# Hollow-fibre biomanufacturing and cell-free engineering of HEK293 extracellular vesicles.

- **DOI:** [10.1016/j.tibtech.2026.05.005](https://doi.org/10.1016/j.tibtech.2026.05.005)
- **Rating:** Sound (3.1 out of 5)
- **Verdict:** An integrated cell-free EV engineering workflow advances laboratory prototyping, but missing attribution controls limit its central claim of successfully engineered vesicles.
- **Framework version:** v5.10
- **Pipeline version:** v0.4
- **Evaluation date:** 2026-10-19
- **Evaluation page:** https://www.nabu.science/paper/10.1016%2Fj.tibtech.2026.05.005

## Key claims

1. up to 4.83 × 10¹¹/ml of HEK293 EVs were successfully cell-free engineered
   - Support: Tentative
   - Type: Primary finding
   - Basis: missing EV-free fusion-expression controls prevent attribution of the fluorescent-particle count specifically to engineered HEK293 EVs
   - Location: Abstract; Results, Cell-free EV engineering (Figure 3D)
2. hollow-fibre-based cell culture is a feasible and scalable way to generate significant amounts of EVs for cell-free EV engineering applications
   - Support: Tentative
   - Type: Secondary
   - Basis: serial harvests from one bioreactor establish feasibility without independently testing scale-up or run-to-run reproducibility
   - Location: Introduction, final paragraph; Results, Hollow-fibre biomanufacturing of HEK293 EVs
3. CD63ITM3-mGL co-localised and EV integrated more efficiently than CD63-mGL
   - Support: Supported
   - Type: Secondary
   - Basis: the topology-assay comparison supports a directional scaffold difference within a single EV-source and lysate system
   - Location: Results, Cell-free EV engineering (Figure S11D)
4. cell-free engineered HEK293 EVs and their luminal protein cargoes remain largely intact
   - Support: Supported
   - Type: Secondary
   - Basis: qualitative TEM and cargo-dye staining support the hedged integrity claim without establishing definitive retention of original luminal cargo
   - Location: Results, Cell-free EV engineering (Figures S10, S12)

## Strengths

- **Detailed methods support laboratory workflow reproduction** STAR Methods specifies constructs, reaction compositions, processing steps, and instrument settings, with whole-plasmid sequencing used to verify constructs.
  - Dimension: methodological rigour
  - Location: STAR Methods, Strain and plasmid construction and Dialysis-mode cell-free EV engineering reactions
- **Complementary assays characterize source vesicle preparations** Particle sizing, EV-marker assays, microscopy, and membrane labeling provide complementary characterization of UC- and TFF-isolated preparations.
  - Dimension: methodological rigour
  - Location: Results, Characterisation of HEK293 EVs; Figures 1–2
- **Explicit readiness framing bounds therapeutic expectations** The platform is placed at TRL3/4, and potency assays, improved isolation, expression yields, and automation are identified as remaining development requirements.
  - Dimension: positioning
  - Location: Technology readiness; Outstanding questions

## Limitations

- **Missing control leaves fluorescent particle attribution unresolved** An EV-free reaction expressing the CD63 fusion and undergoing the same SEC and nano-flow processing is not reported. The evidence therefore does not exclude fusion-bearing particles generated by the cell-free system itself.
  - Dimension: methodological rigour
  - Location: Results, Cell-free EV engineering; STAR Methods, Dialysis-mode cell-free EV engineering reactions
- **Headline concentration exceeds the implied vesicle input** The stated reaction recipe implies 1.5 × 10^9 input EVs/ml, whereas Figure 3D reports approximately 4.83 × 10^11 fluorescent nanoparticles/EVs per ml. The concentration reporting basis is not reconciled sufficiently to establish a corrected engineered-EV count.
  - Dimension: reporting
  - Location: Abstract; Figure 3D; STAR Methods, Dialysis-mode cell-free EV engineering reactions
- **Manufacturing comparison rests on one bioreactor** The 49 batches are serial harvests from a single 26-day culture rather than independent bioreactor runs. Different final isolation volumes and protein normalization also complicate interpretation of UC–TFF yield and contamination comparisons.
  - Dimension: methodological rigour
  - Location: Results, Hollow-fibre biomanufacturing of HEK293 EVs; Figure 1; STAR Methods, Isolation of HEK293 extracellular vesicles

## Reliability checks

Caveats (4 of 4 checks).

- **Results check:** Noted
  - up to 4.83 × 10^11 /ml of HEK293 EVs were successfully cell-free engineered to incorporate CD63ITM3-mGL membrane fusion proteins: "3.6 μl of HEK293 EVs (2.5×1010/ ml of either control" (STAR Methods, Dialysis-mode cell-free EV engineering reactions) vs "which were ~4.83 ± 0.35 ×1011 /ml (Figure 3D)" (Results, Cell-free EV engineering (Figure 3D)). Using the stated 60 μl reaction volume, 3.6/60 × 2.5 × 10^10 gives 1.5 × 10^9 input EVs/ml, approximately 322-fold below the reported fluorescent-particle concentration, although the concentration reporting basis is unresolved. → Effect on result unclear
- **Reporting and transparency:** Clean. Competing interests and resource availability are declared, with plasmids available under an MTA and additional information available on request. No substantive research-conduct finding is established.
  - Conflict of interest: declared · Data availability: declared · Code availability: not stated
- **Reference integrity:** Clean. 102 of 102 checkable references verified
  - 103 references in manuscript
  - 1 could not be checked — not the same as no match found
  - 2 references confirmed by manual review
  - References not verified automatically are resolved manually within 24 hours.

## Post-publication record

- **Status:** Clean
- **Summary:** No retraction or expression of concern found in Crossref (includes Retraction Watch).
- **Last checked:** 2026-10-19
