Axonal transport

In adults with type 2 diabetes diagnosed within five years and no clinical DPN at entry, a persistent defect in neuronal cargo transport precedes physical terminal loss. Normalizing the transport machinery defined below would correct a specified local cargo imbalance and avert at least 20% of five-year net terminal loss, in a baseline-defined group accounting for at least 10% of reference-care incident DPN.

The target is identified human peptidergic unmyelinated sensory neurons, including their somas, peripheral axons and cutaneous terminal arbors. Mixed sural nerve, autonomic fibers, myelinated sensory axons and cultured embryonic neurons inform plausibility without becoming this endpoint.

Structural outcome. Track physical arbors with stable anatomical and molecular subtype identification. Five-year net loss is destruction minus replacement, divided by baseline terminal amount, with equal person weighting. Newly generated terminals subsequently destroyed contribute to both quantities. Define:

B = (mean reference net loss − mean net loss with Q) / mean reference net loss.

The reference mean must be positive. Reference care is contemporary diabetes care without Q. Changes in pain, conduction, reporter intensity or mitochondrial motion do not substitute for physical terminal preservation.

Finite groups. Candidates are all eligible adults, males, females, baseline BMI ≥30 kg/m², baseline HbA1c ≥7.5%, and a transport-positive entry group. The last has the transit abnormality below in at least 10% of mapped neurons during a 30-day entry observation. No extra intersections or future-responder groups qualify.

Incident DPN is new bilateral length-dependent peripheral nerve signs with objective small-fiber loss or abnormal conduction, excluding another cause. Coverage is C = P(group membership | incident DPN under reference care). The same group, cargo branch and connected neurons must satisfy the chain.

Cargo branches and direct measurements. The finite alternatives are mitochondria, ligand-bound NGF–TrkA signaling endosomes, closed autophagosomes, NMNAT2-bearing vesicles, and packets carrying actin, tubulin or neurofilament proteins. Within the last branch, actin, tubulin and neurofilament are separate measured alternatives. This is not an unrestricted search for any cargo discovered after observing benefit.

Track complete trajectories, including pauses, reversals, docking, cargo exchange and disappearance. For mitochondrial and protein delivery, measure first arrival from the soma into the mapped distal arbor. For NGF–TrkA endosomes, measure terminal uptake to somal arrival. For autophagosomes, measure exit from the distal arbor to a degradative lysosome, distinguishing local fusion from retrograde delivery. Local completion without a long journey counts as successful clearance, not failed transport.

The primary quantity is the probability that an individually marked cargo cohort completes the defined journey within the matched nondiabetic median completion time. Normalize by the amount entering that journey, not the amount synthesized by the whole neuron. Record cargo that disappears before arrival as a competing fate, not an infinitely slow moving particle. Resolve motor-dependent passage separately from degradation and local capture.

A qualifying transport abnormality is a completion probability at most half its normal value and below the fifth percentile of the normal person-level distribution. Use age-, sex-, path-length- and anatomical-site-matched nondiabetic references with positive completion probabilities. At least 10% of mapped neurons, averaged with equal person weight in a candidate group, must show the abnormality in 30-day means for 90 consecutive days in year one.

That interval must precede the first ≥5% net physical terminal deficit from entry maintained for 30 days in the connected arbor. Ordinary turnover is distinct from sustained net loss. These boundaries define the priced proposition; they are not observed biological tipping points.

Selective Q. Q restores deficient motor-driven transport capacity for a specified nonempty subset of the cargo branches. Capacity means cargo-specific attachment, force–velocity behavior, processive passage, reversal and release under common ATP concentration, cargo load and path geometry. A capacity below the matched normal fifth percentile is restored to the normal median; normal capacity is left alone.

The ideal intervention changes the transport actions of kinesin/dynein motors, their cargo adaptors and motor–track interactions in the named neurons. It leaves the independent mechanical support supplied by the cytoskeleton unchanged. It does not directly manufacture cargo, stabilize NMNAT2, activate TrkA, repair mitochondria, create lysosomes or supply ATP. This requires selective molecular capabilities beyond current HDAC6 drugs.

Q begins at entry and lasts five years. Native energy, cargo production and competing demands remain free. Normal machinery can therefore still transport poorly when ATP is scarce, cargo is defective or a local destination is absent. Physiological transport consumes resources; those costs and inappropriate removal of useful stationary mitochondria count against benefit. A successful intervention is not defined by its effect on terminals.

Local imbalance. R2 requires Q to correct an independently measured abnormal cargo pool in the same connected neurons: distal mitochondrial protein mass per axon volume; somal abundance of transported ligand-bound NGF–TrkA complexes; distal NMNAT2, actin, tubulin or neurofilament abundance per axon volume; or retained, undegraded distal autophagosome cargo mass per axon volume. For the supply branches, reference-care abundance must be below the normal fifth percentile and Q must bring it to at least that percentile. For retained autophagosomes, the respective boundaries are above and at or below the 95th percentile.

The contrast must persist for a 90-day pre-deficit interval during year one. Measure abundance physically; a change in fluorescent brightness or enzymatic activity alone is insufficient. Local synthesis, docking, degradation and export are measured rather than set to zero. For a combined Q, every included branch must satisfy R1 and R2 in overlapping qualifying intervals. Unrelated cargo correction cannot complete a chain whose transport evidence concerns a different cargo.

Source assessment

Best judgment: about 14%; skeptical–favorable sensitivity: 0.4%–62%. These are subjective probabilities of the complete human proposition, not effect sizes or confidence limits. Transport abnormalities are documented, but a rate-limiting, correctable cargo imbalance before human terminal loss remains uncertain. Reviewed 14 September 2026. Probability method · Collection index.

Reasoning from the biology

For a local cargo pool:

change in stock = incoming delivery + local synthesis − degradation − export.

Motor velocity is only one determinant of delivery. The number of loaded motors, direction, pauses, source input, cargo lifetime and terminal capture also matter. Mitochondria can remain stationary while serving demand; moving more is not invariably better. Reduced reverse transport may retain damaged material, but local degradation can compensate. Reduced NGF-endosome transport can limit somal signaling without producing a distal mitochondrial deficit.

A simple one-way transport model illustrates the ambiguity. With fixed entering flux J, velocity v and first-order loss k, steady density at distance x is c(x)=J/v × exp(−kx/v). Halving v changes local density by 2 exp(−kx/v): it increases density near the source and reduces it sufficiently far away. The sign changes at kx/v = ln(2). This is a dimensional example, not a fitted human axon model.

At a terminal with unchanged local synthesis, halving delivery reduces steady stock by only 5% if transport supplied 10% of total input, and by 50% if it supplied everything. Neither proportion is established for the relevant human cargo pools. Protein half-lives measured in another cell type and instantaneous velocities measured during selected runs cannot identify a human leg's supply limit.

The proposed route is therefore:

impaired transport → altered local cargo balance; normalizing transport → possible terminal preservation.

R2 establishes correction of the cargo imbalance; R3 separately tests structural causality of Q. The proposition does not claim that one measured pool mediates all of Q's benefit. Energy supply, trophic signaling, disposal and structural maintenance can interact. It also does not require SARM1 as the sole downstream executor.

Evidence that moves the judgment

Human transport evidence exists, but concerns established mixed neuropathy. Brimijoin and Dyck measured enzyme accumulation against a ligature in incubated sural biopsies from 20 normal subjects and more than 50 patients with several neuropathies. Diabetic nerves had lower DBH/AChE content and transport estimates. Normal DBH average velocity was 1.8 mm/hour; correcting for a transported fraction of 25% gave 7.2 mm/hour. This directly illustrates why moving fraction and velocity must be separated. The accessible primary abstract does not give the diabetic subgroup's n or adjusted comparison. DBH particularly represents sympathetic cargo, and the experiment cannot establish pre-loss timing in peptidergic terminals. Brimijoin and Dyck, 1979.

Experimental diabetes affects some cargo transport. Rats with four to six weeks of STZ diabetes had slower movement of labeled cytoskeletal proteins, proximal axonal enlargement and distal reduction in cross-sectional area. The changes were cargo-dependent. This is an association between transport and axon caliber, not selective restoration with a skin-terminal endpoint; sample-level structural effects were not recovered from the accessible abstract. Medori et al., 1985.

Adult DRG cultures provide a more specific metabolic input: palmitate impaired mitochondrial movement and depolarized mitochondria, whereas the tested physiologically relevant glucose elevations did not produce the same transport deficit. Co-occurring depolarization leaves open whether transport is the limiting lesion or an energy-dependent consequence. Rumora et al., 2018.

The in vivo counterexample matters. In male mice fed 45%-fat diets for 16 or 36 weeks, direct imaging of myelinated saphenous axons found no general reduction in mitochondrial motion, despite altered polarization, conduction and skin innervation. Some transport quantities increased. The main stimulated comparisons have three mice per group, although they include 77–97 axons; the early sham comparison has only two controls. This cannot establish population equivalence, and it does not directly measure the target C-fibers. It does show why a culture transport defect should not be assumed in every diabetic nerve. Sajic et al., 2021.

Newer motor-expression evidence is weaker than its transport interpretation. A 2026 lipoic-acid study retained 15 HFD/STZ rats per group and measured sciatic morphology and proteins in five per group. Treatment improved functional and morphological measures, increased KIF5A and decreased DYNC1I2; glucose did not differ significantly between diabetic arms. The methods contain no cargo-tracking assay. The accompanying NSC34 experiment used 50 mM glucose plus 250 µM palmitate. Thus motor abundance and neurite length support neither measured mitochondrial flux nor selective transport causality. Nonsignificant glucose differences also do not establish exact metabolic matching. Tian et al., 2026.

The HDAC6 trial is a real translational negative with an unresolved structural outcome. In 282 people randomized to ricolinostat 120 mg/day or placebo for 12 weeks, pain change was approximately −1.2 versus −1.0 points (P=.38), and UENS change −1.51 versus −1.8 (P=.56). Mean diabetes duration was approximately 14 years and DPN duration six years. Fourfold PBMC tubulin acetylation does not establish axonal engagement; axonal acetylation was not measured. Michelson et al., 2023.

The public protocol planned baseline-to-week-24 IENFD in approximately 120 participants at selected sites. By week 24, the original placebo group would also have received 12 weeks of active drug, making this a 24-versus-12-week exposure comparison. Neither the paper nor the accessed posted registry results supplies that structural result. It is unavailable, not negative. The protocol's skin-fiber rescue rationale refers to chemotherapy neuropathy, while its diabetic-rat evidence is chiefly allodynia. Protocol, sections 5.3 and 10.3.6, trial record.

The registry labels UENS spreads as standard deviations although their values closely match the paper's standard errors. Treating them as raw SDs would create an unjustifiably narrow interval. Using the paper's rounded SEs yields an approximate Q-minus-placebo UENS contrast of +0.29, with interval −0.82 to +1.40. This examination-scale uncertainty cannot be converted into terminal preservation.

No accessed study measures the entire early-human sequence from cargo passage to corrected local balance, selective structural benefit and incident-case coverage.

Probabilities of the logical claims

Each row after R1 is conditional on all preceding rows. The percentages express joint belief about one compatible group and intervention subset. Cargo branches are alternatives; their probabilities are not independently multiplied or added.

See the claims and probabilities table above.

R1 receives substantial weight from actual human and experimental transport observations, reduced by stage, cargo and fiber mismatch and the in vivo mitochondrial counterexample. R2 is the central physical uncertainty: machinery normalization need not correct a shortage when production, ATP, local capture or degradation is limiting. R3 is relatively high assuming that a sustained cargo imbalance has already been corrected, not because transport is generally necessary.

R4 and R5 remain separate. Correcting a local deficit can produce a small benefit or matter in too few incident cases. A B of 10% satisfies R3 and fails R4; a 30% benefit in a group containing 3% of incident cases passes R4 and fails R5. The conjunction is the full defined claim.

Ideal experiments that would resolve each claim

These are hypothetical resolving experiments, not a proposed study program. They require continuous noninjurious human cargo tracking, selective adult molecular normalization and durable subtype identification. Humans, not trajectories, are the inference units. Incomplete engagement, inadequate precision or missing physical structure leaves a claim unresolved.

R1 — pre-deficit cargo transit. Sample representative eligible adults and matched nondiabetic controls. From entry through year one, track the specified cargo cohorts along complete connected neuronal routes and map terminal structure. Establish each candidate group's 30-day means, 90-day persistence, ≥10% neuron fraction and ordering before the sustained 5% net-deficit landmark. Control reference comparisons for age, sex, route length and site; distinguish motor passage from destruction of the labeled cargo. YES requires at least one complete qualifying branch/group combination; NO means none reaches all boundaries under adequately resolved measurements. Slow motion only after terminal loss is NO for R1.

R2 — transport correction changes cargo balance. Assuming R1, randomize representative members of each qualifying baseline group to Q or reference care from entry. Verify normalization of transport capacity under common-substrate assays and retain its physiological ATP cost. Directly establish unchanged independent cargo synthesis, stabilization and organelle-repair actions of the intervention, while allowing these quantities to respond downstream. Track the branch-specific local stock, production, delivery and removal in the same neurons. An independent molecular implementation and restoration of the original transport defect distinguish intended transport effects from artifacts. YES requires the specified out-of-reference stock under reference care to cross the stated normal boundary with Q for 90 pre-deficit days. Adequate transport-machinery normalization with no such correction is NO, even if motion looks faster.

R3 — structural causality of transport normalization. Assuming R1–R2, continue the randomized comparison to five years. Hold systemic glucose, lipids, blood pressure, medications and other initiating exposures comparable by design; allow downstream ATP, trophic signaling, disposal, firing and repair to change. Quantify physically destroyed and successfully replaced peptidergic terminals, including resource-related harm from Q. Verify the selective transport implementation independently and reverse its transport action where necessary to distinguish a shared off-target effect. YES is positive reference mean loss and B≥5%; NO is B<5% in every qualifying group/subset. Pain improvement without structural measurement is unresolved, not NO.

R4 — material effect size. Assuming R1–R3, estimate the same person-weighted five-year contrast with enough precision to resolve B=20%. No additional exposure, subgroup or endpoint is substituted. YES is B≥20%; NO is 5%≤B<20% for every remaining candidate. Values near a boundary require greater precision, not a nominal P-value declaring success.

R5 — population coverage. Assuming a candidate passes R1–R4, use representative reference-care follow-up of the full eligible population. Ascertain incident DPN uniformly and calculate the fraction of incident cases belonging to that entry-defined group, accounting for loss to follow-up without assuming it is random. YES requires C≥10% for a group already passing the preceding requirements; NO is C<10% for all such groups. The group's prevalence in all adults or a treated responder series does not answer this question.

Uncertainty, sensitivity and overlap

Best inputs [70,60,75,70,65]% give 14.3325%. Skeptical [35,25,40,35,30]% and favorable [90,90,95,90,90]% give 0.3675% and 62.3295%. These are coordinated judgments, not independent beta distributions, sampling intervals or a measured distribution of human effects.

A buffered-supply scenario, [50,35,60,50,45]%, gives 2.36%. A transport-limited scenario, [85,80,90,85,80]%, gives 41.62%. Neither scenario has an estimated frequency. The favorable case requires early transport abnormalities to persist in the target fibers, a correctable imbalance, and a substantial common structural contribution.

Holding other rows fixed, evidence moving R2 to 85% raises the headline to 20.30%; moving it to 25% lowers it to 5.97%. Resolving R2 YES with certainty would leave 23.89%. These are hypothetical sensitivity calculations. Merging R1 and R2 must preserve their 42% joint probability; splitting them is not an extra biological penalty.

Changing the cargo list, duration, transit boundary, B threshold or case-coverage threshold changes the proposition and requires renewed elicitation. Relaxing a boundary can only expand the qualifying set, but available evidence does not determine the resulting probability numerically.

Transport overlaps with NGF signaling, energy supply, autophagic completion, mitochondrial dynamics and SARM1. Their probabilities cannot be summed as separate fractions of preventable loss. This estimate concerns cargo imbalance correctable through the specified transport machinery; it does not establish that any HDAC6 drug prevents incident DPN.