Oxidized LDL

In adults diagnosed with type 2 diabetes within five years and without clinical DPN, increased local exposure to oxidized LDL activates neuronal LOX-1-dependent NADPH-oxidase superoxide production. Selectively preventing that production increment before injury would reduce five-year net peptidergic skin-terminal loss by at least 20%, in a baseline-describable group accounting for at least 10% of reference-care incident DPN cases.

This is the direct neuronal LOX-1/redox branch. Oxidized particles can also injure vessels, act through TLR4 or other receptors, or change support-cell function. Those alternatives can be real while this proposition is false. LOX-1 is the OLR1-encoded scavenger receptor, not the ordinary LDL receptor LDLR and not a lipoxygenase. Neither high LDL cholesterol nor an antibody assay called “oxLDL” alone establishes the specified exposure.

The relevant unit is a peptidergic sensory neuron's DRG soma, peripheral axon and skin terminals. Exposure and signaling must occur in connected parts of the neurons whose terminals are followed. For this assessment, LDL means apoB100 particles in the LDL density fraction, excluding apo(a)-containing particles. Oxidation must be chemically demonstrated in particle lipids or apoB; glycation alone does not qualify. The operational exposure is their oxidation-dependent binding activity for human LOX-1, measured as binding events per unit receptor-bearing area per second under a fixed, calibrated cell-free assay. Measure the available extracellular particle mixture at the neuronal surface, preserve its concentration, and compare it with molecularly matched reconstituted particles in which identified oxidative modifications have been replaced by their unoxidized precursors. This defines a particle-associated activity without defining it by neuronal injury. It is not interchangeable with HODE content, MDA immunoreactivity or total apoB mass.

Allowed baseline groups are the whole eligible population, fasting triglycerides ≥150 mg/dL, HbA1c ≥7.5%, or local oxidation-dependent LDL binding activity above the matched nondiabetic 95th percentile. R1 requires a permitted group's mean activity above the age-, sex- and ordinary-activity-matched nondiabetic 95th percentile for at least 90 cumulative days in year one, while the implicated neuronal structures remain intact before losses attributed to the route. Include all baseline participants and early losses; do not select future decliners or survivors. The same group, particle activity and connected neurons must satisfy every later row.

For each person, L is terminal disappearances minus newly established terminals over five years, divided by baseline terminal count. Track actual structures with stable peptidergic identity; disappearance of CGRP staining is insufficient. A new terminal that later disappears enters both counts. For positive reference-care mean loss L0 and intervention mean loss L1, define B=(L0−L1)/L0. A smaller structural contribution is B≥0.05; materiality is B≥0.20. Coverage C is the qualifying group's share of all reference-care incident DPN cases, with C≥0.10. Incident DPN requires new bilateral length-dependent sensory signs with objective small-fiber loss or abnormal nerve conduction, excluding other neuropathy causes. These are chosen definitions, not measured biological constants.

Source assessment

Best judgment: about 5%; coordinated sensitivity scenarios: 0.06–42%. The conditional product is 4.5%. These are subjective probabilities of the complete human claim, not treatment effects or confidence intervals. Reviewed 14 September 2026. Probability method · Collection index.

From particle exposure to structural causation

Oxidation changes a particle's molecular cargo and receptor recognition. Production, vascular access, local oxidation, binding, uptake and clearance determine what reaches a neuron. Circulating LDL can be abundant without substantial local oxidized-ligand activity. Conversely, local oxidation can matter without a higher blood assay result. Whole-nerve oxidative products can arise in vessels, macrophages or Schwann cells and need not be carried into neurons on LDL.

Let Q be an ideal adult-onset intervention that prevents the excess oxidation-dependent LDL interaction with neuronal LOX-1, returning this interaction to the matched nondiabetic trajectory. It leaves particle availability, other neuronal receptor interactions, LOX-1 interactions with other ligands, and LOX-1 in vessels and support cells at reference conditions. This is a ligand- and cell-restricted manipulation, not whole-body OLR1 deletion. Its effects on receptor-mediated uptake and downstream signaling may occur. Reversible restoration of the original interaction tests specificity.

R2 asks whether Q reduces superoxide production by neuronal NADPH oxidases during the exposure periods. Production is source-resolved mol O2•− per neuronal volume per time, not total fluorescent probe oxidation, steady-state oxidant concentration, NOX expression or antioxidant depletion. The contrast is reference minus Q; it must be strictly positive. This row tests receptor-to-source coupling without assuming that the oxidants cause injury. It does not require a particular NOX isoform, mitochondrial amplification or a separate depletion of antioxidant defenses.

Let S selectively reproduce the NADPH-oxidase superoxide-production trajectory under Q while retaining reference LDL exposure and LOX-1 interactions. It changes the implicated oxidase electron transfer to oxygen, with electron/substrate balances accounted for; it does not directly remove oxidants from other sources or inhibit axon destruction. Hydrogen peroxide generated downstream, redox signaling, calcium, mitochondrial responses, inflammation induced by the neuron, and replacement may change. Preserving ordinary basal oxidase function is part of this counterfactual. A nonspecific antioxidant or apocynin alone does not implement S.

The decisive sequence is local particle activity → LOX-1-dependent source production → the structural effect of preventing that source increment. An increase in oxidants may be buffered, adaptive or too small to matter. The net effect of S includes any loss of beneficial redox signaling. ATP depletion, soma apoptosis and failed regeneration are possible consequences, not compulsory additional discounts.

Probability decomposition

Each row after R1 assumes every preceding requirement holds in the same permitted group and connected neuronal unit. Retain only groups satisfying all earlier rows. The conjunction includes the initiating exposure, specified receptor/source relation, structural effect, material magnitude and case coverage; it is not merely a list of necessary molecular possibilities.

See the claims and probabilities table above.

R1 is near even odds because oxidized particles and diabetic tissue oxidation are established possibilities, but local pre-loss receptor-binding activity in the human target has not been quantified. R2 remains below even odds despite a favorable adult-DRG experiment: receptor dependence conflicts across preparations, source attribution is weak, and systemic antibody treatment does not locate the effect in neurons. Conditional on R1–R2 being true, R3 discounts whether the measured source increment actually contributes to terminal structure, not whether particles reach neurons again. Acute cell-death rescue supplies possibility rather than a five-year estimate. R4 concerns the 20% magnitude after accepting a smaller contribution. R5 gives moderate weight to common dyslipidemia while retaining uncertainty about how many incident cases share this much narrower local mechanism.

Evidence that bears on the claim

Experimental exposure and structural association. Vincent and colleagues fed mice 45% versus 10% fat diets, with 10 mice per group. Functional deficits appeared at 12 weeks; epidermal fiber loss was measured at 34 weeks. LDL chemistry used only two five-mouse pools per group. HODE, dityrosine and nitrotyrosine ratios in isolated LDL were 1.66, 23.74 and 1.60, respectively. Whole-nerve oxidation and diet-associated denervation do not demonstrate local neuronal particle delivery or LOX-1 mediation. Their estimated 300–400 µg/mL total plasma LDL protein is 10–13 times the 30 µg/mL culture dose, but neither fraction oxidized nor perineuronal exposure was established. This arithmetic cannot establish a physiological culture dose. Vincent et al., 2009.

In the same report, adult female rat DRG cultures received 1–100 µg/mL oxLDL at basal 5.7 mM glucose, with a separate 25.7 mM glucose condition. At 30 µg/mL, roughly 60% of cells were TUNEL-positive after 24 hours; LOX-1 antibody, apocynin and α-lipoic acid reduced injury. The displayed n=9 represents triplicate wells on three culture occasions, not nine independent animals. The article prints the neutralizing-antibody concentration as 100 mg/mL, with other inconsistent units, and does not sufficiently characterize the experimental particles' oxidation for human-dose matching. Fluorescence, FCCP responses and apocynin-sensitive lysate assays do not uniquely identify neuronal NOX or mitochondrial superoxide production. The endpoints are soma-associated caspase/TUNEL, not terminal persistence. This remains favorable mechanistic evidence with substantial measurement limits, not a replicated in vivo structural intervention.

Receptor dependence is disputed. Nowicki and colleagues found that blocking LOX-1 increased cleaved caspase-3 in postnatal rat DRG cultures, whereas TLR4 inhibition reduced it. Their primary abstract therefore challenges a universal injurious LOX-1 interpretation. It does not show that mature human peptidergic neurons behave identically. The full methods were not accessible in this review, so independent n, particle preparation and matched-dose explanations for the discrepancy remain unresolved. Nowicki et al., 2010.

A subsequent systemic LOX-1-antibody experiment treated db/db mice every 48 hours from six to twelve weeks of age. Its primary report describes preservation of sural sensory conduction, without correction of sciatic motor conduction or major glycemic/lipid measures. That is useful intervention evidence beyond culture. However, the accessible abstract supplies neither independent n nor a structural effect estimate, and systemic blockade cannot distinguish neuronal from vascular or glial action, or oxLDL from other LOX-1 ligands. The article belongs to the 2012 journal volume; the publisher's 2019 online date is not a second experiment. The full report remained inaccessible. Vincent et al., 2012.

A different cell/receptor route remains plausible. In IMS32 mouse Schwann cells, Nihei et al. used human LDL oxidized with 20 µM copper for 24 hours. At 150 µg/mL oxLDL and 25 mM glucose, TAK-242 increased the reported MTT value from 56.7% to 86.8% and reduced caspase-3 activity. MTT panels show five wells from representative experiments, with at least three experiments reported; they do not provide five independent donors. TLR4 mRNA contrasts used two independent experiments and protein contrasts three. This supports a Schwann/TLR4 possibility, with acute copper-oxidized particles and no connected axonal structural outcome. It does not resolve neuronal LOX-1. Nihei et al., 2024.

Human blood studies are mixed and poorly matched to local exposure. Rosales-Hernandez et al. studied 70 people with established DPN, 12 with T2 diabetes without neuropathy, 13 with idiopathic neuropathy and 20 healthy controls. Mean diabetes duration was approximately 7–9 years; statin use was common. Reported plasma oxLDL was 1,633±121 versus 1,829±721 pg/mL in DPN versus diabetes-only participants, with SEM. A Welch calculation gives a difference of −196 pg/mL and an approximate 95% interval of −1,795 to +1,403. It excludes neither a substantial increase nor decrease. The printed pooled DPN mean also differs from the 1,650 pg/mL implied by its subgroup means. Clinical severity explained little assay variation. This is neither a precise equivalence result nor a measurement of the defined neuronal particle activity. Rosales-Hernandez et al., 2014.

Al-Nimer et al. likewise found no clear LDL-oxidation association in 51 T2 participants with neuropathy, 33 without and 31 controls. Their assay measured thiobarbituric-acid-reactive material in a precipitated LDL fraction, a different quantity from receptor-binding particles. Al-Nimer et al., 2012. A positive study reported oxLDL 11.38±3.1 ng/mL in a 25-person DPN group versus 7.34±2.95 in 25 healthy controls, and higher values in ulcer/Charcot groups. All 100 diabetic participants already had neuropathy; there was no diabetes-without-neuropathy comparison. Its AUC of 0.855 distinguishes complicated diabetes from health and does not establish incident-DPN prediction. El Erian et al., 2015, primary report.

ApoE deletion on diabetic mouse backgrounds greatly altered lipoproteins without clearly worsening the neuropathy course at 24 weeks. This challenges a simple total-lipid dose explanation. It does not selectively manipulate local oxidized-ligand activity or LOX-1, and ApoE also affects neuronal lipid handling. The accessible primary abstract and figures do not provide a sufficiently precise, source-engaged structural contrast to exclude this route. Hinder et al., 2013.

Human triglyceride associations, LDLR genetics and statin observations address broader lipid exposures. They cannot identify this oxidized-particle/neuronal-receptor path. The retrieved evidence contains no selective human Q or S intervention with neuronal engagement and terminal structure. Public bulk or single-cell OLR1 expression also cannot resolve ligand activity, source production or pre-loss timing.

Ideal experiments that would resolve each claim

These are conceptual truth conditions, not proposed studies. They require noninjurious repeated human perineuronal sampling, source-resolved oxidant measurements, stable longitudinal terminal identification and selective adult interventions. Sample a representative early-T2 population and matched nondiabetic references. Independent humans supply inference. All comparisons preserve the same reference diabetes care, glycemia, oxygen delivery, other lipid inputs, trophic support and other initiating inflammation; downstream responses to the named manipulation may change. Use simultaneous precision across the finite allowed groups.

R1 — local exposure before loss. With no predecessors, measure the defined oxidation-dependent binding activity in extracellular fluid at the connected DRG/axon/terminal surfaces from baseline through year one. Resolve apoB100 LDL from other particles, quantify chemical oxidation and compare matched unoxidized reconstitutions. Preserve free versus sequestered particle availability during sampling. Follow the same baseline terminals, including early losses. YES requires at least one allowed group to meet the above-95th-percentile and 90-day definition before its attributed losses. NO requires every allowed group to fail it with sufficient temporal and anatomical sampling. Blood ELISA alone, sampling damaged tissue only, or uncertainty spanning either boundary remains unresolved. Existing mouse chemistry supplies oxidation, but not this human localization/activity/time comparison. This row does not establish receptor signaling or structural injury.

R2 — neuronal LOX-1 to source production. Assuming R1, randomize Q versus sham in the same group and measure neuronal NADPH-oxidase superoxide production during the R1 exposure periods. Verify selective normalization of the oxidized-LDL/LOX-1 interaction, intact other-ligand and other-cell interactions, and unchanged extracellular particle availability. Resolve oxidase source and compartment directly; reversible restoration and an independent ligand-specific manipulation check causality. YES is a strictly positive mean production contrast, reference minus Q. NO is a nonpositive contrast in every qualifying group. An interval wholly above zero supports YES and one wholly below zero NO; exact zero is NO in principle but usually unresolved with finite precision. A change in dye uptake, removal of glial oxidants, or incomplete receptor engagement is unresolved. Existing adult-culture evidence supports this connection indirectly, while the postnatal receptor result conflicts. Neither settles its human sign or the next row's injury.

R3 — smaller structural contribution. Assuming R1–R2, randomize S versus sham from entry through five years. Target the source-production trajectory measured under Q, while retaining reference particles and LOX-1 interactions. Verify source selectivity and electron/substrate accounting; allow downstream oxidant products, signaling, degeneration and replacement to respond. Use an independent source manipulation and restoration of the original source trajectory to distinguish source causation from apparatus effects. Directly count terminal disappearances and new entries. YES requires B≥0.05; NO is B<0.05 at adequate precision. For example, L0=.30 and L1=.28 gives B=.067 and passes this row. Imprecise effects, nonpositive or poorly estimated reference loss, nonselective antioxidant action, or pain/conduction alone do not establish it. Existing culture rescue and systemic conduction results do not supply S's human structural effect or its size.

R4 — material magnitude. Assuming R1–R3, estimate the same five-year S contrast in the same baseline group, propagating uncertainty in L0. YES is B≥0.20; NO is B<0.20 with the interval wholly on the appropriate side. L0=.30 and L1=.23 passes; .28 fails despite satisfying R3. A true L0≤0 fails the relative-preservation proposition; an uncertain denominator leaves it unresolved. No new mediator is required. Current evidence lacks a human estimate of this magnitude, and a decisive R4 result alone would not establish case coverage.

R5 — case coverage. Assuming at least one group meets R1–R4, estimate C in a representative reference-care population over the same five years. Assign membership from baseline chemistry or the other allowed baseline definitions, without using treatment response. Account for missing follow-up and competing death in the defined incident-case population. YES requires C≥0.10 for at least one otherwise qualifying group; NO requires C<0.10 for all of them with adequate precision. A 3% group fails even with large structural benefit; a 15% group qualifies. An enriched trial, the prevalence of high triglycerides alone, or the fraction of cases remaining after S leaves this unresolved. None of the retrieved studies jointly measures the qualifying neuronal mechanism and incident-case coverage.

Calibration, sensitivity and boundaries

The best inputs [55, 45, 55, 55, 60]% give 4.4921%. Skeptical inputs [25, 15, 25, 25, 25]% give 0.0586%, interpreting the blood/nerve signals as largely compartment-mismatched and receptor-dependent culture toxicity as poorly transported to adults. Favorable inputs [85, 85, 85, 80, 85]% give 41.7605%, assuming common early local ligand activity, adult neuronal LOX-1/NOX coupling and consequential structural mediation. These coordinated interpretations move related rows together; they are not independent probability distributions. The approximately 42-percentage-point span is a sensitivity range, with no justified variance or central 90% interval.

A separate whole-claim assessment is low but appreciably above zero: specific receptor perturbations make a causal route credible, yet neither the human exposure nor the selective structural counterfactual is demonstrated. A few-percent joint judgment is compatible with that assessment. The favorable scenario remains broad because human source-resolved measurements could overturn the main doubts. The evidence cannot distinguish 4% from 6% reliably.

Hypothetically changing only R2 from 45% to 80% raises the complete judgment to 8.0%; changing it to 15% lowers it to 1.5%. A direct human receptor/source result would resolve more than another OLR1 expression association, but it would not itself settle structural materiality. Real evidence spanning multiple steps should update those steps together. Merging R1–R2 into a 24.75% joint condition preserves the product; splitting receptor localization or optional mitochondrial amplification must not impose arbitrary new penalties.

The 90-day, 95th-percentile, 5%, 20% and 10% boundaries define this claim. Relaxing them can only enlarge its truth set when the intervention and population remain fixed; current data do not quantify how much probability should change. A negative LOX-1 result does not exclude TLR4 inflammation, vascular injury, Schwann support failure, or other oxidized-lipid mechanisms. This route overlaps oxidative/nitrosative injury; the probabilities cannot be added. It is distinct from nonesterified fatty-acid entry.