Erythrocyte Dinucleoside Polyphosphates and Time in Suspension among Revived Crisis-Era Hibernators: An Archive Study of 241 Revival-Day Samples

1. Introduction

We find that erythrocyte Ap4A and total Ap4N at the moment of revival rise with the time a person spent in suspension, by 6.4% and 5.6% per doubling of that time, and that the rise is small beside the elevation that revival leaves behind even after the shortest suspensions. The result bears on a question that every revived hibernator raises, the question of a duration at which deep low-temperature suspension stops being recoverable, but it does not answer it.

People who entered suspension in the Crisis Era have been revived since Year 1 of the Deterrence Era, and the Hibernation Register of the Academy's biobank in Beijing, whose inventory we use here, records 1,142 revivals to the end of Deterrence Era, Year 58 (Reconstituted Chinese Academy of Sciences, Archival Series, 59 DE). This study is written in Deterrence Era, Year 59 (DE 59), and no source it uses postdates that year. The Register records dates, protocols and outcomes in detail. It records almost nothing about the physiology of the suspension itself, and we take that to be a whole-body state at deep low temperature and nothing more specific.

Diadenosine 5′,5‴-P¹,P⁴-tetraphosphate (Ap4A) and the related Ap4N species are formed by aminoacyl-tRNA synthetases as a side reaction of amino acid activation, and specific hydrolases remove them (Reconstituted Chinese Academy of Sciences, Archival Series, 12 CE). Their steady-state levels rise in cells under oxidative, thermal and metabolic stress, and that behaviour is the reason that we treat them as candidate markers of cumulative cellular strain.

Suspension durations in the Register run from 10 to 243 years. A range of that width lets duration be handled as a continuous exposure and not as a pair of categories, so a dose-like relation can be estimated with a regression.

Our earlier work uses the same compound class as a comparator for the Trisolaran dehydration cycle, a cycle that is rapid and repeated (Chen, 58 DE). A single suspension of one human body is a different kind of event, and we draw on that earlier work here for contrast only.

2. Methods

The Register lists 1,142 persons revived from DE 1 to DE 58. Of these, 397 have a retained aliquot of packed erythrocytes from the revival-day draw, the first taken after core temperature had returned to normal, because earlier retention schedules discarded the rest. We excluded 96 aliquots for a recorded excursion above -70 °C in the biobank freezer log, 42 for recovery of the internal standard below 70% or an Ap4A signal under the limit of quantification, and 18 for a missing entry age or revival date. That leaves 241 persons.

Duration is the interval in whole years between the recorded vault entry and the revival date. For 50 persons who entered before Crisis Era, Year 40, the entry date comes from registers that were reconstituted after the Great Ravine, and we treat their durations as uncertain by up to three years. Covariates were age at entry in decades, sex, storage time from draw to assay in decades, and the rewarming protocol, first-generation up to DE 30 and second-generation from DE 31 (Haugen-Tan, 58 DE), with 121 persons in the second group. Persons who entered before Crisis Era, Year 120 were suspended in first-generation vaults, 140 persons in all, and the engineering record of the vaults supplies that date (Mbeki-Larsen, 53 DE).

Washed packed cells were deproteinised with perchloric acid and neutralised. Ap4A, Ap4G and Ap4U were quantified by liquid chromatography with tandem mass spectrometry against a stable-isotope-labelled Ap4A standard added before extraction, and ATP by ultraviolet absorbance after ion-pair separation (Valdemar and Soltani, 55 DE). Total Ap4N is the sum of the three species. Plasma protein carbonyls were measured by dinitrophenylhydrazine derivatisation and plasma HSP70 by immunoassay. Platelets store Ap4A in dense granules, so we estimated carryover from the archived blood counts, and it came to below 0.5% of cells. Ap4A hydrolase activity falls during frozen storage, so storage time enters every model (Ferreira-Qian and Okonjo-Reyes, 56 DE). Ninety-four Deterrence-Era blood-bank donors served as controls, and their cells were frozen under the same protocol and assayed in the same batches.

Each marker was log-transformed and regressed on log₂ duration with age at entry, sex, storage time, rewarming protocol and five batch indicators, leaving 230 residual degrees of freedom. The coefficient is reported as the percentage change per doubling of duration, 100(e^b - 1), with a t-based 95% interval. Five primary tests were corrected by the Holm method. The exceedance model, the hinge test and the subset analyses are secondary and are reported uncorrected. Departure from a straight line was tested with a hinge term at 64 years, six doublings from one year, a knot fixed in the analysis plan before any assay was run. A Breusch-Pagan test on the primary model gave p = 0.87. Exceedance of the control 95th percentile was modelled by logistic regression with the same covariates. The 87 persons with a dermal punch biopsy formed a tissue subset analysed with the same model on 76 residual degrees of freedom, and agreement between tissue and erythrocyte values was measured by Pearson correlation of log values with a Fisher interval.

3. Results

The 241 persons had a median suspension of 130 years, with an interquartile range of 72 to 187. Mean age at entry was 41.1 years, SD 10.8, and 115 were women. Duration correlated with storage time at r = -0.30, because the longest suspensions were revived later in the Deterrence Era.

Geometric mean Ap4A was 0.64 µmol per litre of packed cells in the hibernators and 0.41 in the 94 controls, a ratio of 1.56, 95% CI 1.44 to 1.69, p < 0.001, that is unadjusted for age. Within the hibernators, Ap4A rose by 6.4% per doubling of duration, 95% CI 1.7 to 11.2, p = 0.007, and total Ap4N by 5.6%, 95% CI 1.5 to 9.8, p = 0.007. Both kept an adjusted p of 0.034 after Holm correction. The Ap4A to ATP ratio, protein carbonyls and HSP70 did not reach significance. The ratio and HSP70 intervals are wide enough to hold effects as large as the two markers that did, while the carbonyl interval excludes any rise above 5.4% per doubling, so that null is the most informative of the three. Table 1 gives all five markers. Storage time and rewarming protocol also entered the Ap4A model. Each additional decade of storage was associated with 2.2% higher Ap4A, 95% CI -4.7 to 9.7, a direction that may reflect the loss of hydrolase activity, and revival under the second-generation protocol with 7.9% lower Ap4A, 95% CI -22.5 to 9.4, so neither could be dropped without risk to the duration estimate.

The control 95th percentile for Ap4A was 0.70 µmol/L. Of the 241 hibernators, 100 (41.5%) exceeded it, and the share rose from 37.5% in the shortest tertile to 45.0% in the longest, as Table 2 shows. The adjusted odds of exceedance rose by a factor of 1.46 per doubling of duration, 95% CI 1.09 to 1.96, p = 0.011.

Elevation was already common after the shortest suspensions, so duration cannot be the only source of it.

A straight line fit about as well as a plateau. Among the 53 persons suspended for 64 years or less the slope was 11.2% per doubling, 95% CI 0.5 to 23.1, and among the other 188 it was 2.9%, 95% CI -5.0 to 11.6. The test for departure from a straight line gave p = 0.34, so the data fit a plateau and a line equally well, and a larger sample might separate them where these numbers cannot.

Two sensitivity checks barely moved the estimate. Dropping the 50 persons with reconstituted entry dates left 191 persons and a slope of 7.8% per doubling, 95% CI 2.4 to 13.5. Adding an indicator for first-generation vaults, an indicator that correlated with log₂ duration at r = 0.80, left the point estimate at 6.3% but widened the interval to -1.1 to 14.4, p = 0.098. In the tissue subset the slope was 4.8% per doubling, 95% CI -7.8 to 19.1, p = 0.47, and tissue and erythrocyte Ap4A correlated at r = 0.49, 95% CI 0.31 to 0.64.

4. Discussion

Both dinucleotide measures drift upward with time in suspension, and the drift is small. A doubling of the interval, from 65 to 130 years for example, may add about 6% to erythrocyte Ap4A. We write within the Luo-Ji / Ye framework, and its axioms say nothing about the physiology of suspension, so no part of these results depends on them.

The elevation in the shortest suspensions is the larger feature of the data, and it may point to rewarming, an event that every person in the cohort underwent, as a source of dinucleotide stress that does not scale with time. The archive cannot show this. It holds no pre-entry sample for most of the cohort, and the Register lists entry-day aliquots for only a few entrants, so those aliquots might test the idea once they have been examined. The comparison with controls is also crude, since it takes no account of age.

Contrast with our dehydration comparators bears on how much to expect. In the terrestrial systems that we reviewed, Ap4N shifts accompany a state that is entered and left repeatedly (Chen, 58 DE). A single prolonged suspension need not behave in the same way, and the modest slope reported here could weigh against importing the comparator's expectation of large pool shifts into human suspension.

Nothing here locates a limit. The slope is positive, but the shape is unresolved and the vault-adjusted interval includes zero. More basically, the archive contains only persons who were revived. A suspension that exceeded any limit would not appear in it, and that survivorship, more than any choice of model, is what a prospective revival series would have to break before any limit could be located.

5. Limitations

Duration is nearly inseparable from vault generation in this archive, with r = 0.80, so the estimate of 6.4% per doubling could include an effect of older vault engineering that no covariate here can remove. The cohort is a single-centre convenience sample, the retention schedules that produced it were not random, and the 901 revivals without a usable aliquot may differ from the 241 in ways that we cannot see.

Protein carbonyls and HSP70 do not survive long frozen storage intact, so even the narrow carbonyl interval may understate a true rise, and the HSP70 null could be weaker still. The tissue subset is small, and its interval spans slopes from a decline to a rise larger than the erythrocyte estimate, so it may serve as no more than a rough check on the blood result.

These results do not show that suspension has a duration limit, and they do not show that it has none. They also leave open whether the elevation seen at revival fades afterwards. The measurement that could settle that part is a prospective paired draw from the next persons the Register records as revived, taken at revival and again at 1, 7 and 28 days, with Ap4A hydrolase activity measured in fresh cells and any surviving entry-day aliquots assayed alongside them.