- Email: [email protected]
Oxygen diffusion in the trout retina
Erp. Eye Reg.(1985) 41. 607-618
Oxygen PAUL
Diffusion
E. DESROCHERS, AND
~fl Physiology,
Lkpartment
in the Trout
KENNETH @J. RCSSELL
Michigan
A. PRATT.
Retina PAUL
0. FROMM
HOFPERT
State T!niversity. r.s.il.
East Lansing,
MI 438%4-1101.
(Recei~wd 4 March 1985 and accepted 29 August 19X5, IVruj York) Oxygen tensions were measured in viva within the different layers of the rainbow trout retina. Oxygen microelectrodes were advanced in 1Opm increments through the retinas and the Po, measured at each location. Mean retinal F’o, ranged from 124 mmHg at the retinal-vitreal interface to 381 mmHg at the choriocapillaris. These data reaffirm that the cellular layem of the normal trout retina are continually exposed to supra-arterial oxygen tensions that arc known to cause toxicity in other species. The intraretinal Po, gradient was mathematically characterized and results indicate that the in vivo oxygen profile can be described by a two-component exponential function. In order to gain a better understanding of oxygen delivery to the retina. the data were also subjected to an analysis based upon the classical equation for planar diffusion. The trout retina is an excellent model for studying oxygen diffusion in viva since this tissue is supplied with oxygen from a single source, thereby simplifying the mathematical analysis. C?alculations yielded values of 1.86 x 10m5and 0.58 x 10e5 ml 0, min-’ cm- I atm-’ (at 9°C) for the Krogh permeation coefficient (DS) for the photoreceptor region and the remaining neural retina, respectively. fYel/ mordh: hyperoxia: Krogh permeation coefficient; oxygen: oxygen diffusion; rainbow trout retina.
1. Introduction Exposure of animals to oxygen tensions in excess of that found in air may cause both structural damage and cellular dysfunction in various tissues and/or organs. The retina and choriocapillaris of the mammalian eye are especially vulnerable to the toxic effects of hyperoxia. As a consequence. much attention has been directed towards studies of the mechanism of oxygen toxicity and agents which protect against it. In light of the above, it seemed incredible when Wittenberg and Wittenberg (1961) reported that the choroidal rete mirabile of many teleosts. in conjunction with the choriocapillaris, is capable of concentrating oxygen to levels far in excess of those found in arterial blood. Several investigators recognized the potential of the teleost eye preparation as a model for the study of oxygen toxicity and studies have been done to describe the oxygen-concentrating mechanism in the rhoroid and the nature of oxygen delivery to the teleost retina. Fairbanks, Hoffert and Fromm (1974) proposed a mechanism to describe how the choroidal rete mirabile of the rainbow trout (flalmo gairdneri) functions to elevate ocular PO, via countercurrent oxygen multiplication. The mechanism is similar to that proposed for the secretory epithelium of gas bladders of certain physoclistous teleosts (Kuhn, Ramel, Kuhn and Marti, 1963). Measurements of the PO, in the choroid layer of rainbow trout eyes showed it to average more than 400 mmHg (Fairbanks, Hoffert and Fromm, 1969), while it was approximately 100 mmHg at the retinal-vitreal @ Dr Jack R. Hoffert suffered a fatal heart attack on 21 November 1984. This paper is dedicated to his memory. Please address all correspondence and requests for reprints to Paul E. Desrochers. Department of Physiology, Michigan State University, East Lansing, MI 48824-l 101, IT.S.A. 001&4835/85/l
10607 + 12 $03.00/O
@z$ 19X5 Academic- Prrss Inc. (London)
I,imit,rd
m4
F'. E. I)IC:SI
interface. Thus there is a large I<;,, g radient wit,hin the retina of this sl)rc*ic~s. ‘I’bts distribution of PO, within isolated teleost retinas has been reported by, lkujan. Svaetichin and Negishi (1971) and Segishi. Svaeitichin. Laufer and I)ruja,n (19i5). The present’ report however, is the first in which the 16, was measured in retinas ot intact fish which possess a functioning countercurrent oxygen multiplier. The aims of the research reported here were to develop a t’echniyue to measure t,hc Po, of the rainbow trout eye in vivo and determine t,hose areas of the retina that are exposed to oxygen tensions generally expected t.o be boxic. The data obt.ained were used to characterize mathematically the intraretinal pob,gradient and gain insight’ into the nature of delivery and/or utilization of oxygen in this tissue. Values were calculated for the Krogh permeation coefficient for t’he diffusion of oxygen through (a) the pigmented epithelial-photoreceptor cell layer and (b) the remaining neural retina. 2. Materials
and Methods
weighing 15&300 g were obtained from Midwest Rainbow trout (Salmo gairdneri) Fish Farms, Inc., Harrison, MI and held in the laboratory at 9+ 1°C. All studies were carried out in accordance with the ARVO Resolution (1983) on the use of animals in research. Oxygen microelectrodes with tip diameters of approximately 5 pm, constructed according to the method of Whalen, Riley and Nair (1967), were used to measure oxygen tension in the eye. Recess length averaged 50 ,um to ensure an accurate measurement of PO, at the tip. The microelectrodes exhibited the following charat:teristics: (a) rapid response (90% full response obtained within 2.5 set). (b) linear response over a PO, range of &740 mmHg and (c) slight sensitivity to changes in temperature (3.63%“C-l over a range of l&37’%). The electrodes were constructed with a short shank to minimize bending. The oxygen electrodes were connected to a current amplifier (Transidyne General Corp., Arbor, MI, Model 1201) and strip-chart recorder such that ocular PO, could be continuously monitored and recorded. The electrodes were calibrated before and after each experiment using saline solutions into which pure nitrogen or pure oxygen were bubbled. Experimental data were discarded if the initial and final calibration differed by more than 10%. Fish were anesthetized with tricaine methanesulfonate (MS-222, Ayerst Laboratories Inc, NY), paralyzed with 1 mg tubocurarine chloride administered i.p. and anchored on their side in a saline-filled Plexiglass chamber. A catheter was inserted into the buccal cavity and an aerated, isotonic saline solution was directed across the gills at a controlled flow rate of approximately 500 ml min-I, a ventilatory flow deemed adequate to maintain the normal trout electroretinogram (HoEert and Ubels, 1979). Temperature of the saline bath was maintained at 9fl”C. All ocular 1’0, determinations were made under conditions of normal room light. A small incision was made through the nasal aspect of the cornea and the oxygen microelectrode, shielded by a cannula constructed from a 13 gauge hypodermic needle, was inserted through the incision past the lens and into the vitreous chamber. The electrode tip was positioned in the vitreous humor 2-3 mm from the retina and the Ag-AgCl reference anode was placed in the saline bath. The electrode approached the retinal surface at approximately 90” and was advanced through the retina and choroid in increments of 10pm every 15 sec. Polarizing voltage was maintained at. -@7 V. previously
OSYGES
DIFFV’SIOS
IN THE
TROUT
RETINA
tit I!1
determined to be the plateau region of the current-voltage polarogram. Within this region, small changes in voltage did not significantly affect electrode current. A st’rip-chart recorder was calibrated such that current from the microelectrode was recorded as volume percent oxygen vs. electrode penetration depth. The electrode was advanced through the retinal and choroidal layers until a precipitous fall in PO*, to a level approaching 0 mmHg was noted, whereupon penetration was stopped. A similar t,echnique for the measurement of the retinal PO, of cats in situ has prrviousl? been reported (Alder, Cringle and Constable. 1983). The data were transformed and plot’ted as pob, vs. rlect,rode penetration depth for each 1 pm through the retina and choroid. Values of l’o, at points between each 10 pm sampling point were obtained by linear interpolation. Ocular pob, profiles were also obtained from enucleated ey’es of freshly killed fish. -4 relatively large (approximately 2 mm diameter) hole was made in the sclera near the optic nerve, and the isolated eye was placed cornea down in a. Plexiglass chamber through which a buffered nutrient Ringer solution was circulated. Pure oxygen was bubbled through the solution with temperature maintained at 9 + 05°C. The calibrated oxygen microelectrode was directed through the hole in the sclera then advanced through the choriocapillaris and retina in 10 pm steps while 1’0, was monitored and recorded as in the in vivo preparation. Bending of the elect.rode tip was not observed and therefore assumed not to occur in the in vivo determinations. Oxygen consumption ( vo,) of isolated whole retinas with pigment epithelium intact was measured using a YSI Model 53 Biological Oxygen Monitoring System (Yellow Springs Instrument Co., Yellow Springs, OH). A11 determinations were made in 3 ml of modified Krebs-Ringer-phosphate solution containing 5 mM glucose and saturat,ed wit’h room air. Temperature of the media was maintained at 9 +O.Z”C. Calculation of oxygen consumption was based on the solubility coeficient of oxygen in Ringer’s solution (0.048 ml 0, ml-’ fluid at 1 atm 0, and 10°C) (Vmhreit.. Burris and Stauffer. 1961) and corrected to STPD. Protein determinations were made using the Lowr) method (Oyama and Eagle, 1956). 3. Results A typical in vivo PO, profile of the retina and choroid is illustrated in Fig. 1. Zero depth represents the position of the electrode from which the initial POXmeasurements were made, hut does not necessarily represent the anterior margin of the retina. As electrodes were projected through the retina in the direction of the choriocapillaris, each resulting profile contained a region of rapidly rising PO2which terminated in a sharp inflection (see arrow - Fig. 1). It was assumed that the rapidly rising region corresponded to electrode movement through the retina, and that the sharp inflection occurred just as the electrode penetrated Bruch’s membrane and entered the choriocapillaris. These assumptions were based on the following considerations. (1) A very steep PO, gradient between the vitreous body and the choroidal layer has been reported for teleosts (Fairbanks, Hoffert and Fromm, 1969; Negishi et al., 1975). This steep gradient is deemed necessary to ensure an adequate supply of oxygen for the high metabolic activity of the avascular retinas of these fish. A region of rapidly rising PO,would therefore be predicted as the retina is penetrated in a vitreal-choroidal direction. (2) The method used for obtaining retinal PO,profiles in vitro made possible a direct observation of the electrode tip as it first touched the choriocapillaris. There is close
610
P. E. I)ESKO(‘HEKS
ET AI,.
i
sI 500 E
,s 3ooF/ -
;,;( 400
i
,
n
,nn
/:
,I
, 3nn
, 7
,
7nl-T
,
(
,
,
,,
Ann
Depth (pm)
FK. 1. Tissue oxygen tension as a function of electrode location within the retina and choroid of one teleost fish. The abscissa represents electrode depth measured from the point at which PcXz,was seen to increase. The arrow represents a close approximation of the retinal-choriocapillaris interface. The dashed lines represent the PO, at the choriocapillaris Po, and the posterior retinal margin. Data were obtained at 9 * 05°C.
Teleost retinnl Midpoint depth (W 0
Choriocapillaris Po8teosterior retinal
POPprojiles In vitro ‘b, (mm%)
In viva P (m&Ig)
394+X (a = 15) 381 f32
Predicted P (mmoirg)
(n = “1)
381
288+7 (n = 14) 282&27 (n = 21)
276
-
layer
Visual cell and pigmented epitheli urn Ail,~teric~rretinal layer8 Outer nuclear Outer plexiform Inner nuclear Inner plexiform Ganglion cell
64
137 164 193 235 290
r’zYk3 193*7 152f7 ilO& 85&2
(n (n (n (n (n
= = = = =
4) 3) 3) 7) 7)
213*25 192*25 185k27 156&26 129+24
(a (n (n (n (n
= = = = =
PO) 19) 15) 14) 12)
211 193 176 1.54 132
Values are expressed as mean + SE.. n = number of observations. In vitro and in viva profiles reflect the mean PoS measured at the midpoint of each retinal cell layer. The predicted PO, profile is based upon equation (1) (see text). The depth ofeach cell layer was determined from percentage locations in histological preparations.
agreement between values for PO, of the retinal layers obtained in vitro with those obtained from in vivo ocular PO, profiles (Table I). (3) The distance over which the rapidly rising portion of the in vivo profiles was measured correlated closely with the mean retinal thickness. The mean thickness (324558 pm, mean+s.E.) was determined from 242 measurements made on frozen sections from three trout eyes, i.e. approximately 80 measurements per eye. (4) The current theory of the mechanism of oxygen concentration wit,hin the teleost eye predicts that the highest PO, should occur at the retinal-choriocapillaris interface (Fairbanks et al., 1974). That portion of the profile (Fig. 1) corresponding to the retina was selected and
OXYGEN
DIFFUSION
Ih’ THE
TROIJT
RETINA
61 1
reversed with respect to the abscissa such that the retinal-choriocapillaris interface was now assigned a depth of 0,um. Thus PO, at the retinal-choriocapillaris interface(Pcc) represents the value recorded just prior to the inflection (see arrow. Fig. 1). Values from 21 in vivo retinal profiles were used to plot an average profile of each cell (Fig. 2) and to calculate the mean retinal PO, values at the midpoint layer (Table I). The mean maximal retinal PO, of 381 mmHg at the retinal-choriocapillaris interface decreased to 124 mmHg at the retinal-vitreal interface. The highest observed value of PO, recorded in any eye was 753 mmHg.
0
50
100
150
200
250
300
350
Depth (pm) I--
Posterror--I-----Anterm-----
I
FIG. 2. Mean retinal 16, (calculated from the eyes of 21 tish) as a function of electrode location within the retina. Individual profiles have been reversed with respect to the s-axis, setting the posterior retinal margin to zero depth.
Values for PO2 obtained from profiles measured in retinas of enucleated eyes, i.e. in vitro profiles, are also summarized in Table 1. The mean PO, at the retinalchoriocapillaris interface was 394 mmHg and decreased rapidly as the vitreous was approached. The in vitro profiles were similarly characterized by the large PO-gradient from vitreous body to choriocapillaris. The oxygen consumption ( vo,) of isolated whole trout retinas was 6.98 rf: ti39 (10) ,A 0, hr-l mg--’ protein at 9°C. The average protein concentration was 52.7 + 2.5 (10) mg protein g-l wet weight. To facilitate calculations, the IfoZ values were converted t’o ml 0, min-’ cm-3 using the protein concentration of individual retinas and assuming that 1 g wet tissue equals 1 cm3. The average vo, for 10 determinations was calrulated to be (602 + 0.24) x 10m3 ml 0, mine1 cm-z.
4. Discussion A technique using polarographic microelectrodes was developed for the in vivo measurement of PO,within the various layers of the rainbow trout retina. The average PO, in the choriocapillaris of these eyes (381 mmHg), did not differ significantly from the value reported by Fairbanks et al., 1969. The data obtained in the latter report were from rainbow trout using oxygen electrodes of a different design. The similarity of the two values was taken as an indication that t>he oxygen microelectrodes used in the present’ study were functioning properly.
tilt’
P. E. DESROC’HERS
ET AL
Measurements of retinal PO, in enucleated eyes were made in order to determine if penetration of the eye effected any distortion of the microelectrode tip. Bending of the electrode was not observed and therefore assumed not to occur in t’hr in viva determinations. This in vit’ro technique also allowed for the visualization of cahoriocapillary penetration, which aided in the determination of electrode localization. The in vitro profiles were in very good agreement with the in vivo profiles, lending further support to the assumption that peak ocular PO,occurs at the level of the choriocapillaris. In the in vit’ro preparations the choriocapillaris was exposed to saline equilibrated with 100 7; 0, and the gradient observed is representative of oxygen difl’usion through t hesr eyes. Xo physiological significance should be attached t’o the in vitro retinal profiles with respect to the countercurrent multiplication of oxygen, since the mechanism was not operational in these studies. To mathematically characterize the in vivo retinal Po2 profiles, data were plotted as In (P,,) vs. electrode depth. The resultant curve was divided into six segments, each segment corresponding to one of the retinal layers listed in Table 1. We will refer to the region anterior to the photorecept’or layer as the anterior retina. The term ‘posterior retina’ will be used to designate the region comprised of the pigmented epithelial-photoreceptor cell layer. The slope of the PO, profile for the posterior area was significantly greater than that for the anterior retina and the slopes for the anterior layers did not differ from each other. These results suggest that the mean in vivo retinal profile could be described by a two-component exponential function. [‘sing the least squares and Gauss minimization methods the following equation was determined.
PC,, = PC,,, (0207 e-“‘0178z+ 0.793 e~“‘oo2sx).
(1)
where: Pccc, = the oxygen tension of choriocapillary blood in mmHg, x = distance from the choriocapillaris in pm, Pcz. = the oxygen tension of the retina at x .urn from the choriocapillaris. For any given choriocapillary oxygen tension Equation (1) can be used to calculate PO,values for each layer of the teleost retina. Values for the retinal PO,in the individual cell layers, assuming a choriocapillary PO, of 381 mmHg, were calculated and are listed as ‘predicted’ values in the extreme right hand column of Table I. interface was calculated to be 118 mmHg, which The mean PO,at the retinal-vitreal confirms that all cell layers of the trout retina are normally exposed to oxygen tensions in excess of that found in the arterial blood, approximately 80 mmHg. Much of the trout retina, particularly the photoreceptors and pigment epithelial cells, is exposed to oxygen tensions comparable to those known to exert toxic effects on the retinas of other species. It has been hypothesized that resistance to the toxic effects of high oxygen has evolved in the retina and choroid of trout eyes in response to the high 0, levels to which they are continuously exposed (Desrochers and Hoffert. 1983). Research on the presence of protective mechanisms in trout eyes against oxygen toxicity is currently underway in our laboratory. Although equation 1 is useful as a predictor of the tissue PO, at any point within the retina of the trout, its derivation was completely empirical and it sheds little light on the nature of oxygen delivery and/or utilization of oxygen within this tissue. The retina is a complex network of cellular layers and oxygen diffusion will be influenced by many factors, including the lipid composition, solubility of oxygen and rate of 0, consumption within a particular layer. At this time it is not possible to predict the relative contribution of each of these factors in the in vivo situation. However, the
OXYGEN
DIFFUSION
IN THE
TROIJT
RETINA
613
spatial resolution of the PO, electrodes used in this study was sufficient to enable us to subdivide the retina into anterior and posterior regions. To gain some understanding of oxygen delivery and/or utilization within each region. the data were subjected to an analysis based upon the classical equation for planar diffusion (Crank, 1975; Hill, 1928). Alder et al. (1983) have stated that the existence of both an anterior and posterior oxygen supply to the retina, characteristic of most mammalian eyes, complicates the mathematical analysis of the retinal oxygen profile. Alternatively. the trout retina presents itself as an excellent model for studying oxygen diffusion in vivo since this tissue has a single source of oxygen. The mathematical analysis is therefore somewhat simplified and the following solution for this model is presented. Since we measured Paz, appropriate substitutions into the classical equation have been made and the working equation becomes:
d2P dP Sdt = DS d22-
. 110~’
where: DS = the Krogh permeation coefficient? and D and R are the respective diffusion and solubility coefficients for oxygen, P = the partial pressure for oxygen. x = ,um anterior from the choriocapillaris and, $‘o, = oxygen consumption in ml 0, min-’ cmw3. In the steady state condition with which we are dealing, dP/dt = 0 and the general form of the solution (Simon, 1972) is:
pi(x) = K@+Bix+Ci where:
Pi(x) = the partial
pressure for oxygen
(i = 1, 2),
(3)
at any given depth x.
Ki = Volb,(i,/2DS(i, and the parameters Bi and C, are constants dependent on the boundary conditions, with Ci = the PO, at the choriocapillaris. Three assumptions must hold true in order that the solution of this equation be valid: (1) the tissue should be considered a homogeneous medium, (2) poZ should be constant and independent of PO, for all oxygen tensions above zero and (3) the diffusion and solubility coefficients should be constant and independent of PO, (Ganfield, Whalen and Nair, 1970). In the situation that we are investigating, oxygen is considered to be diffusing in only one dimension through a tissue of finite thickness. Therefore at no depth within the tissue will the PO, drop to zero. The condition that PO, does not decrease to zero does not preclude the use of equation (3) as long as a second boundary with a known PO, is established at the anterior margin of the retina. The teleost retina cannot be considered to be homogeneous throughout, a fact supported by the values for oxygen consumption reported below. Our data also indicate that a logarithmic transformation of the PO, profile through the retina is not’ linear. As stated earlier, we have divided the retina into two regions. The posterior region begins at the choriocapillaris and extends to a depth of 127 ym into the retina. while the anterior portion begins at a depth of 128 ym and terminates at the retinal-vitreal interface. The thickness of the posterior region of the retina was determined histologically from the percent thickness of the retinal cell layers using a mean retinal thickness of 324 f 5% pm (IL = 3). The solution for planar diffusion of oxygen in the two regions of the trout retina therefore necessitates the establishment of two sets of boundary conditions and the solution of simultaneous differential equations (see appendix). “I EEH41
(i I 4 The anterior
I'. I':. I)ESSt~O('HEIIS
boundary
of the retina
ET AI.
(retinal-vitreous interface), was determined to profile approaches an asympt.otic value. The d&e&nation of this boundary was based on the vrry low l&, of the vitreous body and the extremely long path for the diffusion of oxygen through the vit,reous. Equation (3) was applied to the posterior and anterior portions of the profiles from value at the retinal- vi&al 16 eyes in which the /:,2 dropped to an asymptotic, interface. The consta.nt,s Ki and K, of equat,ion (3) were determined using a least 1969) for the two regions of squares fit to the second order polynomial (Bevington. the retina. Oxygen consumption for the two regions of t’he retina was estimated from values reported in the literature. For the teleost retina, Santamaria, Drujan. Svaetichin and ’ Negishi (1971) determined @$‘to be4.3 ,~l hY’ mg e-1dry tissue while the ‘receptor-free retina was 2.63 ~1 hr-’ rng-’ dry tissue (the latter is a correction of an apparent’ incorrec:t value in the original article). In order to estimate that fraction oftotal oxygen consumption that can be attributed to the anterior region of the retina we must first t,ake into account the volume occupied by that area. From measurements of retinal depth we determined that the anterior region of the retina, which corresponds to t,he ‘receptor-free’ retina occupies 60 “b of the total volume. Using this SOY,, value and assuming that 1 mm” of tissue is equal to 1 mg dry weight we ohtained a value for the oxygen consumpt’ion of the anterior region of the retina. Per unit volu~ne of retinal tissue. the anterior region ofthe retina consumes oxygen at a rate of l-58 ,uI hr.-‘. which represents 36.7 ok1of Ma1 retinal consumption, i.e. 36.7 O,clof 4.3 ,~l hr. ‘. We measured the rate of oxygen consumption for bhe entire retina and found it to be 6.02 ,~l Inin-’ cme3. Knowing this value plus the relative size and the fraction of oxygen consumption attributed t.o the anterior and posterior regions of the trout retina. we ran now calculate the rates of oxygen consumption for these two regions. Since the post,erior retina accounts for 63.3 ‘:, of t)otal retinal oxygen consumption and it, occupies 40 O,,,of the ret’inal volume, the calculated rate of oxygen consumption for this region is 9.53 ,LLIm&l cm -3. Similarly the rat’e of oxygen csonsumption for the anterior retina, which contributes 36.7 ‘to vf t.otal retinal oxygen consumption and occupies 60 O0 o f retinal volume. is 3.67 ,uI min-’ crnP3, These rates were used to calculate the Krogh permeation coefficient for the respect,ive layers of the trout retina. The average value (16 eyes) of the Krogh permeation coefficient at 9°C’ for the posterior retina (US,,, Table II) was calculated to be l.H6+_cP33 (X IO-‘), which is significantly higher than t,he value of 0.58 + 0.06 ( x 10d5) ml 0, min ’ cm-’ atrn-’ for the anterior retina (DA',,. Table II). It should be emphasized that these estimates of DS are based on data collected at 9”(’ and that the temperature dependence of DA' should be taken into account, when comparing these values with those obtained 1’) others. In his discussion of oxygen transport in tissues. Thews (1968) reported that DS values at 37°C for skeletal and cardiac, muscle. cerebral cortex and plasma from a variety of species, ranged from 0.7 x 10-5--3.6 x 10e5 ml 0, min-’ cm‘.* at,m-‘. A DS value of 129 x 10e5 ml 0, mine1 cm-’ atm-’ has heen reported for cat, cerebral cortex at 37°C (Ganfield et al., 1970). With a Q1, of 1.1, which was calculated using the 1115’ values for frog muscle at 20 and 37V (Altman and Dittmer, 1971). the BS’ values for the retina of the trout when corrected to 37°C are: Ds,, = 243 x 1W5 and DS,,= 0.76 x 10V5 ml 0, min-’ cm-’ atn-‘. These values are within the reported range of IX! for all tissues and are in good agreement with the values reported for the neural tissue of other species. For example. values for the cerebral vortices of rabbit and rat were 1.9 x 1O-5 and 2.1 x 10-j ml 0, min-’ cn-’ atm-’ rrsprctively. A IM be in the region where t!he slope of the oxygen
OXYGEN
DIFFUSIOK
I?r’ THE
TROYT
RETIS.
615
TABLE II
Second orderpolynomialfit to individual in vivo retinal oxygenprofilesfor the determination of the Krogh permeation coeflicientfor the posterior and anterior retina of the rainbow trou,t, DS,, and DSAR respectively Posterior Eyr 1 2 3 4 -5 6 7 8 9 10 11 12 13 l-4 15 16
retina
(1,
6,
441 238 497 261 315 531 341 250 324 359 398 543 206 207 379 188
2.31 4.80 1.49 3.39 6.20 1.50 2.16 1.10 301 415 933 2.20 062 1.14 2.51 2.76
Anterior retina __--__ DS,, x 105 (h K, ns,, x 103 -.-. ~~ -~- __~~~ .-... -~~ .~~1.57 309 24.5 0.57 0.76 39 7.33 0.19 243 328 2.27 0.61 I .07 127 2.29 W6l 0% 12.5 2.56 0.55 ‘41 417 1.92 073 1.68 180 2.98 047 329 69 6% 02”k 1.20 244 1.79 ti78 087 159 3.55 0.37 0.39 251 147 (h95 1.79 370 3.55 0.39 0.58 105 2.60 0.54 032 14X 1.no o-73 014 174 1.15 0.1% 0.13 45 3.13 0.45
P&r) = KiX2+Bix+Ci (i = 1, 2), where e(z) = the partial pressure for oxygen at any given depth I, s = hm anterior from choriocapillaris, ‘fi = $‘ob,Co(2DS)& where Kt above is reported in mmHg pmeY. Vo, = oxygen consumption in ml min-’ cmm3, DS = Krogh permeation coefficient (ml 0, min-’ cm-’ atm-‘), R, and C, = constants with C, equal to the r’oe at the choriocapillaris
value of 3-24 x lop5 ml 0, min-’ cm-’ atm -l has been reported for the retina of the honeybee drone, Apis mellifera, measured at 22°C (Tsacopoulos. Poitry and Borsellino. 1981). The calculated values for the trout retina at this temperature are: DS rR = 2.11 x 10e5 and DS,, = 0.66 x 10m5 ml 0, min-’ cm-’ atm-‘. Differences between DS values for these two species would be expected based on the high degree of structural dissimilarity between the retina of the trout and that of the compound eye of the drone. The data as derived above clearly indicate that the diffusion of oxygen through the teleost retina is not linear. Previous investigators (Tsacopoulos, Baker and Levy, 1976: Zuckerman and Weiter, 1980; Alder et al., 1983) have also provided evidence that oxygen transport rates vary within the retinas of the pig, bullfrog and cat, respectively. Our calculat’ions would indicate that the higher rate of diffusion through the posterior layer is likely, due to the higher rate of oxygen consumption by photoreceptor cells contained in this region, Since no reliable estimate of 0, solubility (S) is available for the retinal cell layers, determinations of the diffusion coefficients (0) have not been made. The lipid content of the different retinal layers would be expected to be important in determining oxygen solubility in these regions. In summary, a technique was developed for measurement of intraretinal PO, of the rainbow trout in vivo. Determination of the intraretinal PO, profile confirmed that all layers of the trout retina are continuously exposed to oxygen tensions in excess of “I.2
1’. E. DESKO(:H
616
ICKS ET AI,
values shown to be toxic to the retinas of other species. The resistancle to osygen toxicity exhibited by this tissue thereby provides us with an excellent model for thr study of oxygen toxicity and the agents which protect. against it. The mammalian retinal PO, (Tsacopoulos et al., 1976; Alder et al., 1983) differs from that reported fol the trout in that the mammalian profile tends to be parabolic in nature. This reflects the fact that the mammalian retinas receive oxygen from both retinal and choroidal circulations. This dual supply ofoxygen precludes calculation ofpermeat’ion coefficients for oxygen diffusion through these retinas, while trout retinas, which are avascular and have a single source of oxygen, are ideal for studies of this nature. For t,he trout retina, analysis of 16 profiles yielded two statistically different values of the Krogh permeation coefficient for oxygen. We obtained values of 1.86 x 10e5 and 0.58 x region and remaining 10M5 ml 0, mine1 cm-l atm-’ at 9°C for th e photoreceptor neural retina respectively. ACKNOWLEDGMENTS This work was supported by NIH grant EY-00009 and Michigan Agricultural Experiment station, Project 1441 H (Journal Article Number 10801). The authors wish to thank Esther Brenke for her expert technical assistance and Dr William F. Jackson. Medical College of Georgia, Augusta, GA for his valuable assistance in the mathematical analyses, APPENDIX The determination of the Krogh permeation coefficient for oxygen under the conditions of unidirectional planar diffusion in steady state is presented below. The model deals with a tissue of finite depth (i.e. oxygen tension does not fall to zero) and comprises two regions with differing vo,. Within each region it is assumed that the vo, is homogeneous, constant and independent of PO,. The posterior region of the retina extends from 5 = 0 at the choriocapillaris to .z, = 127 pm which includes the pigment epithelium and visual cells. The anterior region extends from z1 to z2 located at the retinal-vitreal interface. In region z = 0 to x = zi : d2P,/dx2 = voio,,/nS, = 2K,. (1) In region 5 = xl to 5 = x2: d2P2/dx2 = vo;z2/DS2 = 2K2.
(2)
Boundary conditions: PI(O) = Pee = PO, at choriocapillaris,
and
(3)
Pz(x2) = PAR = PO, at boundary margin of anterior retina,
(4)
Wdx, - 1, = dP,/dz, = rl PI(q) = PZ(rl) (i.e. flux between the two layers must be equal).
69
(6)
Solutions :
Therefore applying boundary Boundary condition 1:
P,(x) = K,x2+Blx+C1,
(7)
Pz(x) = K,x2+B2x+CZ.
(8)
conditions to obtain constants Bi and C, :
P,(O) = Pee = K,(0)2+B1(0)+C1 therefore :
PI(s) = h;(.~)~ + B, xs Pee.
= Cl,
(9) (10)
Boundary condition 2: &(x2) = PAR = K,x,~ + B,x, + Cr,,
(11)
OSYGES therefore :
DIFFUSIOK
IN THE TROUT
P,(x) = K2(x*-x122)+B2(x-x2)
RETJXA
+P*R.
(iii
(II)
condition 3 :
Boundary
at .r = x1’ therefore : Bwndary
condition
dP,/dx = dP,/dx.
(13)
B, = 2K, x1 - 2K, x1 + B,.
(14)
p,b-I) = &4x,).
(15)
4:
Tn addition : therefore :
B,
Then substituting
(16)
= K,(x~~x~)-K~(~~+~~/~~)+(P~~-P~c)I~~.
for B, in equation (14) and rearranging: B, = K,(xf/x,-2x,)+K,(22,-XT/X,-x,)+(P,,-P&/x,.
Now substituting
the appropriate
(17)
constants into equations (7) and (8) and simplifying,
PI(x) = ~‘,X~+~K~(~~/~~-~~~)+K~(~T~-X~/Z~-~~)+(P~~ZR-P~~)/Z~JZ~P~~, Rx)
= K2x2 + [K,W4
-4(x,+x:/x2)
+ (Pa
yields: (18)
-Pcc)/zzlr+P~,-K,t~-K,x:.
(IF))
(i = 1. 2)
(W
Thus both equations are in the form e(x)
= &z~+B~x+C’~
where K, = p001,/2DS,, K, = ~obp2/2DS,, and Bi, Cti are constants. Each region is then fit to the polynominal, qx, = Kxa+Bx+(’
and the coefficient K used to calculate L)S for each region.
REFERENCES Alder, V. A., Cringle, S. J. and Constable. I. J. (1983). The retinal oxygen profile in cats. Invest. Ophthalmol.
Vis. Sci. 24, 30-6.
Altman,
P. 1~. and Dittmer, D. S., Eds. (1971). Biologica Handbooks. Respiration and Circulation. Pp. 21-3. FASEB Press: Bethesda, MD. ARVO resolution on the Use of Animals in Research (1983). Invest. Ophthalmol. l’iu. s’ci. 24. 1156. Bevington, P. R. (1969). In Data Reduction and Error Analysis for the Physical Sciences. Pp. 13444. McGraw-Hill: New York. Crank, J. (1975). In The Mathematics of Diffusion (2nd edn). Pp. 4468. Oxford University Press: Oxford. Desrochers, P. E. and Hoffert, J. R. (1983). Superoxide dismutase provides protection against the hyperoxia in the retina of the rainbow trout. (Sulmo gairdneri). Camp. Biochem,. Physiol.
76B, 241-7.
Drujan, B. D., Svaetichin, G. and Negishi, Ii. (1971). Retinal aerobic metabolism as reflec.ted in S-potential behavior. Vision Res. 3, 151-9. Fairbanks, M. B., Hoffert, J. R. and Fromm, P. 0. (1969). The dependence of the oxygenconcentrating mechanism of the teleost eye (Salmo gairdneri) on the enzyme carbonic anhydrase. J. Gen. Physiol. 54, 203-l 1. Fairbanks, M. B., Hoffert, J. R. and Fromm, P. 0. (1974). Short circuiting the ocular oxygen-concentrating mechanism in the teleost Salmo gairdneri using carbonic anhydrasc inhibitors, J. Gen. Physiol. 64, 263-73. Ganfield, R. A., Whalen, W. J. and Nair, P. (1970). Mass transfer, storage and utilization of 0, in cat cerebral cortex. Am. J. Physiol. 219, 814-21. Hill. A. V. (1928). The diffusion of oxygen and lactic acid through tissues. Proc. R. Sot. London Ser. B. 104, 3996. Hoffert, J. R. and Ubels, J. L. (1979). The intraocular PO%and electroretinogram of the trout as affected by temperature and ventilatory flow. f’omp. Biochem. Physiol. 62A. 563-68.
P. E. DESRO(‘HERS
618
ET AL
Kuhn, W., Ramel, A., Kuhn, H. J. and Marti. E. (1963). The filling mechanism of the swimbladder. Generation of high gas pressures through hairpin countercurrent multiplication. Experientia 19. 497-552. and Negishi, Ii.. Svaetichin. G., Laufer, M. and Drujan, B. 1). (1975). Polarogmphic electrophysiological studies of retinal respiration. Vision Res. 15, 527-33. Oyama. V. I. and Eagle, H. (1956). Measurement of cell growth in tissue culture with a phenol reagent (Folin-Ciocalteau). Proc. Sot. ICxp. Biol. Med. 91, 305-7. Santamaria, L.. Drujan. B. D., Svaetichin, G. and Negishi, K. (1971). Respiration, glycolysis and S-potentials in teleost retina: a comparative study. Vision Rea. 11, 877-87. Simon. W. (1972). Mathematical Techniquesfor Physiology and Medicine. Pp. 41-60. Academic Press: New York. Thews, G. (1968). The theory of oxygen transport and its application to gaseous exchange in the lung. In Oxygen Transport in Blood and Tissue (Eds Lubbers. D. W., Luft. IT. f’.. Thews, G. and Witzleb. E.). Pp. l-20. Georg Thieme Verlag: Stuttgart. Tsacopoulos, M., Baker, R. and Levy, S. (1976). Studies on retinal oxygenation. Adv. Exp. Med. Biol. 75, 413-6. Tsacopoulos. M., Poitry, S. and Borsellino, A. (1981). Diffusion and consumption of oxygen in the superfused retina of the drone (Apis mellifera) in darkness. J. Gen. Physiol. 77. 601-28. Umbreit, W. W., Burris, R. H. and Stauffer, J. F.(Eds) (1964). Manometric Techniques (4th edn). Pp. 5. Burgess: Minneapolis. Whalen, W. J., Riley, J. and Nair, P. (1967). A microelectrode for measuring intracellular Paz. J. Appl.
Physiol.
23, 798401.
Wittenberg, J. B. and Wittenberg, B. A. (1961). Active transport of oxygen and the eye of fish. Biol. Bull. 121, 379. Zuckerman, R., and Weiter, J. J. (1980). Oxygen transport in the bullfrog retina. Exp. Eye Res. 30, 117-27.
Oxygen PAUL
Diffusion
E. DESROCHERS, AND
~fl Physiology,
Lkpartment
in the Trout
KENNETH @J. RCSSELL
Michigan
A. PRATT.
Retina PAUL
0. FROMM
HOFPERT
State T!niversity. r.s.il.
East Lansing,
MI 438%4-1101.
(Recei~wd 4 March 1985 and accepted 29 August 19X5, IVruj York) Oxygen tensions were measured in viva within the different layers of the rainbow trout retina. Oxygen microelectrodes were advanced in 1Opm increments through the retinas and the Po, measured at each location. Mean retinal F’o, ranged from 124 mmHg at the retinal-vitreal interface to 381 mmHg at the choriocapillaris. These data reaffirm that the cellular layem of the normal trout retina are continually exposed to supra-arterial oxygen tensions that arc known to cause toxicity in other species. The intraretinal Po, gradient was mathematically characterized and results indicate that the in vivo oxygen profile can be described by a two-component exponential function. In order to gain a better understanding of oxygen delivery to the retina. the data were also subjected to an analysis based upon the classical equation for planar diffusion. The trout retina is an excellent model for studying oxygen diffusion in viva since this tissue is supplied with oxygen from a single source, thereby simplifying the mathematical analysis. C?alculations yielded values of 1.86 x 10m5and 0.58 x 10e5 ml 0, min-’ cm- I atm-’ (at 9°C) for the Krogh permeation coefficient (DS) for the photoreceptor region and the remaining neural retina, respectively. fYel/ mordh: hyperoxia: Krogh permeation coefficient; oxygen: oxygen diffusion; rainbow trout retina.
1. Introduction Exposure of animals to oxygen tensions in excess of that found in air may cause both structural damage and cellular dysfunction in various tissues and/or organs. The retina and choriocapillaris of the mammalian eye are especially vulnerable to the toxic effects of hyperoxia. As a consequence. much attention has been directed towards studies of the mechanism of oxygen toxicity and agents which protect against it. In light of the above, it seemed incredible when Wittenberg and Wittenberg (1961) reported that the choroidal rete mirabile of many teleosts. in conjunction with the choriocapillaris, is capable of concentrating oxygen to levels far in excess of those found in arterial blood. Several investigators recognized the potential of the teleost eye preparation as a model for the study of oxygen toxicity and studies have been done to describe the oxygen-concentrating mechanism in the rhoroid and the nature of oxygen delivery to the teleost retina. Fairbanks, Hoffert and Fromm (1974) proposed a mechanism to describe how the choroidal rete mirabile of the rainbow trout (flalmo gairdneri) functions to elevate ocular PO, via countercurrent oxygen multiplication. The mechanism is similar to that proposed for the secretory epithelium of gas bladders of certain physoclistous teleosts (Kuhn, Ramel, Kuhn and Marti, 1963). Measurements of the PO, in the choroid layer of rainbow trout eyes showed it to average more than 400 mmHg (Fairbanks, Hoffert and Fromm, 1969), while it was approximately 100 mmHg at the retinal-vitreal @ Dr Jack R. Hoffert suffered a fatal heart attack on 21 November 1984. This paper is dedicated to his memory. Please address all correspondence and requests for reprints to Paul E. Desrochers. Department of Physiology, Michigan State University, East Lansing, MI 48824-l 101, IT.S.A. 001&4835/85/l
10607 + 12 $03.00/O
@z$ 19X5 Academic- Prrss Inc. (London)
I,imit,rd
m4
F'. E. I)IC:SI
interface. Thus there is a large I<;,, g radient wit,hin the retina of this sl)rc*ic~s. ‘I’bts distribution of PO, within isolated teleost retinas has been reported by, lkujan. Svaetichin and Negishi (1971) and Segishi. Svaeitichin. Laufer and I)ruja,n (19i5). The present’ report however, is the first in which the 16, was measured in retinas ot intact fish which possess a functioning countercurrent oxygen multiplier. The aims of the research reported here were to develop a t’echniyue to measure t,hc Po, of the rainbow trout eye in vivo and determine t,hose areas of the retina that are exposed to oxygen tensions generally expected t.o be boxic. The data obt.ained were used to characterize mathematically the intraretinal pob,gradient and gain insight’ into the nature of delivery and/or utilization of oxygen in this tissue. Values were calculated for the Krogh permeation coefficient for t’he diffusion of oxygen through (a) the pigmented epithelial-photoreceptor cell layer and (b) the remaining neural retina. 2. Materials
and Methods
weighing 15&300 g were obtained from Midwest Rainbow trout (Salmo gairdneri) Fish Farms, Inc., Harrison, MI and held in the laboratory at 9+ 1°C. All studies were carried out in accordance with the ARVO Resolution (1983) on the use of animals in research. Oxygen microelectrodes with tip diameters of approximately 5 pm, constructed according to the method of Whalen, Riley and Nair (1967), were used to measure oxygen tension in the eye. Recess length averaged 50 ,um to ensure an accurate measurement of PO, at the tip. The microelectrodes exhibited the following charat:teristics: (a) rapid response (90% full response obtained within 2.5 set). (b) linear response over a PO, range of &740 mmHg and (c) slight sensitivity to changes in temperature (3.63%“C-l over a range of l&37’%). The electrodes were constructed with a short shank to minimize bending. The oxygen electrodes were connected to a current amplifier (Transidyne General Corp., Arbor, MI, Model 1201) and strip-chart recorder such that ocular PO, could be continuously monitored and recorded. The electrodes were calibrated before and after each experiment using saline solutions into which pure nitrogen or pure oxygen were bubbled. Experimental data were discarded if the initial and final calibration differed by more than 10%. Fish were anesthetized with tricaine methanesulfonate (MS-222, Ayerst Laboratories Inc, NY), paralyzed with 1 mg tubocurarine chloride administered i.p. and anchored on their side in a saline-filled Plexiglass chamber. A catheter was inserted into the buccal cavity and an aerated, isotonic saline solution was directed across the gills at a controlled flow rate of approximately 500 ml min-I, a ventilatory flow deemed adequate to maintain the normal trout electroretinogram (HoEert and Ubels, 1979). Temperature of the saline bath was maintained at 9fl”C. All ocular 1’0, determinations were made under conditions of normal room light. A small incision was made through the nasal aspect of the cornea and the oxygen microelectrode, shielded by a cannula constructed from a 13 gauge hypodermic needle, was inserted through the incision past the lens and into the vitreous chamber. The electrode tip was positioned in the vitreous humor 2-3 mm from the retina and the Ag-AgCl reference anode was placed in the saline bath. The electrode approached the retinal surface at approximately 90” and was advanced through the retina and choroid in increments of 10pm every 15 sec. Polarizing voltage was maintained at. -@7 V. previously
OSYGES
DIFFV’SIOS
IN THE
TROUT
RETINA
tit I!1
determined to be the plateau region of the current-voltage polarogram. Within this region, small changes in voltage did not significantly affect electrode current. A st’rip-chart recorder was calibrated such that current from the microelectrode was recorded as volume percent oxygen vs. electrode penetration depth. The electrode was advanced through the retinal and choroidal layers until a precipitous fall in PO*, to a level approaching 0 mmHg was noted, whereupon penetration was stopped. A similar t,echnique for the measurement of the retinal PO, of cats in situ has prrviousl? been reported (Alder, Cringle and Constable. 1983). The data were transformed and plot’ted as pob, vs. rlect,rode penetration depth for each 1 pm through the retina and choroid. Values of l’o, at points between each 10 pm sampling point were obtained by linear interpolation. Ocular pob, profiles were also obtained from enucleated ey’es of freshly killed fish. -4 relatively large (approximately 2 mm diameter) hole was made in the sclera near the optic nerve, and the isolated eye was placed cornea down in a. Plexiglass chamber through which a buffered nutrient Ringer solution was circulated. Pure oxygen was bubbled through the solution with temperature maintained at 9 + 05°C. The calibrated oxygen microelectrode was directed through the hole in the sclera then advanced through the choriocapillaris and retina in 10 pm steps while 1’0, was monitored and recorded as in the in vivo preparation. Bending of the elect.rode tip was not observed and therefore assumed not to occur in the in vivo determinations. Oxygen consumption ( vo,) of isolated whole retinas with pigment epithelium intact was measured using a YSI Model 53 Biological Oxygen Monitoring System (Yellow Springs Instrument Co., Yellow Springs, OH). A11 determinations were made in 3 ml of modified Krebs-Ringer-phosphate solution containing 5 mM glucose and saturat,ed wit’h room air. Temperature of the media was maintained at 9 +O.Z”C. Calculation of oxygen consumption was based on the solubility coeficient of oxygen in Ringer’s solution (0.048 ml 0, ml-’ fluid at 1 atm 0, and 10°C) (Vmhreit.. Burris and Stauffer. 1961) and corrected to STPD. Protein determinations were made using the Lowr) method (Oyama and Eagle, 1956). 3. Results A typical in vivo PO, profile of the retina and choroid is illustrated in Fig. 1. Zero depth represents the position of the electrode from which the initial POXmeasurements were made, hut does not necessarily represent the anterior margin of the retina. As electrodes were projected through the retina in the direction of the choriocapillaris, each resulting profile contained a region of rapidly rising PO2which terminated in a sharp inflection (see arrow - Fig. 1). It was assumed that the rapidly rising region corresponded to electrode movement through the retina, and that the sharp inflection occurred just as the electrode penetrated Bruch’s membrane and entered the choriocapillaris. These assumptions were based on the following considerations. (1) A very steep PO, gradient between the vitreous body and the choroidal layer has been reported for teleosts (Fairbanks, Hoffert and Fromm, 1969; Negishi et al., 1975). This steep gradient is deemed necessary to ensure an adequate supply of oxygen for the high metabolic activity of the avascular retinas of these fish. A region of rapidly rising PO,would therefore be predicted as the retina is penetrated in a vitreal-choroidal direction. (2) The method used for obtaining retinal PO,profiles in vitro made possible a direct observation of the electrode tip as it first touched the choriocapillaris. There is close
610
P. E. I)ESKO(‘HEKS
ET AI,.
i
sI 500 E
,s 3ooF/ -
;,;( 400
i
,
n
,nn
/:
,I
, 3nn
, 7
,
7nl-T
,
(
,
,
,,
Ann
Depth (pm)
FK. 1. Tissue oxygen tension as a function of electrode location within the retina and choroid of one teleost fish. The abscissa represents electrode depth measured from the point at which PcXz,was seen to increase. The arrow represents a close approximation of the retinal-choriocapillaris interface. The dashed lines represent the PO, at the choriocapillaris Po, and the posterior retinal margin. Data were obtained at 9 * 05°C.
Teleost retinnl Midpoint depth (W 0
Choriocapillaris Po8teosterior retinal
POPprojiles In vitro ‘b, (mm%)
In viva P (m&Ig)
394+X (a = 15) 381 f32
Predicted P (mmoirg)
(n = “1)
381
288+7 (n = 14) 282&27 (n = 21)
276
-
layer
Visual cell and pigmented epitheli urn Ail,~teric~rretinal layer8 Outer nuclear Outer plexiform Inner nuclear Inner plexiform Ganglion cell
64
137 164 193 235 290
r’zYk3 193*7 152f7 ilO& 85&2
(n (n (n (n (n
= = = = =
4) 3) 3) 7) 7)
213*25 192*25 185k27 156&26 129+24
(a (n (n (n (n
= = = = =
PO) 19) 15) 14) 12)
211 193 176 1.54 132
Values are expressed as mean + SE.. n = number of observations. In vitro and in viva profiles reflect the mean PoS measured at the midpoint of each retinal cell layer. The predicted PO, profile is based upon equation (1) (see text). The depth ofeach cell layer was determined from percentage locations in histological preparations.
agreement between values for PO, of the retinal layers obtained in vitro with those obtained from in vivo ocular PO, profiles (Table I). (3) The distance over which the rapidly rising portion of the in vivo profiles was measured correlated closely with the mean retinal thickness. The mean thickness (324558 pm, mean+s.E.) was determined from 242 measurements made on frozen sections from three trout eyes, i.e. approximately 80 measurements per eye. (4) The current theory of the mechanism of oxygen concentration wit,hin the teleost eye predicts that the highest PO, should occur at the retinal-choriocapillaris interface (Fairbanks et al., 1974). That portion of the profile (Fig. 1) corresponding to the retina was selected and
OXYGEN
DIFFUSION
Ih’ THE
TROIJT
RETINA
61 1
reversed with respect to the abscissa such that the retinal-choriocapillaris interface was now assigned a depth of 0,um. Thus PO, at the retinal-choriocapillaris interface(Pcc) represents the value recorded just prior to the inflection (see arrow. Fig. 1). Values from 21 in vivo retinal profiles were used to plot an average profile of each cell (Fig. 2) and to calculate the mean retinal PO, values at the midpoint layer (Table I). The mean maximal retinal PO, of 381 mmHg at the retinal-choriocapillaris interface decreased to 124 mmHg at the retinal-vitreal interface. The highest observed value of PO, recorded in any eye was 753 mmHg.
0
50
100
150
200
250
300
350
Depth (pm) I--
Posterror--I-----Anterm-----
I
FIG. 2. Mean retinal 16, (calculated from the eyes of 21 tish) as a function of electrode location within the retina. Individual profiles have been reversed with respect to the s-axis, setting the posterior retinal margin to zero depth.
Values for PO2 obtained from profiles measured in retinas of enucleated eyes, i.e. in vitro profiles, are also summarized in Table 1. The mean PO, at the retinalchoriocapillaris interface was 394 mmHg and decreased rapidly as the vitreous was approached. The in vitro profiles were similarly characterized by the large PO-gradient from vitreous body to choriocapillaris. The oxygen consumption ( vo,) of isolated whole trout retinas was 6.98 rf: ti39 (10) ,A 0, hr-l mg--’ protein at 9°C. The average protein concentration was 52.7 + 2.5 (10) mg protein g-l wet weight. To facilitate calculations, the IfoZ values were converted t’o ml 0, min-’ cm-3 using the protein concentration of individual retinas and assuming that 1 g wet tissue equals 1 cm3. The average vo, for 10 determinations was calrulated to be (602 + 0.24) x 10m3 ml 0, mine1 cm-z.
4. Discussion A technique using polarographic microelectrodes was developed for the in vivo measurement of PO,within the various layers of the rainbow trout retina. The average PO, in the choriocapillaris of these eyes (381 mmHg), did not differ significantly from the value reported by Fairbanks et al., 1969. The data obtained in the latter report were from rainbow trout using oxygen electrodes of a different design. The similarity of the two values was taken as an indication that t>he oxygen microelectrodes used in the present’ study were functioning properly.
tilt’
P. E. DESROC’HERS
ET AL
Measurements of retinal PO, in enucleated eyes were made in order to determine if penetration of the eye effected any distortion of the microelectrode tip. Bending of the electrode was not observed and therefore assumed not to occur in t’hr in viva determinations. This in vit’ro technique also allowed for the visualization of cahoriocapillary penetration, which aided in the determination of electrode localization. The in vitro profiles were in very good agreement with the in vivo profiles, lending further support to the assumption that peak ocular PO,occurs at the level of the choriocapillaris. In the in vit’ro preparations the choriocapillaris was exposed to saline equilibrated with 100 7; 0, and the gradient observed is representative of oxygen difl’usion through t hesr eyes. Xo physiological significance should be attached t’o the in vitro retinal profiles with respect to the countercurrent multiplication of oxygen, since the mechanism was not operational in these studies. To mathematically characterize the in vivo retinal Po2 profiles, data were plotted as In (P,,) vs. electrode depth. The resultant curve was divided into six segments, each segment corresponding to one of the retinal layers listed in Table 1. We will refer to the region anterior to the photorecept’or layer as the anterior retina. The term ‘posterior retina’ will be used to designate the region comprised of the pigmented epithelial-photoreceptor cell layer. The slope of the PO, profile for the posterior area was significantly greater than that for the anterior retina and the slopes for the anterior layers did not differ from each other. These results suggest that the mean in vivo retinal profile could be described by a two-component exponential function. [‘sing the least squares and Gauss minimization methods the following equation was determined.
PC,, = PC,,, (0207 e-“‘0178z+ 0.793 e~“‘oo2sx).
(1)
where: Pccc, = the oxygen tension of choriocapillary blood in mmHg, x = distance from the choriocapillaris in pm, Pcz. = the oxygen tension of the retina at x .urn from the choriocapillaris. For any given choriocapillary oxygen tension Equation (1) can be used to calculate PO,values for each layer of the teleost retina. Values for the retinal PO,in the individual cell layers, assuming a choriocapillary PO, of 381 mmHg, were calculated and are listed as ‘predicted’ values in the extreme right hand column of Table I. interface was calculated to be 118 mmHg, which The mean PO,at the retinal-vitreal confirms that all cell layers of the trout retina are normally exposed to oxygen tensions in excess of that found in the arterial blood, approximately 80 mmHg. Much of the trout retina, particularly the photoreceptors and pigment epithelial cells, is exposed to oxygen tensions comparable to those known to exert toxic effects on the retinas of other species. It has been hypothesized that resistance to the toxic effects of high oxygen has evolved in the retina and choroid of trout eyes in response to the high 0, levels to which they are continuously exposed (Desrochers and Hoffert. 1983). Research on the presence of protective mechanisms in trout eyes against oxygen toxicity is currently underway in our laboratory. Although equation 1 is useful as a predictor of the tissue PO, at any point within the retina of the trout, its derivation was completely empirical and it sheds little light on the nature of oxygen delivery and/or utilization of oxygen within this tissue. The retina is a complex network of cellular layers and oxygen diffusion will be influenced by many factors, including the lipid composition, solubility of oxygen and rate of 0, consumption within a particular layer. At this time it is not possible to predict the relative contribution of each of these factors in the in vivo situation. However, the
OXYGEN
DIFFUSION
IN THE
TROIJT
RETINA
613
spatial resolution of the PO, electrodes used in this study was sufficient to enable us to subdivide the retina into anterior and posterior regions. To gain some understanding of oxygen delivery and/or utilization within each region. the data were subjected to an analysis based upon the classical equation for planar diffusion (Crank, 1975; Hill, 1928). Alder et al. (1983) have stated that the existence of both an anterior and posterior oxygen supply to the retina, characteristic of most mammalian eyes, complicates the mathematical analysis of the retinal oxygen profile. Alternatively. the trout retina presents itself as an excellent model for studying oxygen diffusion in vivo since this tissue has a single source of oxygen. The mathematical analysis is therefore somewhat simplified and the following solution for this model is presented. Since we measured Paz, appropriate substitutions into the classical equation have been made and the working equation becomes:
d2P dP Sdt = DS d22-
. 110~’
where: DS = the Krogh permeation coefficient? and D and R are the respective diffusion and solubility coefficients for oxygen, P = the partial pressure for oxygen. x = ,um anterior from the choriocapillaris and, $‘o, = oxygen consumption in ml 0, min-’ cmw3. In the steady state condition with which we are dealing, dP/dt = 0 and the general form of the solution (Simon, 1972) is:
pi(x) = K@+Bix+Ci where:
Pi(x) = the partial
pressure for oxygen
(i = 1, 2),
(3)
at any given depth x.
Ki = Volb,(i,/2DS(i, and the parameters Bi and C, are constants dependent on the boundary conditions, with Ci = the PO, at the choriocapillaris. Three assumptions must hold true in order that the solution of this equation be valid: (1) the tissue should be considered a homogeneous medium, (2) poZ should be constant and independent of PO, for all oxygen tensions above zero and (3) the diffusion and solubility coefficients should be constant and independent of PO, (Ganfield, Whalen and Nair, 1970). In the situation that we are investigating, oxygen is considered to be diffusing in only one dimension through a tissue of finite thickness. Therefore at no depth within the tissue will the PO, drop to zero. The condition that PO, does not decrease to zero does not preclude the use of equation (3) as long as a second boundary with a known PO, is established at the anterior margin of the retina. The teleost retina cannot be considered to be homogeneous throughout, a fact supported by the values for oxygen consumption reported below. Our data also indicate that a logarithmic transformation of the PO, profile through the retina is not’ linear. As stated earlier, we have divided the retina into two regions. The posterior region begins at the choriocapillaris and extends to a depth of 127 ym into the retina. while the anterior portion begins at a depth of 128 ym and terminates at the retinal-vitreal interface. The thickness of the posterior region of the retina was determined histologically from the percent thickness of the retinal cell layers using a mean retinal thickness of 324 f 5% pm (IL = 3). The solution for planar diffusion of oxygen in the two regions of the trout retina therefore necessitates the establishment of two sets of boundary conditions and the solution of simultaneous differential equations (see appendix). “I EEH41
(i I 4 The anterior
I'. I':. I)ESSt~O('HEIIS
boundary
of the retina
ET AI.
(retinal-vitreous interface), was determined to profile approaches an asympt.otic value. The d&e&nation of this boundary was based on the vrry low l&, of the vitreous body and the extremely long path for the diffusion of oxygen through the vit,reous. Equation (3) was applied to the posterior and anterior portions of the profiles from value at the retinal- vi&al 16 eyes in which the /:,2 dropped to an asymptotic, interface. The consta.nt,s Ki and K, of equat,ion (3) were determined using a least 1969) for the two regions of squares fit to the second order polynomial (Bevington. the retina. Oxygen consumption for the two regions of t’he retina was estimated from values reported in the literature. For the teleost retina, Santamaria, Drujan. Svaetichin and ’ Negishi (1971) determined @$‘to be4.3 ,~l hY’ mg e-1dry tissue while the ‘receptor-free retina was 2.63 ~1 hr-’ rng-’ dry tissue (the latter is a correction of an apparent’ incorrec:t value in the original article). In order to estimate that fraction oftotal oxygen consumption that can be attributed to the anterior region of the retina we must first t,ake into account the volume occupied by that area. From measurements of retinal depth we determined that the anterior region of the retina, which corresponds to t,he ‘receptor-free’ retina occupies 60 “b of the total volume. Using this SOY,, value and assuming that 1 mm” of tissue is equal to 1 mg dry weight we ohtained a value for the oxygen consumpt’ion of the anterior region of the retina. Per unit volu~ne of retinal tissue. the anterior region ofthe retina consumes oxygen at a rate of l-58 ,uI hr.-‘. which represents 36.7 ok1of Ma1 retinal consumption, i.e. 36.7 O,clof 4.3 ,~l hr. ‘. We measured the rate of oxygen consumption for bhe entire retina and found it to be 6.02 ,~l Inin-’ cme3. Knowing this value plus the relative size and the fraction of oxygen consumption attributed t.o the anterior and posterior regions of the trout retina. we ran now calculate the rates of oxygen consumption for these two regions. Since the post,erior retina accounts for 63.3 ‘:, of t)otal retinal oxygen consumption and it, occupies 40 O,,,of the ret’inal volume, the calculated rate of oxygen consumption for this region is 9.53 ,LLIm&l cm -3. Similarly the rat’e of oxygen csonsumption for the anterior retina, which contributes 36.7 ‘to vf t.otal retinal oxygen consumption and occupies 60 O0 o f retinal volume. is 3.67 ,uI min-’ crnP3, These rates were used to calculate the Krogh permeation coefficient for the respect,ive layers of the trout retina. The average value (16 eyes) of the Krogh permeation coefficient at 9°C’ for the posterior retina (US,,, Table II) was calculated to be l.H6+_cP33 (X IO-‘), which is significantly higher than t,he value of 0.58 + 0.06 ( x 10d5) ml 0, min ’ cm-’ atrn-’ for the anterior retina (DA',,. Table II). It should be emphasized that these estimates of DS are based on data collected at 9”(’ and that the temperature dependence of DA' should be taken into account, when comparing these values with those obtained 1’) others. In his discussion of oxygen transport in tissues. Thews (1968) reported that DS values at 37°C for skeletal and cardiac, muscle. cerebral cortex and plasma from a variety of species, ranged from 0.7 x 10-5--3.6 x 10e5 ml 0, min-’ cm‘.* at,m-‘. A DS value of 129 x 10e5 ml 0, mine1 cm-’ atm-’ has heen reported for cat, cerebral cortex at 37°C (Ganfield et al., 1970). With a Q1, of 1.1, which was calculated using the 1115’ values for frog muscle at 20 and 37V (Altman and Dittmer, 1971). the BS’ values for the retina of the trout when corrected to 37°C are: Ds,, = 243 x 1W5 and DS,,= 0.76 x 10V5 ml 0, min-’ cm-’ atn-‘. These values are within the reported range of IX! for all tissues and are in good agreement with the values reported for the neural tissue of other species. For example. values for the cerebral vortices of rabbit and rat were 1.9 x 1O-5 and 2.1 x 10-j ml 0, min-’ cn-’ atm-’ rrsprctively. A IM be in the region where t!he slope of the oxygen
OXYGEN
DIFFUSIOK
I?r’ THE
TROYT
RETIS.
615
TABLE II
Second orderpolynomialfit to individual in vivo retinal oxygenprofilesfor the determination of the Krogh permeation coeflicientfor the posterior and anterior retina of the rainbow trou,t, DS,, and DSAR respectively Posterior Eyr 1 2 3 4 -5 6 7 8 9 10 11 12 13 l-4 15 16
retina
(1,
6,
441 238 497 261 315 531 341 250 324 359 398 543 206 207 379 188
2.31 4.80 1.49 3.39 6.20 1.50 2.16 1.10 301 415 933 2.20 062 1.14 2.51 2.76
Anterior retina __--__ DS,, x 105 (h K, ns,, x 103 -.-. ~~ -~- __~~~ .-... -~~ .~~1.57 309 24.5 0.57 0.76 39 7.33 0.19 243 328 2.27 0.61 I .07 127 2.29 W6l 0% 12.5 2.56 0.55 ‘41 417 1.92 073 1.68 180 2.98 047 329 69 6% 02”k 1.20 244 1.79 ti78 087 159 3.55 0.37 0.39 251 147 (h95 1.79 370 3.55 0.39 0.58 105 2.60 0.54 032 14X 1.no o-73 014 174 1.15 0.1% 0.13 45 3.13 0.45
P&r) = KiX2+Bix+Ci (i = 1, 2), where e(z) = the partial pressure for oxygen at any given depth I, s = hm anterior from choriocapillaris, ‘fi = $‘ob,Co(2DS)& where Kt above is reported in mmHg pmeY. Vo, = oxygen consumption in ml min-’ cmm3, DS = Krogh permeation coefficient (ml 0, min-’ cm-’ atm-‘), R, and C, = constants with C, equal to the r’oe at the choriocapillaris
value of 3-24 x lop5 ml 0, min-’ cm-’ atm -l has been reported for the retina of the honeybee drone, Apis mellifera, measured at 22°C (Tsacopoulos. Poitry and Borsellino. 1981). The calculated values for the trout retina at this temperature are: DS rR = 2.11 x 10e5 and DS,, = 0.66 x 10m5 ml 0, min-’ cm-’ atm-‘. Differences between DS values for these two species would be expected based on the high degree of structural dissimilarity between the retina of the trout and that of the compound eye of the drone. The data as derived above clearly indicate that the diffusion of oxygen through the teleost retina is not linear. Previous investigators (Tsacopoulos, Baker and Levy, 1976: Zuckerman and Weiter, 1980; Alder et al., 1983) have also provided evidence that oxygen transport rates vary within the retinas of the pig, bullfrog and cat, respectively. Our calculat’ions would indicate that the higher rate of diffusion through the posterior layer is likely, due to the higher rate of oxygen consumption by photoreceptor cells contained in this region, Since no reliable estimate of 0, solubility (S) is available for the retinal cell layers, determinations of the diffusion coefficients (0) have not been made. The lipid content of the different retinal layers would be expected to be important in determining oxygen solubility in these regions. In summary, a technique was developed for measurement of intraretinal PO, of the rainbow trout in vivo. Determination of the intraretinal PO, profile confirmed that all layers of the trout retina are continuously exposed to oxygen tensions in excess of “I.2
1’. E. DESKO(:H
616
ICKS ET AI,
values shown to be toxic to the retinas of other species. The resistancle to osygen toxicity exhibited by this tissue thereby provides us with an excellent model for thr study of oxygen toxicity and the agents which protect. against it. The mammalian retinal PO, (Tsacopoulos et al., 1976; Alder et al., 1983) differs from that reported fol the trout in that the mammalian profile tends to be parabolic in nature. This reflects the fact that the mammalian retinas receive oxygen from both retinal and choroidal circulations. This dual supply ofoxygen precludes calculation ofpermeat’ion coefficients for oxygen diffusion through these retinas, while trout retinas, which are avascular and have a single source of oxygen, are ideal for studies of this nature. For t,he trout retina, analysis of 16 profiles yielded two statistically different values of the Krogh permeation coefficient for oxygen. We obtained values of 1.86 x 10e5 and 0.58 x region and remaining 10M5 ml 0, mine1 cm-l atm-’ at 9°C for th e photoreceptor neural retina respectively. ACKNOWLEDGMENTS This work was supported by NIH grant EY-00009 and Michigan Agricultural Experiment station, Project 1441 H (Journal Article Number 10801). The authors wish to thank Esther Brenke for her expert technical assistance and Dr William F. Jackson. Medical College of Georgia, Augusta, GA for his valuable assistance in the mathematical analyses, APPENDIX The determination of the Krogh permeation coefficient for oxygen under the conditions of unidirectional planar diffusion in steady state is presented below. The model deals with a tissue of finite depth (i.e. oxygen tension does not fall to zero) and comprises two regions with differing vo,. Within each region it is assumed that the vo, is homogeneous, constant and independent of PO,. The posterior region of the retina extends from 5 = 0 at the choriocapillaris to .z, = 127 pm which includes the pigment epithelium and visual cells. The anterior region extends from z1 to z2 located at the retinal-vitreal interface. In region z = 0 to x = zi : d2P,/dx2 = voio,,/nS, = 2K,. (1) In region 5 = xl to 5 = x2: d2P2/dx2 = vo;z2/DS2 = 2K2.
(2)
Boundary conditions: PI(O) = Pee = PO, at choriocapillaris,
and
(3)
Pz(x2) = PAR = PO, at boundary margin of anterior retina,
(4)
Wdx, - 1, = dP,/dz, = rl PI(q) = PZ(rl) (i.e. flux between the two layers must be equal).
69
(6)
Solutions :
Therefore applying boundary Boundary condition 1:
P,(x) = K,x2+Blx+C1,
(7)
Pz(x) = K,x2+B2x+CZ.
(8)
conditions to obtain constants Bi and C, :
P,(O) = Pee = K,(0)2+B1(0)+C1 therefore :
PI(s) = h;(.~)~ + B, xs Pee.
= Cl,
(9) (10)
Boundary condition 2: &(x2) = PAR = K,x,~ + B,x, + Cr,,
(11)
OSYGES therefore :
DIFFUSIOK
IN THE TROUT
P,(x) = K2(x*-x122)+B2(x-x2)
RETJXA
+P*R.
(iii
(II)
condition 3 :
Boundary
at .r = x1’ therefore : Bwndary
condition
dP,/dx = dP,/dx.
(13)
B, = 2K, x1 - 2K, x1 + B,.
(14)
p,b-I) = &4x,).
(15)
4:
Tn addition : therefore :
B,
Then substituting
(16)
= K,(x~~x~)-K~(~~+~~/~~)+(P~~-P~c)I~~.
for B, in equation (14) and rearranging: B, = K,(xf/x,-2x,)+K,(22,-XT/X,-x,)+(P,,-P&/x,.
Now substituting
the appropriate
(17)
constants into equations (7) and (8) and simplifying,
PI(x) = ~‘,X~+~K~(~~/~~-~~~)+K~(~T~-X~/Z~-~~)+(P~~ZR-P~~)/Z~JZ~P~~, Rx)
= K2x2 + [K,W4
-4(x,+x:/x2)
+ (Pa
yields: (18)
-Pcc)/zzlr+P~,-K,t~-K,x:.
(IF))
(i = 1. 2)
(W
Thus both equations are in the form e(x)
= &z~+B~x+C’~
where K, = p001,/2DS,, K, = ~obp2/2DS,, and Bi, Cti are constants. Each region is then fit to the polynominal, qx, = Kxa+Bx+(’
and the coefficient K used to calculate L)S for each region.
REFERENCES Alder, V. A., Cringle, S. J. and Constable. I. J. (1983). The retinal oxygen profile in cats. Invest. Ophthalmol.
Vis. Sci. 24, 30-6.
Altman,
P. 1~. and Dittmer, D. S., Eds. (1971). Biologica Handbooks. Respiration and Circulation. Pp. 21-3. FASEB Press: Bethesda, MD. ARVO resolution on the Use of Animals in Research (1983). Invest. Ophthalmol. l’iu. s’ci. 24. 1156. Bevington, P. R. (1969). In Data Reduction and Error Analysis for the Physical Sciences. Pp. 13444. McGraw-Hill: New York. Crank, J. (1975). In The Mathematics of Diffusion (2nd edn). Pp. 4468. Oxford University Press: Oxford. Desrochers, P. E. and Hoffert, J. R. (1983). Superoxide dismutase provides protection against the hyperoxia in the retina of the rainbow trout. (Sulmo gairdneri). Camp. Biochem,. Physiol.
76B, 241-7.
Drujan, B. D., Svaetichin, G. and Negishi, Ii. (1971). Retinal aerobic metabolism as reflec.ted in S-potential behavior. Vision Res. 3, 151-9. Fairbanks, M. B., Hoffert, J. R. and Fromm, P. 0. (1969). The dependence of the oxygenconcentrating mechanism of the teleost eye (Salmo gairdneri) on the enzyme carbonic anhydrase. J. Gen. Physiol. 54, 203-l 1. Fairbanks, M. B., Hoffert, J. R. and Fromm, P. 0. (1974). Short circuiting the ocular oxygen-concentrating mechanism in the teleost Salmo gairdneri using carbonic anhydrasc inhibitors, J. Gen. Physiol. 64, 263-73. Ganfield, R. A., Whalen, W. J. and Nair, P. (1970). Mass transfer, storage and utilization of 0, in cat cerebral cortex. Am. J. Physiol. 219, 814-21. Hill. A. V. (1928). The diffusion of oxygen and lactic acid through tissues. Proc. R. Sot. London Ser. B. 104, 3996. Hoffert, J. R. and Ubels, J. L. (1979). The intraocular PO%and electroretinogram of the trout as affected by temperature and ventilatory flow. f’omp. Biochem. Physiol. 62A. 563-68.
P. E. DESRO(‘HERS
618
ET AL
Kuhn, W., Ramel, A., Kuhn, H. J. and Marti. E. (1963). The filling mechanism of the swimbladder. Generation of high gas pressures through hairpin countercurrent multiplication. Experientia 19. 497-552. and Negishi, Ii.. Svaetichin. G., Laufer, M. and Drujan, B. 1). (1975). Polarogmphic electrophysiological studies of retinal respiration. Vision Res. 15, 527-33. Oyama. V. I. and Eagle, H. (1956). Measurement of cell growth in tissue culture with a phenol reagent (Folin-Ciocalteau). Proc. Sot. ICxp. Biol. Med. 91, 305-7. Santamaria, L.. Drujan. B. D., Svaetichin, G. and Negishi, K. (1971). Respiration, glycolysis and S-potentials in teleost retina: a comparative study. Vision Rea. 11, 877-87. Simon. W. (1972). Mathematical Techniquesfor Physiology and Medicine. Pp. 41-60. Academic Press: New York. Thews, G. (1968). The theory of oxygen transport and its application to gaseous exchange in the lung. In Oxygen Transport in Blood and Tissue (Eds Lubbers. D. W., Luft. IT. f’.. Thews, G. and Witzleb. E.). Pp. l-20. Georg Thieme Verlag: Stuttgart. Tsacopoulos, M., Baker, R. and Levy, S. (1976). Studies on retinal oxygenation. Adv. Exp. Med. Biol. 75, 413-6. Tsacopoulos. M., Poitry, S. and Borsellino, A. (1981). Diffusion and consumption of oxygen in the superfused retina of the drone (Apis mellifera) in darkness. J. Gen. Physiol. 77. 601-28. Umbreit, W. W., Burris, R. H. and Stauffer, J. F.(Eds) (1964). Manometric Techniques (4th edn). Pp. 5. Burgess: Minneapolis. Whalen, W. J., Riley, J. and Nair, P. (1967). A microelectrode for measuring intracellular Paz. J. Appl.
Physiol.
23, 798401.
Wittenberg, J. B. and Wittenberg, B. A. (1961). Active transport of oxygen and the eye of fish. Biol. Bull. 121, 379. Zuckerman, R., and Weiter, J. J. (1980). Oxygen transport in the bullfrog retina. Exp. Eye Res. 30, 117-27.