Growth Hormone and Neuronal Hemoglobin in the Brain—Roles in Neuroprotection and Neurodegenerative Diseases

It is known that GH and IGF-I are associated with circulating hemoglobin (Hb) levels in health and disease states [for reviews, see ( 1 , 2 )]. Recently, it has been discovered that Hb is also locally expressed in the brain ( 3 , 4 ) [for a review, see ( 5 )], which suggests that the oxygen-binding nature of hemoglobin may have a neuroprotective role in the brain. Although each of these topics have been reviewed previously, a comprehensive review of circulating and local Hb in relation to the neuroprotective and cognitive effects of GH and IGF-I has to the best of our knowledge not been presented. Here, we present an updated review of the potential links between these areas of neuroscience, in which we also identify knowledge gaps that warrant further investigation ( Figure 1 ). Figure 1 Model of crosstalk between neuronal hemoglobin (nHb), growth hormone (GH), and insulin-like growth factor 1 (IGF-I), and the known relations to brain diseases. With respect to interactions with neuronal Hb (nHb), the green arrows are beneficial, and the red arrows are detrimental pathways. Please, observe that circulating or free HbA has other effects not included in this figure. In general, GH and IGF-I levels within the normal ranges are optimal for the metabolism of oxygen and iron and for the mitochondria, which is beneficial for the brain. With age or for other reasons (e.g., genes, environment), the levels of GH and IGF-I decline, and the metabolism deteriorates leading to lower levels of nHb. This has adverse effects on degenerative brain diseases (AD, PD, MS), although in some aspects nHb may also respond in a manner that slows the disease progression. References for the depicted relationships are given in the main text. Hemoglobin Biochemistry and Relationship to Other Globins Hemoglobins are widespread in the biosphere and are found in all kingdoms of organisms, including prokaryotes, fungi, plants, and animals. The ancestral Hb gene existed phylogenetically before the divergence of plants and animals ( 6 ). There are at least four different globins in vertebrates ( Table 1 ): Hb, myoglobin (Mb), cytoglobin (Cygb), and neuroglobin (Ngb) [for reviews, see ( 5 , 9 )]. With respect to the brain, it is important to note that neuronal Hb (nHb) is present in parallel with the structurally different Ngb, with which it shares 25% protein sequence identity ( 10 ). Although the presence of both globins has been demonstrated ( 11 ), their relative amounts and the overlaps in their expression profiles are only partly known ( 11 ). Unlike Hb, Ngb is not responsive to erythropoietin (EPO). Nevertheless, Ngb appears to have similar neuroprotective functionalities, as it has been shown to correlate with serum IGF-I in hypoxic-ischemic conditions ( 12 ). Mammalian Hb is a tetramer of 574 amino acids, consisting of two α-globin (Hba; 141 amino acids) subunits and two β-globin (Hbb; 146 amino acids) subunits. Each of these subunits is bound to a heme group ( 13 ), ensuring four oxygen-binding groups per tetramer. The tetramer forms HbA (hemoglobin A) in red blood cells and with a different name nHb in the brain, however identical in structure. The heme group contains four pyrrole rings, the nitrogen atoms of which coordinate a central iron atom that binds weakly to oxygen ( 14 ). The main function of Hb is to transport oxygen from the lungs to other tissues in the body in exchange for carbon dioxide, which is transported back to the lungs. Molecular oxygen, O 2 , is essential for cellular respiration in aerobic organisms. Aerobic respiration takes place in the mitochondria of cells and requires oxygen to create adenosine triphosphate (ATP), which provides the energy to drive numerous processes in living cells. However, Hb also functions as a redox enzyme, owing to its ability to bind to hydrogen peroxide (H 2 O 2 ) ( 15 ). In addition, nHb may play a protective role against oxidative and nitrosative stresses by binding nitric oxide (NO), the strongest known ligand of the ferrous heme iron of Hb, with approximately 500,000-times higher affinity than oxygen ( 16 , 17 ). Although the brain constitutes only ≈2% of bodyweight, it consumes ≈20% of the oxygen in the resting body ( 18 ), demonstrating the high energy demand of the brain. Therefore, it is no surprise that Hb is expressed locally in the brain, and it has been attributed numerous functions therein. Table 1 Cell type-specific expression of vertebrate globins in the body and in the brain. Hemoglobin tetramer nHb/HbA Type of globin hemoglobin α hemoglobin β myoglobin cytoglobin neuroglobin Protein abbreviation HBA HBB MB CYGB NGB Transcript abbreviation used in Ms Hba Hbb Mb Cygb Ngb Identity vs Hbb 42% 100% 30% 33% 26% Identity vs Ngb 30% 26% 25% 30% 100% Cell type expression in body cardiac myocytes cardiac myocytes sympathetic nerves smooth muscle cells skeletal muscles skeletal muscles Relative expression in tissues neurons neurons N/A neurons neurons red blood cells 186,1 52

GH is a 191-amino acid, single-chain polypeptide. It is synthesized by and secreted from somatotropic cells of the pituitary gland and stimulates cell growth, reproduction, and regeneration ( 26 ). It is well-established that GH promotes postnatal growth and metabolism (for reviews, see ( 26 – 28 ). The secretion of GH is regulated by the balanced release of the two hypothalamic peptides: growth hormone-releasing hormone (GHRH) and growth hormone-inhibiting hormone (GHIH or somatostatin). These, in turn, are influenced by many physiologic stimulators (e.g., physical exercise, nutrition, sleep) and inhibitors (e.g., free fatty acids) ( 29 ). GH can act directly on tissues, although many of its effects are mediated by the downstream insulin-like growth factor-I (IGF-I), which is a 70-amino acid polypeptide hormone that is synthesized primarily in the liver ( 30 , 31 ). Furthermore, circulating IGF-I affects the levels of GH through the classic negative feedback loop formed by the hypothalamic-pituitary axis and the liver ( 26 ). The secretion levels of GH and IGF-I are highest during adolescence and decline thereafter in an age-related manner ( 32 ). GH exerts its actions via the GH receptor (GHR) ( 33 ), which is expressed in virtually every tissue of the body, including the brain ( 34 ). From the circulation, GH crosses the blood–brain barrier (BBB) ( 35 , 36 ), mostly through a process of passive diffusion ( 37 ), and binds to the GHR, which is expressed by both neurons and glial cells ( 38 ). The dimerization of GHR, which belongs to the type I cytokine receptor family, results in activation of the associated Janus kinase 2 (JAK2) and Src family kinases ( 34 ). The activation of JAK2 initiates tyrosine phosphorylation of the signal transducer and activator of transcription 5 (STAT5), the key transcription factor for GH ( 39 ). This results in the activation or repression of multiple genes ( 40 ), including the stimulation of IGF-I transcription in the liver ( 30 ). The EPO receptor (EPO-R) shares homology with the GHR, similarly leading to activation of the JAK2/STAT5 signal transduction pathway ( 41 ). It has also been shown that GH induces STAT5 expression in neurons ( 42 ), which is of interest with respect to Hb synthesis, as STAT5 activation independently drives erythropoiesis and myelopoiesis, both in vitro and in vivo , in the absence of EPO-signaling via EPO-R or JAK2 ( 43 ). Furthermore, experiments have revealed positive correlations between JAK2/STAT5 activation and Hb expression ( 44 ). Therefore, STAT5 appears to be a plausible link between GH and the synthesis of the heme portions of Hb, which is known to take place in mitochondria ( 45 ). In conclusion, GH appears to act through its key transcription factor STAT5 to enhance nHb expression, although direct evidence for this process is still lacking ( Figure 2 ). Figure 2 GH and/or IGF-I act as significant determinants of Hb concentrations by adapting adequate levels of Hb to the oxygen demand. STAT5, which is a key transcription factor of GH, is sufficient to allow erythropoiesis and myelopoiesis in vitro and in vivo. The PI3K/AKT/mTOR pathways lie downstream of the IGF-I-receptor. AKT phosphorylation of GATA-1 and its partner FOG-1 coordinates erythroid proliferation and differentiation. HIF-1 induced by IGF-I regulates the supply of oxygen to the cell by creating a balance between oxygen demand and oxygen supply. nHb, Neuronal hemoglobin; GH, growth hormone; IGF-I, insulin-like growth factor 1; STAT5, signal transducer and activator of transcription 5; PI3K/AKT/mTOR, phosphoinositide 3-kinases/protein kinase B/mammalian target of rapamycin; GATA-1, globin transcription factor 1; HIF-1, hypoxia inducible factor 1; FOG-1, friend-of-GATA1. References for the depicted relationships are given in main text.

It is well established that the level of circulating Hb is regulated mainly by EPO [for a review, see ( 69 )]. Moreover, it has been known for quite some time that GH and/or IGF-I, partly independently of EPO, affect to some extent the amount of blood Hb ( 70 – 77 ). For instance, at the end of a 5-year study of GH replacement therapy in children with GHD, the level of Hb was increased and there was also a positive correlation between the serum concentrations of IGF-I and Hb ( 78 ). The Hb-increasing effect of GH in children with GHD appears to be specific for erythropoiesis, since it was shown in another report that there were no effects of GH on white blood cells and platelets, regardless of prior history of anemia ( 79 ). In analogy, administration to healthy young adult men of a long-acting GHRH analog, which stimulates GH release into the circulation, induced upregulation of the Hbb protein by 0.5–1.0 log units, corresponding to 7–10 Hb units ( 80 ). Moreover, GH and/or IGF-I have been reported to associate with Hb in elderly subjects ( 81 ). The various reports of increased Hb following GH administration can be explained by direct stimulatory effects of GH and IGF-I on erythroid cells, in combination with a more generalized indirect effect of GH in facilitating physical activity through increased muscle performance and well-being ( 82 ), which in turn enhances general health and Hb production. Conceptually, this may be mediated by the general anabolic effect of GH, which is considered to stimulate protein and fat metabolism, which in turn requires increases in oxygen transportation and Hb levels ( 83 ). IGF-I has also been shown to stimulate erythropoiesis in vivo independently of GH, which means that it could indirectly mediate GH effects on EPO synthesis, erythropoiesis, and Hb regulation. As compared to EPO, which has a substantial and significant effect on Hb levels, the effect of serum IGF-I has been shown to be less-prominent than but still independent of circulating EPO ( 84 ). Moreover, it has been suggested that IGF-I is a factor that coordinates red blood cell formation with organ and body growth, and this would account for the adaptation of red blood cell mass to whole body mass ( 84 ). From another publication, it can be calculated that an increase of 100 units of serum IGF-I (ng/ml) is associated with an increase of 4.0 units of Hb (g/L) in males and 7.5 units in females ( 81 ). A difference in serum IGF-I of 100 ng/ml is rather large considering that the mean serum IGF-I level in middle-aged men and women is in the range of 150–200 ng/ml. The serum IGF-I is also about 100–200 ng/ml higher in young adults than in middle-aged subjects ( 85 ). Taken together, these findings indicate definite albeit moderate effects of endocrine GH and IGF-I on circulating Hb.

Stroke and Hemoglobin Hb is important in relation to brain injuries in several aspects. The most common brain injuries are ischemic stroke, hemorrhagic stroke, and perinatal asphyxia. Ischemic strokes, regardless of etiology, result in an inadequate blood supply that leads to cell death within the affected tissue [for reviews, see ( 98 – 100 )]. Ischemic pre- or post-conditioning is an experimental stroke paradigm that can reduce the ischemic injury. Interestingly, ischemic preconditioning to hypoxia in vitro robustly increases nHb expression in neuronal cultures from E-17 rats ( 101 ). A plausible explanation for this is that Hb inhibits oxidative stress-induced mitochondrial dysfunction and caspase activation via the protein kinase A, protein kinase C, and mitogen-activated protein (MAP)-kinase signal transduction pathways ( 102 ). In contrast to the intracellular expression of Hb, extracellular Hb is highly toxic to the brain and constitutes a mechanism of damage after ischemic stroke, and more profoundly so after hemorrhagic stroke ( 103 ). Specifically, free Hb induces the widespread and concentration-dependent death of neocortical neurons in vitro ( 104 , 105 ). Similarly, intracerebral injections of free Hb or its degradation products, including heme, induce brain injury in vivo ( 106 ). While the exact mechanisms are unknown, it appears that extracellular Hb contributes to oxidative stress and iron accumulation in neurons, resulting in a deleterious cycle of heme breakdown, Hb denaturation, and, over time, neurodegeneration ( 107 ). Overall, it seems that extracellular free Hb activates deleterious cellular pathways, whereas intracellular Hb is primarily protective ( 20 ). In addition, circulating Hb may affect stroke risk, severity, and outcome, presumably through mechanisms other than the neuronal expression of Hb. For example, circulating Hb seems to influence stroke risk, since anemia increases the risk of stroke and worsens stroke outcomes ( 108 ). Moreover, in acute ischemic stroke, low levels of circulating Hb independently predict short- and long-term mortality ( 109 ), correlate inversely with larger stroke volumes ( 110 ), and are strongly associated with subsequent decreases in the Hb and hematocrit levels ( 111 ). Interestingly, in comparison to anemia, polycytemia with high levels of Hb (≥15 g/L in men, ≥14 g/L in women) has been associated with twice as many cerebral infarctions ( 112 , 113 ). This is thought to be mediated by increased viscosity of the blood and comorbidities, causing an increased risk of thrombosis ( 113 ). Furthermore, once the clot has been formed, it starts to resolve, which in turn causes erythrocyte lysis and the release of free Hb, which leads to an overload of free, toxic iron ( 103 ). Although several studies have shown that GH and IGF-I either protect against injury or promote recovery, these effects have not so far been linked to Hb regulation. It has been shown that GH protects the brain against hypoxic-ischemic injuries (HI) in neonatal rodents ( 114 ) and increases motor and cognitive recovery in adult rodents ( 115 – 117 ). Furthermore, the positive effects on functional recovery of GH treatment also appear to be present in humans who are suffering from chronic stroke, as shown by the increased amplitude of blood oxygen level-dependent signals in the brain, which are dependent upon the different magnetic properties of oxy-Hb and deoxy-Hb ( 118 ). That the effects of systemic GH treatment are mediated by local or endocrine IGF-I with respect to injuries is not clearly elucidated but certainly possible, as the IGF-I system in the brain shows marked changes in response to transient neural injuries. For instance, in asphyxia, there is an increase in the level of Igf1 in glial cells in the region of the injury ( 119 ), suggesting that the IGF-I system is involved in neuroprotection. IGF-I administration has been shown to be both neuroprotective ( 120 ) and recovery-promoting ( 121 ) in studies of experimental ischemic stroke [for a review, see ( 59 )] and the articles ( 116 , 117 , 122 ). However, the role of circulating IGF-I and its associations with functional outcome in human stroke appear to be complex. While high serum levels of IGF-I have been associated with both worse outcomes ( 123 , 124 ) and greater stroke severities, they are also linked to better long-term recovery ( 85 , 125 ). These discrepancies can partially be explained or confounded by the fact that the levels of serum IGF-I are lower in aged, malnourished, and/or sedentary subjects ( 71 ) and may depend on the post-stroke sampling time-point, as suggested previously ( 122 ). Although the absolute levels of serum IGF-I appear to predict functional outcome, post-stroke dynamic changes (decreases) in the levels of serum IGF-I may also be at least as important ( 122 , 126 ). Some studies have shown a weak-to-moderate correlation between serum IGF-I and Hb ( 64 , 65 , 68 ), a relationship tha

Alterations in cerebral oxygen metabolism have been found to be associated with many of the most common degenerative brain disorders, such as Alzheimer’s disease (AD), Parkinson’s disease (PD), and multiple sclerosis (MS), and the CNS expression of Hbb has been implicated in these diseases ( 135 ). Alzheimer’s Disease AD is the most common cause of dementia, and the risk of AD is higher when there is a history of brain lesions, such as TBI ( 136 , 137 ) or stroke [for review, see ( 138 )]. Early clinical symptoms include difficulties in remembering recent events and conversations, and ultimately there are reduced abilities to walk, speak and swallow. The histopathologic hallmarks of AD are the progressive accumulation of beta-amyloid (Aβ) peptide plaques outside neurons in the brain, twisted strands (tangles) of tau protein inside neurons ( 139 ), and a general hyperphosphorylation of the tau protein ( 140 ). Eventually, these changes result in damage to and death of neurons [Alzheimer’s disease facts and figures ( 141 )]. As is the case for AD and brain Hb expression, several studies have indicated a role for Hb in AD. In humans, Hb protein has been detected in the neurons and glial cells of post-mortem AD brains. Furthermore, the level of local Hb protein has been found to be increased in AD brains ( 142 ). This indicates that either the upregulation of Hb expression or leakage of Hb from the circulation into the brain parenchyma (due to disruption of the BBB) is involved in AD pathogenesis ( 142 ). In animals, Hbb is upregulated with normal aging, as well as in a amyloid precursor protein/PS1 transgenic mice (murine model of AD), possibly serving as a compensatory mechanism for hypoxia ( 143 ). In addition, when micro-hemorrhaging occurs as the result of aging, TBI or cerebrovascular disease, Hb from red blood cells can bind to the Aβ peptide via the iron-containing heme moiety and accelerate its aggregation ( 143 ). Furthermore, altered iron metabolism has been observed in the CNS of patients with a variety of neurodegenerative diseases, including AD. The concomitant occurrence of a decrease in local Hb expression in AD and the increased level of free iron, which is highly toxic due to the creation of reactive oxygen species ( 144 ), is possibly coincidental, although it indicates the importance of iron metabolism. Epidemiologic studies have suggested strong associations between anemia, decline of cognitive function, and AD, with an inverse relationship noted between Hb and AD ( 145 – 147 ). Moreover, in a large population-based study, it has been shown that low Hb levels are associated with increased long-term risks of dementia (34%) and AD (41%) ( 148 ). Aβ activates a molecular response pathway that leads to strong downregulation of Hb gene expression in the brain, which could damage cells due to defective oxygen homeostasis ( 144 ). Reduced concentrations of the Hb α-chain and β-chain proteins have been detected in neurons with granular or punctuate hyperphosphorylated tau deposits, in AD neurons with tangles in the hippocampus and frontal cortex, and in AD neurons in the amygdala ( 135 ). These findings are in line with a report showing significantly lower levels of mean oxygen metabolism in the AD cortex ( 149 ). Loss of Hb is specific, as related transcripts, i.e., for Ngb and the EPO receptor, are expressed at normal levels ( 135 ). There are strong indications that IGF-I is of importance in the development and progression of AD. A deficiency of liver-derived circulating IGF-I in mice resulted in debilitated spatial learning and memory ( 150 ), as well as reduced exploratory activity ( 151 ). In experimental AD, the passage of IGF-I through the BBB has been shown to be reduced ( 152 ), while systemic IGF-I infusions increased Aβ clearance, thereby reducing the brain level of Aβ ( 152 ). In a later study, a specific BBB disruption allowing serum IGF-I entrance into the rodent brain was demonstrated, as the choroid plexus receptor megalin/LRP2 induced IGF-I transport across the BBB and protected against AD, and in AD mice, megalin abundance was decreased ( 54 ). Furthermore, IGF-I not only interacts with Aβ metabolism and distribution, but also with tau metabolism, as evidenced by the finding that the phosphorylation of tau is markedly increased in the IGF-I-null brain in mice ( 153 ). In a prospective, population-based study of older persons, low IGF-I status (<9.4 nmol/L) was associated with both lower levels of cognitive ability and a sharper decline in cognition ( 154 ). In a case-control study, lower serum IGF-I levels were significantly related to dementia that was clinically diagnosed as AD ( 155 ). However, discrepant results have been reported with respect to the serum and cerebrospinal fluid (CSF) levels of IGF-I in patients with AD. It has been postulated that these discrepancies can be explained in terms of resistance to IGF-I action early in AD development (reflected as increased s

MS is primarily an inflammatory disorder of the brain and spinal cord, in which focal lymphocytic infiltration leads to damage to myelin and axons. Initially, the inflammation is transient and re-myelination occurs, but this most often does not persist and the result is chronic neurodegeneration and progressive disability [for a review, see ( 179 )]. Interestingly, there are some lines of evidence to suggest a link to Hb expression in the CNS. In a study of eight monozygotic twin pairs discordant for MS ( 180 ), >2-fold upregulation of six gene transcripts was detected in mononuclear cells isolated from the peripheral blood, amongst them Hba2 and Hbb , suggesting some role for these genes in MS pathogenesis ( 180 ). This is in line with a report showing increased expression of local Hbb protein in the cortex, as well as intense Hb staining in the pyramidal neuronal cell bodies of patients with MS ( 181 ). Specifically, in MS, it has been suggested that neuronal Hbb is part of a mechanism linking neuronal energetics with epigenetic changes in the nucleus, providing neuroprotection by supporting neuronal metabolism ( 24 ). Dysregulation of Hbb transport into the nuclei of pyramidal neurons in MS has been suggested by the findings that in patients with MS/experimental MS there is a reduction of the histone marker H3K4me3 and downregulation of genes involved in mitochondrial respiration ( 24 ). In contrast, free Hb originating from dying erythrocytes is detrimental in MS pathogenesis, as it causes damage to the myelin and BBB ( 182 ). In addition, an association between anemia and MS has been described ( 183 ). There are also indications that the GH/IGF-I system is involved in MS. First, the levels of CSF-GH were found to be 2-fold lower in patients with MS ( 184 ). In contrast to CSF levels of GH, the serum levels of GH were unchanged, which suggests normal pituitary function and, instead, indicates malfunction of the BBB or in the local production of GH in the brain. The levels of IGF-I in serum and CSF also showed a trend toward divergence ( 184 ). Furthermore, MS and higher levels of disability seem to be associated with a reduction in the bioavailability of IGF-I, possibly due to high serum concentrations of IGF-binding protein-3 ( 185 ). Taken together, these findings suggest that Hbb dysregulation and GH abnormalities are present in MS, although to our knowledge, comprehensive evidence for a direct linkage between the two factors is still lacking.

MW wrote the first draft of the manuscript. DÅ contributed significantly to the second version. MW, JS, LK, MÅ, RM, GK, JI, and DÅ contributed to manuscript revision, read, and approved the submitted version. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.