White Light Spectroscopy Characteristics and Expansion Dynamic Behavior of Primary T-Cells: A Possibility of Online, Real-Time, and Sampling-Less CAR T-Cell Production Monitoring

Abstract

The production of advanced therapy medicinal products (ATMP) is a long and highly technical process, resulting in a high cost per dose, which reduces the number of eligible patients. There is a critical need for a closed and sample-free monitoring system to perform the numerous quality controls required. Current monitoring methods are not optimal, mainly because they require the system to be opened up for sampling and result in material losses. White light spectroscopy has emerged as a technique for sample-free control compatible with closed systems. We have recently proposed its use to monitor cultures of CEM-C1 cell lines. In this paper, we apply this method to T-cells isolated from healthy donor blood samples. The main differences between cell lines and human primary T-cells lie in the slightly different shape of their absorption spectra and in the dynamics of cell expansion. T-cells do not multiply exponentially, resulting in a non-constant generation time. Cell expansion is described by a power-law model, which allows for the definition of instantaneous generation times. A correlation between the linear asymptotic behavior of these generation times and the initial cell concentration leads to the hypothesis that this could be an early predictive marker of the final culture concentration. To the best of our knowledge, this is the first time that such concepts have been proposed.

Keywords: CAR T-cell; advanced therapy medicinal product; cell concentration monitoring; quality control; white light spectroscopy.

Figure 1

General experiment protocol.

Figure 2

Absorption spectra and corresponding concentrations of dilution ranges of T-cells. (a) Absorption spectra (each color represents one concentration). (b) Corresponding concentrations measured with a Malassez cell. n = 73.

Figure 3

Comparison between T-cell absorption function and experimental dilution ranges spectra. (a) Function vs. experimental data. (b) Difference between absorption function and spectra.

Figure 4

Measuring concentration in dilution ranges using the absorption function. (a) Spectral vs. automated counting. (b) R² values of spectra fittings with absorption function.

Figure 5

Examples of T-cell culture with concentration measurements performed with a Malassez cell, an automated counter and a spectrometer. (a) Donation #1. (b) Donation #6.

Figure 6

Examples of expansion folds obtained for each concentration measurement method. (a) Donation #1. (b) Donation #6.

Figure 7

Examples of expansion folds fitted with either an exponential or a power function with corresponding R² values. (a) Donation #1. (b) Donation #6. (a) and (b) are not drawn to the same scale. Black stars: concentrations spectrally measured, blue curves: exponential fitting and red curves: power fitting.

Figure 8

Expansion folds calculated using Equation (6) and parameters in Table 6.

Figure 9

Instantaneous Generation Times (IGTs) calculated using Equation (7) for all experiments. Vertical scale in hours.

Figure 10

Calculating primary T-cell concentration using the CEM-C1 spectral shape function.

Figure 11

Difference between T-cell and CEM-C1 spectra functions.