Optimizing LNP Feasibility Batches with Tangential Flow Filtration

Downstream processing of LNPs involves removing residual ethanol and exchanging to final storage buffer through tangential flow filtration (TFF). The process requires careful optimization of multiple parameters including membrane surface area (balancing flux against hold-up volume), shear rate (controlled by flow rate to minimize particle damage), and transmembrane pressure (driving force through the membrane). Studies comparing 4,000, 6,000, and 8,000 s⁻¹ shear rates demonstrated significant time reductions from 185 to 125 minutes when increasing from 4,000 to 6,000 s⁻¹, with diminishing returns at higher rates. Typical configurations use 300 kDa cutoff membranes at 4 bar transmembrane pressure. Critical quality attributes (particle size, PDI, encapsulation efficiency) must be monitored throughout to ensure they remain within acceptance criteria, as small variations at small scale become magnified at larger production volumes.

Analysis revealed mRNA concentration most significantly impacts particle size, PDI, and encapsulation efficiency. Flow rate ratio also affects size and PDI but less dramatically. Total flow rate (90-115 μL/min) showed minimal impact, validating manufacturer guidance on minimum flow requirements. Encapsulation efficiency increased during TFF processing. Formulation differences emerged: Formulation 2 showed greater sensitivity to starting material concentrations in encapsulation efficiency compared to Formulation 1. Statistical analysis confirmed mRNA concentration as the primary driver, with formulation effects becoming apparent when analyzing grouped data. These findings guide optimization priorities for LNP manufacturing.

This video introduces an innovative analytical platform called LNP switcher that enables simultaneous monitoring of three critical quality attributes of lipid nanoparticles (LNPs) in a single run: encapsulation efficiency (determined by separating free nucleic acids from encapsulated nucleic acids using hydrophobic interaction and reverse-phase chromatography), LNP size and size distribution (measured using multi-angle light scattering detection), and lipid composition (analyzed using charged aerosol detection). The platform uses two analytical columns connected inline with three detectors—UV for nucleic acids, multi-angle light scatter for particle sizing, and charged aerosol detector for lipids—allowing researchers to evaluate vital LNP characteristics without extensive sample pretreatment, delivering results in under 45 minutes.

Comprehensive analytical monitoring ensures LNP quality throughout TFF processing: sampling at concentration and diafiltration stages (1x, 2x, 4x, 5x volumes) with testing for particle size, PDI, encapsulation efficiency, RNA integrity, and potency. Membrane selection is formulation-dependent—PES membranes suit some formulations while cellulose esters suit others, and hollow fiber versus cassette configurations show different performance characteristics. No universal membrane recommendation exists; empirical testing determines optimal choices. Understanding these relationships enables process troubleshooting and optimization for specific LNP formulations.

Multi-angle dynamic light scattering (DLS) is essential for accurate LNP characterization because it detects aggregation that single-angle DLS misses; Stunner's rotating angle DLS (Rattles) measures at seven angles to provide angle-independent Z-average sizing, enabling reliable detection of aggregates and accurate particle concentration measurements across different LNP formulations and buffer conditions.

Tangential Flow Filtration (TFF) is essential for downstream processing of nanoparticles formed with organic solvents requiring removal to below 10%. The system uses hollow fiber filters with feed, retentate, and permeate streams. Critical parameters include transmembrane pressure (driving force calculated from feed and retentate pressures minus permeate pressure) and shear rate (tangential flow velocity expressed in inverse seconds). Applications include concentration, buffer exchange (diafiltration), and fed-batch processing. Membrane selection follows MWCO rules where retention cutoff should be 3-5 times smaller than target species. Process optimization requires systematic variation of shear rate and transmembrane pressure to identify optimal conditions. Shear rate above 4000 s⁻¹ provides effective membrane sweeping action. Filtrate flux (LMH) is the normalized performance metric. The optimization strategy identifies membrane-controlled regions (where increasing TMP improves flux) versus gel-layer-controlled regions (where TMP increases provide no benefit). Optimal operating ranges are typically selected at 15% or more before the flux plateau occurs.

An independent study conducted at Fiser compared membranes from several major suppliers, including 50 kDa PES membranes. In flux versus TMP experiments, the Tangen membrane demonstrated the highest permeate flux rate compared to other membranes tested. This performance advantage translates to practical benefits in bioprocessing operations, including reduced process time or decreased membrane surface area requirements for equivalent throughput.

This comprehensive section covers the foundational principles of tangential flow filtration. Membrane selection requires evaluating four key factors: application purpose (microfiltration for larger molecules, ultrafiltration for smaller ones), available configurations (cassettes for general use, fibers for high viscosity, ceramic membranes for extreme conditions), molecular weight cut-off (3-5 times smaller than target molecule), and additional factors including molecular shape, charge, conductivity, and concentration. System preparation involves installation, flashing, and sanitization steps that must be factored into total process time, with single-use systems eliminating preconditioning requirements. Integrity testing ensures system validity through air diffusion testing (measuring gas flow against supplier limits) and pressure decay testing (monitoring pressure decrease after stabilization). Two key resistance phenomena affect TFF performance: polarization (gel layer formation, essential in ultrafiltration that can be optimized) and fouling (molecules trapped on surfaces or within pores, primarily affecting microfiltration). In ultrafiltration with protein solutions, permeate flux increases linearly with transmembrane pressure within the membrane-controlled region, transitioning to a plateau controlled by concentration polarization. Stable TMP indicates controlled optimization, while fluctuations signal uncontrolled processes.

Membrane performance depends on the tightness and thickness of the gel layer of retained molecules pressed against the membrane surface. This gel layer formation directly affects both permeate flow rate and selectivity. Even small changes in cross-flow flux, transmembrane pressure, or feed characteristics immediately impact gel layer formation and consequently permeate flow rate. The sensitivity of membrane performance to feed conditions makes diafiltration an excellent method for monitoring feed homogeneity.

Modified PES (polyethersulfone) membranes in hollow fiber format are most frequently used for TFF of LNPs for buffer exchange and concentration. Flat sheet membranes and regenerated cellulose membranes are also options being investigated. The selection of appropriate membrane material can impact both processing efficiency and particle characteristics during downstream purification.

This video explains how to calculate residual solvent limits in pharmaceutical products using two approaches from ICH Q3C and USP general chapter 467: Option 1 (fixed concentration ppm) for products with daily dose ≤10g, calculated as ppm = 1,000 × PD/10, where PD is permitted daily exposure in mg/day; and Option 2 (total daily exposure mg/day) for products with dose >10g or when individual components exceed ppm limits, calculated by summing (ppm × component amount)/1,000 for all ingredients and comparing to PD. The video emphasizes that residual solvents are classified into three classes based on toxicity (Class 1: toxic, Class 2: limited, Class 3: higher PD), and that testing is not required if no solvents are used in manufacturing and all excipients comply with limits.

ICH quality guidelines span 14 guidelines from Q1 to Q14. Q1 covers stability testing including Q1A (stability testing of new drug substances/products), Q1B (photo stability testing), Q1C (new dosage forms), Q1D (bracketing/matrixing designs), Q1E (stability data evaluation), and Q1F (stability data package for climate zones 3/4). Q2 focuses on analytical validation with Q2R1 covering analytical procedure validation and Q2R2 addressing analytical procedure development. Q3 addresses impurities including Q3A (drug substance impurities), Q3B (drug product impurities), Q3C (residual solvents), Q3D (elemental impurities), and Q3E (extractables/leachables control). These guidelines provide standardized frameworks for ensuring pharmaceutical product quality throughout development and manufacturing processes.

The ICH Q3C guideline classifies residual solvents into three categories: Class 1 solvents (benzene, carbon tetrachloride, chloroform, and three others) must be avoided due to unacceptable toxicity; Class 2 solvents (including acetone, methanol, and chloroform) require limitation based on inherent toxicity; and Class 3 solvents (acetic acid, ethanol, isopropanol, and 27 others) have low toxic potential. The ICH Q3D guideline classifies elemental impurities similarly: Class 1 elements (arsenic, cadmium, mercury, and lead) are human toxicons requiring avoidance; Class 2A elements (cobalt, nickel, vanadium) have high probability of occurrence; Class 2B elements (ten elements) have reduced probability of occurrence; and Class 3 elements (barium, chromium, copper, lithium, molybdenum, tin, and lead) have relatively low toxicities. To define exposure limits, three options are available: Option 1 uses a fixed product mass of 10 grams for products administered less than 10 grams daily; Option 2 uses the actual product mass for products administered more than 10 grams daily; and Option 2B (for elemental impurities only) calculates limits by summing contributions from individual components. Option 3 allows finished product analysis to justify limits when processing reduces impurity levels below permitted daily exposures.

ICH Q3C classifies residual solvents into three categories based on toxic potential: Class 1 (carcinogenic, to be avoided with 30% limit), Class 2 (animal carcinogenic/teratogenic, to be limited with 10% limit), and Class 3 (very low toxic potential, acceptable up to 50 mg); the guideline establishes Permitted Daily Exposure (PDE) limits and requires justification for use beyond specified limits, with analytical methods including gas chromatography for Class 1/2 solvents and loss on drying for Class 3 solvents.

ICH guidelines establish specific impurity limits based on dosage and impurity type. For drug substances (ICH Q3A): reporting threshold is 0.05% for dosages under 2g, identification threshold is 0.10% or 1mg, and qualification threshold is 0.15% or 1mg. Unknown impurity limits decrease proportionally above 650mg dosage to maintain 1mg maximum daily intake. For drug products (ICH Q3B): reporting threshold is 0.1% for dosages under 1g and 0.5% for dosages exceeding 1g. Degradation products that are significant metabolites in animal/human studies are generally qualified. Residual solvents follow ICH Q3C classification: Class 1 solvents should be avoided, Class 2 solvents require limitation based on PDE values, and Class 3 solvents have low toxic potential. Elemental impurities are controlled under ICH Q3D with mandatory risk assessment for Class 1 and Class 2A elements.