# 40. Persistent gene regulation sustains injury after the exposure improves

**Base judgment: 3.3%; skeptical–optimistic sensitivity: 0.1%–30.6%.** Subjective probabilities of the causal proposition below, not effect sizes. Decimal places only show arithmetic. Reviewed 13 September 2026. [Probability method](probability_method.md) · [Collection index](README.md).

## The causal claim and its scope

Past diabetic exposure installs a persistent gene-regulatory state in peripheral neurons or supporting cells that continues to drive axon injury after current glycemia improves.

**Scope:** Exception: diabetes after a sustained improvement in glycemia, with a five-year injury/progression endpoint. The subgroup threshold refers to neuropathy cases in this post-improvement population. This is persistence of a harmful mediator, not the mere clinical fact that earlier control has lasting benefits.

**Causal role:** persistent amplification or failed recovery. Unless the scope specifies an exception, materiality means at least 20% less five-year net terminal loss in an identifiable subgroup representing at least 10% of incident DPN cases in the stated population. The target remains human peptidergic C-fibers when a study measures a broader endpoint.

## Reasoning from the biological requirements

Clinical metabolic memory can arise from already lost axons, glycated matrix or accumulated exposure without any persistent epigenetic program. This hypothesis additionally needs a durable regulatory state that remains after the initiating exposure is reduced, produces a harmful effector, and continues causing new loss. A stable methylation difference can be a record of exposure rather than its causal carrier. Reversing a gene mark must change injury through that state, not through unrelated drug effects.

## Probabilities of the necessary claims

Conditional prices are skeptical / base / optimistic. Each price after the first assumes *all earlier requirements are true*. Their order is an order for assessing joint belief, not necessarily a linear biological pathway. The cumulative column uses the base scenario. Temporal order, connected compartments and the causal prevention comparison are included in these requirements, without extra generic multipliers.

| # | Necessary claim | Conditional YES | Cumulative YES | Reason for this judgment |
|---|---|---:|---:|---|
| 1 | Diabetic exposure creates a persistent regulatory state in relevant peripheral cells. | 35 / **65** / 85% | 65.0% | Human blood epigenetics support exposure memory, not target-cell causation. |
| 2 | The state persists independently of current exposure and cell-composition changes. | 25 / **50** / 80% | 32.5% | Cross-sectional methylation cannot establish persistence within the same cells. |
| 3 | It sustains a harmful effector rather than marking history or supporting adaptation. | 20 / **45** / 75% | 14.6% | The relevant neuronal/glial effector has not been isolated. |
| 4 | The effector causes new peptidergic structural loss after glycemia improves. | 20 / **45** / 75% | 6.6% | Legacy damage and reduced reserve are strong alternative explanations. |
| 5 | Removing the persistent state would materially reduce that loss. | 25 / **50** / 80% | 3.3% | Selective human mediation and population magnitude remain unknown. |

## Evidence that moves the judgment

- Selected DCCT/EDIC blood monocytes/lymphocytes showed histone-modification associations with glycemic history and complications. This supports human exposure-related epigenetic differences, but is not a peripheral-neuron perturbation or proof of a neuropathy mediator. [Miao et al., 2014](https://pmc.ncbi.nlm.nih.gov/articles/PMC3994951/).

- The repository's public-evidence search did not identify a nerve structural contrast that cleanly separated a persistent glycation/GLO1 mechanism from ongoing exposure and acute rescue. Its GLO1 result demonstrated buffered acute detoxification, not persistence after washout or glucose normalization. [Prior correction](../../../findings/aug13_glycation_memory_glo1_epistasis_no_go.md).

## Why these odds and how uncertain they are

The broad fact of lasting consequences of diabetes is credible, but epigenetic mediation is a much narrower proposition. The price is low because exposure history, cell composition and irreversible prior damage explain many of the same observations. A compelling molecular mark needs a damaging downstream function and ongoing new injury, not just longitudinal stability.

The skeptical scenario treats marks as records and clinical persistence as legacy damage. The optimistic scenario assumes a reversible harmful state in a definable cell population. Existing human blood evidence offers little precision about which peripheral neuron/glial program, if any, carries the effect.

Conditional inputs: base **[65, 50, 45, 45, 50]%**; skeptical **[35, 25, 20, 20, 25]%**; optimistic **[85, 80, 75, 75, 80]%**. Products give the headline prices. The span measures sensitivity to these interpretations; it is not a statistical confidence interval.

**Numerical update example:** Public evidence linking a persistent regulatory state to a specific harmful effector after exposure normalization would strengthen requirement 3; a stable mark with no functional effect would weaken it. Changing only requirement 3 from 45% to 80% raises the whole price to **5.9%**; reducing it to 10% lowers the price to **0.7%**. These are hypothetical evidence updates. Even resolving that one requirement with certainty would leave a complete price of 7.3% if the others stayed unchanged.

## Overlap and boundary

Distinct from persistent [matrix glycation](03_extracellular_matrix_glycation.md), [senescence](33_senescence_sasp.md) and current [inflammation](19_tnf_tlr4_inflammation.md). The [metabolic-memory topic](../../../knowledge/topics/metabolic-memory.md) records related negative analyses; missing identification is not biological disproof.