기본 콘텐츠로 건너뛰기

HEADLINE DESK

DKBEE.COM 최신 주요 리포트

최신 기사를 불러오는 중입니다...

LATEST INSIGHTS

분야별 심층 분석

PFAS in the buzz: Seasonal biomonitoring with honey bees (Apis mellifera) and bee-collected pollen☆

 

PFAS in the buzz: Seasonal biomonitoring with honey bees (Apis mellifera) and bee-collected pollen☆

Per- and polyfluoroalkyl substances (PFAS) are emerging environmental contaminants with known persistence, bioaccumulation potential, and ecotoxicological relevance. This study investigates the occurrence, distribution, and potential exposure pathways of PFAS in honey bee-related matrices (pollen, forager bees, hive bees). Samples were collected from six sites over a full season representing diverse land uses (urban, semi-urban, rural, and agricultural). A total of nine different PFAS were detected across 90 samples, with short-chain compounds such as 4:2 FTS, PFPS, and PFBS predominating. Hive bees exhibited the highest average sum of target PFAS concentration (5.29 ng gꟷ1), followed by forager bees (1.93 ng gꟷ1), and pollen (1.10 ng gꟷ1). Differences in PFAS occurrence among sample types, especially PFPS between forager and hive bees, suggested various exposure routes, including pollen ingestion, atmospheric deposition, and contact with contaminated surfaces. Temporal and spatial differences were notable. Site-specific PFAS patterns highlighted potential contamination sources, such as fluorinated pesticide use, while seasonal variation suggested influence from plant phenology and rainfall. Principal component analysis (PCA) confirmed the dominance of shared PFAS profiles, with some outlier samples exhibiting unique signatures. PFPS and PFBS were positively correlated with total fluorine (TF) concentrations, while 4:2 FTS showed a strong site-specific association. Extractable organic fluorine (EOF) remained below detection limits (800 ng F gꟷ1), but TF was present in all samples above LOD (500 ng F gꟷ1) confirming the presence of additional unidentified fluorinated compounds. Although EOF and PFAS concentrations varied, we showed that forager bees can introduce PFAS in the hive, either on their bodies, or through contaminated resources such as nectar or pollen. These findings demonstrate the suitability of honey bees as bioindicators for PFAS contamination but emphasise the need for harmonized sampling to fully assess PFAS exposure dynamics.

Keywords

LC-MS/MS
Hive bee
Forager bee
Target analysis
Extractable organofluorine
Total fluorine

1. Introduction

Per - and polyfluoroalkyl substances (PFAS) are of growing concern to the environment due to their strong chemical and thermal stability originating from the strong C-F bonds. This enables PFAS bioaccumulation, which is observed not only in long chain (C > 8), but also short chain (C < 8) and ultrashort chain (C < 3) PFAS as well. The use of sentinel animals, including humans, is an effective way to monitor PFAS contaminants in the environment (Aristizabal-Henao et al., 2021; Andrews et al., 2023; Rock et al., 2023; Khalid et al., 2024; Libenson et al., 2024).
Bees usually forage around a two to three km radius of the hive (Garbuzov et al., 2015, 2020; Danner et al., 2017), thus making them an excellent sentinel animal for local contaminations (Bargańska et al., 2016; Cunningham et al., 2022; Farias et al., 2023; Glinski et al., 2024) of pesticides (Murcia-Morales et al., 2020a, 2020b; Catalano et al., 2024 Donley et al., 2024), microplastics (Edo et al., 2021), heavy metals (Zarić et al., 2022; Bora et al., 2023), polycyclic aromatic hydrocarbons (PAHs) (Lambert et al., 2012; Murcia-Morales et al., 2024; Catalano et al., 2024), dioxins and dioxin-like compounds (Catalano et al., 2024), fine particles (Meza-Figueroa et al., 2024), particulate matter (Marcoccia et al., 2024), explosives (Filipi et al., 2022), but also antimicrobial resistance (Resci et al., 2024). The most studied matrix is honey as it is not only a possible storage for the pollutants but is also a human food (e.g. Bora et al., 2023; Nowak & Nowak, 2021; Smith et al., 2019). Also, live or dead bees are commonly sampled (Zarić et al., 2022). During the active foraging season, honey bee workers live only a few weeks, which increases the time increments that can be analysed. Furthermore, honey bees show an interesting division of labour: Whereas young bees stay in the hive to consume bee bread (stored pollen) and feed larvae, older bees fly out foraging on plants for nectar, honeydew or pollen (Seeley, 2009).
Although there are limited investigations of PFAS in honey bees and bee products, it was shown that honey originating from an industrial region contained 20 % higher concentrations of perfluorocarboxylic acids (PFCAs) than those from a non-industrial origin (Surma et al., 2016), and that bee colonies exposed to perfluorooctansulfonic acid (PFOS) show an adverse effect on the colony activity, hive maintenance and defence, amongst others (Sonter et al., 2021). All of these highlight the importance of investigating PFAS in bees and hive products. However, there are many things to be considered during sampling, as plants tend to accumulate PFAS at different rates based on the environmental conditions (e.g. water/rain). The use of fluorinated pesticides could also influence PFAS concentrations and uptake as these pesticides break down over time to trifluooracetic acid (TFA) (Joerss et al., 2024).
The above-mentioned publications showed the presence of PFAS in bee products and possible adverse effects of PFAS on colonies, however, they do not give a picture of the PFAS contamination of the bees, which eventually are in contact with the environment. Therefore, the aim of this study was to investigate PFAS in different colonies with a multi-platform approach using liquid chromatography coupled to electrospray tandem mass spectrometer (LC-ESI-MS/MS) for targeted analysis, and combustion ion chromatography (CIC) for assessing the total fluorine content and extractable organofluorine (EOF) content in bees and pollen. We also utilised a mass balance approach to give a better overview about the PFAS burden in the environment using bees and bee products as biomonitoring tools.

2. Material and methods

2.1. Sampling areas

Samples were taken from honey bee colonies (Apis mellifera) at six locations in Austria. Samples were taken from one hive per location monthly from April until August 2024. Samples consisted of pollen collected in a pollen trap lined with disposable paper towel, bees flying out or returning to the hive, as well as bees from inside the hive. For the latter, bees were collected from the outermost frame of the upper box (according to van der Steen et al., 2012). Briefly, three sample locations were Styria, in the city of Graz (suburban gardens, suburban meadows and an urban site), two rural sites in upper Austria, upwind of the Burghausen (Germany) industrial zone, and one in Lower Austria, in an agricultural area. Details on sampling dates and description of sampling locations are shown in Table 1 and Fig. S1. Samples in Graz were taken by one of the authors and all others by experienced beekeepers that were instructed and supplied with sample bags. Non-Graz samples were frozen and shipped after the season. All samples were dried at 60 °C for 3–4 days prior to chemical analysis.

Table 1. Details on samplings and locations.

Location codeNameDescription of locationSampling dates 2024
Suburban Gardens – GGraz suburban gardensCemetery, many small gardens and many houses with gardens28.4., missed, 8.6., 15.7., 16.8.
Suburban meadows – GGraz suburban meadowsSuburban area with meadows and forests26.4., 24.5., 28.6., 6.7., 12.8.
Urban – GGraz urbanUniversity campus, surrounded by many buildings, private gardens and public parks29.4., 22.5., 18.6., 15.7., 13.8.
Agricultural – VSt. ValentinAgricultural site12.4., 15.5., 18.6., 11.7., 16.8.
Rural – UÜberackernRural, ca.6 km upwind from industrial zone30.4., 23.5., 20.6., 18.7., 22.8.
Rural – OOttenschwandRural, ca.12 km upwind from industrial zone27.4., 27.5., 19.6., 24.7., 21.8

2.2. Chemicals

PFAS standards include a native PFAS mix (PFAC30PAR) and internal standard solution (MPFAC-HIF-ES) were purchased from Wellington Laboratories (Canada). PFAC30PAR contains perfluoro carboxylic acids (PFCA) ranging from C4 – C14, perfluorosulfonic acids (PFSA) ranging from C4 – C10, C4, C6 and C8 perfluoroalkanesulfonamides, methyl perfluoro octanesulfonamidoethanol, and ethyl perfluorooctanesulfonamidoethanol, GenX, Nadona, 9Cl-PF3ONS, and 11Cl-PF3OUdS. MPFAC-HIF-ES contains 13C labelled C4 – C12, C14 PFCA and C4, C6, C8 PFSA, C4, C6, C8 perfluorosulfonates (FTS), GenX, methyl perfluoro octanesulfonamide, ethyl perfluorooctanesulfonamide, methyl perfluoro octanesulfonamidoethanol, and ethyl perfluorooctanesulfonamidoethanol.Ultrapure water was provided from Millipore Ultrapure Water System (18 MΩ cm, Merck). LC/MS grade ammonium acetate (CH3COONH4) was purchased from Fischer Chemicals (Austria). Methanol (MetOH) and acetonitrile (ACN), both LC/MS grade, were purchased from Bartelt (Austria). More information on chemicals and instruments used can be found in the supporting information tables S1-2.

2.3. Sample preparation

2.3.1. Targeted analysis

0.4 g of dried and ground samples were weighed into centrifuge tubes, then spiked with 20 μL of 16–40 μg L – 1 MPFAC-HIF-ES mix. After 1 h, the samples were extracted with 40 mL of 50 % (v/v) methanol: water with 0.1 % (v/v) ammonium hydroxide. The solution was vortexed then placed into a sonication bath for 1 h. After the sonication the solution was centrifuged at 2500 g force for 5 min. The supernatant was decanted into a new centrifuge tube (methanol fraction). The extracts were evaporated until dryness, then reconstructed into 1 mL 100 % (v/v) methanol, and transferred into Eppendorf vials containing ∼25 mg active carbon for clean-up purposes. The mixture was vortexed and centrifuged at 10,200 g force for 5 min. The supernatant was transferred into muffled LC vials and analysed for targeted analysis with LC-ESI-MS/MS.

2.3.2. Extractable organofluorine (EOF) content

For EOF analysis the same amount was extracted as described above without the addition of the isotopically labelled internal standards.

2.4. Instrumentation

2.4.1. Targeted analysis with LC-ESI-MS/MS

An Agilent 1260 infinity II HPLC (Agilent Technologies, Germany) combined with a BrownLee SPP C18 column (2.7 μm, 3 × 100 mm, PerkinElmer, UK) and a BrownLee SPP guard column (2.7 μm, 3 × 5 mm, PerkinElmer, UK) were used for the separation of the analytes by LC. A gradient was used for the separation of target PFAS, consisting of 5 mmol L−1 CH3COONH4. The LC system was coupled to an Agilent Ultivo LC/TQ (Agilent Technologies, Germany). The MS was used in negative multiple response monitoring (MRM) mode. Two transitions were monitored (quantifier and qualifier ions) for the analytes except PFBA, PFPA, PFdDA, PFOSA, and the ultrashort chain PFAS where only one transition was monitored. Transitions and LC methods are listed in Table S3 in SI. Due to the large number of transitions, the transitions were monitored only around their retention times.

2.4.2. EOF and TF analyses with CIC

EOF analysis was carried out using a Thermo-Mitsubishi Analytech CIC, as described previously (Schröder et al., 2024). Briefly, sample extracts (100 μL) and solids (∼30 mg) were placed in a ceramic sample boat containing glass wool for better dispersion of the fluids. Sample boats were pre-baked 3 times prior to sample combustion to minimize background contamination (Schultes et al., 2019). Samples were introduced into a combustion oven (HF-210, Mitsubishi Analytech) heated to 1100 °C with argon, oxygen, and argon with water. All combustion gases were collected in ultrapure water (GA-210, Mitsubishi Analytech), and the absorption solution (ultrapure water) was adjusted to 10 mL. 0.1 mL was injected onto a Dionex IonPac AS19 column (2 × 250 mm) equipped with a guard column (Dionex IonPac AG16 (2 × 50 mm). A gradient of 4 mM–50 mM potassium hydroxide (KOH) with a flow rate of 0.28 mL/min was used for the chromatographic separation.

2.5. Quality control

2.5.1. Targeted analysis

Randomly chosen bee and pollen samples were spiked with the analytes to calculate the recovery rates of the individual analytes. 10 μL of 1 mg L – 1 PFAS-30-PAR and 20 μL of 2 mg L – 1 PFAS USC mix was added to the samples, then extracted the same was as described above, to assess the recovery rates. Three randomly chosen pollen samples were also subsampled and extracted in triplicates to assess repeatability of the extraction and clean-up process. Three method blanks were processed, and in case of observing PFAS in the blanks, the average area of the method blanks was subtracted from the samples. A calibration curve with seven points, ranging from 0.05 ng g – 1 to 50 ng g – 1 in methanol was measured during the run. To assess instrumentation drift, two calibration solutions (0.5 ng g – 1 and 5 ng g – 1) were measured after every 20 samples. Quantification was performed using the analyte:isotopically labelled internal standard ratio. The instrumental limit of quantification and detection (LOQ and LOD) were calculated as ten times and three times the error of the linear regression divided by the slope, respectively. Method LOQ and LOD were calculated by multiplying the respective value with the dilution factors derived from the sample preparation and dividing this by the average sample mass (0.5 and 1 g respectively). Results of recovery analysis, instrumental and method LODs can be found in Tables S5–6.

2.5.2. EOF and TF

To assess the recovery and repeatability of the extraction and clean up processes, pollen and bee samples were spiked with 20 μL of 1000 mg L – 1 PFOA solution, then extracted as described above. For quantification, a calibration curve from 100 to 10000 μg L−1 in ultrapure water was prepared using single element fluoride standard (ROTI Star, 1000 mg L−1 in water). Ultrapure water was used as a blank. Limits of detection and quantification (LOD and LOQ respectively) were calculated by 3 times and 10 times the standard deviation of the blank divided by the slope of the calibration curve. Results of recovery analysis, instrumental and method LODs can be found in Tables S5–6.

2.5.3. Statistical data analysis

A list with all the calculated concentrations can be found in Tables S7–10. To explore the dataset, principal component analysis (PCA) was carried out in GraphPad Prism version 10.4.1. All analytes with positive detection > LOD have been included in the analysis and in the figures. For all statistical tests and further data processing, concentrations below the LOD were set to zero, while the values between LOD and LOQ remained unmodified. Unless stated otherwise, average concentrations are reported and referring to concentration in dry weight. Significance was tested by the nonparametric Mann-Whitney-Test, with a confidence level of α = 0.05. Additionally, Kruskal-Wallis-test was performed followed by Dunn's Post-Hoc-test to compare the medians of more than two independent groups and identify which groups are the significant ones.

3. Results and discussion

3.1. Targeted analysis

3.1.1. Different sample types

From the 30 monitored PFAS, only a total of 9 were observed in the samples (n = 30 for each type of sample), with short-chain PFAS being predominant. The most frequently detected compounds were 4:2 FTS in pollen (77 %), PFPS in forager bees (77 %) and PFBS as well as 4:2 FTS (30 % each) in hive bees (Fig. 1). 4:2 FTS, PFBS, NaDONA and FBSA were detected across all sample types, suggesting a direct exposure link between pollen – forager bees – hive bees. It is important to acknowledge that the recoveries of 4:2 FTS and FBSA across pollen and hive bees were very low (<2 %). This indicates that the applied analytical method was not able to fully extract these compounds, which could lead to the underestimation of their actual concentrations. As such, the results for these specific PFAS should be considered semi-quantitative. One contributing factor is that the extraction method was optimised using forager bees and a smaller suite of PFAS (C4-C8 PFCA and PFSA), for which the recoveries in forager bees are within acceptable ranges. However, broadening the method to include a wider range of PFAS involves a trade-off for the recoveries, which can be seen in the recoveries of the other PFAS. The use of not PFAS matched internal standard for certain PFAS (e.g. 13C PFOS was used for all the PFSA which did not have a corresponding matched IS) could also contribute to having lower recoveries. Additionally, the observed differences in recoveries between forager bees and the rest of the sample types are likely due to matrix effects. For example, pollen and hive bees show similar recoveries and may share similar matrix characteristics, as hive bees contain undigested pollen in their intestine.
Fig. 1
  1. Download: Download high-res image (288KB)
  2. Download: Download full-size image
Fig. 1
PFHxA and N-EtFOSA were only detected in forager bees and pollen, suggesting a similar exposure route for these two PFAS. PFPS was only found in forager and hive bees while PFOA was only found in pollen. The presence of PFAS in both forager and hive bees but not in pollen indicates that PFPS has a different exposure route, potentially originating from contaminated water, nectar or resin. Another possible exposure is atmospheric deposition, where particles with bound PFAS on them can deposit on the outside of the bee, and are transported back to the hive. Nevertheless, PFAS found exclusively in pollen suggests that although the pollen can be contaminated via atmospheric deposition or plant uptake (Wright et al., 2025), transport from pollen to hive does not necessarily contaminate bees. For example, PFOA (which was only found in pollen) binds strongly to pollen and does not transfer into the bees’ food web.
Average ∑Target PFAS were 1.93 ng gꟷ 1 (median: 0.51 ng gꟷ 1) for forager bees, 5.29 ng gꟷ 1 (median: 0.51 ng gꟷ 1) for hive bees and 1.10 (median: 1.06 ng gꟷ 1) ng gꟷ 1 for pollen. The big difference between the median and mean values in the forager and hive bees indicates that the data distribution is skewed towards the lower end of the distribution, with a few outliers pulling the mean away from the median.

3.1.2. Differences between pollen, forager and hive bees

Fig. 2, Fig. 3 show PFAS profile and concentration in forager, hive bee and pollen from April to August, with one graph corresponding to each location. From the observed PFAS species, only PFPS showed significant differences (p < 0.05, Mann-Whitney test) between forager and hive bees with forager bees having higher PFPS. The other PFAS (4:2 FTS, PFHxA, 4:2FTS, PFNA and PFBS) showed p values above 0.05 (Mann-Whitney test). At the Urban – G site, hive bees showed an increase in PFAS concentration from April to May, followed by a sharp decrease in June. Not only the concentrations have changed but the speciation has also shifted, mirroring the PFAS profile in pollen collected in July. A subsequent increase in August was again followed by a sharp decrease in September in the PFAS concentration in hive bees. There are several explanations for this phenomenon. Firstly, PFAS that entered the hive might accumulate in different bee products (such as honey, wax or propolis), rather than remaining in the hive bees. Secondly, the observed decrease can be attributed to the continuous turnover of hive population, as bees during summer live only up to 6 weeks, leading to the replacement of the older, contaminated bees with newly emerged ones, less contaminated with PFAS.
Fig. 2
  1. Download: Download high-res image (313KB)
  2. Download: Download full-size image
Fig. 2
Fig. 3
  1. Download: Download high-res image (1MB)
  2. Download: Download full-size image
Fig. 3
Interestingly, the observed concentrations for 4:2 FTS, PFBS, Nadona, FBSA and PFNA were lower, although not significantly in the forager bees than in the hive bees. An explanation for this could be the frequent defecating while foraging (Brodschneider et al., 2022), thus removing orally ingested PFAS accumulated in the gut. Foragers also do not consume large amounts of pollen (Crailsheim et al., 1992). Hive bees, on the other hand, do rarely leave the hive for defecation, but consume lots of pollen, which could explain higher concentrations of PFAS in hive bees.
The difference between forager and hive bees was not consistent across all sample locations and type, possibly due to the different exposure routes. On one hand forager bees are exposed to PFAS mostly by atmospheric deposition or from their hive environment before they were foragers. On the other hand, hive bees are exposed to PFAS mostly through feeding on pollen, as a colony consumes kg amounts of pollen through the season (Brodschneider & Crailsheim, 2010). However, a key takeaway message is that the forager bees introduce PFAS in the hive, either on their bodies, or through contaminated resources such as pollen, where PFAS have the possibility to be distributed and/or stored in the hive.

3.1.3. Difference between locations

Fig. 2, Fig. 3 also illustrate differences in PFAS exposure across different sampling locations, demonstrating how land use could influence PFAS contamination of bees and pollen. The observed differences suggest localised pollution sources for example the use of fluorinated pesticide as well as different PFAS accumulation/deposition dynamics over time. Around 70 % of the pesticides used for plant protection are fluorinated (Alexandrino et al., 2022). While many of these are regulated under the European Union (Regulation (EC) No 1107/2009), legacy use and degradation products remain a concern. A notable example is a widely used fluorinated pesticide (Flufenacet) used for crop protection since 2004, which is now banned effectively from June 2025 degrading to TFA. However, other products can also degrade to TFA (Alexandrino et al., 2022; Joerss et al., 2024), which is now widely present in the water cycle globally with concentrations up to 974 ng L− 1 (Freeling et al., 2020; Freeling and Björnsdotter, 2023; Scheurer et al., 2017; Wang et al., 2014), and has been shown to accumulate in plants (Benesch and Gustin, 2002).
With the presence of PFAS in the surface waters, watering systems may influence PFAS profiles at each site, especially agricultural areas where pesticide application may occur before and after flowering, potentially leading to degradation into TFA or other short-chain PFAS.
Not only the watering system can influence the differences between the location, but plant uptake as well. PFAS bioavailability depends on soil properties, such as pH, clay content or soil organic matter (Adu et al., 2023; Bao et al., 2020; Bolan et al., 2021; Bräunig et al., 2019; Shen et al., 2024). For example, longer chain PFAS tends to bind more strongly to soil, reducing their uptake by plants. Indeed, short chain PFAS were predominantly found across all samples, with occasionally a few long chain PFAS present as well. Recently Wright et al. (2025) reported 114 μg kg−1 d.w. in control hemp plant (Cannabis sativa L.) and 43 μg kg−1 d.w. in control hemp pollen. The group also estimated that a bee can consume up to 3.1 μg/kg body weight per day PFOS from this plant alone (Wright et al., 2025). The PFAS species observed in pollen included PFBA, PFPeA, PFHxA, PFOA, PFBS, PFPS and PFOS. It is important to note, that among all PFAS reported in our study and by Wright et al. (2025), only PFOA, PFHxA, and PFBS were found in both datasets.

3.1.4. Difference between months

Fig. 2, Fig. 3 also highlight the temporal variations in PFAS exposure, showing peak contamination in May–June, and declining towards August. This peak in PFAS levels is primarily attributed to elevated levels of FBSA in the Suburban gardens – G location during April and June, which subsequently decreased in the later months. In addition to FBSA, the presence of 4:2 FTS also contributes significantly to the observed temporal differences in PFAS concentrations across the months. These fluctuations suggest that specific compounds, such as FBSA and 4:2 FTS, may exhibit seasonal patterns of accumulation or exposure, influencing overall contamination levels during the study period. Different plants have varying capacities to uptake PFAS from soils. For example, positive correlation has been shown between protein and lipid contents and PFAS uptake through the root (Shen et al., 2024; Wen et al., 2016). Negative correlation between lipid content and PFAS uptake has also been shown (Wen et al., 2016). Uptake is further influenced by the blooming period, flower species, wildflower variety and rainfall patterns. Short chain PFAS are more water soluble, enabling them to be transported in water transportation systems in the xylem (Adu et al., 2023) to the flowers, from where forager bees can collect them. Therefore, both atmospheric deposition and plant uptake influence the PFAS profile in collected pollen. In contrast, forager bees reflect primarily the environmental deposition.

3.1.5. Principal component analysis (PCA)

Fig. 4 represents the PCA analysis, showing that most of the sample cluster together, indicating similarity in PFAS profile across the sample types. A few forager bees and pollen were outside of the main cluster, due to the presence of different PFAS in them, such as PFOA or N-EtFOSAA. PFPS and PFBS concentrations align well with TF concentrations, showing a positive influence, which also aligns with the direction of clustering. 4:2 FTS and location vectors are closely aligning, suggesting that 4:2 FTS levels are influenced by the locations, possibly due to a localised contamination source, although the effect of temporal variation is not that pronounced.
Fig. 4
  1. Download: Download high-res image (206KB)
  2. Download: Download full-size image
Fig. 4

3.2. EOF and TF

Matrix matched recovery of EOF spikes using PFOA varied between 1 and 64 % across the sample types, which highlights the importance of using appropriate matching internal standards for analysis. Instrument recovery using only standard, however, ranged between 93 and 114 %, suggesting reliable quantification despite the poor analyte recoveries. The low analyte recoveries may be attributed to the high variability in matrix composition across the sample types. For example, samples with higher organic content would ideally require adjustments in the combustion method to ensure efficient breaking of the C-F bonds. However, a single combustion method was applied uniformly across all samples, as it is not feasible to optimise the combustion method parameters for individual samples each having different matrix. This lack of adaptability likely contributed to the reduced reproducibility and may explain the high number of samples with concentrations < LOD.
Albeit no samples showed detectable amount of EOF content in the extracts above the method LOD (800 ng F g ꟷ1), TF was observed in the majority of samples above the method LOD (500 ng F g ꟷ1). Spiked bee and pollen showed good recoveries (104 % in both cases), and most quality controls recovery exceeded the 70 % apart from two 500 μg L−1 standards (65 % and 60 %). Sample heterogeneity was high with relative standard deviations (RSD) between the replicates was between 14 % and 109 %. Average TF of pollen, hive and forager bees 1140 ng F g ꟷ1 d.w., 1780 ng F g ꟷ1 d.w., and 3000 ng F g ꟷ1 d.w. respectively (Fig. 5, Table S10). Maximum TF concentrations for pollen, hive and forager bees were 3480 ng F g ꟷ1 d.w., 4710 ng F g ꟷ1 d.w. and 7600 ng F g ꟷ1 d.w.respectively. Comparing the ∑Target PFAS to the TF, it is clear that very little fluorine content can be identified with targeted analysis (Fig. 6), which is common for environmental samples (Schröder et al., 2024; Müller et al., 2023; Robel et al., 2017). ∑Target PFAS accounted for 0.06 %, 0.18 % and 0.04 % in pollen, hive and forager bees respectively. Forager bees and pollen have lower identified PFAS content most likely because they are exposed to precursor PFAS (those PFAS that can degrade to PFCA and PFSA). It is important to note that using targeted analysis, we exclude PFAS that can be in the sample, but not on the target list, or failed the quality control. In the case of this study, although ultrashort-chain PFAS were analysed, the results could not be used due to contamination issues. Even using non-targeted analysis, only those PFAS can be analysed that are ionisable with the given ionisation method (e.g. electrospray). This influences how much targeted and non-targeted analyses can account for the EOF.
Fig. 5
  1. Download: Download high-res image (283KB)
  2. Download: Download full-size image
Fig. 5
Fig. 6
  1. Download: Download high-res image (359KB)
  2. Download: Download full-size image
Fig. 6
No significant difference between forager and hive bees was found at five out of six sites, implying that forager bees may contribute to hive contamination. The exception was Agricultural – V site, where TF content of forager bees was significantly higher than that of hive bees (p = 0.04, t-test) or pollen (p = 0.03 t-test). Significant differences were found as well between pollen and forager bees (p = 0.01 t-test) from Suburban Meadows – G, with forager bees having higher TF. This indicates that in these areas, atmospheric deposition plays a bigger role in PFAS contamination than pollen collection.
Comparing the 5 locations together, location Rural – O found to be significantly different (p < 0.05) from the other locations, apart from location Suburban Gardens – G, showing generally lower TF concentrations than the other areas. This indicates that Rural – O area has a different exposure source to total fluorine. The other areas did not show significant differences based on their TF results.
Although TF analysis showed high levels of fluorine in the bees and pollen, TF analysis includes both inorganic fluoride and organofluorines without a clear division between them. Although EOF analysis failed to detect the presence of organofluorines in the extracts, possibly due to the low levels of organofluorines in the samples combined with the high detection limits of the CIC, the presence of other fluorinated compounds (such as fluorinated pesticides, particles and PFAS) in the samples cannot be excluded.

4. Environmental implications

Compared to other sentinel animals such as horses or wild boars (Schröder et al., 2024), that can live years, bees offer advantages associated with their short life span. Bees live only a few weeks, which allows for the monitoring of acute, short term and seasonal changes in contaminants. Moreover, bees do not accumulate persistent pollutants (such as PFAS) to the same extent as longer-lived biomonitors, reflecting on more recent exposure instead. Honey bees foraging around a 2–3 km radius field, coming into contact with diverse environmental media including air, pollen, water, soil and resin. The relatively small foraging area makes biomonitoring with honey bees well-suited for pointing towards local PFAS point sources. Albeit the good potential, monitoring programs should consider multiple matrices (bees, pollen, hive products) to capture comprehensive contamination profiles. Incorporating environmental metadata, such as land use and meteorological data, will further enhance interpretation of PFAS exposure dynamics.

5. Conclusions

This study demonstrates that PFAS can be detected at quantifiable levels in bee matrices; however, interpreting the data presents significant challenges. Location plays a critical role, as PFAS profiles vary based on proximity to industrial sources and the presence of key fluorinated species and transformation products. Direct comparisons between sites remain complex due to these spatial variations. Forager bees likely introduce PFAS into hives, contributing to contamination of hive bees and potentially hive products. Seasonal variation, including bloom timing and rainfall, influences PFAS uptake into pollen and exposure pathways for bees. These findings highlight the need for careful consideration when employing bees as bio monitors for PFAS. Despite variability, bees serve as sensitive indicators of PFAS presence. However, to improve comparability and reliability, future research must adopt harmonized protocols and incorporate controlled experiments to better understand PFAS uptake, retention, and depuration dynamics in bees. Future research should also explore whether PFAS are absorbed through the gut of the bees or remain in the gut, therefore being removable with faeces.

CRediT authorship contribution statement

Viktoria Müller: Writing – review & editing, Writing – original draft, Visualization, Supervision, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Jörg Feldmann: Writing – review & editing, Supervision, Funding acquisition, Data curation, Conceptualization. Eileen Prieler: Methodology, Investigation, Formal analysis. Robert Brodschneider: Writing – review & editing, Supervision, Resources, Project administration, Investigation, Data curation.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: All authors reports financial support was provided by University of Graz. Viktoria Mueller and Joerg Feldmann report financial support was provided by PARC - CORDIS. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors acknowledge the financial support by the University of Graz. We also thank Roland, Stefan and Karl for their help in sampling their own colonies. The authors thank PARC for financial support.
The authors would like to acknowledge the use of a language model chatbot (ChatGPT based on the GPT-4 architecture, accessed via OpenAI's ChatGPT platform (chat.openai.com)) in order to refine the language of the manuscript. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

댓글