Neoplastic Therapy in Nude Mice Xenografted with lacZ Transduced Human Tumor Cells
Top of pageAbstractGenetic labeling of tumor cells with the Escherichia coli lacZ reporter gene, encoding the enzyme galactosidase, is widely used for histochemical detection of micrometastases in mice. Recently, we have developed a novel, highly sensitive and specific immunocapture chemiluminescence assay for the quantitation of E. coli galactosidase. This assay achieved a detection limit of 0.01 mU of E. coli galactosidase per milliliter, and 97% signal recovery of purified enzyme added to mouse plasma. LacZ transduced MDA MB 231 BAG human breast cancer cells grown in vitro released soluble galactosidase into the culture medium, and the concentration found correlated with cell density. Growth of the same cells in nude mice produced readily measurable levels of E. coli galactosidase enzyme activity in host plasma and a highly significant correlation could be demonstrated between the size of primary tumor xenografts and the host plasma level of E. coli galactosidase activity. When mice bearing MDA MB 231 BAG tumor xenografts were treated intravenously with a single injection of doxorubicin (5 mg/kg), the mean tumor volume after 16 days was reduced 4 fold in the group of doxorubicin treated mice compared with saline treated control mice, and the mean level of plasma E. coli galactosidase was correspondingly reduced 3.8 fold in the doxorubicin treated mice compared with control mice. Sensitive and specific measurement of soluble E. coli galactosidase in blood, using an immunocapture chemiluminescence assay, thus provides objective assessment of tumor burden in mice xenografted with lacZ transduced human tumors. This assay may have important applications as a tool for determining the efficacy of new experimental anti tumor agents.Top of pageIntroductionMethods for objective quantitation of tumor burden in animals are desirable for experimental tumor studies, including the evaluation of anti tumor therapy. Tumor burden in mice with subcutaneously growing tumors is usually measured bidimensionally using calipers, but for several orthotopically grown tumors, such as ovarian cancer (Will et al, 1975), pleural cancer (Astoul et al, 1994), ascites tumors (Low et al, 1996), or colon cancer (Pocard et al, 1996), as well as metastatic tumors, this method is not feasible.Transfection of tumor cell lines with the Escherichia coli lacZ reporter gene, which codes for galactosidase, has often been used as a histochemical marker for detection and quantitation of the disseminated tumor cells (Lin et al, 1990). Thus, galactosidase activity can be detected by staining with the chromogenic substrate 5 bromo 4 chloro 3 indolyl d galactopyranoside (X gal), which results in deposition of an insoluble dark blue reaction product (Br et al, 1992). Recently, we have developed a novel, highly sensitive and specific immunocapture chemiluminescence assay for the quantitation of E. coli galactosidase activity in organ extracts of nude mice with lacZ transduced human tumor xenografts (Holst Hansen et al, 1998).In this study, we show that cancer cells, transduced with the E. coli lacZ reporter gene, release galactosidase into culture media in vitro. When grown in vivo in nude mice these cells release galactosidase into the host circulation. galactosidase activity in host plasma was strongly correlated with tumor volume, and based on an experiment with doxorubicin we propose that this assay can be used to monitor tumor burden in response to experimental anti tumor therapy.Top of pageResultsRelease of Galactosidase Activity from lacZ Transduced Tumor Cells into Culture MediumFor measurements of conditioned medium, specific recovery of galactosidase activity was first demonstrated by addition of increasing concentrations of purified E. coli galactosidase standard to serum free culture medium (SFM) followed by immunocapture assay of the galactosidase activity. In the SFM, 99% recovery of galactosidase activity was obtained (Fig. 1). coli galactosidase added at increasing concentrations to phosphate buffered saline (PBS) (), and recovery of chemiluminescence signal from purified E. coli galactosidase added at increasing concentrations to an undiluted citrate mouse plasma pool () or serum free culture medium (). Values shown are the mean sem of triplicates and are displayed as log log plots. The linear correlation coefficient for each fitted line was > 0.99, pFull figure and legend (44K)When MDA MB 231 BAG cells were grown in monolayer culture, it was found that galactosidase activity could readily be measured in the conditioned medium. To test the relationship between cell density and galactosidase activity in conditioned medium, different densities of MDA MB 231 BAG cells were grown in culture under serum free conditions for 48 hours, and subsequently the conditioned media were assayed for galactosidase activity. In the range of 103 to 106 cells per milliliter of medium, cell number correlated significantly with galactosidase activity measured in the conditioned medium (linear correlation coefficient, r > 0.98, p Fig. 2). MDA MB 231 BAG cells were cultured at different densities for 48 hours under serum free conditions with Eagles MEM containing 0.1% bovine serum albumin (BSA). Aliquots of culture supernatants were used for the measurements of galactosidase activity. At the end of the experiment, the cell number was determined by the MTT method. Results are expressed as the mean sem for three independent determinations. The linear correlation coefficient for the fitted line was > 0.98, pFull figure and legend (42K)In addition, when testing three other lacZ transfected human tumor cell lines, MDA MB 435 BAG breast carcinoma, PC 3 prostate carcinoma, and MV3 melanoma, which stably expressed the lacZ gene, a linear correlation was observed between galactosidase activity measured in the conditioned medium and cell number (r > 0.98 and p E. coli galactosidase in phosphate buffered saline (PBS) with and without addition of doxorubicin, it was verified that doxorubicin in concentrations up to 2 g/ml did not interfere with the measurements of galactosidase activity in the immunocapture assay (data not shown).Recovery of Galactosidase in Mouse PlasmaFor measurements of mouse plasma, specific recovery of galactosidase activity was demonstrated by addition of increasing concentrations of purified E. coli galactosidase standard to an undiluted control citrate mouse plasma pool, and subsequent determination of the galactosidase activity by the immunocapture assay. In the pandora earrings plasma pool, 97% recovery of galactosidase activity was obtained (Fig. 1). Galactosidase Activity in Plasma Correlates with Tumor Volume in Mice Xenografted with lacZ Transduced Human Tumor CellsWhereas plasma from mice without tumors gave negligible galactosidase signal in the immunocapture assay, significant levels of E. coli galactosidase were detected in plasma from mice xenografted with MDA MB 231 BAG tumor cells. A study on the relationship between galactosidase activity and tumor burden was then performed. Fifty mice were inoculated subcutaneously and bilaterally with MDA MB 231 BAG tumor cells. Tumors were allowed to grow for 4 (n = 9 mice), 6 (n = 10 mice), and 8 (n = 31 mice) weeks. At these different time points, tumor volume was measured and blood was taken for citrate plasma by cardiac puncture under anesthesia. The total tumor volume of the left and the right tumors differed for the individual mice in each group, representing the following mean values and ranges: 120 mm3 (28 mm3; 4 weeks of tumor growth); 1830 mm3 (462 mm3; 6 weeks of tumor growth); and 7170 mm3 (1340 mm3; 8 weeks of tumor growth). The individual plasmas (from individual mice) were measured for galactosidase activity. The mean values of plasma galactosidase activity in the three groups was as follows: 0.019 mU/ml (0.0070 mU/ml; 4 weeks of tumor growth); 0.15 mU/ml (0.016 mU/ml; 6 weeks of tumor growth); and 0.65 mU/ml (0.0069 mU/ml; 8 weeks of tumor growth). When the individual galactosidase activity levels were plotted against the corresponding total tumor volume per mouse, it was found that the plasma galactosidase activity values correlated well with the corresponding tumor volumes (Spearman correlation coefficient rho = 0.89, p Fig. 3). From Figure 3, it can be estimated that the minimum tumor volume in MDA MB 231 BAG tumor bearing mice, which produces detectable levels of galactosidase activity in plasma, corresponds to approximately 100 mm3 (representing approximately 5107 cells). The galactosidase activity in individual host plasma samples from mice xenografted with MDA MB 231 BAG tumor cells was plotted against the total tumor volume of the tumors on the left and the right flanks of each mouse. Spearman correlation coefficient rho = 0.89, pFull figure and legend (51K)In a subsequent experiment the other human breast cancer cell line (MDA MB 435 BAG) was inoculated subcutaneously and bilaterally into 28 nu/nu META/Bom mice, seven mice in each group. Tumors were allowed to grow for 6 weeks (Group 1), 8 weeks (Group 2), 10 weeks (Group 3), and 12 weeks (Group 4). Tumor size was measured every week. At the end of the experiment the total tumor volume of the left and the right tumors represented the following mean values and ranges: 140 mm3 (40 mm3; Group 1); 210 mm3 (43 mm3; Group 2); 750 mm3 (83 mm3; Group 3); and 1480 mm3 (340 mm3; Group 4). Blood was taken from individual mice and pooled within each group. Citrate plasma was separated from the blood, and plasma galactosidase activity was measured by the immunocapture assay. The mean values for the pooled plasma of galactosidase activity for each group were: 0.038 mU/ml (Group 1); 0.24 mU/ml (Group 2); 0.41 mU/ml (Group 3); and 1.9 mU/ml (Group 4). Comparison of the level of galactosidase activity in pooled plasma (plasma from the mice within each group were pooled) and mean tumor volumes (all tumors within a group) at the time of animal sacrifice showed a strong correlation (linear correlation coefficient r > 0.98, p = 0.01, not shown).Plasma Galactosidase Activity and Tumor Volume in Response to Doxorubicin TreatmentTo select a potent cytotoxic agent, MDA MB 231 BAG cells were exposed in vitro to the cytotoxic agents BCNU, ara C, cisplatin, VP 16, and doxorubicin in a clonogenic chemosensitivity assay. The different agents showed the following cytotoxicity, expressed as LD50 values: BCNU (1.3 g/ml); ara C (0.080 m); cisplatin (0.30 g/ml); VP 16 (0.15 m); and doxorubicin (0.030 m). Subsequently, doxorubicin was chosen for the anti tumor treatment of mice.To determine whether plasma galactosidase activity could be used in evaluating the effect of anti neoplastic therapy, 33 mice were inoculated subcutaneously and bilaterally with MDA MB 231 BAG tumor cells. After 3 weeks of established tumor growth (mean tumor volume: 420 mm3), mice were randomized to single IV injection treatment of either saline only (vehicle control) or 5 mg/kg of doxorubicin, corresponding to 25% of the LD10 value for doxorubicin in mice. Subsequently, tumor growth was monitored until termination of the experiment at Day 37 after tumor cell inoculation (Fig 4A). A significant inhibition (Mann Whitney U test, p 3) compared with the control group (mean tumor volume per mouse: 4800 mm3), reflecting a 4 fold reduction in mean tumor volume (Fig. 4A). At Day 37 blood samp pandora earrings les were drawn and plasma galactosidase activity was determined. A significantly (Mann Whitney U test, p = 0.001) lower level of plasma galactosidase activity was found in the doxorubicin treated group (mean galactosidase activity: 0.050 mU/ml) compared with the control group (mean galactosidase pandora earrings activity: 0.19 mU/ml), reflecting a 3.8 fold reduction in plasma galactosidase activity (Fig. 4B). Values shown are mean sem of 17 mice (saline treated) or 16 mice (doxorubicin treated). At Day 37 the tumor volumes of doxorubicin treated mice were significantly reduced compared with the saline treated mice (Mann Whitney U test, p B, Box plot (showing median, 25th and 75th percentiles) of plasma galactosidase activity level at Day 37, between MDA MB 231 BAG xenografted nude mice treated with either doxorubicin or saline. A Mann Whitney U test showed that the doxorubicin treated group had reduced galactosidase plasma levels compared with the group treated with saline alone (p = 0.001).Full figure and legend (44K)Top of pageDiscussionWe have previously described the development of a new assay for the measurement of bacterial galactosidase activity in mouse tissue extracts from lacZ transduced human tumor xenografts growing in nude mice, using an immunocapture assay using chemiluminescence detection (Holst Hansen et al, 1998). This assay provides a high sensitivity, with a detection limit of 0.01 mU/ml of purified bacterial galactosidase diluted in PBS and a wide dynamic range. In addition, the assay exhibits high specificity for bacterial galactosidase, showing insignificant cross reactivity with endogenous mammalian lysosomal galactosidase(s). In an experimental metastasis model, this method was well suited for the quantitative determination of the amount of bacterial galactosidase in host mouse lung homogenates, as an estimate of the number of metastatic lacZ transduced tumor cells. In the present study, we report the release in vitro and in vivo of bacterial galactosidase from lacZ transduced human tumor cells, and the correlation of the level of host plasma E. coli galactosidase with xenograft tumor volume.By applying the immunocapture chemiluminescence assay to conditioned media from MDA MB 231 BAG human breast cancer cells stably transduced with the lacZ reporter gene, it was shown that these cells release soluble galactosidase into the medium. The release of galactosidase was cell number dependent and appeared to be a more general phenomenon, because three other human tumor cell lines stably transfected with the lacZ gene released galactosidase into the culture medium as a linear function of viable cell number. It could be speculated that release of intracellular galactosidase from dying and disintegrating cells contributed significantly to the concentrations measured in the medium. However, this is not considered likely because MTT (3 bromide) assays showed that the cultured cells remained viable throughout the monitoring period.In contrast to nontumor bearing animals, plasma from nude mice with the MDA MB 231 BAG or MDA MB 435 BAG tumors tested positive for bacterial galactosidase activity. A statistically significant correlation was found between tumor volume and the level of galactosidase activity in mouse plasma, indicating that plasma levels of galactosidase activity reflect tumor burden in the host mice. This may prove to be a methodological improvement for the estimation of tumor burden in animal models, especially when the tumor cells are growing nonsubcutaneously, eg, orthotopically (Hoffman, 1994), and may as such pr pandora earrings esent a significant advantage over currently used methods, eg, area and volume tumor measurements, histological quantitation of tumor cells, and measurements of specific tumor markers. The application of this new assay is further supported by our previous characterization of a close correlation between galactosidase activity in lung tumor extracts and experimental lung metastases (Holst Hansen et al, 1998).A study in nude mice was then undertaken to assess whether measurements of galactosidase activity in plasma were responsive to reduction in tumor volume brought about by cytotoxic drug therapy. Mice bearing established MDA MB 231 BAG tumor xenografts were treated intravenously with a single injection of doxorubicin, which produced a strong inhibition of primary tumor growth compared with saline treatment. It was found that the mean tumor volume and mean plasma galactosidase in doxorubicin treated mice were significantly reduced by 4 and 3.8 fold, respectively, compared with vehicle treated mice. These results provide evidence that the plasma level of galactosidase in host mice is an indicator of tumor burden, and furthermore that it reflects response to reduction in tumor burden following cytotoxic drug therapy.