{"id":"aef0c763-dbb0-4fcb-8923-edc879bf9f98","arxiv_id":"1908.03686","paper_version":2,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":2.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"A personal memoir describing the author's 1990s hypothesis and unpublished experiments on dendritic cell-mediated immune memory, with a proposal for modern validation.","lead":"An immunologist recounts how he hypothesized in the 1990s that dendritic cells can remember antigens on their own, and how he named his daughter Shuyi after that idea. The essay contains no new data, but it connects his old unpublished experiments to today's trained immunity research.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Residual OVA is not excluded: anti-OVA IgG appeared after reconstitution even without challenge, yet 125I was measured only in the footpad; transferred DCs can carry peptide-MHC for months, confounding antigen persistence with APC memory.","rationale":"The reader's weakest assumption identified residual OVA in tissues other than the footpad; I agree that this is the central confound. I extend it in two ways: the manuscript's own observation of an anti-OVA response without re-challenge is internal evidence that antigen persists somewhere, and the in-vitro-pulsed DC transfer design cannot distinguish long-lived APC function from long-lived antigen carriage, since DCs are themselves a route for antigen and peptide-MHC persistence. This is not an external-consensus disagreement; it is an internal control problem. The author's 125I measurement at the injection site only, followed by a claim of no residual antigen, is insufficient because antibody production without challenge requires an antigen source. The transfer experiments with adherent-cell depletion and pure DCs are likewise consistent with antigen transfer rather than APC memory. Because the manuscript provides no raw data, protocols, or statistics, the claim cannot be verified from the text, matching the reader's UNVERDICTED verdict. My read does not change that verdict, but it sharpens the specific experimental control that would settle the issue. A no-challenge arm in the DC transfer experiment, combined with whole-animal radiotracing, would distinguish residual antigen from APC-intrinsic memory. I therefore leave the verdict unchanged and mark agreement as partial because the reader focused on uninjected tissues, while I also flag the no-challenge response and DC-carried peptide-MHC as equally load-bearing aspects of the same confound.","tokens_in":3712,"tokens_out":5679,"duration_ms":69445,"concrete_test":"Repeat the pure-DC transfer experiment with a no-challenge arm and with 125I-OVA traced over the full three months: transfer OVA-pulsed DCs into SCID mice, then after three months either reconstitute with naive splenocytes and challenge with OVA, or reconstitute and challenge with PBS. If the PBS-challenged recipients still produce anti-OVA IgG, OVA/peptide-MHC persisted and the memory-specific interpretation fails; also perform whole-animal radioactivity counting (spleen, lymph nodes, bone marrow, liver, and blood) to locate any residual depot.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Taking the described experiments at face value, the central inference that APCs/DCs are intrinsically memory cells is underdetermined. The first observation—primed SCID mice produce anti-OVA IgG after receiving naive splenocytes 'even without being challenged with OVA'—is direct evidence that immunogenic OVA-derived material persists somewhere in the animal. The author's exclusion test measured 125I only in the footpad; antigen depot in spleen, lymph nodes, bone marrow, or on APCs is not ruled out. Even if free antigen is below the priming threshold for naive immunocompetent mice, a persistent depot could provide prolonged antigen exposure that mimics a secondary response, especially in mice whose APCs are reportedly primed. The later pure-DC experiment is even more confounded: DCs cultured from bone marrow, pulsed with OVA in vitro, and transferred into SCID mice are a direct vehicle for antigen and peptide-MHC. Their persistence for three months and ability to activate naive T cells after reconstitution is exactly what would be expected from long-lived antigen carriage, not necessarily from altered APC function. Adherent-cell depletion also removes antigen-bearing cells, so it does not separate memory from antigen transport. The manuscript itself asserts both 'independent of priming antigens' and 'present their antigenic epitope loads,' a tension that highlights the ambiguity. Thus the strongest claim—that APCs gained and maintained an enhanced ability to activate antigen-specific T cells—requires ruling out antigen/peptide-MHC persistence; the tests described do not.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript is a personal essay by Yiqiang Wang describing his hypothesis, developed in the late 1990s, that antigen-presenting cells (APCs), including dendritic cells (DCs), constitute a third class of immune memory cells operating independently of T and B lymphocytes. The author recounts experiments in SCID mice in which OVA-primed mice, after a three-month rest and subsequent reconstitution with naive splenocytes, produced a secondary-pattern anti-OVA IgG response without further antigenic challenge. He also describes attempts to exclude residual OVA by 125I-labeling, adherent-cell depletion, and transfer of OVA-pulsed DCs into SCID mice. The essay argues that these findings support a memory-like state in APCs, connects the hypothesis to the recent literature on trained immunity (including Hole et al.), and proposes future experiments using GFP-labeled DCs and single-cell sequencing. It concludes with a personal story about naming his daughter Shuyi after the hypothesis.","tokens_in":3975,"tokens_out":8203,"duration_ms":86226,"significance":"If the central claim were established, it would represent a significant conceptual advance, adding APCs as a third arm of immunological memory independent of T/B cells. The manuscript clearly articulates the hypothesis and, importantly, proposes a concrete and feasible experimental design (GFP-labeled DCs, single-cell sequencing) that could rigorously test it. The connection to recent work on trained immunity and memory-like DC responses gives the hypothesis contemporary relevance. However, the paper provides no verifiable data: all experimental recollections are presented without raw data, statistics, methods details, or independent confirmation, and the original manuscript was never published. The evidentiary basis is anecdotal. The main constructive contribution is the proposal of future experiments; if those experiments are performed, they could provide the needed test.","major_comments":[{"comment":"The exclusion of residual OVA is incomplete, and this incompleteness directly undermines the central inference. The author reports measuring 125I radioactivity only at the footpad injection site, yet the key observation that primed SCID mice produce anti-OVA IgG after reconstitution with naive splenocytes even without OVA challenge indicates that immunogenic OVA-derived material persists somewhere in the animal. No measurements are described for regional lymph nodes, spleen, bone marrow, or cell-associated antigen depots. The alternative explanation that the 'secondary-like' response is a normal primary response boosted by persisting antigen is therefore not ruled out, which contradicts the claim that APCs 'gained and maintained the ability' to activate T cells independently of residual antigen.","section":"Paragraph beginning 'To test if this apparent memory was actually due to potential residual OVA antigens'"},{"comment":"The pure-DC transfer experiment is confounded by antigen carriage. DCs pulsed with OVA and transferred into SCID mice carry both intact antigen and peptide-MHC complexes on their surface, so the enhanced anti-OVA response after three months could be due to long-lived antigen presentation rather than an intrinsic memory phenotype. The author's own statement that APCs 'present their antigenic epitope loads' is in tension with the claim that the memory is 'independent of priming antigens.' Without an experiment that separates antigen carriage from altered APC function, the described data cannot distinguish between these possibilities.","section":"Paragraph beginning 'For explaining the memory-like status observed in OVA-primed SCID mice'"},{"comment":"The manuscript contains no raw data, error bars, statistical tests, or methodological details (e.g., numbers of mice, antigen doses, time courses, ELISA protocols). Because the original manuscript was rejected and never published, the reader has no independent means to verify the described findings. This is load-bearing, since the paper's central claim rests entirely on these unreported experiments; the recent literature on trained immunity and Hole et al. provides plausibility but does not validate the specific observations recounted here.","section":"Full text, experimental descriptions throughout"},{"comment":"The author acknowledges that the earlier manuscript was rejected, the project discontinued, and the hypothesis now needs to be demonstrated 'with more convincingness.' This admission should be reflected in the structure of the essay: the preliminary, confounded observations should be clearly labeled as historical background and hypothesis-generating, and the proposed future experiments should be presented as the definitive test. As written, the manuscript's language overstates the level of support for the memory-DC hypothesis.","section":"Final paragraph before Competing interests"}],"minor_comments":[{"comment":"The abstract contains a typo: 'dendtitic' should be 'dendritic.'","section":"Abstract"},{"comment":"The text uses 'naï ve' with a combining diaeresis and a space; it should be 'naive' in standard English formatting.","section":"Full text"},{"comment":"Figures 1 and 2 are photographs of documents, not data figures. They would be better labeled as plates or supplementary material, and the narrative reference to them should be clarified accordingly.","section":"Figure captions"},{"comment":"The description of the daughter's current university and major is personal information that, while touching, is not relevant to the scientific content and could be moved to an acknowledgments section or omitted.","section":"Final paragraph"},{"comment":"The reference list could usefully include earlier work on antigen persistence and follicular dendritic cells, which would help readers evaluate the residual-antigen confound discussed in the major comments.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is more of a personal recollection than a conventional research paper; the editorial office should decide whether the journal has a suitable format (e.g., Perspective or Reflections) for such a contribution. The central hypothesis is plausible given recent literature, but the evidence described is too weak and confounded to sustain the strong claims as written. A revision that clearly separates hypothesis from evidence and explicitly addresses the residual-antigen confound could make the paper acceptable as a perspective essay."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this so you know what it is: a personal essay, not a research preprint. Wang revisits unpublished experiments from the late 1990s that suggested dendritic cells/APCs can mediate antigen-specific memory in the absence of T and B cells, connects the idea to the trained immunity literature, and tells the story of naming his daughter Shuyi after the hypothesis. There is no new data, no protocols, and no statistics; the experiments cannot be checked from the text. If you expect a research claim, this is thin. But as a documented early hypothesis it is unusually honest: Wang states plainly that the original manuscript was rejected, never resubmitted, and is being described from memory.\n\nCredit where due: the framing against current work (especially Hole et al. on memory-like DCs) is accurate, and the proposed modern experiment with GFP-labeled DCs and single-cell sequencing is concrete and sensible. The personal narrative is not fluff; it records priority and motivation without overclaiming. The reference list is appropriate, and the author does not lean excessively on his own old papers.\n\nThe soft spot is load-bearing. The old data cannot distinguish APC memory from persistent antigen. The author's observation that reconstituted SCID mice produced anti-OVA IgG even without OVA challenge is direct evidence that immunogenic OVA-derived material was still present somewhere. The residual-antigen control measured 125I only in the footpad; it did not rule out antigen depot in spleen, lymph nodes, bone marrow, or on transferred cells. The pure-DC experiment is the most confounded: DCs pulsed with OVA in vitro and transferred into SCID mice are a vehicle for antigen and peptide-MHC, so their persistence and ability to activate naive T cells three months later is exactly what long-lived antigen carriage would predict. Adherent-cell depletion also removes antigen-bearing cells rather than separating memory from transport. The manuscript itself contains the tension: APCs are said to act 'independent of priming antigens' while also presenting 'their antigenic epitope loads.' Those two statements cannot both carry the argument.\n\nSo the central hypothesis remains plausible but unproven, and the tests described do not exclude the obvious alternative. That makes this a perspective/history piece rather than a research result. It could be useful for someone thinking about trained immunity or the history of the idea; it is not citable as evidence. The citation pattern is fine.\n\nMy recommendation: as a research manuscript, desk reject. As a perspective or historical correspondence, one serious referee could help make the limitations explicit. I would not send it out as a primary research claim.","headline":"A memoir-style essay with a plausible but unproven APC-memory hypothesis; residual antigen is never really excluded, so treat it as perspective, not evidence.","tokens_in":4512,"tokens_out":4150,"would_cite":false,"duration_ms":45047,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Dendritic cells may form a third class of immune memory","keywords":["dendritic cells","immunological memory","antigen-presenting cells","SCID mice","trained immunity","innate memory","ovalbumin","adoptive transfer"],"falsifier":"Measure the 125I radioactivity in regional lymph nodes, spleen, liver, and other organs at three months after priming in SCID mice; if intact OVA or immunogenic fragments above the priming threshold exist anywhere, the memory-like response may be a primary response boosted by persistent antigen rather than APC memory. Alternatively, adoptively transfer primed APCs that have been fixed with paraformaldehyde or treated with acid to strip surface peptides; if memory survives such treatment, it would argue against epitope presentation as the mechanism.","tokens_in":3509,"feed_emoji":"🛡️","tokens_out":8399,"duration_ms":74862,"temperature":0.7,"pith_summary":"This essay reports the author's late-1990s experiments suggesting that antigen-presenting cells, especially dendritic cells, can build up a memory of an antigen that lasts at least three months even when no T or B cells are present. In SCID mice primed with ovalbumin, transferring naive lymphocytes months later produced a secondary-like antibody response without a fresh challenge, and the memory could be carried by splenic or lymph-node cells but not bone-marrow cells. The author argues these findings, together with recent evidence for trained immunity, make antigen-presenting cells a plausible third class of immune memory. The personal story about naming his daughter Shuyi frames the hypothesis as an unfinished project that modern tools could now test.","feed_headline":"Dendritic cells may remember invaders without T or B cells","feed_subtitle":"SCID-mouse experiments from the 1990s point to antigen-presenting cells that keep antigen-specific recall for months.","key_machinery":"The key experimental system is the SCID mouse, which has no functional T or B lymphocytes, so any antigen-specific memory observed after reconstitution must originate from the host's antigen-presenting cells. The mechanism proposed is that APCs activated during priming retain antigenic epitopes or an enhanced presentation state for at least three months, and when they encounter naive T cells they present these epitopes more efficiently than newly-challenged APCs would. The effective transfer of the memory by splenic and lymph-node cells, but not bone-marrow cells, and its loss when adherent cells are removed, points to dendritic cells as the operative cell type. The paper also uses 125I-labeled OVA to estimate residual antigen at the injection site, an attempt to rule out the simplest alternative explanation.","core_discovery":"The author claims that antigen-presenting cells (APCs) constitute an independent, antigen-specific memory system. The central evidence comes from SCID mice, which lack functional T and B cells: mice primed with OVA and rested for three months produced a secondary-pattern anti-OVA IgG response after receiving naive splenocytes, even without further OVA challenge. Measured 125I-labeled OVA at the footpad injection site was minimal and below the threshold needed to trigger a de novo response, arguing against residual antigen. Transfer of splenocytes or draining-lymph-node cells from primed SCID mice into naive mice reproduced the memory-like response, whereas bone-marrow cells did not, and removing adherent cells (presumed APCs) abolished it. In a further step, bone-marrow-derived dendritic cells primed in vitro and transferred into SCID mice also conferred higher anti-OVA titers than non-primed dendritic cells. The author reads these results as evidence that dendritic cells and other APCs can present their epitope loads to naive T cells in an enhanced, memory-like manner for months, independent of lymphocytes.","pith_inferences":["The 1990s data, if reproduced with modern lineage tracing and single-cell sequencing, could distinguish genuine antigen-specific APC memory from trained immunity driven by epigenetic or metabolic changes.","A testable extension would determine whether the memory-like state requires retained peptide–MHC complexes on the APC surface or instead reflects a permanently altered activation program.","The author's proposal implies APC memory could be transferred between individuals by cell transfer, which might eventually be exploited therapeutically.","If APC memory exists in humans, it would be relevant to vaccine dosing schedules and to immune monitoring of patients with depleted lymphocytes."],"forward_implications":["Vaccines could be designed to induce long-lived memory in antigen-presenting cells, potentially extending protection beyond T and B cell memory.","Immunological memory would include a third cellular class, changing how immune history is modelled and measured.","APC memory might explain some heterologous immunity and cross-protection against unrelated pathogens.","If APCs carry antigenic epitopes for months, strategies that deplete or modulate dendritic cells could alter long-term vaccine responsiveness.","The SCID-mouse design offers a clean assay for memory that is independent of lymphocytes."],"supporting_citations":[{"why":"Shows metabolic induction of trained immunity in monocytes, a precedent that innate cells can build memory.","marker":"1"},{"why":"Documents alternative activation of macrophages, supporting the idea that innate cells adapt to prior stimuli.","marker":"2"},{"why":"Reports BCG educating hematopoietic stem cells, extending trained immunity to progenitor cells.","marker":"3"},{"why":"Describes modulation of myelopoiesis progenitors in trained immunity, further evidence for innate memory.","marker":"4"},{"why":"Introduces the concept of trained immunity as a memory for innate host defense.","marker":"5"},{"why":"Broadly outlines how training unlocks innate potential, providing a conceptual frame for APC memory.","marker":"6"},{"why":"Demonstrates memory-like dendritic cell responses in vivo with histone modification, the most direct precedent for the paper's APC-memory hypothesis.","marker":"7"}],"fun_headline_variants":["Dendritic cells recall antigens without T or B cells","APCs mount memory-like responses sans lymphocytes","SCID study: DCs keep immune memory for months","Antigen-presenting cells show antigen-specific recall","DC memory persists in SCID mice for months"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the amount of antigen remaining three months after priming is truly negligible everywhere in the body, but only the footpad injection site was measured, so antigen trapped in other tissues could still be present and cause the secondary-like response.","fun_headline_variants_meta":{"raw":{"variants":["Dendritic cells recall antigens without T or B cells","APCs mount memory-like responses sans lymphocytes","SCID study: DCs keep immune memory for months","Antigen-presenting cells show antigen-specific recall","DC memory persists in SCID mice for months"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000188,"raw_usage":{"total_tokens":1382,"prompt_tokens":1042,"completion_tokens":340,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":658,"completion_tokens_details":{"reasoning_tokens":266}},"tokens_in":658,"tokens_out":340,"duration_ms":4292,"temperature":1.0,"reasoning_tokens":266,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:05:17.125634+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the 125I radioactivity in regional lymph nodes, spleen, liver, and other organs at three months after priming in SCID mice; if intact OVA or immunogenic fragments above the priming threshold exist anywhere, the memory-like response may be a primary response boosted by persistent antigen rather than APC memory. Alternatively, adoptively transfer primed APCs that have been fixed with paraformaldehyde or treated with acid to strip surface peptides; if memory survives such treatment, it would argue against epitope presentation as the mechanism.","supporting_citations":[{"cited_title":"Metabolic Induction of Trained Immunity through the Mevalonate Pathway","cited_arxiv_id":null,"evidence_quote":"Shows metabolic induction of trained immunity in monocytes, a precedent that innate cells can build memory."},{"cited_title":"Alternative activation of macrophages: mechanism and functions","cited_arxiv_id":null,"evidence_quote":"Documents alternative activation of macrophages, supporting the idea that innate cells adapt to prior stimuli."},{"cited_title":"BCG Educates Hematopoietic Stem Cells to Generate Protective Innate Immunity against Tuberculosis","cited_arxiv_id":null,"evidence_quote":"Reports BCG educating hematopoietic stem cells, extending trained immunity to progenitor cells."},{"cited_title":"Modulation of Myelopoiesis Progenitors Is an Integral Component of Trained Immunity","cited_arxiv_id":null,"evidence_quote":"Describes modulation of myelopoiesis progenitors in trained immunity, further evidence for innate memory."},{"cited_title":"Trained immunity: a memory for innate host defense","cited_arxiv_id":null,"evidence_quote":"Introduces the concept of trained immunity as a memory for innate host defense."},{"cited_title":"Immune Training Unlocks Innate Potential","cited_arxiv_id":null,"evidence_quote":"Broadly outlines how training unlocks innate potential, providing a conceptual frame for APC memory."},{"cited_title":"Induction of memory-like dendritic cell responses in vivo","cited_arxiv_id":null,"evidence_quote":"Demonstrates memory-like dendritic cell responses in vivo with histone modification, the most direct precedent for the paper's APC-memory hypothesis."}],"review_version":1}