SEC system suitability - proteins

[biochemist]

Dear friends,

Does anyone have examples (with acceptance criteria) of system suitability specifications when testing proteins/mAbs by SEC-HPLC?

Do we inject low levels of our product (to see if the LOD/LOQ is still valid)?
Or a "mixture" which contains some covalent aggregates? Or do we re-run the MW standard curve each time, and check the plate count?

As an example, the E.P. monograph for erythropoietin specifies injecting EPO at 0.2mg/mL, and again at 0.004mg/mL (50X dilution). The peak area for the second injection should be between 1.5 - 2.5% of the first injection.

I assume this is to check that the recovery of protein from the column is ok, even at low levels...

P.S. the aim of the test is to detect HMW aggregates, or fragments (not for accurate MW sizing)

[biochemist]

P.S. Rande, if you're reading this:

thank you for sharing the info about the "stability sample" containing HMW compounds that you inject into the SEC column to condition it (in a previous post). How it this sample prepared?

Is it a case of a clean column needing it's pound of flesh in terms of protein that is adsorbed? And after the "adsorption sites" have been filled, it can finally start behaving like the good SEC column it's supposed to be.. (Not the most technical description, I know)


Thank you very much for any help

[JJG]

If the purpose of the test is to detect HMW and LMW fragments, then I would recommend that your system suitability confirm that you can detect these fragments. This means injecting a standard which has HMW and LMW peaks and evaluating your resolution between the HMW and LMW peaks. You mention having an LOD/LOQ for the method. Was this done well enough that you can trust that the LOQ will not change? This would mean you also know the repeatability of your method. If so, I wouldn't worry about re-establishing this limit because you are really concerned about high levels of aggregate, not low levels. But, you do want to confirm you are seeing all the HMW and fragment peaks you saw previously, and determining resolution would serve this purpose. You can even trend the system suitability parameters and determine if your column is going bad over time.

You could also evaluate peak tailing and %RSD of n=3 injections as well.

The assumption that I have made is that you understand your method well enough to set a limit on those values. In other words, you've run it before and seen a poor enough resolution value that you can say, when it is below some value, you don't get acceptable results, but as long as it is above another value, your results are fine. As for setting a specific value, I think that will determine which industry setting you are in (i.e. research, GLP, GMP, etc.).

[biochemist]

JJG,

Thanks very much for your advice!

Is it usual to run all the MW calibration standards for the column each time I perform the analysis (I don't see why since I am not estimating the MW)?

I much prefer your suggestion of one HMW and one LMW compound..

By the way, I don't think I can trust the LOQ to remain what it was when the column was new. From what I have read on this forum, it sounds like recovery of proteins from an SEC column can be quite variable as the column deteriorates. As an impurity detection method, I guess it's important to confirm the LOQ with every run? E.g. by diluting the sample and evaluating the peak area wrt the original, undiluted peak.

Would love to hear any opinions about this..


(btw, our company is GMP-regulated)

======
Peace!