{"id":"ad272dd9-57eb-443c-ba5f-366b8b4bfa2f","arxiv_id":"2501.00086","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"3D-printed hepatocyte-HUVEC 'discoids' in packed PEG microgels keep albumin, urea, ADME gene expression, and CYP enzyme activity for weeks, and outperform their own printed spheres.","lead":"Liver cells printed into thin discs inside a bed of gel beads stayed alive and made liver proteins for weeks, at a speed of about one disc per minute in standard lab plates. This offers drug developers a simple, perfusion-free human liver model for toxicity testing, if the results hold up in head-to-head drug validation.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The perfusion-free, 21-day claim depends on unverified oxygen transport in 200 µm discoids; Fig. 5c's size-dependent urea drop and the absence of hypoxia/thickness measurements leave this load-bearing assumption unconfirmed.","rationale":"The reader's weakest assumption already identifies oxygen/nutrient diffusion as the load-bearing condition, and I agree. The strongest claim bundles three outcomes: stable albumin/urea, ADME profile, and CYP metabolite formation. All three require living, functional hepatocytes; if a 2.5 mm static disc develops a hypoxic core by D21, the observed functions may come from a peripheral viable shell, so the model would not be 'perfusion-free' in the claimed size range. The paper has real strengths: direct LC-MS metabolite measurements, n=4-5 for function assays, external comparators (liver, FTH, HepatoPac), and a reproducible printing protocol. It also includes a self-identified limitation that perfusion might enhance function (Discussion), which is honest but does not replace direct evidence that static transport is adequate. My proposed test is feasible with standard hypoxia staining and oxygen microsensors; it directly settles whether the diffusion assumption holds. Because this is the same condition the reader flagged, and because the available data are suggestive but not conclusive, the CONDITIONAL verdict should stand unchanged.","tokens_in":21710,"tokens_out":6227,"duration_ms":71170,"concrete_test":"Perform pimonidazole or HIF-1α staining plus O2 microelectrode transects on H:Hu discoids of 1.5, 2.0, and 2.5 mm diameter at D7, D14, and D21, and measure disc thickness/cell density by confocal or two-photon imaging at the same timepoints. If hypoxic regions or central necrosis appear in 2.0/2.5 mm discs, or if thickness increases beyond ~200 µm, the perfusion-free, 21-day claim is not supported for the full stated size range.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that H:Hu discoids (200 µm thick, up to 2.5 mm diameter) remain functional for 21–28 days in static culture without perfusion. This requires oxygen and nutrients to reach all cells by diffusion. The paper's support is indirect: cell numbers from DNA content are stable and albumin/urea remain high, but DNA-based counts do not distinguish live from dead cells, supernatant function could be produced by a viable rim, and no oxygen gradients, hypoxia markers (HIF-1α, pimonidazole), central necrosis, or LDH release are reported. The high hydraulic permeability of packed spherical microgels (Supplementary S1) does not itself establish adequate diffusive transport in static culture; permeability matters for flow, whereas static oxygen delivery is governed by diffusivity and the tissue's geometric and metabolic properties. The size-dependence in Section 2.4/Fig. 5c—urea significantly lower in 2.0 and 2.5 mm versus 1.5 mm discs at D7, D14, and D21—is consistent with an emerging diffusion limitation, yet the paper does not test this interpretation. In addition, HUVEC-containing discoids compact to ~1.5-fold smaller area by D10 (Fig. 4d), and no post-print thickness or density measurements are provided; if compaction increases thickness/local cell density, the effective oxygen diffusion distance may exceed the nominal 200 µm. Should this transport assumption fail at larger diameters, higher densities, or later times, the 'no perfusion needed' claim and the scalable 2.5 mm model would not generalize.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript describes a 3D bioprinting approach for producing human liver tissue models (\"discoids\") — 200 µm thick and 1–3 mm in diameter — printed into a packed PEG microgel support medium. The authors report precise and accurate printing (≈4% diameter error, ~100 tissues/hour), stable albumin and urea synthesis over 3–4 weeks for hepatocyte/HUVEC co-cultures, size- and geometry-dependent functional differences, ADME gene expression clustering close to human liver controls, and LC-MS-measured CYP/UGT enzyme activities after 14 days. The paper argues that the high hydraulic permeability of the spherical microgel pack allows long-term culture without perfusion.","tokens_in":21875,"tokens_out":8478,"duration_ms":78082,"significance":"If the claims are substantiated, the system would provide an accessible, static-culture, millimeter-scale liver model suitable for pharmaceutical ADME and toxicology screening. The strengths include direct LC-MS-based measurement of five enzyme activities, benchmarking against freshly thawed hepatocytes, liver biopsy tissue, and HepatoPac, and a systematic comparison of co-culture compositions. The printing precision statistics are also a useful contribution to embedded bioprinting. However, the central claims of perfusion-free culture and functional equivalence to human liver rest on indirect or incomplete evidence (no oxygen/hypoxia data, DNA-based viability, limited ADME statistics), so the current version does not yet establish these claims at the level asserted.","major_comments":[{"comment":"The central claim that 200-µm-thick discoids up to 2.5 mm in diameter can be cultured without perfusion is not supported by the presented data. The authors attribute the design to diffusion limits but provide no oxygen gradient measurements, hypoxia markers (e.g., HIF-1α, pimonidazole), or evidence of central necrosis; the only viability-related metric is total DNA content. Moreover, the size-dependence in Fig. 5c — urea synthesis significantly lower in 2.0 and 2.5 mm discs than in 1.5 mm discs at D7, D14, and D21 — is consistent with an emerging diffusion limitation, yet this interpretation is not tested. Compaction of H:Hu constructs to ~1.5-fold smaller area by D10 (Fig. 4d) may increase thickness and effective diffusion distance, and no post-print thickness data are reported. The claim that the microgel medium 'does not need to be perfused' therefore requires direct transport evidence or at minimum regional live/dead analysis.","section":"Section 1, Section 2.4, Fig. 5c"},{"comment":"Cell numbers are estimated from PicoGreen DNA content, which cannot distinguish live from dead cells. Stable DNA-based cell counts therefore do not establish that viable cell numbers remain constant for 21 days; dead or lysed cells can contribute to the DNA signal, and proliferation can mask death. The conclusion in Section 2.3 that 'cell number within each group did not significantly change over the culture duration' and the related 'stable function' claims require direct viability measurements (live/dead staining, LDH release, ATP content, or apoptosis markers). Without them, the 21-day functionality claim rests on supernatant activity that could be supported by a viable rim.","section":"Section 4.5, Fig. 4a, Fig. 5d"},{"comment":"The abstract's claim that the discoids are 'outperforming spheroid tissue models' is not supported as stated. The comparison is only against spheroids printed from the same bioink in the same microgel medium, not against established spheroid culture models (e.g., primary hepatocyte spheroids or commercial systems). The sphere constructs also had ~4-fold higher cell numbers, and their lower per-cell albumin/urea synthesis and 46% cell-number decrease by D14 are from DNA-based measurements (see above). The claim should be limited to 'outperforming printed sphere constructs of the same bioink' or supported by a direct comparison with a recognized spheroid model.","section":"Abstract, Section 2.4, Fig. 5d-f"},{"comment":"The claim that the tissues express 'more than 100 genes associated with ADME at levels within the range of human liver' is not demonstrated. Fig. 6a shows relative expression for only a subset of genes; the heatmap (Fig. 6b) is based on hierarchical clustering of 115 probes from n=2 independent experiments, with batch correction and normalization. Clustering 'closer to liver controls than HepatoPac' is not a statistical equivalence test, and no criterion is given for 'within the range of human liver.' Quantitative thresholds (e.g., within 2-fold of donor liver for a defined gene set) and per-gene variability are needed to support this claim.","section":"Section 2.5, Fig. 6, Methods 4.7"},{"comment":"The permeability measurements show that packed spherical PEG microgels are 15–40x more permeable to pressure-driven flow than irregular microgel packs, but this Darcy-scale hydraulic permeability does not directly measure diffusive transport of oxygen or nutrients under static conditions. Static-culture oxygen delivery is governed by the diffusion coefficient and the geometry of the tissue, not by the hydraulic permeability of the surrounding support. The argument that high permeability 'side-steps the need to perfuse' (Discussion) is therefore a non-sequitur unless a link between microgel porosity and effective diffusivity is demonstrated.","section":"Supplementary S1, Discussion"}],"minor_comments":[{"comment":"The displayed equations for the spiral diameter are garbled (e.g., '2 2 44 mm cA d cdπ' and '22 m cAd π= +'); these need proper typesetting and clear definition of variables, including the calibration factor f = 1.1.","section":"Section 2.2"},{"comment":"The probe substrate concentrations are given as 100 mM, 25 mM, 50 mM, 15 mM, and 10 mM; these are almost certainly intended as µM and should be corrected, since mM concentrations would be cytotoxic and outside normal assay ranges.","section":"Methods 4.8"},{"comment":"The caption lists 'Y-2732', which is a typo for Y-27632, as used in the main text (Section 2.6).","section":"Supplementary Fig. S2 caption"},{"comment":"The formula for 'Theoretical Copy #: 2(40-CT value of target gene)/10K' is unclear; define the denominator and explain how copy numbers are normalized and then scaled.","section":"Methods 4.7"},{"comment":"The 'data collapse' rescales each curve by its own fitted slope; this does not by itself demonstrate a linear scaling law. The raw concentration-vs-time data with linear fits would be more informative to the reader.","section":"Section 2.6, Fig. 7b"}],"recommendation":"major_revision","confidential_remarks":"The manuscript has been reviewed on its scientific merits. I note that the abstract's claims (e.g., 'outperforming spheroid tissue models', 'within the range of human liver') are more sweeping than the data support and should be tempered. The central transport concern — the unverified oxygen diffusion argument for perfusion-free culture — is the main reason for major revision; adding direct oxygen/hypoxia/viability data would substantially strengthen the paper."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Sam,\n\nThis one is worth a look. The group's jammed-microgel printing platform was already established (their refs 28-33, and a liver model in ref 39), but the paper adds the systematic functional validation: 200 µm-thick discoids, 1.5-2.5 mm across, printed at ~100/hour into 96-well plates, stable albumin/urea for 3-4 weeks, ADME gene profiles in the human-liver range, and LC-MS-measured CYP/UGT metabolite formation at day 14. The CYP data are direct and reproducible on the page: n=4-5, clear methods, external references (liver biopsy, freshly thawed hepatocytes, HepatoPac). The HUVEC co-culture benefit is consistent, and the geometry comparison, though limited, is honestly described.\n\nSoft spots: the 'outperforming spheroid tissue models' claim overreaches. The comparator is their own printed sphere, not the established spheroid field; that's an overstatement. The bigger issue is the perfusion-free design. A 200 µm disc is right at the edge of reported oxygen diffusion limits, and the paper provides no oxygen gradients, hypoxia markers, or necrosis data. DNA-based cell counts don't distinguish a live rim from a dead core. The size-dependent urea drop at 2.0 and 2.5 mm diameter (Fig. 5c) is consistent with emerging diffusion limitation, and the paper doesn't test that interpretation. That's not fatal—200 µm may well be fine—but it's the load-bearing assumption and it's currently inferred, not measured. I'd want pimonidazole or HIF-1α staining on larger discs before believing the 'no perfusion, up to 2.5 mm' headline. Also, the substrate concentrations in Section 4.8 are off by orders of magnitude if they're meant to be µM (phenacetin at 100 mM would be absurd); a referee should catch that quickly. The ADME 'within range of human liver' rests on two independent experiments; fine for proof-of-concept, but the abstract states it as settled. And there's no validation against known hepatotoxicants, so the DILI framing is marker-based inference, not yet demonstrated.\n\nOverall the central claim holds up as a proof-of-concept. The platform is real, the data mostly solid, and the delivery—standard 96-well plates, no perfusion hardware—is genuinely useful for ADME screening. My recommendation: send it to peer review. A good referee can push for the oxygen-transport evidence and some language tightening without reopening the core results.","headline":"Solid functional validation of a microgel-printed liver discoid platform; the 'no perfusion needed' claim rests on an oxygen-transport assumption that the paper doesn't directly test.","tokens_in":22657,"tokens_out":3067,"would_cite":true,"duration_ms":27131,"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":"Thin 3D-bioprinted liver discs with endothelial cells in a microgel bed keep secreting albumin and urea and metabolizing drugs for 3-4 weeks without perfusion.","keywords":["liver tissue model","discoid","3D bioprinting","microgel support medium","3D cell culture","ADME gene expression","hepatocyte-HUVEC co-culture","drug metabolism assay"],"falsifier":"Look for a hypoxic or necrotic core in a 2.5-mm H:Hu discoid at day 21: stain with pimonidazole or HIF-1alpha, or measure an oxygen gradient with a microelectrode; if the center is hypoxic while albumin and urea output remains high, diffusion alone is not the mechanism keeping the tissue alive.","tokens_in":21341,"feed_emoji":"🧪","tokens_out":10017,"duration_ms":96436,"temperature":0.7,"pith_summary":"The paper sets out to show that a 3D-printed liver model can be made simple enough for routine drug screening without giving up liver-like function. The tissues are 200-micrometer-thick discs (discoids) of primary hepatocytes (liver cells) plus vessel-lining HUVECs, printed into a bed of packed spherical PEG microgels; the same bed that shapes the print then acts as the culture medium. Over 3-4 weeks the discs keep producing albumin and urea at stable rates, express more than 100 genes involved in absorbing, distributing, metabolizing, and excreting drugs (ADME) at levels comparable to human liver tissue, and convert probe drugs into metabolites through CYP and UGT enzymes. The authors report that the process is fast (about 100 tissues per hour) and dimensionally precise (about 4% diameter error), and that tissues can be assayed in standard 96-well plates without perfusion. If these results hold, the model offers a practical, reproducible platform for toxicity and metabolism testing.","feed_headline":"Thin printed liver discs keep working for 21 days, no pump needed","feed_subtitle":"Hepatocyte-endothelial discoids in microgel beds match human liver ADME genes and metabolize test drugs for weeks.","key_machinery":"The central object is the packed-microgel support medium: spherical polyethylene glycol (PEG) beads about 6 micrometers in diameter, packed at 5% w/w into a yield-stress solid, a material that flows only when pushed hard enough, with a storage modulus around 100 Pa and a yield stress around 2 Pa. It holds printed structures in place during fabrication, yet its interstitial pores give it a permeability that follows Kozeny-Carman scaling, $k \\sim d^2$, and is 15-40 times higher than irregular-particle packs. The second piece is the print path: a planar spiral that lays down a 200-micrometer-thick collagen-cell disc, a geometry that keeps the tissue within diffusion distance of the surrounding medium. The third piece is the H:Hu 2:1 cell ratio, whose cell-cell and cell-matrix contacts maintain hepatocyte phenotype.","core_discovery":"The core discovery is that geometry and support material, rather than perfusion, carry the model. A 200-micrometer-thick disc keeps every hepatocyte close to the culture medium, and the loosely packed spherical PEG microgels leave micron-scale pores that make the bed 15-40 times more permeable than the irregular-particle packs used in earlier embedded-printing media. In this environment, hepatocyte-HUVEC (H:Hu) co-culture outperforms hepatocyte-only, cholangiocyte-containing, and spheroid formats: it produces higher albumin and urea, holds cell numbers steady for 21 days, and its ADME transcriptome clusters closest to human liver tissue, closer than the established HepatoPac comparison. After 14 days the printed discs still form metabolites through CYP1A2, CYP2C9, CYP2D6, CYP3A4, and UGT1A1, and adding Y-27632 plus HGF roughly doubles albumin output, raises urea, and boosts UGT1A1 activity by about 2.5-fold.","pith_inferences":["The paper never directly measures oxygen or hypoxia; a day-21 hypoxia stain of the largest (2.5-mm) discs would settle whether the no-perfusion claim holds at the core or only on the outer shell.","The HUVEC benefit is left mechanistically open; an immediate test the authors do not run is whether HUVEC-conditioned medium alone reproduces the albumin and urea boost, which would separate soluble-factor signaling from direct cell-cell contact.","If the ADME profile really sits within the human-liver range, the model should be challenged with a blinded panel of known hepatotoxic and non-toxic drugs to see whether metabolite formation and albumin/urea decline predict toxicity.","The Kozeny-Carman scaling the authors fit implies that larger microgels could support thicker or denser tissues; this is a design rule they observe but do not exploit, and it is directly testable by printing the same discoid in different bead sizes."],"forward_implications":["Because the microgel bed doubles as the culture medium, tissue fabrication, culture, and assaying happen in the same 96-well plate, removing a barrier to industrial adoption.","The size comparison shows 1.5-mm discoids make the most urea; cell density and tissue diameter are therefore controllable inputs for matching a desired metabolic output.","HUVEC co-culture is the composition that works best; adding cholangiocytes suppresses albumin and urea, so cell-ratio optimization is a direct route to tune the model.","Y-27632 plus HGF supplementation increases albumin about 2-fold, urea about 1.5-fold, and UGT1A1 metabolite formation about 2.5-fold, giving a chemically simple way to boost function.","The permeability result, $k \\sim d^2$ with spherical beads 15-40 times more permeable than irregular ones, means the support medium can be adjusted if future designs need perfusion or larger constructs."],"supporting_citations":[{"why":"Supplies the PEG microgel synthesis route and the measured particle-size distribution (median diameter 6.26 micrometers) on which the support-medium formulation is built.","marker":"[27]"},{"why":"Establishes that jammed microgels behave as liquid-like solids that support and hold 3D-printed cellular structures, the founding mechanism of this culture platform.","marker":"[28]"},{"why":"Provides the yielding and flow criteria for microgel packs near jamming, used to choose the PEG concentration and printing conditions.","marker":"[29]"},{"why":"Defines HepatoPac, the established primary-hepatocyte co-culture standard the paper compares its ADME gene profile against.","marker":"[8]"},{"why":"Shows perfused liver models raise viability and albumin/urea output, the benchmark that motivates the claim that perfusion is unnecessary in discoids.","marker":"[22]"},{"why":"Supplies the industrial CYP-induction practices and suspension-cell benchmark used to validate the LC-MS enzyme activity measurements.","marker":"[34]"},{"why":"Shows small molecules including Y-27632 maintain long-term primary hepatocyte function, grounding the supplementation experiments.","marker":"[35]"},{"why":"Identifies HGF as a hepatocyte mitogen and trophic factor, grounding the HGF-supplementation arm.","marker":"[36]"}],"fun_headline_variants":["Printed liver discs beat spheroids: 21 days of albumin, urea, drug metabolism","No pump needed: 3D-printed liver discs stay functional for weeks","Thin printed liver discoids outperform spheroids in long-term function","No perfusion? 3D-printed liver discs function for 21 days","3D-printed liver discs: drug metabolism for weeks without pumps"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire no-perfusion claim rests on the assumption that a 200-micrometer-thick, 1.5-2.5-millimeter-wide cell disc stays within oxygen and nutrient diffusion limits for 21 days, because the study reports stable DNA-based cell counts and function but no direct oxygen, hypoxia, or central-necrosis measurements.","fun_headline_variants_meta":{"raw":{"variants":["Printed liver discs beat spheroids: 21 days of albumin, urea, drug metabolism","No pump needed: 3D-printed liver discs stay functional for weeks","Thin printed liver discoids outperform spheroids in long-term function","No perfusion? 3D-printed liver discs function for 21 days","3D-printed liver discs: drug metabolism for weeks without pumps"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0009,"raw_usage":{"total_tokens":3895,"prompt_tokens":988,"completion_tokens":2907,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":604,"completion_tokens_details":{"reasoning_tokens":2806}},"tokens_in":604,"tokens_out":2907,"duration_ms":20549,"temperature":1.0,"reasoning_tokens":2806,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T23:01:57.062748+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Look for a hypoxic or necrotic core in a 2.5-mm H:Hu discoid at day 21: stain with pimonidazole or HIF-1alpha, or measure an oxygen gradient with a microelectrode; if the center is hypoxic while albumin and urea output remains high, diffusion alone is not the mechanism keeping the tissue alive.","supporting_citations":[],"review_version":1}