Xeno-free MSC harvesting is the step where an otherwise xeno-free expansion process most often reverts to animal material: the medium has been switched to human platelet lysate, the surface coating is recombinant, and then the culture is released with porcine trypsin and quenched with fetal bovine serum. This note covers what the harvest step does to the identity markers a mesenchymal stromal cell product is released on, what documentation an animal-origin-free dissociation enzyme has to carry, and how to qualify a change on your own line — including a clear statement of where the published data stops.
The problem in one sentence. The assay that defines an MSC — a surface-marker panel — is run on cells that have just been through a protease, so the enzyme and the exposure time are variables in your identity and potency data, not neutral background.
The Harvest Step Is a Manufacturing Step, Not Housekeeping
The International Society for Cell & Gene Therapy minimal criteria define human MSC by three things: plastic adherence under standard culture conditions; expression of CD105, CD73 and CD90 with absence of CD45, CD34, CD14 or CD11b, CD79α or CD19 and HLA-DR; and tri-lineage differentiation to osteoblasts, adipocytes and chondroblasts in vitro (Dominici et al., Cytotherapy 2006).
Two of those three criteria are read on the cell surface, and in most laboratories they are read immediately after harvest. The same is true of potency. An ISCT workshop on immune functional assays for MSC as a potency release criterion identified three preferred analytical methods for a matrix approach — quantitative RNA analysis of selected gene products, flow cytometry of functionally relevant surface markers, and protein-based assay of the secretome (Galipeau et al., Cytotherapy 2016). Two of the three are protease-exposed measurements.
So the detachment reagent sits upstream of identity and potency data. It is not a consumable choice made at the bench; it is a process parameter.
What Xeno-Free Has to Mean at the Harvest Step
The medium half of the problem is well covered. Human platelet lysate has been characterised as a clinical-grade replacement for fetal bovine serum across several manufacturing routes, including solvent/detergent virus inactivation, with Wharton jelly MSC maintaining morphology, doubling time, immunophenotype and tri-lineage differentiation over five passages (Chen et al., N Biotechnol 2019), and defined xeno-free platelet-lysate media have outperformed traditional serum-containing media for colony formation, proliferation and CD73, CD90 and CD105 staining in primary cartilage-derived progenitors (Mantripragada & Muschler, Curr Res Transl Med 2022).
The harvest half gets less attention, and it has a trap in it. Porcine trypsin is animal-derived, and the classic way to stop trypsin is to add serum — which reintroduces animal protein at the last step before formulation or cryopreservation. A xeno-free harvest therefore needs both halves solved: a non-animal enzyme, and a quench that is not serum. In practice that means a defined or recombinant inhibitor, dilution and centrifugation with a wash, or an enzyme formulated to be quenched by dilution alone.
This is a solved problem in published protocols rather than a theoretical one. A complete xeno-free protocol for Wharton jelly MSC replaced porcine trypsin with a recombinant enzyme and fetal bovine serum with 5% pooled human serum, then reported matched growth kinetics, differentiation potential, surface marker expression and colony-forming unit potential against FBS-expanded controls, with no significant deviation across a stem-cell and early-lineage gene panel (Venugopal et al., Stem Cells Cloning 2011).
Detachment Enzyme Choice Changes the MSC Immunophenotype You Measure
The most directly useful measurement in this area compared four detachment reagents on human synovial MSC across incubation times from 5 to 120 minutes, then read stem-cell-related surface antigens by flow cytometry (Tsuji et al., Cell Transplant 2017). The findings:
- Trypsin and a recombinant trypsin-like reagent (TrypLE™) both dissociated the cells within 5 minutes; collagenase and a non-enzymatic dissociation reagent needed 60 minutes for maximum yield.
- The non-enzymatic reagent significantly reduced viability at 120 minutes.
- Trypsin significantly reduced CD44+, CD55+, CD73+, CD105+, CD140a+, CD140b+ and CD201+ cell numbers within 30 minutes.
- Collagenase reduced CD140a expression by 30 minutes.
- The recombinant trypsin-like reagent did not affect the expression of any surface antigen tested by 30 minutes.
- Despite significant marker loss after 60 minutes of trypsin, the adverse effects of enzymatic digestion on MSC multipotency were limited.
Read that list against the ISCT panel: CD73 and CD105 — two of the three required positive markers — are among the antigens trypsin reduced within half an hour. The practical consequences are concrete.
- Exposure time is a release-assay variable. If harvest runs long on a Friday and short on a Tuesday, your identity percentages can move without the cells changing.
- Fix the reagent and the clock, then hold them. Time the detachment, record it, and treat a change in either as a change to the assay.
- Distinguish cleaved from lost. Cleaved epitopes re-express. Running the panel immediately after harvest and again after a recovery period separates a proteolytic artefact from a real change in the population.
The same phenomenon is documented in a different tissue. In a systematic comparison of dissociation methods for central nervous system tissue, papain most aggressively reduced antigenicity for CD24, CD133 expression was lowest after papain or Accutase®, and the authors concluded that a purified collagenase/neutral protease cocktail and a purified trypsin replacement gave the best balance of dissociation efficiency, viability and antigen retention (Panchision et al., Stem Cells 2007). Enzyme choice deciding what you can detect is a general property of dissociation, not an MSC quirk.
One boundary on all of this: those studies measured specific named reagents. They describe how trypsin, one recombinant trypsin-like product, collagenase, papain and a non-enzymatic reagent behaved in those systems. They are not evidence about any other product, and we do not extend them to ours.
Documentation an Animal-Origin-Free Dissociation Enzyme Has to Carry
A dissociation enzyme used in cell therapy manufacture is an ancillary or raw material: it contacts the product during manufacture but is not intended to be present in the final therapy. There is no single regulation that governs these materials; the accepted approach is that the user and the supplier jointly qualify each material for source, purity, identity, safety and suitability in the specific application (Solomon et al., Cytotherapy 2016), and that this qualification is revisited on a risk basis as the programme moves through clinical development (Solomon et al., Adv Exp Med Biol 2025).
That translates into a document list. Ask any supplier for all of it before a comparison, not after.
| What to ask for | Why it matters for an MSC process |
|---|---|
| Expression system and host organism, with the safety classification | Establishes that the protein is recombinant and identifies what else the host could contribute |
| Animal-origin-free / TSE-BSE statement covering the fermentation and the formulation | An enzyme can be recombinant and still be formulated with animal-derived components |
| Endotoxin limit and method | Endotoxin carried in at harvest travels with the cells into formulation |
| Sterility per Ph. Eur. 2.6.1 / USP <71> and mycoplasma status | Directly relevant if the reagent contacts an open process step |
| Activity assay with the unit definition and the release range | Without the unit definition, activity numbers between suppliers are not comparable |
| Purity method and specification | Residual proteolytic side activities are a plausible cause of epitope loss |
| pH and osmolality ranges | Harvest buffer that sits outside physiological range costs viability before the enzyme does |
| Quality system: ISO 9001 for research grade, GMP/GDP standard for the manufacturing grade | Determines what you can carry into a regulatory filing |
| Shelf life with the storage condition, plus excursion data | Shipping excursions are the routine failure mode for cold-chain reagents |
| Lot-to-lot data and change-notification agreement | A supplier change you learn about after the fact is a deviation you cannot explain |
No supplier can waive your change control. Swapping a harvest enzyme in a clinical or clinical-track process is a change under your own quality system, and the comparability exercise belongs to you. Treat any vendor claim that a substitution needs no revalidation as a claim about their product, never as a statement about your process.
Where CellTrypase Fits — and What It Has No Data For
CellTrypase from Kerry (formerly c-LEcta) is one animal-origin-free option in this class: a recombinant trypsin-like serine protease from a Fusarium oxysporum gene expressed in a Bacillus sp. host (GMO safety level S1), ~22 kDa, cleaving after lysine and arginine; ≥95% purity by HPLC; formulated in PBS with 1.1 mM EDTA and sterile filtered; pH 7.1–7.6 and 270–320 mOsm/kg; endotoxin ≤1 EU/mL at 1× and ≤10 EU/mL at 10×; mycoplasma negative; sterility per Ph. Eur. 2.6.1 / USP <71>. Activity is 0.6–1.2 kU/L at 1× and 6–12 kU/L at 10×, where one unit releases 1 µmol p-nitroaniline per minute at 37 °C from 8 mM L-AAPA at pH 8.0. It is quenched by dilution in buffer or medium, so no trypsin inhibitor enters the stream — which is what makes it workable in a serum-free harvest. Stability is at least 21 months at 2–8 °C, and brief warm excursions in shipping do not affect activity, although long-term stability is still under investigation by the manufacturer. R&D grade is manufactured under ISO 9001:2015; GMP grade under EXCiPACT® GMP/GDP, from the same manufacturer.
What has been measured — and what has not
The dissociation table in the product information sheet covers four adherent monolayer lines:
| Cell line | Release time (min:s) | Viability | Yield |
|---|---|---|---|
| CHO-K1 | 2:25 | 98% | 106% |
| HEK 293 | 2:27 | 95% | 100% |
| Vero | 4:07 | 99% | 103% |
| MDCK | 23:32 | 99% | 102% |
Source: CellTrypase Product Information Sheet v4.0 (Kerry, 18 August 2026). MDCK is shown at 23:32 — roughly ten times the CHO-K1 incubation at the same viability. Junctional architecture, not enzyme quality, sets that number, which is exactly why a timing figure from one cell type predicts nothing for another.
Kerry has also published five-passage data (P0–P4) for CHO-K1, HEK 293, MDCK and Vero covering viability, relative yield and population doubling time, with doubling times equal to or shorter than the industry-standard control on all four lines; and human iPSC data, where 1× CellTrypase reached a single-cell suspension in 5 minutes against 7 minutes for the industry-standard control at 94% viability for both, 99% relative yield and residual aggregates below 5% in 2D, and about 10 minutes for both enzymes in 3D aggregates at 87% versus 85% viability, 104% relative yield and 12% versus 16% residual aggregates.
Source: Kerry scientific poster #11462, “Advancing cell dissociation in bioprocessing with CellTrypase” (February 2026); n = 3 technical replicates, mean ± SD.
None of that is MSC data. There is no published release time, viability, yield, recovery, CD-marker retention, CFU-F efficiency, doubling time or differentiation figure for bone marrow, adipose, umbilical cord or any other mesenchymal stromal cell source. The manufacturer lists MSC among the intended applications — an intention is not a measurement, and we will not put numbers on that gap. Anyone considering the switch for an MSC process should qualify it side by side against their current reagent, on their own cells, with their own release endpoints.
Free 100 mL sample for qualification
BioHippo supplies a free 100 mL sample of CellTrypase for side-by-side qualification — one per laboratory, while supplies last. That is enough for roughly 100 T-75 harvests at 1 mL per 25 cm², which covers the two-passage comparison below with material to spare. Request a free sample →
A Qualification Study You Can Run Alongside a Normal Passage
Split one flask, harvest each half with a different reagent on the same day by the same operator, and read the endpoints that your release specification already contains. Dosing guidance for CellTrypase is approximately 1 mL per 25 cm², 1× for most applications, pre-warmed to room temperature or 37 °C. Where a supplier describes a 1:1 substitution, treat that as dosing guidance only — verify detachment time and viability in your own cell model before switching.
| Endpoint | How to read it | What it tells you |
|---|---|---|
| Time to >90% release | Microscope check on a clock, recorded to the second | The single number most likely to differ between reagents; sets your new process window |
| Viability | Trypan blue or automated counter, immediately after quench | Detects over-digestion at the new exposure time |
| Recovery yield | Viable cells as a percentage of the parallel control arm | Yield loss shows up here before it shows up in the batch record |
| ISCT positive panel: CD73, CD90, CD105 | Flow immediately after harvest and after a recovery period | Separates a cleaved epitope, which returns, from a real population change, which does not |
| ISCT negative panel: CD45, CD34, CD14 or CD11b, CD79α or CD19, HLA-DR | Same run | Confirms the harvest has not selected a subpopulation |
| CFU-F efficiency | Colony assay at low density from each arm | Clonogenic capacity is more sensitive than viability to harvest damage |
| Population doubling time | Over at least three passages after the switch | Catches a slow accumulating effect a single passage would miss |
| Tri-lineage differentiation | At the final passage of the comparison | Confirms the third ISCT criterion is intact |
| Potency assay | Whichever matrix assay your programme uses | The endpoint a regulator will ask about; do not leave it until after the switch |
Run both arms from the same flask on the same day. A comparison split across two weeks measures your incubator as much as your enzyme.
Related reading: animal-origin-free cell dissociation for cell and gene therapy — what the documentation has to cover, the CellTrypase specifications guide, and the CellTrypase cell dissociation protocol.
Frequently Asked Questions
Does trypsin damage MSC surface markers?
In human synovial MSC, trypsin significantly reduced CD44+, CD55+, CD73+, CD105+, CD140a+, CD140b+ and CD201+ cell numbers within 30 minutes, while a recombinant trypsin-like reagent did not affect any antigen tested at that time point; effects on multipotency were limited even after 60 minutes of trypsin (Tsuji et al. 2017). Treat exposure time as an assay variable and hold it constant.
How do I stop a trypsin-like enzyme without serum?
Three routes: a defined or recombinant inhibitor, dilution and centrifugation with a wash, or an enzyme formulated so that dilution alone is sufficient. CellTrypase is quenched by dilution in buffer or medium and needs no trypsin inhibitor; confirm the quench method against the instructions for whichever reagent you use.
Is an animal-origin-free enzyme required for MSC work?
Not for research use. It becomes a practical requirement when the process is clinical-track or has to satisfy a xeno-free claim, because animal-derived reagents carry TSE/BSE and adventitious-agent risk that has to be assessed and documented for every ancillary material (Solomon et al. 2016).
What performance data exists for CellTrypase on MSC?
None. Published comparative data covers CHO-K1, HEK 293, MDCK, Vero and human iPSC in 2D and 3D. MSC are named by the manufacturer as an intended application with no performance figures attached, so we quote none. The honest route is a side-by-side qualification on your own line, which is what the free 100 mL sample is for.
Do I need to revalidate my process if I change the harvest enzyme?
That decision belongs to your quality system, not to a supplier. A harvest reagent is an ancillary material contacting the product during manufacture, and any change to it is assessed under your own change control with a comparability exercise sized to your development stage.
How much enzyme does a typical MSC harvest need?
Approximately 1 mL per 25 cm² of growth surface, 1× concentration for most applications, pre-warmed to room temperature or 37 °C. A T-175 flask therefore needs about 7 mL. Confirm the release time on your own line rather than carrying over the timing from your previous reagent.
Reagents and Support
Browse dissociation enzymes in the BioHippo enzymes collection, or see the CellTrypase product page for specifications, documents and pricing. For help designing a harvest comparison on your own MSC line, ask a scientist.
For research use only; not for use in diagnostic or therapeutic procedures. The R&D grade of CellTrypase is supplied for research use only. TrypLE™ is a trademark of Thermo Fisher Scientific and Accutase® is a registered trademark of its respective owner; both are referred to here for identification and comparison only, and BioHippo is not affiliated with, endorsed by, or sponsored by those companies. Published findings are attributed to their authors and apply to the reagents and cell types those authors tested; they are not claims about CellTrypase. CellTrypase performance figures are attributed to the CellTrypase Product Information Sheet v4.0 (Kerry, 18 August 2026) for the four-cell-line table and to Kerry scientific poster #11462 (February 2026) for the five-passage and human iPSC data; they apply only to the cell types named, and no performance data is claimed for mesenchymal stromal cells or any other cell type. Figures 1–4 are illustrative summaries of the cited publications and product documentation, drawn to aid interpretation; they are not experimental data generated by BioHippo and contain no values beyond those in the sources named in each figure. Volumes, concentrations and timings given here are starting values to be confirmed on your own model, not specifications. Peer-reviewed sources are cited inline and indexed in PubMed.
