Detailed specifications of used buffers, media, antibodies, and kits please find in the Supplementary Material.

CyanoHAB samples

The preparation of the CyanoHAB biomasses collected from water reservoirs in the Czech Republic is described in Skočková et al. (2023). The phytoplankton cell counts and species composition were determined using microscopic evaluation of sample aliquots fixed with 2% formaldehyde. The cell counts were converted to biovolumes (Hillebrand et al. 1999; Skácelová and Lepš 2014) and expressed as a percentage share of biovolume in a particular sample. The sample codes, dominant groups and species, biovolume percentages for each sample, and identification of sampled localities are summarized in Supplementary Table S1.

LPS preparation and electrophoretic separation

The extraction of CyanoHAB-LPS was performed using a hot phenol−water extraction method, as previously described (Skočková et al. 2023). Isolated and purified LPS were separated by SDS-PAGE followed by silver or Pro-Q™ Emerald 300 staining. Protein contamination was evaluated using SDS-PAGE and Sypro Ruby staining, nucleic acid contamination by agarose gel electrophoresis and ethidium bromide staining. For procedure details see the Supplementary Materials.

DNA extraction

The extraction of DNA from CyanoHAB biomasses was conducted based on previous studies (Sehnal et al. 2022; Skočková et al. 2023). The details of the procedure are described in Supplementary Materials.

qPCR detection of cyanobacteria, heterotrophic bacteria, and Gram-negative heterotrophic bacteria

DNA quantification was performed using qPCR and specific FAM-BHQ1-labeled Taqman probes and primers that targeted 16S rRNA genes as reported previously (Skočková et al. 2023). The primers and probes (Supplementary Table S2) were designed to detect the gene sequences specific for three groups of microorganisms: cyanobacteria, total heterotrophic bacteria, and a subset of G- heterotrophic bacteria. The qPCR was conducted on a LightCycler® 96 Real-time PCR System under the same conditions as reported in a previous study (Lang-Yona et al. 2014). qPCR values (copies of respective 16S rRNA genes) were converted to cell equivalents using a calibration curve constructed using a cultured E. coli CCM 3954 (for total and G- bacteria) or Microcystis aeruginosa PCC7806 (for cyanobacteria). The obtained cell count equivalents were normalized per total isolated DNA, and expressed as a percentage share of each group, with the sum of cyanobacteria and total heterotrophic bacteria considered as 100%. The share of heterotrophic bacteria other than G- bacteria was determined by subtracting G- bacteria from the total heterotrophic bacteria.

Endotoxin activity by PyroGene™ rFC assay

The endotoxin activity of isolated LPS was determined using the PyroGene™ Recombinant Factor C Endpoint Fluorescent Assay (Lonza, Switzerland) following the manufacturer’s instructions, as described previously (Skočková et al. 2023). Fluorescence was measured using a microplate reader (Infinite M200, Tecan, Switzerland; excitation 380 nm, emission 440 nm), and endotoxin activity was calculated according to the standard curve. Data were expressed as endotoxin activity in endotoxin units (EU) per g of biomass dry weight (d. w.) and per mg of the respective LPS, endotoxin activities of CyanoHAB-LPS concentrations used in in vitro experiments were calculated (Table S3).

Cell lines

Caco-2 cells (human colorectal adenocarcinoma cell line; CLS, Germany) were maintained in the Caco-2 medium. The cells were seeded at a density of 2 × 104 cells/cm2 on 24-well cell culture inserts (Corning, USA) and maintained for 21 days to reach a differentiated monolayer. The apical (300 μl) and basolateral (800 μl) culture media were changed three times per week.

HaCaT cells (human keratinocyte cell line; CLS, Germany) were maintained in HaCaT medium. The cells were seeded at a density of 4 × 104 cells/cm2 on a 24-well or 96-well plate (TPP, Switzerland) and maintained for 3 days to reach a confluence.

PBMC isolation

Peripheral blood from healthy donors was obtained based on the collection protocol approved by The Ethics Committee for Research at Masaryk University, Brno, Czech Republic (number EKV-2018-083). The PBMC isolation procedure was based on the work of Meital et al. (2019). The detailed procedure is provided in the Supplementary Materials.

Treatment

HaCaT cells were treated with 50 μg/ml of CyanoHAB-LPS, Caco-2/PBMC 100 μg/ml, and PBMC 1 μg/ml, the non-toxic concentrations (Fig. S1). A detailed description of the procedure is in Supplementary Materials.

Lactate dehydrogenase assay

Cytotoxicity of LPS treatments was evaluated by lactate dehydrogenase (LDH) activity in cell culture medium using the Cytotoxicity Detection Kit PLUS (Roche, USA) as described previously (Binó et al. 2016).

TEER

Trans-epithelial electric resistance (TEER) measures changes in the permeability of the monolayer which increases during inflammation. TEER of the Caco-2 monolayer was measured by ECIS® TEER24 (Applied BioPhysics, USA) at two time points: before the assembly of the co-culture and after 4 days of the exposure to LPS. For the measurement at the end of the experiment, the co-culture was disassembled and only inserts with Caco-2 cell monolayer were put in the TEER measuring plate. As a negative control, two empty inserts with a cultivation medium were used.

Wound healing assay

Wound healing assay studies changes in cell migration and proliferation under pro-inflammatory conditions. Confluent HaCaT cells grown on the 96-well plate were used for the assay. Half of the plates was treated with mitomycin at a final concentration of 1 μg/ml for 2 h to stop the proliferation and observe only migration. After PBS wash, the wounds were made by BioTek AutoScratch Wound Making Tool (Agilent, USA), the cells were washed again, supplemented with HaCaT culture medium, and treated with LPS as mentioned above. The cells were maintained in BioTek BioSpa 8 Automated Incubator (Agilent) and photos of the wounds were acquired by BioTek Cytation 5 Imaging Reader (Agilent) each 6 h for 24 h. The area of the wounds was calculated in the Gen5 program (Agilent).

Flow cytometry

Under pro-inflammatory conditions, the monocytes differentiate into macrophages and change the expression of specific surface molecules CD16 and CD14. After 4 days of exposure, PBMC were collected and the expression of the markers was measured by flow cytometry. Detailed procedure is provided in the Supplementary Materials. The data were analyzed in the FlowJo program (BD Life Sciences, USA). The gating strategy is described in Fig. S2. tSNE analysis included four parameters (FSC, SSC, CD14, CD16).

ELISA

Upon pro-inflammatory stimuli, cells produce cytokines to drive inflammation and regeneration. Concentrations of selected cytokines were measured in the media by commercially available ELISA kits according to the manufacturer’s instructions.

Western blot

Pro-inflammatory activation can change the expression of various proteins by the cell. To study these changes, the cells were lysed, sonicated (Sonopuls, Bandelin, Germany), and boiled (10 min). Protein concentration was measured by BCA assay (PierceTM BCA ProteinAssay kit, ThermoFisher Scientific, USA) according to the manufacturer’s instructions, the samples were adjusted to the same concentration and supplemented with 5 × Laemli buffer. The SDS-PAGE electrophoresis and western blotting were done as described previously (Pekarova et al. 2015). The detailed procedures can be found in Supplementary Materials.

Statistical analysis

Data are presented as a mean ± SD. The number of independently repeated experiments (n) is stated in the figure legend. Statistical analysis was performed using GraphPad Prism version 9.5.1 for Windows (GraphPad Software, USA). Outliers were identified by the ROUT method and filtered out. Significant difference from negative control (named as “control” throughout the text and figures) was determined by unpaired t-test. In case of disagreement of variances, the Mann–Whitney test was used. Data normalized to control were analyzed by one-sample t-test. *p < 0.05, **p < 0.01.