Disaggregation studies of xenograft solid tumors grown from pure or admixed clonal subpopulations from a heterogeneous human colon adenocarcinoma

Two clonal tumor subpopulations (designated as A and D) obtained originally from a heterogeneous human colon adenocarcinoma (DLD-1) were used to produce xenograft solid tumors in nude mice. First, disaggregation studies were performed to determine the optimal choice of enzyme and time of dissociation for the pure A and D neoplasms, using cell yield (cells/mg/min) and colony forming efficiency (CFE) assays. The enzymes investigated were: 0.5 or 0.2% trypsin, and two cocktails containing pronase (0.5 or 0.05%), collagenase (0.02%), and DNAse I (0.02%). For the 0.5% trypsin treatments, the cell yield from A and D tumor fragments increased until about 30 min, at which time a plateau in cell yield was reached. A plateau in CFE was also reached at this time. In contrast, the cell yields for the 0.2% trypsin treatment did not reach a plateau within the time of the dissociation (120 min), and the CFEs were lower than with the 0.5% trypsin. Whereas no differences in cell yield or CFE were found between the enzyme cocktail studies (0.5% trypsin vs. 0.05% pronase), the cell yield and the CFE from the clone D carcinomas were significantly less than that found with the 0.5% trypsin (the cell yield and CFE from clone A tumors were identical for 0.5% trypsin or enzyme cocktail). These data indicate that, while these clonal neoplasms have somewhat different responses to enzyme disaggregation, it is possible to select an enzyme treatment and treatment time that is appropriate for use on both A and D tumors (i.e., 0.5% trypsin). After determination of an acceptable enzyme procedure, 'reconstructed' heterogeneous tumors produced from an initial injection bolus of 50% clone A and 50% clone D cells were disaggregated as a function of time (days 12-83 postinjection). Over this period, we found that the cell yield decreased exponentially, with a half-time (T1/2) of 20.5 +/- 7.3 days (95% confidence limits), with a maximum extrapolated cell yield at time zero of about 1.2 X 10(5) cells/mg. The CFE was essentially constant over the duration of the assay period. Moreover, it was found that the percentage of clone A cells appeared to decrease exponentially (T1/2 = 20.5 +/- 11.5 days, 95% confidence limits) until about 40 days postinjection. After this time an equilibrium mixture consisting of about 10% clone A cells and 90% clone D cells was reached.(ABSTRACT TRUNCATED AT 400 WORDS)     

An enzymatic method for the consistent production of monodispersed viable cell suspensions from human solid tumors

An enzymatic method is described for disaggregation of viable tumor cells from human solid tumors. The enzymatic cocktail consists of 0.1% collagenase, 0.01% hyaluronidase, and 0.002% deoxyribonuclease. After mechanical mincing of the tumor tissue, tumor specimens are dissociated by incubation in the enzymatic cocktail for 12-18 hours at room temperature. In 17 cases of sarcoma, the mean yield was 5 X 10(6) viable cells per gram tumor tissue. Yield was 1 X 10(7) viable cells per gram tumor tissue in 23 cases of gastrointestinal carcinoma. The viabilities of tumor cell suspensions ranged from 50 to 98%, except for low viabilities in four specimens that were grossly composed almost entirely of necrotic tissue. The dissociation procedure is simple and the viable cell yield is sufficient for applications in studies of human cancer immunobiology.     

Inflammatory cells in solid murine neoplasms. I. Tumor disaggregation and identification of constituent inflammatory cells.

Mechanical and enzymatic methods of disaggregating tumors were studied with the goals of (1) minimizing cell losses while (2) maintaining functional and surface membrane markers needed to objectively identify inflammatory cells (IC)1 in resultant suspensions. Application of the principles and methods described makes accurate estimation of the percentage of each IC type present in neoplasms possible for the first time. Compared to purely mechanical means of disaggregating tumors, all enzyme mixtures tested markedly increased yields of viable cells/g neoplasm. Best results were obtained with a combination of collagenase and a protease of broader substrate range (alpha chymotrypsin, papain, pronase or trypsin). The combination of enzymes that gave the highest yields with the least effect on inflammatory cell markers was trypsin, collagenase and DNAse (TCD). Because mechanical injury appeared to be the greatest single cause of cell loss (the enzymes themselves had little direct effect), potential sources were identified and either eliminated or minimized. With TCD, depending on the tumor system, cell recovery (measured as DNA recovered in cell suspensions) was as high as 50% and yields were as much as 6.9 X 10(8) viable cells/g tumor. Complete disaggregation was not required to obtain representative IC populations from tumor fragments. Neutrophils, eosinophils and mast cells from disaggregated neoplasms were counted in Giemsa stained cytocentrifuge preparations based on their unique morphologic appearances. Macrophages were identified by their capacity to phagocytose zymosan, a function which proved highly resistant to the effect of enzymes. Flourescent microscopic identification of brain associated thymus antigen (BATA) allowed quantification of T lymphocytes, since this marker was virtually unchanged by enzyme exposure. Surface immunoglobulin (Ig) was stripped from B lymphocytes most rapidly by pronase and chymotrypsin, slowly by trypsin and papain, and not at all by collagenase. Ig positive cells therefore could be quantified in suspensions generated by collagenase or very short (20 min) exposure of fragments to trypsin.     

Recovery of cell subpopulations from human tumour xenografts following dissociation with different enzymes.

Human epidermoid tumours (Co112, HEp3, A431, ME180) grown in nude mice were dissociated using four different enzyme cocktails: 0.025% collagenase, 0.05% pronase, 0.04% DNase; 0.1% protease IX; 0.14% trypsin, 0.04% DNase; 0.025% collagenase, 0.02% DNase. Using these different enzymatic procedures, the total cell yields, host to tumour cell ratios, plating efficiencies and cell cycle distribution profiles obtained from each tumour model were compared. For all tumours tested, enzyme cocktail 1 was the most effective in releasing the greatest total number of cells g-1 tumour. However, for each tumour the percentage of neoplastic cells recovered, the plating efficiency and the cell cycle distributions varied according to the enzyme cocktail used to dissociate the tumour. For example, for HEp3 tumours, the highest plating efficiency was achieved using enzyme cocktail 4, whereas for ME180 tumours, this enzyme cocktail produced the lowest plating efficiency. Further, the effect of lethally irradiated (HR) feeder cells on the plating efficiency of the various tumours was found to be influenced by the enzymes chosen to dissociate the tumours. These studies indicate that the choice of an enzyme dissociation technique may profoundly influence the results obtained using human tumour xenografts.     

Evaluation of cell subpopulations isolated from human tumor xenografts by centrifugal elutriation

Human epidermoid tumor cells (Coll2, ME180, A431, HEp3) grown as xenografts in nude mice, were dissociated into single cell suspensions using an enzyme cocktail containing 0.025% collagenase, 0.05% pronase, and 0.04% DNase. The dissociated cell suspensions were separated by centrifugal elutriation into fractions containing homogeneous cell subpopulations primarily based on the differences in the rates of sedimentation. The quality of separation was evaluated by several techniques including flow cytometry, cell volume distributions, in vitro colony forming assay and morphological examination of Wright-Giemsa stained cells. In each separated fractions, the host to neoplastic cell ratio, the DNA ploidy, the plating efficiency and the cell cycle distribution were determined. After an initial separation of non-neoplastic host cells from malignant cells, a purity of greater than 95% host cells was obtained from the four xenografts studied. DNA analysis of tumor suspensions showed that neoplastic cells of different xenografts contained aneuploid cells with a DNA index of 1.51 to 1.95. The neoplastic cells were further separated into fractions according to their positions in the cell cycle. Fractions containing greater than 95% G1, 65% S, and 72% G2M cells were obtained from HEp3 xenografts. Less efficient separation with respect to cell cycle was attained with cells derived from Coll2, ME180, and A431 xenografts. Colony forming abilities of the neoplastic cells were determined at different phases of the cell cycle and found to be similar to those of the unseparated cell suspensions after corrections for non-neoplastic host cells were made. These investigations indicate that centrifugal elutriation is an effective technique of obtaining homogeneous subpopulations of cells from human tumor xenografts for various tumor biology and cell kinetics studies.     

The combined effect of interferon beta and radiation on five human tumor cell lines and embryonal lung fibroblasts

PURPOSE: The combined effect of natural Interferon-beta (n-IFN-beta) and ionizing radiation was tested in vitro on 5 different tumor cell lines and 1 embryonal lung fibroblast cell line. MATERIALS AND METHODS: The following cell lines were used: A549 (lung cancer), MCF-7 (breast cancer), CaSki (cervical cancer), WiDr (colon cancer), ZMK-1 (head and neck cancer), and MRC-5 (embryonal lung fibroblast line). Cells were incubated with n-IFN-beta (30 I.U./ml to 3000 I.U./ml) 24 h before irradiation. Irradiation was given as single dose between 1 and 6 Gy. Cell survival was evaluated using a standard colony-forming assay. RESULTS: Incubation with n-IFN-beta enhanced the effect of radiation in all tumor cell lines tested. The maximum sensitizing enhancement ratios (SER) at the 37% survival level were: 1.66 for A549 cells, 1.47 for CaSki cells, 1.56 for MCF-7 cells, 1.40 for WiDr cells, and 1.57 for ZMK-1 cells. In the nonneoplastic MRC-5 cell line, no radiosensitizing effect of n-IFN-beta could be demonstrated. The linear quadratic fit of the survival curves showed an increase of the alpha-component for all tumor cell lines treated with n-IFN-beta. CONCLUSIONS: IFN-beta enhanced the effect of radiation in the tumor cell lines, but not in the nonmalignant lung fibroblasts. The increase of the alpha component in the survival curves indicates that impaired radiation repair or the accumulation of sublethal damage might play a role for the radiosensitizing effect of n-IFN-beta.     

Separation, phenotyping and limiting dilution analysis of T-lymphocytes infiltrating human solid tumors

Tumor-infiltrating lymphocytes (TIL) were obtained by a combination of mechanical release and enzymatic disaggregation from 35 human solid tumors. The number of lymphocytes in TIL-enriched suspensions varied from 1 X 10(4) to 7.6 X 10(6) per wet gram of tumor. The TIL preparations separated by differential centrifugation on Ficoll-Hypaque gradients contained 10-95% of T11+ cells (mean 50%), and tumor cells accounted for the other major cellular component. Macrophages, NK cells, B cells and granulocytes were infrequently seen. Morphologically, TIL-T were small non-activated cells. They expressed the T11 and T3 antigens but not the receptor for IL-2 (IL-2R) or HLA-DR antigens as determined by double immunofluorescence staining. Rare T11+/IL-2R+ cells were recovered only from colon and lung carcinomas. The mean T4/T8 ratio in 12 TIL preparations was 1.1 +/- 0.8. Immunohistology with monoclonal antibodies (MAbs) performed in 31/35 tumors confirmed that the T11+ cells infiltrating solid tumors rarely expressed the IL-2R and that the cell content of suspensions enriched in TIL was comparable to that determined in situ. The recovered TIL were cloned in a microculture system that permits proliferation of nearly all normal peripheral blood T lymphocytes (PBL-T). Under these culture conditions, frequencies of the proliferating T lymphocyte precursors (PTL-P) were depressed in both the TIL preparations (less than 0.01 to 0.39) and patients' PBL-T (0.05 to 0.5). These low frequencies of PTL-P were seen in patients with all tumor types, both primary and metastatic.     

Characterization of cells obtained by mechanical and enzymatic means from human melanoma, sarcoma, and lung tumors

Characterization of cells comprising solid tumors will facilitate the rational design of cancer chemotherapy for individual patients. We have prepared cell suspensions from human melanoma, sarcoma, and lung tumors by thinly slicing the tissue with a microtome and scalpels (mechanical release), followed by treatment with a mixture of collagenase II and DNase I (enzymatic release). This method of disaggregation resulted in two cell suspensions for each tumor specimen, and we characterized these suspensions by assessing their dye exclusion capability, ribonucleoside triphosphate pools, cytological profile and clonogenicity in soft agar. The enzymatic method thus yields cells in addition to those obtainable by a mild mechanical procedure, and these cells are similar in cytological profile and clonogenicity in soft agar to those released mechanically. Furthermore, the enzymatically released population is superior to that released mechanically for purposes requiring large numbers of dye-excluding cells having intact ribonucleotide pools.     

Comparison of glutathione levels in rodent and human tumor cells grown in vitro and in vivo

Cellular glutathione (GSH) levels were compared in human and rodent tumor cells grown both in vivo and in vitro. Three human (A431, HEp3, ME180) and two murine (KHT and RIF-1) tumor cell lines were used. The average GSH contents for exponentially growing human cells in vitro were 14.2, 10.9, and 17.0 fmol/cell for ME180, A431, and HEp3 cells, respectively. These cells also were grown as tumors in nude mice. Following dissociation, greater than 90% pure populations of neoplastic and nonneoplastic cells were isolated by centrifugal elutriation prior to GSH determination. The data showed that the GSH levels of the tumor associated host cells were appreciably lower than those of the neoplastic cells. In addition, in contrast to the values obtained for the exponential cells, neoplastic cells grown in vivo showed a 2- to 3-fold reduction in GSH. However, the values for in vivo cells were similar to those obtained for the same tumor cells grown in vitro in the plateau phase. Compared to the human tumor cells the GSH contents of murine tumor cells always were lower. For example, RIF-1 and KHT cells in the exponential growth phase had GSH contents of 3.3 and 7.5 fmol/cell, respectively. Also, as was observed with the human cells, the GSH content of KHT cells in plateau phase of growth was 2-3 times less than that of cells in the exponential phase of growth. Similarly, the GSH content of KHT cells grown as in situ tumors prior to dissociation and isolation by centrifugal elutriation also was reduced by a factor of 3 compared to exponential phase cells. Although the average volume of tumor cells grown in vivo was less than that of cells grown in vitro, this did not account for the differences in GSH values observed when in vitro and in vivo derived cells were compared. Finally, GSH measurements made on multiple biopsies of individual human tumor xenografts varied by a factor of 2-3 within each tumor type studied. This variation, likely due to host cell fluctuations, may present a complicating feature in the interpretation of solid tumor GSH levels.     

Clonal origin of tumors induced by ultraviolet radiation

Skin tumors were induced in (C3H/HeN X C3H/HeN-PGK-1a)F1 female mice, heterozygous at the X-linked phosphoglycerate kinase-1 (PGK-1) locus, by exposure to the carcinogenic influence of ultraviolet radiation (UVR), 3-methylcholanthrene [(MCA) CAS: 56-49-5], or benz[a]pyrene [(BP) CAS: 50-32-8]. An assessment of the clonal origin of these tumors was accomplished through an analysis of the PGK-1 enzyme phenotype expressed by the transformed cells. In vitro culture was employed as a means of depleting nontransformed cells of host origin from the induced tumors. Cultured lines derived from tumors induced by each of the above agents were found to express only one of the two enzyme forms encoded by the host genotype, consistent with the probability that all the UVR-, MCA-, and BP-induced tumors examined in this study were monoclonal in origin. For further substantiation of the monoclonality of UVR-induced tumors, 2 UVR-induced tumors were enzymatically dissociated immediately following excision from the primary hosts, and the resulting cell suspensions were cloned in soft agar. Upon analysis, each set of clones selected in soft agar expressed only a single PGK-1 enzyme form. To rule out the possibility that the apparent monoclonality of culture-adapted tumor lines was due to in vitro selection, tumors that arose in UVR-treated PGK-1a/b female heterozygote mice were transplanted into PGK-1a and PGK-1b homozygous recipients. These transplanted tumors expressed a single PGK-1 allozyme following growth in recipients that were genetically homozygous for the major PGK-1 enzyme form expressed by the tumor prior to transplantation. These data strongly support the concept that most, if not all, UVR-induced tumors are derived from the progeny of a single transformed cell. This observation is important to the understanding of the nature of tumor-specific transplantation antigens expressed by individual UVR-induced tumors and indicates that such antigens also are clone specific. In addition, the results of this study indicate that polyclonality does not play a significant role in the generation of cellular and phenotypic heterogeneity known to be present within individual UVR-induced tumors.     

An assessment of intratumor phagocytic and surface marker-bearing cells in a series of autochthonous and early passaged chemically induced murine sarcomas.

Single cell suspensions from five different 3-methylcholanthrene-induced tumors in CBA mice were examined in the autochthonous host and sequentially for 5-11 passages. They were also examined for Fc receptor-bearing, phagocytic, theta antigen-positive, and surface immunoglobulin-bearing cells. The preparations contained a high proportion of phagocytic and marker-bearing cells both in the original host and during early passage. This proportion was consistent for any particular tumor and passage. Between different tumors, however, the proportions were sufficiently different to allow the tumor to be identified on this basis; this suggested that various chemically induced tumors may be unique in their tumor-host relationship as measured by the type of cells which infiltrate them. With on-going early passage of the tumors, the proportion of marker-bearing cells decreased to a constant level in most instances, mainly because of a reduction in the percentage of phagocytic cells. The tumor with the least macrophages (MBQA, less than 5%) consistently appeared more rapidly and killed the host more rapidly than did the tumor with the most macrophages (MBQD, 15-30%), but was not significantly different in its growth rate. The theta antigen- and Fc receptor-positive cells within these tumors were derived from the animal receiving the tumor inoculum, and thus represented host cell infiltration of the tumor. The results were discussed with reference to fundamental concepts of the immunology of chemically induced tumors and the importance of host cell infiltration within these tumors.     

Quantifying transient hypoxia in human tumor xenografts by flow cytometry

Transient hypoxia is a poorly understood and potentially important factor that may limit tumor response to various forms of therapy. We assessed transient hypoxia on a global scale in two different human tumor xenografts by sequentially administering two hypoxia markers followed by quantification of hypoxic cells using flow cytometry. High levels of the first hypoxia marker (pimonidazole) were maintained in the circulation over an 8-hour period by multiple hourly injections, providing a "time-integrated" hypoxia measure showing an asymptotic increase in the total number of hypoxic cells. Subsequent administration of a second hypoxia marker (CCI-103F) showed that substantial numbers of the previously pimonidazole-labeled cells were no longer hypoxic during the circulation lifetime of the second marker. The overall fraction of tumor cells that demonstrated changes in hypoxic status with time increased with different kinetics and by different magnitudes in the two xenograft systems. Specifically, up to 20% of the cells in SiHa (human cervical squamous cell carcinoma) tumors and up to 8% of the cells in WiDr (human colon adenocarcinoma) tumors were intermittently hypoxic over an 8-hour period. Also, the tumor cells that demonstrated transient hypoxia were typically not adjacent to functional tumor blood vessels. Similar approaches could be used in the clinic to provide information on the duration of intermittent hypoxia episodes and the fraction of transiently hypoxic tumor cells, which would, in turn, have important implications for the strategic improvement of cancer therapy.     

Immunohistologic evaluation of the lymphoreticular infiltrate of human central nervous system tumors.

Forty-five nervous system tumors (9 glioblastomas, 9 meningiomas, 15 assorted primary neural tumors including 3 medulloblastomas, and 12 brain tumors metastatic to the brain were analyzed for their content of lymphocytes, granulocytes, and macrophages. Cell suspensions were prepared by enzymatic digestion; lymphocytes and granulocytes were quantitated by morphology following cytocentrifugation, and macrophages were quantitated by IgG EAC (erythrocyte-antibody-complement) rosette formation. EA (erythrocyte-antibody) adsorption to sections of tumor was employed to determine the distribution of the IgG Fc receptor-positive cells within the tumors and to serve as quality control for selective release of Fc receptor-positive cells by enzyme digestion. The 9 glioblastomas had a mean macrophage content of 41% (range: 5-78%); the 9 meningiomas, 42% (range: 5-80%); the 3 medulloblastomas, 6% (range: 2-15%); and the metastatic tumors, 21% (range: 2-50%). Lymphocyte contents were variable but generally less than 10%. Most tumors contained less than 10% granulocytes. EA adsorption demonstrated that Fc receptor-positive cells were distributed throughout the tumor mass, although different types of patterns were observed. There was an excellent correlation between the percent EAC rosette-positive cells in suspensions and the extent of EA adsorption to the tumor sections. The significance of the study primarily rests in the demonstration that most nervous system tumors contain high numbers of infiltrating host cells, primarily macrophages.     

Cellular DNA of human neoplastic B-cells measured by flow cytometry

Flow cytometric analysis of DNA of tumor cells rapidly provides information on cell kinetics and tumor ploidy. Human B-cell lymphomas, however, often contain high numbers of nonneoplastic cells, mainly T-lymphocytes, which may hamper the accurate measurement of cell cycle phases and ploidy level of these tumors. The neoplastic cells in each B-cell lymphoma express a single immunoglobulin light chain. Therefore, we labeled surface immunoglobulin light chains to discriminate between predominantly neoplastic B-cells and nonneoplastic cells in the same tissues. Using this label as well as antibodies against nonneoplastic T-cells, we performed multiparameter correlated flow cytometric analysis of 52 human B-cell lymphomas measuring cellular DNA in neoplastic and nonneoplastic populations from the same tissues without physical separation of cells. Comparison of cellular DNA of immunoglobulin light chain-bearing neoplastic cells with that of nonneoplastic cells from the same tumor enabled us to detect DNA changes (aneuploidy) in almost 80% of the lymphomas, an incidence higher than observed previously by conventional DNA analysis of unseparated cells. These ploidy changes were confirmed by comparing in the same tumor the DNA of normal T-cells with that of predominantly neoplastic cells. The proportion of neoplastic cells in the synthetic phase of the cell cycle (S-fraction) varied widely from tumor to tumor. Lymphomas with high neoplastic S-fractions (higher than 10%) were mostly hyperdiploid tumors and histologically corresponded to intermediate- and high-grade unfavorable lymphomas. Tumors with low neoplastic S-fractions (less than 5%) were predominantly diploid and near diploid, histologically low-grade lymphomas. Six lymphomas showed two discrete cell populations bearing the same immunoglobulin light chain but containing different amounts of DNA suggesting the presence of two neoplastic clones in the same tumor (biploidy). In two patients in whom the lymphoma relapsed at 17 and 34 months, respectively, after the initial biopsies, repeat tumor samples were obtained. Despite an increase in the neoplastic S-fraction, no change in ploidy level was observed in either case. Light scatter analysis suggested a relationship between cell size and genomic size; large cells in these tumors were mostly presynthetic aneuploid cells. The ability to measure DNA, antigens, and cell size in individual cells in a rapid, correlated manner is a unique attribute of flow cytometry.(ABSTRACT TRUNCATED AT 400 WORDS)     

Measurement of oxygen diffusion distance in tumor cubes using a fluorescent hypoxia probe.

Oxygen diffusion distance was measured in solid tumor "cubes" prepared by excising the tumor from the mouse and incubating 1-2 mm sided tumor cubes in spinner culture flasks with fluorescent drugs (AF-2 or DM113) which bind to hypoxic cells. After incubation, frozen sections were prepared and examined for AF-2 or DM113 binding using a fluorescence image processing system. Alternatively, cubes were stained with a slowly penetrating fluorescent dye, Hoechst 33342, prior to disaggregation and measurement of AF-2 binding using flow cytometry. The distance from the cube surface to the AF-2 stained region (i.e., the thickness of the unstained rim) was used as an indication of oxygen diffusion distance, which depends on the square root of the tumor oxygen consumption rate. Oxygen diffusion distance was dependent on tumor type, external oxygen concentration, and temperature during incubation with AF-2, but was unaffected by tumor size up to 1.2 gm or position of the cube within the tumor. Oxygen diffusion distances (in planar geometry) measured for cubes prepared from SCCVII, RIF-1, Lewis lung or WiDr tumors were 107, 123, 153 and 193 microns, respectively. For SCCVII and WiDr tumors, the percentage of blood vessels separated by a distance greater than double the mean oxygen diffusion distance was used as an indication of the hypoxic fraction.     

Changes in the transplantability of a virus-induced murine leukemia tumor.

The subcutaneous growth potential of a Friend virus-induced murine leukemia cell line (FBL-3) passaged in the ascites form was found to change, depending on the interval of time spent in the peritoneal cavity. Injection sc of the original stock of FBL-3 (growth continuously in vitro) into C57BL/10 mice produced transient tumors uniformly rejected by days 20-40. The tumor cells passaged in the ascites form for 7 days behaved like the in vitro-grown cells, whereas ascites cells harvested at 14 days produced progressive lethal subcutaneous tumors in 50-70% of the recipients. The change in subcutaneous growth was reversible simply by alteration of the ascites passage schedule. Day 7 ascites cells (regressors) were converted to progressors by ip passage for 14 days, and day 14 ascites cells (progressors) were converted to regressors by ip passage for 7 days. The difference in growth potential between day 7 and day 14 ascites cells was not due to effects of nonneoplastic host cells accompanying the tumor cells in the ascites population, because neither dilution of nonneoplastic cells by in vitro culture nor selective killing of H-2a/b host cells by anti-H-2 serum and complement altered the subcutaneous growth behavior or either day 7 or day 14 ascites cells. These results indicated that the change in the growth potential of FBL-3 occurred at the level of the tumor cells. However, no quantitative differences were observed in the expression of serologically detectable tumor-associated antigens by these two populations. Possible mechanisms for this change in transplantability were considered.     

Fc receptor-bearing cells as a reliable marker for quantitation of host lymphoreticular infiltration of progressively growing solid tumors

Disaggregated cell suspensions made from transplanted solid tumors, either chemically-induced fibrosarcomas, or spontaneous mammary carcinomas, can contain very high numbers of F_c receptor-bearing cells which are of host origin. Because most types of lymphoreticular cells have F_c receptors, and because T cells—most of which are F_c receptor-negative—appear to infiltrate such tumors only to a very limited degree, the possibility that F_c receptor cells could serve as a reliable and simple marker for host lymphoreticular cell infiltration of solid tumors was tested. This was accomplished by comparing the ratios of F_c rosetting cells to serologically detectable host cells in H2^d or H2^k haplotype tumor cell suspensions grown in (H2^d × H2^k) F₁ hybrid mice, where host cells could be distinguished from tumor cells by treatment with the appropriate anti-H2 serum. Ratios of 0.8 to 1.08 were obtained for four different tumors including the SaD/2 fibrosarcoma, a CBA spontaneous fibrosarcoma, and the T1699 and CaD/2 mammary carcinomas. Analysis of the results showed that enumeration of F_c rosettes was a reliable host cell marker for at least three of the four tumors tested. The mean non-malignant host cell content of the various tumors, as assessed by anti-H2 cytotoxicity tests, ranged from 23% to 41%.     

CPT-11 in human colon-cancer cell lines and xenografts: Characterization of cellular sensitivity determinants

CPT-11, a new semisynthetic derivative of camptothecin, is active in a number of tumor types in the clinic, including colon cancer. CPT-11 is a drug that is converted into the active metabolite SN-38 by a carboxylesterase. Experiments were performed to obtain more insight in the cellular characteristics in 5 unselected human colon-cancer cell lines that account for the differential sensitivity to CPT-11 and SN-38. In vitro, the sensitivity to CPT-11 and SN-38 was highest in LS174T and COLO 320 cells, intermediate in SW1398 cells and lowest in COLO 205 and WiDr cells. SN-38 was 130 to 570 times more active than CPT-11. CPT-11 induced complete remissions in 6 out of 12 COLO 320 tumors grown as subcutaneous xenografts, but was not effective in WiDr tumors. The cellular carboxylesterase activity did not relate to the sensitivity to CPT-11. The enzyme activity was higher in normal mouse tissues, i.e., serum and liver, than in COLO 320 or WiDr xenografts, indicating that tumor carboxylesterase is of minor importance for CPT-11 efficacy. The topoisomerase-l mRNA expression in tumor cells was not predictive of the antiproliferative effects of CPT-11 or SN-38. We observed a positive relationship between the DNA topoisomerase-1 activity and the cellular sensitivity to carboxylesterase-activated CPT-11 (r = 0.75, p < 0.1) as well as to SN-38 (r = 0.89, p < 0.05). The higher topoisomerase-1 activity in COLO 320 cells and tumors when compared with that in WiDr cells and tumors reflected the differences in sensitivity to the drug(s). In conclusion, the DNA topoisomerase-1 activity was the best determinant for CPT-11/SN-38 sensitivity in this panel of unselected human colon-cancer cell lines. Int. J. Cancer, 70:335–340, 1997. © 1997 Wiley-Liss, Inc.     

TSU-68 (SU6668) inhibits local tumor growth and liver metastasis of human colon cancer xenografts via anti-angiogenesis

A number of receptor tyrosine kinases (RTKs) are involved in angiogenesis. TSU-68 (SU-6668) was developed as an inhibitor of RTKs involved in VEGF, bFGF and PDGF signaling, which then inhibits endothelial cell proliferation. We investigated the antitumor effects of TSU-68 against human colon cancer xenografts in male SCID mice and its anti-angiogenic activity using a dorsal air-sac (DAS) assay. TSU-68 was administered orally at a dose of 200 mg/kg twice daily. Mice bearing human colon carcinoma, HT-29, or WiDr xenografts were treated for 16 days. To determine the effect on hepatic metastasis, cell suspensions of HT-29 or WAV-I were injected into the spleen of mice on day 0, and mice treated for 28 days starting from day 1. For the DAS assay, HT-29, WiDr or WAV-I cells suspended in PBS at 2 x 10(7) cells/Millipore chamber were implanted subcutaneously into SCID mice, which were then treated from day 0 to 5, On day 6, the anti-angiogenic effects were assessed. Results indicated that TSU-68 significantly inhibited the growth of subcutaneous tumors. In the hepatic metastasis model, liver weights of the TSU-68-treated group were significantly reduced, compared to those of control mice. In the DAS assay, the angiogenic indices of the treated groups were significantly decreased for HT-29, WiDr and WAV-I tumors, with T/C ratios of 13.4, 50 and 35.3%, respectively. As TSU-68 significantly inhibited tumor growth and liver metastasis formation of human colon cancer xenografts, probably through anti-angiogenic activity, this agent may be useful for the treatment of colon cancer.     

Efficient recovery of clonogenic stem cells from solid tumors and occult metastatic deposits

We describe the use of enzymes combined with brief, sequential mechanical disruptions in a Tekmar Stomacher blender for the recovery of clonogenic neoplastic cells from solid tumors, lungs, and livers. The method has yielded 3 X 10(8) to 5 X 10(8) total cells and 1.2 X 10(6) to 17 X 10(6) clonogenic cells per gram of tissue from three different mouse mammary tumor subpopulations growing in the subcutis. The clonogenic cell yields represent a 4- to 13-fold increase over our previous best method of tumor disaggregation. The increase in total cells recovered, while not as dramatic (up to 3-fold), was statistically significant for two of the three tumor lines. We were also able to efficiently recover 125I-iododeoxyuridine labelled neoplastic cells from lungs and livers after injecting the cells intravenously. Over half of the total radiolabel present in these organs prior to disaggregation could be recovered in the cell suspensions obtained.