Animals
Tri-coloured domestic Guinea pigs (Cavia porcellus, n = 29) were reared in plastic boxes with wire tops, illuminated overhead by white light emitting diodes diffused through a 3mm thick Perspex screen located 200 mm above the box lid. Lights operated on a 12-hour light-dark cycle at a constant illuminance of 700lx, measured at eye level within the housing box. Room temperature was maintained at 21(±2)oC and food and water was provided ad libitum. All experiments and experimental procedures were approved by the Animal Care and Ethics Committee of the University of Newcastle and were conducted in accordance with the Australian code for the care and use of animals for scientific purposes.
Experimental Design
Two experiments were undertaken on retinae freshly extracted from dark adapted guinea pigs between 14 and 28 days of age (Table 1). Experiment 1 used 17 retinae that were cut-loaded with molecular tracer for different incubation times, and the imaged tissues were analysed using three different methods to yield the coupling space constant, the coupling rate constant, or the diffusion coefficient. In Experiment 2, retinae were cut-loaded with molecular tracer and incubated for a single duration. Cell coupling was inhibited with increasing concentrations of the general gap-junction inhibitor meclofenamic acid and imaged tissues were analysed using the same three methods as in Experiment 1. In both experiments, a-type horizontal cells were selected as a model system for analyses. In Experiment 3, the generalisability to a different coupled network was studied in retinae from Experiment 2, by co-labelling for neuronal nitric oxide synthase (nNOS).
Table 1
Experimental Design showing the primary variables manipulated. The Experimental Light Exposure refers to the light level during cut-loading at a wavelength of 850nm and was approximately equivalent to total darkness
|
Aim
|
Incubation Time (mins)
|
No. of retinae (No. of cuts)
|
MFA Concentration (µM)
|
Experimental Light Exposure
(log R*.Rod−1.s−1)
|
Experiment 1
|
Test analysis methods across different incubation times
(Horizontal Cells)
|
1
|
2 (5)
|
N/A
|
-4.5
|
3
|
4 (10)
|
5
|
4 (12)
|
10
|
4 (10)
|
20
|
3 (7)
|
Experiment 2
|
Test analysis methods for variable coupling at a fixed time-point
(Horizontal Cells)
|
20
|
2 (4)
|
50
|
-4.5
|
2 (4)
|
100
|
2 (5)
|
150
|
2 (5)
|
200
|
2 (4)
|
250
|
2 (4)
|
500
|
Experiment 3
|
Reanalysing tissues from
Exp. 2, for non-horizontal cells
|
Same as Exp. 2
|
Procedures
Light Adaptation and Euthanasia
Guinea pigs were adapted to complete darkness for one hour prior to euthanasia. Animals were anaesthetised using gaseous isoflurane (5% in 1.5 L/min O2) and euthanised via intracardial injection of Pentobarbitone Sodium (Lethabarb®, Virbac Australia Pty Ltd, NSW, Australia). A dim red head torch (light emitting diode (LED) spectral peak at 635nm, max 20uW.cm− 2 at 20cm away from tissue, luminance at tissue was 0.5 log R*.Rod− 1.s− 1) was used during euthanasia and dissection in an otherwise dark room. During cut-loading, infrared LEDs (spectral peak 850nm, -4.5 log R*.Rod− 1.s− 1) and night vision goggles (Sionyx Aurora Pro) were used to maintain dark-adaptation of the retina during experimentation.
Cut-loading Procedure
The eye was rapidly enucleated and submerged in Ames solution (catalogue: A1420-10X1L, Sigma-Aldrich, MO, USA) at room-temperature (18 to 21oC). Whilst submerged, the cornea and limbus were removed with a circular cut extending along the posterior pars plana (Figure. 1). A small incision was then made along the long ciliary artery to mark the nasal axis using a No.11 scalpel blade. The crystalline lens and vitreous humour were gently removed using No.5 forceps. The remaining eye cup containing the retina, choroid and sclera were transferred into a well plate containing 10mL of Ames solution (36oC) bubbled with carbonox (95% O2 5% CO2).
The retina was separated from the underlying choroid using a blunt dental spatula. The optic nerve joining the retina and sclera was then cut using 3mm curved scissors and the retina orientated with photoreceptor side down was mounted onto 0.22µm pore size Millipore membrane filter paper (catalogue: GSWP04700, Merck & Co. NJ, USA). Tissues were acclimatised for 15 minutes in complete darkness in the bubbled Ames solution. In Experiment 2, the required dilution of meclofenamic acid (MFA) (catalogue: M4531, Merck & Co. NJ, USA) made from stock solution of 100mg/mL MFA sodium salt dissolved in 100% ethanol was added directly to the acclimatisation Ames bath solution and all bath solutions up until fixation.
The retinas were briefly removed from solution and cut along the superior, temporal, and inferior axes with a size 11 scalpel blade that prior to each cut was dipped in 3% the biotin derivative, N-(2-aminoethyl) biotinamide hydrochloride (Neurobiotin™ Tracer, catalogue: SP-1120, Vector Laboratories, CA, USA) diluted in Ames solution. The tissue was returned to the Ames solution and the NeurobiotinTM dye allowed to disperse through the cell network by incubating for either 1, 3, 5, 10 or 20 minutes in Experiment 1, or 20 minutes in Experiment 2 (see Table 1). At the end of each Experiment, tissues were removed from the bubbled bath solution and fixed in 4% paraformaldehyde (4% wt/v, diluted in 0.1M phosphate buffer) at room temperature for 30 minutes. Retinas were washed in 1 x PBS (30 mins) and reacted with Alexa-Fluor 488 conjugated streptavidin (catalogue: S11223, ThermoFisher Scientific, MA, USA, diluted 1:100 in 0.5% Triton-x in PBS) overnight at 4oC. Retinas were washed in 1 x PBS and mounted, photoreceptor side down, onto microscope slides using anti-fade Vectashield aqueous mounting medium (Vector Laboratories, CA, USA, catalogue: H-1000-10).
Immunofluorescence
Retinae from Experiments 1 and 2 were removed from glass slides and placed in PBS, incubated in Triton-x for 30 minutes (1% Triton-x in PBS) and blocked using normal Donkey serum (10% NDS in 0.5% Triton-x, PBS) for one hour. Retinae from Experiments 1 and 2 were then counter labelled with antibodies for Calbindin and Neuronal Nitric Oxide Synthase (nNOS) respectively (See Table 2 for details). Retinae were washed in PBS (3 x 10 mins) and incubated in secondary antibody solution (Table 2). Retinas were washed in 1 x PBS and mounted, photoreceptor side down, onto microscope slides using anti-fade Vectashield aqueous mounting medium.
Table 2
Antibody reagent list. Both primary and secondary antibodies were diluted in 1% NDS and 0.5% Triton-x in PBS. RT, room temperature.
Antibody
|
Catalogue
|
Host
Species
|
Company
|
Dilution
|
Time (Temperature)
|
Neuronal Nitric Oxide Synthase (nNOS)
|
N7280
|
Rabbit
|
Sigma-Aldrich
|
1:1000
|
Overnight (4oC)
|
Calbindin
|
|
Rabbit
|
|
1:400
|
Overnight (4oC)
|
Donkey anti-Rabbit (CY3)
|
715-165-150
|
Donkey
|
Jackson ImmunoResearch Laboratories, Inc.
|
1:400
|
1 hour (RT)
|
Image Collection and Preparation
Images were collected 3mm from the optic nerve head along the superior, temporal and inferior cuts using an Olympus IX91 scanning laser confocal microscope. Each image comprised of 50 optical slices (1µm apart), taken between the outer plexiform later and the retinal ganglion cell layer. Confocal settings were kept consistent across all image acquisition. Composite images from 10 slices were made using the sum-slices z-project function in Fiji (open-source distribution based on ImageJ2 released by National Institutes of Health)24.
Light Intensity Calculations
Reported values of spectral irradiance (\(E\left(\lambda \right),\)µW.cm−2.nm−1) of the home-box during rearing, the red LED head torch used during euthanasia and dissection and the infrared LED lighting used during cut-loading experimentation were measured using an Ocean Optics spectrophotometer (USB-4000, Ocean Optics). Illuminance (lx) was calculated based on the Guinea pig photopic spectral sensitivity curve25. Photometric units were converted to total effective photons.cm−2.s−1 based on the rod spectral sensitivity curve as determined using Govardovskii nomograms 26 with λmax = 496 nm (Jacobs & Deegan, 1994). This was then converted into photoisomerisations.rod−1.s−1 (R*.rod−1.s−1) assuming a 1.0 µm2 effective collection area for each rod 27. At low light levels, irradiance (µW.cm−2) was measured using a Newport optical power meter centred at the spectral peak (main unit: 2936-R, sensor: 818-ST2-UV/DB, Newport).
Modelling NeurobiotinTMDye-Transfer
Standard protocol: The mean fluorescence of the image spanning perpendicular from the cut location was measured using the plot profile function in Fiji24.
Modified protocol: As the true cellular concentration of dye was not known, the relative mean fluorescence of each cell was used to gauge the relative concentration of tracer in cells. The mean fluorescence and position of each cell soma was measured in Fiji using the oval tool. The background fluorescence was measured and subtracted from absolute fluorescence measurements. All measurements were then normalised to the maximum fluorescence measured at the cut location. The distance between adjacent horizontal cells (Experiments 1 and 2) and amacrine cells (Experiment 3) were measured from soma centre to soma centre. Forty measurements for each cell-type per retinal image were used to obtain an average.
Data were analysed as both cell fluorescence per absolute distance (µm) from the cut site and as cell fluorescence per cell-separation from the cut site (calculated by dividing absolute distance by the mean soma-soma distance).
Method 1: The decline in fluorescence intensity per distance was fitted with the exponential decay curve below
$$C={C}_{o}{e}^{\left(\frac{-x}{\text{λ}}\right)}$$
1
Where C is the concentration of the cell at distance x from the cut (either in µm or cells), C0 is the maximum concentration and λ is the space constant (either in µm or cells).
Method 2: Zimmerman and Rose21 described the diffusion kinetics of molecular tracers through a chain of 5-7 coupled giant cells of the Chironomus salivary gland using compartment model analysis. For a series of three coupled cells (C0, C1, C2) the movement of tracer into (qr) and out of the cell (qs) C1 can be approximated using the following equations
$${q}_{R}={-k}_{j}\left({C}_{1}-{C}_{0}\right)$$
2
$${q}_{s}={-k}_{j}({C}_{2}-{C}_{1})$$
3
Where kj is a rate constant of units (distance2.time-1). The net movement of tracer through cell C1 can then be approximated by the combination of these two equations to yield:
$${q}_{R}- {q}_{S}= -{k}_{j}\left({C}_{1}-{C}_{0 }-{C}_{2}+{C}_{1}\right)= {k}_{j}\left({C}_{0}+{C}_{2}-{2C}_{1 }\right)$$
4
The geometric arrangement of HCs differs to that of giant cells used in Zimmerman and Rose’s original paper. The arrangement of HCs can be modelled using a triangular lattice, with each cell connecting to the six neighbouring cells (Fig. 2). Therefore, in cut-loading a line of cells will be initially filled with tracer. If we assume no tracer transfer occurs between cells of equal internal dye concentration, then the number of cells C0 would feed into depends on the angle of the cut with respect to the lattice (Fig. 2). In situation one, each cell at C1 would receive tracer from two cells at C0 and would feed into two cells at C2 (Fig. 2a). In situation two, the cells at C1 would either receive tracer from one or three cells at C0 and would feed into either one or three cells at C2 (Fig. 2b). As these situations alternate in the second example, each layer receives and exports tracer to an average of two cells.
This was incorporated into equation (4) to yield:
$${2q}_{R}- {2q}_{S}= -{k}_{j}\left(2{C}_{1}-{2C}_{0 }-{2C}_{2}+{2C}_{1}\right)= {2k}_{j}\left({C}_{0}+{C}_{2}-{2C}_{1 }\right)$$
5
Which yields the series:
|
\(\frac{d{C}_{1}}{dx}={2k}_{j}\left({C}_{2}-{C}_{1}\right)-{k}_{s}\left({C}_{1}{V}_{1}\right)\)
|
(6)
|
|
\(\frac{d{C}_{n}}{dx}={2k}_{j}\left({C}_{n+1}+{C}_{n-1}-2{C}_{n}\right)-{k}_{s}\left({C}_{n}{V}_{n}\right)\)
|
etc.
|
Where kj is a rate constant describing dye transfer between coupled cells within a network (cells2/s), ks is sequestration or loss of dye as it passes through the tissue (cells2/s) and V is the relative volume of the cell (set to 1 as all HC assumed to have equal volume). These equations were solved in MATLAB (mathworks) by first fitting a 2-parameter Gaussian curve to the original data to calculate the mean relative fluorescence at discrete intervals (Cn). A solution for kj was obtained by fitting the concentration series to the ode45 solver, which solved the above equation series (expanded to n = 1:40) at a defined time point based on the 4th and 5th Runge-Kutta method.
When kj was calculated in terms of absolute distance (cm2.s−1), the geometric path of molecular tracer was not accounted for and the non-normalised cell-cut distances were used for calculation via Zimmerman and Rose’s21 original equation written below:
|
\(\frac{d{C}_{1}}{dx}={k}_{j}\left({C}_{2}-{C}_{1}\right)-{k}_{s}\left({C}_{1}{V}_{1}\right)\)
|
(7)
|
|
\(\frac{d{C}_{n}}{dx}={k}_{j}\left({C}_{n+1}+{C}_{n-1}-2{C}_{n}\right)-{k}_{s}\left({C}_{n}{V}_{n}\right)\)
|
etc
|
Method 3: The diffusion of dye through a homologously coupled cell network (such as horizontal cells) following cut-loading can be described as the diffusion of a substance along one axis using Fick’s second law of diffusion28
$$\frac{\partial C}{\partial t}=D\frac{{\partial }^{2}C}{\partial {x}^{2}}$$
8
Where C is concentration of the diffusing substance, t is time in seconds, D is the diffusion coefficient (cm2.s−1) and x is distance (cm).
One method for modelling dye-diffusion during cut-loading is to consider that at t = 0, all molecular tracer is located within the region -l < x < +l describing the boundary of the initial cut made through the retina. Dye transfer then occurs in one-dimension through the coupled cells at the cut boundary, which can be treated as the sum of an infinite number of line sources with diffusion occurring along the x-axis and modelled using the following equation:
$$C=\frac{C{}_{0}}{2}\left\{erf\left(\frac{l-\left|x\right|}{2\sqrt{Dt}}\right)+erf\left(\frac{l+\left|x\right|}{2\sqrt{Dt}}\right)\right\}$$
9
Where:
$$\text{erf} \left(z\right)= \frac{2}{\sqrt{\pi }} {\int }_{0}^{z}{e}^{-{t}^{2}} d t$$
10
And C0 is the initial concentration, with initial boundary conditions at t = 0:
C = C 0 , x < l and C = 0, x > l
Figure 3 shows an example solution for equation (9). with l = 0.04 cm, D = 1.6 x 10-6 cm2.s-1 and C0 = 6.15.
Statistical Analysis
Data from the five incubation conditions in Experiment 1 were analysed using a 1x5 way ANOVA in SPSS (IBM Statistics 25). In cases where data failed the assumption of homogeneity of variance (Levene’s test, p < 0.05) a Brown-Forsythe ANOVA was used instead. Independent samples t-tests were used when comparing fits (Experiment 1) and when comparing coupling rates between cell types (Experiment 3). In Experiment 2, the reduction in cell-coupling due to MFA (100 µM – 500 µM) was expressed as a percentage reduction from the low-dose (50 µM MFA) experimental condition. Paired samples t-tests were then used to compare the normalised effective response calculated by the three analysis methods (Experiment 2). Wilcoxon signed-rank test was used when data was not normally distributed. All tests were conducted as two-tailed tests. Data from Experiment 2 was fitted with the dose-response Hill curve29:
$$\frac{E}{{E}_{max}}=\frac{1}{1+{\left(\frac{E{C}_{50}}{\left[A\right]}\right)}^{-n}}\left(11\right)$$
Where E is the magnitude of the change in coupling, Emax is the maximum change in coupling, [A] is the concentration of meclofenamic acid, n is the Hill coefficient and EC50 is the concentration at which 50% of the maximum response is observed. Fitting was performed in Matlab (Mathworks) using a modified script by Ritchie Smith30. Power analyses were performed in G*Power (Version 3.1, Universitӓt Kiel and Universitӓt Dusseldorf, Germany) for an independent samples t-test (2-tailed).