Dear Jabdurai Jayapaul,

Thanks for your reply.

We are innoculating 8-10 week old male SCID mice with 4 million LNCap cell sin matrigel. In initial experiments palpable tumors were obtained after about 3 weeks. In a recent experiment we saw no growth for about 2 months then rapid growth, but on excision, this was mostly inflammatory tissue with only a small amount of actual tumor. We have also innoculation m-Cherry-transfected LNCap cells but these also did not produce consistent tumors. Do you have experience with modulating testosterone levels with hormone in order to improve tumor growth?

Michael

The trypsinization of the cells is critical. If you over trypsinize the cells the tumor induction frequency and time to palpable tumor is suboptimal. Washing with CMF-HBSS prior to trypsin/EDTA addition is important and minimal time in trypsin and its rapid neutralization can be helpful.

LNCaP cell behavior varies greatly by source since culture conditions and handling influence the cell behavior greatly. Leland Chung developed a medium called T-medium (available form GIBCO) that makes the cells easier to grow and handle in vtiro. For tumor growth to be optimum in my hands, several hundred animals worth, the cells need to be almost confluent before use. We use cells up to passage 40 and usually not after. I always feed the cells 24 hours before use and inject in an equal volume of standard concentration matrigel. Growth factor reduced or regular matrigel does not matter too much. I use 2 to 2.5 million per site, subcutis at 4 sites/mouse in balb/c nu/nu mice or SCID beige NC17. They grow faster in SCID and have a higher tumor take rate. We routinely get >50% of injection sites growing and about 67% in SCIDs. We have used osteosarcoma cells and Neil Bhowmick and others have used fibroblasts to accelerate LNCaP tumor growth. The growth between injection sites is variable and some sites will grow very fast, but the first measureable tumors show up in by three weeks in nudes. Our mice are intact males, 20 g or more, at least 30 days old. Tumor take goes down as these animals age.

For a cell line which is widely used, LNCaP cells can be remarkably variable. One way to reduce variability, as has been alluded to here, is to use castrated animals supplemented with T or another androgen (even synthetic versions).

Might want to check out this recent reference:

http://onlinelibrary.wiley.com/doi/10.1002/pros.22520/abstract

Describes use of a hemostatic gelatin matrix specifically for LNCaP cell tumor growth in mice.

We have made several attempts to optimise our sc LNCap model in SCID beige mice i.e. using cells 100 days. Depsite all these efforts, , there were huge variations (slow, medium and aggressive tumor development). Perhaps this somehow reflects the clinical situation?

Thanks everyone for your comments. I've not be able to get back to this site until now.

To James: We use a standard trypsinization protocol that odes not affect cell viability nor subsequent cell growth kinetics. Will look at the CMF-HBSS wash though.

To Jabadurai: Generally we do not work with cells after bout the 20th passage. In our initial clibration experiments, we did not see tumors if cells were not innoculated together with matrigel.

To Robert: Thanks for all the details, many of which seem similar to what we are doing. We did not see any growth at all in nu/nu mice. Will look into the T-medium from Gibco.

To Kristen: We have also seen the different pathology of these tumors as I mentioned. Can you say something of the models you are now using?

To Nick: Thanks for the references - will follow it up.

To Jianying: Your experience reflects our own to a lrage extent. Maybe you are correct that this variablity has some clinical relevance but it makes it a very expersive and problematic model to use during drug development!

I am agreeing with Kristen Arnold. Tumor model with LNCaP and C42B cell lines are filled with vasculature/bloody. Exact mechanis not clear. This is may be due to presence of Andorgen Receptor, even though C42B cell line not depending on testasterone for its invitro growth. I tried orthotopic models of above said cell lines and gave up because of inconsistancy and very high vasculature. It depends which kind of study you want to do. If you are not interested in andorgen receptor status... not using any ant-angiogenic drugs, move to PC-3 model. It works very well.

The comment about PC-3 as reliable is interesting. This is in fact why most people use them. However, they reflect less than 5% of prostate cancers since they express no or extremely low levels of androgen receptor and are osteolytic. More than 95% of PCa is osteosclerotic with a net blastic phentoype. For modeling tumor metastasis and bone disease they do not accurately reflect the human disease state. There are bone metastatic variants of PC-3, namely PC-3M, developed by Mark Stearns that will at least give you a pair of cells to compare. Many people actually find LNCaP and the variants to be easier to culture once grown in T-medium as described by Dr. Chung. Of course, LNCaP has only full length AR unlike Cwr22Rv1 and does not have a TMPRSS/ERG rearrangement. The vasculature of many tumors is highly disorganzed and leaky.

It may depend on the time it takes you to collect and inject your cells, and how many injections you're doing at once. Good viability does not mean that the cells will form tumours. In fact I tested the viability of my cells after sitting in a syringe for 1 hour - all cells were viable, but the tumour would grow extremely poorly. In my experience it is best to do 3 injections at once, up to 5 if you're really fast. The longer the cells lie around in the syringe the lower your likelihood of having great tumours. I always resuspended my tumour cells (typically 5e^6) in blank medium and injected 100 ul. It takes far more time to do it this way, but I had 95-100% tumour take with tumours that were comparable.