Protocol Online logo
Top : New Forum Archives (2009-): : ELISA and Immunoassay

Elusive protein refuses to be found in human serum with ELISA - (Jun/24/2013 )


For about 5 months I've struggled to develop, and am near giving up on, an ELISA to measure human OSM (IL-6 related protein) by adapting a kit from R and D biosystems.

Some background info: My boss and I have worked in a monomac 6 cell line and I can routinely measure OSM with this kit. The challenge comes from wanting to measure OSM in human serum/plasma, as we have some samples in the freezer from a few different feeding studies.

The R&D biosystems ELISA was not my first choice to attempt to measure OSM in human samples, as they directly state that it is designed for cell culture experiments. But unfortunately, other companies that claim to measure OSM in human sera are lying, as I have found out both by trying their kits and confronting their sales associates; there is no evidence in the scientific literature that a routine elisa kit exists from a reputable manufacturer to measure OSM in human samples.

So, weary and defeated, I was ready to put the human OSM idea to rest, but fortunately (or unfortunately) in a last chance lit search I came across a research team who claimed to have successfully adapted the protocol from the R&D biosystems kit to measure OSM in human samples (!!). They essentially increased incubation times, altered primary and secondary antibody concentrations, and bought a few reagents outside of the R&D product line. Yet despite several attempts, and my own personal additions of reagents to remove heterophilic Abs, I could not generate a reliable signal following Dr. Tease's protocol. My litmus test was to spike sample wells with standard, to no avail.

There is something binding OSM in the serum. I have filtered samples (.45 micron) and added both super chemiblock, excess IgG, or a combo of the two and still cannot generate a signal with spiking. Do you have any thoughts as to what other steps I could take to remove whatever is blocking my protein from detection?


-More of the Same 66-

If the assay is sensitive enough, maybe you could dilute your serum sample enough to minimize the effects of whatever is blocking your specific signal. By combinining diluting the serum with spiking increasing amounts of OSM, you should be able to reach a point where the protein is detectable and at least you'll have something to go by.
Also, as an outside-the-box idea, you could process your serum sample first with protein g or protein a to scrub out the serum antibody and see if that helps.
If what is binding to your protein is doing so specifically, like one of it's receptors, would it be possible to spike in it's agonist to release your protein?
I'm sorry I don't have any answers but hopefully throwing out ideas might still be helpful to you!


I understand your issues with kits from non-reputable manufacturers. There are two categories of kit supplier: those who are experts in the field who rely on scientific niche application and reputation, and those who rely on providing a kit for any molecule with minimal scientific justification and make their living from single no-repeat sales.

It seems that the assay you have works to detect the OSM in manufactured buffer (TC supt etc) , so playing with incubation times etc is icing on the cake, but does not address fundamental questions of work/work not in serum. Similarly with human anti animal antibodies, this may be a problem in some serum lots but not all.

I agree, there is probably a high abundance binder of OSM in serum, and the litmus test for success is consistent spike recovery in multiple donor lots of human serum/plasma. An alternative is that there may be a blood born protease degrading the protein.

If its a protease, you could add a broad spectrum protease inhibitor (Sigma P8340 for example) to the serum at 1% V/V and repeat the spike in experiments. Although that will give a reason why you can't measure the protein, it doesn't offer a solution, because the protein will be degraded by the protease in serum/plasma samples prior to adding the inhibitor.

If it's a binding protein, you need to find a way to kill the binder, but maintain the eptiopes to which the antibodies bind on the OSM intact. Various armaments can be deployed, but it's a process of trial and error.

1) 1 Modify pH. Combine serum with equal volume of 1 M acetic acid (to hit ~ pH 3) or ~ 250 mM HCl (to hit lower pH) and neutralize in diluent on the coated plate with 0.5 M Tris at an appropriate pH (to neutralize). Don't be afraid to leave the sample to cook in the acid for a while (do a time course) as the OSM is reportably pH stable and the binders may not be. Also, you can play around with the pH of the capture antibody incubation (don't have a neutralization step after acidification for example) as the binders may stay unbound but the antibody-OSM binding may work over a wide pH range (I have seen monoclonals bind well at < pH 2).

2) Modify salt concentration of the sample diluent. 1 M NaCl for example

3) 10% DMSO to the sample diluent

4) Heat the sample to denature the binders but leave the OSM intact. OSM is reportedly quite heat stable. It will be necessary to dilute the serum to prevent it solidifying in the heat, so a 3 fold dilution in diluent (1% casein-PBS from Pierce for example or the R+D systems sample diluent) and a 10 minute heat treatment at 70 degrees may well work for you. It's a case of trial and error with the temperature and the duration.

I have used all of these strategies multiple times with different molecules (with success). My best bet would be heat and second acid.

Good luck!

-Ben Lomond-

Super helpful ideas, Ben. I will pursue these approaches and let you know what works. Thank you for your help and expertise!


-More of the Same 66-