Soft Substrates Train Human T Cells for Sustained Proliferation Through Mechanomemory

Engineered cellular immunotherapies, including adoptive T cell therapies deployed against malignant tumors, require reliable ex vivo culture systems to expand potent immune cells from patient or donor samples. While traditional bioprocesses focus largely on chemical and soluble cues, a research study published in Frontiers in Bioengineering and Biotechnology reveals that human T cells actively process the physical rigidity of their activation substrates to establish a durable, cellular mechanomemory. Researchers at Columbia University demonstrated that primary human T cells stimulated on mechanically soft elastomer substrates sustain rapid cell division across multiple generations long after being removed from the physical material. Rather than acting as a fleeting, immediate response to physical contact, this mechanical sensitivity becomes deeply imprinted into the population over several days. The findings clarify how mechanical environments guide long-term immune cell growth and suggest new strategies for improving cellular immunotherapy manufacturing.

To explore the physical dynamics governing expansion, the investigators engineered planar polydimethylsiloxane substrates using Dow Sylgard 184, systematically modulating mechanical compliance by altering the crosslinker-to-base ratio. Two specific formulations were investigated: a rigid substrate prepared at a standard 1:10 ratio, exhibiting a Young’s compressive modulus of 2 megapascals, and a compliant, soft counterpart prepared at a 1:50 ratio, yielding an elastic modulus of 20 kilopascals. Cylindrical test samples measured by microcompression on an Instron 5848 MicroTester verified these bulk elastic properties under controlled indentation increments. To present physiological activating cues, the team functionalized the sterilized elastomer surfaces with goat-anti-mouse secondary antibodies and bovine serum albumin blocking layers before capturing defined mixtures of mouse-anti-human monoclonal antibodies targeting CD3 and CD28 receptors.

The experimental expansion pipeline utilized primary, mixed-subtype peripheral human T cells isolated by negative selection from healthy donor leukapheresis packs. Cells were rested overnight in complete culture medium consisting of RPMI 1640 supplemented with fetal bovine serum, HEPES, L-glutamine, penicillin-streptomycin, and beta-mercaptoethanol before seeding onto the functionalized elastomer surfaces at a concentration of one million cells per milliliter. After three days of initial activation, the cells were detached from the substrates via pipetting and transferred to non-coated, tissue-culture-treated polystyrene plates, completely removing subsequent physical ligand exposure. Every two days thereafter, the expanding suspensions were enumerated, diluted back to five hundred thousand cells per milliliter in fresh vessels, and tracked until population growth ceased. Across thirteen distinct biological donors, cells initially activated on compliant twenty-kilopascal substrates exhibited an extended period of exponential proliferation, achieving significantly greater cumulative population doublings than cells stimulated on rigid two-megapascal surfaces, despite showing identical initial growth kinetics during the first week.

Because the physical divergent growth trajectories manifested several days after cells had been separated from the activating polymers, the authors investigated whether mechanomemory is driven strictly by cell-autonomous changes or mediated by extrinsic soluble factors shared within the culture. To test the role of paracrine signaling, the researchers executed cross-over media exchange experiments between soft-activated and hard-activated cultures at day one, day three, and day five of culture. For day one transfers, adherent cells were detached from the elastomer surface using a brief three-minute incubation on ice packs before harvesting conditioned supernatants. Transferring media conditioned by soft-substrate cultures onto hard-activated cells at day one or day three successfully transferred the enhanced proliferative capacity, extending their division lifespan. Conversely, exposing soft-activated cells to hard-conditioned media during these early intervals prematurely curtailed their division capacity. Crucially, performing reciprocal media exchanges at day five caused no change in proliferative trajectories, demonstrating that soluble reinforcement operates within a distinct, early temporal window before cellular mechanomemory becomes definitively locked.

To complement the media exchange experiments and isolate intrinsic cellular programming from perpetual soluble accumulation, the team instituted an aggressive media replacement protocol, designated R1+, R3+, and R5+. In these regimens, culture media was completely removed and replaced with entirely fresh, factor-free medium at each serial dilution step starting on day one, day three, or day five post-seeding. While completely replacing media from day one or day three reduced overall cumulative doublings across all conditions by stripping endogenous growth factors, cells stimulated on soft substrates maintained their relative proliferative advantage over hard-stimulated counterparts. Most notably, initiating complete medium replacement at day five yielded no statistically significant deviation in cumulative population doublings compared to unreplaced control cultures. These findings demonstrate that although soluble signals reinforce early mechanotransduction, the cells progressively assimilate these external cues, establishing an autonomous mechanical imprint by day five that no longer depends on the continued presence of conditioned soluble molecules.

To identify the specific secreted molecules responsible for conveying mechanical information between cells, the researchers profiled culture supernatants collected at days one, three, and five across three representative donors using multiplex Luminex assays. A targeted panel revealed that soft-activated T cells secreted substantially higher concentrations of interleukin-2, interferon-gamma, tumor necrosis factor-alpha, interleukin-4, interleukin-6, and interleukin-17F relative to hard-activated cultures, with the most pronounced quantitative divergence occurring at day three. However, when the authors tested whether single cytokines could replicate the mechanomemory phenotype by supplementing hard-activated cultures with exogenous recombinant interleukin-2 or interleukin-6 alone, neither factor restored cellular proliferation to levels observed on soft substrates. The investigators concluded that mechanical memory transfer is not dictated by an individual master cytokine, but instead relies on a multifaceted, coordinated mixture of soluble proteins, potentially accompanied by unmeasured factors or extracellular vesicles released during the early activation window.

A critical consideration in interpreting bulk expansion differences was whether substrate stiffness simply shifted the differentiation balance among major immune lineages during early activation. Using multi-parameter spectral flow cytometry on day five, the researchers stained cell populations for CD4, CD8, CD45RA, CD62L, and CD25 to quantify naïve, central memory, effector memory, and terminal effector subsets. Across the broad panel of human donors, the relative frequencies of these major phenotypic compartments shifted from predominantly naïve to memory phenotypes over five days, but showed no statistically significant differences attributable to substrate stiffness. Extended profiling of helper subsets within the CD4-positive compartment similarly exhibited consistent distributions regardless of whether cells experienced hard or soft activation. Because the overall phenotypic composition remained equivalent at day five, the study established that mechanomemory reflects functional reprogramming within individual cells rather than the selective survival or expansion of a specific T cell subset.

To discover the persistent molecular mechanisms underlying this autonomous behavior, the researchers performed genome-wide promoter DNA methylation profiling on day five using Illumina Infinium MethylationEPIC V2 arrays capturing approximately 250,000 promoter CpG loci. Analyzing CD45RA-positive and CD45RA-negative fractions independently, principal component analysis and per-gene differential methylation analyses revealed that activation on rigid two-megapascal substrates induced widespread promoter hypermethylation compared to compliant substrates. This epigenetic divergence was markedly pronounced in the CD45RA-positive population, which encompasses naïve and early effector cells. Unstimulated control T cells exhibited substantially lower baseline promoter methylation than both stimulated cohorts. These global epigenetic patterns indicated that mechanical rigidity during activation actively remodels the chromatin landscape, preferentially imposing repressive DNA methylation marks that correspond with restricted proliferative longevity.

Gene Set Enrichment Analysis focusing on curated T cell Gene Ontology terms revealed that hard-substrate stimulation drove extensive promoter hypermethylation across pathways governing cell cycle progression and mitotic infrastructure. Key loci showing increased methylation in hard-activated cells included genes essential for positive regulation of activated T cell proliferation, such as PRKCQ, CD3E, and multiple cytokine-related genes including IL2RG, IL18, IL23, TNFSF9, and IGF2. Regulators of cellular quiescence and mitotic spindle alignment, including CDK6, ABL1, WNT, NSFL1C, HTT, and PAFAH1B1, were also hypermethylated following hard activation. By methylating regulatory elements responsible for sustaining cell division and escaping quiescence, cells activated on rigid matrices appear epigenetically primed for premature cell-cycle arrest, providing a molecular explanation for their earlier transition out of exponential growth.

Substrate compliance also generated profound epigenetic divergences across metabolic and signaling pathways that sustain cellular bioenergetics. Cells activated on rigid substrates exhibited targeted hypermethylation at glycolytic regulatory loci, including HK1, PRKACA, and DYRK2, indicating a potential epigenetic constraint on aerobic glycolysis, the primary metabolic engine supporting rapid immune cell proliferation. In sharp contrast, cells activated on soft substrates displayed localized hypermethylation around oxidative metabolic programs associated with quiescent states, including fatty acid alpha-oxidation and oxidative phosphorylation loci such as ETFAE, ETFDH, LONP2, and IVD. Furthermore, hard-activated T cells demonstrated hypermethylation across essential activation cascades, including calcineurin-NFAT, CAMKK-AMPK, and JAK-STAT networks, as well as structural cytoskeletal regulators involving Rho GTPase signaling and actin dynamics, suggesting an integrated mechanical reprogramming linking force generation to cellular longevity.

The authors contextualized these findings within cellular therapy engineering while noting important experimental considerations and limitations. Because individual human donors naturally exhibit wide baseline variations in starting immune composition and expansion capacity, translation of mechanical priming protocols will require accommodating donor-to-donor heterogeneity. Additionally, while major flow cytometry subsets remained balanced at day five, the rapid internalization and dynamic expression of activation markers during exponential expansion constrained the resolution of rare subsets, such as stem cell memory T cells, which could potentially contribute to long-term proliferative fitness. The investigators also emphasized that their genomic interpretations remain prospective correlations that warrant further functional validation. Nonetheless, by proving that mechanical memory takes time and paracrine reinforcement to mature into stable epigenetic programs, this research establishes that biomaterial mechanics can be strategically leveraged to train durable immune responses for clinical applications.

Subject of Research: Mechanomemory and reinforcement of proliferation in human T cells activated on compliant biomaterials

Article Title: T cell mechanomemory takes training: cell-secreted factors reinforce short-term mechanosensing into long-term proliferation

Article References: Sepp, A.-L. I., Ji, Y., & Kam, L. C. (2026). T cell mechanomemory takes training: cell-secreted factors reinforce short-term mechanosensing into long-term proliferation. Frontiers in Bioengineering and Biotechnology, 14, Article 1944620. https://doi.org/10.3389/fbioe.2026.1944620

Image Credits: AI Generated

DOI: 10.3389/fbioe.2026.1944620

Keywords: T cell mechanomemory, mechanosensing, biomaterials, polydimethylsiloxane, cell expansion, cellular immunotherapy, cytokines, epigenetics, DNA methylation, cell proliferation, CD3 CD28 activation, substrate stiffness, Luminex assay, flow cytometry, MethylationEPIC BeadChip, metabolic programming, cytoskeletal dynamics, cell-cell signaling