Chapter 6
Phase 3 of WIA-BIO-020 describes how a cell becomes a functioning graft. The protocol layer has five families: cell preparation, differentiation and reprogramming, implantation and clinical administration, vascularization and engraftment, and tumourigenicity and immunogenicity safety. The specification calls each family an "envelope", meaning a bounded set of parameters that an operator must document and hold fixed for a product. The integration problem lies in the handoffs between envelopes. A passage number set during expansion, a differentiation end-point, and a dose delivered into a scaffold only mean something if each one is traceable to the step before it.
This chapter follows the product in that order. Cell preparation and differentiation come first, then indication-specific implantation. Safety and engraftment follow, and the chapter closes with the controls that tie the whole path together.
Everything downstream inherits the quality of the preparation envelope. The specification breaks it into six stages, each of which has to be documented before the next begins.
Two stages deserve emphasis. Expansion is where biology and logistics collide: cells that are passaged too long drift genetically, a risk revisited under safety in 6.4. Post-thaw recovery is where the bank meets the operating room, because a viability result decides whether a thawed dose can go straight to the patient or must first be cultured back to health. In an integrated system both decisions are recorded against the same product identity that began at procurement.
| Envelope stage | What it fixes | Handoff concern |
|---|---|---|
| Donor screening | Eligibility under US/EU/Korean rules | No procurement without a determination |
| Procurement | Consent, IRB approval, custody | Custody continues into isolation |
| Isolation and expansion | Dissociation method, passage ceiling | Passage count travels with the lot |
| Cryopreservation and thaw | Freeze recipe, viability criteria | Release decision before implantation |
The technical basis of Phase 3 is a quantitative description of what cells do after they leave the bank: proliferate, commit to a lineage, migrate, and build tissue. The specification gives each behavior a simple measurable quantity, so a protocol can state a target and an assay can check it.
The differentiation envelope is defined per target lineage. For MSC the lineages are osteogenic, chondrogenic, and adipogenic. For iPSC-derived products they are cardiomyogenic, neural, hepatic, and pancreatic. Each envelope fixes a chemically defined medium, an inductive cocktail with documented concentrations and time-courses, a growth-factor set, and a maturation end-point. Positive and negative markers are validated at every stage rather than only at the end. Lineage commitment is driven by transcription factors (RUNX2 for bone, PPARG for adipose), by epigenetic state (DNA methylation, histone acetylation), and by Wnt, BMP, Notch, and Hedgehog signaling.
Two cocktails in the specification show how concrete these envelopes are:
Osteogenic:
MSCs + BMP-2 + Dexamethasone + β-glycerophosphate + Ascorbic acid
→ Osteoblasts → Mineralized matrix
Chondrogenic:
MSCs + TGF-β3 + BMP-6 + Dexamethasone → Chondrocytes
→ Collagen II, Aggrecan, SOX9⁺A neural path starts earlier in development: iPSCs form neural rosettes, then NPCs, then neurons, astrocytes, and oligodendrocytes. Reprogramming itself is an envelope. It names the vector (Sendai virus, episomal plasmid, mRNA, or piggyBac) and requires a transgene-clearance step. Afterwards, pluripotency must be validated by teratoma assay, PluriTest, or qPCR with immunostaining of pluripotency markers.
Efficiency is stated as a ratio:
D_eff = (N_differentiated / N_total) × 100%
Target efficiency: >80% for clinical applicationsThree other quantities cover the remaining behaviors. Proliferation follows P = N₀ × e^(rt) under cell-cycle control by cyclins D, E, A, and B, the CDKs, and the inhibitors p21, p27, and p16. Migration follows v = μ × ∇C, driven by homing factors such as the SDF-1/CXCR4 axis, VEGF, selectins, and integrins. Tissue formation is modeled as
dT/dt = α × C × G - β × T
I = (V_integrated / V_total) × F_vascular × F_mechanical
Target integration score: I ≥ 0.8where T is tissue volume, C is cell density, G is growth-factor concentration, α is the regeneration coefficient, and β is the degradation coefficient. The integration score is the useful handoff metric. It multiplies the integrated fraction of the graft by a vascular factor and a mechanical factor, so a graft that is well populated but poorly perfused or mechanically mismatched cannot pass.
The specification does not quote market figures, and this chapter does not supply any. What the protocol layer does reveal is how demand is structured. Autologous products, such as chondrocytes for cartilage repair or keratinocytes and fibroblasts for skin, carry a patient-specific expansion step. Skin constructs need two to three weeks of cell expansion before they can be applied. Allogeneic products shift the burden elsewhere. They need HLA matching, an immunosuppressive regimen where applicable, and per-patient surveillance for anti-donor antibodies. Cell doses also span orders of magnitude, from millions of cells per square centimeter of cartilage to hundreds of millions for the heart. Manufacturing scale, cryo-logistics, and follow-up obligations therefore differ far more between indications than a single "cell therapy" label suggests.
The implantation envelope fixes the route of administration: intra-articular, intra-muscular, intra-coronary, intravenous, intrathecal, topical on a scaffold, or surgical-site implant. It also fixes the dose-escalation plan from the trial protocol and the per-administration monitoring. That monitoring covers vital signs, cytokine-release syndrome where applicable, and immediate immunogenicity. Because some cellular therapies modulate systemic immune responses, the specification also calls for monitoring of sympathetic-nervous-system effects.
The scaffold is the vehicle for most doses, so its design is part of the protocol. The specification asks for biocompatibility, controlled biodegradability matched to tissue growth, porosity of 70-90% for nutrient diffusion, mechanical strength matched to native tissue, and surface chemistry that supports adhesion. Degradation is modeled as M_remaining(t) = M₀ × e^(-kt). The choice of material follows the indication. Fibrin and alginate are used for cardiac delivery, hyaluronic acid and chitosan for neural, hydroxyapatite, β-TCP, and PLGA for bone, collagen type I/III or hyaluronic acid for cartilage, and collagen with fibrin glue for skin.
| Indication | Cell sources | Specified protocol parameters | Success metrics |
|---|---|---|---|
| Cardiac (post-MI) | iPSC cardiomyocytes, cardiac progenitors, MSCs (paracrine) | 2-5 × 10⁸ cells; 2-4 weeks post-MI; VEGF, FGF-2, IGF-1; fibrin or alginate hydrogel | Ejection fraction >5%; infarct size reduction >15%; vessel density increase >30% |
| Neural | NPCs, iPSC neurons, Schwann cells | 5-10 × 10⁶ cells; intraspinal or intracerebral; hyaluronic acid or chitosan; cyclosporine or tacrolimus | Axon regeneration >500 μm; ASIA scale improvement; restored conduction |
| Bone | BM-MSCs, adipose MSCs, periosteal cells | 5 × 10⁶ cells/cm³; BMP-2 100-200 μg; 7-14 days in vitro; direct surgical placement | BV/TV >40%; >70% native strength; bridging in 3-6 months |
| Cartilage | Autologous chondrocytes, MSCs, iPSC chondrocytes | 1-5 × 10⁶ cells/cm²; TGF-β3 10 ng/ml, BMP-6 100 ng/ml; arthroscopic or open surgery | ICRS ≥8/12; collagen II >70%; GAG >60% of native |
| Skin | Keratinocytes, fibroblasts, MSCs | Keratinocytes 5 × 10⁴/cm², fibroblasts 1 × 10⁴/cm²; EGF, KGF, PDGF; bilayer construct | Closure >90% at 3 weeks; Vancouver Scar Scale <5; pigmentation restored in 6 months |
Three patterns stand out. Bone is the only indication where in vitro culture before implantation is specified (7-14 days). Neural is the only one that names an immunosuppressive drug pair. Cardiac is judged by perfusion as much as by function, because vessel density sits beside ejection fraction among its success metrics.
Safety is where the protocol is most conservative. The pre-clinical checklist requires in vitro biocompatibility testing (ISO 10993), a tumorigenicity assay in nude mice, a biodistribution study, toxicology (acute, subacute, and chronic), immunogenicity assessment, an off-target differentiation check, and long-term genomic stability.
For tumourigenicity the specification lays out four layers:
Immunogenicity is handled by HLA matching for allogeneic products, the immunosuppressive regimen where applicable, and per-patient surveillance for anti-donor antibodies. The adverse events to watch are immune rejection, infection, tumor formation, ectopic tissue formation, and thrombosis. Monitoring starts dense and thins out:
Week 1: Daily
Week 2-4: Every 3 days
Month 2-6: Weekly
Month 7-12: Monthly
Year 2+: QuarterlyLaboratory biomarkers include a complete blood count, liver function (ALT, AST), kidney function (creatinine, BUN), inflammation markers (CRP, IL-6), and tumor markers (AFP, CEA) where applicable. First-in-human dosing uses a 3+3 design with four levels of three patients each.
| Level | Dose (cells) | Patients |
|---|---|---|
| 1 | 1 × 10⁶ | 3 |
| 2 | 5 × 10⁶ | 3 |
| 3 | 1 × 10⁷ | 3 |
| 4 | 5 × 10⁷ | 3 |
Dose-limiting toxicity is defined as a grade ≥3 adverse event, any serious adverse event, or death within 30 days. Long-term follow-up runs a minimum of five years for cell therapy (fifteen for gene therapy). It covers disease-specific functional outcomes, quality of life (SF-36, EQ-5D), imaging (MRI, CT, ultrasound), biopsy when indicated, and genetic stability by karyotype and CGH array. The iPSC-specific 15-year cancer surveillance above is the stricter requirement where it applies.
Vascularization is the integration step that most often decides whether an implanted construct survives. The specification models angiogenesis as VEGF binding VEGFR2 and signaling through ERK1/2 and Akt to drive endothelial proliferation. Tip cells follow the VEGF gradient, stalk cells form the lumen, and pericytes stabilize the vessel. Vessel density is counted per square millimeter, and healthy tissue is given as 200-400 vessels/mm². The technical detail of host-vasculature integration is cross-referenced to the tissue-engineering standard WIA-BIO-006 (Phase 3 §A.3). Within WIA-BIO-020 the concern is how engraftment is assessed and how failure is classified.
| Failure mode | Practical meaning |
|---|---|
| Engraftment failure | Cells do not persist or integrate in the host |
| Immune rejection | Host response destroys the graft; links back to HLA matching and immunosuppression |
| Off-target engraftment | Cells settle outside the intended site; links to biodistribution studies |
| Fibrotic encapsulation | Scar tissue walls off the construct and blocks perfusion |
The remaining implementation concern is the control plane that carries registration and clinical-administration events. The specification uses a 96-bit nonce, a 300-second skew window, and a 600-second seen-nonce cache to defend against replay. Audit-event records are signed at registration and anchored in a per-tenant Merkle tree. A chain break invalidates the consent envelope and triggers a forensic-review event, which ties the integrity mechanism directly to the consent chain begun at procurement. Cleanroom telemetry uses mTLS with per-suite monotonic counters.
The direction of the protocol layer follows from where its current envelopes are most demanding. The reprogramming envelope already offers non-integrating routes (Sendai virus, episomal plasmid, mRNA) alongside piggyBac, and it already requires transgene clearance. Pressure on tumourigenicity risk makes it likely that the residual-undifferentiated-cell assays and the genetic-stability checks will receive the most refinement. On the engraftment side, the combination of cell-tracker imaging, circulating donor-DNA signatures, and functional metrics suggests a move toward engraftment evidence that is quantitative and comparable across sites. The 15-year surveillance horizon for iPSC-derived products means that data continuity over decades, not just trial-length records, will shape how these protocols are operated.
Phase 3 treats a regenerative product as a chain of documented envelopes. Preparation fixes the donor, custody, passage ceiling, and freeze-thaw conditions. Differentiation and reprogramming fix the cocktails, the marker checks, and a D_eff target above 80%. Implantation fixes route, dose, scaffold, and the indication-specific success metrics for heart, nerve, bone, cartilage, and skin. Engraftment checks the result through vessel density, imaging, and biomarkers, and classifies failures. Safety wraps the whole chain with tumourigenicity assays, genetic-stability testing, HLA and immunosuppression planning, a thinning monitoring schedule, 3+3 dose escalation, and follow-up lasting from five years to well beyond fifteen. The integration score I ≥ 0.8 and the signed, consent-linked audit chain are what let one stage's output be trusted by the next.
These protocols are defined for a single therapeutic product moving through a single clinical pathway. The next chapter looks beyond that boundary. It takes up Phase 4 of the architecture, where the protocol exchanges described here connect with the wider ecosystem of other standards and the long-running programs that depend on them.