukocyte trafficking in vivo with two-photon-excited endogenous tryptophan fluorescence.’ Optics Express 2010;18:988-999, with permission of The Optical Society.

Fluorescence lifetime imaging microscopy (FLIM) measures the time a molecule spends in an excited state and can be used to discriminate between multiple fluorophores. Measurement of the lifetime of exogenous and endogenous fluorophores within a cell strongly depends on the cellular environment, e.g., pH, viscosity, ion concentration (41). It can be used as a stand-alone technique or in combination with others, such as multiphoton microscopy.

Coherent anti-Stokes Raman scattering (CARS) (46–48) is a label free technique that utilizes two laser pulses (pump and Stokes) whose wavelengths are tuned to match the energy gap between two vibrational states of a molecule of interest. When this condition occurs, a strong anti-Stokes signal is generated, providing contrast to usually CH₂ bands and sensitivity to lipid rich structures, such as myelin and adipocytes. In SCD it may be useful as it can be used to image cell membrane and also water within cells (48). An IVM based on a fast scanning system combining CARS, fluorescence imaging, SHG, and backscattering provides multiple contrast mechanisms and has been used to monitor cell trafficking (70). Different fluorescent labels and one- and two-photon fluorescence were used to separate cancer cells, leukocytes, dermal lymphatic vessels, and blood vessels. Collagen was visualized using SHG. Myelin and a lipid rich serbaceous gland were imaged with CARS.

Although the signal yield is lower than CARS, in vivo Raman spectroscopy (i.e., inelastic scattering due to vibrations of a molecule in its ground state) can be used to uniquely identify molecules. It has been applied in IVM and has been used to monitor the oxygen saturation of microvessels (42, 43). Although conventional transillumination can usually only be applied to larger vessels or organs, it is demonstrated that this approach can be applied to thin tissues as well. Raman microspectroscopy has been applied to monitoring the hydration of cells (44). The application in this case was the monitoring of the abnormally high water content of cancer cells but such an approach could be applied to monitoring the hydration of sickle red blood cells as it is known that cell hydration plays a role in sickle crises (4).

For larger microvessels absorption spectroscopy approaches can be more conveniently applied. For example, hyperspectral imaging can be used to measure vascular oxygen saturation in two dimensions. Oxy- and deoxy-haemoglobin have different absorption spectra and so absorption of light is proportional to the level of oxygen saturation. Hyperspectral imaging involves taking multiple images across a wavelength range to obtain a data cube (65). The data is usually processed using an algorithm that takes into account the effect of light scattering, e.g., the modified Lambert-Beer law (66). The approach is popular in retinal imaging for monitoring the health of the eye for conditions, such as diabetic retinopathy and glaucoma. A typical blood oxygen saturation map obtained from the retina of a diabetic is shown in Figure 2.

Figure 2. Oxygen saturation map of a diabetic retina obtained using hyperspectral imaging (field of view 9.8 mm × 8.8 mm). Courtesy of Prof. Andy Harvey. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Blood flow is often measured by projecting a line of light across a vessel and measuring fluctuations as the blood cells scatter the light (71). Blood flows can also be measured in real time using full field using laser speckle (62) or full field laser Doppler imaging based on high frame rate CMOS sensors (63, 64). These full field techniques have previously been applied to larger microvessels containing multiple cells rather than single capillaries.

Photoacoustic tomography uses pulsed laser light to illuminate tissue. Absorption by haemoglobin or other contrast agents caused absorption and subsequent reradiation of the light as ultrasound because of thermoelastic expansion. The approach has been widely used to image blood vessels in three dimensions in relatively thick tissue (∼3 mm for photoacoustic microscopy) (52). Recent research has involved higher spatial resolution in weakly scattering samples, such as the exposed cortex of the mouse (53). Systems incorporating two or more wavelengths have been used to image the oxygenation of blood vessels (54) and single red blood cells (55). A typical blood oxygenation map obtained using photoacoustic microscopy is shown in Figure 3. Flow can also be obtained using photoacoustic Doppler broadening (56).

Figure 3. Photoacoustic microscopy image of nude mouse ear in vivo (a) structural image at 570 nm (b) Oxygen saturation map based on dual wavelength (570 nm and 578 nm) measurements. A1: representative arteriole, V1: representative venule (54). Reprinted from S. Hu et al., ‘Noninvasive label-free imaging of microhemodynamics by optical-resolution photoacoustic microscopy.’ Optics Express 2009;17:7688-7693, with permission of The Optical Society.[Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

Other methods can be used to image blood vessels in three dimensions. Photothermal imaging (59, 60) involves illumination with pulsed light and detection of infrared emission due to localized heating. Photothermal imaging can image blood vessels but may be better applied in SCD for discriminating between cells for in vivo flow cytometry, possibly in conjunction with nanoparticles (61). Optical coherence tomography (OCT) involves interfering scattered light from the sample with a reference beam. Only when the pathlengths travelled by the light in the sample and along the reference paths are matched within the coherence length of the light source will a signal be detected. Light scattered over longer pathlengths will be rejected allowing the effects of light scattering to be reduced. Variations of OCT developed to analyze flows include Doppler OCT (72), Doppler amplitude OCT (49), and optical microangiography (OMAG) (50). A typical OMAG image of the 3D microstructure of the brain cortex of a live mouse is shown in Figure 4 (51). OMAG provides (50) high spatial resolution imaging (9 μm lateral, 9.6 μm axial) of microcirculatory flows over a range of velocities from 4 μm/s to 23 mm/s. It should be noted that the 4 μm/s lower limit represents the best case as it is dependent on the depth and orientation of the vessel. OMAG has previously been applied to imaging the vascular networks of a mouse brain in 3D with the skull intact. Doppler OCT (providing flow velocity) and Doppler amplitude OCT (providing concentration of red blood cells) have been demonstrated to be capable of discriminating between red blood cells with different rigidity, an important characteristic in SCD (49).

Figure 4. In vivo OMAG image of a mouse cortex (A) 3D rendered microstructural image (2 mm × 2 mm × 2 mm) (B) x-y projection image of directional blood flow within scanned volume in (A). Directional blood flow in the meningeal layer (C) and the cortex layer (D) can be separated. Scale bar = 500 μm (51). Reprinted from R.K. Wang, ′Directional blood flow imaging in volumetric optical microangiography achieved by digital frequency modulation. Optics Letters 2008;33:1878-1880, with permission of The Optical Society. [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

Several of the techniques described here have been adapted to perform in vivo flow cytometry. For example in vivo flow cytometry (73, 74) has been developed based on similar instrumentation to that applied in conventional flow cytometry. A focused line of excitation light is projected across a blood vessel and fluorescently labelled cells that cross the line can be detected in different wavelength channels. The method has been used to characterize the kinetics of red and white blood cells circulating in the vasculature of the mouse ear and also used to monitor circulating prostate cancer cells. Figure 5 shows an example of cell counting with this technique. Figure 5a shows a combined still image of three frames from a movie. The blue, green, and red highlights show DiD-labelled rat prostate MLL cells in each of three consecutive frames as they traverse the vessel. Figure 5b shows typical bursts of fluorescent light received at the detector as labelled cells pass across the illumination line. Figure 5c shows a scatter plot showing the distribution in height and width of the detected fluorescence from a control (red, bottom left) and data acquired after injection of labelled cells (blue, top right). Figure 5d shows that the number of detected cell counts/min correlates well with the number of injected cells.

Figure 5. In vivo flow cytometry example of the ear of a rat injected with DiD labelled rat prostate MLL cells. (a) image showing cell motion in three consecutive frames (blue, green, and red show the same cells in consecutive images) (b) fluorescent bursts observed as labelled cells pass under the excitation line illumination (c) scatter plot showing the height and widths of the fluorescent bursts, the differences between control (red, bottom left) and after injection of labelled cells (blue, top right) can be observed (d) number of detected cells/min correlates well with number of injected cells. The field of view is 800 μm by 1200 μm with a pixel size of 1.47 μm. Adapted and reprinted with permission from the American Association for Cancer Research: Georgakoudi et al., In vivo flow cytometry: A new method for enumerating circulating cancer cells. Cancer Research 2004;64:5044-5047. [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

Multiphoton intravital flow cytometry has been used to quantify circulating tumor cells (40). It was noted that the multiphoton flow cytometer was moderately more sensitive in terms of cell detection than a confocal flow cytometer but significantly more sensitive than a conventional in vivo flow cytometer. A two channel multiphoton flow cytometer has been used to monitor two different populations of circulating cells (75). Fluorescently labelled, circulating red blood cells were also monitored for more than 2 weeks.

Both photoacoustic and photothermal techniques have also been applied to in vivo flow cytometry (57, 76, 77). A system that integrates photothermal, laser speckle, fluorescence, and high speed imaging for label-free transmission cytometry of circulating cells in the rat mesentery has been developed. It was shown that photothermal imaging was particularly sensitive to the absorption properties of red blood cells (76). An in vivo photoacoustic flow cytometer developed provided a sensitivity of 1 cancer cell in a background of 10⁷ normal cells (58).

Discussion

The techniques described in the previous section offer advantages over IVM and have the potential to benefit the understanding and treatment of SCD. Although not an exhaustive or definitive list, some of the advantages of the new techniques and suggestions for application to SCD are proposed in this section.

Imaging in three dimensions.

Many of the techniques described allow measurement of the tissue in three dimensions, e.g., multiphoton microscopy, OCT, CARS. For example in nonlinear techniques such as multiphoton microscopy the nonlinear effect can only be generated in a highly localized region which enables high resolution 3D imaging at the cellular level to be carried out. Microangiography and photoacoustic imaging allow the structure of the microvasculature to be imaged in three dimensions. To my knowledge the 3D structure of the microvasculature of SCD patients has not yet been studied.

Flow measurement.

Measurement of flow is clearly an important parameter in SCD research, and the development of new treatments that aim to reduce the effect of red cell adhesion to the endothelium. Quantitative two dimensional imaging of flows in superficial vessels can be achieved in full field using techniques, such as laser speckle imaging and full field laser Doppler imaging. In general, flow is studied in isolated capillaries in IVM and it may be of interest to study the interaction between the flows in interconnected microvessels in SCD. Quantitative Flows can be measured in 3D using microangiography and Doppler OCT although this is not possible at the single cell level because of the limited spatial resolution.

Label free imaging.

For many of the techniques described (e.g. multiphoton, CARS, FLIM, photoacoustic) it is not essential to use exogenous fluorophores which may affect the natural behavior of the cells within an animal model. Exogenous fluorophores may also preclude the use of the techniques on human volunteers or as clinical tools. Label free techniques therefore may provide a more realistic representation of flow and adhesion in SCD. Although many of the techniques described are often used as tools for the research laboratory there is evidence of their translation into clinical use (39).

In vivo flow cytometry.

Several of the techniques described (e.g., extension of in vitro flow cytometry (73, 74), photothermal (57), photoacoustic (58), and multiphoton (40) have been adapted to perform in vivo flow cytometry. This could allow the cell population of a SCD transgenic mouse to be monitored noninvasively for long periods and be a valuable tool in the development of new treatments. For example it could be used to monitor the effects of stem cell transfusion therapy (4) by monitoring the number of HbS cells. In cases where endogenous markers are present, in vivo flow cytometry could translate into a clinical tool. For example leukocyte rolling, adhesion, and tissue infiltration was demonstrated using a two photon system that excited endogenous tryptophan (38) and the authors suggest that human studies may be possible in future.

Functional imaging.

Many of the techniques applied offer the potential for complementary functional imaging and multimodal systems have been described (38, 76). For example multiphoton microscopy is sensitive to endogenous fluorophores, particularly NADH, NAD(P)H, riboflavins, and tryptophan. SHG sensitive is sensitive to anisotropic tissue such as collagen. Lipid structures can be easily detected using CARS and so could be used to study the role of the cell membrane and lysis in SCD. FLIM can be used in combination with these techniques. As an example, it is known that sickling of red cells can occur at high altitudes and in parts of the body where pH is low and dehydration contributes to an increase in blood viscosity (78). FLIM is capable of monitoring intracellular changes which are dependent on these factors (41). The oxygen saturation of the red blood cells plays an important role in the adhesion and gelation process in SCD. In larger microvessels this could be easily measured using hyperspectral imaging based on the absorption of oxy- and deoxyhaemoglobin (65). In capillaries it may be more appropriate to use photoacoustic (52) or Raman microspectroscopy techniques (42). Raman spectroscopy has been used to monitor heme aggregation and denaturation in red blood cells in vitro for conditions such as SCD (45). Further insight into this process in SCD may be obtained from in vivo Raman measurements. There is also potential for applying in vitro polarization sensitive measurements of polymerization (79) in vivo (80). Dehydration and the effects of red cell rehydration therapies (4) could be investigated using Raman microspectroscopy (44) or CARS (48).

Conclusions

Intravital microscopy is an important tool that has served the sickle cell disease research community well and has helped in the understanding of the sickling process. The understanding of the role of cell adhesion and subsequent logjamming of rigid SCD cells in post capillary venules has been enabled by this approach. Recent advances in in vivo optical imaging and in vivo flow cytometry have had a significant impact on other fields such as cancer research. The range of new optical techniques offers advantages in terms of 3D, label free, functional imaging with the capability of carrying out in vivo flow cytometry. If adopted, such techniques could have a significant impact on sickle cell disease research.

Acknowledgements

The author thanks Dr Martin Pickstone and members of the Sickle Cell Society (UK) for many interesting discussions. Also thanks Prof. Andy Harvey of Heriot-Watt University (UK) for providing a hyperspectral image of the retina and to the Optical Society of America for allowing reproduction of Figures 1, 3, and 4.