If you are unsure of the mechanism of action, one solution is to perform a time course experiment. Apply the treatment at a range of time points before LPS challenge, during, and after LPS challenge. To avoid confusion about whether the experiment has "worked" use a positive control treatment, with a known mechanism of anti-inflammatory action, such as dexamethasone. It is also important to verify cell viability at the end of the experiment; the MTT assay is easy and cost effective. In addition, it might be useful to have a range of readouts, rather than rely on the Griess reagent.

You do whatever works. You can added it before, or after. And if you think it can disrupt LPS binding, you can also added at the same time.

Are you sure the extract is functional? If you are, shorten the time to

If you are unsure of the mechanism of action, one solution is to perform a time course experiment. Apply the treatment at a range of time points before LPS challenge, during, and after LPS challenge. To avoid confusion about whether the experiment has "worked" use a positive control treatment, with a known mechanism of anti-inflammatory action, such as dexamethasone. It is also important to verify cell viability at the end of the experiment; the MTT assay is easy and cost effective. In addition, it might be useful to have a range of readouts, rather than rely on the Griess reagent.

The Iain's suggestion is very appropriate. I also suggest you perform a concomitant treatment with LPS and different concentrations of the extract for 24h. After, measure the IL-6 or IL-8 concentration in the supernatant by ELISA and compare the concentration of the LPS-stimulated cells without the extract supernatant.