INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DELAYED PUBERTY, in association with linear growth retardation, is a frequent extraintestinal manifestation of inflammatory bowel disease (IBD) in children and adolescents (7, 15, 27, 45). When Crohn's disease develops in childhood, puberty may be delayed in both girls and boys. Menarche occurred at age 16 yr or later in 73% of female patients in whom disease onset preceded puberty, and in a few patients menarche was delayed until the early 20s (15). In another study of young female patients with IBD, the mean age of the onset of puberty was 12.6 yr compared with 11.1 yr in healthy controls (7). Similarly, puberty was delayed in boys with IBD (7). Moreover, with frequent relapses or in the absence of a remission, the onset of puberty could be delayed indefinitely with potentially disastrous consequences on the pubertal growth spurt and with a reduction in final adult height. The duration of puberty and its onset may also be prolonged, particularly in patients with frequent disease relapses during this period (7).

Children with severe IBD and delay in puberty have been likened to patients with protein-calorie malnutrition in whom puberty is also delayed or prevented (31, 52). In children with Crohn's disease and growth failure, calorie intake has been documented to be only 43-82% of recommended values (3, 24, 28). Furthermore, nutritional supplementation may be associated with the onset of puberty and an increase in growth velocity (3, 24, 28). However, some patients with persistently active disease do not enter puberty, despite the provision of energy supplements, suggesting that factors other than undernutrition are implicated in the etiology of pubertal delay (7). Induction of a sustained remission is frequently associated with the onset of puberty. The most impressive results are seen after surgical removal of active disease when the first signs of puberty often occur within 1 yr of intestinal resection (7). These studies suggest that inflammatory mediators secreted by the inflamed gut may have a direct adverse influence, independent of undernutrition, on the onset and progression of puberty. In young patients with other chronic inflammatory diseases, such as cystic fibrosis, puberty may be delayed despite normal nutritional status (21). These observations support the hypothesis that inflammatory mediators may directly inhibit puberty or potentiate the effects of undernutrition.

The endocrine mechanisms responsible for diet-induced changes in reproductive function are incompletely understood. In both animals and humans, food deprivation and a reduction in body weight are associated with reduced activity of hypothalamic neurons producing gonadotropin-releasing hormone (GnRH), and hence pituitary gonadotropins, and this is thought to mediate pubertal delay (8, 9). However, in humans there is no simple model of undernutrition, and studies in underweight patients are influenced by confounding variables. For example, basal and GnRH-stimulated plasma concentrations of gonadotropins are reduced in patients with anorexia nervosa for up to 1 yr after normalization of body weight, suggesting that hypogonadism is caused by factors other than nutritional deficiency, including psychological abnormalities (23, 38). In rats, the gonadotropin response to undernutrition has usually only been assessed after periods of extreme calorie deprivation (2, 14, 40, 49), and the results of these studies may not necessarily be applicable to lesser degrees of undernutrition that occur with chronic disease. Furthermore, in peripubertal rats, gonadotropin secretion was normal but seminal vesicle weights were reduced after total food restriction for 4 days, suggesting a mechanism other than reduced gonadotropin secretion for the delay in progression through puberty (6). In patients with IBD and delayed puberty, plasma gonadotropins are reported to be normal or reduced (7).

The purpose of our study was twofold. The first goal was to determine the relative contribution of undernutrition and inflammation to the delay in the onset and progression of puberty. The second goal was to determine the effect of intestinal inflammation on the hypothalamic-pituitary-gonadal axis. In these studies, we have used an experimental model of colitis and a group of healthy animals whose food intake matches that of the colitic group, thus controlling precisely for the effects of undernutrition in the colitic group. Our results confirm that undernutrition is the major mediator of the delay in onset and progression through puberty. However, inflammation appears to potentiate the effects of undernutrition, and colitic rats had a delay in the onset and progression of puberty compared with pair-fed rats. Plasma gonadotropin concentrations were normal in colitic rats, but testosterone concentrations were reduced. The onset of puberty and the weights of secondary sex organs were restored by testosterone treatment in male rats.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All experiments were carried out as regulated procedures with home office approval under the Animal Scientific Act of 1986.

Animals and treatments. Prepubertal (age 24 days) Wistar rats bought from Charles River were divided into three groups matched for sex and weight. The groups consisted of healthy free-feeding controls (15 female, 16 male) and colitic (14 female, 16 male) and pair-fed (14 female, 16 male) rats. Animals were housed individually at an ambient temperature of 22 ± 1°C and humidity of 40-70% and were maintained under a 12:12-h light-dark cycle. Water was available ad libitum, and food, standard laboratory chow (RMI cubed), was provided according to the protocol described below. All of the rats were anesthetized by intramuscular injection of 0.1 ml of Hypnorm (0.315 mg/ml fentanyl citrate and 10 mg/ml fluanisone), and, in the colitic group, a soft lubricated plastic catheter was inserted 5 cm proximal to the anus. In healthy free-feeding controls and pair-fed groups, a catheter was passed in the rectum and removed after 1 min. In the colitic group, colitis was induced by administration, via the catheter, of 2,4,6-trinitrobenzenesulfonic acid (TNBS, 8 mg/100 g body wt; Sigma Chemical) in 40% ethanol at a final volume of 0.2 ml (36). The following two sets of controls were used: 1) healthy free-feeding controls allowed free access to food and 2) a pair-fed group comprising healthy animals whose daily food intake was matched to that of their pair in the colitic group. Body weight and 24-h food and water intake were measured daily in all animals. To maintain distal colitis of moderate severity, animals in the colitic group were retreated with TNBS if food intake increased to >80% that of healthy free-feeding controls. Body length assessed by the measurement of nose-to-tail base distance (4, 10) was measured in anesthetized animals at the induction of colitis and immediately before death. Colitis was induced at age 26 and 32 days in the female and male groups, respectively. Colitis was induced at a later date in the male group so that both females and males experienced a similar time of intestinal inflammation before the onset of puberty; preliminary studies in healthy rats showed that the onset of puberty occurred ~6 days later in male compared with female rats.

Assessment of the onset and progression through puberty. The onset and progression of puberty were assessed by physical examination and a serum hormonal profile (12, 29, 30, 32, 37, 39, 47). To determine maturation of the female reproductive tract, rats were inspected daily for vaginal opening by an independent observer. Vaginal smears from rats with vaginal openings were examined daily, and the onset of vaginal estrus and cycling was determined as described previously (32, 37). Briefly, the onset of cycling and estrus was determined from examination of vaginal smears for cellular morphology. Smears were obtained by placing a drop of distilled water in the vagina, and then the water was suctioned repeatedly and expelled from a pipette tip to collect the desquamated cells. The water drop was transferred to a microscope slide and fixed using cytology fixative. Examination of smears was performed at ×40 light microscopy, and cycling was determined by two independent observers who were blind to the treatment regimen. A normal estrous cycle was defined as the appearance of cornified cells in the smear at intervals of 5 days or less. Prolonged estrous cycles were defined as the appearance of cornified epithelium at intervals of 6 days or more. Cycling was considered to be absent if cornified epithelium failed to be detected at any time after vaginal opening.

Measurement of plasma hormone and IL-6 concentrations. Plasma concentrations of rat FSH, LH, and prolactin were measured by RIA (Amersham International, Bucks, UK) with minimum detection limits of 1.0, 0.1, and 0.7 ng/ml, respectively. Plasma concentrations of testosterone were measured in samples from male rats by ELISA (IDS, Tyne & Wear, UK) with a minimum detection limit of 0.1 ng/ml. Interleukin-6 (IL-6) concentrations were measured by ELISA using immunoaffinity-purified polyclonal antibodies from a sheep anti-rat IL-6 serum (National Institute for Biological Standards and Control, Hertfordshire, UK; see Ref 40). For each hormone, all samples were measured in duplicate in a single assay containing colitic and appropriate control samples.

Measurement of tissue MPO concentrations. Tissue concentrations of MPO were measured to assess the degree of intestinal inflammation by a minor modification (35) of the technique described by Smith and Castro (50).

Effect of testosterone treatment on the onset and progression through puberty. Male rats with TNBS-induced colitis were treated with one of the following: vehicle (0.2 ml sesame oil with 0.5% benzylalcohol/day sc; n = 13), low-dose testosterone (4-androsten-17-ol-3-one, 100 µg/100 g body wt in 0.2 ml sesame oil/day sc; n = 8; Sigma), or high-dose testosterone (250 µg/100 g body wt in 0.2 ml sesame oil/day sc; n = 8). The low-dose testosterone preparation has previously been shown to prevent the decrease in ventral prostate weight and the increase in serum LH in castrated rats (51). The following two control groups were used: healthy free-feeding controls (n = 18) and rats pair fed to the testosterone-treated colitic group (n = 16), both treated with vehicle. The onset of puberty, weights of testicles and secondary sex organs, and change in linear length were assessed as described previously. Tail-vein blood was taken at 5-day intervals for measurement of plasma FSH and testosterone concentrations. In the testosterone-treated group, the interval between testosterone administration and blood sampling was between 2 and 24 h after dosing. The severity of intestinal inflammation was assessed at the end of the experiment, as described previously.

Statistical analysis. Results are presented as means ± SD or as medians (interquartile range) for data that are not normally distributed. Differences in food intake and body weight were compared between the experimental groups by two-way ANOVA. Multiple comparisons were conducted with the Studentized range statistic and evaluated according to the Newman-Keuls procedure. Tissue concentrations of MPO, plasma hormone concentrations, and change in length were compared using one-way ANOVA coupled to a Bonferroni test for comparison of means. The Kruskal-Wallis test was used for data not normally distributed. A P value of <0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the experiment to assess the onset and progression through puberty, two female colitic and two male colitic rats died before the onset of puberty as a result of severe colitis or of anesthesia, and they and their pair-fed controls were excluded from further analysis.

Induction of colitis and its effect on food and water intake, body weight, and composition. In the colitic group, there was evidence of macroscopic inflammation with colonic wall thickening, mesenteric adhesions, and mucosal ulceration. MPO concentrations were significantly higher (P = 0.001) in the colitic group (93.2 ± 71.5 U/g wet tissue) compared with both free-feeding controls (25.1 ± 19.8) and pair-fed (14.5 ± 6.1) groups. Plasma IL-6 concentrations were significantly higher (P = 0.002) in the colitic group [280 (140-315 pg/ml), median (interquartile range)] compared with healthy free-feeding controls [45 (12)] and pair-fed groups [74 (20)].

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Twenty-four hour food intake in healthy free-feeding controls and colitic groups. Values are means + SD. Colitis was induced at day 0. By definition, daily food intake in pair-fed animals (data not shown) was exactly the same as that of their pair in the colitic group. A: 24-h food intake in female healthy free-feeding controls (n = 15) and rats with 2,4,6-trinitrobenzenesulfonic acid (TNBS) colitis (n = 12). P < 0.01. B: 24-h food intake in male healthy free-feeding controls (n = 16) and rats with TNBS colitis (n = 14). P < 0.01.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Body weight in healthy free-feeding controls and pair-fed and colitic groups. Values are means ± SD. Colitis was induced at day 0. A: female rats. Body weight was significantly less (P < 0.001) in the colitic group compared with healthy free-feeding controls at all time points after day 3 in the colitic group. Body weight was similar in colitic and pair-fed groups. B: male rats. Body weight was significantly less (P < 0.001) in the colitic group compared with healthy free-feeding controls at all time points after day 2. Body weight was similar in colitic and pair-fed groups.

Effect of TNBS colitis on linear growth. Linear growth of animals with TNBS colitis was reduced significantly compared with healthy free-feeding controls (Fig. 3). However, by comparison with pair-fed groups, it can be seen that ~60-75% of the growth impairment could be attributed to reduced food intake, with the remainder resulting from a direct effect of inflammation. In both female and male groups, the increase in body length in pair-fed groups (i.e., healthy but food restricted) was significantly less (P < 0.001) than in healthy free-feeding controls. However, there was a further reduction in linear growth in the colitic groups, and the change in length in male and female groups was 66-88% that of the respective pair-fed group.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Change in body length expressed as mm/day during the experimental period in healthy free-feeding controls, TNBS colitis, and pair-fed groups. Values are means ± SD. *P < 0.05 vs. pair fed. P < 0.001 vs. healthy free-feeding controls. P = 0.01 vs. pair fed.

Effect of TNBS colitis on the onset and progression through puberty. The onset and progression of puberty in female rats is shown in Table 1. In female rats, the onset of puberty occurred a mean of 6.2 days later (P < 0.0001) in colitic compared with healthy free-feeding controls. The onset of puberty was also significantly delayed in the pair-fed group compared with healthy controls, occurring a mean of 4.8 days after controls in the pair-fed group (P < 0.0001). The onset of puberty occurred 1.4 days later in the colitic group compared with the pair-fed group, but this just failed to reach statistical significance (P = 0.09). Similarly, food restriction in healthy animals (i.e., pair fed) had a detrimental effect on vaginal cycling. A normal cycle of 5 days or less occurred in only 42% of pair-fed rats compared with 93% of healthy free-feeding controls (P = 0.04). However, none of the colitic rats had normal cycles, with prolonged or absent cycles occurring in all rats.

Effect of TNBS colitis on the weight of the testes and accessory sex organs. Weights of the testes and accessory sex organs are shown in Table 2. Seminal vesicle and ventral prostate weights were markedly reduced in the colitic group (P < 0.001) to only 35 and 42%, respectively, of healthy free-feeding controls. The weights of the accessory sex organs were also reduced in the pair-fed group, but to a lesser degree: organ weights were 63% (P = 0.02) and 74% (P = 0.03) compared with free-feeding controls. Seminal vesicle and ventral prostate weights were reduced in the colitic group compared with pair-fed rats (P = 0.005). Testicular weight, however, was similar in the three experimental groups.

Effect of TNBS colitis on plasma concentrations of 17-estradiol, gonadotropins, prolactin, and androgens. Plasma concentrations of gonadotropins and prolactin in female rats are shown in Fig. 4. Plasma concentrations of LH were below and above the assay limits at age 33 and 47 days, respectively, and are not represented in Fig. 4. In healthy female rats, plasma concentrations of FSH, LH, and prolactin were increased after vaginal opening and the onset of estrus (at 35.6 ± 1.4 days) compared with concentrations in samples taken before vaginal opening, i.e., at age 33 days. In colitic and pair-fed groups, although the onset and progression of puberty were delayed, the profile of plasma hormone concentrations was similar to that seen in healthy free-feeding controls. Sufficient blood was also obtained at age 33 and 45 days after induction of colitis for measurement of plasma concentrations of 17-estradiol. Plasma concentrations were lower in the colitic group (5.9 ± 2.3 pg/ml) compared with healthy controls (8.3 ± 3.7 pg/ml) at age 33 days, but this just failed to reach statistical significance (P = 0.08). Plasma 17-estradiol concentrations were similar at 45 days in healthy free-feeding controls (10.0 ± 2.0 pg/ml) and colitic (10.2 ± 3.8 pg/ml) and pair-fed (12.7 ± 2.7 pg/ml) groups. Plasma concentrations of testosterone in male rats were significantly (P < 0.02) reduced at age 38 and 43 days in rats with TNBS colitis to 40 and 55%, respectively, of control values. However, concentrations were similar in colitic and pair-fed groups (Fig. 5). Similar to the females, plasma concentrations of FSH in the males were similar in healthy controls and colitic and pair-fed groups (data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Plasma concentrations of follicle-stimulating hormone (FSH; A), luteinizing hormone (LH; B), and prolactin (C) in female rats. Values are means ± SD. Data are shown for healthy free-feeding controls, pair-fed rats, and colitic rats at time points after induction of colitis at age 26 days.

View larger version (36K): [in this window] [in a new window]   Fig. 5.   Plasma concentrations of testosterone in healthy free-feeding controls and pair-fed and colitic groups treated with vehicle, 100 µg · 100 mg body wt1 · day1 testosterone (T), or 250 µg · 100 mg body wt1 · day1 testosterone. Values are means ± SD. Data are shown for healthy free-feeding controls, pair-fed rats, and colitic groups at intervals after induction of colitis at age 26 days. *P = 0.02 vs. healthy free-feeding controls.

Effect of testosterone treatment on the onset and progression through puberty and linear length. The onset of puberty in the testosterone-treated colitic group (100 µg · 100 mg body wt¹ · day¹) occurred at 42.8 ± 2.5 days of age and was not significantly different from the onset in healthy free-feeding controls (41.9 ± 2.9 days old). Testosterone treatment (100 µg/100 g body wt) in the colitic group also restored seminal vesicle weight (162 ± 76 mg, P = 0.5) and ventral prostate weight (107 ± 32 mg, P = 0.1) to that of healthy free-feeding controls (seminal vesicle, 189.4 ± 71.4 mg; prostate, 134 ± 35 mg). There was no further increase in organ weight with the higher dose of testosterone (167 ± 28 mg; 105 ± 21 mg). Plasma concentrations in the testosterone-treated colitic group were three- to fourfold those of healthy free-feeding controls (Fig. 5). There was no correlation between the interval from testosterone administration to blood sampling and the plasma testosterone concentration, suggesting that this mode of administration resulted in steady-state levels and did not induce peaks and troughs.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The aim of our study was to examine the relative contribution of reduced food intake and inflammation to the delay in the onset and progression of puberty in the TNBS model of colitis. We have used a group of healthy rats whose daily food intake was exactly matched to that of pair-fed rats in the colitic group; thus, we have controlled precisely for the effects of undernutrition on the onset and progression of puberty in the colitic group. In the colitic group, this has allowed us to separate the effects of undernutrition (occurring equally in colitic and pair-fed groups) from inflammation (occurring only in the colitic group) on pubertal development. The systemic effects of intestinal inflammation were evident in the colitic group from the observations of hypophagia, reduced linear growth, and increased plasma concentrations of IL-6. We and others (4, 34) have previously shown that both hypophagia and reduced linear growth in TNBS colitis are mediated, at least in part, by a systemic action of proinflammatory cytokines. Food intake in the colitic and pair-fed groups was the same; in the colitic group, reduced feeding was the result of anorexia, and, in the pair-fed group, it was the result of restricted feeding. Reduced weight gain was similar in the two groups, and this suggests that, similar to human IBD (5, 11), weight loss in TNBS colitis results entirely from reduced food intake and not from an increase in energy requirements. An increase in the metabolic rate or malabsorption of nutrients would be expected to result in reduced weight in the colitic group compared with the pair-fed group. However, our results show that this is not the case. As a general principle, malabsorption of energy substrates is not a feature of colitis in humans (43, 44). Body composition, however, was different in colitic and pair-fed groups. There was a similar reduction in fat content of both colitic and pair-fed groups compared with free-feeding controls. Lean body mass was reduced in the colitic group compared with pair-fed rats, and this is consistent with the known effects of proinflammatory cytokines on protein metabolism (19). Presumably, energy is diverted toward carbohydrate synthesis in the colitic group because body weight and fat content are similar to pair-fed rats.

We have shown that vaginal opening, as an indicator of puberty onset, is delayed in rats with TNBS colitis compared with healthy free-feeding controls. Vaginal opening was also delayed, but to a lesser degree, in healthy food-restricted animals (i.e., pair-fed rats). Furthermore, in rats with TNBS colitis, estrous cycles were disrupted severely with prolonged or even absent estrous cycles in all rats. However, this pattern was only seen in 62% of pair-fed rats. Vaginal opening, vaginal estrus, and cyclicity have been used in the rat to characterize the onset and progression of puberty (32, 37). These parameters have been shown to correlate with changes in plasma estradiol concentrations. Estradiol concentrations peak 1 day before vaginal opening, decrease thereafter, and increase again during vaginal estrus (46). Thus, based on our findings in female rats, we conclude that there is a delay in the onset and progression though puberty in both colitic and pair-fed groups compared with healthy free-feeding controls. This was not an unexpected finding and demonstrates the previously well-documented adverse influence of undernutrition on pubertal development (16, 25, 31, 52). However, vaginal opening was delayed and estrous cycling disrupted in the colitic group compared with pair-fed rats. These results suggest that inflammatory mediators (that are only present in the colitic group) potentiate the effects of undernutrition and result in a further delay in the onset and progression through puberty. The difference between the onset and progression of puberty in colitic and pair-fed female rats just failed to reach statistical significance. This may suggest that delayed puberty is related largely to undernutrition and that inflammation plays a relatively minor role. However, it may also be related to the reduced sample size as a result of increased mortality in the colitic group, occurring as a direct complication of severe colitis. Therefore, those with severe colitis, whom we would expect to have the greatest derangement in puberty onset and progression, did not enter puberty during the time course of this experiment.

In male rats, the onset of puberty was determined by the time of full separation of the prepuce from the glans penis. The progression of puberty was assessed from the weights of the testes and accessory sex organs. In male rats, undernutrition had a significant detrimental effect on pubertal development, resulting in a delay in the onset of puberty and reduced weight of the ventral prostate and seminal vesicles compared with healthy free-feeding controls. Reduced weights of the accessory sex organs were not merely the result of a global reduction in organ weights, since this difference persisted when weights were compared with liver weight. However, in accordance with previous reports, we found that testicular weight was preserved in underweight rats (14, 20). In rats with TNBS colitis, the onset of puberty was also delayed and the weights of accessory sex organs were reduced compared with healthy free-feeding controls. However, this was not merely a manifestation of undernutrition, since clear differences were apparent between the colitic and pair-fed groups. Thus, in both male and female colitic rats, inflammation per se appears to have a detrimental effect on pubertal development and either directly inhibits puberty or potentiates the effects of undernutrition. Our conclusions regarding a possible direct adverse effect of inflammatory mediators on puberty are supported by observations in female patients with cystic fibrosis who have delayed puberty despite a normal nutritional status (21).

The adverse effects of undernutrition on puberty are thought to be due, at least in part, to a reduced secretion of hypothalamic GnRH and consequently a fall in secretion of gonadotropins (FSH and LH), which normally promote gonadal development (9, 48). We expected that the delay in pubertal development in pair-fed rats would be associated with reduced plasma concentrations of LH and FSH. Prolactin concentrations were also measured because hyperprolactinemia may delay the onset of puberty. In pair-fed rats, we found that plasma concentrations of gonadrotropins and prolactin were maintained within the control range. The degree of undernutrition may explain the discrepancy in gonadotropin concentrations between our study and previous work. In the present study, pair-fed and colitic rats weighed 67-75% that of healthy free-feeding controls at the end of the study period. This compares with relative body weights of 33% (14), calorie restriction to only 12% of controls (49), or periods of complete food restriction for 48 h (2, 40) in previous studies. The relatively mild degree of undernutrition in our study, compared with previous studies (2, 14, 40), may have more relevance to the situation seen in patients with chronic inflammatory disease. Thus the effects of severe undernutrition on pubertal development did not supervene, and this has allowed us to separate the effects of undernutrition from inflammation. We also found that puberty was delayed further in the colitic group relative to pair-fed rats, but plasma concentrations of gonadotropins were similar to pair-fed rats. These results suggest that the inflammation-mediated delay in pubertal development is not by inhibition of gonadotropin production. An alternative explanation is that blood sampling was too infrequent to detect a "peak" in gonadotropin secretion that was different in colitic and healthy control groups. However, at the time of these samples, plasma testosterone concentrations were low in the colitic group, and we may have expected some reduction in gonadotropins.

Plasma concentrations of testosterone were reduced in pair-fed and colitic groups at age 38 and 43 days relative to healthy free-feeding controls. Undernutrition, however, may also have direct effects on the reproductive tract. The accessory sex organs of underfed bull calves showed decreased responsiveness to the administration of testosterone, and underweight boys with human immunodeficiency virus infection were found to have normal serum testosterone concentrations but delayed puberty (17, 33). Thus we also determined the effect of exogenous administration of testosterone on the onset of puberty and the weights of accessory sex organs in rats with TNBS colitis. This treatment completely restored the weights of the seminal vesicles and ventral prostate to control values. The onset of puberty was also similar to that of healthy free-feeding controls.

We have considered that the abnormality in estrous cycling and the weights of the accessory sex organs may be influenced by the close proximity of the reproductive tract to the site of inflammation rather than a systemic effect of inflammation. This alternative explanation, however, seems unlikely in the male group because prostatic inflammation is reported to promote prostatic hyperplasia (26). In females, we counted the number of days between the appearance of cornified epithelium in vaginal smears as an indicator of estrous cycling. Unlike the appearance of neutrophils, this is unlikely to be influenced by local inflammation.

In children with IBD and in the absence of a pair-fed group, it is impossible to accurately determine the effect of inflammation on the onset and progression of puberty. However, published work (13, 31, 53) suggests that the TNBS colitis model, even in the early stages of evolution, is similar to human Crohn's disease, particularly with respect to T cell activation and the cytokine profile. Thus we feel that the results of our study can be extrapolated to human IBD in a meaningful way.

In summary, inflammatory mediators adversely influence the onset and progression of puberty, or at least potentiate the effects of undernutrition in a model of colitis. Plasma concentrations of gonadotropins are normal in pair-fed and colitic rats, suggesting that, similar to the effects of inflammation on linear growth (4), the delay in puberty does not seem to involve inhibition at either the hypothalamic or pituitary level. However, the reduction in linear growth in the colitic group compared with pair-fed was more apparent than the delay in onset and progression of puberty, suggesting a greater influence of inflammatory mediators on the growth axis than the gonadal axis and reproductive tract. The reduction in plasma concentrations of testosterone in pair-fed and colitic groups suggests that inadequate production of androgens may contribute to the delay in puberty. Furthermore, in colitic rats, testosterone treatment normalized the onset and progression of puberty. The increase in plasma testosterone concentrations in treated rats is less than that seen in adolescents who are treated with testosterone for constitutional delay in puberty (2, 22). Thus we conclude that testosterone, at doses similar to those used for constitutional delay, may be useful to induce puberty in young patients with IBD. Furthermore, testosterone was highly efficacious, even in the presence of active inflammation and undernutrition.