Title: The history of acute promyelocytic leukaemia
Abstract: Why should the British Journal of Haematology devote an article to the history of acute promyelocytic leukaemia (APL)? Because APL is probably the best example of a disease where the dialogue between physicians and scientists has provided a chance to make several advances in both clinical practice and basic sciences. Most patients with APL are now cured. Oncogenesis and the reversion of oncogenic events are now better understood. APL was the malignant disease to be treated by cell modulation, using agents which act specifically on oncogenic events. The recent history of APL could, in fact, be subdivided into three periods: before, during and after treatment with all-trans retinoic acid (ATRA). During the first period (1957–1988), the disease was defined; during the second (1988–1993), a specific treatment, later recognized as targeting the oncogenic event, dramatically improved the prognosis of the disease and, during the third post-ATRA period (1991–2002), the main achievements were improved knowledge of cellular and molecular biology, including the control of protein expression and degradation, and a new discovery: the beneficial effect of arsenic. Acute promyelocytic leukaemia (APL) was first described in 1957 by the Swedish author Leif Hillestad (Fig 1). He reported three patients characterized by 'a very rapid fatal course of only a few weeks' duration, a white blood cell picture dominated by promyelocytes, and a severe bleeding tendency due to fibrinolysis and thrombocytopenia′. He noted 'a normal ESR (erythrocyte sedimentation rate), probably caused by the reduced fibrinogen concentration in the plasma'. His conclusion was that the disease 'seems to be the most malignant form of acute leukaemia'. One of the three patients had previously been described by Stormorken (1956). The first description of APL by Leif K. Hillestad. From: Hillestad, L.K. (1957) Acute promyelocytic leukaemia. Acta Medica Scandinavica, 159, 189–194; copyright Blackwell Publishing Ltd. Leif Hillestad also mentioned that previously, Cooperberg and Neiman (1955) had described a patient with acute myelogenous leukaemia with fibrinolytic purpura, which was 'identical to his cases' and that a similar patient had also been described by Pisciotta and Schultz (1955). Leif Hillestad also thought that a patient reported by Risak (1935), with a 'rapid down hill course and the coincident rise of myelocytes in the peripheral blood', might have been APL. Leif Hillestad concluded the introduction to his 1957 report by the statement that: 'a logical name of this type of leukaemia is acute promyelocytic leukaemia'. In the first series of 20 patients recorded during the pre-ATRA period, Bernard et al (1959) described more detailed features of the disease. At the sixth European Congress of Haematology in 1957, Jacques Caen (Caen et al, 1957) reported the occurrence, in haematological malignancies, of a fibrinolytic syndrome which he defined more precisely 2 years later as acquired fibrinopenia (Caen et al, 1959). In fact, the most impressive clinical feature of APL at diagnosis was the occurrence of severe bleeding diathesis. Patients with APL experienced muco-haemorrhages combined with purpura and abundant ecchymotic subcutaneous haemorrhages. A significant proportion of patients (20–30%) died rapidly of cerebral haemorrhage. The disease constituted an individual entity, mainly because of its more hyperacute outcome compared with other forms of acute leukaemia; the two clinical features defining APL were the characteristic morphology of malignant cells and the presence of severe fibrinopenia. Larger numbers of typically promyelocytic malignant cells were found in the bone marrow than in the blood. The peripheral blood white cell count was low, and blasts had a monocytoid appearance. The predominant malignant cells in bone marrow were described as resembling abnormal promyelocytes, with an immature nucleus and a copious cytoplasm filled with several azurophilic granules. An abundance of large granules sometimes covered and masked the nucleus. Auer rods were found in many cells, grouped into what were called faggots. In 1976, the well-characterized morphology of these malignant cells led the French–American–British (FAB) Nomenclature Committee to assign them the specific classification of M3 cells (Bennett J.M. et al, 1976). Four years later, a rare variant form of APL, the hypogranular variant, was officially recognized by this Committee. It was characterized by cell nuclei that were usually bilobed, with no granules visible on light microscopy and a positive myeloperoxidase reaction (Bennett et al, 1980). Patients with this variant form experienced similar coagulation disorders, but often had a high white blood cell (WBC) count. A third, very rare variant form, basophilic microgranular APL, was described in 1982 (McKenna et al, 1982). Thus, APL and its morphological variants were identified within a 15-year timespan. Nevertheless, the clinical management of the disease remained a nightmare for physicians as a result of the unpredictable onset of life-threatening bleeding disorders. Bernard et al (1973) reported that the disease was particularly sensitive to treatment by anthracyclines, which resulted in a high rate of complete remission. However, chemotherapy exacerbated the bleeding diathesis, thus increasing the risk of early death. Two major disorders, thrombocytopenia and fibrinogenopenia, were considered responsible for the coagulopathy. Although everybody agreed on the presence of fibrinogenopenia and high levels of serum fibrinogen–fibrin degradation products, the origins of the fibrinogenopenia gave rise to several controversial discussions, which led to the adoption of different therapeutic approaches: was it due to disseminated intravascular coagulopathy (DIC) with secondary fibrinolysis, to primary fibrinolysis, or both? Apart from platelet transfusion, what was the best treatment: low-dose heparin, antifibrinolytic drugs, or no treatment at all? DIC was most frequently incriminated, because of the presence of D-dimers and the decreased levels of factors V and X. Low-dose heparin was usually included in treatment protocols. However, in contrast to the usual features of DIC, normal survival time of platelet and fibrinogen levels had been shown in APL patients (Bennett M. et al, 1976). Avvisati et al (1988) further reported normal levels of protein C and antithrombin III. The same authors noted an acquired reduction of alpha-2 plasmin inhibitor (α2PI) levels. They concluded that the fibrinolytic process, rather than DIC, was the main mechanism responsible for the haemorrhagic diathesis. In 1988, the problems concerning the mechanism of fibrinogenopenia and the best way of managing the bleeding diathesis were still unsolved. Apart from the description of the two major clinical features, the morphology of malignant cells and the coagulation disorders, APL had been confirmed as a distinct entity by the presence of an abnormal cytogenetic feature. This abnormality was at first considered by Golomb et al (1976) as a partial deletion of chromosome 17. However, a year later, the same group identified it as a balanced reciprocal translocation between the long arms of chromosomes 15 and 17 (Rowley et al, 1977), and APL was consistently associated with this 15;17 translocation by Larson et al (1984). However some APL patients did not have the t(5;17) because of an insertion or complex chromosomal exchanges. In the 1980s, APL was, therefore, defined by three features: the presence of the M3 cell (a morphological subtype of the FAB nomenclature), the occurrence of fibrinogenopenia and the presence of the specific 15;17 translocation. Between 1980 and 1988, studies focused on the management of APL treatment and on prognostic factors, the obsession being to avoid a fatal haemorrhage of the central nervous system during the first days of treatment (Rodighiero et al, 1990). The beneficial effect of the anthracycline treatment initiated by Bernard et al (1973) was confirmed in Europe and the USA. Some prognostic factors of APL were rapidly identified by Bernard et al (1973), such as the intensity of the fibrinogenopenia and the high WBC count. The latter parameter was also assessed by others, using various criteria: high blast counts or an increased level of lactate dehydrogenase. Other prognostic factors investigated included age, fever, and serum creatinine or albumin levels. Failure to obtain complete remission was not only due to resistance to chemotherapy, which was reduced by aggressive daunorubicin treatment (Marty et al, 1984), a reduction later confirmed by the South-west Oncology group (Head et al, 1995), but also to fatal haemorrhages. The dispute as to whether the fibrinogenopenia and haemorrhages were due to DIC or primary fibrinolysis continued, as well as the differences between recommended treatments. The only positive confirmed therapeutic requirement was the intensive need for platelet transfusions during chemotherapy. To increase patient survival, two studies suggested that maintenance therapy using 6-mercaptopurine and methotrexate could result in longer remissions than short consolidation regimens (Marty et al, 1984; Kantarjian et al, 1986). By 1988, the end of this initial period, APL was well characterized, both clinically and cytogenetically, and was being treated with anthracyclines and frequent platelet transfusions. The complete remission rate was around 75%, the early mortality rate 15% and the resistance rate 10%. At 2 years, the relapse rate after complete remission was 35%. Thus, about 25% of patients survived for more than 2 years and were considered cured because late relapses are rare in this type of acute leukaemia. In 1978, Leo Sachs introduced the new concept that certain agents can trigger a differentiation process in leukaemic cells, thus contradicting the dogma of the irreversible status of malignant cells. To understand the disorders affecting cell regulation of blood cell development in leukaemia, he had developed the first culture system in which normal blood cells from mice could be cloned and expanded (Ginsburg & Sachs, 1963). This procedure enabled the identification of the molecules that regulate the differentiation and proliferation of haematopoietic cells. Dr Sachs' group also demonstrated that certain leukaemic cells obtained from mouse cell lines after cloning and culture could be reprogrammed to resume normal differentiation and to become non-dividing mature granulocytes or macrophages as a result of stimulation by various cytokines (Paran et al, 1970). Different clones of myeloid leukaemic cells formed different blocks in this cytokine-induced differentiation (Fibach et al, 1973). Leo Sachs' team also demonstrated that when leukaemic cells from leukaemic mice were injected into mouse embryos they participated in normal haemopoiesis after birth (Gootwine et al, 1982). In addition, the establishment of the human myeloid leukaemic HL-60 cell line by Dalton et al (1988) enabled the identification of agents that were capable of inducing terminal differentiation. More than 100 agents were listed, some of which, e.g. retinoic acid, induced differentiation into mature granulocytes, others, e.g. dimethyl sulphoxide, into monocytes–macrophages, and others again, e.g. anthracyclines, into erythroid cells. One of the best candidates for the treatment of several types of human leukaemia was low-dose cytosine arabinoside (ARA-C), according to the in vitro maturation of two myeloid cell lines, HL-60 and U937, using a low concentration of ARA-C (Chomienne et al, 1986a;Poirier et al, 1986). Accordingly, patients treated with low-dose ARA-C (Housset et al, 1982) achieved a complete remission. In a cohort of elderly patients, 35% achieved a complete remission after 3 weeks of treatment with low-dose ARA-C (Tilly et al, 1985). However, various cytotoxic effects hampered the demonstration of the pure induction of differentiation. In 1980, the HL-60 myeloid cell was considered to be derived from promyelocytic leukaemic cells. However, we now know that this cell is not a promyelocytic leukaemic cell because it does not carry the specific 15;17 translocation and it only possesses one chromosome 17. Breitman et al (1981) showed terminal differentiation in primary cultures of HL-60 cells and of cells from APL patients in the presence of retinoic acid (RA), a derivative of vitamin A that plays a major role in embryonic development. Using the list of potential RA derivatives that can induce differentiation in the myeloid HL-60 and U937 cell lines, Chomienne et al (1986a, 1990a) assayed primary cultures of fresh malignant cells from the bone marrow of leukaemic patients (instead of cell lines). By testing over 60 bone marrow samples from patients with acute myeloid leukaemia (AML), they demonstrated that the differentiating effect of these derivatives was specific to APL and that different retinoid derivatives exhibited different abilities to induce differentiation. Etretinate was not effective (Chomienne et al, 1986b) and ATRA was potentially 10 times more effective than 13-cis (or 4 OXO) retinoids. However, during the 1980s, etretinate was the only derivative available in Europe, and only 13-cis RA was available in the USA. ATRA was not manufactured in western countries. Two patients were treated with 13-cis RA, but this induced no in vivo maturation, despite some in vitro differentiation (Chomienne et al, 1989) In 1985, thanks to travel facilities offered to Chinese medical personalities by Air France, the first informal meeting took place in Paris between Wang Zhen Yi and Laurent Degos. Their initial discussion concerned treatments designed to induce differentiation, using low-dose ARA-C in Paris and low-dose homoharringtonine in Shanghai. They also discussed the specific activity of ATRA in APL. These discussions resulted in a close collaboration between the Shanghai Institute of Haematology (Medical University number 2) and the Institute of Haematology of the Saint Louis Hospital in Paris (University of Paris 7) using ATRA manufactured by the Pharmaceutical Unit number 6 in Shanghai (Fig 2). The first production of all-trans retinoic acid for clinical use, as 10 mg tablets, made by the Shanghai Pharmaceutical Unit. Courtesy of L. Degos. Wang Zhen Yi was educated in Shanghai at Aurore University, which was managed by French Jesuits prior to the Chinese communist revolution. This is why he spoke perfect French, which facilitated communication with his French colleagues. After 1985, he came to Paris several times to visit his young student Chen Zhu (currently Vice-President of the Chinese Academy of Sciences and a prominent figure in the development of arsenic treatment) who spent 5 years of his scientific training in Paris. ATRA was first used to treat APL patients in 1987 at the Rui-Jin Hospital in Shanghai (Fig 3). During a second collaborative meeting in Shanghai in 1987 between the Saint-Louis and Shanghai Institutes of Haematology, the remarkable results of this treatment were reported: they showed that ATRA could induce differentiation of malignant cells in APL patients until they reached the stage of complete clinical remission (Huang et al, 1988). A joint presentation of differentiation therapies using low-dose ARA-C in AML and ATRA in APL was presented at the second conference of Differentiation Therapy of Cancer, in September 1987 (Degos et al, 1988). However, despite the demonstration of these positive results, western pharmaceutical companies refused to manufacture ATRA and the Chinese kindly provided the drug for the treatment of French APL patients. It was transported by Chinese students when they travelled to Paris for their training, the first of whom was Huang Meng Er. Huang Meng Er (first author of the first report on ATRA in APL), Wang Zhen Yi, Degos Laurent and Dr Chang in Shanghai in 1987 when the first APL patients were treated with all-trans retinoic acid. Courtesy of L. Degos. In Shanghai, treatment with 45 mg/m2/d ATRA was proposed for newly diagnosed patients (Huang et al, 1988) and in Paris for patients experiencing a first or subsequent relapse (Castaigne et al, 1990; Degos et al, 1990). The complete remission rates reached 95%. The clinical features observed were unusual for a treatment that induced complete remission. Thus, instead of the initial worsening of coagulopathy and bleeding diathesis usually observed during chemotherapy, patients experienced a rapid improvement. No aplasia, no primary resistance to the drug, no alopecia and few infectious episodes were observed. The most striking feature was the gradual terminal differentiation of malignant cells in the bone marrow, sometimes combined with the presence of Auer rods in mature granulocytes (Castaigne et al, 1990). In June 1989, governmental rules required all French Research Institutions to stop their cooperation with Chinese groups because of the Tien An Men Square events. ATRA from China was, therefore, no longer available, but several French patients were undergoing treatment. Faced with this situation, Laurent Degos asked representatives of Roche France, a subsidiary of Roche Switzerland, to make the drug. They agreed, but restricted its indication to French patients, to avoid going against the policy of the Roche headquarters in Basel. At that time, Roche was developing interferon for the treatment of hairy cell leukaemia. Loretta Itri, the vice-president of Roche Nutley, in the USA, chaired a controversial meeting in Paris on the neutral or neutralizing effect of antibodies against interferon. Laurent Degos took the opportunity to contact Loretta Itri, to obtain a source of ATRA production. She consulted her husband, Raymond Warrell, who was in Paris with her. Raymond Warrell was surprised by the effects of ATRA, as he thought that the induction of a complete remission by in vivo terminal differentiation of malignant cells was innovative, and suggested to Laurent Degos that his results of ATRA treatment be presented at the Memorial Sloan Kettering Cancer Center in August 1989. After the presentation, which showed series of bone marrow samples from several APL patients at various times after treatment, the audience was convinced by this new treatment, and Raymond Warrell immediately asked Loretta Itri to make the drug. As a result of her great efforts at persuasion, Roche Nutley manufactured ATRA tablets, not only for the treatment of APL but also for an extensive clinical trial involving various types of malignancies conducted under the auspices of the National Cancer Institute. French clinical findings for ATRA produced in China and by Roche France, as well as the cellular investigations, were published together in Blood (Castaigne et al, 1990; Chomienne et al, 1990a), in conjunction with an editorial in the same issue (Wiernik, 1990), which asked the questions: is APL treated by vitamin A derivative similar to pernicious anaemia treated by vitamin B12? Does a malignant cell still remain malignant if its status is reversible (Fig 4)? The concept that Leo Sachs had formulated in 1978 had still not been really accepted in 1990. Opinions expressed after the striking results obtained with all-trans retinoic acid. Is acute promyelocytic leukaemia a true leukaemia or a pseudo-leukaemia? Malignancy was not reversible in 1990. From: Wiernik, P.H. (1990) Acute promyelocytic leukaemia: another pseudoleukaemia? Blood, 76, 1675–1677. Copyright American Society of Hematology, used with permission. The specificity of ATRA treatment for APL, and the location of the retinoic acid receptor alpha (RARA) gene by Mattei et al (1988) on the long arm of chromosome 17, at a site close to the one at which cytogeneticists had located the breakpoints, prompted Laurent Degos, Christine Chomienne and Hughes de Thé to undertake further investigations. Using the RARA and probes kindly provided by Martin Petkovich (Petkovich et al, 1987) of the Pierre Chambon Laboratory in Strasbourg (France), and by Anne Dejean and Hugues de Thé of the Pierre Tiollais Laboratory at the Pasteur Institute in Paris, they observed an unusual pattern of RARA mRNA in blast cells from APL patients, a pattern not found in normal individuals or in patients with other types of leukaemia. They also noted multiple abnormalities in genomic DNA, as shown by Southern blotting, and concluded that, in APL, the RARA receptor gene undergoes rearrangement. Early in 1989, an article describing these observations was submitted to a prominent medical journal but was rejected, on the grounds that the results were due to artefacts and polymorphisms at restriction sites. The reviewers quoted a previous study conducted in the USA showing no particular RARA mRNA pattern in APL. The French paper was later published in another journal (Chomienne et al, 1990b). Meanwhile, Christine Chomienne, together with Laurent Degos (St Louis Hospital in Paris), and Hugues de Thé and Anne Dejean at the Pasteur Institute, continued their research, and cloned and sequenced the breakpoint on the RARA gene (de Théet al, 1990) using the NB4 cell line, established by Michel Lanotte from a patient with APL (Lanotte et al, 1991) and fresh APL cells. The partner gene was at first named myl. Two other teams conducted concomitant investigations: one, headed by Ellen Solomon (Borrow et al, 1990), worked on chromosome 17 and, the other, headed by Pier Giuseppe Pelicci (Longo et al, 1990), focused on several candidate partner genes located on chromosome 17, found the same breakpoint on the RARA gene. The three studies were published simultaneously at the end of 1990. One year later, the partner gene of RARA in the 15;17 translocation was completely sequenced simultaneously by de Théet al (1991) and Kakizuka et al (1991), and both teams published its sequence in the same issue of Cell. The gene was renamed PML for promyelocytic leukaemia, instead of myl, which might have led to confusion with a myosin light chain gene. An extremely beneficial effect (1990–1993). The first two reports of ATRA treatment showed a truly beneficial effect with regard to the number of complete remissions obtained: 22 out of 23 de novo APL patients (Huang et al, 1988) and 19 out of 20 patients in first relapse (Degos et al, 1990). The differentiation process and the rapid improvement of coagulopathy were considered to constitute a breakthrough in the treatment of APL. These data were described in greater detail by Castaigne et al (1990) and Chen et al (1991), and confirmed by Warrell et al (1991), who supplied evidence of the differentiation process by serial fluorescence in situ hybridization (FISH) and of the clonality, using X chromosome-linked polymorphism. However, the treatment had one major drawback: patients who achieved complete remission with ATRA, either alone or combined with low-dose maintenance therapy, usually relapsed within a few months (median 5 months during continuous ATRA treatment) (Castaigne et al, 1990; Chen et al, 1991; Warrell et al, 1991). These results led French clinicians to initiate a combination therapy, consisting first of ATRA until complete remission, followed by intensive chemotherapy in order to combine the positive effect of the two treatments, i.e. the high rate of complete remission with rapid improvement of bleeding diathesis obtained by ATRA and the long-term remission obtained by the intensive chemotherapy. In the first non-randomized study (Fenaux et al, 1992), the results for 26 newly diagnosed patients with APL treated with ATRA followed by three courses of daunorubicin and ARA-C were compared with the results of a historic control group treated by chemotherapy alone. ATRA followed by chemotherapy greatly reduced the number of early relapses that occurred within 18 months of complete remission (Fenaux et al, 1992), although a later follow-up showed that the low number of late relapses was similar to that seen after chemotherapy alone (Fenaux et al, 1994). Such dramatic progress prompted the immediate launch, early in 1991, of the first randomized European ATRA trial (APL 91). In this trial, the results of chemotherapy alone, comprising three courses of daunorubicin and ARA-C, were compared with those for ATRA until complete remission followed by the same three courses of chemotherapy. The trial was stopped prematurely after 18 months, at the end of 1992, because event-free survival was significantly better in the ATRA group (Fenaux et al, 1993). The intergroup difference was confirmed in the subsequent 4-year follow-up. By the end of 1992, the European Cooperative Group considered any regimen for the treatment of APL that did not include ATRA to be unethical, and this attitude was supported by the Memorial Sloan Kettering Cancer Center in New York. On the other hand, the American–Australian Cooperative Intergroup elaborated a protocol for a randomized trial in which the results of ATRA plus two courses of chemotherapy were compared with those of three courses of chemotherapy without ATRA. In 1993, the Europeans launched a second trial (APL 93), in which patients were randomly assigned to receive either ATRA followed by chemotherapy (the reference treatment) or ATRA and chemotherapy simultaneously. The resulting relapse rate was far better for the group receiving ATRA and concomitant chemotherapy (Degos et al, 1995; Fenaux et al, 1999). At the same time, the results of the USA Intergroup Study confirmed those of the first European randomized trial (fewer relapses when ATRA is administered during induction treatment) but also included an unexplained lower complete remission rate (68%) in the cohort of patients given ATRA and chemotherapy concomitantly (Tallman et al, 1997). In addition to the non-randomized trials conducted by Wang Zhen Yi in Shanghai, Raymond Warrell in New York, and Akihisa Kanamaru in Japan (Kanamaru et al, 1995), the Italians and Japanese later launched randomized trials. In 1989, the Italian group (Gruppo Italiano Malattie Ematologiche Maligne dell Adulto; GIMEMA) tested the role of ARA-C in the treatment of APL. This trial delayed the following one (Mandelli et al, 1997), which included ATRA. The Japanese tried to enlarge their cooperative group (Japan Adult Leukaemia Study Group; JALSG) in order to have enough patients for a randomized trial. Two of the randomized trials (APL 93) in Europe and the USA Intergroup trial also tested the effect of maintenance therapy. The results strongly indicated that ATRA maintenance treatment was beneficial in newly diagnosed APL. The European trial showed the additive effects of 2 years of intermittent ATRA treatment (for 2 weeks every 3 months) combined with low-dose chemotherapy (6-Mercaptopurine daily plus methotrexate weekly) for 2 years in reducing the risk of relapse, even (and mainly) in the poor-prognosis group with a high WBC count at diagnosis. In the USA, Intergroup Trial, patients who had not received ATRA during induction therapy also benefited from ATRA maintenance therapy, i.e. continuous ATRA treatment for 1 year (Tallman et al, 1997). At the end of this period (1991–1993), the high rate of complete remission with ATRA (reduced risk of early mortality and absence of primary resistance) and the low relapse rate generated the hope of curing the disease, at least in 75% of patients (Degos, 1994). In view of the results obtained with a combination of ATRA and chemotherapy, neither autologous nor allogeneic haemopoietic progenitor transplantation in first remission were recommended. Moreover, they seemed to have a detrimental effect. In 1993, Franco Mandelli organized a meeting in Rome on the theme 'APL: a curable disease'. Reactions of biologists and clinicians to adverse effects (1992). Treatment with ATRA led to three major adverse effects: leucocyte activation during treatment, secondary resistance to ATRA generated by continuous ATRA treatment and some unpredictable relapses. On the other hand, ATRA treatment rapidly eliminated the bleeding disorders. Biologists used their talents to elaborate guidelines and procedures designed to counteract these adverse effects and manage the coagulopathy. Castaigne et al (1990) described patients whose WBC counts increased during treatment, and who also developed fever, dyspnoea, and even renal failure and coma. These severe disorders were reversible in some patients who achieved complete remission, but others died. The disorders were well defined as the 'ATRA syndrome' (Frankel et al, 1992). It included fever, respiratory distress, weight gain, and pleural and pericardial effusion, and sometimes renal failure. The syndrome affected about one-third of patients in western countries and Japan, but was less common in China (Wang et al, 1999). It was often preceded by an increase in the number of white cells in peripheral blood. Biologists found that the syndrome was due to leucocyte activation and appeared to be related to cytokine release by the differentiating cells (Dubois et al, 1994). Two approaches were attempted to counteract this syndrome: (1) the European approach, which was to prevent the syndrome by chemotherapy when the WBC count increased (Fenaux et al, 1993), and (2) the approach in New York, which was to treat the syndrome with high doses of intravenous corticosteroids (Frankel et al, 1992; Warrell et al, 1994). Secondary resistance, which was recognized in the first series of patients, occurred in all patients treated with ATRA for long periods (Warrell et al, 1994). On account of this resistance, patients receiving ATRA alone only achieved complete remission for a short time and then became refractory to ATRA for periods of 6–12 months. After these periods, the acquired resistance was reversible. Some biologists tried to explain this by establishing artificial ATRA-resistant cell lines with additional