Erythropoietin and Its Derivatives: Mechanisms of Neuroprotection and Challenges in Clinical Translation

Neurodegeneration and Neurogenesis After Traumatic Brain Injury

The pathophysiologic damage that results from traumatic brain injury (TBI) is shaped by both primary and secondary mechanisms of injury. Primary injury is a consequence of both the initial mechanical impact sustained by brain tissue and/or the inertial forces that cause cellular strain and membrane damage ( ). Secondary injury includes ionic imbalance, the release of excitatory amino acids, mitochondrial dysfunction/oxidative damage, inflammation, apoptosis, and diffuse axonal injury ( ). A compromised blood–brain barrier (BBB) contributes to the propagation of vasogenic and cytotoxic edema and allows the infiltration of inflammatory cytokines and chemokines into the brain parenchyma, ultimately promoting the infiltration of inflammatory cells ( ) ( Fig. 4.1A and B ). The secondary injury cascade ultimately results in the activation of proteases (eg, calpains and caspases), which contribute to cell death via either apoptosis or necrosis ( ). The mechanisms of secondary injury develop within hours to days after the primary injury and thereby provide a realistic window for clinically relevant therapeutic interventions.

Cellular damage resulting from secondary injury mechanisms in TBI are followed by a restorative phase during which the brain attempts to remodel itself in an effort to compensate for damage. Posttraumatic plasticity has been shown to involve neurogenesis, gliogenensis, angiogenesis, synaptic plasticity, and axonal sprouting ( ). Accordingly, compensatory plasticity is believed to underlie the spontaneous recovery of function that takes place after TBI, with its ultimate extent depending on severity of TBI, age, and other factors ( ). It is prudent to note that the processes underlying plasticity can be augmented by growth factors (and other growth related proteins) and thus may be modified/improved via pharmacologic interventions ( ).

Of prime importance is the widely accepted view that a decrease in cerebral oxygenation is an important pathologic component of secondary brain injury. A study of severe TBI patients showed that a hemoglobin (Hb) level lower than <9 g/dL was associated with decreased brain tissue oxygenation (PbtO ₂ ), while an anemic status (ie, Hb <9 g/dL) and a simultaneously low PbtO ₂ (ie, <20 mmHg) increased the risk of unfavorable clinical outcomes ( ). In line with the aforementioned, a study has shown that patients with severe TBI and low Hb levels (10.7 ± 0.9 g/dL) had lower survival rate than those with higher Hb levels (12.9 ± 0.4 g/dL) ( ). As such, a number of therapies have demonstrated preclinical efficacy via a modulation of tissue oxygenation and have thus improved neurologic function(s) in animal models of TBI ( ). Of these, the agent with the most comprehensive and convincing preclinical data (suggestive of clinical efficacy) in TBI is erythropoietin (EPO) ( ).

Erythropoietin as an Endogenous Neuroprotective Molecule

The EPO gene is located on chromosome 7 and encodes for a ∼30-kDa glycoprotein, which is expressed mainly in the peritubular fibroblasts of the kidney and in liver hepatocytes ( ). When tissue becomes hypoxic, hypoxia-inducible factor-1α (HIF-1α) is stabilized and activates the EPO gene ( ). EPO production and the resultant increase in erythropoiesis leads to improved tissue oxygenation and a subsequent downregulation of HIF ( ). In addition to the kidneys, EPO has been shown to be produced within the brain by both astrocytes and neurons ( ). It is therefore unsurprising that the EPO system plays an important role during normal central nervous system (CNS) development ( ). Mice with targeted deletions of EPO display an increase in CNS apoptosis before the development of severe anemia, suggesting that the EPO required for normal brain development acts in a manner distinct from its role in erythropoiesis ( ).

EPO expression within the brain increases after pathologic insults and injuries. Most notably, local CNS EPO production is increased during hypoxia (eg, experimental anemia and hypoxia induced via carbon monoxide exposure resulted in 20-fold upregulation of EPO mRNA) ( ). Increased local production of endogenous CNS EPO suggests that EPO may act in a paracrine and/or autocrine fashion in an effort to facilitate endogenous neuroprotection.

Neuroprotective effects of EPO have been extensively demonstrated for a number of cell types within the brain. For example, EPO was able to protect hippocampal and cortical neurons from glutamate neurotoxicity and also reverse hypoxia-induced hippocampal neuronal death when applied in conjunction with the hypoxic insult ( ). EPO also aids the maturation of glia (ie, oligodendrocytes and astrocytes) ( ). Finally, EPO has been shown to maintain microglial integrity during oxidative stress induced by oxygen and glucose deprivation, thereby fostering tissue repair and reorganization ( ).

The majority of the neuroprotective and other effects ascribed to EPO have been linked to the binding of EPO to erythropoietin receptors (EPO-R) ( ). EPO-R is a type I cytokine receptor that possesses a single transmembrane domain which lacks an intrinsic tyrosine kinase domain ( ). Binding of EPO to the EPO-R changes the conformation of the receptor to a dimer thereby activating the Janus family tyrosine protein kinase-2 (Jak2) and downstream signal transduction pathways which include phosphatidylinositol-3 kinase (PI3K), mitogen-activated protein kinases (MAPK), and signal transducers and activators of transcription-5 (STAT5) ( ) ( Fig. 4.2 ).

EPO-R expression within the brain has been demonstrated during both development and adulthood in humans and other higher mammals ( ). The highest levels of EPO-R expression are detected in the neural tube of the mouse at embryonic day 10.5 and are comparable to levels found within adult hematopoietic tissue ( ). In the adult brain EPO-R are predominantly located within the midbrain, cortex, hippocampus, and capsula interna ( ). Within these niches, EPO-R are present on various cell types, such as neurons, astrocytes, oligodendrocytes, and brain endothelial cells ( ). Human, rat, and mouse brain capillaries and stellate astrocytes display an intense immunoreactivity for EPO-R, while large vessels and other astroglia do not express EPO-R ( ). Transmission electron microscopy has shown EPO-R on the surface of capillary endothelial cells and within the astrocytic endfeet surrounding them ( ). Interestingly, brain capillary endothelial cells express two forms of EPO-R mRNA, which are differentially activated after the administration of human recombinant EPO (rHu EPO) ( ). Such a multicellular distribution of EPO-R again suggests that EPO may function more generally (ie, beyond driving erythropoiesis) to aid in tissue development and repair.

Beyond normal development/homeostatic biology EPO-R expression is increased via a myriad of pathologic conditions. Mild hypoxia increases EPO-R expression in the ventral limbic region of the rat brain ( ), and both EPO and EPO-R are upregulated in the human brain after ischemic infarcts and/or hypoxic damage ( ). Increased expression of EPO-R has also been noted in astrocytes of the hippocampus in patients with sporadic Alzheimer’s disease (AD) and mild cognitive impairment. This suggests that the upregulation of EPO-R in the temporal, cortical, and hippocampal astrocytes may be an attempt at neuroprotection during the pathogenesis of sporadic AD ( ).

EPO is a large molecule (molecular weight of ∼30 kDa), yet its ability to cross an intact BBB has been confirmed ( ). Murine, human, and darbepoetin alfa (an analog of human EPO in clinical use) EPO are all capable of crossing the murine BBB at the same rate as albumin suggesting that these proteins cross via extracellular pathways. Accordingly, EPO concentration in the brain was shown to peak ∼3 h after intravenous injection ( ). Of note, endogenous EPO has been found in the cerebrospinal fluid of patients with TBI at levels that correlated with the overall degree of BBB dysfunction ( ).

In summary, numerous animal studies have confirmed EPO’s neuroprotective, antiapoptotic, neurogenic, and angiogenic properties following both TBI and ischemic stroke ( ). Further, studies have demonstrated EPO’s ability to decrease microglial activation/activity after trauma possibly via the attenuation of secondary mechanisms of brain injury ( ). As such, preclinical studies in TBI, stroke, and other neurological disorders support the hypothesis that EPO may be therapeutic for both ischemic and traumatic brain injuries in humans.

Derivatives of Erythropoietin

One of the principal concerns underlying EPO therapy in human patients is that the administration of several doses may ultimately lead to potentially harmful increases of Hb with a concordant increase in the incidence of thrombotic sequela ( ). Therefore, a number of EPO derivatives which lack hematopoietic effects have been generated and studied both in vitro and in vivo in an effort to clarify their therapeutic potential ( ).

The most studied derivative thus far is carbamylated erythropoietin (CEPO). Carbamylation of the lysines within EPO interferes with its binding to the EPO-R on hematopoietic cells and thus prevents increases in hematopoietic activity after long-term administration ( ). The precise molecular mechanisms governing the actions of CEPO remain to be elucidated; because CEPO is not able to bind to classical EPO-R it may act by engaging an alternative receptor ( ). Critically, peripherally administered CEPO is able to cross the BBB and has been shown to be within the brain parenchyma and CSF compartments after middle cerebral artery occlusion (MCAO) ( ). A number of experimental studies in both stroke and TBI have demonstrated the neuroprotective effects of CEPO, which are comparable to effects produced by wild-type EPO ( ). For example, CEPO is capable of reducing CA1 pyramidal cell death in the hippocampus after TBI in a manner similar to EPO ( ). Further, postischemic treatments with CEPO reduced infarct size and concordantly improved neurological outcomes after focal cerebral ischemia in a manner analogous to EPO ( ). Several mechanisms associated with CEPO-mediated neuroprotection have been observed. First, similar to EPO, CEPO has antiapoptotic properties with regard to on-damaged neuronal cells ( ). Second, both EPO and CEPO promote angiogenesis via endothelial progenitor cells (EPC), which are derived from hematopoietic CD34 ⁺ progenitors ( ). EPC levels have been associated with stroke severity ( ) ( Fig. 4.3 ), and increases in EPC levels have been linked with functional recovery following ischemic stroke ( ). Studies on human EPC and erythroid cell lines have shown that CEPO prevents EPC death via a reduction in apoptosis ( ). Third, CEPO significantly increases neural progenitor cell differentiation into neurons ( ). Fourth, CEPO has been shown capable of reducing neuroinflammation (as measured by microglia activation and attenuation of polymorphonuclear leukocyte infiltration) after experimental stroke ( ). Despite such promising results in preclinical studies, human studies using CEPO in both stroke (NCT00756249) and Friedreich’s Ataxia (NCT01016366) have not reported any clinical benefits yet have reinforced the drug’s overall safety profile.

Another EPO analog, asialoEPO, was developed under the premise that a continuous presence of EPO is required for the production of red blood cells while only a brief exposure is necessary to induce neuroprotection. Accordingly, intravenously injected asialoEPO exhibits a plasma half-life of 1.4 min and is below the lower limit of detection within the systemic circulation at 1–2 h. This is in stark contrast to rhEPO, which has a plasma half-life of ∼5.6 h; in line with such a hypothesis multiple doses of asialoEPO do not increase hematocrit (Ht) levels ( ). Interestingly, the binding affinity of asialoEPO to an EPO-R-Fc construct falls within a similar range to that of rhEPO, thereby suggesting that asialoEPO is similar to EPO in its mechanism of action ( ); this differs strikingly from aforementioned CEPO ( ). AsialoEPO is able to cross the BBB and has been detected in the CSF of rats having undergone MCAO when it was administered for 4 days through the pump at a dose of 20 μg/kg per 24 h ( ). In studies on experimental stroke and models of traumatic spinal injury, asialoEPO acted as a nonerythropoietic cytokine with a broad range of neuroprotective activities ( ). Neuroprotection by asialoEPO was confirmed via a reduction of infarct size in an MCAO model when administered 90 min after the occlusion; further the administration of asialoEPO also diminished the zone of experimental spinal cord injury ( ). Another putative mechanism of action by asialoEPO (as shown in a model of a spinal cord injury) includes the promotion of glial activation ( ). Finally, both the safety and putative efficacy of asialoEPO were further demonstrated in a model of amyotrophic lateral sclerosis ( ).

A new nonhematopoietic EPO-R agonist ARA290 has displayed promising cytoprotective capacities and efficacy in both preclinical and clinical studies of metabolic neuropathy ( ). Currently administration of ARA290 shows potential in cardiac disorders, in neuropathic pain, and in sarcoidosis-induced chronic neuropathic pain ( ), although, studies focusing on CNS pathology have yet to be published.