I agree with Robert that your question needs details but when considering mouse models you have to keep in mind that there are at least three major types: genetically-engineered mouse models (GEMMs), patient-derived xenograft models (PDXs), and cell line xenograft models.  Like all experimental platforms, they all have pros and cons.  The GEMMs are nice because you can study pathogenesis and drug response in an intact animal (microenvironment, intact immune system, etc.).  The PDXs are nice because they are human tumor cells growing in a mouse.  The hope is that PDXs will be more predictive than cell line xenograft models which have been the mainstay for decades.  I hope this helps.

There are many tumor models - and as a model they should help to understand some aspects of tumorigenesis, progression, metastasis - or treatment and prevention, but as you know only to some extend. I think your question is not specific enough to get a more detailed answer. Many cancer biologists say: "There are many ways to cure cancer, as long as you are a mouse." Some of the very promising candidate drugs may have immunological site effects in humans... One should evaluate the right concentration to start in human, i.e. phase I study. One should not start with 4 people at the same time when there is a chance of severe immune reaction (as a couple of years ago).

I agree with Robert that your question needs details but when considering mouse models you have to keep in mind that there are at least three major types: genetically-engineered mouse models (GEMMs), patient-derived xenograft models (PDXs), and cell line xenograft models.  Like all experimental platforms, they all have pros and cons.  The GEMMs are nice because you can study pathogenesis and drug response in an intact animal (microenvironment, intact immune system, etc.).  The PDXs are nice because they are human tumor cells growing in a mouse.  The hope is that PDXs will be more predictive than cell line xenograft models which have been the mainstay for decades.  I hope this helps.

Really a perfect answer by Ronny! I would like to add to my earlier comment, that preclinical models are necessary, but may need more improvement, since most of the really big money is lost in clinical studies phase I, II and III - and most preclinical "cures" later fail in clinical tests. Many biologists and journalists overinterprete animal experiments, like over 20 years ago, when CD44 splice variants were thought to be the major cause for metastasis - and human therapy would be just a few years away. There were not only Cell, Science papers at that time, but even title pages in the yellow press (german BILD, like german researchers solved the cancer problem) and serious tv news (german Tagesschau) and news agencies beleived in those animal experiments...

Therefore, it is quite important to "fail early" just more often, i.e. preferentially already in preclinical tests in animals, to focus better on those lead substances which can make it through phase III. Therefore, it is good to have some substances and animal models, but one should not be too surprised if they later fail in clinical models. I think for glioblastoma (a malignant brain tumor, not very common among all cancers), there are a few hundred of clinical studies - I wonder why those tumors should be so much more important and better financed than many of the other tumors for which there might be more success with the same amount of money.

There is always an animal model for study of any disease but there is no animal model for every disease.....as said earlier that there are many different animal model but it is very difficult to 

It is difficult to add something since most relevant aspects have already been answered. As a clinical oncologist with a  past experience in preclinicals tudies let me just add that I think that careful evaluation of animal models (already listed in previous comments) would probably lower the "clinical failure" of drugs coming into phase I-II trials.

I am convinced that, for example,animal models could demonstrate that the activity of EGFR inhibitors (gefitinib,e rlitinib) is limited to tumours carrying a specific mutation before clinical evidence.

Xenotransplants in nude mice are certainly useful, and so called "avatars" (studying the tumour of the specific patient implanted in a nude mouse in order to obtain indications on how to treat this patient) is today very fashionable. Still I think we can obtain many valuable informations from "old fashioned" cheap and practical murine tumours growing in immunocompetent mice.

Thanks everyone for your answers! In speaking to the mouse models used, is it correct to say that it is best to have 1) in-vitro cell line data, 2) PDX model before any thought of advancing to clinical trials? As a follow-up, there are so many different cell lines available, how do we choose the right one, especially given the tumor heterogeneity that exists in human tumors. 

In vitro cell line data are the easiest to obtain, and you can perform fast and cheap tests to verify any hypothesis. Once you have an idea try different cell lines to see if it holds, then verify it in animals (if possible in different species [mice, rats, etc] or at least in different strains). Collect pharmacodynamic data on drug kinetics, anabolism, catabolism, drug concentration in blood cells and in the tumour. See in what conditions your drug works, but you can learn a  lot by analysing conditions where your drug does not work.

I think this will let you lay a sound basis for any further clinical evaluation.

Lijing - that is a very good question.  In my field, ovarian cancer, there were a couple of recent papers showing that the commonly used ovarian cancer cell lines (typically available from ATCC) do not recapitulate the genomic complexity of human ovarian cancers (as per the TCGA ans similar efforts).  The paper by Domcke et al. (Nature Commun. 2013) did, however, identify a set of cell lines that are better models of high-grade serous ovarian cancer (the most common type) and people are now starting to use those more often.   I imagine that other diseases have similar issues.

From a mouse model with a mechanism to treat a specific subtype of cancer to "advancing to clinical trials", there are many more requirements to be fulfilled than just an idea of a working mechanism: usually (I just expect it to be similar in any country) there should be a new drug application (NDA), which till take several months to be approved by the federal drug agency (or similar in most countries), before any new substance would be allowed to be tested in humans. I beleive (i.e. I am not an expert in this) that many tests in vitro and in vivo (animals, like rodents, monkeys) need to be done first to check toxicity (genotoxicity, metabolism) - and to get an idea at which concentration the initial phase I study should be performed without a too high risk of severe site effects (but one never knows, especially with the most beautiful immunological weapons). Later on, in phase II and IIb there could be a good idea of the best dose for phase III (if not stopped due to major site effects, or lack of effect).

Concerning cell lines: they usually (my estimated guess, almost 100%) are different from any "normal" tumor developing "naturally", especially mouse and rat tumors may have the identical names as human tumors, but can be quite different. I once developed a rat brain tumor model for primitive neuroectoderman tumors (PNET, like a medulloblastoma), and to the surprise of my two bosses, the tumors looked indistinguishable from a human tumor (wiht all characteristic features of the human counterpart); I was told this is the exception, not the rule. Even the wonderful transplantations of human tissue in some mice will, unfortunately, not fully ressemble a real human tumor, since a human tumor does not only consist of tumor cells, but NON-tumor cells. After transplanting and re-transplantation of a human tumor into animals, those non-tumor cells may not be of human origin any more, only the human tumor cells...

As mentioned in the beginning, each model has to have its limitations.

Hi all, 

these are all great answers and i am certainly less of an expert than  you guys are. I would also think that cell lines are a bit far from what is happening at the real tumor site and they can oly be used for rough estimations. On the other hand more "in vivo" models may be somehow closer to real thing. My assumption is that there are no perfect models out there yet. Some are better but not perfect. 

I personally like 3d model cultures as well, they give a kind of a 3D real life feeling but again they have drawbacks. 

I think that only human trials can give you the real picture on the effectiveness of drugs and that there should be great precautions (all tests done on e.g. toxicity etc, that you all mentioned) before applied to humans.

I just also want to note that i am somehow convinced that boosting the immune system somehow will be a major part of fighting cancer but we do have to find the effective ways to do it.

Nikos

Any tumor model needs to be considered in relation to its objective: Is it going to be used for scientific research for better understanding of the etiology of the cancer of interest? Or is it used for drug screening?

For research purposes, all models have their value. Once we fully understand the molecular mechanism behind one type of cancer, we can start thinking of a screening model for new drugs. Unfortunately, this has not been common practice, hence the high failure rate in clinical trials.

Great answers by all..!!

i want to add " using of orthotopic models with careful evaluation may gives good results for clinical success" 

Thanks. 

I think tumor model   for preclinical studies should be used in larger animals like rabbits or dogs or pigs better than mices.Because these larger  animals are easy to

examined by CT\MRI .And the images are better.

Is it worth thinking through whether a particular model might not adequately represent the key mechanisms that are intended to be tested?  In particular:

1.  Are there specific considerations for the type of treatment planned?  E.g., for immune response suppression checkpoint inhibitors like anti-PD-1/anti-PD-1 drugs, would one need to know the animal model has an intact immune system (but doesn't that mean xenografts would find it difficult to survive in the first place)???

2.  Re: predictive value, that depends on what you are testing.  Some drugs targeted to a driving mutation work very well with high odds of response and very high odds of either response or stability for a while (e.g., crizotinib for ROS1+ driven cancer), but others don't work very well due to various mechanisms of resistance such as emergence of mutational variants of the driving mutation, a driving mutation of a different gene, or up-regulation of alternate pathways.  I don't know much about the subject, but I'd wonder if it could be difficult to expect good predictive potential where transplanting cancer cells might disrupt the up-regulation feedback loops that caused that resistance bypass mechanism.

I like very much the response of Nikos, and I fully share it.

Here attached are some articles that actually merit to be read "in depth".

Best regards

Robert

PS: a model remains a model ...

What do you think are the advantages of 2D models: cell lines? Most people still use these for research!

There are two "advantages" for 2D models, which in fact are not at all advantages. Low costs and easy to use. But totally unrepresentative of what happens in a cancer in a patient.

It is not because "many people are doing something" that this thing is "good".

And most of the people using 2D models while claiming that they obtained "prominsing anticancer effects" should question themselves about the absence of actually active new (since 20 years) compounds against metastatic cancers, which still kill 90% of the cancer patients.

Best regards

Robert

Human xenograft models are essential to modern day drug development. There are a lot (see our collection here: http://altogenlabs.com/xenograft-models/) and they have been used for decades to determine possible side effects and outcomes from given drugs/chemotherapeutics. There are plenty of factors to consider when translating xenograft results to clinical trials, but usually if you have a human cell line xenografted in an immunodeficient mouse, the results from drug treatment will be sufficiently reliable for translation into human trials.