British Journal of Haematology, 1992, 81, 231-234

Effects of anti-TNF monoclonal antibody infusion in patients with hairy cell leukaemia

DAVID HUANG,l JOYCE E. REITTIE,’ SUSAN A. VICTOR HOFFBRAND AND MALCOLM K. BRENNER3

‘Department of Haematology, Royal Free Hospital, London, 2Celltech, Bath Road, Slough, and 3Department of HematoZogy/Oncology, St Jude Children’s Research Hospital, and Departments of Pediatrics and Medicine, University of Tennessee, Memphis, Tennessee, U S . A.

Received 31 August 1991; accepted for publication 16 January 1992

Summary. Tumour necrosis factor (TNF) can act as an autocrine growth factor for hairy cell leukaemia (HCL) cells. The TNF produced by the malignant clone may also inhibit normal haematopoiesis thereby contributing to the cyto- penias observed in patients with the disease. We have studied the effects of infusing a murine monoclonal anti-TNF anti- body in three patients with HCL. In two patients receiving 0.5 mg of antibody/kg on alternate days for 12 d, the drug was well tolerated. The third patient received 2 mg/kg on alternate days and developed symptoms of serum sickness by day 9. In two patients with severe B-lymphocytopenia,

circulating CD19 and CD20 positive, B-cells were restored to normal, the majority of which were negative for the HCL- associated marker CD1 lc. B-lymphocyte recovery was asso- ciated with a rise in serum immunoreactive IL-6 and with an early rise in immunoreactive TNF. These short courses of anti-TNF MAb treatment had modest effect on the tumour burden, producing a reduction in splenomegaly in one patient. Exploration of the effects of more prolonged adminis- tration of higher dose anti-TNF antibody will only be feasible when less immunogenic MAbs are available.

There is now convincing evidence that the B lymphocytes of hairy cell leukaemia (HCL) produce substantial quantities of tumour necrosis factor (TNF); HCL cells contain transcripts for the cytokine (Cordingley et al, 1988; Heslop et al. 1990), bioactive TNF is released into the supernatant when the cells are cultured and serum levels of TNF are elevated in patients with HCL (Lindemann et al, 1989). Tumour cell secretion of TNF may produce a number of effects. Since TNF is a growth factor for B lymphocytes, the cytokine may increase the growth of the tumour cells themselves in an autocrine growth loop (Bianchi et al, 1988; Paganelli et al, 1986; Digel et al, 1989; Barut et a!, 1990). Expansion of this malignant population may in turn competitively inhibit the growth and development of normal B cells, contributing to the reduction in circulating B cell numbers frequently observed in the disease. Finally the TNF produced may inhibit the growth of normal haematopoietic progenitors and cause or exacerbate the cytopenias associated with HCL (Lindemann et al, 1989; Anonymous, 1990; Michalevicz et al, 1991).

Correspondence: Dr M. K. Brenner, Division of Bone Marrow Transplantation, Department of Hematology/Oncology. St Jude Children’s Research Hospital, 332 N. Lauderdale. Memphis, TN 38101, U.S.A.

A recent phase I study of monoclonal anti-TNF antibody has been reported in patients with gram-negative septicaemic shock (Fisher et al, 1991). Even in these severely ill patients, the anti-TNF MAb was well tolerated. Because of the evidence that TNF production has a central role in the pathophysio- logy of HCL, we have studied the effects of infusing a monoclonal anti-TNF antibody in three patients with HCL. We show here that the dominant effect of the antibody in the treatment regimen used, is to augment the percentage of CD19 and CD20 positive B lymphocytes in peripheral blood and produce an increase in immunoreactive IL-6 and TNF in serum.

METHODS

MAb source and administration, CB0006 is a murine monoclonal anti-human TNF (IgG 1) produced by Celltech Ltd. A test dose of 1/1OOth the final dose was reconstituted in 50 ml0.90/, NaCl and given intravenously over 30 min. The full dose in 50 mlO.9% NaCl was then given i.v. over 30 min with pulse and blood pressure monitoring. Two patients received six doses of 0.5 mg/kg on alternate days for 1 2 d, and one patient received four doses at 2 mg/kg on alternate days over 8 d (see Results).

231

2 3 2 David Huang et a1 Table I. Patient details

p 40- - (I] - ._

y Patient AN -MAb Infusion- I % I , * <

I t I , I ,

# I I , I ,

_ _ _ _ _ _ _ _ CD19 HC infiltrate

Spleen Marrow PB Age (yr) Sex Pre Post* Pre Post* Pre Post

RM 48 M Ocm Ocm 26% 21% <2% <2% AN 62 M lOcrn 6 c m 15x1 10% <3% <2% AF 54 M 1Ocrn lOcm 25%$ 23% 5% 5%

? 40

* Post= 7-14 d following completion of course of injection. -) t 50% of which were FMC7 and CD1 Ic positive. $ > 80% of which were FMC7 and CD1 Ic positive.

- MAb Infusion- Patient AF -

TNF I R M A and IL-6 assay. 200 pl of standard control and serum samples were incubated overnight in duplicate with a lZsI tracer antibody in tubes coated with a mixture of monoclonal antibodies to TNF (Medgenix. Fleurus, Belgium). Tubes were then washed and counted. Counts measured were proportional to the amount of TNF in the sample which was determined by comparison with a recombinant TNF standard. Serum IL-6 estimations were carried out using Medgenix Easia IL-6 kit. 100 p1 of standards, controls and serum samples were added in duplicate followed by peroxi- dase coupled anti-IL-6 MAb. The substrate was measured colorimetrically at an OD of 450 nm with a reference filter of 650 nm.

M A b usedfor diagnosis ofHCL. Marrow and peripheral blood were stained using immunofluorescence with monoclonal antibodies to CD3, CD4. CD8 (T cell markers) CD19 and CD20 (B cell markers) and FMC7. and CDllc (for assessment of hairy cell numbers) and analysed by two-colour fluorescence microscopy. Five hundred cells were scanned.

RESULTS

Patient details Three patients with HCL gave written informed consent to enter the study which was approved by the institutional Ethical Practices Subcommittee. Patient details are shown in Table I. All patients had detectable HCL cells in marrow and two had 10 cm splenomegaly. All had thrombocytopenia (< 75 x 109/i) and rieutropenia (< 1-2 x 109/1) and two patients had severe B lymphocytopenia (< 1% of circulating lymphocytes, Fig 1).

Measurement of serum TNF confirmed that all three patients had elevated levels. Normal individuals have serum levels of < 5 pg TNF/ml on IRMA. Of the patients studied here, AN had 20.6pg, AF had 10.2 pg and RM had 9.1 pg/ml of TNF in their serum.

Tolerance of M A b At a dose of 0.5 mg/kg for six doses over 12 d, the antibody was well tolerated. Patient RM developed a transient low- grade fever following the first injection, and mild flu-like symptoms after the fifth and sixth injections. These resolved without further treatment. Patient AN had no symptoms. As

Date of Sample Fig 1 . Anti-TNF MAb infusion produces a rise in circulating B cells in B-lymphocytopenic patients AN and AF. Percentages of CD19 + and CD2O+ B cells in the peripheral blood of the study patients were measured by flow cytometry. Percentages in patients AN and AF rose from < 1% to 26% and 28%, respectively. RM showed no change (see text).

this initial regimen was tolerated, patient AF was given an increased dose of 2 mg/kg of antibody on alternate days. This patient developed joint pains, a rash and fever following the fourth injection, 8 d after therapy began. Antibody infusion was discontinued, and steroids started at 1 mg/kg. The symptoms rapidly resolved and the steroids were tailed over the following 2 weeks. The patient was subsequently shown to have high titre of circulating immune complexes, and a high titre of human anti-mouse antibody.

Effect of TNF M A b on cytokine levels Unexpectedly, in all three patients, there was a rise in the level of serum TNF detected by an IRMA assay (Fig 2a) 48-72 h after the first infusion of anti-TNF MAb. No equivalent rise was detected by bioassay, implying that the TNF detected was not bioactive. A more substantial and sustained apparent rise in TNF levels was seen 10 d or more after treatment began. This increase, however, occurred as the human anti-mouse antibody titre rose (data not shown). Such antibodies produce false-positive results in the TNF-IRMA used. Anti- body infusion also induced an early rise in serum immuno- reactive ILL-6 in all three patients (Fig 2b). In two this rise was short lived, returning to baseline within 7 d of completing the infusion.

Effect ofanti-TNF M A b on B cell numbers and function The antibody-associated rise in TNF/IL-6 correlated with an effect on circulating normal B cell numbers in both patients

Anti-TNF MAb Infusion in HCL 2 3 3

diminished during treatment, from 10 cm to 6 cm, returning to its pre-treatment size over the next 10 weeks. In this patient, infusion of MAb had a limited effect on the proportion of marrow HCL cells. These formed 15% of the nucleated cells pre-treatment and 10% of the total immediately post therapy. Patient AF, who received the (abbreviated) high dose regi- men, showed no change in splenomegaly or in the proportion of marrow HCL cells.

a,

U

...... A .....

“ 0 2 4 6 8 10 12 14 16 18 20 22 24 26

Days from Starting Treatment IL6 pg/ml

1000 (b)

1 100

10

1

n r RM AN

Patient AF

m P r e MAb I 7 0 n MAb Post Mab

Fig 2. (a) Elevation of serum TNF levels in the three patients receiving anti-TNF MAb. A small rise after the first 48 h is followed by a more substantial rise in patients RM and AN. All samples were measured by Medgenix IRMA. No bioactivity was found in any sample. Results plotted are the means of duplicate determinations. (b) Elevation of serum IL-6 levels in MAb recipients. IL-6 levels were measured by EIISA before, mid-way, and 7-1 2 d post anti-TNF antibody infusion. Results plotted are the means of duplicate determinations.

(AF and AN) with B lymphocytopenia. Within 1 week of beginning infusion, there was a rise in circulating CDl9 and CD20 positive B lymphocytes. The elevation was sustained for more than 3 weeks after treatment ended (Fig 1). These B cells were of normal morphology and < 5% were positive for CD1 l c (versus 50% or 85% of hairy cells in marrow: Table I). Total lymphocyte counts were not significantly altered and T cell numbers and CD4: CD8 ratios remained unchanged. In contrast, patient RM whose pretreatment circulating B cell proportions and numbers were within the normal range (CD19 = 14%, CD20= 12%), showed no change on treat- ment with anti-TNF MAb.

In all three patients, serum levels of immunoglobulins M, G and A remained unaltered throughout treatment.

Effect of anti-TNF M A b on normal haemopoiesis No patient showed any significant change in blood neutro- phils or platelet counts up to 4 weeks after discontinuation of therapy.

Effect of anti-TNF M A b on HCL In patient RM there was little change in bone marrow trephine appearance or peripheral blood HCL cell count pre- or 2 weeks post-treatment (Table I). In patient AN, spleen size

DISCUSSION

Anti-TNF antibody was well tolerated at lower dose levels (0.5 mg/kg), but induced serum sickness at higher doses (2 mg). Paradoxically, within 48 h of beginning treatment, anti-TNF MAb infusion modestly increased levels of circulat- ing TNF detected by an IRMA assay. The observation that increased immunoreactive TNF can be produced following infusion of an antibody to the cytokine has been made previously in patients with cerebral malaria treated with a single dose of the same anti-TNF MAb (Kwiatowski et al, submitted), although the mechanism is unknown. The later sustained rise in TNF observed is almost certainly artefactual and is likely due to the presence of human anti-mouse antibodies, which are reactive in the IRMA used to measure TNF.

Infusion of MAb is also followed by a rise in serum IL-6, a cytokine which acts as a growth factor for HCL cells (Barut et ul, 1990; Griffiths & Cawley, 1990; Freeman et al, 1989). Since TNF has been shown to be a potent inducer of IL-6 from HCL cells (Heslop et al, 1990), any increase in the persistence of TNF brought about by anti-TNF antibody may induce IL-6 secretion, which in turn helps maintain the disease status quo.

The increase in serum IL-6 may contribute to the substan- tial increase in circulating CD19 and CD20 B cell numbers in the two patients in whom HCL was associated with severe B lymphocytopenia. While some of these B cells may have originated from the malignant clone itself, they were all morphologically normal, and < 5% were strongly CDl l c positive. Notwithstanding this rise in B cell numbers, there was no associated increment in serum Ig levels of any isotype.

In contrast to its effects on serum cytokines and circulating B cells, anti-TNF MAb had only limited effect on the disease in one patient. Similarly, there was no discernible effect on neutrophil or platelet counts. This lack of influence may in part be due to the brief duration of the infusion course. It is probable that HCL cells are long lived in vivo so that only prolonged interruption of their autocrine growth loop would be expected to moderate HCL cell numbers. Certainly the effects of alpha IFN-which may also work by interrupting HCL cell autocrine growth loops (Heslop et al, 1990)-are detected only after some weeks of treatment (Doane et al, 1990; Martinet ul, 1990; Berman et al. 1990). Alternatively, the limited response of the HCL and the cytopenias may be a consequence of the associated augmentation of basal TNF (and 1L-6) secretion ensuring that local or intracellular concentrations of the cytokines remain sufficient to maintain the survival or growth of the HCL cells and the suppression of normal marrow progenitors. One way of discriminating

234 David Huang et a1 between these possibilities is to use prolonged infusions of the MAb. Unfortunately, as patient AF illustrated, the immuno- genicity of currently available anti-TNF antibodies preclude this approach. With the availability of humanized/chimaeric antibodies, however, such a comparison becomes feasible, and may allow evaluation of the pathogenic role of autocrine growth factors in HCL and related B cell malignancies.

ACKNOWLEDGMENTS

Dr Malcolm Brenner is supported by NIH grants CA 20180, CA 21765 (CORE) and the American Lebanese Syrian Associated Charities (ALSAC). We thank Vicki Gray and Nancy Parnell for word processing.

REFERENCES Anonymous (1990) Treatment of hairy cell leukaemia. Lancet, 336,

Barut. B.A., Cochran, M.K., OHara. C. & Anderson, K.C. (1990) Response patterns of hairy cell leukemia to B-cell mitogens and growth factors. Blood, 76, 2091-2097.

Berman. E., Heller, G.. Kempin, S.. Gee, T., Tran, L.L. & Clarkson. B. (1990) Incidence of response and long-term followup in patients with hairy cell leukemia treated with recombinant interferon alfa- 2a. Blood, 75, 839-845.

Bianchi, A.C.. Heslop, H.E., Drexler. H.G., Cordingley. F.T., Turner, M..DeMel, W.C..Hoffbrand.A.V. &Brenner,M.K. (1988)Effectsof tumour necrosis factor and alpha interferon on chronic B cell malignancies. Nouvelle Revue Francaise d’Hernatologie, 30, 3 1 7- 319.

Cordingley, F.T., Bianchi. A., Hoffbrand, A.V.. Reittie. J.E.. Heslop. H.E., Vyakarnam. A., Turner, M., Meager, A. & Brenner, M.K. (1988) Tumour necrosis factor as an autocrine tumour growth factor for chronic B-cell malignancies. Lancet. i, 969-971.

Digel. W., Stefanic, M., Schoniger, W.. Buck, C., Raghavachar. A,. Frickhofen, N., Heimpel, H. & Potzsolt. F. (1989) Tumour necrosis factor induces proliferation of neoplastic B cells from chronic lymphocytic leukemia. Blood. 73, 1242-1246.

149-1 50.

Doane, L.L., Ratain, M.J. & Golomb, H.M. (1990) Hairy cell leukemia. Current management. Hernatology/Oncology Clinics of North Amer- ica, 4, 489-502.

Fisher. C.J., Opal, S.M.. Ohalnaut, J.P., Zimmerman, J., Nightingale, P.. Harris, S.J., Schein, R.L., Panacek, E.A., Vincent, J.L.. Foulke, G.E., Warren, E.L., Garrard. C.. Park, G., Bodmer, M.W., Stephens, S.. Cohen, J.. van der Linden, G. & Sadoff, J.C. (1 991 ) Anti-TNF IgG monoclonal antibody (CB0006) in sepsis syndrome: pharmaco- kinetics, safety and role of pretreatment IL-6 as a predictor of time of survival. Circulatory Shock, 34, 167-168.

Freeman, G.J.. Freedman, AS.. Rabinowe. S.N.. Segil, J.M., Horowitz, J., Rosen, K., Whitman. J.F. & Nadler. L.M. (1989) Interleukin 6 gene expression in normal and neoplastic B cells. journal ofClinical Investigation, 83, 1 5 1 2-1 5 1 8.

Griffiths, S.D. & Cawley, J.C. (1990) The effect ofcytokines. including IL-1, IL-4, and IL6, in hairy cell proliferation/differentiation. Leukemia, 4, 337-340.

Heslop. H.E.. Bianchi, A.C., Cordingley, F.T., Turner, M.. Chandima, W..De-Mel,C.P..Hoffbrand.A.V.&Brenner. M.K. (1990)Effectsof interferon alpha on autocrine growth factor loops in B lympho- proliferative disorders. Journal of Experimental Medicine, 172,

Lindemann. A., Ludwig, W.D., Oster, W., Mertelsrnann, R. & Herrmann, F. ( 1 989) High-level secretion of tumor necrosis factor- alpha contributes to hematopoietic failure in hairy cell leukemia.

Martin. A., Nerenstone. S.. Urba, W.J., Longo, D.L., Lawrence, J.B., Clark, J.W., Hawkins. M.J., Creekmore, S.P.. Smith, J.W.. 2d & Steis, R.G. (1990) Treatment of hairy cell leukemia with alternat- ing cycles of pentostatin and recombinant leukocyte A interferon: results of a phase I1 study. Journal ofclinical Oncology, 8, 721-730.

Michalevicz. R.. Porat. R.. Vechoropoulos, M., Baron, S.. Yanoov, M.. Cycowitz. 2. & Shibolet. S. (1991) Restoration of in vitro hemato- poiesis in B-chronic lymphocytic leukemia by antibodies to tumor necrosis factor. Leukemia Research, 15, 11 1-120.

Paganelli. K.A., Evans. S.S., Han, T. & Ozer, H. (1986) B cell growth factor-induced proliferation of hairy cell lymphocytes and inhibi- tion by type I interferon in vitro. Blood, 67, 937-942.

1 729-1 734.

Blood. 73, 880-884.