Photoperiodic Tumor Development and Immune Functionmedicine.osu.edu/neuroscience/documents/RandyPaper… ·  · 2012-09-11Photoperiodic Effects on Tumor Development and Immune Function

233

Photoperiodic Effects on Tumor Developmentand Immune Function

Randy J. Nelson1 and Joan M. C. Blom2Department of Psychology, Behavioral Neuroendocrinology Group,The Johns Hopkins University, Baltimore, Maryland 21218

Abstract Seasonal changes in adaptations associated with winter coping strategies have been fre-quently studied. Central among the suite of energy-saving, winter-coping strategies is the suspension ofreproductive activities. The inhibition of reproduction by nontropical rodents is mediated by daylengthchanges. Although balanced annual energy budgets are critical, survival and subsequent reproductivesuccess also require avoiding predators, illness, and early death. Because the stressors of winter couldlead to suppressed immune function, we hypothesized that animals should have evolved survivalstrategies involving immunoenhancement. Short daylengths provide a predictive cue to individualsthat could be used to enhance immune function in advance of stress-induced immunosuppression. InExperiment 1, adult female deer mice (Peromyscus maniculatus) were housed in either long (LD 16:8)or short (LD 8:16) days for 8 weeks, then injected with the chemical carcinogen 9,10-dimethyl-1,2-benzanthracene (DMBA) dissolved in dimethyl sulfoxide (DMSO) or with the DMSO vehicle alone.Animals were evaluated weekly for 8 weeks after injection. None of the animals treated with DMSOdeveloped tumors in any of the experiments. Nearly 90% of the long-day deer mice injected withDMBA developed squamous cell carcinoma. None of the short-day deer mice injected with DMBAdeveloped tumors. Small lesions developed at the site of injection; short-day females had less severelesions and healed faster than long-day females. Immunoglobulin G (IgG) response to i.p. injectionof sheep red blood cells (SRBC) did not differ between photoperiodic conditions. The role of estrogensin the photoperiodic responses was evaluated in Experiment 2: Ovariectomized or sham-ovariectomizeddeer mice received estradiol benzoate replacement therapy or a control procedure in long daylengthsfor 8 weeks prior to injection of DMBA or DMSO, then were monitored for 8 additional weeks.Females treated with DMBA developed tumors at the same rate, regardless of estrogen manipulation.Estrogen did not affect healing rates. In Experiment 3, female deer mice were injected with a slurryof microspheres that either contained bromocriptine or were empty. Suppression of prolactin withbromocryptine resulted in a decrease of tumor incidence from 55.6% to 24% in long-day females 8weeks after injection with DMBA. Healing rates were not affected by prolactin manipulations. Silasticcapsules that were filled with either melatonin or cholesterol were implanted into long-day female deermice in Experiment 4; 8 weeks later, females received an injection of either DMBA or DMSO, thenwere monitored for 8 weeks. Approximately 66% of females implanted with cholesterol and injectedwith DMSO developed histologically verified tumors. None of the melatonin-implanted mice developedtumors. Melatonin did not affect healing rates. Taken together, these results indicate that photoperiodcan exert a functionally significant effect on immune processes and clinical disease.

Key words immunoglobulin, seasonality, daylength, mammary cancer, carcinoma, estrogen,prolactin, melatonin, DMBA, tumorigenesis

1. To whom all correspondence should be addressed.

2. Present address: Center for Neuropharrnacology, Via Balzavett 9, 20133 Milan, Italy.

234

Seasonal changes in energy requirements have been documented for many species of nontrop-ical rodents (Bronson, 1989; Goldman and Nelson, 1993). The energy required for thermoreg-ulation reaches a peak during the winter, when energy availability is typically reduced. Theenergetic &dquo;bottleneck&dquo; created by high energy demands during low energy availability hasled to the evolution of many adaptations to cope with winter (Blank, 1992; Wunder, 1992).Reproduction is energetically expensive, and the curtailment of breeding is central amongthe suite of winter-coping strategies among boreal and temperate zone rodents (Blank, 1992;Wunder, 1992). Successful individuals have evolved strategies to maximize the length of theirbreeding effort without jeopardizing their survival vis-d-vis energy usage. These temporalstrategies require the ability to ascertain the time of year in order to phase energeticallyexpensive activities to coincide with peak energy availability and other local conditions thatpromote survival (Bronson, 1989; Moffatt et al., 1993). Photoperiod can be used by nontropicalanimals as a very precise temporal cue for time of year. This temporal information is oftenemployed by individuals to anticipate predictable seasonal environmental changes, and to initiateor terminate specific seasonal adaptations in order to maintain a positive energy balance.

Maintenance of a positive energy balance is obviously required for survival, and thevast majority of studies of seasonality have focused upon the temporal organization ofenergetic adaptations (for reviews, see Nelson et al., 1990; Bronson and Heideman, 1994).However, other threats to survival must also be met in order for individuals to increasetheir fitness. They must avoid encountering predators, engaging in potentially dangerousinteractions with conspecific competitors, and succumbing to disease. Certainly, energeticscan interact with immune function; a negative energy balance can sufficiently weaken animalsto the extent that they increase their risk of infection and death (Kelley, 1985). Otherconditions perceived as stressful can suppress immune function and promote opportunisticdiseases. Stressful conditions such as reduced food availability, low ambient temperatures,overcrowding, lack of cover, or increased predator pressures can recur on a seasonal basis,leading to seasonal changes in immune function among individuals, as well as seasonalchanges in population-wide disease and death rates (Lee and McDonald, 1985; Lochmilleret al., 1994). Because these stressful conditions are somewhat recurrent, we have hypothe-sized that animals may have evolved mechanisms to combat seasonal stress-induced reduc-

tions in immune function. From an evolutionary and ecological perspective, it is reasonableto expect that rodents have evolved the ability to forecast recurrent conditions associatedwith immunosuppression and to maximize immune function in advance of these challeng-ing conditions.

In addition to the well-studied seasonal cycles of mating and birth, there are alsoseasonal cycles of illness and death among many populations of animals (e.g., Bradley etal., 1980; McDonald et al., 1981; Mihok et al., 1988; Lochmiller et al., 1994). Furthermore,seasonal fluctuations in immune function and the incidence of opportunistic diseaseshave been well documented in a variety of species, including humans (Christian, 1978;McDonald et al., 1981; Maestroni and Conti, 1991; Blom, 1992). For example, thenumbers of circulating leukocytes in mice (Mus), rats (Rattus), rabbits (Lepus), dogs(Canis), ground squirrels (Citellus), voles (Microtus), and humans were reported to beelevated during the autumn and winter; spleen and thymus masses were also reported tobe highest during autumn and winter in deer mice (Peromyscus maniculatus), prairievoles (Microtus ochrogaster), and northern red-backed voles (Clethrionomys rutilus) (fora review, see Blom, 1992).

235

These seasonal changes in immune function appear to be mediated by daylength. Previ-ous studies have indicated that short-day exposure increases splenic masses of deer miceand Syrian hamsters (Mesocricetus auratus); short-day exposure also elevates total lympho-cyte and macrophage counts (Vriend and Lauber, 1973; Brainard et al., 1985, 1988; Champ-ney and McMurray, 1991; Blom et al., 1994). Enhancement of immune cell numbers shouldbe particularly crucial to young bom near the end of the breeding season. Indeed, informationabout short daylengths is transmitted by mothers to their offspring in utero, which acceleratesimmune cell production in the young (Blom et al., 1994). Melatonin appears to mediatephotoperiodic enhancement of immune function both in vivo and in vitro; however, someimmunosuppressive effects of melatonin have also been reported (reviewed in Maestroniand Conti, 1991). Importantly, virtually all of the studies examining the interaction betweenmelatonin and immune function have been conducted on humans or inbred strains of labora-

tory mice and rats-species that do not usually display robust responsiveness to melatoninor photoperiod. Also, many previous studies have not controlled for the suppressive effectsof short days or melatonin treatment on reproductive function. Reproductive steroid hor-mones, as well as prolactin, can directly affect immune function (for reviews, see Grossman,1984, 1985; Bemton et al., 1988; Hartmann et al., 1989). Consequently, the reportedphotoperiodic effects on immune function may merely reflect steroid, prolactin, or melatonineffects on immune function.

Associated with its effects on immune function, melatonin has also been reported topossess oncostatic properties against several types of tumors both in vivo and in vitro

(reviewed in Blask, 1984; Blask et al., 1991). In common with studies of melatonin andimmune function, the majority of studies investigating the effects of melatonin on cancerare based on inbred strains of mice and rats, which are not particularly responsive tomelatonin. Furthermore, previous studies have administered melatonin with little regard forthe temporal characteristics of the melatonin treatment (Blask et al., 1992). Consequently,the adaptive function of this pineal hormone in mediating seasonal changes in immunefunction may have become obscured. The present study investigated the influence ofdaylength on immune function and tumorigenesis in female deer mice (Peromyscus manicula-tus). We hypothesized that short days should enhance immune function and suppress tumordevelopment in this species. The roles of estrogen, prolactin, and melatonin in tumor develop-ment were also assessed.

METHODS

ANIMALS AND HOUSING CONDITIONS

Adult (>50 days of age) female deer mice (Peromyscus maniculatus bairdii) were obtainedfrom our colony, which was originally derived from animals obtained from the PeromyscusStock Center at the University of South Carolina. Animals were maintained with theirlittermates until weaning at 21 days of age (or after their body mass was >7 g), and were

individually housed thereafter in polypropylene cages (28 x 7.5 x 13 cm) at 21° ± 2°Cand a relative humidity of 50% ± 5%. Food (Prolab 1000; Agway, Syracuse, NY) andtapwater were available ad libitum throughout all experiments. Deer mice were maintainedin either long (LD 16:8; lights-on at 0700 hr Eastern Standard Time [EST]) or short (LD8:16; lights-on at 1000 hr EST) photoperiods.

236

GENERAL EXPERIMENTAL PROCEDURES

After 8 weeks in their respective photoperiodic conditions, animals were injected subcutane-ously (s.c.) in the lower left abdominal region with the chemical carcinogen 9,10-dimethyl-1,2-benzanthracene (DMBA) dissolved in dimethyl sulfoxide (DMSO) (50 mg/kg) or withthe DMSO vehicle alone. All animals were palpated weekly to detect tumor onset and growthduring an 8-week postinjection period. After 8 weeks, the animals were anesthetized withmethoxyflurane (Metofane; Pitman Moore, Fort Washington, NJ) vapors and weighed, anda blood sample was obtained; animals were then killed by cervical dislocation. Tumors,spleens, and (in some experiments) uteri were dissected at autopsy and the weights recorded.Tumors were fixed for later histopathological diagnoses.

Animals were examined daily and also monitored by the veterinary staff; during thecourse of the experiments, the injection sites were observed to develop skin irritation resultingin lesions. The severity and rate of healing of these lesions were assessed on a weekly basisin some of the experiments with a subjective 5-point scale: (0) no lesions, (1) barely detectablelesions, (2) mild lesions (<3 mm), (3) moderate lesions (3-5 mm), or (4) severe lesions(>5 mm). Animals with severe lesions or with tumors >5 mm in diameter were sacrificed

on the day that these were discovered (n = 3 for all experiments).

BLOOD SAMPLES

Blood samples (250 ~,1) were collected between 0900 hr and 1200 hr EST from the retro-orbital sinus, stored at room temperature for 1 hr, declotted, and then centrifuged at 3500rpm for 1 hr at 4°C. Serum was separated and stored frozen at - 80°C until assayed forimmunoglobulin G (IgG).

IMMUNE ASSAYS AND PATHOLOGY

Sheep red blood cells (SRBC) were obtained from Truslow Farms (Chestertown, MD). TheSRBC (3 ml) were washed three times in 0.2 M phosphate-buffered saline (PBS; pH 7.2);they were then resuspended in PBS for a final concentration of 0.1%.

IgG levels in the blood samples were determined by means of a sandwich enzyme-linked immunosorbent assay (ELISA) that was developed and validated for use in deer micein our laboratory. Serial dilutions yielded values in parallel to the standard curves. Weincubated 96-well immunoplates (Nuncg, MaxiSorp) overnight at 4°C with 100 Wl/well of agoat polyclonal antibody against house mouse IgG (Cappel) diluted 1:3000 in a carbonate-bicarbonate buffer (0.1 M, pH 9.6). The following day the plates were washed four timeswith PBS (0.05 M, pH 7.4) containing 0.05% Tween-20 and 0.001% of NaN3; an automaticmicroplate washer (BioRad, model 1550) was used. A standard curve (upper limit, 1000fLg/ml; lower limit, 0.001 fLg/ml) was prepared using purified mouse IgG (Sigma) dilutedin standard diluent (PBS [0.05 M, pH 7.4] containing 0.05% Tween-20). The standards (100iLl/well in triplicate), and samples of deer mouse serum (100 ~,1/well in duplicate) diluted1:100 with standard diluent, were placed in wells on the plates. The plates were incubatedovernight at 4°C. The following day the plates were washed, and 100 ~,1 of alkalinephosphatase-conjugated sheep antimouse IgG diluted 1:2500 in standard diluent wereadded to each well. The plates were incubated overnight at 4°C. The following day theplates were again washed, and 100 fLI of substrate buffer (0.1 mMp-nitrophenyl phosphate

237

in diethanolamine buffer [0.1 M, pH 9.5] containing 5 mM MgC’2) was added to eachwell. The plates were incubated for 30 min, and the optical density of the resultingcolored product in each well was measured at 405 nm with a microplate reader (BioRad,model 450). The absolute concentrations of IgG in the samples were determined relativeto the standard curve.

At autopsy, the presence or absence of tumors was recorded. Tumors were excised at

autopsy and preserved in a solution of 10% formalin dissolved in PBS. Tissue samples fromall groups were fixed and stained with hematoxylin and eosin at American Histolab, Inc.(Gaithersburg, MD). Pathological evaluation of slides was performed by Pathco, Inc. (Ijams-town, MD). Most tumors were diagnosed as squamous cell carcinoma. In a few instancesthe squamous cell carcinomas were accompanied by primary mammary carcinomas. Tumorincidence was defined as the number of animals with palpable and histologically verifiedtumors per number of animals injected.

DATA ANALYSES

An overall analysis of variance was used to test comparisons among groups for each experi-mental parameter. Individual pairwise comparisons were analyzed with independent two-tailed t tests. Because these comparisons were planned, no post hoc corrections were used

(Keppel, 1982). Some data were nonparametric, and mean differences were evaluated withx2. The treatment effects were considered statistically significant if p < 0.05.

EXPERIMENT 1

Sixty-four adult female deer mice were housed individually in either LD 16:8 or LD 8:16for 8 weeks as described above. Half of the animals were then injected s.c. with DMBA orthe DMSO vehicle, and were palpated weekly for tumors. Eight weeks after injection, theanimals were treated as described above.

EXPERIMENT 2

Short days decrease plasma estradiol levels. If DMBA tumors are estrogen-dependent, then

exposure to short days may reduce tumor development by suppressing estradiol levels. Todetermine whether the short-day effects on tumorigenesis and immune function are mediated

by estradiol, we maintained 120 adult female deer mice in LD 16:8. Animals were assignedto one of 10 experimental groups (n = 12/group). Females were subjected to either a bilateralovariectomy or a sham operation at 50 days of age. Ovariectomized animals either (1)received no further treatment, (2) received weekly injections of estradiol-17 P (Sigma; 0.1 1

mg) dissolved in 0.1 cc of sesame seed oil, or (3) received weekly injections (0.1 cc) ofthe oil vehicle alone. Sham-ovariectomized females either (1) received no further treatmentor (2) received weekly s.c. injections of sesame seed oil. One week after the injectionregimen began, half of the animals in each of the five groups were injected s.c. in the lowerleft abdominal region with DMBA dissolved in DMSO (50 mg/kg), and half of the animalswere injected with the DMSO vehicle alone. All animals were palpated on a weekly basisfor tumors for 8 weeks, then killed and examined as described above.

238

Two additional groups of adult female deer mice were maintained individually in LD16:8. Both groups of females were ovariectomized at 50 days of age; one group (n = 12)received weekly s.c. injections of estradiol, and the other group (n = 10) received injectionsof the sesame seed oil vehicle as described above. After 8 weeks, all mice were injectedintraperitoneally (i.p.) with SRBC to assess antibody formation. A blood sample was obtained4 days after SRBC injection, and IgG titers were assayed by the ELISA techniques de-scribed above.

EXPERIMENT 3

In addition to reducing levels of sex steroid hormones, exposure to short daylengths alsoreduces blood prolactin levels in every mammalian species thus far examined (Goldman andNelson, 1993). Prolactin has pronounced effects upon immune function in a variety of species(reviewed in Welsch, 1985; Bernton et al., 1988).

The effects of DMBA are believed to be partially dependent on estrogen and, morespecifically, on prolactin. An elevation in the blood level of prolactin potentially increasestumor incidence (Welsch and Nagasawa, 1977). If prolactin effectively stimulates the growthand incidence of tumors, then reduced prolactin levels should result in decreased tumorigen-esis rates. If prolactin does not affect tumorigenesis, then suppression of prolactin shouldnot affect tumor incidence. Furthermore, if suppressed prolactin levels are mediating theshort-day enhancement of immune function in deer mice, then long-day animals in whichprolactin is experimentally suppressed should have elevated antibody production upon thepresentation of a specific antigen, as compared to long-day animals experiencing physiologi-cal levels of circulating prolactin. If suppressed prolactin levels are not causing the short-day enhancement of immune function, then all mice in this experiment should have compara-ble levels of immune function.

Seventy-six adult female deer mice were individually housed for 8 weeks in long days(LD 16:8) in conditions similar to those described above. Because prolactin secretion iseffectively suppressed by dopamine, a dopamine agonist, bromocriptine, was used to lowerthe blood prolactin concentrations (Neill and Nagy, 1994). Daily injection of bromocriptinewas not a satisfactory method of prolactin suppression, because the repeated stress of thedaily injections might affect immune function and potentially confound the results; weeklyinjections were considered preferable, in order to diminish stress responses. After 8 weeksin long days, half of the mice were injected s.c. with a suspension of microspheres providinga slow, regular release of bromocriptine (Sandoz, Basel, Switzerland; see Neidhart, 1989)in a dose of 125 p,g each day, for an initial period of 21 days (i.e., each suspension injectedcontained 125 f.Lg x 21 days = 2.33 mg bromocriptine); the remaining animals in thisstudy received a placebo suspension. The injections were repeated every 3 weeks.

Three days after the initial injection with the microspheres, half of the animals in eachgroup were injected s.c. in the lower left abdominal region with the chemical carcinogenDMBA dissolved in DMSO (50 mg/kg) or with the vehicle alone. Thus, four experimentalgroups were formed with 15 animals per group, receiving either (1) bromocriptine + DMBA,(2) bromocriptine + DMSO, (3) vehicle suspension + DMBA, and (4) vehicle suspension+ DMSO. During an 8-week postinjection period, animals were palpated to detect tumorgrowth on a weekly basis. After 8 weeks a blood sample was obtained and assayed forprolactin by radioimmunoassay (RIA); however, the prolactin RIA data were inadvertently

239

destroyed. Tumors, spleens, and uteri were dissected at autopsy, and histopathology wasperformed to verify the nature of the tumors.

EXPERIMENT 4

Melatonin also influences tumor growth. Experimental and clinical reports indicate that thereis a link between cancer development and pineal function (Bartsch and Bartsch, 1981;

Lapin and Ebels, 1981; Blask, 1984; Hill and Blask, 1988). The data, however, are ofteninconsistent. In some cases, pinealectomy enhances tumor development; in other cases,pinealectomy inhibits tumor growth (for a review, see Blask et al., 1992). The conflictingdata suggesting both anti- and procarcinogenic effects of melatonin and conflicting effectson immune function are reminiscent of the &dquo;antigonadal&dquo; and &dquo;progonadal&dquo; effects of mela-tonin on reproduction (Reiter, 1983).

Experiment 4 investigated the role of melatonin in mediating the photoperiodic influ-ences on immune function. If increased levels of melatonin in short photoperiodic conditionsdecrease the susceptibility to tumorigenesis, then females maintained in long photoperiodicconditions and treated with melatonin implants should also display a decline in the susceptibil-ity to develop tumors. In contrast, in the absence of melatonin treatment, long-day miceshould show an increase in tumor susceptibility.

Thirty-four female deer mice were housed in long days (LD 16:8) at 21° ± 2°C,with uncontrolled humidity (50-60%). Implants of melatonin capsules induce reproductiveregression in deer mice comparable to that observed in animals maintained in short daylengths(Lynch and Epstein, 1976; Petterborg and Reiter, 1983; Carlson et al., 1989).

At 50 days of age, females were implanted s.c. with silastic capsules (15 mm in length,inner diameter = 0.147 cm, outer diameter = 0.195 cm) filled with either melatonin (Sigma)or cholesterol. These capsules provided melatonin that approximated physiological levels ofshort-day animals necessary to induce reproductive regression in deer mice (Lynch andEpstein, 1976). Three days after receiving the melatonin or placebo implant, animals wereinjected s.c. with DMBA (50 mg/kg) or with the DMSO vehicle. Animals were checked atweekly intervals for tumor development, and autopsies were performed 8 weeks after DMBAor DMSO injection as described before.

RESULTS

EXPERIMENT 1

The majority of the animals that developed tumors were diagnosed with squamous cellcarcinoma, but some mice also developed primary mammary carcinoma. None of the micemaintained in short days developed tumors 10 weeks after treatment with DMBA. In contrast,89% (8/9) of the LD 16:8 animals treated with DMBA developed tumors within 3-4 weeks.None of the mice in either daylength developed tumors after DMSO treatment (Fig. 1).

Both DMBA and DMSO irritated the skin, regardless of the photoperiod in which theanimals were maintained. However, short-day females exhibited fewer skin lesions (60%vs. 80%; p < 0.05), as well as accelerated healing of the lesions, compared to long-dayfemales (Fig. 2). Animals injected with DMBA displayed more severe wounds and a slower

240

FIGURE 1. Percentage of female deer mice housed in long (LD 16:8) or short (LD 8:16) days thatdeveloped histologically verified tumors after injection with the chemical carcinogen DMBA dissolvedin DMSO, or injection with DMSO alone. Number of animals in each treatment group is indicatedin parentheses.

pace of healing than animals injected with DMSO (Fig. 3). IgG titers did not differ signifi-cantly between photoperiodic treatment groups (LD 16:8, 0.0065 ± 0.002 fLg/ml; LD 8:16,0.0119 ± 0. 003; p >0.05).

These results indicate that photoperiod affects the susceptibility to chemically in-duced tumorigenesis. However, the nature of the mechanism underlying the photoperiodiceffect on tumorigenesis remains unspecified. One possible explanation may be that thetumors caused by DMBA are prolactin- and/or estrogen-dependent (Tamarkin et al.,1981). Short days could inhibit tumors by suppressing these hormones. Alternatively,melatonin might either be directly or indirectly involved in the photoperiodic modificationof tumorigenesis. Regardless of the mechanism, the results of this study suggest a poten-tially important functional role of photoperiod in disease onset and in healing pro-cesses.

EXPERIMENT 2

The rate of tumorigenesis did not significantly differ as a result of estrogen treatmentamong the groups that were injected with DMBA (p > 0.05 in all cases) (Fig. 4). Theseresults suggest that the presence or absence of estrogen does not differentially affectsusceptibility to DMBA-induced tumorigenesis. The groups that had been exposed toDMBA displayed tumor development in 62.5% to 75% of the cases. None of the animalsinjected with DMSO developed tumors (Fig. 4). Surgical incisions required an averageof 14 days to heal completely in long-day females (p < 0.05). The presence of ovariesdid not affect this rate of healing (p > 0.05).

The changes in uterine mass were consistent with the fact that the absence of estrogenresults in a substantial decrease in uterine mass. All ovariectomized females displayed a

241

FIGURE 2. Percentage of female deer mice that developed lesions after injection with DMBA (hatchedbars) or DMSO (filled bars) in short (top panel) or long (bottom panel) days. Other symbols andconventions as in Figure 1.

marked decline in uterine mass (p < 0.00 in all cases). In contrast, ovariectomized femalesin which estrogen has been replaced through weekly injections maintained uterine massesthat were statistically equivalent to those of ovary-intact females (p > 0.05) (Table 1).

Splenic mass was not significantly affected in ovariectomized deer mice by the presenceor absence of estrogen (p > 0.05). However, the presence of intact ovaries significantlyaffected splenic mass (p < 0.05); intact females had significantly smaller spleens (absoluteand adjusted for body mass) than ovariectomized females (Table 1). Injection with DMBAdid not affect spleen mass in ovariectomized females that received weekly injections with

242

FIGURE 3. Subjective ratings of the lesions of female deer mice that developed after injection withDMBA (hatched bars) or DMSO (filled bars): (0) no lesions, (1) barely detectable lesions, (2) mildlesions (<3 mm), (3) moderate lesions (3-5 mm), or (4) severe lesions (>5 mm).

sesame oil or no injection, or in females that received weekly estradiol injections (p > 0.05

in each case; data not shown).

EXPERIMENT 3

Injection with a suspension of bromocriptine resulted in tumor development among 24% ofDMBA-treated females, whereas females injected with the vehicle suspension alone devel-oped significantly more tumors after exposure to a single dose of DMBA (55.6%; p < 0.05).

243

FIGURE 4. Percentage of ovariectomized (OvX), estradiol-treated ovariectomized (OvX + E), andgonadally intact (S - OvX) deer mice housed in long (LD 16:8) days that developed histologicallyverified tumors in response to DMBA or DMSO treatment. Other symbols and conventions as inFigure 1.

Again, none of the mice treated with DMSO developed tumors (Fig. 5). Treatment with

bromocriptine did not affect uterine or splenic mass in any of the four experimental groups(Table 2). Bromocriptine treatment also failed to affect wound healing (p > 0.05).

EXPERIMENT 4

Mice treated with melatonin did not develop tumors 8 weeks after injection with DMBA.In contrast, 66% of the placebo group displayed tumors within 3 to 4 weeks after DMBAtreatment (p < 0.05). None of the mice in either group developed tumors after treatment

TABLE 1. Mean (± SEM) Body Mass and Absolute and Relative Splenic and Uterine Massof Ovariectomized and Sham-Operated Females

Note. Data of DMBA- and DMSO-injected animals did not differ significantly and were combined for this table.An asterisk (*) indicates statistically significant results.

244

FIGURE 5. Percentage of bromocriptine- or placebo-treated female deer mice housed in LD 16:8 thatdeveloped histologically verified tumors in response to DMBA or DMSO injections. Other symbolsand conventions as in Figure 1.

with DMSO (Fig. 6). Splenic mass was not affected by DMBA or melatonin treatment (p> 0.05) (Table 3). Uterine mass only differed among the melatonin-implanted animals.Females treated with melatonin and exposed to DMBA had significantly higher uterinemasses compared to melatonin-treated females injected with DMSO (p < 0.005) (Table 3).Overall, DMBA treatment resulted in a significant elevation in uterine mass (p < 0.005)(Table 3). Wound healing rate was unaffected by melatonin (p > 0.05).

DISCUSSION

Animals maintained in short daylengths did not develop tumors after treatment with thechemical carcinogen DMBA when given this agent at a dose that produced tumors in 89%of long-day cohorts. Exposure to short daylengths inhibits development of DMBA-inducedtumors. Short days also may enhance immune function. Healing was more rapid in short-day mice as compared to long-day deer mice, suggesting a photoperiodic effect on cytokineactivities; however, a direct test of immune function, IgG responses to SRBC, was notaffected by daylength in the present study. Tumor development was unaffected by manipula-tions of blood levels of estrogen. Suppression of blood prolactin levels and chronic elevationof blood melatonin levels led to a significant reduction in tumor incidence among long-daydeer mice. Manipulation of blood levels of estrogen, prolactin, or melatonin did not affecthealing rates among long-day animals.

These results are consistent with previous findings that melatonin generally enhancesimmune function and suppresses tumorigenesis (Maestroni and Conti, 1991; Blask et al.,1992). The extent to which chemically induced tumorigenesis and immune function arerelated or independent phenomena, and the extent to which compromised immune functionmay affect DMBA-induced tumors, remain unspecified.

The results of the present study suggest that seasonal changes in healing rate could bemediated by photoperiodic regulation of estrogen levels in female deer mice. Further studiesare required to assess this hypothesis, as well as to assess the effects of androgens on healing

245

TABLE 2. Mean (±SE3o Body Mass and Absolute and Relative Splenic and Uterine Mass of IntactFemales That Were Injected with a Slurry of Empty Microspheres or Microspheres Filledwith Bromocriptine

processes. A seasonal change in healing responses might reflect adaptive increases in immunefunction (Blom et al., 1994). Presumably, differential healing rates reflect differential cyto-kine responses. We attempted to assay interleukin-2 (IL-2) and IL-6 levels in the blood ofthese deer mice, using two different commercial kits (one kit was monoclonal for laboratorystrains of house mice, Mus; the other kit used polyclonal antibodies). However, both assaysfailed to detect deer mouse IL levels. Although most of the immune assays are based onMus, this species is not an ideal animal model with which to pursue this research, becauseMus do not show reliable responses to daylength or melatonin. Further studies on themechanisms of photoperiodic changes in healing processes await development of reliablecytokine assays in photoperiod-responsive species.

Seasonal changes in the rate of DMBA-induced tumorigenesis in deer mice are estrogen-independent, but may be mediated by melatonin directly or by melatonin acting upon other

FIGURE 6. Percentage of melatonin- or placebo-treated female deer mice housed in LD 16:8 thatdeveloped histologically verified tumors in response to DMBA or DMSO injections. Other symbolsand conventions as in Figure 1.

246

TABLE 3. Mean (±SE3o Body Mass and Absolute and Relative Splenic and Uterine Mass of IntactFemales Implanted with a Silastic Capsule Filled with Either Melatonin or Cholesterol

Note. An asterisk (*) indicates statistically significant results.

physiological parameters such as prolactin secretion. Thus, melatonin may play a criticalrole in the mechanism through which photoperiod affects the risk of tumor development.This could have clinical significance. For example, several groups have reported a seasonalvariation in the month of initial detection of human breast cancer, with maximum detection

reported in spring and early summer (Lee, 1967; Cohen et al., 1983; Hartveit et al., 1983;Kirkham et al., 1985; Mason et al., 1985; Chlehoun and Gray, 1987). The pattern of tumordetection is most apparent in premenopausal women with estrogen-receptor-positive tumors(Mason et al., 1985). Women secrete quantitatively more melatonin in winter than in summer(Kauppila et al., 1987), and a hormonally mediated increase in tumor rate during the springmight be mediated by the cyclic variation in the pattern of melatonin secretion (Mason etal., 1985). The exact nature of the mechanism of melatonin’s oncostatic properties, however,remains unspecified (Blask et al., 1992).

One way melatonin might affect tumorigenesis is through direct interaction with thetarget tissue by modifying the process of tumor initiation (Tamarkin et al., 1981; Blask etal., 1992). Alternatively, indirect effects of melatonin on the secretion of steroid hormones,or other hormones such as prolactin that are suppressed in response to short days (or thepattern of melatonin secretion), might be involved. Among women, diminished melatoninlevels secreted during the long days of spring and summer might allow an increase of ovariansteroidogenesis (Kauppila et al., 1987) or the production of prolactin (Wright et al., 1986).Consequently, long-day melatonin release profiles may mediate an increase in tumor growthvia higher circulating levels of estradiol or prolactin. This would be particularly likely tooccur in hormone-responsive tumors (Mason et al., 1985). In the present study, low prolactinlevels reduced tumor incidence from 56% to 24%. Assessment of blood prolactin levels willbe required to evaluate whether prolactin levels are equally suppressed in bromocriptine-treated animals as in melatonin-treated females. In the current study, our prolactin RIA datawere inadvertently destroyed.

There was no significant effect of photoperiod or hormone manipulation on IgG titersin the present study. Although IgG levels begin to increase 4 days after SRBC exposure inlaboratory mice, maximal IgG responses are not observed until 8-10 days after antigenexposure. Thus, photoperiod could exert an effect on antibody production that was undetectedin the present study. The results of the present study indicate that no photoperiodic differences

247

in IgG titers occurred within 4 days after exposure to SRBC. Improved IgG and IgM ELISAtechniques for deer mice are currently being developed in our laboratory.

Taken together, our results indicate that the ontogeny of a chemically induced tumorcan be affected by photoperiodic manipulation. Melatonin, either directly or through its

actions on other physiological processes such as prolactin secretion, may mediate photoperi-odic effects on DMBA-induced tumorigenesis. Healing rates of deer mice maintained inshort daylengths were accelerated as compared to those for long-day deer mice, suggestingphotoperiodic mediation of cytokine activities. The enhancement of healing rates and suppres-sion of chemically induced tumorigenesis associated with exposure to simulated winterdaylengths imply an adaptive immune response to short days that would proactivelycounteract winter stress-induced immunosuppression in nature. Further studies are re-quired to assess other functional changes in response to photoperiod in both immuneprocesses and tumor development, using other tumor models. Also, the clinical signifi-cance of photoperiodic mediation of immune function and tumor incidence in humansrequires further study.

ACKNOWLEDGMENTS

We are grateful for the excellent technical support and advice of L. Tamarkin, J. Shiber, C. Moffatt, M. Labbe,J. Fine, and A. Wohn; we thank R. Bungiro for superior animal care. We are also grateful for a generous contributionof bromocryptine microsphere suspensions from Sandoz, Inc., Basel, Switzerland, and thank C. Moffatt fordeveloping the IgG ELISA for deer mice.

This research was supported by National Institute of Child and Human Development (NICHD) Grant No. HD22201; National Cancer Institute Grant No. CA 58168; and a generous donation from Mr. Ted C. Hanf. It wasalso supported in part by NICHD Grant No. P30 HD 06268.

REFERENCES

BARTSCH, C., and H. BARTSCH (1981) Effect of melato-nin on experimental tumors under different photoperi-ods and times of administration. J. Neural Transm.

52: 269-279.

BERNTON, E. W., M. S. MELTZER, and J. W. HOLADAY(1988) Suppression of macrophage activation and T-lymphocyte function in hypoprolactinemic mice. Sci-ence 239: 401-404.

BLANK, J. L. (1992) Phenotypic variation in physiologi-cal response to seasonal environments. In Mammalian

Energetics: Interdisciplinary Views of Metabolismand Reproduction, T. E. Tomasi and T. H. Horton,eds., pp. 186-212, Comstock, Ithaca, NY.

BLASK, D. E. (1984) The pineal: An oncostatic gland?In The Pineal Gland, R. J. Reiter, ed., pp. 253-284,Raven Press, New York.

BLASK, D. E., S. Cos, S. H. HILL, D. M. BURNS, A.LEMUS-WILSON, and D. S. GROSSO (1991) Melatoninaction on oncogenesis. In Role of Melatonin andPineal Peptides in Neuroimmunomodulation, F.

Fraschini and R. J. Reiter, eds., pp. 233-240, Plen-um Press, New York.

BLASK, D. E., A. M. LEMUS-WILSON, S. T. Wnsorr,and S. Cos (1992) Neurohormonal modulation of can-

cer growth by pineal melatonin. In Melatonin andthe Pineal Gland: From Basic Science to Clinical

Application, Y. Touitou, J. Arendt, and P. Pevet,eds., pp. 303-310, Elsevier, Amsterdam.

BLOM, J. M. C. (1992) Environmental Influences onImmune Function and the Susceptibility to Disease:Effects of Photoperiod and Learning. PhD disserta-tion, The Johns Hopkins University, Baltimore,MD.

BLOM, J. M. C., J. M. GERBER, and R. J. NELSON

(1994) Day length affects immune cell numbers indeer mice: Interactions with age, sex, and prenatalphotoperiod. Am. J. Physiol. 267: R596-R601.

BRADLEY, A. J., I. R. MCDONALD, and A. K. LEE

(1980) Stress and mortality in a small marsupial (Ante-chinus stuartii Macleay). Gen. Comp. Endocrinol.40: 188-200.

BRAINARD, G. C., R. L. KNOBLER, P. L. PODOLIN, M.LAVASA, and F. D. LUBLIN (1985) Neuroimmunology:Modulation of the master immune system by photo-period. Life Sci. 40: 1219-1326.

BRAINARD, G. C., M. WATSON-WHITMEYER, R. L. KNOB-LER, and F. D. LUBLIN (1988) Neuroendocrine regula-tion of immune parameters: Photoperiod control of the

248

spleen in Syrian hamsters. Ann. NY Acad. Sci. 540:704-706.

BRONSON, F. H. (1989) Mammalian Reproductive Biol-ogy, University of Chicago Press, Chicago.

BRONSON, F. H., and P. D. HEIDEMAN (1994) Seasonalregulation of seasonal reproduction in mammals. In ThePhysiology of Reproduction, 2nd ed., E. Knobil and J.D. Neill, eds., pp. 541-584, Raven Press, New York.

CARLSON, L. L., A. ZIMMERMAN, and G. R. LYNCH(1989) Geographic differences for delay of sexualmaturation in Peromyscus leucopus: Effects of photo-period, pinealectomy, and melatonin. Biol. Reprod.41: 1004-1013.

CHAMPNEY, T. H. and D. N. MCMURRAY (1991) Spleenmorphology and lympoproliferative activity in shortphotoperiod exposed hamsters. In Role of Melatoninand Pineal Peptides in Neuroimmunomodulation, F.Fraschini and R. J. Reiter, eds., pp. 219-224, Plen-um Press, New York.

CHLEHOUN, J. O., and B. N. GRAY (1987) The profileof breast cancer in Western Australia. Med. J. Aus-tral. 147: 331-334.

CHRISTIAN, J. J. (1978) Neurobehavioral endocrine reg-ulation of small mammal populations. In Populationsof Small Mammals under Natural Conditions, D. P.Snyder, ed., pp. 143-158, University of PittsburghPress, Pittsburgh.

COHEN, P., Y. WAX, and B. MODAN (1983) Seasonalityin the occurrence of breast cancer. Cancer Res. 43:

892-896.

GOLDMAN, B. D., and R. J. NELSON (1993) Melatoninand seasonality in mammals. In Melatonin: Biosyn-thesis, Physiological Effects, and Clinical Applica-tions, H. S. Hu and R. J. Reiter, eds., pp. 225-252,CRC Press, Boca Raton, FL.

GROSSMAN, C. J. (1984) Regulation of the immune sys-tem by sex steroids. Endocrine Rev. 5: 435-455.

GROSSMAN, C. J. (1985) Interactions between the go-nadal steroids and the immune system. Science

227: 257-261.

HARTMANN, D. P., J. W. HOLADAY, and E. W. BERNTON

(1989) Inhibition of lymphocyte proliferation by anti-bodies to prolactin. FASEB J. 3: 2194-2202.

HARTVEIT, F., S. THORESEN, M. TANGEM, and J. F.

HALVORSEN (1983) Variation in histology and oestro-gen receptor content in breast carcinoma related totumour size and time of presentation. Clin. Oncol.9: 233-238.

HILL, S. M., and D. E. BLASK (1988) Effects of the pinealhormone, melatonin, on the proliferation and morpho-logical characteristics of human breast cancer cells

(MCF-7) in culture. Cancer Res. 48: 6121-6126.KAUPPILA, A., A. KIVELA, A. PAKARINEN, and O. VAK-

KURI (1987) Inverse seasonal relationship betweenmelatonin and ovarian activity in humans in a regionwith a strong seasonal contrast in luminosity. J. Clin.Endocrinol. Metab. 65: 823-828.

KELLEY, K. W. (1985) Immunological consequences ofchanging environmental stimuli. In Animal Stress,

G. P. Moberg, ed., pp. 193-223, American Physio-logical Society, Bethesda, MD.

KEPPEL, G. (1982) Design and Analysis: A Researcher’sHandbook, Prentice-Hall, Englewood Cliffs, NJ.

KIRKHAM, N., D. MACHIN, D. COTION, and J. M. PIKE

(1985) Seasonality and breast cancer. Eur. J. Surg.Oncol. 11: 143-146.

LAPIN, V., and I. EBELS (1981) The role of the pinealgland in neuroendocrine control mechanisms of neo-plastic growth. J. Neural Transm. 50: 275-282.

LEE, A. K., and I. R. MCDONALD (1985) Stress andpopulation regulation in small mammals. Oxford Rev.Reprod. Biol. 7: 261-304.

LEE, J. A. (1967) Seasonal alterations and natural historyof malignant neoplasms. Prog. Clin. Cancer 3: 96-106.

LOCHMILLER, R. L., M. R. VESTEY, and S. T.MCMURRY (1994) Temporal variation in humoral andcell-mediated immune response in a Sigmodon his-pidus population. Ecology 75: 236-245.

LYNCH, G. R., and A. L. EPSTEIN (1976) Melatonininduces changes in gonads, pelage, and thermo-genic characters in the white-footed mouse, Pero-myscus leucopus. Comp. Biochem. Physiol. C53:67-68.

MAESTRONI, G. J. M., and A. CONTI (1991) Action ofmelatonin on immune system. In Role of Melatoninand Pineal Peptides in Neuroimmunomodulation, F.Fraschini and R. J. Reiter, eds., pp. 201-210, Plen-um Press, New York.

MASON, B. H., I. M. HOLDAWAY, P. R. MULLINS,R. G. KAY, and S. J. SKINNER (1985) Seasonal varia-tion in breast cancer detection: Correlation with tu-

mour progesterone receptor status. Breast Cancer Res.Treat. 5: 171-176.

MCDONALD, I. R., A. K. LEE, A. J. BRADLEY, andK. A. THAN (1981) Endocrine changes in dasyuridmarsupials with differing mortality patterns. Gen.

Comp. Endocrinol. 44: 292-301.MIHOK, S., T. LAWTON, and B. SWARTZ (1988) Fates

and movements of meadow voles (Microtus pennsyl-vanicus) following a population decline. Can. J. Zool.66: 323-328.

MOFFATT, C. A., A. C. DEVRIES, and R. J. NELSON

(1993) Winter adaptations of male deer mice and prai-rie voles that vary in reproductive responsiveness tophotoperiod. J. Biol. Rhythms 8: 221-232.

NEIDHART, M. (1989) Bromocriptine microcapsules in-hibit ornithine decarboxylase activity induced byFreund’s complete adjuvant in lymphoid tissues ofmale rats. Endocrinology 125: 2846-2852.

NEILL, J. D., and G. M. NAGY (1994) Prolactin and itscontrol. In The Physiology of Reproduction, 2nd ed. ,E. Knobil and J. D. Neill, eds., pp. 1833-1860,Raven Press, New York.

NELSON, R. J., L. L. BADURA, and B. D. GOLDMAN(1990) Mechanisms of seasonal cycles of behavior.Ann. Rev. Psychol. 41: 81-108.

PETTERBORG, L. J., and R. J. REITER (1980) Effect of

photoperiod and melatonin on testicular development

249

in the white-footed mouse, Peromyscus leucopus. J.Reprod. Fertil. 60: 209-212.

REITER, R. J. (1983) The pineal gland: An intermediarybetween the environment and the endocrine system.

Psychoneuroendocrinology 8: 31- 50.TAMARKIN, L., M. COHEN, D. ROSELLE, C. REICHERT,M. LIPPMAN, and B. CHABNER (1981) Melatonin inhi-bition and pineal enhancement of 7,12-dimethyl-benzanthracene-induced mammary tumors in the rat.

Cancer Res. 41: 4432-4436.

VRIEND, J., and J. K. LAUBER (1973) Effects of lightintensity, wavelength and quanta on gonads andspleen of the deer mice. Nature 244: 37-38.

WELSCH, C. W., and H. NAGASAWA (1977) Prolactinand murine mammary tumorigenesis: A review. Can-cer Res. 37: 951-963.

WELSCH, R. M. (1985) Host factors affecting the growthof carcinogen-induced rat mammary carcinomas: Areview and tribute to Charles Brenton Huggins. Can-cer Res. 45: 3415-3443.

WRIGHT, J., M. ALDHOUS, C. FRANEY, J. ENGLISH, andJ. ARENDT (1986) The effects of exogenous melatoninon endocrine function in man. Clin. Endocrinol.24: 375-382.

WUNDER, B. A. (1992) Morphological indicators of theenergy state of small mammals. In Mammalian Ener-

getics: Interdisciplinary Views of Metabolism and Re-production, T. E. Tomasi and T. H. Horton, eds.,pp. 83-104, Comstock, Ithaca, NY.