The limitation of all these technologies comes from their capacity to specifically

quantify viral particles. Although most of them can handle non-purified material at

the first stages of the production process. None of them can discriminate viral

particles from exosomes or extracellular vesicles (EVs). Indeed, most of these

technologies have been applied for the quantification and characterization of EVs.

This is a major limitation if such technologies are to be applied at the very first

stages of the process while viral production is occurring upstream. Indeed, most of

the cell factories are secreting exosomes and EVs during the viral replication

process. In that case, technologies allowing for specific labeling with antibodies

(virometry, HPLC) are still preferred to confirm the identity of the particles.

The selection of dedicated analytics for a specific application and process phase is

highly complex. Most of the development of new quantification technologies has

necessitated correlation/comparison with reference assays. Several comparison stu-

dies show high discrepancies between the different tools used for quantification thus

underlying a strong particularity of the viral samples. First, while comparing in-

fectious titration assays (either TCID50 of pfu) and total particle counting tools

(TPRS, HPLC, flow virometry), a difference of 1 to 3 log between infectious and total

particles is observed ([3], [4], and [27]). On the opposite, the total viral genome

quantified (qPCR, qRT-PCR, ddPCR) are always providing ranges above the total

particle counting methods. Indeed, while the producing cells are dying in the dynamic

production process, as the infection process is not perfect and complete, several viral

genomes not encapsulated in particles are released in the medium. If one relies only

on the qPCR to count the viral particles, this assay might overestimate the production

yield. Consequently, to develop or assess the performances of a viral production

process, at least two or three orthogonal quantification methods should be im-

plemented to generate consistent information about the production. Ideally, the se-

lection of assays targeting (i) the genome, (ii) the particle count, (iii) the infectivity

assays, or the protein content should be mandatory, and these assays should be in-

dicating the same trend of viral product accumulation over time.

8.4

PROCESS ANALYTICAL TECHNOLOGIES AND IN-LINE

ANALYTICS FOR VIRAL PRODUCTION PROCESSES

Process analytical technology (PAT) is a major part of the 2004 FDA guidance

establishing the concept of quality by design (QbD). The aim of quality by design

is to integrate quality assessment through all the processes and not only at the late

stage while assaying the final quality of the pharmaceutical product. Thus, reg-

ulatory agencies supported the concept to develop products where the quality will

be better described along the different process steps. The broad definition of PAT

includes all the tools that could be implemented to gain an improved understanding

and monitoring of the processes and the quality of the product. PAT tools are thus

applied for the monitoring and the detection of key parameters for a viral production

process, ranging from cell biomass quantification, evaluation of metabolite con-

centrations, physicochemical cell environment, etc. These tools can be classified

into four categories depending on their interaction with the process operations.

Thus, you can range analytical tools from off-line, at-line, on-line, and in-line where

Analytics and virus production processes

217