My groups are:

(A) Krebs-Ringer phosphate buffer (no treatment/control) incubated with cells

(B) Krebs-Ringer phosphate buffer (plus treatment - high glucose) incubated with cells

(C) Krebs-Ringer phosphate buffer (treatment plus ROS inhibitor) incubated with cells

(D) Krebs-Ringer phosphate buffer with 10 uM hydrogen peroxide (final 5 uM, positive control)

(E) just Krebs-Ringer phosphate buffer

The problem is that treatment (E) is getting higher RFU values at (540 m/580 nm) than (A), (B), and (C), however, (A), (B), and (C) are following the exact pattern they should.

So, when I try to subtract out the blank/(E) from (A), (B), and (C), I get very negative values. I do not think this should be happening or makes much sense.

I will say that when I subtract the blank but normalize (B) and (C) to (A), there is no problem - the values reflect the unnormalized RFUs pattern between (A), (B), and (C).

However, I am not sure that is actually valid because of the high background?

1) I am measuring the ROS in cell media, so no cells are present in the AR reaction, ruling out autofluorescence of the cells.

2) I do not think it is light. My AR working solution is not exposed to direct light in my hood - I turn it off. Also, I wrap it in aluminum foil. And, lastly, I aliquot out what I need from the stock per timepoint and use a multichannel to apply it to my 96-well plate in the identical fashion.

3) I ensured that all solutions had identical concentrations of DMSO (very low) based on the amount needed for solubilizing the ROS inhibitor

What could be going on here?