The Toxpod

5 in 30 (Ayahuasca, metabolomics and bone marrow)

October 09, 2019 Tim Scott & Peter Stockham Season 2 Episode 10
The Toxpod
5 in 30 (Ayahuasca, metabolomics and bone marrow)
Show Notes Transcript Chapter Markers

In our last episode of the season, we look at 5 recent publications in the field of toxicology.

  1. Steuer, A. et al. Identification of new urinary gamma-hydroxybutyric acid markers applying untargeted metabolomics analysis following placebo-controlled administration to humans. (2019) Drug Testing and Analysis. 11 (6):813-823
  2. Souza, R. et al. Validation of an analytical method for the determination of the main ayahuasca active compounds and application to real ayahuasca samples from Brazil. (2019) Journal of Chromatography B. 1124: 197-203
  3.  Snamina, M. et al. Postmortem analysis of human bone marrow aspirate - Quantitative determination of SSRI and SNRI drugs. (2019) Talanta. 204:607-612
  4. Fabresse, N. et al. Development of a sensitive untargeted liquid chromatography-high resolution mass spectrometry screening devoted to hair analysis through a shared MS2 spectra database: A step toward early detection of new psychoactive substances. (2019) Drug Testing and Analysis. 11 (5):697-708
  5. Sitasuwan, P. et al. Comparison of purified beta-glucuronidases in patient urine samples indicates a lack of correlation between enzyme activity and drugs of abuse metabolite hydrolysis efficiencies leading to potential false negatives.(2019) Journal of Analytical Toxicology. 43 (3):221-227


Contact us at toxpod@tiaft.org

Find out more about TIAFT at www.tiaft.org

The Toxpod is a production of The International Association of Forensic Toxicologists. The opinions expressed by the hosts are their own and do not necessarily reflect the views of TIAFT.

Tim:

Hello and welcome to The Toxpod. I'm Tim Scott.

Peter:

I'm Peter Stockham.

Tim:

And today we're going to do a 5 in 30 episode looking at five recent publications in the field of toxicology and seeing what's interesting about them. So the first article we're going to talk about today is in Drug Testing and Analysis. It's by Andrea Steuer and it's called Identification of new urinary gamma-hydroxybutyric acid markers applying untargeted metabolomics analysis following placebo-controlled administration to humans.

Peter:

I found this quite interesting this paper because you sort of think that after all these years we would have, sort of know, you have a good idea of what metabolites get generated by drugs, but really GHB is, a lot of its metabolites haven't even been discovered yet.

Tim:

It's interesting isn't it? Especially because GHB is one that causes such a problem for forensic toxicologists. It's got such a short half life and so a lot of times when you, even if it's been used, if you're looking for it, you're not going to find it if the samples have been taken many hours later, which often happens in drug facilitated sexual assault cases for example.

Peter:

So it's got a blood detection time of less than six hours or something I think, in the urine of probably less than 12 hours or something.

Tim:

Yeah. And all of the problems as well with endogenous levels of GHB, which just make the interpretation of it hard even if you do detect it.

Peter:

Yeah so it's naturally occurring, causes lots of trouble. But there was um, back in about 2012 or maybe around that period there was, people started looking for GHB glucuronide and that, we thought that could be a good opportunity to look for a longer lasting metabolite, but unfortunately didn't really turn out like that because, uh, the concentrations were really quite low and I don't think they could really tell them apart from endogenous levels, but they looked further at this here in this paper so let's talk about that.

Tim:

So here they're using untargeted metabolomics to find some potential GHB markers. So metabolomics is really taking off. It's, the principle of, it's been around for a while, but I guess the technology is now getting to the stage, both in terms of instrument technology and also computing power, it's getting to the stage where we can do it actually really well.

Peter:

And the high resolution mass spectrometers are becoming more common, so I think it's starting to take off more. So metabolomics is sort of a shotgun approach I suppose where they don't necessarily look for one or two compounds. They look for a whole bunch of compounds and so you end up with a sort of overall features or patterns of compounds in a particular sample type for a population. So in this case we're using one population who's taken a sample of GHB, the other population that hasn't. So then they do a pretty generic extraction, run it through a mass, a high resolution mass spec and try to see differences between the two populations and then focus on those differences to see if they can be used as markers. And we haven't really thought about metabolomics in the toxicology sense a great deal. Often it's been more related to disease types. So they may look at say a population of cancer sufferers, look for particular indicators in their blood samples that may help to diagnose cancer for example, where this is really looking at toxicology stuff, so it's quite interesting. It's in our field.

Tim:

But you get huge amounts of data from this. I mean even from a targeted drug screening method, you get quite a lot of data coming back. But from this untargeted approach, you don't even know what you're looking for. You're sort of just looking for statistical changes in a number of peaks across different groups. You get a huge amount of data.

Peter:

Well shall we talk about what they did? So they administered a therapeutic dose to a number of patients and compared the urine of these patients after four hours or so, uh, with one another using various extraction methods, like really generic extraction methods, just a dilution and then several mass spec and chromatography methods to make sure they picked up as many compounds as possible.

Tim:

And then they identified 42 compounds out of all the compounds that they were looking at, all the peaks. They identified 42 that seemed like they might be relevant, that they might be different between the control group and the group that was administered GHB.

Peter:

When you can consider how small GHB is, it's a tiny molecule, but you still get that many metabolites. That's surprising, isn't it?

Tim:

It's pretty amazing.

Peter:

It's amazing what we don't know.

Tim:

Ironically, those metabolites we were talking about before, GHB glucuronide and then also GHB sulfate weren't on that list of 42 compounds.

Peter:

Didn't they detect GHB glucuronide?

Tim:

Yeah, but the difference in the concentration wasn't significant enough between the control group and the administered group.

Peter:

Yes, so they were the same concentration whether they took GHB or not. So it's further confirms that it's not really much good looking for GHB glucuronide.

Tim:

Yeah, which, I mean GHB is in endogenous, so I guess that's where it's coming from. But you would never predict that before you did this study. I would certainly not have predicted that.

Peter:

No.

Tim:

So then of these 42 compounds, they managed to identify eight or at least presumptively identify them because you don't necessarily have standards and things that you can match these to, but you do have an accurate mass, M+H, and you also do have the fragments so you can sort of piece together the structure as best you can. So because they'd only presumptively identified these compounds, they used criteria which is in the metabolomic standard initiative, which just specifies criteria for how well you've identified something I suppose.

Peter:

So there's similar ones in other fields. I can't recall exactly what they are now, but depending on whether you've got an authentic standard and whether you're just identifying it from a library mass spectrum. And the third level is identification by, um, just by fragmentation pattern and accurate mass, for example.

Tim:

Yeah, there's a few of these things out there. People have published recommendations on this is how you should classify tentative identifications. It's not really, there's not really a right or wrong approach, but it's good to be consistent about these things. So, they're adhering to the standard which is most appropriate here. And so from 42, they narrowed it down to eight, which they tentatively identified and then ended up with five that they thought might be particularly useful in terms of the concentration difference between the GHB administered group to the control group. And these included things like GHB carnitine, a couple of amino acid conjugates, and then a couple of other GHB related compounds as well.

Peter:

So they didn't actually determine the concentration of these compounds. They were just looking at the uh, relative area of these compounds to creatinine, and GHB carnitine is actually relatively high, it's the highest of the lot in terms of its relative area. So that may be quite a useful marker.

Tim:

And in this particular study, they were only looking at a fairly short window, four and a half hours from administration of GHB to when they collected the samples. So normally if it's that kind of window, you'd probably expect to find GHB in the urine anyway, which they do.

Peter:

So this study wasn't really to prolonged GHB detection. It was more to find out what metabolites are there as a first study.

Tim:

Yeah. And but maybe the next step will be seeing, okay, so how long can we detect then these identification markers of GHB afterwards, maybe they can be detected for much longer than GHB.

Peter:

Yeah, and to sort of top this paper off, they did a brief proof of concept where they had 30 authentic samples with 20 were negative GHB and 10 were positive and these markers were detected in just about all of the positive ones, but not in the negative one. So that's a promising start.

Tim:

One thing I have to say Pete, in this paper, they used what they call IDA, information dependent analysis. And that's the acronym that they use here, which is obviously what this manufacturer calls it. But other manufacturers have very similar acronyms, which mean other things, like...

Peter:

The opposite.

Tim:

Yeah, like data independent analysis, so DIA, which is what one manufacturer uses, is the opposite of IDA. It's extremely confusing. All the manufacturers have different acronyms for these things, which basically mean the same things. It's worse than the NPS situation.

Peter:

I don't think it'd be worse than the NPS situation, could it?

Tim:

Yeah, you could have one NPS which is called about five different things. But can't we all just agree on one acronym to use for everything? I guess we can't cause it's all sort of patented technology, but it just makes life very confusing for people with different instruments to be talking about, they're basically talking about the same thing, but they're using different names to describe it.

Peter:

Well the brand of instrument they're using's one of the foremost manufacturers, maybe they were the first people to do it. So maybe IDA is correct?

Tim:

Maybe they were.

Peter:

Who knows?

Tim:

Well, let's move on, shall we? The next paper is from journal of Chromatography B. It's by Souza et al and it's titled Validation of an Analytical Method for the determination of the main ayahuasca active compounds and application to real ayahuasca samples from Brazil.

Peter:

Have you been practicing that?

Tim:

No, I've actually, I've dealt with some cases which involved ayahuasca so I've had some practice before, even though it's not widely used in Australia. It's, ayahuasca is a drink, which is widely used in South America and some other places as well. It's kind of like a brew from some natural plants.

Peter:

And its main active ingredient's N,N-dimethyltryptamine, which gives you hallucinations and stuff.

Tim:

Well, this is the really clever thing about it. I mean, they've been using this for centuries. I don't know how long. The clever thing about it is that one of the plants that they use to make this brew has the dimethyltryptamine in it, another of the plant has monoamine oxidase inhibitors in it because dimethyltryptamine is, although i t's very psychoactive, it's extensively metabolized in first pass metabolism and so orally it doesn't really get into the bloodstream. But if you take at the same time, these monoamine oxidase inhibitors, it allows it to get into the b loodstream and then it can be active.

Peter:

So those inhibitors are harmine, harmaline, tetrahydroharmine.

Tim:

It's been traditionally used in religious ceremonies and things like that because it's a psychoactive compound and as forensic toxicologists, usually we're analyzing samples that are taken from humans, blood and urine, but sometimes we need to analyze other things as well. You might have to analyze some foodstuffs or drinks that were found at a scene or other household materials. We've had all kinds of stuff.

Peter:

Yeah, cakes, coffee cups, but in this case they've got a natural product which are very variable from plant to plant and depending on who's making it, they're making it a bit differently, how long they mashed the ingredients together. You can get different levels of DMT, different levels of contaminants, so it's really going to be quite a tricky analysis to perform with a, especially with a LCMS system.

Tim:

So they analyzed four compounds and they used one internal standard, which was diphenhydramine. One of the problems with doing this kind of analysis on these plant extracts is that the matrix is very complex and it differs from brew to brew. They're never going to be quite the same because people are making them in slightly different ways, different amounts of steeping and so on.

Peter:

And there's no blank. You can never find a valid DMT free matrix to use for matrix effects.

Tim:

Yeah, that's right. So that makes it really difficult to take into account matrix effects. But even just to validate, it just adds a layer of complexity because normally when you're validating matrix effects, you're going to be able to get some blank oral fluid or blank blood or blank urine, whatever, that's available in plentiful supply. But here you just can't get it. So what they did was they got some pooled matrix, which whether, I'm not sure whether they'd source that themselves or it came from other cases, but it had quite low concentrations of the drugs that they were looking for. So then they used that, spiked in their compounds that they were looking at and assessed those against curves that they generated just in solvent. And so they've got these two curves and what they were looking for was to see if the curves are parallel. So the slopes are the same. You obviously had a bit of an intercept in the one that was from the matrix because...

Peter:

Yep, that's gonna be a bit higher.

Tim:

Yeah, there was a little bit of stuff in there, but if the curves are parallel then it shows that there's not significant quantitative effects for the matrix.

Peter:

Sounds reasonable.

Tim:

One of the other things they had to assess in the validation was the robustness in terms of the chromatography, so we talked about this in a previous episode, just how hard is to assess ruggedness in a method. There's just so many little things that you can change and it's difficult to know where to stop in all of this, but they did adjust various LC parameters like column temperature and flow rate. They just adjusted it a bit each way just to see how much that would change the chromatography.

Peter:

Yeah, so I guess if you change the chromatography slightly, you might change the chromatography of a contaminant or coeluting compounds, so that might affect your detection a bit. So as you'd expect with preparation of this stuff, it varies quite a bit and in fact they talk about how it's prepared traditionally and they usually have, during the preparation they usually have a volunteer who has a sample of it to see how strong it is.

Tim:

See if it needs a bit more brewing.

Peter:

Or a bit less perhaps, but of dilution.

Tim:

I wonder if they got a lot of volunteers for that job.

Peter:

Yeah, it'd be the short straw I think. So in this study they found, this is in milligrams per liter. So this is a tea type substance they are drinking and they were talking concentrations in milligrams per liter of between 62 and 340 mg/L, which is quite a decent amount. So there's a huge amount of variation between different concoctions.

Tim:

I wonder if there's some parallels here, Pete, to cannabis products that are coming out now, especially in America with the legalization of cannabis?

Peter:

Yeah and Canada.

Tim:

And in some other places. Yeah, other countries around the world. There's so many of these cannabis products, not, not medicines. I assume the medicines are all pretty robust in terms of having the same concentration.

Peter:

You'd hope so.

Tim:

You'd hope so, but I'm talking about the, you know, the edibles and the oils and things that are coming out, which seem to be just flooding the market. And there's lots of reports that have come out saying, you know, we tested this one and it was supposed to have CBD and it didn't have any, or it was not supposed to have THC and it had lots of THC.

Peter:

Or it had a synthetic cannabinoid.

Tim:

Yeah, that's happened too.

Peter:

So not quite sure what the quality control is on, on those legalized products, but I hope there's some there.

Tim:

You would hope so.

Peter:

So the next article we're gonna talk about is from Talanta and it's on the postmortem analysis of human bone marrow aspirate and quantitative determination of SSRI and SNRI drugs. And it's by Snamina et al from Poland. So I, a few years ago I saw a paper about analysis of earwax and I thought that's got to be the last alternative substance we can find surely after toenails, hairs, everything else. But no, apparently now we can do bone marrow aspirate. In fact, this has been done a little bit before. Not a great deal, there's only a few papers around where they're looking at bone marrow and why would you look at it? Well in some cases you might only ever find a skeleton and in those cases is probably the only liquid material remaining in that whole skeleton is the bone marrow.

Tim:

Yeah. If you work for long enough in the forensic tox field, you're going to come across a case where a body is found and it's severely decomposed and there basically is just a skeleton left. Normally tox in those cases isn't the most important thing. They're probably trying to identify who the person is. So DNA might be important or maybe just the analysis of the skeleton itself to ID the person, but sometimes tox can be useful if you're trying to work out whether the person died from a drug overdose. That could be important, but sometimes also in terms of trying to work out the identity, finding out what drugs the person was taking may aid in that. I mean if you find paracetamol in their bone marrow, that's not really going to help much because most people take that. But if you find something unusual that might give you a clue as to who the person was.

Peter:

I found this interesting because you know you often think about bone marrow being that yummy solid stuff inside of a bone when you're, when you're eating lamb roast or something. But in life it's a sort of solidy, liquidy solid stuff that you can actually sample with a syringe.

Tim:

But difficult to work with, right? Because it's, in terms of analysis, cause it's fairly viscous and a bit fatty.

Peter:

Well according to this, as you're younger you get, there's a larger proportion of red fluidy type bone marrow, but as you get older this gets replaced it to be a bit more fatty so it's going to be a fair mixture of fat and aqueous liquid, which is going to cause trouble and it did cause trouble. So they had to invent specialist sort of extraction techniques.

Tim:

Yeah, so they had quite a long extraction procedure here, they're looking at five selective serotonin reuptake inhibitors, and they're sort of aiming it around therapeutic serum ranges, making the assumption that it's going to be similar in bone marrow, which may not always be the case, but we'll get to that maybe later in the paper. But the extraction was very complex just because it is such a difficult matrix to work with.

Peter:

They had to add a fat removal step and they validated this as well to show that they didn't have too much of the drugs that they wanted to detect weren't extracted in the fatty layer.

Tim:

And at the start of the extraction step, they had to homogenize it by sonicating it. Because I, it sounds like there's some solid stuff in there which may be fatty bits of solid or, or something else. I'm not quite sure. But they sonicated it for a while just to break it up to make it as homogenous as possible before actually extracting it then. And then they ran it on a LC-TOF. So in terms of internal standards, they used three deuterated internal standards, so five drugs, three of them they chose a deuterated internal standard. And for the two which didn't have them, which was citalopram and sertraline, they initially chose the internal standard based on what had the closest retention time, which is a good choice because that might limit matrix effects. If you've got an internal standard which is coming out very close in retention time and you've got some kind of coeluting peak from the matrix, it might affect them both in the same way, so it's good from that point of view, but they found that that didn't actually work very well because that's not the only consideration. It's also how they're behaving throughout that extraction and because this was quite a long extraction procedure with lots of different steps in it, if they're chemically behaving a bit different at any of those steps or multiple steps, then they're not going to extract to the same extent. So when you're setting up a method and you're trying to decide on an internal standard, assuming you don't have a deuterated internal standard, it's hard to know what kind of internal standard to choose, right? You want to choose one that's similar in the extraction and on the instrumentation. That's going to give a similar response on the instrument but also have similar retention time.

Peter:

Yep. It's pretty tricky to do. So in this case, what happened with the one that they had?

Tim:

They got poor process recoveries using the ones that they chose, and so then they switched it and they used a, another one which it didn't elute quite as close, but uh, seem to behave a bit more similarly. I mean in the validation, you can validate using any internal standards you like, but the problem is validation is often done in ideal conditions.

Peter:

Oh yeah.

Tim:

So you can validate a method, you could even have no internal standard when you validate a method and it's probably going to work just fine, you know, you'll get linear curves and so on. But it's when you actually come to do analyzing case samples, especially in forensic toxicology where the samples are so varied and some of them are just really poor and you're gonna get a lot of other stuff coming out. That's when it really starts to matter that your internal standard behaves the same way as your analyte.

Peter:

And the other problem is of course its difficult to get hold of a hold of blank material, so it's not a very common uh, matrix to use. So to, to do things like matrix effect, they were limited in the amount of material they could use for method development. So that meant when they did matrix effects, they only had enough to do one concentration, which is sort of less than what you normally do for tox methods. But, it is appropriate for this cause it's, you're never going to prosecute anyone on this data. It's just uh information that may assist the pathologist rather than getting an accurate concentration, isn't it?

Tim:

Yeah. Usually when you're assessing matrix effects, you've got to do a lot of different extractions. You're doing pre extraction spikes and then post extraction spikes where you're adding your drugs to it after you've already extracted the matrix. And so there's a lot of different extractions and you need quite a lot of blank material to do that if you want to do it on in a number of different samples, which usually do you want do replicates and so on. But there are some other ways that you can do it with, where you can minimize the amount of matrix that you're using.

Peter:

So the other way of doing a post extraction spike is to do a co-injection. Uh, we did this few years ago as a paper if you want to go and read it and it actually, it's a very quick way to get a lot of different matrices assessed using a co injection method. So you just set the auto sampler up to co inject with your blank matrix. So that's another way to do it.

Tim:

So they applied this method to a few real cases and they found some of the drugs that they were looking at. In one case they found paroxetine which was at a really high concentration, much higher than what would be a therapeutic level in serum. So again, if you're making that assumption that the levels are going to be similar, uh, they suggest maybe that was a toxic concentration, but we don't really have a lot of data on that comparison.

Peter:

So their concentration they found was 3,600 nanograms per mL in the bone marrow aspirate.

Tim:

That's high. That's massive if you found that in a serum. And maybe the concentrations are similar for some drugs, but then you might think there's some drugs which the concentrations wouldn't be similar in serum to bone marrow because it's quite a fatty substance. Maybe something like THC would be quite high in the bone marrow. So they conclude here by saying that more work should be done on bone marrow basically, there's, as you said at the start, there's not many papers that are looking at this.

Peter:

There was, the group in Osaka was doing some work and they compared ethanol and a number of other drugs to, in a few, in a few papers, and they found that there was a reasonable correlation between serum and bone marrow.

Tim:

Well, that's interesting about the ethanol cause that would be one I would have thought maybe wouldn't partition well into bone marrow cause it's so fatty.

Peter:

Well, this is just from my reading a couple of days ago so maybe if someone wants to go and check it properly they can.

Tim:

Okay. The next paper is from a group in France. The first author is Nicolas Fabresse and it's entitled Development of a sensitive untargeted liquid chromatography, high resolution mass spectrometry screening devoted to hair analysis through a shared MS2 spectra database.

Peter:

A step towards early detection of new psychoactive substances.

Tim:

Why do all journal articles have such long names Pete, they make it difficult for us to read on the podcast don't they? Actually let me ask you a question. Some people think with journal articles you should have, you should have really long titles basically because you should be as descriptive as you can so that people know exactly what they're getting when they see it. Other people think that you should be a little more creative and maybe...

Peter:

Catchy.

Tim:

Catchy. Exactly.

Peter:

I think...

Tim:

Any thoughts?

Peter:

I like the former really. But if it's got any substance it should be catchy by itself.

Tim:

Long and boring. Is that, is that what you're in favor of?

Peter:

Maybe.

Tim:

Okay.

Peter:

No if it's long, it's got to be divided up so that it makes sense.

Tim:

Put a few colons and semi-colons in there huh?

Peter:

Yeah, and a dash, they put a dash in this one. No criticism.

Tim:

So anyway, this paper is really talking about how to do untargeted LCMS screening using a database that's available, that's sort of shared between different laboratories and multiple people can upload things to this database and really build up a strong database. Cause it's hard for one lab to bring in all the materials that they need to build up a really good database on their own.

Peter:

Yeah. So this is a, we're talking tens of thousands of compounds or at least thousands of compounds of high resolution mass spectra in, it's in mzCloud, which is actually a good thing cause it's freely accessible to anyone. You don't need to own a particular instrument to actually get into, to search a database.

Tim:

This is really, it's based on one main platform. Other people can access it and, who are using different types of instruments can access it and analyze their spectra and so on. But it's not as easy as just incorporating it into your instrument's software so you can search it, you know from your software that you're using.

Peter:

Yeah. It's based on the Thermo platform, the Orbitrap.

Tim:

And, untargeted LCMS is extremely powerful. But, uh...

Peter:

But it's complex, very, very difficult to analyze an untargeted mass spec run, especially with, or even with high resolution mass spectrometry, it's very complicated. So untargeted meaning they're not focusing on a specific group of compounds like benzodiazepines or like a routine screening method where you might focus on a hundred or 500 particular drugs. They're trying to look for as many things as possible.

Tim:

So you get an M+H obviously for each peak and some fragment ions. But working out the structure from first principles is really difficult. We were talking about that like with the metabolomics before, it's very difficult. So if you've got a database which has got lots and lots of compounds on there, that just makes this process so much easier.

Peter:

Yeah, it's a valuable resource.

Tim:

So they're using an Orbitrap here in data dependent analysis mode.

Peter:

So not IDA?

Tim:

Uh, yeah, you're going to confuse me with all the, uh...

Peter:

Data dependent. So what's that mean? So some manufacturers call it AutoMSMS or...

Tim:

Or IDA, which is that paper that we just read before.

Peter:

Information dependent. Yes.

Tim:

Yep. But basically the point is that the instrument is monitoring what's coming out and then deciding to fragment based on what it sees coming out in the Mass Spec, the MS level.

Peter:

Yeah, so it chooses its MSMS based on the previous MS level.

Tim:

Yeah, and so they do, they talk a bit here about inclusion lists and exclusion lists.

Peter:

They're basically just to make this DDA process work properly. Because if you don't have those lists in it will start spending all of its MSMS time doing stuff that's irrelevant. So they had some blank hair that they got all the um, native compounds in the blank hair, they entered that into the exclusion list. So that means that the, the first MS will detect them, but the MSMS would just ignore them. So then they have a higher likelihood that it's going to detect drugs that they are looking for or drugs that they're interested in.

Tim:

Yeah, and then you can have inclusion lists as well, which they didn't use here. That would be more useful if you are doing more sort of targeted analysis where you know what you're looking for and you know, okay, I definitely want to make sure that I get an MSMS of fentanyl, for example, when it sees that coming out. But where you're not, where you don't know what you're looking for, there's really no point in having an inclusion list.

Peter:

No. But I think you probably would include one anyway, if you do know, you'd have a inclusion list of compounds that you know are there and you do have to follow up as well as getting the mass spectrometer to also look at unknown peaks too.

Tim:

The risk of having an exclusion list is that it's possible there could be something that's got the same mass and a similar retention time, you might be setting it up according to retention times, which I think they did here for at least some of them. There could be a, an NPS for example, coming out which has the same mass, retention time and then it excludes it and doesn't get MSMS, but the risk is probably small.

Peter:

That's right. So the big danger of, of uh, what am I going to call it now, data dependent acquisition?

Tim:

Yep.

Peter:

Is that if MSMS is not acquired, then that compound will be undetectable because it won't be able to search an MSMS spectra against the library. So as it says in the title, it's a way that you may be able to look for NPS. So often, no one ever has all of the NPS that are available in their library or in their, their drug collection in their lab. So it's an opportunity to use other people's spectra to screen your samples.

Tim:

That's right. And there are other, um, databases around as well.

Peter:

Yeah, there's a few around, but a lot of them are very difficult to incorporate into your software. And I think that this one here can be incorporated into the software of the, of the Thermo platform. So another pretty generic high-resolution NPS database, HighResNPS, uh, that enables you to incorporate their drug list into your software with fragments. That's quite useful. I was involved in a paper where we evaluated that.

Tim:

Yeah, that's across multiple platforms, right?

Peter:

Yeah, it looks like it works.

Tim:

So this is, the fact that there are multiple databases like this being set up really just shows the need for it. Forensic toxicologists really want these kind of databases and really want them to work in a way that is very simple to use with their software. And that's probably where it's, we're still coming to that, you know, trying to integrate these, we can get databases of MSMS, but integrating it with software is the really next big key step.

Peter:

That's it, that's a tricky bit. And to get a database that's actually got fragment ion data, that's very useful.

Tim:

So one of the downsides of data dependent analysis, which they mentioned here, cause they had an experience with this, is that if you do have isobaric compounds coeluting or almost coeluting, so these are compounds with the same mass, and they had a couple here amitriptyline and maprotiline, what happens with DDA is it'll see the first mass, let's say the amitriptyline, and then it will acquire MSMS, but then there's a delay that it has before it will acquire MSMS again because it doesn't want to just keep acquiring MSMS of the same peak over and over.

Peter:

Yep.

Tim:

But then it can miss that next peak if they're very close.

Peter:

And the other thing about, so even though you're using a high resolution mass spec, you can't discriminate at the first ms level between isomers of course. They're using 70,000 resolution, which if you're not familiar with uh high res mass spec instruments that's very high. But even if you had a compound that differs by half a mass unit, the, the first part of the ms is where the, where they isolate that ion and that's got a resolution of probably a thousand or something. So you can actually only isolate plus or minus 1 amu. So if you've got two compounds that have got similar mass or even two compounds whose isotopes have got similar mass, then they can interfere with your msms spectrum.

Tim:

So they're applying it in this case to hair. And in a way this is kind of an ideal match because you've got untargeted screening, which gives you the best chance to find any kind of NPS that might be there. And you've got hair which has a really long window of detection. So you can find what NPS someone's been using, you know, in the past month or couple of months and so on.

Peter:

Yeah, that's a useful way of looking at it. So I guess one way to assess the reliability of your method is to analyze past proficiency samples. And that's what they did here, in this case, they analyzed some society of hair testing uh quality control samples. Are they proficiency tests Tim? I think, is that the same thing?

Tim:

Yeah, they're proficiency tests. So normally you're looking for about 20 or so compounds, I think, in these.

Peter:

Yeah. So in this case they found 18 additional compounds. So it looks like it could be effective as a much broader technique than a lot of, uh, current more targeted methods.

Tim:

And they applied it as well to some case hair samples and found a whole bunch of different things.

Peter:

And they also showed that you could use retrospective analysis. So they found that they weren't detecting a few compounds they expected then realized that those compounds weren't in the database on the website. So then they went and added in those compounds into the database and then they found they could detect them

Tim:

Alright, our last paper for today is by Sitasuwan et al, it's in Journal of Analytical Toxicology and it's titled Comparison of purified beta-glucuronidases in patient urine samples, and then the title goes on a bit from there. But basically what they're doing is comparing several different glucuronidase enzymes to see which is the most effective.

Peter:

Yeah. So they compared three, but I guess we should say that they are actually the manufacturer of one of those enzymes. So um, how you regard this paper is up to you.

Tim:

Yeah, I think those, you should always sort of take that into consideration. There's quite a lot of papers that get published by someone who is trying to either sell something or promote something that they're involved in and that's okay. It's okay that people have their own agendas and we need people to be making new things to sell so that we can buy them and use them if they're good products. But I guess you just always keep that in mind, take it into account, and they've got some good data in this paper, but you'd always want to try this out in your own laboratory first before making any kind of commitment to one or another of these enzymes or any other consumables, really.

Peter:

Yeah. Well at least this is a, a peer reviewed journal article, which is much better than a glossy pamphlet I think.

Tim:

Yeah, true. So you can measure glucuronides intact but...

Peter:

You need the standards for them, don't you?

Tim:

Yeah, that's right. And so what most laboratories will do with a broad urine screen is have a hydrolysis step first, which will cleave that glucuronide off of your drug molecule. And then you can just measure the drug itself.

Peter:

So you can hydrolyze in a number of ways depending on the drug you looking at. So like acid hydrolysis is good. It will definitely hydrolyze your glucuronide but it might destroy some drugs. So the enzymatic process is generally very non-destructive, but it doesn't really hydrolyze all glucuronides though I don't think.

Tim:

Yeah, some are better than others. Some are easier to cleave, let's say, than others. And so this can be really time consuming step at the start of your extraction. In the past you would have incubations of a day.

Peter:

Mm, but improved enzyme technology and isolation techniques are helping that I think.

Tim:

So normally these enzymes, well, they're naturally occurring, so they're usually sourced from various animals, including abalone.

Peter:

Abalone, snails, limpets, that's the main three I think. But you can get the genes that make those enzymes and put them into some bacteria, get the bacteria to cook you up a nice little solution of that enzyme. In this paper, they're actually calling their enzyme a genetically modified enzyme. I'm not exactly sure how they've genetically modified it, I'm not big on that side of science, so rather than purifying a whole abalone, they're purifying a vat full of bacteria, which is, I'm assuming is going to be bit cleaner than purifying an abalone.

Tim:

And good for the abalone, right?

Peter:

And good for the abalones, yes.

Tim:

Not so good for the bacteria.

Peter:

This is where I'm not sure Tim, there's other enzymes available. They call them recombinant, which is a similar sort of thing I think, where they put the gene into a bacteria, recombine the gene of the enzyme with bacteria DNA, or RNA, whatever, and then cook that up. That's called recombinant enzymes.

Tim:

Right.

Peter:

Not quite sure the difference between those recombinant ones and the one they're talking about in this paper. Maybe it's the same.

Tim:

Yeah, if you want to learn about gene splicing, this is the wrong podcast, but I'm sure there's one out there you can go to. So here they're comparing their genetically modified glucuronidase with a couple of others, which have been sourced from abalone, and they...

Peter:

They're not comparing it to another recombinant enzyme, by the way.

Tim:

No. And so they picked five different drugs of abuse, which sort of represent a few different classes of drugs, and they monitored them every 15 minutes over the course of an hour. And this enzyme that they're manufacturing here seemed to work quite well, especially for amitriptyline and codeine. Codeine is a difficult one to cleave.

Peter:

Yeah, that's the one they normally talk about being the, the real test compound, if you can, everything else will be hydrolyzed as long as codeine's hydrolyzed. And so sometimes people put d3-codeine glucuronide into their urine before they hydrolyze it. And then to determine whether the hydrolysis has been effective, they look at how much d3-codeine's left.

Tim:

So with urine hydrolysis methods, you obviously do need an internal standard or more than one perhaps, which you're adding in as the glucuronide prior to hydrolysis to make sure that hydrolysis step is working okay. But then you might also add some other internal standards either before or after the hydrolysis to then monitor the extraction and that you're then going to use for quantitation as well. So if your enzyme's not very efficient, this is the problem, if your enzyme's not efficient, you can get a false negative cause you're just not going to cleave that glucuronide. It'll get lost through the extraction or you won't, it won't appear on your column.

Peter:

The other big worry there might be something in your urine sample which is occupying the enzyme all the time. So it doesn't actually get to hydrolyze the compounds that you are interested in. So it might spend all it's time on Paracetamol glucuronide instead of codeine glucuronide.

Tim:

Yeah, so you've got to...

Peter:

The selectivity of the enzyme's very important.

Tim:

So for their enzyme, they got cleavage of all the glucuronide drugs that they looked at in 30 minutes, which is good. That's very quick. So hopefully in the future this will get even faster. Wouldn't it be great if you could have like a five minute incubation time and cleave all the glucuronides.

Peter:

Yeah. And be guaranteed to cleave them all.

Tim:

Yeah.

Peter:

I don't think that'll happen, but...

Tim:

Well keep working on it guys. All right, so that's it for this episode and that's it for this season Pete.

Peter:

It's been a good one, I think. A bit more variation in what we're doing.

Tim:

Yeah, I think we've tried to insert a few different things, a couple of live episodes, and...

Peter:

Taken some feedback from listeners.

Tim:

Yeah. So keep those emails coming in. If you've got emails, thetoxpod@ sa.gov.au.

Peter:

It's good to hear from other toxicologists around the world, see what they're doing. It's great.

Tim:

Yeah. And hopefully this podcast is stimulating some discussion amongst your own groups.

Peter:

That's all we're trying to do.

Tim:

So we'll be back for a third season, probably in 2020 I guess.

Peter:

You didn't tell me that.

Tim:

Yeah. You signed a 10 year contract Pete, you're in.

Peter:

All right, see you next year.

Tim:

Thanks for listening.

Steuer, A. et al
Souza, R. et al
Snamina, M. et al
Fabresse, N. et al
Sitasuwan, P. et al