We demonstrated the feasibility of blood cell assisted in vivo Particle Image Velocimetry using confocal microscopy. Blood flow of skin vessel in a mouse was non-invasively imaged in vivo using a confocal microscopy. The video-rate confocal microscope was used to monitor the motion of the blood cells in the capillary of a live mouse ear. The home-built confocal laser scanning microscopy allowed us to take images at the acquisition rate of 30 frames per second. The individual blood cells could be distinguished from other cells and the trajectory of the each cell could be followed in the sequential images. The acquired confocal images were used to get the velocity profile of the in vivo blood flow in conjunction with the Particle Image Velocimetry (PIV), without injecting any exogenous nano/micro particles into the mouse. We were able to measure the blood velocity up to a few hundreds µm/sec for various vessels in a live mouse. Because there is no need for the injection of the exogenous tracing particles, it is expected that we could apply the current technology to the study of human capillary blood stream.
©2011 Optical Society of America
The blood is one of the most essential components in the human body especially for the circulation system that includes the blood and lymphatic vessels. The main components of blood include white blood cells, red blood cells, a liquid component called plasma and the platelets. The volumetric concentration of red blood cells is the highest among other components and the hematocrit (the portion of blood volume that is occupied by red blood cells) reaches 38-48%. Because of the high concentration, the biophysical behavior of red blood cells is one of the key parameters that decide the fluid dynamics of the blood flow. The blood flow is characterized as the solid-liquid phase flow, elastic flow and non-Newtonian flow . The rheological features, such as the aggregation and deformation of blood cells, determine the characteristics of the micro-vascular circulation in the arterioles, venules, and capillaries. It is well known that the micro-vascular flow is crucial to maintain the metabolic balance of the human body and the characteristics of blood flow could be used as the indicators for the circulation system diseases such as diabetics, stroke, and hypertension [2–6].
Because of the biophysical significance of the blood flow in the circulation system, a few technologies have been employed to measure the in vivo and in vitro velocity profile and the volume of blood flow. The ultrasonic technique has been used in the clinics and laboratories to measure the velocity field of the blood stream . The photometric slit method was proposed by Wayland in 1960s and it has been commonly used to measure the flow in conjunction with the intravital microcopy . Laser Doppler velocimetry (LDV) has been used to measure in vivo blood flow [9,10] and makes use of the Doppler shift of the spatial frequency for the reflected light which is determined by the blood velocity. Recently, Optical Doppler Tomography (ODT) was introduced to measure the blood velocity using the laser interference signal [11,12].
Along with the above mentioned technologies, Particle Image Velocimetry (PIV) has been used for the monitoring of the blood flow. The PIV is an optical method for the fluid visualization and PIV provide the instantaneous two dimensional velocity vector field [13,14]. The PIV is based on the tracer particles seeded into the fluid and assumed to follow the flow dynamics. A typical PIV system consists of a pulsed laser source and a CCD-based detector that can acquire two-dimensional images of tracing particles in a thin laser light sheet. The micro/nano scale particles and bio-inert liposomes are typically used as the tracing particles [15,16]. The velocity field of a cross-sectioned area can be calculated by estimating the displacement vector of groups of particles in two temporally consecutive images with a help of signal processing and autocorrelation or cross-correlation techniques. The displacement vector can be converted to a velocity vector using the physical distance of image pixel and time interval between two sequential images. Recently, due to the small geometry of MEMS devices especially in bio-technology field, the micro-PIV technique has been employed for the velocity measurement of the micro-scale fluid motion. The micro-PIV was implemented using a microscopic imaging system and the micro-PIV has been used for the velocity of the blood flow [17–19]. Sugii et al. and Lee et al. demonstrated that PIV measurement can be performed by employing blood cells as the tracing particles [20–22]. They measured the flow velocity in the micro-vessels using RBCs as their tracer without injecting exogenous tracing particles. However, it should be noted that their in vivo studies were limited to vessels in the highly transparent tissues that includes the mesentery and the chicken embryo. Their use of conventional wide-field microscopy for PIV is not adequate for sub-surface imaging of turbid media.
The Confocal Laser Scanning Microscopy (CLSM) was first introduced in 1950s in order to overcome the limitation of the wide field microscopy that could not acquire the sub-surface image of turbid samples [23,24]. The CLSM uses a laser source that is focused to a single point in the image plan and the back-scattered light is reflected back to the optical detector. By placing a confocal pinhole (which is located at the confocal point of the laser source) in front of the detector, the detector could exclusively collect the back-scattered light from the focal point of the laser source. The name ‘Confocal Microscopy’ stems from this geometrical configuration which uses two conjugated confocal points. The described confocal imaging scheme makes it possible to image features beneath the surface of a thick tissue sample. Because the CLSM is based on the point detection, either the laser source or the sample needs to be scanned for the 2D imaging.
Kenneth Kihm et al. first employed the confocal microscopy for the micro-PIV in the micro-channel . They imaged the fluorescence nano-particles in the flow using the confocal microscopy. By analyzing the sequence of raw confocal images, they were able to get the velocity profile of the flow at the different depth of the micro-channel. Compared to conventional PIV, the flow field measurement at various depths is the main advantage of the confocal microscopy-based PIV. By placing the image plane at a certain depth from the surface, the optically-sectioned sub-surface region can be imaged exclusively. This confocal-based optical sectioning allows us to measure the velocity field of a sub-surface region. Following Khim’s confocal PIV study, a few groups employed confocal microscopy as the imaging modality of PIV. The 3D measurement of the moving droplet was conducted with the confocal microscopy by Kaneda group . Lima et al. demonstrated the confocal microscopy based PIV for the blood flow in a micro-channel. They used the micron-size sphere as the tracing particle for the blood flow PIV .
Recently, our group demonstrated “blood cell assisted PIV using CLSM” for the flow in a micro-channel . We used the video-rate CLSM to observe the motion of blood cells inside the micro-channel. The acquired confocal images were used to perform PIV that estimates the velocity profile of the blood stream in the micro-channel. While the previous confocal microscopy PIV used the exogenous tracing particles, we employed the blood cells as the tracing particles in the reflection imaging mode.
In this study, we demonstrated “the blood cell assisted in vivo Particle Image Velocimetry using CLSM.” The previous blood cell assisted in vivo PIV measurements were limited to the vessels in transparent tissues because of the limitation of the conventional wide-field microscopy. We employed CLSM to image the in vivo blood stream of the mouse skin non-invasively. In vivo mouse skin was imaged from the skin surface to the upper dermis where the capillary blood vessels reside. The motion of the blood cells in the mouse skin was monitored and the velocity field of the in vivo blood stream was measured using the two frames cross correlation PIV method. The intrinsic blood cells were used as the tracing particles without injecting the exogenous nano/micro particles into the mouse. The feasibility of “the blood cell assisted in vivo confocal microscopy PIV” was tested for various capillaries in the live mouse ear.
2.1 Video-rate Confocal Laser scanning microscopy
The home-built Confocal Laser Scanning Microscopy was employed as the imaging apparatus to observe the optically sectioned image of the mouse skin (Fig. 1 ). The home-built CLSM is based on the confocal detection scheme that utilizes a single pinhole in front of the PhotoMultiplier Tube (PMT). The diameter of the pinhole was set to be 75µm. The CW Ti:Sapphire laser with the wavelength of 800 nm was used as the imaging light sources. The raster scanning was conducted using a polygon mirror that can produce as many as 17,000 line scans per seconds. The slow axis scanning was implemented with a galvanometer mirror with 33Hz repetition rate. Upon the irradiation of the laser source on the sample, the back-scattered light from the sample was collected by the objective lens and delivered to the PMT. We used a water-immersing 40X objective lens with the NA of 1.2. The axial resolution of the CLSM can be defined as the axial position at which the detected intensity drops as the half of the focal plane. The measured axial resolution was about 4µm; i.e. the CLSM is optically sectioning the image plane with about 4µm thickness. The analog signal from the PMT was fed into an A/D converter and the digitally converted signal was used to acquire the snapshot and the movie clip with the home-written GUI software. More details on the home-built CLSM can be found in the literature .
2.2 Animal preparation
BALB/c mice 6weeks old or younger were used for imaging after the administration of 1.2% solution of Tribromoethanol. The mouse was placed on the 3-Dimenstional stage and the ear of the mouse was flattened on a 100m thick cover glass using a 2% methocel gel.
2.3 Particle image velocimetry
We have employed the two-frame cross-correlation method for the Particle Image Velocimetry. The underlying idea of the two-frame cross-correlation is described in Fig. 2 . The acquired raw image is divided into sub-areas called “the interrogation window.” The interrogation windows from two consecutive image frames are cross-correlated with each other, pixel by pixel. The cross-correlation function produces the signal peak that identifies the common displacement between two frames. The velocity for this window can be calculated by dividing the common displacement with the time delay between two frames. A velocity field over the whole image is obtained by repeating the cross-correlation for each interrogation window over the two consecutive image frames. Details of the PIV and the cross-correlation method are described in the literature . The size of the interrogation window is one of the critical factors that should be adjusted in accordance with the desired spatial resolution and flow velocity. We used the interrogation windows ranging from 25x25 pixels to 40x40 pixels and 150 image pairs were used for the ensemble averaging. We used the SNR (Signal to Noise Ratio) validation and about 5% of data were removed in the validation step.
3.1 CLSM imaging of in vivo mouse skin
We observed the upper layer of the mouse skin in vivo using the CLSM which operates at the imaging rate of 30 frames per second. The ear of the mouse was place on the cover glass and the skin morphology at various depths was observed. The acquired images of the mouse skin are presented in Fig. 3 . All images were taken using the reflection mode of the CLSM, i.e. the back scattered light from the skin is used to make the cross-sectional images. The field of view was set to be 200µm by 200µm. The acquired images clearly demonstrate the different layers of the upper mouse skin. The stratum corneum, the outmost layer of the epidermis, is presented in Fig. 3(a). The non-viable stratum corneum cells do not contain a nucleus and they have the polygonal shape. This layer produces the strongest back-scattered signal due to the mismatch of the refractive index between the air and the stratum corneum. The hairs of the mouse (bright thin lines) were also observed on this imaging plane. Viable epidermal cells were observed beneath the stratum corneum (Fig. 3(b)). The cytoplasm and nucleus of the cells appear dark while the border between cells looks brighter. In the upper dermal region, we were able to image various features including the fibrous structures, hair follicles and fat cells (Fig. 3(c) and 3(d)). The fibrous matrix is clearly visualized as the bright mesh-like structure. The hair follicles are observed along with the fibrous matrix in the dermis. We were able to image fat cells in the depth ranging from 70µm to 100µm.
3.2 CLSM imaging of blood cells in the capillary of the live mouse ear
Along with various components in the dermis, the blood vessel and blood cells are clearly imaged with the CLSM in the reflection mode (Fig. 4 ). The contrast of the blood flow is provided by the strong back-scattering at the blood cells’ membrane. The blood cells appeared bright due to the strong back-scattering while the blood vessel and the plasma components appeared dark. Since the blood cell images were taken in the reflection mode without tagging exogenous fluorescence particles to the blood cells, the type of the cell was not distinguishable from the acquired images.
The continuous images of the blood flow were taken by fixing the image plane at a certain depth from the skin surface. The typical sequential images are presented in Fig. 4. and the time delay between each image is 132 msec. In the sequential images, the individual cells could be distinguishable and the dynamic behavior of the cells could be monitored. Some cells (marked with dark arrow) remained in the image plane while they move from one side to other side of the vessel. However, some cells (marked with open arrow) move out of the image plane and disappear from the field of view. The cells in the middle of the vessel travel faster than cells near the wall as expected. The presented images were taken at the image plane placed near the upper wall of blood vessel and the width of the imaged vessel is about 70 µm. As the imaging plane gets closer to the center of the blood vessel, the cell image gets blurry due to the reduced intensity of the laser source and the back scattered light.
3.3 Blood Cell assisted in vivo Particle Image Velocimetry
We imaged the dynamic motion of the blood cell in the live mouse ear using the video-rate CLSM. The velocity profile of the vessel presented in Fig. 4 was calculated with the PIV and presented in Fig. 5 . Figure 5(a) presents the confocal image and Fig. 5(b) shows the stream wise velocity vector field and contour plots. Figure 5(c) shows the cross-sectional velocity profiles near the center of the imaged vessel (marked as the dashed line). As predicted from the sequential snap-shot images, the flow velocity near the center of the vessel shows the maximum velocity of about 40µm/sec and the flow decreases away from the center. The velocity profile presents a parabolic pattern which is the characteristic feature of a laminar flow at the low speed regime.
The blood flow in the vessel with a small diameter was observed using the CLSM and the velocity profile was calculated using the PIV (Fig. 6 ). The center of the blood vessel was located about 50 µm down from the skin surface and the diameter of the vessel was approximately 30 µm. In contrast to the large blood vessel, the motion of blood cells on the central cross-section (in terms of the z axis that moves from the top of the skin to the dermal side) could be imaged and the velocity field could be calculated for the smaller vessel.
The maximum velocity of the blood flow in Fig. 5 and 6 is around a few tens of micrometers per second. We observed blood vessels where the flow velocity was higher than 100 µm/sec. The CLSM image of such blood vessel and the corresponding PIV result is presented in Fig. 7 . The maximum velocity of the center stream measures as high as 134 µm/sec.
The blood flow in the branched vessels was imaged and the corresponding velocity profile was presented in Fig. 8 . The blood flow moves from left to the right of the image and the vessel on the right side is branched out into two vessels. The PIV data shows that the flow to the lower vessel gets suddenly reduced near the branch out point. This is because the flow to the lower vessel is partially blocked at the branch out point and the reservoir-like flow is formed near the beginning part of the vessel. We were able to confirm the reduced flow speed near the beginning part of the lower vessel using the raw image of the CLSM.
4. Discussion and conclusion
In order to observe the in vivo blood stream of the mouse skin, we employed the confocal laser scanning microscopy which allows us to image the sub-surface skin structure non-invasively. The home-built confocal laser scanning microscopy has the sectioning depth of 4µm and the frame rate of 33 frames/second. The individual blood cells could be distinguished from other cells and the trajectory of the each cell could be followed in the sequential images. This observation indicates that the frame rate of the employed CLSM is fast enough to trace the cells in the raw reflection images. The reflection images of blood cells were used to calculate the velocity profile of the blood stream using the micro-Particle Image Velocimetry. The two-frame cross-correlation method was employed in order to evaluate the shift of the particles between frames. The effect of the out-of-focus blood cells increases with the thickness of the optical sectioning. Because the measured thickness of the optical sectioning is equivalent to the dimension of a single red blood cell, we expect the influence of the out-of-focus blood cells to be minimal in the acquired images.
The blood stream of various vessels was imaged and the corresponding velocity profile was determined using the PIV. In case of the small vessels with a diameter of about 30 µm, we could image the movement of the cells in the central cross-sectional plane (in term of the z axis) of the blood vessel (Fig. 6). However, the imaging depth was limited to the upper plane for the vessels with larger diameter (>50µm). The limited imaging depth of the blood flow can contribute to the absorption and scattering of the blood constitutions. We expect that the longer wavelength for the imaging laser source is beneficial for the deeper imaging because of the reduced absorption and scattering by the red blood cells. The maximum velocity near the center of vessel reaches a few hundred µm/sec and the velocity near the wall is about tens of µm/sec. The upper limit of the measurable velocity of the current technique is determined by the imaging speed i.e. the frame rate. We could follow the faster moving cells by enhancing the frame rate with a faster mechanical scanner. It is expected that we can measure the blood stream as fast as 1mm/sec with a polygon mirror that has more facets than the current system. The blood flow in the branched vessels was imaged and the corresponding velocity field was determined using the PIV. This result demonstrates that the current technique can be applied for various applications that include the complex blood vessel networks.
The velocity profile of all measured vessels presented a smooth stream line that follows the geometry of the vessel. We can observe that there exist some irregular velocity patterns with a relatively small disturbance compared to the main stream lines. These irregular velocity patterns are attributed to a couple of factors that includes the irregularity of the vessels and the three dimensional motion of the blood cells. The blood vessels are different from artificial pipes or channels, because they have non-homogenous wall surface and an irregular geometry with the curvy or abruptly changing paths. These irregularities can induce locally randomized velocity patterns that deviate from the smooth main stream. The motion of the blood cell in the vessel has three dimensional features such as the tumbling, the deformation of the cell shape, and the travelling through the different z-planes. These three-dimensional motions of the blood cell can be another cause of the locally randomized pattern.
Because of the significance of the blood flow, various non-invasive technologies have been employed to measure the in vivo and in vitro velocity profile of the blood stream. Among various technologies, the ultrasonic Doppler flow meter and, the Optical Doppler Tomography (ODT), Laser Doppler Velocimetry (LDV) are popularly tested and used in the field. These apparatus measure the velocity based on the Doppler shift of the applying frequency determined by the velocity of the blood cell. Compared to above mentioned technologies, the “blood cell assisted in vivo confocal microscopy PIV” has unique features. While the measurable depth of PIV is limited by the imaging depth of the CLSM, the spatial resolution of PIV is superior to the other technologies. The blood cell assisted in vivo confocal microscopy PIV calculates the velocity of the blood stream using the sequential microscopic images of the blood cells. Because of the high spatial resolution of the CLSM, the current PIV technique can be applied for the velocity measurement with the single cellular level accuracy.
In conclusion, we examined the feasibility of “blood cell assisted in vivo PIV using confocal microscope.” The confocal microscope allowed us to image the movement of the blood cells in the live mouse skin. The sequential images of blood cells were used to estimate the in vivo blood velocity of the mouse capillaries based on the PIV technique. The current technique can be applied for the study of blood stream in human skin and retina capillary because (1) it employs a non-invasive optical imaging and (2) the intrinsic blood cells are used as the tracing particles for the PIV measurement. It is expected that we can apply the current technique to monitor the risk of circulatory diseases such as diabetics, hypertension and stroke.
This research was financially supported by the Ministry of Education, Science Technology (MEST), Korea Institute for Advancement of Technology (KIAT) through the Human Resource Training Project for Regional Innovation and the Korea Research Foundation (KRF) Grant funded by the Korea Government (MEST) (No. 20090069912).
References and links
1. D. N. Ku, “Blood flow in arteries,” Annu. Rev. Fluid Mech. 29(1), 399–434 (1997). [CrossRef]
3. E. Fossum, A. Høieggen, A. Moan, G. Nordby, T. L. Velund, and S. E. Kjeldsen, “Whole blood viscosity, blood pressure and cardiovascular risk factors in healthy blood donors,” Blood Press. 6(3), 161–165 (1997). [CrossRef] [PubMed]
4. A. J. Lee, P. I. Mowbray, G. D. Lowe, A. Rumley, F. G. Fowkes, and P. L. Allan, “Blood viscosity and elevated carotid intima-media thickness in men and women: the Edinburgh Artery Study,” Circulation 97(15), 1467–1473 (1998). [PubMed]
5. K. Toth, G. Kesmarky, J. Vekasi, J. Nemes, L. Czopf, P. Kapronczay, R. Halmosi, E. Papp, and I. Juricskay, “Hemorheological and hemodynamic parameters in patients with essential hypertension,” Clin. Hemorheol. Microcirc. 21(3-4), 209–216 (1999).
6. K. L. Resch, E. Ernst, A. Matrai, and H. F. Paulsen, “Fibrinogen and viscosity as risk factors for subsequent cardiovascular events in stroke survivors,” Ann. Intern. Med. 117(5), 371–375 (1992). [PubMed]
7. J. R. Haywood, R. A. Shaffer, C. Fastenow, G. D. Fink, and M. J. Brody, “Regional blood flow measurement with pulsed Doppler flowmeter in conscious rat,” Am. J. Physiol. 241(2), H273–H278 (1981). [PubMed]
8. H. Wayland and P. C. Johnson, “Erythrocyte velocity measurement in microvessels by a two-slit photometric method,” J. Appl. Physiol. 22(2), 333–337 (1967). [PubMed]
10. J. Seki, Y. Sasaki, T. Oyama, and J. Yamamoto, “Fiber-optic laser-Doppler anemometer microscope applied to the cerebral microcirculation in rats,” Biorheology 33(6), 463–470 (1996). [CrossRef] [PubMed]
11. Y. Zhao, Z. Chen, C. Saxer, S. Xiang, J. F. de Boer, and J. S. Nelson, “Phase-resolved optical coherence tomography and optical Doppler tomography for imaging blood flow in human skin with fast scanning speed and high velocity sensitivity,” Opt. Lett. 25(2), 114–116 (2000). [CrossRef]
12. Z. Chen, T. E. Milner, S. Srinivas, X. Wang, A. Malekafzali, M. J. van Gemert, and J. S. Nelson, “Noninvasive imaging of in vivo blood flow velocity using optical Doppler tomography,” Opt. Lett. 22(14), 1119–1121 (1997). [CrossRef] [PubMed]
14. R. J. Adrian, “Particle-imaging techniques for experimental fluid mechanics,” Annu. Rev. Fluid Mech. 23(1), 261–304 (1991). [CrossRef]
15. C. Poelma, P. Vennemann, R. Lindken, and J. Westerweel, “In vivo blood flow and wall shear stress measurements in the vitelline network,” Exp. Fluids 45(4), 703–713 (2008). [CrossRef]
16. P. Vennemann, K. T. Kiger, R. Lindken, B. C. Groenendijk, S. Stekelenburg-de Vos, T. L. ten Hagen, N. T. Ursem, R. E. Poelmann, J. Westerweel, and B. P. Hierck, “In vivo micro particle image velocimetry measurements of blood-plasma in the embryonic avian heart,” J. Biomech. 39(7), 1191–1200 (2006). [CrossRef]
17. C. D. Meinhart, S. T. Wereley, and J. G. Santiago, “PIV measurements of a microchannel flow,” Exp. Fluids 27(5), 414–419 (1999). [CrossRef]
18. M. R. Bown, J. M. MacInnes, and R. W. K. Allen, “Micro-PIV simulation and measurement in complex microchannel geometries,” Meas. Sci. Technol. 16(3), 619–626 (2005). [CrossRef]
19. R. Okuda, Y. Sugii, and K. Okamoto, “Velocity measurement of blood flow in a microtube using micro PIV system,” in Proceedings of PSFVIP-4, (Chamonix. France. Jun. 2003), pp. 1–7.
20. Y. Sugii, S. Nishio, and K. Okamoto, “In vivo PIV measurement of red blood cell velocity field in microvessels considering mesentery motion,” Physiol. Meas. 23(2), 403–416 (2002). [CrossRef] [PubMed]
21. A. Nakano, Y. Sugii, M. Minamiyama, and H. Niimi, “Measurement of red cell velocity in microvessels using particle image velocimetry (PIV),” Clin. Hemorheol. Microcirc. 29(3-4), 445–455 (2003).
23. M. Minsky, “Microscopy apparatus,” US patent 3013467 (1961).
24. T. Wilson, and T. Confocal Microscopy, Wilson, ed. (Academic Press, 1990).
25. J. S. Park, C. K. Choi, and K. D. Kihm, “Optically sliced micro-PIV using confocal laser scanning microscopy (CLSM),” Exp. Fluids 37(1), 105–119 (2004). [CrossRef]
26. H. Kinoshita, S. Kaneda, T. Fujii, and M. Oshima, “Three-dimensional measurement and visualization of internal flow of a moving droplet using confocal micro-PIV,” Lab Chip 7(3), 338–346 (2007). [CrossRef] [PubMed]
27. R. Lima, S. Wada, M. Takeda, K. Tsubota, and T. Yamaguchi, “In vitro confocal micro-PIV measurements of blood flow in a square microchannel: the effect of the haematocrit on instantaneous velocity profiles,” J. Biomech. 40(12), 2752–2757 (2007). [CrossRef] [PubMed]
28. W. H. Kim, C. I. Kim, S. W. Lee, S. H. Lim, C. W. Park, H. Lee, and M. K. Park, “Particle image velocimetry of the blood flow in a micro-tube using the confocal laser scanning microscope,” J. Opt. Soc. Korea 14(1), 42–48 (2010). [CrossRef]
29. I. Veilleux, J. A. Spencer, D. P. Biss, D. Côté, and C. P. Lin, “In vivo cell tracking with multimodal video rate microscopy,” IEEE J. Sel. Top. Quantum Electron. 14(1), 10–18 (2008). [CrossRef]
30. R. D. Keane and R. J. Adrian, “Theory of cross-correlation analysis of PIV images,” Appl. Sci. Res. 49(3), 191–215 (1992). [CrossRef]