I guess the better way is after growing cells enough is to separate cells by FACS or magnetic beads. Do you know a specific antibody for your bladder cancer cells?
You can try selecting them for differential adhesion, which is doesn't require many resources. Typically, your fibroblastic/stromal cells will be stably attached to the tissue culture substrate much faster than bladder cancer (epithelial) cells when freshly plated. So, you can try the following: after you trypsinize your culture, wash cells and plate them in fresh medium.Now, plate three flasks and allow 10 min, 20 and 30 min (should vary time depdning on how fast your cells attach) after plating, then recover the medium (and still floating cells) and transfer to a fresh tissue culture plates. With a couple of trials you should find a time spot in which you get rid of most mesenchymal cells and end up with a relatively pure cancer cell culture (make sure that you extensively characterize it (and repeat it periodically!) before you declare it is a bladder cancer cell line.
You can also try using Percoll gradients. Below is a method we used (https://bmcgastroenterol.biomedcentral.com/articles/10.1186/1471-230X-6-13) to separate cell types when isolating mouse billiary cells, but the principle is the same.
Isotonic Percoll solution (IPS) was made with 10X phosphate buffered saline (PBS) pH 7.4 and Percoll (Pharmacia, Uppsala, Sweden) (9:1 v/v). The IPS was diluted in DMEM containing 1% FBS. Pre-formed Percoll gradients were made by stepwise addition of 40, 50 and 60% IPS in the centrifuge tube. The cell suspension (1 × 10E7 cells/ml) was slowly placed on top of the gradient and centrifuged for 20 min at 1500 rpm in a refrigerated centrifuge at 22°C. Then, each fraction was collected, washed three times in BSS and plated in 60 mm tissue culture dishes coated with type I collagen film (may or may not be needed).