Researchers have unveiled a new laboratory platform that could change how enzyme-activated anticancer drugs are tested, combining a 3D tumor model with dynamic fluid flow to better mimic conditions inside the human body. The system, developed for screening a class of therapeutics known as enzyme-induced self-assembling peptides (EISAPs), addresses a persistent problem in cancer drug development: traditional two-dimensional cell cultures fail to capture the complexity of real tumors, including drug penetration, extracellular matrix interactions, and the nutrient gradients that shape how cells respond to treatment. By growing breast cancer spheroids on tiny hydrogel-coated pillars and exposing them to continuous media flow, the team has created a tool that is both physiologically relevant and suitable for high-throughput screening.
The peptides at the heart of the study are designed to exploit a distinctive weakness of tumors. In their inactive state, the molecules circulate harmlessly through the body. Only when they encounter enzymes that are overexpressed in the tumor microenvironment—such as alkaline phosphatases or the Eyes Absent (EYA) tyrosine phosphatase—do they lose their protective phosphate caps. Dephosphorylation reduces the molecule’s hydrophilicity, triggering self-assembly into nanostructures through π–π stacking and hydrophobic interactions. These in situ nanofibers enhance drug accumulation and cellular uptake at the disease site while sparing healthy tissue, offering a potential route around the multidrug resistance that plagues aggressive breast cancers such as triple-negative breast cancer (TNBC). Previous work by the same group had shown that EYA-activated peptides localize in the cytoplasm and nucleus of breast cancer cells, persist for days, and trigger DNA damage responses—but only in simplified 2D settings.
To bring those findings into a more realistic context, the researchers grew spheroids from two breast cancer cell lines representing very different disease biology: MCF-7, an estrogen-receptor-positive line with wild-type TP53, and MDA-MB-231, an aggressive triple-negative line carrying a gain-of-function mutant p53 (R280K). Spheroids were formed in ultra-low attachment plates using a liquid-overlay technique, with a small amount of Matrigel added to promote extracellular matrix deposition and produce compact, uniform aggregates. Careful characterization showed that metabolic activity rose steadily after seeding, peaked on day 7, and declined sharply by day 8, while spheroids larger than roughly 500 micrometers began developing necrotic cores as oxygen diffusion fell behind demand. Days 3 to 7 emerged as the optimal testing window, ensuring that drug responses measured in the platform reflected therapeutic effects rather than the artifacts of a dying culture.
The centerpiece of the platform is a 36-pillar plate system in which spheroids are transferred and immobilized on individual hydrogel-coated posts, then submerged in a companion deep-well plate. For dynamic culture, the pillar plate pairs with a perfusion plate placed on a rocker that tilts at 10 degrees in one-minute intervals, generating bidirectional flow that mimics interstitial fluid movement in tumors. The critical technical hurdle was finding the right hydrogel to anchor the spheroids. Matrigel coatings achieved a 91.6 percent transfer rate but induced invasive outgrowth, a behavior the team eventually harnessed as a feature rather than a flaw. Pure alginate gels performed poorly across concentrations from 0.75 to 1.75 percent, with transfer rates ranging from zero to roughly 44 percent and frequent necrosis. Self-assembling Fmoc-Phe-Phe-OH hydrogels fared worse still, releasing spheroids within 48 hours at every concentration tested.
The winning formulation turned out to be a hybrid of 2 percent gelatin and 1 percent alginate, crosslinked with calcium chloride. Gelatin enhances cell–extracellular matrix interactions and improves the viscoelastic properties of alginate, while alginate provides mechanical stability through ionic crosslinking. This combination achieved a 91.6 percent transfer efficiency with no detachment during handling or perfusion, and spheroids remained compact, non-invasive, and viable over extended culture. The result is a reproducible baseline model free of the confounding effects of invasion, against which therapeutic responses can be cleanly measured.
Rather than discarding the invasive phenotype observed with Matrigel, the investigators exploited it to model metastatic behavior. Spheroids embedded in 75 percent Matrigel on the pillars developed invasive protrusions within two days of transfer, recapitulating early steps of the metastatic cascade, including loss of cell–cell adhesion and migration through a basement membrane-like matrix. This is particularly relevant for TNBC, where epithelial-to-mesenchymal transition drives aggressive spread; MDA-MB-231 spheroids in 3D culture are known to show suppressed E-cadherin, elevated N-cadherin, and heightened invasive capacity compared with flat cultures. A refined Matrigel–alginate mixture produced a more confined, flower-like invasion pattern that proved remarkably informative during drug testing.
With the platform established, the team screened a panel of six peptides—P1 (Fmoc-FF-pTyr), P2 (Fmoc-FF-pThr), P3 (RGD-FF-pTyr), P4 (NBD-FF-pTyr), P5 (Nap-FF-pTyr), and P6 (Nap-FF-pThr)—whose molecular weights were confirmed by mass spectrometry. All six reduced spheroid viability in a dose-dependent manner across concentrations from 10 to 200 micromolar, but two stood out. P1 reduced MDA-MB-231 spheroid viability to about 58 percent at its optimal dose of 100 micromolar, and P5 showed the strongest cytotoxicity at 200 micromolar. Critically, when P1 was tested under dynamic perfusion, viability dropped to approximately 55 percent—evidence that continuous flow enhances the peptide’s activity, plausibly by improving compound delivery and waste removal and by generating the concentration gradients and shear forces that cells encounter in living tumors.
The most striking results came from combination therapy. When P1 was co-administered with 5 micromolar doxorubicin, the standard chemotherapy, viability fell significantly below levels seen with doxorubicin alone, with effects intensifying at 48 hours and under dynamic flow. In the invasive Matrigel–alginate model, the combination did not merely slow tumor spread—it completely suppressed invasive outgrowth beyond the spheroid boundary. Notably, spheroids in the Matrigel–alginate matrix showed lower viability than those in ULA plates or Matrigel alone at 10, 50, and 100 micromolar of the combination, a statistically significant difference suggesting that the extracellular environment itself modulates drug efficacy, likely through effects on penetration and cell–matrix interactions. This synergy echoes earlier reports that sodium alginate incorporated into Fmoc-FF peptide networks stabilizes nanofiber architectures and slows peptide degradation.
Visualizing where the drugs actually go within a tumor was made possible by the fluorescently tagged peptide NBD-FF-pTyr. Over five days of treatment, fluorescence microscopy revealed a progressive journey: the signal initially concentrated at the spheroid periphery, then penetrated inward with growing intensity, peaking around day 3 as measured by relative fluorescence units, before slightly declining—possibly reflecting redistribution, clearance, or partial degradation. This real-time confirmation of intratumoral penetration and core accumulation is a feature that flat cultures simply cannot offer, and it demonstrates that the enzyme-activated peptides can reach the nutrient-starved interior regions of a tumor where many chemotherapies fail.
At the molecular level, RT-qPCR analysis revealed that the peptides’ effects are both peptide-specific and cell-line-specific, a finding with important implications for personalized oncology. In MCF-7 spheroids, P1 significantly downregulated BCL2, BRCA2, and TP53—a systemic collapse of the survival and DNA-repair axes that primes cells for apoptosis, undermining anti-apoptotic signaling and impairing homologous recombination-based genomic maintenance. The triple-negative MDA-MB-231 cells, protected by their oncogenic mutant p53, instead mounted a compensatory upregulation of all three genes in response to P1, signaling potential resistance. However, P5 (Nap-FF-pTyr) broke through this defense, achieving the strongest repression of survival and DNA-repair pathways even in the aggressive TNBC background. By contrast, P2 and P6 elevated BCL2 in MDA-MB-231 cells, suggesting they may inadvertently activate resistance pathways. Together, the viability data, imaging results, and transcriptional profiles converge on a clear message: structural variations among EISAP designs strongly influence cytotoxic potential, tumor penetration, and resistance profiles, and P1 and P5 emerge as the most promising candidates for further development.
The work, funded by the National Institute of General Medical Sciences, establishes a dynamic, high-throughput 3D screening platform that brings laboratory drug testing measurably closer to the complexity of the clinic. The researchers note that future studies should investigate the mechanisms underlying EISAP-mediated cytotoxicity and extend the platform to heterogeneous spheroids containing stromal and immune cells, advancing enzyme-responsive peptide therapies toward preclinical evaluation. For a class of drugs whose entire therapeutic logic depends on the biology of the tumor microenvironment, being tested within a system that finally reproduces that environment’s architecture, matrix, and flow may prove to be the decisive step forward.
Subject of Research: Screening of enzyme-induced self-assembling peptide therapeutics in 3D breast cancer spheroids using a dynamic pillar-perfusion platform
Subject of Research: Technology and Engineering
Article Title: Dynamic pillar-perfusion platform for screening enzyme-induced self-assembling peptide therapeutics in 3D breast cancer spheroids
Article References: Escobar Martinez, A., Joshi, P., Carney, E., Fouladgar, F., Powell, R., Vanga, M. G., Gnenema, V., Hripko, S., Lee, M.-Y., & Habibi, N. (2026). Dynamic pillar–perfusion platform for screening enzyme‐induced self‐assembling peptide therapeutics in 3D breast cancer spheroids. Bioengineering & Translational Medicine, 11(4), Article e70135. https://doi.org/10.1002/btm2.70135
Image Credits: AI Generated
DOI: 10.1002/btm2.70135
Keywords: enzyme-induced self-assembling peptides, breast cancer spheroids, pillar-perfusion platform, 3D tumor model, triple-negative breast cancer, doxorubicin co-treatment, hydrogel optimization, tumor microenvironment, drug penetration, RT-qPCR, BCL2, BRCA2
Cite Scienmag News
APA MLA Chicago
Nathaniel Bowman. (September 3, 2026). Perfusion platform screens self-assembling peptide drugs in 3D breast tumor models. Scienmag. https://scienmag.com/perfusion-platform-screens-self-assembling-peptide-drugs-in-3d-breast-tumor-models/
Nathaniel Bowman. “Perfusion platform screens self-assembling peptide drugs in 3D breast tumor models.” Scienmag, 3 September 2026, https://scienmag.com/perfusion-platform-screens-self-assembling-peptide-drugs-in-3d-breast-tumor-models/. Accessed 3 September 2026.
Nathaniel Bowman. “Perfusion platform screens self-assembling peptide drugs in 3D breast tumor models.” Scienmag. September 3, 2026. https://scienmag.com/perfusion-platform-screens-self-assembling-peptide-drugs-in-3d-breast-tumor-models/
Copy citation Download RIS
Tags: 3D breast tumor models3D tumor models for drug testingbreast cancer spheroid culturecancer drug screeningdynamic fluid flow in tumor modelsdynamic fluid flow tumor screeningenzyme-activated self-assembling peptide drugsenzyme-activated self-assembling peptide therapeuticsenzyme-responsive nanostructures for drug deliveryenzyme-responsive peptide nanostructuresextracellular matrix interaction in drug screeningextracellular matrix interactions in cancerhigh-throughput cancer drug screening platformshigh-throughput cancer drug testinghydrogel-coated pillar tumor cultureshydrogel-coated pillars in cancer researchmimicking in vivo tumor conditions for drug developmentnanoparticle self-assembly in cancer therapynutrient gradient modeling in cancer researchphysiologically relevant tumor microenvironmentphysiologically relevant tumor modelstumor drug penetration simulationtumor microenvironment mimicking platformstumor penetration and drug accumulation mechanisms
