Pharmacokinetics of centhaquin citrate in a rat model

Abstract

Objective

Centhaquin citrate is a novel agent being developed for use in the treatment of haemorrhagic shock. It has decreased mortality in rat, rabbit and pig models of hypovolaemic shock compared to hypertonic saline and lactated Ringer's resuscitation. The pharmacokinetics of centhaquin citrate have not been described to date.

Methods

Sixteen male Sprague Dawley rats were given an intravenous bolus of 0.45 mg/kg centhaquin citrate. Rats were divided into two groups; plasma concentrations were measured at five time points for each group within 24 h after administration. Competing compartmental pharmacokinetic models were assessed. The nonparametric adaptive grid function within the Pmetrics package for R was used for parameter estimation. Predicted concentrations were calculated using population median and individual Bayesian posterior parameters.

Key Findings

A two-compartment model of centhaquin citrate best fit the data. Median (IQR) values for elimination coefficient (K_e), volume of distribution (V) and intercompartmental transfer rates (K_cp,K_pc) were 8.8 (5.2–12.8) h⁻¹, 6.4 (2.8–10.4) l, 11.9 (4.6–15.0) h⁻¹ and 3.7 (2.3–9.1) h⁻¹, respectively.

Conclusion

This is the first report of the pharmacokinetic parameters of centhaquin citrate in a rat model. Centhaquin citrate was found to have a short half-life with a large volume of distribution.

centhaquin citrate, pharmacokinetics, rat

Introduction

Hypovolaemic shock, due to excessive blood loss, accounts for a large portion of post-traumatic deaths. The majority of deaths due to hypovolaemic shock occur within the first 6 h after trauma,^[1] and many of these deaths are preventable.^[2,3] Although fluid resuscitation with crystalloid, colloid or packed red blood cells is recommended in the Surviving Sepsis Campaign Guidelines^[4] to prevent mortality, aggressive fluid resuscitation is known to dilute plasma coagulation factors and exacerbate haemorrhage.^[5,6] Administration of large volumes of normal saline (NS) in trauma patients with shock may also contribute to metabolic acidosis, which can worsen coagulopathy^[7] and lead to further blood loss. Patients resuscitated with crystalloids have been reported to develop irreversible loss of capillary bed perfusion, coagulopathy, hypothermia, acidosis, immune suppression, systemic inflammation, oxidative stress, multiple organ failure and death.^[5,8–16] When volume resuscitation alone is not sufficient to sustain perfusion, vasopressors are administered.^[17] Intravenous catecholamine infusion enhances cardiac contractility as well as vascular tone and thus influences overall arterial, venous, and capillary pressures and blood flow.^[18] While the importance of prompt resuscitation is unquestioned, the optimal means of resuscitation has been vigorously debated.^[19] Clearly, an unmet need exists for novel resuscitative agents that minimize coagulopathy and maximize perfusion pressures.

As we have previously described,^[20] low doses of centhaquin (2-[2-(4-(3-methylphenyl)-1-piperazinyl)]ethyl-quinoline) citrate, when added to either lactated Ringer's solution (LR) or hypertonic saline (HS), significantly decreased blood lactate and increased mean arterial pressure (MAP), pulse pressure (PP) and cardiac output (CO) compared to LR and HS alone. In a rat model of fixed pressure blood loss, we demonstrated that centhaquin citrate, a cardiovascular active agent, is highly effective in reducing mortality following hypovolaemic shock.^[21–23] Animals treated with centhaquin citrate experienced decreased heart rates (HR) and no change or decrease in vascular resistance; however, animals that received centhaquin citrate experienced markedly increased MAP due to increased CO compared to controls. It is noteworthy that unlike vasopressors, centhaquin can increase MAP and tissue perfusion and decrease work load on the heart.

The aim of this study was to describe the pharmacokinetics of centhaquin citrate in the rat, a novel agent that has shown mortality benefit in animal models of hypovolaemic shock. The ultimate goal was to develop centhaquin as a novel agent to reduce mortality and morbidity among patients with hypovolaemic shock.

Methods

Animals

Male Sprague Dawley rats (mean 240 g, range 220–250 g) (NIPER, S.A.S. Nagar, Mohali, India) were housed for at least 4 days before being used in a temperature (23 ± 1 °C), humidity (50 ± 10%) and light (6:00 a.m. to 6:00 p.m.) controlled room. Food and water were made available continuously. All procedures and animal care protocols were approved by the Institutional Animal Ethics Committee (IAEC), NIPER, S.A.S. Nagar.

Drugs and chemicals

Centhaquin citrate was synthesized by Pharmazz India Private Limited (Greater Noida, UP, India) as previously described.^[24,25]. HPLC-grade acetonitrile, methanol, trifluoroacetic acid and water were obtained from Merck (Merck Specialities Pvt. Ltd., Mumbai, India). All the other reagents and materials used were of analytical grade.

Sampling and quantification

Rats received an intravenous bolus dose of 0.45 mg/kg of centhaquin citrate through tail vein injections. Rats were anesthetized with halothane (3%) (Halothane Anesthetic System, Harvard Apparatus, USA), and blood samples were withdrawn from retro orbital plexus using sterile capillaries. Sixteen rats were divided into two groups of eight. Samples (0.2–0.3 ml) were drawn from group A at 5 and 15 min, as well as 1, 4 and 12 h. Samples were drawn from group B at 10 and 30 min, as well as 2, 8 and 24 h after the dose. All samples were collected into chilled K3 EDTA vacutainers. Blood volume removed was replaced with equal volumes of normal saline. The blood samples were centrifuged at 1600g for 15 min at 0 °C, and separated plasma was frozen at −70 °C until it was assayed. All assays were conducted within 15 days of sampling. Samples were found to be stable for 45 days during the method validation studies (data not shown). Before analysis, plasma proteins were separated by methanol precipitation. Concentrations of centhaquin citrate were measured using prominence ultrafast liquid chromatography system (Shimadzu Corporation, Kyoto, Japan) consisting of solvent delivery units (LC-20 AD; Shimadzu) and variable wavelength photo-diode array detector (SPD-M20A; Shimadzu). Chromatographic separation was achieved using C₁₈ column (250 mm × 4.6 mm, 5 μm, 120Å pore diameter, Enable C₁₈ G; Spinco Biotech, Chennai, Tamil Nadu, India) with isocratic elution within a run time of 21.0 ± 1.0 min. The mobile phase consisted of 0.1% trifluoroacetic acid in water (A) and acetonitrile : methanol [50 : 50 v/v] (B) in ratio of A : B, 42 : 58. The mobile phase was delivered at a flow rate of 0.6 ml/min, and the detection of centhaquin citrate was carried out at 239 nm. Analyses were performed at 22 °C utilizing an injection volume of 20 μl. Data acquisition was performed using LC solutions software (Shimadzu Corporation). Assay error was determined by fitting various polynomials (i.e. first order, second order and third order) to five spiked concentrations ranging from the lower limit of quantitation to the upper limit of quantification, performed in replicates of 5. The R² and a rule of parsimony (i.e. the simplest yet most explanatory model) were used to select the final equation.^[26–28]

Compartmental model building

One- and two-compartment models were fit using the nonparametric adaptive grid (NPAG) algorithm within the Pmetrics package for R (Los Angeles, CA, USA).^[27,29] NPAG was used to create nonparametric mixed-effects models, with each model parameter being described as a set of discrete support points and the associated probability of each value.^[27] The covariate of animal weight was considered as a modifier of parameter estimates. Input (IV bolus) and output components were modelled as mass transit, and compartmental transfer was fitted using differential equations. Each measured concentration was weighted by 1/error² where error was defined as: error = SD × gamma. Model parameters included elimination coefficient (K_e), volume of distribution (V) and intercompartmental transfer rates (i.e. K_cp, K_pc).

Goodness of fit was assessed by regression of the observed–predicted plots, coefficients of determination and log-likelihood values, as previously described.^[30,31] Predictive performance was evaluated using mean prediction error and the mean bias-adjusted squared prediction error of the population and individual prediction models. The best fit model was determined by the rule of parsimony and use of the Akaike's information criteria (AIC) as calculated in Pmetrics.^[27]

Predicted centhaquin citrate concentrations were calculated at 1-min intervals based on median population parameters as well as individual Bayesian posterior parameter estimates for each patient.

To estimate pharmacokinetic parameters, a noncompartmental analysis was completed. Median individual Bayesian posterior predicted time observation profiles were used for calculations. Pharmacokinetic values were estimated using the ‘makeNCA’ command within the Pmetrics package for R.

Results

A representative chromatogram of detection of centhaquin is displayed in Figure 1. The assay was found to be accurate and illustrated linearity from 8.16 to 81.6 ng/ml with a regression coefficient (R²) of 0.9976 (Figure 2). The method was validated using three different concentration levels of centhaquin citrate (low, medium and high) as quality controls (QCs; n = 6/concentration) and the LLOQ (n = 6) prepared in rat plasma. The intra- and interday precision (RSD) ranged from 1.20 to 1.61% at LLOQ concentration and 0.59 to 4.70% at other QC levels. The intra- and interday assay accuracy ranged from 86.06 to 116.4% for LLOQ and 91.9 to 108.1% for other QCs. The mean extraction recovery of centhaquin citrate from plasma was 98.6% at LLOQ and ranged from 96.9 to 100.6% at other QCs. The RSD value of matrix effect was 1.61% at LLOQ concentration and 0.59–4.70% at other QCs. Assay error was well fit by a second order polynomial (i.e. SD = C₀ + C₁Y + C₂Y² with inputs of 0.000028, 0.014 and −0.097; R² = 0.977).

Figure 1

Chromatograms of centhaquin citrate acquired using prominence ultrafast liquid chromatography system. (a) Blank plasma spiked with centhaquin citrate and (b) plasma sample at 5 mins after intravenous injection of centhaquin citrate.

Figure 2

Standard curve of centhaquin citrate in plasma (Concentrations range used: 8.16–81.6 ng/ml).

Concentration–time curves for each of the rats (n = 16) based on the two-compartment model are shown in Figure 3a. Measured concentrations and fitted Bayesian-predicted concentrations are shown in Figure 3b. As the drug has a rapid elimination, Bayesian-predicted C_max values are much higher than measured values when backextrapolating. Weight did not explain variation in volume (R² = 0.01) and therefore was not incorporated into the final model. One- and two-compartment models each produced reasonable fits for the observed data. The two-compartment model was superior to the one-compartment model in terms of improved fit [as quantified with the observed vs predicted plot of the population (R² = 0.758 vs 0.71) and the Bayesian individual predictions (R² = 0.999 vs 0.971)]. Additionally, the two-compartment model exhibited a lower AIC compared to the one-compartment model (−1176 vs −1020). Thus, the two-compartment model was accepted as the final structural model. In the final population pharmacokinetic model, gamma was 2.04.

Figure 3

Concentration–time profiles for (n = 16) rats. (a) Measured concentrations plotted and directly connected. (b) Bayesian fit with x-axis truncated at 0.3 h.

Figure 4 shows the observed vs predicted plots for both the median population parameters and the median individual Bayesian posterior parameter values. Bias and imprecision for the population model were −3.17 ng/ml and 63.6 ng²/ml², respectively. The Bayesian posterior predicted model had bias and imprecision values of 0.121 ng/ml and 1.03 ng²/ml², respectively.

Figure 4

Observed vs predicted plots for (a) the population and (b) Bayesian-predicted individual values. Two-compartment models of observed vs predicted plasma concentrations of centhaquin citrate (R² = 0.755 and 0.999). Predictions are based on population parameters (top) and Bayesian posterior individual values (bottom).

Sixteen discrete support points for the parameter values were identified. Measures of central tendency and precision of the parameter estimates are shown in Table 1. Physiologic parameters (i.e. K_e and V) exhibited reasonable %CVs (<60%) and were thus considered acceptable as population estimates. Likewise, the interparameter covariance is shown in Table 2 as the lower-triangular form of the covariance matrix (describing the covariance of the individual pharmacokinetic parameters).^[26,27]

Table 1

Measures of central tendency and precision for parameter estimates

	Mean	Standard error of the mean (SEM)	Coefficient of variation	Variance	Median
Ke (h−1)	8.87	1.06	47.74	17.92	8.80
V (l)	6.45	0.96	59.17	14.56	6.37
Kcp (h−1)	9.97	1.31	52.59	27.46	11.94
Kpc (h−1)	5.98	1.32	88.19	27.84	3.73

V, volume of distribution; K_e, elimination constant; K_cp, transfer constant from central to peripheral compartment; K_pc, transfer constant from peripheral to central compartment.

Table 2

Covariance matrix for parameters

	Ke (h−1)	V0 (l)	Kcp (h−1)	Kpc (h−1)
Ke	17.92
V	−7.07	14.56
Kcp	−2.22	9.87	27.46
Kpc	−5.59	9.68	5.11	27.84

K_e, elimination constant; K_cp, transfer constant from central to peripheral compartment; K_pc, transfer constant from peripheral to central compartment; V, volume of distribution.

A noncompartmental analysis using the Bayesian posteriors is shown in Table 3. The median (IQR) K_e was 8.8 (5.2–12.8) h⁻¹. Additionally, the volume of distribution (V) was estimated to be relatively large (median: 6.4 l; IQR: 2.8–10.4 l). In the model, drug left the central compartment (median: 11.9 h⁻¹; IQR 4.6–15.0) quicker than it re-entered (median: 3.7 h⁻¹; IQR 2.3–9.1). The median (IQR) predicted terminal half-life was short [0.6 (0.35–1.0) h], rapid time to maximum concentration, and the median (IQR) population predicted volume of distribution at steady state was high [17.6 (13.0–31.0) l] (Table 3).

Table 3

Noncompartmental analysis derived pharmacokinetic properties

	Median (IQR)
AUC (ng/ml h)	3.0 (1.6–4.1)
AUC (0-inf) (ng/ml h)	3.0 (1.6–4.1)
Cmax (ng/ml)	14.8 (6.8–27.8)
Tmax (h)	0.017 (0.017–0.017)
Cl (l/h)	35.2 (26.2–68.7)
Vdss (l)	17.6 (13.0–31.0)
Thalf (h)	0.55 (0.26–0.94)

AUC, area under the concentration–time curve; AUC (0–inf), area under the concentration–time curve from time 0 to infinity; C_max, maximum plasma concentration; Cl, clearance, calculated as dose/AUC(0–inf); IQR, interquartile range; Thalf, half-life; T_max, time to C_max; Vdss, volume of distribution at steady state.

Discussion

Centhaquin citrate is a novel resuscitative agent in development for haemorrhagic shock. Previously, we conducted efficacy studies in rats that were bled to a fixed pressure of 35 mmHg for 30 min before initiation of resuscitation with centhaquin citrate.^[20,22,23] The present study is the first to our knowledge to describe the pharmacokinetic properties of centhaquin citrate in rats and demonstrated that a two-compartment structural model best fit the data.

Using parametric and nonparametric pharmacokinetic modelling, we were able to estimate population median and individual Bayesian pharmacokinetic parameters with reasonable precision. Notably, a R² = 0.758 for the parametric population model is high and represents a controlled laboratory experiment.^{[26,27,30,31]} Bayesian predictions utilizing the nonparametric support points improved fit to R² = 0.999. Given the improvements in model fit, even for very similar exbred rats, greater variation may be seen in more heterogeneous populations. Thus, titrating to effect or completing Bayesian modelling may be needed.

Centhaquin citrate displayed a relatively short half-life in noncompartmental analysis (33 min) as well as a large volume of distribution from model parameter estimates (median 17.6 l/kg; IQR 13.0–31.0 l/kg). It is worth noting that the apparent half-life observed from the graphical data is much more rapid than the 33-min terminal half-life calculated with noncompartmental analysis. Changes in drug concentration for centhaquin are triphasic in the rat and explained well from our two-compartment model. When centhaquin is first given, the first phase of drug concentration decline is rapid and is explained by our model as K_e + K_cp. At steady state (i.e. the second phase), drug concentration decline is defined by the sum of the elimination and distributive properties (i.e. transfer from plasma to tissue and vice versa). That is, drug concentration change is the sum of the elimination constant (K_e) plus the peripheral distribution constant (K_cp) minus the distribution to central compartment constant (K_pc). The sum of the mean estimates for our study is 12.85 h⁻¹. Thus, at steady state, centhaquin will have a half-life of 3.24 min (i.e. 0.693/12.85 h⁻¹ × 60 min). Finally during the third phase, presumed tissue release occurs, and prolonged low concentrations are defined by K_pc. It is this last phase that results in a higher noncompartmental analysis terminal half-life estimate of 33 min. After considering the drug concentration change at steady state for centhaquin (i.e. half-life of ~3 min), parameter value estimates are similar to those seen with other vasoactive agents such as norepinephrine and dopamine.^[32,33] Further study of the pharmacodynamic impact of the triphasic concentration change for centhaquin will be important. If higher concentrations are needed for prolonged maintenance of MAP and CO during haemorrhagic shock, frequent dosing or continuous infusion may be desirable. Additionally, it will be important to learn if drug concentration decline is secondary to metabolism or elimination.

This study is limited in that it represents a single-species pharmacokinetic study utilizing a total of 16 rats. As this was the first pharmacokinetic study performed in rats, a single-dose design was chosen; however, the rapid half-life identified in this study likely predicts that drug accumulation will be minimal. Additional pharmacokinetic and pharmacodynamic studies in other species are in progress. While the number of rats studied was somewhat limited, there is an impetus to minimize repeated studies in animals^[34] and rich pharmacokinetic sampling was performed. The coefficients of variation demonstrated that parameter estimates were reasonably precise (i.e. <60% for all parameters except K_pc, which was 79%). However, more heterogeneous populations (e.g. critically ill patients) may display even greater variability. Thus, dose titration and individualization on the basis of pharmacokinetic variability may be necessary. It is highly reassuring that Bayesian posterior estimates resulted in excellent predictions (R² = 0.999). Future studies should continue to explore whether the addition of other covariates improves population estimates through the inclusion of fixed effects.

In summary, the population pharmacokinetic of centhaquin citrate in rats is now more clearly defined. Once optimal target exposures have been defined for pharmacodynamic effect, optimal dosing regimens can be derived.

Declarations

Conflict of interest

Anil Gulati has issued and pending patents and Manish Lavhale is employed by Pharmazz, Inc. having rights to those patents. Drs O'Donnell, Rhodes, Sharma, Scheetz and Patel have no relevant conflicts of interest to report.

Funding

Anil Gulati has issued and pending patents and Manish Lavhale is employed by Pharmazz, Inc. having rights to those patents. No relevant funding related to this work was supported by Drs O'Donnell, Rhodes, Sharma, Scheetz and Patel.