Discussion
This was the first manuscript investigating the role of diabetes on GBP pharmacokinetics in humans. A population pharmacokinetic model was developed to investigate the influence of type 2 diabetes and glycaemic control and other potential covariates on GBP kinetic disposition. The pharmacokinetic estimates presented here are similar to parameter values previously reported for GBP, except for the lower values of volume of distribution and ka when compared to clinical trials with non-diabetic participants [44,48]. Our data showed that the volume of distribution of GBP was affected by body height and serum levels of glucose, while the total clearance was affected by eGFR (Table 3). GBP is primarily eliminated unchanged in urine and associations between eGFR and GBP pharmacokinetics have been reported previously [53-56].
Increased renal clearance of GBP was observed in rats with experimental diabetes induced by streptozotocin [57], suggesting that the effects of diabetes on the kinetic disposition of GBP occurred by inducing glomerular hyperfiltration [58]. While the experimental model of diabetes follows a strict protocol in rats in terms of duration of the disease, this clinical study includes patients with different levels of renal function and duration of diabetes. The results presented here have shown that eGFR is a covariate on renal clearance of GBP. This finding means that GBP kinetic disposition depends on the nephropathy level, which is indirectly related to the glycaemic levels [59]. Among type 2 diabetic patients, 20 to 40% develop diabetic nephropathy (DN) [60], which consists of 5 steps: 1. Increase in eGFR and glomerular hypertrophy; 2. Hyperfiltration and microalbuminuria (> 30 mg/24 h); 3. Higher microalbuminuria (> 300 mg/24 h) and hypertension; 4. Microalbuminuria (> 300 mg/24 h), decrease on eGFR and increase in creatinine and blood urea nitrogen; 5. eGFR < 10 mL/min, which leads to haemodialysis [61]. A well-accepted theory for DN is that hyperglycaemia increases reactive oxygen species and pro-inflammatory cytokines [62-64].
Diabetes typically alters the expression and function of transporters for organic cations in mice with experimentally induced type 2 diabetes, probably due to the accumulation of end products of advanced glycation and inflammation [7]. Drug transporters for organic cations such as OCT2, MATE 1 and 2-K, and OCNT1 have been described to contribute to GBP renal excretion [43,44]. Although GBP has been described as an OCT2 substrate [42,43], the interaction with OCT2 is not relevant at therapeutic drug concentrations [44]. No significant changes in GBP kinetic disposition were observed after the coadministration of cimetidine (a known inhibitor of OCT2) or metformin (a known substrate of OCT2) in rats [57]. Moreover, cetirizine, an inhibitor of OCT2, MATE1 and 2-K [65,66], reduced the systemic exposure to GBP with no changes in renal clearance in patients with neuropathic pain, suggesting an interaction in the oral absorption process mediated by active transport (probably OCTN1) and not by renal drug transporters [44].
The effect of glycaemic control on the clinical pharmacokinetics of GBP observed here could be explained by the saturation of absorption processes, since the apparent volume of distribution is dependent on oral bioavailability [67,68]. OCTN1 is expressed in gut cells and might be responsible for the saturable absorption of GBP [69,70]. In type 1 diabetic mice, the renal protein expression of Octn1 is decreased [71]. The protein level of intestinal Octn1 follows a circadian rhythm, both in mice with diabetes induced by streptozotocin and in mice without diabetes [72]. However, there is no information in the literature about the impact of high glycaemic levels on the gut levels of OCTN1.
GBP intestinal uptake is mediated by L-type transporter (LAT) 2 [73-76] and by the system b0,+ together with peptide transporter (PEPT) 1. These transporters might be associated with the saturable absorption of GBP [77]. Rats with experimental diabetes showed a reduction in the expression and activity of the pept1 transporter mRNA [78]. In rabbits with maternal diabetes induced by alloxan, the transcripts of LAT2 were increased in blastocysts, when compared to blastocysts of non-diabetic rabbits [79]. The reduction in the activity of PEPT1 by hyperglycaemia [76] seems to be a reasonable explanation for our findings which showed glycaemic level was a covariate in GBP apparent volume of distribution.
In therapeutic concentrations, the distribution of GBP to the central nervous system is regulated by LAT1 uptake [80]. The high concentration of glucose reduced LAT1 expression by 80%, compared to cells without the excess of glucose [81]. In opposition to the expected reduction on LAT1 expression, the effect of hyperglycaemia on LAT1 does not seem to influence GBP plasma concentrations since higher glycaemic levels are associated with lowering GBP plasma concentrations. Despite the lack of effect of LAT1 in GBP plasma levels, patients with uncontrolled diabetes could have an impact on drug concentrations in the effect compartment [75,82].
This work has some limitations. Firstly, participants were not genotyped for genetic polymorphisms of drug transporters involved in the absorption process of GBP. The polymorphisms c.438C>G (rs1060253) of the gene SLC7A5 (LAT1) and 1347T>C (rs1339067) of the gene SLC15A1 (PEPT1) are involved in risperidone and sirolimus pharmacokinetics [83,84]. Secondly, although the basal pain score on VAS was considered for inclusion of the participants, this study focused only on GBP pharmacokinetics. A population model relating pharmacokinetics to pharmacodynamics would be of great importance. Thirdly, the diabetes-induced epigenetic modifications on targets related to drug pharmacokinetics were not investigated. The possibility of a maintained lesion secondary to the hyperglycaemic memory [10,11] could not be ruled out to explain the changes in GBP clearance and volume of distribution in hyperglycaemic patients.
In conclusion, GBP population pharmacokinetics was influenced by renal function and by the serum levels of glucose. Genetic polymorphisms of OCT2 and OCTN1 transporters, sex, age, weight or BMI did not influence GBP population pharmacokinetics. Our data suggest normal glycaemic levels in diabetic patients, achieved either by adherence to diabetes pharmacotherapy and changes in lifestyle (diet and physical exercise), reduces the variability in the kinetic disposition of GBP. In addition to the benefits related to diabetes comorbidities and patients’ quality of life, the control of glycaemic levels is also relevant to achieve positive pharmacotherapy outcomes.