Decoding the Key Role of Plasma Protein Binding in Drug-Drug Interactions

In modern drug development, accurately predicting a drug's pharmacokinetic properties is essential for ensuring its safety and efficacy. Plasma protein binding (PPB) is a key factor that influences these properties, particularly in the assessment of drug-drug interactions (DDIs). This article examines the importance of PPB in DDI prediction, discusses the challenges of measuring PPB for highly bound drugs, and highlights how updated regulatory guidelines support more precise DDI risk assessments.

What Is Plasma Protein Binding and Why Does It Matter to Drug-Drug Interactions?

Plasma protein binding refers to the reversible interaction between drug molecules and proteins in the blood, primarily albumin and α1-acid glycoprotein. The unbound (free) fraction of a drug (f_u) is the portion not bound to these proteins. Only free drugs can cross cell membranes, interact with therapeutic targets, and undergo metabolism or excretion.

Since PPB affects a drug’s distribution and clearance, especially hepatic clearance, accurately measuring f_u is critical for predicting drug-drug interactions involving metabolizing enzymes (e.g., CYP450) and transporters (e.g., OATP1B1). Historically, conservative defaults (e.g., assuming f_u = 0.01 for highly bound drugs) were used in DDI assessments, but recent guidelines like ICH M12 now emphasize experimentally determined f_u values for more accurate predictions.

Methodological Challenges and Strategies in Measuring Plasma Protein Binding (PPB)

Plasma protein binding is a significant area of research in drug discovery and development. The principle is that drug molecules can reversibly bind to proteins and lipids in plasma and tissues. Unbound drug molecules can freely diffuse across cell membranes and interact with therapeutic targets or other biomolecules (Figure 1). While a high degree of compound binding does not necessarily indicate good or bad outcomes, the fraction unbound (f_u) measured in PPB experiments is a crucial parameter that needs accurate determination for predicting drug interactions^[2]. For highly bound compounds, accurately measuring f_u is critical. WuXi AppTec DMPK has established various methods for determining highly bound compounds, such as pre-saturation, ultracentrifugation, dilution method, and flux dialysis (see the article: Plasma Protein Binding (PPB) Measurement of High Protein-binding Drugs: Principles and Advantages of the Flux Dialysis Methods), which can be combined to obtain accurate f_u values.

Figure 1. Free drug hypothesis ^[3]

Accuracy Assessment of High Plasma Protein Binding Compounds Measurement

Researchers selected two typical high plasma protein binding drugs—warfarin (anticoagulant) and itraconazole (antifungal) as model compounds. Using a multi-laboratory collaborative research system, they assessed the accuracy of different methods for measuring high-binding drug PPB. Researchers from nine laboratories measured warfarin's f_u using various methods, including equilibrium dialysis (HTD and RED), ultracentrifugation, and ultrafiltration under different experimental conditions. The data showed that the average f_u of warfarin in human plasma was 0.011 (range: 0.005-0.017), closely matching the literature value of 0.014. Meanwhile, researchers from eleven laboratories optimized methods to measure itraconazole's average f_u at 0.0015 (range: 0.0007-0.0022), also consistent with the literature value of 0.0020^[4]. These results confirm that optimized experimental conditions can yield f_u values closely matching literature values, providing a reliable methodological basis for drug interaction studies of highly bound drugs.

Experimental Strategy Optimization for Challenging Compounds

There are multiple methods for measuring the fraction unbound of compounds in biological matrices, such as equilibrium dialysis, ultracentrifugation, and ultrafiltration. Equilibrium dialysis is the most commonly used method, but its reliability is limited by factors such as whether the system reaches full equilibrium, the solubility properties of the test compound, and whether the concentration of free samples in the buffer meets analytical sensitivity requirements. For some challenging compounds, appropriate methods should be chosen based on their properties. For example, highly bound lipophilic acids, poorly water-soluble lipophilic compounds, high molecular weight, and slow-diffusing compounds can use pre-saturation, dilution method, and flux dialysis strategies. However, for certain compounds like covalent inhibitors, ultrafiltration or ultracentrifugation may be more suitable. Although precisely measuring compound plasma protein binding is challenging, optimized methods can accurately measure f_u < 0.01 for challenging compounds (See Figure 2).

Figure 2. Decision tree for measuring highly bound compounds^[2]

Application of Plasma Protein Binding (PPB) in Predicting Drug-Drug Interactions

Accurate measurement of PPB is crucial for predicting human pharmacokinetics (PK), drug interaction, toxicity index (TI) estimation, and the development of PK/PD relationships. Systematic analyses show that PPB significantly impacts understanding the key characteristics of candidate drugs, especially in DDIs, as it provides insights into clinical DDI risk and study design. For highly protein-bound drugs (f_u < 1%), the ICH M12 guidelines allow the use of experimentally measured f_u values for prediction, significantly improving prediction accuracy.

General Principles of Clinical DDI Risk Assessment

Regulatory guidelines provide a systematic framework for assessing DDI potential, recommending a progressive prediction strategy from simple static models to complex mechanistic models to assess potential clinical impacts more accurately. When extrapolating in vitro inhibition/induction data to in vivo DDI predictions, plasma protein binding plays a crucial role in estimating clinical DDI effects. Historically, regulatory agencies have commonly recommended using 0.01 as the lower limit of f_u in DDI predictions, but this conservative setting can overestimate the clinical drug interaction risk of highly bound drugs ^[5]. Four case studies compare the impact of using the 0.01 f_u lower limit versus actual measured f_u values on DDI prediction magnitude, further verifying this viewpoint. 

Case 1:

Fahmi et al. (2009) ^[7] demonstrated that itraconazole (with the unbound > 99.7%, measured f_u = 0.001-0.003), as a strong CYP3A4 inhibitor, has DDI predictions highly dependent on accurate estimation of free drug concentration. Using actual measured f_u values (0.001-0.003) for prediction, the interaction between itraconazole and midazolam (CYP3A4 substrate) predicted an AUC increase of 5-10 times, closely matching clinical observations. However, using the previous regulatory default 0.01 f_u lower limit, the predicted AUC increase was exaggerated to 10-30 times, significantly deviating from the clinical reality. This comparison confirms that mechanically using the 0.01 f_u lower limit for extremely high protein-binding drugs (f_u < 0.01) introduces conservative bias, potentially causing overestimation of clinical DDI risk. The study recommends using measured f_u values or PBPK modeling for accurate prediction of such special drugs. 

Table 1. Itraconazole (measured f_u) predicted drug interaction results

Case 2:

In a case of an oncology drug causing drug interactions by inhibiting renal organic anion transporters OAT1/OAT3 ^[4], using a measured f_u value of 0.008 and a static model to assess DDI risk, the free drug concentration did not reach the EMA guideline-recommended K_i/50 threshold, thus not predicting clinically meaningful interactions. However, using the 0.01 f_u lower limit combined with a 50-fold safety threshold suggested further clinical DDI risk evaluation. Due to practical difficulties in conducting clinical drug interaction studies for oncology drugs, the drug label included a warning about potential interactions with OAT1/OAT3 substrates, possibly preventing some cancer patients who could have benefited from using the drug. Evaluating DDI risk for highly bound compounds with the 0.01 f_u lower limit (instead of a lower measured value) leads to more conservative drug label requirements.

Case 3:

As shown in Table 2 ^[4], an early-stage discovery compound was evaluated for DDI risk due to inhibition of hepatic OATP1B1 transport protein. Using the formula of the 2012 FDA drug interaction guideline to calculate the R value, researchers evaluated clinical DDI risk using both the measured f_u value and the 0.01 lower limit. Using the 0.01 f_u lower limit predicted significant clinical interactions (AUCR=6.1, assuming FT, OATP=1, meaning the drug completely depends on OATP1B1 to enter hepatocytes), but this result significantly overestimated the actual clinical observation (AUCR=1.8). Using the measured f_u value for prediction provided a more accurate result (AUCR=2.0). This case further emphasizes that using the 0.01 f_u lower limit is overly conservative and may overestimate clinical DDI risk.

I_in,max=C_max+(K_a X Dose X F_aF_g) / Q_h

R-value=1+(f_uX I_in,max/IC₅₀)

Note: R is the AUC ratio of the substrate drug; f_u is the fraction unbound of the inhibitor in blood; I_{in, max} is the estimated maximum concentration of the inhibitor in the portal vein (assuming k_a = 0.1 min^-1 and F_aF_g = 1); IC₅₀ is the half-maximal inhibitory concentration of the inhibitor; C_max is the maximum concentration of the inhibitor in plasma; k_a is the absorption rate constant (assumed to be 0.1 min^-1, a common assumption); Dose is the administered dose of the inhibitor; F_aF_g is the absorbed fraction of the inhibitor dose (assumed to be 1, meaning complete absorption with no intestinal metabolism); Q_h is the hepatic blood flow.

Table 2. Comparison of predicted drug interaction results using measured f_u and 0.01 f_u lower limit

Case 4:

Montelukast is a highly bound drug, and the fraction unbound (f_u) value is 0.000051. In the past, regulatory agencies conservatively set the f_u value for all highly bound drugs at 1% for DDI predictions. For montelukast, this setting would lead to a CYP2C8 inhibition AUC ratio (AUCR) of 1.89, implying a potential clinically relevant drug interaction. However, in clinical oral administration practices, montelukast did not show significant DDIs with rosiglitazone (CYP2C8 substrate) and repaglinide (CYP2C8 substrate). This indicates that using the 1% f_u,p lower limit might overestimate the CYP2C8 DDI risk of montelukast. Conversely, using the experimentally obtained actual f_u value (0.000051) for prediction showed no DDI occurrence, consistent with clinical observations^[6]. This example highlights the importance of experimental f_u measurements in accurately predicting DDIs.

Special Considerations for Displacement Interactions

Plasma protein binding displacement can directly affect the safety and efficacy of drug therapy. When multiple drugs are used simultaneously, they may compete for the same plasma protein binding sites, altering free drug concentrations, affecting efficacy, or increasing the risk of adverse reactions. Accurate plasma protein binding measurement is crucial for predicting drug interactions, guiding clinical medication decisions, and drug development. When co-administering drugs that bind the same plasma protein, understanding their PPB characteristics can help predict potential interactions, allowing for dose adjustments or alternative drug choices. Although PPB displacement usually has limited clinical impact, it can cause significant consequences for specific drugs (e.g., f_up < 0.1, narrow therapeutic window, high hepatic extraction rate, or intravenous administration). For example, warfarin with phenylbutazone causes coagulation abnormalities, and sulfonamides with tolbutamide cause severe hypoglycemia and are believed to be related to plasma protein displacement ^[2,5].

Figure 3. Decision tree for assessing the clinical significance of plasma protein binding displacement interactions ^[2]

Case Summary

Accurately measuring plasma protein binding rates is a core aspect of drug development, directly impacting the evaluation of drug interactions. Studies show that under strictly optimized experimental conditions, current plasma protein binding measurement methods can stably measure the fraction unbound as low as < 0.01. However, the suitability and limitations of each method must be considered comprehensively. Historically, regulatory agencies set the f_u lower limit at 1% for predicting highly bound compound DDIs, leading to high false positives. The ICH M12 drug interaction guidelines, published in May 2024, allow using experimental f_u values to predict highly bound compound DDIs. Will using measured f_u values lead to false negatives? To further enhance confidence in using experimental f_u values for predicting highly bound compound DDIs^[6], research institutions evaluated nine highly bound drugs (f_u < 1%). They found that using measured f_u values for prediction with mechanistic models successfully identified all clinically relevant DDIs (effect size > 20%) (no false negatives). Only almorexan's inhibition of CYP2D6 was not flagged with basic models, but even with the 1% lower limit, the risk was not identified, indicating the issue was unrelated to f_u measurement and possibly involved other mechanisms like metabolite inhibition. The study confirmed that using measured f_u values not only avoids false negatives but also significantly reduces false positive predictions caused by the traditional 1% lower limit. These findings validate the foresight of the ICH M12 guidelines and provide methodological support for improving DDI prediction for highly bound drugs.

The Bottom Line

Plasma protein binding plays a crucial role in drug development. It affects drug distribution and excretion and relates to predicting drug interactions. Accurate plasma protein binding measurement is central to optimizing DDI predictions, improving drug development efficiency, reducing unnecessary clinical trials, lowering patient risk, and ultimately enabling the development of more efficient and safer drug therapies. WuXi AppTec DMPK offers various methods for accurately measuring compound PPB values. In compliance with the ICH M12 guideline, we provide clients with reliable solutions for DDI risk assessment of highly bound drugs, delivering more dependable references for estimating in vivo parameters and predicting DDIs—ultimately helping to mitigate development risks.

Authors: Chunhong Lu, Jie Wang, Xiangling Wang, Genfu Chen

Talk to a WuXi AppTec expert today to get the support you need to achieve your drug development goals.

Committed to accelerating drug discovery and development, we offer a full range of discovery screening, preclinical development, clinical drug metabolism, and pharmacokinetic (DMPK) platforms and services. With research facilities in the United States (New Jersey) and China (Shanghai, Suzhou, Nanjing, and Nantong), 1,000+ scientists, and over fifteen years of experience in Investigational New Drug (IND) application, our DMPK team at WuXi AppTec are serving 1,600+ global clients, and have successfully supported 1,700+ IND applications.