[CANCER RESEARCH 47, 1724-1730, March 15, 1987) Expression of GD3 in Childhood T-Cell Lymphoblastic Malignancies1 William D. Merritt,2 James T. Casper, Stephen J. Lauer, and Gregory H. Reaman

The Departments of [W. D. M.] and Child Health and Development [G. H. R.J, The George Washington University School of Medicine and Health Sciences, Washington, DC 20037; The Children's Hospital National Medical Center ¡G.H. KJ, Washington, DC 200JO; and The Midwest Children's Cancer Center, Department of Pediatrics, The Medical College of Wisconsin and Milwaukee Children 's Hospital [J. T. C., S. J. L.¡,Milwaukee, Wisconsin 53233

ABSTRACT composition in malignant blasts of six children with acute T- cell lymphoblastic leukemia (T-Cell ALL), one with T-cell Lymphoblasts from seven children with T-cell lymphoblastic malig lymphoblastic lymphoma, and three with non-T, non-B ALL. nancies and three children with non-T, non-B acute lymphoblastic leu We have found that GD3 was a major ganglioside in each of kemia (ALL) were analyzed for ganglioside content. Nonmalignant I - the T-cell malignant lymphoblastic samples studied, whereas cells from thymus served as controls. Both ganglioside and glycoprotein sialic acid were increased approximately 3-3.5-fold in T-cell disease this ganglioside was barely detectable or absent in the non-T, non-B leukemias and normal thymus that we evaluated. compared to thymic tissue when expressed on a per cell basis, but not on a per milligram protein basis. Thin-layer chromatography of the isolated ganglioside fraction from T-cell lymphoblasts revealed two major resor- cinol-positive bands. One ganglioside comigrated with lI'-a-/V-acetylneu- MATERIALS AND METHODS raminosyllactosylceramide (GM3), the major ganglioside in normal lymphoid tissue, and the other ganglioside comigrated with authentic II3- Isolation and Characterization of Lymphocytes from Human Lymph o-A'-acetylneuraminosyl-a2—»8-jV-acetylneuraminosyllactosylceramide oid Tissue and Lymphoblasts from Patients with Acute Lymphoid Ma lignancies. Fresh tonsils and adenoid tissue or lymphomas were ob (GDj) in three different solvent systems. Neuraminidase treatment of the tained immediately after surgical removal; any highly inflammed tissue latter ganglioside yielded (;M

Table 1 Characteristics of lymphoblastic malignancies analyzed PatientDiagnosis1 (yrs)4.51112111635952.5Surfacephenotype"ER*(77%),T-Ag*(70%),CALLA-ER-"(57%),T-Ag*(72%),CALLA-ER*(63%),T-Ag*(70%),CALLA-ER*(85%),T-Ag*(95%),CALLA-ER-(2%),T-Ag*(40%),CALLA-ER*(96%),OKT1 T-ALL2 T-ALL3 T-ALL4 T-lymphoblasticlymphoma5

T-ALL6 T-ALL7 *(97%),B4-,CALLA-ER*(95%),OKT11 T-ALL8 r(97%),B4~,CALLA-ER-.OKT1 ALL9 non-T, non-B l-,B4*,sIg-,CALLA-ER-,T-Ag-,sIg-,CALLA-ER-,T-Ag-,sIg-,CALLA* ALL10 non-T, non-B non-T, non-B ALLSexMFMMMMFFMMAge " KK. percentage of sheep erythrocyte rosette formation at 4'C; T-Ag, percentage of expression of thymocyte antigen; CALLA, expression of common acute lymphoblastic leukemia antigen; OKI 11 and B4, expression of Tl 1 and B4 antigens, respectively; slg, expression of surface membrane immunoglobulin. All samples were positive for terminal deoxytidyl transferase. 1725 Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1987 American Association for Cancer Research. GDj IN MALIGNANT T-LYMPHOBLASTS

TLC and Neuraminidase Treatment of Purified Ganglioside from Malignant T-Lymphoblasts. Based on the migration of the two major gangliosides of the mixed ganglioside fractions in the GM3 and GD3 regions, purified upper and lower doublet gangliosides were tested for comigration with GM3 and GD3 standards in three different solvent systems. As seen in Fig. 3, the two major T-lymphoblast gangliosides chromatographed with identical Rf values to GM3 and GD3, respectively, in each solvent system. The results also indicate that the doublet gan gliosides are indeed single ganglioside species, with differences in the ceramide portion accounting for slight differences in the migration of each band of the doublets. Standard GD3 and the lower doublet ganglioside of T-lym- phoblasts were treated with C. perfringens neuraminidase to compare the ganglioside and neutral neura

Br St Ly Th T-1 T-2 T-3 minidase products of these gangliosides (Fig. 4). Incubation of Fig. 1. Gangliosides of T cd l ALL. Thin-layer chromatogram of gangliosides brain GD3 standard with 150 munits neuraminidase for short from normal human lymphoid tissue and blasts from T-cell acute lymphoblastic times (10-40 min) resulted in loss of GD3 and appearance of leukemia patients. The separation was on 20- x 20-cm silica gel G TLC plates, both GM3 and lactosyl ceramide. Longer periods of time (2 and using solvent system C. Br, total bovine brain gangliosides; St, codeveloped purified ganglioside standards (a, GM3; h, ( ,M.,; c, GM|); Ly. tonsil and adenoid 24 h) were required to hydrolyze GM3 completely to lactosyl lymphocytes; Th, thymocytes; T-l-3, cells from three patients with T-cell ALL; ceramide. Neuraminidase treatment of the ganglioside from T- /:';•.humanred blood cells. A double-band ganglioside (double arrow) is increased lymphoblasts also resulted in glycolipids which migrated with in patients T-l-3 relative to thymus, while GM3 (single arrow) is decreased. For tissue samples and standards, 10 nmol and S nmol ganglioside sialic acid was GM3 and lactosyl ceramide, although this ganglioside was more spotted, respectively. sensitive to neuraminidase than brain GD3: 150 munits treat ment for 10 min resulted in complete hydrolysis of the ganglio side to lactosyl ceramide, 15 munits for 10 min resulted in appearance of GM3 with no lactosyl ceramide, and 50 munits resulted in both GM3 and lactosyl ceramide (not shown). These results are consistent with the preliminary identification of the lower doublet ganglioside of T-lymphoblasts, based on the three-solvent TLC, as GD3. Comparison of Gangliosides of Malignant T-Lymphoblasts to Gangliosides of Non-T, Non-B Lymphoblasts. Gangliosides were extracted from cells obtained from three non-T, non-B ALL patients (two CALLA" and one CALLA*) to determine if high levels of GD3 were common to childhood lymphoblastic leu- kemias and lymphomas in general. The results of the TLC chromatogram of the purified gangliosides showed that the lower double-band ganglioside observed in malignant T-lym phoblasts (GD3) was not found in these non-T, non-B leukemic blasts. Rather, a ganglioside which cochromatographed with the major ganglioside of human red blood cells, SPG, was the major ganglioside other than GM3 in these cells (Fig. 5). Quantitative Scanning Densitometry of Ganglioside Profiles of Childhood Lymphoblastic Malignancies. Results of scanning densitometry of the ganglioside profiles of thymocytes, blasts from six T-ALL patients and one T-cell lymphoma patient, and blasts from three non-T, non-B ALL patients are shown in Fig. 6. ( ¡I) i was a major ganglioside of cells from all seven patients with T-cell disease. In contrast to thymus, in which 97% of the gangliosides was GM3 and 3% was GD3, T-cell lymphoblasts St T-ALL contained from 2 to 77.7% GM3 and from 22 to 86% GD3, with the median ganglioside content of 37% GM3 and 63% Fig. 2. Gangliosides of T-cell ALL. Thin-layer chromatogram of gangliosides GD3. Non-T, non-B ALL, as indicated above, did not contain from a T-cell ALL patient. The separation was on 10- x 10-cm silica gel G HPTLC plates, using solvent system B. ,S"r.codeveloped purified standards (a, detectable GD3, but rather appeared to contain SPG as a major GM3; b, GDj); T-ALL, cells from a patient with T-cell ALL. (Ganglioside sialic ganglioside. acid, S nmol.) Gangliosides of T-Cell Leukemia Cell Lines. Gangliosides were extracted from two cultured ALL cell lines, CCRF-HSB2 cytes but not in red blood cells (Fig. 1) and cochromatographed and RPMI 8402. Unlike the results from patient-derived T- with a purified GD3 standard (Fig. 2). Further characterization lymphoblasts, these cells did not contain increased levels of of the nature of the gangliosides of T-malignant lymphoblasts GD3 relative to GM3 (Fig. 7 and Table 2). Rather, GD3 was followed isolation of the upper and lower doublet gangliosides only a minor ganglioside in HSB2 and was not detectable in from purified total ganglioside by preparative TLC. 8402. These cell lines contained high levels of three other 1726 Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1987 American Association for Cancer Research. GD3 IN MALIGNANT T-LYMPHOBLASTS

4A 3A

GD3 T,

3B St

4B

GD3 'u GM- 3C

St 1 2

Fig. 4. Sialidase treatment of GD3 gangliosides. Thin-layer chromatograms of the products of sialidase hydrolysis of purified brain GD3 (.I) and purified lower band ganglioside from T-ALL cells (B). Separations were on silica gel G HPTLC plates, utilizing the solvent system, CHC13:CH3OH:H:2O, 50:20:3.3 (by volume). In A, GD3 was treated with 1SOmunits sialidase for 10 min (lane 1), 20 min (lane 2), 40 min (lane 3), 2 h (lane 4), or 24 h (lane 5). In B, T-cell ALL ganglioside was treated for 10 min with either ISO munits sialidase (lane 1) or IS munits sialidase (lane 2). Standard lanes include: a, lactosyl ceramide; b, GM3; c. ( .1),.

I u *•*M3 gangliosides which were not observed in the patients with T- cell malignancies.

Fig. 3. Purified gangliosides of T-cell ALL. Thin-layer chromatograms of isolated individual gangliosides of T-cell ALL together with purified standards. DISCUSSION Separations were on silica gel G HPTLC plates, utilizing solvent system A (A), solvent system B (B), and solvent system C (Q. Purified ganglioside standards T. . i_-ui j T • are indicated; Tt, purified lower band ganglioside; r„,purified upper band gan- The results show that childhood T-cell acute lymphoblaStlC giioside.Gangliosidesialicacid(3nmol)wasspottedineachlane. malignancies consistently expressed ( ìDi as the major gangiio- 1727 Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1987 American Association for Cancer Research. GD3 IN MALIGNANT T-LYMPHOBLASTS

side standard was observed in lipid extracts of blasts from two adult T-cell leukemia patients (31). GD3 has been identified by anti-GD3 antibody immunostaining in several preparations of acute and chronic lymphocytic leukemia cells (11), although the cell surface antigen phenotypes were not further subclassi- tied immunologically. In our study of the gangliosides of ten childhood lymphoblastic malignancies, GD3 was expressed in all seven T-cell malignancies, whereas no GD3 was found in the a three non-T, non-B leukemias which were studied. Although very low levels of GD3 that are detectable only by sensitive anti- b GD3 but not resorcinol staining may be present in non-T, non- B ALL, the results suggest that large quantitative differences c in GD3 levels may distinguish these subtypes of childhood lymphoblastic malignancies. d A major ganglioside of the non-T, non-B (CALLA") blasts, besides GM3, was a ganglioside tentatively identified as SPG. SPG was not found exclusively in these non-T blasts, since low amounts were also observed in two of seven T-ALL cells. Westrick et al. (10) observed in one non-T, non-B lymphoblastic leukemia the expression of high SPG by TLC analysis. SPG was also observed as one of two major gangliosides of acute myelomonoblastic leukemia (32). St N-1 N-2 Unlike the results obtained with freshly isolated patient T- cell leukemic lymphoblasts, two cultured T-ALL cell lines did Fig. 5. Gangliosides of non-T, non-B acute lymphoblastic leukemias. Thin- layer chromatogram of gangliosides of T-ALL compared to non-T, non-B acute not contain a high relative concentration of GD3. Rather, only lymphoblastic (CALLA') leukemias. Gangliosides, separated on HPTLC silica trace amounts were observed, in contrast to high levels of three gel G plates in solvent B, were from T-cell ALL cells (7) and from two patients with non-T, non-B ALL (N-1. N-2). Standards (St) are: a, GM3; b, GM2; c, SPG; other gangliosides which were not found in the patient samples. d, GDj. Several explanations are possible for this observation. One is that these two leukemias were established from two patients side. The purified leukemia ganglioside was identified as GD3 with rare T-cell leukemias that can be established in culture and by similar R, values to authentic (il), in three solvent systems, coincidently contained this unique ganglioside pattern. It is well and by the products of neuraminidasc treatment. GD3 has been known that blasts from T-cell ALL are very difficult to establish observed in both childhood and adult leukemias in other labo in culture (33). Since this ganglioside pattern was observed in ratories. In particular, chronic lymphocytic leukemia blasts both cultured T-ALL lines, a relationship between establish were shown by Goff et al. to also contain predominantly GM3 ment of these cells in culture and the ganglioside content is and GD3 (9), unlike chronic myelogenous leukemia cells (30). suggested. Secondly, it is possible that the ganglioside compo In two of three T-cell ALL patients, and one acute myeloblastic sition of T-cell leukemia blasts is altered during establishment leukemia patient studied by Westrick et al. (IO), a double-band of the cells in culture. In this regard, the ganglioside composi ganglioside was visualized on TLC plates between IV3 NeuAc- tion of cells in culture is dependent upon the culture stage (34). nLcOse4Cer (SPG) and VI3 NeuAc-nLcOse6Cer, but this gan Third, these cultured leukemias may represent a stage of leu- glioside was not identified as GD3 in this paper. Among T-cell kemogenesis that is different from any stage of T-cell lymphoid malignancies, GD3 may not be restricted to childhood disease: malignancy in our current patient population, although this is a ganglioside which migrated in thin-layer chromatograms in unlikely due to the relatively large number of T-cell malignan the expected GD3 region compared to a human brain ganglio cies studied. Saito et al. (35), have tested various cultured

•OUHo1Z

Fig. 6. Major gangliosides of T-cell lym phoblastic malignancies compared to thymus and non-T, non-B ALL. The gangliosides iso 60 lated from 10 acute childhood malignancies Oíoo* and pooled thymus tissue were separated by TLC or HPTLC. The plates were sprayed with resorcinol reagent, and the separated ganglio sides were quantitated as to the relative 40200.•_ g20n67 amount of GM3, GD,, SPG, and total other gangliosides using a TLC densitometer. The T-cell and non-T, non-B malignancies are numbered according to patient number from Table 1. m. GM3; •.GD5; B, SPG; M, other gangliosides.

1i 13_¡:;THYMUSEI 12345T-CELL 8 10NON-T. 9

MALIGNANCIESi11In10080*seoi>p840 NON B-ALL 1728 Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1987 American Association for Cancer Research. CD, IN MALIGNANT T-LYMPHOBLASTS

an anti-GDj antibody to investigate the surface expression of 7A GD3 in large numbers of patients with childhood lymphoblastic malignancies. We conclude that the immunological subclassification of childhood lymphoid malignancies appears to correlate with the ganglioside composition, particularly the expression of ( ¡D,. Since GD3 is increased in T-cell malignancies relative to normal b thymocytes, GD3 may be a leukemia/lymphoma-specific alter ation rather than an alteration related to normal T-lymphocyte d ontogeny, although the possibility that GD3 is expressed in very early stem cells of normal bone marrow and again in these malignancies cannot be discounted.

ACKNOWLEDGMENTS

We would like to thank Dr. Robert Yu, Yale University Medical HSB2 8402 St St School, for providing samples of GD3 and GM? standards, and Dr. Warner Henle for performing Epstein-Barr virus analysis of our leu 7B kemia cell lines. We also thank Susan Adams and Phyllis Kirchner (The Medical College of Wisconsin), and Vicki Cipriano (The George Washington University Medical Center) for excellent technical assist ance.

REFERENCES

1. Hakomori, S. in cellular interaction, differentiation, and oncogenesis. in: J. N. Kanfer and S. Hakomori (eds.). Handbook of Lipid Research, Vol. 3, pp. 327-379. New York: Plenum Press, 1983. 2. Hakomori, S. Glycosphingolipids in cellular interaction, differentiation, and oncogenesis. Ann. Rev. Biochem., 33: 733-764, 1981. 3. Hakomori, S., and Kannagi, R. Glycosphingolipids as tumor-associated and differentiation markers. J. Nati. Cancer Inst., 71: 231-251, 1983. 4. Moiré,D.J., Klöppel,T. M., Merritt, W. D., and Keenan, T. W. Giycolipids as indicators of tumorigenesis. J. Supramol. Sinici., 9:157-177, 1978. 5. Merritt, W. D., Richardson, C. L., Keenan, T. W., and Moire, D. J. Gangliosides of liver tumors induced by JV-2-fluorenylacelamide. I. Ganglio side alterations in liver tumorigenesis and normal development. J. Nat). St HSB2 8402 St Cancer Inst., 60: 1313-1326, 1978. 6. Merritt, W. D., Moiré,D.J., and Keenan, T. W. Gangliosides of liver tumors Fig. 7. Gangliosides of IMI cell lines. Thin-layer chromatograms of gan- induced by /V-2-fluorenylacetamide. II. Alterations in biosynthetic enzymes. gliosides of two T-ALL cell lines compared to primary T cell ALL cells. Ganglio J. Nail. Cancer Insl., 60:1329-1337, 1978. sides from CCRF-HSB2 cells (USUI) or RPMI 8402 (8402) cells were separated 7. Pukel, C. S., Lloyd, K. O., Travassos, L. R., Dippold, W. G., Oellgen, H. F., on silica gel G HPTLC plates, and developed in either solvent system A (A) or solvent system B (B). T, cells from a T-cell ALL patient; .Vi.standards (a, GM3; and Old, L. O. GDj, a prominenl ganglioside of human melanoma. Detection and characterization by mouse monoclonal antibody. J. Exp. Med., 155: 6, GM2; c, SPG;

leukemias and malignant lymphomas in leukemic phase. Leuk. Res., 2: 57- noma variant lines selected in vivo for enhanced lung implantation. Cancer 69, 1978. Res., 38: 1336-1344, 1978. 19. Minowada, .1., Ohnuma, T., and Moore, G. E. Rosette-forming human 28. Coakley, W. T., and James, C. J. A simple linear transform for the folin- lymphoid cell lines. I. Establishment and evidence for origen of thymus- Lowry protein calibration curve to 1.0 mg/ml. Anal. Biochem., 85: 90-97, derived lymphocytes. J. Nati. Cancer Inst., 49: 891-895, 1972. 1978. 20. Minowada, J., and Moore, G. E. T-lymphocyte cell lines derived from 29. Wenger, D. A., and Wardell, S. Action of neuraminidase (EC 3.2.1.18) from Clostridium perfringens on brain gangliosides in the presence of bile salts. J. patients with acute lymphoblastic leukemia. In: Y. Ito and R. M. Dutcher Neurochem., 20:607-612, 1973. (eds.). Comparative Leukemia Research, 1973, pp. 251-261. Tokyo: Univer 30. Westrick, M., Lee, W. M. F., and Macher, B. A. Isolation and characteriza sity of Tokyo Press, 1975. tion of gangliosides from chronic myelogenous leukemia cells. Cancer Res., 21. Ledeen, R. W., Yu, R. K.. and Eng, L. F. Gangliosides of human myelin: «.•5890-5894,1983. sialosylgalactosylceramide (G7) as a major component. J. Neurochem., 21: 31. Kuriyama, M., Nomura, K., Tara, M., Matsubara, H., and Igata, A. Glyco- 829-839, 1973. of leukemic cells in adult T-cell leukemia-lymphoma. Biochim. 22. Vance, D. E., and Sweeley, C. C. Quantitative determination of the neutral Biophys. Acta, 834: 396-401, 1985. glycosyl in human blood. J. Lipid Res., 8:621-630, 1967. 32. Lee, W. M. F., Westrick, M., Klock, J. C., and Macher, B. A. Isolation and 23. Seyfried, T. N., Ando, S. and Yu, R. K. Isolation and characterization of characterization of glycosphingolipids from human leukocytes. A unique human liver hematoside. J. Lipid Res., 19: 538-543, 1978. glycosphingolipid pattern in a case of acute myelomonoblastic leukemia. 24. Siddiqui, B., and Malumori. S. A ceramide tetrasaccharide of human eryth- Biochim. Biophys. Acta, 711:166-175, 1982. rocyte membrane reacting with anti-type XIV pneumococcal polysaccharide 33. Karpas, A. Long-term culture of leukaemic cells and the expression of new antiserum. Biochim. Biophys. Acta, 330: 147-155, 1973. viruses: possible role in leukaemogenesis. In: D. Quaglino and F. G. J. Hayhoe (eds.). The cytobiology of leukemias and lymphomas. Serona Sym 25. Svennerholm, L. Quantitative estimation of sialic acids. II. A colorimetrie posia Publications from Raven Press, Vol. 20, pp. 409-429. New York: resorcinol-hydrochloric acid method. Biochim. Biophys. Acta, 24:604-610, Raven Press, 1985. 1957. 34. Langenbach, R., and Kennedy, S. Gangliosides and their density-dependent 26. Hammond, K. S., and Papermaster, D. S. Fluorometric assay of sialic acid changes in control and chemically transformed C3H/10T1/2 cells. Exp. Cell in the picomole range. A modification of the thiobarbituric acid assay. Anal. Res., 112: 361-372, 1978. Biochem., 74: 292-297, 1976. 35. Saito, M., Nojiri, H., Takaku, F., and Minowada, J. Distinctive characteris 27. Yogeeswaren, G., Stein, B. S., and Sebastian, H. Altered cell surface orga tics of ganglioside-profiles in human leukemia-lymphoma cell lines. Adv. nization of gangliosides and sialylglycoproteins of mouse metastatic mela Exp. Med. Biol., ¡52:369-385, 1982.

1730 Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1987 American Association for Cancer Research. Expression of GD3 Ganglioside in Childhood T-Cell Lymphoblastic Malignancies

William D. Merritt, James T. Casper, Stephen J. Lauer, et al.

Cancer Res 1987;47:1724-1730.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/47/6/1724

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/47/6/1724. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 1987 American Association for Cancer Research.