• Open Access

The Effect of Geographic Location, Breed, and Pituitary Dysfunction on Seasonal Adrenocorticotropin and α-Melanocyte-Stimulating Hormone Plasma Concentrations in Horses

Authors


  • Dr Vainio is presently affiliated with University of Helsinki Equine Hospital, Finland.

Corresponding author: D. McFarlane, Department of Physiological Sciences, Center for Veterinary Health Sciences, Oklahoma State University, Stillwater, OK 74078; e-mail: diannem@okstate.edu.

Abstract

Background: Plasma α-melanocyte-stimulating hormone (α-MSH) and adrenocorticotropin (ACTH) concentrations in horses vary with season, confounding diagnostic testing for pituitary pars intermedia dysfunction (PPID).

Hypothesis: The goals of this study were to determine whether seasonal variation in plasma α-MSH and ACTH concentrations in horses is influenced by geographic location, breed, or PPID.

Animals: Healthy light breed horses residing in Florida, Massachusetts, and Finland (n = 12 per group); healthy Morgan horses (n = 13); healthy ponies (n = 9) and horses with PPID (n = 8).

Methods: Monthly plasma α-MSH and ACTH concentrations were measured by radioimmunoassay. Nonlinear regression analysis was used to estimate the time of peak hormone concentrations. Mean hormone concentrations in fall and nonfall months were compared.

Results: The fall peak plasma α-MSH concentration occurred earlier in horses residing at more northern locations. Mean seasonal α-MSH concentrations were similar in all healthy groups at all locations, but in the fall, plasma ACTH concentrations were higher in horses living in more southern locations. Plasma ACTH but not α-MSH concentrations were higher in Morgan horses compared with light breed horses from the same location. Hormone concentrations of ponies did not differ from those of horses during either season. Concentrations of both hormones were high in the fall compared with the spring in horses with PPID.

Conclusions and Clinical Importance: These findings suggest geographic location of residence and breed may affect the onset, amplitude, or both of the seasonal peak of pars intermedia (PI) hormones and should be considered when performing diagnostic testing for PPID. Horses with PPID maintain seasonal regulation of PI hormone output.

Abbreviations:
ACTH

adrenocorticotropin

α-MSH

α-melanocyte-stimulating hormone

FIN

Finland

FL

Florida

MA

Massachusetts

PI

pars intermedia

PPID

pituitary pars intermedia dysfunction

POMC

proopiomelanocortin

Equine pituitary pars intermedia dysfunction (PPID) is a debilitating disease that affects approximately 15–25% of all aged horses, with ponies and Morgan horses affected more frequently.a,1–4 PPID is associated with several serious and life-threatening conditions, including laminitis, secondary infections, and insulin resistance. Early recognition and therapeutic intervention may help prevent horses with PPID from developing these highly debilitating complications.

The clinical signs of PPID include hirsutism, which alone may have a high diagnostic sensitivity and specificity.4 However, development of hirsutism occurs relatively late in the course of the disease, often after the development of laminitis.3 This suggests hirsutism alone is not a useful diagnostic indicator if early intervention is the goal. Therefore, specific endocrine diagnostic testing is advised in aged horses with other clinical signs compatible with PPID.

Currently, the available diagnostic strategies for equine PPID include dynamic tests assessing the pituitary-adrenal axis (eg, dexamethasone suppression test, thyrotropin-releasing hormone stimulation test, domperidone stimulation test) or measurement of endogenous plasma concentrations of pituitary hormones (adrenocorticotropin [ACTH], α-melanocyte-stimulating hormone [α-MSH]). Interpretation of all PPID diagnostic tests is complicated by seasonal variation in the activity of the pituitary-adrenal axis in the horse.5,6

Similar to several other species such as hamsters and sheep, horses have a seasonal rhythm to the regulation of the pars intermedia (PI).5,7–11 As a result, plasma concentrations of the PI hormones, including α-MSH and ACTH, are higher in the fall than in the winter or spring,5,6 an adaptation that has been suggested to prepare the animal for the metabolic and nutritional pressures of the approaching winter.9–11

Increased activity of the PI is associated with false-positive results when healthy horses and ponies are tested for PPID in the fall.6,12 Plasma concentrations of ACTH are increased with 80–97% of healthy horses and ponies having ACTH concentrations above the reference range when tested in September, despite having normal concentrations in the spring.6 Dynamic testing of the pituitary-adrenal axis also is affected with a false-positive rate of 40% in horses and 21% in ponies when the dexamethasone suppression test is performed in September compared with May.6

Several studies5,12 have suggested that ponies may have a greater seasonal PI response than horses; however, these studies did not control for potentially confounding factors such as different geographic locations of residence of the animals5 and unmatched body condition scores (BCS).12 If ponies do have a more robust seasonal PI response, this may be a result of their thrifty nature. Thrifty animals are those that have a genetic predisposition to be overweight in the absence of excessive calories, the so-called “easy-keeper.” If circannual PI hormonal rhythm is an adaptation that coordinates metabolism with season, then an exaggerated seasonal hormonal response may elicit an enhanced ability to survive the metabolic and nutritional challenges of winter. If so, ponies and other thrifty breeds, such as Morgan horses, would be expected to exhibit higher fall hormone concentrations when compared with nonthrifty, light breed horses.

It is unclear if horses with PPID maintain seasonal responsiveness of PI hormone secretion. If PPID horses have a blunted or shifted seasonal hormone peak, it would be clinically important to determine when in disease development the alteration occurs. If loss of seasonal response were an early disease manifestation, it could provide a diagnostic opportunity for early disease recognition.

Although previous studies suggested a seasonal variation in PI activity,5,6 several critical questions remain. All studies to date have examined animals residing in a very restricted geographic locations. Therefore, it is unclear if seasonal variation is geographically biased and at what time of the year the increase in PI activity begins, making it difficult to interpret PPID diagnostic test results obtained in July and August. Although the preliminary data suggests ponies may have a greater seasonal increase in plasma PI hormone concentration in the fall than light horses, it is unknown if this is a typical finding in all thrifty equids from all locations. Therefore, the objective of this study was to characterize the month-to-month fluctuation in PI hormone expression in healthy horses and to determine the role of geographic location of residence, breed, and PPID in this physiological process. We hypothesized that hormone release from the healthy equine pituitary PI is regulated by seasonal change in length of day and therefore an increase in plasma concentrations of α-MSH and ACTH would be observed starting in July, as day length begins to shorten. We also expected horses residing in latitudes farther from the equator (where the seasons are more extreme) to have a more pronounced seasonal hormone increase compared with those residing closer to the equator. Furthermore, we anticipated that ponies and Morgan horses (thrifty horses) would experience greater seasonal changes whereas animals with PPID would show blunted changes in plasma α-MSH and ACTH concentrations.

Methods

Animals

All samples were collected in accordance with IACUC guidelines following approval of Oklahoma State University and the appropriate local (Tufts Cummings School of Veterinary Medicine, University of Florida, Finland) animal care committees and with informed consent of the owners.

Adult (5–17 years) healthy light horses were recruited from the equine population of Hyvinkää, Finland (Hyvinkää Horse Hospital, 60°N latitude, n = 12), North Grafton, MA (Tufts Veterinary College, 42°N latitude, n = 12), and Gainesville, FL (University of Florida, 29°N latitude, n = 12). Horses were kept outdoors as weather permitted (see “Housing”). Health status was confirmed by history, physical examination, CBC, serum biochemistry, and normal plasma ACTH (<50 pg/mL) and α-MSH (<35 pmol/L) concentrations. Blood analysis and physical examinations for study inclusion were performed in April. Horses with abnormal health examination findings including evidence of PPID (eg, abnormal haircoat, muscle wasting), equine metabolic syndrome (eg, cresty neck, gross obesity, or laminitis), or those pregnant or lactating were excluded.

Healthy ponies (n = 9) were selected from the referral equine population of Hyvinkää Horse Hospital, Finland. Healthy Morgan horses (n = 13) were selected from the equine referral population of Tufts Cummings School of Veterinary Medicine. Selection criteria were as described for the light breed horses, although two 18-year-old and one 19-year-old Morgan horses were included. Horses with PPID (n = 8) were selected from the North Grafton, MA area between April and June, based on the presence of clinical signs, including hair coat abnormalities in all cases, and increased plasma ACTH or α-MSH concentrations (>50 pg/mL or >35 pmol/L, respectively13,14). Horses with clinical signs of PPID that had plasma ACTH and α-MSH concentrations within the normal reference range were excluded from the PPID group, but monthly blood collection was continued in these horses, and the group (designated as “clinically suspect”) was considered separately for statistical analysis (n = 5).

Housing

The horses from Florida (FL) were maintained solely on pasture throughout the study. In Massachusetts (MA), 5/12 normal light breed horses were kept outside with a run-in shed year-round. The remaining 7 horses in the normal group were housed in a barn with daily turnout of a few hours except in severe weather. Most of their stalls were close to an outside door or window. The Morgan horses all came from a show barn that did not provide turnout. Many of the stalls had windows. The PPID horses in the study were housed either in paddocks with run-in sheds or in barns with daily turnout. In Finland (FIN), the horses were turned out on green pasture during the day starting in late May to early June and stayed on pasture fulltime starting around June 10–15. The horses then were moved indoors for the night beginning around August 31 and were off pasture entirely by the end of September. For the remainder of the year the horses had access to dirt pasture, usually from 0700 to 1500 hours barring severe weather. Daily light exposure (intensity and duration) was not recorded while horses were housed indoors.

Sample Collection, Processing, and Storage

Blood (10 mL) was collected on the 15th day of the month (± 4 days) by jugular venipuncture into tubes containing EDTA as an anticoagulant. Samples were collected for all assays between 0600 and 1200 hours. Blood was transported on ice and plasma separated by centrifugation within 3 hours of collection and stored at −80°C until shipped for assay. Samples were shipped frozen on dry ice to Oklahoma State University for endocrine assays (see “Methods”). Hours of daylight (photoperiod) were recorded from the US Navy Oceanography website, http://aa.usno.navy.mil/data/docs/RS_One Year.php for each sample collection date. Mean monthly temperatures were recorded for the airport nearest to each location (Gainesville, FL; Worcester, MA; Helsinki, FI) using data from the website http://www.wunderground.com/history/

Assays

α-MSHa and ACTHb concentrations were determined by radioimmunoassays validated previously for use in horse plasma.5,13 The lower limit of detection for the α-MSH assay was 3 pmol/L and 5.7 pg/mL for the ACTH assay.

Statistics

The timing of the seasonal peak hormone concentration among healthy horses from different locations was estimated by visual inspection of a nonlinear regression analysis using log-transformed monthly plasma α-MSH data. Amplitude was calculated as the difference between peak and trough mean monthly log hormone concentration. When separated by season, the data were normally distributed (Kolmogorov-Smirnov test) and expressed as mean ± SD. Mean hormone concentration was determined for the fall (August–October) and nonfall (November–July) season for each group. Multiple group comparisons were made by analysis of variance (ANOVA) with Bonferroni's correction for multiple comparisons for post hoc analysis. Two point comparisons were made by t-test. The number of samples for each group that was greater than the reference range (ACTH>50 pg/mL, α-MSH>35 pmol/L) was compared by Fisher's exact test. For all comparisons a P < .05 was considered significant. Data analyses were performed by a statistical software package.c

Results

Animal Group Characteristics

The control group from FL consisted of 5 geldings, 2 stallions, and 5 mares with 4 Quarter horses, 7 Thoroughbreds, and 1 Warmblood, with a mean age of 12.3 ± 2.1 years. The control group from MA consisted of 7 mares and 5 geldings, with 2 Standardbreds, 1 Thoroughbred, 2 Warmbloods, 4 Quarter horse crosses, and 3 grade horses with a mean age of 11.0 ± 3.7 years. The Morgan horse group consisted of 13 geldings, with a mean age of 12.1 ± 4.8 years. The MA control horses had a slight but significantly (P= .02) greater mean BCS than the Morgan horses (5.6 ± 0.2 versus 5.3 ± 0.1, respectively). The FIN control horse group consisted of 7 geldings and 5 mares, with 2 Quarter horses, 1 Paint horse, 7 Warmbloods, and 2 Finnhorses, with a mean age of 13.8 ± 2.5 years. The pony group was comprised of 8 geldings and 1 mare, with 3 Shetland ponies, 2 Ponies of Estonia, 1 Gotland Russ, and 3 large mixed-breed ponies with a mean age of 12.2 ± 1.6 years. The ponies had a significantly (P < .05) greater mean BCS of 6.2 ± 0.2 compared with the FIN horses (5.0 ± 0.4). There was no difference in mean age among these 5 healthy groups (P= .3).

Thirteen horses were identified that had typical hair coat abnormalities of PPID and were not receiving PPID treatment. Of the 13 identified, only 8 horses were PPID positive using plasma α-MSH or ACTH plasma concentration. The PPID horses included 4 mares and 4 geldings, with 3 Arabians, 1 Tennessee Walking Horse, 1 Appaloosa, 1 Thoroughbred, 1 Morgan horse/Quarter horse cross, and 1 Selle Francais/Quarter horse cross with a mean age of 26.3 ± 6.0 years. Four horses failed to shed, 1 had hirsutism that shed slowly in the summer, and 3 had mild but generalized changes in hair coat including thicker, curlier summer coats. There were 5 horses deemed clinically suspect but that tested negative for PPID, including 2 mares and 3 geldings consisting of 3 Quarter horses, 1 Appaloosa, and 1 Arabian with a mean age of 25.0 ± 3.4 years. Of the 5 clinically suspect horses, 2 did not shed at all, 2 exhibited delayed shedding, and 1 had an excessively heavy coat.

Environmental Conditions

Mean ambient temperatures differed among the 3 locations (Table 1), with the mean temperature in FL significantly greater than in both MA and FIN (P < .001, repeated measures ANOVA) and the mean temperature in MA greater than in FIN (P < .01, repeated measures ANOVA). Day length was significantly shorter in November, December, January, and February and significantly longer in May, June, July, and August as latitude increased (FL or MA versus FIN: P < .001; FL versus MA: P < .05, Table 1).

Table 1.   Mean monthly temperature and day length.
 Gainesville, FLWorcester, MAHelsinki, Finland
Temperature (°C)Day lengthTemperature (°C)Day lengthTemperature (°C)Day length
January12.210 hours 28 minutes−7.29 hours 27 minutes−4.46 hours 42 minutes
February12.211 hours 10 minutes−2.210 hours 27 minutes−6.19 hours 12 minutes
March17.712 hours 59 minutes2.211 hours 55 minutes−2.211 hours 45 minutes
April19.413 hours 53 minutes9.413 hours 23 minutes4.414 hours 36 minutes
May24.414 hours 39 minutes12.814 hours 36 minutes1017 hours 13 minutes
June26.715 hours 2 minutes2015 hours 16 minutes13.918 hours 53 minutes
July27.215 hours 52 minutes22.814 hours 58 minutes16.118 hours 6 minutes
August26.715 hours 12 minutes2013 hours 52 minutes14.415 hours 39 minutes
September26.113 hours 20 minutes17.812 hours 28 minutes9.412 hours 53 minutes
October20.612 hours 27 minutes1011 hours 5 minutes7.810 hours 10 minutes
November14.410 hours 40 minutes3.99 hours 37 minutes2.27 hours 29 minutes
December15.510 hours 15 minutes−1.19 hours 7 minutes05 hours 52 minutes

Effect of Geographic Location of Residence on Timing and Amplitude of Seasonal Hormone Increase in Healthy Horses

There was a significant seasonal rhythm for plasma α-MSH concentration at all locations with peak concentrations occurring in the fall. ACTH rhythm was difficult to ascertain because of the relatively minimal seasonal fluctuation (Fig 1) and the presence of outliers (eg, December in MA; October in FIN). Horses in FIN had an earlier estimated fall peak of α-MSH (September 10) compared with horses in FL (September 22, Fig 1). The estimated fall peak of horses in MA (September 13) was intermediate from that of either FL or FIN (Fig 1). In addition, none of the horses from FL and only 1/12 of the horses from MA had plasma α-MSH concentrations higher than the normal reference range in July, compared with 5/12 of the horses from FIN. No attempt was made to determine timing of peak ACTH concentration because of the absence of a clear seasonal rhythm. For all locations, the time of maximal α-MSH concentration (Tmax) was September, whereas Tmax for ACTH concentration was in September in FL, October in MA, and December in FIN. There was no difference between the mean α-MSH concentration in September (Cmax) for the 3 locations (FL, 79.4 ± 36.9 pmol/L; MA, 75.1 ± 31.9 pmol/L; FIN, 69.9 ± 53.2 pmol/L). Mean ACTH concentration was greater in FL (49.7 ± 14.0 pg/mL, September) and FIN (49.2 ± 16.0 pg/mL, December) compared with MA (32.0 ± 8.0 pg/mL, October) when peak monthly concentrations were compared (P < .05). Mean α-MSH and ACTH amplitude was greatest in FL (α-MSH: FL > MA > FIN; ACTH: FL > FIN > MA).

Figure 1.

 Analysis of seasonal hormonal pattern. Nonlinear model (solid line) predicted from the means and SD of log plasma α-melanocyte-stimulating hormone (α-MSH) (A–C) and adrenocorticotropin (ACTH) (D–F) concentration in horses residing in Florida (A, D), Massachusetts (B, E), and Helsinki, Finland (C, F) in samples collected from January (1) through December (12). The vertical dashed lines mark the timing of the circannual hormonal peak, the dotted horizontal lines mark the upper reference range cut-off value. Plasma α-MSH concentration shows a greater seasonal rhythm compared with plasma ACTH concentration.

Effect of Breed and PPID on Hormone Concentrations in the Fall and Nonfall Seasons

For all groups of animals, mean plasma α-MSH concentration was higher in the fall than nonfall whereas plasma ACTH concentration was higher in the fall for horses in FL (P < .0001) and MA (controls, P < .05; Morgan horses, P < .001; Clinically suspect, P < .05; PPID, P < .01), but not FIN (horses, P= .3; ponies, P= .7). Mean α-MSH concentrations were similar in control horses in both fall and nonfall months in all 3 locations, whereas ACTH concentration was higher in the fall in FL compared with MA (P < .01) and FIN (P < .05) and higher in FIN compared with in MA (P < .05) during the nonfall months (Fig 2). Clinically suspect horses had increased α-MSH concentration compared with Morgan horses and control horses from the same location during the fall and nonfall months (Fig 3), whereas ACTH concentration increased in both the Morgan horses and clinically suspect horses in the fall. There was no difference in mean ACTH concentration during the nonfall months. In FIN, mean α-MSH and ACTH concentration in the fall (July–October) and nonfall (November–June) months were similar in horses and ponies (Fig 4). In horses with PPID both α-MSH and ACTH concentrations were significantly higher in the fall compared with nonfall months (repeated measures ANOVA, Fig 5, P < .01).

Figure 2.

 Mean ± SD plasma α-melanocyte-stimulating hormone (α-MSH) concentration (A–B) and adrenocorticotropin (ACTH) concentration (C–D) from horses residing in Florida (FL), Massachusetts (MA), and Helsinki, Finland (FIN) during the fall (A and C; August–October) and nonfall months (B and D; November–June); n = 12/group. One-way analysis of variance, *P < .05, **P < .01. FL, white bars; MA, gray bars; FIN, black bars.

Figure 3.

 Mean ± SD plasma α-melanocyte-stimulating hormone (α-MSH) concentration from healthy light breed horses (control, n = 12), Morgan horses (n = 13), and horses that were clinically suspect of having pituitary pars intermedia dysfunction (PPID) (clinical, n = 5) residing in Massachusetts during the fall (August–October, A) and nonfall months (November–July, B). Mean ± SD plasma adrenocorticotropin (ACTH) concentration from healthy light breed horses, Morgan horses, and horses that were clinically suspect of having PPID residing in Massachusetts during the fall (August–October, C) and nonfall months (November–July, D). One-way analysis of variance, *P < .05, **P < .01, ***P < .001. Control group, white bars; Morgan horses, gray bars; clinical suspect group, black bars.

Figure 4.

 Mean ± SD plasma α-melanocyte-stimulating hormone (α-MSH) concentration from healthy horses (n = 12) and ponies (n = 9) residing in Helsinki, Finland during the fall (July–October, A) and nonfall months (November–June, B). Mean ± SD plasma adrenocorticotropin (ACTH) concentration from healthy horses and ponies residing in Helsinki, Finland during the fall (July–October, C) and nonfall months (November–June, D). Horses, gray bars; ponies, black bars.

Figure 5.

 Monthly mean ± SD plasma α-melanocyte-stimulating hormone (α-MSH) (A–C) and adrenocorticotropin (ACTH) (D–F) concentration in horses with equine pituitary pars intermedia dysfunction (PPID, n = 8, A and D), horses that were clinically suspect of having PPID (clinical, n = 5, B and E), and healthy light breed horses (control, n = 12, C and F), all from Massachusetts. There was a significant increase in the plasma concentration of both hormones in fall (August–October) compared with nonfall (November–July) months (n = 8, repeated measures analysis of variance, P < .05) in PPID horses. The fall hormone concentration was greatest in the PPID group, intermediate in the clinical group and least in healthy controls.

Effect of Housing on Seasonal Hormone Increase

For the purpose of assessing the effect of housing conditions (stabled versus pastured) on seasonal hormone response, plasma α-MSH concentrations from the Morgan horses, and control horses housed indoors with daily turnout were compared with those of control horses that were maintained year-round on pasture. There was no difference in plasma α-MSH concentrations in the fall (76.4 ± 7.3 pmol/L, n = 13 versus 56.4 ± 5.2 pmol/L, n = 5; P= .12) or nonfall months (13.2 ± 1.1 pmol/L, n = 12 versus 10.6 ± 1.0 pmol/L, n = 5; P= .2) between stabled Morgan horses and pastured MA control horses, respectively. Similarly, control horses from MA that lived outside year-round experienced a decrease in plasma α-MSH concentration that did not differ from that of horses housed indoors (56.4 ± 5.2 pmol/L, n = 5 versus 54.6 ± 9.1 pmol/L, n = 7; P= .9), whereas plasma α-MSH concentration was modestly increased in indoor horses during the nonfall months compared with those housed outdoors (15.0 ± 1.4 pmol/L, n = 7 versus 10.6 ± 1.0 pmol/L, n = 5; P= .04).

Percentage of Horses with Hormone Concentrations above the Reference Range

From November through June (nonfall months) the percentage of samples from healthy animals with an ACTH concentration above reference range (>50 pg/mL) and an α-MSH concentration within reference range was higher than the number of samples that had an α-MSH above reference range (>35 pmol/L) and ACTH within the reference range (8.4 versus 1.4%; P < .001). There were no samples from healthy horses collected between November and June in which both ACTH and α-MSH were above reference range. In horses with PPID, 58% of nonfall samples (n = 60) had both ACTH and α-MSH concentrations above the normal reference range; 8.3% had only ACTH increased, and 25% had only an increase in α-MSH concentration. 8.3% of the nonfall samples from horses with PPID had both ACTH and α-MSH concentrations within the normal reference range. During the nonfall months, all samples from clinically suspect horses had normal ACTH concentrations, however, 17.5% of samples had α-MSH concentrations above the normal reference range. In the fall (P < .0001) and nonfall months (P < .001), Morgan horses were more likely to have a plasma ACTH concentration higher than the normal reference range than clinically normal horses of other breeds from the same location. α-MSH was higher than reference range more frequently in Morgan horses than in non-Morgan horses only in fall (P < .05). There was no difference in the frequency of hormone concentrations above the reference ranges when ponies and horses from FIN were compared.

Discussion

Plasma concentrations of the pituitary proopiomelanocortin (POMC) peptide hormones, including α-MSH and ACTH. recently have been shown to have a seasonal rhythm in the horse, with peak concentrations occurring in the fall.5,6 Seasonal biorhythms of hormones are typically regulated by changes in photoperiod, mediated by night time production of melatonin from the pineal gland. Circulating melatonin then coordinates seasonally regulated processes either directly or indirectly through interactions at melatonin responsive cells in the pituitary pars tuberalis. Therefore, one might expect the timing and the amplitude of the annual peak of the equine POMC hormones to vary according to where the animal resides, with those living where the change in photoperiod is the most extreme having the earliest and highest magnitude peak.

As anticipated, in the present study the fall hormone peak occurred earlier (approximately 11 days) in horses that resided farthest from the equator (FIN) compared with those nearest to the equator (FL). Surprisingly, however, horses from FL had the greatest increase in fall hormone concentrations; both the amplitude and the mean fall hormone concentration were highest in FL despite similar concentrations among horses from all 3 locations during the nonfall months. Compared with nonfall samples, plasma ACTH concentration was markedly higher in the fall in horses in FL, modestly increased in the fall in horses in MA, and not increased in animals in FIN. Therefore testing for PPID may not be affected equally by season in all locations, and it may be possible to test those animals residing farther from the equator year-round without using seasonal-specific reference ranges. Regional reference ranges will need to be established as the current study was not designed with adequate healthy animal numbers to provide these data, and reference ranges, particularly for ACTH, are assay dependent.

The biological relevance of the relationship between magnitude of seasonal response and latitude is not clear, although similar findings have been reported for luteinizing hormone response in birds.15 Birds captured from Norway (latitude 69°40′) had a low amplitude LH peak after photostimulation compared with birds captured from Northern Italy (latitude 45°26′). If seasonal cues are mediated by multiple physiological and environmental imputs, it is possible that in more temperate climates a stronger hormonal response is necessary to signal change of season. Another environmental input that may contribute to seasonal response, in addition to photoperiod, is ambient temperature. In the current study, it was not possible to separate the effect of photoperiod and ambient temperature on hormone response.

One potential confounder in the design of this study was that the horses that lived farther north were housed indoors part of the year whereas the horses from FL were maintained solely on pasture throughout the study. In MA, there were horses in the control group that lived outside year-round but there also were horses that were stabled with a few hours of turnout throughout the year. The Morgan horses were stabled year-round with little to no turnout. Although artificial light of the appropriate intensity and wavelength is fully capable of inducing a seasonal hormone response, the intensity, duration, and wavelength of light exposure in the horses housed indoors in this study were not recorded. Therefore, it was uncertain if horses housed inside were under the same influence of light as those housed outdoors. Despite having less direct access to the either sunlight or weather, the Morgan horses had similar plasma α-MSH concentrations as the control horses that lived outdoors in MA year-round, and the MA control horses that were housed indoors in the winter had a slight increase in mean plasma α-MSH concentration during the nonfall period. Therefore, it is unlikely the differences in housing resulted in a blunting of a seasonal effect on plasma hormone concentrations in this study.

In this study, α-MSH concentration appeared more predictive of the presence or absence of PPID than ACTH. Provided all horses were accurately diagnosed, a significantly higher number of normal horses had a false-positive ACTH test in the nonfall months than had a false-positive α-MSH test and more PPID horses had a false-negative ACTH test than α-MSH test. Diagnosis in this study was based heavily on clinical signs, with plasma ACTH and α-MSH concentrations used to confirm the diagnosis. In addition, diagnosis confidence was enhanced because horses were monitored for a minimum of 15 months during the study and disease development was not observed in any of the control animals. However, it is not presently possible to identify preclinical PPID horses accurately, and it is conceivable that some of the control horses might have been inaccurately assigned.

Although some previous studies have found that the magnitude of increase in plasma α-MSH and ACTH concentrations observed in the fall is higher in ponies compared with horses,5,12 reports in the literature are inconsistent and the role of breed in seasonal hormonal variation remains unclear. In a small study, no significant difference in mean plasma α-MSH concentration from horses or ponies was observed during any of the nonfall months, however, ponies had significantly higher plasma α-MSH concentrations in September (mean 96.5 ± 49.2 pmol/L, n = 13) compared with horses (mean 20.5 ± 15.2 pmol/L, n = 11).5 However, in this study, the ponies resided in Chester County, PA (latitude 39°59′) whereas the horses were from Prince Edward Island, Canada (latitude 46°20′). There was no difference in plasma ACTH concentration in these same ponies when compared with horses from the same location.6 In a second study,12 ponies had increased plasma α-MSH and ACTH concentrations in the fall compared with horses, however, the ponies had significantly higher BCS. BCS has been shown to positively correlate with plasma α-MSH concentration.12,16

Because ponies typically are more metabolically thrifty than light breed horses and seasonal increase in PI POMC hormones is proposed to serve to metabolically prepare animals for winter, we hypothesized that genetically thrifty equids might have a more robust seasonal PI response resulting in a more metabolically efficient winter state. Two groups of thrifty equids (ponies from FIN and Morgan horses from MA) were evaluated in the present study. Horses for the thrifty study groups were not selected from FL, because we anticipated the seasonal PI response may be blunted or absent in horses at this temperate latitude. In this study, ponies did not differ from horses from the same location. It is possible that all breeds of ponies are not genetically similar in hormonal response to season; the present study included only 9 ponies, 2/3 of which were Northern (Scandinavian) type, large mixed-breed ponies. Alternatively, it is possible that only ponies living in more temperate climates have high magnitude fall hormone peaks. In contrast to the ponies, Morgan horses had a greater fall ACTH peak and were more likely to have an α-MSH or ACTH concentration above reference range compared with the site-matched control group, although the difference in α-MSH concentration was far more modest than that of ACTH. In both the present and previous studies, plasma α-MSH has been shown to have a more robust seasonal rhythm.12 Therefore, it is reasonable to assume that the increase in ACTH observed in the Morgan horse group was not the result solely of change in photoperiod or thriftiness but rather a response to other environmental events. The Morgan horses in this study were from a show barn. During the show season (spring, summer, and fall) several of these horses would be transported to horse shows, housed on the show grounds, and shown in various performance events. ACTH has been shown to increase after physiological stress such as transport,17 exercise,18 and novelty stress.19 The effect of acute or chronic disease on plasma ACTH concentration has not been extensively studied in the adult horse.13 Although both ACTH and α-MSH increase with season, ACTH is primarily produced by the corticotropes of the pars distalis whereas α-MSH is a product of the melanotropes of the PI. Therefore it is reasonable to assume they respond differently to physiological and pathological events that have yet to be elucidated. Although the present study suggests that thrifty horses do not have a more robust season-driven PI hormone response, further work is needed to clarify the factors that contribute to breed-related variations in seasonal response.

One question this study sought to address was whether horses with PPID maintain a seasonal hormonal rhythm. Similar to the control horses, assignment of horses to the diseased groups was not straightforward. Admittedly, the assignment of clinically affected horses as PPID or clinically suspect was unlikely to be accurate because it was based on a single sample hormone concentration. By this method, an equal number of horses that had advanced clinical signs, such as hirsutism and failure to shed were assigned to the clinically suspect group as to the PPID group. A provocative test such as the dexamethasone suppression test may have performed better at identifying PPID in clinically suspect horses. However, the performance of the dexamethasone suppression test has not been directly compared with endogenous ACTH concentration. False positive and negative test results occur with both testing procedures and it is unclear which test is superior in early cases of disease. In addition, because of the need to collect samples on 2 consecutive days and the number of horses in this study that resided at client-owned facilities, use of the dexamethasone suppression test was not feasible because of the limited resources of the study. Despite the difficulty assigning the horses, the study design was adequate to ascertain that seasonal rhythm is intact in horses with PPID because all horses in both clinical groups showed a fall increase in α-MSH and all but 1 had an increase in ACTH concentration. The high fall concentrations of both PI hormones in clinically suspect horses suggests that an exaggerated seasonal response may be an early disease manifestation of PPID. With further characterization of normal fall reference ranges and diagnostic cut-off values, measurement of fall PI hormone concentrations may prove useful in providing a unique diagnostic approach for early recognition of PPID in clinical or preclinical cases.

Maintenance of a circannual PI hormonal rhythm in horses with PPID suggests that seasonal hormonal rhythm is not regulated by dopamine because the underlying lesion in PPID is a loss of functional dopaminergic neuronal input to the PI. Understanding the regulatory pathways that modulate seasonal activity of PI melanotropes may lead to design of a unique diagnostic method for identifying early cases of PPID.

In summary, geographic location of residence appears to affect onset and amplitude of the seasonal increase in PI hormones, with the fall increase occurring earlier in horses residing farther north and a greater amplitude in seasonal hormone variation occurring in horses farther south. Therefore, seasonal and photoperiod-specific reference ranges are needed when interpreting PI hormone concentrations in horses. Mean fall and nonfall ACTH concentrations were not different in horses residing in FIN and only slightly different in horses in MA, suggesting it may be possible to assess ACTH concentration in horses residing in far northern locations year-round without using seasonally specific reference ranges. Although the results of this study do not support a strong breed predisposition toward a more robust seasonal hormone response, these results should be interpreted cautiously because several confounding factors may have obscured a seasonal response. Horses with PPID continue to have seasonal variation in PI hormone concentrations, whereas those horses suspected of having early PPID had a more exaggerated seasonal hormone increase than did normal horses. These data indicate a more complete characterization of the seasonal pattern of hormonal secretion in development of PPID is needed because monitoring the magnitude of seasonal hormone rhythm may provide a novel aid in early and accurate PPID diagnosis.

Footnotes

a Eurodiagnostica, Alpco, NH

b MP Biomedical, Orangeburg, NY

c Graph Pad Prism version 5.01, GraphPad Software Inc, La Jolla, CA

Acknowledgment

The funding for this study was provided by the American College of Veterinary Internal Medicine Foundation. The authors thank Dr Lara Maxwell for her assistance with preparation of Figure 1.

Ancillary