Tumor Cloning Pathobiology 1990;58:323-328
Cloning of Primary Human Tumors in Capillary Tube versus Petri Dish: A Head-to-Head Comparison Xlange Peng, Martin J. Murphy, Jr. Hippie Cancer Research Center, Dayton, Ohio, USA
Key Words. Capillary cloning • Two-layer agar plating • Human tumors • Plating efficiency
Introduction Since 1977, when Hamburger and Salmon [1] re ported the cloning of primary human tumors in the Petri dish, numerous studies have reported its application in the cloning of primary human tumors for chemosensitivity testing [2-4]. Comprehensive reports confirm that the successful growth rate for unselected primary human tumors is only 35-50%, where ‘successful’ is defined as those tumors forming a sufficient number of colonies for drug evaluation [5, 6]. This promising technique, there fore, requires further research to realize its full potential [7-9]. Various modifications have been made in at tempts to increase the tumor plating efficiency (PE) [10,
11]. Among these, the use of capillary tubes has the potential to circumvent the inherent disadvantages of the two-layer agar system [12-14]. Von Hoff et al. [14] reported that the median PE was 5-fold higher in capil lary tubes than in Petri dishes for the 183 humor tumors examined (18 different histological types). In the same paper by Von Hoff et al. [14], emphasis was placed on cloning tumor cells of breast, ovary and lung origins. The results in this paper confirm the data of Von Hoff et al., and extend the usefulness of capillary cloning for lymphoma, melanoma, and pancreatic can cer cells. In addition, our results taken together with those of Von Hoff et al. show that only stomach carci noma cells clone better in Petri dishes.
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
Abstract. Plating efficiencies (PEs) of primary human tumors cloned in a capillary assay system were compared to those derived from the conventional two-layer agar Petri dish assay system. A total of 143 consecutively received primary human tumors of 24 different pathologies were simultaneously tested in both assay systems. The successful clonal growth rate in the capillary assay was 82.7%, while in the Petri dish it was 64.7% (p < 0.001). The median PE was 0.017% in the capillary assay demonstrating a 4.25-fold increase over the 0.004% PE of the Petri dish system. The data confirmed previous results showing that cancer cells of ovarian, breast, and lung origins clone with higher PE in capillary tubes. In contrast, we confirmed that stomach carcinoma cells were the only tumor type that showed a higher PE with the Petri dish method. In addition, this study shows for the first time that lymphomas and renal cell carcinoma, when they survive in vitro, clone equally well in both methods. However, the capillary cloning method resulted in a 66% success rate for lymphoma cell cloning, but Petri dish cloning resulted in only a 33% success rate. Thus, for some types of cancers (i.e., lymphoma), capillary cloning may be advantageous because it improves the probability of obtaining evaluable results. In other cases, the advantage of capillary cloning may be only the decreased amount of specimen and reagents needed for the assay.
324
Peng/Murphy
Prior to introduction into the two different culture systems, han dling of all samples and cell preparations was performed identically and simultaneously. Collection and Preparation o f Tumor Cells Tumor specimens were obtained incidental to surgery in accor dance with federal and institutional guidelines. The specimens were placed in McCoy’s 5A medium (Gibco, Grand Island, N.Y.) con taining 15% fetal bovine serum (FBS) (Hyclone, Logan, Utah) and preservative-free heparin, and transported to our laboratory within 2-24 h of surgical removal. None of the patients had a history of previous chemotherapy. Preparation o f Single Cell Suspensions A total of 143 consecutively received primary human tumors of 24 different pathologies were simultaneously tested in both assays. Solid tumors were minced into 2- to 5-mm fragments and trans ferred to a trypsinizing flask (Bellco Glass, Vineland, N.J.). An enzyme mixture consisting of 0.2% neutral protease (Cat. No. P0384; Sigma Chemical Co., St. Louis. Mo.), 0.02% DNAse (Cat. No. D-4527 from bovine pancreas. Lot 66F-96601; Sigma) and 0.4% collagenase (Type IA, Cat. No. C-9891, Lot 87F-6843; Sigma) was added at 15 ml/g of tumor tissue [15]. A sterile magnetic bar was placed in the mixture, which was incubated at 37 °C with gentle stirring for 45-60 min and filtered through a 60-pm sterile mesh. If cell clusters of 40 pm or more were present in the culture medium, the mixture was filtered through a 30-pm mesh to eliminate the cell clusters, or the same volume of enzyme mixture was added for a further 30-min digestion, then washed with McCoy’s 5A medium (Gibco) as described previously [12]. Ascitic, pleural and peritoneal effusions were placed in sterile containers with preservative-free heparin (lOU/ml; Elkins-sinn, Inc., Cherry Hill, N.J.). The fluids were centrifuged at 1,500 rpm for 10 min and the cells washed twice with McCoy’s 5A medium. Cell viability was determined in a hema cytometer with trypan blue. Sample viabilities below 30% were excluded from the cloning assay, since at low cell viabilities colonies will not grow. Culture o f Cells in the Two-Layer Agar Petri Dish System Feeder layers consisted of McCoy’s 5A (Gibco) plus 15% heatinactivated FBS (Hyclone), 1 mM sodium pyruvate, 42 pg/ml Lserine, 1 mM L-glutamine, 20mA/4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid (all from Gibco), 9 mg/ml tryptic soy broth, 117 pg/ml L-asparagine and 0.5% agarose (Seaplaque; Marine Col loids, Rockland. Me.). One milliliter of this feeder layer was added to Lux 35 X 10 mm tissue culture dishes (Nunc, Inc., Naperville, III.) and kept at 4 °C for no more than 2 weeks. The upper agar layer consisted of 0.3 % agarose in enriched Con naught Medical Research Laboratories Medium 1066 (CMRL 1066; Gibco), supplemented with 15% heat-inactivated FBS, 3 mM ascor bic acid, 2.4 pg/ml transferrin (Sigma), 2 m M glutamine (Gibco), 2 U/ml preservative-free insulin (Lilly, Indianapolis Ind.), 20 mM 4-(2-hydroxyethyl-l-piperazine-ethanesulfonic acid), 3 pg/ml cata lase (Sigma), 0.004 pg/ml hydrocortisone (Sigma), 100 pg/ml strep tomycin and 100 U/ml penicillin G (Gibco). Tumor cells were sus pended in this plating medium at a final density of 4 X 105 viable cells/ml and 1 ml of the resultant mixture was pipetted onto the feeder layer. After addition of the upper agarose layer containing the
cells, the agarose was allowed to gel at 4 °C for 5 min and then the plates were incubated at 37 °C in a 5% COi humidified atmosphere for 14 days. Colonics (clusters of > 5 0 cells or 60 pm in diameter) usually appeared by day 14 of culture, and the number of colonies in each plate were enumerated with an Omnicon FAS III automated image analyzer (Bausch & Lomb, Rochester, N.Y.) [16], Culture o f Cells in Capillary Tubes Round 10-pl capillary tubes (Clay Adams, Parsippany, N.J.) were cut to 97.5 mm in length, washed for 2 h in an ultrasonic bath with deionized water containing a nontoxic detergent (Liqui-Nox; Alconox Inc., New York, N.Y.) and finally rinsed with five ex changes of deionized water. The cleaned capillary tubes were dryheat sterilized for 2 h before use. Tumor cells were collected and prepared exactly as described above. Cells to be cloned were suspended in 0.2% agarose in the same medium as that for the Petri dish system. The final plating mixture consisted of enriched CMRL 1066 medium (0.6 ml), cell suspension of 3 X 106 viable cells/ml (0.1 ml), drug solution at peak plasma concentration (0.1 ml) and 1% agarose (0.2 ml) [12], (Note that the number of cells/ml is lower in capillary cloning than in Petri dish cloning, because PE is higher, and < 200 colonies per capillary are desirable for accurate counting.) These ingredients were vortexed for thorough mixing, and 50 pi of the resultant mixture (i.e., containing 15,000 viable cells) was introduced into each capillary tube in triplicate, using an Eppendorf pipette. The capillary tubes were then horizontally placed in a capillary cassette specially de signed for this purpose. The agarose was allowed to gel at 4 °C for 5 min and then incubated at 3 7°C in a 5% CCL humidified air atmosphere. Both ends of each capillary tube were left open to the incubator atmosphere. A set of three capillary tubes was examined immediately prior to incubation for the presence of cell clusters (i.e., cell aggregates of 40 pm or greater in diameter. Those capillary tubes which contained more than four clusters were replaced by freshly prepared capillary culture tubes. Control groups were exam ined on the 5th day of incubation to evaluate colony growth, and capillary tube colonies (clusters of > 5 0 cells or > 6 0 pm in diame ter) were counted on the 7th day using an inverted microscope (X 150) except when growth was slow, in which case these capillary tubes were counted on the 14th day of culture. Data Analysis and Statistical Considerations The percentage of cloning efficiency in each system was calcu lated: __ Number of colonies,'vessel PE % = ---------------------------------- X 100 Number of cells plated Student’s paired t tests were used to statistically compare results between different groups.
Results Among the 143 different tumor samples, 4 lacked suf ficient viable cells to attempt culture in the capillary cloning system, rendering a total of 139 evaluable tu mors. In the Petri dish assay, which requires more cells
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
Materials and Methods
Cloning Human Tumors in Capillary Tube vs. Petri Dish
Discussion Capillary tubes have only recently [14] been used for the cloning of primary human tumor cells, although this vessel has been used to clone hematopoietic progenitor cells for the past decade [17-19]. The cloning of primary human tumor cells in Petri dishes has been expanded to predict the chemosensitivity response of individual pa tient tumors to anticancer drugs [20-23], Even though the conventional Petri dish methodology has been widely used. It has inherent disadvantages. For example: (1) only 35-50% of human tumors form enough colonies to enable drug sensitivity evaluation; (2) the PEs are very low, and (3) a large number of tumor cells is required to perform drug sensitivity testing. Since many tumor biop-
Table 1. PE (%) in capillary tube system versus Petri dish1
Mean PE, % Median PE, % Range PE, % Average colonies per 15,000cells plated Median colonies per 15,000cells plated Range of colonies per 15,000cells plated
Capillary
Petri dish
Capillary/ Petri dish
0.0307 0.017 0-3.0
0.0122 0.004 0-0.54
2.52 4.25 NA
56.4
18.3
3.08
6
4.66
28 0-450
0-162
0-2.77
NA » Not applicable. 1 Based on 143 different human primary malignancies of 24 dif ferent histological types.
sies are small, it is often difficult to test many chemo therapeutic drugs. Attempts have been made to increase the Petri dish PE including modification of the culture medium and supplementation of growth-promoting factors [9-11, 24, 25]. None of them, however, were found to be totally satisfactory. Cells have been cultured in square and round capillary tubes with an increase in both the suc cess rate of cloning and in the PEs when compared with Petri dish-derived data [26, 27], We have optimized the culture conditions, cell densi ty, supplemental growth-promoting factors, and the length of incubation [13]. The rate of successful growth in the capillary cloning system is appreciably higher than that of the conventional two-layer agar cloning system. A study from our laboratory reported a success rate of 77% [13], and with further improvements we have recently reported an 88% success rate [12], Von Hoff et al. [14] reported a higher PE for breast, ovarian and lung cancers cloned in capillary tubes. We obtained similar results indicating that this method of cloning can be reproduced in other laboratories. Lymphoma is usually difficult to clone in vitro. In the paper of Von Hoff et al., only 2 cases were reported, and both showed poor growth in Petri dishes and capillary tubes. We obtained a success rate of 66% in capillary cloning and 33% in Petri dish cloning. Of the 33% growth in both systems, a head-tohead comparison yielded a PE that was identical. We also observed that there were no differences in PE for pancreatic and renal cancers cloned in Petri dish or cap illary tube. Of importance is the finding that colon can-
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
than the capillary assay, 7 tumor samples could not be plated due to insufficient viable cells. Therefore, the total number of evaluable tumors plated in the Petri dish system was 136. The successful growth rate in the capillary cloning system was 82.7% (115/139) compared with 64.7% (88/136) in the Petri dish cloning system. The difference between these two cloning systems was significant (p < 0.001). Table 1 details the comparison of PE in the cap illary tube versus the Petri dish. The median number of colonies per 15,000 cells was 28 (mean 47), while the median number of colonies per 15,000 cells per Petri dish was 6 (mean 18). The median PE was 0.017% for tumors cloned in capillary tubes and 0.004% for tumors cloned in Petri dishes (i.e., a 4.25-fold increase in me dian percent of PE with capillary tubes). In a comparison of 51 consecutively received breast cancers, 4 tumors were excluded from evaluation since they did not grow in either culture system. The PE ratio of capillary cloning versus Petri dish cloning was >1 in 34 of the cases (72%), a further indication of the superior PE in capil lary tubes. To determine if there were differential tumor types in which the PEs were the most favorable in the capillarytubes, an analysis of mean PE by tumor type was per formed (table 2). To test the significance of these differ ences for each tumor type, paired t tests were run on the PE for both culture methods. As shown in table 2, the most significance differences were noted for bladder can cer (p < 0.001), breast cancer (p < 0.01), colon cancer (p < 0.01), lung cancer (p < 0.01), renal cancer (p < 0.02) and melanoma (p < 0.05).
325
Peng/Murphy
326
Table 2. Comparison of mean PE in capillary versus Petri dish by tumor type Tumor type
Total
Breast
51
Ovary
25
Lymphoma
10
Lung
8
Melanoma
8
Renal
7
Colon
5
Pancreas
5
Endometrium
4
Lymph node
3
Sarcoma Bladder Neck mass Stomach Unknown Brain Thyroid Testis Wilm’s Salivary gland Mesothelioma Bone marrow
2 2 2 2 2 1 1 1 1 1 1 1
Total
Colonies per 15,000 cells plated ( ± SD)
PE per 15,000 cells plated
Range per 15,000 cells plated
Number of no growth
cap.
dish
cap.
dish
cap.
dish
cap.
dish
44.0 (33.9) 57.0 (68.9) 9.0 (7.8) 72.0 (108) 12.0 (5.8) 58.0 (51.3) 79.0 (11.3) 47.0 (4.9) 63.0 (25.2) 5.0 (6.4) 57.0 165.0 23.0 13.0 23.0 40.0 49.0 0 37.0 8.0 0 0
20.0 (24.1) 44.0 (52.3) 8.0 (9.4) 18.0 (24.8) 6.0 (6.4) 17.0 (21.5) 33.0 (15.2) 20.0 (1L7) 45.0 (19.6) 18.0 (5.9) 23.0 9.0 0 27.0 20.0 3.0 84.0 15.0 0 0 0 0
0.029
0.013**
0-225
0-127
10
13
0.038
0.029
0-297
0-147
2
9
0.007
0.005
0-25
0-33
3
6
0.048
0.012**
0-450
0-102
1
4
0.008
0.004**
0-36
0-18
4
4
0.047
0.013
0-131
0-41
1
0
0.053
0.022*
10-163
0-11
0
2
0.031
0.013
0-152
0-13
2
1
0.042
0.030
0-133
0-162
0
2
0.003
0.012
0-8
0-23
1
1
0.038 0.105 0.015 0.005 0.015 0.027 0.032 0 0.024 0.005 0 0
0.015 0.060 0 0.018 0.013 0.002 0.056 0.010 0 0 0 0
0-9 0-226 0-26 0-16 43 40 49 0 37 8 0 0
0-24 0-9 0 0-27 39 2 84 15 0 0 0 0
0 0 0 0 1 0 0 1 0 0 1 1
0 0 1 0 1 0 0 0 1 1 1 1
28
48
143
cer cells clone higher in capillary tubes, which is in con trast to previous data [14]. This difference cannot be explained, but it is not due to the composition of the nutrient medium. In this present report, the success rate for the growth of 139 unselected human tumors was 82.6% in capillar ies compared to 64.7% for the same tumor cells when cultured in Petri dishes (p < 0.001). These higher PEs in capillary cultures were helpful for in vitro drug sensitiv ity testing, since more tumors achieved evaluable growth
and fewer tumor cells were required per drug test. Only 15,000 viable cells per capillary tube produced growth of sufficient colonies for drug evaluation, and only 7 days of culture was required. In our unpublished data, we observed that anticancer drugs were stable in in vitro capillary cloning assay, except nitrogen mustard solu tion, which had been inactivated when incubated at 37 °C for 3 h. The median PE is 4.25-fold greater for the capillary system (i.e., 0.017 for the capillary system and 0.004
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
*p < 0.05; **p < 0.01.
327
Cloning Human Tumors in Capillary Tube vs. Petri Dish
Acknowledgments The authors gratefully acknowledge the support of the NCI (CM77829) and N. Greenberg who supplied the anticancer drugs, and S.H. Woody, Y.F. Lu and D.L. Braet for their technical assistance. We are also grateful to our consortium of consulting surgeons who have provided the tumor tissue which enabled these studies to be conducted.
References 1 Hamburger AW, Salmon SE: Primary bioassay of human tumor
stem cells. Science 1977;197:461-643.
2 Zee-Cheng RK. Cheng CC: Screening and evaluation of anti cancer agents. Methods Find Exp Clin Pharmacol 1988; 10:67— 101. 3 Hanauske AR, Hanauske U, Von Hoff DD: The human tumor cloning assay in cancer research and therapy: a review with clin ical correlation. Curr Probl Cancer 1985;9:1-66. 4 Von Hoff DD: Human tumor cloning assays: applications in clinical oncology and new antineoplastic agent development. Cancer Metastasis Rev 1988;7:357-371. 5 Von Hoff DD, Casper J, Bradley E, Sandbach J, Johnes D, Makuch R: Association between human tumor colony-formationassay results and response of an individual patients’ tumor to chemotherapy. Am J Med 1981;70:1027-1041. 6 Kern DH, Campbell MA, Cochran AJ. Burk MW, Morton DL: Cloning of human solid tumors in soft agar. Int J Cancer 1982; 30:725-729. 7 Selbey P, Buick RN, Tannock I: A critical appraisal of the ‘hu man tumor stem cell assay’. N Engl J Med 1983;308:129-134. 8 Rosenblum ML, Dougherty DV, Deen DF, Wilson CB: Poten tials and limitations of clonogenic cell assay for human brain tumors. Cancer Treat Rep 1981 ;65(suppl 2):61—66. 9 Scheithauer W, Temsch EM, Moyer MP. Grabner G: Search for improved culture conditions for clonogenic growth of human colorectal cancer cells in vitro. Int J Cell Cloning 1987:5:5570. 10 Citron ML. Jaffe ND. Hamburger AW, Lindblad AL. Banda FP, Yenson A, Nathan KA, Cohen MH: Improvement of human tumor cloning assay by suspension of fibroblasts into the bottom layer of agarose. Cancer 1986;57:2357-2362. 11 Eliason JF, Fekete A, Odartchenko N: Improving techniques for clonogenic assays. Recent Results Cancer Res 1984:94:267— 275. 12 Peng X, Chen GZ, Lu YF, Murphy MJ Jr: Capillary cloning of primary human tumor cells: Assay miniaturization for drug effi cacy testing. Int J Cell Cloning 1989;7:322-329. 13 Ali-Osman F, Beltz PA: Optimization and characterization of the capillary human tumor clonogenic cell assay. Cancer Res 1988;48:715-724. 14 Von Hoff DD, Forseth BJ, Mai Huong, Buchock JB, Lathan B: Improved plating efficiencies for human tumors cloned in capil lary tubes versus Petri dish. Cancer Res 1986;46:4012-4017. 15 Moezzi J, Murphy MJ Jr: Experience with the human tumor cloning assay; in Salmon SE, Trent JM (eds): Human Tumor Cloning. Orlando, Grune & Stratton. 1984. pp 229-243. 16 Salmon SE, Young L, Lebowitz J, Thomson S, Einsphar J, Tong T. Moon TE: Evaluation of an automated image analysis system for counting human tumor colonies. Int J Cell Cloning 1984;2: 142-160. 17 Abrams L, Carmeci P, Bull JM, Carbone P: Capillary tube scan ning applied to in vitro mouse marrow granulocyte growth. J Natl Cancer Inst 1973:50:267-270. 18 Maurer HR, Henry R: Colony growth of mouse bone marrow cells in agar contained in glass capillaries. Blut 1979;33:11-22. 19 Maurer HR, Henry R: Drug evaluation on haemopoietic cells in vitro. 1. A micro agar colony assay with semi-automatic optical monitoring. Arzneimittelforschung 1978;28:601-605.
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
for the Petri dish) and is similar to that reported by Von Hoff et al. [14] (0.015 and 0.003, respectively) and previous work in our laboratory [13] (0.018 and 0.004 respectively). Von Hoff et al. [14] compared the PE ratio of different pathological tumor types, and noticed that among 22 tumors tested the ratio was always greater than 1, indicating a superior PE in the capillary tube. We compared the PE of 51 unselected breast can cers which were simultaneously cloned in capillary tubes and in Petri dishes; 72% had a ratio greater than 1. For different histological tumor types, we found that breast, colon, and lung cancer (p < 0.01), and mela noma (p < 0.05) showed the most significant differ ences. These results also correlated well with the report by Von Hoff et al. [14], in which the most significant differences were noted for cancers of the bladder, breast, sarcoma, and unknown samples (p = 0.02 0.07). We also observed that in vitro the number of anti cancer drugs assayed as ‘effective’ (defined in terms of tumor cell ‘sensitive’) was higher in the capillary assay. This may be related to the higher PE with concomitantly more evaluable data which is routinely obtained in the capillary tube versus the Petri dish. Although one report questioned whether tumor colony growth in agar-containing glass capillaries might be an alternative to the Petri dish method for chemosensitivity testing of human malignancies [9], these authors were unable to routinely achieve a growth of 30 colonies per capillary in control cultures. This indicates a significant difference in tech nique, since we routinely average 58 colonies per control capillary tube. The advantages of the capillary cloning system have the potential to circumvent some of the inherent disad vantages of the conventional two-layer agar cloning sys tem. For example, with the capillary assay (1) there is a higher rate of successful colony growth; (2) small tumor samples are evaluable with a higher PE; (3) fewer num ber of viable nucleated cells are required per drug tests; (4) there is shorter turnaround time, and (5) less reagents and incubation space is needed.
328
26 Maurer HR, Ali-Osman F: Tumor stem cell cloning in agar-con taining capillaries. Naturwissenschaften 1981,68:381-383. 27 Von Hoff DD: Plating efficiencies of human tumors in capillar ies versus Petri dishes: in Salmon SE, Trent JM (eds): Human Tumor Cloning. Orlando, Grune & Stratton. 1984, pp 153— 161.
Received: April 2, 1990 Accepted: June 21, 1990 Dr. Martin J. Murphy, Jr. Hippie Cancer Research Center 4100 South Kettering Boulevard Dayton, OH 45439-2092 (USA)
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
20 Von Hoff DD: 'Send this patient’s tumor for culture assay sen sitivity’. N Engl J Med 1983;308:154-155. 21 Agre P, Williams TE: The human tumor cloning assay in cancer drug development. Invest New Drugs 1983;1:33-45. 22 Von Hoff DD: In vitro predictive testing: the sulfonamide era. Int J Cell Cloning 1987;5:179-190. 23 Mirabelli CK, Sung CM, McCabe FL, Faucette LF, Crooke ST, Johnson RK: A murine model to evaluate the ability of in vitro clonogenic assays to predict the response to tumors in vivo. Can cer Res 1988;48:5447-5454. 24 Hug V, Haynes M, Rashid R, Spitzer G, Blumenschen G, Hortobagyi G: Improved culture conditions for clonogenic growth of primary human breast tumors. Br J Cancer 1984;50:207-213. 25 Hamburger AW, White CP, Brown RW: Effect of epidermal growth factor on proliferation of human tumor cells in soft agar. J Natl Cancer Inst 1981;67:825-830.
Peng/Murphy
Cloning of Primary Human Tumors in Capillary Tube versus Petri Dish: A Head-to-Head Comparison Xlange Peng, Martin J. Murphy, Jr. Hippie Cancer Research Center, Dayton, Ohio, USA
Key Words. Capillary cloning • Two-layer agar plating • Human tumors • Plating efficiency
Introduction Since 1977, when Hamburger and Salmon [1] re ported the cloning of primary human tumors in the Petri dish, numerous studies have reported its application in the cloning of primary human tumors for chemosensitivity testing [2-4]. Comprehensive reports confirm that the successful growth rate for unselected primary human tumors is only 35-50%, where ‘successful’ is defined as those tumors forming a sufficient number of colonies for drug evaluation [5, 6]. This promising technique, there fore, requires further research to realize its full potential [7-9]. Various modifications have been made in at tempts to increase the tumor plating efficiency (PE) [10,
11]. Among these, the use of capillary tubes has the potential to circumvent the inherent disadvantages of the two-layer agar system [12-14]. Von Hoff et al. [14] reported that the median PE was 5-fold higher in capil lary tubes than in Petri dishes for the 183 humor tumors examined (18 different histological types). In the same paper by Von Hoff et al. [14], emphasis was placed on cloning tumor cells of breast, ovary and lung origins. The results in this paper confirm the data of Von Hoff et al., and extend the usefulness of capillary cloning for lymphoma, melanoma, and pancreatic can cer cells. In addition, our results taken together with those of Von Hoff et al. show that only stomach carci noma cells clone better in Petri dishes.
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
Abstract. Plating efficiencies (PEs) of primary human tumors cloned in a capillary assay system were compared to those derived from the conventional two-layer agar Petri dish assay system. A total of 143 consecutively received primary human tumors of 24 different pathologies were simultaneously tested in both assay systems. The successful clonal growth rate in the capillary assay was 82.7%, while in the Petri dish it was 64.7% (p < 0.001). The median PE was 0.017% in the capillary assay demonstrating a 4.25-fold increase over the 0.004% PE of the Petri dish system. The data confirmed previous results showing that cancer cells of ovarian, breast, and lung origins clone with higher PE in capillary tubes. In contrast, we confirmed that stomach carcinoma cells were the only tumor type that showed a higher PE with the Petri dish method. In addition, this study shows for the first time that lymphomas and renal cell carcinoma, when they survive in vitro, clone equally well in both methods. However, the capillary cloning method resulted in a 66% success rate for lymphoma cell cloning, but Petri dish cloning resulted in only a 33% success rate. Thus, for some types of cancers (i.e., lymphoma), capillary cloning may be advantageous because it improves the probability of obtaining evaluable results. In other cases, the advantage of capillary cloning may be only the decreased amount of specimen and reagents needed for the assay.
324
Peng/Murphy
Prior to introduction into the two different culture systems, han dling of all samples and cell preparations was performed identically and simultaneously. Collection and Preparation o f Tumor Cells Tumor specimens were obtained incidental to surgery in accor dance with federal and institutional guidelines. The specimens were placed in McCoy’s 5A medium (Gibco, Grand Island, N.Y.) con taining 15% fetal bovine serum (FBS) (Hyclone, Logan, Utah) and preservative-free heparin, and transported to our laboratory within 2-24 h of surgical removal. None of the patients had a history of previous chemotherapy. Preparation o f Single Cell Suspensions A total of 143 consecutively received primary human tumors of 24 different pathologies were simultaneously tested in both assays. Solid tumors were minced into 2- to 5-mm fragments and trans ferred to a trypsinizing flask (Bellco Glass, Vineland, N.J.). An enzyme mixture consisting of 0.2% neutral protease (Cat. No. P0384; Sigma Chemical Co., St. Louis. Mo.), 0.02% DNAse (Cat. No. D-4527 from bovine pancreas. Lot 66F-96601; Sigma) and 0.4% collagenase (Type IA, Cat. No. C-9891, Lot 87F-6843; Sigma) was added at 15 ml/g of tumor tissue [15]. A sterile magnetic bar was placed in the mixture, which was incubated at 37 °C with gentle stirring for 45-60 min and filtered through a 60-pm sterile mesh. If cell clusters of 40 pm or more were present in the culture medium, the mixture was filtered through a 30-pm mesh to eliminate the cell clusters, or the same volume of enzyme mixture was added for a further 30-min digestion, then washed with McCoy’s 5A medium (Gibco) as described previously [12]. Ascitic, pleural and peritoneal effusions were placed in sterile containers with preservative-free heparin (lOU/ml; Elkins-sinn, Inc., Cherry Hill, N.J.). The fluids were centrifuged at 1,500 rpm for 10 min and the cells washed twice with McCoy’s 5A medium. Cell viability was determined in a hema cytometer with trypan blue. Sample viabilities below 30% were excluded from the cloning assay, since at low cell viabilities colonies will not grow. Culture o f Cells in the Two-Layer Agar Petri Dish System Feeder layers consisted of McCoy’s 5A (Gibco) plus 15% heatinactivated FBS (Hyclone), 1 mM sodium pyruvate, 42 pg/ml Lserine, 1 mM L-glutamine, 20mA/4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid (all from Gibco), 9 mg/ml tryptic soy broth, 117 pg/ml L-asparagine and 0.5% agarose (Seaplaque; Marine Col loids, Rockland. Me.). One milliliter of this feeder layer was added to Lux 35 X 10 mm tissue culture dishes (Nunc, Inc., Naperville, III.) and kept at 4 °C for no more than 2 weeks. The upper agar layer consisted of 0.3 % agarose in enriched Con naught Medical Research Laboratories Medium 1066 (CMRL 1066; Gibco), supplemented with 15% heat-inactivated FBS, 3 mM ascor bic acid, 2.4 pg/ml transferrin (Sigma), 2 m M glutamine (Gibco), 2 U/ml preservative-free insulin (Lilly, Indianapolis Ind.), 20 mM 4-(2-hydroxyethyl-l-piperazine-ethanesulfonic acid), 3 pg/ml cata lase (Sigma), 0.004 pg/ml hydrocortisone (Sigma), 100 pg/ml strep tomycin and 100 U/ml penicillin G (Gibco). Tumor cells were sus pended in this plating medium at a final density of 4 X 105 viable cells/ml and 1 ml of the resultant mixture was pipetted onto the feeder layer. After addition of the upper agarose layer containing the
cells, the agarose was allowed to gel at 4 °C for 5 min and then the plates were incubated at 37 °C in a 5% COi humidified atmosphere for 14 days. Colonics (clusters of > 5 0 cells or 60 pm in diameter) usually appeared by day 14 of culture, and the number of colonies in each plate were enumerated with an Omnicon FAS III automated image analyzer (Bausch & Lomb, Rochester, N.Y.) [16], Culture o f Cells in Capillary Tubes Round 10-pl capillary tubes (Clay Adams, Parsippany, N.J.) were cut to 97.5 mm in length, washed for 2 h in an ultrasonic bath with deionized water containing a nontoxic detergent (Liqui-Nox; Alconox Inc., New York, N.Y.) and finally rinsed with five ex changes of deionized water. The cleaned capillary tubes were dryheat sterilized for 2 h before use. Tumor cells were collected and prepared exactly as described above. Cells to be cloned were suspended in 0.2% agarose in the same medium as that for the Petri dish system. The final plating mixture consisted of enriched CMRL 1066 medium (0.6 ml), cell suspension of 3 X 106 viable cells/ml (0.1 ml), drug solution at peak plasma concentration (0.1 ml) and 1% agarose (0.2 ml) [12], (Note that the number of cells/ml is lower in capillary cloning than in Petri dish cloning, because PE is higher, and < 200 colonies per capillary are desirable for accurate counting.) These ingredients were vortexed for thorough mixing, and 50 pi of the resultant mixture (i.e., containing 15,000 viable cells) was introduced into each capillary tube in triplicate, using an Eppendorf pipette. The capillary tubes were then horizontally placed in a capillary cassette specially de signed for this purpose. The agarose was allowed to gel at 4 °C for 5 min and then incubated at 3 7°C in a 5% CCL humidified air atmosphere. Both ends of each capillary tube were left open to the incubator atmosphere. A set of three capillary tubes was examined immediately prior to incubation for the presence of cell clusters (i.e., cell aggregates of 40 pm or greater in diameter. Those capillary tubes which contained more than four clusters were replaced by freshly prepared capillary culture tubes. Control groups were exam ined on the 5th day of incubation to evaluate colony growth, and capillary tube colonies (clusters of > 5 0 cells or > 6 0 pm in diame ter) were counted on the 7th day using an inverted microscope (X 150) except when growth was slow, in which case these capillary tubes were counted on the 14th day of culture. Data Analysis and Statistical Considerations The percentage of cloning efficiency in each system was calcu lated: __ Number of colonies,'vessel PE % = ---------------------------------- X 100 Number of cells plated Student’s paired t tests were used to statistically compare results between different groups.
Results Among the 143 different tumor samples, 4 lacked suf ficient viable cells to attempt culture in the capillary cloning system, rendering a total of 139 evaluable tu mors. In the Petri dish assay, which requires more cells
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
Materials and Methods
Cloning Human Tumors in Capillary Tube vs. Petri Dish
Discussion Capillary tubes have only recently [14] been used for the cloning of primary human tumor cells, although this vessel has been used to clone hematopoietic progenitor cells for the past decade [17-19]. The cloning of primary human tumor cells in Petri dishes has been expanded to predict the chemosensitivity response of individual pa tient tumors to anticancer drugs [20-23], Even though the conventional Petri dish methodology has been widely used. It has inherent disadvantages. For example: (1) only 35-50% of human tumors form enough colonies to enable drug sensitivity evaluation; (2) the PEs are very low, and (3) a large number of tumor cells is required to perform drug sensitivity testing. Since many tumor biop-
Table 1. PE (%) in capillary tube system versus Petri dish1
Mean PE, % Median PE, % Range PE, % Average colonies per 15,000cells plated Median colonies per 15,000cells plated Range of colonies per 15,000cells plated
Capillary
Petri dish
Capillary/ Petri dish
0.0307 0.017 0-3.0
0.0122 0.004 0-0.54
2.52 4.25 NA
56.4
18.3
3.08
6
4.66
28 0-450
0-162
0-2.77
NA » Not applicable. 1 Based on 143 different human primary malignancies of 24 dif ferent histological types.
sies are small, it is often difficult to test many chemo therapeutic drugs. Attempts have been made to increase the Petri dish PE including modification of the culture medium and supplementation of growth-promoting factors [9-11, 24, 25]. None of them, however, were found to be totally satisfactory. Cells have been cultured in square and round capillary tubes with an increase in both the suc cess rate of cloning and in the PEs when compared with Petri dish-derived data [26, 27], We have optimized the culture conditions, cell densi ty, supplemental growth-promoting factors, and the length of incubation [13]. The rate of successful growth in the capillary cloning system is appreciably higher than that of the conventional two-layer agar cloning system. A study from our laboratory reported a success rate of 77% [13], and with further improvements we have recently reported an 88% success rate [12], Von Hoff et al. [14] reported a higher PE for breast, ovarian and lung cancers cloned in capillary tubes. We obtained similar results indicating that this method of cloning can be reproduced in other laboratories. Lymphoma is usually difficult to clone in vitro. In the paper of Von Hoff et al., only 2 cases were reported, and both showed poor growth in Petri dishes and capillary tubes. We obtained a success rate of 66% in capillary cloning and 33% in Petri dish cloning. Of the 33% growth in both systems, a head-tohead comparison yielded a PE that was identical. We also observed that there were no differences in PE for pancreatic and renal cancers cloned in Petri dish or cap illary tube. Of importance is the finding that colon can-
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
than the capillary assay, 7 tumor samples could not be plated due to insufficient viable cells. Therefore, the total number of evaluable tumors plated in the Petri dish system was 136. The successful growth rate in the capillary cloning system was 82.7% (115/139) compared with 64.7% (88/136) in the Petri dish cloning system. The difference between these two cloning systems was significant (p < 0.001). Table 1 details the comparison of PE in the cap illary tube versus the Petri dish. The median number of colonies per 15,000 cells was 28 (mean 47), while the median number of colonies per 15,000 cells per Petri dish was 6 (mean 18). The median PE was 0.017% for tumors cloned in capillary tubes and 0.004% for tumors cloned in Petri dishes (i.e., a 4.25-fold increase in me dian percent of PE with capillary tubes). In a comparison of 51 consecutively received breast cancers, 4 tumors were excluded from evaluation since they did not grow in either culture system. The PE ratio of capillary cloning versus Petri dish cloning was >1 in 34 of the cases (72%), a further indication of the superior PE in capil lary tubes. To determine if there were differential tumor types in which the PEs were the most favorable in the capillarytubes, an analysis of mean PE by tumor type was per formed (table 2). To test the significance of these differ ences for each tumor type, paired t tests were run on the PE for both culture methods. As shown in table 2, the most significance differences were noted for bladder can cer (p < 0.001), breast cancer (p < 0.01), colon cancer (p < 0.01), lung cancer (p < 0.01), renal cancer (p < 0.02) and melanoma (p < 0.05).
325
Peng/Murphy
326
Table 2. Comparison of mean PE in capillary versus Petri dish by tumor type Tumor type
Total
Breast
51
Ovary
25
Lymphoma
10
Lung
8
Melanoma
8
Renal
7
Colon
5
Pancreas
5
Endometrium
4
Lymph node
3
Sarcoma Bladder Neck mass Stomach Unknown Brain Thyroid Testis Wilm’s Salivary gland Mesothelioma Bone marrow
2 2 2 2 2 1 1 1 1 1 1 1
Total
Colonies per 15,000 cells plated ( ± SD)
PE per 15,000 cells plated
Range per 15,000 cells plated
Number of no growth
cap.
dish
cap.
dish
cap.
dish
cap.
dish
44.0 (33.9) 57.0 (68.9) 9.0 (7.8) 72.0 (108) 12.0 (5.8) 58.0 (51.3) 79.0 (11.3) 47.0 (4.9) 63.0 (25.2) 5.0 (6.4) 57.0 165.0 23.0 13.0 23.0 40.0 49.0 0 37.0 8.0 0 0
20.0 (24.1) 44.0 (52.3) 8.0 (9.4) 18.0 (24.8) 6.0 (6.4) 17.0 (21.5) 33.0 (15.2) 20.0 (1L7) 45.0 (19.6) 18.0 (5.9) 23.0 9.0 0 27.0 20.0 3.0 84.0 15.0 0 0 0 0
0.029
0.013**
0-225
0-127
10
13
0.038
0.029
0-297
0-147
2
9
0.007
0.005
0-25
0-33
3
6
0.048
0.012**
0-450
0-102
1
4
0.008
0.004**
0-36
0-18
4
4
0.047
0.013
0-131
0-41
1
0
0.053
0.022*
10-163
0-11
0
2
0.031
0.013
0-152
0-13
2
1
0.042
0.030
0-133
0-162
0
2
0.003
0.012
0-8
0-23
1
1
0.038 0.105 0.015 0.005 0.015 0.027 0.032 0 0.024 0.005 0 0
0.015 0.060 0 0.018 0.013 0.002 0.056 0.010 0 0 0 0
0-9 0-226 0-26 0-16 43 40 49 0 37 8 0 0
0-24 0-9 0 0-27 39 2 84 15 0 0 0 0
0 0 0 0 1 0 0 1 0 0 1 1
0 0 1 0 1 0 0 0 1 1 1 1
28
48
143
cer cells clone higher in capillary tubes, which is in con trast to previous data [14]. This difference cannot be explained, but it is not due to the composition of the nutrient medium. In this present report, the success rate for the growth of 139 unselected human tumors was 82.6% in capillar ies compared to 64.7% for the same tumor cells when cultured in Petri dishes (p < 0.001). These higher PEs in capillary cultures were helpful for in vitro drug sensitiv ity testing, since more tumors achieved evaluable growth
and fewer tumor cells were required per drug test. Only 15,000 viable cells per capillary tube produced growth of sufficient colonies for drug evaluation, and only 7 days of culture was required. In our unpublished data, we observed that anticancer drugs were stable in in vitro capillary cloning assay, except nitrogen mustard solu tion, which had been inactivated when incubated at 37 °C for 3 h. The median PE is 4.25-fold greater for the capillary system (i.e., 0.017 for the capillary system and 0.004
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
*p < 0.05; **p < 0.01.
327
Cloning Human Tumors in Capillary Tube vs. Petri Dish
Acknowledgments The authors gratefully acknowledge the support of the NCI (CM77829) and N. Greenberg who supplied the anticancer drugs, and S.H. Woody, Y.F. Lu and D.L. Braet for their technical assistance. We are also grateful to our consortium of consulting surgeons who have provided the tumor tissue which enabled these studies to be conducted.
References 1 Hamburger AW, Salmon SE: Primary bioassay of human tumor
stem cells. Science 1977;197:461-643.
2 Zee-Cheng RK. Cheng CC: Screening and evaluation of anti cancer agents. Methods Find Exp Clin Pharmacol 1988; 10:67— 101. 3 Hanauske AR, Hanauske U, Von Hoff DD: The human tumor cloning assay in cancer research and therapy: a review with clin ical correlation. Curr Probl Cancer 1985;9:1-66. 4 Von Hoff DD: Human tumor cloning assays: applications in clinical oncology and new antineoplastic agent development. Cancer Metastasis Rev 1988;7:357-371. 5 Von Hoff DD, Casper J, Bradley E, Sandbach J, Johnes D, Makuch R: Association between human tumor colony-formationassay results and response of an individual patients’ tumor to chemotherapy. Am J Med 1981;70:1027-1041. 6 Kern DH, Campbell MA, Cochran AJ. Burk MW, Morton DL: Cloning of human solid tumors in soft agar. Int J Cancer 1982; 30:725-729. 7 Selbey P, Buick RN, Tannock I: A critical appraisal of the ‘hu man tumor stem cell assay’. N Engl J Med 1983;308:129-134. 8 Rosenblum ML, Dougherty DV, Deen DF, Wilson CB: Poten tials and limitations of clonogenic cell assay for human brain tumors. Cancer Treat Rep 1981 ;65(suppl 2):61—66. 9 Scheithauer W, Temsch EM, Moyer MP. Grabner G: Search for improved culture conditions for clonogenic growth of human colorectal cancer cells in vitro. Int J Cell Cloning 1987:5:5570. 10 Citron ML. Jaffe ND. Hamburger AW, Lindblad AL. Banda FP, Yenson A, Nathan KA, Cohen MH: Improvement of human tumor cloning assay by suspension of fibroblasts into the bottom layer of agarose. Cancer 1986;57:2357-2362. 11 Eliason JF, Fekete A, Odartchenko N: Improving techniques for clonogenic assays. Recent Results Cancer Res 1984:94:267— 275. 12 Peng X, Chen GZ, Lu YF, Murphy MJ Jr: Capillary cloning of primary human tumor cells: Assay miniaturization for drug effi cacy testing. Int J Cell Cloning 1989;7:322-329. 13 Ali-Osman F, Beltz PA: Optimization and characterization of the capillary human tumor clonogenic cell assay. Cancer Res 1988;48:715-724. 14 Von Hoff DD, Forseth BJ, Mai Huong, Buchock JB, Lathan B: Improved plating efficiencies for human tumors cloned in capil lary tubes versus Petri dish. Cancer Res 1986;46:4012-4017. 15 Moezzi J, Murphy MJ Jr: Experience with the human tumor cloning assay; in Salmon SE, Trent JM (eds): Human Tumor Cloning. Orlando, Grune & Stratton. 1984. pp 229-243. 16 Salmon SE, Young L, Lebowitz J, Thomson S, Einsphar J, Tong T. Moon TE: Evaluation of an automated image analysis system for counting human tumor colonies. Int J Cell Cloning 1984;2: 142-160. 17 Abrams L, Carmeci P, Bull JM, Carbone P: Capillary tube scan ning applied to in vitro mouse marrow granulocyte growth. J Natl Cancer Inst 1973:50:267-270. 18 Maurer HR, Henry R: Colony growth of mouse bone marrow cells in agar contained in glass capillaries. Blut 1979;33:11-22. 19 Maurer HR, Henry R: Drug evaluation on haemopoietic cells in vitro. 1. A micro agar colony assay with semi-automatic optical monitoring. Arzneimittelforschung 1978;28:601-605.
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
for the Petri dish) and is similar to that reported by Von Hoff et al. [14] (0.015 and 0.003, respectively) and previous work in our laboratory [13] (0.018 and 0.004 respectively). Von Hoff et al. [14] compared the PE ratio of different pathological tumor types, and noticed that among 22 tumors tested the ratio was always greater than 1, indicating a superior PE in the capillary tube. We compared the PE of 51 unselected breast can cers which were simultaneously cloned in capillary tubes and in Petri dishes; 72% had a ratio greater than 1. For different histological tumor types, we found that breast, colon, and lung cancer (p < 0.01), and mela noma (p < 0.05) showed the most significant differ ences. These results also correlated well with the report by Von Hoff et al. [14], in which the most significant differences were noted for cancers of the bladder, breast, sarcoma, and unknown samples (p = 0.02 0.07). We also observed that in vitro the number of anti cancer drugs assayed as ‘effective’ (defined in terms of tumor cell ‘sensitive’) was higher in the capillary assay. This may be related to the higher PE with concomitantly more evaluable data which is routinely obtained in the capillary tube versus the Petri dish. Although one report questioned whether tumor colony growth in agar-containing glass capillaries might be an alternative to the Petri dish method for chemosensitivity testing of human malignancies [9], these authors were unable to routinely achieve a growth of 30 colonies per capillary in control cultures. This indicates a significant difference in tech nique, since we routinely average 58 colonies per control capillary tube. The advantages of the capillary cloning system have the potential to circumvent some of the inherent disad vantages of the conventional two-layer agar cloning sys tem. For example, with the capillary assay (1) there is a higher rate of successful colony growth; (2) small tumor samples are evaluable with a higher PE; (3) fewer num ber of viable nucleated cells are required per drug tests; (4) there is shorter turnaround time, and (5) less reagents and incubation space is needed.
328
26 Maurer HR, Ali-Osman F: Tumor stem cell cloning in agar-con taining capillaries. Naturwissenschaften 1981,68:381-383. 27 Von Hoff DD: Plating efficiencies of human tumors in capillar ies versus Petri dishes: in Salmon SE, Trent JM (eds): Human Tumor Cloning. Orlando, Grune & Stratton. 1984, pp 153— 161.
Received: April 2, 1990 Accepted: June 21, 1990 Dr. Martin J. Murphy, Jr. Hippie Cancer Research Center 4100 South Kettering Boulevard Dayton, OH 45439-2092 (USA)
Downloaded by: Karolinska Institutet, University Library 130.237.122.245 - 1/17/2019 7:15:15 PM
20 Von Hoff DD: 'Send this patient’s tumor for culture assay sen sitivity’. N Engl J Med 1983;308:154-155. 21 Agre P, Williams TE: The human tumor cloning assay in cancer drug development. Invest New Drugs 1983;1:33-45. 22 Von Hoff DD: In vitro predictive testing: the sulfonamide era. Int J Cell Cloning 1987;5:179-190. 23 Mirabelli CK, Sung CM, McCabe FL, Faucette LF, Crooke ST, Johnson RK: A murine model to evaluate the ability of in vitro clonogenic assays to predict the response to tumors in vivo. Can cer Res 1988;48:5447-5454. 24 Hug V, Haynes M, Rashid R, Spitzer G, Blumenschen G, Hortobagyi G: Improved culture conditions for clonogenic growth of primary human breast tumors. Br J Cancer 1984;50:207-213. 25 Hamburger AW, White CP, Brown RW: Effect of epidermal growth factor on proliferation of human tumor cells in soft agar. J Natl Cancer Inst 1981;67:825-830.
Peng/Murphy
