Macrophage migration inhibitory factor (MIF) in the development and progression of pulmonary arterial hypertension

Macrophage migration inhibitory factor (MIF) has been described as a pro-inflammatory cytokine and regulator of neuro-endocrine function. It plays an important upstream role in the inflammatory cascade by promoting the release of other inflammatory cytokines such as TNF-alpha and IL-6, ultimately triggering a chronic inflammatory immune response. As lungs can synthesize and release MIF, many studies have investigated the potential role of MIF as a biomarker in assessment of patients with pulmonary arterial hypertension (PAH) and using anti-MIFs as a new therapeutic modality for PAH.


PULMONARY ARTERIAL HYPERTENSION (PAH)
PAH is a devastating disease that leads to progressive systemic hypoxemia, right ventricular failure and death 1 . PAH demonstrates rapid deterioration after diagnosis, with an average survival time for primary pulmonary hypertension of only 2.8 years, and an estimated 5 year survival rate of between 21-34% 2,3 . Although previously considered a rare disease, over the last two decades, there has been an increase in the diagnosis possibly due to the increased awareness of the physician and improved diagnosis methods 4 . PAH can occur in association with chronic lung disorders, with hypoxia playing a pivotal role in the etiology. Hypoxia induces pulmonary vessel constriction and persistent hypoxia results in pulmonary vascular remodeling resulting in increased vessel wall thickness and narrowing of the vessel 5,6 ( Figure 1). Pulmonary vascular remodeling chronically increases pulmonary vascular resistance (PVR), leading to right ventricular failure, decreased left ventricular preload and reduced cardiac output. The remodeling also causes mismatch of blood flow and ventilation (V/Q), which, together with decreased cardiac output and possible cardiac shunt, lead to further hypoxia. A major factor in the rapid progression of PAH symptoms may be due in part to the creation of a vicious cycle: PAH can be initiated by hypoxia, itself causes hypoxia, and hypoxia in return exacerbates PAH.
Current therapies for PAH are relatively ineffective and their benefits limited to improving ability to exercise. They include prostacyclin and its analogues, endothelin receptor antagonists, and phosphodiesterase type 5 inhibitors [7][8][9] . Notably, current therapies do not significantly improve mortality, rate of clinical progression of disease, or WHO functional class 10,11 . The limitation of current treatment suggests the need for a better understanding of the pathogenesis and identification of new therapeutic targets for this lethal disease.

VASCULAR REMODELING
Vascular remodeling leads to increased vessel wall thickness and narrowing of the vessel lumen 5,6 . Endothelial, fibroblasts and smooth muscle (SMC) are the principal intimal, adventitial and medial cells of the vascular wall, respectively. Chronic hypoxia induces pulmonary vascular cell proliferation and remodeling, but the mechanisms involved remain unclear. While morphological changes to the intima are observed, they are usually minimal 12 . Chronic hypoxia induces structural changes to the pulmonary arteries including the appearance of SMC-like myofibroblasts expressing α-smooth muscle actin, in previously non-muscularized vessels. While hypoxia-induced remodeling is associated with medial hypertrophy, direct stimulation of SMC proliferation by hypoxia remains controversial 13 .
There are reports of hypoxia-driven smooth muscle cell proliferation 14,15 , but several in vitro studies have shown that hypoxia does not directly increase SMC proliferation [16][17][18] or may actually decrease proliferation 19,20 . However, fibroblasts which are less differentiated than the other two cell types, have a greater proliferative response to hypoxia than either endothelial cells or SMC [20][21][22][23] . Fibroblasts are uniquely positioned in the scheme of remodeling being able to rapidly proliferate, contract, migrate, synthesize cytokines and other mediators, and transdifferentiate into other cell types such as the SMC-like myofibroblast 24 . Animal models indicate that the adventitia undergoes the earliest and most profound changes under hypoxic conditions 25,26 and in vitro, hypoxia induces fibroblast proliferation in the absence of any exogenous co-mitogen 27 .

MACROPHAGE MIGRATION INHIBITORY FACTOR (MIF)
MIF was originally described over 50 years ago in studies of delayed hypersensitivity in which it was suggested that lymphoid cells released soluble materials that inhibited the random migration of peritoneal exudate cells in vitro 28,29 . This resulted in the original nomenclature, but since that time, a multiplicity of functions have been assigned to the molecule has led to the less formal epithet of simply MIF 30 .
MIF is a conserved protein of 12.5 kDa, of which homologues can be found in plants, nematodes and vertebrates. In its active form, it is a homotrimer which is associated with two distinct enzymatic activities. The first, a phenylpyruvate tautomerase 31 , residing within the hydrophobic pocket between adjacent monomers. This tautomerase activity and structural relationship are similar to a second human protein, D-dopachrome tautomerase (D-DT) 32 , with which human MIF shares around 34% sequence homology. The similarities between structures and activities of the two molecules have led to the D-DT designation as 'MIF-2' 33 .
While the relevance of the tautomerase activity to human disease remains unclear, the hydrophobic pocket within which the activity is located binds to cell surface CD74 molecules 34 thereby activating its CD44 co-receptor initiating cellular activation, cell proliferation and inhibition of apoptosis 35,36 . The second MIF catalytic activity is a thiol-protein oxidoreductase 37 within a conserved [Cys 57 -Ala-Leu-Cys 60 ] region of the molecule. This antioxidant activity has been shown to be particularly relevant to situations of cellular oxidative stress [38][39][40][41] . Thus, these two characteristics of MIF alone make it a molecule of interest in the development and progression of pulmonary arterial hypertension.
Furthermore, MIF is a potent proinflammatory cytokine involved in both chronic and late-stage acute inflammation and plays a key role in inflammatory cell proliferation. It is stored within cytoplasmic pools and can be released to extracellular compartments following interaction with the Golgi complex-associated protein p115 42 . Extracellular release of MIF is initiated by a series of factors including cytokines, materials from Gram-positive or Gram-negative bacteria 43,44 oxidative stress 45 , and steroids 30 . MIF is involved in wound healing 46 , synoviocyte proliferation 47 , and transformation from chronic inflammation to tumorigenesis 48,49 .
MIF increases proliferation of many cell types 48,50-54 including fibroblasts 36,46,54-57 , endothelial cells 58 , and SMCs 59,60 . MIF also appears to be involved in systemic vascular remodeling, including carotid intima-media thickening 60,61 , and restenosis after vascular injury 60 . One possible mechanism of MIF's proliferative effects is through the inhibition of p53, an endogenous cell cycling inhibitor that induces G1 stage arrest or apoptosis 62,63 , and is also involved in pulmonary arterial smooth muscle cell proliferation 50 .

MIF AND THE LUNG
While early studies showed the anterior pituitary as a possible source of circulating MIF 64 , more recent studies have shown that in severe acute inflammation, the lungs are a major source of the circulating MIF, which can have a profound effect on cardiovascular function 65,66 . While MIF is a normal component in the epithelial lining fluid of the lung there is a significant increase in accumulation in the alveolar airspaces in the acute respiratory distress syndrome (ARDS) 67 . The increased lung MIF is both at the gene expression and protein levels and can be associated with haplotypes located in the 3 end of the MIF gene 68 . Furthermore, the increased MIF concentrations due to a particular challenge, in both the alveolar spaces and the plasma, are age-dependent 69 . Thus, while increased extracellular MIF emanating from the adult lung appear to be detrimental, a recent study by Roger et al showed that in very preterm infants, low levels of MIF on postnatal day 6 were associated with an increased risk of developing bronchopulmonary dysplasia and late-onset neonatal sepsis 70 .
There are also several chronic inflammatory lung-associated pathologies that have been noted to be associated with changes in MIF. In particular, in idiopathic pulmonary fibrosis there is increased MIF expression in areas of remodeling, bronchiolar and alveolar epithelium, and ongoing fibrosis 71,72 . In systemic sclerosis, where MIF may be involved in the amplifying proinflammatory loop leading to scleroderma tissue remodeling 73 , an MIF promoter polymorphism is associated with susceptibility to pulmonary arterial hypertension in diffuse cutaneous systemic sclerosis (SSc) 74 . Functional promoter polymorphisms in the MIF gene, such as the high-expression MIF haplotype, C7, which is defined by −173*C and −794 with 7 CATT repeats, can also affect the clinical presentation of SSc 75 .
In addition, recent studies in a cohort of individuals with chronic obstructive pulmonary disease (COPD), demonstrate an association with increased plasma MIF and its acute exacerbations 76 , although others have suggested that MIF and its receptor are required for the preservation of normal alveolar structure and normal pulmonary endothelial cell apoptosis 77,78 .

RELATIONSHIP BETWEEN HYPOXIA AND MIF
Several studies describe a clear link between MIF and the presence of hypoxia. Hypoxia can lead to the secretion and elevation of MIF in fibroblasts 79 , cardiac myocyte 80 , monocytes 81 , and endothelia 81 . Hypoxia induces the stabilization of the transcription factor hypoxia inducible factor-1 alpha (HIF-1 α). When stabilized, HIF-1 α binds with aryl hydrocarbon receptor nuclear translocator (ARNT)/HIF-1 β. This hetero-dimer binds to elements found in the promoters of many hypoxia-responsive genes, leading to the expression of these target proteins such as vascular endothelial growth factor (VEGF), endothelin-1 (ET-1), and erythropoietin (EPO) 82 .
In 2006, Welford et al. demonstrated that the MIF gene contains an hypoxia response element within its promoter, which could explain the hypoxia induced MIF elevation 79 . This mechanism also has been suggested by others [83][84][85][86] . Furthermore, hypoxia in the presence of increased TNF α leads to an increase of the MIF receptor molecule CD44 on the surface of monocytes 87 .
Once released, MIF can contribute to hypoxic pulmonary vasoconstriction, which if maintained can lead to pulmonary vascular remodeling 88 . Zhang et al have shown that MIF affects delayed hypoxia-induced pulmonary hypertension and suggest that the action is via agonist enhancement on smooth muscle cells. However, while much is discussed about the apparently detrimental aspects of MIF inflammatory activity, it must be remembered that MIF has been shown to be protective in the early stages of ischemia.
Under ischemic conditions the extracellular release of MIF and its interaction with the CD74 receptor activates AMPK, thereby promoting glucose uptake and protects the cardiomyocyte 39 . In addition, the anti-oxidant activities of MIF can reduce intracellular oxidative stress and reduce injury in the post-ischemic heart 38 .

RELATIONSHIP BETWEEN HYPOXIA, MIF, AND PULMONARY VASCULAR CELL PROLIFERATION
Pulmonary vascular cell proliferation is the major pathological change during hypoxiainduced remodeling. The pulmonary vascular wall is composed of three layers of different cells: endothelial cells in the intima, SMCs in the media, and fibroblasts in the adventitia. Hypoxia in vivo induces proliferation of all of these cells, but only fibroblast proliferation is induced by hypoxia in vitro 13 in the absence of exogenous co-mitogens. In addition, fibroblast proliferation takes place earlier after hypoxic exposure than SMCs 13 , and hypoxia induces SMC proliferation only in co-culture with fibroblasts 20 .
Therefore, it appears that fibroblasts are essential to trigger the vascular remodeling process, perhaps because they are less differentiated and prepared for local injury repair 27 . Fibroblasts are remarkably plastic, allowing for rapid migration, proliferation, cytokine expression, and differentiation 27 . Fibroblasts differentiation to myofibroblasts 24 is a critical step for vascular remodeling and hypoxia alone can induce fibroblast proliferation 20,27,89 and differentiation to myofibroblasts 90,91 . Studies indicate that the lung is a major source of MIF 92 , is released from the lungs in patients with PAH 93 and plays a key role in hypoxia-induced cell proliferation 45 .

MIF AND PULMONARY ARTERIAL HYPERTENSION
PAH is a critical, and potentially devastating, clinical syndrome. The disorder, is particularly affects the small pulmonary arteries, and is characterized by vascular narrowing due to high-tone and abnormal vaso-reactivity. These abnormalities, if not corrected, lead to pulmonary vascular remodeling and intraluminal obstruction. Thus, the blood leaving the right side of the heart encounters an increased resistance to flow. While this can occur at any stage of life, it is particularly important in neonatal and adult pulmonary medicine.
In the neonatal setting, PH is associated with several conditions including, congenital heart disease 94 , connective tissue disease 95 or sickle cell disease 96 , stenosis 97,98 , and chronic lung disease of prematurity 99,100 ; and in adults PH is commonly seen in chronic obstructive pulmonary disease (COPD) [101][102][103] , sleep disordered breathing (sleep apnea) 104,105 and sickle cell disease 106,107 .
While hypoxia plays a pivotal role in the etiology, inducing pulmonary vessel constriction, and persistent hypoxia results in pulmonary vascular remodeling, leading further narrowing of the vessel, not all individuals subjected to hypoxia or hypoxemia develop pulmonary hypertension or other sequelae.

INHIBITION OF MIF ACTIVITY
The importance of MIF in the pathogenesis of disease has led to the development of inhibitory strategies to try to disrupt these processes. Early studies used polyclonal antibodies to inhibit MIF inflammatory activities to prevent the lethality in rodent models of acute hepatic failure 108 and septic peritonitis 109 .
Since that time, Phase 1 clinical trials have assessed the possible use of anti-MIF antibodies cases of malignant solid tumors, metastatic adenocarcinoma 110 , and lupus nephritis 111 . However, a disadvantage of the monoclonal antibody approach as a therapeutic pathway, is the possible development of local and systemic inflammatory reactions during administration. Therefore, a second approach whereby endogenous anti-MIF antibodies are generated, has also been advanced. This method involves active immunization with an MIF/tetanus toxin DNA vaccine and has been shown to protect against acute lung injury resulting from endotoxemia or a septic peritonitis 112 .
In 2002, Leng et al demonstrated that MIF interacts with the extracellular domain of the HLA class II histocompatibility antigen gamma chain (CD74), initiating activation of cell proliferation, and prostaglandin E2 production 34 . Since that time there has been considerable effort expended on developing small molecule inhibitors that could block the MIF-CD74 interaction. Studies have suggested that an imino-quinone metabolite of acetaminophen, N-acetyl-p-benzoquinone imine (NAPQI), can inhibit both the isomerase and inflammatory activities of MIF 113 .
Perhaps the most studied of these molecules is (S,R)-3-(4-hydroxyphenyl)-4,5-dihydro-5-isoxazole acetic acid methyl ester, or 'ISO-1' 114,115 . There have been many modifications on this and other chemical scaffolds to try to develop more effective and selective inhibitory molecules [116][117][118][119] . Recently, iguratimod, a novel antirheumatic drug 120 used in China and Japan, was found to selectively inhibit MIF inflammatory activity 121 . In many cases these molecules have been found to modify the pathology of animal models.

THYROXINE AS AN MIF INHIBITOR
The thyroid hormone, thyroxine, 3,5,3 ,5 -tetraiodothyronine (T4), has been identified as a potent inhibitor of MIF proinflammatory activities 122 . T4 is produced in the thyroid at around 100 mg/day 123 . While previously considered solely as a prohormone and parent molecule for T3 (3,3 ,5-triiodo-L-thyronine), both molecules have been shown to elicit profound effects on myocardial activity 124,125 , and extra thyroidal conversion of T4 to T3, by specific deiodinases, occurs in a variety of tissues and precedes many thyroid hormone actions 126 .
Once released from the thyroid, thyroxine circulates in the blood, bound with thyroxine-binding globulin, transthyretin and albumin with only around 0.05% in the unbound, free-thyroxine (fT4) form 127 . Cellular actions of thyroxine occur at the plasma membrane, in the cytoplasm in the mitochondria or the cell nucleus 128 .
Non-genomic action of T4, which can be initiated at approximate physiological concentrations of free T4 of around 10 −10 M 129 , are initiated by interaction with the integrin av β3 plasma membrane receptor, or in the cytoplasm 130 . Interaction with the integrin receptor on the cell surface stimulates MAPK (ERK1/2) activation 125,129,130 , leading to a series of downstream events including nuclear-trafficking of specific proteins and serine phosphorylation of nucleoproteins 128 , including estrogen receptor-α, thyroid hormone nuclear receptor β1, and signal transducing and activator of transcription (STAT)-1 α.
While T4, and its hormonally inactive dextro-rotary isomer D-T4, are effective inhibitors of MIF activity, triiodothyronine (T3), a T4 metabolite, is not 122 . In a clinical situation such as sepsis, plasma concentrations of T4 and MIF are inversely correlated, suggesting a clinically-relevant interaction between these two molecules 122 . In addition, a potential role of MIF-T4 interactions in the pathogenesis of PAH has been suggested since T4 inhibits MIF-induced ERK 1/2 phosphorylation in macrophages; T4 inhibits MIF activation of NF κB RelA/p65 in fibroblasts; and MIF inhibits T4-induced CXCR2 mRNA accumulation in vascular smooth muscle cells 131 . Thus, a schematic of potential mechanisms in vascular cells in hypoxia-induced altered MIF-T4 interactions is shown in Figure 2.

INHIBITION OF MIF IN PULMONARY HYPERTENSION
Studies from our group have found significantly increased plasma MIF concentrations in individuals with primary PH or PH secondary to interstitial lung disease, compared to control subjects 45 . Therefore, we examined the effect of MIF lung fibroblast growth and showed that the increased hypoxia-induced proliferation was MIF dependent. Furthermore, in a mouse model of hypoxia-induced pulmonary hypertension, the pulmonary vascular remodeling, increased right ventricular systolic pressures and right sided cardiac hypertrophy were all significantly decreased in the presence of a small molecule inhibitor of MIF. This suggests that MIF plays a significant role the development of PH.   135,136 . STAT5 is the regulator of multiple genes. There is a functional interdependence between MIF and HIF1 α, which can also impinge on STAT5 activity.
It should be noted, however, while in human idiopathic pulmonary arterial hypertension there is a two-to four-fold higher prevalence in postpubertal women than in men 132 , there is an opposite male-sex bias in the hypoxia, or monocrotaline rodent models of pulmonary hypertension 5,133,134 .
In associated studies we examined the role of the signal transducer and activator of transcription 5 (STAT 5) in the development of PH, and showed decreased STAT5 expression in the obliterative lesions of human idiopathic PAH, and that deletion of STAT5 from the vascular smooth muscle cells abrogated the male bias 135 . This is of relevance since STAT5 can act as a mediator in hypoxia-mediated gene expression 136 , and that, at least in some cells, MIF can promote intracellular signaling by STAT5 137 (Figure 3).
In addition to the pulmonary pathology, clinical studies suggest that PAH is also associated with cognitive impairment, depression and anxiety. Recent studies have shown, in the hypoxic mouse model, that hypoxia induces increased MIF accumulation within the hippocampus (at both level of mRNA and protein), and metabolic, biochemical, and electro-physiological changes within the hippocampus were associated with cognitive dysfunction. 138,139 . These hypoxia-driven changes were corrected by the administration of an MIF inhibitor 138,140 . However, it remains unclear whether the cognitive dysfunction was corrected by direct inhibition of the increased MIF within the hippocampus, or as a result of the improvement in pulmonary function.

MIF INHIBITION IN CONGENITAL DIAPHRAGMATIC HERNIA (CDH)
CDH is identified in around 1/3000 newborns 141 . It results from a defective diaphragm allowing the protrusion of abdominal tissues into the thorax, thereby interfering with normal lung development leading to lung hypoplasia and persistent pulmonary hypertension of the newborn 141 . Rodent models of the condition have been developed, and often use administration of a single dose of 2,4-dichlorophenyl-P-nitrophenyl (Nitrofen) on the tenth gestational day 142 . Poor vascular growth in the CDH rat model is associated with poor lung growth. The exact mechanism of defective angiogenesis associated with CDH is not fully understood. However, studies show that inhibiting the MIF activity in the rat CDH model results in higher expression of VEGF and Tie-2 receptor while normalizing Sflt-1. Together, these molecular changes lead to a significant improvement in pulmonary angiogenesis as well as lung development as shown by CT and histological studies 143 .

MIF AND OTHER NEW PATHWAYS
In addition to being a potent vascular vasodilator, nitric oxide (NO) enhances angiogenesis by activating endothelial cell growth and tube formation 144 . Neonatal rats with CDH were treated with ISO-92 and we observed a significant increase of phosphorylated eNOS (P-eNOS), which is known to increase NO production.
This was also associated with decreases in both arginase-1 and -2 expression. Arginase is a urea cycle enzyme which competes with endothelial NO synthetase (e NOS) and inhibits NO synthesis via a common substrate, L-arginine. Hypoxia upregulates the expression of both arginase enzymes 145 .
In our study, among neonates with CDH, both arginase-1 and -2 enzymes were overexpressed significantly in comparison to healthy control neonates. This may suggest that, both NO production and its bioavailability were significantly compromised among neonates with CDH.
We have shown that treating pregnant adult rats with ISO-92 after inducing CDH around day 8-9 of gestation, significantly decreased both arginase-1 and -2 expression, which is known to eventually increase NO production and its bioavailability. Accordingly, we postulate that inhibition of both arginase-1 and -2, could be the mechanism through which inhibition of MIF activity (using ISO-92 in our model), can lead to increase NO bioavailability in utero, thereby improving pulmonary angiogenesis and lung development 146 .

CONCLUSION AND FUTURE STUDIES
Macrophage migration inhibitory factor is a key mediator of inflammatory responses and innate immunity and has been implicated in the pathogenesis of several inflammatory and autoimmune diseases. MIF's role in the pathogenesis of PAH, induced by chronic hypoxia, or associated with chronic lung diseases, or idiopathic, has been explored in many studies. The link between endothelial dysfunction and MIF in animals models with chronic PAH has been established. Other studies highlighted MIF role as a biomarker for the assessment of PAH associated with chronic obstructive lung diseases. Discovering the role played by T4 as a natural ligand inhibitor of MIF's inflammatory activity opens the door for new therapeutic role of anti-MIF's, as shown in preclinical and clinical data, which suggest that blocking the inflammatory active site of MIF may both reduce inflammatory responses and improve the availability of T4. Preclinical data using different anti-MIF's in different animal models with chronic and severe forms of pulmonary hypertension are very promising. Whether inhibition of MIF or its oxidized forms may offer promising therapy in PAH, needs to be elaborated in future human interventional studies.