Journal Club: Validation of a clot lysis assay for potency

Written by Anindya Ghosh Roy Posted in Method validation

Introduction and background

Recombinant human tissue Plasminogen Activator (rt-PA), the synthetic version of Tissue plasminogen activator (tPA), is a serine protease involved in degradation of blood clots. As a drug it is used for the treatment of stroke. It functions by converting plasminogen into plasmin, which in turn degrades insoluble fibrin into soluble by-products. This process of converting fibrin into soluble by-products is called fibrinolysis.

The activity of rt-PA is usually determined by calculating the rate of lysis of a synthetic fibrin clot. Multiple kinetic potency assays such as semi-automated or plate-based have already been developed. In case of being a potency assay for a pharmaceutical product to be released on the market by the quality control (QC) laboratory, it is mandatory to be validated before with complying with the ICH guideline Q2(R1).

The ICH Q2(R1) guideline requires for an assay test specificity, trueness, precision, linearity, and range and check its eligibility within predefined values. In this article, we will discuss the validation of an automated clot lysis activity assay using ACL TOP analyzer described by Lichun Huang with assessment of critical parameters as per the ICH Q2(R1) requirements. This automated method was also compared to the previously validated manual method already accepted by the authorities.

 

Test principle and analytical method

The test is based on the creation of a synthetic fibrin clot by mixing fibrinogen, plasminogen, thrombin and tPA, which afterwards is lysed due to the activity of tPA. Photometrically, this can be monitored: a previously turbid solution is getting clearer as time goes by. To measure potency, the time required to reach a predetermined level of turbidity reduction is used. Thus, the final lysis time is defined as the 50% threshold of the absorbance reading. The potency of the sample is calculated by comparison with a reference standard analyzed in parallel, using the parallel line analysis model described in USP <1034>. Both axis (lysis time and tPA concentration) have been logarithizmed before.

The ACL TOP instrument was used for its capability to be programmed and applying the pre-dilution function. Additionally, it can read at two different wavelengths. After analyzing the clot lysis at 405 nm and 671 nm, a better absorption was seen at 405 nm and hence was selected for the measurements. A standard curve was prepared using 5 different concentrations, while the samples, controls and blank are measured at the 3 middle dilution levels. Standards, controls and samples were mixed with thrombin and the reaction was started by adding a previously mixed solution of plasminogen and fibrinogen. For the suitability of the system (system suitability test, SST), the criteria defined for the manual assay were taken. They are defined as:

  • Correlation coefficient Rstandard between -0.995 and -1.0
  • Value of the control (a known sample): within the established mean ± 15%
  • Relative standard deviation RSDsample replicates and RSDcontrol replicates: ≤ 15%

These criteria were compared against the automated test results. As the automated method results showed greater precision compared to the manual assay, the range of the control (upper and lower control limit) could be tightened to mean ± 8%.

 

Analytical method validation and discussion

For analytical method validation, the trueness was determined by performing the assay 8 times by 3 different analysts (thus n = 24 in total), analyzing 5 different concentrations (70%, 85%, 100%, 115% and 130% of target potency). The mean recovery values obtained in all the cases were within 98% - 99%, which is in accordance with the predefined acceptance criterion of 95% - 105%.

Precision was evaluated in terms of repeatability and intermediate precision. The repeatability was performed using the assay control as sample and evaluating intra-assay RSD values. One operator performed the assay 3 times using each time 6 sample positions in the same instrument. The RSD values obtained were 1%, 1% and 2%, which is well below the acceptance criterion of 5%. For the intermediate precision (also called inter-day precision) analysis, the variation of the factors assay-to-assay, analyst-to-analyst, day-to-day, and instrument-to-instrument was evaluated by taking the data of the trueness experiments. All the RSD values obtained were below the permissible limit of 5%.

For linearity, the mean of the measured potencies (taking the data of the trueness experiments) were plotted against the expected potencies and the correlation coefficient was shown to be 0.997. Slope, y-intercept and residual sum of squares as well as the figure are presented. Additionally, for the standard linearity was evaluated using 5 concentrations (5 mg/mL, 3.3 mg/mL, 2.2 mg/mL, 1.1 mg/mL and 0.99 mg/mL) and correlation coefficients ranging from -0.9997 to -1.0 were obtained after logarithmic transformation of the lysis time in seconds and the concentration in ng/mL. Previous experiments already showed a broad linear range of 400 ng/mL to 10 µg/mL. Considering the results from trueness, precision and linearity, the range was determined as 4.06 – 7.54 x 105 IU/mg.

The specificity was first evaluated analyzing various commercial products (including enzymes, hormone and antibody products) in the assay and second, by understanding their effect on clot lysis activity in spiking experiments. The idea behind such an approach is to establish whether the clot lysis method is unaffected by any other reagents. For the first approach, no lysis was observed for all products, except for one closely related enzyme. For the second approach, potencies in the middle of the range (corresponding to about 100% target potency, although no recovery was mentioned) were observed, of course except for the closely related enzyme, where an additive effect was seen. Nevertheless, the specificity parameters were found to be within defined acceptance criteria.

Furthermore, the stability of the sample was evaluated applying several stress conditions (heat, light exposure, high and low pH and oxidation). The activity of the enzyme was found to be different when treated at 40°C for 4 and 7 days, with intense light and, subjecting to pH 4.0. These results show that the method is able to detect changes in the potency.

One interesting study performed was the comparison between the previously validated manual methods against the automated method. 44 samples including drug substance, drug product, correctly stored and stressed samples were compared. All results were comparable with an absolute mean difference of only 0.03 x 105 IU/mg. The automated method was however found to be more precise and thus able to detect any loss of potency (as seen in the stability samples) quicker.

 

In conclusion, the automated method developed by the ACL TOP is comparable to the manual method. For validation, a bracketing approach for trueness, linearity and precision was applied. All ICH Q2(R1) parameters for a potency assay were successfully evaluated with very good results. One single typing error in the paper was observed (in the method section for linearity evaluation a coefficient of determination R2 is mentioned, while in all other sections a correlation coefficient R is cited), which is of minor importance. This method was later successfully transferred into QC routine analysis as it showed better precision and robustness when compared to the manual assay. Additionally, the mixing and dilution steps of this assay are critical, thus an automatic system can generally guarantee more robust results.