Lymphocytes activated by lectin exposure in vitro produce the lymphokine -interferon [13], and -interferon production is a feature of T-cell activation in animals and humans. -Interferon suppresses hematopoiesis, both in tissue culture [14] and in patients treated with this cytokine [15, 16]. Lymphocytes from the peripheral blood of patients with aplastic anemia secrete -interferon spontaneously or make excessive quantities of it after being stimulated with lectin [8]. Recently, -interferon gene expression in the bone marrow of patients with aplastic anemia was found to be highly predictive of a response to cyclosporine therapy [17]. A surprising result of this study [17] was the apparent absence of -interferon messenger RNA (mRNA) in normal bone marrow. We have now examined -interferon in newly diagnosed patients, previously treated patients, and controls.

Methods

Top Methods Results Discussion Author & Article Info References

Patients with aplastic anemia were evaluated at the National Institutes of Health; the diagnosis was established by bone marrow biopsy, and severity was categorized according to the criteria of Camitta and colleagues [18] (Table 1). Eleven patients had received no previous therapy; in other instances, the only previous treatments were brief courses of corticosteroids or growth factors, from which all patients had been withdrawn before being admitted to the Clinical Center. Patients with other hematologic diseases served as controls. Bone marrow aspiration samples were obtained when the procedure was done for diagnostic purposes. Approximately 0.5 mL of marrow was withdrawn under negative pressure into 1 mL of Iscove's modification of Dulbecco medium and 100 units of preservative-free heparin (Lymphomed; Deerfield, Illinois). Aspiration of marrow from normal volunteers was done according to a protocol approved by our institutional review board.

Detection of -Interferon Messenger RNA by the Polymerase Chain Reaction

-Interferon was detected by the reverse polymerase chain reaction method, in which sequences between two olignonucleotide primers (chosen from two exons of the -interferon gene sequence) were amplified using a heat-stable DNA polymerase [19]. To determine the gene expression, total RNA was first prepared from bone marrow cells and a complementary DNA (cDNA) synthesized [20]. In a typical experiment, RNA was extracted from 1 to 3 x 10⁶ mononuclear cells and treated with the enzyme reverse transcriptase (MLV, Gibco BRL; Gaithersburg, Maryland) to generate cDNA; the cDNA was then amplified by the polymerase chain reaction by the DNA polymerase Taq (Perkin-Elmer; Norwalk Connecticut) using the following oligonucleotides: 5' GGCTTTTCAGCTCTGCAT 3' (5' primer) and 3' GGATGCTCTTCGACCTCC 5' (3' primer) [21]. Gene amplification products were electrophoresed in 2% agarose, transferred to nitrocellulose, and hybridized with a Phosphorus-32 oligonucleotide internal probe. The sequence of the -interferon internal probe was CTGTTACTGCCAGGACC CAT. The expected product of gene amplification was 462 bp, determined by comparing it with standard molecular weight markers electrophoresed in agarose. As a control for the variable quantities of RNA obtained after extraction, a similar procedure was done concurrently for a "housekeeping" gene, glyceraldehyde phosphate dehydrogenase (GAPDH), which is constitutively expressed in all cells. Phytohemagglutinin-stimulated normal peripheral blood mononuclear cells were a positive control for these experiments. The sensitivity of this method for a -interferon-producing cell was determined to be 1 cell/10 (⁵), on the basis of results after mixing genetically engineered cells that constitutively produce this cytokine with nonproducing cells. A signal could be detected from as little as 100 ng of RNA from positive controls (lectin-stimulated blood mononuclear cells).

Serum -interferon, soluble CD8, and soluble interleukin-2 receptor levels were determined using commercial assay kits (Cell Free T8 and Cell Free IL2R Test Kits, T Cell Diagnostic; Cambridge Massachusetts; IFNg ELISA Kit, Endogen; Boston, Massachusetts). Flow cytometry of peripheral blood cells was done using standard methods and an EPICS Elite microflow cytometer (Coulter; Hialeah, Florida).

Results

Top Methods Results Discussion Author & Article Info References

Detection of -Interferon Messenger RNA in Aplastic and Control Bone Marrow

When the reverse polymerase chain reaction was used to detect -interferon mRNA, a signal was observed in most samples from patients with newly diagnosed aplastic anemia (Figure 1); Table 1 and Table 2. Interferon mRNA was also present in four patients with moderate aplastic anemia and in one of two patients with aplasia associated with paroxysmal nocturnal hemoglobinuria. A signal was not observed in patients with other hematologic diseases that produce pancytopenia or single lineage marrow failure, such as Fanconi anemia and myelodysplasia, including four patients with hypocellular myelodysplasia. (Other categories included three patients with refractory anemia with ringed sideroblasts, two patients with 5q-syndrome, one patient with refractory anemia, one patient with refractory anemia and excess blasts, and three patients with chronic myeloproliferative syndromes.)

View larger version (60K): [in this window] [in a new window]   Figure 1. Southern analysis for -interferon gene expression in marrow samples from patients with various hematologic diseases. Controls included RNA from stimulated normal peripheral blood cells. After RNA extraction, we simultaneously amplified sequences from a housekeeping gene, glyceraldehyde phosphate dehydrogenase (GAPDH), constitutively expressed in all human cells and the -interferon gene. -Interferon gene expression was detected only in the bone marrow of patients with aplastic anemia (see Table 1 for summary).

-Interferon gene expression was not related to transfusion history. Most of the patients in the hematologic control group had received many transfusions. In contrast, of the patients with aplastic anemia, two of two patients who were untransfused and five of six patients who had received fewer than 10 units of blood products were positive for -interferon signal in their marrow.

Although two patients with aplastic anemia were receiving antibiotics as empiric treatment of fever in the setting of neutropenia, none were clinically infected at the time of marrow sampling. -Interferon mRNA was not detected in marrow from patients with dengue hemorrhagic fever, although this cytokine has been implicated by in vitro studies of the immune response in dengue [22].

Correlation of -Interferon Expression with Patient Course and Other Measurements

We also tested marrow from patients who had previously had immunosuppressive therapy. A signal was not detected in patients who had received treatment with antithymocyte globulin, alone or in combination with cyclosporine, 3 to 6 months before sampling, regardless of hematologic response (Table 1). -Interferon RNA was present in the marrows of patients who had had relapses after remissions of up to 12 months after immunosuppression. In three patients who had combined immunosuppression treatment, followed serially from the time of first presentation, -interferon expression was observed to decrease or disappear in samples obtained 3 to 6 months after the initiation of therapy, concurrent with improvement to transfusion independence (data not shown). All patients whose marrow contained -interferon mRNA ultimately showed a hematologic response to immunosuppression. However, at least two patients (patients 11 and 13), who later responded to immunosuppression, did not show -interferon cytokine mRNA in their marrows at diagnosis.

There were no statistically significant differences in mean total leukocyte counts, absolute neutrophil counts, or platelet counts, used as a measure of the severity of pancytopenia, between patients with severe aplastic anemia and those with other forms of marrow failure (Figure 2). Also, no correlation was observed between -interferon expression in marrow and either serum levels of the interleukin-2 receptor (elevated in six of eight patients with aplastic anemia), serum interferon (elevated in three of eight patients), or increased circulating activated lymphocytes (data not shown).

View larger version (10K): [in this window] [in a new window]   Figure 2. Neutrophil and platelet counts in marrow donors. Only total leukocyte counts were available in the patients who had had chemotherapy (open circles). AA = aplastic anemia; IFN = -interferon; FA = Fanconi anemia; MDS = myelodysplasia.

Chromosome analysis could be done in nine patients with aplastic anemia who showed marrow -interferon expression but was abnormal in only one (patient 9). This patient was reported to have normal cytogenetic results when she was referred; her first bone marrow biopsy at our institution showed trisomy 8 in a few mitoses; 3 months later, trisomy 8 was not present but a few mitotic figures showed deletions in chromosome 7. Her bone marrow was markedly hypocellular without dysplastic features. Despite the cytogenetic abnormalities in the bone marrow, this patient became transfusion independent for erythrocytes and showed a marked increase in platelets after 3 months of cyclosporine therapy. Among patients with a diagnosis of myelodysplasia, three of eight had cytogenetic abnormalities (del 5q [q22 or 23 q31]; del 5q [q14 q23]; del 7q).

Titration of Messenger RNA in Normal and Aplastic Samples

Inability to detect -interferon mRNA in normal marrow, when RNA was extracted from the usual 1 x 10⁶ cells, might be caused by the higher proportion of lymphocytes in aplastic compared with normal marrow specimens. In our series, the percentage of lymphocytes in aspirates from patients with aplastic anemia averaged 35% (range, 9% to 90%). In the normal bone marrow aspirates, lymphocytes averaged 13% (range, 8% to 17%). To adjust for this difference, we compared RNA extracted from a larger quantity of normal marrow with diluted samples of RNA from patients with aplastic anemia. No signal was present in three normal marrow specimens, even when the total amount of RNA used was tripled, in which the lymphocyte number was approximately equivalent to a patient's sample from a smaller number of total cells, and the -interferon signal was similarly preserved with threefold dilution of RNA from patients' cells. RNA was extracted from 5 x 10⁶ to 1 x 10⁷ cells from three normal donors; these samples also were negative for interferon expression (data not shown). These results suggested that the presence of -interferon gene expression was not caused simply by the difference in lymphocyte number between normal marrow samples and aplastic samples.

Comparison of -Interferon Messenger RNA in Blood and Marrow

Because -interferon production by peripheral blood mononuclear cells or lymphocytes has been abnormal in previous studies of these cells in vitro, we also did concurrent gene amplification for this cytokine's mRNA in circulating nucleated blood cells of some patients and controls. Peripheral blood cells showed -interferon gene expression in a few normal persons whose concurrent marrow samples were negative by the same gene amplification method. Conversely, in patients with aplastic anemia, only 9 of 14 patients in whom -interferon expression was detected in marrow showed concurrent blood cell gene expression.

Discussion

Top Methods Results Discussion Author & Article Info References

Our results corroborate our earlier claims [8] and those by others [9, 11, 17] of aberrant and elevated -interferon production by lymphocytes in aplastic anemia. In particular, our data confirm Nakao and colleagues' study [17] of patients with refractory aplastic anemia. In contrast to their study, we were not able to correlate the presence of -interferon expression and clinical response to therapy in our patients, partly because of the high rate of responsiveness to immunosuppression. Among previously treated patients, we detected -interferon expression only during frank relapse. Antithymocyte globulin and cyclosporine therapy appeared to reduce markedly or to abrogate entirely interferon expression, sometimes regardless of clinical outcome. Fanconi anemia and myelodysplasia are syndromes that are readily confused with acquired aplastic anemia, and our data suggest that the specificity of -interferon gene expression may make this finding diagnostically useful. Measurement of -interferon by gene amplification, a rapid and inexpensive method, may be a helpful adjunct to chromosome analysis, particularly for defining immune-mediated marrow failure syndromes.

Marrow expression of the -interferon gene in aplastic anemia suggests a qualitative, local abnormality of lymphocytes in this disease. Although -interferon can be produced by natural killer cells and B lymphocytes as well by T cells, T cells have been implicated in the pathogenesis of aplastic anemia by various cell culture and immunophenotyping studies. Unfortunately, we could not formally show that T cells produced this lymphokine in our aplastic anemia specimens because of the scarcity of cells in these hypocellular samples even when we used flow cytometric fractionation procedures followed by the reverse polymerase chain reaction or in situ hybridization. Although we considered whether the presence of this cytokine mRNA might only reflect a higher proportion of lymphocytes present in marrow from patients with aplastic anemia, it appears unlikely. Gene amplification is an extremely sensitive technique, capable of detecting only a few producing cells in a large sample in our experiments, yet we could not detect -interferon mRNA in normal samples and in other controls. In the case of several normal marrow samples, -interferon was not detected even when the quantity of RNA assayed was increased severalfold to approximate the number of lymphocytes usually present in samples from patients with aplastic anemia. Local interferon production was inferred from comparison with peripheral blood: For some patients, -interferon expression was found only in marrow and not in blood and, conversely, some controls showed -interferon expression in blood but never in marrow.

Several possible explanations for -interferon expression in the bone marrow in aplastic anemia may be offered. First, -interferon may be a marker of lymphocyte activation that is pathophysiologically important in acquired disease, regardless of the inciting cause. This hypothesis is consistent with the independent clinical and laboratory data reviewed earlier [1]. A more trivial interpretation assigns -interferon expression to a cascade of lymphokine production triggered by pancytopenia but only conspicuous in aplastic anemia because of the particular severity of blood count depression in this disorder. -Interferon is usually regarded as a potent inhibitor of hematopoiesis but it has been reported to increase marrow cell proliferation in tissue culture under some circumstances [23, 24], and -interferon production by circulating lymphocytes has been observed in some patients after cytotoxic chemotherapy [25]. Our failure to detect interferon expression in patients with pancytopenia other than aplastic anemia and the absence of any correlation of interferon expression with neutrophil or platelet counts argue against this possibility. Third, -interferon expression might be a consequence of infection. -Interferon usually has been associated with viral infections, to which patients with aplastic anemia are not especially susceptible. Few of our patients were clinically infected at the time of sampling. We were surprised not to find -interferon RNA in dengue, because this flavivirus has been shown to provoke lymphocyte production of the cytokine [26]. Testing of a larger number and variety of clinical syndromes using gene amplification should elucidate other circumstances of -interferon expression. Our data are most consistent with a pathophysiologic role for -interferon in aplastic anemia, in which a cellular immune response harmful to hematopoiesis is induced by various antigenic stimuli, including drugs, chemicals, and viruses.

Cytotoxic lymphocytes and -interferon have been implicated in the pathogenesis of other human diseases, such as type I diabetes mellitus [27, 28], multiple sclerosis [29], myasthenia gravis [30], and inflammatory eye disease [31]. These syndromes are "autoimmune" but may be incited by viral infections. -Interferon is a protein with many functions [32, 33] and a central role in the host immune response to viruses [34]. The interferons can both restrict virus infection and mediate viral pathophysiology [35], dual effects well demonstrated by lymphocytic choriomeningitis virus infection, in which cytotoxic lymphocytes can passively transfer disease to recipient mice and cytopathologic findings can be blocked by administering anti--interferon antibodies [36]. -Interferon is not only directly cytostatic or cytolytic, but it acts to stimulate the release of other negative regulators like tumor necrosis factor and lymphotoxin [37]. It enhances the susceptibility of treated cells to attack by cytotoxic lymphocytes by increasing major histocompatibility gene expression [38]. In murine diabetes, for example, cytotoxic lymphocyte activation and interferon production have been implicated in the early pathogenesis of ss-islet cell failure. In transgenic mice, either increased HLA expression [39, 40] in the pancreas or simple insertion of a -interferon gene under the regulation of a pancreas-specific promoter [40] is sufficient to reproduce the pathologic findings of natural disease. Whether -interferon in aplastic anemia is solely a marker of immune activation or itself the mediator of stem-cell suppression may be revealed by similar experiments, now in progress, to localize cytokine production to the murine bone marrow.