Corticosteroid-binding globulin (CBG) was discovered in the human plasma in the 1950s by three different groups (13–15). The latter group called it transcortin by analogy to transferrin, the iron-binding protein (14). This CBG presents low capacity but high affinity for corticosteroids, which is in contrast to albumin, another plasma transporter protein, which has a high capacity but low affinity for these hormones (16, 17).

CBG is a protein consisting of 405 amino acids (aa) with a molecular weight of 45 kDa. Before its secretion into plasma, its first 22 aa, which represent the cellular export peptide signal, are excised. Thus, circulating CBG is composed of 383 aa with a theoretical weight of 42 kDa (18). However, since CBG is glycosylated, it is generally detected with a molecular weight between 60 and 70 kDa. CBG sequence contains 6 asparagines (Asn) corresponding to 6 N-glycosylation putative sites: Asn9, Asn74, Asn154, Asn238, Asn308 and Asn347 (the position of the Asn is indicated without the signal peptide). These glycosylations appear to have an important role in the function of the protein. Glycosylation is commonly considered as a message for intracellular trafficking and exportation. In the case of CBG, glycosylation may also be important for GC binding and delivery. For example, glycosylation at position 238, may ensure CBG binding to GC (19, 20) by allowing proper folding (19), while glycosylation at position 347 may be important in modulating GC delivery (21, 22). Thus, variations in protein glycosylation may cause variations in GC binding affinity and may protect CBG against proteolysis (23, 24).

Among vertebrates, mammalian CBG were the first to be characterized [human (18); rat (25); rabbit (26); sheep (27)]. The primary structure of human CBG defines it as a serine proteinase inhibitor (serpin) family member (18). Its gene (SERPINA6) is located in a cluster of related clade A SERPIN genes (28, 29) with synteny across other mammalian genomes (30). Phylogenetic studies showed that CBG is part of the SERPIN family (SERPINA6) (31–33). In birds, when looking for proteins able to bind steroid hormones in 23 avian species, Wingfield and colleagues discovered that no specific sex hormone-binding protein/globulin (SHBG) with high affinity and low capacity could be characterized in any of the species investigated (34). They also found that the high affinity binding protein for corticosterone binds progesterone with identical affinity, whereas testosterone and estradiol are bound with lower affinity (34). Birds were shown to lack SHBG gene and thus CBG is likely the substitute for SHBG in this vertebrate group (35). CBG characterization in white-throated sparrow (Zonotrichia albicollis) plasma showed that CBG binds corticosterone with over 6-fold higher affinity than testosterone (36). In amphibians, CBG binds GC and sex steroids with equivalently high affinity [Ambystoma tigrinum (37); Bufo terrestris (38)]. In reptiles, two steroid-binding globulins were identified in male tree lizard Urosaurus ornatus: one binds androgens and estradiol with high affinity and is similar to other vertebrate SHBG; the other binds androgens and C21 steroids (steroids containing 21 carbon atoms) including progesterone and corticosterone with higher specificity than estradiol and is named androgen-glucocorticoid-binding globulin (AGBG) (39). Genes for both plasma AGBG (named CBG in the article) and SHBG were identified in a study performing phylogenetic comparisons in anole lizard Anolis carolinensis and Chinese softshell turtle Pelodiscus sinensis (40). In teleost fish, little evidence for a plasma CBG-like protein exists (41), as most of the few data available reported a lack of specific plasma proteins able to bind cortisol (42–45). Only Caldwell and colleagues observed that mature female rainbow trout had greater cortisol bound to a CBG-like (48%) than mature males (16%) and immature fish (19.5%) (46), but this study relies on binding assay and not gene characterization, which is not sufficient for demonstrating the existence of CBG in teleosts. In contrast, SHBG genes have been characterized in this vertebrate group (47).

Mainly synthetized by the liver (48), the CBG is released in the blood ( Figure 2 ). CBG was measured, by means of gel-filtration method, in the plasma of 131 species representing all the vertebrate classes and was detected in every vertebrate species studied, even fish species (7, 49). In fact, at that time, it was the % of steroid (cortisol and corticosterone) bound which was assayed, and up to now no CBG has been characterized in teleost fish. Desantis and colleagues demonstrate a dichotomous pattern among vertebrates with respect to maximum corticosteroid binding capacity (MCBC) and total CBG levels (50): a dominant branch, where high levels of CBG bind most of the GC, applies to the majority of vertebrates; a secondary branch, comprising lamprey, flying squirrel or New World monkey, in which low levels of CBG bind almost none of the GC, applies to a very small subset. For the latter, the critical unknown is how these species mitigate the impact of the high free GC levels and how such a dramatic trait shift could evolve (50). Other comparative studies, this time among birds and mammals, also show a great variation in the binding capacity and affinity of CBG within and among the avian and marine mammalian species studied (51, 52). Beyl and collaborators recently evidence the importance of assay temperature when measuring CBG, as colder temperatures maximize specific binding but likely underestimate GC affinity for CBG (53).

Although the regulation of CBG by GC has been widely studied in the literature, no consensus on this subject has emerged. Adrenalectomy induced a CBG increase in rats Rattus norvegicus (7, 54), while having no effect in the other mammalian species tested (human, Homo sapiens; guinea pig, Cavia porcellus; rabbit, Oryctolagus cuniculus; dog, Felis familairis; cat, Felis domesticus) (7, 55). No change in plasma CBG levels was noticed after injection of dexamethasone, a synthetic GC, followed by ACTH administration, in snowshoe hare Lepus americanus and laboratory rabbit (56) or of corticosterone in rat (57). A rapid increase of plasma CBG levels is observed after ACTH injection in rat (58) and laboratory rabbit (56). In male rat, subcutaneous injection of dexamethasone was, however, able to decrease both plasma and hepatic mRNA levels of CBG (59). Using in vitro perfused liver slices from adrenalectomized male rats injected intraperitoneally in vivo with GC, Feldman and colleagues demonstrated that both prednisolone and dexamethasone also induced a decrease in CBG basal concentration and production rate (60). A decrease of CBG mRNA levels was also observed using mouse hepatoma cell line (BWTG3) treated by dexamethasone (61). In contrast, dexamethasone treatment had no effect on CBG mRNA levels and release by mouse Leydig tumor cell line (mLTC-1) (62). In human, both endogenous and exogenous glucocorticoids can modulate circulating levels of CBG: plasma levels of CBG are suppressed during prolonged exposure to GC, whether endogenous, as in Cushing’s syndrome, or exogenous, as after administration of synthetic GC (55, 63). A decrease of CBG (this time mRNA levels) was also observed in vitro after treatment by dexamethasone of hepatoma cell lines from human (HepG2) (61). However, another study, also carried out on HepG2, showed no effect of dexamethasone on hepatic CBG synthesis, regardless of the dose used (64). Jung and colleagues also reported that the administration of hydrocortisone to healthy subjects, either orally or intravenously, does not change plasma CBG levels (65). Interestingly, one study showed that dexamethasone increased CBG production during fetal life, while inhibiting it in adulthood (27). Studies on the effect of glucocorticoids on CBG expression seem therefore dependent on the model, age, dose, and type of glucocorticoid used, with synthetic ones being more effective in regulating CBG.

In mammals, castration leads to either increase (rat; mouse, Mus musculus; cow, Bos taurus; goat, Caprya hircus), or decrease (rabbit; horse, Equus caballus) and even no change (human; cat; pig, Sus scrofa; hamster, Mesocricetus auratus) (7). Estrogen treatment was reported to increase CBG concentration in men and women (49, 55), as well as in some other mammals [rats, mice, rabbits and guinea-pigs (49)]. However, this was not the case in dogs (66) and sheep (67). The in vitro basal concentration and production rate of CBG by liver slices was also increased in adrenalectomized male rats injected intramuscularly with estradiol (60). In the rat, sex difference was observed concerning CBG activity, as administration of estradiol (E2) did not change it in females but increased it in males (68). Recently, the gonadotropin, luteinizing hormone, was shown to differentially regulate CBG mRNA levels (decrease) and release (increase) by mLTC-1 (62).

A decrease of CBG levels is observed in human hyperthyroidism (55). In contrast, in male rat, the thyroid hormone T4 administrated subcutaneously increased serum CBG and hepatic CBG mRNA levels, but did not influence the rate of CBG gene transcription, suggesting that T4 increases CBG mRNA stability (59).

As one of the main functions of CBG is to sequester GC in plasma by high affinity binding, it has long been considered under the prism of the free hormone hypothesis. This hypothesis asserts that the only biologically active part of the hormone is that which is free, i.e. not bound to a protein in plasma (176). Thus, the concentration of a hormone in a tissue is determined solely by its free plasma concentration, rather than by the concentration bound to its transport protein. This hypothesis is one of the best explanations for the clinical manifestations seen in patients suffering from hormone deficiency or excess. According to this hypothesis, the main function of CBG is to regulate the bioavailability and metabolic clearance of GC. A clinical study showed that patients with high plasma CBG levels cleared injected radiolabeled cortisol more slowly than those with lower levels (177). CBG therefore increases the half-life of plasma GC. Three studies in 2011 gave strong experimental evidence for this ‘free hormone hypothesis’. The first one was carried out in patients with a CBG mutation rendering it unable to bind its ligands (178). These patients had increased metabolic clearance and a decreased cortisol half-life. The second one assayed free GC levels in rats during and after various stressors (80). Forced swim stress induced an elevation of total plasma GC within 30 min, while the elevation of free plasma and tissue GC level was observed 30 min later (80). The third one showed that salivary GC levels mimic free plasma GC levels and both correspond to about 10% of plasma total GC (179). Thus, total GC is not what is available to tissue, but free GC (not bound to CBG) is.

In their review in 2013, Breuner and colleagues refer to a ‘reservoir hormone hypothesis’, a complimentary to the ‘free hormone hypothesis’, when citing Malisch and Breuner data on steroid-binding protein and free steroids in birds from 2010 (69). These authors suggest that the GC remaining bound serves as a reservoir of GC in the blood to be used as needed (180). In her correspondence paper in Nature reviews, Marie-Pierre Moisan gives different arguments for considering CBG as a cortisol reservoir rather than a transporter (181). First, cortisol can circulate in CBG-deficient patients (134) or in vertebrate species lacking CBG (133, 182), and the presence of CBG is necessary to mount a normal stress response as also shown by studies in CBG-deficient mice (167, 168). Signs of hypocortisolism rather than hypercortisolism are reported in case of CBG deficiency (please refer to section 2.2.2). She also mentioned that the mineralocorticoid, aldosterone, is equally hydrophobic as cortisol and does not possess a specific binding protein, as at that time only binding studies have reported a potential plasma binding proteins for this corticosteroid [human (183, 184)], which is still true twenty years after. For Marie-Pierre Moisan, it is albumin which ensures the transport of GC, as GC are mainly bound to albumin in CBG-deficient patients (141) or mice (167), and albumin is present in the blood of all vertebrates. Thus, CBG would have appeared during evolution to be the retention in blood of a circulating GC reserve readily available in case of an emergency (reservoir).

Various parameters can influence hormone distribution at tissue level. For example, the location of the target cell, endothelial permeability, the composition of the extracellular matrix and the juxtaposition of different cell types within the same tissue can influence a cell’s accessibility to a given hormone (185). The level of free cortisol is therefore not the only parameter determining its tissue concentration. CBG could play an active role in determining this concentration, notably through its addressing function.

In human, CBG circulates in low quantities (30 to 52 pg/mL) in plasma compared to albumin (40 g/L). However, its high affinity for GC means that it plays an important role in determining plasma GC concentration. CBG binds 80% to 90% of circulating cortisol, while 7% to 15% is bound to albumin and less than 5% is free (186).

CBG binds cortisol at a surface pocket located between the β B leaflet and the helix A and H. Like other serpins, CBG establishes a covalent bond with its ligand, which stabilizes CBG in its S conformation and exposes its reactive center loop (RCL). When cleaved by neutrophil elastase ( Figure 2 ) between Val344 and Thr345, a conformational change in CBG occurs. The cleaved segment of RCL is then inserted into the β A leaflet, the protein is stabilized in its R conformation and loses its affinity for cortisol (187). With a 10-fold decrease in affinity, CBG releases cortisol. It has been shown that Pseudomonas aeruginosa elastase (LasB) can also cleave CBG at a site a few amino acids distant from the neutrophil elastase site (188). LasB is released by the bacterium at the sites it infects, while neutrophils release elastase at inflammatory zones. It is therefore considered that, in addition to its role as a transporter, CBG is able to specifically address glucocorticoids to the site of inflammation and infection. Thus, CBG could represent an interesting target to control exclusively GC effects on inflammation.

The binding of CBG to cortisol is also temperature-dependent. Several studies have shown a decrease in affinity between these two molecules as temperature rises from 37 °C to 42 °C (189, 190). This is consistent with the addressing role of CBG. Patients with inflammation or fever will have an increase in body temperature, allowing them to release more cortisol and resolve the inflammation as best as they can.

It is worth noting that chymotrypsin, a protease secreted by the pancreas, has been shown to be able to cleave the RCL of CBG (191). The discovery of this cleavage site is to this day still misunderstood, and its physiological significance remains to be elucidated.

In addition to the liver, CBG is expressed (at lower levels) by various other organs and tissues [for reviews in mammals (192, 193)]. In mammals, its protein or transcripts have thus been detected in the kidney [rat (194); rhesus monkey (18); mouse (195)], the lung [rabbit (26); mouse (164)], the heart [human (196)], the spleen [rabbit (26)], the white adipose tissue [rat (197)], the ovary [rhesus monkey (18); rabbit (26)], the female genital tract [rat (194) and human (198) uterus; human Fallopian tubes (199)], the placenta [human (200)] and the testis [rhesus monkey (18) mouse (201)]. CBG is also detected in the adrenal [mouse (165) and thyroid [rat (194)] glands, as well as the central nervous system [rat (202–204); mouse (205)] and the pituitary [guinea pig (206); rat (194)]. In a bird, the zebra finch Taeniopygia guttata, the CBG mRNAs were quantified in spleen, lung, kidney and gonads of males and females (while being undetectable in skeletal muscle), but their levels were more than 300-fold lower than those in liver (40).

Few studies have examined the role of these extrahepatic CBGs. It is commonly assumed that these CBGs do not contribute to plasma CBG levels, as they are intracellular. Despite their low level of expression, these CBGs could be finely regulated at the cellular level and play an important local role. In human lung, the levels of CBG transcripts seem to be differentially expressed among the airways and regulated in lung disease situation [cystic fibrosis (207)]. Concerning local CBG function, it is hypothesized that GC released from plasma would enter the cell and, at this level, bind intracellular CBG. Blocked in this way, the GC would no longer be able to bind its receptor, the main mediator of its intracellular actions, and would no longer be able to act.

This hypothetical role of extrahepatic CBG is supported by the better-known regulation of other hormones by their binding proteins. This is the case for IGF (insulin-like growth factor), which is transported in biological fluids by proteins known as IGFBPs (IGF binding proteins). They are capable of regulating the bioavailability of IGF, but also of modulating its activity by binding it locally to target tissues (208). CBG may enable equally fine regulation of GC. Depending on the tissues expressing these CBGs, limiting the intracellular actions of GC may have wider consequences. In the brain, CBG is expressed by numerous cell types, including astrocytes (209). GC are capable of inducing remodeling of these cells and eventually their death. Locally produced CBG could therefore act as a buffer to protect the cell from the deleterious effects of these molecules. CBG also appears to have a role in behavior and memory, particularly in response to stress (168). The role of CBG at the cerebral level in this emotional response remains to be determined, but it could modulate and control this stress response. In the lung, CBG could influence fetal maturation of this organ, particularly at the alveolar level. The differentiation of alveolar epithelial cells and the control of cell proliferation are indeed dependent on the action of GC (210). Pulmonary CBG could therefore also play a role in lung development and maturation. The putative roles of these local productions, however, remain to be confirmed in order to fully understand their biological relevance.

In the early 1980s, the main site of CBG synthesis was established in the liver. Several groups subsequently showed that the protein was detected in other tissues, even within cells, although it was unclear whether this was local production or not. A hypothesis was then put forward regarding the potential internalization of circulating CBG by target cells. This hypothesis is based on the existence of a receptor capable of binding CBG to enable its endocytosis. In 1983, Strel’chyonok and Avvakumov were the first to suggest the presence of this receptor in the plasma membrane of human liver cells (211). Hryb and his colleagues confirmed this result on cell membranes from human prostate. Their study underlines the specificity of the binding of the receptor to CBG since albumin or transferrin are incapable of shifting the binding equilibrium. They also demonstrate that this binding varies depending on time and temperature (212). The presence of this receptor has also been demonstrated in the endometrium (213). It should be noted that in vitro studies show that only 2% to 5% of CBG bind to its receptor (212, 214). The binding of CBG to its receptor also appears to be dependent on prior binding of CBG to one of its ligands. Data in the literature are contradictory on this subject. Strel’chyonok and Avvakumov show that CBG binds to its receptor only if it is already bound to a GC, while Maitra and colleagues observe that binding CBG to a GC prevents it from binding to its receptor (214, 215). Since CBG is thermosensitive, these differences may be explained by differences in protocol, the first study having been carried out at 4 °C while the second was at 37 °C.

Three decades after these studies, a molecular characterization of such a receptor is still needed and one may question the reliability of the data. Only in vitro studies have indirectly shown its existence and a partial characterization has been carried out in rats (214). The few studies of this receptor have, however, led to the emergence of two hypotheses on its role. The first indicates that CBG, bound to its ligand, would be internalized in the cell and would thus allow finer control of the distribution of its ligands in its target tissues. This hypothesis thus gives CBG an active role in the addressing of its ligands. The second hypothesis considers CBG as a pro-hormone with intrinsic activity, the binding to its receptor allowing the activation of a second messenger. Nakhla and his colleagues demonstrate, in fact, that the binding of CBG to the plasma membrane leads to an activation of adenylate cyclase and an increase in cAMP in the cell (216). CBG would thus induce rapid intracellular signaling cascades and allow glucocorticoids to act rapidly in a non-transcriptional manner.

Thyroid diseases can influence corticosteroid metabolism, with hyperthyroidism being associated with an increase in their catabolism (217). Consequences have also been observed on CBG, the plasma level of which decreases in patients with hyperthyroidism (218). The opposite has also been observed in patients with hypothyroidism (219). After treatment allowing a return to euthyroidism in these two types of patients, the CBG concentration normalizes (218–220). Hormones thus appear to negatively regulate CBG expression since in vitro studies show that long exposure of HepG2 cells to triiodothyronine causes it to decrease (221).

Changes in plasma CBG levels have also been observed in pathologies related to altered cortisol levels. Thus, in patients suffering from Cushing’s disease, a disease defined by chronic hypercortisolism, a decrease in plasma CBG levels is observed, accompanied by an increase in free cortisol levels (222, 223). The decrease in CBG concentration is associated with a decrease in binding capacity of up to 40% (63). The decrease in CBG could be the consequence of the regulation of its gene expression by cortisol. However, CBG levels appear to be normal in patients with Addison’s disease, a disease characterized in particular by a defect in cortisol secretion (222, 224). It should be noted that this disease is also associated with a defect in mineralocorticoid secretion. Therefore, the regulation of CBG in this context cannot be explained solely through the prism of cortisol levels.

CBG levels also appear to be decreased during obesity (225). As previously seen, some polymorphisms in the SERPINA6 gene have been associated with decreased plasma CBG concentrations and certain obesity-related parameters, such as insulin resistance (156). However, several studies have produced conflicting results on plasma CBG levels and insulin sensitivity (159, 225, 226). In vitro, in HepG2 cells, insulin is able to decrease CBG secretion and mRNA production (227). In vivo, in lean subjects, insulin injection causes a brief decrease in plasma CBG, but not in obese subjects (225). The regulation of plasma CBG in the context of obesity therefore appears to be more complex since obesity is also a chronic inflammatory pathology, a type of pathology that also causes variations in CBG levels.

Given the role of CBG in inflammation, several studies have focused on its regulation in a pathological and inflammatory context. Savu and collaborators show a depletion of cortisol binding activity by CBG in patients with septic shock, reflecting a decrease in plasma levels (228). Pugeat and colleagues then directly demonstrate a drastic decrease in plasma CBG concentration in patients with septic shock (229). Monitoring a patient 9 days after shock reveals a progressive return to normal of CBG levels. However, the decrease in CBG was not observed for toxic, hemorrhagic or cardiogenic shock. The decrease in CBG during septic shock has been attributed to the regulation of CBG by inflammation, with interleukin-6 (IL-6), a central cytokine in inflammation, being able to inhibit its expression and secretion. Plasma IL-6 levels have indeed been correlated with plasma CBG levels in these patients: the higher the IL-6 level, the lower the CBG level (230). CBG levels have more recently been directly correlated with shock severity: patients who did not survive shock had the lowest plasma CBG levels (231). These findings have highlighted the importance of CBG in septic shock, with finally CBG deficiency independently associated with mortality (232).This same decrease in CBG was found in patients with burns or necrotizing pancreatitis (233, 234). In burn patients, this decrease was also correlated with an increase in IL-6. All the data in the literature on diseases with an inflammatory component suggest that CBG could be used as a biomarker of the degree of inflammation. It should be noted that, in the case of pancreatitis, the decrease in CBG in the first 48 hours has even been proposed as a predictive marker of future infection, with a positive predictive value of 100% and a negative predictive value of 87.5% (234). Concerning rheumatoid arthritis, contradictory data exist. Patients with rheumatoid arthritis either do not present change in plasma CBG levels compared to healthy controls (235), or show higher total and high-affinity CBG, reflecting reduced CBG cleavage in this pathology (236). Reduced plasma levels of CBG have been also recently reported in patients with coronavirus-19 disease (237).

To better understand the regulation of CBG in the inflammatory context of diseases, in vitro studies were conducted on the regulation of CBG in inflammatory conditions. They tend to show an inhibitory effect of inflammation on CBG expression. In 1993, Bartalena and colleagues first showed that IL-6 decreased CBG synthesis by HepG2 (238). The authors observed a decrease in CBG secretion in the media as well as a decrease in mRNA, in a dose- and time-dependent manner. However, they did not observe variations in the transcription rate and then hypothesized that IL-6 would decrease the stability of CBG mRNA. This in vitro effect of IL-6 on CBG expression was confirmed a few years later by another team who showed a decrease in CBG of up to 30% to 40% (64). Interestingly, this decrease is even more significant when the cells are treated with a combination of IL-6 and dexamethasone. The combination of these two molecules increases the expression of a subunit of the IL-6 receptor and thus potentiates the effects of IL-6. These two articles are at the origin of the classification of CBG as a negative acute phase protein in humans, a classification already carried out in rats in 1980 (239). This makes it possible to measure the severity of inflammation by the extent of the decrease in CBG: CBG could therefore be a biomarker of inflammation (240). The effect of IL-6 has also been studied in vivo on healthy subjects. A high-dose injection of IL-6 (3.0 μg/kg) decreases plasma CBG levels, which only return to normal after 7 days (241). However, Emptoz-Bonneton and colleagues also show that another pro-inflammatory molecule, IL-1β, causes an increase in CBG secretion at the same time as its decrease in mRNA (64). IL-1β would thus act post-transcriptionally and/or directly on the CBG secretion mechanism. Inflammation could therefore act on CBG according to different mechanisms that remain to be elucidated.

Tinnikov and colleagues were the first to show, in children undergoing cardiac surgery, that plasma CBG levels decreased by half during the procedure, while cortisol levels increased (199). CBG levels also remained low the day following surgery. This decrease has been confirmed in other cohorts (200). Roth-Isigkeit and colleagues, however, noted a decrease in hematocrit percentage during surgery and over the following two days (201). Correction by hemodilution shows that the CBG level is only slightly decreased, and this only on the day of surgery. The corrected total and free cortisol concentration is still increased for several postoperative days. Their study thus shows that, in this context, cortisol secretion seems to be the main determinant of the free cortisol level, the CBG level not being altered.

Since CBG is mainly produced by the liver, studies have focused on its plasma level in the context of liver diseases. Shortly after its discovery, Doe and colleagues showed that patients suffering from cirrhosis had a lowered plasma CBG concentration (242). Several studies then validated this result on different types of cirrhosis including cirrhosis caused by biliary atresia, hepatitis B or C virus, and autoimmune disease (55, 243). The decrease in plasma CBG concentration seems to be correlated with the severity of cirrhosis: the more pronounced the liver damage, the more the CBG level falls (244). Interestingly, another study investigating new markers of liver fibrosis identified CBG as a potential biomarker, with its plasma level also decreasing with increasing severity of the disease (245). The decrease in plasma CBG has been explained by a probable decrease in synthesis at the hepatic level, although no study has been able to demonstrate this to date.

Cystic fibrosis (CF) is a genetic disease with both pro-inflammatory profile and, for some patients, liver condition; some CF patients requiring liver transplant. We expected that, in CF patients suffering from liver hepatic disease (cirrhosis), the levels of both liver and serum CBG would be decreased, as a pioneer study reported a slight decrease in plasma CBG capacity in CF patients with low liver condition (246). We obtained exactly the opposite with significant increase of both transcripts and protein in cirrhotic liver from CF patients compared to healthy donors and cirrhotic non-CF patients (207). It is unlikely that the hepatic increase in CBG transcripts and protein is a direct consequence of a CFTR channel dysfunction within the cell, as CBG is produced by hepatocytes (247) when CFTR is exclusively expressed by cholangiocytes (248). The cause is rather the specific microenvironment of CF liver with inflammation due to toxic bile acids accumulation (249). In our study, the plasma levels of CBG were unchanged among patients analyzed (207). On the other hand, this could be due to a hepatic retention, but the plasma levels of other proteins produced and secreted by the liver are not disrupted in CF patients (250). On the other hand, an increased cleavage of CBG in plasma can be considered as we also observed an increase of elastase/α-1-antitrypsin complex in the plasma of CF patients (207). When one knows that elastase is sustainably released in CF patients due to an exacerbated neutrophilic activity (187) and to the early colonization of patients by Pseudomonas aeruginosa (188), this enzyme could target CBG and cleave it, leading to its irreversible inactivation (187, 192).