Aging changes in synaptology of luteinizing hormone-releasing hormone neurons in male rat preoptic area

0306-4522/8753.00+0.00

Nrurhcclcnce Vol. 22, No. 3, pp. 1003-1013,1987 Printedin Great Britain

PergamonJoamah Ltd 0 1987 IBRO

AGING CHANGES IN SY-NAPTOLOGY OF LUTEINIZING HORMONE-RELEASING HORMONE NEURONS IN MALE RAT PREOPTIC AREA J. W. WITKIN Department of Anatomy and Cell Biology, Columbia University College P. & S., 630 West 168th Street, New York, NY 10032, U.S.A.

Ahatrac-This study was undertaken to examine some aspects of the anatomical substrate for reproductive senescence. Immmmcytochemically identified luteinizing hormone-releasing hormone neurons and their processes in the male rat brain preoptic area were compared in young adult (2-4 months), middle-aged (12-14 months) and old (20-23 months) animals. At the light microscopic level there were no age-dependent differences in total numbers or sixes of LHRH neurons nor in their distribution in the brain, Examination of these neurons at the electron microscopic level did teveal sign&ant differences in certain organelles and in the degree and kind of synaptic input. Random sections of middle-aged luteiniaing hormone-releasing hormone neurons more frequently passed through the nucleohts and the incidence of nematosomes was higher than in luteinixing hormone-releasing hormone neurons from the young and old animals. Quantitative measures of synaptic input to luteini&g hormont-rcltasing hormone soma and dendrites as well as to unidentified neurons in the same thin section were made. These are reported as percent of membrane that showed synaptic structure. Dendrites of both luteinixing hormone-releasing hormone and nonidentified neurons were more densely innervated than perikarya. The density of synaptic input to luteinixing hormoncreleasing hormone neurons was significantly greater than that to nonidentitied neurons in young and middle-aged animals, but was equal to that of nonidentified neurons by old age. Age-related changes were noted in synaptic organization with the most significant

change being an increased input to luteinixing hormone-releasing hormone perikarya. Indeed, synaptic input to luteinizing hormoncreleasing hormone perikaryal membrane was increased three-fold by middle age and ten-fold by old age. Density of synaptic input to luteinizing hormone-releasing hormone dendritic membrane did not change with age. There were no aging changes in percentage of membrane with synaptic structure in nonidentified elements. Synapses were also classified on the basis of their synaptic vesicle content. There were proportionately more synaptic boutons containing round clear than pleomorphic vesicles in the young sample. The proportion of synapses with pleomorphic vesicles increased with age onto both luteinizing hormone-releasing hormone perikarya and their dendrites. The proportion of boutons containing some electron dense-core vesicles along with clear vesicles decreased with age onto both luteinizing hormone-releasing hormone and nonidentified neurons and their processes.

preoptic arear and of losses in dendritic arborization.*’ Matsumoto et al.“’ found decreases of more than 50% between 3- and 24month-old male rats in total numbers of both axosomatic and axodendritic synapses in the arcuate nucleus (through which many LHRH processes course “en route” to the median eminence). No electron microscopic studies of the synaptology of immunocytochemically identified LHRH neurons in the preoptic area in aging animals have been reported. The purpose of this study was to examine LHRH neurons and their processes in the preoptic area in male rats: young adult (2-4 months), middle-aged (12-14 months) and old (20-23 months), to see whether age-related differences in their synaptic input accompany decrements in the affective milieu of their peptide hormone.

There are aging changes in the hypothalamicpituitary-gonadal axis in male rats4ga which result in decreased serum levels of testosterones*4’~s’ and decreased amplitude and frequency of luteinizing hormone pulses in gonadectomixed rats.” These declines may be attributable, at least in part, to alterations in the release of luteinizing hormonereleasing hormone (LHRH). The LHRH content of the median eminence has been found by radioimmunoassayM and by immunocytochemistry12 to be diminished in aged male rats. The source of LHRH to the median eminence of rats is neurons located in several anterior regions, including the preoptic area, the diagonal band of Broca and medial septum. 47,48The largest population of LHRH neurons is that in the preoptic area.57 There were early reports of age-related neuron losses in the

EXPERIMENTAL PROCEDURES

Abbreviations: LHRH, luteinixing hormone-releasing hormone.

The animals used in this study were male F344 rats, purchased from Harlan-Sprague Dawley and housed, 2 per

1003

1004

J. W.

cage. They were maintained in a temperature controlled facility with 14:lO L:D and food and water ad libirum. Three age groups were studied, young adults (N = 7) (2-4 months), middle-aged (N = 7) (12 -14 months) and old (N = 5) (20-23 months). Rats were anesthetized with methoxyfhuane (metofane, Pitman-Moore) and then with pentobarbitol (4Omg/kg body weight i.p.) before perfusion via the left cardiac ventricle with 500 ml of fixative (4% pamformaldehyde and 0.1% glutaraldehyde in 0.1 M pH 7.3 phoqhate buffer). Brains were removed. cut into blocks containing the preoptic area and postied at room temperature with agitation in the same fixative for about 2 h. Sixty-pm coronal sections were cut on a Vibratome and immunocytochemical procedures for the demonstration of sites of LHRH were carried out as described earliep on all sections. The antiserum to LHRH was LRI (code 20-679, Salk Institute, kindly provided by Dr R. Benoit) used at 1: 20,000 in 0.02% saponin in phosphate buffer. (This antibody was prepared in rabbits using a @Lys6] LRF glutaraldehyde ovalbumine conjugate. The antiserum recognizes amino acids 3, 4, 7, 8, 9 and IO of the decapeptide.) Incubations in primary antibody were for 72 h at 4°C. Visualization of sites of immunoreactivity was effected by the use of the avidin biotin horseradish peroxidase method (Vectastain) with 3,3’diaminobenzidine tetrachloride as the chromogen, enhanced with 0.5% cobalt chl~ride.‘~ Controls in which the primary antiserum was omitted or absorbed with synthetic LHRH resulted in no reaction product. In addition, this antibody does not recognize any neural structures in the hypogonadal (hpg) mouse9 which is genetically deficient in LHRH. Pituitaries and adrenal glands of all aging animals were examined histologically for the presence of tumors. None was found. For electron microscopic study, areas of interest were removed from tissue sections. (The area chosen for study was the preoptic area which is not a unitary structure but contains several subgroups.’ The region from which blocks were taken was the preoptic area within I mm from the organum vasculosum of the lamina terminalis. This region encompasses the periventricular, medial and lateral preoptic areas. Because of the techniques used, sections could not be counterstained before removing electron microscope sections, so precise cytoarchitectonic localization was not possible.) These tissue sections were osmicated (2% osmium with I .5% potassium ferricyanide*’ in 0.9% saline solution for I h), dehydrated and embedded in EPON. Thin sections (70 nm) were collected on formvar-coated slot grids, stained with uranyl acetate and lead citrate55 and examined with the Jeol IOOS or Jeol IZOOEX electron microscope. Sections were photographed at x 10,000 and printed with final magnifications of about ~25,000. For neurons, a single profile of each perikaryon was used. In general this entailed 5 micrographs. A total of 20 electron micrographic fields was used for each animal for the study of synaptic input lo dendrites. Five LHRH and five nonidentified neurons from the same electron microscope blocks were measured per animal for perikaryal input. The nonidentified neurons comprise a mixed population and therefore are questionable as “controls”. However, to date, all studies of m~rphologid and ultrastructuraPp” aging change=3have been performed on nonidentiaed mixed populations of neurons. The present observations can thus be compared with earlier Andings. ne number of neurons analyaed in each category for each animal is small (N = 5) because the time required to locate, thin section, photograph, and measure each cell is considerable. However, by using strict statistical criteria, accounting for sample size (see Results), statements CM be made about differences among the populations which were measured. Measulements of lengths of dendritas, *al mem-

branes and synapses were made using a Houston In&u-

WITKIN

ments Bioquant program, an X-Y digitizing pad and an Apple II plus computer. LHRH elements were recognized on the basis of the presence of black reaction product. Nonidentified neurons were selected by photographing the first neuron with a nucleus containing a nucleolus seen in an electron microscope section. The perikaryon was defined as the area within IOpm of the nuclear membrane. Synaptic length was taken to be membrane along which there was any recognizable synaptic structure: a synaptic cleft, a pre- or postsynaptic density or. in the case of “en face” sections, a density across both pre- and postsynaptic elements. Comparisons of synaptic input were made using percentage of membrane containing synaptic structure and analysed using analysis of variance on log transformed values, followed by an analysis of covariance. At the light microscopic level, numbers of LHRH neurons in the diagonal band of Broca, medial septum and preoptic and anterior hypothalamic areas were counted in each animal. (Neurons cut out of sections for electron microscope examination were added to the totals.) In addition, the areas and perimeters of LHRH neurons (except those removed for electron microscope examination) were measured using the Bioquant program and digitizing pad.

RESULTS Ligh! microscopy Total numbers of LHRH neurons in the medial septum, diagonal band of Broca and pnoptic and anterior hypothalamic areas for each animal are shown in the scatterplot (Fig. I). There were no age-associated differences in numbers of neurons (by analysis of variance). Total LHRH neurons counted in this region ranged between less than 100 and more than 250. Although all sections were counted, this is an underestimation of the total population as the immunological reagents penetrate to a maximum depth of IOpm, due in part to the gentle detergem treatment. In a 6Opm section, 2/3 of the tissue is thus unavailable lo the immunocytochemistry and presumably 213 of the neurons would not be identified. Underestimations may be even higher if the Vibratome section happens to fall such that the

Aging LHRH neurons dense population of LHRH neurons in the diagonal b~d/preop~~ regions7 is not exposed at either surface of the section and so is unavailable for immunocytochemicai demonstration. This probably accounts for the lowest numbers counted and for the variability in numbers. The areas and perimeters of all LHRH neurons in each of the three age groups were compared by analysis of variance and were found not to differ by age.

1005

had fewer synapses than did nonidentifi~

Electron microscopy Sites of LHRH immunore~tivity in male rat brain preoptic area were recognized by the presence of the chromogen, rendered electron dense by o~ication. The distribution of reaction product was indistinguishable in young adult, middle-aged and old neurons (Fig. 2). That is, it was found scattered in the cytoplasm, associated with ribosomes and the rough endoplasmic reticulum, and in vesicles, The Golgi apparatus was consistently free of reaction product. The rough endoplasmic reticulum of the young and middle-aged LHRH neurons had a similar appearance, with no indication of hypertrophy; however, the diameter of the cistemae was not measured. There was a higher incidence of nucleoli and of the ~topiasmic organelle, the nematosome, in LHRH neurons from the middle-aged than either the young or old groups (Fig. 2B). Total nucleoli/total LHRH neurons =0.39 in the young group, 0.64 in the middle-aged and 0.29 in the old group. Total nematosomes/total LHRH neurons =0.06 in the young group, 0.20 in the middle-aged, and 0.11 in the old group. In that nonidentified neurons were selected randomly by photographing the first neuron with a nucleolus in the plane of section, only nematosomes could be counted in this population. Their incidence was 0.15 in young, 0.03 in middle-aged and 0.03 in old animals. Sy~~~~e~. Synaptic inputs were observed both on imm~orea~tive and nonreactive neurons and their processes. Numbers of synapses per neuronal profile are shown in Table 1. (These numbers are based on synapses per single micrograph and so represent only a sample of the input to any neuron.) LHRH neurons

Table 1. Numbers of synapses per neuronal profile expressed as percent of total profiles in age

Middle-aged Old Nonid.* Young Old

41 5

33 4.5

22 9

4 18

5

5

13

2.5 25 12

35 20 19

23 28 26

I 10 12

I 10 4

0 4 19

3 2 8

*Nonid. = nonidentified neurons.

observations are summarized in Table 2. Dendrites versus soma. Dendrites were more densely innervated than perikarya in both .LHRH and nonidentified neurons in young and groups. However, in the old sample, while input to dendrites of nonidentified neurons was still greater than that to their soma, the input to soma of LHRH neurons was greater than that to LHRH dendrites. Luteinbing hormone-releasing

middle-aged animals but was equal in old animals. Aging changes. There was an increase in synaptic input to the perikarya of LHRH neurons with age (Table 2). The input was increased three-fold (P -C0.05) by middle age and ten-fold by old age (P c 0.001). In nonidentified neurons, there was no change by middle age. By old age, there was a slight increase in input over that of young animals; however, this was not (P > 0.05). Types of synapses. Synaptic boutons were classified on the basis of their vesicle content. Vesicles were either clear and round with diameters ranging from 30 to 80 nm, or or had electron-dense

pleomorphic; (3) clear round with 3 or less dense core vesicles; (4) with 3 or less dense core vesicles; and (5) many dense core vesicles. Due to the density of chromogen deposited at sites of immunoreactivity, postsynaptic density could be made. In analysing the data, some categories of bouton type were combined in order to make comparisons: (designated “round”) were compared with all those containing vesicles “pleomorphic”); (designated “clear”) were compared with those containing some dense core vesicles “dense”). These are shown in Figs 4

Fig. 2. LHRH neurons from (A) young adult (2 months); (B) middle-aged (12 months); and (C) old (20 months) male rat brain preoptic area. These portions of immunoreactive neurons demonstrate distribution of reaction product. It is scattered in the cytoplasm (arrowheads) and associated with rough endoplasmic reticulum (RER) but not with Golgi (G). A higher incidence of nucleoli and of nematosomes (Ne) was found in LHRH neurons of middle-aged rats. N, nucleus. 1006

Aging LHRH neurons Table 2. Percent of membrane with synaptic modification Age (months)

LHRH

Nonid.*

Perikarya Bassdan5ncuronain~~~~animals$*~~~~P). 2-4 0:70 f 0:16t 10-12 2.40 f 0.28% 20-23

2:17 f 0:35 2.90 f 0.58

Dendrites Rased on 1500~m2 per animal. 1.04 f 0.23 2-4 (N=7) 0.84 f 0.39 10-12 (N = 5) 1.87 f 0.51 20-23 (N = 5)

5.47 f 0.34 5.52 f 0.46 5.65 f 0.24

*Nonid. = nonidentified neurons. -iGreater than 2-4 months (ANCOVA, P < 0.05). $Greater than 24 months and 10-12 months (ANOVA, P < 0.001). Mean f S.E.M. Note on statistical analyses: comparisons of data from middle-aged animals and young animals were made in two steps to ensure that the apparent differences were statistically significant. An analysis of variance (ANOVA) with repeated measures was performed upon the log transform of the synaptic length data presented in the table (Perikarya) for neuronal cell bodies. The data were log transformed because the variance appeared to be proportional to the value of the mean. Hence variability differed for measurements on LHRH and nonidentified neurons. In the ANOVA the neuron type was the repeated measures factor, and age was the indenendent variable. The ANOVA revealed a sianificant effect of age upon synaptic density (F = 6.297; d? 1,8; P = 0.0351) and a highly significant difference between LHRH and nonidentified neurons in terms of synaptic input to their soma (F = 64.306, df 1,8; P = 0.0002). In addition,

and 5. It should be noted that these comparisons are of total numbers of boutons of various types in all animals in each age group. The numbers for each category for each animal were so small that no statistical analysis was performed. We found two aging changes in the contents of synaptic boutons: (1) there was an increase in the percentage of synapses containing pleomorphic vesicles onto LHRH neurons (Fig. 4A) and their dendrites (Fig. 4C); (2) there was a general, though slight, decrease in the occurrence of dense core vesicles in terminals on all elements (Fig. 5). As a final note to these observations, very little LHRH-LHRH synaptic interaction in the preoptic area was found. In all micrographs studied, there were only two examples of such synapses, one axodendritic and one dendro-somatic, both of these were in the young adult sample.

In

male months.” decreased reduced

by

reproductive is middle (10-12 there of testosterone**4’*‘1 and of

1007

the analysisrevealed a trend towards a siguiticant age-byneuron-type interaction (F = 2.322, df 1,S;0.20 > P > 0.10). This potentially selective effect of age upon LHRH neurons exammed more closely with an analysis of covariance &a&WA), as will be dexribed below. An equivalent ANOVA with repeated measwas performed upon the log transform of the synaptic density data presented in the table (De&i&s) for dendrites. (Separate ANOVAs were performed for perikarya and dendrites because of the different number of subjects contributing to each data set.) As with the cortmponding data for cell soma, there was a highly -- sit&cant difference in the amount of synaptic input to LHRH as opposed to nonidentified dendrites (F = 158.0: df 1.11: P
P = 0.22). Potential age-related changes in the synaptic input to LHRH neurons were examined more closely with an analysis of covariance (ANCOVA), with synaptic input to nonidentitied neurons serving as the covariate. This analysis was intended to control statistically for the effects upon the LHRH data of individual diKetences in overall synaptic density in the preoptic area as indexed by the data drawn upon the nonidentified neurons. An ANCGVA performed upon the LHRH soma data revealed that synaptic input to LHRH cell bodies was &&antly increased in aged subjects (F = 5.732; df 1.7; P = 0.046). On the other hand, synaptic input to LHRH dendrites was unchanged in aged subjects (F = 1.305; df 1,lO; P = 0.280, not significant). This pair of results suggests that age selectively atfects synaptic input upon the cell bodies of LHRH neurons, and not upon their dendrites. Comparisons of data from old animals with those from the young and middle-aged samples were performed using analysis of variance alone as the F values were highly significant.

hormone in changes age-related generator drives LHRH were to whether number synaptology be downturn reproductive Light

electron

animals.” in LHRH system. this immunocytoin with

investigations

neurons their were in adult months), (12-14 and (20-23 male at light electron levels. the microlevel age in or (areas perimeters) cells an including diagonal of the septum the and hypothalamic were Similar on numbers reported Sladek a1.52 parallel earlier in aging macaque which were agedifferences numbers LHRH in sample animals from to 20 Hoffman FinchlO reported change LHRH number female up middle In rats, increases LHRH-

J. W. WITKIN

Fig 3. Types of synapses onto LHRH elements (D, dendrite; P, perikaryon). (A) Bouton contains clear round vesicles (30-50 nm diameter) (2-month-old rat). (B) Bouton contains pleomorphic (ellipsoidal and flattened) clear vesicles (12-month-old rat). (C) Synapse on perikaryon contains a few electron-dense core vesicles (100 nm diameter) and clear round vesicles (50-60 nm diameter). No section directly through synaptic cleft but grazing section of synaptic densities (asterisks) (4-month-old rat). (D) Bouton contains mainly electron-dense cored vesicles (IO&130 nm diameter). Synaptic cleft visible only at arrow (Cmonth-old rat).

i~rnunor~ac~v~

neurons

have been reported.32 These

increases were interpreted to mean that the peptide had accumulated in the soma in aged animals because it was being released at a slower rate at the terminals than in young animals. The apparent stability of the LHRH neuron population throughout the lifetime of both male and female animals suggests that aging deficits in brain content of the decapeptide52~‘2must be traced to differences in its manufacture, processing or release. In the present study immunoreactive material was found to be distributed similarly in~a~ll~arly in young, ~~-~ and old brains. These findings wrrohorated a study by Rubin et at.“in q&g fcslnale rats in which immnn~yt~ist~ using antibodies which recognized different portions of the LRRH molecule gave no evidence for diflFerences in processing of the decapeptide in aging animals. Our observation of an increase in the occurrence of nucleoli may suggest elevated protein synthesis in the middle-age group, but further experimental evidence is required to test this. There was also an increase in the

frequency of nematosomes in LHRH neurons in the middie-aged sample. These cytoplasmic i&Ausions have been seen in LHRH neurons by other investigators, .16however, then function is unknown.” Synaptic input to hormone neurons

luteinizing

hormone-releadq

The lack of age-related changes in LHRH Ml number or size and sub&ular distribution of immunoreactivity led us to examine the sym@~ input to LHRH neurons. The preoptie %TB Is an important processing station for netam@ and ho~~al~y transmitted info~a~on. Projections to the region have been traced from higher eor&&, olfactory, diencephalic and brains&em reg$@~.” Hence there is potential for broad iofiuences, some of which may be directly synaptically transmitti, and some of which may be wordmated through interneurons. Electrical stirnAtion of the pre&#& region results in increased levels of LHRH in the hypophysial portal blood.‘” Considerabie &&SW suggests the importance of biogenic amines in

Aging LHRH neurons release (for reviews, see Refs 33 and 39 and for critical evaluation of experimentation, see Ref. 6). Juxtaposition of LHRH neurons with catecholaminergic and serotinergic projections has been reported at the light microscopic level. 13'18 In addition, an electron microscopic study has suggested that terminals which take up [3H]5-hydroxytryptamine make synaptic contact with LHRH-immunoreactive dendrites and neurons in rat preoptic area. 2~ In the present study two features of synaptology were studied: (1) the density of synaptic input as defined by the percentage of membrane apposed by a synaptic bouton with synaptic cleft or pre- or postsynaptic density; and (2) the proportions of various categories of synaptic boutons, classified according to their vesicle morphology. In both nonidentified and LHRH neurons there was less synaptic input to perikarya than to dendrites. (It should be noted that grouping all "nonidentified" neurons very likely masks differences among the subpopulations.) This has also been shown for

1009

neurons in the lateral hypothalamus35 and in the paraventricular nucleus.22 LHRH neurons and their processes were strikingly underinnervated compared to their neighbors, at least through middle age, receiving 1/5 as much synaptic input as nonidentified neurons in both young and middle-aged groups. Indeed, a thorough study of one LHRH neuron (in a young adult), employing 50 semiserial sections, revealed no synapses on its soma (unpublished observations). There was a range of between 0 and 6 synapses per neuronal profile. For LHRH neurons there were no profiles with more than 2 synapses in the young group, or more than 3 synapses in the middle-aged group. But in the old group, there was a substantial number of LHRH neuronal profiles with up to 6 synapses. In contrast, input to neurons (presumably neurosecretory) and their processes in the paraventricular nucleus revealed that approximately 6% of perikaryal membrane was postsynaptic (unpublished observations). This is about the same density of input

Nonid perika~a

LHRH perikarya

A

B

Young

Young

Middle

Old

t

0

10

20

30

40

50

60

70

80

90

0

100

10

20

30

40

50

60

Percent vesicle shape

Percent vesicle shape

LHRH dendrites

Nonid dendrites

70

80

90

100

D

C

Young

Young

~liddle

10

20

30

40

50

60

70

80

90

100

0

Percentvesicfe shape

10

20

30

40

50

60

70

80

Percentvesicle shape

Round Pteomorphic

Fig. 4. Bar graphs of proportions (expressed as percent) of synaptic boutons with clear round ("round") or pleomorphic ("pleomorphic") vesicles. (A) Synapses onto LHRH neuronal perikarya. (B) Synapses onto nonidentified perikarya. (C) Synapses onto LHRH dendrites. (D) Synapses onto nonidentified dendrites.

go

100

1010

J.W.

WITKIN

A

g .

Young

Young

OlO

10

20

30

40

50

60

70

80

90

100

O

10

20

30

Percent vesicle content

C

D.

Young

Young

Old

Old

10

20

30

40

50

60

40

50

60

70

80

70

80

Percent vesicle content

70

80

90

1go

0

10

20

Percent vesicle content

30

40

50

60

Percent vesicle content

Fig. 5. Bar graphs of proportions (expressed as percent) of synaptic boutons containing no dense core vesicles ("clear") or containing a few dense core vesicles along with clear vesicles ("dense"). (A) Synapses onto LHRH neuronal perikarya. (B) Synapses onto nonidentified perikarya. (C) Synapses onto LHRH dendrites. (D) Synapses onto nonidentified dendrites.

as to nonidentified dendrites in the preoptic region, but is three times as great as that to nonidentified perikarya in the preoptic area and 25 times as great as that to L H R H perikarya of young animals. These observations strongly suggest: (1) that the synaptic input to the L H R H neuron at the level of both the soma and the dendrite is slight compared to the neurosecretory cells in the paraventricular nucleus; (2) that though there may be influences onto the preoptic area from widespread regions, this input may be minimal or at least is effected through minimal synaptic transmission; and (3) it may be profitable to study the region of the median eminence for evidence of synaptic mediation. 34 There are species differences in the distribution of L H R H neurons and in their innervation. Whereas there is a considerable population of L H R H neurons in the medial basal hypothalamus in some mammals (guinea-pig, 46 sheep, 24 bats, 2° primates 1'2°'44'45'56) in mouse, 11 hamster 25 and rat, 57 L H R H cell bodies are limited to regions anterior to the medial basal

hypothalamus. There are also species differences in the relative synaptic input to L H R H neurons in the preoptic area. There is considerable input to these neurons in the guinea-pig 43'49 and in the sheep (Lehman, unpublished observations), but a sparse input in hamster 25 and rat. 58 These observations suggest that the modulation of L H R H manufacture and its release are mediated differently among species. Relation to aging The synaptic input to L H R H perikarya showed a three-fold increase in the middle-aged group and ten-fold in the old group, while that to nonidentified neurons and their processes did not change significantly with age. (The present study, which combined all nonidentified neurons as a single group, may have masked aging changes within subgroups of these neurons.) The fact that the size of L H R H nerons did not change with age lends further credence to the aging increase in synaptic density since this increase cannot be attributed to a loss in surface area. (It

1011

Aging LHRH neurons

might be added that there may well be differences within the LHRH population as sampled as it included neurons from periventricular, medial and lateral preoptic areas. However, LHRH neurons from all regions of the preoptic area have projections to the median eminence.‘7~47~‘s) At present we do not know whether or how this change in input to LHRH neurons is related to aging changes in the release of the decapeptide. It is likely that the changing steroid environment has an important effect on the synaptology of these neurons. Treatment with estradial benzoate has been found to increase synaptic density in the deafferented hypothalamic arcuate nucleus of young ovariectomized~ and aged” female rats. The relative number of synapses with electrondense core vesicles in young adult animals in the present study was in good agreement with a report by Kiss et al.22 on the paraventricular nucleus of the hypothalamus in which roughly l/3 of the boutons contained some dense core vesicles. We found a slight decrease with aging in the proportion of synapses containing dense core vesicles onto all dendrites and soma. Vesicles which are ellipsoidal or flattened have been postulated to be inhibitory on the basis of their distribution in the cerebellum” and on their neurotransmitter content. 31 The shift toward relatively more flattened vesicle-containing terminals would suggest that with aging there is relatively more inhibitory input to all neurons in the male rat preoptic area. Proportionately the shift toward “inhibitory” input is greater in the LHRH population. Although we have classified synapses on the basis of the shape of vesicles contained in the bouton, these observations must be interpreted with great caution. It has been amply demonstrated that vesicle shape is

subject to tixation procedures.” Furthermore, vesicles with similar size and shape may well carry different (or multiple) transmitters.~37 Another pitfall may be difTerences in susceptibility to fixation procedures between young and middle-aged animals. However, our material contained an internal control in that proportions of vesicle type in synapses on neurons and dendrites were differentially affected by aging (see particularly the consistent proportion of “round” to “pleomorphic” onto all dendrites as compared to the aging differences in their proportions onto neurons). As in a previous report,” we found rare evidence of LHRH-LHRH interaction. This is in contrast to a report by Leranth et ~1.2~using female rats in which colchicine was employed to maximize neuronal immunoreactivity. We must conclude from our observations that such contact is not a major factor in the coordination of pulsatile discharge of the decapeptide in male rats. Further study is required to determine whether an LHRH network in the preoptic area underlies the proestrus surge of the hormone. CONCLUSIONS This study has demonstrated some age-related changes in the synaptology of LHRH neurons in the male rat proptic area. The functional significance of these shifts in site and kind of input remains to be determined.

Acknowledgements-The author is grateful to Dr AnnJudith Silverman for critical reading of the manuscript, to Dr Dennis Kelly for statistical analysis of the data, and to Katharine Bock and Kafm Demasio for technical help. This study was supported by USPHS grants AGO5290 and AGO5366 (to the author) and HDlO665 (to Dr A.-J. Silverman).

REFEltF#NCES

1. Barry J. and Carette B. (1975) Immunofluorescence study of LRF neurons in primates. Cell Tips. Res. 164, 163-178. 2. Berk M. L. and Finkelstein J. A. (1981) AfTeremprojections to the preoptic area and the hypothalamic regions in the rat brain. Neuroscience 6, 1601-1624. 3. Bleier R., Cohn P. and Siggelkow I. R. (1979) A cytoarchitectonic atlas of the hypothalamus and hypothalamic third ventricle of the rat. In Hand&ookof& HypothafumuP,Vol. 1 (eds Morgane P. J. and Panksepp J.), pp. 137-220. Marcel Dekker, New York. 4. Bremner W. J., Karpas A. E., Matsumoto A. M., Steiner R. A., Clifton D. K. and Dorm D. M. (1983) Reproductive neuroendocrinology of the aging male. In Neuroendocrine Aspects of Reproduction(cd. Norman R. L.), pp. 379-392. Academic Press, New York. 5. Chiba T. and Murata Y. (1985) Afferent and efferent connections of the medial preoptic area in the rat: A WGA-HRP study. Brain Rer. Bull. 14, 261-272. 6. Coen C. W., Coombs M. C., Wilson P. M. J., Clement E. M. and MacKinnon P. C. B. (1983) Possible resolution of a paradox concerning the use of p-chlorophenylalanine and 5-hydroxytryptopham evidence for a mode of action involving adrenaline in manipulating the surge of luteinixing hormone in rats. Neuroscience g, 583-591. 7. Eskay R. L., Mica1 R. S. and Porter J. C. (1977) Relationship between luteinixing hormone releasing hormone concentration in hypophysial portal blood and luteinixing hormone release in intact, castrated, and electrochemicallystimulated rats. Enaiwinology 100, 263-270. 8. Ghanadian R., Lewis J. G. and Chisholm G. D. (1975) Serum testosterone and dihydrotestosterone changes with age in rat. Steroids Zs, 753-762. 9. Gibson M. J., Krieger D. T., Charlton H. M., Zimmerman E. A., Silverman A. J. and Perlow M. J. (1983) Mating and pregnancy can occur in genetically hypogonadal mice with preoptic area brain grafts. Science 225,949951. 10. Hoffman G. E. and Finch C. E. (1986) LHRH neurons in the female C57BL/6J mouse brain during reproductive aging: no loss up to middle age. Neurobtol. Agng 7,45&L 11. Hoffman G. E., Knigge K. M., Moynihan J. A., Melnyk V. and Arimura A. (1978) Neuronal fields containing luteinixing hormone releasing hormone (LHRH) in mouse brain. Neuroscience 3, 219-231.

1012

J. W.

WITKIS

12. Hoffman G. E. and Sladek J. R., Jr (1980) Age-related changes in dopamine, LHRH and somatostatin in the rat hypothalamus. Neurobiol. Agng 1, 27-37. 13. Hoffmann G. E. and Wray S. (1982) Relationship of catecholamines and LHRH: light microscopic study. Brain Rrs. Bull. 9, 417430. 14. Hsu H. K. and Peng M. T. (1978) Hypothalamic neuron number of old female rats. Geronrology 2.4, 436440. IS. Itoh K., Konishi A., Nomura S., Mizuno N., Nakamura Y. and Sugimoto 1. (1979) Application of coupled oxidation reaction to electron microscopic demonstration of horseradish peroxidase: cobalt-glucose oxidase method. Brain Res. 175, 34-346.

16. Jennes L., Stumpf W. E. and Sheedy M. E. (1985) Ultrastructural characterization of gonadotropin-releasing hormone (GnRH)-producing neurons. J. camp. Neural. 232, 534547. 17. Jennes L. and Stumpf W. E. (1986) Gonadotropin-releasing hormone immunoreactive neurons with access to fenestrated capillaries in mouse brain. Neuroscience 18, 403416. 18. Jennes L., Beckman W. C., Stumpf W. E. and Grzanna R. (1982) Anatomical relationships of serotoninergic and noradrenalinergic projections with the GnRH system in septum and hypothalamus. Expl Brain Res. 46, 331-338. 19. Karpas A. E., Bremner W. J., Clifton D. K., Steiner R. A. and Dorm D. M. (1983) Diminished luteinizing hormone pulse. frequency and amplitude with aging in the male rat. Endocrinology 112, 788-792. in humans and other mammals: snecies 20. Kina J. C. and Anthony E. L. (1984) LHRH neurons and their proiections . _ comparisons. Peptides g, 195-207. 21. Kiss J. and Halasz B. (1985) Demonstration of serotoninergic axons terminating on luteinixing hormone-releasing hormone neurons in the preoptic area of the rat using a combination of immunocytochemistry and high resolution autoradiography. Neuroscience 14, 69-78. 22. Kiss J. Z., Palkovits M., Zaborsxky L., Tribollet E., S&o D. and Makara G. B. (1983) Quantitative histological studies on the hypothalamic paraventrictdar nucleus in rats: I. Number of cells and synaptic boutons. Bruin Res. 262.217-224. 23. Langford L. A. and Coggeshall R. E. (1980) The use of potassium ferricyanide in neural fixation. Anar. Rec. 19, 297-303.

24. Lehman M. L., Robinson J. E., Karsch F. J. and Silverman A.-J. (1986) lmmunocytochemical locahxation of luteinizing hormone-releasing hormone (LHRH) pathways in the sheep brain during anestnts and mid-luteal phase of the estrow cycle. 1. camp. Neural. 244, 1%35. hormone (LHRH) neurons 25. Lehman M. N. and Silverman A.-J. (1985) Ultrastructure of luteinixing hormonorekasing in intact and castrated male hamsters. Sot. Neurosci. 11, 351. 26. Leranth Cs., Segura L. M. G., Palkovits M., MacLusky N. J., Shanabrough M. and NaBolin F. (1965) The LH-RHcontaining neuronal network in the preoptic atea of the rat: demonstration of LH-RH-containing nerve terminals in synaptic contact with LH-RH neurons. Brain Res. 345, 332-336. 27. Machado-Salas J., Scheibel M. E. and Scheibel A. B. (1977) Morphologic changes in the hypothalamus of the old mouse. Expl Neural. 57, 102-I I I. 28. Matsumoto A., Arai Y. and Gsanai M. (1981) Neuronal plasticity in the deaIIetented hypothahunk arcuate nucleus of adult female rats and its enhancement by treatment with estrogen. J. camp. Neural. 197, 197-205. 29. Matusmoto A., Arai Y. and Osanai M. (1985) Estrogen stimulates neuronal plasticity in the deafferented hypothalamic arcuate nucleus in aged female rats. Neurosci. Res. 2, 412-418. 30. Matsumoto A., Okada R. and Arai Y. (1982) Synaptic changes in the hypothalamic arcuate nucleus of old male rats. Expl Neural. 78, 583-590. 31. Matus A. I. and Dennison M. E. (1971) Autoradiographic localisation of tritiated glycine at “flat-vesicle” synapses in spinal cord. Bruin Res. 32, 195197. 32. Merchenthaler I., Lengvari I., Horvath J. and Setalo G. (1980) Immunohistochemical study of the LHRH-synthesizing neuron system of aged female rats. Cell Tin. Res. 209, 499-503. 33. Negro-Vilar A. and DePatdo L. (1983) Monoaminergic and steroidal regulation of luteinixing hormone-releasing hormone release. In Male Reoroduction and Fertilitv (ad. Netuo-Vilar A.). DD. 13-25. Raven Press. New York. 34. Palkovits M., Leranth C., Jew’J. Y. and Williams T.-H. (1982~Simultaneo’us~~haracterization of pre- and postsynaptic neuron contact sites in brain. Proc. natn. Acad. Sci., U.S.A. f9, 2705-2708. 35. Palkovits M. and van Cut H. (1980) Quantitative tight and electron m icromopic studies on the lateral hypothalamus in rat. Cell and synaptic densities. Brain Res. BUN. 5, 643647. 36. Pelietier G. (1983) Identification of endings containing dopamine and vasopresain in the rat posterior pituitary by a combination of radioautography and immunocyto&emistry at the ultrastntctural kvel. J. Hisrochem. Cytochem. 312, 562-564.

37. Pelletier G.. Steinbusch H. W. M. and Verhofatad A. A. J. (1981) Immunomaotive substance P and aerotonin present in the same dense-co re vesicles. Norlot 1)3. 71-72. 38. Peters A., Palay S. L. and Webster H. deF. (1976) The Fine Structure of the Nerooau System. W. B. Saunders, Philadelphia. 39. Ramirez V. D., Feder H. H. and Sawyer C. H. (1984) The role of brain catechoiamines in the mgtdation of LH secretion: a critical inquiry. In Frontiers in Neuroendwinology, Vol. 8 (eds Martini L. and Ganong W. F.), pp. 27-84. Raven __ __ _ Press, New York.

40 Rubin B. S., King J. C. and Bridges R. S. (1984) Immunorcactive forms of iuteiadxing bormone-rekasing hormone in the brains of aging rata exhibiting per&tent va&ual *rtrus. &I. Reprod 31, 343-35 I. 41. Saksena S. K. and Lau I. F. (1979) Vuidon~ in aannn androgolre, eatrogaaa, pro@ina, gonadotropins and prolactin levels in male rats from prepubertal to advanced age. Expl Agog Res. 5, 179-194. 42. Sherwood N. M., Chiappa S. A. and Fink G. (1976) Innnunomaetive lutekdxing hormone rekaaing factor in pituitary stalk blood from female rats: sex steroid modulation of response to electrical stimulation of pnaptic ama or msdiart eminence. J. Endocr. 78, 501-511. hormone containing aynas=e in the diagonal band and pmoptic 43. Silverman A.-J. (1984) Luteinixinn borm~ area of the guinea pig. J. camp. -Neural. 227, 452-458. 44. Silverman A.-J., Antunea J. t., Abrarm G. hf., Nilaver G., Tbau. R., Robinron J. A., Ferin M. and Krey L. C. (1982) The luteinixing hormone-mkaaing hormone pathways in rhesus (“Aleeoea mukma”) and pi@ibd (“Maraca nemesrrina”) monkeys: new observations on thick, unembcdded a&ions. J. camp. Neural. 211, X%317.

Aging LHRH neurons

1013

45. Silverman A.-J., Antunes J. L., Fern M. and Zimmerman

46. 41. 48. 49. so. 51. 52. 53. 54.

E. A. (1977) The distribution of luteinizing hormonereleasing hormone (LHRH) in the hypothalamus of the rhesus monkey. Light microscopic studies using immunoperoxidase technique. En&crinology 101, 135-142. Silverman A.-J. and Kmy L. C. (1978) The luteinizing hormonMeleasing hormone (LH-RI-I) neuronal networks of the guinea pig brain. I. Intra- and extra-hypothalamic projections. Brain Res. 157, 233246. Silverman A.-J. and Renaud L. (1985) Which LHRH neurons project to the median eminence? A combined retrograde tracing and immtmocytochemical analysis. Sot. Neurosci. 11, 10.7. Silverman A.-J. and Renaud L. P. (In press) Localization of luteinizing hormone-releasing hormone (LHRH) neurons that project to the median eminence. J. Neurosci. Silverman A.-J. and Witkin J. W. (1985) Synaptic interactions of luteinizing hormone-releasing hormone (LHRH) neurons in guinea pig preoptic area. J. Hktochem. Cytochem. 33, 69-72. Simpkins J. W., Estes K. S., Kalra P. S. and Kalra S. P. (1983) Alterations in hypothalamic neurotransmitters contribute to age-related decline in reproductive function in the male rat. In Mule Reproduction and Fertiky (ed. Negro-Vilar A.), pp. 91-109. Raven Press, New York. Simpkins J. W., Kalra P. S. and Kalra S. P. (1981) Alterations in daily rhythms of testosterone and progesterone in old male rats. Expl Agng Res. 7, 25-32. Sladek J. R., Jr, Khachaturian H., Hoffman G. E. and Scholer J. (1980) Aging of central endocrine neurons and their aminergic aBerents. Peptides 1, 141-157. Uchizono K. (1965) Characteristics of excitatory and inhibitory synapses in the central nervous system of the cat. Nuture 207,642. Valdivia 0. (1965) Characteristics of excitatory and inhibitory synapses in the central nervous system of the cat.

Nature 207, 642-643. 5s. Venable J. H. and Coggeshall R. A. (1965) A simplified lead citrate stain for use in electron microscopy. J. Cell Biol. 2!5,4OS-408. hormone (LHRH) neurons in aging female rhesus macaques. 56. Witkin J. W. (1986) Luteinizing hormone-releasing Neurobiol. Agng 7, 259-263. 57. Witkin J. W., Paden C. M. and Silverman A. J. (1982) The luteinizing hormone-releasing hormone (LHRH) systems in the rat brain. Neuroendocrinology 35, 429-438. hormone neurons in rat preoptic 58. Witkin J. W. and Silverman A. J. (1985) Synaptology of luteinizing hormonereleasing area. Peptides 6, 263-271. (Accepted 20 December

1986)