{"id":"dd3e7541-edc9-48c6-b169-3f7584a1db56","arxiv_id":"1908.03725","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Single mouse fibroblasts compression stiffen under uniaxial squeeze, and a semiflexible cortex loop around an incompressible interior matches the measured stress-strain curve to about 35% strain.","lead":"Mouse embryo fibroblasts stiffen, not soften, when squeezed by an atomic force microscope bead, and the authors propose that an incompressible cell interior wrapped in a bendable cortex explains the response up to about 35 percent compression. The result offers a possible cell-level origin for compression stiffening seen in tissues and suggests that organelle-packed cells carry a mechanical fingerprint.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The mEF stiffening claim rests on an unvalidated spherical-cap stress normalization; a linear-elastic control could show the curvature is geometric.","rationale":"The paper's central experimental observation is that mEFs compression stiffen. I read the protocol in Section II as defining stress by F/(2πrh), which is an assumption the authors state explicitly. The load-bearing question is whether this normalization produces the reported curvature even for a passive material. The authors' own PAA-gel control is a different geometry (rheometer plates), so it cannot rule out a geometric artifact in the AFM measurement. This is a more basic concern than the reader's: even if the cell has no bulk cytoskeletal network, the loop-model comparison in Fig. 6 is only meaningful if the experimental stress-strain curve is a material property rather than an artifact of the area model. I am not asserting the experiment is wrong; I am proposing a specific reanalysis and control that would settle it. The reader's CONDITIONAL verdict is appropriate, but the condition should include validating the stress conversion. The analytic calculations in Appendix A and the fibrin/bead experiments provide independent support for the general mechanisms, so I would not reject the paper.","tokens_in":122,"tokens_out":10806,"duration_ms":185628,"concrete_test":"Recompute the mEF stress–strain curves from the raw AFM force–distance data using an alternative, physically motivated contact area in place of A = 2πrh — e.g., the projected contact area a = π(2rh − h²) or the cell cross-sectional area measured from bright-field images — and test whether the upward curvature beyond ~20% strain persists. If the curvature disappears or changes sign under a plausible area model, the compression-stiffening claim is an artifact of the spherical-cap normalization; if it persists across area models, the material-level stiffening is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section II converts AFM force–distance data to stress as σ = F/(2πrh), where h is indentation depth and r = 12.5 µm is the sphere radius, and strain as ε = h/H0. Thus the reported stress–strain curve is proportional to F(ε)/ε. The spherical-cap area 2πrh is the area the indenter would penetrate into a flat rigid surface; it is not the actual contact area of a compliant rounded cell deformed between the sphere and the dish. For a linear-elastic half-space, Hertz contact gives F ∝ ε^{3/2}, so the reported σ ∝ ε^{1/2} would have decreasing slope; the observed upward curvature can therefore arise from the choice of area model, from finite-thickness/confinement effects, or from material stiffening. The paper provides no control in which the identical AFM sphere protocol and the σ = F/(2πrh) analysis are applied to a passive linear-elastic material of comparable size. The PAA-gel control in Appendix B3 is a rheometer plate compression, not the AFM spherical-indenter geometry, so it does not validate the stress conversion. If the upward curvature in σ(ε) is an artifact of the area normalization, then the claimed mEF compression stiffening—the paper's starting point—does not hold, and the loop-model comparison in Fig. 6 is fitting an artifact. The weakest assumption is therefore not only whether a bulk cytoskeletal network exists, but whether the measured stress–strain curve represents material response at all.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports AFM uniaxial compression experiments on rounded mouse embryonic fibroblasts, finding a compressive stress–strain curve that bends upward with stiffening onset near γc ≈ 20%. To explain this, the authors study three mechanical models: (i) a two-dimensional semiflexible polymer loop enclosing an incompressible area, representing the actomyosin cortex with viscous interior; (ii) a diluted triangular fiber network with embedded area-conserving loops, representing organelles and vesicles; and (iii) a fiber network with angle-constraining crosslinks. They report that the single-loop model matches the mEF experiment up to about 35% strain with one dimensionless parameter κ̃, and that the fiber-network-with-loops model captures compression stiffening observed in fibrin networks with adherent beads. Analytical calculations in the appendices support the numerical energy minimizations.","tokens_in":23199,"tokens_out":5235,"duration_ms":54995,"significance":"If the mEF result is a genuine material response, it establishes single-cell compression stiffening on fast timescales, in contrast to expectations from semiflexible polymer networks, and it would be directly relevant to tissue-scale compression stiffening. The study's strengths include explicit numerical simulations with error bars and system-size checks, analytical treatments of the loop model in Appendix A, a separate fibrin-bead experiment that provides a falsifiable test of the area-conserving-loop mechanism, and a PAA-gel control in Appendix B3 showing no stiffening when fiber bending is negligible. However, the load-bearing experimental stress conversion and the number of effective fitting parameters in the central model–experiment comparison must be addressed before the quantitative claims are secure.","major_comments":[{"comment":"The reported stress is computed as σ = F/(2πrh), with r = 12.5 µm and h the indentation depth, while the strain is ε = h/H0. This makes the reported stress proportional to F(ε)/ε. For a linear-elastic half-space indented by a sphere, Hertz contact gives F ∝ ε^{3/2} and hence the reported stress would scale as ε^{1/2}, which has decreasing slope. The observed upward curvature could therefore arise from the assumed contact-area model, finite-thickness or confinement effects, or genuine material stiffening. The PAA-gel control in Appendix B3 uses rheometer plate compression, not the AFM spherical-indenter geometry, so it does not validate the stress conversion used for the cells. A control with the identical AFM sphere protocol and the same σ = F/(2πrh) analysis applied to a passive linear-elastic material of comparable size is needed to support the paper's central experimental claim.","section":"Section II (Whole cell compression)"},{"comment":"The text states that the loop-model comparison has 'only one free parameter, κ̃,' but the experimental curve is transformed by subtracting the pre-stress and normalizing the stress by that same value. The subtracted offset and the stress scale are determined from the data being explained, and the model stress is separately normalized to the experimental scale. Therefore the comparison effectively uses more than one fitted quantity, and the agreement in Fig. 6 is a weaker test than claimed. The authors should either report the fit with a pre-specified offset and scale, or explicitly count the offset and scale as fitted parameters when assessing the agreement.","section":"Section V, Fig. 6"},{"comment":"The selection of the single-loop cortex model as the explanation up to 35% strain rests on the stated assumption: 'Since we do not know directly whether or not there is a bulk, rigid cytoskeletal network, let us assume there is not.' This assumption is load-bearing: if a spanning fiber network contributes appreciably to the mEF compression response, the fitted κ̃ and the loop-model agreement could be coincidental. The manuscript does not present any experimental test of this assumption, such as comparing wild-type cells with vimentin-null or actin-disrupted cells under the same compression protocol. Given that the authors themselves identify this as an unknown, the conclusion that the cortex is the dominant contributor up to 35% strain should be either explicitly tested or substantially weakened.","section":"Section V (Comparison with Experiments)"}],"minor_comments":[{"comment":"In the paragraph following the definition of Ho+fn, the text says 'See Fig. 3b for the stress-strain curves' when discussing the fiber network without loops; the relevant panel appears to be Fig. 4b, not Fig. 3b.","section":"Section IV"},{"comment":"The sentence 'The data was obtained from 10 cells and averaged over with the error bar denoting the standard deviation' should be rephrased as 'Data were averaged over 10 cells, with error bars denoting the standard deviation.'","section":"Section II"},{"comment":"There are two typographical issues: 'mouse embryonic fibroblastas' should be 'mouse embryonic fibroblasts,' and 'experimental experimental techniques' contains a duplicated word.","section":"Section VI"},{"comment":"The caption of Fig. 4b states that the analytical fits 'are scaled here to fit with the numerical data,' so the comparison between the numerical network compression and the analytical model is not parameter-free; the scaling factor should be reported.","section":"Appendix B1 and Fig. 4b"}],"recommendation":"major_revision","confidential_remarks":"The paper has a strong modeling component and a plausible experimental observation, but the core experimental claim needs a control with the same AFM spherical-indenter protocol on a linear-elastic material. The 'one free parameter' statement in the Fig. 6 comparison should be corrected. These issues are fixable within the scope of the manuscript, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper reports a new AFM finding—mouse embryonic fibroblasts compression stiffen under spherical-indenter compression—and it offers two new model mechanisms for compression stiffening: area-conserving loops embedded in a fiber network, and angle-constraining crosslinks. That is a real contribution. The modeling is careful, the analytic calculations in the appendices are credible, and the authors are unusually honest about what they cannot rule out.\n\nThe thing to know before reading closely is that the experimental stress–strain curve, which anchors the whole paper, is built on a stress conversion that is not validated. The AFM data are converted to stress as σ = F/(2πrh), where r is the sphere radius and h is indentation depth. That is the area of the indenter cap, not the actual contact area for a compliant rounded cell squeezed between a sphere and a dish. For a linear-elastic half-space, Hertz contact gives F ∝ h^{3/2} and therefore σ ∝ h^{1/2}—a concave-down curve with decreasing slope. The paper's concave-up stress–strain curve could be real material stiffening, but it could also come from the area assumption or from finite-thickness/confinement effects. There is no control run with the same AFM sphere and the same analysis on a passive linear-elastic material. The PAA-gel control in Appendix B3 is a rheometer plate compression, not the AFM geometry, so it does not settle this.\n\nThe loop-model comparison in Figure 6 is suggestive, not conclusive. It uses one fitted stiffness ratio and normalizes by the experimental pre-stress; that can absorb a fair amount of mismatch. The authors also assume away a bulk cytoskeletal network—Section V says 'let us assume there is not'—and they explicitly say the other two mechanisms cannot be ruled out. Those are limitations, not fatal flaws, but they keep the paper at the level of a plausible mechanism rather than a demonstrated one.\n\nWhere the paper earns its keep: the mEF measurement is new, the two network mechanisms are new, and the analytic small-strain calculations in the appendices give the reader a way to check the model behavior independently. The fibrin-bead comparisons are a good attempt to test the organelle-loop idea in a controlled system.\n\nI would send this to a serious referee. The novelty and modeling effort justify the time. But the referee should require either a control experiment that validates the stress conversion or a much more careful treatment of the contact mechanics before the central claim is taken as established. I would not cite it in my own work until that is resolved.","headline":"New AFM compression result and two new model mechanisms, but the stress conversion is unvalidated, so the central claim needs a control experiment.","tokens_in":23786,"tokens_out":4304,"would_cite":false,"duration_ms":47525,"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":"Mouse embryonic fibroblasts compression stiffen under uniaxial compression, and a single semiflexible polymer loop model captures the measured stress-strain curve up to about 35% strain.","keywords":["compression stiffening","single-cell mechanics","mouse embryonic fibroblasts","atomic force microscopy","semiflexible polymer loop","actomyosin cortex","area-conserving loops","fiber network mechanics"],"falsifier":"Compress mouse embryonic fibroblasts after selectively disrupting the actomyosin cortex, for instance with an actin-depolymerizing drug, leaving the interior intact: the single-loop model predicts that compression stiffening should be abolished or shifted to substantially larger strains, whereas survival of stiffening would show that a bulk fiber network is load-bearing.","tokens_in":1759,"feed_emoji":"🧫","tokens_out":1846,"duration_ms":115568,"temperature":0.7,"pith_summary":"Single living cells are generally expected to soften when squeezed, because the semiflexible biopolymer networks that give them structure buckle under compression. This paper reports the opposite for mouse embryonic fibroblasts: atomic force microscopy compression shows their compressive modulus rising with strain, with stiffening beginning around 20% strain. To explain the result, the authors model the cell as a semiflexible polymer loop, the actomyosin cortex, enclosing an incompressible viscous interior, and show that this single-loop model matches the measured stress-strain curve up to roughly 35% compressive strain with one free dimensionless parameter. They also show that fiber networks containing area-conserving loops, representing fluid organelles and vesicles, stiffen by forcing fibers to bend around the loops, and that the stress-strain response is sensitive to the loops' packing fraction and size distribution. The finding matters because it identifies a fast, non-remodeling mechanical response that protects cells under compression and offers a possible building-block mechanism for tissue compression stiffening.","feed_headline":"Cells stiffen when squeezed; a loop model fits the data to 35 percent","feed_subtitle":"AFM compression data on mouse embryo cells match a semiflexible cortex loop around an incompressible interior up to 35% strain.","key_machinery":"The central object is the semiflexible polymer loop Hamiltonian\n$$H_{v+c}=\\frac12 K_{cf}\\sum_{\\langle ij\\rangle}(l_{ij}-l_0)^2+\\frac12 K_{sf}\\sum_{\\langle ijk\\rangle}(\\theta_{ijk}-\\theta_0)^2+\\$\\lambda$(A-A_0),$$\na closed chain of stretchable springs and bending springs whose enclosed area is held fixed by a Lagrange multiplier. The area constraint models an incompressible viscous cytoplasm and supplies an effective outward pressure that prevents the loop from buckling in plane, which is why the loop compression stiffens instead of softening. The dimensionless ratio $\\tilde{\\kappa}=K_{cf}l_0^2/K_{sf}$ controls whether stretching or bending dominates: as $\\tilde{\\kappa}\\to 0$ the stress grows as $\\gamma^3$ with no linear regime, while finite bending produces the linear small-strain response and delayed stiffening seen in the mEF data at $\\tilde{\\kappa}=0.768$. Transplanting the same loop concept into a two-dimensional fiber network, as area-conserving inclusions and as angular constraints at the mesh scale, generates the bending-driven stiffening mechanisms that apply when a bulk cytoskeletal network or an in vitro fibrin network is present.","core_discovery":"Mouse embryonic fibroblasts compression stiffen under uniaxial compression imposed by an atomic force microscope, with stiffening setting in at $\\gamma_c \\approx 20\\%$ compressive strain, in contrast to the compression softening expected from semiflexible biopolymer networks. The paper's central quantitative claim is that, up to about 35% compressive strain, the measured stress-strain curve is well described by a single semiflexible polymer loop model of the actomyosin cortex enclosing an incompressible fluid, with only one free parameter, $\\tilde{\\kappa} = 0.768$, the ratio of stretching to bending stiffness of the loop. In this model, the incompressible interior acts as an outward pressure that suppresses in-plane buckling, so the loop stiffens rather than collapses; with finite bending stiffness the stress-strain curve has a linear small-strain regime followed by nonlinear stiffening, as observed. The paper additionally establishes two bulk-network mechanisms for compression stiffening: area-conserving loops, modeling fluid organelles and vesicles, interspersed in a fiber network force fibers into non-affine bending and convert the network's compression softening into stiffening even at small packing fractions, and angularly constrained crosslinks distort under compression to produce stiffening without organelles. For fibrin networks embedded with adherent beads, the area-conserving-loop model captures the stiffening and stress magnitude, with a later onset of stiffening in the model than in the experiment.","pith_inferences":["Editorial inference: if the single-loop description is the operative mechanism, the onset strain $\\gamma_c$ and fitted $\\tilde{\\kappa}$ should shift systematically when cortex composition is altered, so compression tests on cells with more or less actin-crosslinking protein would provide a direct check.","Editorial inference: the mechanical-fingerprint result suggests compression tests could serve as a label-free assay for organelle size and vesicle content; changing organelle size osmotically should move $\\gamma_c$ and the stress magnitude in a way the area-conserving-loop model can predict.","Editorial inference: the paper's two-dimensional-to-three-dimensional argument implies a direct three-dimensional loop-shell model should produce quantitatively different stress at large strain; comparing enucleated cells with whole cells would isolate whether the deviation above 35% strain comes from the nucleus or from the percolating bulk network."],"forward_implications":["Single cells can stiffen under compression on short timescales, without waiting for cytoskeletal reorganization, giving them a fast mechanical defense against large homeostatic pressures.","For cortex-dominated cells, uniaxial compression curves collapse onto a one-parameter family; measuring $\\tilde{\\kappa}$ from a compression test gives a readout of the balance between cortex stretching and bending.","Only a small number of fluid-like organelles or vesicles, with packing fraction around 0.04, can flip a fiber network from compression softening to compression stiffening by forcing fibers to bend.","Because the stress-strain curve depends on the packing fraction and size distribution of area-conserving loops, compression rheology can distinguish cell types or states by their internal organelle geometry.","Cellular compression stiffening is a plausible contributor to tissue compression stiffening, consistent with the observation that decellularized tissue loses much of its stiffening response."],"supporting_citations":[{"why":"Supplies the AFM whole-cell compression protocol, including constant-force indentation of rounded cells, from which the stress-strain data are obtained.","marker":"[23]"},{"why":"Provides the fibrin-network experiments with embedded dextran beads and shows that compression stiffening appears with adherent beads but not inert beads, the direct composite-network comparison.","marker":"[24]"},{"why":"Documents that in vitro semiflexible biopolymer networks typically compression soften, the baseline expectation the cell measurements overturn.","marker":"[15]"},{"why":"Supplies the standard semiflexible-polymer-network modeling framework with freely rotating crosslinks that the fiber-network models extend.","marker":"[17]"},{"why":"Provides the Maxwell constraint-counting logic used to explain why area conservation creates rigidity in an otherwise floppy loop of central-force springs.","marker":"[27]"},{"why":"Reports cubic force-strain behavior in compressed T-lymphoma cells, which the no-bending ($K_{sf}=0$) loop limit reproduces.","marker":"[28]"}],"fun_headline_variants":["Squeezed cells stiffen, not soften: loop model fits to 35%","Single loop model matches cell stiffening to 35% strain","Compression stiffening in cells explained by loop model up to 35%","Cells defy softening: loop model fits compression stiffening to 35%"],"cache_read_input_tokens":25856,"weakest_assumption_plain":"The quantitative fit assumes that a squeezed cell behaves like independent flat slices of an incompressible fluid wrapped in a cortex, with no load-bearing fiber network spanning the cell's interior.","fun_headline_variants_meta":{"raw":{"variants":["Squeezed cells stiffen, not soften: loop model fits to 35%","Single loop model matches cell stiffening to 35% strain","Compression stiffening in cells explained by loop model up to 35%","Cells defy softening: loop model fits compression stiffening to 35%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00122,"raw_usage":{"total_tokens":5125,"prompt_tokens":1162,"completion_tokens":3963,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":778,"completion_tokens_details":{"reasoning_tokens":3882}},"tokens_in":778,"tokens_out":3963,"duration_ms":31031,"temperature":1.0,"reasoning_tokens":3882,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:03:57.281999+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compress mouse embryonic fibroblasts after selectively disrupting the actomyosin cortex, for instance with an actin-depolymerizing drug, leaving the interior intact: the single-loop model predicts that compression stiffening should be abolished or shifted to substantially larger strains, whereas survival of stiffening would show that a bulk fiber network is load-bearing.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the AFM whole-cell compression protocol, including constant-force indentation of rounded cells, from which the stress-strain data are obtained."},{"cited_title":"Pilyugina, B","cited_arxiv_id":null,"evidence_quote":"Provides the fibrin-network experiments with embedded dextran beads and shows that compression stiffening appears with adherent beads but not inert beads, the direct composite-network comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents that in vitro semiflexible biopolymer networks typically compression soften, the baseline expectation the cell measurements overturn."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the standard semiflexible-polymer-network modeling framework with freely rotating crosslinks that the fiber-network models extend."},{"cited_title":"Hatami-Marbini, Phys","cited_arxiv_id":null,"evidence_quote":"Provides the Maxwell constraint-counting logic used to explain why area conservation creates rigidity in an otherwise floppy loop of central-force springs."},{"cited_title":"Picariello and et al., PNAS , 201902847 (2019)","cited_arxiv_id":null,"evidence_quote":"Reports cubic force-strain behavior in compressed T-lymphoma cells, which the no-bending ($K_{sf}=0$) loop limit reproduces."}],"review_version":1}