Detection of DMSO

[Postovka]

Greetings

If anyone would be able to help me with any advice concerning this, I would be very grateful.

I am trying to develop a method to detect and quantify residual DMSO in drug substance. Drug is a complex biopharmaceutical (it has protein, DNA and short peptides in it - very complex matrix).

I am currently using column packed with Carbon coated Zr particles with water for isocratic elution. It works very well with my control (which is just water spiked with DMSO). However, when I try to use it with my samples, peaks are not reproducible at all. I get a complete junk of chromatogram. And after I inject my sample, I can see lots of peaks even in water blanks.

I can detect DMSO peak in my sample chromatograms, however it is not separated well from other junk and as I said, peaks are not reproducible at all even when I inject the same sample twice.

Any input would be welcome. What type of column and mobile phases would you use to detect DMSO in samples which also contain protein, DNA and other junk.

Thanks in advance!

[Mattias]

Do you perform any kind of sample preparation?

I assume most of your junk would adsorb to a C18 SPE column, and DMSO would not. You should see some improvement if you just let your sample run through a SPE cartridge before you put it in a vial.

The other option could be to use a so called RAM-column. That is a column that do not retain large molecules. But I doubt you can find a RAM column that can retain DMSO.

[Postovka]

I am trying to get rid of protein by filtering the samples through Microcon 30K filters. But it seems like still enough junk comes through. Would it help if I put any buffers or organic modifier in water for elution?

[Mattias]

I am afraid that I do not know a lot about protein analysis, and I have never used a Microcon 30K filter. From the name I assume that it is some kind of membrane that only allows molecules <30 kDa to pass?

All your proteins and peptide "junk" should be positively charged at a lower pH, while DMSO is a uncharged molecule. This difference could be used in the sample prep. I change my mind from SPE-C18 column, to a SPE cation-exchange (like Oasis MCX) to get rid of the interferences.

Otherwise, I guess GC is the preferred technique for DMSO. But since you post here, I guess you don't have access to a GC.

[Postovka]

Yes, indeed. Only molecules < than 30 kDa get through the filter. But there is still too much interference. Thank you in any case...you gave me some ideas what to do.

So, lowering the pH of mobile phase and using cation exchange guard column would make sense? I also think all proteins and peptides in the sample matrix are getting stuck on the column forever and slowly destroying it. I do use flushing step with 50% tetrahydrofuran to clean the column, but Im not sure if it works.

Thanks again for the input.

[Mattias]

To me, that sounds like a good way to go. I would do the <30 kDa filtration first and then hook up a cation-exchange precolumn before the other column. Don't make the mobile phase too acidic, or the precolumn may loose its negative surface. You will probably need to wash or replace the precolumn quite often.

I have never tested anything like this myself, maybe it doesn't work. There should be some more ideas out there (it is only early morning in US...)

[Peter Apps]

If DMSO is dimethyl sulphoxide then this analysis is straightforward by headspace or purge and trap GC, and the muck in the matrix will not interfere.

Peter

[Postovka]

Yeah, I know GC would be much better choice for this, but unfortunately our lab doesnt have access to it.

[Postovka]

Another question concerning this. I am using flushing steps between analysis where I flush the column with 50% tetrahydrofuran in water. Should I worry about my buffer precipitating in the system if I use for instance ammonium carbonate, or phosphate buffer?

Thanks for the answer in advance!

[Mattias]

If you think of "our" cation-exchange idea, I think it should be enough to add some acetic acid to the mobile phase (0.1%?). If you add to much ions, you will elute the molecules from the SCX. Acetic acid is safe to use with THF.

If it is a general question, I think you need to test if it works or do a quick rinse with water before you wash the column with THF.

[JMB]

This sounds somewhat similar to detection of small molecules in serum--if the drug product is in aq. solution, you can ppt. the protein etc. by mixing with an excess of acetonitrile; spin down, decant the supernatant from the solid pellet and analyze for DMSO.

As previously stated, GC methods are best for DMSO--but only if you do have a GC.

Would any LC experts care to speculate on indirect UV detection ???

[Bruce Hamilton]

I'd look at the detection wavelength you are using as well, trying to move higher if possible. DMSO absorbs up to about 270nm - sensitivity may be reduced, but hopefully the interference will reduce even more.

I'd first probably just dilute with mobile phase, spin, and inject, and then look at spectra of impurities to try and optimise detection wavelength. If that doesn't work I'd probably suggest to management they contract out the GC analysis, ao purchase a GC with headspace.

I can't see the need for indirect UV, when DMSO absorbs, and there are trace DMSO in water published papers using direct UV to ppm levels.

Bruce Hamilton