Microsatellite instability and cytogenetic survey in myeloid leukemias

Microsatellites are short tandem repeat sequences dispersed throughout the genome. Their instability at multiple genetic loci may result from mismatch repair errors and it occurs in hereditary nonpolyposis colorectal cancer. This instability is also found in many sporadic cancers. In order to evaluate the importance of this process in myeloid leukemias, we studied five loci in different chromosomes of 43 patients, 22 with chronic myelocytic leukemia (CML) in the chronic phase, 7 with CML in blast crisis, and 14 with acute myeloid leukemia (AML), by comparing leukemic DNA extracted from bone marrow and constitutional DNA obtained from buccal epithelial cells. Only one of the 43 patients (2.1%), with relapsed AML, showed an alteration in the allele length at a single locus. Cytogenetic analysis was performed in order to improve the characterization of leukemic subtypes and to determine if specific chromosome aberrations were associated with the presence of microsatellite instability. Several chromosome aberrations were observed, most of them detected at diagnosis and during follow-up of the patients, according to current literature. These findings suggest that microsatellite instability is an infrequent genetic event in myeloid leukemias, adding support to the current view that the mechanisms of genomic instability in solid tumors differ from those observed in leukemias, where specific chromosome aberrations seem to play a major role. Correspondence


Introduction
Microsatellite instability (MSI) constitutes a recognized mechanism of mutation involved in human cancer.Differences in nucleotide repeats, representing either an expansion or a reduction of these sequences, were initially described in hereditary nonpolyposis colorectal cancer (HNPCC) (1) and have been found in HNPCC-related cancer and sporadic tumors in other organs (reviewed in 2 and 3).MSI appears to reflect multiple replication errors because of defective mismatch repair genes.Five products have been shown to be involved in human cells (MSH2, MSH6, MSH3, MLH1 and PMS2) (4) and data regarding MSI in human hematologic diseases are unclear.Wada et al. (5) reported a high frequency (53%) of MSI in the evolution of chronic myelocytic leukemia (CML) from chronic phase to blast crisis and concluded that this loss of fidelity in replication and repair mechanism may be associated with CML evolution.However, other studies have shown that instability is infrequent in the evolution of CML to blast crisis (6,7) and in acute leukemias (8)(9)(10)(11)(12).On the other hand, cytogenetic analysis is of recognized importance in leukemias since the years before the availability of banding techniques, but few studies have discussed both mechanisms together.We report here a study of MSI and cytogenetics in 43 patients with different types of myeloid leukemias.

Clinical samples
Bone marrow samples were obtained from 36 untreated patients, 22 with CML in the chronic phase and 14 with acute myeloid leukemia (AML), and from 7 patients with CML in blast crisis previously treated for the chronic phase.Thirty samples were collected at the Department of Hematology, University Hospital, Federal University of Paraná (UFPR), Curitiba, PR, 10 were collected at the Department of Hematology-Oncology, University Hospital, Santa Maria, RS, both in South of Brazil, and 3 were collected at the Blood Center of Recife, PE, Northeast Brazil.The samples were collected in 1996 and 1997.During this period, all patients whose bone marrow samples were sent to the Cytogenetics Unit (UFPR) were considered.On the occasion of the next ambulatory visit of the patient, following approval by the referring physician, informed consent was obtained and buccal epithelial cells were collected as a source of constitutional DNA.For the patients admitted to the hospital, consent was obtained on the ward.The study was approved by the Hospital Ethics Committee (UFPR).
Samples for which it was not possible to obtain amplification products of a minimum of two loci were excluded from the study (two from patients with blast crisis and one from a patient with AML).We believe that, within the defined period of time for sample collection from patients referred to the Hematology Department, the data were fully randomized, without any bias.At the other two centers where samples were obtained the same selection criteria were used.Fifteen cases were received from the Blood Center of Pernambuco; however, only three cases were feasible for DNA extraction.
The types of leukemias were diagnosed according to the French-American-British (FAB) classification (13) using immunohistochemistry, immunophenotyping and cytogenetics.

Cytogenetic analysis
Bone marrow aspirates were processed by the method of Williams et al. (14), G banding was performed by the method of Scheres (15), and karyotypes were described according to the International System for Human Cytogenetic Nomenclature (16).

DNA isolation
Genomic DNA from bone marrow samples and buccal epithelial cells was obtained by phenol-chloroform extraction following proteinase K treatment, as previously described (17).

Single-strand conformation polymorphism
Amplified PCR fragments were diluted 1:1 in denaturing loading buffer consisting of 98% formamide, 10 mM EDTA, pH 8.0, 0.02% bromophenol blue, and 0.02% xylene cyanol FF, and denatured at 94ºC for 5 min.The denatured and non-denatured reactions were subjected to electrophoresis side by side in a vertical discontinuous polyacrylamide gel system (18).A non-denatured sample was used to control for the presence of technical artifacts, which could lead to the false conclusion of the presence of new alleles.Electrophoresis was run overnight at 10 W (room temperature) and bands were detected by silver staining (19).

Analysis of microsatellite instability
We examined the extracted DNA for genetic alterations at five separate loci containing nucleotide repeat sequences.Information on repeat sequences, chromosome localization, length of amplified fragments and primer sequences, obtained from the Genome Database (20), are indicated in Table 1.MSI was analyzed by comparison of DNAbanding patterns of PCR products of micro-satellite sequences in leukemic and matched normal buccal epithelial cells.

Statistical analysis
The detection of MSI as a function of the number of analyzed loci was evaluated by the regression coefficient.
Most reports did not show instability during the chronic phase of CML (5-7,9,21) and our results agree with them.Although the instability of microsatellites is implicated in the evolution from chronic phase to blast crisis (5), other authors (6,7) have not confirmed these results, and the present report is (-) = no microsatellite instability, (+) = microsatellite instability, F = failure in amplification, ND = not done, CML = chronic myeloid leukemia, CP = chronic phase, BC = blast crisis, AML = acute myeloid leukemia, ALL = acute lymphocytic leukemia, Ct = chemotherapy, BMT = bone marrow transplantation.
in the same direction, since no alterations in electrophoresis pattern were observed among 7 patients.With respect to the patients with acute leukemias, we studied 14 patients with AML (10 de novo, 1 relapse, 1 secondary to previous acute lymphoblastic leukemia and 2 post-myelodysplastic syndrome).The only case with alteration in the allele length at a single locus (SE33) was the patient in relapse (patient number 30) (7.14%).In view of that we analyzed a small number of markers, we checked if this factor could be responsible for the absence of instability in our study.Using the information from 19 papers (Table 3) and our results, the regression coefficient showed a b value = -0.0011,t = -0.786,and P>0.40, showing that the frequencies observed are independent of the number of loci analyzed.This is in agreement with several investigators, who think that for an MSI+ pattern to be defined, alterations in just one microsatellite marker may be sufficient.Despite the presence of thousands of microsatellite loci throughout the genome, there is no indication from any  published report to date that studies based on screening of three microsatellite markers are any less accurate than those using four or more markers (3).Another fact to be considered is the type of marker.Boland et al. (33) defined a panel of five microsatellite markers recommended as a reference for future MSI studies in colorectal cancer.Unfortunately, there is no similar recommendation for leukemias and more data will be necessary to define a basic panel.
With respect to cytogenetic analysis, all the 22 patients with CML in the chronic phase showed the t(9;22) mutation (Ph+) and only one had additional chromosome aberrations (number 22).This patient was clinically and hematologically in the chronic phase, but cytogenetic analysis showed trisomy of chromosome 8 and he progressed to blast crisis and died within a short period of time.As expected, additional cytogenetic aberrations were frequently observed in the group of patients with blast crisis.Among the 14 patients with AML several chromosome aberrations were detected, some of them related to the FAB subgroup (Table 2).
Independent of microsatellite analysis, we could establish relations between chro-mosome aberrations and follow-up for most of the patients.For example, in AML, trisomy of chromosome 8 (patient number 32) has been reported to be associated with a low number of complete remissions.The t(15;17) (q22;q12) translocation is related to a high number of remissions with the combination of ATRA and chemotherapy, but the patients are at high risk for early death from intracranial hemorrhage, which was the cause of death of patient number 31.Monosomy 7 (patient number 36) and t(9;11)(p21-22;q23) (patient number 38) have been reported to be associated with a poor prognosis.These correlations have been extensively described in the literature and are used to guide the choice of the best treatment.Based on these considerations, we conclude that cytogenetics is already the method of choice for prognosis and for treatment indications at the time of diagnosis.For follow-up, molecular cytogenetics and the detection of genic rearrangements by PCR are recognized as better methods than traditional cytogenetics to evaluate minimal residual disease.In contrast, microsatellite analysis does not seem to contribute to the diagnosis or follow-up of leukemic patients.

Figure 1 .
Figure 1.Alterations in the electrophoretic pattern observed in patient number 30, locus SE33.Arrows show shift or extra bands.BM = bone marrow; N = normal.

Table 1 .
Synthetic oligonucleotides used in this study.

Table 2 .
Patient characteristics, microsatellite instability, cytogenetic analysis and follow-up.

Table 3 .
Frequencies of microsatellite instability observed in 20 different studies.
N = number of the study; X = number of loci studied; Y = frequency of instability observed/ number of reactions performed.