TFF is sufficient only when the impurity problem is size-based
Tangential flow filtration (TFF) can concentrate an ADC, exchange buffer, and remove permeable species such as salts, solvents, quench reagents, and monomeric linker-payload. It cannot provide high-resolution separation among ADC molecules that differ mainly in drug-to-antibody ratio (DAR), surface charge, conformation, or aggregation state. Those species are generally retained together by a membrane selected to retain the antibody.
The first process question is therefore not whether TFF is simpler than chromatography. It is whether every material that must be removed is actually present in a membrane-permeable state. If the answer is no, a property-selective step is needed.
Conjugation chemistry defines the starting impurity map
Random lysine conjugation can distribute payload across multiple sites and loading states. Two batches can share the same average DAR while differing in site occupancy and in the proportions of unconjugated, target-DAR, and high-DAR species. Average DAR alone does not establish distributional consistency.
Interchain cysteine conjugation limits the loading pattern to more predictable even-numbered DAR families, but reduction, reoxidation, maleimide conversion, and structural recovery still affect the product mixture. Higher loading generally increases hydrophobic character and can increase aggregation propensity.
Engineered cysteine, glycan remodeling, enzymatic tags, and other site-selective approaches narrow the number of accessible conjugation states. This shifts part of the purification burden into reaction control: antibody-intermediate quality, activation, stoichiometry, and endpoint control must be strong enough that downstream processing is not expected to repair a broad distribution.
Use TFF for measurable clearance, not hidden distribution correction
Under an ideal constant-volume diafiltration model, the concentration of a freely permeable solute falls exponentially with diavolume. With a sieving coefficient near one, approximately ten diavolumes corresponds to about four logs of theoretical clearance. A real process should be evaluated against its own clearance curve rather than this ideal endpoint alone.
A curve that flattens can indicate system hold-up, membrane adsorption, reversible association, or formation of colloidal or protein-bound payload species. A low-molecular-weight linker-payload may therefore clear poorly even when its molecular mass is far below the membrane cutoff. Membrane chemistry, protein concentration, pH, ionic strength, temperature, and cosolvent level can all change the observed behavior.
This is why a TFF-only route is credible only when the conjugation produces little aggregate, the DAR profile is already acceptable, free payload remains permeable, and recovery plus clearance are demonstrated across a defined operating window.
HIC, CEX, and membranes answer different selectivity questions
HIC separates by hydrophobic expression
Hydrophobic interaction chromatography (HIC) is a direct option when payload loading creates a useful hydrophobicity gradient. Unconjugated or lower-DAR species typically retain less strongly than higher-DAR species, allowing a preparative method to remove both ends of an unwanted distribution. Published HIC methods also show why resin ligand, salt identity, temperature, load, and elution design must be developed together rather than transferred as a single universal condition.
CEX targets charge and multivalent binding differences
Cation exchange chromatography (CEX) is useful when aggregation, conformation, or linker-payload ionization creates a charge-based separation window. It may operate in bind-and-elute mode or as flow-through polishing in which the desired ADC passes while more strongly interacting impurities are retained. A published patent family describes using related CEX knowledge before conjugation and again after conjugation for aggregate-focused flow-through polishing.
Membrane chromatography combines selectivity with short flow paths
Ion-exchange and hydrophobic membrane adsorbers provide convective transport and can be arranged as complementary operations. A published PBD-ADC study used an S-type cation-exchange membrane for free linker-payload clearance and a phenyl membrane for high-DAR and aggregate reduction. This is still chromatography: the format changes, but separation still depends on surface-property differences rather than molecular size alone.
Development data should close the mass balance
A defensible purification route explains where each impurity is formed, what state it occupies, why the selected unit operation recognizes it, and where it exits the process. Four measurements are especially useful:
- Quantify feed, retentate, permeate, flush, and chromatography fractions. Low residual payload is not proof of clearance if overall recovery is also missing.
- Plot free payload, solvent, and salt against diavolume. Compare the profiles with the expected sieving behavior and investigate any plateau.
- Use stability-indicating LC-MS to distinguish parent linker-payload, quench adducts, hydrolysis or oxidation products, and released payload rather than reporting one undifferentiated small-molecule signal.
- Pair an average-DAR result with a distribution-sensitive method. HIC or native MS can describe loading families, while reduced LC-MS and peptide mapping provide complementary chain- and site-level information.
SEC should also track monomer, aggregate, and fragment behavior through concentration and polishing. Conjugation can change aggregation behavior, so the antibody intermediate and the conjugated product need separate evidence.
Practical route patterns for process teams
When reaction control already delivers a narrow DAR distribution and free payload behaves as a true permeant, a pre-purified antibody intermediate can proceed through conjugation, TFF, final concentration, and formulation exchange.
When conjugation introduces aggregate or an unacceptable DAR tail, TFF can first lower solvent and small-molecule load, followed by HIC or CEX polishing and a final TFF step. If free payload shows strong adsorption or association, an ion-exchange, hydrophobic, or mixed-selectivity adsorber may be more informative than simply adding diavolumes.
The shortest process is the one that meets the impurity target with demonstrated recovery and a closed analytical account. Removing a column is useful only when reaction and clearance data have already removed the need for its selectivity.
FAQ
Can membrane cutoff alone predict free payload clearance?
No. Molecular mass is only one factor. Association, adsorption, protein binding, membrane material, and solution conditions determine the effective sieving behavior.
Is average DAR enough to select a purification route?
No. The same average can arise from different proportions of DAR species or different conjugation sites. Distribution-sensitive and site-sensitive methods are needed to understand the product family.
When is HIC preferred over CEX?
HIC is a logical starting point when payload loading creates the clearest hydrophobicity difference. CEX is more useful when charge, conformation, or multivalent aggregate binding provides the stronger contrast.
Why is TFF often used after chromatography?
Chromatographic elution can leave the ADC in a high-salt or otherwise non-final buffer. TFF then supports buffer exchange and concentration without being asked to perform the earlier selective separation.
References
- Cordova JC, Sun S, Bos J, et al. Development of a Single-Step Antibody–Drug Conjugate Purification Process with Membrane Chromatography. Journal of Clinical Medicine. 2021;10(3):552.
- Becker CL, Duffy RJ, Gandarilla J, Richter SM. Purification of ADCs by Hydrophobic Interaction Chromatography. Methods in Molecular Biology. 2020;2078:273–290. First published online 24 October 2019.
- Matsuda Y, Leung M, Okuzumi T, Mendelsohn BA. A Purification Strategy Utilizing Hydrophobic Interaction Chromatography to Obtain Homogeneous Species from a Site-Specific Antibody Drug Conjugate Produced by AJICAP™ First Generation. Antibodies. 2020;9(2):16.
- Chen T, Chen Y, Stella C, Medley CD, Gruenhagen JA, Zhang K. Antibody-drug conjugate characterization by chromatographic and electrophoretic techniques. Journal of Chromatography B. 2016;1032:39–50.
- Gandhi AV, Randolph TW, Carpenter JF. Conjugation of Emtansine Onto Trastuzumab Promotes Aggregation of the Antibody–Drug Conjugate by Reducing Repulsive Electrostatic Interactions and Increasing Hydrophobic Interactions. Journal of Pharmaceutical Sciences. 2019;108(6):1973–1983.
- Hendricks R, Hutchinson MH, Fedesco MF, Tran BP. Flow through cation exchange chromatography purification processes for antibody drug conjugates. Patent record.