Category Archives: Editorial regarding previously-published work

Unexpected results with combining “targeted” drugs in colorectal cancer

Two recent clinical trials showed that empiric combinations of bevacizumab and ant-EGFR antibodies, combined with chemotherapy, did not improve therapeutic outcomes in colorectal cancer  – in fact, outcomes were inferior to  single “targeted” therapy combined with chemotherapy.

Hecht JR, Mitchell E, Chidiac T, et al. A randomized phase IIIB trial of chemotherapy, bevacizumab, and panitumumab compared with chemotherapy and bevacizumab alone for metastatic colorectal cancer. J Clin Oncol 2009;27:672-680

Tol J, Koopman M, Cats A, et al. Chemotherapy, bevacizumab, and cetuximab in metastatic colorectal cancer. N Engl J Med 2009;360:563-572

In his video blog, Dr. Maurie Markman of MD Anderson editorializes that all new concepts must be proven in prospective, randomized clinical trials:

http://boards.medscape.com/forums/.29f0fd39 (may require free registration to view)

In the late 1980s, the NCI, aided and abetted by herd mentality study sections, effectively closed down research into fresh human tumor cell culture methods for testing and optimizing chemotherapy. The proof of this is the complete lack of NIH-funded studies relating to this topic appearing in PubMed for the last 15 years. Instead, we have put all of our clinical trials resources into trying to identify the best treatment for the average patient — in a disease notorious for heterogeneity. Drug screening (including therapy screening) belongs in the laboratory, not in the clinic. All of a sudden, there is a belated recognition that “personalized” therapy is worthy goal — yet 100% of the effort is going into static profiling of molecular markers, as opposed to dynamic, functional profiling of tumor response ex vivo. It’s crazy/nuts, and, down the road, tomorrow’s translational researchers will shake their heads and say “what on earth were they thinking?”

How to make progress in combined “targeted” therapy?

Here’s an example: http://tinyurl.com/weisenthal-breast-lapatinib But there’s absolutely no support for work such as this. And now there are efforts under way to make it impossible to do this work in grass roots, private sector laboratories, for example: http://www.cancertest.org/?cat=11

One relevant irony is that, in my opinion, few individuals have done more to extinguish support for the development of cell culture methodologies in clinical cancer testing than the afore-mentioned Dr. Markman.

I intend on discussing this further in the ongoing consideration of precisely what type of evidence should be required to “validate” a clinical cancer test.

- Larry Weisenthal/Huntington Beach, CA, USA

Tumor response by depth of invasion: a probable artifact

A very misleading paper was published online very recently in the European Journal of Surgical Oncology by Y.B. Cho and co-authors, from the Samsung Medical Center and the Isu Abxis Company in Seoul, South Korea.

There are a number of generally misleading and/or incorrect statements in the article (some noted below), but the most serious is the implication that there are intrinsic differences between the respective drug resistances of colorectal cancer specimens obtained from different locations in the same tumor. I believe that their findings most likely relate to laboratory artifacts, as opposed to intrinsic differences in tumor biology.

Cho and colleagues isolate tumor cells from superficial (mucosa and submucosa) and deep (muscle, subserosa, serosa) regions of surgical specimens from primary colorectal tumors and compare the percent cell death induced in (1) superficial and (2) deep regions by four different drugs (5FU, irinotecan, oxaliplatin, and mitomycin c), using the ATP endpoint.

They report significant differences between drug activities between superficial and deep regions of the same tumor and they provide an elaborate explanation of why this may represent true biological heterogeneity in the different regions of the same tumor.

Their work may be criticized on many levels — for example, they present no data regarding the consistency of results from two different “superficial” and two different “deep” regions of the same tumor.  It is not clear that the variability they see is relating to true differences between tumor regions or simply variability in results between segments of the tumor processed individually.

An even more serious concern is that the differences they report may relate to such basic issues as differences between tumor viability, tumor three dimensionality in culture, and/or differences in the quantity of “contaminating” normal (non-tumor) cells present at the end of the culture period (normal cells being, most prominently, macrophages, lymphocytes, connective tissue cells, and normal intestinal epithelial and endothelial cells).

With regard to the latter point, the authors only use a 2 day drug incubation before testing the ATP levels. We and others prefer a longer duration of cell culture time, in part because this allows for more selective death of normal cells relative to tumor cells by the end of the culture period and in part because the peak signal for apoptotic caspase expression may range between 36 and 66 hours following exposure to different drugs in different specimens.

To address first the issue of the effect of normal cells on assay results, we have compared the results between the DISC and MTT endpoints performed simultaneously on the same specimens, in thousands of tumors.  The DISC endpoint is relatively specific for tumor cells, while the MTT endpoint, like the ATP endpoint, is a general metabolic signal which is generated by both normal cells and tumor cells.

Shown below is the correlation coefficient (r squared) as a function of the percentage of cells which are tumor cells, measured in control cultures at the end of the incubation period (which is when the assay measurements are made).

picture-31

Note that there is a good correlation when the percentage of cells which are tumor cells at the end of the culture is greater than 70% and that most solid specimens do have greater than 70% tumor cells, following 96 hours of anchorage-independent culture, when extensive procedures are employed to initially “purify” tumor cell clusters, as we utilize in our laboratory.  Cho and co-workers apparently did not have quality controls in place to measure the cell composition post-culture, at the time the cultures were tested for ATP content.

A second issue has to do with the fact that the results of cell culture assays are often profoundly affected by (1) cell numbers present in culture (“plating density”), (2) metabolic “robustness” of the tumor cells, and (3) degree of tumor three dimensionality in culture .

The chart below shows the relationship between the activity of the same four drugs tested by Cho and co-workers (5FU, irinotecan, oxaliplatin, mitomycin c) and two measurable cell culture parameters:

(1) the magnitude of the MTT formazan signal in control cultures (which is an index of both viable cell number at the end of the culture and also of the metabolic “robustness” of the tumor cells post-culture (where “healthy” tumor cells will produce a greater signal than will “sick” tumor cells — again, these measurements are made in the control cultures)).

(2) the degree of tumor three dimensionality, as measured both prior to culture and at the end of culture.  We determine, for each specimen, a “cluster index,” comprised of measurements of (a) percentage of total tumor cells in three-dimensional clusters, at both the beginning and end of cultures, (b) average size (occular micrometer units) of tumor clusters, both pre and post culture, and (c) average density (“loose,”, “medium,” and “tight”) of the tumor clusters, both pre and post culture.

Shown below is the relationship (two sided P value, Fisher’s exact test, performed in hundreds of colon cancer specimens per P value) between control MTT formazan signal and drug activity and also between “three dimensional cluster index” and drug activity.

picture-7As a general rule, with many drugs, there is a significant relationship between the drug activity, as measured in culture, and the metabolic “robustness” on one hand (more metabolically “robust” specimens, as determined by strength of MTT formazan signal in control cultures, tending to be more drug resistant) and tumor three dimensionality on the other hand (more three dimensionally clustered specimens also tending to be more drug resistant).

Because of the varying effects of these parameters (metabolic “robustness” and tumor three dimensionality), it is important to make “apples to apples” comparisons, as opposed to making “apples to oranges” comparisons.  For example, one should be wary about comparing the results obtained in a metabolically weak, relatively discohesive cell culture with a database derived from predominately metabolically strong, largely three dimensional cell cultures.

With regard to the Cho study, it is not at all clear that their observed results reflect true tumor heterogeneity of drug resistance, as opposed to simply reflecting different degrees of metabolic robustness and/or three dimensionality between cells obtained from different regions of the same tumor.

Cho and colleagues also make additional statements which are misleading and harmful to the general field of individualized tumor response testing (cell culture drug response testing) in human tumors.

For example, they state:

chemosensitivity testing is not commonly used to evaluate the tumor response prior to treatment, mainly because of the low reliability, low evaluability rates, high cost, and poor correlation between the assay results and the clinical response.12,13 The low reliability of these conventional in vitro assay systems can be largely attributed to contamination by non-malignant cells, such as fibroblasts and lymphocytes.14,15 This situation has changed with the introduction of an ATP-based chemosensitivity test.16,17

These statements are extremely misleading. In the first place, their references number 12 and 13 pertain to the old “human tumor clonogenic” assay, which was abandoned more than 15 years ago, as a clinical test, used to assist in drug selection. Secondly, while fibroblast and lymphocytes contamination (as well as normal epithelial and macrophage contamination) can be a problem in many assay systems, including the two day ATP system used by the authors, these were not prominent problems in the human tumor clonogenic assay systems used in the authors’ cited references (12,13).  And there is nothing at all with the ATP endpoint which “changes” the situations described by the authors.

The ATP endpoint is simply a cell death endpoint (used also as one of the endpoints in my own laboratory) which gives very similar results to those obtained with other cell death endpoints, when applied correctly to relatively “pure” tumor cell cultures.

There are a wide variety of cell death endpoints, each of which has specific advantages and disadvantages for different drugs and in different tumors. In point of fact, there are vastly more data to support the MTT endpoint for application in gastrointestinal neoplasms than there are data to support the ATP endpoint, although both endpoints are usefully valid for many drugs, when applied properly.

I do feel that the MTT endpoint may be uniquely more valuable (and more “accurate”) for testing fluoropyrimidines than are other cell death endpoints, however. I’ll discuss the reasons for this in a subsequent post.

- Larry Weisenthal/Huntington Beach, CA