Standard curve confusion - problem in biological matrices

[PMDOC]

Hi!

This might be quite obvious to some of you, but I just don't get it.
I am developing a HPLC-MS/MS method on steroids for saliva and serum(for starters). I use an deutered IS to my analyte. The method involves online extraction using Oasis and an analytical column.

Now...my problem is my standard curves. They are quite linear, but the slope varies a lot depending on the matrices I prepare it in.

In H2O and H2O:ACN (1:1) the ratio Analyte:IS is about 100% higher than i a phosphat-buffered albumin solution.

Likewise, the same phenomena I guess occurs in real matrices as my results are to low.

I have repeated the experiments over and over with consisent results. I have also tried different analytical columns. Furthermore I have tried increasing equlibration times after IS is added up to 24 hours. None of this seems to influence.

Can anyone here explain this to me?
Is this due to spesific ionsuppression / enchasment in the ion source (using turbospray on API 4000))? Or can this be explained by different chemical/physical properties between IS and analyte affecting online extraction og chromatography?

Best regards!

[Jim Shen]

Hi!

This might be quite obvious to some of you, but I just don't get it.
I am developing a HPLC-MS/MS method on steroids for saliva and serum(for starters). I use an deutered IS to my analyte. The method involves online extraction using Oasis and an analytical column.

Now...my problem is my standard curves. They are quite linear, but the slope varies a lot depending on the matrices I prepare it in.

In H2O and H2O:ACN (1:1) the ratio Analyte:IS is about 100% higher than i a phosphat-buffered albumin solution.

Likewise, the same phenomena I guess occurs in real matrices as my results are to low.

I have repeated the experiments over and over with consisent results. I have also tried different analytical columns. Furthermore I have tried increasing equlibration times after IS is added up to 24 hours. None of this seems to influence.

Can anyone here explain this to me?
Is this due to spesific ionsuppression / enchasment in the ion source (using turbospray on API 4000))? Or can this be explained by different chemical/physical properties between IS and analyte affecting online extraction og chromatography?

Best regards!

A superquick observation, since you are working with such none-polar compounds shouldn't you be using APCI or APPI instead of ESI?

[sassman]

I am running steroids extracted from manure/soil on an API 3000. We also typically see ion suppression in highly concentrated manure/soil extracts. I suspect that this can be attributed to the analyte combining with something else in the source forming an adduct of different mass from the analyte. Ion suppression can be avoided if you can separate the analyte from the species causing the suppression. What is your mobile phase gradient?

[bert]

PMDOC,

how did you prepare your standard curves? Spiking of blank matrix?

regards Bert

[PMDOC]

Thx for replies!

I have prepared standard curves in H2O, H2O:ACN(1:1) and a PBS-buffered albumin solution. Furthermore, I have done recovery studies by spiking real saliva samples with known amounts of my analytes(that is actually my standard curve in H2O).

The PBS-buffered albumin and spiked saliva samples gives consistenly lower respons, approx 50% lower.

Now...today I tried using APCI ion source, and preliminary results indicates higher concentrations in the spiked saliva samples - close to what I expect. I have to confirm this though.

Anyway, if this occurs to be true - the explanation is that my deutered IS and the analyte is behaving differently in my TurboIonSpray source? It's hard to accept, but I have read about it. Still...this isn't a normal phenomen?

My method is not optimal for seperation, so I guess I have a lot of coeluting quite similar substances. My problem is that increasing seperation gives lower sensitivity. Using APCI also results in to low sensitivity - the respons is about 25 % of TIS.

My setup is an Oasis HBL extraction column coupled with an Atlantis T3 2.1x50mm. I am running a ACN:H2O gradient on the analytical column from 20-95% from t=2-8.5 minutes. My RT is 8.0 min.

So what do you think? Is this related to ionsuppression/enchasment? How can I get around this problem? All suggestions are wellcome!

Best regards!

[sassman]

Also be aware that steroids don't like to be dissolved in water. They will find anyway possible to get out of solution. Even though you may be well below solubility limit, you can still have sorption to glassware and solid particulates in solution. Plastic containers are particularly sorptive of hydrophobic components when using 100% aqueous samples. Biotic degradation is also a concern in aqueous samples. Adding some organic solvent (maybe 10% ACN) to the samples might help.

[PMDOC]

Some of my samples are prepared by adding 1:1 ACN as a precipitation step, but that doesn't seem to matter. And if glass or plastic absorbs my analytes, why isn't this happening in my standard samples in H2O?

Anyway, I have ran some more samples. They all show consisently higher analytt:IS ratio under ACPI then turboionspray. About 70-90% higher.

The hypothesis that this difference is due to different ionsuppression / enchasement of the analyte og the deutered IS seems a bit far fetched to me. But I have very little experience in this, so I am on thin ice here.

Could anyone advice how to proceed to further eluciate? (Or confirm this to be a very common phenomen?) What is the best way to avoid it, given to turboionspray is the only choice when sensitivity matters?

There is one thing that strikes me though. I have experienced some peak broadening when running biological matrices. This is true for both analyte and IS, but not in my standard curves in H2O. Could there be something here?

[sassman]

Is your internal standard added after Oasis extraction? If yes, maybe the problem is low recovery in the Oasis extraction. Not sure why APCI would give different results though. The most likely explanation seems to be some interference that only shows up in ESI and is causing signal suppression in the analyte but not the internal standard.

[Tony]

1) Is your deuterated IS a deuterated version of your analyte with the same RT? This may sound like a stupid question, but I wanted to make sure your deuterated IS is actually the deuterated version of the analyte you are measuring, and not another deuterated steroid that may behave differently in terms of extraction efficiency, chromatography, etc.

2) Are you adding your IS before your extraction/ppt step? It must be added before to take into account any matrix efffects on extraction.

3) Is the lower response due to higher IS area in matrix, or lower analyte area (assuming by response you mean analyte/IS ratio at a given concentration)? This will detemine whether the anlyte or IS is behaving badly in matrix.

4) Assumming curves are linear for both standards and spikes in matrix, are the slope values for the plot of concentration against analyte/IS ratio the same? The absolute ratio values may be different (as you suggest), but if the slopes are the same, then a standard curve made without matrix will accurately predict analyte concentration in matrix. However, if the slopes are different, then it indicates that there is a matrix effect on your standard curve that needs addressing. You need to consider this, and not individual response ratios alone.


good luck,
Tony

[PMDOC]

Hi...thx for responses!

1. My IS is a deutered version of the my anayte. RT is exacly the same.
2. IS is added as the first step - ACN containing IS as protein precipitation.
3. Actually, it seems that IS area is slightly increased whereas analyte area is decreased.
4. Both curves for standards and spikes in matrix are linear. The slope values are NOT, they differ from 0.033 vs 0.061.

Now...I have tried a different column(3x250mm C18). The funny thing is that this problem is GONE! Anyway...just to check i reverted back to my old analytical column (2.1x50mm C18) and there it was again! After a lot and trial and error I changed the chromographical conditions applied to this column and it seems (for now) that this ion suppression/enchasement thing is eliminated.

Maybe better separation did the trick? I don't know.

I have one last question....The slopes on the standard curves differs slightly but concistently when prepared in different matrices. I know this might be expected....but WHY??? (when utilising deutered IS)

Regards!

[sassman]

Likely separation of your analyte and IS from whatever was causing the signal suppression/enhancement solved your problem.

I can think of two things that might cause different response when your standards are dissolved in different matrices:

1) Changes in the chromatographic separation can affect area counts. This is because for wider peaks, the outside edges of the peak can get "lost" in the noise.

2) Different matrices will affect the degree of ionization of your compound. For example, in organic matrix the degree of ionization will be less than for water. This is somewhat negated by mixing with the mobile phase, but I believe that the sample matrix does indeed have an affect on ionization in the source (correct me if I am wrong).

[Uwe Neue]

Sampling rate maybe? If your sampling rate is too slow on the short column, you could get differences in the area counts between analyte and IS.