B. Modulation of growth by systemic mechanisms

One possible explanation for coordinated growth of different organs is that body growth is orchestrated by a hormonal or other systemic mechanism. In fact, a hormonal mechanism plays a significant role in growth deceleration of insects. When the larvae reach a “critical size,” juvenile hormone levels fall and ecdysone levels rise, which signals juvenile growth to stop and metamorphosis to commence (40). Might GH play an analogous role in mammalian growth deceleration? Several observations appear to support this hypothesis. Certainly, GH is a potent regulator of postnatal growth and can affect final body size. Mice lacking functional GH receptors reach an adult body mass approximately 60% that of wild-type mice, and mice lacking IGF-I reach an adult body mass approximately 30% of normal (34). Conversely, overexpression of GH and IGF-I in mice leads to an increase in body size (41, 42). Analogous effects on body size are seen in humans with decreased GH, increased GH, or decreased IGF-I. GH deficiency affects primarily postnatal growth, whereas IGF-I deficiency affects both pre- and postnatal somatic growth (43). Interestingly, size differences among different breeds of domestic dogs are associated with different IGF-I alleles (44) and with serum IGF-I levels (45). Furthermore, the normal rapid proliferation rate in early life is temporally associated with a high level of GH. Circulating GH levels in the human, ovine, and rodent fetus are substantially higher than in the adult organism (46). These levels then decline sharply during the neonatal period (46–48). Similarly, the human pubertal growth spurt is driven in large part by an increase in serum GH levels.

However, multiple lines of evidence suggest that growth deceleration is not primarily driven by declining GH levels. First, the elevated fetal levels of GH are not required for the normal rapid pace of fetal growth, which is maintained even in the absence of functional GH receptors in both mice and humans (34, 49). Furthermore, in these mice and humans lacking functional GH receptors, overall postnatal growth is diminished, but the normal pattern of deceleration is maintained (34, 49). Second, in GH-deficient humans, administration of a fixed dose of GH per kilogram of body weight produces a growth curve similar to that of normal individuals, including the normal decline and cessation in growth. Third, overexpression of GH or IGF-I in transgenic mice leads to a substantial increase in body and organ size, but postnatal growth still slows progressively (42, 50, 51). Fourth, the decline in GH levels during early infancy is not accompanied by declines in circulating IGF-I or of IGF binding protein-3 (IGFBP-3) (47, 52), both of which reflect GH action on the liver. Indeed, IGF-I, which mediates most of the growth-promoting effect of GH (34), generally increases in concentration from infancy through early adolescence, as growth is slowing (53) (Fig. 1). This increase in IGF-I is coupled to an increase in IGFBP-3 (47, 52), which stabilizes circulating IGF-I but may also decrease IGF-I bioavailability. However, free IGF-I rises with age (52) (Fig. 1), suggesting that bioavailable IGF-I also likely rises and thus could not account for a declining growth rate. Fifth, both circulating GH and IGF-I concentrations are quite elevated during mid to late adolescence, as growth slows and the individual approaches adult size (47, 52, 54, 55). Serum GH and IGF-I levels do reach a peak during midpuberty and then decline (55, 56), but even when body growth is ceasing, levels are still considerably higher than levels during the rapid growth of earlier childhood.

These findings argue that declining GH and IGF-I, acting in an endocrine manner, do not play major roles in growth deceleration. However, none of these arguments exclude the possibility that IGF-I, acting in a paracrine fashion, plays a role in growth deceleration. Indeed, IGF-I is expressed by many tissues under complex regulation (including, but not limited to, GH) and in mice, IGF-I promotes growth through both endocrine actions (57) and paracrine actions (58). Thus, the role of paracrine IGF-I in growth deceleration remains to be determined.

Nutritional intake is another systemic factor that strongly modulates growth but does not appear to be the predominant mechanism driving normal growth deceleration. Although undernutrition inhibits juvenile growth (29, 59–62), overnutrition does not prevent growth deceleration. In rats (63, 64), pigs (65), and chickens (66, 67), forced overfeeding accelerates weight gain. However, with increasing age, more of the weight gain is attributed to fat deposition rather than protein content (63, 64). Similarly, in transgenic animal models that exhibit hyperphagia and obesity, such as leptin-deficient mice (68) and melanocortin-4 receptor (Mc4r)-deficient mice (69), the majority of weight difference is due to increased fat mass, rather than changes in lean mass or visceral organ mass (69, 70).

A systemic factor, estrogen, does plays a role in growth deceleration, specifically in the growth plate. In humans, the 20-fold decline in linear growth rate that occurs between fetal life and preadolescence appears to be largely independent of sex steroids. However, with puberty, in both males and females, estrogen levels increase, causing linear growth acceleration (due in part to increased GH secretion). Soon afterward, the decline in growth resumes, so that the growth rate approaches zero at approximately 15 yr of age in girls and 18 yr of age in boys. Once linear growth has essentially ceased, the inactive growth plates are converted into bone (71), a process called epiphyseal fusion. In the absence of estrogen effect, this latter phase of linear growth deceleration occurs more gradually. In a man with mutations in estrogen receptor-α, and thus estrogen resistance, persistent growth plate function was observed well into adulthood (72). Importantly, however, in this individual, linear growth did progressively slow and eventually cease. At 30 yr of age, the growth rate had fallen to approximately 0.5 cm/yr (73), and at 35 yr of age, ossification of the growth plates finally occurred (74). Similar slowly progressive declines in linear growth rate have been observed in men with estrogen deficiency due to aromatase mutations (75) and, indeed, in normal rodents whose growth plates appear not to respond to estrogen in a manner similar to humans (76). Studies in rabbits indicate that the normal developmental program of growth plate senescence, which includes a decline in chondrocyte proliferation rate, is accelerated by estrogen, causing earlier exhaustion of the proliferative capacity and consequently earlier fusion (77). Thus, estrogen promotes growth deceleration in the growth plate but is not required for this process.

C. Progressive suppression of growth by local mechanisms

Might some unknown systemic factor other than GH, IGF, or nutritional status drive postnatal growth deceleration? Transplantation experiments between animals of different ages argue against this conjecture. When juvenile organs, including intestine (78, 79), kidney (80), heart (81), thymus (82, 83), spleen (84), and the growth plate (85, 86) are transplanted into adult recipients, these organs continue to grow rapidly as if they were still situated in a juvenile host, and then later show growth deceleration when approaching their adult sizes (78, 80, 82, 84), suggesting that growth deceleration is an intrinsic property of the organ rather than the systemic environment. In contrast, parabiosis between young (2–3 months old) and old (19–26 months old) mice restores the proliferation and regenerative capacity of aged muscle satellite cells and hepatocytes in the older mice (87). However, it is unclear how these studies on regenerative capacity of adult cell proliferation apply to juvenile growth.

In summary, existing evidence does not exclude the possibility that systemic mechanisms contribute to physiological growth deceleration but does indicate that local mechanisms play a central role in this process. Thus, growth deceleration appears to be primarily an intrinsic property of growing organs. It remains possible that such local mechanisms may depend on a systemic signal. For example, although circulating GH and IGF-I concentrations do not decline substantially during human growth deceleration, the local downstream effectors of GH or IGF-I at the organ level, such as receptors and the associated signaling pathways, could be involved. Similarly, declining nutritional intake does not appear to be a principal mechanism driving growth deceleration, but it is possible that local, downstream mediators of this nutritional effect might be involved. For instance, nutritional availability regulates cell growth in part through the mammalian target of rapamycin (mTOR) signaling pathway at the molecular level (88–90). Activation of mTOR kinase under high nutrient/energy conditions leads to the phosphorylation of two major targets, S6K and 4E-BP (91). Phosphorylated S6K in turn acts on a variety of effectors (92) to promote cell growth, cell-cycle progression, and protein synthesis (93–96). Phosphorylated 4E-BP is released from the translation initiation factor eIF4E, allowing its binding with 5′-capped mRNAs to initiate translation (90, 97, 98). Conversely, nutrient deprivation leads to suppression of mTOR kinase activity, which in turn causes inhibition of protein synthesis and cell growth (90, 97, 98). It is possible that the mTOR kinase activity is gradually suppressed in growing organs, thus contributing to growth deceleration. This hypothesis has been tested recently by injecting rapamycin into fetal (embryonic d 19) and postnatal (10-d-old) mice and examining the effects on proliferation (99). Surprising, cellular proliferation in tissues from both ages was resistant to rapamycin treatment and thus occurs independently of mTOR/S6K signaling (99). These findings therefore argue against a role of mTOR signaling in regulating normal organ growth with age.

Although the mechanisms responsible for limiting organ growth appear to be local, rather than systemic, they may not be cell autonomous but rather may involve local interactions among cells, such as paracrine signals. In muscle, experiments involving coculture of myofiber explants with satellite cells suggest that the proliferative capacity of satellite cells depends on the age of the myofiber rather than the satellite cells themselves, suggesting a local, niche-dependent mechanism (100). Similarly, resting zone chondrocytes, which are thought to act as progenitor cells for the growth plate (16), show marked declines in cell proliferation in vivo (17), but when these cells are placed in cell culture, their proliferative capacity is independent of the age of the donor animal (101), suggesting that the decline in proliferative activity is not cell-autonomous but depends on interaction of the cell with neighboring cells, extracellular matrix, or other extracellular factors.

A. Mechanisms that assess time: biological clocks

The existence of a biological clock would allow the body to grow for a defined period of time before slowing, thus achieving a certain size. Theoretically, under this mechanism, if growth is transiently inhibited, the final body size would be reduced. Such timing mechanisms may be employed during embryonic development in certain cell types. Embryonic rat cardiac myocytes cultured under different conditions tend to stop dividing at a set time, despite having undergone different numbers of cell division (114). Similar studies in oligodendrocyte precursor cells in the central nervous system (115–117) suggest that time, rather than the number of cell divisions, signals the start of differentiation, indicating the presence of a cell-intrinsic timing mechanism. In particular, cyclin-dependent kinase (Cdk) inhibitor p27/Kip1 has been proposed to act as the timing mechanism because it accumulates progressively with time in cultured oligodendrocyte precursor cells (117). However, when these cells are isolated from p27-deficient mice, they only undergo one or two more cell divisions than wild-type cells before differentiation (118), suggesting that p27/Kip1 is only part of a more complex regulatory mechanism that consists of multiple components with overlapping functions, such that defects in one component do not totally abolish the timing function (119).

B. Catch-up growth: evidence for mechanisms that assess growth

An alternative mechanism for organs and organisms to grow to a reproducible size is by gathering information about growth itself. It is perhaps best demonstrated by the phenomenon of catch-up growth. In humans and other mammals, juvenile growth is inhibited by conditions such as GH deficiency, hypothyroidism, or malnutrition. If such conditions transiently inhibit growth but then resolve, the growth rate generally does not just return to normal but rather exceeds the normal rate for chronological age (120). Even when postnatal growth is delayed for more than 1 yr in rats fed a tryptophan-deficient (Trp⁻) diet (121) or a corn gluten-based diet (122), switching to a replete diet induces marked catch-up growth, although the animals are well into adulthood. As ingeniously suggested in 1914, this finding suggests that the loss of growth capacity results from the execution of growth itself, rather than increasing age (122).

The mechanisms responsible for catch-up growth have been best studied in the growth plate. Previously, the prevailing explanation for catch-up growth was that it results from a mechanism within the central nervous system that compares the actual body size to an age-appropriate set-point and then adjusts the growth rate accordingly. Thus, if a child is too small for age, the mechanism would sense this deficit and turn up the growth rate, causing catch-up growth (123). However, subsequent studies showed that local growth inhibition, in a single growth plate, is followed by local catch-up growth within that growth plate. This anatomic specificity suggests that the mechanism responsible for catch-up resides not in the central nervous system but rather within the growth plates themselves (124). Studies in rats, rabbits, and humans, after various growth-inhibiting conditions, suggest that this local catch-up growth occurs because growth-inhibiting conditions slow the normal loss of proliferative capacity in the growth plate (125–128). Thus, the loss of growth capacity in the skeletal growth plate is driven, not simply by time, but rather by growth itself.

As discussed above, current evidence suggests that the normal process of growth deceleration is driven, at a physiological level, by growth itself and, at a molecular level, by a multiorgan genetic program. Thus, one attractive hypothesis is that growth drives the juvenile multiorgan genetic program, which in turn drives growth deceleration. This hypothesis predicts that experimentally slowing growth by imposing growth-inhibiting conditions would slow the progression of the genetic program and thereby conserve future growth potential. This prediction has recently been tested by slowing growth experimentally in newborn rats for 4 wk using a Trp⁻ diet that suppresses food intake and therefore induces malnutrition (113). At 5 wk of age, 1 wk after release from the Trp⁻ diet, these rats showed increased proliferation in kidney, lung, and liver compared with control rats that had received a replete diet, consistent with the concept that growth-inhibiting conditions slow the loss of growth capacity (Fig. 3). In addition, expression microarray analysis demonstrated that the putative growth-limiting genetic program had been delayed; for a large set of growth-promoting genes that are normally down-regulated with age, the expression levels in rats with prior tryptophan deficiency were higher than in control rats of the same age (113) (Fig. 3). These findings are further supported by a previous study suggesting that hypothyroidism-induced growth retardation also delays down-regulation of growth-promoting genes, although far fewer genes were studied (111). These findings therefore suggest that the putative growth-limiting genetic program is not simply a function of age, but rather depends on the extent of organ growth undergone.

Open in a separate window

Fig. 3.

Transient inhibition of growth slows the multiorgan juvenile genetic program and also slows the normal loss of growth capacity. Left column represent conceptual diagrams, whereas middle and right columns represent experimental data. During normal mammalian juvenile growth (black curves), multiple growth-promoting genes show declining expression (top left panel). This decline is associated with declining proliferation rate (middle left panel), and therefore plateauing of body mass (bottom left panel). The gray curves in these panels depict the hypothesis that transient growth inhibition (gray box just above x-axis) would delay the normal decline in expression of growth-promoting genes and consequently delay the decline in proliferation rate and thus the increase in body mass. The middle and right columns show data from newborn rats that were placed on a Trp⁻ diet for 4 wk (gray bars) to restrict growth (113). In kidney and lung, transient tryptophan deficiency delayed the decline in expression of many genes, including Igf2 (top row), delayed the decline in proliferation (middle row), and therefore the gain in organ mass (bottom row). [Middle and right panels of top two rows reproduced from J. C. Lui et al.: FASEB J 24:3038–3092, 2010 (113).]

Thus, current findings suggest that mammalian body growth is limited, at least in part, by a negative feedback loop (Fig. 4). Growth causes the down-regulation of a large set of growth-promoting genes, which in turn causes body growth to progressively slow and eventually cease, thus setting a fundamental limit on adult body size. This negative feedback appears to act in multiple organs, providing a possible explanation for how growth slows concordantly throughout the body, thus maintaining body proportions.

Open in a separate window

Fig. 4.

Model for a mechanism that restricts juvenile body growth. In early life, multiple growth-promoting genes are well expressed, leading to rapid growth. However, growth causes down-regulation of these growth-promoting genes (perhaps through epigenetic mechanisms) which causes growth to slow. Progression of this negative feedback loop would eventually cause the growth rate to approach zero.

As discussed above, organs appear to stop growing based not simply on information about the amount of time elapsed, but based at least in part on the amount of growth undergone. This information about prior growth could involve any of several possible variables. One possibility is that a mechanism exists to assess organ size, either the tissue mass or the number of cells that have accumulated in the organ. A second possibility is that growth continues until the organ achieves a certain level of function. A third possibility is that a mechanism exists to count the number of cell divisions undergone. There is an important distinction between counting the number of cells present and counting the number of cell divisions that have occurred. If growth deceleration is determined by counting the number of cells that have accumulated, it would imply that if we experimentally remove a certain fraction of cells from an organ during the period of growth, the organ will continue to grow until the normal size is achieved. In contrast, if growth deceleration is determined by counting the number of cell divisions undergone, the organ will not sense any experimental removal of cells and would thereafter attain a smaller than normal size. Whether growth is limited by tissue mass, number of cells accumulated, organ function, or the number of cell divisions may depend on the organ involved. Even for a particular organ, more than one type of information might be employed to limit growth.

C. Mechanisms that assess organ mass: growth-inhibiting chalones

Classical studies in salamanders suggest that body growth in these amphibians is dependent on the ultimate tissue mass (129). A tetraploid salamander has cells that are twice as large as those of a diploid salamander, but the tetraploid adult has half as many cells, resulting in a similar body size (129). Similarly, in mice, generation of tetraploids by electrofusion (130) results in embryos with cells of approximately twice the normal size but half the normal cell number, thus achieving an overall size comparable to normal embryos (131). Conversely, in Drosophila, when the cell cycle pace is accelerated or slowed in an imaginal disc by overexpression or ablation of dE2f, the changes in proliferation rate are offset by changes in cell size to achieve a constant overall organ size (132). Recent studies suggest that an additional mechanism called cell competition helps maintain a constant organ mass (133). When growth-promoting dMyc is expressed in some but not all cells in the imaginal wing disc, cell competition occurs such that more slowly proliferating wild-type cells are selectively eliminated by apoptosis, thus, compensating for the increased number of dMyc-expressing cells (103, 134) to maintain a constant cell number.

How might organ size regulate growth? One old theory postulates the existence of “chalones”—soluble factors that are secreted by cells in an organ and negatively regulate growth of that organ (135) (Table 1). According to this theory, because the concentration of a chalone is directly proportional to the cell number in an organ, when the organ reaches a certain size, the concentration of the chalone is sufficient to stop growth (135). Such growth-inhibiting signals could act locally or could circulate and act systemically. Elegant studies demonstrate that myostatin serves as a chalone (136). Myostatin is a TGF-β family protein, which is secreted by muscle cells and acts to inhibit muscle growth (137–139) and myogenic satellite cell renewal (140, 141). Homozygous deletion of the myostatin gene (Mstn) in mice results in dramatic increases in skeletal muscle mass, attributed to both muscle cell hyperplasia and hypertrophy (137, 139, 142). This effect is not limited to the embryonic formation of muscle fibers because postnatal loss or blockade of myostatin has the same effect (143–149). Similar studies in dogs (150), pig (151), sheep (152, 153), cattle (154–156), and human (157) suggest that myostatin is an evolutionarily conserved negative regulator of skeletal muscle growth. Although myostatin was initially reported to be expressed exclusively in skeletal muscle (137), more recent evidence suggests that it is also expressed in cardiomyocytes, with its expression up-regulated postnatally (158), and serves to inhibit cardiomyocyte proliferation (158, 159). These findings therefore suggest that myostatin might also be involved in regulating the size of the heart.

Another member of the TGF-β family, activin A, has recently been shown to function alongside myostatin to negatively regulate muscle growth (160). Interestingly, the activity of both activin A and myostatin could be inhibited by follistatin (161, 162), which makes follistatin a potential therapeutic agent for improving muscle strength (163, 164). It is yet unclear how muscle mass itself might regulate the expression levels of activin A and follistatin as it regulates myostatin.

For many major organs, such as liver, kidney, and lung, it is not known whether chalones are involved in growth regulation. However, recent studies have identified chalone-like growth-inhibiting signals in a limited number of cell types. For example, GDF11 is produced by cells of the neuronal lineage of the olfactory epithelium and inhibit the production of olfactory receptor neurons (165, 166). Similarly, BMP2 and BMP4 serve as local refractory signals to inhibit proliferation of hair follicle stem cells, thus coordinating waves of hair regeneration (167, 168).

Leptin is another hormone that functions similarly to a chalone but acts through a systemic negative-feedback loop. Leptin is primarily produced by adipocytes, and the level of leptin in the circulation is directly proportional to whole-body adipose tissue mass (169, 170). At the hypothalamus, leptin then acts on proopiomelanocortin neurons to activate MSH secretion (171–173) and inhibits neuropeptide Y (NPY) and agouti-related peptide (AgRP) secretion by the AgRP/NPY neurons (174, 175). Collectively, these changes suppress appetite, stimulate energy expenditure, and therefore inhibit accumulation of adipose tissue (169, 170).

In the recently identified multiorgan juvenile genetic program (111), some of the growth-promoting genes that are down-regulated with increasing body growth are growth factors, such as IGF-II, midkine, and pleiotrophin. These findings suggest that local paracrine factors are involved in deceleration of juvenile growth. However, these soluble factors are functionally distinct from chalones. Although chalones are proposed to be growth-inhibitory factors that increase with growth, these are growth-promoting factors that decrease expression with growth, and therefore, in contrast to chalones, their abundance does not reflect the number of cells. However, it is possible that chalones in different organs drive the common genetic program.

D. Mechanisms that assess organ function

Another strategy to target the appropriate adult size of an organ is based on the function of the organ. For example, if the organ is too small to provide adequate function, a mechanism to stimulate growth might be activated. When the organ achieves a size sufficient to provide adequate function, the growth-stimulatory function would cease. Conversely, when the organ achieves adequate size and therefore function, a growth-inhibiting mechanism might be activated.

The existence of a mechanism that senses organ function is suggested by the phenomenon of liver regeneration after partial hepatectomy in mammals (176–178). This regeneration involves a brief burst of cell replication by a large number of differentiated hepatocytes (25). Similarly, the liver also adjusts its size relative to the overall body size after orthotopic transplantation from a small dog to a larger dog (38). Furthermore, in rats connected by parabiosis, partial hepatectomy in one animal stimulates hepatocyte proliferation in the other (179), indicating that liver regeneration involves circulating factors. The concept that liver size may be regulated by function is further supported by its ability to undergo hyperplasia upon increased xenobiotic stress, such as administration of phenobarbital and 1,4-Bis[2-(3,5-dichloropyridyloxy)]benzene (180), without liver injury (181, 182).

More recent studies suggest that the liver senses organ functionality, at least in part, through the flux of bile acids (183). Bile acids are synthesized in the liver and secreted into the small intestine, where they facilitate digestion. The bile acids are subsequently reabsorbed and returned to the liver, where they are taken up by hepatocytes, thus completing the enterohepatic circulation. When exposed to a mild increase in bile acid flux, hepatocyte proliferation is induced, causing an increase in liver size (183). Similarly, liver regeneration after partial hepatectomy is potentiated in the presence of bile acids (183, 184). These findings were further supported by a complementary study, showing that disruption of the enterohepatic circulation by draining the common bile duct externally during the 7 d before hepatectomy delays subsequent liver regeneration (185). Thus, collectively compelling evidence suggests that liver regeneration and adult liver size can be modulated by organ function. One possibility is that this same mechanism alone could be used to drive rapid juvenile growth of the liver and its subsequent growth deceleration, because the mechanism would cause liver size to keep pace with the growth rate of the rest of the body. Alternatively, organ function might be used only to fine-tune the adult liver size, and a separate mechanism might be used to suppress proliferation in early juvenile life.

For some endocrine organs, such as the thyroid, adrenal, and parathyroid glands, size is also regulated in part by organ function. If the organ produces insufficient hormone, a negative feedback loop serves to stimulate not only increased secretion but also proliferation and growth of the endocrine gland. Although this mechanism may be used to precisely target the optimal adult organ size based on function, other mechanisms may also contribute to juvenile growth deceleration and size determination. For example, the size of the thyroid gland is regulated by a negative feedback loop involving thyroid hormone and TSH, and yet, in mice lacking a functional TSH receptor, the thyroid still grows to approximately 50% of the normal size (186).

Kidney size also seems to be regulated in part by function. Unilateral nephrectomy causes an increase in the size of the remaining kidney (187), perhaps involving a temporary increase in serum creatinine that signals the increased functional need (187). However, this increase in kidney size is primarily due to hypertrophy (cell enlargement) rather than hyperplasia (cell proliferation) and thus may involve mechanisms different from those primarily regulating juvenile growth (188).

This work was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health.

Disclosure Summary: The authors have nothing to disclose.