(M1030-01-03) High-Throughput Viscometry with Injectability Screening of High Concentration Protein Formulations by Multiplexed Microfluidic Quartz Resonators

Purpose: By 2026, there will be ~3,900 projects on protein-based therapeutics for chronic conditions in the drug discovery and preclinical development phase.¹ The preferred mode of administration of these therapeutics is subcutaneous or intramuscular injection, which enables home administration, short treatment times, and sustainable release. However, the requirement of small injection volumes necessitates high concentration protein formulations (HCFs), >100 mg/ml, presenting a significant challenge to manufacturability and injectability.² Due to scarcity of material at early stages, most viscometry is performed during formulation development after preclinical scale-up. Projects may exceed 500 formulations, incurring significant time and material costs in sample preparation and testing with conventional viscometry. Opportunity cost is also associated with lack of testing pre-scale-up, when such information could be used to select optimal drug candidates. HCFs are shear-thinning solutions which experience a broad range of shear rates during handling and delivery. Viscometry at a comparable shear rate range is necessary to predict manufacturability and injectability of HCF drug formulations. Existing techniques typically measure on the order of 20,000 s^-1,³ though shear-rates during injections can reach 2,000,000 s^-1. Direct Injection force measurement using a syringe/needle equipped with a load-cell requires ~1.2 ml HCF ( >120 mg protein) for a single test.⁴ Current “high-throughput systems” typically cycle through the instrument one-by-one. While reducing manual labor, they are slow and have exceedingly high material consumption. The purpose of this work was to develop a high-throughput microviscometer capable of broad shear-rate range with minimal sample volume, and to correlate these viscosities to injection force. The development of such a technique will enable viscosity and injection force characterization before preclinical scale-up and better capture shear behavior over the life of the drug. We hypothesize that a combination of microfluidics and quartz crystal microbalance (QCM) with dissipation monitoring will enable such measurements.

Methods: We developed nanovisQ™, a QCM-based microfluidic viscometer, which combines two orthogonal sensing techniques in a single measurement. Quartz resonances are used to measure viscosity at ultra-high shear ( >5,000,000 s^-1). These resonances also determine capillary flow position accurately with high time and position resolution. The position vs. time information is converted to viscosity at “low” shear-rates (100-70,000 s^-1), thereby acting as an extremely small capillary viscometer. The nanovisQ™ channel height is ~20x smaller than comparable capillary viscometers, enabling wider shear-rate ranges. The combined shear-rate range proffered by these complimentary techniques is particularly helpful when determining injectability of HCFs, which are often non-Newtonian. We developed single-test sensors and a multiplex array with four devices on a single crystal for high-throughput screening. Newtonian glycerol solutions and non-Newtonian protein formulations were created to characterize the instrument. Protein solution concentration was verified by nanodrop. Each solution’s viscosity was validated by rotational rheometry and with a commercially available capillary viscometer. Viscosities were measured using 3-5 µL of sample using the nanovisQ™ single-test and/or 4x1 multiplex sensor. Injection forces were measured using The Nano Plug and Play hydraulic testing system.

Results: Using the single-test nanovisQ™, high concentration human IgG was observed to have higher viscosity at low shear rates (1,000-50,000 s^-1) than low concentration, as expected.  Histidine formulations exhibited higher viscosity than PBS in this range.  However, at ultra-high shear rates (5,000,000 s^-1), all samples exhibited similar viscosity, regardless of buffer formulation or concentration.  This phenomenon would not be observable with comparable measurement techniques, especially with such small sample volumes.  These solutions, as well as BSA in identical buffers, were plotted against a calibration curve of Newtonian glycerol solutions to predict injection force.  These predictions were confirmed by measuring the injection force directly.⁵ We tested multiplex nanovisQ™ sensors with four glycerol solutions ranging from 3.2-24 cP. The time response of the nanovisQ™ sensors to the fluid is nearly identical, meaning that the low-shear-rate viscosity calculations (100-70,000 s^-1) are very similar, and nanovisQ™ arrays were able to deliver both their low and high shear-rate viscosity data within 10% of the known sample viscosity. We measured the viscosity of three BGG solutions in PBS at 90, 150, and 230 mg/ml in a multiplex array, using the fourth sensor to monitor crosstalk. The measured viscosities at low shear-rates agree with independent capillary viscometry measurements. These experiments consumed only 5 μl of each solution and three solutions were characterized simultaneously in 200 seconds, yielding viscosity information faster and over a wider range than possible with competing technologies.

Conclusion: In conclusion, we have created a technique and device for measuring viscosity over a broad shear-rate range with minimal sample volume. We have demonstrated that these measurements can be mapped to injection force. Taken together, these experiments demonstrate high-throughput testing of protein formulations is possible with nanovisQ™, which can be used to select optimal drug candidates and accelerate formulation development. We next aim to combine these 4x1 sensors into a cartridge containing 24 devices total, which could be run in as little as 30 minutes for low-viscosity samples.

References: 1. Lloyd, I., Shimmings, A., Scrip, P.S. Pharma r&d annual review 2022. Available at: pharmaintelligence informa com/resources/product-content/pharma-rd-annual-review-2022 2022 [cited 2022].
2. Zhang, Z., Liu, Y. Recent progresses of understanding the viscosity of concentrated protein solutions. Current Opinion in Chemical Engineering; 16:48-55 (2017).
3. Accurate Estimation of Injectability from High Shear Viscosity Measurement. In: Rheosense, editor. Rheosensecom. California
4. Zhang, Q., Fassihi, M.A., Fassihi, R. Delivery Considerations of Highly Viscous Polymeric Fluids Mimicking Concentrated Biopharmaceuticals: Assessment of Injectability via Measurement of Total Work Done “W T”. AAPS PharmSciTech:1-9 (2018).
5. French, D. Optimizing SubQ Delivery: NanovisQTM’s insight into viscosity and injection forces for biologics. Presented at the Annual Meeting of the American Chemical Society, New Orleans, LA; 20 March, 2024. Paper ID 3987220.

Acknowledgements: This work was supported by funding provided by NSF, NIST, and NIH. Travel support was provided by NC Biotech. The authors are affiliated with QATCH Technologies, which develops and commercializes the nanovisQTM.

Figure 1. Viscosity and Injection Force of High Concentration Formulations. A) Viscosity of human IgG in Histidine and PBS buffers. B) Injection force mapping of high concentration protein formulations to a curve standardized for a single syringe-needle combination.

Figure 2. Multiplexed Viscosity of Standard Newtonian Solutions. A) Raw trace of a single solution applied to four devices in a multiplex array of viscosity sensors. Similarity of the curves leads to consistent viscosities with little cross-talk. B) Viscosities at low and high shear of four standard solutions run simultaneously in about three minutes compared to the same samples run independently with a commercially available microrheometer.

Figure 3. Simultaneous Viscosity Measurements of High-Concentration Protein Formulations. A) Raw traces and B) processed viscosity measurements for bovine gamma globulin at three concentrations. Parentheses show comparable measurements with an alternative microrheometer.