Human growth hormone (GH) is a pleiotropic polypeptide hormone that mediates a myriad of metabolic and growth processes (e.g., it impacts lipid, carbohydrate, and protein metabolism in nearly all body tissues) (2,3). Specific functional effects of GH on local tissues include a protein anabolic effect (i.e., increased DNA, RNA, and protein synthesis) contributing to a positive nitrogen balance, stimulation of growth and calcification of cartilage, an increased mobilization of fats and use of fats as an energy source, a decreased use of carbohydrates, and stimulation of insulin-like growth factor I release from the liver (2,3). The exercise-induced increase in GH is thought to mediate, either directly or indirectly, many of the somatotrophic events surrounding tissue remodeling (2,7).
Since the initial observation of Hunter et al. (4) in Science in 1965, it is well recognized that physical activity is a naturally occurring stimulator of GH release into the circulation. The literature is voluminous with studies that have reported that exercise of sufficient intensity and duration elicit significant increases in GH immediately after exercise (14). The insight these studies have provided concerning GH secretory dynamics has been limited because most studies have measured only the 22 kDa molecular isoform of GH.
A unique characteristic of GH that has been considered only recently with regard to exercise is the fact that GH exists as a family of molecular isoforms. In fact, Baumann (1) postulated that there are more than 100 forms of GH in circulation. The importance of possessing a greater understanding on the exercise-mediated influences on GH molecular heterogeneity resides in the fact that the GH isoforms have diverse downstream metabolic and anabolic actions in target tissues. The purpose of this review is to discuss recent work that has examined the effects of exercise on GH molecular heterogeneity as well as to provide a theoretical framework to understand how these results may represent a potential regulatory mechanism for exercise-induced modulation of GH somatotrophic influences.
GROWTH HORMONE MOLECULAR HETEROGENEITY
The hGH-N gene expresses the main pituitary molecular weight variant in the GH family, which is the 22 kDa form. Monomeric 22 kDa GH is a 191 amino acid sequence and represents approximately 21% of all circulating plasma GH. The next most prevalent form is the monomeric 20 kDa molecule, representing approximately 6% of all circulating plasma GH, formed through alternative mRNA splicing, during which amino acid residues 32 through 46 are cleaved out (see Fig. 1). Growth hormone also can undergo posttranslational modification and peripheral tissue proteolytic cleavage at its site of action to form variants and aggregates (i.e., dimers, trimers, pentamers, oligomers) and fragments that exist in the circulation. The existence of high- and low-affinity GH binding proteins, which are released from the pituitary, cleaved from the extracellular domain of the GH receptor, or both, adds further complexity to the nature and spatial arrangement of circulating GH moieties (1,3,11,12).
Growth hormone no longer should be regarded as a single hormone, but rather as a family of related polypeptides, all of which are derived from one gene (see Table 1). This molecular heterogeneity appears to have physiologic significance: the different forms have been shown to possess different biological activities (e.g., relative potency in bioassays), as well as different immunodetectabilities (1,3).
The 22 kDa GH isoform mediates growth stimulation, stimulation of insulin-like growth factor I production, nitrogen, phosphate, and sodium retention, anabolic action in many tissues, antiinsulin activity, as well as “insulin-like activity.” Monomeric 22 kDa is the reference GH against which the biological potencies of all other variants are measured. The 20 kDa form has a propensity to dimerize and forms homodimers and heterodimers with 22 kDa GH. The 20 kDa form appears to have reduced affinity for GH receptors; however, its in vivo growth-promoting, insulin-like growth factor I–generating, and lactogenic activity are similar to 22 kDa in rats. The 20 kDa form also differs in some metabolic actions from the 22 kDa (i.e., has diminished insulin-like activity) (1,3).
Large molecular weight variant oligomers may exist via covalent and noncovalent bonded complexes of homodimers, heterodimers, and binding protein. The available data suggest that oligomeric GH may possess different biological and immunoreactive abilities when compared with the monomeric GH. Table 1 lists the relative proportions of circulating GH at rest (1).
It is critical to consider the molecular heterogeneity when interpreting the meaning of GH concentrations and to understand that the proportions of GH molecules potentially can be influenced by exercise and physical activity. To date, few people have attempted to characterize GH concentrations using different assays before and after exercise. Exercise alterations in GH isoforms may represent an important regulatory step in the adaptational process by which downstream target tissues are somatogenically modulated by physical activity patterns.
MEASUREMENT OF GROWTH HORMONE
A wide variety of “immunoassays,” including both isotopic- and nonisotopic-based assays, are commercially available that can detect circulating GH concentrations. Because these immunoassays use different antibodies (monoclonal vs. polyclonal) directed at specific epitopes on the GH molecule, and because the variant molecular forms of GH may or may not have these intact epitopes physically accessible to these antibodies, very different results can be obtained for GH concentrations from a given sample. Factors affecting the proportions of various immunoreactive GH forms in the circulation include differences in binding to high- and low-affinity binding proteins, binding to GH receptors, metabolic clearance rates, likelihood of aggregation, proteolytic processing in tissues resulting in circulating fragments, as well as non–GH-released “immunoreactivity” (8,9,11,12). Other factors that can influence the absolute immunoreactivity of hormonal measurement can include matrix effects, purity of the standard used, and relative concentrations of buffer, antiserum, and tracer used (15). Recent work evaluating the impact of exercise on in vivo bioassayable versus immunoassayable GH also have shown a disparity between the two for GH concentrations (7). Thus, the selection of an appropriately relevant assay method for detecting circulating GH concentrations becomes an important decision for the exercise scientist whose interest is in obtaining the most robust and meaningful biological information.
IMMUNOFUNCTIONAL GROWTH HORMONE
In an attempt to facilitate movement toward a possible consensus for quantitative measurement of GH, Strasburger and Dattani (11) and Strasburger et al. (12) developed a unique enzyme-linked immunoabsorbent assay (ELISA) based on the molecular interaction between the hormone and its receptor, necessary for receptor dimerization and subsequent induction of signal transduction (see Fig. 2). This immunofunctional ELISA uses an anti–human growth hormone (hGH) monoclonal antibody and a biotinylated recombinant hGH binding protein (hGHBP) that bind to hGH receptor binding sites 2 and 1, respectively (see Fig. 3). Strasburger et al. termed the GH measured by this assay immunofunctional (IF) GH. Strasburger et al. reported that the IF GH ELISA had a higher correlation than a conventional polyclonal assay when compared with the in vitro NB2 bioassay, which is based on the incorporation of labeled thymidine into the NB2 cells. This lends credence to the notion that the IF GH ELISA may be more biologically relevant than other assays. One potential benefit of using the IF GH ELISA within exercise study paradigms would be in yielding greater insight into GH-mediated tissue remodeling adaptations of the physical training process.
EXERCISE AND IMMUNOFUNCTIONAL GROWTH HORMONE
Our laboratory was the first to examine the effects of exercise on IF versus immunoreactive (IR) GH. In our first study, we compared the IF versus IR GH concentrations in men and women before and after acute resistance exercise (i.e., 6 sets of 10 repetition maximum squats separated by 2-min rest periods) (9). IF GH concentrations were determined by an ELISA purchased from Diagnostics Systems Laboratories (DSL, Webster, TX), which was based on the work of Strasburger et al. (12), and IR GH concentrations were determined by a monoclonal immunoradiometric assay (IRMA) purchased from Nichols (San Juan Capistrano, CA). In this study, both men and women demonstrated similar increases for IR (men, 1.47 vs. 25.0 ng·mL−1; women, 4.0 vs. 25.4 ng·mL−1) and IF (men, 0.55 vs. 11.7 ng·mL−1; women, 1.94 vs. 10.4 ng·mL−1) GH after exercise. However, postexercise IF GH was significantly less than IR GH for both men and women. The ratio of IR to IF after exercise was approximately 2 and was similar for both men and women. The correlation between postexercise IR and IF GH was r = 0.83. This study initially suggested that approximately half of the GH isoforms measured by the Nichols IRMA released after exercise did not possess intact sites 1 and 2 required for receptor dimerization, thus suggesting biological inactivity.
Our next study considered the fact that GH is released in an episodic, pulsatile manner. IF GH was measured in 10 men who underwent two overnight blood draws with sampling every 10 min from 5:00 p.m. to 6:00 a.m. The overnight serial sampling was performed in both a control and an acute heavy resistance exercise condition (8). For the exercise condition, subjects performed a high-volume, multiset resistance exercise bout from 3:00 p.m. to 5:00 p.m. IF GH was compared with the Nichols IRMA and National Institute of Diabetes and Digestive and Kidney Diseases’ (NIDDK) polyclonal radioimmunoassay. The Pulsar peak detection system was used to evaluate the pulsatility profile characteristics of GH. Even though the results from all three immunoassays were highly correlated (correlations ranged from 0.85 to 0.95), the Nichols IRMA again yielded higher mean GH concentrations than did IF GH (3.98 vs. 1.83 ng·mL−1, respectively). The results from the Nichols IRMA also yielded higher pulse amplitudes compared with IF GH (8.0 vs. 4.63 ng·mL−1, respectively).
The consistent finding in both of these studies (8,9) was that IF GH measured from the same sample was approximately one half that of the Nichols IRMA, one of the most widely used GH assays in clinical practice in the United States. Because the IF assay purports to measure only biologically active forms of GH (i.e., only those forms of GH capable of inducing receptor dimerization are translated to an assay signal), the additional GH isoforms detected by the Nichols IRMA likely are fragments whose potential biological actions are not mediated by the GH receptor. It has been reported that the GH fragment 44 to 191 is detectable in substantial levels in human serum and may even antagonize GH action (10). Because this fragment lacks part of the N-terminus, it is unlikely to be detected by the IF assay; however, the fragment could be detected by the IRMA, depending on the targeted epitopes. Alternatively, the additional GH isoforms detected by the Nichols IRMA also could be high molecular weight variant forms of GH (1,6).
Our findings conclusively show that at least some of the molecules released during secretory bursts are able to dimerize GH receptors. In that sense, these molecules are biologically active. However, our data also demonstrated that GH isoforms are released, both in exercise and nonexercise conditions, that are not capable of initiating signal transduction through the GH receptor. Based on the high correlations among the immunoassays and the similar detection of the number of peaks and interpeak intervals, it appears that the immunoassays report qualitatively comparable pictures of the GH response. The quantitative differences among the immunoassays have yet to be explained fully, but are likely the result of the existence of various molecular isoforms. Other factors that contribute to the differences in GH measurement may include the assay equilibrium conditions, buffer, tracer, and standard used (15).
It is important to consider the impact of binding proteins (BPs) in the immunofunctional assay (IFA) (8,11). The IFA uses a recombinant hGH BP to bind site 1. One could infer that a GH molecule already complexed to a GH BP would not be detected in this assay system, because site 1 would not be freely accessible. Also, an hGH–BP complex may be configured such that site 2 is not exposed to the monoclonal antibody (mAb7B11). It has been reported that the high-affinity GH BP inhibits GH binding to receptors and in vitro bioactivity via competition for ligand (12). If it is true that GH complexed to BP is too large a molecule to traverse the capillary endothelium to bind to cellular receptors, the lack of detection of the GH complexed in the IFA provides further support for the functional selectivity of the IFA.
FURTHER CHARACTERIZATION OF EXERCISE EFFECTS ON GROWTH HORMONE MOLECULAR ISOFORMS
Wallace et al. (13) used seven different assays to measure GH in 17 aerobically trained men before and after 20 min of cycle ergometry at 80% o2 max to characterize further the response of GH molecular isoforms to exercise. Serum was assayed with antibodies specific for total, pituitary, 22 kDa, recombinant, non-22 kDa, 20 kDa, and IF GH. Salient findings from this study were (a) all forms of GH increased during and at the end of exercise, (b) 22 kDa GH was the predominant isoform (73%) at the cessation of exercise, (c) the ratios of non-22 kDa per total GH and 20 kDa per total GH increased, and those of recombinant and pituitary GH decreased. Wallace et al. (13) attributed the increase in non-22 kDa isoforms to slower disappearance rates of 20 kDa and perhaps non-22 kDa GH isoforms. Collectively, Wallace et al.’s findings demonstrate that the proportion of GH isoform changed across acute exercise and into recovery. Although the 22 kDa was the predominant isoform detected at peak concentrations, isoforms of GH other than 22 kDa increased during the postexercise period. These results suggest that the proportion of 20 kDa, 17 kDa, and possibly other non-22 kDa isoforms (dimers, oligomers, and GH bound to serum proteins) increase after exercise. The authors postulated that the increase in the proportion of isoforms other than 22 kDa after exercise may be attributed to differential pituitary isoform secretion, the appearance of isoforms from nonpituitary sources, generation of fragments, dimers, and oligomers in the circulation, and differences in clearance rates of the different isoforms. The authors also speculated that the biological consequences of their findings may reside in enhanced diabetogenic effects of smaller GH isoforms, which may prevent postexercise hypoglycemia.
Extending on Wallace et al.’s work, we next conducted a study in which we fractionated human plasma in 35 women before and after acute resistance exercise (six sets of 10 repetition maximum squats, separated by 2-min rest periods) using size exclusion chromatography into three size classes (5). Fraction A contained molecules > 60 kDa (presumably oligomers, monomeric GH bound to receptor, or both); fraction B contained molecules 30 through 60 kDa (presumably homodimers and heterodimers); and fraction C contained GH molecules < 30 kDa (presumably a mixture of 22, 20, 16, 12, and 5 kDa forms). All samples were then assayed using the Diagnostic Systems Laboratories IFA, the Nichols IRMA, and the NIDDK radioimmunoassay. Additionally, we assayed all samples before and after glutathione treatment to determine the effects of chemical reduction of disulfide linked bonds. Recovered immunoreactivities were 4% to 11% in fraction A, 22% to 45% in fraction B, and 44% to 72% in fraction C. Significant exercise-induced increases were observed for the lower molecular weight GH moieties (30–60 kDa and <30 kDa isoforms), but not for the higher molecular weight GH moieties (>60 kDa). Another important finding was that chemical reduction of the postexercise samples increased immunoassayable GH as measured by the Nichols and NIDDK assays more than preexercise samples, suggesting that exercise may increase specifically the release of disulfide-linked hormone molecules, fragments, or both. From these data, the most important effect of acute resistance exercise appears to be on dimeric hormone. Because complexes of GH and binding proteins have longer half-lives than free GH, it is possible that dimeric GH also may have a longer half-life. Therefore, the net effect of the increase in GH isoforms within this molecular weight range would be to prolong the biological activities of these forms after exercise.
SUMMARY AND PERSPECTIVE
From the few studies that have examined the effects of exercise on GH molecular heterogeneity, the concept is emerging that acute exercise, both resistance and aerobic, alters the relative proportions of circulating GH molecular isoforms (see Fig. 4). Wallace et al. (13) demonstrated that aerobic exercise lengthened the half-life of the 20 kDa and other non-22 kDa isoforms, and Hymer et al. (5) reported that disulfide-linked GH aggregates were increased preferentially by resistance exercise. Taken together, these data suggest that acute exercise may potentiate the bioactivity of GH by inducing the release of molecular isoforms with extended half-lives and thereby sustaining biological action.
Aggregation states of monomeric GH variants occur in the pituitary (1,3). These aggregates can be homodimers or heterodimers of monomeric GH or may be aggregates of monomer associated with a binding protein. The 20 kDa actually is preferentially stored as a dimer. These aggregated and nonaggregated GH molecules are both released from the pituitary and may be differentially regulated. The physiological significance and precise characterization (monomer complexed to monomer, or monomer complexed to binding protein) remain unknown. That there are different releasable pools of GH within the pituitary gland raises the question of whether exercise preferentially stimulates release from one pool over another. Future studies aimed at further separating different GH isoforms during longitudinal physical training and exploring the mechanisms and biological consequences of such responses would seem prudent.
1. Baumman, G. Growth hormone heterogeneity: genes, isohormones, variants, and binding proteins. Endocr. Rev. 12( 4): 424–449, 1991.
2. Giustina, A., and Veldhuis. J.D. Pathophysiology of the neuroregulation of growth hormone secretion in experimental animals and humans. Endocr. Rev. 19( 6): 717–797, 1998.
3. Harvey, S., Scanes, C.G.and Daughaday. W.H. Growth Hormone. Boca Raton, FL: CRC Press, 1995.
4. Hunter, W.M., Foneska, C.C.and Passmore. R. Growth hormone: important role in muscular exercise in adults. Science 150: 1051–1053, 1965.
5. Hymer, W.C., Kraemer, W.J. Nindl, B.C. Marx, J.O. Benson, D.E. Welsch, J.R. Mazzetti, S.A. Volek, J.S.and Deaver. D.R. Characteristics of circulating growth hormone in women after acute heavy resistance exercise. Am. J. Physiol. Endocrinol. Metab. 281: E878–E887, 2001.
6. Lewis, U.J., Shaw, Y.N.and Lewis. G.P. Structure and properties of members of the hGH family: a review. Endocr. J. 47: S1–S8, 2000.
7. McCall, G.E., Gosselink, K.L. Bigbee, A.J. Roy, R.R. Grindland, R.E.and Edgerton. V.R. Muscle afferent-pituitary axis: a novel pathway for modulating the secretion of a pituitary growth factor. Exerc. Sport Sci. Rev. 29( 4): 164–169, 2001.
8. Nindl, B.C., Hymer, W.C. Deaver, D.R.and Kraemer. W.J. Growth hormone pulsatility profile characteristics following acute heavy resistance exercise. J. Appl. Physiol. 91: 163–172, 2001.
9. Nindl, B.C., Kraemer, W.J.and Hymer. W.C. Immunofunctional vs. immunoreactive growth hormone responses after resistance exercise in men and women. Growth Horm. IGF Res. 10: 99–103, 2000.
10. Rowlinson, S.W., Waters, M.J. Lewis, U.J.and Barnard. R. Human growth hormone fragments 1–43 and 44–191: in vitro somatogenic activity and receptor binding characteristics in human and nonprimate systems. Endocrinology 137: 90–95, 1996.
11. Strasburger, C.J., and Dattani. M.T. New growth hormone assays: potential benefits. Acta Paediatr. Suppl. 423: 5–11, 1997.
12. Strasburger, C.J., Wu, Z. Pflaum, C.D.and Dressendorfer. R.A. Immunofunctional assay of human growth hormone (hGH) in serum: a possible consensus for quantitative hGH measurement. J. Clin. Endocrinol. Metab. 81: 2613–2620, 1996.
13. Wallace, J.D., Cuneo, R.C. Bidlingmaier, M. Lundberg, P.A. Carlsson, L. Boguszewski, C.L. Hay, J. Healy, M. Napoli, R. Dall, R. Rosen, T.and Strasburger. C.J. The response of molecular isoforms
of growth hormone to acute exercise in trained adult males. J. Clin. Endocrinol. Metab. 86: 200–206, 2001.
14. Wideman, L., Weltman, J.Y. Hartman, M.L. Veldhuis, J.D.and Weltman. A. Growth hormone release during acute and chronic aerobic and resistance exercise. Sports Med. 32( 15): 987–1004, 2002.
15. Wood, P. Growth hormone: its measurement and the need for assay harmonization. Ann. Clin. Biochem. 38: 471–482, 2001.
Keywords:©2003The Amercian College of Sports Medicine
isoforms; immunofunctional growth hormone; immunoassays