Different neural melatonin-target tissues are critical for encoding and retrieving day length information in Siberian hamsters.

Siberian hamsters exhibit several seasonal rhythms in physiology and behaviour, including reproduction, energy balance, body mass, and pelage colouration. Unambiguous long- and short day lengths stimulate and inhibit reproduction, respectively. Whether gonadal growth or regression occurs in an intermediate day length (e.g. 14 h L : 10 h D; 14L), depends on whether the antecedent day lengths were shorter (10L) or longer (16L). Variations in day length are encoded by the duration of nocturnal pineal melatonin secretion, which is decoded at several neural melatonin target tissues to control testicular structure and function. We assessed participation of three such structures in the acquisition and retrieval of day length information. Elimination of melatonin signalling to the nucleus reuniens (NRe), but not to the suprachiasmatic nucleus (SCN) or paraventricular thalamus (PVt), interfered with the acquisition of a long day reproductive response, whereas the obscuring of melatonin signals to the SCN and the NRe, but not to the PVt, interfered with the photoperiod history response. The SCN and NRe contribute in different ways to the melatonin-based system that mediates seasonal rhythms in male reproduction.

The low-density lipoprotein pathway of cultured Leydig tumor cells. Utilization of low-density lipoprotein-derived cholesterol for steroidogenesis.

We have previously reported that low-density lipoprotein (LDL) enhances and prolongs steroidogenesis in human choriogonadotropin (CG)-stimulated Leydig tumor cells (MA-10). The studies described herein elucidate the mechanisms by which LDL increases human CG stimulated steroidogenesis. Our results show that the MA-10 cells express the classic LDL pathway. LDL is bound to specific surface binding sites which are regulated by the level of intracellular cholesterol. The cellular processing of bound LDL is temperature-dependent and is inhibited by blocking lysosomal function. By using an LDL derivative in which the core cholesteryl esters have been replaced with [3H]cholesteryl linoleate, we show that LDL cholesterol is rapidly utilized for steroid hormone synthesis. The utilization of LDL cholesterol quantitatively accounts for the LDL-induced augmentation of steroidogenesis. We also show that the addition of LDL to human CG-stimulated MA-10 cells maintains cellular free and esterified cholesterol levels and increases progesterone biosynthesis. The addition of LDL does not, however, affect the cellular utilization of preexisting cholesterol stores for steroidogenesis.

Perinatal influences of melatonin on testicular development and photoperiodic memory in Siberian hamsters.

We assessed the influence of perinatal melatonin on reproductive development and adult responsiveness to melatonin. Testicular growth in an intermediate day length (14 : 10 h light/dark cycle) was substantially reduced in Siberian hamsters gestated by pinealectomised compared to pineal-intact females; gonadal development was normalised in offspring of pinealectomised dams that were pinealectomised at 3-4 days of age. Hamsters deprived of melatonin only during gestation, or both pre- and postnatally, underwent testicular involution during treatment with melatonin in adulthood. Photoperiodic histories acquired prenatally did not endure as long as those acquired by adult hamsters. Hamsters first exposed to melatonin in adulthood were not more proficient in acquiring photoperiodic histories than were normal males. These findings indicate that pre- versus postnatal differences in melatonin signal duration determine rates of testicular development. Exposure to melatonin perinatally does not appear to organise the neuroendocrine substrate that mediates effects of day length and melatonin on the gonads of adult hamsters.

Evidence that the circadian system mediates photoperiodic nonresponsiveness in Siberian hamsters: the effect of running wheel access on photoperiodic responsiveness.

Juvenile male Siberian hamsters from a line of hamsters selected for nonresponsiveness to short photoperiod (PNRj) and animals from the general colony (UNS) were separated at weaning into two groups. Group 1 males were moved into short days (10 h light:14 h dark [10L:14D]) with free access to running wheels (RW). Group 2 animals were the male siblings of Group 1 hamsters; they were moved at the same time into the same room, but were housed in cages without access to RW. Group 2 hamsters only had access to RW for the final week of short-day exposure (Week 8). Animals were blood sampled at the time of sacrifice for analysis of serum prolactin (PRL) and follicle-stimulating hormone (FSH) concentrations. At sacrifice, paired testis weights were obtained and pelage color was scored. Animals from the UNS line showed the expected declines in testis weight, body weight, and serum concentrations of both PRL and FSH, regardless of the presence or absence of RW. These animals also exhibited a high proportion of individuals molting to winter-type pelage. By contrast, a marked difference was noted between siblings from the PNRj line depending on whether RW access was provided at the time of weaning. Animals with access to RW exhibited identical responses to those of the UNS responder animals, whereas PNRj animals without access to RW showed no adjustments to short days (i.e., testis regression, pelage molt, expansion of alpha). In a second experiment, PNRj and UNS males were placed in constant darkness (DD), with or without RW access. The results of this experiment indicated that PNRj animals respond to DD regardless of the presence or absence of RW. In DD, PNRj hamsters also exhibited significantly longer free-running period lengths (taus) than did UNS hamsters; all the PNRj hamsters had taus > 24 h, whereas none of the UNS hamsters had a tau > 24 h. These results indicate that PNRj hamsters retain the proper neural pathways for responding to short day lengths and establish a role for locomotor activity feedback in modulating the circadian system and, subsequently, photoperiodic responsiveness in PNRj hamsters.

Photoperiod nonresponsive Siberian hamsters: effect of age on the probability of nonresponsiveness.

Groups from three different breeding lines of Siberian hamsters (UNS = general colony animals, PNRa = selected for photoperiod nonresponsiveness as adults, PNRj = selected for photoperiod nonresponsiveness as juveniles) were exposed to short days at weaning and again as adults (Experiment 1) or only as adults (Experiment 2). The proportion of photoperiod nonresponsive individuals in each line was determined by measuring testis length after 6 weeks of exposure to short days (juveniles) or by paired testis weights after 12 weeks in short photoperiod (adults). Adults were blood sampled on the day of sacrifice (Experiment 1) or on Weeks 3, 4, 5, 6, 8, 10, and 12 (Experiment 2) for determination of serum prolactin (PRL) and follicle-stimulating hormone (FSH) concentrations. Nonresponsive individuals were present in all three lines of hamsters. Furthermore, all three lines of hamsters showed an increase in the proportion of nonresponders with age; some individuals are responsive to short days as juveniles, but become nonresponsive in adulthood. The two PNR lines exhibited a greater proportion of nonresponders at both ages compared to the UNS line, with the PNRj line exhibiting the greatest proportion of nonresponders at each age. During exposure to short days, nonresponders exhibited significantly higher serum PRL and FSH concentrations that did the UNS line; nonresponders also exhibited larger testis size, and fewer animals molted to winter-type pelage. The results indicate that (a) in all three lines, a significantly higher proportion of animals are nonresponsive to short photoperiod as adults than as juveniles; (b) selection for nonresponsiveness as juveniles can produce a line of hamsters that, as adults, are nearly all nonresponsive to short days; and (c) some individuals from each line are responsive to short photoperiod early in life, but become nonresponsive as adults.

Temperature-independence of circannual variations in circadian rhythms of golden-mantled ground squirrels.

In golden-mantled ground squirrels, phase angles of entrainment of circadian locomotor activity to a fixed light-dark cycle differ markedly between subjective summer and winter. A change in ambient temperature affects entrainment only during subjective winter when it also produces pronounced effects on body temperature (Tb). It was previously proposed that variations in Tb are causally related to the circannual rhythm in circadian entrainment. To test this hypothesis, wheel-running activity and Tb were monitored for 12 to 14 months in castrated male ground squirrels housed in a 14:10 LD photocycle at 21 degrees C. Animals were treated with testosterone implants that eliminated hibernation and prevented the marked winter decline in Tb; these squirrels manifested circannual changes in circadian entrainment indistinguishable from those of untreated animals. Both groups exhibited pronounced changes in phase angle and alpha of circadian wheel-running and Tb rhythms. Seasonal variation in Tb is not necessary for circannual changes in circadian organization of golden-mantled ground squirrels.

Pineal-independent regulation of photo-nonresponsiveness in the Siberian hamster (Phodopus sungorus).

The pineal hormone melatonin influences circadian rhythms and also mediates reproductive responses to photoperiod. The authors tested whether pinealectomy influences circadian oscillators responsible for induction of nonresponsiveness to short day lengths by preventing normal short-day patterns of circadian entrainment. Adult male Siberian hamsters were pinealectomized or sham operated, maintained in either 18 h light per day (18L) or 15L for 10 weeks, and then tested for responsiveness to 10L. Because pinealectomized hamsters do not show gonadal regression in short day lengths, responsiveness was assessed by measuring phase angle of entrainment and the length of the nightly activity period following transfer to 10L. The incidence of nonresponsiveness was significantly higher in 18L hamsters than in 15L hamsters but was unaffected by pineal status. Fully 88% of 18L hamsters failed to entrain to 10L in the normal short-day manner; the duration of nightly activity remained compressed, and the phase angle of entrainment was large and negative relative to lights off. The 15L hamsters entrained normally to 10L. Exposure to constant light after 10L treatment was equally effective in inducing arrhythmicity in pinealectomized and intact hamsters. Changes in the period of morning and evening circadian oscillators subsequent to 18L treatment did not predict circadian responsiveness to short photoperiod. Long-day induction of photo-nonresponsiveness, which prevents winter responses to short day lengths, occurs independently of pineal melatonin feedback on the circadian system.

Photoperiodism in hamsters: abrupt versus gradual changes in day length differentially entrain morning and evening circadian oscillators.

In studies of photoperiodism, animals typically are transferred abruptly from a long (e.g., 16 h light per day [16L]) to a short (8L) photoperiod, and circadian oscillators that regulate pineal melatonin secretion are presumed to reentrain rapidly to the new photocycle. Among rats and Siberian hamsters, however, reentrainment rates vary depending on whether additional darkness is added to morning or evening, and a subset of hamsters (nonresponders) fails ever to reentrain normally to short photoperiods. The authors assessed whether several short-day responses occurred at different rates when darkness was extended into morning versus evening hours and the effectiveness of abrupt versus gradual shortening in day lengths (DLs). Entrainment patterns of photoresponsive hamsters also were compared to those of photononresponsive hamsters. Responsive hamsters transferred on a single day from 16L to 8L underwent more rapid gonadal regression, weight loss, decreases in follicle-stimulating hormone titers, and expansion of nocturnal locomotor activity when darkness was added to morning versus evening. When the dark phase was extended gradually by 8 h over 16 weeks, short-day responses occurred at the same rate whether darkness was appended to morning or evening or was added symmetrically. Darkness added to evening promoted more rapid short-day responses when it was added gradually rather than abruptly, despite the fact that average DLs were significantly shorter for the latter group. Among nonresponders, morning extensions of darkness transiently increased activity duration, whereas evening extensions did not. Gradual and abrupt decreases in DL differentially affect entrainment of evening and morning circadian oscillators. The authors argue for the incorporation of simulated natural photoperiods in studies of photoperiodism.

Low ambient temperature accelerates short-day responses in Siberian hamsters by altering responsiveness to melatonin.

Exposure to low ambient temperatures (Ta) accelerates appearance of the winter phenotype in Siberian hamsters transferred from long to short day lengths. Because melatonin transduces the effects of day length on the neuroendocrine axis, the authors assessed whether low Ta promotes the transition to winterlike traits by accelerating the onset of increased nocturnal melatonin secretion or by enhancing responsiveness to melatonin in short day lengths. Male hamsters were transferred from 16L (16 h light/day) to 8L (8 h light/day) photoperiods and held at 5 degrees C or 22 degrees C. Locomotor activity was recorded continuously, and body mass, testis size, and pelage color were determined biweekly for 8 weeks. The duration of nocturnal locomotion (alpha), a reliable indicator of the duration of nocturnal melatonin secretion, lengthened significantly earlier in hamsters exposed to a Ta of 5 degrees C than 22 degrees C. Cold exposure increased the proportion of hamsters that were photoresponsive: gonadal regression in short days increased from 44% at 22 degrees C to 81% at 5 degrees C (p < 0.05); low Ta did not, however, accelerate testicular regression in animals that were photoresponsive. Nonphotoresponsive animals at 5 degrees C temporarily had longer alphas during the first 4 weeks in short days and significant decreases in body mass and testicular size that were reversed during the ensuing weeks when alpha decreased. In a 2nd experiment, pinealectomized male hamsters infused for 10 h/day with melatonin for 2 weeks had significantly lower body and testes masses when maintained at 5 degrees C but not 22 degrees C. Low-ambient temperature appears to accelerate the appearance of the winter phenotype primarily by increasing target tissue responsiveness to melatonin and to a lesser extent by augmenting the rate at which the duration of nocturnal melatonin secretion increases in short day lengths.

Estradiol acts as a competitive inhibitor of the 3 beta-hydroxysteroid dehydrogenase/delta 5-delta 4 isomerase enzyme of cultured Leydig tumor cells.

To study the local regulatory mechanisms involving steroid hormones in steroidogenic cells, the effect of estradiol on steroidogenesis was investigated using MA-10 Leydig tumor cells. Estradiol inhibited progesterone biosynthesis in MA-10 cells in a dose-dependent manner. Inhibition of progesterone biosynthesis by estradiol was associated with a concomitant accumulation of pregnenolone in the incubation medium. Estradiol inhibited the activity of 3 beta-hydroxysteroid dehydrogenase/delta 5-delta 4 isomerase by a chemical mechanism which is not mediated through the cellular estrogen receptor. Thus, the estrogen receptor agonist diethylstilbestrol did not inhibit this enzyme activity, nor could this agent block the effect of estradiol on the enzyme. Furthermore, estradiol inhibited enzyme activity in isolated microsomes which do not contain estradiol receptor protein. Kinetic analysis of the inhibitory effect of estradiol on 3 beta-hydroxysteroid dehydrogenase/delta 5-delta 4 isomerase revealed that this steroid hormone functions as a competitive inhibitor of the enzyme, with an average apparent Ki of 1.8 microM.

Constitutive steroidogenesis in the R2C Leydig tumor cell line is maintained by the adenosine 3',5'-cyclic monophosphate-independent production of a cycloheximide-sensitive factor that enhances mitochondrial pregnenolone biosynthesis.

These studies were designed to characterize constitutive steroidogenesis in Leydig tumor cells. Constitutive steroidogenesis was investigated by comparing constitutively active R2C Leydig tumor cells to trophic hormone-responsive MA-10 Leydig tumor cells. Unlike the MA-10 cells, R2C cells appeared to synthesize steroid hormones independently of the cAMP-protein kinase pathway. Although the adenylate cyclase of R2C cells could be stimulated in the expected manner by cholera toxin, cAMP concentrations in these cells were low, and R2C cell steroidogenesis could be dissociated from other cAMP-dependent processes. Two cAMP-dependent processes in steroidogenic cells, protein kinase activation and lactate formation, showed low basal activities in R2C cells and could be stimulated by (Bu)2cAMP with a dose dependence similar to that detected in MA-10 cells. Steroid hormone biosynthesis parallelled these other cAMP-dependent processes in MA-10 cells, but not in R2C cells. Cycloheximide, however, caused similar dose-dependent inhibition of steroidogenesis in both the R2C and MA-10 cells. Using a cell component bioassay, it was shown that R2C cells constitutively synthesize an extramitochondrial cycloheximide-sensitive factor that is functionally identical to the factor produced in response to hCG in MA-10 cells. This factor enhanced mitochondrial pregnenolone biosynthesis. Thus, constitutive steroidogenesis in R2C cells could be explained by the cAMP-independent but cycloheximide-sensitive constitutive production of an extramitochondrial factor that activated mitochondrial pregnenolone biosynthesis.

Plasma membrane cholesterol: removal and insertion into the membrane and utilization as substrate for steroidogenesis.

The plasma membrane cholesterol content of MA-10 Leydig tumor cells is depleted by trophic hormone stimulation and repleted by incubating the cells with low density lipoprotein. The present studies used subcellular fractionation to investigate the membranes involved in steroid hormone synthesis. The results showed that the plasma membrane was the major source of cholesterol substrate and that the cholesterol content changed independently of any mass changes in membrane protein or phospholipid. Membrane phospholipid composition also did not change as membrane cholesterol content decreased or increased, a finding inconsistent with the proposal that phospholipid composition dictates the amount of cholesterol contained in a membrane. The mitochondria of the MA-10 cells were cholesterol rich, containing more cholesterol per unit protein or phospholipid than the plasma membrane. This cholesterol was presumably in the outer mitochondrial membrane, since virtually all of the cholesterol of intact mitochondria was accessible to cholesterol oxidase. Although there was a high concentration of mitochondrial cholesterol, this cholesterol was largely inert as a substrate for steroidogenesis, and plasma membrane cholesterol was incorporated into steroid hormones without ever equilibrating with the mitochondrial cholesterol pool.

Effect of cholesterol transport inhibitors on steroidogenesis and plasma membrane cholesterol transport in cultured MA-10 Leydig tumor cells.

These studies were directed toward understanding the cellular actions of inhibitor drugs that affect steroidogenesis and cholesterol transport. We investigated the microfilament inhibitor cytochalasin-D, the microtubule inhibitor colchicine, the calmodulin antagonist trifluoperazine, and the inhibitor of acidic vesicle function nigericin. We found that all of these compounds caused dose-dependent inhibition of progesterone synthesis in the MA-10 cells. Each compound also inhibited (Bu)2cAMP-stimulated pregnenolone synthesis, indicating that each inhibited a fundamental process required for steroidogenesis. Each compound was next evaluated for inhibitory actions on cholesterol transport to and from the plasma membrane. On the basis of inhibitor sensitivity, two different categories of cholesterol transport were defined. Transport of newly synthesized or low density lipoprotein-derived cholesterol from the cell interior to the plasma membrane was inhibitor insensitive. Plasma membrane cholesterol internalization, however, was sensitive to all of the inhibitors and did not result because of any drug effect on the acyl-coenzyme-A-cholesterol acyl transferase. Cycling of cholesteryl ester-derived cholesterol through the plasma membrane appeared to occur before its use for steroidogenesis. Thus, inhibition of plasma membrane internalization would prevent utilization of both plasma membrane cholesterol and cholesteryl ester-derived cholesterol, the two major substrate sources for steroid hormone synthesis. Consistent with this interpretation was the finding that inhibition of plasma membrane cholesterol internalization by each inhibitor paralleled the inhibitor's effect on steroidogenesis.

A cholesteryl ester hydrolase inhibitor blocks cholesterol translocation into the mitochondria of MA-10 Leydig tumor cells.

The present studies describe an unexpected action of a cholesteryl ester hydrolase inhibitor on MA-10 Leydig tumor cells. These studies were initially intended to use the inhibitor, diethylumbelliferyl phosphate, to block cholesteryl ester hydrolysis and, thus, determine the contributions of this form of cholesterol to steroidogenesis and reveal any direct hormone effects on cholesterol esterification. Although this compound acted as an effective inhibitor of the cholesteryl ester hydrolase in intact MA-10 cells, it inhibited steroidogenesis at lower concentrations and to a greater extent than could be explained by simple inhibition of the ester hydrolase enzyme. This compound proved not to be generally toxic, but blocked some process occurring between cAMP formation and cholesterol side-chain cleavage. The diethylumbelliferyl phosphate block of steroidogenesis was readily bypassed by 22-hydroxycholesterol. These data indicated that the compound inhibited cholesterol transport. The lesion in cholesterol transport was not general, but very specific; cholesterol translocation to the mitochondrial site of cholesterol side-chain cleavage was blocked by this organophosphate compound.

Finasteride blocks progesterone synthesis in MA-10 Leydig tumor cells.

The 5 alpha-reductase inhibitor, finasteride, inhibits progesterone synthesis in the MA-10 Leydig tumor cells. Inhibition is dose-dependent with half maximal inhibition occurring at 10 ng/ml, a concentration significantly less than serum concentrations detected in finasteride-treated patients. Experiments to localize the site of inhibition by this compound revealed that the 3 beta-hydroxysteroid dehydrogenase delta 5-->delta 4 isomerase enzyme was not blocked by finasteride, but that cholesterol side-chain cleavage was inhibited. Thus, both dibutyryl-cAMP-stimulated and 22-hydroxycholesterol-stimulated steroidogenesis were inhibited by finasteride. This effect of finasteride to block cholesterol side-chain cleavage may be species-specific. Inhibition is readily detected in the mouse-derived MA-10 cells; however, human granulosa cell steroidogenesis is finasteride-insensitive while rat Leydig cell steroidogenesis is only minimally effected by finasteride.

Inhibition of steroidogenesis in cultured Leydig tumor cells by 22-amino-23,24-bisnor-5-cholen-3 beta-O1 and (20R) 20-phenyl-5-pregnene-3 beta,20-diol.

Using the cholesterol side-chain cleavage enzyme purified from bovine adrenals, we have previously shown that 22-amino-23,24-bisnor-5-cholen-3-beta-ol (22-ABC) and (20R) 20-phenyl-5-pregnene-3 beta, 20 diol (20-PPD) are potent competitive inhibitors of this enzyme. The studies presented herein were designed to characterize the effects of these new inhibitors on steroid production by intact cells. Using cultured Leydig tumor cells, we compared the effects of 22-ABC, 20-PPD, and aminoglutethimide (a well-known inhibitor of the cholesterol side-chain cleavage enzyme) on hormone-stimulated steroidogenesis. Our results show that these compounds inhibit steroid production in a dose-dependent manner and that 20-PPD and 22-ABC are about 100 and 4 times more active, respectively, than aminoglutethimide. In these cells, the inhibitory effects of the new compounds seem to be localized exclusively at the conversion of cholesterol to pregnenolone.

Steroid hormone-producing tumors of the adrenal, ovary, and testes.

Steroid hormone-synthesizing tumors are rare and may be difficult to diagnose and treat. This article focuses on adrenal tumors as models for the even less common tumors of the ovary and testes, and emphasizes useful diagnostic procedures and pitfalls as well as treatment options. Ovarian and testicular steroidogenic tumors are categorized and discussed in terms of the generalizations that can be derived from adrenal neoplasia.

Cellular internalization, transport, and esterification of iodine-125-NP59 by MA-10 Leydig tumor cells.

The present studies were directed toward understanding the cellular processing of the cholesterol analogue, NP59. NP59 readily entered MA-10 Leydig tumor cells. The cholesterol analogue entered the cells by binding to the plasma membrane and becoming internalized along with plasma membrane cholesterol. Internalized NP59 was readily esterified to NP59 ester. Transport of NP59 within the cell was indistinguishable from transport of cholesterol. Cholesterol and NP59 transport were under the control of cAMP, however, only cholesterol entered the mitochondria and was converted into progesterone. Thus, internalized NP59 could not be removed from the cell by conversion into steroid hormones. Esterified NP59 was metabolically inert and could not be converted back to free NP59 and free fatty acid. Since NP59 was not a substrate for the cholesteryl ester hydrolase, it became trapped in the cell as NP59 ester.

Immunohistochemical localization of pregnancy-related placental protein 4 in human placenta, umbilical cord and adult human female genital tissues.

Placental protein 4 (PP4) is a soluble placental tissue protein which was isolated from human placenta. The aim of the present study was to demonstrate the localization of cells containing PP4 in human placenta and in various female genital tissues under normal conditions. PP4 immunoreactive structures were demonstrated by using the peroxidase-antiperoxidase immunohistochemical technique. The samples were obtained from normal human placenta, umbilical cord, uterine cervix, endometrium, ovary and vulva. The most differentiated trophoblastic cells, the syncytiotrophoblasts, as well as the intermediate trophoblast cells contained PP4. PP4 immunoreactivity was present in umbilical cord as well. Occasionally PP4 was detected in normal ovarian, endometrial or vulvar tissue samples. Cervix and myometrium were free of PP4 immunoreactive material. PP4 staining was cytoplasmic. Our findings indicate that PP4 cannot be considered specific for the placenta since it is present in some human adult tissues as well.

Termination of neuroendocrine refractoriness to melatonin in Siberian hamsters (Phodopus sungorus).

Siberian hamsters maintained from birth in a short day length (DL), unlike their long-day counterparts, fail to undergo reproductive development by 5 weeks of age. Instead, reproductive maturation of short-day males is delayed for approximately 20 weeks, at which point neuroendocrine refractoriness to the inhibitory effects of short DLs develops, resulting in growth of the gonads. To terminate refractoriness and re-establish responsiveness to short photoperiods, 10-15 weeks of long-day exposure is required. We assessed whether continuous exposure to long days is necessary to terminate refractoriness or whether the first few weeks of long days initiate a process that culminates several months later in the breaking of refractoriness. Male hamsters refractory to short DLs were transferred to a long-day photoperiod, pinealectomized (PINx) after 0, 3, 6 or 15 weeks, and subsequently infused for 6 weeks with a short-day melatonin signal. This melatonin treatment induces gonadal regression in photosensitive but not in photorefractory hamsters. Six percent of males PINx at week 0 and 88% of those PINx at week 15 underwent gonadal atrophy by the end of the melatonin infusion treatment initiated on week 15. Among hamsters PINx on week 6, 17% versus 76% underwent testicular involution in response to melatonin infusions initiated on week 6 and week 15, respectively. This finding indicates that a fraction of the long days that hamsters experience during spring and summer are sufficient to trigger the processes that restore responsiveness to short DLs. Additional groups of pineal-intact photorefractory animals were given 3, 6 or 15 weeks of long-day exposure and then returned to a short DL for several months; only those treated for 15 weeks terminated refractoriness. The breaking of refractoriness, once triggered by long-day melatonin signals, proceeds to completion only in the absence of short-day melatonin signals.