Is there, or will there be, sufficient commonality of toxicological endpoints in conventional in vivo studies (with similar or different durations of exposure), or endpoints in NAMs (new assessment methodologies, such as zebrafish, cell culture studies, ‘omics assays) that will enable robust relative potency factors (RPFs) to be developed for PFAS? What caveats would need to be placeds on such an approach?

Answer Explanations

• No, sufficient commonality does not/will not support support RPFs

  Expert 4

  It’s likely that there will be some commonality. However, the information is still too limited to know whether it will be sufficient to support RPFs.

• Expert 11

  I am not able to answer this question. Not my area of expertise.

• Yes, sufficient commonality is or will support support RPFs

  Expert 2

  In the past 20 years of toxicology research on PFAS, there is no one single mode of toxicity has been identified. Nevertheless, with new assessment methodologies, some commonalities among PFAS can be achieved. I think this needs to be started with grouping/subgrouping of PFAS class and approach from there. For example, group all sulfonates and perform toxicity and develop a RPF. Similarly, carboxylates, fluorotelomers etc - each can be grouped and RPF be developed.

• Yes, sufficient commonality is or will support support RPFs

  Expert 10

  This was a tricky question because PFAS are hugely structurally diverse. For subgroups of PFAS (e.g. the perfluoroalkyl acids), yes, sufficient commonality is or will support relative potency factors (RPFs). However, this will not be the case for all PFAS (these substances are hugely diverse in structure and thus behaviour; e.g. they are gases, liquids and solids with diverse exposure pathways and toxicities).

• No, sufficient commonality does not/will not support support RPFs

  Expert 8

  No. To date, the pk, toxicity and MOA data do not support this.

• Yes, sufficient commonality is or will support support RPFs

  Expert 3

  Caveats include ensuring validation of assays, articulation of clear guidance in interpretation (of positive and negative, that the assay meets validation criteria, and so on), establishment of relevance to human toxicity pathways, providing access to data.

• Yes, sufficient commonality is or will support support RPFs

  Expert 9

  There are already some common toxicological endpoints shared among PFAS evaluated in in vivo studies. One endpoint is increased liver weight. However, some toxicologists have argued against this endpoint as they consider it as "adaptive" rather than "adverse." Effects on development also are shared in common among PFAS that have been evaluated for developmental toxicity. However, not all PFAS produce the same types of developmental toxicities, so again, insistence on extreme granularity for defining a toxicological outcome may impair the possibility of defining a common toxicological outcome. It seems like the problem is not whether or not there is commonality of toxicological endpoints but agreements among toxicologists on what constitutes a "commonality of toxicological endpoints."

• Yes, sufficient commonality is or will support support RPFs

  Expert 1

  This information is not yet available, but it could become available in the next few years. In vitro tests on PFAS contributing significantly to the organofluorine load in drinking water could allow generating information on a series of endpoints relatively quickly (e.g., ToxCast). The main caveat is that in vitro assays do not account for toxicokinetics, and high-throughput TK would need to be carried out with potentially great uncertainty (PFAS half-lives may be difficult to estimate using in vitro methods). Where both potency and TK can be estimated adequately using in vitro tests and biological modeling, estimated administered equivalent doses (AEDs) could be used to calculate RFPs.

• Yes, sufficient commonality is or will support support RPFs

  Expert 5

  In time, sufficient toxicity data may be developed to identify health endpoints expected to occur at low exposures. Endpoints arising from non-physiological, high animal doses should be treated as suspect until the likelihood of humans attaining similar or comparable (extrapolated) doses can be ascertained. At that point, NAMs can be applied to determine observable, quantifiable changes that are or can/should be linked via a MOA analysis (AOP or Toxicity Pathway) to the health endpoint. At that point, in vitro studies may be conducted to test for additivity and to develop RPF. It would be incumbent to demonstrate the relationship between the in vitro event and the health endpoint. This approach would have extra value in that if studies are properly conducted in vitro using appropriate human and test species samples, concentration-response data may be used to derive non-default extrapolation factors (data derived extrapolation factors; USEPA, 2014) to account for animal to human differences in the toxicodynamic component of the animal to human uncertainty factor. But, it is paramount that the relationship between the event observed in vitro and the health endpoint be demonstrated.

• Yes, sufficient commonality is or will support support RPFs

  Expert 6

  This is a question I posed in Round 1 and I am interested in the views of other panel members. I do believe this is the direction in which we are heading. There will be some difficulties in determining how differences in exposure duration and systemic clearance kinetics might influence the estimation of RPFs (since an RPF reflects differences in effective DOSAGE). RPFs developed from these types of NAMs will presumably better reflect differences in effective concentrations at receptor site(s) so pharmacokinetic factors will influence the conversion to exposures/doses. It is my belief that this approach will ultimately offers the best option for assessing the aggregate toxicity of mixed PFAS exposures.