Skip to main content
NCBI home page
Drug Metabolism and Disposition logoLink to Drug Metabolism and Disposition
. 2011 Jul;39(7):1214–1220. doi: 10.1124/dmd.110.036855

Receptor-Based Dosing Optimization of Erythropoietin in Juvenile Sheep after Phlebotomy

Matthew Rosebraugh 1, John A Widness 1, Peter Veng-Pedersen 1,✉
PMCID: PMC3127239  PMID: 21460232

Abstract

The primary objective of this work was to determine the optimal time for administration of an erythropoietin (Epo) dose to maximize the erythropoietic effect using a simulation study based on a young sheep pharmacodynamic model. The dosing optimization was accomplished by extending a Hb production pharmacodynamic model, which evaluates the complex dynamic changes in the Epo receptor (EpoR) pool from the changes in Epo clearance. Fourteen healthy 2-month-old sheep were phlebotomized to Hb levels of 3 to 4 g/dl. Epo clearance was evaluated longitudinally in each animal by administering tracer doses of 125I-recombinant human Epo multiple times during the experiment. Kinetic parameters were estimated by simultaneously fitting to Hb data and Epo clearance data. The phlebotomy caused a rapid temporary increase in the endogenous Epo plasma level. The Hb began to increase after the increased in the Epo level with a lag time of 1.13 ± 0.79 days. The average correlation coefficients for the fit of the model to the Hb and clearance data were 0.953 ± 0.018 and 0.876 ± 0.077, respectively. A simulation study was done in each sheep with fixed individual estimated model parameters to determine the optimal time to administer a 100 U/kg intravenous bolus Epo dose. The optimal dose administration time was 11.4 ± 6.2 days after phlebotomy. This study suggests that the Hb produced from Epo administration can be optimized by considering the dynamic changes in the EpoR pool.

Introduction

Erythropoietin (Epo) is a glycoprotein hormone responsible for regulating erythrocyte production. Epo exerts its mechanism of action by binding to erythropoietin receptors (EpoRs) located on erythroid progenitor burst-forming units and colony-forming units found primarily in the bone marrow(Casadevall, 1995).

Information regarding the in vivo clearance of Epo remains incomplete. Some studies have shown that Epo is cleared by the liver (Fukuda et al., 1989; Spivak and Hogans, 1989) and the kidney (Jensen et al., 1994); however, their effects do not contribute significantly to the total clearance (Jelkmann, 2002). It has been hypothesized that in vivo desialidation is the rate-limiting step for Epo metabolism by the liver (Nielsen et al., 1990). Receptor-mediated endocytosis in the bone marrow by erythroid progenitors followed by lysosomal degradation is the primary mechanism of Epo elimination from the body (Sawyer et al., 1987). Additional evidence for this elimination mechanism has been shown by studies investigating different degrees of bone marrow activity (Beguin et al., 1993; Cazzola et al., 1998). Several preclinical and clinical studies have reported an increase in Epo clearance within 1 month after Epo treatment (Kinoshita et al., 1992; Ohls et al., 1996; Widness et al., 1996; Sans et al., 2000), whereas others found no statistical difference (Salmonson et al., 1990; Kampf et al., 1992). An additional study modeled the change in the EpoR level over time after the induction of anemia (Chapel et al., 2001). In a recent study by our group, we reported up-regulation of EpoR mRNA levels of 4.97 ± 3.92 times baseline at 9 days after the induction of anemia (Nalbant et al., 2010). Although EpoR mRNA levels do not necessarily correlate directly with the number of receptors, an increase in the mRNA level is probably associated with an increase in the quantity of EpoRs. An important aspect of Epo dosing is that receptor-mediated Epo clearance (CLR) leads to Hb production, whereas the nonreceptor-mediated clearance (CLL) does not produce Hb. Thus, the greatest efficacy in Epo dosing is achieved when the biggest fraction of the dose is eliminated via the erythropoietic elimination pathway. One study was able to characterize and quantify both types of clearances through a chemical bone marrow ablation method (Veng-Pedersen et al., 2004). With two different types of clearances, optimal Epo dosing involves using the receptor-mediated clearance pathway as much as possible and minimizing the fraction of Epo eliminated via the nonerythropoietic elimination pathway. Thus, the primary objective of this study was to determine the optimal Epo dosing time in neonatal sheep with phlebotomy-induced anemia, considering the complex, receptor-mediated elimination mechanism of Epo. The dosing optimization was based on a mechanistic, pharmacodynamic model for predicting Hb production, which considers the dynamic Epo-dependent changes in the EpoR pool that determines erythropoietic efficacy.

Pharmacodynamic analysis studies have modeled different variables in response to the erythropoietic stimulatory effect of Epo. Prior studies have examined the effect of Epo on reticulocytes (Chapel et al., 2000) and Hb (Al-Huniti et al., 2004) individually, and reticulocytes and Hb together (Veng-Pedersen et al., 2002). An additional mechanistic PK/PD study took into consideration the Epo, reticulocyte, and EpoR levels to determine the Hb response to Epo (Woo et al., 2007). Although it is commonly stated that the binding of Epo to EpoR accounts for erythrocyte production and target-mediated disposition, the quantity of EpoRs has not previously been used to predict the pharmacodynamic response of Epo. Nonetheless, target-mediated drug disposition pharmacokinetics has been considered in different contexts previously (Shepherd, 1989; Levy, 1994; Mager and Jusko, 2001). The modeling-based methodology used in this work is not limited to Epo but has been applied to other therapeutic agents with target-mediated disposition such as recombinant tumor necrosis factor (Blick et al., 1987) and selected angiotensin-converting enzyme inhibitors (Till et al., 1984). The fundamental modeling objective was to determine the kinetic mechanism governing the complex relationship between the Epo concentration and the EpoR pool size, both of which are determinants of Hb production. The analysis we have chosen proposes a feedback regulation mechanism for EpoRs and is based on the assumption that the receptor-mediated Epo elimination pathway results in an Epo clearance that is linearly related to the quantity of EpoRs.

Epo therapy has been well established in patients with renal failure (Salmonson et al., 1990) and patients undergoing chemotherapy treatment (Rizzo et al., 2010). However, the benefit of Epo therapy in anemic preterm infants is still under investigation. Clinical trials involving Epo administration in preterm infants have shown some significant results, although the clinical significance is inconsistent (Ohlsson and Aher, 2006). Young sheep have previously been used to model human anemia (Veng-Pedersen et al., 1999) and may provide guidelines for how to optimally administer Epo in preterm infants.

Materials and Methods

All surgical and experimental procedures were approved before the study by the local animal care review committee. Fourteen healthy sheep ages 21 to 60 days were studied. All animals were housed in an indoor light- and temperature-controlled environment. Jugular venous catheters were inserted into the sheep for undisturbed plasma sampling. To prevent infection, intravenous ampicillin (1 g) was administered for the first 3 days after insertion of the venous catheters. Anemia was induced by an exchange transfusion using 0.9% NaCl. The procedure for the exchange transfusion included removing blood from the venous catheter until Hb levels had fallen to 3 to 4 g/dl. While blood was being removed from the animal, an equal volume of 0.9% NaCl was infused back into the animal to keep a constant total volume of blood in the animal. Hb concentrations were determined using a veterinary flow cytometer (XT-2000; Sysmex, Kobe, Japan).

To determine the Epo clearance, tracer doses of biologically active 125I-rhEpo equivalent to 0.1 U/kg Epo were given intravenously over less than 30 s. Ten to 15 plasma samples were drawn over the next 7- to 8-h period after Epo dosing. The concentration of 125I-rhEpo was determined by immunoprecipitation as described previously (Widness et al., 1992). To evaluate dynamic changes in the Epo clearance, 125I-rhEpo clearance studies were completed before and after the phlebotomy. In addition to the tracer doses of Epo, endogenous plasma Epo concentrations were measured at multiple time points during the study by using a radioimmunoassay as described previously (Widness et al., 1992).

Pharmacokinetic Analysis.

The Epo plasma concentration data from a single intravenous 125I-rhEpo tracer bolus dose was fitted using a biexponential function to determine the Epo clearance based on the dose administered and the resulting area under the curve. Curve fitting and parameter estimation were performed using WINFUNFIT, a general nonlinear regression program evolved from the FUNFIT program (Veng-Pedersen, 1977). The dosing optimization algorithms used with WINFUNFIT were programmed using FORTRAN.

Pharmacodynamic Model for Hb Response.

A pharmacodynamic model was developed to predict the Hb response to induced anemia based on current understanding of the mechanism of action of Epo. In particular, the model assumes that Hb is produced only after the binding of Epo to its receptor, EpoR. It was also assumed that the receptor-based clearance of Epo (CLR) is proportional to the EpoR pool size (Chapel et al., 2001). Finally, it was assumed that before the start of the study the system was at steady state; i.e., the Hb production was initially equal to the Hb removal resulting from red blood cell (RBC) senescence. RBC senescence was accounted for in the model by assuming that no senescence occurred among newly produced RBCs after the phlebotomy within the time span of the experiment. This assumption is justified on the basis of the assumption of a 120-day lifespan of RBCs reported for sheep (Mock et al., 1997), which is longer than the approximately 30-day duration of the kinetic experiment. The model incorporated a lag time (tlag) between the Epo-EpoR binding and the production of Hb that corresponded to the time from the binding of Epo to EpoRs in the bone marrow and the appearance of newly formed RBCs, i.e., reticulocytes in the peripheral bloodstream.

Modeling for the CLR is similar to our group's previous work (Chapel et al., 2001) in that the feedback mechanism for the CLR is modeled through the use of an intermediate variable denoted M. The intermediate variable in this model provides a mathematical method to account for the clearance rebounding above the baseline value. A representative fitted Hb profile with the Epo concentrations is illustrated in Fig. 1. The following numerically integrated delay-type differential model equations (eq. 1) were simultaneously fit in all animals to the Hb and clearance versus time data:

graphic file with name zdd00711-5344-m01.jpg

where Hb is the Hb amount, CLR is the receptor-based clearance of Epo assumed to be proportional to the quantity of EpoR, CLR, SS is steady-state value of CLR, M is the intermediate variable used to account for the feedback, Epo is the Epo plasma concentration represented with a linear spline, τ is the lifespan of RBCs fixed to 120 days (Mock et al., 1997), k3 is a first-order elimination rate constant for M, k1 is the rate constant for Epo-EpoR binding; k0 is the zero-order rate constant for M formation, and k2 is a first-order feedback rate constant. The + subscript denotes the truncation function (eq. 2); i.e.,

graphic file with name zdd00711-5344-m02.jpg

E(t) is the Hb production given by a modified Hill equation model (eq. 3):

graphic file with name zdd00711-5344-m03.jpg

where tlag is the lag time between Epo-EpoR binding and production of Hb, EC50 is the concentration of Epo resulting in half of the maximum Hb production rate, and Emax(t) is the maximum Hb production possible at the current time. Emax, SS is the initial SS value of Emax; i.e., Emax, SS = Emax(0).

Fig. 1.

Fig. 1.

Open in a new tab

Representative fit of pharmacodynamic model to Hb (top) and CL data (bottom). Erythropoietin data are represented with a linear spline function (top).

In the differential equations presented, Hb is the amount of Hb, but the experimental data represent the Hb concentration. Thus, the output variable of the differential equation must be divided be V(t) (eq. 4):

graphic file with name zdd00711-5344-m04.jpg

where Vkg is the volume of distribution per kilogram, which is considered a constant and W(t) is the weight of the animal at time t. The weight function is represented by a linear spline fit to the weight data of each animal.

Previous literature has shown that the red cell lifespan (τ) for sheep is approximately 120 days (Mock et al., 1997). For the 14 sheep used in this study, the average study time was 34.6 ± 5.7 days. Therefore, the lifespan (τ) of RBCs is much longer than the study length. Thus, the above differential equations become (eq. 5):

graphic file with name zdd00711-5344-m05.jpg

subject to the SS initial conditions in eq. 6:

graphic file with name zdd00711-5344-m06.jpg

The Epo concentration was used as a forcing function for the differential equations by representing the Epo response by a linear spline fit to the plasma Epo data. The tracer kinetic data enable the total clearance (CL) of Epo to be obtained. Previous studies have shown that the total clearance includes a second linear clearance (CLL) that is assumed to be unaffected by anemia and independent of the Epo concentration (Veng-Pedersen et al., 2004). Therefore, the total Epo clearance may be expressed as eq. 7:

graphic file with name zdd00711-5344-m07.jpg

where CLL is an estimated constant and ClR(t) is the receptor-based clearance that changes according to the dynamic changes in the receptor pool. The above total clearance equation and the numerically integrated differential equations define the model that was fit to the clearance and Hb data.

Dosing Optimization.

The objective for the dosing optimization was to find the optimal time to administer a single intravenous 100 U/kg bolus Epo dose to anemic neonatal sheep based on a simulation study. To accomplish this, an additional differential equation (eq. 8) was added to the model:

graphic file with name zdd00711-5344-m08.jpg

where Epox is the exogenous Epo concentration resulting from an intravenous bolus injection and V(t) is the weight-adjusted volume of distribution at time t. The objective function for the optimization is based on the total amount of postphlebotomy Hb produced to the end of the study when a fixed 100 U/kg intravenous bolus dose of Epo is administered at a certain time (eq. 9):

graphic file with name zdd00711-5344-m09.jpg

where Hb End is the amount of Hb in the animal at the end of the study, Hb Start is the amount of Hb in the animal immediately after the phlebotomy, and Hb Senescence is the Hb amount removed from the circulation because of natural erythrocyte senescence. It is assumed that CLL is constant, that CLR is proportional to the EpoR pool, and that the Epo dosing does not significantly change the endogenous Epo production.

Simulation Study.

The simulation study involved a dose simulation for each of the 14 sheep with fixed individual PK/PD parameters that were determined individually from the model based on eq. 5. The individual Emax, EC50, k1, k2, k3, τ, and lag-time parameters were fixed for each individual dose simulation. The software developed for the analysis determined the postphlebotomy Hb response for a sheep when a single 100 U/kg intravenous bolus dose is given at a particular time. For each sheep and each simulated Epo dosing, the software determined the postphlebotomy Hb produced on the basis of eq. 9. The postphlebotomy dosing time that produced the most Hb was considered to be the optimal dosing time. This optimal dosing time was determined by numerical optimization using the derivative-free optimization algorithm developed by Nelder and Mead (1965), commonly referred to as the simplex method. Likewise, the simplex optimization algorithm was also set up to determine the lowest postphlebotomy Hb produced according to eq. 9 for a single 100 U/kg intravenous Epo bolus dose. The dosing time with the lowest Hb produced was considered to be the worst possible dosing time.

Results

Study animals began with a mean prephlebotomy Hb of 10.5 ± 2.22 g/dl and a mean starting weight of 14.4 ± 6.26 kg; mean ending weight was 21.8 ± 7.07 kg. All animals were phlebotomized to a similar Hb value of 3.87 ± 0.450 g/dl. Figure 1 shows a representative fit of the proposed model to the clearance and the Hb data. As expected, in all cases there was a marked increase in the plasma Epo concentration after the phlebotomy followed by an increase in Hb. Furthermore, every animal showed a decrease in Epo clearance immediately after the phlebotomy followed by an increase in clearance above each animal's baseline clearance once plasma Epo levels had declined to nearly normal levels. The sheep reached an average peak clearance of 110.3 ± 25.90 ml/h · kg at 5.57 ± 2.12 days after the phlebotomy.

The simultaneous curve fitting was able to capture the behavior of the data for all animals. Although there was an age difference of up to 39 days for some animals, there was no statistical difference (p > 0.05) in baseline clearance values or parameter values for the different age groups. Table 1 summarizes the parameters and goodness-of-fit values. The curve fitting showed a trend toward a rapid increase in Hb level of production followed by a more gradual increase in Hb concentration. In all cases, the Hb levels continued to increase when plasma Epo concentrations were at baseline as a result of the EpoR pool size being above the baseline value. The average lag time for the study was 1.13 ± 0.794 days. The average Epo steady-state clearance value for all of the animals was 62.7 ± 17.5 ml · h−1 · kg−1. The average maximum and minimum clearance values for all of the animals were 94.3 ± 28.8 and 24.7 ± 9.8 ml · h−1 · kg−1, respectively.

TABLE 1.

Parameter value summary for pharmacodynamic model

Parameter Mean Value S.D.
Emax (g/day) 0.333 0.412
EC50 (mU/ml) 578 1160
Lag time (days) 1.13 0.794
k3 (l/day) 0.807 0.359
k1 (ml/mU per day) 9.75 × 10−3 4.63 × 10−3
k2 (1/day2) 16.3 10.9
Hb correlation r 0.953 0.0180
Clearance correlation r 0.876 0.0770
Open in a new tab

The simulation of exogenous Epo administration caused a variable increase in postphlebotomy Hb production, which was dependent on the time the Epo dose was administered (Fig. 3). A representative plot of this increased Hb production is shown in Fig. 2, top. These Hb profiles are based on the Epo profiles shown in Fig. 2, bottom. The two Hb profiles in Fig. 2 are superimposable until the Epo dose is given. In all cases, there were two local solutions for the dosing optimization algorithm, which are summarized in Tables 2 and 3. Figure 3 shows the objective function values (Hb produced) as a function of the intravenous bolus dosing time. The plot demonstrates two local solutions found for a particular animal. For 11 of the 14 animals, the later dosing time was more optimal than the first. The average for the optimal dose administration time was 11.4 ± 6.2 days postphlebotomy, which corresponded to an average increase in Hb production of 21.1 ± 9.7% over the worst possible dosing time. The average for the second local solution for the optimal dosing time was 4.6 ± 4.4 days postphlebotomy, which corresponded to an average increase in Hb production of 10.2 ± 5.4% over the most unfavorable possible dosing time. The averages for the two optimal dosing times could be estimated in a different manner as dealt with under Discussion. The average worst possible dosing time was 6.6 ± 3.5 days postphlebotomy. The average Hb produced was 186.2 ± 50.5 g for the optimal dose time, which was significantly larger (p < 0.05) than for the worst possible dose time (152.1 ± 33.6 g).

Fig. 3.

Fig. 3.

Open in a new tab

Representative plot of amount of Hb produced as a function of simulated Epo dosing time for one animal.

Fig. 2.

Fig. 2.

Open in a new tab

Representative estimated Hb profile (top) for a model with a simulated intravenous bolus Epo dosing (broken curve) and without any Epo dosing (continuous curve). Representative Epo profiles (bottom) with an intravenous bolus Epo dosing simulation.

TABLE 2.

Summary of dosing optimization for best possible solution

Subject No. Optimal Postphlebotomy Dose Time 1 Increase in Objective Function
days %
1 1.40 23.0
2 13.2 20.9
3 2.20 24.3
4 13.7 33.5
5 14.1 30.0
6 7.90 26.5
7 13.0 31.3
8 16.8 12.3
9 18.1 26.0
10 18.0 7.80
11 19.1 6.50
12 1.90 6.10
13 12.9 15.6
14 7.40 31.2
Mean 11.4 21.1
S.D. 6.20 9.70
Open in a new tab

TABLE 3.

Summary of second dosing optimization solution

Subject No. Optimal Postphlebotomy Dose Time 2 Percentage Increase in Objective Function
days %
1 13.0 15.6
2 4.10 13.7
3 12.0 4.50
4 2.00 14.8
5 2.40 11.5
6 1.50 12.2
7 1.20 12.9
8 2.10 3.30
9 4.10 6.70
10 0.900 6.80
11 3.40 4.30
12 12.8 4.30
13 3.00 10.4
14 2.50 21.7
Mean 4.60 10.2
S.D. 4.40 5.40
Open in a new tab

Discussion

Judged from the goodness-of-fit parameters, the proposed kinetic model was able to accurately predict the Hb production in phlebotomized sheep on the basis of the Epo concentration and Epo clearance values. Furthermore, the model was extended to determine the optimal time to administer an intravenous bolus Epo dose to anemic sheep. Because the Epo CLR is assumed to be proportional to the change in receptors and because the linear clearance is assumed to be constant (Veng-Pedersen et al., 2004), any change in the total clearance, CL, is due to a change in the quantity of EpoRs. In addition, because the EpoRs are located primarily in the bone marrow, phlebotomy does not significantly change the EpoR pool. Our group recently showed that EpoR mRNA is up-regulated after the induction of anemia in sheep, which is consistent with the model used in this study (Nalbant et al., 2010). Because this study used the GAPDH gene as a control, it can be suggested that the number of EpoRs per cell increased rather than the total number of cells. Another study released recently shows that the Epo receptors decrease after large plasma Epo concentrations, demonstrating consistency with our model (Becker et al., 2010). However, this study does not show EpoRs increasing above the baseline value, which is most likely due to the short time scale (5 h) of the study. The present simulation study provides additional evidence that the increased total clearance of Epo is probably due to an increase in the number of EpoRs present in the bone marrow. The agreement of the present model provides additional indirect evidence for a receptor-mediated clearance mechanism for Epo. In particular, the decrease in the total Epo clearance immediately after the endogenous Epo surge suggests that the EpoR is consumed in a receptor-mediated endocytic process.

There are several limitations of this study that may be improved in future experimental work. One limitation is that an indirect method was used to determine the EpoR pool. In addition, the lifespan (τ) of red blood cells is fixed to 120 days to account for senescence. In reality, the red blood cell lifespan could be variable throughout the course of the animal's life. Although the simulation study suggested a potential Hb production benefit based on consideration of the dynamic EpoR state, experimental work needs to be done to verify the simulations. The study is also based on the assumption of an initial EpoR steady state, which may not be true.

Work by others has shown the potential for scaling preclinical animal Epo studies to humans (Mager et al., 2009). Thus, with proper allometric pharmacodynamic scaling, it should be possible to make use of the sheep model for pharmacodynamic dosing predictions useful in planning mechanism-based human efficacy studies.

Dosing Optimization Solutions.

The dosing optimization algorithm obtained two solutions for the optimal time to administer an intravenous bolus Epo dose. The average dose administration time was reported as the average for the best solution and the second best solution. To better analyze the optimal dosing times, it is more informative to average the solutions by the early solution and the late solution. If the averages are computed in this manner, then the average early postphlebotomy Epo administration solution is 2.3 ± 1.0 days and the average late solution is 13.7 ± 3.5 days postphlebotomy. The later dosing time was expected and is probably due to the EpoR state being up-regulated around the optimal dosing time. The earlier solution for the dosing optimization was not expected at first. However, this solution may be explained by an earlier Epo dosing, allowing the endogenous Epo to make use of the up-regulated EpoRs, resulting in a greater efficacy of the endogenous Epo.

The optimization algorithm showed a 21.1 ± 9.7% increase in the amount of Hb produced for the best dosing time above that of the worst dosing time. This percentage corresponds to an average difference in Hb of 34.0 g for the best dosing time compared with the worst dosing time. The estimate for sheep blood volume for our study is 98.1 ml/kg. With this value we can estimate, on the basis of the average weight of the sheep in this study, that a Hb difference of 34.0 g would correspond to an improvement in Hb concentration of 1.57 g/dl.

Linear Clearance Mechanism.

Understanding of the linear nonerythropoietic clearance component of Epo remains incomplete. Our previous bone marrow ablation study in nonanemic lambs showed that the linear clearance of Epo accounts for only 7.38 ± 2.7% of the total Epo clearance and that this clearance does not change under anemic conditions (Veng-Pedersen et al., 2004). Even though the total linear Epo clearance does not change under anemic conditions, the proportional contribution of the linear clearance to the total clearance does change. The linear clearance in the present study was shown to contribute an average of 26.8% when the EpoRs were the most down-regulated.

Although this study does not attempt to identify a mechanism for the linear Epo clearance, its significant contribution to the total clearance when the EpoRs are down-regulated merits further examination. From previous research, it has been well established that EpoRs are found in all tissues of the body and not just in the bone marrow (Yamaji et al., 1996; Juul et al., 1998, 2001; Carlini et al., 1999). It is possible that these receptors have different Epo binding characteristics in mediating a linear (nonsaturable) clearance of Epo (Veng-Pedersen et al., 2004). However, it is not known whether these additional receptors are responsible for the linear Epo clearance.

Lag Time in the Present Study.

In the present study, the lag time between receptor stimulation and the corresponding newly formed reticulocytes appearing in the circulation was calculated to be 1.13 ± 0.794 days. Other studies have looked at the lag time between Epo presence and reticulocyte production. One study found this lag time to be 1.7 days in adult humans (Krzyzanski and Perez-Ruixo, 2007), whereas another study calculated the lag time as 0.873 days in young sheep (Veng-Pedersen et al., 2002). These results seem to be similar to the results found in this study, although the variability in the lag time in this study was high. One publication has shown that under anemic conditions reticulocytes may be prematurely released into the bloodstream (Freise et al., 2007). Early release of reticulocytes will lead to earlier detection of new Hb in the bloodstream, which could contribute to study-to-study variability in lag time.

Potential Clinical Application.

There have been a number of clinical trials in which Epo was administered to preterm infants with the goal of reducing or eliminating the need for blood transfusions. A review study done on 27 different Epo clinical trials with preterm infants concluded that although Epo administration reduces the use of RBC transfusions and the quantity of RBCs transfused, these reductions were claimed to be of limited clinical importance (Ohlsson and Aher, 2006). However, because these published clinical trials used a “trial and error” methodology that does not consider the complex PK/PD of Epo, it is not clear how well they can be used as a guide to the real benefit of Epo when Epo is optimally administered to anemic infants.

The results from the current study show that the Epo receptor state plays an important role in Hb production. Furthermore, in the sheep studied, the endogenous Epo drops sharply at a certain point (Fig. 1, day 9). In a similar way, preterm infants' endogenous Epo concentrations drop sharply in the first few weeks of life (Freise et al., 2010). This similarity in the drop in Epo indicates a similarity in the Epo clearance/receptor pool dynamics in sheep and humans. Thus, knowledge of the efficacy of Epo relative to the Epo receptor pool dynamics gained from sheep experiments may provide a guide for how to optimally administer Epo in anemic infants. The hypothesis based on this similar kinetic mechanism is that a larger Epo receptor pool provides a greater potential for Hb production. Thus, Epo dosing should optimally be “in sync” with the variable status of the Epo receptor pool so that the peak Epo concentration coincides with the maximally unregulated quantity of Epo receptors.

In conclusion, a substantial increase in the efficacy of Epo is possible by a dosing optimization that considers the dynamic receptor regulation mechanism. This study was conducted under anemia induced by a major phlebotomy, but the basic approach may be extended to other anemic conditions for which similar large improvements in erythropoietic efficacy may be possible by considering the EpoR dynamics.

Acknowledgments

The recombinant human Epo used in the Epo RIA was a gift from Dr. H. Kinoshita of Chugai Pharmaceutical Company, Ltd. (Tokyo, Japan). The rabbit Epo antiserum used in the Epo radioimmunoassay was a gift from Dr. Gisela K. Clemens (University of California, Berkeley, Berkeley, CA). We appreciate all of the hard work by our laboratory team: Robert Schmidt, Demet Nalbant, Earl Gingerich, and Jessica Goehring.

This work was supported by the National Institutes of Health National Heart, Lung, and Blood Institute [Program Project Grant 2 P01-HL046925-11A1]; and the Children's Miracle Network Telethon of Iowa.

Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.

doi:10.1124/dmd.110.036855.

ABBREVIATIONS:
Epo
erythropoietin
EpoR
erythropoietin receptor
PK/PD
pharmacokinetic/pharmacodynamic
rhEpo
recombinant human erythropoietin
RBC
red blood cell.

Authorship Contributions

Participated in research design: Rosebraugh, Widness, and Veng-Pedersen.

Performed data analysis: Rosebraugh and Veng-Pedersen.

Wrote or contributed to the writing of the manuscript: Rosebraugh, Widness, and Veng-Pedersen.

References

  1. Al-Huniti NH, Widness JA, Schmidt RL, Veng-Pedersen P. (2004) Pharmacokinetic/pharmacodynamic analysis of paradoxal regulation of erythropoietin production in acute anemia. J Pharmacol Exp Ther 310:202–208 [DOI] [PubMed] [Google Scholar]
  2. Becker V, Schilling M, Bachmann J, Baumann U, Raue A, Maiwald T, Timmer J, Klingmüller U. (2010) Covering a broad dynamic range: information processing at the erythropoietin receptor. Science 328:1404–1408 [DOI] [PubMed] [Google Scholar]
  3. Beguin Y, Clemons GK, Pootrakul P, Fillet G. (1993) Quantitative assessment of erythropoiesis and functional classification of anemia based on measurements of serum transferrin receptor and erythropoietin. Blood 81:1067–1076 [PubMed] [Google Scholar]
  4. Blick M, Sherwin SA, Rosenblum M, Gutterman J. (1987) Phase I study of recombinant tumor necrosis factor in cancer patients. Cancer Res 47:2986–2989 [PubMed] [Google Scholar]
  5. Carlini RG, Alonzo EJ, Dominguez J, Blanca I, Weisinger JR, Rothstein M, Bellorin-Font E. (1999) Effect of recombinant human erythropoietin on endothelial cell apoptosis. Kidney Int 55:546–553 [DOI] [PubMed] [Google Scholar]
  6. Casadevall N. (1995) Cellular mechanism of resistance to erythropoietin. Nephrol Dial Transplant 10 (Suppl 6):27–30 [DOI] [PubMed] [Google Scholar]
  7. Cazzola M, Guarnone R, Cerani P, Centenara E, Rovati A, Beguin Y. (1998) Red blood cell precursor mass as an independent determinant of serum erythropoietin level. Blood 91:2139–2145 [PubMed] [Google Scholar]
  8. Chapel SH, Veng-Pedersen P, Schmidt RL, Widness JA. (2000) A pharmacodynamic analysis of erythropoietin-stimulated reticulocyte response in phlebotomized sheep. J Pharmacol Exp Ther 295:346–351 [PubMed] [Google Scholar]
  9. Chapel SH, Veng-Pedersen P, Schmidt RL, Widness JA. (2001) Receptor-based model accounts for phlebotomy-induced changes in erythropoietin pharmacokinetics. Exp Hematol 29:425–431 [DOI] [PubMed] [Google Scholar]
  10. Freise KJ, Widness JA, Schmidt RL, Veng-Pedersen P. (2007) Pharmacodynamic analysis of time-variant cellular disposition: reticulocyte disposition changes in phlebotomized sheep. J Pharmacokinet Pharmacodyn 34:519–547 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Freise KJ, Widness JA, Veng-Pedersen P. (2010) Erythropoietic response to endogenous erythropoietin in premature very low birth weight infants. J Pharmacol Exp Ther 332:229–237 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fukuda MN, Sasaki H, Lopez L, Fukuda M. (1989) Survival of recombinant erythropoietin in the circulation: the role of carbohydrates. Blood 73:84–89 [PubMed] [Google Scholar]
  13. Jelkmann W. (2002) The enigma of the metabolic fate of circulating erythropoietin (Epo) in view of the pharmacokinetics of the recombinant drugs rhEpo and NESP. Eur J Haematol 69:265–274 [DOI] [PubMed] [Google Scholar]
  14. Jensen JD, Madsen JK, Jensen LW, Pedersen EB. (1994) Reduced production, absorption, and elimination of erythropoietin in uremia compared with healthy volunteers. J Am Soc Nephrol 5:177–185 [DOI] [PubMed] [Google Scholar]
  15. Juul SE, Ledbetter DJ, Joyce AE, Dame C, Christensen RD, Zhao Y, DeMarco V. (2001) Erythropoietin acts as a trophic factor in neonatal rat intestine. Gut 49:182–189 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Juul SE, Yachnis AT, Christensen RD. (1998) Tissue distribution of erythropoietin and erythropoietin receptor in the developing human fetus. Early Hum Dev 52:235–249 [DOI] [PubMed] [Google Scholar]
  17. Kampf D, Eckardt KU, Fischer HC, Schmalisch C, Ehmer B, Schostak M. (1992) Pharmacokinetics of recombinant human erythropoietin in dialysis patients after single and multiple subcutaneous administrations. Nephron 61:393–398 [DOI] [PubMed] [Google Scholar]
  18. Kinoshita H, Ohishi N, Kato M, Tokura S, Okazaki A. (1992) Pharmacokinetics and distribution of recombinant erythropoietin in rats. Arzneimittelforschung 42:174–178 [PubMed] [Google Scholar]
  19. Krzyzanski W, Perez-Ruixo JJ. (2007) An assessment of recombinant human erythropoietin effect on reticulocyte production rate and lifespan distribution in healthy subjects. Pharm Res 24:758–772 [DOI] [PubMed] [Google Scholar]
  20. Levy G. (1994) Pharmacologic target-mediated drug disposition. Clin Pharmacol Ther 56:248–252 [DOI] [PubMed] [Google Scholar]
  21. Mager DE, Jusko WJ. (2001) General pharmacokinetic model for drugs exhibiting target-mediated drug disposition. J Pharmacokinet Pharmacodyn 28:507–532 [DOI] [PubMed] [Google Scholar]
  22. Mager DE, Woo S, Jusko WJ. (2009) Scaling pharmacodynamics from in vitro and preclinical animal studies to humans. Drug Metab Pharmacokinet 24:16–24 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Mock DM, Lankford GL, Burmeister LF, Strauss RG. (1997) Circulating red cell volume and red cell survival can be accurately determined in sheep using the [14C]cyanate label. Pediatr Res 41:916–921 [DOI] [PubMed] [Google Scholar]
  24. Nalbant D, Saleh M, Goldman FD, Widness JA, Veng-Pedersen P. (2010) Evidence of receptor-mediated elimination of erythropoietin by analysis of erythropoietin receptor mRNA expression in bone marrow and erythropoietin clearance during anemia. J Pharmacol Exp Ther 333:528–532 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Nelder JA, Mead R. (1965) A simplex-method for function minimization. Comput J 7:308–313 [Google Scholar]
  26. Nielsen OJ, Egfjord M, Hirth P. (1990) The metabolism of recombinant erythropoietin in the isolated perfused rat liver. Liver 10:343–349 [DOI] [PubMed] [Google Scholar]
  27. Ohls RK, Veerman MW, Christensen RD. (1996) Pharmacokinetics and effectiveness of recombinant erythropoietin administered to preterm infants by continuous infusion in total parenteral nutrition solution. J Pediatr 128:518–523 [DOI] [PubMed] [Google Scholar]
  28. Ohlsson A, Aher SM. (2006) Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants. Cochrane Database Syst Rev 3:CD004863. [DOI] [PubMed] [Google Scholar]
  29. Rizzo JD, Brouwers M, Hurley P, Seidenfeld J, Arcasoy MO, Spivak JL, Bennett CL, Bohlius J, Evanchuk D, Goode MJ, et al. (2010) American Society of Clinical Oncology/American Society of Hematology clinical practice guideline update on the use of epoetin and darbepoetin in adult patients with cancer. J Clin Oncol 28:4996–5010 [DOI] [PubMed] [Google Scholar]
  30. Salmonson T, Danielson BG, Grahnén A, Wikström B. (1990) Pharmacokinetics of intravenous recombinant human erythropoietin in patients with chronic renal failure. J Intern Med 228:53–57 [DOI] [PubMed] [Google Scholar]
  31. Sans T, Joven J, Vilella E, Masdeu G, Farrè M. (2000) Pharmacokinetics of several subcutaneous doses of erythropoietin: potential implications for blood transfusion. Clin Exp Pharmacol Physiol 27:179–184 [PubMed] [Google Scholar]
  32. Sawyer ST, Krantz SB, Goldwasser E. (1987) Binding and receptor-mediated endocytosis of erythropoietin in Friend virus-infected erythroid cells. J Biol Chem 262:5554–5562 [PubMed] [Google Scholar]
  33. Shepherd VL. (1989) Intracellular pathways and mechanisms of sorting in receptor-mediated endocytosis. Trends Pharmacol Sci 10:458–462 [DOI] [PubMed] [Google Scholar]
  34. Spivak JL, Hogans BB. (1989) The in vivo metabolism of recombinant human erythropoietin in the rat. Blood 73:90–99 [PubMed] [Google Scholar]
  35. Till AE, Gomez HJ, Hichens M, Bolognese JA, McNabb WR, Brooks BA, Noormohamed F, Lant AF. (1984) Pharmacokinetics of repeated single oral doses of enalapril maleate (MK-421) in normal volunteers. Biopharm Drug Dispos 5:273–280 [DOI] [PubMed] [Google Scholar]
  36. Pedersen PV. (1977) Curve fitting and modeling in pharmacokinetics and some practical experiences with NONLIN and a new program FUNFIT. J Pharmacokinet Biopharm 5:513–531 [DOI] [PubMed] [Google Scholar]
  37. Veng-Pedersen P, Chapel S, Al-Huniti NH, Schmidt RL, Sedars EM, Hohl RJ, Widness JA. (2004) Pharmacokinetic tracer kinetics analysis of changes in erythropoietin receptor population in phlebotomy-induced anemia and bone marrow ablation. Biopharm Drug Dispos 25:149–156 [DOI] [PubMed] [Google Scholar]
  38. Veng-Pedersen P, Chapel S, Schmidt RL, Al-Huniti NH, Cook RT, Widness JA. (2002) An integrated pharmacodynamic analysis of erythropoietin, reticulocyte, and hemoglobin responses in acute anemia. Pharm Res 19:1630–1635 [DOI] [PubMed] [Google Scholar]
  39. Veng-Pedersen P, Widness JA, Pereira LM, Schmidt RL, Lowe LS. (1999) A comparison of nonlinear pharmacokinetics of erythropoietin in sheep and humans. Biopharm Drug Dispos 20:217–223 [DOI] [PubMed] [Google Scholar]
  40. Widness JA, Schmidt RL, Veng-Pedersen P, Modi NB, Sawyer ST. (1992) A sensitive and specific erythropoietin immunoprecipitation assay: application to pharmacokinetic studies. J Lab Clin Med 119:285–294 [PubMed] [Google Scholar]
  41. Widness JA, Veng-Pedersen P, Peters C, Pereira LM, Schmidt RL, Lowe LS. (1996) Erythropoietin pharmacokinetics in premature infants: developmental, nonlinearity, and treatment effects. J Appl Physiol 80:140–148 [DOI] [PubMed] [Google Scholar]
  42. Woo S, Krzyzanski W, Jusko WJ. (2007) Target-mediated pharmacokinetic and pharmacodynamic model of recombinant human erythropoietin (rHuEPO). J Pharmacokinet Pharmacodyn 34:849–868 [DOI] [PubMed] [Google Scholar]
  43. Yamaji R, Okada T, Moriya M, Naito M, Tsuruo T, Miyatake K, Nakano Y. (1996) Brain capillary endothelial cells express two forms of erythropoietin receptor mRNA. Eur J Biochem 239:494–500 [DOI] [PubMed] [Google Scholar]

Articles from Drug Metabolism and Disposition are provided here courtesy of American Society for Pharmacology and Experimental Therapeutics

RESOURCES