Granulocyte function in myeloblastic leukaemia.

A study of granulocyte function in myeloblastic leukaemia is reported. Function was assessed by the ability of peripheral blood granulocytes to ingest and kill Candida albicans in bitro. Depressed cidal activity was observed in 11 patients with smouldering leukaemia and in 19 patients with acute myeloid leukaemia. Cidal activity was lowest in the untreated acute disease; this improved during cytoreduction therapy and was maintained when remission occurred. Leukaemic plasma depressed the function of control granulocytes; the possible role of a plasma "factor" is discussed.

Summary.-A study of granulocyte function in myeloblastic leukaemia is reported. Function was assessed by the ability of peripheral blood granulocytes to ingest and kill Candida albicans in vitro. Depressed cidal activity was observed in 11 patients with smouldering leukaemia and in 19 patients with acute myeloid leukaemia. Cidal activity was lowest in the untreated acute disease; this improved during cytoreduction therapy and was maintained when remission occurred. Leukaemic plasma depressed the function of control granulocytes; the possible role of a plasma " factor " is discussed.
THE NATURAL history of acute leusixth MRC trial, but had received no therapy kaemia is characterized by the develop-for 7 days before the assay. ment of frequent and severe infections, The term " smouldering leukaemia " (SL) and infection is a major cause of death (Drayfus, Sultan and Manonni, 1970) was (Hersch et al., 1965). Various theories applied to those patients with abnormalities have been proposed to explain the in-of the peripheral blood and/or bone marrow creased incidence of infection. Hirschen in which the percentage of blast cells in the ceres incidencastheofinfection.epo Hirse marrow, whilst similar in morphology, was berg (1939) was the first to report impaired substantially less than in AML and did not phagocytosis in acute leukaemia, an account for more than 30 % of the total observation that has since been confirmed marrow cells. Most of these patients were by many workers. anaemic, thrombocytopenic and many granu- The purpose of this study was to locytopenic. Two patients were receiving examine granulocyte function in myelo-daily prednisolone and blood transfusions blastic leukaemia at various stages of the when indicated. The remainder were not disease. The results are presented and the receiving chemotherapy or immunotherapy. findings discussed.
Control samples were obtained from healthy hospital personnel.
The effect of drug therapy on granulocyte PATIENTS AND METHODS function was assessed by incubating sepa-Patients and controls.-All patients studied rately control cells with hydrocortisone had an unequivocal diagnosis of myeloblastic (30 ,tg/ml), daunorubicin (3 ,ug/ml) and cytoleukaemia on the basis of clinical, peripheral sine arabinoside (0.6 ,ug/ml).
blood and bone marrow examination. The Student's " t " test was employed to term acute myeloid leukaemia (AML) emassess differences between the mean scores. braces all the morphological forms. Patients Assay.-The technique for assay was that were investigated at presentation, during described by Lehrer and Cline (1969) as induction therapy and when in remission. modified by Goldman and Th'ng (1973). Those receiving drug therapy had no chemo-Candida albicans was maintained on therapy for 5 days before the assay. Patients glucose-peptone agar at room temperature.
in remission were concurrently receiving Only cultures containing more than 95% of immunotherapy with weekly injections of viable cells were used for the assay. Equal

autologous plasma and Hanks' balanced salt
The values obtained in granulocytic solution (HBSS) were placed in sterile conleukaemia are shown in Table I. Patients tainers and a suspension of Candida albicans with both smouldering leukaemia and added to provide a neutrophil : Candida ratio AML had a significantly reduced index of 1 : 1. Each estimation was performed compared with control subjects. The with control leucocytes and a blank tube cmaedw it control subjets th containing plasma HBSS and Candida albi-index was most impaired in patients with cans, but no leucocytes. untreated AML (3.1 i 0t6%); but im-Tubes were incubated at 37°C for 1 h. proved during induction therapy (6.6 ± After 30 min a small sample was withdrawn 1-5%), and later when in remission to confirm that all the added organisms had (8.7 + 1.1%). The difference in Candibeen ingested. At 1 h Na deoxycholate dacidal index between the 4 groups was (pH 8.7) was added to lyse the leucocytes and statistically significant (Table II). 5 min later methylene blue (0.25%) was The effects of leukaemic plasma on added and the suspension centrifuged. From normal granulocyte function are shown a wet preparation of each cell sediment 500 in Table III. In all groups, the incubation Candida cells were counted (dead cells stain of normal granulocytes with leukaemic a uniform intense blue). The Candidacidal plasma significantly depressed the index. index (%) was calculated as follows: pa wit unt reated the the In patients with untreated AML, the Total number dead cellsaddition of normal plasma to leukaemic number dead in control tube. cells significantly increased the index.

5
Cytosine arabinoside and hydrocortisone did not depress the Candidacidal RESULTS index in control subjects.-The index fell Sixteen control subjects and 30 patients to 8% when cells were incubated with with leukaemia were studied. Nineteen daunorubicin. patients had AML (7 untreated, 6 were receiving drug therapy and 6 immuno-DISCUSSION therapy) and 11 patients smouldering Various in vitro techniques are availleukaemia.
able to assess granulocyte function. The The Candidacidal index in control method used in this study was chosen subjects was 12-4 ± 2.4% (range 10-18) because of the reproducibility of the test and was highly reproducible. and because Candidiasis is the most  6-6±1i4 8-5±2i7 1.9 11 8±12 8 8±1 6 3.0* AML (immunotherapy) 6 8 7±1i1 9 4+1i0 0 7 12 0±16 10 3±2i1 1 7* *P = 0 * 05 two-tailed. frequent fungal complication in leukaemia range, implying that the cell is still (Bodey, 1966). Because of the ratio of functionally abnormal and may support cells to organisms employed, the test is of the view that granulocytes in AML in value in assessing cidal but not phagocytic remission are derived from persistently properties of granulocytes.
abnormal stem cells. Hudson and Wilson Granulocyte function in leukaemia has (1960) also observed that leucocytes from been studied extensively. The observapatients in remission had decreased baction that granulocytes obtained from tericidal activity. This functional abnorpatients with AML have an impaired mality may be due to persistent enzyme ability to kill Candida albicans confirms defects within the cell; such defects have the results of previous workers (Lehrer been observed by other workers (Skeel, and Cline, 1971;Goldman and Th'ng, Yankee and Henderson, 1971;Strauss 1973). There is, however, little work on et al., 1970). granulocyte function in smouldeting leu-Other factors not specific to leukaemia kaemia, though these patients also show a may depress the Candidacidal index, e.g. high incidence of infection. In these iron deficiency (Chandra, 1973) and abpatients, specific therapy is not normally normal immunoglobulin status (de Cree indicated during the " smouldering " et al., 1974). There was no evidence of phase of the disease and transfusion iron deficiency in the patients studied and requirements are often minimal. Bacin most instances immunoglobulin levels terial infections, however, are common were within the normal range. Compleand impaired bactericidal activity has ment levels were not assessed but Fahey been demonstrated in vitro (Lehrer et al., and Boggs (1960) found no consistent 1973). This study provides further eviabnormality in complement levels in dence of defective granulocyte function. acute leukaemia. The presence of anti- The functional capacity of the granulocyte bodies against Candida albicans would not in SL is greater than that observed in the produce a low index as the anti-Candida acute phase of the disease. The Candidaantibody does not kill or lyse Candida cidal index in the 2 patients in whom the organisms in vitro (Lehrer and Cline, acute phase was imminent fell to that 1969). We found no evidence that blast observed in untreated AML, suggesting cells or immature granulocytes had sigthat there is a gradual impairment of nificant cidal activity, confirming the function as the disease progresses.
observations of other workers (Silver et al., In AML granulocyte function im-1957;Lichtman and Weed, 1972). While proved significantly during induction daunorubicin was found to depress normal therapy (Table II) and this was main-granulocyte function, because of the tained when remission occurred (Table I). interval between administration and per-The index did not return to the normal forming the assay (10 days) the drug residue remaining in the body was negligible and unlikely to invalidate the assay. Thus, the observed defect in granulocyte function must reflect an intrinsic abnormality.
While previous studies have generally been concerned with the granulocyte, little attention has beeil paid to extrinsic factors. The effects of leukaemic plasma on normal granulocyte function was first noted by Braude, Feltes and Brooks (1954) and later by Goldman and Th'ng (1973).
Plasma from patients with acute AML depressed the cidal capacity of normal granulocytes. The explanation of this phenomenon must remain a matter for conjecture for the present. There is some evidence to suggest that it be due in part to a factor released from the leukaemic blast cell. The Candidacidal index in patients with untreated AML improved when granulocytes were incubated with control plasma (Table III) but not when remission occurred. There was no evidence that granulocyte function in smouldering leukaemia was related to the number of blast cells in the blood. Mir and Bobinski (1975) observed that a blast cell extract impaired the sodium " pump " of normal red blood corpuscles. These observations suggest the presence of an enzyme inhibitor, probably a protein, which may have protean effects on cell function.
Leukaemic plasma from patients with AML in remission and with smouldering leukaemia also depressed normal function. Patients in remission had received several blood transfusions and weekly injections of BCG and allogeneic blast cells; the resulting antibody production may have impaired normal function. Patients with smouldering leukaemia, however, were receiving neither blood transfusions nor specific drug therapy. The ability of plasma to depress normal function in these patients may be due to the development of an abnormal protein which further impairs granulocyte function when the acute stage develops.