Ponto-geniculo-occipital-wave suppression amplifies lateral geniculate nucleus cell-size changes in monocularly deprived kittens

Developmental Brain Research 114 Ž1999. 109–119

Research report

Ponto-geniculo-occipital-wave suppression amplifies lateral geniculate nucleus cell-size changes in monocularly deprived kittens James P. Shaffery a

a,)

, Howard P. Roffwarg a , Samuel G. Speciale b, Gerald A. Marks

b

Department of Psychiatry and Human BehaÕior, DiÕision of Neurobiology and BehaÕior Research, UniÕersity of Mississippi Medical Center, 2500 N. State Street, Jackson, MS 39216-4505, USA b Department of Psychiatry, UniÕersity of Texas Southwestern, Medical Center at Dallas, 5323 Harry Hines BouleÕard, Dallas, TX 75235-9070, USA Accepted 2 February 1999

Abstract We have previously shown that during the post-natal critical period of development of the cat visual system, 1 week of instrumental rapid eye movement ŽREM. sleep deprivation ŽIRSD. during 2 weeks of monocular deprivation ŽMD. results in significant amplification of the effects of solely the 2-week MD on cell-size in the binocular segment of the lateral geniculate nucleus ŽLGN. w36,40x. In this study, we examined whether elimination of ponto-geniculo-occipital ŽPGO.-wave phasic activity in the LGN during REM sleep ŽREMS., rather than suppression of all REMS state-related activity, would similarly yield enhanced plasticity effects on cell-size in LGN. PGO-activity was eliminated in LGN by bilateral pontomesencephalic lesions w8,32x. This method of removing phasic activation at the level of the LGN preserved sleep and wake proportions as well as the tonic activities Žlow voltage, fast frequency ECoG and low amplitude EMG. that characterize REM sleep. The lesions were performed in kittens on post-natal day 42, at the end of the first week of the 2-week period of MD, the same age when IRSD was started in the earlier study. LGN interlaminar cell-size disparity increased in the PGO-wave-suppressed animals as it had in behaviorally REM sleep-deprived animals. Smaller A1rA-interlaminar ratios reflect the increased disparity effect in both the REM sleep- and PGO-suppressed groups compared to animals subjected to MD-alone. With IRSD, the effect was achieved because the occluded eye-related, LGN A1-lamina cells tended to be smaller relative to their size after MD-alone, whereas after PGO-suppressing lesions, the A1-lamina cells retained their size and the non-occluded eye-related, A-lamina cells tended to be larger than after MD-alone. Despite this difference, for which several possible explanations are offered, these A1rA-interlaminar ratio data indicate that in conjunction either with suppression of the whole of the REMS state or selective removal of REM sleep phasic activity at the LGN, altered visual input evokes more LGN cell plasticity during the developmental period than it would otherwise. These data further support involvement of the REM sleep state in reducing susceptibility to plasticity changes and undesirable variability in the course of normative CNS growth and maturation. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Function; Rapid eye movement sleep; Activity-dependent; Development; Visual cortex; Pontomesencephalon

1. Introduction Despite nearly four decades of investigation into the neurophysiological and neurochemical control of rapid eye movement ŽREM. sleep and the active processes involved in its initiation and maintenance, few data have been uncovered that explain the basic biological function of this major organismic state. In the mid-sixties, Roffwarg et al. w43x proposed that REM sleep plays an important role, as does waking stimulation, in the development of the central

)

Corresponding author. E-mail: [email protected]

nervous system ŽCNS.. Noting that REM sleep-related CNS activation compares in intensity with that in waking and that infant humans exhibit the highest amount of REM sleep throughout the life span, these workers suggested that the extensive neural discharge generated in brainstem during REM sleep and directed toward higher centers may, like waking sensory activity, promote maturational processes. They theorized that REM sleep-related endogenous activation constitutes a pre-programmed process affecting CNS development. In agreement with this idea, recent observations have shown that neural connectivity in the immature CNS is directed by an interaction of preexisting neural circuitry not only with specific, exogenous sensory

0165-3806r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 3 8 0 6 Ž 9 9 . 0 0 0 2 7 - 9

110

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119

inputs but also with spontaneously generated, endogenous activity in sensory pathways w28,38,50,57,63x. In many mammals, the visual system undergoes a postnatal ‘critical period’ of development, a time when this system is maximally reactive to altered visual experience of the type imposed by monocular deprivation ŽMD.. The effect of short periods of MD on LGN cell size during the critical period of visual development is well documented and is the empirical basis for the notion of ‘activity-dependent development’ in visual CNS w23,33,51,62x. In our previous investigations, the structural alterations in LGN cell size due to MD were exploited as a baseline against which to assess the effects of eliminating REM sleep in MD animals w36,40,49x. In those studies, kittens were subjected to instrumental REM sleep deprivation ŽIRSD. by the multiple platform-over-water method, spanning the second week of a 2-week MD interval Žpost-natal day ŽPN. 35 to 49. during the visual critical period. In agreement with previous MD studies, we determined that in the activity-dependent, LGN binocular segment ŽBS. of MDonly control animals, the cells in laminae receiving input from the occluded eye were smaller than normal, while cells in the lamina receiving input from the non-occluded eye were larger than normal w36,40x. Our findings with IRSD in the monocularly occluded kitten indicated a greater magnitude of interlaminar alteration in LGN cell size than the changes induced in control animals that were monocularly patched but not deprived of REM sleep. Furthermore, the effects of IRSD in the MD kittens were not limited to the BS of the LGN. Cells in the monocular segment ŽMS. of the LGN, which receive retinal input solely from that portion of the contralateral eye’s visual field Žthe lateral periphery. not seen by the other eye, were also more affected, like the BS cells, by the combination of MD and IRSD than by MD alone. Specifically, the MS cells of the MD q IRSD kittens to which the patched retina projects were significantly smaller than the comparable MS cells in the MD-alone animals w36,49x. MS-neuron soma size is not usually reported to be affected by the relatively short periods of MD that we employed w17,19,25,45,51x. Though the IRSD data offered support for an ontogenetic effect of REM sleep on CNS structure, they did not provide insights into the mechanisms responsible for the additional cell-size changes in the MD q IRSD kittens. A candidate mechanism is the effect of REM sleep phasic activity, indexed by the ponto-geniculo-occipital ŽPGO. wave. PGO-waves are episodic electrophysiological events that are superimposed on tonic LGN or cortex activations that are continuous during REM sleep w20,55x. The waves are observable in several brain structures, including the pons, LGN and visual cortex. Simultaneous, phasic, single-neuron discharges occur in widespread regions of the brain w44x. In LGN, PGO-waves are isolated field potentials appearing in the 30–90 s of slow wave sleep ŽSWS. just preceding onsets of REM sleep and continuing as either single- or grouped bursts of waves in the body of

REM sleep w2,9x. In young kittens, PGO-waves are not present until the beginning of the post-natal critical period of visual system development w5,7x. Laurent et al. demonstrated that without disturbing other electrophysiological signs of REM sleep, rostrally appearing PGO-waves can be eliminated selectively by interruption of the ascending fibers that project from the pontomesencephalic region Žwhere the PGO-waves are generated. w32x. Davenne and Adrien also showed in their studies in very young ŽPN13. and visually unmanipulated kittens that bilateral lesions of the pontomesencephalic isthmus successfully block PGO-waves in the LGN and also significantly slow normal maturational growth as reflected in reduced LGN volume, smaller A-lamina cell size, and lower LGN unit activity w7–9x. In the experiment described here, we paired MD with the Davenne and Adrien PGO-deprivation protocol Žbilateral pontomesencephalic lesions. w7x in place of our previously utilized MD q IRSD paradigm. Our objective was to confirm the finding that in kittens experiencing MD, deprivation of solely the PGO-component of REM sleep, rather than the complete REM state, is sufficient to affect LGN cell growth during visual system development. If suppression of phasic PGO-activation of LGN is the mechanism altering expected LGN cell size, the combination of MD and bilateral brainstem lesions that remove PGO activation from LGN should elicit interlaminar ratio alterations similar to those found when MD is paired with IRSD, namely, compared to non-lesioned, MD-alone kittens, a smaller A1rA-interlaminar ratio. A smaller ratio may represent either or both a reduction in size of visually deprived, BS A1-lamina cells and an increase in size of visually experienced BS A-lamina cells. This result also would implicate elimination of PGO-wave activation in LGN as a basic mechanism underlying the greater interlaminar cell-size disparity found previously in MD q IRSD kittens w36,40x.

2. Materials and methods All procedures were carried out in conformity with the guidelines of the NIH and local Internal Review Board for the Care and Use of Animals. Methods employed in this study, with several exceptions noted below, followed our previously published protocols for related studies and are briefly described here. Pregnant cats were obtained commercially and gave birth to litters in our animal care unit. PN zero is defined as the day of birth. Some 15 kittens of both sexes were taken from seven litters for use in this study. The kittens were housed with their mothers on a 12:12 h, light:dark ŽLD. schedule. At the height of the critical period of visual-system development in cats ŽPN34 or 35. w22,23x, the right eye of all animals was surgically patched to effect MD w54x.

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119

Aseptic technique was utilized under deep anesthesia Žpreanesthetic, atropine, 0.05 mgrkg, s.c.; anesthetic agent, pentobarbital, 40-mgrkg, i.p., supplemented as required.. Upper and lower eyelids were sectioned, a contact lensshaped piece of soft, opaque, neoprene plastic was inserted between skin and underlying tarsal layer, and the lids were sutured together. This technique reduces light transmission more effectively than simple eyelid suture w34x. Recording electrodes were implanted at this time and consisted of two small stainless-steel screws tapped into the skull to monitor the electrocorticogram ŽECoG., two steel spring electrodes sutured bilaterally to trapezius muscles to register the nuchal electromyogram ŽEMG., and a twisted tripolar macroelectrode Ž30-ga. nichrome. stereotaxically placed into the animal’s left LGN ŽHorsley–Clarke coordinates: A 2.5–3.5; L 7.5–9.5; H 13.0–15.0. contralateral to the patched eye to record PGO-waves. During this surgery, a pair of multipolar electrodes was implanted bilaterally and utilized later to produce electrolytic lesions in brainstem Žsee below.. After regaining consciousness from surgical anesthesia, kittens were returned to home cages for 6 to 7 days to recover in routine contact with mothers and littermates. A broad-spectrum antibiotic Žamoxicillin, 50 mgrml, oral suspension. was administered prophylactically before surgery Ž0.5 ml porday. and also during the remainder of the study Ž0.2 ml porday.. On PN40, all animals were housed individually in recording chambers and allowed to adapt to the chambers for 24 h. Food and water were available ad libitum during adaptation. Each recording chamber resided in a soundproofed, temperature Ž20– 228C.- and light-controlled cabinet Ž1 = 2 = 3 m.. An electrically shielded cable connected each animal through a commutator ŽAirFlite. to electrophysiological recording equipment in an adjacent room. On PN41, the baseline ŽBL. recording quality of PGO-wave tracings was assessed. Animals showing high-quality PGO-waves in their polygraphic recordings Ž n s 8. were assigned to the lesion group ŽLESIONED.. Animals in whom electrode placement yielded either no recordable- or only partially discernable PGO-waves Ž n s 7. were assigned to the control group ŽSHAM.. At the start of PN42, when IRSD had commenced in our earlier studies w36,40,49x, the animals assigned during baseline monitoring to the LESIONED group were anesthetized Ž60 mgrkg ketaminer30 mgrkg xylazine, i.p.. and DC electrolytic lesions Žeither 5-mA, 30-s, 30-ga. nichrome, P2.0; L2.0,3.0,4.0; H6.2 or 5-mA, 20-s, 30-ga. nichrome, P2.0; L2.5,3.5; H6.2. were made according to the procedures of Davenne and Adrien w7x. The SHAM group also was anesthetized but no electrical current was applied through the pontine electrodes. After a short recovery period Žusually 4 to 6 h., kittens were again placed in their individual chambers and recorded for 7 days until sacrifice. During this period, they were continued on a 12L:12D photoperiod and recorded for 23 h each day. The

111

recording chambers were monitored by closed circuit television cameras, and visual observations were made both in light and dark with the aid of an infrared light source to which the cameras were sensitive. Food and water were always available in the chambers where the animals were comfortably tethered to the recording equipment through a head-mounted electrode connector. Two times during the day Žat 0600–0700 and 1800–1900 h., each chamber was visited to insure that adequate food, water and dry bedding material were available. At the end of each day’s recording Žbetween 1200–1300 h., animals were removed from the chambers for an hour and body weights and temperatures were obtained. The scalp wounds were cleaned and antibiotic ointment applied, if required. During this period, the animals were kept awake by engaging them in play activities, which they eagerly pursued. Electrophysiological indicators of sleep and wake ŽECoG, EMG, PGO-waves. were amplified ŽGrass P511 amplifiers., digitized at 125-Hz ŽDAP1200r4 AD board, Microstar Laboratories., and stored on the hard disk of a personal computer. Digitized polygraphic records of individual kittens were scored automatically for state ŽWAKE, SWS, REM. in accordance with modified sleep-scoring criteria in cats w61x. The computer program stage-scored kitten sleep in 15-s epochs. Based upon the relative amplitude of the ECoG and EMG traces in each animal, the software algorithm scored each second of digitized electrophysiological data as REM, SWS or WAKE. Then, utilizing a majority criterion, the program assigned every 15-s epoch to only one stage. For example, REM sleep was assigned to 15-s epochs that contained eight or more 1-s intervals in which the digitized voltages from both ECoG and EMG traces were at or below a calibrated minimum; WAKE was assigned to epochs in which eight or more periods contained digitized ECoG values as low as in REM sleep but in which EMG voltages were higher than in REM sleep; SWS was assigned to 15-s epochs when both the digitized EMG and ECoG values were above the REM sleep criterion for 8 s or more. Epochs in which no state claimed a majority of periods were assigned to the MIXED category. This computer-aided, sleep-staging algorithm has been verified to have greater than 90% agreement with visual scoring w48x. Following the post-lesion period of state recording ŽPN42 to PN49, during the last week of MD., all kittens were deeply anesthetized with three times the anesthetic dose of sodium pentobarbital Ž120 mgrkg, IP. and sacrificed by means of intracardiac perfusion with 0.9% saline and 0.2% procaine followed by phosphate buffer Ž0.26 M NaH 2 PO4 , 0.077 M Na 2 HPO4 , pH 7.2. containing 2% paraformaldehyde and 2% glutaraldehyde. The brains were removed, blocked, cryoprotected Ž40% sucrose in fixative solution for 48 to 72 h., and then returned to buffered fixative to postfix until they were sectioned coronally through the LGN on a freezing microtome. One of every five 35-mm sections was mounted on a microscope slide

112

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119

and stained for Nissl substance with Cresyl violet, then dehydrated in ethanol and cleared in xylene. Coverslips were placed over these sections before projected planar areas of individual-cell outlines were measured. The investigator who performed cell-sizing ŽJPS. was blind to the experimental group from which the sections were drawn. Cell size in the BS of the LGN were obtained following the procedures of Kalil w25x, Sanderson w45x, Hubel and Wiesel w23x and Wiesel and Hubel w62x. A total of 200 cells from each animal were measured in double 100-cell sets, one from lamina-A Žreceiving input exclusively from the non-occluded eye. and one from lamina-A1

Žreceiving input exclusively from the occluded eye. within a single, coronal section that was taken from the rostrocaudal midportion of the right LGN ipsilateral to the occluded right eye at approximately the 58-isoazimuth projection line. Cells in the left LGN were not measured because of the presence of the PGO-wave recording electrode. Camera–lucida drawings of all complete, perikaryal outlines that contained a visualized nucleus and well-defined nucleolus were traced in the plane of the nucleolus onto a digitizing tablet connected to a personal computer at a final magnification of X985. The cell-sizing procedure started with visualization of a 7000-mm2 field at the

Fig. 1. Electrophysiological tracings from one kitten pre ŽA.- and post ŽB.-lesion. ŽA. A polygraphic recording from post-natal day 41 ŽPN41., the baseline recording of kitten KL07. The large arrow indicates the transition to REM sleep. The small arrows point to examples of individual PGO-waves registered by the lateral geniculate nucleus ŽLGN. electrode just at the end of a slow wave sleep ŽSWS. bout. Note the reduction in amplitude of the PGO-waves after the transition to REM sleep. ŽB. A polygraphic recording from post-lesion day 7 ŽPN49. in the same kitten at a similar SWS-to-REM-sleep transition Žlarge arrow. showing complete elimination of PGO-waves. Amplification factors and time base were unchanged between the recordings. A 5-s scale bar is shown. ECoG: electrocorticogram, EMG: electromyogram.

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119

113

Fig. 2. Photomicrograph of a frontal section from the pons of an animal that ceased exhibiting PGO-waves following bilateral lesions. This photomicrograph is of a section at approximately P2.0 from an animal with smaller lesions Ždue to 20- vs. 30-s currents.. A 1-mm scale bar is drawn.

medial and uppermost dorsal aspect of the A-lamina. Cellsizing proceeded ventrally to the bottom of the A1-lamina before shifting medially to the adjacent 7000-mm2 field, at which point, the direction of movement turned dorsal. Cell-sizing usually required two to four passes in each direction until 100 cells in each laminae were represented in perikaryal drawings. Morphometric software ŽSigmaScan 3.0, Jandel. calculated the digitized cross-sectional areas of the individual somas. Mean soma size for each animal was calculated for the 100-cell sets in the right LGN, one set from lamina-A1 and one from lamina-A. Interlaminar cell-size disparity was expressed as the ratio of the A1rA cell-size means w3x. The A1rA-lamina ratio utilized here and in previous studies effectively factors out individual-animal variability inherent in absolute cell sizes w3,40x. Differences in this ratio between the two groups of kittens were tested for statistical significance with t-tests for independent means, whereas tests between the two laminae within each group were tested with paired t-tests Žtwo-tailed, As 0.05.. At the end of the experimental period, the brains were removed. The spleen, thymus and both adrenal glands were excised and stored in the perfusate. Wet weights of these organs were obtained later to assess the degree of stress induced by either or both the lesions and the periodic isolation of the animals w31x. To serve as control

values, the same organs were obtained from seven similarly aged kittens, reared in their home cages with mothers and siblings, that had received no other treatment. Multivariate analysis of covariance ŽMANCOVA. based on the group-mean weights of the organs ŽAs 0.05. was used to indicate stress-related effects. In this analysis, body weight at sacrifice was used as a covariate factor to control

Table 1 Stress measuresa Adrenals

Spleen

Thymus

Body weight b

0.108 0.027

2.094 0.687

1.112U 0.598

577.8 116.6

LESIONED Ž ns 7. 0.095 S.D. 0.022

1.448 0.546

0.589U 0.407

626.5 95.5

Cage controls Ž ns 5. 0.1246 S.D. 0.023

1.995 0.728

1.748 0.431

529.2 148.7

SHAM Ž ns 7. S.D.

U

Significantly different from Cage controls on Bonferroni corrected, post-hoc t-tests Ž ps 0.001.. a Uncorrected, mean wet weights in grams. b Taken at the end of experiment.

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119

114 Table 2 Mean LGN cell sizes Žmm2 .

3. Results

A1

A

A1rA

SHAM Ž N s 7. S.D.

221.48 32.81

248.69U 48.09

0.90 0.07

LESIONED Ž N s 7. S.D.

215.10 26.52

274.63U 31.43

0.79UU 0.10

UU U

Significantly different from SHAM Ž p- 0.05.. Significantly different from A1-lamina Ž p- 0.05..

statistically for the possible effects of different body sizes on organ weights. Post-hoc comparisons of pairs of group means were made using Bonferroni-corrected t-tests.

3.1. Sleep amounts and effectiÕeness of PGO-waÕe elimination Fig. 1A and B illustrates the effectiveness of bilateral pontomesencephalic lesions in eliminating PGO-waves from the electrophysiological tracing of a kitten that had exhibited distinctive PGO-activity in its baseline record. Lesion sizes ranged from large to small and varied according to the amount of current passed. Irrespective of the size of the lesions Žsee Fig. 2., PGO-waves were eliminated from all but one lesioned animal in the post-lesion sleep recordings. This animal was excluded from cell-sizing analysis because PGO-waves returned after the first postlesion day, leaving the LESIONED group with an n of 7.

Fig. 3. Mean frequency distributions ŽqS.E.M.. of cell-size Žmm2 . class for the LESIONED and SHAM groups generated from individual-animal sets of 100-LGN cells sampled within the A- and A1-laminae of the LGN ipsilateral to the occluded eye. The mean cross-sectional area within the A- and A1-laminae are indicated on each graph by the arrow and also by the number Ž"S.E.M...

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119

Visual inspection of the electrophysiological recordings verified that the lesions did not alter either the tonic desynchronization of the ECoG or the inhibition of the nuchal EMG that are characteristic in REM sleep. In addition, similar quantities of all three vigilance states were observed during baseline and post-lesion recording sessions in both the LESIONED and SHAM groups. In the LESIONED group, mean REM sleep percent Žof total recording time. remained at the baseline level throughout the post-lesion period, and mean SWS percent was slightly increased Žabout 5%., which was not statistically significant. In no individual case did sleep or wake amounts significantly change following a lesion. 3.2. Stress measures: whole body and internal organ weights The mean weights of the spleen, thymus and both adrenal glands in the LESIONED, SHAM and same-age control kittens were analyzed by multivariate analysis of covariance ŽMANCOVA, SPSS 8.0, SPSS. with body weight as the covariate ŽTable 1.. A significant group-byorgan weight effect Ž F6,24 s 3.83, p s 0.008. was found. Only thymus weights differed among the three groups on the post-hoc, Bonferroni-corrected t-tests. Spleen, adrenal and body weights did not distinguish among the three groups. Mean thymus weights in both LESIONED Ž p s 0.001. and SHAM Ž p s 0.038. groups were significantly smaller than in the cage-reared controls. Thymus weight in the SHAM group was numerically larger than in the LESIONED group but this difference was not statistically different. Most of the LESIONED animals showed no post-lesion side effects but several animals exhibited motor ataxia, stereotypic biting and circling Žcf. Ref. w18x.. No LESIONED animals exhibited side effects persisting for more than two days or interfering with normal weight gain Žsee Table 1.. 3.3. Cell-size changes The ratio of the mean sizes of the A1- and A-lamina cells expresses the degree of interlaminar cell-size disparity. A slight size disparity normally exists between the A1and A-laminae cells in both LGNs that favors the A1-cells, which receive only ipsilateral retinal input w19,51,62x. Accordingly, the A1rA ratio in untreated animals is typically greater than 1.0 w3,36,40x. In this study, as in our previous MD studies in which the right eye was patched, the right A1-lamina cells were smaller than the A-lamina cells, causing a reversal of the normative A1rA ratio to less than 1.0 w3,36,40x. The coupled effect of pontomesencephalic lesions and 2 weeks of MD in the LESIONED animals resulted in a significantly smaller A1rA ratio than in the SHAM Ž2 week MD-only. animals Žtwo-tailed t-test, t s 2.31, df s 12, p s 0.039.. Only numerical differences were found in

115

mean A1-cell sizes and mean A-cell sizes across groups, which did not attain statistical significance ŽTable 2.. Within each group, the non-deprived A-lamina cells were statistically larger than the deprived A1-lamina cells Žpaired, two-tailed t-tests, LESIONS; t s 4.68, df s 6, p s 0.034; SHAMS t s 3.061, p s 0.011; see Table 2 and Fig. 3..

4. Discussion Bilateral electrolytic lesions were made in the pontomesencephalic isthmus to block the occurrence of thalamic PGO-waves during the post-natal critical period of visual-system development in kittens. The kittens were lesioned at the halfway point of a 2-week period of MD occurring at the height of the critical period. Pontomesencephalic lesions effectively suppressed PGO-waves in LGN for the duration of the post-lesion survival period in all but one animal ŽFig. 1.. Data from continuous recording confirmed earlier reports that similarly placed, bilateral pontomesencephalic lesions do not affect sleeprwake tracings or stage amounts in either adults or kittens w9,32,47x. Both LESIONED and SHAM animals gained weight at similar rates. Their weights compared favorably at the end of the experiment with cage-reared, unmanipulated controls. In terms of stress, the two treatment groups did not differ significantly on weights of spleen, thymus and adrenal glands w31x. Both groups had lower thymus weights than the unmanipulated cage-control animals, indicating increased stress in the two treatment groups ŽTable 1.. The mean LGN A1rA-lamina cell-size ratio found in the SHAM group was comparable to both the A1rA-ratio that our group previously reported in MD-only animals in an IRSD study w40x and the ratio found by Bear and Coleman in MD-only kittens w3x. The PGO-wave-suppressing lesions, in combination with MD, apparently elicited a lower A1rA measure Ž0.79. than that observed in the MD-alone, SHAM lesion animals Ž0.90. Žsee Table 2.. The ratio in the LESIONED group was numerically similar to the one found in our earlier studies of IRSD in MD kittens Ž0.78., though smaller cell size in the A1-lamina Žreceiving afference from the occluded eye., as reported in our earlier work w36,40x, was not found. The low A1rA-lamina ratio in this study’s LESIONED animals is primarily accounted for by larger A-lamina cells than in the SHAM group ŽTable 2.. Nevertheless, these data support our prediction that selective suppression of PGO-waves in LGN while the REM state is otherwise holistically in progress will alter the pattern of LGN cell growth, i.e., an expression of increased plasticity in visual-system development. The LGN, lamina-specific, increase in cell size also suggests that the lesions do not cause degenerative, antitrophic presynaptic effects on cell growth. After carrying out pontomesencephalic lesions in kittens younger than those used in this study, Davenne and Adrien

116

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119

reported that cells were smaller than normal in the A-lamina of an LGN and also found an overall reduction in the LGN volume w7–9x. Though a degree of presynaptic degenerative effect of the lesions can not be ruled out entirely, the results are consistent with what might be expected from removal of activation in both afferent eye channels. PGOwave suppression leads to equal removal of REM sleep phasic activation in all laminae of both LGNs. This suppression is analogous to the reduction in activation in both eye-specific pathways during dark rearing. The latter has been shown to bilaterally reduce activation of both sets of LGN relay cells and decrease cell growth in all LGN lamina compared to normally sighted kittens w19,51x, as well as to extend the period of plasticity w39x. In our studies, unlike solely pontomesencephalic lesion suppression of LGN PGO-waves w7–9x or singular dark-rearing suppression of visual activation in both eye channels w19,51x, LGN PGO-wave suppression was coupled with single-eye occlusion Ževoking visual-system competitive mechanisms.. Nevertheless, our findings, like those of Davenne and Adrien, implicate brainstem PGO-wave-generation mechanisms as influencing central visual-system development via control of activation reaching the LGN. 4.1. PGO-waÕe suppression effects on LGN cell size We have argued previously that the effects of IRSD on cell-size may be due, in part, to the symmetrical Žboth laminae. removal of activation that would otherwise buffer LGN cells from the effects of the extremely unbalanced visual input of MD Žsee below. w35,36,40,49x. The main hypothesis of this study was that elimination of PGO-wave activation in the LGN of the MD kitten will induce LGN cell-size changes in the A- and A1-laminae that are essentially indistinguishable from the changes observed after suppression of REM sleep by IRSD in MD kittens. Such an outcome would support the suggestion that removal of REM sleep PGO-activation of LGN is sufficient Žand perhaps also necessary. to cause the cell-size rearrangement that we have observed in MD kittens after IRSD w36,40x. The data largely supported this conclusion; namely, both IRSD and bilateral pontomesencephalic lesions in MD kittens suppress the REM sleep PGO-waves that normally reach the LGN, leading to similar changes in interlaminar cell-size disparity in the BS. Though the two experimental approaches in MD animals yielded almost exactly the same A1rA-interlaminar cell-size ratio, this proportional relationship was achieved in a dissimilar way in the two studies. Mean cell size in the non-patched-eye-related A-lamina of the LESIONED group, though not statistically different from that in the SHAM control group Ž p s 0.29., tended to be larger, whereas in the IRSD study, A-lamina cells were similarly sized in the experimental and control groups w40x. Counter to our expectations, the mean cell sizes in the patchedeye-related A1-laminae in LESIONED and SHAM animals

were hardly different, the LESIONED group reflecting only slightly smaller A1-lamina cells than those in the SHAM animals. This was not the case in the IRSD study in which the mean A1-lamina cell size of the MD q IRSD group was substantially smaller than in the MD-only group, though not significantly Ž p s 0.062.. In view of virtually the same A1rA-cell-size ratio results in the two studies, the lack of statistically significant, same-lamina differences between the IRSD q MD and PGO-suppressedq MD experimental groups may render it unnecessary to explain the apparent directional differences in particular lamina cell-size changes in the two groups. The fact, however, that in the IRSD study the patched-eye-related A1-lamina cells tend to be smaller than- and the non-occluded-eye A-lamina cells are virtually the same size as in the MD-only controls, whereas in the PGO-suppressed animals, the A-lamina cells tend to be larger than- and the A1-lamina cells are the same size as in the controls deserves some discussion. Elementary differences exist between the effects of IRSD and bilateral brainstem lesions on appearance and patterning of PGO-waves as well as on tonic activation of REM sleep. These divergences may affect the profile of LGN cell-size changes in response to the two manipulations. First, it should be restated that pontomesencephalic lesions eliminate only the phasic PGO-activation not the tonic activation of the REM state at the LGN. IRSD, on the other hand, eliminates both activations throughout the CNS. It is possible that the different-laminae cell-size results are attributable to the aggregate visual, phasic-REM sleep and tonic-REM sleep activations reaching the two laminae. In the IRSD study, A1-lamina cells are deprived of all three sources of activation: visual input, phasic REM sleep activation, and tonic REM sleep activation. In contrast, in the lesion study tonic activation at the LGN is retained. As a result, the A1-lamina cells in the lesion study may be spared the full weight of the deactivation effects that occur in the IRSD study. On the other hand, in the lesion study, A-lamina cells receive both visual and tonic REM sleep activations, whereas in the IRSD study only visual activation is retained. Accordingly, the additional activation of the A-lamina cells in the lesion study could account for the larger size than in the IRSD study. Secondly, with IRSD some PGO-waves are expressed in SWS, and they occur as well in the pre-arousal intervals of REM sleep during REM sleep deprivation. The PGO-waves also appear to increase throughout the deprivation period w11x. ŽPGOwaves appearing in SWS, however, may have little effect on neuronal plasticity because they occur against a nonactivated cortical background w53,57,58x.. Lastly, unlike IRSD, pontomesencephalic lesions disrupt the putative cholinergic fiber tract relaying PGO-wave activation to LGN w4,7,32x. The lesions employed in this study not only suppress PGO-waves in the LGN in all sleep states but also disrupt the LGN’s cholinergic activation in the wak-

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119

ing state. Suppression of cholinergic innervation of LGN in the lesion paradigm but not with IRSD could also possibly result in differential lamina cell-size effects. 4.2. Effects of PGO-waÕe suppression on LGN cell plasticity What mechanisms might account for heightened plastic changes in the visual system of the MD kitten after PGO-wave elimination? Thalamocortical relay cells in the geniculate appear to give rise to a common visual-sensory pathway through which activation due to either visual experience or endogenous REM-sleep processes impinge upon the central visual system and influence developmental processes. REM sleep-related phasic activity in LGN associated with PGO-waves is propagated by a cholinergic pathway from the brainstem w21,52x. Brainstem cholinergic synapses are anatomically distributed upon neural elements in the LGN in the same way as synapses of retinal origin w10x. In addition, the majority of LGN neurons in the cat show facilitation of discharge rate associated with PGOwaves, as is observed with certain types of visual stimulation w2,44x. However, retinal and brainstem cholinergic innervations of the LGN differ in significant aspects of their organization: retinal input from each eye segregates into eye-specific LGN laminae whereas individual cholinergic brainstem axons appear to reach both eye-specific ŽA and A1. laminae alike w51,56,60x. Extracellular unit recording in our laboratory has also shown that cells in the Aand A1-laminae are synchronously facilitated as PGOwaves occur w35,36x. This synchronous, cross-laminar, REM-sleep phasic activity in the LGN parallels binocular visual experience in terms of concurrent A1- and A-laminae activation, and may have similar activity-dependent effects in the visual system, opposing the effect of aberrant visual input such as produced by MD. Accordingly, in the present study, the greater than 20% of the day that the MD-only ŽSHAM. kittens were in REM-sleep may partially offset the effects of the 25% of the day that the animals were awake in the light, experiencing unbalanced visual input. We have seen that suppression of PGO-waves at the LGN, either by means of pontomesencephalic lesions or IRSD, may remove this asymmetry-opposing and bilateral type of activating signal, resulting in amplification of the interlaminar-disparity effects of MD. Altered cortical excitability is another mechanism that may be related to the effects on LGN cell sizes of removal of REM-specific phasic activity. Data along these lines has emerged in recent investigations into the phenomenon of cortical long-term depression ŽLTD. w13,14x. Manipulations that lower cortical excitability antagonize the effects of aberrant visual experience, and, apparently, the potential for plastic change diminishes with reduced cortical excitability w3,26,27,57x. The increased excitability associated with cortical desynchrony, for example, has been shown to be necessary Žor perhaps sufficient. for plastic changes to occur in anesthetized animals in response to visual stimula-

117

tion w53x. Several reports indicate that one effect of IRSD, and possibly also of PGO-wave suppression, is an increase in cortical excitability w41,46x. LTD can be induced in visual cortex by trains of low frequency stimulation w14,30x. This phenomenon appears to be age-restricted, operating during the time of the critical period of visual development but not in adulthood w13,14,30x. Accordingly, LTD has been proposed to play a role in the mechanisms mediating plasticity in the developing visual cortex w29,30x. Inasmuch as PGO-waves appear in visual cortex at a rate Žapproximately 60 per min. corresponding to the frequency of stimulation that induces LTD, the periods of PGO-activity in young animals may operate to reinforce chronic synaptic depression w12,24x. Either IRSD or brainstem lesions, both of which remove this PGO-wave influence, should increase cortical excitability and thereby enhance the effects of MD. Increased levels of cortical excitability may also contribute to the observed changes in LGN cell sizes after PGO-wave suppression through effects on the expression of neurotrophins. Recent reports support the idea that neurotrophic factors and their tyrosine kinase receptors are critically involved in visual-system maturation w1,6,15, 16,28,42,59x. For example, brain derived neurotrophic factor ŽBDNF. synthesis and secretion is regulated in visual cortex by activity-dependent mechanisms w37,59x. We propose that increased cortical excitability due to PGO-wave suppression may result in increased BDNF expression in visual cortex, which in conjunction with altered visual input may produce the observed cell-size changes in LGN. In conclusion, this study provides additional empirical support for our hypothesis that a primary, early function of the plentiful amounts of REM sleep in younger mammals and birds is to provide necessary, endogenously generated, neuronal activation for normative, activity-dependent, ontogenetic development of the CNS. Like removal of the whole of the REM state, experimental suppression during development of just the phasic REM-sleep process, PGOwave activation at the level of the LGN, exaggerates the CNS plasticity of LGN cells. It appears that at least some effects of IRSD on LGN cell size are mediated solely by the phasic processes of REM sleep. Acknowledgements The authors thank Christian Birabil and Zizhuang Li for their technical assistance. This work was supported by NIH Grant NS31720. References w1x K.L. Allendoerfer, R.J. Cabelli, E. Escandon, D.R. Kaplan, K. Nikolics, C.J. Shatz, Regulation of neurotrophin receptors during the maturation of the mammalian visual system, J. Neurosci. 14 Ž1999. 1795–1811. w2x G. Aston-Jones, C. Chiang, T. Alexinsky, Discharge of noradrener-

118

w3x

w4x

w5x

w6x

w7x

w8x

w9x

w10x

w11x

w12x

w13x

w14x

w15x

w16x

w17x

w18x

w19x

w20x

w21x

w22x

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119 gic locus coeruleus neurons in behaving rats and monkeys suggests a role in vigilance, Prog. Brain Res. 88 Ž1991. 501–520. M.F. Bear, H. Coleman, Binocular competition in the control of geniculate cell size depends upon visual cortical N-methyl-D-aspartate receptor activation, Proc. Natl. Acad. Sci. U.S.A. 87 Ž1990. 9246–9249. E. Bizzi, D.C. Brooks, Functional connections between pontine reticular formation and lateral geniculate nucleus during deep sleep, Arch. Ital. Biol. 101 Ž1963. 666–680. C. Bowe-Anders, J. Adrien, H.P. Roffwarg, Ontogenesis of activity in the lateral geniculate nucleus of the kitten, Exp. Neurol. 43 Ž1974. 242–260. R.J. Cabelli, A. Hohn, C.J. Shatz, Inhibition of ocular dominance column formation by infusion of NT-4r5 or BDNF, Science 267 Ž1995. 1662–1666. D. Davenne, J. Adrien, Suppression of PGO waves in the kitten: anatomical effects on the lateral geniculate nucleus, Neurosci. Lett. 45 Ž1984. 33–38. D. Davenne, J. Adrien, Electrophysiological maturation of the lateral geniculate neurons after lesions of the ponto-geniculo-occipital ŽPGO. pathways in kittens, Comp. Rend. Acad. Sci. ŽParis. 300 Ž1985. 59–64. D. Davenne, J. Adrien, Lesion of the ponto-geniculo-occipital pathways in kittens: I. Effects on sleep and unitary discharge of the lateral geniculate nucleus, Brain Res. 409 Ž1988. 1–9. A.D. De Lima, V.M. Montero, W. Singer, The cholinergic innervation of the visual thalamus: an EM immunocytochemical study, Exp. Brain Res. 59 Ž1985. 206–212. W. Dement, The biological role of REMS sleep Žcirca 1968., in: A. Kales ŽEd.., Sleep: Physiology and Pathology, Lippincott, PA, 1969, pp. 245–265. W.C. Dement, The biological role of REMS sleep Žcirca 1968., in: A. Kales ŽEd.. Sleep: Physiology and Pathology, Lippincott, PA, 1969, pp. 245–265. S.M. Dudek, M.F. Bear, Bidirectional long-term modification of synaptic effectiveness in the adult and immature hippocampus, J. Neurosci. 13 Ž1993. 2910–2918. S.M. Dudek, M.J. Friedlander, Developmentally dependent long-term depression of neocortical synapses in vitro, Neurosci. Abstr. 19 Ž1993. 370.23. J.P. Fawcett, S.X. Bamji, C.G. Causing, R. Aloyz, A.R. Ase, T.A. Reader, J.H. McLean, F.D. Miller, Functional evidence that BDNF is an Anterograde neuronal trophic factor in the CNS, J. Neurosci. 18 Ž1998. 2808–2821. R.A.W. Galuske, D.S. Kim, E. Castren, H. Thoenen, W. Singer, Brain-derived neurotrophic factor reverses experience-dependent synaptic modification in kitten visual cortex, Eur. J. Neurosci. 8 Ž1996. 1554–1559. R.W. Guillery, D.J. Stelzner, The differential effects of unilateral lid closure upon the monocular and binocular segments of the dorsal lateral geniculate nucleus in the cat, J. Comp. Neurol. 139 Ž1970. 413–422. J.B. Harrison, N.J. Woolf, J.S. Buchwald, Cholinergic neurons of the feline pontomesencephalon: I. Essential role in Wave A generation, Brain Res. 520 Ž1990. 43–54. T.L. Hickey, P.D. Spear, K.E. Kratz, Quantitative studies of cell size in the cat’s dorsal lateral geniculate nucleus following visual deprivation, J. Comp. Neurol. 172 Ž1977. 265–282. J.A. Hobson, R.W. McCarley, Neuronal Activity in Sleep: An Annotated Bibliography, UCLA Brain Information Service, Los Angles, CA, 1971. B. Hu, M. Steriade, M. Deschenes, The cellular mechanisms of thalamic ponto-geniculo-occipital ŽPGO. waves, Neuroscience 31 Ž1989. 25–35. D.H. Hubel, T.N. Wiesel, Single-cell responses in striate cortex of kittens deprived of vision in one eye, J. Neurophysiol. 26 Ž1963. 1003–1017.

w23x D.H. Hubel, T.N. Wiesel, The period of susceptibility to the physiological effects of unilateral eye closure in kittens, J. Physiol. ŽParis. 206 Ž1970. 419–436. w24x M. Jouvet, The role of monoamines and acetylcholine-containing neurons in the regulation of the sleep–waking cycle, Ergebn. Physiol. 64 Ž1972. 165–307. w25x R.E. Kalil, A quantitative study of the effects of monocular enucleation and deprivation on cell growth in the dorsal lateral geniculate nucleus of the cat, J. Comp. Neurol. 189 Ž1980. 483–524. w26x T. Kasamatsu, T. Ohashi, K. Imamura, Lithium reduces ocular dominance plasticity in kitten visual cortex, Brain Res. 558 Ž1991. 157–162. w27x T. Kasamatsu, J.D. Pettigrew, Depletion of brain catecholamines: failure of ocular dominance shift after monocular occlusion in kittens, Science 194 Ž1976. 206–209. w28x L.C. Katz, C.J. Shatz, Synaptic activity and the construction of cortical circuits, Science 274 Ž1996. 1133–1138. w29x A. Kirkwood, M.F. Bear, Homosynaptic long-term depression in the visual cortex, J. Neurosci. 14 Ž1993. 3404–3412. w30x A. Kirkwood, S.M. Dudek, J.T. Gold, C.D. Aizenman, M.F. Bear, Common forms of synaptic plasticity in the hippocampus and neocortex in vitro, Science 260 Ž1993. 1518–1521. w31x V.M. Kovalzon, V.L. Tsibulsky, REM-sleep deprivation, stress and emotional behavior in rats, Behav. Brain Res. 14 Ž1984. 235–245. w32x J.P. Laurent, R. Cespuglio, M. Jouvet, Delimitation des voies ascendantes de l’activite PGO ches le chas le chat, Brain Res. 65 Ž1973. 29–52. w33x S. LeVay, M.P. Stryker, C.J. Shatz, Ocular dominance columns and their development in layer IV of the cat’s visual cortex: a quantitative study, J. Comp. Neurol. 179 Ž1978. 223–244. w34x M.S. Loop, S.M. Sherman, Visual discrimination during eyelid closure in the cat, Brain Res. 128 Ž1978. 329–339. w35x G.A. Marks, H.P. Roffwarg, J.P. Shaffery, Neuronal activity in the lateral geniculate nucleus associated with ponto-geniculo-occipital waves lacks lamina specificity, Brain Res. 815 Ž1998. 21–28. w36x G.A. Marks, J.P. Shaffery, A. Oksenberg, S.G. Speciale, H.P. Roffwarg, A functional role for REM sleep in brain maturation, Behav. Brain Res. 69 Ž1996. 1–11. w37x S. Marty, M.P. Berzaghi, B. Berninger, Neurotrophins and activitydependent plasticity of cortical interneurons, Trends Neurosci. 20 Ž5. Ž1997. 198–202. w38x M. Meister, R.O. Wong, C.J. Shatz, Synchronous bursts of action potentials in ganglion cells of the developing mammalian retina, Science 252 Ž1991. 934–943. w39x G.D. Mower, The effect of dark rearing on the time course of the critical period in cat visual coretex, Dev. Brain Res. 58 Ž2. Ž1991. 151–158. w40x A. Oksenberg, J.P. Shaffery, G.A. Marks, G.A. Speciale, G. Mihailoff, H.P. Roffwarg, Rapid eye movement sleep deprivation in kittens amplifies LGN cell-size disparity induced by monocular deprivation, Dev. Brain Res. 97 Ž1996. 51–61. w41x M. Owen, E.L. Bliss, Sleep loss and cerebral excitability, Am. J. Physiol. 218 Ž1969. 171–173. w42x D.R. Riddle, D.C. Lo, L.C. Katz, NT-4-mediated rescue of lateral geniculate neurons from effects of monocular deprivation, Nature 378 Ž1995. 189–191. w43x H.P. Roffwarg, J.N. Muzio, W.C. Dement, Ontogenetic development of the human sleep–dream cycle, Science 152 Ž1966. 604–619. w44x H. Sakakura, Spontaneous and evoked unitary activities of cat lateral geniculate neurons in sleep and wakefulness, Jpn. J. Physiol. 18 Ž1968. 23–42. w45x K.S. Sanderson, The projection of the visual field to the lateral geniculate and medial interlaminar nuclei in the cat, J. Comp. Neurol. 143 Ž1971. 101–118. w46x E. Satinoff, R.R. Drucker-Colin, R. Hernandez-Peon, Paleocortical excitability and sensory filtering during REM sleep deprivation, Physiol. Behav. 7 Ž1971. 103–106.

J.P. Shaffery et al.r DeÕelopmental Brain Research 114 (1999) 109–119 w47x J.P. Shaffery, T. Fitch, S.G. Speciale, H.P. Roffwarg, R. Armitage, G.A. Marks, Sleep architecture in kittens after bilateral mesencephalic lesions that eliminate ponto-geniculo-occipital ŽPGO. waves in the lateral geniculate nucleus ŽLGN., Soc. Neurosci. Abstr. 21 Ž1995.. w48x J.P. Shaffery, G.A. Marks, S.G. Speciale, H.P. Roffwarg, Computer-aided sleep staging in young kittens, Sleep Res. 21 Ž1992.. w49x J.P. Shaffery, A. Oksenberg, G.A. Marks, S.G. Speciale, G. Mihailoff, H.P. Roffwarg, REM sleep deprivation in monocularly occluded kittens reduces the size of cells in the LGN monocular segment, Sleep 21 Ž1998. 837–845. w50x C.J. Shatz, Viktor Hamburger Award review. Role for spontaneous neural activity in the patterning of connections between retina and LGN during visual system development, Int. J. Dev. Neurosci. 12 Ž1994. 531–546. w51x S.M. Sherman, P.D. Spear, Organization of visual pathways in normal and visually deprived cats, Physiol. Rev. 62 Ž1982. 738–855. w52x P.J. Shiromani, K.M. Armstrong, A. Berkowitz, D.V. Jeste, Distribution of choline acetyltransferase immunoreactive somata in the felline brainstem: implications for REM sleep generation, Sleep 11 Ž1988. 1–16. w53x W. Singer, J.P. Rauschecker, Central core control of the developmental plasticity in the kitten visual cortex: II. Electrical activation of mesencephalic and diencephalic projections, Exp. Brain Res. 47 Ž1982. 223–233. w54x R.H. Spiro, G.S. Kolberg, A new technique for functional visual deprivation, Clin. Neurophysiol. 37 Ž1974. 654–656.

119

w55x M. Steriade, J.A. Hobson, Neuronal activity during the sleep–waking cycle, J. Neurobiol. 6 Ž1976. 155–176. w56x M. Steriade, D. Pare, A. Parent, S. Smith, Projections of cholinergic and non-cholinergic neruons of the brainstem core to relay and associational thalamic nuclei in the cat and macaque monkey, Neuroscience 25 Ž1988. 47–67. w57x M.P. Stryker, The role of neural activity in rearranging connections in the central visual system, in: R.J. Ruben, E.W. Van De Water, E.W. Rubel ŽEds.., The Biology of Change in Otolaryngology, Elsevier, Amsterdam, 1986, pp. 211–224. w58x M.P. Stryker, W.A. Harris, Binocular impulse blockade prevents the formation of ocular dominance columns in cat visual cortex, J. Neurosci. 6 Ž1986. 2117–2133. w59x H. Thoenen, Neurotrophins and neruonal plasticity, Science 270 Ž1995. 593–598. w60x D.J. Uhlrich, J.B. Cucchiaro, S.M. Sherman, The projection of individual axons from the parabrachial region of the brainstem to the dorsal lateral geniculate nucleus in the cat, J. Neurosci. 8 Ž1988. 4565–4575. w61x R. Ursin, M.B. Sterman, A Manual for Standardized Scoring of Sleep and Waking States in the Adult Cat, Brain Information Service, Los Angeles, 1981, pp. 103. w62x T.N. Wiesel, D.H. Hubel, Effects of visual deprivation on morphology and physiology of cells in the cat’s lateral geniculate body, J. Neurophysiol. 26 Ž1963. 978–993. w63x R.O. Wong, M. Meister, C.J. Shatz, Transient period of correlated bursting activity during development of the mammalian retina, Neuron 11 Ž1993. 923–938.