BritishJournul of Haematology, 1981,47,563-569.

Heterogeneous Blast Cell Crises in Philadelphia Negative Chronic Granulocytic Leukaemia

ANDREW HUGHES, B. A. MCVERRY, H. WALKER, K. F. BRADSTOCK,* A. V. HOFFBRAND* AND G. JANOSSY*

Department of Haernatology, University College Hospital, and *Departments of Immunology and Haematology, Roya l Free Hospital, London

(Received 3 March 1980; accepted for publication 9 September 1980)

SUMMARY. A case of Philadelphia negative chronic granulocytic leukaemia (Ph'- CGL) is described showing features only previously demonstrated in Ph'+ disease. These features include: (1) lymphoid blast crisis, determined by morphology and immunological marker analysis; (2) dual blast cell populations that can be dis- tinguished both morphologically and by immunological markers; (3) clonal evolu- tion, as shown by the emergence of chromosome markers and in one of the cell lines a change in membrane phenotype. These changes were apparently associated with the emergence of a relatively drug resistant subclone of leukaemic cells. This study demonstrates that the lymphoid blast crisis of CGL, and its sequelae, can occur in Ph'- cases. It is similar in respect to morphology, enzyme, and membrane markers and responsiveness to vincristine and prednisolone therapy to the lymphoid blast crisis seen in Ph'+ CGL. This suggests that the Philadelphia chromosome is a clonal marker only, and its presence is not directly related to the subsequent clinical course of the disease.

Most patients with Philadelphia positive chronic granulocytic leukaemia (Ph'- CGL) progress to a terminal blastic phase (Canellos, 1977). Immunological membrane marker studies and measurement of terminal deoxynucleotidyl transferase (TdT) have shown that approximately 20% of these blastic transformations have the phenotype of common type acute lymphoblastic leukaemia (CALL), the blast cells demonstrating surface positivity with anti-ALL serum and showing the presence of surface Ia-like (p28,33) antigens and the intranuclear enzyme TdT; ALL+, Ia+, TdT+ (Marks et al, 1978; Janossy et al , 1979). They also respond to anti-lymphoid treatment, vincristine and prednisone (Marks et al, 1978; Janossy et al, 1979). About 5510% of patients with CGL are Ph'- (Rowley, 1978) and in adults this is often atypical and carries a worse prognosis than Ph'+ disease (Ezdinli et al, 1970; Canellos et al , 1976). Up to now only myeloblastic type acute transformations have been described with Ph'- CGL. We now describe a patient with Ph'- CGL who developed blast transformation of the lymphoid type. In addition, this patient showed evidence of two distinct blast cell types, and clonal evolution in one of theni with the appearance of chromosome markers and a change in membrane

Correspondence: Dr Andrew Hughes, The Katharine Dormandy Haemophilia Centre and Haemostasis Unit, The Royal Free Hospital, Pond Street, Hampstead, London NW3 2QG.

0007-1048/81/0400-0563$02.00 0 1981 Blackwell Scientific Publications

563

564 Atidrew Hughes et al

phenotypc. This was associated with the emergence of a relatively drug resistant subclone of leukaemic cells.

CASE REPORT

A 55-year-old male was admitted to hospital with a 3 month history of fatigue, weight loss, and breathlessness. O n examination he was pale, had minimal ankle oedema, 6 cm hepato- megaly, and 20 cm non-tender splenomegaly. Investigations were as follows: Hb 8.4 g/dl, WBC 264 x 10”/1 (neutrophils 46%, myelocytes 37%, blasts l%), platelets 315 x 109/1, leuco- cyte alkaline phosphatase score 0. Bone marrow aspiration showed marked myeloid hyper-. plasia with a preponderance of neutrophils and myelocytes and <5% blasts (Fig 1). Chromo- some analysis showed normal male karyotype and no evidence of a Philadelphia abnormality. He was diagnosed as Philadelphia negative chronic granulocytic leukaemia, and treatment with continuous low dose busulphan (4 mg daily) was begun. Clinically he was much improved, though haematological response was slower. After 1 year of treatment his WBC was 9.3 x 10y/l and his spleen size had decreased to 3 cm below the costal margin. 14 months after diagnosis his WBC began to rise, and his spleen size increased. Bone marrow aspiration showed a homogeneous population of small blasts (Fig 2). Chromosome analysis was normal as before. He was treated with vincristine and prednisolone, which readily induced a bone marrow remission, and was maintained on 6-mercaptopurine and methotrexate. This was followed 6 months later by a relapse of acute disease. Bone marrow aspiration showed a dual population of small and large blasts (Fig 3). Chromosome analysis now showed the presence of three marker chromosomes in 64% of the cells examined, but no Philadelphia abnormality (Fig 5). He was treated with vincristine, prednisolone and razoxane, and a further bone marrow remission was induced. He was maintained on razoxane for 24 months before a second relapse supervened. Although his bone marrow still showed a dual blast cell population, it was dominated by the smaller blasts (Fig 4). The marker chromosomes were now present in 90% of the cells examined. Again no Philadelphia abnormality was noted. He was re-treated with vincristine, prednisolone and razoxane, and a partial bone marrow remission was maintained for a further 3 months. A third frank relapse then occurred. The majority of the blasts present in the bone marrow were of the small type, though some large blasts were still present. The marker chromosomes were now seen in all cells examined, and he remained Philadelphia negative. No further attempt at active treatment was made, and he died 30 months after the original diagnosis had been made.

MATERIALS AND METHODS

Studies were performed on samples of fresh bone marrow. Romanowsky, Sudan black, and periodic acid-Schiff (PAS) stains were performed by standard techniques. Leucocytes were prepared by Ficoll-Isopaque density gradient centrifugation. The leukaemic cells were charac- terized by a panel of membrane and enzyme markers. The membrane markers used included: anti-ALL serum (made against the common non-T non-B form of ALL (Greaves et at, 1975), and anti-Ia-like serum made against purified human p28,33 antigens (Janossy et al, 1977b). Terminal deoxynucieotidyl transferase (TdT) was detected by both immunofluorescence and a

Blast Cell Crises in Ph'- CGL

FIG 1. Bone marrow aspiration at presentation showing predominantly neutrophils and myelocytes.

FIG 2. Bone marrow aspiration at initial blast crisis showing small, compact blast cells of lymphoid morphology.

(Facing p . 564)

Aildrew Hughes et a1

FIG 3 Bone marrov acpiration at first relapse thoniiig both small (s) and large (4) blasts.

FIG 4. Bone rnarroLv aspiration a t second relapte, thowing predominantly small blasts and occasional larger blast cell (w).

Blast Ce l l Crises in Ph'- CGL 565

biochemical technique (Bollum, 1975; H o a r a n d et al , 1977). Direct bone marrow prep- arations for chromosomal analysis were made by a modification of the technique of Tjio & Whang (1962). Giemsa banding techniques were performed using a modification of the Summer method (Summer et al, 1971). Slides were incubated at 60°C for 90 min in 2 x SSC, followed by rinsing in distilled water and Giemsa staining.

RESULTS

At presentation this patient showed the clinical features, peripheral blood and bone marrow findings typical of CGL. Chromosome analysis revealed a normal karyotype and no Ph' abnormality. His subsequent clinical course was characterized by an acute blastic tranforma- tion followed by three clinically distinct relapses. The morphological features of these blast crises can be seen in Figs 2-4. In the initial transformation the blasts were a homogeneous population of small cells with little cytoplasm, giving the appearance of lymphoblasts (Fig 2). The first relapse showed two distinct blast cell populations, one small as before and the other large cells with abundant cytoplasm giving the appearance of myeloblasts, in equal propor- tions (Fig 3). The second and third relapses also showed two populations with the small cell type showing increasing predominance (Fig 4). Throughout, all blast cells were cytochemi- cally unreactive (Sudan black and PAS negative).

Immunological marker analysis results are shown in Table I. They demonstrate that the first blast crisis was homogeneous and lymphoid in type (ALL+, Ia+, TdT+), while the subsequent three relapses were heterogeneous, the small cells being predominantly TdT positive and the larger one negative (Fig 6). Biochemical analysis showed markedly raised bone marrow TdT in all four crises (Table I). There was also evolution of membrane phenotype, with the loss of ALL positivity in the small blasts in the third relapse. Chromosomal analyses are shown in Table I. They demonstrate that the chromosomal karyotype in the first blast crisis was normal, but the three subsequent relapses showed an increasing proportion of cells with marker chromosomes (Fig 5). In the preparation obtained during the first relapse, one B and one C group chromosome was missing and three marker chromosomes were present which could not be identified. These could have represented the missing chromosomes with rearrange- ments and/or deletions of the genetic material, or could possibly represent loss of normal chromosomes and the de novo appearance of abnormal additional chromosomes. At no time could the typical Philadelphia abnormality (22q-) or translocation (9qf,22q-) be demon- strated using banding techniques (Fig 5).

DISCUSSION

This study shows several interesting features. Firstly, to our knowledge, this is the first demonstration of lymphoid blast transformation of Phi- CGL, though an incompletely documented case has been reported by French (1979). This shows that the Philadelphia chromosome is not exclusively associated with lymphoid transformation during a pre- dominantly myeloid disease. Lymphoid blast crisis in Phi- CGL is, however, in all other respects similar to that seen in Phl+ disease, in that it shows the same membrane phenotype and enzymatic marker and is sensitive to treatment with vincristine and prednisolone (Marks et al,

FIG 5. 5 and (22q-

Atidrew Hughes et a1

5

I I i

5

K 34,

12 X

Chromosome preparation (Giemsa banded) showing the two absent chromosomes from pairs 6, and the additional three marker chromosomes (1). There is no Philadelphia abnormality ,9q+) present.

T A B l t I. Morphology, immunological marker and chromosome analysis on bone marrow specimens in four blast crises of Ph'- chronic granulocytic leukaemia

Immiitrolqiralt Morphology* nietiibrarie

markers Biochemical Chromosomes Sttiall Large TdT blast, blasts ALL+ TdT+ (U/7@ cells) Karyotype Abnormal$

Blast cri\is I >95 < 3 83s 83% 76 46,XY 0

Blast crisis I 1 5W 50 ND 4F 38 46,XY 47,XY+markers 64

Blast crisis 111 80 20 ND 50 78 46,XY 47,XY+markers 90

Blast crisis IV 95 5 0 50 164 47,XY+markers 100

ND = Not done. Figures represent percentages of blasts showing the appropriate morphology.

t Figures represent percentages of positive cells within the blast cell population $ Figures repres

Blast Cel l Crises in Ph'- CGL 567

FIG 6. Bone marrow preparations in first relapse. The small blasts (seen in the phase contrast photomicrograph on the left) are positive with both the anti-ALL serum (centre) and TdT inimuriofluorescence (right). The large blast shown is negative.

1978; Janossy et al , 1979). Two studies have shown that Ph'- CGL in adults carries a worse prognosis than Phl+ disease (median survivals of 14 and 8 months as against 44 and 40 months respectively) (Ezdinli et al , 1970; Canellos et a l , 1976). Although our patient's response to chronic phase therapy was slow, and his first blast crisis occurred only 14 months from the time of diagnosis, he survived overall for 30 months. This may only reflect upon the lymphoid component of this particular leukaemia that was responsive to the relatively non-myelosup- presive therapy used.

Secondly, we have evidence to show the presence of two distinct blast cell types; one small with little cytoplasm and a lymphoid phenotype (ALL+, Ia+, TdT+), the other large with abundant cytoplasm and showing no reactivity with the markers for lymphoid type leukaemic blasts cells (Fig 6). Dual blast cell populations have been described before in both acute leukaemia (Mertelsmann et al, 1978; Prentice et al, 1980); and in the blast crisis of Ph'+CGL (Forman et al , 1977; Janossy et al, 1978). The present study demonstrates that this phenomenon can also occur in Ph'- disease.

Thirdly, we have evidence from the chromosome studies for clonal evolution throughout the course of this patient's blast crisis and subsequent relapses. Drug resistant leukaemic subclones have been previously described in acute lymphoblastic leukaemia (ALL) (Goldstone et af, 1979), and a similar situation is seen in our patient. Initially the chromosomally normal lymphoid blast cells were sensitive to vincristine and prednisolone, and a remission was readily established. The appearance of chromosome markers in the lymphoid blast cell line during his subsequent relapses was associated with relative drug resistance and increasing difficulty in establishing and maintaining further remissions. That the chromosome markers lay in the smaller lymphoid blasts was indicated by the fact that in the second and third relapses almost all the cells examined had abnormal chromosomes when the dominant cell was the small

568 Aildrew Hughes et a1

lymphoid type blast. Interestingly, there was also evolution of membrane phenotype at the latter stages of the disease. During the third relapse small lymphoid cells were observed which, though they still exhibited the nuclear TdT enzyme, failed to react with the anti-ALL serum (ALL-, TdT+). The development of drug resistant subclones with ALL- phenotype is seen in approximately 10% of cases of relapsing childhood ALL (Greaves et al , 1980) and are apparently more frequent in relapsing cases of Ph'+ lymphoid blast crisis (Janossy et al, 1977a). In summary, Ph'- CGL can behave in a very similar fashion to Phi+ disease. The similarities include the features of lymphoid blast crisis, dual blast cell populations, and the evolution of relatively drug-resistant leukaemic subclones. The Philadelphia chromosome appears to be a clonal marker only, and its presence is not necessarily directly related to the subsequent clinical course of the disease.

ACKNOWLEDGMENT

We thank Dr F. J. Bollum for the anti-terminal transferase antibody, Mr Peter Luton for the photography, and Mrs R. E. Wenzerul for typing the manuscript.

REFERENCES

BOLIUM, FJ. (1975) Antibody to terminal deoxynuc- leotidyl transferare. Proceedings ofthe National Aca- demy of Sciences of the United Stares of America, 72, 41 19-4122.

CANELLOS, G.P. (1977) The treatment of chronic granulocytic leukaemia. Clinics i n Haematology, 6 ,

CANELLOS, G.P., WHANG-PENG, J. & D E V ~ A , V.T. (1976) Chronic granulocytic leukemia without the Philadelphia chromosome. American _lourrial i$

Clinical Pathology, 65, 467-470. EZDINLI, E.Z., SOKAL, J.E., CROSSWHITE, L. & SAND-

BERG, A.A. (1970) Philadelphia chromosome posi- tive and negative chronic myelocytic leukaemia. Annals of Internal Medicine, 72, 175-182.

BARKER, B.E. S. FARNES, P. (1977) Ph' positive leukemias. Spectrum of lymphoid-myeloid expres- sions. Blood, 49, 54F558.

FRENCH, E.A. (1979) Lymphobtastic transformation in Ph'- chronic granulocytic leukaemia. Trent Region Adult and Childhood Leukaemia Groups Annual Combined Meeting, 1979 (Abstract).

GOLDSTONE, A.H., MCVERRY, B.A., JANOSSY, G. & WALKER, H. (1979) Clonal identification in acute lymphoblastic leukemia. Blood, 53,892-898.

GREAVES, M.F., BROWN, G., RAPSON, N.T. & LISTER, T.A. (1975) Antiserum to acute lymphoblastic leukaemia cells. Clinical Immurrology and Immuno- pathology, 4,67-84.

GREAVES, M . F . , PAXTON, A., JANOSSY, G., PAIN, C., JOHNSON, S. ti LISTER, T.A. (1980) Acute lympho-

( l) , 113-128.

FORMAN, E.N., PADRE-MENDOZA, T., SMKH, P.S.,

blastic leukaemia associated antigen. 111. Alterations in expression during treatment and in relapse. Leuk- aemia Research, 4, 1-14.

HOFFBRAND, A.V., GANESHAGURU, K., JANOSSY, G., GREAVES, M.F., CATOVSKY, D. & WOODRUFF, R.K. (1977) Terminal deoxynucleotidyl transferase levels and membrane phenotypes in diagnosis of acute leukaemia. Lancet, ii, 52e523.

JANOSSY, G., GOLDSTONE, A.H., CAPELLARO, D., GREAVES, M.F., KULENKAMPFF, J., PIPPARD, M. & WELSH, K. (1977b) Differentiation linked expres- sion of p28,33 (Ia-like) structures on human leukae- mic cells. British Journal of Haematology, 37,

JANOSSY, G., ROBERTS, M., GREAVES, M.F., WOOD- RUI.F, R., PIPPARD, M., PRENTICE, G. & HOFFBRAND, A.V. (1977a) Lymphoid blast crisis in chronic mye- loid leukaemia and Philadelphia positive acute lym- phoid leukaemia. Bollettino dell'lstituto Sieroterapico Milanese, 57,355-359.

JANOSSY, G., WOODRUFF, R.K., PAXTON, A., GREAVES, M.F., CAPELLARO, D., KIRK, B., INNES, E.M., EDEN, O.B., LEWIS, C., CATOVSKY, D. & HOFFBRAND, A.V. (1978) Membrane marker and cell separation studies in Ph' positive leukemia. Blood, 51,861-877.

JANOSSY, G., WOODRUFF, R.K., PIPPARD, M.J., PREN- TICE, G., HOFFBRAND, A.V., PAXTON, A., BUNCH, C. & GREAVES, M.F. (1979) Relation of 'lymphoid' phenotype and response to chemotherapy incor- porating Vincristine-Prednisolone in the acute phase of Ph' positive leukaemia. Cancer, 43, 426-434.

391-402.

Blast Cell Crises in Ph’- CGL 569

MARKS, S.M., BALTIMORE, D. 81 MCCAFFREY, R. (1978) Terminal transferase as predictor of initial responsiveness to Vincritine-Prednisolone therapy in blastic chronic myelogenous leukemia. New Eng- landJournal ofMedicine, 298,812-814.

MERTELSMANN, R., MERTELSMANN, I., KOZINER, B., MOORE, M.A.S. & CLARKSON, J. (1978) Improved biochemical assay for terminal deoxynucleotidyl transferease in human blood cells: results in 89 adult patients with lymphoid leukemias and malignant lymphomas in leukemic phase. Blood, 51,

PREMICE, A.G., SMKH, A. 81 BRADSTOCK, K.F. (1980) 105 1-1056.

Mixed lymphoblastic-myeloblastic leukemia in treated Hodgkin disease. Blood, 56, 12%133.

ROWLEY, J.D. (1978) Chromosomes in leukaemia and lymphoma. Seminars in Haematology, 15, (3) ,

SUMMER, A.T., EVANS, H.G. & BUCKLAND, R.A. (1971) New technique for distinguishing between human chromosomes. Nature: New Biology, 232,

TJIO, J.H. & WHANG, J. (1962) Chromosome pre- paration of bone marrow cells without prior in vitro culture or in vivo colchicine administration. Stain Technology, 37, 17-20.

301-319.

31-32.