Speed, Simplicity and Accuracy of Protein Binding Determination using BioSPME

M. James Ross, Senior R&D Scientist¹, Olga Shimelis, R&D Manager¹

1 Merck

• Introduction
• Experimental
  • Protein Binding Determination by BioSPME
  • Determination of %Free Fraction (FU) of Drug by BioSPME
  • Plasma Protein Binding Determination by Rapid Equilibrium Dialysis
• Results & Discussion
  • Accuracy of BioSPME Method for Protein Binding for Single Compound across Multiple Lots of Pin Tools
  • Precision across a Well Plate
  • Comparison of BioSPME and Rapid Equilibrium Dialysis Results for Multiple Compounds
• Conclusion
  • Merits of BioSPME Method over Rapid Equilibrium Dialysis
     

Introduction

An important aspect of drug discovery is understanding the interaction of the drug with plasma proteins and lipids. The portion bound to proteins and lipids is referred to as the plasma protein binding (PPB or F_b). The type of drug may also provide insight on the number of interactions with protein. In general, organic acids have a single binding site with albumin, whereas, organic bases will have multiple binding sites associated with glycoproteins.¹ In addition to albumin, other commonly associated proteins in drug binding are alpha-1-acid glycoprotein (AAG) and lipoproteins such as very high-density lipoprotein (VHDL) and low-density lipoprotein (LDL).² When it comes to the pharmacological effectiveness of a drug, it is the free fraction or unbound fraction (F_u) of a drug that is generally responsible for drug activity as described by the free drug hypothesis.^3,4,5

Protein binding properties of drugs, therefore, are important in understanding the amount of drug available in blood. Historical methods to measure protein binding include equilibration of the spiked plasma with the drug-free buffer across the membrane preventing the transfer of larger protein molecules into the buffer. This is an equilibrium dialysis method. Additional improvements in equilibrium dialysis products advertise reducing the equilibration time down from the traditional 24 hours to 4 hours using specifically designed devices.

In this work, Supel™ BioSPME C18 high-throughput devices were used to measure the protein binding and found to reduce the sample preparation time further to less than one hour for the extraction step. BioSPME devices are compatible with robotic liquid handling systems ensuring that the extraction can be fully automated for truly high-throughput methodology.⁶ The BioSPME, bioanalytical solid phase microextraction, method was applied to the selected compound, carbamazepine, using a high-throughput method for 88 plasma samples and 8 buffer samples. This method was evaluated across multiple lots of Supel™ BioSPME C18 96-pin tools and for multiple compounds and showed good accuracy and precision.
 

EXPERIMENTAL

Protein Binding Determination by BioSPME

Human plasma and buffer were spiked at a therapeutically relevant concentration and incubated for one hour at 37 ^oC while shaking at 300 rpm. After the incubation, 200 µL plasma and buffer were loaded into separate columns onto the extraction well plate (n = 8 to 88). The determination of the protein binding was performed using an automated robotic method and the BioSPME C18 96-pin tool.

Briefly described in Figure 1, the pin tool is conditioned statically for twenty minutes in isopropanol, then transferred into a new well plate for 10 seconds in water (wash step), followed by the extraction step. The pin tool is transferred into the preloaded extraction plate described earlier. Here, the pin tool extracts the analytes while shaking at 1200-1250 rpm at 37 ^oC for 15 minutes. The pin tool is returned to the water solution for a 60 seconds wash and finally transferred into a desorption plate. The desorption solution is 80:20 methanol:water and desorption is carried out for 20 minutes under static conditions. Samples were analyzed using methods described in Tables 1 and 2.

Condition
Solution: isopropanol	Time: 20 min	Condition: Static
⮟
Wash
Solution: water	Time: 10 sec	Condition: Static
⮟
Extraction
Solution: plasma	Time: 15 min	Condition: agitation
⮟
Wash
Solution: water	Time: 60 sec	Condition: static
⮟
Desorption
Solution: 80% methanol	Time: 20 min	Condition: static
	⮟
Analyze
Using LC-MS/MS

Figure 1. Schematic diagram for BioSPME analysis of free fraction of drug in human plasma.

Table 1. LC-MS/MS Conditions for monitoring analytes for free fraction determination

Column:	Ascentis® Express Biphenyl, 10 cm x 2.1 mm I.D., 2.7 µm (64065-U)
Mobile Phase:	[A] 5 mM Ammonium acetate, 0.1% acetic acid in 95% water and 5 % acetonitrile; [B] 5 mM Ammonium acetate, 0.1% acetic acid in 95% acetonitrile and 5% water
Gradient:	Initially start at 10% B and hold for 0.5 min, increase to 90% over 2.5 min, hold at 90% for 2 min, decrease to 10% in 0.1 min and hold for 2 min at 10%
Flow Rate:	0.4 mL/min
Column Temperature:	40 oC
Detector:	MS, ESI(+) scheduled MRM (See Table 1 and 2)
Injection:	5 µL

Table 2. Analyte description and LC-MS/MS parameters

Analyte	MW​	Log P​	pKa​	Charge​ at pH 7.4	Calibration Curve (ng/mL)	q1	q3	dwell (ms)	DP (V)	EP (V)	CE (V)	CXP (V)
Carbamazepine	236​	2.45	13.9​	0	0.5 – 100	237.1	194.2	40	21	10	29	26
Diazepam	284.7​	2.82	2.92​	0	1 – 100	285.0	193.2	40	91	10	43	28
Imipramine	280.4​	4.8	9.4  ​	1	0.5 -100	281.1	58.2	40	41	10	61	10
Propranolol	259​	3.48	9.42​	1	2.5 - 100	260.2	183.1	40	66	10	25	34
Prednisolone	360.4	1.6	12.59	0	5 - 250	361.1	147.1	50	36	10	33	10
Warfarin	308.3	5	6.33	-1	0.05 - 10	309.1	163	40	256	10	21	12
Zolpidem	307.4​	3.15	5.65​	0	0.5 – 100	308.2	235.2	40	36	10	49	16

Determination of %Free Fraction (F_U) of Drug by BioSPME

The BioSPME method determines the free concentration of analyte in plasma by comparing it with the extraction of the analyte from buffer samples where 100% of the analyte is considered to be free of protein binding.

The percent free or percent unbound is determined in Equation 1:

Equation 1:

 

where concentration free represents the unbound concentration of the analyte in the matrix, in this case plasma, and concentration total represents the total concentration of analyte. The amount extracted is independent of units and can be applied using preferred quantities (e.g. nanograms or moles) M_free, and extraction volume of plasma, V_plasma. The concentration of analyte in the desorption solution is quantified by an external calibration curve, and if the desorption volume is equal to the plasma and buffer extraction volumes, the concentration from desorption will be equal to the extracted concentration as shown in Equation 2:

Equation 2:
Equation 3:

The bound fraction, F_B, can be determined from the extracted concentrations as shown in Equation 6.

Equation 4:
Equation 5:
Equation 6:
Equation 7:

Figure 2. Representation of the extraction step (left) removing free analytes (yellow) from plasma (pink) and buffer (blue) and the analytes releasing into the desorption solution (right). The amount extracted does not greatly impact the concentration of free analyte which is termed non-depletive and does not disturb the equilibrium of the portion bound to plasma proteins (purple) (Equation 6 and Equation 7). As the buffer solution is considered 100% protein free, BioSPME will extract more from buffer than from the plasma.

In cases where depletion of compounds from plasma was determined by BioSPME extraction (extraction exceeded 5% of total spiked analyte), a correction to the calculated Bound Fraction was required as described below:

Equation 8:

where B and P, represent the respective amounts extracted from buffer, B, and plasma, P. B⁰ and P⁰ represents the concentration the samples were spiked or the total concentration in the buffer or biological sample. Equation 8 accounts for the concentration in solution after extraction on the tip; the depletion of the analyte from sample.⁷ Equation 6 and Equation 7, do not take this factor into consideration. However, they provide accurate values when the extracted amount is less than 5% the total amount.

Plasma Protein Binding Determination by Rapid Equilibrium Dialysis

Equilbrum dialysis was performed as directed by the accompying instruction sheet. Briefly, 200 µL of “spiked” human plasma at a therapeutically relevant concentration and 400 µL of phosphate buffered saline (PBS) were loaded in the corresponding chambers in at least triplicates. The dialysis proceeded for at least 4 hours while covered and shaking at 300 rpm and 37 ^oC on an Eppendorf shaker. At the end of dialysis, 50 µL of the spiked plasma was mixed with 50 µL of clean (unspiked) PBS and 50 µL of the dialysate (buffer compartment) was mixed with 50 µL of clean plasma. This was achieved to ensure matrix consistency. Next, 300 µL of ice-cold acetonitrile was added to each sample before centrifugation at 5,000 rpm for 10 minutes at 4 ^oC. Finally, the supernatant was transferred into glass vials for analysis by LC-MS/MS as described in Tables 1 and 2 using an AB Sciex 6500 with Agilent 1290 LC using a matrix-matched external calibration in the desorption solution.

Results and Discussion

Accuracy of BioSPME Method for Protein Binding for Single Compound across Multiple Lots of Pin Tools

To determine protein binding of carbamazepine, spiked buffer and plasma samples were deposited into 8 wells each of 3 lots of pins tools. In addition, six pin tools across three batches from each lot were tested. As can be seen in Figure 3, the same protein binding value was obtained for all lots with 1-2% CV. The average protein binding value across all 54 plates was 76.8 ± 2.2% which corresponded well to 70-80% range of protein binding found in literature for carbamazepine.

Figure 3. Protein binding values for carbamazepine obtained using 3 different lots of BioSPME pin tools. Each batch represents six pin tools. In green, represents the average across the 18 pin tools for each lot tested. Not shown is the overall average for the 54 pin tools (76.8 ± 2.2%). The yellow lines represent the boundaries of the literature range for protein binding (70 – 80%).

Precision across a Well Plate

In addition to showing consistency across multiple pin tools, an entire plate was tested in what is termed “8 + 88”. The 8 represents a column for buffer extraction, and the 88 represents the remaining 11 columns for plasma extraction. The results across the plate can be seen in Figure 4 (analyte extraction). For the extraction of the analyte, the average extracted amount across 88 plasma samples was 7.31 ± 0.4 ng/mL (with CV 4.8%). The average percent binding for the 88 samples was 75.9 ± 1.3%. As can be seen from Figure 5 there was no bias across rows or columns of the 96-well plate.

Figure 4. Concentration of analyte (ng/mL) extracted from plasma samples at each well position.

Figure 5. Percent bound with the standard deviation for each row and column. The yellow lines represent the boundaries of the literature range for protein binding (70 –80%).

COMPARISON OF BioSPME AND RAPID EQUILIBRIUM DIALYSIS RESULTS FOR MULTIPLE COMPOUNDS

Using the equations, Equation 5 and Equation 6, the values in Table 3 for analyte-protein bindings were determined from BioSPME extractions. These values are in good agreement with values determined using rapid equilibrium dialysis devices and the reported literature values as shown in Figure 6. 

Table 3. Binding values for the seven compounds from plasma using BioSPME and 200 µL sample volumes (n=8)

Analyte	Conc. Spiked (ng/mL)	Ave. FB (%)	CV FB (%)	Ave. Buffer Extracted (ng/mL)	Buffer CV (%)	Ave. Plasma Extracted (ng/mL)	Plasma CV (%)
Carbamazepine	100	76	3	35	7	8	9
Diazepam	100	97	0.4	71	1	2	15
Imipramine	100	93	0.6	65	6	12	6
Prednisolone	100	78	3	62	5	14	8
Propranolol	100	91	1.3	18	13	2	19
Warfarin	2500	99.8	<0.1	10	6	<0.1	22
Zolpidem	100	97	0.4	68	12	2	9

Figure 6. Comparison of protein binding values between equilbirum dialysis (yellow) and BioSPME (green) methods. The blue lines indicate the protein binding literature values interval. Compounds with stars are charged at physiological pH. Note: zolpidem has a single value reported in literature and not a range.

The BioSPME technique also provides a timesaving of over 50% as shown in Table 4. The longest step in the BioSPME process is the initial incubation of the analyte with the plasma (60 minutes). This is considerably shorter than the minimum four-hour incubation time required by equilbirum dialysis devices. With this BioSPME method, the total number of samples that can be processed is almost quadrupled compared to the rapid equilibrium dialysis method.

Table 4. Comparison of methods, number of samples and time commitment (* 88 corresponds to using the remaining 8 pins for buffer extraction).

Rapid Equilibrium Dialysis Method	Step Time (min)	BioSPME Method	Step Time (min)
Dialysis	240	Condition	20
Post sample preparation	5	Wash	0.2
Centrifugation	10	Extraction	15
Transfer for into vials for analysis	5	Wash	1
		Desorption	15
Total	260	Total	51.2
Max # of Samples processed	54		88*

CONCLUSION

Merits of BioSPME Method Over Rapid Equilibrium Dialysis

The Supel™ BioSPME C18 technique offered 50% timesaving and increased the number of samples for protein binding determination when compared with a commercially available equilibrium dialysis product. BioSPME allows the method to be fully automated with a standard robotic system. The protein binding values obtained compare well with those from the rapid equilibrium dialysis method as demonstrated with 7 compounds with log P values in the range 1.6 to 5.0. The precision of extraction using the pin tool was excellent across 88 samples with a CV on extracted concentration from plasma of 4.8%.

References

1. Cervelli MJ, Russ GR. 2010. Principles of Drug Therapy, Dosing, and Prescribing in Chronic Kidney Disease and Renal Replacement Therapy.871-893. https://doi.org/10.1016/b978-0-323-05876-6.00073-3

2. Scheife RT. 1989. Protein Binding: What Does it Mean?. DICP. 23(7-8):S27-S31. https://doi.org/10.1177/106002808902300706

3. Eichel E. Protein binding changes and drug interactions: What do we know? Current Psychiatry. 2018;38-4.

4. Smith DA, Di L, Kerns EH. 2010. The effect of plasma protein binding on in vivo efficacy: misconceptions in drug discovery. Nat Rev Drug Discov. 9(12):929-939. https://doi.org/10.1038/nrd3287

5. Wanat K. 2020. Biological barriers, and the influence of protein binding on the passage of drugs across them. Mol Biol Rep. 47(4):3221-3231. https://doi.org/10.1007/s11033-020-05361-2

6. Roy KS, Nazdrajić E, Shimelis OI, Ross MJ, Chen Y, Cramer H, Pawliszyn J. 2021. Optimizing a High-Throughput Solid-Phase Microextraction System to Determine the Plasma Protein Binding of Drugs in Human Plasma. Anal. Chem.. 93(32):11061-11065. https://doi.org/10.1021/acs.analchem.1c01986

7. Huq M, Tascon M, Nazdrajic E, Roszkowska A, Pawliszyn J. 2019. Measurement of Free Drug Concentration from Biological Tissue by Solid-Phase Microextraction: In Silico and Experimental Study. Anal. Chem.. 91(12):7719-7728. https://doi.org/10.1021/acs.analchem.9b00983