Neurotransmitter systems in the outer plexiform layer of mammalian retina

Neuroscience Research, Suppl. 8 (1988) $127-S136 Elsevier Scientific Publishers Ireland Ltd.

$127

NEUROTRANSMITTER SYSTEMS IN THE OUTERPLEXIFORMLAYEROF MAMMALIANRETINA DIANNA A. REDBURN Department of Neurobiology and Anatomy, University of Texas Medical School at Houston, P.O. Box 20708, Houston, TX

77225

GLUTAMATE There is a growing consensus from many different disciplines that glutamate is the neurotransmitter used by both rods and cones to signal bipolar cells 1-4. I t is released from photoreceptors when they become depolarized by darkness. Since neurons in the outer retina do not produce action potentials, the amount of glutamate released would be a direct function or analogue of the degree of photoreceptor depolarization at any given point in time. In other words, there would be tonic release under most conditions, and changing levels of darkness and l i g h t would lead to increases and decreases respectively in that tonic release rate.

This type of analogue output system provides a high degree of

f i d e l i t y in the transfer of information which would be an obvious requirement for the relay of highly discrete visual events. Much of the more recent information about glutamate transmission in retina has come from analyses of post-synaptic elements of the glutamate system. Four different subclasses of glutamate receptors have been described in retina 5. They are classified and named according to the potency and specificity with which they interact with a series of glutamate agonists.

NMDAand quisqualate

sensitive glutamate receptors may be present in the OPL but to date they appear to have only a minor role in light-activated pathways. The most prominent type of glutamate receptor in the OPL is of the kainate variety. Glutamate and its agonist, kainate, cause depolarization of cells which contain these receptors, namely OFF cone bipolar cells, horizontal cells, and perhaps rod bipolar cells which are

all

sign-conserving.

The glutamate receptor

present

on the

sign-inverting pathway from cones to ON bipolars appears to be a unique in its preferential sensitivity to the agonist AP4 (2-amino-phosphonobutyric acid, formally abbreviated APB)6. In addition to i t s unique pharmacological sensitivity, i t is also unique in its actions on the cell. Activation of AP4 receptors on ON bipolar cells leads to hyperpolarization of the cell and is thus responsible for the sign-inverting characteristics of this pathway. AP4 is a very selective ligand and has only subtle effects on elements of the OFF pathway. For this reason AP4 represents a valuable pharmacological tool for functionally separating elements of the ON pathway from those of the OFF pathway. I t has been our intent to determine i f retinal glutamate receptors could be Presented at the 10th Taniguchi International Symposium on Visual Science, November 23-27, 1987 0168-0102/88/$03.50 © Elsevier Scientific Publishers Ireland Ltd.

$128

further characterized

through biochemical analysis such as in vitro binding

assays which have been developed for subclasses of glutamate receptors in other parts of the CNS. These assays can be used to estimate total

numbers of

receptors, screen potential ligands, examine effects of ions on the recognition site of the receptor, and confirm the existence of distinct receptor classes. We have examined the

characteristics

of

the

AP4-sensitive

3H-glutamate

binding site in retinal membranes under conditions established as optimal for brain preparations 7-9.

We used rabbit retina for these studies in order to

allow a direct comparison of our binding data with the electrophysiological data from other AP4-sensitive displays

labs

using the

same species.

3H-glutamate binding

characteristics

in

which are

Our results

suggest that

synaptosomal membranes from retina

virtually

reported in other areas of the CNS7,10,11.

indistinguishable

from those

The number of sites in freshly

prepared membranes was decreased by 90% when tissues were freeze-thawed prior to the assay. Calcium and chloride were necessary in the assay in order to obtain maximal binding.

The ability of various glutamate analogues to displace

3H-glutamate binding in retinal membranes under optimal conditions was similar to that reported

in brain7,10, 11.

We conclude that

an identical

or very

similar AP4-sensitive, glutamate binding site is widely distributed throughout the CNS including retina and that the maximal number of binding sites are seen in freshly prepared membranes in the presence of chloride and calcium. Some investigators have suggested that the susceptibility to freeze-thawing and the requirement for chloride are more characteristic of glutamate transport sites than glutamate receptors12, 13.

Furthermore, the number of

sites

is

relatively large compared to other types of glutamate receptors and is perhaps more consistant with the number expected for glutamate transport sites.

In

order to assess the validity of these arguments in retina, we examined the chloride-dependent

uptake of 3H-glutamate in retinal

synaptosomes and its

sensitivity to AP4 and other glutamate analogues14. By u t i l i z i n g buffers that were deficient in either sodium or chloride, we were able to uptake.

distinguish

between sodium-dependent and chloride-dependent

B o t h sodium- and chloride-dependent uptake into

rabbit

retinal

synaptosomes were temperature sensitive; however, the amount of chloride-dependent uptake was only ten to twenty percent of that seen in the presence of sodium. The pharmacological sensitivity differences.

of

the

two uptake sites

showed marked

D-aspartate was a potent inhibitor

of sodium-dependent uptake

whereas chloride-dependent uptake was unaffected.

In contrast, AP4 was more

potent

in

inhibiting

chloride-dependent uptake, although sodium-dependent

uptake was also significantly inhibited.

We conclude that AP4 does inhibit

S129

chloride-dependent

and

to

a

lesser

degree, sodium-dependent uptake of

glutamate. We attempted to further characterize the AP4-sensitive and chloride-dependent glutamate transport site by analyzing the release that occurs from the pool which i t labels. Release of 3H-glutamate from tissue preloaded and perfused in chloride buffer was significantly stimulated by depolarizing levels of potassium. The rate of evoked release was approximately double that of resting release. In contrast, release after pretreatment with AP4 was increased by potassium depolarization only s l i g h t l y (<20%). In addition, the tonic release (leakage) rate calculated on a percentage basis was two and one-half times higher after AP4 treatment. These d a t a demonstrate that AP4 inhibits chloride-dependent uptake into a synaptosomal pool which can be released by potassium stimulation and therefore, is at least theoretically, part of the transmitter pool of glutamate in retina. In summary, b o t h electrophysiologica115

and biochemical techniques7,14

provide substantial evidence for AP4-sensitive recognition sites for glutamate in rabbit retina. I t is now important to determine how many AP4-sensitive sites are there and what is their function. can

be

addressed by

analyzing

the

The f i r s t part of this question

pharmacological specificity

of

the

AP4-sensitive recognition site under all three assay conditions. Table I compares the potency of a number of glutamate analogues to i n h i b i t receptor binding and uptake. Not only is the rank order of potency the same for both assays, but the actual values of the inhibitory constants (KI) are also quite similar. There is also good agreement between the data and those reported from electrophysiological studies on ON bipolar cells.

Of particular interest

is the preference for the L-isomer of both AP4 and o-phosphoserine in the binding assay, the uptake assay and electrophysiological studies by Slaughter and Miller 15.

It

is clear from this comparison that the pharmacological

characteristics of these three AP4-sensitive sites are very similar.

We

suggest that the recognition sites may be identical and therefore in vitro binding assays would detect recognition sites for both post-synaptic receptors and uptake sites. The functional importance of AP4-sensitive, chloride-dependent glutamate uptake is s t i l l unresolved. Recently, reports by Ayoub et al. 16, and by Arkin and Miller 17, suggest that some minor effects of AP4 on OFF circuits can be observed. One of several possible explanations for these effects might be that AP4 inhibits glutamate re-uptake and leads to an overall increase in extracellular glutamate concentrations. This in turn would stimulate OFF as well as ON pathways. The small magnitude of the overall effect is consistent with the small capacity of the chloride dependent uptake system, compared to

S130

Table I.

Druq

Pharmacological Specificity of Chloride-Dependent 3H-Glutamate Binding and Uptake Specific 3H-glutamate binding in fresh tissues in the presence of calcium and chloride

L-2-AP4 D,L-2-AP4 D-2-AP4 D,L-(AP5) D,L-(AP6) D,L-(APT) NMDA O-Phospho-L-Serine O-Phospho-D-Serine D-Aspartate

Chloride-dependent 3H-glutamate uptake in retinal synaptosomes

6 5 58 >1000 >1000 51 >1000 15 60 ......

12 30 57 250 >1000 48 >1000 20 75 >1000

The Ki values (MM) were calculated as follows: Ki = (ICso [concentration of drug required to displace 5 ~ of the ligand])/([concentration of ligand used/Kd of ligand] + I ) .

that of the sodium-dependent uptake which is less sensitive to AP4. In summary, the unique attributes of glutamatergic transmission in the OPL result in part from the existence of multiple classes of glutamate receptors. The sign conserving second order neurons contain primarily kainic acid-sensitive glutamate receptors which bring about depolarizataion when activated. The sign inverting, ON bipolar cells possess a unique type of glutamate

receptor

which

preferentially

binds

AP4

and

leads

to

hyperpolarization when activated. Our demonstration of other potential binding sites for AP4 does not negate the importance of this drug in providing a tool for i n h i b i t i n g elements of the ON channel while leaving elements of the OFF channel r e l a t i v e l y intact. d is tr i b u ti o n of this unique

However, i t does raise questions about the glutamate binding site and i t s possible

relationship to glutamate transport systems which may play important role in glutamatergic transmission in the OPL.

an

equally

MELATONIN In addition to the discrete point-to-point transfer of information along the primary sensory pathway that leads to visual perception, there is evidence for collateral pathways which do not encode visual images, but rather monitor changes in total levels of illumination irrespective of the shape, size or movement of objects in the visual f i e l d . In higher centers, there are probably many such secondary pathways; for example, the suprachiasmatic nucleus projection to the pineal gland which provides information to

$131

light-entrain certain circadian rhythmic events. There are also light-entrained circadian rhythmic events in the retina such as shedding of outer segments of photoreceptors, cone elongation and contraction, and pigment migration in the pigment epithelium18. There is increasing evidence that the signal which controls these events may be generated in the retina i t s e l f and not via centripetal fibers from relay centers elsewhere in brain sent to the retina. The chemical messenger encoding for total illumination may be an indole hormone such as melatonin and i t is probably generated in the photoreceptor i t s e l f . Our f i r s t evidence of an indole system in the vertebrate retina came from autoradiographic studies which showed the accumulation of 3H-serotonin in photoreceptor terminals of the Long Evans rat retina 19. In addition, this accumulation appears to be l i g h t sensitive. When incubated in the l i g h t , clear labeling is observed; however, after a brief exposure to darkness there is a dramatic loss of v i r t u a l l y all labeling of the terminals. We have established that endogenous levels of retinal serotonin show similar light/dark fluctuations20. The highest level of endogenous serotonin (78 ng/g protein) was detected in rats housed in l i g h t for 24 hr; i t was non-detectable in rats housed in the dark for 24 hr; and cyclic (12 hr/12 hr) light-dark conditions resulted in intermediate levels (37 ng/g protein). Subsequent experiments have addressed the mechanism(s) which might be responsible for the dark-induced decrease in retinal serotonin21. Since the photoreceptor cell depolarizes and releases its neurotransmitter in response to darkness, the lack of observed accumulation could result from a dark-stimulated release of serotonin. In support of this notion, our data demonstrate that in the l i g h t , depolarizing levels of potassium stimulate the release of 3H-serotonin. However, mounting evidence for a melatonin system in retina may provide an alternative hypothesis. Several reports over the past two decades have documented the presence of melatonin in retinas from various species22-24. As in the pineal gland, melatonin synthesis in retina is known to increase in the dark and decrease in the light25, 26. The increase in both retinal and pineal melatonin synthesis results from dark activation of N-acetyl transferase (NAT), the rate-limiting enzyme for melatonin synthesis27"29. Melatonin synthesis in the pineal gland has been localized to pinealocytes 24.

Since pinealocytes in the pineal gland

and photoreceptors in the retina are analogous in many ways, photoreceptors are implicated as the source of retinal melatonin. Hydroxy indole-O-methyl transferase (HIOMT), one of the non-rate-limiting enzymes for melatonin synthesis, has been localized to photoreceptor terminals30, 31. In a preliminary report, we showed that 3H-serotonin accumulated in rat retina in

S132

the light

is rapidly converted to melatonin when the retina is subsequently

exposed to darkness20.

Therefore, the decrease in 3H-serotonin accumulation

seen in darkness may be due, at least in part, to dark-stimulated the

NAT enzyme, converting

serotonin

to

activation of

melatonin which is

subsequently

released. In order to further test this hypothesis, we have examined the synthesis and release

characteristics

of

both serotonin

and melatonin21.

The primary

precursor in the neuronal indole pathway is tryptophan and the retina has an active

uptake system for

this

compound. 3H-tryptophan is

taken up in

a

temperature and sodium dependent manner by homogenized rat retina and kinetic analysis reveals a Km of 50 I~M and a Vmax of 66 fmoles/mg protein/min. Chromatographic further

analysis

confirms

the

intact

different

light/dark

the three

precursor

retinas

chromatographed.

radiolabeled

metabolites

endogenous synthesis

3H-melatonin from this isolated

of

in

rat

were f i r s t

conditions.

tryptophan,

3H-tryptophan

b o t h 3H-serotonin

retina.

For

incubated with

serotonin

and

these experiments

3H-tryptophan under

R e t i n a s were t h e n homogenized

A clear precursor-product

indoles

of

of

and

relationship was apparent among and melatonin.

After

a ten min

incubation period with retinas exposed to room light, approximately half of the 3H-tryptophan was converted to 3H-serotonin. virtually

all

depleted.

of the label

resides with

After

3H-serotonin and 3H-tryptophan is

Incubations up to 60 min in the light

percentage being converted

to

labeled

30 min in the light,

results in only a small

melatonin.

However, exposure to

darkness causes a dramatic rise in melatonin synthesis and a concomitant loss of labeled serotonin within the retinal

tissues.

Thus i t

photoreceptor terminals, there is synthesis of serotonin

appears that

in

from tryptophan in

the light and a dark-stimulated conversion from serotonin to melatonin. In the next series of experiments we analyzed release characteristics of endogenously lighting light,

synthesized 3H-serotonin

conditions.

and

3H-melatonin under different

Retinas were incubated with 3H-tryptophan under room

subsequently perfused in the light and perfusates were collected for

analyses.

After a relatively stable basal efflux was established

in control

buffer, depolarizing levels of potassium added to the perfusion buffer caused a significant increase in the amount of radioactivity released. to control

buffer,

a second pulse of potassium did not e l i c i t

Chromatographic analysis

of the releasates

show that

After return a release.

the majority of the

radioactivity released was associated with endogenously synthesized 3H-serotonin (Fig. la).

We conclude that in the light,

serotonin is synthesized and

stored in a pool which can be released by potassium depolarization.

$133

I f we change the lighting conditions during incubation

from room light to

total darkness, which stimulates the conversion of serotonin to melatonin, the amount of radioactivity released is similar but there is a significant increase in the relative amount of melatonin released (Fig. release s t i l l

predominates.

underestimate.

Ib).

However, serotonin

We questioned whether or not this might be an

Perhaps melatonin is not stored after its synthesis because i t

is lipophilic and diffuses out of the terminal as soon as i t is synthesized. In order to test this hypothesis, we included soadenosyl-homocysteine in the incubation medium to inhibit HIOMT activity. should be a dark stimulation of n-acetyl

Under these conditions,

serotonin

conversion to melatonin would be blocked.

there

synthesis but its

Since n-acetyl

serotonin

final is not

lipophilic, i t should remain in the tissue. Our results show that this assumption was correct.

Tissue f i r s t

incubated

in the dark with 3H-tryptophan plus s-adenosyl-homocysteine, and then perfused in the

light

showed a decrease in total

number of

counts taken up and

released; however, a large proportion of the radioactivity was in the form of n-acetyl

serotonin

(Fig.

Ic).

These results suggest that melatonin is not

stored but released very quickly after i t is synthesized. In the next experiments, we incubated retinas in 3H-tryptophan in the light so that the majority of the label would be converted to 3H-serotonin. perfused in the dark or in the light. dark compared to light.

We then

Release rates were very high in the

In the light, the radioactivity remained sequestered,

and was released at a much slower rate. We suggest that the high rates of release in the dark are due to diffusion of newly synthesized melatonin. this conclusion.

We have two lines of evidence which support

First, the chromatographic analysis of these perfusates show

that the dark-released

radioactivity

is

almost exclusively in the form of

endogenously synthesized 3H-melatonin (Fig. ld).

Second, the high levels of

release in the dark were blocked by the HIOMT inhibitor. Finally, we incubated retinas in 3H-serotonin rather than 3H-tryptophan in order to look at the pulse labeling of endogenously synthesized melatonin from a more immediate precursor.

Under these conditions, we obtain an even more

pronounced dark stimulation

of release.

Furthermore, there

is

a dramatic

shift in the relative distribution of label in the perfusates collected in the light versus those collected within the f i r s t 3 min after exposure to darkness (Fig. le).

There is a relatively high basal release of labeled serotonin in

the light but melatonin release predominates in the dark. a minimal response but melatonin,

the

label

released is

Potassium produces

primarily

serotonin,

not

further confirming the fact that melatonin release is insensitive

to potassium.

$134

[]

IO0

[]

[]

8O

6O

40 >I,-

20

fJ ,( O a ..I <{ IO I,u. O

TRY

5HT

TRY

MEL

5HT

MEL

TRY

[]

100

5HT

NAB

MEL

[]

8O

6o

40

20

0

._

I

TRY

5HT

NAS

MEL

5HT

NAS LIGHT

MEL

5HT

NAS DARK

MEL

5HT

Lmm NAS KCI

MEL

FIGURE I. Chromatographic analysis of labeled indole metabolites released from rat retinas. Perfusate samples were taken after a ten min perfusion in control buffer followed by a 3 min exposure to the stimulus. (a) Perfusates from retinas incubated in 3H-tryptophan for 10 min in the light, perfused in the light and stimulated with potassium, contained primarily labeled serotonin and negligible amounts of melatonin. (b) Perfusates obtained from retinas under similar conditions but in darkness, contained only slightly increased levels of melatonin. (c) When an inhibitor of melatonin synthesis was included under conditions of darkness, another metabolite, n-acetyl serotonin, was released in highest concentrations. (d) Melatonin accounted for virtually all of the label released when retinas were preincubated in the light, perfused in the light and theE "stimulated" by exposure to darkness. (e) When retinas were preincubated in ~H-serotonin in the light, both light and potassium stimulated serotonin release whereas exposure to darkness stimulated primarily the release of melatonin.

$135

In summary, we conclude that melatonin synthesis in the retina is under exquisite control

of the rate-limiting enzyme NAT, and that its

rate of

synthesis from tryptophan via serotonin in the dark is very rapid. Furthermore, melatonin is not stored, but appears to be released immediately after synthesis probably, by simple diffusion.

Since we have localized the sites of

3H-serotonin uptake to photoreceptor terminals, there can be l i t t l e doubt that sites of 3H-melatonin synthesis and release also reside there. SUMMARY Melatonin

represents

a second type

of

chemical signal

released from

photoreceptors in response to increased darkness, one with characteristics which are significantly different from those of glutamate. Concise spatial and temporal aspects of the photoreceptor signal are conserved through discrete glutamatergic synapses. Different classes of post-synaptic neurons each have appropriate subclasses of glutamate receptors which transmit sign conserving or sign melatonin,

inverting images of

because of

its

the visual mosaic.

highly l i p o p h i l i c

nature

is

not

In contrast, released by

stimulus-coupled secretion mechanisms, but rather by simple diffusion. Thus control of melatonin "release" may be less concise t h a n glutamate. In addition, melatonin may diffuse beyond the confines of the synaptic area to target cells throughout the retina. Effects of melatonin in retina are not well understood; however, current hypotheses suggest that, perhaps via its control of dopamine systems in the inner retina, melatonin plays an important role in dark adaptation and in various retinal processes which exhibit a circadian rhythm. Melatonin and glutamate may represent "co-transmitters" which provide the visual pathway with two types of signals, with melatonin providing widespread modulatory influences on the discrete visual information conveyed via glutamatergic circuits. ACKNOWLEDGEMENTS I am particularly grateful for the help of Ms. Cheryl Mitchell, a Research Associate, who is largely responsible for producing the data on glutamate and melatonin, and Ms. Diana Parker, who prepared the manuscript in camera-ready format. This work was supported by NEI grant EYO 1655-11 to D.A.R. REFERENCES I.

RodieckRW (1973) The Vertebrate Retina, WH Freeman, San Francisco

2. 3.

MurakamiM, Ohtsuka T, Shimazaki H (1975) Vision Res 15:456-458 Miller RF, Slaughter MM (1984) In: Morgan W (ed) Retinal Transmitters and Modulators: Models for the Brain, CRC Press, Cleveland, pp 107-122

$136

4. 5. 6.

Bloomfield SA, Dowling, JE (1985) J Neurophysiol 53:699-713 Coleman PA, Massey SC, Miller RF (1986) Brain Res 381:172-175 Slaughter MM, Miller RF (1981) Science 211:182-185

7.

Mitchell CK, Redburn DA (1982) Neurosci Lett 28:241-246

8. 9.

Butcher S, Roberts PJ, Collins JF (1984) J Neurochem 43:1039-1045 MonaghanDT, McMills MC, Chamberlin AR, Cotman CW (1983) Brain Res 278:137-144

10. 11.

Nadler JV, Wang A, Werling LL (1985) J Neurochem 44:1791-1798 Fagg GE, Foster AC, Mena EE, Cotman CW (1982) J Neurosci 2:958-965

12. 13.

Pin J-P, Bockart J, Recasesn M (1984) FEBS Lett 175:31-36 Zaczek R, A r lis S, Markl A, Murphy T, Drucker H, Coyle JT (1987)

14.

Neuropharmacology 26:281-287 Mitchell CK, Redburn DA (1987) AP4 i nhi bi t s chloride-dependent binding and uptake of 3H-glutamate in rabbit retina.

Brain Res (In Press)

15. 16.

Slaughter MM, Miller RF (1985) J Neurosci 5:224-233 Ayoub GS, Nawy S, Copenhagen DR (1987) Invest Ophthalmol Vis Sci (Suppl)

17.

28:352 Arkin MS, Miller RF Subtle actions of 2-amino-phosphonobutyrate (APB) on the OFF pathway in the mudpuppy retina.

Brain Res (In Press)

18.

Pierce ME, Besharse JC (1986) In: O'Brien PJ, Klein DC (eds) Pineal and Retinal Relationships. Academic Press, Orlando, pp 219-238

19.

Redburn DA, Churchill L (1987) J Neurosci 7:319-329

20.

Redburn DA, Mitchell, CK (1986) Invest Ophthalmol Vis Sci (Suppl) 27:183

21.

Redburn DA, Mitchell CK Synthesis and release of melatonin from rat

22.

retinal photoreceptor terminals. Manuscript in preparation Quay WB (1965) Life Sciences 4:983-991

23.

Cardinali DP, Rosner JM (1971a) J Neurochem 18:1769-1770

24. 25.

Bubenik GA, Brown GM, Uhlir L, Grota LJ (1974) Brain Res 81:233-242 Pang S, Brown GM, Grota LJ, Chambers JW, Rodman RL (1977) Neuroendocrinology 23:1-13 Bubenik GA, P u r t i l l RA, Brown GM, Grota LJ (1978) Exp Eye Res 27:323-333

26. 27. 28.

Binkley S, Hryshchyshyn M, Reilly M (1979) Nature 281:479-481 Binkley S, Reilly KB, Hryshchyshyn M (1980) J Comp Physiol Biochem Syst Environ Physiol 139:103-108

29. 30.

luvone PM, Besharse JC (1983) Brain Res 273:111-119 Cardinali DP, Rosner JM (1971b) Endocrinology 89:301-303

31.

WiechmannAF, Bok D, Horowitz J (1985) Invest Ophthalmol Vis Sci (Suppl) 26:253