Structural and Hemodynamic Analysis of the Mouse Retinal Microcirculation

Abstract

purpose. In the holangiotic retina, little is known about the connections between and the circulation within microvessel layers. The goal of the present study was to explore the three-dimensional arrangement and hemodynamics of mouse retinal microvessels.

methods. Confocal microscopy was performed on fluorescein dextran-filled retinal flatmounts. Capillary velocity in the deep layer was measured by epifluorescence intravital microscopy. The changes in the studied parameters after branch retinal vein occlusion were evaluated.

results. The superficial and intermediate layers are both asymmetric crossroads for capillary blood flow, with approximately 70% of the capillary connections directing the flow from the arterioles into the deep layer. The venous flow from the deep layer joins the major veins in the superficial layer through transverse venules, indicating that major veins are directly connected to the deep layer. Red and white blood cell velocities ± SD in the deep layer were 1.26 ± 0.34 and 0.8 ± 0.32 mm/sec respectively. After branch vein occlusion, venule dilation and decreased velocity were observed in the deep layer.

conclusions. In the mouse retina, a tridimensional model of retinal microcirculation was established, showing that most microvessel connections on the arteriolar side direct the flow from the superficial to the deep layer, and vice versa on the venular side. However, the presence of direct arteriovenous connections in the superficial layer and the longer vessel length in the deep layer offer the possibility of actively modulating intraretinal flow. Compared with other capillary beds, both the capillary velocity and microhematocrit are high, a situation that favors nutrient delivery to the inner retina.

In the holangiotic retina, there are, schematically speaking, three layers of microvessels: a superficial layer in the ganglion cell layer, an intermediate layer in the inner plexiform layer, and a deep layer in the outer plexiform layer. ^{1 2} The microvessels have been described as predominantly arteriolar in the superficial layer and predominantly venular in the deep layer. ^{2 3 4 5} However, the three-dimensional microvascular arrangement has not been fully determined, which limits the interpretation of the data obtained by techniques of blood flow examination such as scanning laser ophthalmoscopy (SLO) ^{6 7 8} and laser Doppler flowmetry. ⁹ The establishment of a complete three-dimensional model of the microcirculation and the selective hemodynamic analysis of a given microvascular layer is therefore of interest. In the present study, we used confocal microscopy to explore the three-dimensional angioarchitecture of the mouse retina and intravital microscopy to measure the capillary velocity. 

Materials and Methods

Male C57 BLKS mice were purchased from Janvier (Saint-Ile le Genest, France). All manipulations were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Confocal Microscopy

Flatmounted fluorescein dextran-filled retinal vessels were prepared according to a modified version of a previously reported method. ¹⁰ Plasma was labeled in eight anesthetized mice by intracavernous injection of 0.2 mL of fluorescein isothiocyanate (FITC)-dextran (molecular weight: 2000 kDa, 30 mg in 1 mL of PBS). One minute after injection, mice were killed by an overdose of pentobarbital, and the eyes were immediately enucleated. The globes were placed for 10 minutes in 4% paraformaldehyde. The retinas were then flatmounted with gelatin-glycerol and examined with a confocal microscope (500 MRC; Bio-Rad, Hercules, CA) equipped with an argon laser. In each field examined, successive scans were performed through the thickness of the retina, with a ×10 or ×40 lens. The step in the z-axis varied from 0.5 to 5 μm. To visualize the vertical capillaries joining one layer to another, reduction of the pinhole and the use of the ×40 lens to obtain maximum confocality allowed the elimination of the background noise from adjacent layers. 

For image analysis, images were converted to the TIFF format on computer (Graphic Converter software; Lemke Software, Peine, Germany). The definition of each image was 768 × 512 pixels. The vascular density of each layer of microvessels was calculated on ×10 images, using a method derived from Robison et al. ¹¹ Briefly, each layer was analyzed as a two-dimensional image. Vessels were manually outlined on computer (PhotoShop, ver. 4.0; (Adobe Corp., Mountain View, CA). NIH 1.62 software allowed the drawings to be converted into their central axes one pixel wide (skeletonization) and also enabled the number of residual pixels to be counted (NIH Image is available by ftp at zippy.nimh.nih.gov/ or at http://rsb.info.nih.gov/nih-image; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD). This number was proportional to the length of the blood vessels. The corresponding surface was measured in square pixels. 

Intravital Microscopy

Red blood cells (RBCs) were labeled with FITC according to the method of Sarelius and Duling. ¹² Briefly, arterial blood from a donor mouse was centrifuged, and RBCs were isolated and washed in PBS (pH 7.4) containing 100 mg/L EDTA. RBCs were incubated for 2 hours with FITC at pH 8.0, and washed with PBS-EDTA (pH 7.4). Hematocrit was adjusted to 20%. 

Intravital microscopy was performed with a modified epifluorescence microscope (Leica, Heidelberg, Germany), as described elsewhere. ¹³ Anesthesia was induced by intraperitoneal injection of 100 mg/kg ketamine and 25 mg/kg xylazine (both from Sigma-Aldrich, Lyon, France). Body temperature was kept at 37°C throughout the experiment by a homeothermic blanket control unit (Harvard Apparatus, Les Ulis, France). Topical tropicamide (CibaVision, Blagnac, France) was administered for pupil dilation. Twenty microliters of the FITC-labeled red cell preparation and 30 μL of filtered rhodamine 6G (1 mg/mL, Sigma-Aldrich) ¹⁴ was injected into the corpus cavernosum, after which the mouse was left to rest for approximately 5 minutes. The mouse was then placed on its side under the objective. The head was supported so that the iris was perpendicular to the illumination axis. After administration of topical oxybuprocaine (CibaVision), a coverslip was applied on a ring surrounding the globe. Methylcellulose (Goniosol; Allergan, Mougins, France) was applied as contact medium. Care was taken not to exert pressure on the globe and to keep the coverslip horizontal. Epi-illumination was delivered by a mercury lamp through appropriate dichroic filters. Focus was adjusted with the ×10 lens (screen magnification, ×960). In the nasal retina, a capillary path in the deep layer was chosen for analysis. Capillary pathways were easily distinguished from postcapillary venules, on the basis of the presence of a tortuous pathway, a single file of circulating cells, and relatively lower velocity. Switching the filter of the microscope allowed concomitant visualization of RBCs and white blood cells (WBCs). Images of labeled cells were recorded with a digital camera (model 4912-5000; Cohu, Inc., San Diego, CA) and an S-VHS video recorder (Sony, Tokyo, Japan) at a frame rate of 25 images per second. A video timer signal was superimposed on the images. Retinal illumination lasted less than 10 seconds for each series. At the end of the examination, while the mouse was still under the microscope, 100 μL of 10% sodium fluorescein was injected subcutaneously to record the angiogram. Epifluorescence examination of the blood smears obtained after labeled cell injection indicated that the respective emission spectra of FITC-labeled red cells and rhodamine-labeled white cells were not superimposed (data not shown). 

For motion analysis and velocity measurement, the videotapes were reviewed, either at normal speed or in slow motion, to delineate the pathways of the cells. Because velocity was not synchronized with the cardiac cycle, the values measured were averaged over several cardiac cycles. On successive video frames, the center of each cell was marked on a transparent sheet, and its velocity was calculated by measuring the linear distance between its successive positions. Viewed in a noninterlaced mode, the frame rate was 50 images per second. Between 100 and 200 velocity measurements were performed in each mouse. The RBC flux in a single vessel was calculated from the number of labeled RBCs passing along a given capillary path per second, corrected by the proportion of labeled cells among all the RBCs, as measured on blood smears. The theoretical microvessel hematocrit was then calculated according to the following formula ¹²  

\[\frac{F_{\mathrm{m}}\ {\times}\ \mathrm{MCV}}{V_{\mathrm{m}}\ {\times}\ {\Pi}\ {\times}\ (\mathrm{D}/2)^{2}}\]

where F _m is the mean flux of RBCs; V _m, the mean velocity of RBCs; D is the mean diameter of deep capillaries; and MCV, the mean corpuscular volume (48 fl; Paques, personal data, 2000). Wilcoxon’s test was used to compare red and white cell velocities. The magnification factor of the lens and cornea (mean 1.6) was calculated by comparing the length of the optic nerve head circumference in vivo and on retinal flatmounts. 

Branch Retinal Vein Occlusion

The effects of branch retinal vein occlusion (BRVO) on hemodynamic and structural parameters upstream of the occlusion site were evaluated. In anesthetized mice, the fundus was visualized with a slit-lamp through a noncontact lens (Superpupil Lens; Volk, Mentor, OH). The occlusion site was placed on a nasal vein, two to three disc diameters from the disc. Two to six argon laser impacts (power 0.1 mW, duration 1 second, spot size 50 μm; Alcon Crystal focus, Fort Worth, TX) were sufficient to obtain occlusion, which was due to compression of the vein by the surrounding retinal edema. No dye enhancement was used. Mice in which retinal detachment or intravitreal hemorrhage developed were excluded. Twelve hours after photocoagulation, each eye underwent intravital microscopy and then confocal microscopy in the retinal area upstream of the occlusion site. 

Results

Confocal Microscopy

In the following description, the direction parallel to the vitreoretinal interface will be referred to as horizontal and the perpendicular direction as vertical. Superficial signifies close to the vitreous side, and deep, close to the photoreceptor side. Major arteries and veins are the vessels directly arising from the optic nerve in a radial pattern. On confocal microscopy images, the three horizontal microvessel layers were clearly delineated, as well as the vertical connecting capillaries, without background noise from adjacent layers (Fig. 1) . In the superficial layer, arterioles were identified by their numerous dichotomous branchings in the horizontal plane before abruptly changing direction to perpendicular. The relative paucity of arterioles in the vicinity of the arteries gave the superficial layer, but not the intermediate or deep layer, the aspect of a capillary-free zone. Each vertical capillary between the superficial and intermediate layers stemmed from the arterioles, except in areas very close to major veins. The intermediate layer consisted of short capillary segments in the horizontal plane that again abruptly changed to a vertical direction running toward the deep layer. The deep layer was largely anastomotic (Fig. 2) . 

The venous system consisted of postcapillary venules formed in and running along the deep layer, which abruptly changed to a transverse direction to join a major vein with a short vertical or oblique path. Optical sectioning in the z-axis under a major vein shows the confluence of these transverse venules with major veins (Fig. 3) . Along the course of major veins, transverse venules underlying the veins were separated from each other by 150 to 200 μm. Variation in this basic pattern of venous drainage included direct connections between major arteries and veins in the superficial layer through capillary-like microvessels, apparently without postcapillary venules (Fig. 4) . There was approximately one direct arteriovenous capillary connection for 10 to 12 “dipping” capillaries. Another frequent arrangement was the presence of connections between an intermediate layer capillary and either a transverse venule (as shown in Fig. 3D ) or a major vein, leading in both cases to a shunt of the deep layer. Therefore, the microvessels in the superficial layer were almost exclusively on the arteriolar side, whereas the vessels in the deep layer were all in the venular side. Several counts at randomly chosen sites showed that the proportion of those directing the flow from the intermediary to the deep layer was approximately 80%. Therefore, the majority of microvessel pathways (approximately 70%) tended to direct the arteriolar flow from the superficial layer toward the deep layer. Symmetrically, the venous drainage starts for the most part in the deep layer and joins the major veins through transverse venules, with few connections in the intermediary or superficial layer. A model of microvessel arrangement based on these measures is shown in Figure 5 . The vascular density of each layer is indicated in Table 1 . 

Intravital Microscopy

The retinal field examined by intravital microscopy was situated on the nasal side of the optic disc. Its surface area was 650 × 495 μm as measured with a 4× lens (Fig. 6) . The labeled cells circulating in the deep vessel layer were visualized with a 10× lens (Fig. 7) , leaving the other microvessel layers out of focus. The labeled cells were less clearly visible in the superficial and intermediate layers than in the deep layer, probably because of the higher blood velocity in the superficial layer and the shorter capillary length in the intermediate layer. Blurring due to the systolic pulse lasted less than 0.12 second (i.e., three frames per cardiac cycle), and the retinal image remained stable between pulses. 

As already stated, switching the filter of the microscope allowed concomitant observation of circulating red and white cells. WBCs had a significantly lower velocity than RBCs (Table 2 and Fig. 8 ). Their pathways were not congruent, as WBC pathways were fewer and straighter. After subcutaneous sodium fluorescein injection, gradual filling of the retinal vessels with fluorescein allowed red cell transit to be superimposed on the angiogram. When the filling of vessels with fluorescein was maximal, switching the filter allowed successive examination of the angiogram and of WBC pathways. This indicated that all capillaries were perfused with RBCs but not all with WBCs. However, specific WBC pathways could not be clearly identified. Sticking WBCs were rare and could not be reliably counted because of the small field of examination. 

Branch Retinal Vein Occlusion

After BRVO, there was a significant decrease in the mean capillary velocity of the red cells in the deep layer (0.86 ± 0.36 mm/sec; P = 0.01, Mann-Whitney test). Confocal microscopy showed elective dilation and increased tortuosity in the deep layer venules (Fig. 9) . In all layers, the number of perfused capillaries was similar to the number in control mice (Table 3) . 

Discussion

Despite numerous studies, the three-dimensional disposition of microvessels in the holangiotic retina is still not clear. The presence of three layers of microvessels in the mouse retina, with predominantly precapillary arterioles in the superficial layer and postcapillary venules in the deep layer, has already been reported. The presence of transverse venules from the deep to the superficial layer has been noted in retinas of the mouse, ^{1 2} rat, ^{4 5} pig, ³ and monkey. ¹⁵ The resultant three-dimensional arrangement has been described as a series of “capillary hammocks” between major arteries and veins. ¹⁶ However, the connections between the microvessel layers remains incompletely known, partly because most imaging techniques do not allow clear visualization of these connections. Indeed, with conventional optical histology techniques, the imaging of connecting vessels requires the examination of many sections. Trypsin digests alters the three-dimensional arrangement of the vessels. Last, vascular corrosion casts are tedious, subject to pressure artifacts, and do not allow optical sectioning. Confocal microscopy of fluorescein-filled vessels, which does not alter the arrangement of the vessels in the retinal tissue, appears, on the contrary to be a convenient, precise, and rapid technique for the three-dimensional imaging of retinal capillaries. 

Spatial analysis of the intraretinal microvessel connections showed in the present study that the superficial and intermediate layers act as asymmetric crossroads for capillary flow, with approximately 70% of their connections driving the flow from arterioles to the deep layer. In parallel, the venous flow from the deep layer directly joins the superficial layers through vertical transverse venules that have few connections with the intermediate layer. Within the superficial and intermediate layers there are capillaries connected to major veins, but their diameter is very much smaller than that of the transverse venules. The functional importance of the connections between major veins and deep layer venules was also suggested by the elective dilation of deep layer venules after BRVO, whereas venules in the other layers were relatively unaffected (i.e., there was no capillary dilation or nonperfusion). 

Overall, these results indicate that the three microvessel layers are fed by a common arterial system that successively traverses these layers. In the same way, the venous system starts from the deep layer plexus up to major veins in the inner retina. This arrangement, which theoretically directs most of the arteriolar flow toward the deep layer, is associated to the presence of microvessel pathways that shunt the deep layer—that is, direct connections between major arteries and veins in the superficial layer and connections between intermediate vessels and transverse (vertical) venules. As regards blood distribution within the retina, the blood that transits through the deep layer has a longer capillary and postcapillary pathway, resulting in higher impedance. 

Retinal blood flow, as opposed to the choroidal flow, is known to be autoregulated by the modulation of retinal vessel compliance in response to variations in blood pressure and intraocular pressure. ¹⁷ Theoretically, the microvessel arrangement within the inner retina provides an efficient way of altering the distribution of flow within the retina, by modulating the impedance of the deeper layer microvessels relative to that of the superficial layer. For instance, the capillaries in the superficial layer directly connected to the major veins probably have a relatively low resistance to flow, because the pressure in the major vein is lower than in the postcapillary venules. The superficial shunt may thus participate in the autoregulation process, for instance by responding to a sudden increase in perfusion pressure by opening up the superficial channels. Ben-nun et al. ¹⁸ reported that in the cat, the elevation of blood pressure reduces the proportion of flow that transits through the deep layer. In contrast, in the case of an increase in metabolic demand in the area of the outer plexiform layer, there may be dilation of deep layer vessels, thus diminishing the deep layer impedance and thus increasing the blood flow in the deep layer. Overall, this suggests that, in addition to a global autoregulation mechanism (i.e., due to the vasomotoricity of major arteries), there may be a local autoregulation of retinal flow—that is, at the level of microvessels. Additional investigations of the alteration of retinal flow distribution during changes in perfusion pressure and/or in oxygen consumption are needed to confirm this hypothesis. 

It is not yet known whether this three-dimensional model can be applied to other species with holangiotic retinas. In the rat retina, we have found a trilaminar serial organization similar to that of the mouse (Paques et al., manuscript in preparation), unlike Pannarale et al. ⁴ and Bhutto and Amemiya ⁵ who did not identify an intermediate layer. In the porcine retina, the available data support the presence of a similar microvascular disposition. ³ In the posterior pole of nonhuman primates, a multilayered pattern was reported by Snodderly et al., ¹⁵ but was not found by Foreman et al. ¹⁹ in the human retina. Indeed, in the primate, the presence of the peripapillary and macular capillaries, which are absent in rodents, complicates the description of the retinal microcirculation. 

The present description of the capillary connections within the holangiotic retina may contribute to the understanding of the results of retinal blood flow examination techniques. Fundoscopic methods such as SLO analyzes chiefly the superficial layer, because the signal from the deepest layers is blurred. ²⁰ The laser Doppler signal ⁹ may correlate with the displacement of red blood cells in the microvessels connecting the different layers (i.e., in the vertical capillaries and venules). Targeted dye delivery allows selective repeated angiography in a given arteriolar territory, but does not allow clear separation of each microvessel layer. ^{21 22} The advantage of intravital microscopy is that it provides direct measurement of velocity in the deep layer capillaries, as the contours of the cells circulating in the adjacent layers are out of focus. Each of these techniques thus explores different compartments of the retinal circulation, and consequently their respective results are complementary but not equivalent. The limitations of intravital microscopy is the narrow field of view that restricts the sampling area and does not allow reliable measurement of the flow in major vessels. 

Our mean velocity of 1.26 mm/sec is within the value measured in other species: 0.7 mm/sec in the cat, ²³ 1.64 mm/sec in the monkey, ⁷ and 4.8 mm/sec in the rat. ²⁴ The higher velocities in the latter two studies may be due to either interspecies variation or to the fact that SLO is more likely to analyze the flow in the superficial layer, which may have a higher velocity due to its arteriolar nature. Nevertheless, these velocities are higher than those reported in other rodent capillaries, which were between 0.12 and 0.19 mm/sec in the subcutaneous vessels of nude mice, ^{25 26} 0.2 mm/sec in the rat cremaster muscle, ¹³ 0.6 mm/sec in the rat mesentery (E. Vicaut, personal data, 1993), and less than 0.7 mm/sec in the rat brain cortex. ²⁷ In choroidal capillaries, the reported values vary between 0.4 mm/sec in the mouse ²⁸ and 1.1 mm/sec in the rat. ²⁴ In addition, in most microcirculatory beds, stagnant red cells and unperfused capillaries are physiologically present, but we did not find them in the retina. Microhematocrit is an important parameter for oxygen delivery to tissues. Is has been shown previously that microhematocrit is approximately 40% to 50% lower than peripheral hematocrit. ¹² In the choroid, the normal microvessel hematocrit was found to be 0.43. ²⁷ To our knowledge, there is no previous report on the microhematocrit of retinal capillaries. Our mean value of 0.49 appears higher than that measured in other tissues. Thus, the higher capillary velocity in the retina cannot be explained by the difference in microhematocrit, because a higher hematocrit is expected to be associated with an increased blood viscosity and hence with a lower velocity. Overall, the present results support the notion that blood velocity and microhematocrit are higher in the retinal capillaries than in other capillary beds. The high velocity may reduce RBC aggregation and WBC plugging, thus helping to maintain a uniform blood supply throughout the retina, despite a relatively high microvessel hematocrit. Also, the high microhematocrit combined with a high velocity increases oxygen delivery to the inner retina, which has a high level of oxygen consumption. ²⁹ The rich anastomotic plexus, that resembles that of the choroid, also favors a uniform repartition of flow in this metabolically active area. 

As regards leukocytes, Kinukawa et al. ³⁰ reported a mean leukocyte velocity of 1.53 mm/sec in the rat retina. We and others have reported a comparable leukocyte velocity in the experimental ⁶ and human retina. ^{31 32} As in other capillary beds, WBCs had lower velocity than RBCs and a differential pattern of pathways. Red and white cell velocities have been compared in cats, ³³ in which the ratio of WBC-to-RBC velocity was between 66% and 88%, comparable to the 72% in the present study. This ratio has also been measured in humans by indirect methods, ³⁴ which produced a ratio of 41%. However, in the latter study this ratio was measured by two different methods: SLO for red cell velocity and the blue field entoptic phenomenon for white cell velocity. This differential velocity is attributable to the different behavior of these cells in normal retinal vessels (i.e., to the stronger interaction between WBCs and the endothelium). This interaction may be due either to higher stiffness of WBCs and/or to receptor-ligand interaction with endothelial cells. 

In the hours after BRVO, we found, in addition to venule dilation, a decrease in capillary velocity that was only moderate and no capillary closure. This suggests that derivation of venous flow is an efficient process during the initial phase of BRVO. The rich network of anastomosis in the deep layer may indeed facilitate the derivation of venous blood, in agreement with the results of other studies of experimental acute BRVO. ^{35 36 37 38} The efficiency of venous flow derivation seems limited in time, as capillary nonperfusion has been described in other models of long-standing BRVO. ^{39 40} After BRVO, collateral vessel formation that derives the venous blood from one territory to a neighboring one is frequently observed. We suggest that deep layer capillaries connecting transverse venules might be the initiators of these collateral vessels. However, the time course of the onset of microvascular changes after BRVO, which was beyond the scope of our study, remains to be determined in greater detail. 

 Submitted for publication July 19, 2002; revised February 7 and May 14, 2003; accepted May 23, 2003.

 Disclosure: M. Paques, None; R. Tadayoni, None; R. Sercombe, None; P. Laurent, None; O. Genevois, None; A. Gaudric, None; E. Vicaut, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Successive confocal microscopy scans showing capillary connections within the retina. The pathways of microvessels branching from arterioles were observed on successive scans in the z-axis, which shows: (A) the superficial layer with arterioles; (B–D) the capillaries just below the superficial layer; (E) the intermediate layer; (F, G) the capillaries just below the intermediate layer; (H) the deep layer. The distance from the superficial layer is indicated in each image. (B, E, F, arrows) A capillary that joins the superficial to the deep layer but is not connected to the intermediate layer. The irregular fluorescence in the vessel lumen is due to the presence of red cells.

Figure 2.

 

Wide-field confocal microscopy scan of the deep microvessel layer. Note the presence of venules (arrows).

Figure 3.

 

Successive scans from the deep microvessel layer (A) to the superficial layer (E) illustrating the course of a vertical collecting venule up to a major vein (arrowhead on successive images). (C, D) An oblique connection ( ) is shown between the major vein and an intermediate layer capillary. (D) Vertical capillaries are shown joining the major vein (arrowhead).

Figure 4.

 

Direct arteriovenous connections (arrows) in the superficial microvessel layer A, artery. The vein is in the top left corner of the image. Note the apparent absence of a postcapillary venule.

Figure 5.

 

Schematic representation of the microvessel array between an arteriovenous couple. The arteriole is in the top left of the image, the venule in the top right.

Table 1.

 

Morphometric Characteristics of Mouse Retinal Microvessels

Density of superficial microvessels	132 ± 16 μm/104 μm2
Density of intermediate horizontal capillaries	142 ± 31 μm/104 μm2
Density of deep layer vessels	317 ± 50 μm/104 μm2
Mean diameter of deep layer venules	10.2 ± 5 μm
Mean inner diameter deep layer capillaries	4.98 ± 0.96 μm

Figure 6.

 

Fluorescein in vivo angiogram of a mouse obtained during intravital microscopy focused on the deep microvessel layer.

Figure 7.

 

Intravital microscopy video frames of labeled red blood cells circulating in the deep microvessel layer of the retina. Arrow in top left image indicates the general direction in which the cells were flowing. Bar, 100 μm.

Table 2.

 

Hemodynamic Parameters in the Deep Capillary Layer of the Retina

	1	2	3	4	5	6	7	8	Mean ± SD
RBC velocity(mm/sec)	1.28 ± 0.28	1.47 ± 0.25	1.15 ± 0.44	1.3 ± 0.44	1.35 ± 0.37	1.35 ± 0.35	1.25 ± 0.43	0.91 ± 0.38	1.26 ± 0.34
RBC flow per capillary (cell/sec)	228 ± 86	264 ± 81	ND	298 ± 138	202 ± 65	264 ± 81	ND	ND	251 ± 98
Calculated tissue hematocrit	0.44	0.44	ND	0.56	0.48	0.37	ND	ND	0.49
WBC velocity(mm/sec)	0.95 ± 0.35	0.95 ± 0.28	0.83 ± 0.46	0.71 ± 0.33	0.54 ± 0.16	0.89 ± 0.51	0.66 ± 0.37	0.88 ± 0.39	0.8 ± 0.36

Figure 8.

 

Example of a histogram of red cell velocity (top) and white cell velocity (bottom) in the deep microvessel layer of the retina of a mouse (number 8 in Table 2 ). x-Axis, velocity of cells (mm/sec); y-axis, percentage of cells.

Figure 9.

 

Confocal microscopy imaging in the deep microvessel layer of a mouse retina upstream of branch retinal vein occlusion. The occlusion site is in the lower left corner ( ). Note the dilation and tortuosity of the venules. Bar, 150 μm.

Table 3.

 

Morphometric and Dynamic Parameters after Branch Retinal Vein Occlusion in the Mouse

	9	10	11	12
Capillary density (μm/104 μm2)
superficial layer	150	144	150	145
intermediate layer	ND	209	198	195
deep layer	361	346	354	343
Vein diameter in deep layer (μm)	15.2	12.6	13.3	14.6
Mean capillary velocity (mm/sec)	0.72 ± 0.39	0.83 ± 0.37	1.07 ± 0.42	0.61 ± 0.43