In Vitro Evaluation of the Carcinogenic Potential of Perfluorinated Chemicals

Perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) are the major components of long-chain per- and polyfluorinated alkyl substances (PFAS), known for their chemical stability and environmental persistence. Even if PFOA and PFOS have been phased out or are limited in use, they still represent a concern for human and environmental health. Several studies have been performed to highlight the toxicological behavior of these chemicals and their mode of action (MoA). Data suggested the causal association between PFOA or PFOS exposure and carcinogenicity in humans, but the outcomes of epidemiological studies showed some inconsistency. Moreover, the hypothesized MoA based on animal studies is considered not relevant for human cancer. In order to improve the knowledge on PFAS toxicology and contribute to the weight of evidence for the regulatory classification of PFAS, we used the BALB/c 3T3 cell transformation assay (CTA), an in vitro model under consideration to be included in an integrated approach to testing and assessment for non-genotoxic carcinogens (NGTxCs). PFOS and PFOA were tested at several concentrations by using a validated experimental protocol. Our results demonstrated that PFOA is not able to induce cell transformation, whereas PFOS exposure led to a concentration-related increase of type-III foci. Malignant foci formation is triggered at PFOS concentrations equal to or higher than 50 ppm. It is not directly associated with cytotoxicity or proliferation induction. The divergent CTA outcomes suggest that different molecular events could be responsible for the toxicological profiles of PFOS and PFOA, which were not completely captured in our study.


Introduction
Perfluorinated chemicals, such as perfluoroalkyl and polyfluoroalkyl substances (PFAS), represent a complex group of manufactured and widely used compounds, which pose concern for the environment and human health.Current information on their toxicological activity and environmental behavior is still insufficient to support quantitative risk assessment, especially at environmental relevant exposures.PFASs are characterized by the highly stable carbon-fluorine bond that confers them unique physicochemical properties, such as dielectrical properties, resistance to heat and chemical agents, low surface energy and low friction properties.The number of chemicals differing for chain length, chemical structures and containing at least one perfluoroalkylic moiety, therefore belonging to the PFAS family, has been reported to be over 8,000, among which about 600 are currently in use (EPA, 2019).Their persistence in the environment has been confirmed by the inclusion of PFOS (perfluorooctanesulfonic acid) (Stockholm Convention, 2009) and PFOA (perfluorooctanoic acid) (Stockholm Convention, 2019)  Materials and methods

Cells
The original stock of mouse embryo BALB/c 3T3 fibroblasts, clone A31-1-1, was obtained from the Health Science Research Resource Bank (Osaka, Japan), subcultured and stored in liquid nitrogen.After thawing, cells were seeded at 5-10 x 10 4 cells/T75 flask in M10F, prepared by supplementing Minimum Essential Medium (MEM) with 10% Fetal Bovine Serum (FBS, Gibco BRL) and 1% 10,000 U/mL penicillin -10 mg/mL streptomycin solution.The cells were grown in M10F and used within two passages from thawing.The seeding density was established in order to ensure that cells did not overcome 70% confluence.At thawing, aliquots of the cell suspension (2.5 x 10 4 and 5 x 10 4 cells) were seeded in 60 mm diameter dishes and maintained in culture without antibiotic supplementation for at least 48 hours.Then they were checked for mycoplasma contamination by Hoechst 33258 DNA staining.
Starting from the concentrated solutions, serial dilutions of the test items in DMSO were prepared.A rapid vortexing phase (1-2') was needed to obtain the complete solubilization of the test items.The working solutions of the test items were obtained by 1:1,000 dilution of the DMSO stock solutions in M10F immediately before use.At the end of the procedure, the solutions were clear and no turbidity or clumps or precipitates could be seen by the naked eye.The test item working solutions were administered to the test system without prior sterilization by filtration.
The CTA reference item 3-methylcholanthrene (MCA, CAS number 56-49-5, purity 99%, Ultra Scientific Italia) was diluted at 4 mg/mL in DMSO and then at the final concentration (MCA working solution, 4 µg/mL) in M10F.DMSO, which was used as the solvent vehicle for all the chemicals, was administered to cell cultures at a final concentration of 0.1%.

In vitro cytotoxicity assay
Cells in the logarithmic phase of growth were seeded (200 cells/60 mm diameter-dish, 5 dishes for each treatment) in M10F culture media and incubated under cell culture standard conditions.After 24 ± 2 hours from cell seeding, the culture medium was replaced with suitable volumes of working solutions (4 mL/60 mm diameter dish).The treatment lasted 72 ± 2 hours.Untreated cells (UC) and solventtreated cells were used as negative controls.
At the end of the 72-hours period of incubation, the cells were maintained in M10F for 8-9 days from seeding.Then plates were fixed with methanol (CAS 67-56-1, ≥99.8% purity, Sigma Aldrich) and stained with 0.04% Giemsa (CAS 51811-82-6, Sigma Aldrich).Only colonies containing more than 50 cells were counted by using the optical microscope.
The results were expressed as: i) mean number of colonies/plate ± standard error; ii) absolute clonal efficiency (ACE), i.e. the ratio between the number of colonies/plate and the number of seeded cells; iii) relative clonal efficiency (RCE), which estimates the survival rate compared to the vehicle-treated cells.
The differences between the mean number of colonies in the treated group compared to the control group were evaluated by the Student t-test.The z test for comparison of two proportions was applied to confirm the significant differences in the ACE of cells exposed to the chemical treatments compared to the control group.
The following acceptance criteria should be fulfilled: i) at least (n-1) evaluable plates for each treatment group are present (n=total number of plates for treatment group); ii) at least 3 evaluable test item treatment groups are present; iii) the clonal efficiency in the UC treatment group is >0.3; iv) the vehicle does not induce any significant reduction of the ACE in the UC treatment group; v) the reference item 4 µg/mL MCA induces a significant reduction in the mean number of colonies/plate compared to the control-vehicle.

2.4
In vitro cell transformation assay Cells in the logarithmic phase of growth were seeded at 1x10 4 cells/60 mm diameter-dish density, 10 dishes for each treatment, in M10F culture media.After 24 ± 2 hours, cells were exposed to the test item for 72 ±2 hours.Negative controls were set up as for the cytotoxicity assay, whereas cells treated with MCA represented the positive controls.
Starting from day 8 from seeding, the culture medium was replaced twice a week with DF2I2F, containing FBS at low concentration (2%) and insulin (final concentration 2 µg/mL).After 24 days from seeding, plates were maintained without any further media change.At 31-32 day from seeding, plates were fixed with methanol and stained with 0.04% Giemsa.Plates were scored by using the optical microscope, to assess the occurrence of transformed Type III foci, which were characterized by deep basophilic staining, random cell orientation, dense multilayering of cells and invasion into the surrounding contact-inhibited monolayer (Sasaki et al., 2012a).
The concurrent cytotoxicity assay was performed as previously described.Results were reported as: i) mean number foci/plate ± standard error; ii) number of positive plates (plates with foci)/total plates; iii) transformation frequency (TF), i.e. the number of transformed cells (foci) divided by the number of cells surviving the chemical treatment.
The significant percentage of positive plates with respect to total number of plates was calculated according to the onetailed exact Fisher test of significance in 2 x 2 contingency tables.The statistical analysis of foci distribution was performed by the onetailed Mann-Whitney test (Mann-Whitney unpaired t-test).The one-tailed Poisson test was used to analyze the differences between the transformation frequencies calculated in the treatment groups and the solvent group.
The following acceptance criteria should be fulfilled to assess the reliability of CTA data: i) at least (n-1) evaluable plates for each treatment group are present (n=total number of plates for treatment group); ii) at least 3 evaluable test item treatment groups are present; iii) the monolayer appears regular; iv) no significant increase in the cell transformation was observed in the negative control (DMSO 0.1%) compared to the UC group; v) the mean foci number/plate in the solvent vehicle group is < 0.5; vi) the reference item induce a significant increase in the cellular transformation with respect to the vehicle-treated group; vii) the total number of transformed foci detected in the reference item treatment group is at least 3 fold higher than that observed in the vehicle-treated group.

2.5
Assessment criteria The test item was considered positive (able to induce cell transformation) when: i) a concentration-dependent relationship between increasing concentrations of the chemical and the increase in the mean number of foci/plate and/or the transformation frequency is established; ii) the increase in the mean number of foci/plate and TF is statistically significant at the 99% confidence level, when observed only at a single dose of treatment, or at 95% confidence level, when two or more concentrations induced positive effects; iii) the cytotoxicity in the treatment groups where the significant increase in transformation was detected was less than 90%, i.e. the cells surviving the chemical treatment are more than 10% of the treated population.The test item was considered negative (not able to induce cell transformation) when: i) no statistically significant increase in the mean number of foci/plate compared to the control-vehicle was observed at any assayed concentrations; ii) a) the increase in the mean number of foci/plate is statistically significant compared to the control-vehicle at the 95% confidence level for only one of the intermediate tested concentrations and b) a concentration-dependent increase in the mean number of foci/plate and/or the transformation frequency is not established.The assay should be considered inconclusive and should be repeated when: i) the concentration-dependence is observed but the increase of the mean number of foci/plate for each treatment group is statistically significant compared to the control-vehicle at the 95% confidence level for only one concentration ii) the increase of the mean number of foci/plate for each treatment group is statistically significant compared to the controlvehicle at the 95% confidence level for two non-consecutive concentrations

2.6
Test and reference items determination in stock and working solutions The concentrations of the test and reference items were evaluated before and after dilutions in the cell culture medium.The working solutions maintained under test conditions (37°C, 5% CO2, 90% relative humidity, 72 hours) were also analyzed to assess the stability of the chemicals during the test system exposure.Aliquots of the solutions were sampled immediately after the preparation for the test system treatment or at the end of the exposure time and then stored at -20°C until they were analyzed.The resulting measured concentrations were used for cytotoxicity and CTA data processing.

2.6.1
Test item The analytical procedure for the determination of PFOS in the vehicle or in the cell culture media is derived from the ISO 25101:2009 method 2009 "Water quality -Determination of perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) -Method for unfiltered samples using solid phase extraction and liquid chromatography/mass spectrometry" with appropriate adjustments.The experimental procedure entailed the dilution of the culture media samples with methanol followed by LC/MS/MS analysis (1200 series Agilent Technologies HPLC connected to the 6410 Agilent Technologies triple quadrupole mass spectrometer).The solid-phase extraction/concentration step recommended in the ISO method was not carried out, due to the high PFOS levels in the samples.The qualitative determinations were performed by using the retention time and the % abundance of the target and qualifier ions, compared to the standard.The quantitative analysis was performed by using a 6-points fitting calibration curve (1-5-10-25-50-100 ng/mL) constructed by using PFOS standard solutions in methanol.

2.6.2
Reference item The concentrations of the reference item MCA were determined by using the EPA method 8270 D 2014 Rev5 with minor modifications.The samples were extracted using the HS-SPME procedure and analyzed with SIM mode using GC/MS.The chemical identification was performed by using the retention time and the ratio between the isotopic amounts of the target ion (to be quantified) and the qualifier ion.The chemical quantification was performed by a calibration curve established using deuterated PAHs as the internal standard (perylene d 12 or chrysene d 12 ).

Preliminary cytotoxicity assays
Preliminary studies were performed in order to highlight the toxicity of the selected chemicals and choose the most appropriate concentrations for further experiments.
PFOS and PFOA and their conjugate bases were tested in three different experiments, at a broad range of chemical concentrations.Initially, the effects of concentrations ranging from 20 to 200 µg/mL was explored.Then the concentration range was extended to include low and very low concentrations (0.001-100 µg/mL).Cells were treated with the test item after 24 hours from seeding and the exposure lasted for 72 hours.Results from the preliminary studies were summarized in Fig. 1 (see Tab. S1-S41 for detailed results).Cells in the logarithmic phase of growth were seeded (200 cells/60 mm diameter-dish, 5 dishes for each treatment) in M10F culture media.After 24 ± 2 hours, cells were treated with the test item for 72 hours.Untreated cells (UC) and solvent-treated cells were used as negative controls.At day 8-9 from seeding, plates were fixed with methanol and stained with 0.04% Giemsa.Only colonies containing more than 50 cells were counted by using the optical microscope.Results summarize the finding from three cytotoxicity assays, where concentrations ranging from 20 to 200 µg/mL or 0.001-100 µg/mL were tested.Results are reported as relative clonal efficiency (RCE), calculated as percentage of reduced efficiency in forming colonies with respect to solvent-treated cells.A = PFOA, acid form; B = PFOA; ammonium salt; C = PFOS, acid form; D = PFOS, potassium salt.
Both PFOA and its ammonium salt did not show toxic effects in the concentration range 0.001-20 µg/mL.Higher concentrations induced a strong reduction in the number of colonies/plate, as demonstrated by the steep slopes of the curves (Figure 1 A and B).The growth inhibition induced by PFOS and its potassium salt followed a more gradual slope, resulting in complete absence of colonies at the higher concentration (Figure 1 C and D).
The half maximal inhibitory concentration (IC50), representing the concentration of a chemical that is required for 50% inhibition of cell growth in vitro, was calculated to measure the effectiveness of the tested chemicals in inhibiting the cells' ability to form colonies.Based on the concentration-response curves obtained by all the experiments, the IC50 values were obtained by fitting data to a sigmoidal curve using a non-linear regression model (GraphPad Prism, version 8.0).The IC90 and IC10, i.e. the concentration inhibiting the 90% and the 10% of cell growth, respectively, were calculates for all the tested chemicals by interpolation from the concentration-response curves (Tab. 1, Fig. 2).According to the IC50, the cytotoxicity ranking is PFOS < PFOS K + < PFOA < PFOA NH4 + .

PFAS CTA
The effects of the acid forms of both PFOA and PFOS on the transformation frequency of BALB/c 3T3 A31-1-1 cells were assessed, according to the protocol recommended by ECVAM (Sasaki et al., 2012b).The concentrations of the chemicals have been established according to the results from the preliminary cytotoxicity assays and the current available scientific publications (Gimenez-Bastida et al., 2015;Jacquet et al., 2011;Jacquet et al., 2012).The need to support the Regulatory Authorities in the definition of concentrations thresholds to be used for the toxicological profile of the test item was also taken into account.The carcinogen MCA was used as the positive control.Untreated cells and solvent-treated cells, i.e. cells treated with dimethylsulfoxide (DMSO), which had been used to vehicle the tested chemical and dissolve it in the culture medium, served as negative controls.
The combined execution of the in vitro CTA and the cytotoxicity assay allowed the assessment of the cytotoxic and carcinogenic activity of the test item during the same study, facilitating the calculation of the transformation frequency in cells exposed to the chemical as a function of the carcinogenic properties.To minimize the bias introduced in CTA results by strong cytotoxicity induction, the selected ranges included concentrations ≤ IC50.

PFOA
The effects induced by PFOA on cytotoxicity (Fig 3 A) and cell transformation (Tab. 2 and Fig. 3C) were evaluated in the concentration range 5-75 µg/mL, which was selected taking into account the calculated IC50 and the slope of the concentration-effect curve.
PFOA concentrations ranging from 10 to 75 µg/mL induced a statistically significant reduction (p<0.01)both in the absolute clonal efficiency and in the mean number of colonies/plate compared to the vehicle control.A minor but significant (p<0.05)cytotoxic effect was induced by PFOA treatment at the concentration 5 µg/mL (Tab.S6 1 ).The induction of cytotoxicity was not associated with an increase in cell transformation in PFOA treated cells.Even if the mean foci number/plate in the solvent vehicle group is higher than the established threshold of 0.5, significant differences both in the mean number of foci/plate or in the transformation frequency compared to the control vehicle were not detected in any of the PFOA assayed concentrations.

PFOS
The PFOS CTA study was performed according to the Principles of Good Laboratory Practice (GLP), to allow the mutual acceptance of data within the international regulatory community.The highest concentration tested was about the IC50 calculated in the preliminary cytotoxicity assays.The remaining four concentrations were selected in the range 10-100 µg/mL.The analytical results confirmed the solubility of PFOS in DMSO and its stability in cell culture medium (M10F) during 3T3 cell exposure.The results from the PFOS GLP study were summarized in Tab. 3, Fig. 3B and 3D.The CTA fulfilled all the acceptance criteria.The positive control MCA (4 µg/mL) induced a statistically significant increase in the number of transformed type III foci, which were almost absent in untreated and solvent-treated cells.PFOS treatment induced transformation in BALB/c 3T3 cells, as shown by the concentration-related increase in the mean number of foci/plate, starting from the 50 µg/mL nominal concentration.The increase was significant compared to the solvent control at the 95% confidence level for the intermediate concentration (50 µg/mL) and at the 99% confidence level for the highest tested concentrations (75 and 100 µg/mL).The transformation frequency exhibited a concentrationrelated trend as well, showing a significant increase compared to the solvent control at the 99% confidence level for the three highest tested concentrations.In the concurrent cytotoxicity assay, the exposure to the lowest concentrations of PFOS (10 and 25 µg/mL) did not affect cell growth and survival, whereas higher concentrations reduced both the absolute clonal efficiency and the mean number of colonies/plate (Tab.S5 1 ).Even if the cytotoxic effects were concentration-dependent, the decrease of the RCE did not exceed the 20%, except for 100 µg/mL PFOS (Fig. 3B).According to the assessment criteria, PFOS was classified as positive in the cell transformation assay.

PFOS and MCA determination in stock and working solutions
The actual concentrations of PFOS were measured both in the stock and working solutions (Fig. 4).There was a high degree of concordance between nominal and measured concentrations of PFOS.The maximal deviations from the expected values were observed at low concentrations and were about 30%.At the end of the treatment (72 hours), the measured concentrations did not show differences higher than 20% compared to the concentrations measured at the starting of the administration.
The concentrations of PFOS in M10F were lower than the detection limit, except for the cell culture medium incubated for 72 hours, where the measured concentration was 60 ng/mL (Tab.S7 1 ).PFOS was also quantifiable in the vehicle DMSO, used for preparing the PFOS stock solutions.Its concentration (20 ng/mL) was about 6 order of magnitude lower compared to the stock solution at the lower concentration (10 mg/mL) and was undetectable after the 1:1,000 dilution requested for preparing the working solution (Tab.S7 1 ).The impact of the PFOS presence in cell culture media or solvent-vehicle was not relevant for the purpose of the present study.
Regarding the reference item MCA, the nominal and measured concentrations matched, and the amount of the chemical did not decrease after 72 hours incubation.

Discussion
The widespread presence of PFAS in the environment and wildlife has raised concern for the possible adverse impact of these substances on environment and human health.
The ingestion of contaminated food and drinking water has been identified as the main source of PFAS exposure for the general population, leading to the consistent detection of PFAS in human serum samples (Ingelido et al., 2010;Kato et al., 2018;Kaiser et al., 2021).Serum PFAS levels, usually below 10 ng/mL, rise up to 1000 times in people living near production sites or exposed in occupational settings (IARC, 2017).However, each type of PFAS exhibits unique chemical attributes and distinct behavioral patterns in the blood serum.For PFOS, average serum concentrations have been reported to range from approximately 1 to 10 ng/mL in the general population.For PFOA, average serum concentrations have been reported to range from approximately 2 to 20 ng/mL in the general population.
Moreover, demographic factors such as race/ethnicity, age, and sex influence serum PFAS levels over time.A recent study analyzed trends in serum PFAS compounds in US general population, showing serum PFAS levels declined continuously in the studied population from 1999 to 2018 (Sonnenberg et al., 2023).In this study, PFOS tends to have higher concentrations compared to other PFAS compounds, with averages ranging from 4.3 to 30.4 µg/L.PFOA levels tend to be higher in populations with known exposure to sources of PFOA contamination (Frisbee et al., 2009;Frisbee et al., 2010).In a study conducted in the exposed population in Veneto, Italy, PFOA was found at the highest serum concentration (median 44.4 ng/mL), with respect to PFOS (median 3.9 ng/mL) (Pitter et al., 2020).
Systemic effects, such as increased levels of cholesterol and liver enzymes, reduced immune response following vaccination, thyroid disorders, pregnancy complications and increased risk of testis and kidney cancer, have been reported in PFOA-exposed population (Steenland et al., 2010).
To date, the causal association between PFOA or PFOS exposure and carcinogenicity in humans is not entirely supported by the results from the available epidemiological studies, due to study design limitations, heterogeneity and inconsistency in outcomes (Chang et al., 2014;Steenland and Winquist, 2021).
A summary of epidemiological, rodent and in vitro studies examining the association between PFOA and PFOS and cancer is reported in Table 4. Available studies from occupational or environmental exposure of general population highlighted suggestive but not consistent associations between PFOS or PFOA exposure, usually measured by serum levels, and increased risk of cancer (Steenland and Winquist, 2021).Data suggested an association of PFOA with testicular or kidney cancer (Steenland et al., 2020).Notably, a recent report indicated an elevated incidence of liver cancer mortality among a group of Italian workers who were exposed to PFOA.However, the utilization of mortality data may restrict the assessment of non-fatal outcomes that are likely associated with PFAS exposures  ( Girardi and Merler, 2019).Recent reports have also shown positive associations between kidney cancer and PFOA (Shearer et al., 2021), and between ER+ breast cancer and PFOS serum concentration in general population (Mancini et al., 2020;Tsai et al., 2020).However, research on PFAS has produced mixed results regarding their carcinogenic effects, with limited evidence demonstrated in animal studies (Table 4).These findings are often considered to have limited applicability to humans, because of differences in kinetic mechanisms, particularly the significant slower urinary excretion observed in humans than in rats, resulting in PFOA elimination rates 5000-9000 times slower in humans than in rats (ATSDR, 2021).The results from a recent study performed by the National Toxicology Program (NTP, 2023) showed an increase in the incidence of hepatocellular and pancreatic acinar cellular neoplasms in rats following gestational and chronic exposure, reopening the discussion on the carcinogenic effects by PFAS.These results confirmed the findings of previous studies that reported testicular Leydig cell adenomas in addition to liver and pancreas tumors (Caverly Rae et al., 2014;Butenhoff et al., 2012b).
The International Agency for Research on Cancer (IARC) initially classified PFOA in Group 2 B (possibly carcinogenic to humans) in their 2017 evaluation (IARC, 2017).However, subsequent re-evaluation by IARC has led to PFOA being reclassified as a confirmed carcinogen to humans (Zahm et al., 2024) and the US EPA concluded that there is suggestive evidence of the carcinogenic potential of both PFOA and PFOS in humans (EPA, 2016 a,b).Initially, PFOS was not classified as a carcinogen according to the criteria set by IARC for chemical classifications, because of the low statistical power of several studies, inadequate experimental designs and presence of bias and confounders (Arrieta-Cortes et al., 2017).However, recent evaluations have prompted its classification as possibly carcinogenic to humans (Group 2B) due to accumulating mechanistic evidence, including observations of epigenetic alterations and immunosuppression in exposed individuals, consistent with characteristics of carcinogens (Zahm et al., 2024).Our findings unequivocally demonstrate that PFOS but not PFOA is capable of stimulating cell transformation in 3T3 CTA model, likely through non-genotoxic mechanisms.The transforming potential of PFOS exhibited a dose-dependent relationship, with the emergence of type III foci detected at concentrations of 50 µg/mL (50 ppm) and above, extending to the highest assessable concentration of 100 µg/mL.The impact of PFOS on cellular transformation was independent of any cytotoxic effects or changes in cell proliferation.Our results align with those of Jacquet et al (2012), who reported the induction of cell transformation by PFOS at non-cytotoxic concentrations (0.2 and 2 µg/mL) in SHE CTA (Jacquet et al., 2012).In the same experimental model, PFOA exhibited positive results only when administered to SHE cells after exposure to benzo(a)pyrene, suggesting that it acts as a promoter of cells already initiated in the multistep carcinogenesis process (Jacquet et al., 2011).It is unlikely that PFOA is carcinogenic due to its direct interaction with DNA, as evidence from mutagenicity and genotoxicity tests does not support this mechanism of action.PFAS have been observed to have a lack of functional groups that could potentially lead to genotoxicity and are not metabolically transformed, resulting in the formation of reactive intermediates (IARC, 2017;EFSA, 2018).PFAS have also been reported to affect oxidative stress markers in both in vivo and in vitro studies (Temkin et al., 2020) and to induce DNA damage, such as strand breaks, and other genotoxic effects, secondary to oxidative stress (EFSA, 2018).However, these effects are usually observed only at high doses, and genotoxic effects are not discernible in mouse somatic and germ cells upon repeated exposure to PFOA, as determined by a sensitive mouse model that is responsive to PFOA-induced oxidative stress (Crebelli et al., 2019).Therefore, available experimental data suggest that severe toxicological effects, such as cell oncotransformation and genetic damages, are potentially initiated by non-genotoxic events at doses or concentrations exceeding the range of human exposure.
Establishing an effective dose or concentration for PFAS through experimental studies is a challenging endeavor because of the intricate nature of these molecules.Repeated exposure to PFAS results in their accumulation in tissues and organs, as they possess the ability to bind proteins, including plasma proteins.Consequently, exposure doses or concentrations may not accurately reflect the effective internal dose in human and animal studies, particularly after acute exposure.Moreover, the pronounced hydrophobic or lipophobic nature of PFAS could potentially influence the results of in vitro investigations using cell cultures, owing to their ability to interact with bovine serum albumin (BSA) (Beesoon and Martin, 2015;Butenhoff et al., 2012c;Wang et al., 2016).The use of calf serum (FCS) in cell cultures has been a longstanding practice in biomedical research and biotechnology duo to its rich composition of growth factors, hormones, proteins and other essential nutrients that support cell growth and proliferation.This procedure also applies to BALB/c 3T3 cells that undergo apoptosis in the absence of serum.However, the sourcing of FCS from animal-derived material raises ethical concerns and poses challenges related to variability, batch-to-batch consistency, and the potential for contamination with adventitious agents (Verma et al., 2020;Lang et al., 2018;Colacci et al., 2023).
Efforts to develop non-animal-derived alternatives for FCS are ongoing and motivated by concerns regarding animal welfare and the need for more standardized and defined culture conditions.Although some progress has been made in the development of serum-free or defined media formulations that utilize recombinant growth factors and synthetic substitutes, the complete replacement of FCS remains a challenging task (Lang et al., 2018;Subbiahanadar Chelladurai et al., 2021;Jacobs et al., 2023;Colacci et al., 2023).Despite advancements in cell culture technology, non-animal-derived alternatives often fail to fully replicate the growth-promoting properties of FCS, resulting in suboptimal cell growth, reduced cell viability, and altered cellular behavior (Colacci et al., 2023).
One of the primary challenges in identifying viable alternatives to FCS is the difficulty in reproducing its intricate and multifaceted composition, which is crucial for fostering a supportive microenvironment for diverse cell types.However, this complexity also introduces limitations when incorporating FCS into cell cultures.One limitation of using FCS in cell culture is the presence of BSA, which can bind various chemicals and compounds present in the culture medium.This binding capacity may interfere with experimental outcomes by sequestering or modifying the availability of these substances, potentially affecting cell behavior, signaling pathways, and experimental reproducibility (Subbiahanadar Chelladurai et al., 2021).
Both PFOA and PFOS have been shown to interact with BSA.It has been reported that the PFOS fraction, which is not bound to plasma proteins, is associated with cells through cell membrane binding (Sanchez Garcia et al., 2018), which raises the question of the possible PFAS sequestration by serum proteins during cell exposure.Moreover, the higher hydrophobicity of PFOS can result in higher affinity for serum proteins, leading to the hypothesis that PFOS bioavailability in living cells can be lower than that of PFOA (Bischel et al., 2011).
We tested the concentration of PFOS and MCA working solutions before and after dilution to prove the solubility and stability of both the test and reference items under the experimental conditions.The results from our study show that the bioavailability of both PFOS and PFOA in the cell culture is good enough to elicit concentration-related cytotoxicity in the cells exposed to the concentration range of 50-100 µg/mL.Furthermore, the increased toxicity of PFOA at lower concentrations than PFOS suggests that the inability of PFOA to induce cell transformation cannot be attributes solely to reduced availability during cell exposure.Finally, the slight difference in cytotoxicity between PFOS and its potassium salt indicates that the acidity of PFOS does not affect the performance of in vitro cell assays (Wang et al., 2019).
The divergent CTA outcomes suggest that different molecular events could be responsible for the toxicological profiles of PFOS and PFOA, which were not completely captured in our study.
PFAS carcinogenesis in rodents is sustained by the activation of the peroxisome proliferator-activated receptor α (PPARα), which is considered to be a key event in the putative MoA supporting rat liver tumor.Among nuclear receptors, PPARs, which include PPARα, PPARβ/δ and PPAR, exhibit species-specific variations in expression and functional roles in different organisms.This disparity between species highlights the differences in the response to peroxisome proliferator chemicals, such as PFAS or fibrate drugs, which are hepatocarcinogenic in rodents but not in humans (Lalloyer and Staels, 2010).PPARs show the highest level of speciesspecific sensitivity to PPARs agonists (Albrecht et al., 2013;Kersten and Stienstra, 2017;Corton et al., 2018;Yang et al., 2008).Moreover, their levels vary from tissue to tissue, reflecting their diverse roles in the regulation of metabolism, inflammation, and cellular differentiation across different physiological contexts.
Alternative mechanisms and MoA have been proposed to foster carcinogenic effects after the treatment with PFAS, including mitochondrial dysfunction, stress, inhibition of gap junctional intercellular communication and decreased tumor cell surveillance by the immune system (IARC, 2017).In vivo gene expression in PPARα-null (Rosen et al., 2008(Rosen et al., , 2017) ) or in Sprague Dawley rats (NTP, 2023) as well as in vitro reporter assay studies (Vanden Heuvel et al., 2006;Ren et al., 2009;Zhang et al., 2014;Ren et al., 2015;Zhang et al., 2017) have demonstrated that, besides PPARα, several nuclear receptors, including other isoforms of PPAR, such as γ and β/δ, the constitutive activated receptor (CAR), the pregnane X receptor (PXR) and the thyroid receptor, are activated by PFAS.The analysis of ToxCast results aiming at investigating associations between in vitro assays outcomes and liver disease progression showed that PPARγ is associated with preneoplastic and neoplastic levels in the liver disease progression, and PPARα is additionally associated with neoplastic lesions (Judson et al., 2010).
Our results suggest the existence of an alternative mechanism for activating PPAR, potentially explaining the transformative activity of PFOS observed in our model, but not applicable to PFOA.
The differential toxicity between PFOS and PFOA has been previously reported, particularly regarding the activation of PPARα, where PFOA exhibits notably stronger activity (Tsuda, 2016).The PFOA concentration required to induce 20% of the overall maximal response of PPARα is approximately 20 times lower than that of PFOS, in both mouse and human (Tsuda, 2016).The concentrations tested in our model were significantly higher than the minimal required concentration to activate PPARα for both compounds.The lowest effective concentration to support the full process of oncotransformation in PFOS treated cells (50 µg/mL) exceeds by one order of magnitude the benchmark dose (BMDL) reported by the European Food Safety Authority, on the basis of the effects elicited by a group of PFAS on immune response markers in children (Knutsen et al., 2018;Schrenk et al., 2020).
The formation of malignant foci is the phenotyping anchoring of molecular events leading to in vitro cell transformation (Jacobs et al., 2016;Mascolo et al., 2018;Jacobs et al., 2020).It can be considered the final adverse outcome of the in vitro transformation process.In vitro oncotransformation is marked by a committed step, an irreversible event sustained by a critical concentration, which represents the transition from cell adaptive to maladaptive response (Mascolo et al., 2018).As previously described for MCA, which is classified as a genotoxin, carcinogen concentration differentiates cell capacity to activate detoxification pathways from the ability to elude the control mechanisms (Mascolo et al., 2018).All processes are characterized by markers related to the immune-mediated response.
It is possible that lower concentrations of PFOS stimulate an adaptive response which is overcome at higher concentrations.
It is recognized that the evaluation of the whole transcriptome can provide a deep understanding of the events underlying the cellular response to chemical exposure, allowing not only the assessment of changes in the expression levels of individual genes but also the identification of biological signaling pathways deregulations that could be responsible for the shift from a physiological or adaptive feedback to pathological outcomes (Dean et al., 2017).The finding derived from human primary hepatocytes suggest that exposure to various PFAS, such as PFOA and PFOS, in vitro has an impact on genes, pathways and disease outcomes similar to those observed in in vivo studies and human observations (Rowan-Carroll et al., 2021).This further supports the utility of employing integrated omics-based approaches to provide valuable insights into NGTxC IATA.
Indeed, to support the testing strategy for carcinogenesis and fill the critical gaps in the use of CTA, as a stand-alone test, CTA can be coupled with a transcriptomics approach, using the so-called transformics assay, allowing the evaluation of whole-genome transcriptional changes at different times along the process leading to foci formation (Mascolo et al., 2018;Pillo et al., 2022;Colacci et al., 2023).The use of this integrated approach would provide information of the key events at the molecular and cellular level leading to oncotransformation, allowing a better understanding of the MoA of these chemicals, and allowing the identification of an effective threshold.
In conclusion, PFOS, but not PFOA, can induce cell transformation in vitro, depending on the tested concentration.The different toxicological behavior of PFOS and PFOA suggests that more than one mechanism is possible for chemicals using the same receptor to trigger the biological response, strongly suggesting the need to understand better the complex molecular interplay downstream of the path to the adverse outcome.Our result supports the evidence that CTA can highlight the ability of non-genotoxic chemicals to support the process of oncotransformation, and that it is possible to identify a threshold for this adverse outcome.Indeed, CTA emerges as an effective screening tool for further assessing PFAS compounds.Furthermore, the refinement of results becomes feasible through the identification of mechanistic information, enhancing the accuracy of assessments and insights into potential health implications.
in the list of Persistent Organic Pollutants (POPs) regulated by the Stockholm Convention and has led to the phase-out of production of long-chain PFASs, at least in developed countries.The EU Commission, in

Fig. 1 :
Fig. 1: Cytotoxic activity of PFOS and PFOA and their conjugate bases in BALB/c 3T3 cellsCells in the logarithmic phase of growth were seeded (200 cells/60 mm diameter-dish, 5 dishes for each treatment) in M10F culture media.After 24 ± 2 hours, cells were treated with the test item for 72 hours.Untreated cells (UC) and solvent-treated cells were used as negative controls.At day 8-9 from seeding, plates were fixed with methanol and stained with 0.04% Giemsa.Only colonies containing more than 50 cells were counted by using the optical microscope.Results summarize the finding from three cytotoxicity assays, where concentrations ranging from 20 to 200 µg/mL or 0.001-100 µg/mL were tested.Results are reported as relative clonal efficiency (RCE), calculated as percentage of reduced efficiency in forming colonies with respect to solvent-treated cells.A = PFOA, acid form; B = PFOA; ammonium salt; C = PFOS, acid form; D = PFOS, potassium salt.

Tab. 1 :Fig. 2 :
Fig. 2: Inhibitory concentrations IC50, IC10 and IC90 and their corresponding upper and lower limitThe IC values were obtained analysing the RCE data (n=5) in a non-linear regression model (non lin fit -log(inhibitor) vs responsevariable slope (four parameters), least square fit).It was not possible to retrieve by interpolation the IC90s for both PFOS and its salt.

Fig. 4 :
Fig. 4: Concentrations of PFOS measured in treatment solutions A = stock solutions; B = working solution at time 0; C = working solutions at time 72.The uncertainty of measurements has been estimated by using the Horwitz equation.

PFOS -Cell transformation assay -Summary of results n
.d. = not determinable a p<0.05, one-tailed Fisher exact test compared to the vehicle (DMSO 0.1%) b p < 0,05, one-tailed Mann-Whitney U test compared to the vehicle (DMSO 0.1%) c p<0.01, one-tailed Poisson test compared to the vehicle (DMSO 0.1%) d p<0.01, one-tailed Fisher exact test compared to the vehicle (DMSO 0.1%) e p < 0.01, one-tailed Mann-Whitney U test compared to the vehicle (DMSO 0.1%)