Xenotransplantation of Human Lymphoid Malignancies Is Optimized in Mice with Multiple Immunologic Defects WA Hudson1,Qli1,Cle2 and JH Kersey1

Xenotransplantation of human lymphoid malignancies is optimized in mice with multiple immunologic defects WA Hudson1,QLi1,CLe2 and JH Kersey1

1Cancer Center, Departments of Pediatrics and Laboratory Medicine/Pathology and 2Division of Biostatistics, University of Minnesota, Minneapolis, MN, USA

While it is known that mice with genetic immune defects are phomas.6–11 A mouse strain with another genetic mutation useful for establishing durable engraftment of human tumor involving the Rag-1 gene has been developed12 and most xenografts, the relative role of components of host innate and adoptive immunity in engraftment has not been determined. We recently the SCID mutation was backcrossed into the NOD/Lt directly compared the ability of four strains of genetically strain resulting in an animal with multiple immunologic immunodeficient mice (NOD/SCID, SCID, Nude and Rag-1- defects.13 Several studies have compared growth of human deficient) to successfully engraft and support the human cell tumors in SCID and Nude mice.14–19 To our knowledge, to lines Daudi, Raji, Namalwa and Molt-4 as subcutaneous tumors. date no studies have directly compared xenogeneic transplan- We additionally examined the effect of further immunosuppres- tation in NOD/SCID, SCID, nude and Rag-1 mice. The present sion of the mice by whole body irradiation at a dose of 600 cGy for Nude and Rag-1 and 300 cGy for SCID mice and by adminis- study’s purpose was to compare the ability of mice with tration of anti-natural killer (asialo-GM1) antibody on tumor NOD/SCID, SCID, Nude, or Rag-1 genetic defect to permit growth. Mice with each of the defects supported xenografts to the growth of selected human lymphoma and leukemia cell varying degrees. We found differences in growth character- lines. This study permitted further evaluation of the addition istics in the cell lines tested, with Namalwa consistently pro- of whole body irradiation and additional immunosuppression ducing the largest tumors. With all cell lines studied, optimal with asialo-GM1, an antibody directed against murine natural growth was achieved using NOD/SCID mice. Overall, tumor growth was somewhat enhanced by pretreatment with radiation killer (NK) cells. with little additional benefit from the addition of anti-asialo-GM1 antibody. The importance of multiple components of the innate and adoptive immune system in xenotransplantation were best Materials and methods demonstrated when results in untreated NOD/SCID mice were compared to SCID, nude and RAG-1-deficient mice. The NOD/SCID mouse with or without additional immunosuppres- Animals sion provides the optimal model for the study of the biology and treatment of human leukemias and lymphomas. Four strains of immune-deficient mice were used for these Keywords: NOD/SCID mice; xenografts; leukemia/lymphoma; studies: athymic nude (NCr nu/nu) male mice from the immunosuppression National Cancer Institute; CB-17 severe combined immunodeficient (C.B-17/IcrTac-scidfDF) male mice from Taconic (Germantown, NY, USA); Rag-1 (B6, 129-Rag1tm1Mom) male Introduction and female mice from Jackson Laboratories (Bar Harbor, ME,

USA) were obtained at 5–7 weeks of age; and NOD/LtS2-scid Mice with genetically determined immunodeficiencies are mice were originally obtained from the Jackson Laboratory potentially useful for studying the growth and behavior of and locally bred. All mice were maintained in autoclaved foreign tissues, including human tumors. Since the 1960s, micro-isolator cages in a specific pathogen-free area. mice with the Nude mutation have been the standard for establishing in vivo models of human malignancies. This strain’s value in growing human leukemias and lymphomas Human cell lines has been well described,1,2 but residual immunity can inter- fere with transplantability. A mouse strain with a genetic The human B cell lines, Raji, Daudi and Namalwa (Burkitt’s defect causing severe combined immunodeficiency (SCID) lymphomas), one T cell line, Molt-4 (acute lymphoblastic was developed in the 1980s.3–5 Previous studies from this and other laboratories have confirmed the ability of the SCID leukemia) and the fibrosarcoma HT1080 were obtained from mouse to grow a variety of human leukemias and lym- the American Type Culture Collection, Rockville, MD, USA and placed in culture in our laboratory. The Raji, Daudi and Namalwa lines were maintained in RPMI 1640 (Gibco BRL, Grand Island, NY, USA) with 10% fetal bovine serum (FBS) Correspondence: JH Kersey, University of Minnesota Cancer Center, University of Minnesota, Box 86, 420 Delaware St SE, Minneapolis, from Hyclone (Logan, UT, USA), 50 U/ml penicillin and MN 55455, USA; Fax: 612 626 3069 50 ␮g/ml streptomycin (Gibco). Molt-4 was cultured in the Received 2 October 1997; accepted 9 September 1998 same media, with the addition of 1% sodium pyruvate and Bio-technical methods section (BTS) WA Hudson et al 2030 1% non-essential amino acids (Gibco). HT1080 was main- Tumor measurement tained in Earl’s MEM (Gibco) with 10% FBS. All cultures were ° incubated at 37 C in 5% carbon dioxide. Mice were observed to note the beginning of visible tumor growth. From that point on, measurements of two diameters of the tumor (length and width) were made with calipers every Pretreatment of mice 2 days until the mice were moribund. Total tumor volume was estimated using the previously reported formula length × 1/2 Treatment groups were made up of five to nine mice each. width2.22 The mean tumor volume was calculated for each For precise numbers in each group, see Table 1. Nude and group, and these numbers were arrayed over time. Morbidity Rag-1 mice were pretreated with three doses of 200 cGy in mice given the Namalwa tumor determined the experi- 137 whole body irradiation from a Cs source 1 week apart prior ment’s endpoint. to cell injection. SCID and NOD/SCID mice were given only 100 cGy per week due to their known increased sensitivity to radiation.20,21 Half of the mice receiving radiation were also treated with injections of anti-asialo-GM1 (Wako, Richmond, Statistical analysis VA, USA), a rabbit anti-NK cell antibody. The antibody was administered in doses of 25 ␮l per mouse, i.p., immediately Statistical analysis of results was performed using analysis of prior to cell injection, and again on days 8 and 16 following variance for repeated measures as implemented in the SAS injection. One-third of the mice in each strain was given no system’s PROCGLIM.23 pretreatment.

Results Cell injection

In order to optimize tumor growth, our tumor cells of interest Influence of cell type on tumor growth were mixed with HT1080 fibrosarcoma cells prior to injection.1 The HT1080 cells were trypsinized, spun and Study results are shown graphically in Figure 1 and in tabular resuspended. They were then given a lethal dose of 6000 cGy form in Table 1. As expected, the results demonstrated highly radiation from a 137Cs source and re-suspended in serum-free significant time effects (Wilk’s lambda P = 0.0001). As shown media at a concentration of 50 × 106 cells/ml. Daudi, Raji, in Figure 1 and Table 1, cell type significantly influences cell Molt-4 and Namalwa were checked for viability by trypan growth (P Ͻ 0.0001). Note that Namalwa tumors grew to the blue dye exclusion, spun, and re-suspended at a concen- largest size in every treatment group, and that the scale for tration of 50 × 106 cells/ml. Each cell line was mixed 2:1 with Namalwa is therefore different from that of the other cell lines. HT1080 and injected subcutaneously into the flanks of mice, In general, Raji tumors grew larger than Molt-4 tumors, which for a total dose of 10 × 106 cells of interest and 5 × 106 in turn were larger than Daudi tumors although differences HT1080 cells in 0.3 ml media. between each of the three latter tumors were not significant.

Table 1 Effect of cell type, mouse strain, and additional immunosuppression on mean tumor growth

Mouse Ab + Radiation Radiation No treatment strains

Cell type: Daudi SCID 390.9 (302.9, 478.9) (n = 8) 270.4 (202.6, 338.2) (8) 72.9 (50.2, 95.7) (8) Nude 46.3 (17.2, 75.5) (8) 77.3 (28.4, 126.2) (8) 26.9 (22.9, 30.9) (8) RAG1 24.9 (6.3, 43.6) (5) 57.6 (11.5, 103.6) (6) 82.1 (62.4, 101.7) (6) NSCID 427.1 (328.2, 526.1) (6) 444.3 (223.2, 665.5) (6) 338.6 (182.6, 494.6) (6) Cell type: Molt-4 SCID 807.2 (353.6, 1260) (8) 569.2 (348.2, 790.2) (8) 115.9 (61.9, 169.9) (5) Nude 61.3 (45.4, 77.1) (8) 51 (31.3, 70.8) (7) 39.4 (25.2, 53.6) (9) RAG1 30.7 (19.1, 42.4) (6) 43.3 (16.1, 70.6) (6) 31.3 (21.1, 41.6) (6) NSCID 364.2 (273.7, 454.8) (5) 354.2 (298.7, 409.6) (6) 214.8 (129.5, 300.2) (6) Cell type: Namalwa SCID 2346 (1956, 2736) (8) 1702.3 (1434, 1970) (8) 555.8 (237.3, 874.4) (5) Nude 504.3 (263.5, 745.1) (8) 836.2 (513.1, 1159.3) (7) 394.1 (146.1, 642) (9) RAG1 685.9 (325.4, 1046) (6) 386.3 (141.8, 630.8) (6) 187.7 (25.1, 350.4) (6) NSCID 2885 (2403, 3368) (6) 3784 (2703, 4865) (6) 3688 (2438, 4939) (6) Cell type: Raji SCID 547.2 (385.2, 709.4) (8) 334.3 (237.5, 431.2) (8) 100.3 (62.5, 138.2) (8) Nude 260.9 (124.9, 397.1) (8) 133.4 (56.2, 210.7) (8) 27.2 (11.6, 42.7) (8) RAG1 68.35 (41.2, 95.5) (5) 53.98 (18.8, 89.1) (6) 44.2 (34.6, 53.7) (6) NSCID 511.3 (276.9, 745.7) (6) 363.7 (236.5, 490.9) (6) 433.5 (222.4, 644.8) (6)

Number shown is means of tumor volumes followed by 95% conﬁdence intervals in parentheses. The second parentheses indicates numbers of animals in each group. Bio-technical methods section (BTS) WA Hudson et al 2031

Figure 1 This ﬁgure shows the growth of each of four tumors in NOD/SCID, SCID, Rag-1 and Nude mice; ࡯—࡯, NOD/SCID irradiated plus asialoGM1 antibody; ˿—˿, NOD/SCID irradiated; ࡯—࡯, NOD/SCID no pretreatment; ¼—¼, SCID irradiated plus asialoGM1 antibody; ᭺—᭺, SCID irradiated; ᭹—᭹, SCID no pretreatment; ̆—̆, Nude irradiated plus asialoGM1 antibody; ̅—̅, Nude irradiated; ̆—̆, Nude no pretreatment; *—*, Rag-1 irradiated plus asialoGM1 antibody; ×–×, Rag-1 irradiated; +—+, Rag-1 no pretreatment.

Effect of mouse strain on tumor growth treatments was significant (P = 0.0275). The overall difference is due largely to the inferior performance of the ‘no treatment’ As noted in Figure 1 and Table 1, we found the differences group as the ‘radiation’ and ‘radiation plus antibody’ group Ͻ between the four strains of mice to be highly significant (P were virtually tied (P = 0.9643). This lack of additional anti- 0.0001). Tumors grew larger and at the fastest rate in body effect is very evident in the profoundly immunodeficient NOD/SCID followed by SCID, Nude and finally Rag-1 mice. NOD/SCID mouse, where addition of radiation or antibody In our measurements of mean tumor volume, as seen in Table plus irradiation does not enhance tumor growth in a signifi- 1, tumor size in the three treatment groups of NOD/SCID mice cant way. was largest followed by SCID and then other two strains. Stat- istically, the difference between the mouse strains in determining tumor growth was significant overall (P = Discussion 0.0001), and taken separately, the differences between NOD/SCID and each of the other strains individually were The present study was developed to compare the ability of significant (P = 0.0001). The differences between the genetically compromised mice to support human lymphoma NOD/SCID mice with multiple immune defects compared to and leukemia cell growth. To our knowledge, this represents other mice with more selective immune deficiencies were the first direct comparison of NOD/SCID, SCID, RAG-1 and very significant in untreated mice. nude mice. Because of our interest in the biology and therapy of xenografted human tumors, our focus was to find a model Effect of irradiation and anti-asialo-GM1 on tumor to compare mice with various immunodeficiencies for suc- growth cessful xenotransplantation and maximize durable tumor engraftment and growth of implanted tumors. We examined the effects of additional immunusuppression in Results from these experiments demonstrate, for the first the four groups and discovered that the difference between time to our knowledge, that mice with the Rag-1 mutation do Bio-technical methods section (BTS) WA Hudson et al 2032 support human tumor growth. However, they are less capable immunity13 are the animal of choice for xenotransplanted of supporting human leukemia and lymphoma xenografts than human leukemias and lymphomas. are NOD/SCID and SCID mice and are even somewhat inferior to Nude mice in support of growth of Raji, Daudi, Namalwa or Molt-4 cells. These results suggest that the Acknowledgements immune defects in T and B cells in Rag-1-deficient mice are not as profound as those in SCID mice and even the addition This research was supported in part by an Outstanding Investi- of radiation or antibody treatment does not render Rag- gator Award (CA 49721) from the National Institutes of Health deficient mice profoundly immunodeficient. to John H Kersey. As shown in previous studies, human malignancies will grow in NOD/SCID, SCID and Nude mouse models.14,16,19,24,25 To our knowledge, there has been no previous References comparison of NOD/SCID, SCID and Nude mice in their 1 Faguet GB, Agee JF. Transplantation of human hairy cell leukemia ability to permit growth of Raji, Daudi or Namalwa lymphoma in radiation-preconditioned nude mice: characterization of the cells. Previous studies of human leukemias comparing Nude model by histological, histochemical, phenotypic and tumor kin- and SCID mice demonstrated the superiority of SCID mice in etic studies. Blood 1988; 71: 1511–1517. supporting the growth of BALL-1, Molt-4, CEM and A-1 2 McManaway ME, Marti GE, Tosato G, Liu AK, Al-Nasser AA, human leukemia cell lines.14,18,26 The results of these studies Kiwanuka J, Magrath IT. Heterotransplantation of human Burkitt’s lymphoma cell lines in athymic nude mice: tumor–host relation- were very similar to those of the present investigations which ships. Pathobiology 1993; 61: 164–172. showed the superiority of SCID over Nude mice in the growth 3 Custer RP, Bosma GC, Bosma MJ. Severe combined immunodefi- of each of the cell lines examined. ciency (SCID) in the mouse. Am J Pathol 1985; 120: 464–477. A majority of mice from all three strains successfully sup- 4 Bosma GC, Davisson MT, Ruetsch NR, Sweet HO, Schultz LD, ported the growth of the tumors examined in this study, but Bosma MJ. The mouse mutation severe combined immune there were some differences which could effect their useful- deficiency (scid) is on chromosome 16. Immunogenetics 1989; 29: 54–57. ness in future therapeutic models. First, the strains differed 5 Bosma GC, Custer RP, Bosma MJ. A severe combined immuno- slightly in their rates of tumor uptake. NOD/SCID and SCID deficiency mutation in the mouse. Nature 1983; 301: 527–530. mice engrafted 100% of the tumors implanted. Nude mice 6 Blase L, Daniel PT, Koretz K, Schwartz-Albiez R, Moller P. The had a 97% rate of uptake, and Rag-1 mice engrafted 96% of capacity of human malignant B-lymphocytes to disseminate in tumors. Some mice had their tumors regress to the point SCID mice is correlated with functional expression of the fib- ronectin receptor a5b1 (CD49e/CD29). Int J Cancer 1995; 60: where they could no longer be palpated. In some cases, this 860–866. was permanent, and in others the tumors re-emerged after a 7 Cattan AR, Douglas E. The C.B.17 SCID mouse as a model for period of days. Regression was seen only rarely in NOD/SCID human disseminated leukaemia and myeloma in vivo. Leukemia and SCID mice, but occurred in 17% of Nude mice and 31% Res 1994; 18: 513–522. of Rag-1 mice. 8 Jansen B, Kersey JH, Jaszcz WB, Gunther R, Nguyen DP, Chel- The literature suggests that immunosuppression of the host strom LM, Tuel-Ahlgren L, Uckun FM. Effective immunochemo- therapy of human t(4;11) leukemia in mice with severe combined prior to implantation can improve transplantability of foreign immunodeficiency (SCID) using B43 (anti-CD19)-pokeweed anti- tissues even in mouse strains with genetic immunodefi- viral protein immunotoxin plus cyclophosphamide. Leukemia ciencies. While Nude, SCID and Rag-1 mice lack some or all 1993; 7: 290–297. of their primary cell and antibody-mediated immune 9 Chang H, Leeder S, Cook VA, Patterson B, Dosch M, Minden MD, response, all three strains maintain robust NK cell Messner HA. Growth of human lymphoma cells in SCID mice. 3,12,19 Lymphoma 1992; 8: 129–136. activity. Previous studies have demonstrated the efficacy 10 Jansen B, Vallera DA, Jasczc WB, Nguyen D, Kersey JH. Success- of two methods of residual immune system deactivation: ful treatment of human acute T-cell leukemia in SCID mice using whole body irradiation2,15,27–30 and injection of asialo-GM1 the anti-CD7-deglycosylated ricin A-chain immunotoxin DA7. antibody.24,28,31,32 Some research has suggested a combined Cancer Res 1992; 52: 1314–1321. positive effect with both therapies,30 but this result has been 11 Ghetie MA, Richardson J, Tucker T, Jones D, Uhr JW, Vitetta ES. 18 Disseminated or localized growth of a human B-cell tumor disputed. Our results demonstrate that the use of radiation (Daudi) in SCID mice. Int J Cancer 1990; 45: 481–485. alone or with antibody to asialo-GM1 enhanced the ability of 12 Mombaerts P, Iacomini J, Johnson RS, Herrup K, Tonegawa S, SCID mice to support selected leukemia and lymphoma cell Papaioannou VE. Rag-1 deficient mice have no mature B and T growth. It is notable that radiation alone, even at the lower lymphocytes. Cell 1992; 68: 869–877. levels tolerated by SCID mice, had a positive effect on 13 Shultz LD, Schweitzer PA, Christianson SW, Gott B, Schweitzer IB, Tennent B, McKenna S, Mobraaten L, Rajan TV, Greiner DL, tumor growth. Leiter EH. Multiple defects in innate and adaptive immunologic Since NOD/SCID mice are known to have significant function in NOD/LtSz-scid mice. J Immunol 1995; 154: 180–191. additional immunologic defects beyond those seen in SCID 14 Kawata A, Yoshida M, Okazaki M, Yokota S, Barcos M, Seon BK. mice,13 comparison of this group seemed important. Interest- Establishment of new SCID and nude mouse models of human ingly, NOD/SCID mice showed the best tumor growth and, B leukemia/lymphoma and effective therapy of the tumors with in contrast to results seen in SCID, Nude and RAG-1 mice, immunotoxin and monoclonal antibody: marked difference between the SCID and nude mouse models in the antitumor effi- NOD/SCID mice showed little additional benefit from the cacy of monoclonal antibody. Cancer Res 1994; 54: 2688–2694. addition of radiation or radiation and antibody. The most 15 Taghian A, Budach W, Zietman A, Freeman J, Gioioso D, Ruka W, likely explanation is that NOD/SCID mice are already pro- Suit HD. Quantitative comparison between the transplantability of foundly immunodeficient in multiple aspects of the immune human and murine tumors into the subcutaneous tissue of system and the addition of radiation alone or radiation plus NCr/Sed-nu/nu nude and severe combined immunodeficient mice. Cancer Res 1993; 53: 5012–5017. antibody does not significantly enhance the pre-existing 16 Williams SS, Alosco TR, Croy BA, Bankert RB. The study of human immunodeficiency. In conclusion, NOD/SCID mice with pre- neoplastic disease in severe combined immunodeficient mice. Lab viously demonstrated multiple defects of innate and adoptive Anim Sci 1993; 43: 139–146. Bio-technical methods section (BTS) WA Hudson et al 2033 17 Furukawa T, Kubota T, Watanabe M, Nishibori H, Kuo TH, Sai- of human hematopoietic cells in severe combined immuno- kawa Y, Kase S, Tanino H, Teramoto T, Ishibiki K, Kitajima M. A deficient mice following a single injection of CL2MDP liposomes. suitable model for experimental liver metastasis of human colon Leukemia 1997; 11: 1049–1054. cancer xenografts using mice with severe combined immunodefi- 26 Kamel-Reid S, Letarte M, Sirard C, Doedens M, Grunberger T, ciency. J Surg Oncol 1993; 52: 64–67. Fulop G, Freedman MH, Phillips RA, Dick JE. A model of human 18 Cavacini LA, Giles-Komar J, Kennel M, Quinn A. Effect of acute lymphoblastic leukemia in immune-deficient SCID mice. immunosuppressive therapy on cytolytic activity of immuno- Science 1989; 246: 1597–1600. deficient mice: implications for xenogeneic transplantation. Cell 27 Zubair AC, Ali SA, Rees RC, Goepel JR, Winfield DA, Goyns MH. Immunol 1992; 144: 296–310. Analysis of the colonization of unirradiated and irradiated SCID 19 Xie X, Brunner N, Jensen G, Albrectsen J, Gotthardsen B, Rygaard mice by human lymphoma and non-malignant lymphoid cells. J. Comparative studies between nude and SCID mice on the Leuk Lymphoma 1996; 22: 463–471. growth and metastatic behavior of xenografted human tumors. 28 Huang P, Taghian A, Allam A, Freeman J, Duffy M, Suit HD. The Clin Exp Metastasis 1992; 10: 201–210. effect of whole-body irradiation of nude mice on the tumor trans- 20 Budach W, Hartford A, Gioioso D, Freeman J, Taghian A, Suit plantability and control probability of a human soft tissue sarcoma xenograft. Radiat Res 1996; 145: 337–342. HD. Tumors arising in SCID mice share enhanced radiation sensi- 29 Barry TS, Jones DM, Richter CB, Haynes BF. Successful tivity of SCID normal tissues. Cancer Res 1992; 52: 6292–6296. engraftment of human postnatal thymus in severe combined 21 Fulop GM, Phillips RA. The SCID mutation in mice causes a gen- immune deficient (SCID) mice: differential engraftment of thymic eral defect in DNA repair. Nature 1990; 347: 479–482. components with irradiation versus anti-asialo GM-1 immuno- 22 Tomayko MM, Reynolds CP. Determination of subcutaneous suppressive regimens. J Exp Med 1991; 173: 167–180. tumor size in athymic (nude) mice. Cancer Chemother Pharmacol 30 Sudo K. Improvement of transplantability of human neoplastic 1989; 24: 148–154. cells to nude mice. Jpn J Exp Med 1987; 57: 189–192. 23 Glantz SA, Slinker BK. Repeated Measures. In: Day W, Navrozov 31 Shimizu T, Pelletier H, Hammann A, Olsson NO, Martin MS, Mar- M (eds). Primer of Applied Regression and Analysis of Variance. tin F. Effects of a single injection of anti-asialo GM1 serum on McGraw-Hill: New York, 1990, pp 381–463. natural cytotoxicity and the growth of a regressive colonic tumor 24 Garofalo A, Chirivi RGS, Scanziani E, Mayo JG, Vecchi A, Giav- in syngeneic rats. Int J Cancer 1987; 40: 676–680. azzi R. Comparative study on the metastatic behavior of human 32 Shpitz B, Chambers CA, Singhal AB, Hozumi N, Fernandes BJ, tumors in nude, Beige/Nude/Xid and severe combined immuno- Roifman CM, Weiner LM, Roder JC, Gallinger S. High level func- deficient mice. Invasion Metastasis 1993; 13: 82–91. tional engraftment of severe combined immunodeficient mice 25 Terpstra W, Leenen PJM, van den Bos C, Prins A, Loenem WAM, with human peripheral blood lymphocytes following pretreatment Verstegen MMA, van Wyngaart S, van Rooijen N, Wognum AW, with radiation and anti-asialo GM1. J Immunol Meth 1994; 169: Wagemaker G, Wielengz JJ, Lowenberg B. Facilitated engraftment 1–15.