I'm willing to bet that your cell lysates aren't made up in the same volume as your culture supernatants so that's more of less your reason right there. You're comparing apples and oranges. The second thing to bear in mind is the profile of release of TNFa is usually short 4-6hrs after induciton.

I'm not up on the release pathway for TNFa but I know that for some cytokines like IL-1 you can induce the denovo produciton intracellularly without stimulating its release into the medium. That might also be something to check on.

I am wondered if you concentrated your culture medium before performing the western blot.  I ask this question because i will do it before western blot and I will not do it before ELISA. But, this difference of sample treatment might explain the absence of detection in ELISA. You might be under the detection limit in the ELISA.

The second point might be the efficiency of your ELISA antibodies versus the WB antibody.

Your TNFa ELISA levels are fairly low. What is the limit of detection of your assay? You could always try a positive control eg LPS for your ELISA.

Thank you for the replies! The ELISA was performed on cell culture medium since we were expecting the cells to produce  a large amount when treated. The westerns were on whole cell lysates. The minimum detectable dose is less than 5 pg/ml.

I'm willing to bet that your cell lysates aren't made up in the same volume as your culture supernatants so that's more of less your reason right there. You're comparing apples and oranges. The second thing to bear in mind is the profile of release of TNFa is usually short 4-6hrs after induciton.

I'm not up on the release pathway for TNFa but I know that for some cytokines like IL-1 you can induce the denovo produciton intracellularly without stimulating its release into the medium. That might also be something to check on.

I have to agree with Peter that you're comparing apples and oranges. It may be that a fair amount of the TNF alpha being released is being quickly taken up by TNF receptors on the cells (I believe the binding affinity is quite high). This would mean that the levels in the media would only increase slightly. Additionally, a quick search turned up this paper: "Change in TNF-alpha receptor expression is a relevant event in doxorubicin-induced H9c2 cardiomyocyte cell death.", further supporting that the TNF alpha is likely going to be more concentrated in the cell lysate as opposed to the media. 

You have to check with detection limit of protein for ELISA. 

Hi toni,

Yes, you cannot compare using cell lysate in one case and cell supernatant in the other. Try performing an ELISA using the cell lysate.

Also, you absolutely must have a positive control (LPS at 10-100ng/ml for 4 hours should be good). You cannot even know that the ELISA is working properly without the positive control, right?

And the differences in ELISA plot that you are talking about isn't really significant (between 3.5pg/ml and 4.5 pg/ml there isn't really much of a difference).

Hope this helps.

Thank you, everyone, for the replies. We did want to perform the ELISA on cell lysates but the manufacturers havent optimised the kit for lysates, only supernatants. I unfortunately did not have time to play around with optimisation as this was at the end of my Msc. Is it wrong to present both of these results in my thesis?  Should I exclude one? Or can I present them both but indicate that one was lysates and one was supernatants?  Thank you again.

What you could probably do is determine the protein content in the cell lysate and the cell supernatant, and then normalize each data set with the respective protein content.

I am not sure whether you even have time for that, but if you can manage that would be great way to represent you data.

Also, I do not understand what you mean by not optimized for cell lysates. If you have the ELISA optimized such that you can detect the signal, I think it would just work fine if you used, say like a 100ul of the resuspended cell lysate.

Again, I would emphasize on the inclusion of a positive control. Have you detected signal using a known concentration of antigen standard?

By the way, you are sure that TNF alpha is released in the cell line that you are using, right? Sometimes cytokines are not released in certain cell lines, and in that case you wouldn't get any TNF-alpha in the supernatant even if they are expressed by the cells.

Tasfia. .. the manufacturers of the ELISA told me that they had never performed it on cell lysates so would not be able to say whether I should dilute samples/how long to incubate etc. It has only ever been performed on supernatants. Also,  the kit included a TNF positive control. I do think my cells produce TNF but I will be sure to check up on that.

Hi Toni, yes the manufacturers do sometimes respond like that. However, it is not a big deal to perform elisa on the lysate using the same optimizing conditions that you are otherwise using. I had similar problems of detecting the cytokine in the cell supernatant, and I used the cell lysate and it gave beautiful results.

I am assuming that you get a pretty good signal using the antigen standard standard provided (that would prove that your elisa conditions are fine). Did you try the LPS as positive control (that would ensure that the cytokine is indeed released)?

Also, about whether cells produce TNF; remember they can produce it, but the important question would be is it released?

I like Benjamin's idea of opening the cells to release the contents; especially since you are more comfortable working with cell supernatant.

Hi, Toni. Can you give me some information about the TNF antibody you used (manufacture and category No.)? I cannot detect any right size band by WB with my TNF Ab purchased from Abcam.

Hi Peizhou...

I used a Cell Signaling antibody for western blots. The catalogue number is 11948S. I really struggled to get bands.

Thanks a lot.

ELISA is the best option 

Dear Toni,

may I ask you which band you are quantifying by WB? As you probably know, TNF is expressed as a transmembrane form (about 26kDa). Upon activation of TACE peptidase, it can get released in the medium (17kDa). To make things more complicated, it can form dimers, trimers, and can be glycosylated or connected to a lipid chain, which will give you a number of bands in the WB. Also, apparently different antibodies give different stainings... I would definitely use a positive control cell line (for example, RAW264.7 or other immune cell lines which are compatible with your antibodies) and stimulate them with LPS to validate your western blot first.

Regrading the release of TNF: maybe it is worth looking into the acivity of TACE change after stimulation with DOX to correlate it with TNF release kinetics?

Good luck!

Elena.