Continuous Bioprocessing — Perfusion Culture and Continuous Chromatography CMC Integration in BLA Submissions

Continuous bioprocessing delivers higher volumetric productivity than fed-batch, a smaller facility footprint, and improved quality attribute consistency from steady-state operation. The biopharmaceutical industry knows this. FDA’s biologics reviewers know this. And yet the BLA deficiency rate for continuous bioprocess submissions remains elevated — not because the continuous process is less well-characterized than a fed-batch process, but because the CMC package translating that continuous process into a regulatory submission does not map the process description, in-process control strategy, and lot definition onto the discrete-operations framework a 3.2.S.2 section is built around.

The process runs continuously; the regulatory submission has to be understood discretely. The gap between those two requirements is where deficiencies are born.

Perfusion Culture Steady-State Window, IPC Strategy, and the Lot Boundary Definition That Translates Continuous Operation Into Discrete Regulatory Language

A perfusion process’s steady state is not a single parameter but a convergence of several simultaneously: viable cell density holding within a defined operating band, viability remaining above a defined floor, and cell-specific perfusion rate — the perfusion volume delivered per cell per day — staying within a range that governs glycosylation and charge variant consistency as directly as any other process parameter in the entire manufacturing train. In-process monitoring for this steady state combines daily offline sampling for cell density, viability, and key metabolites with in-line capacitance and electrode-based monitoring filling the gaps between samples, but the regulatory translation challenge sits one level up from the monitoring itself: defining where one lot ends and the next begins. A time-based lot definition — commonly a full day of production at confirmed steady-state conditions — requires that the in-process control data from that period actually confirm steady state was maintained throughout, with any out-of-specification reading during the period triggering the diversion logic for that specific lot rather than being absorbed silently into an undifferentiated continuous stream. A 3.2.S.2.2 process description that narrates the continuous operation without ever specifying where these lot boundaries fall, what in-process data confirms quality within each boundary, and how off-specification material gets diverted and disposed of within an otherwise continuous flow leaves the reviewer with no way to evaluate material traceability or lot disposition logic at all — a gap that has drawn deficiency letters requesting an explicit lot boundary definition and a flow diagram connecting continuous operation to discrete disposition decisions.

MCC Continuous Chromatography in 3.2.S.2.2 — Load Challenge, Per-Cycle CPPs, and the Discrete Process Step Description the Reviewer Can Evaluate

Multi-column continuous chromatography achieves higher resin utilization than single-column batch capture by running overlapping load-wash-elute cycles across several columns operating in series, and the loading strategy itself is defined by a load challenge expressed as a percentage of the resin’s dynamic binding capacity — commonly held around 110% at a defined breakthrough threshold — cycling through wash, elution, and regeneration steps on a tight per-column time budget, often under 45 minutes per cycle. The regulatory translation challenge here isn’t the underlying chemistry, which is the same Protein A capture and low-pH elution any batch process uses — it’s describing a continuously cycling operation in a format a reviewer evaluating discrete process steps can actually assess. That means treating each complete MCC cycle as its own discrete process step with its own defined critical process parameters — load challenge, flow rate, buffer volumes, elution pH — and its own in-process control, typically in-line UV absorbance monitoring load breakthrough in real time with a defined acceptance threshold at end-of-load. A process description that instead presents the MCC operation as a continuous, undifferentiated block of cycling activity, without mapping it onto this discrete per-cycle structure, has described the technology accurately while failing to give the reviewer the reviewable process step table a 3.2.S.2.2 section actually requires.

Viral Clearance Validation for Continuous Chromatography — Scaled-Down MCC Cycling Model, Per-Cycle LRV, and the Cumulative Biosafety Package

Viral clearance validation for a continuous multi-column chromatography step cannot simply borrow the validation data generated for an equivalent single-column batch operation, because that batch data was never designed to demonstrate that clearance performance holds up across repeated cycling — and FDA reviewers have specifically rejected LRV claims for a continuous capture step supported only by single-column batch validation data, since a batch model doesn’t represent what actually happens across many consecutive cycles in a real continuous campaign. The defensible validation approach uses a scaled-down model of the actual multi-column system, cycling repeatedly to represent a full production campaign’s worth of consecutive cycles, spiked with a model virus at the load step and assessed for log reduction value at each individual cycle’s elution pool — with the critical demonstration being that clearance performance doesn’t decline across the cycling sequence, holding consistently above a defined minimum log reduction value cycle after cycle rather than degrading as the resin and system see repeated use. That per-cycle demonstration then integrates into the cumulative viral clearance claim across the full downstream process, which for a well-designed continuous capture step combined with orthogonal downstream clearance steps can reach substantial cumulative log reduction values for both enveloped and non-enveloped model viruses. A biosafety section presenting only aggregate viral clearance figures without the per-cycle validation data demonstrating consistency across repeated cycling has not actually validated the continuous chromatography step — it has validated a different, batch-mode process and assumed the result transfers.

The XGene Continuous Bioprocessing BLA CMC Architecture — Steady-State Window, Lot Definition, IPC, MCC Process Description, Viral Clearance, and Comparability

The XGene Continuous Bioprocessing BLA CMC Architecture is a structured framework built around the recognition that a continuous bioprocess’s regulatory submission has to translate continuous operating logic into the discrete process-step language a 3.2.S.2 section requires, at every level from lot definition to viral clearance validation.

1. Steady-State Window and Lot Boundary Definition — Define the perfusion process’s steady-state operating parameters and translate them into explicit, IPC-confirmed lot boundaries. 2. Discrete MCC Process Step Description — Present each continuous chromatography cycle as its own defined process step with dedicated CPPs and in-line IPC, not as an undifferentiated continuous block. 3. Per-Cycle Viral Clearance Validation — Validate viral clearance using a scaled-down model of the actual cycling system across a representative number of consecutive cycles, demonstrating consistency rather than borrowing batch-mode data. 4. Comparability to Reference Fed-Batch Material — Build a pre-specified ICH Q5E comparability protocol with defined acceptance criteria before generating the comparability dataset. 5. Design Space and Pre-BLA Meeting Strategy — Document the steady-state operating window and process dynamics per ICH Q8(R2), engaging FDA early through Type A or B meetings to align on the CMC translation strategy.

The output is the continuous bioprocessing BLA package that gives a reviewer evaluating discrete process steps the reviewable structure they need, without sacrificing the technical accuracy of the underlying continuous process.

The FDA CBER review record for Roctavian (valoctocogene roxaparvovec-rvox, BioMarin, BLA 125720, approved June 29, 2023), the first approved gene therapy for severe hemophilia A, illustrates FDA’s willingness to review advanced, non-conventional upstream production processes for biological products, offering a relevant reference point for continuous bioprocessing programs extending beyond monoclonal antibody platforms. Published industry data on perfusion-plus-continuous-chromatography integrated platforms has documented steady-state consistency in critical quality attributes across extended continuous campaigns, providing the benchmark this article’s framework builds from.

For your continuous bioprocessing BLA CMC package, can you confirm today that your 3.2.S.2.2 process description includes explicit lot boundaries with IPC steady-state confirmation criteria, and that your viral clearance validation for the continuous chromatography step was generated using a scaled-down cycling model demonstrating consistent LRV across multiple consecutive cycles rather than a single-column batch model?