Interpulse growth hormone secretion in the episodic plasma profile causes the sex reversal of cytochrome P450s in senescent male rats

Abstract

Humans as well as other mammals experience an aging-related decline in drug metabolism as well as a diminution in growth hormone secretion. In the case of rats, these events are more pronounced in senescent males, whose expression of male-specific isoforms of cytochrome P450, the major drug-metabolizing enzymes and constituting ≈60-70% of the total cytochrome P450 in male rat liver, is completely suppressed, whereas female-dependent isoforms are remarkably induced to female-like levels. Overlooked in these independently reported studies is the fact that “signals” inherent in the masculine episodic and female continuous growth hormone profiles regulate expression and/or suppression of the dozen or so sex-dependent cytochrome P450 isoforms in rat liver. Whereas previous studies identified profound reductions in the pulse amplitudes of the masculine growth hormone profile as the cause for the diminished hormone secretion during aging, pulse heights are not recognized by the cytochromes as regulatory signals. Instead, we have shown that just a nominal secretion of growth hormone during the usual growth hormone-devoid interpulse period in the masculine episodic profile can explain the complete repression of male-specific CYP2C11, CYP3A2, and CYP2A2 and induction of female-dependent CYP2C12, CYP2C6, and CYP2A1 observed in senescent male rats.

Two presumably unrelated aging-induced events were first identified in humans: a decline in drug-metabolizing capacity and a reduction in growth hormone secretion. Regarding the former, it is commonly known that the elderly experience abnormally high incidences of adverse drug reactions and toxicities (1, 2), resulting, at least in part, from a decline in drug-metabolizing enzymes (3) and requiring that doses of therapeutic drugs often be reduced in older patients (4). Concerning the latter, an aging-induced reduction in growth hormone secretion is well characterized, secretion declining by ≈14% per decade from late puberty as a result of continuously decreasing pulse amplitudes (5, 6). Unfortunately, direct studies on liver drug metabolism and growth hormone regulation in elderly people are scant, and most evidence is derived from animal studies.

Although animal research has been much more thorough and invasive than studies conducted in humans, results with animals have been surprisingly consistent with findings from humans. Early investigations using laboratory animals had to limit their analyses to nonspecific drug-metabolizing enzymes whose activities are dependent on the contributions of multiple isoforms of cytochrome P450 (P450). Moreover, the importance of sex was often overlooked by using only the “preferred” sex in pharmacotoxicological studies, the male rat, or worse yet, not even recording the sex of the animal. Nevertheless, in agreement with limited human findings (4, 7-9), rodent studies observed an aging- and sex-related decline in the in vivo and in vitro metabolism of different classes of drugs and subnormal baseline concentrations of various components of the hepatic drug-metabolizing enzyme system (10-12). With the identification in the 1980s of the multiple isoforms of P450 (designated with the initial letters CYP) it became apparent that males were more susceptible to aging-induced declines in drug metabolism than were females. Depending upon strain (13), the isoforms of P450 in senescent male liver were either demasculinized, i.e., expression of male-dependent† P450 isoforms was suppressed (14), or demasculinized as well as feminized, i.e., expression of the female-dependent isoforms was induced (15, 16).

Regarding growth hormone secretion, as in elderly humans, pulse amplitudes are dramatically reduced in senescent rats of both sexes (17-20). What has been overlooked, however, is that growth hormone regulates the expression of P450 isoforms in all species examined, likely including humans (21-23). Circulating growth hormone profiles in rats as well as other species have been shown to be sexually dimorphic (21). Male rats secrete growth hormone in episodic bursts (≈200-300 ng/ml of plasma) about every 3.5 hr. Between the peaks, growth hormone levels are undetectable. In females the hormone pulses are more frequent and irregular and are of lower magnitude than those in males, whereas the interpulse concentrations of growth hormone are always measurable. Sexual differences in the circulating growth hormone profiles, and not sexual differences in growth hormone concentrations, per se, are responsible for sexual dimorphisms ranging from body growth to the expression of hepatic enzymes observed in rats (21, 24), mice (21), and likely humans (25). In this regard, rat liver contains at least a dozen sex-dependent isoforms of P450 that are regulated by the sex-dependent profiles of circulating growth hormone (21, 24, 26). In fact, suppression of each isoform of P450 is regulated by a different “signal,” or perhaps a differential sensitivity to the signal, in the sexually dimorphic growth hormone profiles. These signals may be recognized by the hepatocyte in the frequencies and/or durations of the pulse and interpulse periods. Alternatively, the hepatocyte can monitor the mean plasma concentration of the hormone (26-30). Relevant to aging studies is the finding that the actual heights of the growth hormone pulses that are so profoundly diminished during senescence have little effect on the expression of sex-dependent, constituent P450s (26-30). In other words, were the pulse amplitudes reduced even further than normally observed in the senescent rats, it could not produce the feminization and/or demasculinization of the sex-dependent isoforms of P450 observed in the old rats. In the present report, we demonstrate that nominal aging-induced abnormalities in the interpulse periods can explain all of the expressional changes in the male- and female-dependent P450 isoforms of old age.

Materials and Methods

Animals. Animals were housed in the University of Pennsylvania Laboratory Animal Resources facility, under the supervision of veterinarians certified by the American College of Laboratory Animal Medicine and were treated according to a research protocol approved by the University Institutional Animal Care and Use Committee. Fisher rats [CDF (F-344)/CrlBR)] were received from the vendor at 7 weeks of age and observed in our facility for an additional 4 weeks. At 11 weeks of age, male and female rats were outfitted with our mobile catheterization apparatus (31), designed to obtain repetitive blood samples from unrestrained, unstressed, and completely conscious animals. Some of the male rats were simultaneously implanted with osmotic minipumps (Alza, Palo Alto, CA) anchored with two stitches in the peritoneal wall (i.p.) calibrated to deliver 10 μg of human growth hormone per kg of body weight per hr or an equivalent volume of vehicle, 0.5 μl/hr. Four days after surgery, serial blood samples (12 μl) were collected every 15 min for 8 continuous hours. Growth hormone patterns were determined by using radioimmunoassays that were highly specific for either rat or human growth hormone and had sensitivities of 1-3ng/ml. Procedural details and statistical validation of the assays have been reported previously (32). Seven days after serial blood collections, the rats were decapitated, the livers were quickly removed and minced, and portions reserved for mRNA determinations were plunged into liquid nitrogen and stored at -70°C. The rest of the minced liver was used for preparation of microsomes.

The remaining rats were maintained under conventional “clean” housing conditions in filter-top cages until 24 months of age. Animals that behaved and appeared healthy (subsequently confirmed at necropsy) were similarly outfitted with our mobile catheterization apparatus and after 4 days serial blood samples were collected for growth hormone determinations. Seven days later, the rats were killed and liver was prepared as above.

RNA Analysis. Total hepatic RNA was isolated and prepared as previously described (14). Basically, 10 μg of RNA was electrophoresed under formaldehyde denaturing conditions on 1% agarose and transferred to Nytran N nylon membranes (Schleicher & Schuell). The Northern blots were probed and occasionally reprobed with ³²P-labeled oligonucleotides by using hybridization and high-stringency washing conditions. The nucleotide sequence of oligonucleotide probes for CYP2A1, -2A2, -2C6, -2C11, -2C12 (33), and -3A2 (34) have been reported. The consistency of RNA loadings between samples was confirmed by ethidium bromide staining of 18S and 28S rRNAs and was verified with an 18S oligonucleotide probe (35). The hybridized mRNA signals were quantified by scanning the autoradiographs by using an Alpha Innotech FluorChem 8800 and normalized to the 18S rRNA signals in each lane.

Western Blotting. Hepatic microsomes were prepared from individual rat livers (32) and then assayed for P450 isoforms by Western blotting. Briefly, 10 μg of microsomal protein was electrophoresed on 0.75-mm-thick SDS/polyacrylamide (7.5%) gels and electroblotted onto nitrocellulose filters. The blots were probed with monoclonal anti-rat CYP2C11 (Oxford Biomedical Research, Oxford, MI) and anti-rat CYP2C12 (kindly provided by Marika Rönnholm, Huddinge University Hospital, Huddinge, Sweden) and polyclonal anti-rat CYP2A1 and -2A2 (kindly provided by Susumu Imaoka, Osaka City University Medical School, Osaka) and anti-rat CYP3A2 (BD Gentest, Woburn, MA) and detected with an enhanced chemiluminescence kit (Amersham Pharmacia Biosciences; ref. 36). Quantification of the relative protein levels was performed with the Alpha Innotech FluorChem 8800 gel documentation system and normalized to protein signals of two control samples repeatedly run on all blots.

Statistics. The characteristics of plasma growth hormone pulses were analyzed with the aid of the CLUSTER ANALYSIS computer program (37) as we reported previously (38). All data, including those obtained from the CLUSTER ANALYSIS program, were subjected to analysis of variance, and differences were determined with t statistics and the Bonferroni procedure for multiple comparison.

Results

The established sexually dimorphic circulating growth hormone profiles in the young adult rats graphically presented in Fig. 1 are distinguished by mathematically analyzable characteristics (Table 1). In young adult males, growth hormone was released in episodic pulses every 3-3.5 hr, resulting in peaks of ≈225 ng/ml, followed by ≈2.5 hr of undetectable nadir levels (<2-3 ng/ml). In contrast, growth hormone was released in same-age females (3 months old) in a more continuous pattern. Frequent lower-amplitude pulses of the hormone resulted in peaks of about half the male heights, followed by short-lived, <60-min, troughs that were always measurable and averaged 15 ng/ml. There was no sex difference in the mean plasma concentration of the hormone. Associated with the sexually dimorphic growth hormone profiles were dramatic sex differences in expression levels of P450 isoforms. As expected, we found that only young adult males expressed hepatic CYP2C11, -2A2, and -3A2, whereas only females expressed CYP2C12. Although both sexes expressed hepatic CYP2C6 and -2A1, levels were two to three times greater in females (Fig. 1). Expression levels of mRNA were reflected by comparable protein concentrations for each isoform (Fig. 2).

Fig. 1.

Representative Northern blot analyses of CYP2C11, -2A2, -3A2, -2C12, -2C6, and -2A1 mRNAs and circulating growth hormone profiles from individual, undisturbed, catheterized young (3-month-old) and senescent (24-month-old) male and female rats. (A) Young male. (B) Senescent male. (C) Young female. (D) Senescent female. Similar findings (see Fig. 2 and Table 1) were obtained from eight additional animals in each group.

Table 1. Analysis of sexually dimorphic growth hormone profiles in young and senescent male and female rats.

				Peak			Interpeak
Sex	Age, mo	Mean conc., ng/ml	Peak interval, min	Width, min	Height, ng/ml	Area, μg/min/ml	Width, min	Nadir, ng/ml
M	3	31 ± 7	193 ± 16	43 ± 3	224 ± 38	4.8 ± 0.9	144 ± 11	ND
	24	35 ± 13	167 ± 11†	47 ± 7	158 ± 21†	4.0 ± 0.9	120 ± 9†	12.3 ± 5.1†
F	3	35 ± 6	91 ± 9	38 ± 7	128 ± 25	2.5 ± 0.6	56 ± 9	14.5 ± 4.5
	24	31 ± 8	111 ± 14†	40 ± 6	93 ± 13†	1.9 ± 0.3*	74 ± 12†	13.7 ± 2.2

Mean concentration is calculated for the entire 8-hr collection period and represents a mean of the means. Peak interval (time period between peaks), width (duration of growth hormone peaks or interpeak periods), height (amplitude of hormone peaks), area (integrated area under growth hormone peaks, concentration × duration), and nadir (mean baseline growth hormone concentration) were all calculated with the aid of the cluster analysis program for hormone pulse detection (33) from serial blood samples collected over 8 hr at 15-min intervals (27). Each value is a mean ± SD (n = 9 rats per group). ND, not detectable. ^*, P < 0.05; †, P < 0.01 compared with young rats of the same sex.

Fig. 2.

Hepatic P450 mRNA and protein isoforms from young (3-month-old) and old (24-month-old) male and female rats as well as young males continuously infused, by i.p. implanted osmotic minipumps, with either vehicle (0.5 μl/hr) or human growth hormone (hGH, 10 μg/kg per hr). Procedural details are described in Materials and Methods. ND, not detectable. Each data column is a mean ± SD with n ≥ 5 and expressed as a percentage of the mean value of the group with the highest concentration (designated 100%) of the isoform.

The sexual dimorphism in the ultradian profiles of plasma growth hormone observed in senescent (24-month-old) rats was similar, although not identical, to that seen in younger rats (Fig. 1). In female rats of both ages, growth hormone was released in a continuous pattern characterized by frequent low-amplitude pulses followed by short-lived interpulse periods (Fig. 1). However, the senescent females secreted fewer pulses with lower amplitudes (≈25%) resulting in longer trough periods than the younger females (Table 1). Although both young and old male rats secreted growth hormone in the masculine episodic profile (Fig. 1), aging caused an increased number of pulses with lower amplitudes (≈30%), resulting in significantly shorter interpulse periods (≈15%). Moreover, the nadirs that contained no measurable hormone in the younger rats consistently had growth hormone concentrations indistinguishable from that observed in females (Table 1).

Whereas the age-dependent changes in the female growth hormone profile were associated with a small decline (≈20%) in CYP2C12 (P < 0.05) and -2C6 (P < 0.01) expression (Fig. 2), the effects in the male were far more profound (Fig. 1). Male-specific CYP2C11, -2A2, and -3A2 were generally completely suppressed, female-predominant CYP2C6 and -2A1 were induced to female-like levels, and CYP2C12, normally female-specific, was induced to ≈50% of female-like levels in the livers of the senescent males. [Protein levels were in agreement with mRNAs (Fig. 2).]

As in females, growth hormone was continuously present in the circulation of the 24-month-old male rats (Table 1 and Fig. 1). Because the masculinization of hepatic P450 isoforms depends on the presence of a minimum growth hormone-devoid interpulse characteristic of the normal episodic profile (30, 39), we replicated the elevated nadirs observed in the senescent males into young males by continuously releasing human growth hormone by means of i.p. implanted osmotic minipumps. This treatment resulted in interpulse hormone concentrations of 12.3 ± 3.0 ng/ml in otherwise normal male growth hormone profiles (Fig. 3). Elevation of the nadirs alone in the young male rats dramatically suppressed male-specific CYP2C11, -2A2, and -3A2 mRNAs and proteins, and it increased expression of female-dependent CYP2C12, -2C6, and 2A1 mRNA and protein levels to those found in senescent male rats (Figs. 2 and 3).

Fig. 3.

Representative Northern blot analyses of CYP2C11, -2A2, -3A2, -2C12, -2C6, and -2A1 mRNAs and circulating growth hormone profiles (•) from individual, undisturbed, catheterized young (3-month-old) male rats continuously infused with vehicle (0.5 μl/hr) (A) or human growth hormone (10 μg/kg per hr) (B), depicted by ▴. Equivalent findings (see Table 1) were seen in at least four additional animals in each group.

Discussion

Our results are in agreement with an earlier study reporting both demasculinization and feminization of hepatic isoforms of P450 in senescent male Fischer rats (16). That is, male-specific CYP2C11, -3A2, and -2A2, constituting 60-70% of the total P450 expressed in male rat liver (24, 29), are completely suppressed, whereas female-specific CYP2C12, not observed in young males, is expressed at ≈50% of female-like levels in old male rats. Even the female-predominant isoforms CYP2C6 and -2A1 (male/female, 1:2-3) are elevated to female-like levels in senescent male rats. In contrast, old female Fischer rats exhibit a small suppression in CYP2C6 and -2C12 expression.

We also found, in agreement with previous reports (17-20), a dramatic decline of the pulse amplitudes in both the male and female growth hormone profiles of senescent Fischer rats. Although the diminished pulse heights may be the most conspicuous changes in the senescent growth hormone profiles, they have little relevance to the aging-induced changes in P450 expression observed in the male rats. Physiologic pulse amplitudes may be requisite signals maintaining male growth rates (30, 31, 39), but even a 95% reduction of the pulse heights in the episodic male growth hormone profile will not suppress normal levels of CYP2C11, -3A2, and -2A2 (29, 32). Moreover, similar or even greater decreases in the growth hormone pulse amplitudes found in the old male rats have no inductive effects on CYP2C12, -2C6, or -2A1 (29). On the other hand, normal male-like expression levels of CYP2C11, -3A2, and -2A2 are absolutely dependent on a minimum growth hormone-devoid interpulse length (30, 39). In fact, the male-specific P450s are responsive not just to the duration of the interpulse but also to the concentration of hormone secreted during this period. When the interpulse length approaches ≈2 hr or fewer, male-specific P450 expression starts to decline (29), which may have contributed to the changes observed in the isoforms of the aged male rats. More importantly, however, continuous growth hormone secretion of as little as 1 ng/ml of plasma (<10 times the amount found in senescent males) can completely suppress CYP2C11, -3A2, and -2A2 while inducing expression of CYP2C12 and -2A1 (27, 28). Accordingly, we proposed that both the demasculinization and feminization of P450 isoforms in the senescent males could be explained solely by the loss of the growth hormone-devoid interpulse by a continuous secretion of just nominal levels of the hormone.

To test our hypothesis, we continuously infused into young male rats human growth hormone at concentrations similar to that secreted in the interpulses of senescent males. Although human growth hormone is less effective than the species-specific form in rats (40), its use allowed us to monitor its circulating levels independent of the endogenous growth hormone profile. Superimposing a continuous, nominal concentration of growth hormone upon an otherwise normal male growth hormone profile eliminated the growth-hormone devoid interpulse and consequently transformed the hepatic P450 profile of a young adult male into that of an old rat.

In summary, our findings demonstrate that the loss of the growth hormone-devoid interpulse by just minimal secretory levels of the hormone can solely explain the dramatic demasculinization and feminization of the hepatic P450 isoforms in senescent male rats. In this regard, it is possible that aging induced changes in the rates and/or patterns of hypothalamic growth hormone-releasing hormone and somatostatin secretion, beginning in middle age (41), may be responsible for the changes in the growth hormone profile of senescent animals (17, 20, 42).