Covert actions of growth hormone: fibrosis, cardiovascular diseases and cancer

Abstract

Since its discovery nearly a century ago, over 100,000 studies of growth hormone (GH) have investigated its structure, how it interacts with the GH receptor and its multiple actions. These include effects on growth, substrate metabolism, body composition, bone mineral density, the cardiovascular system and brain function, among many others. Recombinant human GH is approved for use to promote growth in children with GH deficiency (GHD), along with several additional clinical indications. Studies of humans and animals with altered levels of GH, from complete or partial GHD to GH excess, have revealed several covert or hidden actions of GH, such as effects on fibrosis, cardiovascular function and cancer. In this Review, we do not concentrate on the classic and controversial indications for GH therapy, nor do we cover all covert actions of GH. Instead, we stress the importance of the relationship between GH and fibrosis, and how fibrosis (or lack thereof) might be an emerging factor in both cardiovascular and cancer pathologies. We highlight clinical data from patients with acromegaly or GHD, alongside data from cellular and animal studies, to reveal novel phenotypes and molecular pathways responsible for these actions of GH in fibrosis, cardiovascular function and cancer.

Recombinant human growth hormone (hGH) is used primarily to treat children with short stature due to GH deficiency (GHD). In addition, hGH is approved for use in children born small for gestational age, as well as children with idiopathic short stature, Turner syndrome, Noonan syndrome, Prader–Willi syndrome, short stature homeobox-containing gene deficiency and chronic renal insufficiency, and in adults with GH deficiency. Despite this success, controversial issues remain, and not all indications are approved in all countries^1,2.

The Argentine clinician Bernardo Houssay discovered an adverse effect of GH nearly 90 years ago; namely its ability to inhibit insulin action, which is now known as its diabetogenic effect³. Since that time and despite the astonishing success of hGH, several puzzling actions of GH have been and continue to be described. Here, we review three interrelated covert actions reported in adult mice and humans: fibrosis, cardiovascular disease and cancer. Many have joked that ‘too much of a good thing is bad’, which in this context is too much GH. Interestingly, the absence of GH action in animals and humans, despite short stature, results in resistance to type 2 diabetes mellitus (T2DM) and cancer, and improvements in other indicators of healthspan⁴. With regard to longevity, mice with a reduction or absence in GH action have a robust and reproducible increase in lifespan⁵. Although data from humans is insufficient to draw firm conclusions, some cohorts of individuals with isolated GHD (for example, in the Brazilian Itabaianinha cohort) have attained extreme longevity despite representing a fairly small proportion of the population. Such observations suggest that the findings in rodents are relevant to humans⁵. The effects of the GH–insulin-like growth factor 1 (IGF1) axis on human ageing, how GH levels change during ageing and the relationship to age-related diseases are discussed in detail elsewhere⁶.

In this Review, we explore the possibility that too much, too little or inappropriately expressed GH might provoke covert physiological actions related to fibrosis, cardiovascular function and cancer development.

GH and fibrosis

Fibrosis is characterized by excessive accumulation of the extracellular matrix (ECM) within tissues. Although fibrosis is initially a protective and adaptive response to wound healing and tissue repair, it can become uncontrolled and lead to tissue dysfunction and pathology. Almost every tissue type can be affected by pathological fibrosis; however, heart, adipose tissue, liver, kidney, intestine and skin are among the best studied. Tissue fibrosis is thought to be a major underlying cause of death in humans, with some studies estimating that it is associated with as much as 45% of all-cause mortality in industrialized nations⁷. As GH has an important role in growth and collagen turnover, it is a likely factor in ECM remodelling and fibrosis formation.

Influence of GH on collagens

A compelling reason for a GH–fibrosis link is the close correlation between GH exposure and collagen production. For example, in vitro studies of GH treatment of cells indicate increases in Colla1, Col3a1 and Col6a1 expression in differentiated adipocyte-like 3T3-L1 cells (a mouse cell line)⁸ and type I collagen mRNA and protein in rodent intestinal primary myofibroblasts⁹. Animal studies have shown a similar positive correlation between collagen and GH. For example, GH treatment increases collagen IV protein in mouse neural tissue damaged by stroke¹⁰, and type I collagen RNA and protein in the jejunum of GH-treated rats⁹. Chronic exposure to excess GH in rodent models also increases collagen expression in other tissues such as kidneys and tendons^11,12.

The data in adult humans are similar. Both acute and chronic GH treatment in older men increases collagen expression in tendon and muscle¹³. Furthermore, the same study failed to detect an effect of GH on muscle fibre synthesis¹³. Acromegaly is a rare disease usually caused by a benign GH-producing pituitary tumour and elevated serum levels of IGF1 (REF.¹⁴). Adult patients with acromegaly have increased collagen turnover and serum levels of type I collagen and procollagen III amino-terminal propeptide (PIIINP)¹⁵, which decline with remission of the disease^16,17. The close association among GH, collagen synthesis and turnover, and fibrosis in clinical populations is further illustrated by several examples, including the use of validated tests to detect GH doping in sport that are based on elevated serum levels of IGF1 and PIIINP levels¹⁸. Moreover, a marker of growth plate activity and overall rate of linear bone growth (the intact trimeric non-collagenous domain of type X collagen) shows promise for reflecting GH action in children and adults with GHD¹⁹. Finally, hGH is used in a combination therapy to treat osteogenesis imperfecta, an inherited connective tissue disorder characterized by a quantitative or qualitative defect in collagen synthesis²⁰.

GH and other fibrosis-promoting factors

GH influences many factors besides collagens that contribute to fibrosis, including ECM-modifying proteins, several proteins and pathways implicated in fibrosis (for example, the transforming growth factor-β (TGFβ) pathway and mitogen-activated protein kinase (MAPK) pathways), senescence, immune cell function, and fibroblast activation or plasticity. In general, fibrosis is also often preceded by and closely associated with inflammation⁷. These fibrotic effects of GH are frequently context-specific and tissue-specific and not always in a direction that favours fibrosis. For example, an ECM-degrading endopeptidase, matrix metalloproteinase 2 (MMP2), is decreased after GH treatment of individuals with GHD²¹, but increased in patients with active acromegaly²². Likewise, TGFβ, a master regulator and promoter of fibrosis in multiple tissues, is regulated distinctly by GH in a tissue-specific manner. For example, TGFβ expression is decreased by GH in primary cardiac fibroblasts²³. Furthermore, TGFβ expression is increased in the glomeruli of bovine GH (bGH) transgenic mice, a model of GH overexpression¹¹. Interestingly, GH action itself might also be influenced by ECM-modifying proteins. Tissue inhibitor of metalloproteinase 3 (TIMP3) is unique among the TIMP family because it alone has a high affinity for ECM proteoglycans and possesses the broadest range of substrates. TIMP3 modulates GH receptor (GHR) abundance on the cell surface in human cell lines that stably express GHR and JAK2, and dampens the GH-induced intracellular signalling cascade, revealing an interesting interplay between GH and the ECM²⁴.

Senescence.

Cellular senescence is characterized by irreversible cell cycle arrest and a senescence-associated secretory phenotype (SASP). The SASP is the main feature of senescent cells and includes pro-inflammatory cytokines and chemokines, growth factors and proteases, which contribute to a harmful tissue microenvironment. However, whether senescence is beneficial or detrimental for the development of fibrosis remains debated. For example, senescence might be beneficial by preventing fibroblast differentiation and the ability of these fibroblasts to contribute to liver fibrosis, based on studies using senolytic chimeric antigen receptor T cells that efficiently target and ablate senescent cells in mice²⁵. However, senescent cells could be detrimental, as removal of senescent cells in rodents either pharmacologically or genetically can reverse fibrosis at least in some tissues such as lung²⁶, and can extend median lifespan²⁷. Of note, GH and/or IGF1 levels are positively correlated with senescence and SASP in white adipose tissue in mice²⁸, as well as in cardiomyocytes and skin fibroblasts from patients with acromegaly²⁹, and in human primary fibroblasts and mouse embryonic fibroblasts³⁰. In fact, GH can also be secreted by senescent cells, thus constituting a component of the SASP³¹. By contrast, however, GH treatment attenuated senescence of primary human endothelial progenitor cells³², and IGF1 overexpression in rats improved stress-induced senescence in the liver³³.

Fibroblasts.

Dysregulation of fibroblasts has long been considered a root cause of fibrosis. Data from the past 5 years show marked heterogeneity in fibroblast subtypes, which localize to unique anatomical sites and have distinct physiological functions³⁴. Many studies have shown that GH and/or IGF1 influence the differentiation and proliferation of fibroblasts^35,36. This association has been best studied in wound healing. For example, GH was identified as an inhibitor of TGFβ-induced myofibroblast differentiation in the skin of bGH transgenic mice³⁷. A 2020 study showed that fibroblast activation protein (FAP) has been linked to GH action in humans¹⁶. FAP is a cell surface protease that has an important role in the degradation of ECM and with known substrates that include type I and III collagens³⁸. FAP expression is limited after birth except in activated fibroblasts and in conditions associated with notable ECM remodelling, such as liver fibrosis³⁹. Interestingly, serum levels of FAP are elevated in patients with untreated acromegaly and become considerably reduced after disease management, which closely correlates with reductions in markers of collagen turnover¹⁶. Additional studies are needed to determine if GH directly influences FAP abundance and action to influence ECM remodelling, as well as the ability of FAP to serve as a biomarker for GH-induced fibrosis.

Fibrosis and GH: the yin and the yang

As outlined above, GH has a complex relationship with many factors that contribute to fibrosis and ECM remodelling. Whether GH promotes or reduces pathological tissue fibrosis depends on the experimental conditions, the tissues analysed and the degree of GH signalling.

The yin: acromegaly and GH-transgenic mice.

Acromegaly is accompanied by soft-tissue enlargement and increases in ECM constituents; for example, glycosaminoglycan deposits that contribute to generalized oedema are apparent in the skin of patients with acromegaly⁴⁰. Both early autopsy studies done in the 1980s⁴¹ as well as in vivo studies from 2015 using cardiac MRI have revealed myocardial fibrosis in adult patients with acromegaly⁴². Furthermore, increased hepatic fibrosis has also been shown in subpopulations of adult patients with acromegaly, particularly those that are already at genetic risk of developing hepatic steatosis⁴³. Finally, hard thyroid nodules are common in patients with acromegaly and are also thought to be caused by increased nodular fibrosis⁴⁴. However, overdiagnosis due to surveillance bias is inherent with regard to thyroid nodules, and further studies are needed⁴⁵.

Data derived from mice are more definitive of the actions of excess GH and IGF1 in inducing fibrosis and organ dysfunction. In fact, most tissues assessed for fibrosis in bGH mice show some increase in fibrosis at older ages (FIG. 1) compared with wild-type mice, which might reflect an unexplored benefit of somatopause (the gradual decline in GH secretion that occurs with ageing). The older bGH mice have an enlarged heart with impaired function and marked distinct perivascular and interstitial fibrosis^46,47. These older transgenic mice also have renal damage with increased fibrosis and glomerular lesions⁴⁷. Marked fibrosis is also apparent in their adipose tissue, more prominently in subcutaneous depots⁸, as well as in the intestines⁴⁸. Severe skin fibrosis has also been found in bGH mice, which is more prominent in males than females³⁷. Overall, increased fibrosis has been observed in many tissues assessed in bGH mice relative to controls, with males often exhibiting a more robust phenotype. By contrast, several mouse lines with decreased GH action (such as GHR antagonist mice⁸, mice with GHR disruption in adipocytes⁴⁹ or mice deficient in GH⁵⁰) have decreased fibrosis, at least in adipose tissue. Collectively, these data in humans with acromegaly and GH transgenic mice provide strong support of fibrosis being an underlying component of the organ dysfunction that is common with pathological increases in GH and/or IGF1 action.

Fig. 1 |. GH overexpression in mice induces tissue fibrosis.

Histology of assorted tissues in aged male mice (10–13 months of age) overexpressing bovine growth hormone (bGH) and wild-type control mice. White adipose tissue from inguinal or epididymal adipose depots was stained with Sirius red (a non-specific red collagen stain), Swiss rolls of the small intestine were stained using Sirius Red and Fast Green (which stains non-collagenous protein), heart was stained with Masson’s trichrome (which stains connective tissue blue) and kidney sections were stained with periodic acid Schiff stain (which stains connective tissue a purple–magenta colour). The tibial articular cartilage images were generated via immunostaining with a type X collagen (a bone-specific collagen)-specific antibody (orange–red colour). In bGH mice, the increased GH activity results in increased fibrosis in the tissues shown. The histological images of tibial articular cartilage are courtesy of S. Zhu, Ohio University.

The yang: therapeutic potential of GH.

The benefits of GH in promoting collagen deposition are best illustrated by its effect on promoting longitudinal growth and bone acquisition⁵¹. However, GH might also have therapeutic potential for selected fibrotic diseases. These benefits might relate to restored GH levels in deficient states or to indirect influences on other disease pathologies or severe conditions. For example, studies have suggested that the dysregulation of the GH–IGF1 axis in patients with morbid obesity or in adult patients with hypopituitarism and GHD might contribute to the severity of hepatic fibrosis^52,53. Furthermore, in at least one small study of adult patients with GHD, GH replacement therapy improved hepatic fibrosis⁵⁴. Another example of GH action to attenuate fibrosis was provided by a rat model of intestinal inflammation, in which GH improved rather than exacerbated intestinal fibrosis⁵⁵, Moreover, in a randomized controlled trial, GH therapy was beneficial in the treatment of severely burned children, in whom it promoted healing without the development of hypertrophic scarring caused by tissue fibrosis⁵⁶. Initial small, randomized studies in patients with large burns treated with GH have shown promising clinical benefit. However, treatment with GH also induces hyperglycaemia in adults with severe burns, which raises questions for its utility in clinical practice in the treatment of large burns⁵⁷.

Importantly, several limitations exist for the use of GH therapy for fibrotic disease. First, in most studies that showed some benefit of GH therapy on fibrosis, GH was restored to normal levels in patients with GHD. Second, many studies assessed acute or short-term GH therapy without evaluating the long-term cardiometabolic effects. Third, attenuation of fibrosis, treatment of burns and improvement in obesity and its complications are non-approved uses of GH.

Summary

Overall, the available data suggest an intricate balance between having sufficient GH action to promote favourable ECM remodelling and avoiding excess GH action that promotes ECM deposition and scarring, thereby resembling a ‘Goldilocks effect’. Although the focus of this Review is GH, disentangling the direct effects of GH versus indirect effects exerted via IGF1 is difficult in vivo. Endocrine and paracrine IGF1 also has an important role in fibrosis development. For example, local IGF1 administration stimulates in vivo tendon collagen synthesis in healthy, sedentary humans¹³. In addition, overexpression of IGF1 receptor (IGF1R) in human primary dermal fibroblasts suggests a role for IGF1 signalling in the development of keloid and hypertrophic scarring⁵⁸. However, excess levels of GH, but not IGF1, have been shown to promote kidney glomerulosclerosis⁵⁹. Together, these findings illustrate the need for additional studies or experimental systems that can help delineate the contributions of each individual hormone to fibrosis in specific tissues. In addition, further studies are needed to evaluate whether GH and IGF1 are primary drivers of fibrosis or if fibrosis occurs secondary to organ dysfunction.

GH and cardiovascular disease

GH hypersecretion: acromegaly

The onset of acromegaly is gradual, with a diagnostic delay of 5–10 years in most patients, which has negative therapeutic and prognostic implications. Patients with acromegaly are exposed to long-standing GH and IGF1 excess and have increased cardiovascular mortality⁶⁰ that is associated with hypertension and heart failure^14,61. Interestingly, atherosclerosis is not always present in acromegaly despite the presence of classic risk factors such as hypertension, insulin resistance and T2DM⁶². By contrast, typical cardiomyopathy present in active acromegaly is cardiomegaly and, in particular, hypertrophy of the left ventricle^14,62. Histopathology of cardiac tissue from patients with acromegaly reveals pronounced interstitial fibrosis (FIGS. 1,2), myocardial hypertrophy and evidence of myocarditis with infiltration of inflammatory cells⁴¹. Importantly, 85% of patients with acromegaly have been observed to have interstitial fibrosis⁴¹. In addition, regurgitation of the aortic and mitral valves, as well as diastolic dysfunction, are described in acromegaly, whereas systolic dysfunction is a rare and late-onset occurrence^14,62.

Fig. 2 |. The pleiotropic actions of GH–IGF1 on the cardiovascular system, and electrolyte and water balance in humans.

Growth hormone (GH) and insulin-like growth factor 1 (IGF1) promote renal reabsorption of sodium and water and thereby increase the extracellular volume (ECV) and plasma volume. This effect is accompanied by nitric oxide-mediated vasodilation and increased myocardial contractility, which translate into increased cardiac output. This increase results in hyperperfusion and reduced total peripheral resistance. Pathological GH–IGF1 excess (such as in uncontrolled acromegaly) results in hypertension and cardiac hypertrophy. We hypothesize that GH–IGF1-induced fibrosis eventually contributes to the development of cardiomyopathy and heart failure (systolic dysfunction). Pathological processes are highlighted in the figure in red.

Due to the gradual onset of the disease and the paucity of long-term prospective studies, the pathogenesis of the cardiovascular complications of overt acromegaly remains elusive. Individual and combined effects can also be exerted by GH and IGF1, further complicating the picture. However, short-term experimental studies in humans could provide clues. For example, 1 week of supra-physiological GH administration in healthy men aged ~30 years substantially increased fractional shortening of the left ventricle, which was mainly attributed to a reduced end-systolic diameter⁶³. In concomitance with a significant 15% increase in resting heart rate, cardiac output increased by ~13% without a change in mean arterial blood pressure^63,64, which suggests a simultaneous reduction in peripheral vascular resistance. Increased blood flow and reduced vascular resistance via a nitric oxide-dependent mechanism are also observed after short-term, high-dose GH infusion into the brachial artery of healthy individuals⁶⁵. In addition, the GH–IGF1 axis is well recognized to acutely induce fluid and sodium retention⁶⁶ and increases heart rate without a change in blood pressure⁶⁷. Similar findings of increased myocardial contractility and cardiac output are also encountered in patients with active acromegaly⁶⁸ together with sodium and fluid retention⁶⁹. Taken together, sustained fluid and sodium retention, in combination with direct inotropic effects of GH–IGF1, are hypothesized to promote hyperperfusion, cardiac hypertrophy and ultimately systemic hypertension in acromegaly⁶⁶. In combination with other risk factors such as insulin resistance⁷⁰, T2DM and obstructive sleep apnoea, these effects might eventually cause cardiomyopathy and heart failure in the absence of coronary atherosclerosis (FIG. 2).

Treatment of acromegaly that achieves biochemical remission⁷¹, translates into a life expectancy that is close to that in the general population⁶¹. Regarding cardiovascular function, successful surgical resection of the GH-producing adenoma improves diastolic function and reduces left ventricular mass, together with reductions in heart rate and blood pressure⁷². Similar beneficial effects have been reported after disease control obtained by pharmacological treatment with a somatostatin analogue alone⁷³ as well as with a GHR antagonist (pegvisomant)¹⁴. In addition, disease control in acromegaly improves certain risk factors for cardiovascular disease, such as insulin resistance and obstructive sleep apnoea; however, body composition changes towards increased adipose mass and reduced lean body mass¹⁴.

GH deficiency

GHD in adults is caused by diseases affecting the hypothalamic–pituitary region, the most common being benign tumours. The diagnosis of GHD in adults must be made in the correct clinical context (that is, in patients with a well-defined hypothalamic–pituitary disorder) using a validated GH stimulation test, unless four or more anterior pituitary hormone deficiencies exist together with a low serum concentration of IGF1 (REF.⁷⁴). GHD in adults can occur in isolation but is commonly associated with other pituitary hormone deficiencies. These deficiencies can also affect patient outcomes; for example, inadequate or untreated sex steroid deficiencies in women with hypopituitarism can affect their GHD outcome, inadequate treatment of central hypothyroidism might have negative effects on many metabolic processes and inadequate treatment of adrenal insufficiency has negative effects on metabolism as well as on risk of premature mortality^75–77. Cardiovascular disease is a complication in adults with GHD; adults with GHD have increased blood pressure, a phenotype resembling the metabolic syndrome⁷⁸ and increased cardiovascular mortality⁷⁹. Several cardiovascular risk factors are increased in adults with GHD, which can be explained by many of the effects of GH and IGF1 on the cardiovascular system (which expresses both GHR and IGF1R⁸⁰), as well as effects on lipid and lipoprotein metabolism and systemic inflammation⁸¹.

Adults with GHD often have overweight or obesity, with accumulation of visceral adipose tissue^82,83. Studies on their body composition show that their total body adipose mass is increased (predominantly abdominal), muscle mass is reduced and extracellular fluid is reduced. GH antagonizes the effects of insulin in several important tissues such as liver, muscle and adipose tissue^83,84. Studies in children and young adults with GHD have clearly shown that these individuals have increased insulin sensitivity, and young children in particular might even experience hypoglycaemia⁸⁵. Surprisingly, adults with long-standing untreated GHD deficiency showed markedly reduced insulin sensitivity, measured using the hyperinsulinaemic glucose clamp technique^86,87. Their abdominal adiposity together with reduced muscle mass, reduced serum concentration of IGF1 and physical inactivity due to reduced exercise capacity might help explain this finding^83,86,88. Abdominal adiposity and associated insulin resistance are therefore important mediators of premature atherosclerosis and increased risk of cardiovascular disease seen in adults with GHD.

Adults with GHD have reduced nitric oxide urinary excretion, increased peripheral vascular resistance⁸⁹, increased sympathetic nervous system activity⁹⁰ and increased systemic inflammation⁸¹ (FIG. 2). A consistent finding is that the extracellular fluid volume and plasma volume are decreased in GHD⁹¹. Untreated acromegaly is associated with hypertension, therefore blood pressure would be expected to increase with GH replacement in adults with GHD. However, the initial placebo-controlled trials with adult GH replacement showed that diastolic blood pressure decreases⁹² despite a sustained increase in extracellular fluid and plasma volume⁹¹. The mechanisms for reduced diastolic blood pressure seem to be nitric oxide-mediated vasodilation and reduced peripheral resistance^89,93, reduced sympathetic nervous system activity⁹³ and improved endothelial function⁹⁴. Also, MMP and vascular endothelial growth factors, which are markers of endothelial function, are increased in adults with GHD and decline with GH replacement²¹. Of note, individuals with isolated GHD that is caused by an inactivating mutation of the GH-releasing hormone RH receptor (GHRHR) have obesity, elevated levels of LDL cholesterol, C-reactive protein (CRP) and mild systolic hypertension, but an absence of atherosclerosis or heart failure⁹⁵.

Long-term GH replacement therapy in patients with GHD induces sustained sodium and water retention^96,97. This effect is due to increased sodium and water reabsorption from the distal renal tubuli, mediated by the direct actions of GH and IGF1, but also indirectly through stimulation of the renin–angiotensin–aldosterone system^96–98. This sustained increase in extracellular water is in contrast to the effects of mineralocorticoids, which after administration increase the extracellular volume for days, where after renal sodium escape occurs, extracellular volume returns to baseline levels⁹⁹. A probable explanation for the sustained sodium and water retention induced by GH is the reduction of natriuretic peptides induced by GH treatment, which mitigates renal sodium escape^96,100, together with increased glycosaminoglycan deposition in peripheral tissues⁴⁰.

Reduced exercise capacity has consistently been shown in adults with GHD, which improves with GH replacement^101,102. The mechanism responsible could be related to effects on the heart and cardiovascular system, but also due to effects on muscle size and muscle function¹⁰³. GHD is associated with reduced left ventricular mass index, reduced cardiac output with both impaired diastolic and systolic left ventricular function and reduced fractional shortening¹⁰⁴. With GH replacement, an increase occurs in the left ventricular mass index as well as improvements in systolic and diastolic function that might contribute to improving exercise capacity¹⁰⁴.

GH has important regulatory effects on various aspects of lipid and lipoprotein metabolism. Lipid metabolism is strongly linked to atherosclerosis and local endothelial inflammation¹⁰⁵. GH upregulates LDL receptors in the liver¹⁰⁶, stimulates lipolysis¹⁰⁷ and increases the synthesis and secretion of VLDL from the liver¹⁰⁸. In addition, GH increases the ratio of plasma lecithin to cholesterol acyltransferase concentration and lipid transfer protein activities, which might explain the increase in HDL cholesterol levels seen in some GH replacement studies in adults with GHD¹⁰⁹. Adults with GHD have increased concentrations of total cholesterol and LDL cholesterol, and in some studies also reduced HDL cholesterol^110,111. Likewise, controlled studies of adult GH replacement have revealed decreased circulating levels of total cholesterol and LDL cholesterol⁹², whereas effects on HDL cholesterol and triglycerides are smaller and less robust. Lipoprotein(a) is a lipoprotein that is similar to LDL in terms of its lipid and apolipoprotein B-100 content, but also contains apolipoprotein(a) covalently linked to apolipoprotein B-100. Increased plasma concentration of lipoprotein(a) is associated with increased risk of cardiovascular disease and its plasma level is mainly genetically determined, with an inverse relationship between circulating levels of lipoprotein(a) and the size polymorphism of apolipoprotein(a). Adults with GH deficiency have similar lipoprotein(a) concentrations to healthy matched control individuals, but GH replacement increases the concentration of lipoprotein(a)¹¹². However, the net outcome of these changes in lipid and lipoprotein metabolism during GH replacement cannot be determined from available data.

Systemic inflammation markers such as CRP are well recognized cardiovascular risk factors¹¹³. Adults with GHD have increased circulating markers of inflammation such as CRP and IL-6 (REF.¹¹⁴). In one controlled study, GH replacement reduced the circulating levels of both CRP and IL-6 (REF.¹¹⁵). Also, in studies in women with abdominal obesity, GH treatment in comparison with placebo reduced biomarkers of systemic inflammation¹¹⁶. Thus, GHD is associated with a pro-inflammatory state but not with tissue fibrosis⁸¹. Taken together, studies of GH replacement in adults with GHD have shown beneficial effects on cardiovascular risk factors, cardiovascular function and surrogate variables for cardiovascular morbidity and mortality. However, no prospective controlled studies are available demonstrating reduced vascular morbidity and mortality. Such a study is unlikely to be performed due to the low prevalence of the disorder. However, in a retrospective meta-analysis of mortality in patients with hypopituitarism (including patients receiving GH replacement), GH replacement was associated with reduced mortality, particularly in men¹¹⁷.

GH and endothelial cellular function

The vascular endothelium, which is the largest organ in the body, responds to various circulatory growth factors including GH. Angiogenesis is essential for organogenesis and successful embryonic and fetal development¹¹⁸. Disruption of the mechanisms controlling physiological angiogenesis underlies the pathophysiology and pathogenesis of various diseases including cancer, psoriasis, arthritis, retinopathies, obesity, asthma and cardiovascular disease. During angiogenesis, endothelial cells are regulated by an interplay between cells, multiple soluble factors and the ECM. For example, vascular endothelial growth factor A (VEGFA) binds to its receptor, VEGFR2, to stimulate downstream signalling to activate endothelial nitric oxide synthase (eNOS) and nitric oxide release¹¹⁹ (FIG. 2). Nitric oxide has a crucial role in angiogenesis and vasodilation, and also acts as an inhibitor of platelet adhesion and aggregation, monocyte adhesion and vascular smooth muscle cell growth. Endothelial cells express GHR, and GH, in part, regulates endothelial cell function and angiogenesis through VEGFA¹²⁰.

As discussed earlier in the article, both in vivo and in vitro studies have shown that treatment with IGF1 stimulates eNOS expression and nitric oxide release¹²¹ (FIG. 2). Interestingly, age-dependent impairment of endothelial progenitor cells in middle-aged and older humans is corrected by treatment with hGH-mediated increases of IGF1 (REF.³²), which supports the role of declining GH levels in ageing-associated cardiovascular dysfunction. Similarly, studies in adults with GHD have shown improved endothelial function and reduced vascular risk after GH replacement therapy¹²². As patients with GHD have reduced nitric oxide production⁸⁹, a feasible explanation of increased blood flow after GH replacement could be the improvement in endothelial function¹²³. In addition, an improvement in the arterial response to induced vasodilation is observed in adolescents with GHD after GH replacement therapy¹²⁴.

Although atherosclerosis is not prevalent in acromegaly⁶², GH has been found in experimental studies to directly stimulate the development of atherosclerosis in endothelial cells. For example, GH stimulates VCAM1 and SELE transcripts in human umbilical vein endothelial cells (HUVECs) via the MAPK pathway, which results in augmented adhesion of a human leukaemia monocyte cell line (THP-1) and primary monocytes to HUVECs¹²⁵. As the endothelium has a key role in the pathogenesis of atherosclerotic plaques, excess GH or IGF1 could play an active or passive part in atherosclerosis and cardiovascular dysfunction via effects on endothelial pathophysiology. These effects include endothelial proliferation, endothelial progenitor cell dysfunction or endothelial oxidative stress¹²⁶.

The effect of GH–IGF1 on endothelial cells has also been noted in other tissues and organs. In mouse endothelioma cells, GH has mitogenic effects¹²⁷. GH also affects endothelial cell morphology and augments the deposition of the ECM molecules, laminin and fibronectin, on the cell surface¹²⁷. In addition, human GH at physiological concentrations stimulates human retinal microvascular endothelial cells in vitro, thereby enhancing their proliferation¹²⁸. In retinopathy, the proliferative form of retinal endothelial cells is observed during a more advanced stage of the disease and is characterized by retinal neovascularization. Although now abandoned, pituitary ablation was a method to suppress GH secretion as a potential treatment for proliferative diabetic retinopathy¹²⁹. Interestingly, GH has an essential role in ischaemia-induced retinal neovascularization in mice¹³⁰ and a GHR antagonist prevents this effect¹³⁰. Finally, topical application of GH accelerates the closure of skin wounds by accelerating re-epithelialization and collagen deposition, and stimulating angiogenesis¹²⁷, which occurs primarily via local production of IGF1 in the tissue¹³¹. Overall, these studies highlight the importance of the GH–IGF system in vascularization and angiogenesis.

Regulation of cardiac function

Studies in vitro and in rats have shown that GH and IGF1 increase cardiomyocyte gene expression and protein synthesis, which translates into cardiac hypertrophy and remodelling, but they also inhibit apoptosis¹³². GH transgenic mice that have elevated levels of GH and IGF1 show increased organ sizes including the heart, while heart-specific IGF1 expression markedly promotes myocyte proliferation¹³³. Furthermore, the removal of GH action in the heart of adult mice affects neither the local levels of IGF1 nor cardiac function, even though endocrine IGF1 levels are altered¹³⁴. As for IGF1, loss-of-function studies using cardiac-specific Igf1r-knockout mice have shown that autocrine and paracrine IGF1 promotes heart repair in response to injury and conservation of cardiac function¹³⁵.

Endothelial-specific human IGF1R-overexpressing transgenic mice show reduced basal and insulin-stimulated eNOS activity, with no change in size or weight or whole-body glucose homeostasis¹³⁶. These mice show normal blood pressure, an enhanced aortic response and increased endothelial cell migration and regeneration. By contrast, the Igf1r-knockout mice show normal glucose homeostasis with enhanced basal and insulin-stimulated eNOS phosphorylation¹³⁷ and vascular permeability in the endothelial lining¹³⁸. These studies support an important role for IGF1R in regulating nitric oxide bioavailability and vascular repair, which are hallmarks of several human diseases involving tissue growth and vascularization. In addition, macrophage IGF1 signalling exerts anti-atherogenic effects through reducing macrophage activities, decreased atherosclerotic lesion formation and reduced plaque vulnerability^139,140.

Summary

Overall, evidence indicates that GH–IGF1 pathways have an important role in endothelial cell metabolism. Both elevated (acromegaly) and low (GHD) levels of GH are associated with cardiovascular disease. GH is a critical regulator of inflammation and immune activation. Considering that inflammation and fibrosis regulate progressive cardiac dysfunction during ischaemic heart failure¹⁴¹, targeting mediators of GH action could provide an exciting therapeutic avenue for this disease. GH could also have a ‘Goldilocks effect’, where too little or too much can lead to insufficient or dysregulated immune activation and fibrotic mechanisms that lead to the exacerbation of cardiovascular disease. Comprehensive studies are warranted to fully investigate the involvement of the GH–IGF1 axis in the pathogenesis and pathophysiology of cardiovascular disease and its complications.

GH and cancer

Numerous studies in multiple cancer types since 1950 have shown that the intrinsic growth-promoting action of GH drives the growth and proliferation of cancer cells both in vitro and in vivo¹⁴². However, in light of our rapidly evolving understanding of different aspects of cancer as well as of GH action, a more ‘covert’ role of GH in cancer, beyond just promoting tumour growth, has also emerged. Pituitary secreted endocrine GH, critical for normal growth and development, is well known to decrease steadily in adults with age. By contrast, local or non-pituitary GH production from several non-pituitary sites (including peripheral tissues and multiple tumour types) seems to stay fairly steady or even increase with age^143,144 and exerts a profound autocrine and/or paracrine action. Fibrosis seems to be a consistent underlying theme in this action, wherein GH induces extensive ECM remodelling by inducing expression of proteases, collagen and various cytokines¹⁴⁵. Studies in cultured cells and animal models have revealed an extensive array of molecular mechanisms by which non-pituitary GH is now known to be a critical driver of a tumour supportive microenvironment and cancer therapy resistance (FIG. 3).

Fig. 3 |. Covert actions of GH in cancer.

The emerging covert actions of growth hormone (GH) in cancer as evidenced from reports in the literature to date. a | GH promotes a tumour-supportive microenvironment via the crosstalk of tumour cells and multiple types of cells in the tumour microenvironment (TME), such as immune cells, cancer-associated adipocytes (CAA) and cancer-associated fibroblasts (CAF). This crosstalk occurs via autocrine and/or paracrine GH actions, which promote fibrosis and extracellular matrix (ECM) remodelling via matrix metalloproteases and collagen turnover, production of pro-inflammatory cytokines and immune-suppressive molecules such as transforming growth factor-β (TGFβ). Parts b–d illustrate the process of tumour therapy resistance through various mechanisms involving GH. b | Suppression of apoptosis due to radiotherapy-induced or chemotherapy-induced DNA damage, via downregulation of p53 and apoptotic mediators including caspases and upregulation of anti-apoptotic factors including BCL-2. c | Increased ATP-binding cassette (ABC) transporter multidrug efflux pump expression (ABCB1, ABCC1, ABCC2 and ABCG2), which actively removes multiple anti-cancer drugs out of the tumour cells. d | Induction of the metastatic process of epithelial-to-mesenchymal transition (EMT) by GH acting to upregulate EMT mediators, including vimentin, N-cadherin, ZEB1, SLUG and the microRNA (miRNA) cluster comprising miRNA-96, miRNA-182 and miRNA-183 (miRNA-96–182-183), which altogether enable a phenotype switch of the tumour cells. The above processes are induced by autocrine and/or paracrine GH binding to GH receptor (GHR)-expressing cells in the tumour and TME, activating downstream signalling mediators JAK2, STAT5, SRC and ERK1/2. Therefore, a combination of a GHR antagonist with anticancer drugs is a transformative approach for highly effective tumour clearance, which has already been validated in preclinical models.

GH in the tumour microenvironment

Local GH production in specific tissues exerts an autocrine and/or paracrine effect, distinct from its endocrine role. In tissues such as the colon and breast, local GH production increase with age¹⁴³, forming a cellular niche in which oncogenic transformations could occur^143,146. Furthermore, local GH production in these tissues could support the growth and survival of a tumour by affecting the tumour microenvironment (TME)¹⁴⁷ (TABLE 1). Melmed and colleagues have elegantly elucidated the role of autocrine and/or paracrine GH in modulating the TME in support of tumour growth and survival through the ‘field cancerization’ paradigm¹⁴⁷. This model is a much overlooked but increasingly appreciated aspect of local GH action in oncology via the TME that is intuitive and critical, and supported by several observations. First, extra-pituitary sites of GH production are found in multiple tissues¹⁴⁴. Second, GHR is expressed in several cell types in the TME, including cancer-associated fibroblasts, adipocytes, immune cells, stromal cells¹⁴⁷ and endothelial cells¹⁴⁸, wherein GH exerts differential effects that are supportive of tumour growth^148–151. Third, endocrine GH-induced hepatic IGF1 has well-studied mitogenic effects that act on the tumour and TME^152,153.

Table 1 |.

Molecular actions of GH in human cancers

Molecular action	Effect	Cancer type	Refs.
Oncogenesis and tumour microenvironment
GH is a p53 target and in turn blocks p53 by a negative feedback loop	Tumour initiation	Colon cancer	158
GH induces DNA damage and suppresses DNA damage repair in normal colon tissue	Tumour initiation	Colon cancer	156,157,198
GH resistance reduces number and size of neoplasms in C3(1)/Tag-GHRKO mice	Tumour initiation	Breast cancer, prostate cancer	202,203
GH deficiency suppress DMBA-induced tumour development	Tumour initiation	Breast cancer	204
bGH mice have a higher DEN-induced hepatoma incidence while lit/lit (Ghrhr-knockout) mice have a lower one	Tumour initiation	Liver cancer	205,206
GH supplementation enables MNU-induced tumours in GH-deficient rats; tumours regress on GH withdrawal	Tumour initiation	Breast cancer	175
GH stabilizes hTERT through αCP1 and αCP2	Tumour initiation	Breast cancer	207
GH supports spontaneous neoplastic growth with age	Tumour initiation	Liver cancer	208
GH expression in mammary epithelia promotes oncogenic transformation via HOXA1	Tumour initiation	Breast cancer	209,210
GH produced in ageing colon accumulates DNA damage	Tumour initiation	Colon cancer	143
Invasive tumour growth
GH increases miRNA-96, miRNA-182 and miRNA-183 cluster targeting BRMS1L	EMT induction	Breast cancer	186,211
GH induces CHOP expression via p38 MAPK activation	EMT induction	Breast cancer	212
GH supports tumour angiogenesis	Invasive growth	Breast cancer	120
GH drives migration, invasion and metastasis in tumour xenografts	Invasive growth and EMT induction	Breast cancer	213–215
GH deficiency reduces tumour growth rate in lit/lit mice	Tumour progression	Prostate cancer	216
GH increases migration, invasion, proliferation and MMP levels	Invasive growth	Prostate cancer	217
GH blockade reduces xenograft growth rate	Invasive growth	Colon cancer, meningioma	218,219
GH blockade reduces migration, invasion and markers of EMT	EMT induction	Melanoma, pancreatic cancer	180,181
GH production in lungs guides melanoma metastases in mice	Metastasis	Melanoma	187
GH inhibits tumour apoptosis via PI3K–AKT signalling	Invasive growth	Gastric cancer	220
Therapy resistance
GH increases ABC transporter expressions via JAK2-STAT5 and SRC signalling	Multidrug resistance	Melanoma	172
GH increases MITF-dependent melanogenesis and drug sequestration via JAK2–STAT5 and SRC	Multidrug resistance	Melanoma	154
GH increases ABCG2 levels via JAK2–STAT5	Docetaxel resistance	Breast cancer	174
GH increases PI3K–AKT–MAPK and JAK2 signalling	Ruxolitinib resistance	Breast cancer	145
GH reduces tumour apoptosis and increase proliferation	Doxorubicin resistance	Breast cancer	175
GH increases post-irradiation clonogenicity and reduces radiation-induced DNA damage	Radiation resistance	Colorectal cancer, breast cancer, endometrial cancer	171,221–223
GH induces FOS expression	Doxorubicin resistance	Breast cancer	224,225
GH suppress DNA damage and apoptosis	Mitomycin-C resistance	Breast cancer, endometrial cancer	226
GH activates ERK1/2 and PKC and suppresses caspase activation	Multidrug resistance	Endometrial cancer	227
GH reduces pro-apoptotic BAX and PPARγ via STAT5B activation	PPARγ ligand resistance	Colon cancer	228
GH reduces BAX, BAD, caspase 3, caspase 8 and caspase 9	Methylmethanosulfonate resistance	Lymphoma	229
GH increases ABCG2 and markers of cancer stem cells (NANOG, ALDH1, CD24 and CD44)	Increased stemness	Liver cancer, breast cancer, colon cancer	163,177,178,184,213
GH is a part of senescence-associated secretory phenotype	Therapy evasion and relapse	Colon cancer	31

ABC, ATP-binding cassette; bGH, bovine growth hormone; C3(1)/Tag-GHRKO mice, mice from a cross between Ghr-knockout (GHRKO) mice and C3(1)/Tag mice, where females develop spontaneous mammary tumours; DEN, diethylnitrosamine; DMBA, dimethylbenz[a]anthracene; EMT, epithelial-to-mesenchymal transition; GH, growth hormone; hTERT, human telomerase reverse transcriptase; miRNA, microRNA; MMP, matrix metalloproteinases; MNU, N-methyl-N-nitrosourea.

Ageing or chemically inflicted DNA damage leads to activation of the tumour suppressor p53, which results in either apoptosis or induction of p53–p21 senescent pathway or a DNA damage repair (DDR) pathway¹⁴³. Chromatin immunoprecipitation assays reveal GH as a target for p53 binding, whereas transcriptomic and proteomic analyses confirm DNA damage induces local GH production in normal colon and tumour cells^154–156. In turn, locally produced GH exerts a feedback inhibition on p53 expression and diverts cellular commitment from senescence or apoptosis to proliferative survival¹⁵⁵. Additionally, autocrine and/or paracrine GH signalling abrogates the DDR pathway, which increases the risk of oncogenic mutations, as is observed in human colonic epithelial transformation^156–158. Increased colonic p53 expression was observed in Ghr-knockout (GHRKO) mice and Ames mice (which lack GH, prolactin and thyroid-stimulating hormone) compared with corresponding age-matched controls¹⁵⁸. Furthermore, elevated expression of DDR genes were observed in γ-irradiated primary fibroblasts of GH-deficient Lewis dwarf rats compared with those from wild-type rats¹⁵⁹. Elevated DDR gene expression was also observed in Snell mice (deficient in GH, prolactin and thyroid-stimulating hormone), GHRKO mice and Pappa-knockout mice (a model with alterations to the GH–IGF1 axis) compared with their respective wild-type controls¹⁶⁰. Together, these findings in GH-deficient animal models support the ‘onco-promoting’ role of GH in peripheral tissues.

An analysis of the National Cancer Institute genome-wide association study identified that out of 421 pathways containing 3,962 genes, GH signalling is the third most associated pathway with breast cancer susceptibility¹⁶¹. Seminal work highlighted a concerted role of GH and IGF1 in facilitating the functions of oestrogen and progesterone in normal mammary development¹⁶². The neoplastic effects of non-pituitary GH in breast cancer are particularly important, given that human GH binds to both GHR and prolactin receptor (PRLR), which are highly expressed together in ductal endothelium. These binding events result in a hyperplastic signalling cascade, which normally governs mammary development but can become an onco-driver under appropriate conditions¹⁶³. This signalling cascade is of particular importance in the TME in the context of fibrosis, one of the hallmarks of cancer, which enables invasive and metastatic growth and compromises antitumour immunity^164,165. As discussed above, GH is a potent inducer of fibrosis in multiple tissues, whereas the activation of the PRLR is also known to promote fibrosis in cancer via STAT3-dependent pathways^166,167. Downstream effectors of GH action, IGF1 and TGFβ, are also strongly implicated as autocrine and/or paracrine drivers of fibrosis¹⁶⁸, and can therefore have profound effects in the TME. Additionally, autocrine and/or paracrine GH has been established as a prominent component of SASP, wherein it promotes DNA damage accumulation in bystander cells and predisposes senescent cells to cell cycle re-entry and neoplastic transformation³¹. Of note, congenital GHRKO mice and adult-onset GHRKO (at age 6 months) mice have markedly reduced fatal neoplasms in both sexes, compared with wild-type littermates^169,170.

GH in cancer therapy resistance

A set of key pathways observed across all types of cancer that are unresponsive to therapy are inhibition of apoptosis, active drug efflux via ATP-binding cassette (ABC) transporters and a phenotype switch via epithelial-to-mesenchymal transition (EMT). GH expression induces resistance against apoptosis in mammary and endometrial tumour cells following irradiation¹⁷¹ and induces resistance against several chemotherapy treatments (mitomycin-C, doxorubicin, cisplatin, arsenic trioxide and ruxolitinib) in the same model¹⁴⁵. Similar effects of GH in driving refractoriness against chemotherapy (doxorubicin, cisplatin and paclitaxel) and targeted (vemurafenib) therapies are also observed in melanoma¹⁷². The mechanistic validation comes from the identification that tumoural GHR activation induces a STAT5–SRC–ERK1/2-mediated upregulation of multidrug ABC transporters in human melanoma^172,173. The ABC transporters impart resistance to a wide range of anticancer therapeutics by limiting their cytosolic retention through active efflux. Studies in human melanoma cells have shown autocrine GH signalling-induced upregulation of multidrug efflux transporters of the ABCB, ABCC and ABCG subtypes, which could be effectively blocked by GHR suppression¹⁷². Mouse xenograft models of human oestrogen receptor-negative breast cancer further confirm that GHR silencing reverses docetaxel resistance via downregulation of ABCG2 transporter expression¹⁷⁴. Furthermore, chemically induced mammary tumour establishment was possible in spontaneous dwarf rats (GH-deficient) only when they were supplemented with exogenous GH. Moreover, when GH supplementation was stopped and the rats were treated with doxorubicin, tumours regressed in the dwarf animals but not in GH-sufficient wild-type animals, which confirms a GH-dependent chemoresistance¹⁷⁵. Importantly, ABC transporters (such as ABCB1, ABCB5 and ABCG2) are known biomarkers for cancer stem cells (CSCs)¹⁷⁶, which are highly drug-resistant and are responsible for cancer relapse. Forced GH expression in liver cancer cells increased ABCG2 expression and conferred CSC properties by a JAK2–STAT3-mediated suppression of the tight junction protein claudin 1 (REF.¹⁷⁷). In addition forced GH expression stably increased CSC markers such as NANOG and SALL4 in both liver and colorectal cancer cells^177,178.

In cancers of the breast¹⁷⁹, colon¹⁵⁸, liver¹⁷⁷ and pancreas¹⁸⁰ and in melanoma¹⁸¹, GH promotes successful metastasis from a primary tumour by the process of EMT. This process enables a switch from a well-defined, polarized, basement membrane-adherent cell phenotype to a depolarized, migrating invasive tumour cell phenotype¹⁸². In addition to its angiogenic and lymphangiogenic effects, GH is known to induce EMT in the overlapping pathways of tissue fibrosis and cancer metastases¹⁸³. Suppression of the epithelial marker E-cadherin and upregulation of the mesenchymal markers N-cadherin and vimentin, and transcription factors ZEB1 and SLUG by recombinant hGH treatment occur in melanoma^173,181 and pancreatic¹⁸⁰ cancer cells and by autocrine GH in colorectal¹⁷⁸ and breast cancer cells¹⁸⁴. Attenuating the GHR reverses these effects. The most extensive evidence of GH in promoting EMT is from breast cancer, where autocrine human GH expression in mammary tumour cells leads to robust EMT induction and massive ECM remodelling. This effect occurs by GH downregulating adherence factors such as plakoglobin via hypermethylation mediated by DNA methyltransferases 3A and 3B¹⁸⁵, by GH increasing the production of ECM-degrading MMP2 and MMP9 (REF.¹⁸⁴) and by GH upregulating the microRNA (miRNA) cluster comprising miRNA-96, miRNA-182 and miRNA-183 leading to elevated ZEB1 and suppressed BRMS1L¹⁸⁶. In fact, in Dj1-knockout mice, which have incidental high production of GH in the lungs, this local GH enhances the metastasis of disseminated melanoma cells¹⁸⁷. The pronounced effect of GH in inducing fibrosis, as a part of ECM remodelling, further emphasizes a unique and critical role in promoting cancer therapy resistance¹⁸⁸ (TABLE 1).

Inhibition of GH action in cancer

Laron syndrome is congenital GH insensitivity due to inactivating mutations of GHR¹⁸⁹. Independent long-term follow-up studies on two of the largest cohorts of individuals with Laron syndrome in Israel and in Ecuador have shown no cases of malignancy, while the incidence rate in first-degree relatives is >20%^4,190 in both cohorts. Similarly, among patients with secondary GHD due to a GHRHR defect and patients with primary congenital isolated GH deficiency, cancer incidence is considerably suppressed compared with their relatives⁹⁵. The cohort of patients with isolated GH deficiency type 1B (arising due to GHRHR insufficiency), in Itabaianinha, Brazil, also show an absence of colorectal, prostate and breast cancers, unlike their relatives⁹⁵. Altogether, these epidemiological findings indicate that a congenitally absent or reduced GH action seems to be onco-protective. By contrast, large-scale, long-term follow-up studies in children with GHD treated with recombinant GH do not indicate an elevated risk of neoplastic developments.

Numerous studies estimating the risks of cancer and benign neoplasia in patients with acromegaly have shown elevated risks for specific cancer types and establish cancer as a major cause of mortality in these patients¹⁹¹. However, confounding factors exist, including surveillance bias, normalization of circulating levels of IGF1 due to various treatment options and difficulty in adequately comparing the cause of death in individuals with well-controlled acromegaly versus healthy control individuals.

Overall, the above discussion does provoke the question: can pharmacological suppression of GH action in the ageing population help tackle oncogenicity or improve cancer prognoses? Several GHR inhibitors are currently in development, which reflects a heightened pharmaceutical interest in targeting GH action in human disease^143,192,193. So far, targeting GH action in cancer using somatostatin analogues has shown no objective response in humans, whereas GHRH antagonists so far show promising effects in several preclinical models¹⁹⁴. Of note, somatostatin–GHRH control of GH production in the pituitary is seldom maintained in the context of the tumour. Pegvisomant, the first and only FDA-approved GHR antagonist, is highly successful in efficiently reducing serum concentrations of IGF1 in 70–90% of patients with acromegaly¹⁹⁵. Moreover, pegvisomant has shown efficacy in attenuating growth of multiple GH-expressing and GHR-expressing human cancers in preclinical models¹⁹⁶.

Although clinical validation of the efficacy of GHR antagonists in monotherapy or in combination with other anticancer therapies is awaited, more clarity is needed in understanding the mutually exclusive roles of GH and IGF1, as well as the ratio of GH-induced IGF1 and insulin-like growth factor-binding protein 3 (IGFBP3)¹⁹⁷ in a cancer-specific manner. Importantly, several studies have revealed that endocrine GH-induced hepatic production of IGF1 and IGFBP3 is not always observed in cultured tumour cells and mouse tumour xenografts and TME, wherein autocrine and/or paracrine GH has IGF1-independent oncogenic effects^143,198. A clinical trial in patients with cancer published in 2013 for a combination of pegvisomant and figitumumab (an anti-IGF1R monoclonal antibody) against solid tumours (breast, lung, prostate and colorectal tumours, and sarcoma) was initiated. Unfortunately, the study was terminated prematurely “due to lack of operational feasibility and halt of figitumumab development”¹⁹⁹. Therefore, whether targeting GHR in cancer is a clinically relevant option remains an open question, given the large volume of provocative studies mentioned above alongside the failure of IGF1R inhibitors and monoclonal antibodies in the treatment of cancer. Patient pre-screening for tumoural GHR overexpression and serum levels of IGF1–IGFBP3 might offer an important precision factor in this approach.

Conclusions

In this Review we emphasize the importance of GH in promoting fibrosis and its association with cardiovascular and cancer pathologies. GH induces the expression of a variety of genes encoding collagen. This is not surprising, as one of the most important actions of the hormone is stimulation of longitudinal growth in children, which involves the concerted actions of systemic GH and local IGF1 on the growth plate, a specialized connective tissue⁵¹. Moreover, GH administration also stimulates collagen synthesis and turnover in adult humans to such an extent that quantification of IGF1 and PIIINP in serum is now an approved assay for the determination of GH doping in sport¹⁸. Whether GH doping causes excessive fibrosis formation in, for example, muscle is unknown. Of note, however, GH administration in healthy adults does not increase either muscle strength or aerobic exercise capacity²⁰⁰. Indeed, after completion of longitudinal growth and somatic maturation, the major anabolic effect of GH administration seems to relate to collagen rather than muscle fibres¹³. When present in excess, GH results in fibrosis in a tissue-specific manner (FIG. 1) and the subsequent alterations to normal tissue function represent an added caution to be considered when utilizing GH supplementation in adults.

Data derived primarily from humans with elevated levels of GH (acromegaly) have shown a correlation with cardiovascular pathology, wherein fibrosis was identified as a contributor. In addition, patients with GHD often show a pro-inflammatory state without tissue fibrosis⁸¹. Together, the results suggest that a ‘normal’ amount of GH action is needed to ensure proper heart structure and cardiovascular function and either abnormally high or low levels of GH lead to cardiac pathology. An effect of the GH–IGF1 axis on vascular endothelial cells might be a common denominator of these effects that is reflected in adults with GHD with premature atherosclerosis.

In cancer, the effect of GH on fibrosis and endothelial cells makes it an important factor towards development of a detrimental TME. Clinically, cancer risks associated with GH replacement so far do not suggest any elevated risk of de novo neoplasm. However, since 1950 (REF.¹⁴²), experimental and animal data clearly demonstrate the mechanistic links of how GH, and its partner IGF1, can influence the development, progression, therapy resistance and metastases of multiple human and animal cancers that express the GHR and, thus, depict a definite ‘oncodriver’ role. Therefore, the rational approach of targeting GH action in patients with cancer who have high tumoural GHR expression should be complemented with appropriate screening for the status of tumoural GH action, including testing tumour biopsy samples for GH or GHR expression.

Future studies that consider the surrounding tissue milieu and inherent subcellular differences associated with the covert actions of GH will ultimately define the molecular processes involved. Additionally, the question of local versus endocrine GH and IGF1 participation in these phenotypes must be determined. In terms of growth, GH and IGF1 have both independent and overlapping functions²⁰¹. In this Review, we have updated the hidden or covert pathophysiological effects of GH and attempted to describe new molecular and cellular mechanisms responsible for them. Fibrosis is a common molecular ‘theme’ that seems to link these adverse phenotypes. Other fields of endocrinology clearly show that too much or too little exposure, or an abnormal time of exposure to a hormone, causes different adverse phenotypes. In the context of GH–IGF1, we conclude that either too much or too little of a good thing (GH) is ‘bad’ — a typical ‘Goldilocks effect’.

Key points.

• Growth hormone (GH) is important for growth and tissue remodelling, extracellular matrix formation and fibrosis.

• Patients with acromegaly, which is characterized by excessive circulating levels of GH, have increased cardiovascular mortality that is associated with hypertension and heart failure.

• Patients with GH deficiency have an increased risk of cardiovascular morbidity and mortality that is associated with cardiovascular risk factors and premature atherosclerosis.

• GH actions in cancer are particularly implicated in mechanisms of therapy resistance; for example, active drug efflux, the epithelial-to-mesenchymal transition, apoptosis inhibition and development of a tumour-supportive microenvironment.

• GH has a ‘Goldilocks effect’, where too little or too much can lead to poor clinical outcomes.