The Vectorizing Excipient Hypothesis: Polysorbate 80, Putative SV40-Containing Residual DNA, and Differential Lymphoma Reporting Disproportionality in Anti-TNF Biologics

Abstract Background Among approved anti-TNF biologics, infliximab shows the highest lymphoma reporting disproportionality in pharmacovigilance databases, with a consistent gradient placing it above golimumab, adalimumab, and certolizumab pegol — despite comparable TNF-a blockade potency across all four agents. The mechanism underlying this differential signal has not been identified. Infliximab is the principal intravenous anti-TNF agent, produced in Sp2/0 murine myeloma cells; a possible contribution of this production line's genomic background is examined separately in the main text as an unverified secondary branch, not summarized here. Golimumab also has an approved IV formulation (Simponi Aria) containing polysorbate 80, though available pharmacovigilance data do not permit reliable separation of IV from SC golimumab reports. The Vectorizing Excipient Hypothesis We propose that polysorbate 80 (PS80), present in infliximab formulations at ~3× its critical micelle concentration in the reconstituted vial, functions as an unintended vectorizing agent for residual plasmid-derived DNA fragments that may co-present in the finished drug, and which — if derived from expression vectors carrying SV40 regulatory elements — would constitute a plausible mechanism for sinusoidal macrophage activation. Whether SV40-containing sequences are present in the residual DNA fraction of commercial infliximab has not been publicly demonstrated; this is the central testable assumption of the hypothesis (Prediction P1). The proposed mechanism involves four steps, each with variable indirect support drawn from distinct experimental systems rather than a single demonstrated chain: (i) electrostatic association of residual DNA with cationic patches on the infliximab variable domains, potentially forming a protein-DNA complex within the vial; (ii) a proposed recruitment of plasma ApoE by this complex via PS80 — inferred from engineered nanoparticle systems rather than demonstrated for a free-monomer PS80-mAb-DNA complex — that would facilitate LDL-R-mediated uptake by hepatic Kupffer cells and splenic macrophages if it occurs; and, for DNA reaching the cytoplasm, two alternative, non-sequential fates depending on subcellular localization: (iii) cGAS-STING innate immune sensing for the fraction retained in the cytosol, or (iv) sequence-specific nuclear import via the importin-a/b mechanism for the fraction reaching the nucleus — a DNA nuclear targeting sequence active in every mammalian cell type tested to date, including post-mitotic macrophages, and not shared by several other common viral regulatory elements tested for comparison. Illustrative quantitative estimates of fragment counts and PS80:DNA molar ratios under stated working assumptions are developed in Table 1 and the main text; they are scenario calculations intended to establish physical plausibility, not validated measurements. Epidemiological coherence A disproportionality analysis of FAERS data (2009Q1–2026Q1; n=1,835,068 unique cases) conducted by the authors yields a consistent lymphoma reporting-disproportionality gradient by active substance: infliximab (ROR 8.61, 95% CI 7.86–9.44) > golimumab (4.59, 3.07–6.87) > adalimumab (2.04, 1.83–2.28) > certolizumab (1.28, 0.91–1.80; no signal). This active-substance gradient is the primary epidemiological finding and carries the greatest statistical power. The outcome is any MedDRA Preferred Term containing "lymphoma" and is not specific to hepatosplenic T-cell lymphoma (HSTCL) or any single histological subtype; this broad signal cannot by itself establish that HSTCL specifically is overrepresented, and the mechanistic hypothesis below concerns, at most, an unquantified fraction of this broader gradient. A route-of-administration sensitivity analysis confirms the infliximab signal remains high among reports explicitly coded as intravenous, but the number of route-coded IV golimumab reports is too small (n=49, zero lymphoma cases) to test whether IV administration itself explains the gradient; this is a data limitation, not a negative result. The observed gradient is compatible with, but does not discriminate between, multiple non-mutually-exclusive explanations, including route-of-administration effects, host-cell DNA content, and clinical confounding not captured by spontaneous reports. Falsifiable predictions and regulatory implications The hypothesis generates six experimentally testable predictions, four addressable immediately with commercially available materials and standard methodology, including NGS detection of SV40-containing sequences in the protein-associated residual DNA fraction of commercial infliximab vials. Current residual DNA regulatory limits are quantity-only and route-independent. A tiered characterization framework and linkage of lot-specific quality data to pharmacovigilance outcomes are proposed as actionable regulatory responses.