| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Consultations in Molecular Diagnostics |
From the Departments of Pathology
* and Medicine,
University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma
Abstract
In vitro amplification of polymorphic genetic markers, especially short tandem repeats (STRs), has become standard laboratory practice in the monitoring of allogeneic bone marrow transplant patients. After initial analysis of donor and recipient samples at multiple loci before transplantation, one or more loci are used to follow engraftment status in subsequent specimens. We describe an unusual pattern of STRs in a transplanted patient with a prior history of refractory acute myelogenous leukemia. DNA chimerism studies showed a lack of engraftment at 1 and 2 months after transplantation. Atypical minor peaks occurred at each of three STR loci in the pre-transplant and 2-month post-transplant recipient samples. However, these peaks were of equal amplitude as the major corresponding allele in the 1-month post-transplant sample. A history of myelodysplasia with specific chromosomal deletions before the patient’s acute myelogenous leukemia diagnosis appears to explain the spurious peaks. STR analysis of blood and archival paraffin-embedded tissues collected from the patient at various time points before transplantation reflected the evolution, progression, and response to therapy of the myelodysplasia. The case illustrates the need for comprehensive evaluation of pertinent clinical and laboratory data during engraftment monitoring to identify potential sources for error in interpretation of STR analysis.
Bone marrow engraftment monitoring using polymerase chain reaction (PCR) amplification targeting various polymorphic loci followed by capillary electrophoresis of fluorescently labeled PCR products has become a standard laboratory procedure used in the management of patients receiving allogeneic bone marrow transplants.1, 2 However, the choice of loci to be examined largely remains laboratory specific. Several commercially available kits, targeting either multiple short tandem repeats (STRs) or variable number of tandem repeats (VNTRs), have been used extensively for human identification purposes in the forensic community. These kits have also been applied to engraftment monitoring of patients after allogeneic bone marrow transplantation. In addition, a number of "home-brew" monoplex or multiplex amplification assays using STRs, VNTRs, or other loci have been developed for engraftment monitoring.3 Previously, Southern blot hybridization was the only nucleic acid-based technology available for these analyses; however, the majority of these assays are now based on amplification technologies.
The general application of these assays for allogeneic bone marrow engraftment monitoring or chimerism studies involves initial analysis of DNA, obtained from buccal swabs or peripheral blood from the donor and recipient before the transplant procedure, in an attempt to identify at least one informative locus that can be used to discriminate the recipient from the donor. Then at various time points after transplantation, peripheral blood and/or bone marrow samples from the recipient are analyzed at the previously identified informative locus. Post-transplant sample collection is generally performed on day 21 or 30, day 60, and day 90 and then as clinically indicated. Relative amplification of donor and recipient alleles (as determined from peak heights or areas under the peaks) is then used to assess the relative cellular contributions from the donor and recipient in post-transplant samples. Commonly, once an informative locus or loci have been established for a donor/recipient pair, these are specifically targeted to determine the transplantation status of the patient in subsequent specimens submitted to the laboratory.
Materials and Methods
Case Description
A 43 year-old Caucasian woman diagnosed with refractory secondary acute myelogenous leukemia (AML) was admitted to the OU Medical Center for an allogeneic cord blood transplant. The patient had failed to achieve hematological remission with induction therapy consisting of idarubicin and cytarabine (7 + 3) and re-induction with high-dose cytarabine arabinoside around 2 weeks later. Her bone marrow showed an increasing number of myeloblasts, reaching 85% 4 months later. In the absence of a human leukocyte antigen (HLA)-matched sibling or an unrelated bone marrow donor, she received a sex-matched, umbilical cord blood transplant (cell dose = 2.08 x 107/kg; five of six HLA antigens matched). Her conditioning regimen consisted of cyclophosphamide at 60 mg/kg/day for 2 days and total body irradiation with a total dose of 12 Gy over 4 days and anti-thymocyte globulin at 15 mg/kg/day. Graft-versus-host disease prophylaxis included a short course of methotrexate (days 1, 3, and 6 at 5 mg/m2) and cyclosporine dosing. Peripheral blood counts remained low after transplantation with total white blood cell count of <0.4 K/mm3 (normal = 4.0 to 11.0 K/mm3), hemoglobin of 10.9 g/dl (normal = 12.0 to 16.0 g/dl), and platelet count of 18 K/mm3 (normal = 140 to 440 K/mm3) on day 31. Pathological examination of bone marrow (BM) aspirate and biopsy performed on day 31 revealed a severely hypocellular marrow with no evidence of engraftment and rare myeloblasts. Repeat bone marrow biopsy on day 54 demonstrated a hypocellular marrow with abnormal hematopoiesis and clusters of myeloblasts. A chronology of clinical events for this patient and specimen acquisitions for DNA chimerism studies is provided in Figure 1
.
|
Results
Nine loci were determined to be informative out of a total of 12 examined; however, in 4 of these informative loci, the donor and recipient shared common alleles and were separated by only 1 repeat, making calculations to determine donor/recipient contributions in the post-transplant sample only semiquantitative (Table 1)
. The vWA locus (on chromosome 12) was selected as the optimal informative locus to make quantitative assessments of recipient alleles in the post-transplant samples. The patient had a bi-allelic pattern at this locus with peaks at 14 and 15 repeats, whereas the donor demonstrated a single peak at 16 repeats (Figure 2 A and B)
. The presence of a recipient-specific peak at least 2 repeats smaller than the donor-specific peak and the presence of two unique recipient-specific peaks made the choice of this locus optimal.1
However, selection of an optimum locus for chimerism analysis of this patient was restricted to loci that would generally be regarded as having low sensitivity. Calculations using this locus indicated that there were 100% recipient cells in the 31- and 54-day post-transplant samples (Table 1)
.
|
|
(panels 1 and 2) shows that the donor and recipient shared one peak (at 5 repeats) at the F13A01 locus. The donor also had a peak at 4 repeats at this locus. In the recipient’s pre-transplant sample, a minor peak at 6 repeats, which was 26% of the area represented by the major allelic peak at 5 repeats, was also present, but the origin of this peak was initially unclear because of its smaller relative size. These same peaks (at 5 and 6 repeats) were the only peaks present in the 31- the 54-day post-transplant samples; however, in the day-31 sample, they were of equal size, whereas in the day-54 sample, the area of the 6-repeat peak was again diminished and was only 29% of the 5-repeat peak.
|
, panels 1 and 2). In the patient’s pre-transplant sample, a minor peak was noted at 10 repeats at the CSF1PO locus, but this was only 24% of the 11-repeat peak. On initial analysis, it was unclear whether this minor peak was a true recipient peak or artifact of amplification. In both the day-31 and -54 post-transplant samples, these two peaks (10 and 11 repeats) were the only peaks evident at this locus. However, in the day-31 sample, these peaks were of equal height and area, whereas in the day-54 sample, the area of the 10-repeat peak was diminished to 37% relative to the 11-repeat peak. A similar skewed pattern, with a minor peak at 8 repeats and a major peak at 9 repeats was present in the recipient pre-transplant sample at the D5S818 locus. The minor peak represented 15% of the major peak in this sample (data not shown), leading to concern about its true identity. By contrast, these peaks were the only peaks present in the 31-day post-transplant sample (the donor-specific 10-repeat peak being absent) and were of similar amplitude and area. The 54-day post-transplant sample was not analyzed for this locus.
Discussion
On initial analysis of the post-transplant samples of this patient using the most informative locus, vWA, we determined that there was little or no engraftment at 31 and 54 days. The DNA chimerism studies were consistent with pathological examination of bone marrow aspirate and biopsies at these time points, demonstrating severe hypocellularity, no evidence of engraftment, and rare (31 day) or clustered (54 day) myeloblasts. Other STRs were run concurrent with the vWA locus and gave similar results. However, we noted atypical patterns of amplified products for two STR loci (F13A01 and CSF1PO) in the pre-transplant and the 54-day post-transplant samples and in D5S818 for the pre-transplant sample (day-54 sample was not analyzed for this locus). Each of the loci revealed one major peak (of similar amplitude and area as other peaks at other loci) and a minor peak which was severely diminished in amplitude and area: in the pre-transplant sample, 26% in F13A01, 24% in CSF1PO, and 15% in D5S818, relative to the major allele; and in the 54-day post-transplant sample, 29% in F13A01 and 37% in CSF1PO, relative to the major allele. By contrast, the same sized peaks as those found at both loci in the pre-transplant and 54-day post-transplant samples were also present in the day-31 samples, but the peaks were of equal amplitude and area at each locus and did not show the skewed pattern. Typically, for any bi-allelic locus that has alleles differing by 1 repeat (4 bp), the resultant allelic peaks would be expected to be of similar height and area; however, these minor peaks were only one-fourth the area of the corresponding major peaks at each locus. Therefore, the exact origin of these minor peaks was initially unclear; were they truly recipient-specific peaks or amplification artifacts?
Occasionally, alleles will produce "stutter peaks" that are 1 repeat smaller (N – 1) or (less frequently) greater (N + 1) in size than the main allelic peak. These stutter peaks are most likely due to "slippage" of DNA polymerase during amplification of the repetitive STR sequence. In our experience, these stutter peaks are typically about 5 to 10% of the main peak. Because the minor peaks were 15 to 25% of the major peaks in the recipient pre-transplant and 54 day samples, they were not viewed as typical of N – 1 or N + 1 stutter.
In addition to stutter peaks, nonspecific peaks, which are occasionally produced during amplification, can compromise the use of certain loci.1 So, were these nonspecific peaks? These minor peaks were not present in the donor sample for either locus, and all reactions (donor, recipient, and post-transplant samples) were run concurrently, so it was considered unlikely that these minor peaks were artifacts of nonspecific priming during amplification. Also, because donor-specific peaks for F13A01, CSF1PO, and D5S818 loci were absent in both post-transplant samples, the minor peaks were most likely contributed by the patient.
So, why should these minor peaks apparently resolve in the 31-day post-transplant samples relative to that observed in the pre-transplant sample and again emerge in the 54-day post-transplant sample? A review of the clinical history of the patient revealed that her diagnosis of AML was preceded by myelodysplasia (MDS) for 2 years and a diagnosis of essential thrombocytosis for 7 years (Figure 1)
. Cytogenetic analysis had been performed at the time of AML diagnosis, which was 6 months before transplantation. Examination of 20 metaphases showed monosomies and deletions of chromosomes in 19 cells and a normal karyotype in one cell; composite karyotype 45,XX,–5,del(5)(q13q31),del(6)(p21.3),–9,del(17)(p11.2)[cp]. The karyotypes of individual clones were not indicated in the report, although a descriptive text referred to the frequency of some specific deletions. An interstitial deletion of the long arm of chromosome 5 at breakpoints 5q13 and 5q31 [del(5)(q13q31)] and missing chromosomes 5 (–5) or 9 (–9) were observed in some abnormal cells. By contrast, terminal deletions of 6p [del(6)(p21.3)] and 17p [del(17)(p11.2)] were noted in all abnormal cells. It is worth noting that 5q deletions are commonly associated with AML and MDS, but 5q deletions were not observed in all of the abnormal cells in this case. By contrast, deletions of 6p and 17p were seen in every abnormal cell; therefore, the latter chromosomal changes likely represent the initial events in the disease process for this patient.
These cytogenetic changes underlying the patient’s MDS and subsequent AML before transplantation were strikingly reflected in the results of the DNA chimerism studies. The monosomy 5 that was seen in a portion of cells would effectively remove one of the alleles at each of CSF1PO and D5S818 and produce a skewed pattern for these loci. One allele was observed to be 24 and 15% of the other allele, respectively, in the patient’s pre-transplant sample by DNA analysis (Figure 2B)
. Because the CSF1PO locus resides at 5q33.3–34, the interstitial deletion at del(5)(q13q31) that was also observed in some of the patient’s cells would not affect this locus; however, this interstitial deletion would affect the D5S818 locus that resides at 5q23.3–32 (Table 1)
. This may account for the slightly decreased amount of the minor allele at D5S818 (15%) relative to the minor peak at CSF1PO (24%). By contrast, the 6p deletion [del(6)(p21.3)] that was observed in each abnormal cell would be expected virtually to eliminate the signal for one of the F13A01 alleles. Consistent with this, one allele for this locus was calculated to be present at only 26% relative to the other allele in the patient’s pre-transplant sample by DNA chimerism studies.
We can only speculate as to why the relative ratio of the alleles at F13A01, CSF1PO, and D5S818 went from skewed patterns in the pre-transplant sample to an equitable distribution of alleles at each locus in the 31-day post-transplant sample and then reverted back to skewed patterns in the 54-day post-transplant sample. We presume that the relative increase in the 6-repeat allele at F13A01, the 10-repeat allele at CSF1PO, and 8-repeat allele at D5S818 in the 31-day post-transplant sample is due to the elimination of many of the myelodysplastic cells as a consequence of the conditioning regimen before transplantation. Therefore, in the 31-day post-transplant sample, the recipient’s normal cells, without the 6p (F13A01) and –5 (CSF1PO and D5S818) and 5q13-q31 (D5S818) deletions associated with the MDS, predominate. However, by 54 days post-transplantation, the shift in the ratio of alleles at these loci suggests that the myelodysplastic cells, with their associated chromosome 5 and 6p deletions, are assuming a greater proportion of the sample. Unfortunately, cytogenetic analysis was not performed on any post-transplant samples to confirm our speculation.
To confirm the "true" allelic STR pattern for the patient without the superimposed MDS cytogenetic changes, the molecular laboratory requested a nonhematopoeitic sample from the patient. However, before a buccal swab sample could be acquired, the patient developed clinical sepsis with vancomycin-resistant enterococcus bacteremia, candidemia, and a hospital-acquired pneumonia. She became comatose and died 3 months after her transplantation. However, we subsequently acquired some DNA from our HLA laboratory that had been isolated from peripheral blood from this patient almost 2 months before her BM transplant. In addition, the laboratory acquired several archived formalin-fixed paraffin-embedded specimens (gall bladder and skin) arising from prior surgical procedures, including one (gall bladder) that pre-dated her MDS diagnosis (Figure 1)
. These surgical specimens were deparaffinized, and DNA was extracted using a DNeasy Tissue kit (Qiagen) per the manufacturer’s instructions. All samples were tested using CTTv and FFFL kits (Promega) as indicated by the manufacturer. The allelic patterns are presented in Figure 4 A and B
. These analyses confirmed that the minor peaks that were observed in the initial pre-transplant and 54-day post-transplant samples are patient specific. All samples showed amplification of 5 and 6 repeats at the F13A01 locus and 10 and 11 repeats at the CSF1PO locus. All alleles at these loci were amplified equally in the gall bladder and HLA specimens, but in the skin biopsy, one allele at each locus (again, the 6-repeat allele at F13A01 and 11-repeat allele at CSF1PO) was reduced to 66% (F13A01) or 73% (CSF1PO) relative to that of the major allele. The skin biopsy, diagnosed as erythema multiforme, was taken 1 month before the patient’s diagnosis of AML. It is likely that the same cytogenetic changes observed subsequently in the myeloblasts of the blood and bone marrow of the patient, and causing F13A01 and CSF1PO allelic deletions, were already evident in the skin biopsy. Curiously, the HLA specimen that was collected almost 8 weeks before her transplant and 6 weeks before receipt of the initial recipient pre-transplant sample showed equal amplification of all alleles at the F13A01 and CSF1PO loci (Figure 4 A and B)
. This suggests the absence of myeloblasts with the –5 and 6p deletions that were responsible for producing the skewed amplification pattern for these alleles in other specimens. A bone marrow biopsy immediately after the second induction attempt showed persistent myeloblasts (4%) and an excess of atypical megakaryocytes. We can only speculate that the clones that emerged immediately after the second induction were not those with –5 and 6p deletions that were evident later in her disease.
|
Footnotes
Address reprint requests to S. Terence Dunn, Ph.D., Department of Pathology, University of Oklahoma Health Sciences Center, Biomedical Sciences Building, Room 451, 940 Stanton L. Young Blvd., Oklahoma City, Oklahoma, 73104. E-mail: terry-dunn{at}ouhsc.edu
Accepted for publication December 9, 2005.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
