In vitro methods to predict for patient response to chemotherapy

Dose Issues in Cancer Chemotherapy

In this review, human methotrexate dosing regimens, as well as their relationship to data from in vitro cell culture and in vivo animal and human studies, are discussed. Low-dose, intermediate-dose, and high-dose therapies are covered. Since in vitro and in vivo screenings of potential cancer drugs are commonplace in the development of cancer chemotherapy, comparisons of the three criteria for effectiveness are important.

An Efficient Method for Culturing Human Breast Epithelium: Analysis of Results

We have examined the possibility of obtaining primary cultures from breast tissue utilizing a method especially developed for breast epithelium. The number of specimens able to grow in culture was very high: 82.8%, 64.3 %, 75.0 % and 77.8 %, respectively, for primary breast cancer, skin recurrences, inflammatory breast cancer and normal breast tissue. In our experience, growth was not related to menopausal status or histopathologic type, whereas for skin recurrences, a prior pharmacologic treatment (chemotherapy) of the patient enhanced the growth capacity of the tissue. This culture method could help to study the basic biology of breast epithelia and to improve the chemotherapy approach of breast cancer patients.

Multidrug resistance in ovarian cancer: comparing an immunocytochemical study and ATP-tumor chemosensitivity assay

The aim of our study was to evaluate the possible prognostic and predictive significance of the expression of P-glycoprotein, a transmembrane transport protein related to multidrug resistance, in previously untreated patients with FIGO stage III ovarian cancer; to compare the results of immunocytochemical analysis of tissue sections of tumors to the in vitro chemosensitivity to cytotoxic drug of fresh samples of the same tumors; and to evaluate survival in women who underwent the same surgical treatment and the same adjuvant chemotherapy.

Relationship between in vivo drug exposure of the tumor interstitium and inhibition of tumor cell growth in vitro: a study in breast cancer patients

A novel approach is described to simulate effect site pharmacodynamics of anticancer drugs. This approach is based on (i) the in vivo measurement of unbound, interstitial drug pharmacokinetics (PK) in solid tumor lesions in patients and (ii) a subsequent pharmacodynamic (PD) simulation of the time versus drug concentration profile in an in vitro setting. For this purpose, breast cancer cells (MCF-7) were exposed in vitro to the time versus interstitial tumor concentration profiles of 5-fluorouracil (5-FU) and methotrexate (MTX) from primary breast cancer lesions in patients. This led to a maximal reduction in the viable cell count of 69% on day 4, and of 71% on day 7 for 5-FU and MTX, respectively. This effect was dependent on the initial cell count and was characterized by a high interindividual variability. For 5-FU there was a significant correlation between the maximum antitumor effect and the intratumoral AUC (r = -0.82, p = 0.0005), whereas no correlation could be shown for MTX (r = 0.05, p = 0.88). We conclude, that the in-vivo-PK / in-vitro-PD model presented in this study may provide a rational approach for describing and predicting pharmacodynamics of cytotoxic drugs at the target site. Data derived from this approach support the concept that tumor penetration of 5-FU may be a response-limiting event, while the response to MTX may be determined by events beyond interstitial fluid kinetics.

Prediction of response to drug therapy of cancer. A review of in vitro assays

Cancer chemotherapy has witnessed a great deal of progress since the introduction of the nitrogen mustards in the 1940s. Unfortunately, individual patients with apparently identical tumour histologies do not always respond identically to the same drug regimen. Determining the sensitivity and resistance of an organism before treatment has been the standard of care in infectious diseases for many years, while in oncology treatment has been initiated according to tumour histology rather than the tumour's sensitivity to a given agent. Attempts to individualise therapy have been the goal of oncologists since the 1950s. Since that time a number of in vitro assays have been developed to predict therapeutic outcome prior to the start of therapy. In the 1970s, with the introduction of the human tumour stem cell assay, it was generally believed that oncology was on the threshold of entering an era of predictive in vitro chemosensitivity testing. Unfortunately, this assay was shown to have a number of technical drawbacks including the low plating efficiencies of many primary tumour samples which thus limits the percentage which can be evaluated, leaving us still at this threshold today. Several recent developments, such as the Kern assay, which measures inhibition of radioactive precursors into tumour cells in the presence of antineoplastic agents, ATP bioluminescence assays, and the fluorescent cytoprint assay offer the promise of rapid and sensitive results. Other assays, such as the tetrazolium-based MTT and the sulphorhodamine blue assay appear to hold more promise in the screening and evaluation of potential new agents in established tumour cell lines than for evaluating chemosensitivity of clinical specimens. However, before a particular assay can be considered as an in vitro test of chemosensitivity or resistance, controlled prospective studies must be carried out to validate the assay in a number of different tumour types.

Primary screening and inhibition of macromolecular biosynthesis in Ehrlich ascites cells by benzo(C)fluorene derivatives

The main objective of the present investigation was to screen a series of new benzo(c)fluorene compounds for in vitro activity. It can be stated that each of the 9 newly synthesized benzo(c)fluorene derivatives was about 10 times as active as tilorone. To elucidate the biochemical mode of action, the effects of 2 new compounds (13468 and 14200) on biosynthesis of macromolecules indicated by the incorporation rate of [14C]adenine (DNA, RNA), [14C]-thymidine (DNA), [14C]uridine (RNA) and [14C]valine (protein) were studied in concentration and time dependence. Both compounds inhibited the incorporation of the 4 precursors into the TCA-insoluble fraction of Ehrlich ascites carcinoma cells.

In vitro sensitivity assays in cancer: a review, analysis, and prognosis

Tumors are complex systems consisting of heterogeneous cancer cells as well as normal cells with each exhibiting unique drug sensitivity spectra. There have been many attempts to design in vitro systems to determine drug response to tumors. The most widely used system is the clonogenic assay. which has demonstrated some clinical predictivity. However, the clonogenic assay has been shown to have negative aspects, including low frequency of evaluation, clump artifacts, lack of cytotoxic end-points and disruption of normal cell-cell interactions existing in a true tissue environment. Newer models are described utilizing cytotoxic as well as cell-proliferation end-points, and maintenance of three-dimensional tissue architecture in vitro. It is concluded that less artifactual, more realistic models can be used to select more tumor-specific drugs which themselves in turn will make in vitro chemosensitivity assays more useful for cancer patients.

In vitro drug testing in patients with acute leukemia with incubations mimicking in vivo intracellular drug concentrations

The differential staining cytotoxicity (DiSC) assay was evaluated as a predictive test for response to therapy in patients with acute non-lymphoblastic leukemia. Incubations were designed in such a way that the intracellular concentrations of cytostatic drugs in vitro paralleled those in vivo. Leukemic cells were isolated from 53 patients with acute non-lymphocytic leukemia. 13 of these patients died early due to supportive care failure and were not evaluable for the predictive drug testing. Of the remaining 40 patients, 25 entered a complete remission (CR) and 15 had a resistant disease (RD). According to the patients randomization to therapy the cells were incubated with anthracyclines and Ara-C separately and in combination. After 4 days of culturing in liquid medium the in vitro cytotoxicity was determined by dye exclusion according to Weisenthal. The cytotoxic effect in vitro was significantly higher on cells from patients who achieved a CR compared to patients with RD after incubations with anthracyclines 0.2 mumol/l (p less than or equal to 0.005), Ara-C 0.5 mumol/l (p less than or equal to 0.05) and with the combination of anthracyclines with Ara-C (p less than or equal to 0.0005). The best predictive value was achieved when incubations with 0.2 mumol/l anthracyclines and 0.5 mumol/l Ara-C were analyzed together. With these incubations cells from 20 out of 21 patients who achieved CR showed either less than or equal to 60% surviving cells after the anthracycline incubation or less than or equal to 35% surviving cells after the Ara-C incubation. Cells from 11 out of 13 patients with RD did not fulfill either of these criteria. In vitro drug sensitivity was significantly correlated to a prolonged survival (p less than 0.01). We conclude that, when performed with incubations that mimic in vivo tumor cell exposure to cytostatic drugs, the DiSC assay shows a high correlation to clinical outcome for patients with acute non-lymphocytic leukemia.

Individual human tumors in short-term micro-organ cultures: chemosensitivity testing by fluorescent cytoprinting

Using innovative approaches, we addressed several problems often associated with in vitro chemosensitivity testing of individual human tumors: 1) obtaining a high rate of evaluability; 2) excluding participation of nonmalignant stromal and vascular components usually present in tumor specimens; 3) preserving cell-to-cell interactions present in the original tumor; 4) assessing drug-induced cytotoxicity without sacrificing the tumor culture. To circumvent these problems, tumor specimens were processed as follows: i) tissue (fresh or cryopreserved) was mechanically or enzymatically dissociated under mild conditions into cellular clusters (termed micro-organs); ii) large micro-organs were separated by a brief decantation, resuspended, and then exposed to fluorescein acetate to visualize (under naked eye) viable micro-organs; iii) fluorescent (i.e., viable) micro-organs were collected using a Pasteur pipette, and then planted on a solid support made of cellulose fibers impregnated with collagen. Since tumor micro-organs have been previously shown to consist solely of malignant cells, the procedure described here not only preserves a critical portion of the tumor architecture but eliminates at the onset necrotic tissue and nonmalignant cellular components that could interfere with the chemosensitivity testing. Drug-induced cytotoxicity was measured by "fluorescent cytoprinting", a novel, nondestructive procedure for assessing micro-organ viability in situ. The key feature of fluorescent cytoprinting is that cytotoxic effects are not measured against control cultures but against a baseline provided by a cytoprint of the same culture before drug addition. Using three experimental designs, we tested the potential of the method for clinical applications. The results using 469 human malignant tumors showed that the micro-organ culture assay can distinguish individual tumor chemosensitivity profiles with an overall success rate of 96%. For three commonly used chemotherapeutic drugs, the observed frequency of responding tumors was found to be comparable to previously reported clinical results using single agents.

Experimental chemotherapy and concepts related to the cell cycle

Scheduling of chemotherapy is limited by damage to normal tissues, and tolerated schedules are dependent on normal tissue recovery. Most anticancer drugs are more toxic to proliferating cells and the fall and recovery of granulocyte counts after chemotherapy may be explained by the effect of drugs on rapidly proliferating precursor cells in the bone marrow. It is argued that serious toxicity due to myelosuppression most often occurs because of damage to proliferating precursors that may be recognized in bone marrow rather than to stem cells. In contrast, therapy that is aimed at producing cure or long-term remission of tumours must be directed at killing tumour stem cells. The evidence that tumours contain a limited population of cells which can repopulate the tumour after treatment (and are therefore tumour stem cells) is reviewed critically. While there is quite strong evidence for a limited population of target cells, evidence from studies on metastases suggests that the tumour cells which may express this stem cell property may change with time. The stem cell concept has major implications for predictive assays. Although colony-forming assays appear to have a sound biological background for predicting tumour response, technical problems prevent them from being used routinely in patient management. Cells in tumours are known to be heterogeneous and at least three types of heterogeneity may influence tumour response to drug treatment: the development of subclones with differing properties including drug resistance; variation in cellular properties due to differentiation during clonal expansion; and variation in properties due to nutritional status and micro-anatomy. Heterogeneity in drug distribution within solid tumours may occur because of limited drug penetration from blood vessels, and nutrient-deprived cells in solid tumours may be expected to escape the toxicity of some anticancer drugs as well as being resistant to radiation because of hypoxia. This may occur both because nutrient-deprived cells have a low rate of cell proliferation, and also because of poor drug penetration to them. There is a need for improved understanding of the mechanisms that lead to cell death in tumours. If these mechanisms were understood, it might be possible to simulate them by therapeutic manoeuvres. Recent research from our laboratory suggests that the combination of low extracellular pH and hypoxia may be very toxic to cells in nutrient-deprived regions. Drugs which limit the cell's ability to survive in regions of acid pH may provide strategy for therapy of nutrient-deprived cells.

Semi-micro adaptation of a 4-day differential staining cytotoxicity (DiSC) assay for determining the in-vitro chemosensitivity of haematological malignancies

A semi-micro differential staining cytotoxicity (micro-DiSC) assay has been developed to determine the in-vitro chemosensitivity of haematological cancers. The method comprised isolation of leukocytes from blood or bone marrow, drug exposure and culture for 4 days in 1 ml tubes arranged in the microtitre format. Drug-induced tumour cell kill was determined by differential staining of live and dead cells, such that the former could be morphologically identified. Tumour cell viability was calculated by reference to an internal standard of fixed duck red blood cells. Up to 15 drugs at 5-6 concentrations each could be set up at a time in the assay within one hour of receipt of a sample, using only 10(7) viable cells. A result was obtained in 38 of 40 samples received. The assay is of potential use for the routine prediction of clinical response to cytotoxic drugs in haematological cancers and warrants wider investigation.

Advances in rational combination chemotherapy

Due to the high frequency of micro- or macrometastatic disease at the time of diagnosis of cancer, and to the increasing prevalence of cancer in this country, the use of chemotherapy to evoke cure or prolongation of survival has become critically important. In addition, the growth kinetics of large tumor burdens and the high probability of drug-resistant cells in a tumor mass at the time of diagnosis necessitate combinations of chemotherapeutic agents rather than single agents as the most effective mode of treatment. Since there are 40 to 50 active anticancer drugs now utilized, and since synergy between drug combinations is often dose and/or schedule dependent, the number of possible drug combinations and permutations is vast. Thus, screening for effective drug combinations requires a rational approach which will allow for accurate predictions of synergy. Most advances in this scientific approach have utilized biochemical modulation in conjunction with in vitro cytotoxicity assays, in particular, clonogenic assays. Such an approach has generated a number of drug combinations, such as sequential MTX-5FU, with widely applicable clinical efficacy. The continued use of biochemical modulation should rapidly generate new effective drug combinations which will, hopefully, allow us to cure even those cancers presently considered incurable.

Mullerian inhibiting substance reduction of colony growth of human gynecologic cancers in a stem cell assay

Mullerian Inhibiting Substance (MIS), a fetal testicular product that causes regression of the Mullerian duct in the male mammalian embryo, was evaluated for its antitumor effect on the premise that a substance active against this genital precursor in the fetus might also be active against tumors derived from these tissues. Increasingly pure fractions of biologically active MIS, prepared from newborn calf testes, were tested in the soft agar colony inhibition assay against single cell suspensions of fresh tumors derived in ascitic or solid form from patients with gynecologic malignancies. Twenty-eight tumor specimens placed in soft agar culture have provided sufficient growth to assess an MIS effect. Twenty-five of these 28 tumors showed significant colony inhibition after incubation with MIS. Increased antitumor response correlated with increased purification of MIS when the same tumor was treated with preparations of different purity. Samples obtained from the same patient at different times, from both ascites and solid tumor sources, produced nearly identical responses to MIS. MIS preparations, previously shown to be active in microcytotoxicity and colony inhibition assays against established human ovarian and endometrial carcinoma lines demonstrate consistent antitumor activity against fresh human gynecologic cancers removed at surgery.

Predictive testing in cancer chemotherapy. II. In vitro

During the last thirty years several in vitro techniques have been developed to predict sensitivity of individual tumours. When the results of these techniques were correlated with the clinical response in larger groups of patients, the accuracy for predicting resistance was greater than for predicting sensitivity. Amongst the culture techniques the colony-forming assays have received much attention. Research with tumour cell lines and the sound biological basis do support this preference on other techniques. Studies on these assays have come from several independent laboratories, who report comparable results. Improvement of the culture technique and more insight into the in vitro pharmacology is needed, before application on wider scale is justified. Colony-forming culture techniques have not only been propagated for individualized chemotherapy, but also for drug screening. New antitumour agents and analogous can be screened in a short time for their sensitivity in many histologic tumour types.

In vitro determination of tumour chemosensitivity in haematological malignancies

A four-day tumour chemosensitivity assay of potential use in predicting tumour response to cytotoxic drugs has been developed for haematological cancers. The method comprised isolation of white cells from peripheral blood or bone marrow aspirates, drug exposure and incubation for 4 days. Drug-induced tumour cell kill was assessed by differential staining of live and dead cells such that the former could be morphologically identified. Tumour cell viability was subsequently calculated by reference to an internal standard of fixed duck red blood cells. Over 30 drugs have been tested in vitro, all of which have shown a dose response relationship in vitro and given a good scatter of sensitivities from patient to patient within the concentration ranges tested. In 27 cases where the in vitro chemosensitivity could be compared with the in vivo response, there were 7 true positive comparisons (sensitive in vitro and in vivo), 17 true negative comparisons (resistant both in vitro and in vivo) and 3 false positive comparisons (sensitive in vitro, resistant in vivo). A result was obtained in 38 of 50 samples received, comprising 16 of 18 chronic lymphocytic leukaemias, 11 of 20 acute lymphoblastic leukaemias, 5 of 5 acute myeloid leukaemias and 6 of 7 myelomas. The assay appears to show considerable promise as a tumour chemosensitivity test and warrants wider investigations.

Prediction of the response to chemotherapy in acute leukemia by a short-term test in vitro

To detect sensitivity or resistance of leukemic cells to chemotherapy prior to treatment, a short-term incubation method was employed. Blast cells from the peripheral blood or bone marrow of adult patients presenting with different forms of acute leukemia were analyzed for in vitro responsiveness to anticancer drugs in terms of suppression of 3H-uridine incorporation into cellular nucleic acids. The following agents were tested over a wide range of concentrations: Adriamycin, cytosine arabinoside, thioguanine, 6-mercaptopurine, prednisone, and 4-hydroperoxycyclophosphamide. Retrospectively, the in vitro data were compared to the clinical response of the patients to polychemotherapy. In the majority of the patients, in vitro cytotoxic effectiveness of Adriamycin (doxorubicin) and cytosine arabinoside reflected the in vivo situation. The levels of in vitro inhibition that could distinguish between drug-sensitive and drug-resistant diseases appeared to be 30% for Adriamycin and 20% for cytosine arabinoside. No correlation was found between the Adriamycin effect in vitro and the proliferative state (rate of 3H-thymidine incorporation) of the leukemic cell population. Serial in vitro sensitivity testing during the course of disease of various patients proved the ability of the test system to detect acquired resistance to chemotherapeutic agents. Therefore, the assay might serve as a reliable tool in the selection of effective chemotherapy in individual patients suffering from acute leukemia.

Short-term in vitro sensitivity testing in acute leukemia

To detect sensitivity or resistance of leukemic cells to chemotherapy prior to treatment, a short-term incubation method was employed. Blast cells from the peripheral blood or bone marrow of adult patients with different forms of acute leukemia were analyzed for in vitro responsiveness to cytostatic agents in terms of suppression of nucleoside precursor (3H-uridine, 3H-deoxyuridine) incorporation into the cells. Retrospectively, the in vitro data were compared to the clinical response of the patients to polychemotherapy. In the majority of patients, in vitro cytotoxic effectiveness of doxorubicin and ara-C reflected the in vivo situation. The levels of in vitro inhibition that could distinguish between drug-sensitive and drug-resistant diseases appeared to be 30% for doxorubicin and 20% for ara-C. No correlation was found between the doxorubicin effect in vitro and the proliferative state of the leukemic cell populations. Serial in vitro testing during the course of the disease of various patients proved the ability of the test system to detect acquired resistance to chemotherapeutic agents. Studies with established cell lines (HL-60, Raji) indicated that the short-term test in its present form is probably not suited to monitor the effect of drug combinations in vitro.

Prediction of response to cancer chemotherapy

Cytotoxic chemotherapy plays a key role in the treatment of carcinoma for thousands of patients annually who either present with metastatic disease or relapse after surgical excision of apparently localised disease. Unfortunately, there is such a wide range of responsiveness to drug therapy within individual tumour types that response of an individual patient's tumour to cytotoxic therapy cannot be accurately predicted. Intensive efforts to increase the accuracy of selection of effective chemotherapy have recently culminated in an in vitro system which employs soft agar and standard laboratory tissue culture techniques to predict drug sensitivity and resistance for an individual patient's tumour with reasonable accuracy. Research in this system is being actively pursued at several centres and further modifications and refinements may well make cancer chemotherapy more precise than previously possible. This review surveys methods of studying in vitro drug sensitivity which have been tested and for which clinical correlations are available. The technique and results of the more recently developed human tumour stem cell assay and the potential applications of this system are also discussed.

Clinical correlates of in vitro drug sensitivities of ovarian cancer cells

Of 89 samples of cancer cells from ovarian cancer patients primary cultures representative of the cancer cell population could be established in 17. The clinical response to polychemotherapy was studied in relation to the inhibition of thymidine uptake by the cultured cells. Cultures of each patient's tumour were exposed to concentrations of the drugs the patients had been given for long enough to reproduce the area under the curve (AUC) of the plasma levels resulting from in vivo dosage. Full agreement was observed between the degree of thymidine uptake inhibition induced by at least one of the drugs administered to the cultured cells and the degree of clinical response. This approach may prove useful in pharmacological studies as a means of obtaining ovarian cancer cell populations representative of human tumours, even though the number of tumours that can be successfully evaluated in vitro is still too small to serve as a sound basis for prediction.

An assessment of a short-term tumour chemosensitivity assay in chronic lymphocytic leukaemia

A 4-day tumour sensitivity assay of potential use in predicting tumour response to cytotoxic drugs has been investigated in patients with chronic lymphocytic leukaemia. The method comprised isolation of white cells from peripheral blood, drug exposure and incubation for 4 days. Drug-induced tumour cell kill was assessed by differential staining of dead and live cells such that the latter could be morphologically identified, with subsequent calculation of tumour cell viability. Concentrations of drug for use in the assay were chosen for chlorambucil (2 micrograms ml-1), 4-hydroperoxy-cyclophosphamide (2 micrograms ml-1)--which was used in vitro in place of cyclophosphamide--prednisolone (0.5 microgram ml-1) and vincristine (0.1 microgram ml-1), to give a scatter of values which was in good agreement with clinical expectations. In 21 cases where the in vitro result could be compared with the in vivo response, there were 4 true positive comparisons (sensitive in vitro, sensitive in vivo), 15 true negative comparisons (resistant both in vitro and in vivo) and 2 false positive comparisons (sensitive in vitro, resistant in vivo). A result was obtained in 86% (65/76) of samples received. The assay appears to show considerable promise as a tumour chemosensitivity test and warrants wider investigation, including prospective in vivo/in vitro correlations that could be based on the results presented here.

A logistic model based on enzyme activities for the prediction of response of breast cancer patients to chemotherapy

A logistic regression model, utilizing the activities of certain selected glycolytic enzymes and ER status measured on primary or recurrent lesions, has been applied to predict for response to combination chemotherapy regimens administered to women with advanced breast cancer. The clinical outcome of response or no response was evaluated retrospectively using criteria employed by cooperative group protocols. In 93 cases, 58/61 patients classified as nonresponders and 22/32 patients demonstrating objective responses would have been correctly designated, based on the 50% estimated probability as the level for separation of responders from nonresponders. The overall predictive accuracy of this model was 86%, with apparently greater accuracy for prediction of lack of response. Addition of estrogen receptor status to the model imparted no gain in accuracy of prediction. Application of this model to a prospective study is warranted.

Experimental systems for analysis of the malignant phenotype

Identification of the cellular and subcellular alterations responsible for the metastatic behavior of malignant tumor cells and development of reliable screening programs for detecting new therapeutic agents for improved treatment of metastatic disease both depend crucially on the availability of experimental systems that can serve as relevant models of human cancer. Recent advances in our understanding of the pathogenesis of cancer metastasis have raised serious doubts about the usefulness of many of the experimental approaches that have long been used in the study of metastasis. Recent findings showing that metastases are caused by specific subpopulations of metastatic tumor cells, and that not all cells in a malignant primary tumor possess metastatic properties, are of profound importance for experimental efforts to understand the mechanism of metastatic phenotype among cells from the same tumor means that the traditional, and widely used, approach of analyzing primary tumors and cultured cell lines containing multiple, phenotypically heterogeneous, subpopulations of cells may provide little or no insight into the properties of the metastatic subpopulations, particularly if they represent only a minor fraction of the entire population. Similarly, the practice of screening potential therapeutic modalities for their ability to reduce the mass and/or growth rate of a primary tumor may be inadequate in predicting the responsiveness of metastatic lesions. Solution of these problems requires that new methods must be devised to isolate and characterize the specific subpopulations of tumor cells endowed with metastatic potential. In addition, knowledge of how the extraordinary phenotypic diversity found in tumor cell subpopulations from the same tumor is generated and how subpopulation diversity is regulated during progressive growth of both the primary tumor and its metastases are of fundamental importance if we are to design meaningful experimental systems for studying the metastatic process. This article reviews our current understanding of these complex issues and their implications for the experimental analysis of the malignant phenotype. The merits and shortcomings of different experimental systems are discussed in detail together with the identification of areas in which new experimental strategies and models are now needed.'