Appraising Ex Vivo Culture and Molecular Cell State Against Defined Graft Endpoints in Human Myogenic Cell Transplantation

1. Introduction

Muscle satellite cells occupy the niche between the myofiber membrane and basal lamina and provide the principal stem-cell source for postnatal skeletal muscle repair [1]. Their quiescence, activation, self-renewal, and differentiation are shaped by molecular signals from the niche and by the physical and metabolic conditions encountered after isolation. Human satellite-cell preparations can generate donor-derived fibers in conditioned recipient muscle and can leave donor-derived precursors capable of later regenerative activity [2,3]. Therapeutic development is constrained by the small amount of starting tissue, which makes ex vivo expansion attractive. Expansion changes signaling, metabolism, adhesion, cell-cycle state, and population composition. The relevant unit is therefore the final administered preparation.

The central translational problem is that a molecular state favoring cell accumulation in culture may not support engraftment. A manipulation can increase yield, maintain a marker, or delay differentiation without improving repair. In mouse models, directly isolated satellite cells can outperform expanded cells under defined transplantation conditions [4]. Lineage and single-cell graft experiments further separate immediate fiber formation from precursor maintenance and self-renewal [5,6]. Human product assessment requires the same distinctions. Persistent myopathy requires donor activity after later injury, whereas reconstruction of a focal defect requires repair across the treated volume and an appropriate functional endpoint.

Previous reviews have summarized myogenic cell transplantation in muscle disease [7], examined links between engraftment and functional recovery [8], discussed recipient conditioning and graft heterogeneity [9], and surveyed culture strategies across mouse and human systems [10]. The unresolved issue is which molecular or culture-induced changes improve a defined human graft endpoint when dose, host context, and outcome differ. Selected human graft studies are therefore compared at the administered-product level. Production yield is separated from performance at a matched dose, marker retention from repair, donor persistence from later regenerative activity, and local histology from recipient function.

This endpoint-specific approach tests how far each molecular finding can be extended. More human-derived fibers at a matched final dose support progression to tests of later regeneration or function. Increased culture yield with an unresolved graft dose requires dose attribution before the performance of the administered population can be assigned. Table 1 summarizes the transplant conditions underlying these judgments. Figure 1 links the principal molecular and experimental findings to the next product-level comparisons.

2. Review Scope and Evidence Appraisal

We conducted a critical narrative review of tissue-derived human satellite cells and their myogenic progeny. Study selection was purposive. A targeted PubMed search was rerun on 30 October 2026 and identified no additional eligible primary graft report beyond the reserve-cell study identified in the 11 October search [11]. Among 82 cited sources, 28 primary human-cell graft reports were reappraised using their Results, Methods, figure legends, and relevant supplementary files. Supplementary File S1 provides the search records, source dispositions, extraction rules, and reproducible coding files. Because the original discovery history is incomplete, this review does not claim exhaustive or systematic screening.

The 28 reports yielded 39 question-specific comparison or demonstration records. For each record, we retained the cell preparation, administered dose when available, denominator, recipient conditions, timing, statistics, and source location. Records can share animals, cohorts, or controls and were not treated as independent effect sizes. We distinguished observed benefit, no detected improvement, untested function, and unresolved attribution. A nonsignificant result was not interpreted as equivalence, and values were extracted from source text and legends without graph digitization.

Donor provenance was assessed at tissue donation. The primary adult restriction required a documented entirely adult cohort or an explicitly adult graft source without contradictory information. A cohort mean age or a parental relationship contributed only to a contextual sensitivity analysis. Mixed-age cohorts were excluded from the primary adult restriction when adult results could not be separated. These rules evaluate applicability to adult-derived products and do not test a biological age effect. The accompanying code reproduces the sensitivity-analysis outputs but does not replace independent source adjudication.

3. Molecular and Cellular Determinants of Graft Performance after Ex Vivo Culture

3.1. Matched-Dose Evidence for p38 MAPK, Extracellular Matrix, and SETD7 Modulation

Administering the same number of cells in a within-study comparison excludes a larger delivered dose as the sole explanation for greater graft output. It does not equalize postgraft survival, spatial spread, population composition, or differentiation, each of which may mediate a useful culture effect. A fiber count therefore characterizes the administered population but does not isolate activity per surviving cell. Comparisons that fix the starting culture instead combine production yield with the behavior of the resulting preparation.

Charville and colleagues studied a donor cohort aged 35–82 years and used 10 μM SB203580 to reversibly delay differentiation. Fresh, untreated activated, and inhibitor-treated preparations were transplanted at 30,000 cells per graft into NOD scid gamma (NSG) tibialis anterior muscle injured with barium chloride 48 hours earlier. At four weeks, treated cells produced 221 ± 20 human-containing fibers per section, compared with 49 ± 14 for fresh cells and 21 ± 13 for untreated activated cells. Errors are standard deviations, and each treated-versus-control comparison had P < 0.001. Donor-derived sublaminar cells expressing the satellite cell transcription factor PAX7 were also observed [12]. Equal dose excludes a larger administered graft as the explanation. Supplemental images show human dystrophin contribution at 56 days, extending the observation window without testing a second injury or direct force.

Zhao and colleagues studied human satellite cells from donors aged 40–56 years. After a five-day comparison on decellularized extracellular matrix or uncoated surfaces, they grafted 25,000 cells from each condition into athymic nude tibialis anterior muscle injured with cardiotoxin one day earlier. At four weeks, matrix-cultured cells produced more human dystrophin-positive fibers and donor-derived PAX7-positive cells. Methods specify six tibialis anterior muscles per group [13]. Together with the p38 experiment, this is direct evidence that a culture intervention can improve the output of an equal-sized human graft.

SETD7 inhibition connects production with a human reporter endpoint. Judson and colleagues started cultures with 15,000 cells from donors aged 15–53 years and compared seven days of 4 μM PFI-2 with vehicle. Cells were reporter-transduced and delivered to NSG tibialis anterior muscle in a 1.2% barium chloride suspension. Bioluminescence was followed for eleven days [14]. Human cell accumulation and postgraft signal increased. The published Methods, Figure 7, and supplementary Figure S1–S5 of that report do not identify the final administered human cell number. Consequently, production and performance per administered cell cannot be separated. Matched-dose fiber, recovered stem cell, and force experiments elsewhere in the paper used mouse donors and cannot supply those human conclusions.

The SETD7 result calls for a human graft comparison with a stated, matched final dose and a parallel measurement of usable yield. By contrast, the p38 and matrix results already support an improvement in histological output of an equal-sized graft. Their next tests concern later regeneration or indication-specific function. Fiber counts from different host preparations cannot rank the two culture processes.

3.2. Cell Selection, Senescence, and Expansion History

Population composition can change graft output without an expansion intervention. Equal grafts of 500 caveolin-1 (CAV1)-positive or CAV1-negative human satellite cells produced different fiber outputs in an irradiated and injured NSG recipient, favoring the positive fraction [18]. The result supports selection of a functionally favorable population, but induction of CAV1 in an individual cell remains untested. An equal-dose culture benefit can similarly reflect enrichment of a favorable fraction or altered behavior of its constituent cells.

Conte and colleagues removed a senescent fraction from myotonic dystrophy type 1 (DM1) myoblast cultures using A1155463. Equal grafts of 200,000 cells from three patient lines produced greater human dystrophin-positive fiber contribution in NSG muscle (the source paper's Figure 6j, P = 0.014). Cardiotoxin preceded grafting by one day, and harvest occurred 21 days after injury, or twenty days after grafting [20]. The equal dose attributes the improvement to the treated population and excludes a larger administered graft as its sole explanation. Because the Results and Methods differ in exposure duration, a single exact treatment interval is not assigned. Selection against defective cells remains compatible with the outcome. Rejuvenation of every surviving cell was not demonstrated.

CAV1 selection favors a fresh fraction, while senolytic treatment removes defective cultured cells. Both results concern the retained preparation. Manufacturing assessment must also retain yield lost during selection.

Culture duration must be linked to the transplanted preparation. In the p38 study, a three-week growth curve and a two-week expansion followed by drug withdrawal addressed proliferation and reversible differentiation, respectively. The matched-dose transplantation comparison used 30,000 cells and a four-week endpoint, but its specific culture interval is not stated in the examined main and supplemental procedures. A separate GFP-labeling graft workflow began with ten days of expansion, adenoviral labeling, and more than three additional passages [12]. This supports graft activity after extended processing without assigning that duration to every experiment in the report.

Cooper and colleagues provide a within-donor example of a culture-history effect. Each freeze-injured immunodeficient muscle received 500,000 myoblasts. At eight weeks, 84% of remaining human nuclei were located within fibers after early infant culture, compared with 21% after extensive amplification of the same infant source. The 71-year-old donor population reached 37%, exceeding the late infant population [26]. These percentages describe nuclear location, not retention of that fraction of the injected cells. One donor at each age and three recipient mice per condition prevent generalization of an age effect. The infant comparison makes processing history a product-selection variable. Cross-age differences remain confounded by donor identity.

Human age comparisons also depend on the preparation and comparator. Corbu and colleagues studied cells from six donors aged 74–76 years against three pediatric controls aged four months to sixteen years. Growth, passage tolerance, and differentiation kinetics differed. This was not an older-versus-young-adult graft comparison [37]. Human tissue xenografts retained regeneration across donor ages extending into later life, preserving architecture and associated cells [38]. Serially passaged human cultures showed reduced fusion and several metabolic responses. Insulin responsiveness was endpoint-specific: glucose-uptake fold response did not change detectably, whereas glycogen-synthesis and oxidation responses declined [39]. These findings distinguish chronological age, culture history, and recipient-supported regeneration without assigning an intrinsic age ranking to final products.

Surface-marker work separates CD82/CD318-associated myogenic populations from CD201-associated interstitial cells and tests effects of CD82 perturbation on cell behavior [40]. Source identity still requires a graft-specific readout. Orbicularis oculi cells from an adult female cohort were induced for three weeks and compared with noninduced cells at 200,000 cells per nude-mouse tibialis anterior. At one week, CM-DiI-positive fibers were detected in induced grafts [31]. The reported recovery measure counts labeled fibers, not surviving donor cells. The source/induction result warrants human-specific donor identification and later functional testing before it supports lasting replacement.

3.3. Mechanobiology, Metabolism, and Manufacturing Scale

Mouse mechanisms motivate human culture candidates. Soft substrates preserved self-renewal and regeneration [41]. Combined soft support and transient p38 inhibition improved aged-mouse regeneration and strength [42]. Independent work linked altered p38 activity to self-renewal defects [43]. Compression and Piezo1 perturbation connect mechanical sensing to cell state [44,45]. Mouse strength findings do not supply a human graft endpoint.

Human myoblast media altered cell accumulation, marker expression, and in vitro fusion across passages [46]. Greater usable yield from a biopsy remains valuable even without improved function per administered cell. These production findings do not determine graft reserve or recipient force.

Insulin-like growth factor 2 (IGF2) perturbation altered proliferation, differentiation, and mitochondrial bioenergetics in human cultures [47]. PGC-1α overexpression improved early tissue formation in vivo [48]. These adjustable properties do not establish a metabolic threshold predicting reserve or strength across manufacturing routes.

CDK4 and telomerase expression produced clonal human myogenic lines for disease research [49]. Scalable production comparisons did not uniformly retain proliferation and differentiation [50]. Laboratory expansion remains useful, while clinical manufacture requires validation of the specific population delivered.

Microcarrier production provides a positive example of processing compatibility with a narrow inferential boundary. Cells from a thirteen-month-old donor grown on microcarriers or in static flasks were administered at two million per contralateral tibialis anterior in two conditioned SCID mice. Both routes produced human dystrophin-positive fibers at four weeks. Similar numbers were described without quantitative counts or an equivalence analysis [30]. The comparison is compatible with microcarrier manufacture and leaves equivalence to an adult product unresolved. Calpeptin-mediated fusion control was tested in culture, and the examined report does not identify a graft comparison of calpeptin-treated material. A clinical-grade flask process instead demonstrates animal-free production with human platelet lysate and defined release measurements [51]. Those measurements identify usable material, while matched-dose functional comparisons determine its repair claim.

Across culture strategies, molecular perturbation and product composition act at different levels. p38 inhibition and extracellular matrix culture improved matched-dose histological output [12,13], whereas SETD7 inhibition combined increased production with an unresolved final dose [14]. CAV1 selection and senolytic treatment changed the composition of the transplanted preparation [18,20]. Soft matrices and p38 signaling in mouse systems identify plausible state-control variables [41,42,43]. Compression and Piezo1 studies further implicate mechanosensing [44,45]. Together, these findings support a hierarchy from molecular state to product attribute to graft endpoint. Movement across levels requires a controlled human comparison.

4. Molecular Cell-State Markers and Donor Persistence as Incomplete Surrogates for Repair

4.1. Notch Signaling and the Dissociation between Cell State and Fiber Output

Sakai and colleagues treated expanded, previously frozen human myoblasts with Jagged 1 (JAG1) or Delta-like ligand 4 (DLL4) for 24 hours. MYOD decreased, and the PAX7-positive/MYOD-negative fraction increased, although human PAX7 expression did not increase. Each NOD-SCID tibialis anterior received 100,000 cells after cardiotoxin injury. At two weeks, no increase was detected in the maximum number of human spectrin-positive fibers per injected muscle. The JAG1 comparison used five muscles per condition, and the DLL4 comparison used three [16].

The brief Notch-ligand regimen changed a measured cell-state profile without a detected gain in human fiber output. The small graft groups do not establish equivalence, but the marker result alone does not justify selecting this treatment for fiber replacement. Sustained support from isolation is a distinct intervention that would require its own graft comparison.

Mouse loss-of-function experiments established a requirement for canonical Notch signaling in resting satellite cell maintenance [52,53]. A Notch-linked collagen V circuit further supported niche maintenance [54]. The associated 2025 correction replaced a misassembled receptor-screening image. The authors reported no change in interpretation [55]. These necessity experiments motivate culture strategies. They address a different question from whether brief ligand treatment is sufficient to improve an already expanded human graft.

4.2. Reserve Cells, Reporter Persistence, and Donor Compartment

Reserve cells are the nonfusing fraction recovered after two days of differentiation of expanded myoblasts. Their selection changes graft endpoints unequally. Laumonier and colleagues grafted 500,000 reserve cells or myoblasts into lacerated NOG gastrocnemius. At day twenty-one, reporter activity was 54.1 ± 5.2% versus 31.5 ± 4.8% of each graft's three-hour baseline (n = 12; P < 0.05). The spectrin-positive fiber count divided by human lamin A/C-positive nuclei was 27.1 ± 4.6% versus 23.1 ± 7.0% (n = 6; P = 0.6); errors are standard errors. The human PAX7-positive fraction was also higher [17].

Both groups in the 2017 study underwent reporter and histological evaluation. Higher late signal and a larger PAX7-positive fraction favor the reserve-cell route for further evaluation, but neither substitutes for a direct later-regeneration endpoint. The nonsignificant fiber-to-nucleus ratio cannot demonstrate equivalent absolute fiber production. Yield after generating and selecting the reserve fraction remains part of the manufacturing comparison.

A later reserve-cell study compares autofluorescence-high and -low fractions, a different comparison from reserve cells versus myoblasts. Equal grafts of 200,000 cells showed no detected difference in day-twenty-eight reporter signal or human PAX7-positive fraction. Cardiotoxin reinjury on day twenty-eight was followed by reporter imaging on day thirty-two [11]. The challenge was performed, but the reporter increase does not directly measure new fiber repair or recipient force. The Results describe nonsignificance while that report's Figure 6C legend reports P < 0.05. Donor ages are not reported in the examined main text. The study extends the tested trajectory while leaving direct adult-source later repair and equivalence between fractions unresolved.

A controlled Duchenne study directly separates donor protein contribution from strength. Twelve boys received 110 million donor myoblasts per biceps in each of six monthly procedures, with contralateral sham injections. Exon-specific staining confirmed donor-derived dystrophin in four recipients. One biopsy contained 420 positive fibers among 4,080 examined, or 10.3%, without a treated-arm strength advantage. No strength advantage was detected at the measured times through twelve months [33]. The biopsy fraction does not quantify repaired whole-muscle volume. Protein restoration supplies a necessary molecular observation for this strategy, but it cannot replace the physiological endpoint.

Longitudinal reporter measurements and tissue localization provide complementary information. Bioluminescence follows donor-associated reporter activity and is not a direct cell count. Human-specific DNA establishes human material in a sample, while staining locates nuclei in fibers or precursor compartments. Combined molecular and histological assays have been applied to human muscle xenografts [56]. This combination is particularly useful when increased signal might reflect more fused nuclei instead of a larger unfused reserve. A higher reporter signal can reflect different donor compartments, so normalization and tissue localization are needed before assigning a precursor-persistence claim.

Quarta and colleagues tested human cells after approximately one day in quiescence-supporting conditions. They followed equal grafts of 1,000 cells with reporter imaging for thirty days, with greater signal under optimized conditions [15]. Reporter activity integrates viable reporter-expressing material, proliferation, differentiation, and reporter regulation. The experiment supports brief preservation but leaves extensive human expansion untested. Its positive reporter result and the limiting Notch comparison involve different starting material, processing histories, and measured functions.

4.3. Quiescence, Senescence, and Reversible Activation

Serum-replacement conditions induced a quiescence-associated cell-cycle profile in expanded human myoblasts while retaining subsequent differentiation [57]. Retinoic acid suppressed differentiation and increased specification-associated gene expression in human progenitors [58]. These culture results identify controllable states. Their translational value depends on whether the resulting preparations retain productive repair or later precursor activity after delivery, which cannot be assigned from reversible differentiation alone.

Isolation itself begins the change in cell state. Mouse studies using in situ fixation or in vivo transcript labeling showed that dissociation can induce transcriptional responses before cells are placed in an experimental culture [59,60]. A human preparation described as fresh is therefore an operational category, not proof that every transcript retains its resting in vivo abundance. In a study of dual p38 and JNK inhibition during purification, the transplantation improvement was established in mouse cells, while human experiments supported improved recovery and a less activated transcriptional profile [61]. This species distinction matters when evaluating preservation from the earliest processing step.

Noncycling cells are not necessarily reversibly quiescent. Geriatric mouse satellite cells acquired a senescence program that impaired subsequent activation. Silencing p16INK4a restored activation and regenerative function in mouse experiments [62]. In human primary cultures, p16INK4a silencing increased BrdU incorporation and reduced senescence-associated β-galactosidase positivity [62]. The human arm did not test graft rescue. Basal autophagy helped prevent senescence in mouse cells, with similar observations in human geriatric samples [63]. Nutrient utilization also influenced SIRT1-dependent epigenetic regulation during mouse activation [64]. For a culture signal that suppresses proliferation, subsequent re-entry and differentiation are therefore necessary interpretive tests. Marker retention alone cannot distinguish preserved reserve from impaired activation.

Hypoxic culture improved later graft performance in a mouse myoblast model [65]. In human myoblast xenografts, delaying differentiation was associated with greater proliferation, distribution, and later fiber contribution [66]. These results motivate a precursor interval that permits expansion and spread before fusion. They do not establish an optimal timing sequence for durable adult human repair. The relevant unresolved comparison would hold cell dose and delivery constant while measuring both first repair and subsequent precursor availability.

5. Durable Regenerative Activity and Functional Recovery

5.1. Later Repair, Reinjury, and Secondary Transplantation

Durable repair is a separate product requirement from initial engraftment. Reinjury tests whether donor contribution can recur, whereas secondary transplantation tests whether graft-derived cells can be recovered and remain graft active. Fresh human satellite cells and expanded postnatal myoblasts have supported these endpoints under different conditions. The available experiments demonstrate later activity but do not constitute a controlled fresh-versus-expanded efficacy comparison.

Fresh-cell experiments define functions that an expanded product can be tested against. Xu and colleagues isolated CD56-positive/CD29-positive human cells that formed fibers and occupied the satellite cell niche [2]. Garcia and colleagues used lineage exclusion and CXCR4-based selection and recovered cells capable of secondary transplantation [3]. These are distinct preparations, not interchangeable standards. Their benchmark value lies in the functions demonstrated under specified recipient conditions.

Xu and colleagues demonstrated human contribution after reinjury of conditioned recipient muscle [2]. Garcia and colleagues recovered 200–1,400 human cells per four-mouse group ten weeks after initial grafts of 2,500–5,000 cells. The recovered population formed human fibers and occupied the satellite compartment in secondary recipients [3]. Recovery averaged approximately 7% of the starting graft and did not establish net expansion in the host. The authors considered persistent quiescent cells as an alternative explanation. Their five-month cryopreservation comparison also retained graft activity, with no significant fiber-count difference (P = 0.361). This supports the preservation route but does not prove equivalence or applicability to heavily expanded material. Reinjury and secondary transplantation demonstrate later competence but leave the self-renewing lineage history unresolved.

Expanded postnatal myoblasts also show evidence of later regeneration. Skuk and colleagues used cells from a 13-month-old infant and adults aged 39 and 43 years, with two to three million cells delivered to irradiated SCID muscles. Recovered cells formed human fibers after further culture and secondary transplantation, and human contribution was present after injury two months following the initial graft [22]. The reinjured muscles had a higher mean maximum human fiber count. However, the paired comparison did not reach the conventional significance threshold (P = 0.06). The report therefore supports later regenerative activity but does not demonstrate a quantitative advantage over uninjured grafts or fresh cells. Initial culture duration differs between Methods and Discussion. Product-specific replication at a defined dose and culture history is required before extending this finding to a durable-repair strategy.

These experiments keep both fresh and expanded preparations in consideration for continuing repair. A fair product comparison would specify donor provenance and final processing, match the administered dose and host conditions, and report initial and later donor contribution separately. Table 2 compares the distinct evidence supplied by reinjury, secondary transplantation, and recipient-force measurements.

5.2. Preservation, Bioengineering, and Disease-Corrected Products

Marg and colleagues transplanted human muscle fiber fragments after five to seven days of culture. The graft subset comprised six donors aged 44–64 years, and researchers observed human fibers and niche-associated PAX7-positive cells at three or seven weeks in locally irradiated NOG recipients [25]. The broader tissue cohort was aged 20–80 years. The preparation retained native structures and associated cells, and robust graft formation depended on recipient conditioning. This is a positive adult culture result for a fiber-associated preparation. It does not establish equivalent activity after purification and numerical expansion of the cells, because both the retained tissue support and recipient environment contribute to the tested route.

Preservation acts on different starting materials. Fresh human satellite cell purification and preservation protocols build on validated graft preparations [67]. Cultures derived from frozen biopsies instead address cell recovery from archived tissue [68]. A frozen biopsy, a cryopreserved sorted population, and a frozen expanded product differ in composition and processing stress. A preservation result supports the tested route. Changing the material requires a corresponding functional comparison.

Wu and colleagues demonstrated a physiological outcome with freshly isolated human cells in a 7 × 2 × 2 mm volumetric tibialis anterior defect. A decellularized matrix scaffold releasing basic fibroblast growth factor and insulin-like growth factor 1 was compared with a plain dECM scaffold carrying human cells. Recipient contractile performance improved at six weeks, with four replicates across three donor populations [24]. The comparator supports a benefit of programmed support in the human-cell combination. It does not isolate the force generated by human fibers or establish the necessity of human cells through a complete cell-free factorial comparison. Because the cells were delivered immediately after isolation, the result serves as a benchmark for an expanded product but provides no evidence that expansion preserves force-restoring capacity. Human loading dose and donor ages were not identified in the examined main text, Supplementary Information, or Reporting Summary.

Improved vascular and neuromuscular junction-associated features accompanied the scaffold's recipient-force benefit [24]. Their separate causal contributions were not isolated. For an expanded product, the informative comparison is fresh versus expanded human cells at matched loading in the same growth-factor-releasing scaffold and recipient model. Component controls would address the added human-cell contribution. These comparisons test whether the supported reconstruction route preserves the force gain without attributing it to expanded cells or human fibers alone.

Genetic correction adds a disease-specific manufacturing requirement. Escobar and colleagues grafted SGCA-corrected cells from a 13-year-old male patient into irradiated NSG tibialis anterior muscle. Two injections delivered 25,000 cells each, for a total of 50,000 cells per muscle. Irradiation preceded grafting by two days, and evaluation occurred at 19 days. Corrected protein was present in human-derived fibers, and human PAX7-positive compartment cells were observed [21]. The examined human graft arm demonstrates editing-compatible activity. Superiority over a matched unedited graft was not tested in this arm. The pediatric donor, short observation window, and absence of a direct force endpoint limit extrapolation to an adult product that supports continuing repair.

mRNA delivery of gene-editing tools subsequently enabled editing in primary human cells while retaining tested myogenic attributes [69]. Template-free repair of a CAPN3 mutation supplied another disease-specific correction route [70]. These results should be connected to their individual endpoints: molecular correction, retained culture phenotype, and in vivo repair are different validation steps. For an autologous product, the final corrected and expanded preparation is the relevant unit, including any selection, freezing, and recovery that occur between editing and delivery.

Engineered human myobundles can generate measurable electrically evoked force and drug responses [71]. Engineered muscles from pluripotent-cell-derived cells have also recovered contractility after injury [72]. These assays evaluate tissue formation under controlled conditions but do not reproduce the delivery and recipient barriers of tissue-derived grafts. Mini-MEndR uses a human myoblast template with mouse muscle stem cells. A regenerative response in that mixed system cannot be attributed solely to adult human precursors [73]. Nonhuman engineered-muscle implantation studies further demonstrate integration and contractile maturation [74]. Each platform provides a defined functional assay but cannot substitute for transplantation testing of the intended product.

Macroporous matrix hydrogels improved mouse cell expansion and functional recovery, with scale-up examined in rats [75]. Aligned constructs containing murine myoblasts and endothelial cells improved vascular organization and perfusion after implantation [76]. These studies motivate delivery architecture for human products. Their donor species and multicellular composition prevent attribution of an equivalent force-restoring effect to expanded adult human satellite cells.

Histological contribution by a corrected product, recoverable donor precursors, and whole-recipient force are distinct results. None substitutes for testing the final preparation under the recipient conditions of its intended use. Figure 2 summarizes how donor provenance and product processing narrow the conclusions available from the graft literature.

6. Delivery Geometry and the Recipient Microenvironment

6.1. Dose Distribution and Early Donor Retention

Spatial coverage can change the preferred intervention. Garcia and colleagues found no significant fiber-count improvement when a single-site fresh-cell dose rose from 1,000 to 5,000 or 10,000. At the same total dose of 2,000 cells, multiple sites yielded a reported mean of 155 human fibers versus 40 after one-site delivery [3]. This finding supports evaluating spatial distribution before further single-site dose escalation in this model. It establishes neither a universal saturation dose nor equivalence across the tested doses.

Bentzinger and colleagues treated primary human myoblasts with Wnt7a for three hours, washed the cells, and transplanted 100,000 cells into injured, FK506-immunosuppressed mdx muscle. The source paper's Figure 7 specifies assessment at 21 days. Human fiber contribution, fiber size, and maximum cluster distance increased [19]. This matched-dose result tests spatial deployment after a short pretreatment and does not address preservation during prolonged expansion. The strength experiments used mouse donor cells. Its interpretation complements the p38 and matrix results while identifying distribution as another possible source of greater human fiber output.

Garcia isolated the effect of delivery geometry at a fixed total cell dose. Wnt7a pretreatment altered a different variable before transplantation. Both findings make spatial coverage a plausible development target, but neither establishes whole-muscle benefit from an expanded human product.

AG490 provides a direct comparison of delivery routes. Each graft contained 500,000 expanded myoblasts from an adult donor. Local coinjection increased day-five residual radiolabeled DNA from 5.5% of day-zero signal in controls to 8.0%, 7.3%, and 7.8% at the tested concentrations. Twenty-four-hour pretreatment and systemic recipient treatment did not improve this endpoint. The highest pretreatment concentration reduced it [27]. The readout is retained radioactivity, not a direct live-cell count. Local coinjection is the supported route for an early retention claim, and later repair remains to be tested.

Dextran sulfate defines a temporal boundary. In pediatric myoblast grafts, bioluminescence normalized to each graft's three-hour baseline was 77.9 ± 10.1% versus 39.4 ± 4.9% at day one (P = 0.0009). At day three, normalized signal was 38.1 ± 8.5% versus 15.1 ± 3.4% (P = 0.01). Errors are standard errors, and n = 7. The day-seven difference was not significant (P = 0.06), and both signals reached assay background at day fourteen [28]. Early reporter protection is supported. The result establishes neither lasting function nor death of every donor cell at the limit of detection.

6.2. Host Conditioning and Stromal Support

The grafted cell state is expressed within a recipient niche that can amplify or suppress its regenerative potential. Fresh-cell benchmark experiments used immunodeficient muscle with acute injury and, in several cases, irradiation [2,3,18]. Within-study comparisons remain informative under these conditions, but irradiation changes competition from endogenous precursors and acute injury supplies a regenerative opportunity. Similar four- or five-week fiber endpoints therefore do not make the p38, matrix, and CAV1 experiments equivalent dose-response assays.

Striedinger and colleagues administered recipient Bindarit for seven days after grafting 2,000 fresh human cells into irradiated, bupivacaine-injured NSG tibialis anterior muscle. Human fiber formation and size improved at three weeks [23]. Because the drug was delivered to the recipient, the experiment identifies an environmental intervention and does not test improved ex vivo maintenance. Its broad pharmacological action leaves several donor and host processes within the effect. The small fresh-cell dose and conditioned host also distinguish it from expanded myoblast grafts administered at hundreds of thousands of cells.

A dominant-negative activin type IIB receptor expressed in human myoblasts increased donor-derived dystrophin-positive fibers in a dystrophic recipient [77]. Systemic delivery of a soluble receptor targeting related ligands also improved human myoblast transplantation [78]. The two approaches intervene in donor signaling and the broader recipient environment, respectively. They support modulation of postgraft growth and differentiation without isolating a shared donor-intrinsic mechanism.

Mouse studies identified fibro-adipogenic progenitors that support myogenic differentiation after injury and mesenchymal progenitors that can contribute to ectopic fat [79,80]. Host stromal response can therefore influence repair without generating donor muscle fibers. This supports separating myogenic purity from associated-cell support and from the fibrosis or fatty replacement encountered in a recipient. It does not establish which stromal process explains a particular human culture comparison.

6.3. Anatomical Scale and Clinical Interpretation

Local histology and whole-muscle function sample different extents of repair. A biopsy near an injection tract can show substantial donor contribution while representing only a small part of the treated muscle. Conversely, a whole-muscle measurement can dilute a useful local effect. This is an interpretive consequence of spatially restricted delivery, not a measured explanation for a particular trial. Dose comparisons should therefore retain the treated volume, injection distribution, and anatomical scale wherever those parameters are available.

Adult oculopharyngeal muscular dystrophy (OPMD) transplantation links donor-site selection to a local therapeutic goal. Cultures from affected cricopharyngeal muscle showed premature arrest, whereas clinically unaffected donor muscle was a more favorable source [81]. In the subsequent 12-patient study, autologous myoblasts underwent 21 days of good manufacturing practice expansion, and a median of 178 million cells was delivered during cricopharyngeal myotomy [29]. Swallowing assessments showed favorable changes, but pharyngeal propulsion was unchanged overall at two years. The study had no myotomy-only comparator, and the nonrandomized dose association does not isolate a causal cell benefit. The study supports adult manufacturing and local delivery feasibility and illustrates why procedural improvement and cell-mediated restoration of muscle function require separate attribution.

Earlier controlled studies contain both limiting and favorable findings. Karpati and colleagues delivered 55 million father-derived cells after 16–18 population doublings to each of eight treated biceps. Three recipients showed a statistically significant relative treated-side force advantage, and two also increased from their own baseline. No perceptible daily-function benefit was reported [32]. Low overall efficiency therefore coexists with individual local responses. In the San Francisco study, ten boys received 100 million cells in one tibialis anterior and placebo in the other. Force increased in both legs, beginning after cyclosporine treatment and before implantation, without a detected between-leg difference [34]. The timing limits attribution to cells but does not establish drug causality without a drug-free comparison.

In a five-boy triple-blind trial without immunosuppression, HLA-matched sibling cells were clonally expanded through approximately 24 divisions. Treated biceps received 102–240 million cells. The contralateral biceps received sham injections. No strength advantage was detected at two to six months. At sixteen to eighteen months, the reported mean decline was larger on the cell side than on the sham side (29.2% versus 5.0%; P < 0.05) [35]. Early tibialis anterior biopsies showed 0–36% dystrophin-positive fibers, generally with weak expression. Later biceps biopsies sampled different muscle and could not locate the exact injection sites. These observations cannot establish longitudinal protein loss within one muscle. Serum reactivity and complement-dependent lysis in two recipients support an immune concern, while the cause of strength loss and antigen identity remain unresolved. Donor ages were not reported.

A genetically identical twin case further shows anatomical dependence. A fourteen-year-old donor supplied approximately 704 million expanded cells for biceps and extensor carpi radialis grafts. Reported wrist-extension gains at four observations from one to six months exceeded pretreatment intervals, while elbow flexion did not improve [36]. This is positive local evidence in a single pediatric case, not a group estimate of efficacy. The muscle, baseline variability, and paired control define the claim. Taken together, these clinical reports preserve possible local responses while requiring a controlled functional difference before attributing therapeutic benefit to a manufactured product.

The clinical studies show why manufacturing feasibility, local donor expression, and functional efficacy require separate conclusions. When myotomy and cells are delivered together, an improvement in swallowing cannot establish the added benefit of cells without an appropriate surgery-only comparison. Local donor protein or force responses should be interpreted at the measured anatomical scale.

7. Applicability to Adult Donor-Derived Products

Applying the documented donor-provenance restriction retained seven of 28 reports and 13 of 39 question-specific records. Five reports contained mixed-age sources, four were pediatric, and 12 did not establish an entirely adult source for the relevant comparison. Table 3 shows which conclusions survive this restriction. These counts describe the selected corpus and should not be interpreted as the prevalence of adult evidence in the field or as evidence of a biological age effect.

The matched-dose p38 and matrix findings remain adult-source culture results [12,13]. CAV1 selection retains a human histological advantage, although its adult RNA-sequenced cohort cannot be assigned to the separately described graft arm [18]. The orbicularis induction study remains limited by adult-source feasibility and dye-based identity [31]. The separable arm from the 71-year-old donor further showed donor nuclei within fibers after expansion [26]. That descriptive arm is retained without counting the mixed-age comparison as an adult-only age effect. Adult restriction therefore supports selected graft-active cultured preparations but does not identify a superior manufacturing route.

Garcia's report [3] still supports adult-source secondary transplantation and fixed-dose distribution findings. Its primary source explicitly describes adult tissue and adult-source serial grafts, although individual arm ages are unavailable. The mixed-age Xu and Skuk reports remain useful demonstrations of later human activity [2,22]. They do not separately establish an adult expanded-product result in this extraction. The 2026 reserve-cell challenge also lacks confirmed adult provenance [11]. The broader conclusion that both fresh and expanded human preparations can retain later activity consequently survives, while its adult-only reach narrows.

The adult AG490 route comparison retains its early-window benefit [27]. Dextran sulfate and microcarrier manufacture remain pediatric evidence [28,30]. Wu's human-cell scaffold provides a physiological finding but lacks donor demographics, so it cannot support a confirmed adult-source rescue [24]. Adult OPMD manufacture and combined clinical treatment remain supported. The surgery-only comparator is absent [29]. These distinctions locate an unmet adult-product claim without converting unconfirmed age or an unmeasured function into biological failure.

The contextual sensitivity adds the reserve-cell cohort reported only by mean age and the father-derived Karpati grafts [17,32]. It retains nine reports and fifteen records without verifying individual donor ages. The Notch observations remain outside the primary restriction [16]. Mendell, mixed-age Miller, and the pediatric twin case also retain their broader human observations [33,34,36]. This separation makes the adult applicability of a finding explicit while preserving broader human evidence for candidate selection. Unknown age requires better source documentation, not an inference of biological failure.

8. Molecular-to-Functional Translation and Evidence Gaps

8.1. Matched-Dose and Component-Controlled Comparisons

The next experiment should isolate the unresolved variable in a defined final product (Table 4). Matched-dose histological gains warrant later functional testing. Increased yield with an unknown administered dose requires a dose-controlled graft comparison. A fixed-dose distribution effect calls for spatial-delivery testing, whereas a force result from a cell-scaffold combination requires component controls. Culture, delivery, and host support can interact, but the present evidence neither ranks these processes across studies nor validates a clinical treatment algorithm.

Recent multi-omic analysis identified promoter- and isoform-specific NFIX programs associated with quiescent, committed, and regenerative myogenic states. Corresponding promoter-accessibility patterns were also examined in human satellite-cell and adult-muscle datasets [82]. These data provide a molecular state map, but the study did not test an ex vivo product after transplantation. A candidate state marker becomes translationally informative only when it predicts a defined graft endpoint under controlled dose and host conditions.

8.2. Indication-Specific Validation

A product intended for persistent myopathy needs evidence of donor activity after later injury. Adult-source fresh cells provide a serial-transplant benchmark [3]. Expanded human preparations have also shown later activity, but their adult-specific applicability remains unresolved [22]. Testing defined adult final products under a common later challenge would separate initial graft output from subsequent donor activity.

In a focal defect, fresh human cells combined with a growth-factor-releasing matrix improved recipient force compared with cells in a plain matrix [24]. The result supports testing the complete delivery construct. It does not isolate the added effect of human cells or establish that an expanded adult product will work in the same construct. Matched loading, component controls, and donor-age documentation would address those questions.

Disease-corrected autologous cells must retain the relevant function after editing, expansion, selection, preservation, and delivery. Molecular correction and graft histology validate particular steps, not the entire product. The adult oculopharyngeal muscular dystrophy study established manufacture and combined treatment with myotomy [29]. A surgery-only comparator is needed to attribute a swallowing benefit to the transplanted cells.

8.3. Limitations of the Evidence Reappraisal

This reappraisal is limited by purposive selection, a partly unrecoverable discovery history, shared cohorts, incomplete donor-to-arm mapping, heterogeneous recipient conditions, and source-reporting gaps. Supplementary File S1 reproduces the coding and aggregation, but no second adjudicator independently re-extracted the primary-source data. We make no field-wide effect estimate or cross-product efficacy ranking.

9. Conclusions

Human graft studies support a limited set of molecular and cellular strategies. Within the reviewed evidence, no single state marker was shown to predict durable function. In documented adult donor sources, transient p38 mitogen-activated protein kinase inhibition and extracellular matrix culture improved histological graft output at matched final doses. Cell selection and senolytic removal altered the performance of retained populations, whereas Notch-associated marker changes alone did not predict greater fiber production. Evidence for later donor activity exists in fresh and expanded preparations, but durable adult-source functional recovery remains insufficiently tested. Future studies should define the final cell product, report production yield and administered dose separately, control delivery geometry and host support, and measure both later regenerative activity and indication-specific function. Molecular modulation becomes translationally informative when it is connected to the endpoint that a human myogenic therapy is intended to achieve.