{"id":"1fac1ff6-7aa7-46d2-85d2-3a1e256e6eaf","arxiv_id":"1908.02889","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":10,"one_line_summary":"A multiphase PDE model with chemotaxis and haptotaxis shows that a small smooth muscle cell population can form a stable collagen cap and that TGF-beta is critical for effective cap formation.","lead":"This paper develops a mathematical model of how smooth muscle cells build the protective fibrous cap over fatty plaques in arteries, driven by the growth factors PDGF and TGF-beta. It shows that a small population of cells can form a stable cap when TGF-beta is present, and that the model reproduces several observations from mouse experiments.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantitative 'small SMC population' claim depends on baseline ECM parameters calibrated to the same Reifenberg dataset used for validation; global sensitivity check needed.","rationale":"I read the paper as an internally consistent multiphase model whose main contribution is a mechanistic, spatially resolved account of SMC-driven fibrous cap formation, with a useful local steady-state analysis and a clear qualitative prediction about TGF-beta. The derivation of the PDE system is careful, and the analytical expressions for the optimal SMC fraction, despite a typographical issue in Eq. (53), yield a plausible and interpretable result. The fixed-domain assumption and the acknowledged ill-posedness for larger chi_rho are limitations but are explicitly discussed and do not, by themselves, threaten the central qualitative claims. The most load-bearing weakness is the parameterisation of the ECM remodelling rates. In particular, rs is stated to have been chosen to match the timescale of collagen accumulation reported in Reifenberg et al. (2012), and that same study is later used as the quantitative validation target. The numerical agreement in Section 3.2.1 is therefore not an independent confirmation of the model. Moreover, the analytical 'small SMC population' conclusion depends on ratios involving rs, rd, and beta_rho, all of which are listed in Table 1 without experimental references. If the true in vivo baseline synthesis rate were lower, the optimal SMC fraction could be substantially higher, weakening the paper's headline claim. This concern does not warrant rejection: the model's qualitative TGF-beta result is corroborated by independent experimental work, and the structure of the model is sound. But it does justify the reader's CONDITIONAL verdict, because the quantitative strength of the central claim has not been tested against parameter uncertainty. My proposed global sensitivity analysis is a concrete, feasible check that would either confirm that the conclusions are robust or show that they are artefacts of the base-case calibration.","tokens_in":34649,"tokens_out":8675,"duration_ms":85309,"concrete_test":"Run a global sensitivity analysis over the unmeasured parameters, for example rs in [0.9, 3.6], rd in [0.75, 3], beta_rho in [0.375, 1.5], alpha_P in [0.3, 1.1], alpha_T in [0, 5], and chi_rho in [0, 0.8], using Latin hypercube sampling while holding the reference-based parameters fixed. For each sample, simulate to t = 8 and record (i) total SMC and ECM volume fractions, (ii) cap-region V_rho(8; 0.2), and (iii) the change in V_rho when alpha_T is set to 0. Then compute the fraction of samples that reproduce the qualitative claims: ECM fraction at or above roughly 20% while SMC fraction is at or below roughly 10%, and a substantial reduction in cap ECM upon TGF-beta removal.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim has two legs: (a) a small SMC population suffices for a stable collagenous cap, and (b) TGF-beta is critical for cap formation and maintenance. Leg (b) is independently supported by the alpha_T = 0 simulation and by the cited experimental blockade studies (Mallat et al. 2001; Lutgens et al. 2002), so I do not see a serious objection there. Leg (a) is more fragile. The analytical optimum m_hat in Eq. (54) is decreasing in mu = Rs/Brho, and thus in the unmeasured baseline synthesis rate rs; Table 1 lists no experimental reference for rs, rd, or beta_rho. Section 2.6 states that rs was chosen 'sufficiently large to allow plaque collagen to accumulate on a timescale similar to that reported in Reifenberg et al. (2012)', and Section 3.2.1 then cites the same Reifenberg study as quantitative validation. The reported agreement of total SMC and ECM fractions (about 8.6% and 21.4% vs. 7% and 24%) is therefore partially baked into the parameter choice. The problem is not that the selected values are implausible; it is that the paper's most load-bearing quantitative prediction, that fewer than about 15-20% SMCs are optimal, has not been shown robust to the uncertainty in these baseline rates. This is a calibration-validation circularity that directly affects the headline claim, not merely a cosmetic weakness.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a one-dimensional multiphase PDE model of fibrous cap formation in atherosclerosis. It tracks SMC volume fraction, collagenous ECM volume fraction, a generic tissue phase, and quasi-steady PDGF and TGF-beta concentrations, with non-standard boundary conditions modeling endothelial growth-factor influx and SMC influx at the medial boundary. After nondimensionalization, the authors derive a local steady-state ECM relation, yielding explicit formulas for the SMC fraction that maximizes ECM at a point (Eqs. 54-55). Numerical simulations with base-case parameters reproduce the temporal pattern of SMC and collagen accumulation in ApoE mouse plaques, and sensitivity studies probe the roles of SMC haptotaxis and growth-factor influx. The paper claims that a relatively small SMC population can form a stable cap and that TGF-beta is critical for effective cap formation and maintenance.","tokens_in":34875,"tokens_out":5323,"duration_ms":51204,"significance":"If the quantitative claims are robust, this would be a valuable mechanistic contribution to atherosclerosis modeling: it combines multiphase mechanics with growth-factor-regulated ECM remodeling and provides explicit analytical expressions for the optimal SMC fraction. The qualitative TGF-beta result is supported by independent experimental studies (Mallat et al.; Lutgens et al.), and the local steady-state analysis is a useful interpretative tool. Strengths include the explicit analytical reduction (Eqs. 50-55), the detailed parameterization from in vitro and in vivo data, and the systematic sensitivity studies over alpha_P, alpha_T, and chi_rho. However, the central quantitative claim about small SMC populations rests on unmeasured baseline ECM parameters whose values are calibrated to the same dataset later used for validation; this needs to be addressed before the claim is fully supported.","major_comments":[{"comment":"The base-case parameter rs is set 'sufficiently large to allow plaque collagen to accumulate on a timescale similar to that reported in Reifenberg et al. (2012)' (Section 2.6, Table 1), and Section 3.2.1 then presents the agreement of total SMC and ECM fractions (8.6% and 21.4% versus 7% and 24%) as quantitative validation using the same Reifenberg et al. data. Because rs, rd, and beta_rho are not measured, the agreement is partly built into the parameter choice. Please add a global sensitivity analysis over rs, rd, and beta_rho (and ideally beta_m, rm, and the growth-factor uptake rates) and report the resulting ranges of steady-state SMC/ECM totals and of the cap-region V_m and V_rho, so that the 'small SMC population' conclusion can be checked against parameter uncertainty.","section":"Section 2.6 and Section 3.2.1"},{"comment":"The claim that a relatively small SMC volume fraction (0.1-0.2) maximizes ECM deposition is read off Figure 4 using base-case values. Equation (54) shows that m_hat depends on mu=R_s/B_rho and lambda=R_d/B_rho, and the optimum shifts upward as mu decreases; since mu is proportional to rs/beta_rho, the 'small SMC' conclusion is not robust without specifying the uncertainty in these baseline rates. I recommend either plotting m_hat and rho_hat over the plausible ranges of (rs, rd, beta_rho) or explicitly restricting the claim to the base-case parameter set.","section":"Section 3.1, Eqs. (54)-(55), and Figure 4"}],"minor_comments":[{"comment":"The final sentence, 'an important step towards the development of a comprehensive in silico plaque', is missing a noun; it should read 'comprehensive in silico plaque model'.","section":"Abstract"},{"comment":"Immediately after Eq. (17), 'immune cell ECM degration' should be 'degradation'.","section":"Section 2.1.3"},{"comment":"The assertion that the discriminant is 'trivially' strictly positive for all admissible parameter values is not demonstrated; a short derivation or reference would help.","section":"Section 3.1, Eq. (50)"},{"comment":"The statement that time points for the ECM panel do not correspond exactly to the other panels is vague; please list the actual time values used for each panel or clarify the convention.","section":"Section 3.2.1, Figure 6 caption"},{"comment":"The reported sensitivity simulation with beta_rho=1.5 (32% vs 38% ECM in the cap region) is not shown; either include it in the results or mark it consistently as 'results not shown'.","section":"Section 4.6"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of a mathematical biology journal. The main issue is the calibration-validation circularity for the baseline ECM rates; this is fixable with an additional sensitivity study and more careful wording. I would not reject on the basis of the TGF-beta conclusions, which are consistent with independent experiments and with the structure of the model."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nHere's my take on Watson et al. 1908.02889. It's a careful extension of the authors' two-phase model to include TGF-beta, explicit collagen dynamics, and haptotaxis. The genuinely new piece is the local steady-state analysis: for fixed growth factor concentrations, the ECM volume fraction has a biphasic dependence on SMC fraction, and the closed-form optimum m_hat (Eq. 54) is a useful tool for interpreting simulations. That is a real contribution, not just another PDE simulation.\n\nThe paper also does several things well. The PDE derivation is internally consistent and the reduction to two coupled equations is transparent. The authors are unusually honest about limitations: the fixed-domain issue, the acknowledged ill-posedness for strong haptotaxis, and the qualitative mismatch with the TGF-beta blockade experiments (they report 15% decrease vs 50% in vivo). The sensitivity studies over alpha_P and alpha_T are informative.\n\nThe soft spot is the one the stress-test flags. The baseline ECM synthesis rate rs is chosen in Section 2.6 so that collagen accumulates on the timescale reported in Reifenberg et al., and then the same Reifenberg data is used in Section 3.2.1 as quantitative validation. The quantitative match (8.6% SMC, 21.4% ECM vs 7%, 24%) is therefore partly baked in. This does not destroy the paper because the central qualitative claims—TGF-beta is critical, and a relatively small SMC population can support a cap—are robust to reasonable parameter variation, and the TGF-beta leg is independently supported by the alpha_T=0 simulation and by Mallat/Lutgens. But the precise \"15-20% SMC\" optimum is conditional on the unmeasured baseline rates rs, rd, beta_rho. A global sensitivity analysis or a parameter scan over the baseline ECM parameters would have materially strengthened the claim. The absence of code/data also makes it harder to probe that robustness.\n\nOverall: serious, honest work. The right referee would push for a sensitivity analysis of the baseline ECM parameters and a clearer separation of calibration from validation. I'd send it to peer review. I'd bring it to reading group, and if I worked in this area I'd cite the analytical result.","headline":"Solid model paper with a genuinely new analytical result; the quantitative validation is partly calibrated to the comparison data, but the qualitative TGF-beta story is independently supported and worth referee time.","tokens_in":35505,"tokens_out":1863,"would_cite":true,"duration_ms":18618,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35Q92","92C50"],"pacs":[],"model":"deepseek-v4-flash","headline":"A multiphase model of atherosclerotic cap formation claims that a stable collagenous fibrous cap can be built by a small population of smooth muscle cells, with TGF-β the decisive regulatory factor.","keywords":["atherosclerosis","fibrous cap","smooth muscle cells","TGF-β","PDGF","multiphase model","plaque stability","ApoE mouse"],"falsifier":"Measure the time course of SMC and collagen content in fibrous caps of ApoE-deficient mice after TGF-β blockade: the model predicts cap collagen falls roughly 40% while SMC content rises, whereas the cited TGF-β blockade experiments report about 50% collagen loss with no change in SMC content; the SMC trajectory is the distinguishing observable.","tokens_in":34353,"feed_emoji":"🫀","tokens_out":9907,"duration_ms":92136,"temperature":0.7,"pith_summary":"This paper claims that a stable, protective fibrous cap over an atherosclerotic plaque does not require a large population of vascular smooth muscle cells (SMCs); a relatively small cohort of SMCs, recruited by PDGF and driven by TGF-β, can deposit and maintain the collagen cap. The authors build a multiphase partial-differential-equation model of the intima with SMC, collagen, and generic tissue phases, coupled to diffusing growth factors and non-standard boundary conditions that let SMCs enter from the media in response to a PDGF gradient. Their analytical reduction shows that the collagen level at any point is capped by the local growth-factor concentrations, and that this cap is reached at a low SMC density; simulations reproduce the timing and final amounts of SMC and collagen measured in ApoE-deficient mouse plaques. The claim matters because plaque rupture, a trigger of heart attack and stroke, is governed by cap stability, and the model identifies TGF-β as the decisive factor in both forming and keeping the cap.","feed_headline":"Few smooth muscle cells can build a stable fibrous cap","feed_subtitle":"Multiphase model of mouse plaque growth pins TGF-β as the decisive factor for cap formation and stability.","key_machinery":"The load-bearing object is the reduced three-phase model, equations (39)–(42): volume fractions $m$ (SMC), $\\rho$ (collagenous ECM), and $w=1-m-\\rho$ (generic tissue), with quasi-steady reaction–diffusion equations for PDGF $P$ and TGF-$\\beta$ $T$, and non-standard boundary conditions in which SMC influx at the medial edge is proportional to the PDGF gradient times a fixed medial SMC fraction. At a fixed location with fixed local growth-factor concentrations $P^*$, $T^*$, the ECM equation reduces to a quadratic in $\\rho^*$ whose admissible root is the negative branch; differentiating that root yields the closed-form optimal SMC fraction $\\hat{m}^*$ and maximum ECM fraction $\\hat{\\rho}^*$ in terms of $\\mu$ and $\\lambda$. The fact that $\\rho^*(m^*)$ is flat near its maximum is what lets a small SMC population produce near-maximal cap collagen, and it also explains why haptotaxis (the $\\chi_\\rho$ coupling in $\\psi$) changes cap shape but not cap density.","core_discovery":"The paper's central claim is that the collagenous fibrous cap that stabilises an atherosclerotic plaque can be produced and maintained by a surprisingly small population of vascular smooth muscle cells, provided TGF-β is present; TGF-β is the dominant control of cap density because it simultaneously stimulates collagen synthesis and inhibits both SMC-mediated and immune-cell-mediated degradation. In the analytical steady-state reduction, the ECM volume fraction $\\rho^*$ at a fixed plaque location depends biphasically on the SMC volume fraction $m^*$: the optimal SMC fraction $\\hat{m}^* = \\frac{1+\\sqrt{\\lambda}}{1+\\mu+\\lambda+2\\sqrt{\\lambda}}$ maximises the ECM fraction $\\hat{\\rho}^* = \\frac{\\mu}{1+\\mu+\\lambda+2\\sqrt{\\lambda}}$, where $\\mu$ is the ratio of SMC-driven synthesis to immune-cell-driven degradation and $\\lambda$ is the ratio of SMC-driven degradation to immune-cell-driven degradation. With the model's parameter values the optimum lies at a small SMC fraction, below 15–20% even at moderate TGF-β, and the $\\rho^*(m^*)$ curve is flat around the maximum, so cap ECM stays near maximal across a wide range of SMC densities. The base-case simulation ends with about 8.6% SMC and 21.4% collagen, matching ApoE mouse measurements, and removing TGF-β influx cuts cap collagen by roughly 40% while raising SMC numbers.","pith_inferences":["Editorial inference: the flatness of $\\rho^*(m^*)$ around its maximum implies a threshold-like robustness rule—cap thickness stays near-maximal until SMC density falls below a fairly sharp lower bound, after which degradation accelerates; quantifying that threshold against in vivo SMC-density data would be a direct test.","Editorial inference: transplanting the same steady-state machinery to human arteries, which have resident intimal SMCs absent in mice, would likely push plaque SMC densities above the optimum and make human caps depend more on TGF-β responsiveness than on SMC number.","Editorial inference: the paper's observation that cholesterol loading attenuates cellular TGF-β responsiveness translates in this model into a lower $\\mu$ and higher $\\lambda$, which lowers the maximal ECM ceiling; a quantitative simulation of that coupling is a natural extension the paper leaves implicit."],"forward_implications":["If the model is right, preserving TGF-β signalling is the most direct route to cap stability, because TGF-β both raises collagen synthesis and suppresses two separate degradation pathways.","Cap collagen density is not a monotone function of SMC density: beyond the optimal SMC fraction, extra SMCs reduce ECM by occupying space and degrading collagen, so SMC-rich caps can be thinner than SMC-poor ones.","A moderate excess of SMCs above the optimum makes the cap robust to later SMC loss or falling TGF-β, whereas a cap formed below the optimum degrades quickly under the same perturbations.","The model predicts that lowering PDGF influx delays, but does not prevent, cap formation; a sparse SMC population still builds a substantial cap, just more slowly."],"supporting_citations":[{"why":"Provides the two-phase multiphase model, non-standard boundary conditions, and a large part of the parameter set ($\\chi_P$, $\\kappa$, $n_P$) that this model extends.","marker":"Watson et al. (2018)"},{"why":"Supplies the intimal thickness, SMC content, and collagen content time courses in ApoE-deficient mice used to set baseline rates and validate simulation outputs.","marker":"Reifenberg et al. (2012)"},{"why":"Experimental TGF-β blockade study whose reported collagen reduction is the comparison for the $\\alpha_T=0$ simulation.","marker":"Mallat et al. (2001)"},{"why":"Second TGF-β blockade study; together with Mallat it anchors the paper's claim that TGF-β is critical to cap stability.","marker":"Lutgens et al. (2002)"},{"why":"In vitro data on TGF-β-stimulated collagen synthesis by human plaque SMCs sets $A_s$ and $c_s$.","marker":"Kubota et al. (2003)"},{"why":"PDGF/TGF-β modulation of SMC metalloproteinase release sets $A_d$, $c_d$, and $\\gamma_d$ in the ECM degradation term.","marker":"Borrelli et al. (2006)"},{"why":"Diphtheria-toxin SMC apoptosis in ApoE mice shows cap collagen falling roughly two-fold after a four-fold SMC reduction, supporting the small-SMC-population claim.","marker":"Clarke et al. (2006)"},{"why":"Lineage-tracing proliferative index of about 4% motivates the raised baseline SMC proliferation rate $r_m$.","marker":"Chappell et al. (2016)"},{"why":"Data on TGF-β suppression of MMP-9 in monocytes set the immune-cell ECM degradation parameters $\\varepsilon$ and $\\gamma_\\rho$.","marker":"Vaday et al. (2001)"}],"fun_headline_variants":["Small muscle cell pool can still cap plaques","A few muscle cells cap plaques if TGF-β is around","Stable cap from few cells hinges on TGF-β","A handful of muscle cells cap plaque with TGF-β"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Quantitative predictions stand on parameter values chosen, where direct measurements were unavailable, to produce biologically realistic results, so the numerical agreement with the ApoE mouse data is not a fully independent test.","fun_headline_variants_meta":{"raw":{"variants":["Small muscle cell pool can still cap plaques","A few muscle cells cap plaques if TGF-β is around","Stable cap from few cells hinges on TGF-β","A handful of muscle cells cap plaque with TGF-β"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001054,"raw_usage":{"total_tokens":4532,"prompt_tokens":1158,"completion_tokens":3374,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":774,"completion_tokens_details":{"reasoning_tokens":3309}},"tokens_in":774,"tokens_out":3374,"duration_ms":26113,"temperature":1.0,"reasoning_tokens":3309,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:30:55.716843+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the time course of SMC and collagen content in fibrous caps of ApoE-deficient mice after TGF-β blockade: the model predicts cap collagen falls roughly 40% while SMC content rises, whereas the cited TGF-β blockade experiments report about 50% collagen loss with no change in SMC content; the SMC trajectory is the distinguishing observable.","supporting_citations":[],"review_version":1}