A quiet corner of the Baltic Sea has yielded a surprise for cancer researchers. A strain of soil-dwelling cyanobacteria, Nostoc edaphicum CCNP1411, isolated in 2010 from the Gulf of Gdańsk, has spent years in the spotlight for its unusual peptide chemistry. Now a team of Polish scientists has shown that the compounds most responsible for killing cancer cells in laboratory assays are not the famous peptides at all, but a family of humble lipids that had largely been overlooked. The findings, published in Molecular Diversity, combine classic bioassay-guided fractionation with cutting-edge computational mass spectrometry to reveal that lysophospholipids, monoacylglycerols, and free fatty acids produced by this cyanobacterium carry a substantial share of its cytotoxic punch.
Cyanobacteria of the genus Nostoc are famous among natural product chemists for their enormous genomes, packed with biosynthetic gene clusters that churn out structurally exotic secondary metabolites. N. edaphicum CCNP1411 is a particularly prolific producer. Previous work from the same research groups had catalogued at least three classes of non-ribosomal peptides from this strain: nostocyclopeptides, of which ten structural analogues are made exclusively by Nostoc species and which can inhibit the human 20S proteasome; anabaenopeptins, four analogues with moderate activity against serine proteases and carboxypeptidase A; and cyanopeptolins, the dominant group, with a staggering 96 structural analogues that show strong, selective inhibition of trypsin and chymotrypsin and even potent antiviral activity against SARS-CoV-2 in earlier studies. Cyanopeptolins from this strain had also been shown to be cytotoxic against HeLa cervical cancer cells.
The problem was scale. Out of roughly 120 metabolites identified from 100 grams of freeze-dried biomass, only 13 had ever been tested against cancer cells, because isolating milligram quantities of pure compounds is slow, expensive, and technically demanding. To get around that bottleneck, the team, led by Robert Konkel of the University of Gdansk together with colleagues at the Polish Academy of Sciences’ Institute of Oceanology and the University of Gdansk’s Department of Molecular Biology, turned to a computational strategy built on the Global Natural Products Social Molecular Networking (GNPS) platform.
The workflow began with cultivation. The strain was grown in liquid Z8 medium at a salinity of 7.3 for three weeks under continuous low light, and the biomass was harvested, freeze-dried, and extracted with 75% methanol. The crude extract was then split on a flash chromatography column using a step gradient from 20% to 100% methanol, yielding nine fractions of increasing hydrophobicity. Each fraction was subjected to high-resolution liquid chromatography coupled with tandem mass spectrometry (LC–MS/MS) on a Waters Synapt XS instrument, with data acquired in positive electrospray ionization mode across an m/z range of 200 to 2000. Raw data were processed in MZmine and uploaded to GNPS for feature-based molecular networking (FBMN), which organizes tandem mass spectra into clusters based on spectral similarity, so that structurally related molecules appear as interconnected nodes in a network map. Where library matches were absent, the Network Annotation Propagation tool and the CANOPUS classifier within SIRIUS predicted compound classes from fragmentation patterns alone.
The biological side of the study used the MTT assay, a colorimetric test in which living cells convert a tetrazolium dye into purple formazan crystals, allowing cell viability to be quantified by absorbance. Eleven cell models were tested: seven epithelial cancer lines (A549 lung, C33-A, SiHa, CaSki, and HeLa cervical, PC3 prostate, and T47D breast cancer), two neuronal cancer lines (SH-SY5Y and IMR-32 neuroblastoma), and normal human adult and child dermal fibroblasts as non-cancerous controls. In the initial screen, all samples were tested at 200 micrograms per milliliter, with active fractions then examined across a dose range of 25 to 200 micrograms per milliliter.
The results told a clear story. Fractions eluted with 80% and 90% methanol, the most hydrophobic cuts of the extract, showed broad cytotoxicity against every cancer cell line tested, with relative viability generally falling below 45% at the screening dose. In SH-SY5Y neuronal cells, the effect climbed steadily with solvent strength, from 50.4% viability in the 40% methanol fraction down to 13.8% in the 90% fraction. The most active fractions showed IC₅₀ values, the concentration killing half the cells, ranging from 58.5 to 240.1 micrograms per milliliter, with the 90% fraction the strongest performer overall. Notably, activity against the neuronal lines appeared across nearly all fractions.
The crucial insight came when the researchers overlaid bioactivity onto the molecular network. The peptides that made this strain famous were conspicuously bad candidates for the observed killing. Nostocyclopeptides turned up in every fraction, so they could not explain why only some fractions were potently toxic. Cyanopeptolins were concentrated in the 50–80% fractions, while anabaenopeptins appeared only in trace amounts in the 50–70% fractions. But the 80% and 90% fractions, the ones doing the real damage, were dominated by lipid derivatives: monoacylglycerols in one cluster, lysophospholipids in another, and a steroid derivative bearing a sterane backbone with an oxetane ring, identified as 12-OH-3,7-diAc-17-oxetane-sterane, in a third. In total, the network mapped 638 nodes connected by 1126 edges across 42 clusters, with 19 clusters of unidentified compounds in the most active fractions predicted by CANOPUS to belong primarily to glycerolipids and lipopeptides.
To validate the computational prediction, the team purchased seven of the identified lipids that were commercially available: three lysophospholipids, two monoacylglycerols, and two free fatty acids. Because these hydrophobic molecules dissolve poorly in standard assay conditions, the researchers developed a bovine serum albumin-assisted solubilization protocol, forming thin lipid films and reconstituting them in fatty acid-free BSA at 37 degrees Celsius overnight. The purified compounds were then tested individually. At the screening concentration, lysophospholipids were active against every cell line. The standout was PE(18:1(9Z)/0:0), which inhibited proliferation by 38 to 98%, with the strongest effect on IMR-32 cells. Among the monoacylglycerols, monolinolein, the linoleic-acid-containing species, was powerfully cytotoxic across the board, inhibiting 78 to 98% of cell growth, while monoolein, its monounsaturated cousin, barely exceeded 22% inhibition. Palmitic acid knocked down T47D and IMR-32 viability by 88% and 98%, respectively.
Dose–response experiments, however, tempered the enthusiasm. Most compounds showed full activity only at higher doses, and IC₅₀ values typically exceeded 100 micromolar, which the authors classify as limited potency. Still, ten compound–cell line combinations dipped below the 50 micromolar threshold for high potency, and a few were striking: the lysophospholipid PE(16:0/0:0) hit T47D breast cancer cells with an IC₅₀ of just 2.4 micromolar, and palmitic acid achieved 9.0 micromolar against IMR-32 neuroblastoma cells. Monolinolein was effective against the cervical cancer lines CaSki and C33-A at 28.8 and 40.1 micromolar, while monoolein selectively targeted the neuroblastoma lines SH-SY5Y and IMR-32.
The structural nuances in these data are striking. Palmitic acid, a fully saturated C16 fatty acid, and palmitelaidic acid, which differs only by a single trans double bond, behaved completely differently, with the latter showing moderate activity only against SH-SY5Y cells. This is consistent with a large body of literature showing that saturated fatty acids induce lipotoxicity through membrane disruption, endoplasmic reticulum stress, and mitochondrial dysfunction, while polyunsaturated species such as linoleic acid often trigger oxidative stress in cancer cells. The authors also note an interesting tension with earlier reports: some studies have found lysophosphatidylethanolamines stimulating the growth of breast cancer cells and even driving tumor progression in mouse models of renal cell carcinoma, while elevated LPE levels have been observed in cancer tissue. Resolving these inconsistencies, the researchers argue, will require mechanistic work on how these lipids interact with cellular membranes and metabolic pathways in different cell types.
Several caveats temper the clinical promise. Some compounds also reduced the viability of healthy fibroblasts, a selectivity problem that the authors acknowledge would need to be addressed through therapeutic-window analysis, chemical optimization of lead structures, or targeted delivery strategies such as antibody–drug conjugates or nanoparticles. The lipids’ low water solubility and high lipophilicity pose further pharmacokinetic hurdles, and in vitro selectivity does not always translate into animal or human outcomes. Still, the selective potency of several compounds against neuronal and epithelial cancer lines, combined with their low molecular weight of roughly 200 daltons, makes them worth pursuing as lead structures.
Perhaps the most significant contribution of the study is methodological. By coupling bioassay-guided fractionation with feature-based molecular networking, the team demonstrated that computational metabolomics can shortcut the traditional grind of isolation and purification, directly linking chemical composition to biological effect in a complex cyanobacterial extract. Lipids, long treated as background noise in natural product drug discovery compared with flashy peptides and polyketides, have now been shown to make a measurable contribution to the anticancer potential of Nostoc edaphicum. The work was funded by the National Science Centre Poland through the CYANOCRAB project, and the underlying mass spectrometry datasets have been deposited in the Zenodo repository for future analyses. For a strain first fished out of the Baltic more than fifteen years ago, CCNP1411 continues to prove that even the humblest pond scum can keep a secret worth finding.
Subject of Research: Cytotoxic lipid metabolites produced by the cyanobacterium Nostoc edaphicum CCNP1411
Subject of Research: Medicine
Article Title: Molecular-networking-based characterization of cytotoxic metabolites produced by the cyanobacterium Nostoc edaphicum CCNP1411
Article References: Konkel, R., Cegłowska, M., Grabowski, Ł., Jeleniewska, W., Zielenkiewicz, M., Gonera, K., Węgrzyn, G., & Mazur-Marzec, H. (2026). Molecular-networking-based characterization of cytotoxic metabolites produced by the cyanobacterium Nostoc edaphicum CCNP1411. Molecular Diversity. https://doi.org/10.1007/s11030-026-11717-w
Image Credits: AI Generated
DOI: 10.1007/s11030-026-11717-w
Keywords: cyanobacteria, Nostoc edaphicum, molecular networking, cytotoxicity, lysophospholipids, monoacylglycerols, free fatty acids, LC–MS/MS, bioassay-guided fractionation, GNPS, anticancer, natural products
Cite Scienmag News
APA MLA Chicago
Ophelia Keating. (September 10, 2026). Cytotoxic metabolites from cyanobacterium Nostoc edaphicum characterized via molecular networking. Scienmag. https://scienmag.com/cytotoxic-metabolites-from-cyanobacterium-nostoc-edaphicum-characterized-via-molecular-networking/
Ophelia Keating. “Cytotoxic metabolites from cyanobacterium Nostoc edaphicum characterized via molecular networking.” Scienmag, 10 September 2026, https://scienmag.com/cytotoxic-metabolites-from-cyanobacterium-nostoc-edaphicum-characterized-via-molecular-networking/. Accessed 10 September 2026.
Ophelia Keating. “Cytotoxic metabolites from cyanobacterium Nostoc edaphicum characterized via molecular networking.” Scienmag. September 10, 2026. https://scienmag.com/cytotoxic-metabolites-from-cyanobacterium-nostoc-edaphicum-characterized-via-molecular-networking/
Copy citation Download RIS
Tags: bioassay-guided fractionation of cyanobacteriabioassay-guided fractionation of cyanobacterial compoundsbiosynthetic gene clusters in Nostoc speciescomputational mass spectrometry in natural product analysiscyanobacteria-derived cytotoxic lipidsfatty acids and monoacylglycerols from cyanobacterialipid-based cytotoxic compounds from cyanobacterialysophospholipids and cancer cell cytotoxicitylysophospholipids in cancer therapymarine cyanobacteria bioactive compoundsmass spectrometry in cyanobacterial metabolite analysismolecular networking in natural product discoverynatural product chemistry of Baltic Sea cyanobacterianatural products with anticancer activityNostoc edaphicum secondary metabolitesnovel lipids with anticancer activitystructural diversity of cyanobacterial secondary metabolites
