Graft-versus-leukemia activity may overcome therapeutic resistance of chronic lymphocytic leukemia with unmutated immunoglobulin variable heavy-chain gene status: implications of minimal residual disease measurement with quantitative PCR

Recently, the mutational status of the variable region of the immunoglobulin heavy-chain gene (V_H) has been established as a crucial prognostic factor in chronic lymphocytic leukemia (CLL). Whereas patients with somatic mutations of the V_H gene are characterized by stable disease and long survival, an unmutated germ line V_H configuration is associated with a dismal prognosis.^1-4  Allogeneic stem cell transplantation (allo-SCT) can result in long-term disease control in a proportion of patients with resistant CLL,^5-9  and there is some evidence for graft-versus-leukemia (GVL) activity after allo-SCT.^10-12  To date, however, it is unknown whether GVL can be effective in patients with unmutated V_H status. Information on this issue is particularly important, as allo-SCT in CLL is increasingly performed using nonmyeloablative conditioning,^13-16  implying that the contribution of GVL for disease control becomes even more essential. The purpose of the present study was to investigate whether the therapeutic resistance of unmutated CLL can be overcome by allo-SCT following nonmyeloablative conditioning (NST), and to compare the relative contributions of conditioning regimen and GVL to tumor control.

Included in this study were consecutive patients from an ongoing prospective trial on NST in CLL¹⁷  who had an unmutated V_H status and diagnostic material available for longitudinal quantitative molecular monitoring. Minimum follow-up time required was 12 months. Eligibility criteria were as follows: refractoriness or early relapse (within 12 months) after fludarabine-containing treatment or after autologous SCT (auto-SCT); Richter transformation; or high-risk disease as defined by an unfavorable genetic profile (del11 q23 and/or del 17p13 and/or unmutated V_H status)¹⁸  in the presence of progressive disease. Allogeneic peripheral blood stem cells from granulocyte colony-stimulating factor (G-CSF)–mobilized HLA-identical donors were infused after conditioning by daily fludarabine (30 mg/m²) and cyclophosphamide (500 mg/m²) over 5 days, followed by cyclosporin A (CSA) and short-course methotrexate for graft-versus-host disease (GVHD) prophylaxis. In the case of unrelated donors, antilymphocyte globulin was added. CSA was tapered between day +60 and +100 after transplantation, in the absence of GVHD. Donor lymphocyte infusions (DLIs) were administered from day +120 onwards in the case of incomplete chimerism or residual disease after CSA withdrawal. Patients from a prospective protocol on the value of auto-SCT after myeloablative conditioning in CLL who fulfilled similar criteria served as controls.¹⁹  Eligibility and treatment details of auto-SCT have been described elsewhere.^4,20  Treatment protocols were approved by the responsible institutional review boards.

Immunoglobulin heavy chain (IgH) sequencing, mutational status assessment,^3,4  interphase cytogenetics,¹⁸  and chimerism studies²¹  were done as described previously. IgH real-time polymerase chain reaction (RQ-PCR) for MRD quantification was performed with an ABI PRISM 7700 thermal cycler (Applied Biosystems, Weiterstadt, Germany).²²  DNA from samples with known CLL cell content as assessed by standard flow cytometry (CD19/CD5/CD23/CD20 coexpression) were serially diluted in pooled polyclonal DNA to generate at least 5 standards. Specificity and sensitivity was defined as the last dilution with target DNA detectable or, in case of nonspecific amplification, the dilution with cycle threshold (C_T) values at least one cycle lower than the lowest C_T value found in polyclonal DNA. Each PCR showed a standard curve correlation coefficient of at least 0.95 with a slope of 3.0 to 3.9, minimum sensitivity was 10^–4, and MRD level calculation was based on comparative C_T analysis between follow-up samples and standards. Normalization by parallel amplification of the albumin gene²²  allowed calculating of sensitivity levels for negative samples.

Altogether, 9 patients with unmutated high-risk CLL who had undergone NST and 26 control patients who had undergone auto-SCT for unmutated CLL met the inclusion criteria for this study. NST patients had pretransplantation characteristics similar to the auto-SCT group in terms of sex, age, V_H homology, unfavorable fluorescence in situ hybridization (FISH) karyotype, lymphocyte doubling time (LDT), and stage, but were more heavily pretreated and had a longer time from diagnosis to transplantation (Table 1).

Peripheral blood MRD kinetics in the NST patients were characterized basically by a biphasic pattern with an only modest decrease in the early posttransplantation phase (median MRD level: 1.9 × 10^–1 [range, 2.9 × 10^–3-1.9 × 10⁰] before NST versus 1.1 × 10^–2 [range, 1.8 × 10^–4-2.3 × 10^–2] at day +100; P = .002), followed by a marked reduction of more than 2 logs in the MRD levels between day +100 and day +200 (8.4 × 10^–5 [range, 1.0 × 10^–5-1.1 × 10^–2]; P =.005). During follow-up, MRD levels stayed low and became durably undetectable in 7 of 9 patients (78%; Figure 1). In 5 of the 9 NST patients, disappearance of RQ-PCR–measurable MRD occurred subsequent to development of extensive chronic GVHD. Three patients remained free of GVHD. Two of these patients converted to sustained MRD negativity after DLI without any GVHD, one without conversion to complete chimerism (Figure 1), whereas one patient failed to eliminate MRD despite DLI. The remaining patient, who was the only carrier of a 17p13 deletion in this series, showed evidence of early clinical relapse with progressive bulky lymphadenopathy prior to the onset of extensive chronic GVHD 4 months after allo-NST. Although GVHD was associated with declining but positive MRD levels, the patient died in progressive disease at month +19.All other 8 patients are in ongoing complete clinical remission with a median follow-up of 29 months (range, 14-41 months) after transplantation.

Auto-SCT resulted in an immediate reduction of MRD levels (6.4 × 10^–3 [range, 4.0 × 10^–4-3.9 × 10^–1] before SCT versus 2.3 × 10^–4 [range, 5 × 10^–6-2.5 × 10^–1]atday +100; P = .0002) without further decrease during longer follow-up (median clone-specific DNA copy number at 2 years after transplantation was 5.0 × 10^–4 [range, 1 × 10^–5-6 × 10^–1]; P = .43 if compared to day +100). Although 6 of 26 patients (23%) eventually became PCR negative, there were only 2 patients whose MRD levels remained below the threshold of detection until their most recent follow-up at 28 and 29 months after transplantation, respectively (Figure 1). Clinical relapse, progression, or death was observed in 11 patients after auto-SCT, translating into a median progression-free survival of 48 months. Occurrence of clinical relapse was not predictable by day +100 MRD levels, day +360 MRD levels, or by achievement of MRD negativity at any time after transplantation.

There is circumstantial evidence that GVL mechanisms contribute to disease control after allo-SCT in CLL.^6,8-12,16,17  However, using sensitive longitudinal MRD quantification, the present study is the first to demonstrate a clear correlation between CLL load kinetics and immune modulating maneuvers after NST. Our results show that profound and sustained complete molecular responses occur only after establishing chronic GVHD or DLI, whereas the influence of the conditioning regimen on the tumor cell load is limited.

Of note, all patients studied had an unmutated V_H gene status, showing for the first time that unmutated CLL is sensitive to GVL. Since the GVL-mediated antileukemic activity observed here is clearly superior to that of the intensified fludarabine-cyclophosphamide regimen used for conditioning as well as to the fully myeloablative treatment administered for auto-SCT, allo-SCT appears to be the most effective and yet only potentially curative modality available for treatment of unmutated CLL. This conclusion is confirmed by a recent preliminary report of clinical observations.²³  In contrast, auto-SCT can reduce but not eradicate PCR-measurable tumor load in the majority of patients with unmutated CLL. This observation is in line with previous studies showing that complete responses documented by clinical criteria and by less-sensitive CDR3 consensus primer PCR assays are not durable after auto-SCT.^4,12,20,23 

In summary, this study demonstrates that allo-SCT can provide sustained disease control in unmutated CLL. Complete elimination of measurable MRD after allo-SCT is largely dependent on GVL, whereas the contribution of the conditioning regimen to treatment success appears to be limited. These results may help to define the future role of allo-SCT within the growing therapeutic arsenal for high-risk CLL.

The authors are grateful to the German CLL Study Group (DCLLSG; chairman Michael Hallek) and all the contributing centers taking care of patients being enrolled in this study. As in all studies of the DCLLSG, the cytogenetic studies were performed by S. Stilgenbauer, A. Kröber, and H. Döhner at the University of Ulm, the MRD and chimerism analyses were performed by M. Ritgen,A. Humpe, and M. Kneba at the University of Schleswig-Holstein in Kiel.