Counting Algae Is Not Enough: Gill-Toxic Potency Could Transform Fish-Kill Warnings

Fish-killing harmful algal blooms are among the most economically devastating hazards in marine aquaculture, and the monitoring systems designed to warn of them may be measuring the wrong thing. A new Perspective published in Discover Ecology by Jorge I. Mardones of Chile’s Institute for Fisheries Development argues that the cell-counting paradigm inherited from shellfish-toxin surveillance is structurally incapable of anticipating acute fish mortality, and proposes a radical reframing: monitoring should target what he calls waterborne gill-toxic potency, the measurable capacity of a water mass to damage fish-gill tissue under environmentally realistic exposure. The argument draws on two decades of ecotoxicology and on hard lessons from some of the worst aquaculture disasters on record.

The scale of the problem is enormous. The 2016 bloom of the flagellate Pseudochattonella verruculosa in southern Chile generated losses exceeding 800 million US dollars. Three years later, blooms of Chrysochromulina leadbeateri in Norwegian fjords produced the largest harmful-algal-bloom impact ever recorded on northern European aquaculture. The culprit taxa span a phylogenetically distant assemblage of raphidophytes, kareniacean dinoflagellates, prymnesiophytes and dictyochophytes that share no common toxin class, yet their damage converges on a single organ: the fish gill, whose epithelium maintains a continuous, unbuffered interface with the surrounding water.

The surveillance architecture in use today was built for a different problem. In shellfish monitoring, filter-feeding bivalves integrate exposure to intracellular phycotoxins, certified analytical standards allow direct quantification of toxin burden in tissue, and that tissue burden, not cell abundance in seawater, triggers regulatory action. Transferred to fish-killing blooms, this logic carries a hidden premise: that cell abundance translates roughly into proportional mortality risk. Field evidence says otherwise. During the 2016 Chilean event, routine cell counts showed no statistically significant correlation with mortality across affected farms, partly because thin layers concentrated the bloom vertically and fragile cells lost integrity under standard fixation. Nearly a decade later, a Prymnesium parvum bloom in the Oder River reached densities comparable to the catastrophic 2022 fish kill yet caused no documented mortality, with temperature, river flow and salinity acting as conditional drivers of toxin production independent of cell numbers.

Laboratory work reinforces the decoupling between abundance and hazard. Screening of Chilean fish-killing flagellates with the RTgill-W1 rainbow trout gill cell line revealed that species differed by more than an order of magnitude in cytotoxic potency under standardised conditions, and that mechanically disrupted cells consistently produced stronger responses than intact suspensions. The likely explanation combines active secretion of defensive metabolites across the plasma membrane with experimental release of polyunsaturated fatty acids and reactive oxygen species during cell rupture. In other words, what fish gills actually encounter at cage depth depends on bloom physiology and cell integrity, not on how many cells are floating in a sample.

The Perspective dissects why the shellfish triad of intracellular toxin, bivalve sentinel and certified standard cannot simply be completed for fish. The relevant bioactive agents in most ichthyotoxic blooms reach the water through active export, stress-induced production or cell lysis; no sentinel organism integrates ichthyotoxin exposure on behalf of fish before injury occurs; and for most fish-killing taxa the chemical identities of the harmful compounds remain unknown, leaving the field without certified analytical standards. Even the genus Alexandrium exposes the divergence: the same species can produce paralytic shellfish toxins that fit neatly into certified workflows while simultaneously releasing structurally unrelated extracellular compounds that remain largely uncharacterised. In Chilean Alexandrium catenella, paralytic toxin quotas failed to explain gill-cell cytotoxicity, which was instead driven by polyunsaturated fatty acids acting synergistically with cell-derived superoxide.

Where the chemistry has been resolved, a common functional theme emerges. Karlotoxin 2 from Karlodinium veneficum is a large amphipathic polyketide that non-selectively increases plasma-membrane permeability, producing ionic imbalance and osmotic lysis. Prymnesins from P. parvum disrupt chloride and calcium regulation and kill gill cells at nanomolar concentrations. Chilean A. catenella lysates induce net potassium efflux from gill cells while purified paralytic toxins from the same strain produce no measurable ionic disturbance. For Chattonella marina, superoxide alone is insufficient to kill fish and fatty acid alone acts only at high concentrations, but together they triple the mortality rate. These findings argue against treating any fish-killing bloom as a single-target toxin problem.

The proposed alternative is a tiered, effect-based framework. Tier 1 retains taxonomic surveillance but redesigns sampling around fish exposure, with depth-resolved sampling at cage depth paired with hydrographic measurements, because vertically structured blooms can expose fish to concentrations that surface samples systematically miss. The extreme Heterosigma akashiwo bloom that killed more than 6,000 tonnes of farmed salmon in Chile’s Comau fjord in 2021 reached 70,000 cells per millilitre concentrated in the upper three metres under strong stratification. Tier 2 activates when a known fish-killing taxon is detected, measuring gill-toxic potency in paired water fractions: whole water, cell-free filtrate, resuspended cell-associated material and standardised lysate. The RTgill-W1 assay, which reads metabolic activity, membrane integrity and lysosomal integrity on the same cell population, showed EC50 values within a factor of five of whole-fish acute toxicity for 73 percent of 35 benchmark chemicals, with interlaboratory variation around 31 percent.

Tier 3 pursues chemical discovery through effect-directed analysis coupled with non-target high-resolution mass spectrometry, fractionating active samples, retesting each fraction and prioritising only those retaining activity. Two recent demonstrations prove the workflow works for fish-killing taxa. Bioassay-guided fractionation of the 2019 Norwegian C. leadbeateri blooms concentrated gill-toxic activity into a single fraction and yielded a candidate compound, proposed as leadbeaterin-1, detected in situ in bloom water. During the 2022 Oder River P. parvum disaster, in which several hundred tonnes of fish died, effect-based analysis detected B-type prymnesins alongside more than 120 organic micropollutants, with mixture modelling showing the neurotoxicity was predominantly explained by the prymnesins. The institutional urgency is documented: the IOC-UNESCO biotoxin database held only one formally catalogued ichthyotoxin as of March 2025, against more than 120 fish-killing microalgal taxa reported worldwide.

Mardones is explicit that the framework is a proposal requiring validation, not a system ready for regulation. No study has yet compared gill-cell potency in bloom water with mortality outcomes in caged fish at the same site and time, and the assay has documented blind spots, including compounds requiring metabolic activation and the mismatch between a freshwater-derived trout cell line and marine salmonid aquaculture. Marine matrix effects pose practical problems too: karlotoxins adsorb quantitatively onto standard filter membranes, meaning routine sample handling can erase the very signal being measured. Universal toxicity thresholds are premature, and the first generation of action levels should be calibrated locally against paired observations of blooms, fish behaviour, gill pathology and mortality. Even mitigation carries risk, since hydrogen peroxide treatment of dinoflagellate blooms has been shown to increase gill-cell cytotoxicity by generating persistent toxic aldehydes.

The deeper message is a convergence of fields. Fish-killing HAB science, the Perspective argues, should formally adopt the methodological infrastructure that the ecotoxicology of emerging contaminants built over two decades to assess hazard from chemically unresolved mixtures: effect-directed analysis, high-resolution mass spectrometry, adverse outcome pathways and effect-based monitoring. Cell counts remain indispensable for tracking blooms and forecasting exposure potential, but they answer only whether a hazard may be present. Waterborne gill-toxic potency, once validated, would answer the question that matters to a salmon farmer watching the water: whether that hazard is being expressed right now, at the depth where the fish are breathing.

Subject of Research: Effect-based monitoring of fish-killing harmful algal blooms using waterborne gill-toxic potency as an endpoint

Article Title: Waterborne gill-toxic potency as an effect-based endpoint for monitoring fish-killing harmful algal blooms

Article References: Mardones, J. I. (2026). Waterborne gill-toxic potency as an effect-based endpoint for monitoring fish-killing harmful algal blooms. Discover Ecology, 2(1), Article 28. https://doi.org/10.1007/s44396-026-00045-w

Image Credits: AI Generated

DOI: 10.1007/s44396-026-00045-w

Keywords: harmful algal blooms, ichthyotoxicity, aquaculture, gill-toxic potency, RTgill-W1, effect-directed analysis, high-resolution mass spectrometry, Pseudochattonella verruculosa, Prymnesium parvum, ecotoxicology, fish kills, monitoring