This content is current only at the time of printing. This document was printed on 17 November 2018. A current copy is located at https://apvma.gov.au/node/19441
You are here
Transcript for Professor Michael Roberts
Dermal absorption related to chemical products
This presentation was delivered at the APVMA’s science feature session on 15 October 2015. The full video is available on our YouTube channel.
Good afternoon everybody. It's a pleasure to be here. I need to thank Phil and everybody for having me and inviting me to talk. I have to be honest with you, I actually completely changed my talk over the last 24 hours after talking to Phil and others who said to me we really need to try and make it much more relevant with some of the factors and other aspects. I've been digging in the literature, trying to come to terms with some of the areas that you probably deal with. I'll maybe say some of the controversial things in my talk and you may know more about some of the things I'm saying than I do. We welcome that for some of the discussion. Part of the reason we've done it is to try and encourage some lively discussion and to put some views to you.
Dermal absorption is really one of my areas. I've been working in it now for 40 years. In fact people in my group say I'm an old‑timer because they were born 20 years after I started in this space. In fact, at least we're recognised by the FDA. In fact, I lead two FDA grants in the area of dermal absorption which is a pretty big honour for Australians to be doing. It's quite special to be in that space.
I thought to start off with I would start off with a slide which summarises everything that I think we're trying to cover in this symposium to some extent. The two aspects we have is really the area of exposure science which you see here, which in fact is going to be covered a bit later, and the other area we have is ADMET. Probably what's missing in this is what some of the effects are, the responses.
So far what we've had is we've had Jim has tried to cover this pretty well. I don't have to really go through that in any great depth because Jim, you've done it for us. I've been told that Mark Gen is going to cover this area. Also what you'll notice is that Nancy has already covered nanosystems. So what am I left with? I'm going to cover this area of dermal absorption. It's going to be pretty, you'd probably think, confined and it's obviously of chemicals. That's the focus I'm going to have with the talk today.
I have to emphasise that one of my interests has been human safety. In fact Phil reminded me last night, I probably should tell the story again, Phil, because some people may not have heard it. I was called before the New South Wales Senate Committee on Nanotoxicology. They said, "Look, we've been told by everybody that you're one of Australia's most eminent nanotoxicologists. Do you think you're one of Australia's most nanotoxicologists?" I looked and then and I thought, should I? Should I? And I thought, yes, I would. I said to them, "Do I look like a nanotoxicologist to you because I certainly don't feel it." So it goes on.
One of the questions that we're interested in, in some of our work is, can we treat poisonings better and can we use our clinical data to improve regulatory toxicology? Some of the work we do we think can actually help you and we've just got to try and work with you to do that. We have an NHMRC Program Grant which is in fact one of the hardest grants to get but we have one. We've been looking as pesticide poisonings among others who have been looking at overdoses, all sorts of therapeutic drugs, but we're also looking at pesticides. What we've found in looking at the poisonings that we've come across and are presented to us, there's about a case fatality of about 10%. Ten per cent of people die. We're also now looking at this in terms of what the various side effects and adverse effects are. I'm sure you can probably see where this comes to in regulatory toxicology.
One of the things we did do is we went and looked at what the LD50s were because that's the only data we could really find easily. I have to say I don't believe in LD50s but that's the data we found and we tried to rate the case fatality. In general if you have a low LD50 you've got a greater chance of dying. What it's really saying is a key issue you have to think about with some of the pesticides in principles is in fact that those which are more potent can sometimes cause more problems. Maybe part of that reason is the fact that you've got less control of them. I don't know but that's what the data says to us.
The current focus we have is actually how do we treat our poisoned patients. You've already had Jim talk a lot about physiological PK, population PK, that's what we use in humans. A lot of the mixed model analysis, that's been our routine bread and butter for some time. Certainly TGA which I also help a lot with, most of the drug evaluations we have are all based upon Pop PK, everything we do.
The phase and the question I want to ask you is can we actually look at what the actual safety margins are based upon patients? We will see what I call the biomarkers. We might have what's called a NOAEL or a LOAEL, in other words what's the lowest observed dose or concentration we see causing toxicity and what are the values that cause no toxicity and how do they relate to what we think are in fact the safety margins based upon animal studies. Most animal studies that I'm away of work upon a safety margin of about 100. Sometimes they work on a value of about 1000, that's generally the values.
The problem we have when we think about that, when we start looking at it, is in fact most of our patients we don't know what the dose is because the patients who come to us have already been poisoned and we don't know what their exposure was. In many animal studies, it's starting to change a little bit, a lot of it is based upon doses and they actually don't measure plasma levels. We measure plasma levels in our patients, what we really need to have is what the plasma levels are in the animals so we can apples compared to apples. Once we've got that we can start really comparing what's going on and looking at which particular types of agents, that might be pesticides, it could be any sort of drug, do we have to think about what the safety margins are. Because I don't think they should be the same, 100‑fold safety factor for every drug and poison there is. We should have things based more upon mechanistic class, we should have it based much more upon evidence than upon a hypotheses.
The other thing which I want to do before I move on to dermal absorption is just to emphasise it's not just what gets in, it's actually what happens when it gets into the body. This is a plot showing the body surface area over body weight versus age. The reason I'm showing that is if you have a child and that child is exposed to the drug, they will have a higher concentration in their blood because they've got a lesser body weight for the same sort of exposure, because they've got a larger BSA to body weight ratio. Glomerular filtration rate tends to be lower in very young children. it goes up and then falls off for all of us with time. As we get older of course our metabolism tends to fall away.
This sort of allometry really is very useful and what we've learnt from some of the work of Nick Holford who works a bit with me has been playing around with, it's not just allometry, it's also maturation. What happens in children is a process of enzyme maturation, glomerular filtration maturation and if you try and do it by allometry alone, and you'll see in this diagram here, this would be the actual space, the composite of intervals, based on allometry alone. What you can see, this allometry alone tends to be for this population here, it actually overpredicts for the toddlers and for the neonates it doesn't predict at all. If you put a combination of allometry maturation you can predict the whole lot with that space.
That's where in fact a lot of us are going is how can we get a much better and more understanding of how to dose patients, how to protect them and how to treat them if they're actually poisoned.
My focus is going to be upon dermal absorption today. The first thing I need to highlight is of course it depends upon what the compound is, it depends on its formulation, how you actually apply it and what the state of the skin is. They are the key parameters you always have to think about. There are some assumptions made about some of these and some of those should be challenged but they are the issues.
The second question which in fact was put to me and which I'm presenting today is what should we be using the quantify this absorption? The general view for most here is percentage absorption is easy, it's the best way of doing it. But I want to challenge that and in fact look at how and when we can use it. Jim presented some work on permeability coefficients, permeability coefficients have value but they also have limitations. I just want to talk about those briefly.
I've been pushing maximum flux. Maximum flux have got limitations as well. It gives you an idea of maximum exposure but doesn't give you a measure of actual absorption. I suppose what I want to push is the use of amount absorbed, actually quantify what the amount is, because that's actually what's going to tell you toxicity.
The last thing which I think I'm going to concentrate on a little bit is talk about the dose applied. A lot of the work that's been used to establish kinetics of drugs applied to the skin has been based upon large doses. In France, effusion cells. They've done it to make life simple for themselves. They try to make the mathematics easy. That's a problem. You actually create artefacts in doing that. What you're really best doing is using finite in use conditions even though it's more difficult. A lot of the work we're doing related to FDA relates in that area. What actually happens within use, finite‑type conditions.
We have to also think about roots of dermal absorption and in fact Nancy has covered some of that so I will do this very briefly. She has talked a bit about the hair follicles. Our work with multi‑photon microscopy is we tend to see the highest fluorescence actually in the intercellular region. It looks to me that most penetration occurs between the cells and not through the cells. If you lipidise the skin you'll see it through the cells but by and large we see it occurring between the cells. That says to us really the compounds will go through the skin best should have some lipo‑follicity dissolving those lipids. That you'll find is a key criterion in terms of some of the dermal absorption factors.
In terms of follicles, our observation is, and this actually after finite dosing, this is using a compound caffeine as a marker, and this was some work done with Jurgen Ledermann in Berlin where he went and carefully sort of forensically closed all the hair follicles, just simply closed them all up and let them have both open and closed and then I went and moulded it. What you can find here is if you do this the hair follicle absorption is very rapid and transient. People often talk about hair follicle absorption being very important but it's usually a transient process. Most of the absorption which you see over a long time is occurring in the stratum corneum. This is in vivo work done in volunteers.
There are a range of factors in skin penetration. Some of these will be covered a bit in the next talk but I just want to highlight sub‑sensitivity. This is in fact where material is retained by the skin. Some of the residue work which Jim and others referred to. Volatility, release from the vehicle, wash, rub effects, the kinetics which I'll talk about a little bit, this issue of residues again. I'll talk a little bit about anatomy, not so much on natural spread. Vascular perfusion has been one of my favourite areas of late but I see in the pesticide area people tend to ignore it. Cutaneous metabolism I took out because of time but that's also a big area. Obviously excretion.
Probably the most important consideration to think about is to keep it simple and say that the actual sort of agent, you're going to have both external exposure and internal exposure. External exposure is what you put on the skin. Internal exposure is what gets in the body. When I'm talking about toxicity I'm talking about local toxicity, that is toxicity on the skin which relates to external exposure and I might talk about internal exposure. Some of my therapeutic work was aimed at how do I treat melanomas by regional treatment where you try and get it while it's still in that local area.
The exposure depends on what I call an effective concentration. Not the actual concentration because you'll find the effective concentration can differ between different vehicles and we'll come back to that. The area, the larger the area the larger the dose that can be absorbed. Time, the longer it's on the body, the more that you'll have absorption.
This is some work, I was speaking to Howard about this before I talked today because Howard is one of my old friends who is well known in this area. He referred me to his particular chapter in this book by Hayes, which is actually not a bad read. I replotted some of his data and the first one I plotted was really looking at the dose absorbed for different sites of the body. What you see is that males are particularly vulnerable but for the rest of us we've really got to watch what we put around the face and around the neck. That's where it's a key area. For the working people that's the sort of thing to think about, the fact, the neck, those general areas.
Howard referred me to this particular diagram. I've been worried about it because basically I think Parathion is still in the US market but it's withdrawn here to my knowledge, at least that's what everyone's telling me. So this Parathion but it actually does show you the issue about toxicity and poisoning. Basically what happens is the LD54. Parathion is estimated at about 980mg. If you in fact apply a dose of about 980mg obviously you're going to lead to half your people dying. In this case it was a dose of 4µg per centimetre square over the whole body and applied for 24 hours. The shorter this is basically if you actually apply this over the whole body you'll have death within eight hours. If you have it just to the neck and to the head you could have it within 24 hours. What this does show, it's important to actually get this right, to recognise there is an issue and we need to take this seriously.
There's a whole heap of derivations with dermal absorption factors. I don't intend to go through them all here because you can read them yourself. Essentially what they've been doing in this area, and this is a fairly recent paper, is they were really assuming most of these things don't matter. I think that's to be proven. When people tell me damaged skin doesn't matter, I have trouble believing that for chemicals. I understand that, Nancy, for some of your nanoparticles but for chemicals I don't quite believe it particularly when I'm working with burns and working with ulcers and other areas, I don't think you can quite jump to that conclusion so quickly.
In terms of blood flow, blood flow is a key thing and I think in fact when we come to in vitro, in vivo correlation, it's actually been underestimated. I think some of the low absorptions we've seen in vitro really does reflect there's actually poor perfusion of the dermis and people have got very thick dermis. There's a whole pile of stuff that's been assumed here which I think we have to really, those of us in academia, has to critically appraise and try and refine. That probably doesn't help you guys as regulators or as the industry.
The simplest method we use is the Franz cell. We've heard this already from a couple of people. This is in fact some human skin. I'm just showing you the peeling off of the epidermis which we can do. You can either use the epidermis which in my view is like a maxima evasa of the latter state or you can use dermatome skin which to me is like a maxima evasa constricted state because the dermatome is actually quite thick. Most people will try and use a very thick solution because you get a beautiful sort of profile of concentration, the receptor or amount of absorbed versus time, just being a nice straight line after a lag time which is the time it takes to get through the skin.
The equation for this is fairly straightforward and you end u with a permeability coefficient which is what most people like to use. This particular model makes some assumptions. It says the stratum corneum is rate limiting. It is in some cases but for very lipophilic compounds it probably isn't and in fact that sort of becomes a real issue. It assumes steady state. Rarely do I ever see steady state in most of the work I deal with. It assumes an infinite sink, it assumes everything is perfectly soluble. I'm going to show you a case study in a minute and I'm going to ask you whether you think they actually used an infinite sink and if so has that made the study flawed. That's the sort of question we have to ask.
They often use an aqueous vehicle and I'd ask you, how many products do you know that are out there which are purely aqueous solutions? There's no other excipients. That becomes the issue with a lot of the work we have to work with.
There's been some nice work to show that the permeability coefficient is higher with higher partition coefficient and Gordon Flynn who in fact has just done a book with me which is about to come out showed very beautifully how in fact there's a combination of size and octanol water partition coefficient which can control a lot of these compounds' penetration.
The problem is this is probably all flawed and so I'm showing you the state of the art, I was brought up on this, and the reason it occurred is that most people in the area are physiologists. In physiology you work with dilute solutions because inside the body, but most of those of us who work on skin, we're working with fairly high concentrations and it's actually not relevant to us. What we need to know is what is actually occurring in high concentrations and therefore it's a different scenario.
If you look at the KPs and I'm just showing you here some KPs versus simple alcohols, these are just carbon chains. You see how the alcohol with salines, actually the permeability coefficient goes up. If you look at other vehicles, say an oil, it goes the other way. So what does this all mean? It means that permeability coefficient is an artefact of those that work in physiology with aqueous solutions. It's not a good thing for us to be using in terms of any of our dermal assessment. Although all the academics push it and you'll see people talking about the pots‑guy relationship and everything else, it's actually not necessarily the most relevant for us. It really is in fact the measure or it relates to infinite dilute solutions.
A more reasonable measure which I've been pushing is what's called maximum flux which is really actually the flux that you get from a saturated solution. If you've actually got a solution in oil, we have in water base, they'll all be in equilibrium with a saturated solution. In theory, provided that particular product doesn't affect the skin, they will all have the same flux. This is old data of ours where you can see the flux with a whole range of different vehicles across an inert membrane, they are all perfectly flat. It doesn't matter what the vehicle is, and so what you can now do is you can actually start to be able to transfer data from one product to another by understanding maximum fluxus. That's very useful.
Then we took the maximum flux, we then looked at all the data we could find. These were based mainly on aqueous solutions and we were trying to do some work like Jim. This is how do you plot it in data. What really amazed me when we did it, we plotted it against molecule waves. This explained 70% of our data. A simple relationship between the actually flux, maximum flux, size explained most of the data. What that's saying is the big molecules will never go through, the small molecules will and so people talk about this cut off of 500. You can see the cut off at 500, you've got it really low. This is a log scale for flux, it's a really low flux. The first criteria you have, if you've got a compound with a molecular weight of greater than 500, it's unlikely to go through the skin unless you have some enhancers in your vehicle or other things which will in fact try and force it through.
You can also plot this against melting point. If you see it against melting point which defines, the melting point gets higher the flux gets lower. Why is it? What happens is a lot of these things draw covariant. The very large compounds all sat at the very high melting points. The two are related. Then we also looked at octanol water partition coefficient and lo and behold there wasn't much of a relationship. It's dominated in fact by molecular weight. If you do take compounds of similar molecular weight you find there's actually quite a good relationship, a parabolic relationship with in fact log optimal water and our data suggested to us it's all about solubility in the lipids. That's what really dictates it.
One of the things we published quite recently, this is an invited article, we actually took hold of all the data that's been published for transdermal patches. It's relevant to us because the same principles apply. What we use is an algorithm. The algorithm takes into account both the melting point and the actual molecular weight. This is all the data there is. This is all the patches available. Everything you find fits in here including the very small compounds like nicotine and nitroglycerin and some of the very large ones down here. These ones here which you have more than one sort of dot for a particular compound, that occurs because when people make these products some have enhancers, some don't have quite the saturated solutions and that variability just reflects the different formulations out on the market. These are all the marketed products out there. I think where we're going to in a lot of this area will be nomograms. You don't see too much of it now but I suspect the future will be nomograms which is much easier to use than trying to use complex equations.
And so right now there's a suggestion that the actual values should be POW of less than minus‑1, greater than 4, and a molecular weight greater than 500, and maybe if you do that then you can use a default absorption value of 10%. I'm going to show you a case, an example in the minute which has got log Ps of about 6 with quite good absorption. So what does that all mean? It means a lot of this is a bit hairy. People have made some sort of examples, some guidelines, and they haven't been rigorously tested and validated. That's for us in academia an opportunity.
The current guidance for pesticide dermal absorption basically says that octanol water partition coefficient and molecular weight are not good predictors in general. They are suggesting that you can use this default value. There's a question about concentration and really if you read this carefully they don't quite know what's going on and they then have sort of various criteria which then also relates to what you must do for formulations and dilution.
There's various flowcharts now available. I actually gave Health Canada a lot of help trying to work out their various flowcharts to know, how do we actually work out what sort of criteria. We have now flowcharts for different scenarios. This is a range of them here. One example is this one and what you would do with this one is you would have a product. You might be able to take hold of the data for aqueous solutions. You might then sort of say yes‑no, and you'll come up to a sort of final solution. I'm not going to go through that now because you guys can go and read it up as well as I can.
I thought I might talk about the pyrethroid pesticides. I got interested in this because one of my staff is actually an organic chemist who has been making pyrethroid‑type insecticides. I decided to have a quick look at it. It comes from a chrysanthemum. These are actually synthetic forms of that. The three particular compounds which I'm going to look at are these which are in a paper that I'm about to talk about. One is Bifenthrin which is actually on the market. I was going to do Parathion and Malathion and Phil took me out of that so that's why I've moved to these. You'll see the first thing is, if you look at the actual product composition that this is in. is it water? It won't dissolve in water so a KP based in water is a meaningless exercise. Then you look at what that particular thing is. M‑methol 2 peroladone is quite an effective skin penetration enhancer. To what extent has that been taken into account by any of the people in their analysis? I suspect they haven't even thought about it. it's one of the problems I've had related to some of this mismade to NICNAS]. The combinations of product could actually make something which is fairly innocuous become in fact a bit of a concern because all of a sudden you've got it being absorbed. You need to find a way to understand what's the potency of that compound that's a toxin and if you actually did get it being absorbed when does it become an issue?
If you look at say Deltamethrin, you find this is actually used all over the place but in fact it has a problem because it's actually highly bound to soil. One of my colleagues Annette Bunning has spent a lot of time trying to understand this binding to soil and how compounds can be absorbed through the skin after binding to soil. Also Ronald Westock which he was with us was doing the same sort of work.
There's about 1400 products containing permethrin so this is clearly a big area and look at the formulations: liquids, solids, dust, aerosol solutions, sprays and clothing. Can you use aqueous solutions as your guide? The answer is not really. Then you see it's used in other things like cattle ear tags, flea collars. You can't necessarily use aqueous solutions again. You have to really start to reapply our knowledge of chemistry to put that together.
The other interesting thing which I was talking last night to Phil about, this is also regulated by FDA because these are used for human products so clearly there's a need for the APVMA and presumably Health which I presume is happening to be interacting which you're sharing things which are in common. Big opportunities in terms of those things going forward. The comment made in this particular thing I read was less than 1% of permethrin on human skin was absorbed. The first paper I went to had something like 2% or 3% absorbed so obviously someone sort of has been choosing their literature selectively to help their case.
This is a particular paper I was looking at and in fact I gave it to my team to have a look at as well. This was a lovely study and it's looking at the in vitro dermal absorption of these three compounds on human and rat skin. It's from the EPA therefore it should be very reputable. These are the compounds. The first thing I note if you look at them, they're all very large. We talked about a cut off of 500. These are all in that area. They're all pretty large. Melting points are quite low, log Ps very high. So this less than 500, one to four rule, you're in trouble already. Aqueous solubility, that's tiny, really tiny. When they did this work the first thing which really caught me is they used human cadaver skin and they dermatomed it 300 µm, 350 to 400 µgm. For those of you who know anatomy of the skin, the stratum corneum is about 10 µm. The viable epidermis is about 100 µm. You could say 10 to 30 µm for the stratum corneum, a bit thicker of course if it's in the plantar region. Then it's about less than 100 µm going from the viable epidermis to the actual dermal capillaries or papillaries. This is very deep and of course this is, in fact, not well perfused.
Of course, rat skin was done the same. Then they used this, and this is important, they used Hanks Buffet Solution with 10% bovine serum. The question to you is do you think that's actually dissolved at all? Has it carried it all the way? If you then look at the results they do have penetration in humans of about 2%. They then sort of said, we used acetone as it's a good vehicle as it's volatile and many chemicals are soluble in it. Does it reflect some of the products?
I was just going to comment on Jim's stuff. Jim did some nice work in this area and he in fact didn't find so much penetration but he did find stuff with the pig study.
Time to reflect, and this is where I'll finish off with my last slide. The percentage absorption will vary with how much you apply. Permeability coefficient, little value. Maximum flux, probably useful but difficult at the maximum dose. Probably the best assessment is amount absorbed. Then we come on to dose, product absorption factors and the last question is, is there any relevance to in vivo? I was going to show you some data on that but we'll do that another time.
Errors and omissions excepted; check against delivery.