Previous experiments on ferrets which were designed to determine the way in

425 J. Physiol. (I95I) II3, 425-433 RELATION OF RETINAL STIMULATION TO OESTRUS IN THE FERRET BY A. P. D. THOMSON From the Department of Anatomy, University of Birmingham (Received 31 July 1950) Previous experiments on ferrets which were designed to determine the way in which light stimulation of the retina becomes translated into stimulation of the pituitary, with the consequent release of gonadotrophic hormone, have led to equivocal results. On the assumption that the stimulus is conveyed along the optic nerve, Clark, McKeown & Zuckerman (1939) attempted to find out whether or not anoestrous ferrets whose visual pathways had been interrupted at different levels came into heat under the influence of additional daily illumination. The experiments showed that animals whose optic nerves had been -divided either did not come into heat at all, or came into heat much later than control animals. They showed, in addition, that the gonadal response could also occur: (a) in the absence of the superior colliculi; (b) when all retinal impulses to the dorsal nucleus of the lateral geniculate body and the visual cortex had been completely interrupted. Unfortunately, attempts to divide the optic tract at the level of the ventral border of the lateral geniculate body were only partially successful, and in each of two experiments some cells of the dorsal nucleus remained connected with the optic tract of one side. It was argued, however, that unless a few isolated normal cells in the dorsal nucleus of the lateral geniculate body were sufficient for adequate cortical stimulation, the experiments showed that a normal oestrous response could occur in the absence of the visual cortex and the superior colliculi, and even after interruption of retinal impulses passing to the dorsal nucleus of the lateral geniculate body, visual cortex, superior colliculi, and pretectal area. This conclusion suggested that the visual response depended on impulses passing either to the ventral nucleus of the lateral geniculate body or to the subthalamus by way of accessory tracts. The brains of the animals used in these experiments were later studied by Jefferson (1940). He failed to demonstrate either the existence of accessory optic tracts or of any other connexion between the main optic pathway and the hypothalamus, and consequently concluded 'that there may be no special

426 A. P. D. THOMSON pathway mediating this response, and that the latter may be an indirect response to the changes in the total activity of the animal'. In spite of this negative conclusion, it remains a fact that there is some functional connexion between the eye and the pituitary in the ferret. As a first step, it was decided to see whether the retinal stimulation which produces light-induced oestrus is conveyed along the optic nerve fibres themselves, or by way of either the retinal blood vessels or the long and short ciliary nerves. In all previous experiments in which the optic nerves were divided, the operation involved all these structures. MATERIAL AND METHODS The animals were a mixed population of albino and polecat ferrets obtained from dealers. They were housed singly in cages x 11 x 14 in., which were arranged in four tiers around three sides of a room 9 x 15 ft. The room was divided into two parts by a heavy light-proof canvas partition. One part of the room was exposed to normal day-lighting conditions through a glass door and fanlight (about 28 sq.ft. of glass) and accommodated forty-eight ferrets. The other part of the room had a single window (about thirty-six sq.ft.) and held sixty animals. In this part of the room a 40 W. electric bulb was suspended over the centre of each cage, 15 in. above its floor. Both normal and operated animals were distributed at random on racks in the'illuminated' and 'non-illuminated' halves of the room. Each animal received about lb. fresh horse meat and pint of milk daily. The ferrets were examined approximately every 5 days, and the following points recorded: weight and general condition; degree of distension of the vulva; and the size of the pupils. The animals were tested at the same time for form vision. Operative procedures. Operations described by Wagenmann (1890) for the study of effects of cutting the ciliary arteries of rabbits, and by McCrea, Eadie & Morgan (1942) for cutting the optic nerve fibres of dogs without causing appreciable damage to the autonomic nerve fibres in the optic nerve sheath, were modified for use in the present experiment. In the ferret the long posterior ciliary arteries course along a plane which is continuous with the long axis of the elliptical pupil, and their position can readily be identified. The central artery to the retina takes up its intraneural course about1 mm. from the cribriform plate. Operations were carried out under pentobarbitone anaesthesia. The hair in the region of the outer canthi was removed and the animal placed on its side' with its nose towards the operator. The outer canthus was slit (Fig. 1, 1) and the edge of the wound held apart in a frame retractor (Fig. 1, 2). The bulbar conjectiva overlying the external long posterior ciliary artery was picked up and cut parallel to the corneo-sclerotic margin (Fig. 1, 3). The exposed insertion of the lateral rectus muscle was also divided, allowing the eyeball to be rotated nasally. The optic nerve was then brought into view by pushing a small pledget of cotton-wool between the nerve and the orbital ligament (Figs. 1, 4 and 2, 6). A small curved hook was inserted into the sheath of the optic nerve (Fig. 2, 5) and passed round the optic nerve fibres, which were then divided within the hook by means of a small knife (Fig. 2, 8). Alternatively, the sheath of the nerve could be crushed in special forceps placed around the hook (Fig. 2, 7) or the animal could be left as a control-operated animal. In those animals whose retinal circulation was interrupted, the sheath and its contents were crushed for two periods of 2 min. separated by an interval of 2 min. Complete division of the entire sheath and all the fibres of the optic nerves was effected in the usual way with scissors. All operations were carried out during July, August and September 1947. The animals were separated into a control and an experimental group, each containing: (a) animals in which only the optic nerve fibres of both sides were divided, and in which the central artery to the retina and the long and the short posterior ciliary vessels were not disturbed; (b) animals in which the optic nerve fibres themselves were not damaged and in which the other

FERRET RETINA AND OESTRUS 427 structures named above were crushed; (c) animals in which both sets of connexions were divided; (d) ferrets with control operations; and (e) unoperated normal animals. The 'control' group was placed in the half of the experimental room which was illuminated by daylight only. The 'experimental' group was arranged in cages which, as already observed, were illuminated by means of an electric bulb. Beginning on 22 October 1947, and until about the middle of April 1948 (a period during which the normal female ferret is in anoestrus in this country), the experimental group of animals was exposed to extra artificial illumination for 6 hr. daily (4.30-10.30 p.m.). The lights were controlled by a Venner time switch. Collar of forceps must fit accurately around hook Cutting edge Each half of /1In sni. enough collar soldered 1 mm. eocut insmde hook to forceps Special Instruments,tissue Knilfe (or scissors) cutting and lateral rectus m. exposing ext. long p $Illary a. Fig. 1. Operative technique, 1-4. Histology. Sections of the eye and optic nerves were prepared to determine the effects of the various operations on these structures. Sections of the eye in which the retina, choroid, and sclera remained in their normal anatomical relationship were difficult to prepare. The following method was most suitable. The freshly excised eye was fixed in warm Helly's fluid for 24 hr. It was then washed in running water overnight and dehydrated in alcohol. The cornea and lens were then removed. After paraffin embedding, sections were cut at 10,1. and stained with Weigert's haematoxylin. Optic nerves from normal and operated animals were treated with osmic acid to stain myelin and with Bodian's technique to demonstrate nerve fibres.

428 A. P. D. THOMSON Electroretinogram8. These were recorded from normal and operated animals to assess the functional state of the retina. On the evening before each recording, 1/200 gr. atropine sulphate was instilled into the eye. Immediately before the electroretinogram (e.r.g.) was taken, the animal was light-adapted for at least half an hour in a dark room which was artificially illuminated. The animal was anaesthetized with intraperitoneal pentobarbitone and the hair over the upper eyelid removed. One drop of 2 % cocaine hydrochloride was instilled into the eye before a wire speculum was inserted between the eyelids. When the pentobarbitone anaesthesia was satisfactory, usually after 10-15 min., small cotton-wool electrodes were positioned on the cornea and eyelid, and e.r.g.'s were obtained to a variety of stimuli, viz. from the lights of the room, operated by the wall switch; OritalimentA6 Cotton-wool /~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~" gpost. ~E cliafa ofhook Hook in position inside sheath and around optic n. fibres 7 Forceps In position ready to crush sheath and Its contents Knife cutting pptic n. fibres Fig. 2. Operative technique (schematic) 5-8. from a brief flash of white light lasting about -2-sec.; and from red flashes of similar duration through Wratten filters nos. 70 and 29. The no. 70 filter neutralizes any response from rod elements, and the no. 29 filter acts as a neutral filter cutting down the intensity of the flash. Both eyes of about forty-five animals were examined, many of them more than once, and at varying intervals after the operations. The e.r.g.'s were observed on an oscilloscope and recorded on a pen-recorder. Ophthalmoscopic examination of the eyes. Normal and operated animals were examined with an ophthalmoscope at intervals of about a month. Tests for form Vi8ion. A normal ferret in a cage will closely follow a moving hand outside the cage by moving its head. If placed on top of one of a pair of adjacent cages it will soon bridge the gap

FERRET RETINA AND OESTRUS 429 between them even if this means jumping. If the upper part of the bottomless cage of a ferret is suddenly and silently removed from its tray a normal animal will soon wander off. Animals judged blind behave in a very different way. They do not respond to hand movements outside the cage, nor do they appear to be aware of an approaching hand. They do not succeed in bridging the gap between two cages but either remain on top of one of them or climb down off it backwards. Sudden removal of the upper part of its cage will sometimes find a blind animal remaining motionless in its tray. At other times the animal will lift its front legs as though expecting to feel the walls of the cage. RESULTS The animals whose optic nerve fibres had been divided, and whose optic nerve sheaths had been crushed, behaved differently from normal unoperated or control-operated animals. TABLE 1. Ferrets in oestrus and anoestrus after 16 weeks of added illuimination, 1947-8. (Yatess correction for continuity in small samples was used in each x2 test) Operation group (1) Animals with optic nerve fibres only divided Animals with sheath of optic nerves crushed Animals with complete optic nerves divided Sub-total for operated groups Animals with control operation on eyes Normal ferrets Sub-total for control groups Animals exposed to daylight only No. in No. in oestrus anoestrus (2) (3) 1 5 Animals exposed to normal daylight plus artificial light No. in oestrus (4) 0 No. in anoestrus (5) 12 x2 comparing the oestrous state of the animals in 0 8 2 10 0-208 0 3 0 5 1 16 2 27 0-234 0 5 4 0 5-406 0 30 30 6 37-73 0 35 34 6 46-24 each group in Iboth parts of'the ferret Degrees of room freedom (6) (7) 1-132 1 p (8) 0-8-0-7 1 0-7-0-5 1 0-7-0-5 1 0-02 1 < 0-001 1 <0-001 Table 1 shows the number of animals in oestrus and anoestrus 112 days after the beginning of the experiment in the control (columns 2 and 3), and the experimental halves of the room (columns 4 and 5). The proportion of blinded ferrets in oestrus (i.e. animals in the following operated groups: complete section of the optic nerves; division of the optic nerve fibres alone; and the interruption of the components of the optic nerve sheaths) was the same in two halves of the room, a x2 test revealing no statistical difference. On the other hand, there were significantly more normal and control-operated animals in oestrus in the illuminated half of the room than in the other half of the room. In the whole experiment there were forty-six ferrets, some or all of whose

430 A. P. D. THOMSON optic nerve components had been interrupted. Forty-three of them were not in heat after 112 days. Thirty-two of these came into heat at varying times between 120 and 200 days. The remaining animals were sacrificed between 180 and 250 days after the beginning of the experiment, while still anoestrous. The times at which the operated animals (other than control-operated) came into heat did not vary in any systematic way according to the character of the operation. Correspondingly, the animals which did not show any signs of heat after 200 days included ferrets whose optic nerves were divided, ferrets whose optic nerve fibres had been sectioned, and ferrets whose optic nerve sheaths had been crushed. Analysis also showed that the distance of the cage from the window or door of the room bore no relation to the speed with which normal animals illuminated and non-illuminated, came into oestrus. These observations show that there are no differences in the effects of complete nerve section, division of the optic nerve fibres alone, and crushing ofthe sheaths of the optic nerves, so far as the gonadal response of ferrets to additional light is concerned. Two animals in the group of ferrets whose optic nerve sheaths were crushed and which were exposed to added light came into heat at the same time as did normal animals given additional illumination. One of these animals had form vision-as judged by the tests described. The third operated animal which came into heat within 112 days from the start of the experiment had had its optic nerve fibres sectioned. This animal was exposed to normal daylight only; its response to visual tests suggested that it was blind. It was in fact the only animal in heat after 112 days in the non-illuminated part of the room. This general picture of uniformity in the response of the operated animals to added light contrasts with the following differences in the appearance of the eyes: the histological picture of the retina; and the e.r.g. responses. Ophthalmoscopic examination. The chief conditions attributable to the operations were: (a) fixation of the pupils in dilatation, (b) the formation of cataract, and (c) pallor of the optic disks. It is known that oestrus may not occur or may occur at irregular intervals in ferrets which spontaneously develop bilateral cataract (Bissonnette, 1935; and others). In the present experiment cataract developed in both eyes, significantly more often (as shown by x2 tests) in the group of animals in which all components of the optic nerve trunk were divided, than in those in which the optic nerve fibres alone, or the components of the optic nerve sheath alone, had been interrupted. Nevertheless, oestrus occurred no more frequently in these two groups than in the first. This observation suggests that bilateral cataract per se does not play a major part in the arrest of the sexual cycle. Similarly, the retinal blood vessels in the first group were usually thinner and more scanty than in the other operated groups. It may be observed that in

FERRET RETINA AND OESTRUS 431 general no change was observed in the fundal blood vessels of animals whose optic nerve fibres alone, or whose optic nerve sheaths, had been interrupted. Histology of the retina. The only reference to the histology of the ferret eye found in the literature is a short report by Rochon-Duvigneaud (1943), who states that the Mustelidae in general, and the ferret in particular, have a retina which is rich in cones and ganglion cells, '...bref un ceil de qualites visuelle et accommodative bien superieures'. In the present experiment it was found that normal ferrets may suffer from a congenital retinal condition characterized by thinning and disappearance of the outer retinal layers (Tansley, 1948). The retinae of the animals whose optic nerve fibres only had been cut were practically normal; only the ganglion cell layer appeared to have degenerated. Animals whose optic nerves were completely divided in general showed the greatest amount of damage; in some- of them large areas of all layers of the retina had completely degenerated. These eyes were usually associated with cataract. The retinae of the animals with crushed sheaths tended to show an intermediate amount of damage. The one change that was common to all these groups was degeneration of the ganglioi cell layer. Electraretinograms. The responses varied somewhat among normal eyes and even in the same eye from day to day, but e.r.g.'s from operated animals recorded 7 months after operation were hardly distinguishable from normal. This was especially so in the animals whose optic nerve fibres alone had been sectioned and in the animals whose optic nerve sheaths had been crushed. Altered responses were obtained from some animals whose optic nerves were completely severed. The presence of a cataract invariably prevented an e.r.g. from being obtained. No e.r.g.'s were obtained from animals whose rods and cones had degenerated. Testsforform vision. All the animals whose optic nerves had been completely divided or whose optic nerve fibres alone had been sectioned were judged blind on the basis of the tests used for form vision. With the exception of one animal, the group in which the optic nerve sheaths had been crushed was also judged blind. This animal was one of the two in this group which came into early oestrus during the winter illumination. DISCUSSION Earlier observations on the stimulation effect of artificial illumination on gonadal function in the normal anoestrous female ferret indicate that the primary receptor is the eye. The present experiments show that any one or a combination of a number of ocular defects may be associated with an absence of the response. Neither ferrets with their optic nerve fibres sectioned, nor with the blooct supply to the retina interrupted, came into heat precociously because of added PH. CXIII. 28

432 A. P. D. THOMSON illumination. It would appear, therefore, that the experiment failed to differentiate between the possible role of the optic nerve fibres and the retinal blood supply in the response of the ferret to light-induced oestrus. But as far as the blood supply to the retina was concerned there was no difference in these two groups of animals as indicated by the ophthalmoscope, the histological picture and the injection of the blood vessels with indian ink at death. The significant common feature was degeneration of the ganglion cell layer. Presumably the disappearance of this layer was responsible for the absence of the gonadal response. In so far as the e.r.g. is a measure of physiological function, it is clear from the results that the gonadal response is dependent on some event occurring in the neurological chain of visual processes subsequent to the local response of the retina to light stimulation. It is of interest that the retina can retain some physiological activity, as indicated by the e.r.g., when the ganglion cell layer has degenerated and the remainder of the retina appears healthy. A related observation has been made by Campos (1936), who sectioned the optic nerve in rabbits, taking care that the central vessels of the retina and the ciliary arteries were left undamaged, and then measured the respiration of the retina by the Warburg technique at intervals up to 158 days after the operation. He found that the metabolism of the retina did not alter even when the optic nerve fibres and the ganglion cells were completely degenerated. On the other hand, Santoni (1939) reported that the retina consumed more oxygen if its circulation is interrupted for less than 30 min., and less oxygen if the interruption is continued for longer. He attributes the increased consumption to the liberation of substances such as fatty acids which are said to stimulate oxygen utilization. The fall in respiration following prolonged interruption to the retinal circulation was attributed to the progressive fall in the vitality of the retinal elements. Return of the retinal circulation, as seen by the ophthalmoscope and by the injection of indian ink at death, occurred in all animals whose retinal circulation had been interrupted. The same observation was made by Sherrington (1886), who tied a ligature as tightly as possible around the optic nerve of a rabbit, and 9 weeks later reported that the retinal circulation had become re-established, although the animal appeared to be blind on that side. Blindness and degeneration of the optic nerve also occur clinically after a massive haemorrhage (Levatin, 1947). Since the integrity of the ganglion layer appears to be necessary for the occurrence of the gonadal response, it is safe to infer that the initial mechanism in the transference of light stimuli from the eye to the ferret hypophysis is a nervous one.

FERRET RETINA AND OESTRUS 433 SUMMARY 1. On the assumption that light-induced oestrus in the female ferret operates through the eye, the present experiment was designed to determine which component of the optic nerve was concerned in the response. 2. Four groups of animals were used: (a) animals in which only the optic nerve fibres were divided, and in which the central artery to the retina, and the long and short ciliary vessels and nerves, were left intact; (b) animals in which the optic nerve fibres were not disturbed, and all the other structures named above were crushed; (c) animals in which both sets of connexions were divided, i.e. division of the complete optic nerve; and (d) controls. In this last group were control-operated and normal unoperated animals. 3. The animals were separated into a 'control' and an 'experimental' group each containing animals from (a), (b), (c) and (d). The control group was placed in half of an experimental room which was illuminated by daylight only, while the experimental group was placed in the other part of the room, which, in addition to normal daylight, was artificially illuminated for 6 hr. daily, from October 1947 until April 1948. 4. The experiment showed that the gonadal response to added light appeared to depend on the integrity of the ganglion cell layer of the retina and that the initial mechanism in the transference of light stimuli from the eye to the ferret hypophysis is a nervous one. It is a great pleasure to thank Prof. S. Zuckerman, F.R.S. for his sustained interest and helpfulness. My thanks are also due to Mr W. J. Pardoe, of this Department, for making the special instruments used in the operations; to Mr H. M. F. Asher, B.A., of the Physiology Department for his skill in obtaining electroretinograms; and to Dr K. Tansley, D.Sc. (Institute of Ophthalmology, Judd Street, London), for much interpretation of retinal histology. REFERENCES Bissonnette, T. H. (1935). Anat. Rec. 63, 159. Campos, R. (1936). Ann. Ottalm. 64, 577. Clark, W. E. Le Gros, McKeown, T. & Zuckerman, S. (1939). Proc. Roy. Soc. B, 126, 449. Jefferson, J. M. (1940). J. Anat., Lond., 75, 106. Levatin, P. (1947). Arch. Ophthal., N.Y., 37, 18. McCrea, F. D., Eadie, G. S. & Morgan, J. E. (1942). J. Pharmacol. 74, 239. Rochon-Duvigneaud, A. (1943). Les, Yeux et la Vi.sion des Verte&bri8, 1st ed., p. 669. Paris: Masson. Santoni, A. (1939). Ann. Otm. 67, 299. Sherrington, C. S. (1886). J. Physiol. 7, xvi-xvii. Tansley, K. (1948). Personal communication. WVagenmann, A. (1890). v. Graefes Arch. Ophthal. 38, 1. 28-2