| Introduction Chronic neutrophilic leukemia (CNL) is a rare chronic myeloproliferative disorder that presents as a sustained, mature neutrophilic leukocytosis with few or no circulating immature granulocytes, the absence of peripheral blood monocytosis, basophilia, or eosinophilia, and a normal platelet count. Clinical evaluation fails to disclose an underlying disease process, such as malignancy or infection, capable of causing a leukemoid reaction. Splenomegaly, primarily due to neutrophilic infiltration, is present, and the bone marrow biopsy demonstrates a granulocytic proliferation without evidence of morphologic dysplasia or striking reticulin fibrosis.1 Cytogenetic and molecular analyses are negative for the Philadelphia chromosome and its molecular counterpart, the BCR/ABL fusion gene. These features distinguish CNL from chronic myelogenous leukemia (CML), atypical chronic myeloid leukemia, and chronic myelomonocytic leukemia, as defined by the French–American–British Cooperative Leukemia Group (FAB).2 CNL should also be differentiated from the recently described neutrophilic-CML (CML-N), in which the Philadelphia chromosome is present but is associated with a novel BCR/ABL molecular rearrangement (c3/a2 junction) that encodes for a 230 kDa fusion protein rather than the usual 210 kDa fusion protein associated with classic CML. We present herein a single institution's recent experience with six patients meeting the diagnostic criteria of CNL. Materials and methodsClinical records from all patients with features consistent with the diagnosis of CNL evaluated at our institution from 1992 to 1998 were reviewed. Wright–Giemsa-stained slides of peripheral blood and bone marrow aspirate smears were reviewed in all six cases. Cytochemical staining for iron, myeloperoxidase, chloroacetate esterase, and alpha-napthyl butyrate esterase was performed with standard techniques and was available in five of the six cases. Bone marrow trephine biopsy tissue was obtained in all cases, fixed in B5 fixative, decalcified, embedded in paraffin, and stained with hematoxylin and eosin for morphologic assessment. A reticulin stain of the marrow biopsy was available in all cases. Immunoperoxidase staining for kappa, lambda, myeloperoxidase, CD68 (PGM1), CD34, and tryptase was done in all cases by using previously described methods. Special testing, using fluorescent in situ hybridization (FISH) techniques to analyze interphase nuclei for common cytogenetic abnormalities characteristic of myeloid disorders, was performed. We limited this study to cases seen from 1992 onward, because the necessary molecular testing for BCR/ABL was not consistently performed before this time. Thus, appropriate material for additional confirmatory and exploratory testing would not have been available. Relevant clinical and laboratory data were abstracted from the medical records. D-FISH and FISH analysis techniques were used to screen for BCR/ABL and common myeloid cytogenetic abnormalities, respectively.3 D-FISH for BCR/ABL rearrangement BCR/ABL fusion was studied by using D-FISH, which detects double or two BCR/ABL fusion signals in cells with a t(9;22)(q34q11.2).3 D-FISH was performed with directly labeled, commercially available BCR and ABL1 probes (Ventana, Tucson, AZ, USA) to show two BCR/ABL fusion signals in cells with a t(9;22)(q34q11.2), one on the abnormal chromosome 9 and the other on the abnormal chromosome 22. The ABL1 (400 kb) probe set was directly labeled with rhodamine green and included several DNA sequences that hybridize to 9q34 and span the 200-kb breakpoint region of ABL. The BCR (300 kb) probe-set was directly labeled with Texas red and included several DNA sequences that hybridize to 22q11.2 and span breakpoints in both the major and minor BCR. D-FISH detects all molecular forms of the Philadelphia chromosome, including those with breakpoints in the c2/a3 junction. FISH for common molecular abnormalities associated with myeloid disorders For each patient, interphase nuclei from bone marrow cells were studied with DNA probes and FISH to detect anomalies of 5, 7, 8, 11, and 20. The following commercial DNA probes were used in this study: chromosome 5 anomalies, probes for 5p12 (D5S23) and 5q31 (EGR1); chromosome 7 anomalies, probes for the centromere (CEP-7) and 7q31 (D7S486); trisomy 8, probes for the centromere (CEP-8), with a centromere probe for chromosome 12 (CEP-12) used as a control; translocations involving 11q23, probes for the MLL locus; and 20q- anomalies, probes for 20q13.2 (breast cancer amplicon). ResultsClinical and laboratory features at diagnosis The patients' age at diagnosis ranged from 54 to 86 years (median 66 years). Four patients presented with weight loss of 4.5–9 kg one presented with fatigue and one with pruritus. None of the patients had fever, night sweats, or infection. The patients all had splenomegaly, palpable at a median of 4.5 cm (range 2 to 20 cm) below the left costal margin. All patients demonstrated peripheral leukocytosis, with a median white blood cell count at the time of CNL diagnosis of 67.9  109/l (range 15 to 125 109/l). This increase in white blood cells was due to neutrophilia (median 62.2 109/l, range 14.1 to 110.0 109/l), almost exclusively at the segmented and band stage of development. The percentage of white blood cells at or before the metamyelocyte stage was never greater than 2%. No absolute or relative monocytosis, eosinophilia, or basophilia was identified. Four of the patients were anemic at presentation, with hemoglobin levels ranging from 9.7 to 12.1 g/dl (median 11.6 g/dl). Two patients had mild thrombocytopenia at diagnosis, with platelet counts of 80 and 117 109/l. None of the patients demonstrated thrombocytosis (median platelet count 232 109/l, range 80 to 362 109/l). The leukocyte alkaline phosphatase score was increased in all patients in whom it was tested (range 167 to 385, normal range 40 to 100). Serum protein electrophoretic studies were available in five of the six patients. One patient had a small IgG kappa monoclonal serum protein at diagnosis. None of the patients had any significant co-morbidities, such as infection, malignancy, or systemic inflammatory disease, capable of causing a reactive neutrophilia or leukemoid reaction. The clinical features and laboratory data at diagnosis and during follow-up are presented in Table 1. Peripheral blood and bone marrow features at presentation The peripheral blood smear was most remarkable for a prominent neutrophilia without a left shift (Figure 1). Although the cells were primarily at the segmented stage of development, occasional band forms were seen. No blast cells or micromegakaryocytes were seen in any of the cases. Importantly, no dysplastic features were noted in the granulocytic series. The red blood cells showed, at most, mild anisopoikilocytosis. Dacrocytic, or teardrop-shaped, cells were not identified in five of the patients, and only one patient's blood smear showed a few scattered dacrocytes. Platelets were unremarkable, and no hypogranular or large and atypical forms were seen. The bone marrow aspirate smears were adequate in all six cases and consistently showed marked granulocytic hyperplasia, with more than 80% granulocytic precursors (Figure 2). These smears showed effective granulocytic maturation, the majority of granulocytes being at the metamyelocyte to segmented stages with no increase in blasts. No dysplastic features such as hypogranulation or pseudo-Pelger–Huët changes were seen, and no Auer rods were identified. Erythroid precursors had normoblastic maturation, and no significant monocyte, eosinophil, or basophil population was seen. Megakaryocytes were quantitatively normal in all cases and had a normal morphologic appearance in four of the six cases. The remaining two cases had, as a minor subpopulation, occasional small, monolobated megakaryocytes. However, the presence of megakaryocyte microforms was not nearly as prominent as one would expect in CML. Iron stains showed normal to increased storage iron. No ringed sideroblasts were seen in five cases; in one case, rare (<1%) ringed sideroblasts were present. Cytochemical staining with chloroacetate esterase and alpha-napthyl butyrate esterase showed no significant monocyte population and no dual-esterase staining pattern. The bone marrow biopsies were uniformly hypercellular, with all cases showing an essentially packed marrow. The biopsies were consistent with the aspirate findings, showing a marked granulocytic proliferation. No unusual distributional pattern was seen. As compared with other chronic myeloproliferative disorders, no megakaryocyte hyperplasia and no clusters of large atypical megakaryocytes were seen. Fibrosis was not apparent by hematoxylin and eosin stains, and reticulin stain showed 1+ reticulin fibrosis in all six cases (on a scale of 0 to 4). Immunoperoxidase staining for myeloperoxidase, CD68 (PGM1), CD34, and tryptase was noncontributory and showed no significant monocyte, blast, or mast cell population. Kappa and lambda immunostains showed polyclonal plasma cells in all cases, including the one case with the small serum monoclonal protein. Cytogenetic and molecular features Each of the six patients had normal cytogenetics at the time of diagnosis of CNL. None had the Philadelphia chromosome t(9;22) or the BCR/ABL fusion gene. In addition, D-FISH analysis did not detect a c3/a2 BCR/ABL rearrangement in any of the six cases. At the time of initial diagnosis, molecular analysis for BCR abnormalities was also performed with the use of Southern blotting. FISH was also used to screen interphase nuclei for common anomalies found in myeloid disorders: specifically, susceptible loci on chromosomes 5 (5p12, 5q31), 7 (7q31), 8 (trisomy 8), 11 (11q23), and 20 (20q-) were analyzed. None of the patients demonstrated molecular abnormalities at these sites. Clonal evolution occurred during the course of the disease in two patients. One of these patients had an abnormal clone with trisomy 21 (18/20 metaphases) 15 months after diagnosis. The other patient had an abnormal clone with deletion of the short arm of chromosome 12 [del(12)(p11.2)] (20/20 metaphases) 23 months after diagnosis of CNL. Both patients had been treated with hydroxyurea for the duration of their disease prior to detection of these abnormalities. Clinical course No underlying disease process as a cause of reactive leukocytosis became apparent during a median follow-up of 27.5 months (range 18 to 54 months). Progressive neutrophilia and progressive splenomegaly were found in all patients during follow-up, despite the use of cytoreductive therapy when indicated. No significant changes in the peripheral blood or bone marrow morphologic features were noted as compared with the original presenting blood and bone marrow material. The one patient in whom a small monoclonal protein was identified at diagnosis showed no evidence of a progressive plasma cell proliferative disorder. The monoclonal protein remained stable during prolonged follow-up (54 months), and bone marrow biopsy immunostaining for kappa and lambda light chains failed to identify a clonal plasma cell population. The median maximal spleen size palpable below the left costal margin was 15.5 cm (range 10 to 20 cm), and the median maximal white blood cell count was 113.4 109/l (range 52.7 to 297 109/l). All patients responded initially to therapy with hydroxyurea, with satisfactory control of leukocytosis and reduction in splenomegaly, when therapy to control these manifestations was required. Three patients eventually became refractory to hydroxyurea, manifesting progressive neutrophilia without blast transformation at 6, 12, and 13 months, with maximal leukocyte counts of 156, 187, and 297 109/l, respectively. Any response to subsequent therapy with various agents, including low-dose cytarabine, 6-thioguanine, 2-chlorodeoxyadenosine, and interferon-alpha, was short-lived. Systemic acute myeloid leukemia induction-type chemotherapy, consisting of an idarubicin/cytarabine combination or intermediate-dose cytarabine, was used in an attempt to control refractory leukocytosis in two of these patients. This resulted in death in both cases because of severe prolonged cytopenias following induction therapy, at 18 and 32 months after the initial diagnosis of CNL. The third patient died from complications of progressive disease at 16.5 months, despite the continued use of less intensive outpatient chemotherapy with 6-thioguanine and subcutaneous cytarabine. The final documented leukocyte count was 297 109/l, and the cells consisted largely of mature neutrophils without blastic transformation. Another patient's disease transformed into undifferentiated acute myeloid leukemia (FAB-M0) 21 months after the diagnosis of CNL; this patient's disease had previously been well controlled on a stable daily dose of hydroxyurea. The patient declined aggressive induction chemotherapy in favor of supportive measures and survived 8 weeks after the diagnosis of acute leukemia. Two patients, age 75 and 70 years, remain alive with stable disease on hydroxyurea, 12 and 54 months, respectively, from initial diagnosis of CNL (Table 1). DiscussionAlthough CNL is commonly included as part of the chronic myeloproliferative disorders, fewer than 100 cases have been reported since its original description in 1920, with most being case reports.4 Moreover, it is likely that not all of the reported cases represented true CNL. Molecular studies for the BCR/ABL rearrangement, which must be negative before this diagnosis can be made, have been performed only in cases described since 1992, accounting for fewer than 10 reported cases.5,6,7,8,9,10,11,12 This study is the largest reported series of well-characterized patients with CNL and allows a more detailed description and understanding of this disease. CNL is characterized by a sustained, mature neutrophilia with few or no circulating immature granulocytes, splenomegaly, and bone marrow hypercellularity with marked granulocytic hyperplasia and mild reticulin fibrosis. No myelodysplasia or characteristics of the more common chronic myeloproliferative disorders were present in the six patients from this study. It is also essential to note that cytogenetic and molecular analyses were negative for the Philadelphia chromosome and its molecular counterpart, the BCR/ABL fusion gene. The characteristics seen in cases of chronic neutrophilic leukemia can overlap with, and thus must be distinguished from, other chronic processes, both benign and malignant, that have a neutrophilic component. These would include reactive leukocytosis, leukemoid reaction, chronic myelogenous leukemia, other chronic myeloproliferative disorders, and myelodysplastic syndromes, including chronic myelomonocytic leukemia. Looking for the presence or absence of particular morphologic features is crucial when one considers the differential diagnosis of CNL. All six of our cases showed exclusive neutrophilia without granulocytic immaturity or evidence of proliferation of any other myeloid lineage. Subtle features of myelodysplasia must be sought in order to exclude an unusual presentation of myelodysplasia. Dysgranulopoiesis and dyserythropoiesis should not be present; cytochemical stains for ringed sideroblasts, marrow monocytes, and dual esterase-positive cells must be negative in a bone marrow specimen for a myelodysplastic process to be excluded. More importantly in considering CNL, other chronic myeloproliferative disorders must be excluded. From a morphologic basis, the absence of dacrocytes, basophilia, and immature granulocytes was a striking feature in the peripheral blood from our patients. In contrast to the bone marrow findings in agnogenic myeloid metaplasia, polycythemia vera, and essential thrombocythemia, the bone marrow from patients with CNL consistently showed no large, atypical megakaryocytes and megakaryocytic clustering and no significant reticulin fibrosis. Indeed, only two of the cases showed any megakaryocytic abnormalities, at most having an occasional monolobated megakaryocyte present. It may also sometimes be difficult to confidently distinguish CNL from reactive leukocytosis or a leukemoid reaction. The peripheral blood in CNL shows mature neutrophilia, while that of a leukemoid reaction can be left shifted. Left-shifted granulocytosis associated with eosinophilia or basophilia characterizes CML and helps to distinguish it from CNL. Most of the cases from this study showed a packed bone marrow biopsy together with a slight increase in reticulin fibrosis; these were typically the only morphologic findings that differentiated CNL from a benign, reactive process. Practically, it may be impossible to make a diagnosis of CNL solely on morphologic grounds if the marrow is not packed and if the degree of reticulin fibrosis is not convincing. Thus, correlation with clinical history and presentation, together with a 'tincture of time', is often needed before one can make a diagnosis of this malignant disorder with certainty. Some previously reported cases of supposed CNL were diagnosed in association with other disease processes, such as polycythemia vera, agnogenic myeloid metaplasia, myelodysplastic syndromes, monoclonal gammopathy, and multiple myeloma.13,14,15,16 These may not represent true cases of CNL, as the presence of another clonal hematologic malignancy would certainly confound this diagnosis. In such cases it is difficult to know whether the 'CNL' picture is a reactive leukocytosis rather than a discrete disease. Indeed, Standen et al.,17 using restriction fragment length polymorphisms and the pattern of X inactivation of the X-linked probe M27  , demonstrated polyclonal leukocytes in a reported case of concurrent CNL and multiple myeloma, suggesting that the neutrophilia in this patient was reactive. Resolution of the leukocytosis following treatment of the myeloma with melphalan and prednisone further supported this hypothesis.17 Thus, the clonal nature of CNL has been the subject of controversy. Unlike CML, no characteristic clonal chromosomal or molecular marker has been identified. Similar to other chronic myeloproliferative disorders, eg, polycythemia vera and essential thrombocythemia, distinguishing a clonal hematologic disorder from a reactive process may be difficult, particularly early in the course of the disease. A sustained, progressive neutrophilia, splenomegaly, a packed bone marrow with marked granulocytic proliferation, absence of an underlying disorder, and evolution to a terminal course, due either to blastic transformation or a refractory leukocytosis, as seen in the cases presented here, argue against a reactive process and are consistent with CNL representing a clonal, myeloproliferative disorder. The clonal nature of CNL has been confirmed in individual cases by several investigators on the basis of cytogenetic abnormalities, progenitor cell assays, and X-inactivation patterns.6,7,9 Most well-documented cases of CNL have normal cytogenetics. Among cases with cytogenetic abnormalities, chromosome anomalies are heterogeneous, and no consistent abnormality has been associated with CNL. Most likely, the primary genetic event that leads to CNL is submicroscopic and is not detectable by conventional cytogenetic studies. If that is so, the chromosome abnormalities reported so far may be related to cytogenetic evolution and represent secondary abnormalities in the pathogenesis of CNL. Reported isolated abnormalities have included trisomy 8, trisomy 9 with deletion of the long arm of chromosome 20 (20q-), isolated 20q-, deletion of part of the long arm of chromosome 11 with breakpoints at 11q14, trisomy 21, and complex karyotypes.7,9,10,11,18 Kwong and Cheng6 reported a case with clinical and hematologic features consistent with CNL, with normal cytogenetics and absence of the BCR rearrangement. The clonal nature of the myeloid proliferation was demonstrated by the methylation pattern of the X-linked hypoxanthine phosphoribosyl transferase (HPRT) gene. Froberg et al9 localized the abnormal clone, as recognized by a chromosomal deletion, to the granulocytic population in a patient with CNL. By cytogenetic analysis, this patient's peripheral blood and bone marrow had an 11q14 deletion. FISH was performed on peripheral blood smears using a chromosome 11-specific probe, and monosomy was localized to only the granulocyte population. The origin of the neoplastic clone in CNL has long been considered to be from granulocyte-committed progenitors,19 and this was recently demonstrated by Yanagisawa et al7 in an elegant study using colony-forming progenitor cell assays combined with cytogenetic analysis. They reported a patient with CNL who had complex chromosomal abnormalities without evidence of t(9;22) translocation or the BCR/ABL rearrangement. The patient's peripheral blood and bone marrow mononuclear cells were isolated and in culture were found to spontaneously form colonies consisting of large numbers of mature granulocytes. Cytogenetic analysis of these spontaneously formed colonies revealed chromosomal abnormalities identical to those in the patient's bone marrow. Granulocyte colony-stimulating factor-enhanced granulocyte colony formation also demonstrated these abnormalities. By contrast, cytogenetic analysis of granulocyte–macrophage and macrophage colony-forming units, formed by exposure to interleukin-3 and granulocyte–macrophage colony-stimulating factor, and burst-forming unit-erythroid induced by erythropoietin showed a normal karyotype, as did the patient's phytohemagglutinin-stimulated T cells and Epstein–Barr virus-transformed B lymphocytes. Thus, it appears that the neoplastic process, at least in this case, originated in progenitors capable only of differentiating into granulocytes, and that some of these progenitors acquired the ability to form colonies spontaneously, thus providing further evidence in support of a distinct granulocyte-specific myeloproliferative disorder.7 In the present series, all patients demonstrated a normal karyotype at diagnosis. To exclude the possibility of molecular rearrangements that may be missed by the conventional cytogenetic techniques, FISH analysis was performed to screen for some of the more common abnormalities characteristic of chronic myeloid disorders. This analysis did not reveal any masked or variant chromosome abnormalities. The recently described entity of CML-N, having an associated novel BCR/ABL fusion with c3/a2 junction, may be confused with CNL.20 However, CML-N should be considered as a variant of CML, separate from CNL. CML-N has to date been described only in association with the Philadelphia chromosome, which, by definition, precludes classification as CNL. Molecular analysis in this series did not detect the c3/a2 BCR/ABL rearrangement in any of the six cases and thus we excluded the possibility of a molecular rearrangement in the absence of a cytogenetically recognizable t(9;22). This underscores the importance of performing cytogenetic and molecular studies in the chronic myeloproliferative disorders. During the disease course, secondary chromosomal abnormalities were detected in two cases. Whether these represented spontaneous clonal evolution due to disease progression or mutations induced by chronic cytotoxic therapy remains unknown. Optimal therapy of CNL remains to be defined. In the earliest reported cases, splenic irradiation and splenectomy were used to reduce tumor bulk and relieve abdominal discomfort.1 Splenectomy has resulted in worsening of the degree of neutrophilic leukocytosis.1,11 Treatment of CNL to date has consisted largely of therapy with agents such as oral hydroxyurea, busulfan, and 6-thioguanine.1,6,11 These agents have demonstrated efficacy in controlling leukocytosis and splenomegaly and in inducing a prolonged stable phase. The successful use of interferon-alpha has also been reported, with similar therapeutic effects on leukocytosis and splenomegaly.8 None of these treatment modalities has demonstrated curative potential. Like that of other chronic myeloproliferative disorders, the clinical course of CNL is heterogeneous. Some patients exhibit a prolonged stable phase with minimal or no intervention, but there is a definite risk of death from either leukemic transformation or progressive, refractory neutrophilic leukocytosis, as further demonstrated by this report.1,8,10,12 Thus, the use of aggressive therapeutic strategies for patients who can tolerate such an approach appears warranted. A palliative approach may be reasonable in older patients with a poor performance status or significant co-morbidities, but for patients without these features alternative strategies should be considered, such as allogeneic bone marrow or peripheral stem cell transplantation. Allogeneic bone marrow transplantation has been reported in one case of CNL which met the diagnostic criteria for this disorder.11 The patient underwent a sibling donor bone marrow transplant, disparate for one B locus and one DR locus. Conditioning consisted of total body irradiation, cyclophosphamide, and anti-thymocyte globulin. The patient remains alive and disease-free more than 6 years after the procedure. It must be noted that this patient was only 15 years of age at the time of diagnosis, considerably younger than most reported cases of CNL, which tends to occur in older patients.1 In the series presented here, four patients were older than 66 years and the youngest was 54 years of age at the time of initial diagnosis. ConclusionCNL is an uncommon chronic myeloproliferative disorder of the granulocytic lineage which must be distinguished from other chronic processes that have a neutrophilic component, such as reactive leukocytosis, leukemoid reaction, CML, CML-N, other chronic myeloproliferative disorders, and myelodysplastic syndromes, including chronic myelomonocytic leukemia. The optimal therapy for CNL remains to be determined. Given the rarity of this disorder, it is unlikely that randomized, controlled trials will be feasible. Therefore, management decisions will have to be based on anecdotal reports or extrapolated from therapeutic strategies effective in similar chronic clonal myeloid disorders. Continued reporting of cases of CNL and responses to therapeutic strategies must be encouraged. Other than anecdotal experience with allogeneic bone marrow transplantation, no therapy has demonstrated curative potential in CNL. Given the potential for blastic transformation and progressive refractory neutrophilia, an aggressive approach may be appropriate for younger patients who meet the criteria for this diagnosis. Continued study of these patients is warranted to better determine the prognosis and the most appropriate therapy of this disease.
|
