April 1, 2015 (Vol. 35, No. 7)

Assessing Injectability during the Critical Phase of Bioprocess Development Activities

Pharmaceutical companies have increasingly focused on developing high-concentration protein therapeutics that enable patients to treat themselves at home by self injection. However, the formulation of therapeutic proteins has many challenges.

Formulations with high protein concentrations have inherent viscosity problems.  Such formulation require high injection forces, which hinder injectability. A therapeutically promising candidate drug might be discarded because of high viscosity or require development of an alternative delivery method.

In addition, major efforts are geared toward elegant molecular synthesis of drug candidates with high efficacy, and it is known that minute changes in structure can also have marked effects on viscosity.

For example, variation in the small, but important complementarity determining regions (CDRs) of therapeutic antibodies can affect the viscosity of antibody formulations, causing striking concentration-dependent effects on viscosity. Figure 1 clearly shows that the viscosity behavior of monoclonal antibodies (mAbs) with small variations can differ dramatically with regard to concentration.

If the viscosity of a candidate therapeutic protein is greater than 30 to 50 mPa·s (cP) at its dose concentration, more often than not it would be discarded from the development pipeline. Therefore, it is essential to measure viscosity during the early stages of drug development. But this introduces an additional problem.

During those early stages usually only small amounts of candidate drugs are available; therefore, the instruments that measure viscosity must accurately measure small sample volumes. RheoSense manufactures a rectangular slit viscometer, a VROC® (viscometer/rheometer-on-a-chip), that measures viscosity from the pressure drop of a test liquid flowing through a glass rectangular slit at a controlled shear rate.

The method is a well-known scientific application, and the VROC technology is now a public standard in the U.S. Pharmacopeia. The advantage of the VROC design is that it can accurately measure samples as small as 50 microliters. 

Figure 1. The viscosity of mAb-h rapidly increases with rising concentration, and mAb-h would thus be the least favorable candidate as an injectable. (Modified from Yadav et al., 2010, J Pharm Sci 99:4812-29.)

Air / Solution Interface

While viscosities of protein solutions have been measured with various viscometers or rheometers, another problem inherent to producing accurate viscosity measurements of protein solutions has been recently clarified by a series of scientific reports.

Elaborate experiments have shown that traditional viscometers and rheometers, which require an air/solution interface, produce erroneously high-viscosity measurements. Proteins are amphiphilic molecules that in aqueous solutions tend to move to the air/liquid interface to minimize potential energy, as shown in Figure 2A. The accumulated protein molecules at the interface form a viscoelastic film, which affects the apparent viscosity of the solution.

Figure 2B shows that viscosity measured with a cone and plate rheometer (schematic shown in Figure 2A) erroneously produces high viscosity measurements of samples subjected to low shear rates, because it assesses the interfacial layer of protein and not the protein solution itself. In addition, the viscosity values determined at low shear rates erroneously appear to be shear-rate dependent.

However, there are a few viscometers that do not have an air/liquid interface, including glass capillary viscometers and slit viscometers. Figure 2B also shows the viscosities of the same samples measured with the m-VROC closed-system slit viscometer (schematic shown in Figure 2C).

Figure 2. (A) Schematic of cone and plate rheometer with a protein solution exposed to the air and protein molecules accumulating at the air/liquid interface. (B) Varying concentrations of protein solutions measured with a cone and plate rheometer show erroneously high viscosities (? [Pa·s]) that are shear-rate ( [s-1]) dependent at low shear rates, whereas the same concentrations measured with a VROC viscometer show shear-rate independent viscosities. (C) Schematic of the flow channel of a slit viscometer showing that the sample is not exposed to the air, where Q is the flow rate and P1 and P2 are local pressures measured by corresponding pressure sensors. (2A and 2B modified from Sharma et al., 2011, Soft Matter 7, 5150-5160.)

Shear Rate

Thus far we have described the problems measuring the viscosity of proteins that are candidates for injectable formulations. However, the most important variable for accurate estimation of injectability is the viscosity of the protein solution measured at the shear rate imposed inside the needle during injection.

The shear rate (ġa) and pressure drop (DP) inside the needle and syringe barrel can be estimated accurately using respectively, Eq. 1 and the Hagen-Poiseuille equation (Eq. 2) where Q is the injection rate, R the radius, l the length, and η the viscosity. The Hagen-Poiseuille equation describes flow through circular tubes such as a the needle and syringe barrel, as shown in Figure 3.

Figure 3. Schematic of the prefilled syringe. F is the force required to push the piston; Fv is the resistance force caused by the pressure P1 acting on the piston.

The injection force (F) is equal to the pressure force applied to the piston (Fv) and the frictional force between the barrel and the piston. The pressure caused by the inherent resistance of the liquid to flow, which is the viscosity, is the cumulative sum of the pressure drops in the barrel along the contraction zone from the barrel to the needle, and the needle.

The ratio of the pressure drop in the needle to that in the barrel is estimated using Eq. 2 to be 300,000 to 1 for a typical prefilled syringe, because of the marked differences between the radius of the needle and radius of the barrel. The high ratio suggests that the pressure drop in the needle is the major contributor to the resistance to injection of a highly viscous fluid. Thus, it is essential to assess the dynamics of flow in the needle.

To understand the dynamics in the needle, the shear rate in the needle for typical injection rates must be estimated using Eq. 1. For both 26- and 27-gauge needles, with injection rates ranging from 0.06 mL/s to 0.2 mL/s, the shear rate ranges from 50,000s-1 to 200,000s-1. To estimate the pressure drop accurately, it is critical to measure the viscosity of the solution at the shear rate in the needle, because the viscosity at low shear rates may not be relevant.

While many protein solutions show Newtonian behavior, recent work has indicated that there may be non-Newtonian concentrated protein solutions with shear-rate-dependent viscosities. These formulations will have viscosities at high shear rates that differ from their viscosities at the low shear rates accessible by most viscometers.

The m-VROC viscometer, however, allows viscosity measurements at high shear rates without incurring the flow instabilities faced by other instruments, since the small depth of the flow channel ensures a stable and laminar flow. The viscosity of a non-Newtonian formulation that may decrease with high shear rates that are greater than 1,000,000s-1 can be measured using an m-VROC viscometer. After the viscosity at the shear rate in the needle is measured, the injection force can be estimated using the following equation:

(where b denotes the barrel and n denotes the needle. For most injection applications using a small needle, the contribution from the piston friction is negligible.)

Therapeutic protein formulations that are candidates for injectable drugs have increased viscosity, which impedes injectability. Only small samples may be available for assessment early in the development pipeline. The protein candidates can exhibit non-Newtonian behavior and form viscoelastic films at fluid/air interfaces. The m-VROC closed-system slit viscometer can accommodate these problematic features and provide viscosity measurements at high shear rates imposed in the needle.

Seonggi Baek, Ph.D. (SBaek@RheoSense.com), is president and CEO of RheoSense.