Delayed Tolerance in Non-Autologous Bioware Grafts: A Prospective 12-Month Marker Cohort of 44 Recipients in Three Ninsei Black Clinics, Enrolled 2043–2045

1. Introduction

Bioware here means engineered or cultured biological tissue implanted to replace or extend a recipient's own organs, as distinct from hardware implants such as retractable blades, lens replacements or cranial storage. In Chiba City this work is done mostly in the black clinics of Night City and Ninsei, alongside the nerve-splicing and organ trade for which the city is known. Replacement livers and pancreases are among the most commonly purchased grafts, and some are engineered to change how the recipient metabolises particular drugs. The tissue comes from vat culture, from engineered lines sold by corporate suppliers, and from the less well-documented channels of the trade itself.

Licensed transplant medicine works under a well-described model in which grafts are kept under continuous immunosuppression. Acute rejection episodes occur in a proportion of recipients, and many resolve with treatment. The licensed literature says little about what happens when no immunosuppression is given at all, because licensed centres do not operate that way. The black clinics do. Consortium case reports and earlier registry summaries describe recipients whose marker panels rose sharply in the second month and then fell again without intervention. Those reports were uncontrolled and retrospective, and none followed the failures closely enough to show whether the rise carried prognostic weight.

Outcome work on unlicensed augmentation has shown that rejection-type complications dominate the failure profile of black-clinic procedures. In one case series of combined blade and lens implantation, the single clinic operating a written post-operative monitoring schedule recorded the fewest rejections; that difference did not reach significance, and the authors put a common schedule forward only as a proposal for prospective comparison. Rejection surveillance remains an open practical question for the sector. The present study followed a prospective cohort of bioware graft recipients through the first post-operative year, with three aims: to describe marker trajectories under definitions fixed in advance, to relate trajectory to 12-month outcome, and to identify the earliest point at which a trajectory can be read with prognostic value.

2. Methods

Recipients were enrolled between March 2043 and August 2045 at three cooperating clinics in Ninsei, with follow-up running to 2046. Each clinic contributes to the Chiba City Black Clinic Case Registry, whose consent protocol governed enrolment: recipients agree to anonymised follow-up and blood sampling. The clinics supplied graft provenance from their procurement records, which one of us (Y.B.) audited against surviving supplier batch labels. Adults receiving a first non-autologous bioware graft were eligible, and recipients given any immunosuppressive agent in the six months before grafting were excluded. The final cohort of 44 comprised 11 hepatic replacements, 9 pancreatic replacements, 14 composite dermal-muscle grafts and 10 nerve-splice conduits. On the basis of supplier documentation, 29 grafts were classed as engineered provenance, meaning tissue from commercially engineered lines, including Maas Biolabs lines, with a documented antigen-modification step. The other 15 were classed as unengineered: generic vat culture or trade-sourced tissue with no such record.

The marker panel is a composite 0–100 score adapted from an earlier low-resource system and first validated in unlicensed graft monitoring by one of us. It combines four equally weighted components: donor-specific antibody titre, the graft-derived fraction of cell-free DNA, the proportion of activated circulating T cells, and a pair of interferon-inducible chemokines. Samples were drawn at weeks 1, 2, 4, 6, 8, 10, 14, 20, 28, 40 and 52. Component thresholds and a composite threshold of 40 were fixed before enrolment closed. A recipient had an active marker profile at a visit when the composite score exceeded that threshold.

Four trajectory classes were specified before any outcome was adjudicated, and each is defined on recorded marker behaviour alone. Early rejection: an active profile at any visit in weeks 1–6 together with graft loss, or a graft site no longer sampleable, by week 8. Stable tolerance: no active profile at any visit across the 12 months. Delayed tolerance: an active profile at week 8 or week 10 with a composite score below threshold at week 14. Persistent activation: an active profile at week 8 or week 10 with a score still above threshold at week 14. The rules were applied in that order of precedence, so a recipient meeting the early-rejection definition was not reconsidered under the later rules; the classes are therefore mutually exclusive. Week 14 status was the sole criterion separating delayed tolerance from persistent activation; readings from week 20 onward were recorded but took no part in assignment. Three patterns lie outside the rules — activity confined to weeks 1–6 that resolves with the graft intact, graft loss or an unsampleable site by week 8 without prior activity, and first activity at week 20 or later — and any would have been reported as a residual class. None occurred, and all 44 recipients fall into one of the four classes. All recipients active in weeks 8–10 were sampled at week 14. Graft outcome was function at 12 months, adjudicated from clinic follow-up notes by both authors independently, without blinding to marker data; they agreed on 43 of the 44 cases; the remaining case was resolved by consensus.

Two analyses were specified in advance. To test whether the week 8–10 marker level itself discriminated outcomes, peak composite score in that window was compared between later successes and later failures with the Mann–Whitney U test, a rank-based comparison chosen because the peak-score distribution is right-skewed. To test the resolution signal, the sample was restricted to the 37 recipients with an active week 8–10 profile, and week 14 status (resolved against persistent) was related to 12-month failure by logistic regression. Because one cell of the resulting table was small, the model used Firth's penalised likelihood, and the association was cross-checked with Fisher's exact test. Provenance comparisons were exploratory, outside the lodged plan, and used Fisher's exact test; proportions carry Wilson 95% confidence intervals. The analysis plan was lodged with the registry before outcome adjudication.

3. Results

At 12 months, 31 of 44 grafts (70.5%; 95% CI 55.8–81.8%) were functioning and 13 had failed. Of the 44, 28 followed a delayed-tolerance trajectory, 9 persistent activation, 4 early rejection and 3 stable tolerance. Among the 31 successful grafts, 26 (83.9%; 95% CI 67.4–92.9%) were delayed-tolerance cases; the remaining five successes were the three stable-tolerance grafts that never crossed threshold and two persistent-activation grafts that survived despite unresolved markers. The 13 failures comprised the four early-rejection grafts, seven of the nine persistent-activation grafts, and two delayed-tolerance grafts that failed later in the year despite resolved markers. The table gives the full cross-classification, with row and column totals summing to 44.

Peak week 8–10 marker score did not separate the outcome groups. Among the 37 recipients with an active profile in that window, the 28 whose grafts were functioning at 12 months had a median peak composite of 58 (IQR 51–66) and the 9 whose grafts failed a median of 62 (IQR 54–71); the distributions overlapped substantially (Mann–Whitney U, p = .48). This is unsurprising, since every recipient in this subsample crossed threshold in that window by definition; with 9 failures the comparison would in any case miss a modest difference in level.

The prognostic information lay in what happened next. Within the same 37 recipients, markers resolved by week 14 in 28 and persisted in 9. Failure occurred in 2 of the 28 who resolved (7.1%; 95% CI 2.0–22.6%) and in 7 of the 9 in whom markers persisted (77.8%; 95% CI 45.3–93.7%). The Firth-penalised odds ratio for failure given persistence was 31.8 (95% CI 4.3–236.0; p < .001), agreeing with Fisher's exact test (p < .001). Week 14 was the earliest visit at which resolution could be read, both groups being, by definition, active at week 8 or 10.

In exploratory analysis, delayed tolerance tracked graft provenance. Of 29 engineered-provenance grafts, 22 (75.9%; 95% CI 57.9–87.8%) followed the delayed-tolerance trajectory, against 6 of 15 unengineered grafts (40.0%; 95% CI 19.8–64.3%; Fisher's exact p = .026). The provenance difference in 12-month outcome pointed the same way but was weaker: 23 of 29 engineered grafts were functioning at 12 months (79.3%; 95% CI 61.6–90.2%) against 8 of 15 unengineered grafts (53.3%; 95% CI 30.1–75.2%; Fisher's exact p = .09). Provenance and trajectory are correlated here, and the cohort is too small to estimate their contributions separately.

4. Discussion

Under black-clinic conditions, a large share of bioware grafts that ultimately succeed pass through a phase of measurable immune activity in the second month and then quiet down on their own. Licensed practice would treat that activity and would rarely observe its untreated course, immunosuppression being continuous there. Resolution of treated acute episodes is already recorded; what is new is that resolution without pharmacological support is common in this graft population and was the modal path to a functioning graft.

Earlier Consortium case reports treated the second-month rise itself as the warning sign, and our data do not support that reading. Marker presence in weeks 8–10 fails to separate the outcome groups: successes and failures both show it, and peak level did not discriminate detectably. The signal is carried by week 14 status, resolved below threshold or persistent above it, and persistence at that visit was strongly associated with failure. Any monitoring protocol built on this work must key on the week 14 reading.

A mechanism can be offered, though the cohort cannot test it. Delayed tolerance was concentrated in engineered-provenance grafts, including tissue from suppliers documenting an antigen-modification step. One plausible account is that engineered tissue down-regulates its surface antigens over the first weeks in the host, so that initial immune recognition, seen as the week 8–10 activity, is not sustained once antigen presentation falls, leaving time for adaptation to complete. Maas Biolabs' documented antigen-modification work describes this kind of down-regulation in its engineered lines. Our data fit that pathway without establishing it, and provenance may simply mark better-documented tissue from better-resourced clinics.

For clinics, the implication is a question for prospective testing rather than a protocol ready to adopt. If the week 14 signal holds up in a new cohort, clinics without immunosuppressive capacity could distinguish adapting grafts from failing ones and concentrate scarce interventions on the latter. That is the harm-reduction logic under which a common rejection-monitoring schedule has been proposed for unlicensed augmentation surgery, where the case likewise awaits prospective comparison. Moving from this cohort to a protocol requires validation with decision thresholds committed in advance.

5. Limitations

The cohort is small, especially in the failure stratum: the persistence-failure estimate rests on 9 recipients, with a correspondingly wide interval. The registry draws only from three cooperating Ninsei clinics, so grafts done elsewhere in Chiba and across the BAMA sprawl are unrepresented, and provenance classing depended on procurement records that the trade does not always furnish. Grafts that fail after 12 months are not captured. Outcome adjudicators saw the marker data, although function at 12 months is a fairly hard endpoint.

Conditioning is a further concern: the resolution analysis rests on survival to the week 14 draw, since a graft lost before then can have no week 14 status. No recipient in the restricted sample was lost that early, so no case was dropped here, but in a larger or sicker cohort the same rule would exclude early losses and bias the resolved group toward better outcomes. A prospective design should handle such losses explicitly, for instance by counting them as failures in a sensitivity analysis. Most seriously, the resolution signal was identified and tested in the same cohort, so it remains a hypothesis awaiting prospective validation; withholding intervention on the strength of an expected resolution should not be practised until that validation exists, since a mistaken expectation could cost a graft that intervention would have saved.