Introduction: A breast tumor is more stroma than malignant epithelium, and tumor-associated macrophages are among the most abundant residents of that stroma. These macrophages act on the cancer cells around them largely at a distance, through the extracellular vesicles they release, and what those vesicles contain determines what the recipient cell does next. Much of that content is non-coding RNA. Circular RNAs, in particular, can bind microRNAs and keep them from their targets. One such molecule, circ_0005519, is oncogenic across several human cancers and is predicted to bind miR-98-5p, a tumor suppressor whose targets include interleukin-10, the cytokine that holds the tumor environment in an immunosuppressive state. Exotoxin A from Pseudomonas aeruginosa has been valued for killing cells outright, by ADP-ribosylating elongation factor 2. Here we asked a different question: whether exotoxin A also alters the circRNA cargo that surviving macrophages release, and whether that shift reaches the miR-98-5p/IL-10 axis in recipient breast cancer cells.
Methods: Primary human tumor-associated macrophages and MCF-7 cells were cultured in DMEM with 10% fetal bovine serum at 37 °C under 5% CO2. Macrophage viability was measured by MTT in 96-well plates after 24 and 48 h exposure to exotoxin A (Merck), added at 10–40 µL per 200 µL well. Macrophages were then treated for 48 h alongside untreated controls, and conditioned medium from each was cleared at 300 ×g before the vesicles were precipitated with a polymer-based kit (15,000 ×g, 30 min, 4 °C). Vesicle size was verified by dynamic light scattering. MCF-7 cells received 150 µL of either exosome preparation before harvest. RNA was extracted with TRIzol, assessed on a NanoDrop and treated with DNase I. cDNA was primed with random hexamers for circ_0005519 and IL-10, and by poly(A) tailing with oligo-dT for miR-98-5p. LNA primers and SYBR Green were run on an ABI StepOne with melt-curve analysis. GAPDH normalized circ_0005519 and IL-10; U6 snRNA normalized miR-98-5p. The experiment was performed once in duplicate wells and analyzed by 2^−ΔΔCt with an independent t-test (SPSS 19; P<0.05)
Results: Exotoxin A reduced macrophage viability in a volume-dependent manner over both intervals. Viability fell from 84.7% at the lowest volume tested to 72.6%, 54.2% and 34.6% as the volume was raised, and the point at which roughly half the population remained viable was carried forward into all subsequent work. Vesicles recovered from the conditioned medium of these cultures gave a single narrow peak near 100 nm by dynamic light scattering, within the size range accepted for exosomes. RNA yields ranged from 220 to 287 ng/µL with A260/A280 ratios of 1.81 to 1.98, and melt-curve analysis returned one peak per target at its expected melting temperature.
Exosomes shed by exotoxin A-treated macrophages carried roughly half the circ_0005519 of those from untreated macrophages (0.53 versus 1.00 relative units; P<0.05). The depletion therefore occurred in the cargo itself, not only in the number of cells releasing it.
MCF-7 cells that received the circ_0005519-poor exosomes showed a 2.5-fold rise in miR-98-5p over cells given control exosomes (2.55 versus 1.00; P<0.01). IL-10 transcript in the same cells fell to approximately one third of control (0.35 versus 1.00; P<0.01). The two molecules moved in opposite directions within a single cell population, which is the pattern a sponging relationship would produce.
Conclusion: Exotoxin A did more to these macrophages than kill a fraction of them. The vesicles released by the survivors carried substantially less circ_0005519, and breast cancer cells that took up those vesicles responded with higher miR-98-5p and lower IL-10 transcript, despite never meeting the toxin themselves. The stromal macrophage, rather than the malignant cell, is therefore a viable point of entry into this axis. Two limitations bound the claim: the experiment was run once in duplicate wells, and circ_0005519 was measured in the vesicles rather than in the recipient cells, so the sponging step is inferred from the direction of change rather than demonstrated. Confirming IL-10 at protein level and tracking circ_0005519 inside MCF-7 would close both gaps.