Bina Desai

Quantifying interactions in biological tissues requires determining whether an interaction exists, the distance over which it operates, and its strength. Existing spatial-statistical approaches can characterize spatial organization but often require interaction range or interaction models to be specified a priori, making them less suited to settings where these quantities are themselves unknown. Here, we present a practical framework that recovers the presence, range, and association strength of spatial bias toward a tissue structure directly from segmented histology. The framework combines four complementary statistics, the resource selection function, nearest-structure distance distributions, the cross J-function, and the radial distribution function, with specimen-specific reference states that preserve tissue stromal architecture while spanning a continuum of known spatial biases. Across controlled in silico tissues, we show that tissue stromal architecture imposes geometric limits on the range of recoverable interactions and that event fraction independently constrains statistical sensitivity. These spatial associations establish fundamental bounds on inference and demonstrate that spatial statistics must be interpreted relative to tissue-specific reference states rather than absolute scales. We apply the framework to proliferating tumor cells relative to CAF-rich stroma during tyrosine kinase inhibitor therapy in an ALK+ NSCLC model and identify a strong stroma-associated proliferation bias extending approximately 30-60 µm. The framework is broadly applicable to any histological setting in which events and reference structures can be defined, providing a practical approach for quantifying localized interactions directly from tissue stromal architecture.

Relationship between inflection point and IC50. (A) As the steepness parameter, n, increases the true inflection point (x*; blue) approaches the IC50 value (red). (B) The relation between inflection and IC50 is shown here as a function of n.

Extending fragility to a continuous measure (A) Three example functions: convex function where f(x) = x2 (blue), concave function f(x) = x1/2 (red) and a linear function where f(x) = x (black). (B) Example input distribution, p(x) that is normally distributed with mean μ = 0.4 and standard deviation σ = 0.12. (C) Probability density function, q(y), of outcomes, y = f(x) for the corresponding convex (blue), concave (red), and linear (black) in A. Convex functions expand the right tail, driving the mean of the outcome distribution up so that E(f(x)) > f(E(x)). Concave functions expand the left tail, driving the mean down, so that E(f(x)) < f(E(x)). For linear functions, the outcome distribution is symmetric without any tail in either direction.

A population of cells within a tumor can be described as antifragile (the opposite of fragile) if they derive a benefit from fluctuations or perturbations in environmental conditions induced by treatment. Treatment fluctuations could either promote (antifragile tumor) or inhibit (fragile tumor) the evolution of treatment resistance to targeted therapies. In this study, we showed that analysis of the convexity of dose–response curves provided a direct prediction of response to prescribed fluctuations in treatment. Convexity predicted that continuous treatment protocols (i.e., zero prescribed fluctuations in treatment) would outperform uneven protocols when dose response is convex. This theory was applied to predict in vivo response to targeted therapy, and the predictions were validated by experimentally testing high/low intermittent dosing (uneven dosing) and continuous dosing (even dosing) schedules. Convexity (and its inverse, concavity) explained 2 phenomena: Dose response is a convex (fragile) function, but the resistance onset rate is a concave (antifragile) function. Thus, design and validation of alternative treatment schedules maximized response while maintaining prolonged sensitivity to treatment. Together, these analyses provide supporting evidence that fluctuations alter the evolutionary trajectory of tumors in response to targeted therapy, even without altering the cumulative dose. By using this insight to design alternative treatment protocols to limit the evolution of resistance by careful analysis of the dose–response curvature, the study supports the potential of “evolutionary antifragile therapy” as a subtype in the broad class of evolution-based treatment strategies. Significance:The mathematical method using dose response curvature to determine appropriate treatment dosing that limits the evolution of resistance could facilitate designing dose schedules that optimize tumor response.See related commentary by Pomeroy and Palmer, p. 4195

The second-derivative is an approximation for fragility for low values of h. (A) Hill function, H (x) shown for n = 10, E0 = 100, E1 = 20, C = 10. Analytically derived second-derivative is shown in the bottom panel. (B) Difference between fragility and second-derivative at various dose values (red to blue) corresponding to panel A. As h→0, the error approaches zero: F(x, h) - h2 d2H/dx2 → 0.

Targeted therapies induce strong clinical responses but fail to eliminate advanced cancers, as a subset of tumor cells survives within residual disease and eventually develops resistance. While numerous cell-intrinsic and microenvironmental mechanisms have been implicated in this survival, their relative contributions remain poorly defined. Using spatial histological inferences from ALK + NSCLC models, we show that peristromal niches protect tumor cells from elimination, enabling in vivo persistence. This spatially restricted sheltering provides an ecological rescue mechanism that sustains residual populations, enabling their eventual evolutionary escape. Mechanistically, this protective effect reflects an integrated action of multiple juxtacrine and paracrine signals. This complexity limits the utility of targeting individual mechanisms of protection, favoring a shift towards exploiting orthogonal collateral sensitivities of residual disease. We find that adaptive HER2 upregulation, associated with both cell-intrinsic and stroma-mediated persistence, can be exploited by the antibody-drug conjugate T-DXd to dramatically enhance therapeutic responses and suppress tumor relapse.

A population of cells within a tumor can be described as antifragile (the opposite of fragile) if they derive a benefit from fluctuations or perturbations in environmental conditions induced by treatment. Treatment fluctuations could either promote (antifragile tumor) or inhibit (fragile tumor) the evolution of treatment resistance to targeted therapies. Here, we showed that analysis of the convexity of dose response curves provided a direct prediction of response to prescribed fluctuations in treatment. Convexity predicted that continuous treatment protocols (i.e. zero prescribed fluctuations in treatment) would outperform uneven protocols when dose response is convex. This theory was applied to predict in vivo response to targeted therapy, and the predictions were validated by experimentally testing high/low intermittent dosing (uneven dosing) and continuous dosing (even dosing) schedules. Convexity (and its inverse, concavity) explained two phenomena: dose response is a convex (fragile) function but resistance onset rate is a concave (antifragile) function. Thus, design and validation of alternative treatment schedules maximized response while maintaining prolonged sensitivity to treatment. Together, these analyses provide supporting evidence that fluctuations alter the evolutionary trajectory of tumors in response to targeted therapy, even without altering the cumulative dose. By using this insight to design alternative treatment protocols to limit the evolution of resistance by careful analysis of the dose response curvature, the study supports the potential of "Evolutionary Antifragile Therapy" as a subtype in the broad class of evolution-based treatment strategies.

Cancer-associated fibroblasts (CAFs) are associated with tumor progression and drug resistance. Here, we investigated the mechanisms underlying the cross-talk between CAFs and tumor cells in non–small cell lung cancer (NSCLC). In NSCLC cell lines with EML4-ALK fusions, we observed substantial CAF-mediated drug resistance to clinically used inhibitors of the tyrosine kinase ALK. Array-based cytokine profiling of CAF-derived conditioned medium indicated that a major contributor to the phenomenon was the secreted growth factor HGF, and blocking its receptor MET overcame paracrine resistance to ALK inhibition. However, cell-selective labeling of the proteome in cocultures also revealed an equally important contribution by the fibronectin-integrin pathway, specifically integrin β ₁ , which was confirmed through pharmacological inhibition and cell-specific silencing or knockout. Concurrent targeting of MET and integrin signaling effectively abrogated ALK inhibitor resistance in EML4-ALK ⁺ NSCLC cells cocultured with CAFs. Moreover, the combination of the ALK inhibitor alectinib with the MET inhibitor capmatinib and/or the integrin inhibitor cilengitide was more effective than single-agent treatment in suppressing tumor growth in allografted mice. The findings illustrate a previously unappreciated complex nature of concurrent paracrine and juxtacrine mechanisms of CAF-driven resistance that may inform the development of more effective therapeutic approaches.

Small-molecule KRASG12C(OFF) inhibitors that bind to the inactive GDP-bound state of KRAS have demonstrated efficacy in patients with KRASG12C-mutant tumors, yet responses tend to be transient because of emergence of on-treatment resistance. Recently, RAS(ON) G12C–selective inhibitors, which bind to the active GTP-bound state of RAS, were described, and elironrasib is undergoing evaluation in multiple clinical trials. In this study, we generated resistant cell lines and patient-derived xenograft models to KRASG12C(OFF) and RAS(ON) G12C–selective inhibitors and interrogated resistance mechanisms using a multiomics strategy consisting of phosphoproteomics, whole-exome sequencing, and RNA sequencing combined with functional testing using small-molecule and CRISPR screens and RAS(ON) inhibitors being evaluated in clinical trials. Two models reactivated RAS signaling, either via KRASG12C gene amplification or NRASG13R mutation, and were vulnerable to dual inhibition by RAS(ON) G12C–selective and RAS(ON) multiselective inhibitors, RMC-4998 and RMC-7977. Two models, which lacked any discernable genomic alteration, acquired resistance associated with increased receptor tyrosine kinase activity and downstream persistent RAS activity and were sensitive to RAS-GTP inhibition by RMC-7977. Finally, one model displayed epithelial–mesenchymal transition, loss of RAS dependence, and acquired reliance on cell-cycle kinases and proteins associated with DNA damage response. This work highlights KRASG12C-selective inhibitor resistant states that parallel and complement clinical findings and demonstrate that a large subset could be overcome with a RAS(ON) multi-selective inhibitor as a stand-alone agent or in combination with other therapies. Significance: Multi-omic characterization of resistance mechanisms to KRASG12C-selective inhibitors in non–small cell lung cancer provides insights that could inform precision medicine-based therapeutic approaches for improving the treatment of KRASG12C mutant tumors. See related article by Stern et al., p. 485