problems w/ precipitation&extraction of hydrophilic drug

[neuger]

Anyone had any problems with using precipitation of plasma with hydrophilic compounds when it comes to inconsistent recovery? I have a tricyclic nucleotide mono phosphate that just isn't reproducable/linear on LCMS from plasma. My instrument conditions are very robust, I can perform this in acidic or basic conditions on the YMC-AQ as discussed from a previous thread when use just mobile phase as the matrix. It is very linear in water, but just not from extracted supernatant. Does it have to be SPE? Could I be losing drug in the dry down step or when I reconstitute. I see a small pellet when I reconstitute, but I don't pick up that part to inject, only the clear reconstitution after a quick spin, I know I may be actually precip'ing again without realizing it at the reconst step, but what else can I do.....except SPE mixed mode. Any suggestions are more than welcome, this has been very frustrating.

[Uwe Neue]

I suspect that the issue is not inconsistent recovery, but inconsistent response due to ion suppression. Ion suppression can vary from plasma sample to plasma sample, and possibly also with small details in the sample preparation, especially when the ion suppression is large.

Have you tested for ion suppression? If I am correct, then mixed-mode ion-exchange SPE is the only solution.

[neuger]

I don't have the capabilities for an infusion pump to T - in a stream of drug while I inject the MP or the reconstituted supernatant. However, my baseline is always pretty low (for a single quad), usually starts at or near the same intensity on the y axis per injection and there is clear baseline on either side of the drug peak. Does ion suppression always affect only the peaks of interest? Wouldn't it affect my baseline of a monitored mol wt. in a plasma matrix also and not just the drug? I know the phosphate group isn't helping either when it comes to suppression, but why don't plasma components suppress the same for each sample?

I think I need to run a post extracted spike versus my spike in mobile phase. Might that be another way to check ion suppresion? I have been looking at recovery theoretically instead of checking the response of the sample. I was simply assuming I wouldn't recover it all from plasma, but I would deal with what I got and maybe not checking this step has ruined what I thought I knew. But if it varies from sample to sample, what is the point...can you tell I've been working on this method a while...

I also simply don't understand why plasma can create such large differences from sample to sample in MSD if you don't have tandem MS to get a daughter ion. If I treat the sample the same each time, what varies in plasma that is being carried over the the MS? What a rip off this single quad seems!

[neuger]

I must add, I am using the same pooled blank plasma each time whether I make a blank or spiked sample.

[Uwe Neue]

Ion suppression is rather ubiquitous with plasma samples, or with other "dirty" samples in general.

There is a way in which you can figure out what is happening, and it is similar to the protocol that you have suggested. Please read Matuszewski et al., Anal. Chem 75 (2003), 3019. They have a complete protocol worked out that tells you exactly what is going on. In essence, it is a comparison of a non-plasma standard to your prepared standard, but it is a bit more involved due to other concerns. The matrix can affect the analyte and the internal standard in different ways, and the source of the problem may have little to do with proteins per se.

Do the experiment that you have proposed and you will see what is happening.

The protocol suggested by Matuszewski does the following: 1. inject your analyte and internal standard in neat solution; 2. do your standard sample preparation on a blank plasma (or other biological) sample, then add your analyte and the internal standard to the prepared plasma sample; 3. spike blank plasma with the analyte and internal standard, and do your sample preparation protocol.

The difference between the results of 1 and 2 tells you about ion suppression from the residual stuff in the cleaned sample. The difference between 1 and 3 tells you additionally, if there is loss of analyte due to the sample preparation protocol, for example due to strong protein binding of either analyte or standard.

Now you know what is going on, and what you need to fix. They suggest that one is doing this ALL the time, not only when you think you have a problem. I think that it is some work, but I also know that it is worth it.

[neuger]

Yup, similar stuff. I think we are left no choice, it seems interesting that it can even happen in tandem MS conditions...

Hey, perhaps we will learn that it isn't even the sample clean up step at all and that it may be the ESI source set to a condition too extreme for the analyte/matrix combination. I thought we had that optimized when we carried out the chromatography of the analyte in mobile phase only, but maybe the ESI isn't optimized yet. I need to get my APCI source back from a colleage, this could actually solve the entire problem, but I think we will carry out the tests no matter what.

Either way, Matuszewski et al. is a good read along with some of the other things I found. Thanks again for making the effort, hopefully this week is more successful than last.

[neuger]

More news....we believe a large problem is that we have 95% aqueous for our MP. This could be bad for ESI. If we use a waters atlantis HILIC we propose better results, but aren't sure.

We have the separation under acidic conditions (ph 3), we have a negatively charged phosphate functional group and a positively charged amine group that is part of a ring structure....would the HILIC column still be successful even though it specifies to be used with basic polar analytes? If not what else is available? We feel strongly that ionization or lack there of, is a main source of our reproducibility problems and need to move away from high AQ MPs.

[Uwe Neue]

HILIC could work. It works for all very polar compounds, and silica HILIC works best for amines. But it does not have to be a pure amine...

You could get better or different separation from the interferences, but this is a shot in the dark. I think it will be better to do a better sample cleanup to start with.

[SIELC_Tech]

Neuger,

I think that I can offer you alternative solution when your phosphate will retain based on mixed mode mechanism but you plazma components will come before void (size and ion-exclusion mechanisms). You can probably end up with minimum sample prep.
Something like this ("dronic" acids):

http://www.sielc.com/application_138.html

In this case you can use higher ACN and have better sensitivity.

Contact me if you have questions. My email is on the bottom of this message.

regards,

Vlad

[HW Mueller]

Vlad, you are showing examples under super ideal conditions there, aren´t you? Would love to see the mess one gets if one injected plasma directly on that mixed up mode column.

[SIELC_Tech]

Dear Hans,

We never tried "dronic" acids in plasma. But here is application for nicotinic acid in plasma. We were able to do 200 injections of sample with plasma (1:1):

http://www.sielc.com/Technology_DirectP ... lysis.html

Primesep SB and Primesep D are similar in terms of surface chemistry-both positively charge, although we have not done recovery study for this application.

[neuger]

I guess this thread has taken a life of it's own and I think the direction is lost in regards to figuring out what was wrong. I am using single quad MS, there is a paper that talks about the kinds of ion suppression that can be obtained from precipitation of plasma, insane amounts if you take a read...believable if you actually run the tests, I did it as we talked about Uwe!

I can develop a separation method 16 different ways if you would like when it comes to chromatography, but when it comes to ionization in an ESI spray chamber combined with plasma components from a precip, I think I have learned my lesson...quick and easy precip to use with DAD or FLD may work, but not in single MS - there is simply too much variability in plasma. The columns, the separation, etc has little to do with it, all bets are off in the spray chamber. I appreciate all the input in any case, but this seems to be random stuff at this point.

[neuger]

The High AQ mobile phase wasn't the problem, so the column change doesn't matter....forgot to add that part. I guess I was considering change in that route, but again, the separation isn't the problem we were all discussing originally, it was suppression, recovery, etc.

Uwe, I got what I needed from your input and unfortunately the extraction method is a failure, but my chromatography is perfect!

[HW Mueller]

Neuger, I am confused. What caused your suppression? Coinjected "dirt" underlying your sample peak or mobile phase? If it was the former than why would changing to a wizzard column like that of Vlad not theoretically solve your problem? (I didn´t mean to digress above, just thought a little scepticism could be helpful).
Vlad, you have enormously high concentration of your analytes. I did a nearly perfect sep many years ago after a few hours of trying with a Pinkerton (restricted access), injecting undiluted plasma spiked with gobs of cortisol. Then I tried it with unspiked plasma and needed several years and additional columns before something useful came out (described before).

[neuger]

it is all related to ion suppression, not a co-eluting endogenous "dirty" peak. we checked a post extracted spike vs a mobile phase spike and saw over a 96% suppressed signal. column won't help that, sample clean up in SPE is my proposed solution....mixed mode anion exchange looks like our best solution to dictate what kind of interaction we want and to reduce the plasma matrix suppression by removing many interferences.