To the Editor:
We read with interest the work reported recently by Baurmann et al1 on the importance of defining the kinetics of the graft-versus-leukemia (GVL) response after donor leukocyte infusions (DLI) for relapsed chronic myeloid leukemia (CML) after allogeneic bone marrow transplantation (BMT). Baurmann et al investigated the chimeric profile in different hematopoietic lineages in 4 patients at various timepoints who received DLI and/or stem cell infusions for relapsed CML. Lineage-specific analysis was performed using chromosome-specific fluorescent in situ hybridization (FISH) with simultaneous immunophenotyping of interphase cells (FICTION) to define chimeric status in the various lineages studied. Their results are similar to previous work that we have reported assessing lineage-specific chimerism by short tandem repeat polymerase chain reaction (STR-PCR) of immunomagnetically separated fractions after DLI alone for relapsed CML.2 At the time, we had performed longitudinal chimerism studies on 4 patients who relapsed 1, 3, 6.5, and 7 years post-BMT and who were treated with DLI from the original BMT donor. We observed that the conversion of the peripheral blood (PB) T-cell subset from a mixed chimeric profile (∼50% recipient cells) to a predominantly donor profile (<5% recipient cells) was the earliest indicator of a response to DLI. Molecular monitoring defined subtle changes in the degree of mixed chimerism before a response was seen in the other PB and bone marrow (BM) fractions, ie, during the critical response time. The granulocytes were the last lineage to convert to complete donor chimerism. We have now performed longitudinal STR-PCR in 6 patients and have seen similar response patterns.
The FICTION assay has the advantage of allowing dual staining of the individual cells for CD and FISH markers; however, FISH analysis has low sensitivity and the analysis of at least 200 cells is not always possible in the post-DLI situation due to low cell numbers or pancytopenia. Baurmann et al state that chimerism studies to date have been hampered by the use of different tools to assess chimerism post-BMT. Although we agree with this assertion in general, we feel that our experience with the use of STR-PCR (which we first reported in BLOOD in 19913) in more than 300 patients and the use of lineage-specific STR-PCR allow us to provide reliable results on chimeric profiles post-BMT. The separation of BM and PB into lineage-specific fractions using immunomagnetic bead separation with a high degree of purity allows analysis of a larger number of cells by our radioactive STR-PCR (which has a sensitivity of detection of 1 in 105 of the minor cell population), therefore enabling subtle changes in lymphopoietic chimerism to be detected. As indicated above, the subsequent conversion of other lineages to donor chimerism can be reliably detected using this technique, and the observation of the slower conversion of the granulocyte fraction has been reported by ourselves and in other studies.4
In one of our patients, the CD34+ fraction was predominantly donor (>90% donor) at the time of relapse when 40% donor cells were detected in the total BM fraction, a finding similar to that observed in patient no. 3 described by Baurmann et al. The CD34+ fraction became predominantly recipient, during which time the patient failed to show a clinical response to an initial DLI dose of 0.9 × 107 CD3 cells/kg and disease progression was seen. After a second DLI dose (1 × 108 CD3 cells/kg), the patient showed conversion of the CD34 fraction to a fully donor profile. It would appear that the degree of mixed chimerism and the kinetics of the GVL response are certainly different in those patients with molecular relapse compared with patients with full hematological relapse. The monitoring of the degree of chimerism at the time of relapse is vital to predict the possibility of aplasia after DLI; however, more detailed lineage-specific chimeric analysis should give more insight into the possible outcome after DLI and explain the failure of response in other patients. A larger scale study with more patients receiving a single treatment regimen and detailed monitoring of specific lineages is required to further elucidate the kinetics of the GVL response after DLI and is currently in progress with the help of the Chronic Leukaemia working party of the European Blood and Marrow Transplant (EBMT) group.
In the April 1998 issue of Bone Marrow Transplantation, Gardiner et al1-1 published their results on monitoring lineage-specific chimerism in 4 cases of relapsed CML after DLI. Because their article appeared after our work was submitted for publication to BLOOD,1-2 both groups reached their conclusions independently, and we are pleased to state that the data on the kinetics of the GVL response are quite similar. Our statement about the existence of a critical switch period during which hematopoiesis suddenly reverts to complete donor type is especially supported by the analysis of their cases. Interestingly, Gardiner et al report a sequential switch of T cells and granulocytes in 3 of their 4 cases, a feature not described in our report. However, in this context, it is important to note that, in contrast to all 8 cases of our series, the initial BMT was T-cell depleted in their 3 patients. T-cell chimerism before DLI, therefore, was mixed and not already predominantly of donor type, as in our cases.
Gardiner et al state that their technique using lineage-specific STR-PCR provides reliable results on the chimeric profile post-BMT and may be more useful than the FICTION assay in situations with low cell numbers. Although the important aspect of morphology and individual cell by cell analysis is lost with the PCR approach, we agree that STR-PCR is a most valuable tool in this context and therefore complementary to our FICTION analysis. However, we doubt that subtle changes in the degree of mixed chimerism can be seen with a semiquantitative technique according to which mixed chimerism is defined as a percentage of recipient cells between less than 90% and greater than 10%. Indeed, in the work of Gardiner et al, significant changes in every case were detected only concomitantly to the onset of clinical graft-versus-host disease (GVHD), a result similar to that obtained in our cases when using the AmpliType Polymarker PCR kit to document non–lineage-specific DNA chimerism. In contrast, when using our quantitative techniques, ie, FICTION and competitive differential bcr-abl RT-PCR as an additional disease-specific marker, the beginning of the critical switch period could be detected several weeks before the onset of clinical GVHD. Although we therefore very much support a larger scale study to better understand the kinetics of the GVL response after DLI in T-cell–depleted and non—T-cell–depleted BMT recipients, we strongly suggest the use of true quantitative measures of lineage-specific chimerism. The FICTION method is especially helpful in small centers, because it can be performed without larger equipment or in situations in which morphological control of the selected population is of additional value. Another attractive tool is the combination of FACS sorting and quantitative multiplex STR-PCR with fluorescent primers.1-3 This technique has been successfully applied by members of our group to study subtle changes of subset chimerism in patients undergoing nonmyeloablative stem cell transplantation and was shown to be predictive of response, GVHD, and disease recurrence.1-4