MTT or MTS assays are used to assess cell metabolism, viability and proliferation, but they are really only good for the first of these. If you want to express your results regarding metabolic state as "per cell" you have to do a cell count; if you want to express your results as "per mg protein" you have to lyse cells and do a protein assay. Of course you need an untreated control, but again, if you want to assess cell metabolism, you have to correct for changes in proliferation, which is an entirely different issue. You need to decide what question you actually want to answer.

Your untreated cells sample is the control. Generally for MTT, 10,000 cells are taken in each well before performing the assay. Also depending on the solvent in which the drug or the test compound is dissolved, you need to make an appropriate control for your assay. For ex: If the compound/drug is dissolved in DMSO, Control cells+DMSO is the actual control, as DMSO is also known to show some cytotoxic effect.

Yes, Anjana say it well. 

1. You have to compare the measured absorbance values of treated cells to values of untreated control cells. It is important. When you add a drug solution in the medium, after an incubation time your cells started die if your compound has cytostatic effect. If your absorbance value of treated cells (in the case of MTT assay) is lower than the abs. value of control cells, it is OK. It means that your cells started to die becouse of the drug effect. When you add the MTT to the cells, the enzymes of the viable cells started to convert MTT to formazan product (formazan can be measured). So the abs. value is proportion to the number of the viable cells. For ex: you take 10.000 cells/well in a plate, than you treat it with a compound. After the MTT assay the abs. value of the control untreated cells is 2000, while your abs. value of your treated cells is 1000. It means that there is half amount of the cells are alive (which produce formazan), and the half amunt of the cells died becouse of your drug. In the case of this expamle, there are about 10.000 formazan producing cells in the control wells, while there are about 5000 formazan producing cells in the treated wells. 

2. If your drug has strong antiproliferative effect, you have to choose more used concentration for treat cells. You can make a stock, and after it you can dilute it more time. For ex: if your compound solution in stock is 10 mM, you can dilute it in medium to 100, 10, 1, 0,1 microM etc. So you can treat the cells with many concentration of your compound in one time. You can find the appropriate concentration which has effect, while dont kill all of the cells. The aother dvantage of this method that you can detect the dose-response effect, and if it is necessary you can calculate IC50 or IC90 value.  

MTT or MTS assays are used to assess cell metabolism, viability and proliferation, but they are really only good for the first of these. If you want to express your results regarding metabolic state as "per cell" you have to do a cell count; if you want to express your results as "per mg protein" you have to lyse cells and do a protein assay. Of course you need an untreated control, but again, if you want to assess cell metabolism, you have to correct for changes in proliferation, which is an entirely different issue. You need to decide what question you actually want to answer.

Anjana has answered your query. 

You need to have control wells. According to protocols you need to have control wells (absence of test drug) and test wells (test drug dilutions). The control well usually consist of only medium or DMSO (This serve as 100% viability post measuring the absorbance). You need to compare the absorbance of test wells with control wells to measure the cell viability post drug treatement. 

I agree with Dr. Kartner. You should use other test system like protein assay or if it is difficult - use more sensitive simple test, e.g. resasurin assay. Basically last one our customers like very much.

In addition to the comments of Norbert Kartner: With the MTT your are measuring a metabolic parameter. So to know whether this parmeter changes with the drug addition, I suggest simply dividing  the measured OD value by the respective cell number - the result may ( or may not be) be completely different from the changes in MTT values of the total cell number, depending on wether the drug has such a metabolic effect or not. To dicriminate between changes in cell number and cell metabolism, I have been using the WST-1 assay, which allows counting the cells in the same well following the photometric measurement. This procedure has given me unexpected but thought provoking results in my experiments, showing that cell number and tetrazolium reduction do not always correlate.

Look at the Promega catalog. They have a variety of assay kits that can differentiate between inhibition of cell growth and cytotoxicity.

The MTT test reflects a variation in cell metabolism (mitochondria-dependent) but is highly dependent on cell number, cell proliferation, induction or not of programmed cell death, and so on....  A variation in OD between different conditions prompts you to further characterize the effects of your drugs, using specific cell proliferation/cell death assays.

Anyway, you can "normalize" to cell number (average) if you seed a second series of wells in parallel, treated in the same way, fixed when you perform the MTT test and stained with crystal violet solution. If your replicates are ok, you can divide the average of MTT OD to the average of crystal violet OD.

You can either measure total protein content in those parallel wells.

The MTT test reflects a variation in cell metabolism (mitochondria-dependent) but is highly dependent on cell number, cell proliferation, induction or not of programmed cell death, and so on....  A variation in OD between different conditions prompts you to further characterize the effects of your drugs, using specific cell proliferation/cell death assays.

Anyway, you can "normalize" to cell number (average) if you seed a second series of wells in parallel, treated in the same way, fixed when you perform the MTT test and stained with crystal violet solution. If your replicates are ok, you can divide the average of MTT OD to the average of crystal violet OD.

You can either measure total protein content in those parallel wells.

I agree with both Dr Kartner and Kello. The MTT method needs to be used with another assay (cell counting so far).

Thank you all for your valuable suggestions. I will do the experiments and post the results soon !