Evaluation of Drug Efficacy of GLP-1 Receptor Agonists and DPP-4 Inhibitors Based on Target Molecular Binding Occupancy

Abstract

Glucagon-like peptide-1 (GLP-1) receptor agonists (liraglutide, exenatide, lixisenatide) have recently been used as anti-diabetes drugs. We examined relationships of the binding occupancy of GLP-1 receptors (Φ) and their clinical efficacy after administration of GLP-1 receptor agonists. Next, by focusing on changes of GLP-1 concentration after administration of dipeptidyl peptidase-4 (DPP-4) inhibitors (vildagliptin, alogliptin, sitagliptin, linagliptin), we analyzed the relationship between Φ and clinical efficacy. Furthermore, using Φ as a common parameter, we compared the clinical efficacy elicited by GLP-1 receptor agonists and DPP-4 inhibitors using a theoretical analysis method. The present results showed that GLP-1 receptor agonists produced their clinical effect at a relatively low level of Φ (1.1–10.7%) at a usual dose. Furthermore, it was suggested that the drugs might achieve their full effect at an extraordinarily low level of Φ. It was also revealed that the Φ value of DPP-4 inhibitors (0.83–1.3%) was at the lower end or lower than that of GLP-1 receptor agonists at a usual dose. Accordingly, the predicted value for hemoglobin A_1c (HbA_1c) reduction after administration of the GLP-1 receptor agonists was higher than that of DPP-4 inhibitors. We clarified the differences between the therapeutic effects associated with GLP-1 receptor agonists and DPP-4 inhibitors theoretically. Together, the present findings provide a useful methodology for proper usage of GLP-1 receptor agonists and DPP-4 inhibitors.

Glucagon-like peptide-1 (GLP-1) receptor agonists have recently been used as anti-diabetes drugs with a new action mechanism. GLP-1, a type of gastrointestinal hormone termed incretin, is secreted from the digestive tract into the bloodstream along with dietary intake and then enhances insulin secretion by binding to the GLP-1 receptor in pancreatic β cells. However, it has been shown that the insulin secretion-enhancing effect of GLP-1 is reduced in type 2 diabetic patients.^1,2) Administration of GLP-1 alone is difficult, as it becomes quickly degraded in vivo by dipeptidyl peptidase-4 (DPP-4). Consequently, GLP-1-like agents with resistance to DPP-4 have been developed as GLP-1 receptor agonists. In addition, DPP-4 inhibitors that selectively inhibit DPP-4 have also been developed, and found to retard degradation of GLP-1 and increase GLP-1 concentration in vivo. Both GLP-1 receptor agonists and DPP-4 inhibitors are drugs that exhibit their effectiveness via GLP-1 receptor mediation. A previous study reported that greater clinical efficacy might be expected with a GLP-1 receptor agonist as compared to a DPP-4 inhibitor.³⁾ However, neither appropriate dosages for these drugs nor their potency have been demonstrated theoretically. Our previous study showed that the efficacy profiles of the DPP-4 inhibitors could be demonstrated on the basis of increase in GLP-1 level.⁴⁾ In the present study, we examined that a comparison of the differences in clinical efficacy between GLP-1 receptor agonists and DPP-4 inhibitors on the basis of the binding occupancy of GLP-1 receptors allows for quantitative evaluation of clinical efficacy. Then, we compared the clinical efficacy elicited by GLP-1 receptor agonists and DPP-4 inhibitors using a theoretical analysis method.

MATERIALS AND METHODS

Relationships of the binding occupancy of GLP-1 receptors and their clinical efficacy after administration of GLP-1 receptor agonists was examined. Next, by focusing on changes of GLP-1 concentration after administration of DPP-4 inhibitors, the relationship between GLP-1 receptor occupancy and clinical efficacy was analyzed. Furthermore, using GLP-1 receptor occupancy as a common parameter, the potency of these drugs was compared in order to evaluate appropriate clinical dosage. Pharmacokinetic and pharmacodynamic data necessary for the present analyses were obtained from published studies, including those related to new drug applications.

Relationship between GLP-1 Receptor Occupancy and Clinical Efficacy after Administration of GLP-1 Receptor Agonists

For GLP-1 receptor agonists, in the present study we used liraglutide, exenatide, and lixisenatide, each of which has an established daily dosage regimen. Maximum GLP-1 receptor occupancy (Φ) was calculated by substituting the maximum plasma unbound drug concentration (C^f_max), which was obtained by multiplying the maximum plasma concentration at a steady state (C_max) after repeated administration of each drug, by the plasma unbound fraction (f_u) and dissociation constant of the GLP-1 receptor (K_i) using Eq. 1, as follows.   

(1)

The relationship between the Φ and clinical efficacy (E) values obtained after administration of the GLP-1 receptor agonists was examined. For clinical efficacy, hemoglobin A_1c (HbA_1c) reduction, which is the difference between the placebo group and agonist group after 12–14 weeks of pharmacotherapy, was used. We considered that the effect is produced in accordance with the E_max model after GLP-1 receptor agonist binding to the receptor, with E expressed by Eq. 2 using maximum clinical efficacy (E_max) and Φ exhibiting half of E_max (Φ_E50), as follows.   

(2)

By substituting the values for Φ and HbA_1c reduction after administration of each drug with a various dosage in Eq. 2, E_max and Φ_E50 were obtained by fitting with a nonlinear least squares method.

Relationship between GLP-1 Receptor Occupancy and Clinical Efficacy after Administration of DPP-4 Inhibitors

Vildagliptin, alogliptin, sitagliptin, and linagliptin, each of which has an established daily dosage regimen, were used as DPP-4 inhibitors. Data obtained in clinical trials regarding plasma GLP-1 concentration and HbA_1c reduction after repeated administrations of the subject drugs at a usual dose were obtained.^5–9) The value for Φ was calculated by substituting the GLP-1 concentration (C_G) and K_i of GLP-1 (K_i^G: 1500 pM) in Eq. 3.   

(3)

GLP-1 increase (ΔC_G) was calculated based on the DPP-4 inhibition rate (I) after repeated administrations of the drugs using Eq. 4, which we previously reported.⁴⁾   

(4)

where C_iG,max was the maximum level of inactive GLP-1 (20.4 pM), and K was the ratio of DPP-4 activity (0.041).⁴⁾

Then, C_G was obtained by adding the value to the GLP-1 basal concentration under fasted conditions (C_Gbase: 8.24 pM)^10–13) to calculate Φ with Eq. 3. HbA_1c reduction (E) was calculated by using ΔC_G in Eq. 5, which we previously reported.⁴⁾   

(5)

Finally, we examined whether the data for Φ and E after administration of the DPP-4 inhibitors were applicable to the relationship between Φ and E after administration of the GLP-1 receptor agonists noted above in section 1.

Comparison of Potency between GLP-1 Receptor Agonists and DPP-4 Inhibitors, and Dosage Evaluations

HbA_1c reduction was determined on the basis of C_max after administration of a GLP-1 receptor agonist at the usual dose using Eqs. 1 and 2. The relationship between dose and clinical efficacy was simulated. Moreover, on the basis of I after administration of a DPP-4 inhibitor at the usual dose,⁴⁾ HbA_1c reduction was determined using Eqs. 4 and 5 for simulating the relationship between dose and efficacy. Based on our findings regarding these relationships, findings for clinical potency of the examined GLP-1 receptor agonists and DPP-4 inhibitors were compared and evaluated. From those results, clinical efficacy in association with changes in the usual dose of the GLP-1 receptor agonists and DPP-4 inhibitors was predicted in order to determine appropriate dosage.

For all data analyses, we employed MLAB (Civilized Software Inc.) program.

RESULTS

Relationship between GLP-1 Receptor Occupancy and Clinical Efficacy after Administration of GLP-1 Receptor Agonists

Table 1 shows the pharmacokinetic and pharmacodynamic parameters used for calculating Φ for the GLP-1 receptor agonists,^5–9,14) as well as the calculated values of Φ after administration of each drug at the doses shown in the table and the values for HbA_1c reduction obtained after 12–14 weeks of treatment in clinical trials.^5–7)

Table 1. Pharmacokinetic and Pharmacodynamic Parameters, Φ, and HbA_1c Reduction of GLP-1 Receptor Agonists^4–8,14)

Drug	Dose/day	Cmax (pM)	fu	Ki (pM)	Φ (%)	HbA1c reduction (%)
Liraglutide	0.1 mg	2511	0.008	1500	1.3	0.81
	0.3 mg	7691			3.9	1.16
	0.6 mg	15269			7.5	1.59
	0.9 mga)	22414			10.7	1.76
Exenatide	5 µg	12.8	1	900	1.4	0.95
	10 µga)	28.9			3.1	1.34
	20 µga)	68.3			7.1	1.52
Lixisenatide	20 µga)	33.6	0.45	1330	1.1	0.73

a) Usual dose.

The Φ values after administration of the drugs ranged from approximately 1.1–10.7%, whereas those for HbA_1c reduction were about 0.73–1.76%. The relationship between Φ and HbA_1c reduction is shown in Fig. 1 together with a fitted line obtained with Eq. 2.

Fig. 1. Relationship between Φ and HbA_1c Reduction of GLP-1 Receptor Agonists

The E_max value (mean±S.D.) obtained was 2.01±0.03%, while Φ_E50 was 1.62±0.08%. (Symbols indicate actual values, solid line indicates fitted line, r²=0.91).

The relationship between Φ and HbA_1c reduction corresponded well with the fitted curve for all of the tested agonists (r²=0.91). The E_max value (mean±standard deviation (S.D.)) obtained was 2.01±0.03%, while Φ_E50 was 1.62±0.08%.

Relationship between GLP-1 Receptor Occupancy and Clinical Efficacy after Administration of DPP-4 Inhibitors

Of the tested DPP-4 inhibitors, clinical trial data for the GLP-1 concentration in plasma and HbA_1c reduction after repeated administrations at a usual dose could be obtained only for vildagliptin and alogliptin, with those values shown in Table 2.^10–13) As for the 4 drugs examined, ΔC_G calculated from the DPP-4 inhibition rate (I)⁴⁾ and Φ also calculated from that, as well as HbA_1c reduction are shown together in Table 2. Following administration of each drug at a usual dose, the Φ value ranged from approximately 0.83–1.3%, while HbA_1c reduction was about 0.82–1.1%. The values for both were at the lower limit or lower as compared to those for the GLP-1 receptor agonists.

Table 2. Pharmacokinetic and Pharmacodynamic Parameters, Φ, and HbA_1c Reduction of DPP-4 Inhibitors^10–13)

Drug	Dose/day (mg/d)	Iss (%)	CGbase (pM)	ΔCG (pM)	Φ (%)	HbA1c reduction (%)
Vildagliptin	100a)	97.5	8.24b)	12	1.3	1.0
					1.2c)	1.1c)
Alogliptin	25a)	90.4		5.3	0.89	0.89
					0.90c)	0.82c)
Sitagliptin	75a)	90.6		5.4	0.90	0.89
Linagliptin	5a)	87.9		4.4	0.83	0.85

a) Usual dose. b) Basal concentration of GLP-1 under fasted conditions. c) Actual value.

Figure 2 shows Φ and HbA_1c reduction values after DPP-4 inhibitor administration at the usual dose, and plotted with the fitted line from Fig. 1. The Φ and HbA_1c reduction values fit well with that line (r²=0.53).

Fig. 2. Relationship between Φ and HbA_1c Reduction of DPP-4 Inhibitors

Solid line indicates simulated line, open symbols indicate actual values, closed symbols indicate predicted values, r²=0.53.

Comparison of Potency of GLP-1 Receptor Agonists and DPP-4 Inhibitors, and Dose Evaluation

Figure 3 shows the relationship between dose and HbA_1c reduction following administration of the GLP-1 receptor agonists and DPP-4 inhibitors.

Fig. 3. Relationship between Dose and HbA_1c Reduction of GLP-1 Receptor Agonists (A) and DPP-4 Inhibitors (B)

Gray area indicate usual dose.

The predicted values for HbA_1c reduction obtained after administration of the GLP-1 receptor agonists at a usual dose ranged from 1.73–1.98%. (Fig. 3A, gray colored portion). Meanwhile, the predicted values for HbA_1c reduction obtained after administration of the DPP-4 inhibitors at a usual dose were from 0.77–1.0% (Fig. 3B, gray colored portion). These findings suggested that the HbA_1c reducing effect of the GLP-1 receptor agonists was higher than that of the DPP-4 inhibitors. Our results were theoretically supported by findings showing that the E_max value of the GLP-1 receptor agonists was calculated to be 2.01% with Eqs. 1 and 2, while that of the DPP-4 inhibitors was calculated to be 1.06% with Eqs. 3 and 4. Moreover, it was revealed that the usual dose for both the GLP-1 receptor agonists and DPP-4 inhibitors was at a level adequate for obtaining maximum effect.

DISCUSSION

When the pharmacological effect is mediated by the target molecule, it is important to quantitatively examine the relationship of the target drug near the site of action and its binding to that molecule. In our previous studies, we clarified that occupancy for the target molecule can be useful as a parameter for evaluating clinical efficacy.^15,16) As for GLP-1 receptor agonists, those examined in the present study exhibit a clinical effect after binding to the GLP-1 receptor, thus we considered that our analysis based on GLP-1 receptor occupancy rate would be applicable. DPP-4 inhibitors are considered to elicit their clinical effect by GLP-1, which is increased by selective DPP-4 inhibition. Accordingly, we considered that a quantitative comparison with the present GLP-1 receptor agonists could be made by use of a similar analysis method.

For calculating the GLP-1 receptor occupancy of the agonists, we speculated that the drug concentration near the receptor is equivalent to the plasma unbound drug concentration. K_i values, used for calculating Φ, were collected from the same trial results in order to prevent diffusion caused by differences in trial conditions. The K_i value for liraglutide was set to be comparable to the value for GLP-1, as it has a 97% homology with GLP-1 and showed an equal effect with GLP-1 in an in vitro trial, though no reported value was available.¹⁷⁾ Assuming that the number of GLP-1 receptors is not change, at over dose (C^f_max>>K_i), Φ becomes an upper limit (100%) theoretically from Eq. 1. Since the calculated Φ value of each drug at a usual dose ranged from 1.1–10.7%, we considered that GLP-1 receptor agonists elicit their clinical effect at an extraordinarily low level of occupancy.

For examining the relationship between GLP-1 receptor occupancy and clinical efficacy, we used HbA_1c reduction after 12 to 14 weeks of treatment as a parameter for clinical effect. At over dose, it is thought that HbA_1c reduction as well as Φ is high level. However, HbA_1c is not to be lower than zero. Therefore, Eq. 2 which has upper limit of HbA_1c reduction (E_max) is valid to explain the relationship between Φ and HbA_1c reduction. When collecting related clinical trial data, we confirmed that there was not a great difference in patient background findings. Our analysis showed that the relationship between Φ and HbA_1c reduction fit well with the theoretically constructed model (r²=0.91). Thus, we propose that Φ can be used as a common parameter for the clinical efficacy of GLP-1 receptor agonists (Fig. 1). Moreover, since the Φ_E50 value was 1.62±0.08%, it is suggested that the drugs might achieve their full effect at an extraordinarily low level of occupancy. It is reported that the average receptor occupancy rate of agonist at usual dose is about 1–2%.¹⁵⁾ Thus, in case of agonist, the low receptor occupancy rate is effective.

For the purpose of comparing the present GLP-1 receptor agonists with the DPP-4 inhibitors, Φ was calculated similarly for the DPP-4 inhibitors. The Φ value ranged from 0.83–1.3% when each drug was given at a usual dose, which was at the lower end or lower than that of the GLP-1 receptor agonists. The relationship between the Φ value of the DPP-4 inhibitors and HbA_1c reduction after DPP-4 inhibitor administration at a usual dose corresponded well with the fitted line in Fig. 1 (r²=0.53), suggesting that the Φ value for a GLP-1 receptor can be used to predict HbA_1c reduction for a DPP-4 inhibitor (Fig. 2).

In Fig. 3A, we showed the relationship between Dose and HbA_1c reduction of GLP-1 receptor agonists on the basis of the estimated E_max and Φ_E50 using Eqs. 1 and 2. On the other hand, in Fig. 3B, we showed the relationship of DPP-4 inhibitors on the basis of the estimated I after administration of a DPP-4 inhibitor at the usual dose using Eqs. 4 and 5. Therefore, Figs. 3A and B are the conceptual graphs of clinical efficacy of both drugs. The predicted value for HbA_1c reduction after administration of the GLP-1 receptor agonists at a usual dose ranged from 1.73–1.98% (Fig. 3A). Meanwhile, that value after administration of the DPP-4 inhibitors at a usual dose was from 0.77–1.0% (Fig. 3B). Accordingly, we concluded that the HbA_1c reducing effect was higher with the GLP-1 receptor agonists than with the DPP-4 inhibitors. In addition, the E_max value calculated by use of Eqs. 1 and 2 was 2.01% for the GLP-1 receptor agonists, and that calculated by Eqs. 3 and 4 was 1.06% for the DPP-4 inhibitors, thus it was theoretically suggested that DPP-4 inhibitors could not equal the effect of the GLP-1 receptor agonists. Moreover, for both the GLP-1 receptor agonists and DPP-4 inhibitors, the usual dose was set at a level for attaining maximum effect. Our results clarified that an intensification of the effect should not be expected even if the dose is increased over the usual dose, whereas the effect will become drastically lower when the dose is reduced.

GLP-1 enhanced satiety and reduced energy intake^18,19) and thus treatment with GLP-1 receptor agonists leads to weight loss.^3,20) It is thought that the appetite suppressing effect leads to HbA_1c reduction. Although we did not analyze the appetite suppressing effect in this study, it is thought that we evaluated HbA_1c reduction covering the appetite suppressing effect using Φ calculated from serum drug concentration.

The present results show that GLP-1 receptor agonists produce their clinical effect at a relatively low level of GLP-1 receptor occupancy (Φ), suggesting that Φ can be utilized as a common parameter for evaluating clinical efficacy irrespective of the kind of drugs being examined. At this point, when this study is applied to the individual patient, based on this study and data of the patient, it would be possible to individually determine the parameters and evaluate the clinical efficacy. Furthermore, they suggest that a comparison of the differences in clinical efficacy between GLP-1 receptor agonists and DPP-4 inhibitors on the basis of Φ allows for quantitative evaluation of clinical efficacy. It was also revealed that the examined GLP-1 receptor agonists have a higher HbA_1c reducing effect than the DPP-4 inhibitors when administered at a usual clinical dose. Together, the present findings provide a useful methodology for proper usage of GLP-1 receptor agonists and DPP-4 inhibitors.

Conflict of Interest

The authors declare no conflict of interest.

REFERENCES

• 1) Nauck M, Stockmann F, Ebert R, Creutzfeldt W. Reduced incretin effect in type 2 (non-insulin-dependent) diabetes. Diabetologia, 29, 46–52 (1986).
• 2) Toft-Nielsen MB, Damholt MB, Madsbad S, Hilsted LM, Hughes TE, Michelsen BK, Holst JJ. Determinants of the impaired secretion of glucagon-like peptide-1 in type 2 diabetic patients. J. Clin. Endocrinol. Metab., 86, 3717–3723 (2001).
• 3) Aroda VR, Henry RR, Han J, Huang W, DeYoung MB, Darsow T, Hoogwerf BJ. Efficacy of GLP-1 receptor agonists and DPP-4 inhibitors: meta-analysis and systematic review. Clin. Ther., 34, 1247–1258.e22 (2012).
• 4) Takayanagi R, Uchida T, Kimura K, Yamada Y. Evaluation of drug efficacy of DPP-4 inhibitors based on theoretical analysis with pharmacokinetics and pharmacodynamics. Biopharm. Drug Dispos., 38, 273–279 (2017).
• 5) “Pharmaceuticals and Medical Devices Agency. New drug application of VICTOZA.”: ‹http://www.pmda.go.jp/drugs/2010/P201000012/index.html›, cited 11 March, 2017.
• 6) “Pharmaceuticals and Medical Devices Agency. New drug application of BYETTA.”: ‹http://www.pmda.go.jp/drugs/2010/P201000057/index.html›, cited 11 March, 2017.
• 7) “Pharmaceuticals and Medical Devices Agency. New drug application of LYXUMIA.”: ‹http://www.pmda.go.jp/drugs/2013/P201300098/index.html›, cited 11 March, 2017.
• 8) “Pharmaceuticals and Medical Devices Agency. Interview form of VICTOZA (8th edition).”: ‹http://www.info.pmda.go.jp/go/interview/1/620023_2499410G1021_1_008_1F›, cited 11 March, 2017.
• 9) “Pharmaceuticals and Medical Devices Agency. Interview form of BYETTA (9th edition).”: ‹http://www.info.pmda.go.jp/go/interview/2/670227_2499411G1026_2_091_1F›, cited 11 March, 2017.
• 10) “Pharmaceuticals and Medical Devices Agency. New drug application of NESINA.”: ‹http://www.pmda.go.jp/drugs/2010/P201000029/index.html›, cited 11 March, 2017.
• 11) “Pharmaceuticals and Medical Devices Agency. New drug application of EQUA.”: ‹http://www.pmda.go.jp/drugs/2010/P201000008/index.html›, cited 11 March, 2017.
• 12) “Pharmaceuticals and Medical Devices Agency. New drug application of TRAZENTA.”: ‹http://www.pmda.go.jp/drugs/2011/P201100128/index.html›, cited 11 March, 2017.
• 13) “Pharmaceuticals and Medical Devices Agency. New drug application of JANUVIA.”: ‹http://www.pmda.go.jp/drugs/2009/P200900043/index.html›, cited 11 March, 2017.
• 14) “Pharmaceuticals and Medical Devices Agency. Interview form of LYXUMIA (6th edition).”: ‹http://www.info.pmda.go.jp/go/interview/1/780069_2499415G1024_1_006_1F›, cited 11 March, 2017.
• 15) Sawada Y, Yamada Y, Iga T. Quantitative evaluation of pharmacological effects and adverse effects based on receptor occupancy theory. Yakugaku Zasshi, 117, 65–90 (1997).
• 16) Yamada Y, Irizuki N, Takayanagi R, Yamamoto K, Sawada Y, Iga T. Pharmacokinetic/pharmacodynamic analysis of anti-hyperprolactinemic effect of terguride based on dopamine D2 receptor occupancy. Yakugaku Zasshi, 123, 255–260 (2003).
• 17) Gupta V. Glucagon-like peptide-1 analogues: an overview. Indian Journal of Endocrinology and Metabolism, 17, 413–421 (2013).
• 18) Flint A, Raben A, Astrup A, Holst JJ. Glucagon-like peptide 1 promotes satiety and suppresses energy intake in humans. J. Clin. Invest., 101, 515–520 (1998).
• 19) Gutzwiller JP, Drewe J, Goke B, Schmidt H, Rohrer B, Lareida J, Beglinger C. Glucagon-like peptide-1 promotes satiety and reduces food intake in patients with diabetes mellitus type 2. Am. J. Physiol., 276, R1541–R1544 (1999).
• 20) Vilsbøll T, Christensen M, Junker AE, Knop FK, Gluud LL. Effects of glucagon-like peptide-1 receptor agonists on weight loss: systematic review and meta-analyses of randomised controlled trials. BMJ, 344 (jan10 2), d7771 (2012).