Dose- and time-dependent effects of actomyosin inhibition on live mouse outflow resistance and aqueous drainage tissues

Actomyosin contractility modulates outflow resistance of the aqueous drainage tissues and intraocular pressure, a key pathogenic factor of glaucoma. We established methodology to reliably analyze the effect of latrunculin-B (Lat-B)-induced actin depolymerization on outflow physiology in live mice. A voltage-controlled microperfusion system for delivering drugs and simultaneously analyzing outflow resistance was tested in live C57BL/6 mice. Flow rate and perfusion pressure were reproducible within a coefficient of variation of 2%. Outflow facility for phosphate-buffered saline (0.0027 ± 0.00036 μL/min/mmHg; mean ± SD) and 0.02% ethanol perfusions (Lat-B vehicle; 0.0023 ± 0.0005 μL/min/mmHg) were similar and stable over 2 hours (p > 0.1 for change), indicating absence of a ‘washout’ artifact seen in larger mammals. Outflow resistance changed in graded fashion, decreasing dose- and time-dependently over 2 hours for Lat-B doses of 2.5 μM (p = 0.29), 5 μM (p = 0.039) and 10 μM (p = 0.001). Resulting outflow resistance was about 10 times lower with 10 μM Lat-B than vehicle control. The filamentous actin network was decreased and structurally altered in the ciliary muscle (46 ± 5.6%) and trabecular meshwork (37 ± 8.3%) of treated eyes relative to vehicle controls (p < 0.005; 5 μM Lat-B). Mouse actomyosin contractile mechanisms are important to modulating aqueous outflow resistance, mirroring mechanisms in primates. We describe approaches to reliably probe these mechanisms in vivo.

Aqueous drainage tissue actomyosin contractility 1-4 modulates intraocular pressure (IOP) and aqueous humor outflow resistance, which becomes abnormal in glaucoma. Elucidating contractility-mediated mechanisms of outflow resistance is expected to provide important clues to glaucoma pathogenesis and therapeutic development, but currently there is a scarcity of accessible and validated live animal model platforms to support such studies. The aqueous outflow system is complex in organization and function and important aspects of this complexity are lost when studies are performed ex vivo or in vitro. Traditionally, in vivo studies have been performed by anterior chamber perfusion in non-human primates [5][6][7] . An alternative is to use live mice, which are versatile as a model, and accessible. The mouse aqueous drainage system closely resembles that of primates in structure and function, and mouse molecular tools or genetic strains with elevated IOP or outflow resistance are available [8][9][10][11][12][13][14] .
To exploit the mouse for this purpose, it is necessary to stably perfuse the mouse anterior chamber, pharmacologically or molecularly probe the drainage tissue, and analyze physiology in vivo, but this is challenging to achieve by traditional methods used in larger mammals. To a large part this is because the mouse anterior chamber is so tiny with a volume of only about 5 μ l 15 , making it technically challenging to achieve reliable measurement.
The extent to which aqueous drainage tissue actomyosin contractility modulates outflow resistance in mice is unknown. Mouse models represent powerful platforms on which to molecularly dissect poorly understood IOP regulatory mechanisms and treatment targets. Studies in the model are expected to inform on similarities between mouse and primate aqueous drainage systems, and applicability of mouse findings to humans. Establishing techniques to do so reliably should enhance our capacity to answer these questions.
We have dealt with the challenge of analyzing live mouse aqueous dynamics by appropriately scaling microperfusion techniques to the smaller mouse eye and coupling fluidic control to an electronic feedback system that yields stable and reproducible measurements in vivo 16 . Capacity to simultaneously deliver pharmacologic probes to the live mouse anterior chamber and drainage tissues and measure physiologic effect in real time has many potential applications. A further advantage of using live mice is the possibility that it is not plagued by the experimental 'washout' artifact of anterior chamber perfusion seen in larger non-human mammals, as mouse eyes ex vivo seem to be free of it 11,[17][18][19][20][21] . The washout artifact describes a time-dependent decrease in outflow resistance of unclear etiology that confounds physiologic measurements. Whether live mice are free of this washout artifact has not been determined.
We used latrunculin-B (Lat-B) as a probe to analyze the in vivo effect of actomyosin inhibition on outflow resistance of the mouse drainage tissues [22][23][24][25][26] . A classical approach, 2-level alternating constant pressure perfusion [17][18][19] , was adapted and fine-tuned for mice. Outflow facility representing hydraulic conductivity; functions describing outflow resistance; presence of a washout artifact; and dose-and time-dependent drug responses were determined in vivo. The corresponding state of drainage tissue actin polymerization reflecting contractility in the mouse aqueous outflow system was then analyzed. Our findings are relevant to probing actomyosin outflow regulatory mechanisms, establishing structural-physiologic correlates, and developing novel glaucoma drugs in live mice as a representative model of the human system.

Results
Stable and reproducible physiologic measurement by 2-level constant pressure perfusion. Alternating 2-level constant pressure perfusions yielded stable and reproducible measurements, as shown in representative pressure and flow rate tracings in Fig. 1. This pattern was seen over perfusions lasting 2 hours, confirming our previous reproducibility studies over a shorter period 16 . Coefficients of variation for repeat measurements over 2-hour perfusions were under 2% for pressure (range 0.2-1.3%) and flow rate (range 0.3-1.7%).
Absence of washout artifact in live mice. Serial outflow facility (representing hydraulic conductivity) estimates during alternating 2-level constant pressure perfusion of Dulbecco's phosphate buffered saline (DPBS) for 10 cycles each showed no significant facility change over 2 hours, as shown in Fig. 2 (17 total outflow facility estimates over 2 hours). Linear regression of serial outflow facility estimates over 2 hours showed a slope of zero (− 0.000001; function: y = − 0.000001x + 0.0026; n = 7 mice). A p-value of 0.76 for difference from zero slope indicated rejection of the null hypothesis that a time-dependent facility change occurred. A power calculation indicated that a sample size of n = 7 was sufficient to determine if perfusion over 120 minutes alone artificially elevated outflow facility ("washout effect") by one and a third standard deviations; an increase of 0.00048 μ L/min/ mmHg or 17.8% above the mean of 0.0027 μ L/min/mmHg (80% power; alpha = 0.05, one-tail). Thus live mouse outflow facility did not significantly increase (ie., 'washout' artifact absent) or decrease over 2-hour perfusions with DPBS (or 0.02% ethanol vehicle control). Mean total outflow facility for 2-level constant pressure perfusion over 2 hours was 0.0027 ± 0.00036 μ L/min/mmHg.

Lat-B dose-and time-dependently increased outflow facility in live mice. A protocol for Lat-B
perfusion delivery to the live mouse anterior chamber is summarized in Fig. 3A. Briefly, Lat-B at different doses (2.5, 5, and 10 μ M) was infused to a total volume of 5 μ l into the anterior chamber of each eye by constant flow perfusion (0.63 μ L/min). After anterior chamber Lat-B incubation for 1 hour, 2-level constant pressure perfusion was performed. Perfusion pressures were alternated between 15 and 25 mmHg for 8 cycles each. After this, the mouse eye was subjected to a step-wise, incremental elevation of constant pressure levels from 15, 20, 25, 30, to 35 mmHg, with each eye stably perfused at each pressure for at least 3 minutes, and corresponding flow rates were recorded, as previously described 16 . Vehicle control animals received ethanol 0.02% in DPBS and underwent outflow facility measurement according to the same protocol as animals receiving Lat-B. An example of perfusion pressure and flow tracings is shown in Fig. 3B.
Following delivery of 5 μ M and 10 μ M Lat-B, outflow facility significantly and time-dependently increased relative to baseline facility, which was determined from the first two outflow facility measurements following 1-hour drug incubation and ethanol 0.02% vehicle control facility, as shown in Fig. 4. A trend toward increased outflow facility was seen for 2.5 μ M Lat-B relative to baseline and vehicle control facility, but this was not significantly  different over the 2-hour perfusion period (p = 0.2). Onset of effect was observed earlier with higher doses, particularly for the 10 μ M dose compared with 5 μ M and 2.5 μ M doses, as outflow facility with the former was already significantly higher right after drug incubation (p = 0.02).
Statistical analysis of outflow facility is summarized in Fig. 4B. Baseline outflow facility immediately after 1-hour incubation with 2.5 μ M and 5 μ M Lat-B was not significantly increased (p > 0.2) compared with vehicle control. Baseline facility following 10 μ M Lat-B perfusion was 6.4 fold higher (p = 0.02) compared with vehicle controls.
The time-dependent rate of outflow facility increase due to Lat-B was analyzed over 3 equal 30-minute perfusion periods, early (A; 70-100 min), middle (B; 100-130 min) and late (C; 130-160 min). The rate of facility increase (slope) for 10 μ M Lat-B was highest throughout and at least an order of magnitude higher than the increase due to 2.5 μ M Lat-B and vehicle controls. Compared with 10 μ M Lat-B, the rate of 5 μ M Lat-B-induced facility increase was 7 times lower in period A (1 × 10 −4 (for 5 μ M) vs. 7 × 10 −4 (for 10 μ M) μ L/min/mmHg/ min), but the rates were equivalent in period C (8 × 10 −4 (for 5 μ M) vs. 7 × 10 −4 (for 10 μ M) μ L/min/mmHg/ min). Compared with 2.5 μ M Lat-B, the rate of 5 μ M Lat-B-induced facility increase was double in period A (5 × 10 −5 (for 5 μ M) vs. 1 × 10 −4 (for 2.5 μ M) μ L/min/mmHg/min) but an order of magnitude higher in the later B and C periods (5 × 10 −5 (for 5 μ M) vs. 5 × 10 −4 (for 2.5 μ M); and 4 × 10 −5 (for 5 μ M) vs. 8 × 10 −4 (for 2.5 μ M) μ L/min/mmHg/min, respectively). Similar to control animals perfused with DPBS only, ethanol 0.02% vehicle control animals did not show a significant time-dependent facility change relative to baseline facility (p = 0.1), confirming the absence of 'washout' .  Lat-B effect on aqueous drainage tissue actin polymerization. Polymerized actin (filamentous actin; F-actin) served as an indicator of contractile state. In vehicle control tissue, the density and distribution of labeled F-actin was prominent in the ciliary muscle (yellow arrow) and ciliary body (white arrow), and somewhat less in the trabecular meshwork (TM, green arrow), representing a basal state of contractile tone as shown in Fig. 7A(a-c). The distribution and intensity of F-actin labeling in vehicle control tissue was similar to that of eyes perfused with DPBS only (data not shown). Anterior chamber perfusion of 2.5, 5 and 10 μ M Lat-B caused significantly reduced F-actin fluorescence labeling intensity in the drainage structures of ciliary muscle (p = 0.0007) and TM (p = 0.004), as shown in Fig. 7A(d-f) for 5 μ M Lat-B as a representative finding. F-actin labeling intensity was not different between different Lat-B doses of 2.5, 5 and 10 μ M (data not shown).
Following 5 μ M Lat-B anterior chamber perfusion, mean fluorescence intensity of F-actin labeling in ciliary muscle and TM was decreased by 48% and 46%, respectively (both p < 0.005) compared with vehicle controls in quantitative immunohistochemistry image analysis, as shown in Fig. 7B.
High resolution in situ transscleral 2-photon excitation fluorescence imaging (TPEF) further confirmed the effect of Lat-B on F-actin in the mouse aqueous drainage tissues, as shown in Fig. 8. In control mice, a curvilinear network of F-actin that was denser and more intense in the ciliary muscle than in the adjacent TM was seen. Following exposure to 10 μ M Lat-B as a representative dose, F-actin labeling that was initially arranged as a curvilinear network rearranged as aggregates.

Discussion
Our system incorporating electronically controlled microperfusion fluidics coupled to real-time pressure feedback permitted stable and reproducible 2-level constant pressure perfusion over prolonged periods in live mice. The washout artifact that characterizes many large mammal perfusions was not seen in live mice, with either DPBS or 0.02% ethanol vehicle. A dose-and time-dependent increase in total outflow facility representing a graded increase in total outflow facility was seen in response to Lat-B-induced inhibition of actomyosin contractility. Our facility measurement system was sensitive enough to detect and quantify this change. Lat-B exposure was associated with reduced actin polymerization in the mouse drainage tissues of TM and ciliary muscle. These findings reflect the importance of actomyosin contractility to aqueous outflow modulation in live mice, mirroring functional aspects of the primate drainage system.
Our previous report of in vivo mean outflow facility (0.0066 μ L/min/mmHg) was based on linear regression of flow rate data obtained at multiple pressures between 15-35 mmHg 16 . While our previous data 16   Thus using lower perfusion pressures in vivo appears to explain the lower facility findings reported here, as compared with previously reported facility values derived by higher perfusion pressures. Based on our and foregoing reports, we conclude that facility findings may vary according to (1) differences in methodology (constant pressure vs. constant flow) and tissue state (live animal vs. postmortem tissue), but also (2) choice of perfusion pressure.
Lat-B dose-and time-dependently increased total outflow facility in live mice. Facility increased by up to 12and 21-fold for 5 μ M and 10 μ M Lat-B respectively compared with vehicle controls. While 2.5 μ M Lat-B did not significantly increase facility over that of vehicle controls over the perfusion timeframe, a trend toward gradually increased facility over time was seen that perhaps was not captured within the duration of our perfusions. The onset of action of 10 μ M Lat-B was quicker than that of 5 μ M, and the onset for 5 μ M was quicker than that of 2. Lat-B delivery in live mice was performed by constant flow perfusion to a volume of 5 μ l at each dose. The 5 microliter delivery volume, which approximates the anterior chamber volume of mice, was standardized over all dosages to allow for comparison of dose-dependent effect. Standardized perfusion duration and volume allowed valid comparisons of the different drug doses. Drug perfusion flow rates were chosen to maintain physiologically relevant conditions, with transduced perfusion pressure being kept in the range of 15-20 mmHg for the roughly 8-minute drug delivery perfusions. We are presently validating a promising method 37 to exchange anterior chamber contents and correctly estimate drug dose in the anterior chamber that could be used in the future to achieve a desired anterior chamber drug concentration.
The facility responses to Lat-B we observed may not be directly comparable to the latrunculin dose-responsiveness reported in live non-human primates 24,27,38,39 as we did not directly exchange the anterior chamber contents. For example, while a 4-5 fold facility increase following 2 μ M of Lat-B was seen 90 min after exchange perfusion in monkeys, we did not find a significant increase in facility for the same dose over the same period in mice. The exact anterior chamber drug concentration following initial drug delivery was unknown in the in vivo mouse model, but would have been lower than the dose delivered due to anterior chamber drug dilution and drug escape through the drainage pathway. Anterior chamber drug concentration would have increased with ongoing drug perfusion, however, and likely was reflected in the significant facility increase seen over time with our perfusions of 5 and 10 μ M Lat-B.
Absence of the washout artifact in live C57BL/6 mice confirms ex vivo studies in enucleated eyes of the same mouse strain 11 . Thus unlike studies in larger mammals, perfusion data obtained in live mice do not need to be statistically corrected or adjusted for the confounding washout artifact over the perfusion duration we used. The washout artifact is not seen in human eyes either 17 , further indicating the attractiveness and relevance of mice for modeling human outflow dynamics in vivo.
A conceptual difference between traditional techniques of perfusion under constant hydrostatic pressure and our electronically controlled microperfusion system is that our system delivers nano to microliter-scale pulses of fluid to maintain a constant pressure in the mouse anterior chamber. The pump perfusion is intermittent, activating only when pressure dips below a preset threshold to restore the constant pressure. Conversely, no perfusion occurs while pressure remains above the threshold. By contrast, traditional perfusion systems subject the anterior chamber to the constant pressure of a hydrostatic column as long as the tubing is open. Whether our constant pressure perfusion methodology contributed to the stable and reproducible measurements and lack of washout artifact that we observed remains to be determined.
Lat-B sequesters monomeric actin to prevent polymerization and formation of actin filaments, the contractile form of actin [38][39][40] . Not surprisingly, Lat-B delivery to the anterior chamber was associated with decreased actin polymerization in the aqueous drainage tissues. This was seen as significantly reduced F-actin labeling intensity and rearrangement of the curvilinear F-actin network in the mouse drainage tissues. The finding that reduced actin polymerization in the drainage tissue correlated with increased outflow facility following Lat-B exposure agreed with reports in non-human primates 23,24 . These findings indicate that actomyosin mechanisms modulating outflow facility are common to primates and mice.
We compared the effects of different concentrations of Lat-B (2.5, 5, and 10 μ M) on F-actin fluorescence. We found a significant difference in F-actin intensity/distribution between Lat-B treatment and vehicle, but we did not find F-actin differences between the different Lat-B dosages (data not shown). Subtle differences in actomyosin contractility corresponding to differences in Lat-B dose-related facility change may not have been evident by F-actin analysis alone, or perhaps our techniques could not resolve them.
While we expected to find decreased actin polymerization in the TM following Lat-B perfusion, we did not expect to see this in ciliary muscle. Perhaps this finding is not surprising, however, given that the ciliary muscle forms the anterior part of the uveoscleral drainage pathway exposed to aqueous humor contents as aqueous exits the anterior chamber. Ciliary muscle extensions also intermingle with the TM and affect conventional outflow facility. Ciliary muscle is smooth muscle 41 that is subject to actin polymerization. Primate uveoscleral outflow is considered to be only minimally pressure sensitive under conditions where IOP is less than episcleral venous pressure 42 . It is possible, however, that the uveoscleral pathway becomes more pressure sensitive under the influence of factors such as intraocular inflammation and prostaglandins [43][44][45][46][47][48] . Relaxation of the ciliary muscle itself mediates increased uveoscleral outflow [49][50][51][52] , but whether this also alters the pressure sensitivity of the pathway is unclear. To what extent outflow resistance is influenced by relative contributions and crosstalk between different contractile and extracellular matrix elements in different regions of a complex drainage pathway is still being worked out 53 . To our best knowledge, increased outflow facility by latrunculin or other actomyosin-altering agents has not previously been attributed to altered ciliary muscle actomyosin contractility 54 . Our findings in live mice suggest that such a mechanism is worth exploring further. Two-level Constant Pressure Perfusion. Anterior chamber perfusion was performed using previously described apparatus and methods 16 . With our two-level constant-pressure perfusion in live C57BL/6 mice (n = 7 mice), the perfusion pressure set-point was alternated between 15 mmHg and 25 mmHg. The pump had infusion and withdrawal modes that allowed perfusion pressure to be increased from 15 to 25 mmHg, decreased from 25 to 15 mmHg, then increased again to 25 mmHg and so on, alternating between 15 and 25 mmHg. Typically, perfusion at each pressure started with a 30-second equilibration period followed by perfusion for 3 minutes at that pressure. Each perfusion cycle comprised data collected during perfusion at 15 then 25 mmHg. Alternating 15/25 mmHg perfusion was repeated 8-10 times. This yielded 8-10 sequential 15/25 mmHg perfusion cycle data sets representing total data from 16-20 3-minute perfusion periods per eye (performed 10 sets for Figs 2 and 8 sets for Fig. 4). One eye per animal was perfused. Outflow facility (C) was calculated by the standard method of weighted successive averages, as classically performed in primate perfusions 17,19 . Briefly, outflow facility (C) following Lat-B incubation was calculated from data sets of pressure (P1 − P16) and corresponding flow rates (F1 − F16), as illustrated in Supplementary Fig. 1. For successive 15/25 mmHg perfusion cycles, C1 through C15 was calculated by equation (1).

Methods
The first total outflow facility value (Ct 1 ) was calculated as the average of C from the first three successive perfusion cycles (C1 − C3). To calculate the next Ct value (Ct 2 ), the oldest data point from the first perfusion cycle (C1) was excluded and the subsequent C value (C4) added to the prior two overlapping periods (C2 and C3), as described by equation (2). Outflow resistance was calculated as the inverse of facility measurements, having units of mmHg/min/μ L. Following this method of two-level constant-pressure perfusion, total outflow facility was further determined in the same live mice by constant pressure perfusion for at least 3 minutes at ("step-wise") incrementally higher pressure levels between 15-35 mmHg (15,20,25,30 and 35 mmHg), with total outflow facility given by the regression slope of the flow rate (F; μ L/min) vs. pressure (P; mmHg) function.
To determine total outflow facility in 2-level or step-wise perfusion, pressure and flow rate data were extracted from the software at a rate of 1 per 10 milliseconds for 15,000 consecutive data points for each pressure condition. Data from the first 30 seconds (3000 data points) were taken to occur during system stabilization and excluded. The pressure transducer and pump were calibrated as previously described 16  Drug delivery of Lat-B into the anterior chamber of live C57BL/6 mice was performed at a constant flow perfusion rate of 0.63 μ L/min for 7-8 minutes approximately, depending on pump calibration, for a total of 5 μ l volume delivered. This allowed for standardized volume delivery of drug at the different concentrations into the anterior chamber of all mice. During perfusion drug delivery, transduced anterior chamber pressure did not exceed 25 mmHg. After drug delivery was completed, perfusion was stopped and drug left to incubate in the anterior chamber for 1 hour. After drug incubation, the same apparatus was used to conduct alternating 2-level constant pressure perfusion (at 15 and 25 mmHg) as described above. This was then followed by step-wise, incremental constant pressure elevation . Independent mice received ethanol 0.02% (vehicle control, n = 5 mice) or DPBS (non-vehicle control; n = 5 mice) in the same manner as mice receiving drug (Lat-B).
Fluorescence Microscopy and Quantitative Analysis. After perfusion, enucleated eyes were quickly embedded in Tissue-Tek Optimum Cutting Temperature compound. Cryosections (7 μ m thickness) were fixed with 4% paraformaldehyde (PFA) and further permeablized/blocked in the blocking solution (5% Bovine serum albumin and 0.3% Triton X-100) for 1 hour at room temperature (RT). To visualize the effect of Lat-B on filamentous actin (F-actin), the contractile form of actin, sections were incubated with Alexa 568-phalloidin (Life Technologies, Grand Island, NY) for 1 hour at RT and then mounted using ProLong Gold Anti-fade reagent with 4′,6-diamidino-2-phenylindole (DAPI, Life Technologies). Negative, non-specific labeling was established with normal IgG isotypes. Sections were analyzed with a Leica SP5 high-speed spectral confocal laser-scanning microscope (Leica Microsystems, Wetzlar, Germany) or a Zeiss LSM 710 confocal microscope (Carl Zeiss, Oberkochen, Germany). Immunofluorescence staining for phalloidin staining was performed in randomly selected slides (4-5 slides per each eye, n = 4 mice) containing 4 sections per slide and examined under the confocal microscope.
Specific fluorescence from tissue labeling in histological sections was captured by confocal microscopy with exposure time kept constant across all images. Image sections were imported as 16 bit images and analyzed by NIH Image J software. Fluorescence intensity of pixel grey values in 8 separate regions of interest per region of TM and CM was calculated and averaged across each tissue region. Fluorescence intensity for F-actin was measured in TM and CM separately, averaged across data from 4 mice, and then compared using ANOVA and Tukey's and Sidak's comparison tests 38 .
Additional frozen sections were stained with hematoxylin and eosin. Drainage tissue integrity and effect of Lat-B on tissue histological features of mouse eye after perfusion were examined by light microscopy (data not shown). Two-photon excitation fluorescence imaging. Ex vivo mouse eyes were fixed with 4% PFA and labeled intact, without dissection or sectioning, with Alexa 568-conjugated phalloidin (filamentous actin) and imaged by TPEF (Zeiss 710NLO [Oberkochen, Germany] 55,56 supplemented with a BiG non-descanned detector [Zeiss]) to capture red fluorescence and second harmonic generation (SHG) images. SHG of mouse iridocorneal angle (not shown) was used to identify the region of the TM just deep to Schlemm's canal by a transscleral approach 57,58 . In mice, the ciliary muscle lies just deep to the TM 15 . The eyes of non-pigmented Balb/c mice permitted unimpeded views of the deeper aqueous drainage tissues, which we have characterized previously 41 . Mice that were exposed and unexposed to 10 μ M Lat-B (n = 3 mice) were imaged.