The establishment of relationships between circulating corticosteroids (and, consequently, of CBG), pose a considerable problem because of their rhythms 
, variations in binding vs. free proportions 
, tissue interconversion between 17-hydroxy- (active) or 17-keto forms (less active) of corticosteroids, and the intrinsic problems of evaluating steroid hormones. This question has become critical in the context of the study of inflammatory processes related with the metabolic syndrome. Obesity is widely accepted to be directly related to increased glucocorticoid activity 
, but direct measurement of cortisol or corticosterone (or those of their main excretion products) do not show definite increases in their circulating levels 
. Since we have observed the existence of an additional steroid hormone reservoir pool in blood, i.e. hormones carried on the blood cells 
, we investigated whether compartmentation played a significant role in the increased effect of glucocorticoids observed in the metabolic syndrome.
Testosterone is known to bind (with low affinity) CBG 
, but we have displaced corticosterone from its binding sites in plasma proteins only when using very high (non-physiological) concentrations. We can conclude that the differences observed in corticosterone binding effectiveness by sex could not be explained by the (limited) differences in testosterone levels or by testosterone competitive binding.
The proportion of corticosterone bound to RBCs was not specific, and was dependent on the concentration of the hormone, especially on the proportion carried by the plasma. This is in agreement with the results obtained with sex hormones 
, but in the case of corticosterone, the proportion of blood cell-carried hormone was lower, probably because corticosterone is more hydrophilic than androgens or estrogens. In spite of its relatively small size (in the range of 10% of total blood hormone), the cells' compartment was larger (twice) than that of free corticosterone.
Gender has a marked influence in blood (and plasma) corticosterone, with females showing higher values than males, as described previously 
. These differences affect all compartments, whilst diet increased (in males) the proportion of hormone bound to blood cells; there were no other significant effects of diet on corticosterone compartmentation. These results are in line with the paradox of increased glucocorticoid effects observed in inflammation 
with relatively unchanged plasma or serum levels 
The modulation of free cortisol or corticosterone is widely accepted to rest on two cooperative processes: conversion in tissues through 11β-hydroxysteroid dehydrogenases 
and modulation of CBG levels and affinity 
. In the present study, we have observed a marked difference in the ability of CBG to bind corticosterone in females (measured CBG levels justifying about 1/6th of plasma corticosterone binding) and males (about 1/2). This difference was also observed in the hepatic expression of the gene coding for CBG (Serpina6
), in agreement with previous studies 
. However, the low justification of binding by CBG levels measured with a specific monoclonal antibody ELISA was also in disagreement with the analysis of CBG in plasma by Western plot using a polyclonal antibody. The differences observed between both methods were considerable. In fact, the gene expression tendency, the Western blot results and the total plasma binding capacity were in agreement: females had higher corticosterone-binding activity, CBG expression and CBG levels in plasma. But the ELISA data showed the opposite.
Thus we intended to first determine whether in both cases we were actually measuring CBG: the ELISA method was inhibited by the antigen peptide used to produce the polyclonal antibody, albeit with a lower potency (about half) than that of the complete molecule. But, evidently, the monoclonal ELISA antibody reacted fully with the antigen used to produce the antibody for Western blots. Perhaps the question was the reverse: this antibody may react to a number of other serpins 
or bind to other non-CBG proteins. However, the inhibitory peptide suppressed all bands in the Western blots, thus showing that all were also CBG. The analysis of Western blots at a higher titer showed that probably there were (at least) two immunoreactive bands. This may be a consequence of the shortening of CBG chain by elastases 
, since the difference in molecular weight was small. However, even if that were the case, this could not explain the differences observed, since these bands were present in both females and males.
The possibility that the ELISA kit data were wrong could not be sustained, first because we repeated the test using different lots, second because the standards were purified native rat CBG, third since the ELISA antibodies were raised against a 273-peptide containing the 14-residue peptide used to raise the polyclonal antibody, and fourth because a lower response would affect both female and male, and not discriminate between them in so marked (opposite) way. The rat has a single transcript for the Serpina6 gene and the antibody with lower overall sensitivity was that raised on the longer peptide. We checked that the 14-residue sequence used to obtain the antibody used for Western blots was unique −in rats− for CBG. Thus, both antibodies were recognizing the same protein coded sequences, i.e. CBG.
Since, in addition, not even in males the concentration of CBG justified more than 50–60% of the specific plasma binding (in itself in the range of 95% of total binding), the question remains: what protein was responsible for the remaining specifically-bound carried corticosterone? This figure is raised to perhaps 85% in the case of female rats. Our preliminary conclusion is that in plasma there may be more than one CBG isoforms, differentiated by the very specific monoclonal antibody of the ELISA kit but not by the polyclonal one: i.e. the peptide segments used to obtain them may represent a possible sequence difference between them. However, the gene was only one, thus the eventual differences should be post-translational.
Serpins are a large family of proteins 
, and a few of them are close to CBG. It is unclear whether the usual methods for gene expression and protein measurement are enough to discriminate between two possible forms of the transporter. In fact, most of the studies have been carried out using a single transcript or polyclonal antibody, generally using female rats (for their higher CBG content) or humans. However, it is highly probable that the differences between the two populations of CBG we postulate are of post-translational nature, since the data presented under Results
also show that the increase in liver expression of Serpina6
is closely related to plasma binding capacity and both are enhanced by cafeteria feeding: i.e. Serpina6
expression is a direct response to inflammation, as are CBG 
and glucocorticoid activity 
The possible existence of two forms of CBGs with different affinities for corticosteroids (and perhaps for testosterone too) may add an additional step of modulation of free corticosteroids' levels. It may be indicative that exposure to a cafeteria diet practically did not change the levels of CBG nor binding in males, but deeply altered these parameters in females. Perhaps the ultimate factors controlling this conversion are the estrogens, powerful antiinflammatory agents 
, which may act in coordination with glucocorticoids 
. Estrogens also antagonize glucocorticoids in the development of stress or the metabolic syndrome 
, and have been considered a powerful protection against the damages caused by glucocorticoids in women and female rats 
. On the other side, males are more sensitive to glucocorticoids, which provoke hypoandrogenism 
In conclusion, the data presented show marked sex differences in corticosterone compartmentation in rats, largely due to different affinities or forms of CBG. These differences are magnified in rats which are overweight because of exposure to a cafeteria diet and some degree of inflammation. The glucocorticoid response was intense, both changing the expression of CBG gene in liver, the levels of binding in plasma and the proportion of blood cell-bound corticosterone, but not free corticosterone. The main compartment of corticosterone in rat blood is bound to plasma proteins (essentially through specific binding), with smaller compartments of hormone bound to blood cells or free.