Pillar-perfusion platform screens enzyme-responsive peptide therapies in 3D breast cancer spheroids

In a development that could reshape how experimental cancer drugs are vetted before they ever reach a patient, researchers have built a miniature tumor-testing factory: living breast cancer spheroids grown on tiny hydrogel-coated pillars, continuously bathed in a gentle, rocking flow of nutrient medium that mimics the fluid currents inside a real tumor. Writing in Bioengineering & Translational Medicine, a team funded by the U.S. National Institutes of Health describes how this pillar–perfusion platform allowed them to screen a family of smart anticancer peptides — molecules that circulate in an inert state and assemble into toxic nanostructures only after tumor enzymes switch them on. Under flowing conditions, a lead peptide cut the viability of triple-negative breast cancer spheroids to roughly 55 percent, and pairing it with the chemotherapy drug doxorubicin pushed cell death deeper still. The work confronts one of oncology’s most stubborn problems: laboratory models that look convincing in a flat dish but collapse when faced with the complexity of a living tumor.

Part of the blame lies with the models themselves. Although molecular profiling and targeted therapies have improved outcomes for many breast cancer patients, survival remains poor for aggressive subtypes such as HER2-driven tumors and triple-negative breast cancer, or TNBC, which lacks the three receptors most modern drugs aim at. Conventional two-dimensional cultures fail to capture the extracellular matrix that scaffolds real tumors, the oxygen-poor gradients that build up inside dense tissue, the constant signaling between neighboring cells, and the genetic heterogeneity that lets some cells shrug off a treatment while their neighbors die. Animal models fill some of those gaps but are slow, costly, and often poor predictors of human response, and the field has lacked feasible, high-throughput 3D systems that reproduce both native tissue architecture and blood flow. Tumor spheroids — self-organizing spheres of cancer cells a few hundred micrometers across — offer a compelling middle ground, reproducing nutrient diffusion limits, drug penetration barriers, and the hypoxic, sometimes necrotic cores that define solid tumors, yet standard ways of making them suffer from erratic sizing, poor control over microenvironmental conditions, and restricted transport of oxygen and therapeutics.

The therapeutics at the center of the study belong to a class called enzyme-induced self-assembling peptides, or EISAPs. These short molecules carry phosphorylated amino acids — phosphotyrosine or phosphothreonine — that keep them soluble as they travel. Inside the tumor microenvironment, enzymes overexpressed by cancer cells, such as alkaline phosphatase and the Eyes Absent (EYA) tyrosine phosphatase, clip off those phosphate groups. Stripped of their hydrophilic shields, the molecules suddenly prefer one another’s company, stacking through aromatic π–π interactions between protective caps such as Fmoc and Nap and through hydrophobic forces, spinning themselves into β-sheet nanofibers right at the disease site. The team’s earlier work validated this chemistry in fine detail: researchers directly measured phosphate released when EYA2 dephosphorylated the peptides in vitro and imaged the resulting nanofibers with electron microscopy. Once inside breast cancer cells, the assemblies spread through both cytoplasm and nucleus, remained detectable for up to five days, and accumulated in a subset of mitochondria — while conspicuously sparing normal epithelial cells and igniting the DNA damage response in malignant ones.

To test the compounds in something resembling a real tumor, the researchers first grew spheroids from MDA-MB-231 cells, an aggressive TNBC line, and MCF-7 cells, a hormone-responsive, less invasive line, in ultra-low-attachment plates dosed with three percent Matrigel to encourage compact, uniform aggregation. Daily metabolic readouts showed activity climbing through day seven before falling sharply on day eight, while diameters grew steadily; because spheres larger than about 500 micrometers develop necrotic cores, the team settled on days three through seven — and diameters of 350 to 420 micrometers — as the optimal testing window. Fluorescence stains told the same story over time: at day four, spheroids glowed with living cells; by day eight, death had crept into the core, mitochondrial activity persisted mainly at the rim, and caspase-3/7 apoptotic signaling had risen. Transfer to the pillar platform relied on a stamping maneuver: a 36-pillar plate coated with hydrogel is pressed onto the spheroid plate and inverted, letting gravity deposit one sphere on each pillar. For dynamic culture, the pillar plate couples to a companion perfusion plate riding an OrganoFlow rocker tilted at 10 degrees, swinging at one-minute intervals so medium sloshes bidirectionally across the spheroids — a setup the team found markedly more robust with the rounder, sturdier MDA-MB-231 spheroids.

A large share of the effort went into an unglamorous but decisive question: which hydrogel best glues living spheroids to the pillars without wrecking them? The candidates included Matrigel, a basement-membrane extract rich in laminin and collagen; alginate, a seaweed-derived polymer crosslinked with calcium chloride; blends of the two; gelatin-alginate mixtures; and even Fmoc-Phe-Phe, a peptide gel related to the therapeutics themselves. Each carried trade-offs. Alginate alone, gelled with 2 or 3 millimolar calcium chloride, transferred as few as none and at most 44.4 percent of spheroids, and survivors darkened and developed necrotic cores as nutrients struggled through the dense gel. Fmoc-FF gels peaked at a 33.3 percent transfer rate and then shed their cargo within 48 hours, victims of their own softness. Matrigel at 75 percent concentration achieved a 91.6 percent transfer rate — but it also unleashed invasive outgrowth, with cells streaming outward from dense 5,000-cell spheroids within two days. The eventual winner was a blend of 1 percent alginate and 2 percent gelatin: 91.6 percent transfer, no detachment, preserved sphericity, and no invasion, thanks to gelatin’s reinforcement of cell-matrix interactions and the composite’s sturdier viscoelasticity under dynamic culture.

In a twist that turned a nuisance into an asset, the researchers realized that Matrigel’s talent for provoking invasion was itself an opportunity. TNBC invasion rides on the epithelial-to-mesenchymal transition, the molecular program that equips cancer cells with migratory powers by suppressing E-cadherin and elevating N-cadherin, and a platform that reliably reproduces invasive behavior is rare and valuable. Spheroids embedded in Matrigel plus 0.75 percent alginate invaded in a more confined, flower-like pattern, giving the team a controllable model of the metastatic cascade’s opening step: local invasion through basement membrane. When these invasive spheroids were co-treated with the lead peptide P1 and 5 micromolar doxorubicin, viability dropped most dramatically in the Matrigel–alginate setting, and at higher doses the invasive outgrowth beyond the spheroid boundary vanished entirely. Strikingly, spheroids in the Matrigel–alginate matrix fared worse than those in plain Matrigel or in ordinary suspension even though the Matrigel itself had been diluted to half strength — evidence that alginate’s confinement acts synergistically with the peptide, consistent with earlier reports that sodium alginate stabilizes Fmoc-FF peptide networks and slows their degradation.

With the platform tuned, the team screened six peptide variants ranging from 662 to 885 daltons in molecular weight: Fmoc-FF-pTyr (P1), Fmoc-FF-pThr (P2), RGD-FF-pTyr (P3), the fluorescently tagged NBD-FF-pTyr (P4), Nap-FF-pTyr (P5), and Nap-FF-pThr (P6). Spheroids of both cell lines were soaked in concentrations from 10 to 200 micromolar for 72 hours, and every peptide killed in a dose-dependent fashion, though the profiles diverged tellingly. In TNBC spheroids, the Nap-capped P5 and P6 proved most lethal at the top dose, while P1 hit its sweet spot at 100 micromolar, reducing spheroid viability to 58 percent in static culture. Under dynamic perfusion, P1 performed even better: at the same 100-micromolar dose, viability fell to 55 percent, suggesting that continuous flow — which mirrors interstitial fluid movement in tumors, sharpening drug gradients and sweeping away metabolic waste — boosts the peptide’s self-assembly and cytotoxic reach. In MCF-7 spheroids, P1 was most potent within its dose range at 100 micromolar, while P2 and P5 achieved their greatest kill at 200 micromolar, underscoring how each design carries its own dose–response fingerprint. Combining P1 with doxorubicin outperformed doxorubicin alone, and the dynamic advantage widened with exposure time, becoming most pronounced at 48 hours.

A fluorescent version of the lead chemistry answered a question static assays cannot: does the peptide actually reach the tumor’s interior? The team tagged NBD-FF-pTyr with its built-in 7-nitrobenz-2-oxo-1,3-diazol fluorophore, whose glow signals peptide presence and self-assembly, then exposed spheroids to 250 micromolar of the construct and photographed them daily for five days. NBD labeling is a standard probe for following internalization and the morphological shift from soluble molecules to β-sheet nanofibers inside cells. On day one, green fluorescence hugged the spheroid’s rim. Over the next two days the signal spread inward and intensified, flooding the core — direct visual evidence that these enzyme-activated nanostructures can penetrate the dense, diffusion-limited interior of a 3D tumor, a property many conventional drugs lack. Quantified as relative fluorescence units, the signal peaked on day three and then eased slightly, a pattern the authors attribute to redistribution within the spheroid, clearance, or partial degradation. The imaging confirmed that internalization happens largely in the first days of exposure, information that will shape dosing schedules in future studies.

The most clinically provocative findings emerged from gene expression analysis. After 72-hour treatments, the team measured transcript levels of BCL2, the gene behind a major anti-apoptotic shield; BRCA2, central to high-fidelity DNA repair by homologous recombination; and TP53, the genome’s guardian, by quantitative PCR normalized to the housekeeping gene GAPDH. The responses split cleanly along cell-line lines. In MCF-7 cells, which carry wild-type p53, peptides P1 and P5 jointly suppressed all three genes — a systemic collapse of survival signaling and DNA-repair capacity that primes cells for apoptosis. In MDA-MB-231 cells, which harbor a gain-of-function mutant p53 known as R280K, P1 instead triggered a compensatory upregulation of all three genes, a transcriptional signature of incipient treatment resistance characteristic of this notoriously drug-resistant line. Yet P5 cut through that defense, suppressing BCL2 and BRCA2 even in the resistant background, apparently bypassing the oncogenic protection that mutant p53 normally confers — an effect echoing earlier findings that P5 diminishes mitochondrial activity. Meanwhile, P2 and P6 raised BCL2 in TNBC cells, signaling stress responses entangled with repair and survival programs, a reminder that modest structural changes, such as swapping tyrosine for threonine or Fmoc caps for Nap, can flip the biological outcome.

The authors conclude that 3D spheroids cultured on pillar plates, run under both static and flowing conditions, provide a physiologically relevant, high-throughput way to evaluate cancer therapies before animal studies, and they position EISAPs as targeted adjuvants that could make standard chemotherapy more effective and help overcome drug resistance within the hostile terrain of the tumor microenvironment. The project, supported by the National Institute of General Medical Sciences under award R16GM150848, leaves clear next steps: dissecting the mechanisms by which these peptides kill, and expanding the platform to heterogeneous spheroids that include stromal and immune cells to move EISAP-based therapies toward preclinical evaluation. For a field in which promising molecules routinely die in translation, a system that lets researchers watch a smart drug assemble itself inside a living tumor — under flow, embedded in matrix, in three dimensions — marks a meaningful step toward drug tests that predict what will actually happen in patients.