Journal Logo

Clinical and Translational Research

Clonal Evolution Including Partial Loss of Human Leukocyte Antigen Genes Favoring Extramedullary Acute Myeloid Leukemia Relapse After Matched Related Allogeneic Hematopoietic Stem Cell Transplantation

Stölzel, Friedrich1,7; Hackmann, Karl2; Kuithan, Friederike3; Mohr, Brigitte1; Füssel, Monika4; Oelschlägel, Uta1; Thiede, Christian1; Röllig, Christoph1; Platzbecker, Uwe1; Schetelig, Johannes1; Illmer, Thomas1; Schaich, Markus1; Seliger, Barbara5; Hartmann, Arndt6; Baretton, Gustavo3; Zietz, Christian3; Ehninger, Gerhard1; Schrock, Evelin2; Bornhäuser, Martin1

Author Information
doi: 10.1097/TP.0b013e3182481113

Depending on the individual risk category, disease relapse may occur in 30% to 50% of acute myeloid leukemia (AML) patients undergoing allogeneic hematopoietic stem cell transplantation (HSCT) (1). Although reduced-intensity conditioning regimens for allogeneic HSCT are more feasible in elderly AML patients and in case of induction chemotherapy failure, relapse remains a major challenge requiring novel therapeutic strategies (2). Deletion of one copy of chromosome 6p, including deletion of human leukocyte antigen (HLA) alleles, is a possible mechanism for immune evasion in BCR-ABL1-positive acute lymphoblastic leukemia and solid tumors (3, 4). Other involved mechanisms discussed in this context are down-regulated major histocompatibility (MHC) antigen expression which has been demonstrated in dogs with transmissible venereal sarcomas and lack of MHC diversity in Tasmanian devils with transmissible facial tumors. These examples suggest the susceptibility of tissues with reduced immunologic surveillance for cancer immune evasion (5, 6). Furthermore, acquired uniparental disomy of chromosome 6p was demonstrated in relapsed patients with AML and myelodysplastic syndrome undergoing haploidentical allogeneic HSCT and donor T-cell infusion by Vago et al. The authors demonstrated that loss of mismatched HLA leads to loss of allorecognition of leukemic cell HLA by T cells, thus leading to relapse (7). The aim of this work was to prove that this mechanism of immune escape is also involved in the setting of relapse after matched related allogeneic HSCT, and we hypothesized that it might occur frequently in tissues with reduced immunologic surveillance such as extramedullary AML relapse which precedes systemic hematologic relapse and which is reported to occur more frequently after allogeneic HSCT (8). We report that clonal evolution with partial loss of HLA genes may be involved in escape from graft-versus-leukemia (GVL) alloreaction, subsequently causing extramedullary AML relapse after matched related donor (MRD) allogeneic HSCT which has not been reported in the literature so far.


A 44-year-old woman was diagnosed with a mass in the left mammary gland parenchyma on biopsy, which was consistent with extramedullary manifestation (EM) of AML (Fig. 1A, B). Nine years earlier, she had been diagnosed with AML (FAB M1), harboring a complex aberrant karyotype: 46,XX,del(5)(q31q33),del(12)(p11.2)[9]/49,XX,+5,del(5)(q31q33),+6,+8,del(12)(p11.2)[10]/46,XX[4].

Formalin-fixed, paraffin-embedded immunohistochemical staining of extramedullary manifestation in mammary gland tissue with anti-CD33 (A) and with anti-CD34 (B) showing infiltration of CD33-positive and CD34-positive blasts, surrounding mammary gland ducts (magnification ×40).

Molecular analyses revealed a mutated nucleophosmin 1 (NPM1) gene and wild-type Fms-like tyrosine kinase 3 (FLT3) gene. Because of high-risk cytogenetic features and induction chemotherapy failure, allogeneic HSCT from a MRD had been performed after a second cycle of chemotherapy during aplasia following fludarabine-based reduced-intensity conditioning with subsequent stable engraftment and complete donor chimerism (9). Until relapse, the patient remained in complete hematologic remission with complete donor chimerism. No donor T-cell infusion was performed, and no signs of graft-versus-host disease were observed. Thus, immunosuppressive therapy was tapered rapidly after allogeneic HSCT. CD4/CD8 ratios were within normal ranges determined at days +91, +241, +828, +1072, and +3125. Fluorescence in situ hybridization (FISH) analyses of the recurring EM mass revealed deletion of 5q in 82% AML blasts. Regarding the trisomy of chromosome 6 which was observed in one of the subclones at initial diagnosis, staining of the EM mass with a FISH probe directed against the centromere region of chromosome 6 revealed subclones with 9% of AML blasts with a +6, 30% of AML blasts with a monosomy 6, and 59% of AML blasts with a normal diploidity of chromosome 6 (data not shown). At the same time, neither bone marrow (BM) morphology nor cytogenetic and molecular analyses could detect systemic relapse of AML. Donor chimerism in the BM and the EM mass was 100% and 5.4%, respectively, excluding the existence of donor-derived leukemia. Staining for T- and natural killer (NK) cells showed that the extramedullary tumor was infiltrated by T cells but not by NK cells (Fig. 2A, B).

Formalin-fixed, paraffin-embedded immunohistochemical staining of extramedullary manifestation in mammary gland tissue with anti-CD3 (A) and with anti-CD56 (B) showing infiltration by CD3-positive T cells and no infiltration by CD56 NK cells (magnification ×20).

To understand the mechanism of how this leukemia clone had escaped GVL effects, array comparative genomic hybridization (aCGH) of the DNA isolated from the EM mass was performed. Various previously diagnosed features and a new, small chromosome 6p deletion were detected (among others was, e.g., a chromosome 12p13 deletion including the whole ETV6 gene region). The involved 6p deletion was located at 6p21.32-p21.33 (Fig. 3) including several HLA class I genes such as major histocompatibility complex class I chain-related gene B/A (MICB/MICA) HLA-B, HLA-C, HLA-E, HLA-L, and HLA-J. The intensity of the deletion signal was marginal above a heterozygous state, leading to the assumption that in the majority of the leukemic blasts a loss of heterozygosity exists, whereas some blasts harbor a homozygous deletion of this 6p locus. To characterize the secondary occurrence of the 6p deletion, we performed aCGH of formalin-fixed, paraffin-embedded (FFPE) BM and from cryopreserved BM blasts, which were obtained at the time of initial diagnosis 9 years earlier. In these samples, the 6p deletion was not observed, suggesting that a distinct leukemic subclone may be responsible for the late occurrence of EM. This hypothesis is supported by the fact that a NPM1 gene mutation, detectable at primary diagnosis of AML 9 years ago, was not observed in DNA isolated from the EM mass at relapse. During relapse, the clone harboring the +5 karyotype, which initially carried the mutated NPM1 gene (which is encoded on chromosome 5q) and dominated the analyses, was missing. The FLT3-internal tandem duplication mutation found in the EM site at relapse could be a secondary event occurring in this subclone. Immunohistochemical MICA (encoded inside the 6p deletion) staining in the EM was homogenous positive and was compared with staining of TAP2 which is encoded outside of the 6p deletion (see Figures, SDC 1,—the observation that overall expression of cell-surface class I HLAs is not affected in escape mechanisms, which rely on uniparental disomy of chromosome 6p, has been reported recently in patients' blasts with AML relapse after haploidentical HSCT by flow cytometry (7). To unravel whether the loss of HLA-class I and MIC genes might be associated with an altered susceptibility of AML blasts toward NK cells, we performed killer cell immunoglobulin-like receptor (KIR) genotyping which revealed (in concordance with the HLA type: HLA-A *02:01,*11:01; HLA-B *51:01,*35:01; HLA-C *14:02,*04:01) the existence of activating KIRs 2DS1 and 2DS4 and inhibitory KIR 2DL1 for HLA-Cw*04 and inhibitory 2DL3 for HLA-Cw*14. Furthermore, inhibitory 3DL1 was observed for HLA-Bw4.

Array comparative genomic hybridization (CGH) of chromosome 6 of formalin-fixed, paraffin-embedded material of extramedullary manifestation acute myeloid leukemia (AML) cells at relapse (“paraffin,” above) and of cryopreserved blasts from bone marrow (BM) aspiration at the initial diagnosis of AML 9 years earlier (below). A schematic representation of chromosome 6 is included below both plots of the array CGH results. Red brackets indicate centromeric and telomeric boundaries of the involved 6p21 deletion; deleted human leukocyte antigen (HLA) class I genes includemajor histocompatibility complex class I chain-related gene B/A (MICB/MICA), HLA-B, HLA-C, HLA-E, HLA-L, and HLA-J; deleted genes (official symbols) are shown at the bottom.


Allogeneic HSCT is the only potential curative treatment for refractory or relapsed AML patients (1). Extramedullary relapse—as a solitary event—subsequently preceding hematologic relapse occurs frequently and remains a clinically challenging situation with few therapeutic options (10). In our patient, we hypothesize that partial loss of HLA genes might be a pivotal step in loss of an antileukemic response in the case of allogeneic HSCT, heralding relapse. Since the subclones that were observed in the extramedullary AML relapse had a monosomy of chromosome 6 in 30% of AML blasts, normal diploidity of chromosome 6 in 59% of AML blasts, and a trisomy of chromosome 6 in 9%, a clonal evolution is supposable. However, because the array CGH signal intensity suggests a heterozygous partial 6p loss in almost all AML blasts and in a small fraction of blasts a homozygous partial 6p deletion, we think it is worth speculating that the subclones with a diploid and triploid signal for chromosome 6 in the EM relapse had a heterozygous partial 6p deletion and the subclone with the monosomy of chromosome 6 had a homozygous partial 6p deletion. Furthermore, GVL reactions are less effective in tissues with reduced immunologic surveillance in which the numbers of patrolling donor T- and NK cells are limited. A potential mechanism involved in the escape of AML blasts with loss of MHC-I expression in the HLA-matched setting might be the reduced capability of AML blasts to act as antigen-presenting cells, including the inability to present MHC antigens through HLA-B and HLA-C to donor CD8+ T cells (11). Moreover, lower expression of ligands MICA and MICB might interfere with the antileukemic efficacy of allogeneic NK cells by reducing the ligation of activating NK cell receptors, especially for MICA and MICB, the activating NK receptor NKG2D (12). Furthermore, the loss of HLA-C and -B could theoretically be associated with a differentially regulated NK-cell reactivity because loss of activating activity (2DS1 and 2DS4) and loss of inhibitory activity (2DL1, 2DL3, and 3DL1) occurred. Additionally, the balance between activating NKG2C and inhibitory NKG2A that form heterodimers with CD94 and bind to the nonclassical HLA-E might be imbalanced through the loss of HLA-E. The detection of HLA-E by CD94/NKG2 receptors (except NKG2D) is described as a sensitive mechanism for the immunosurveillance and reactivity against allogeneic cells because HLA-E loads, for example, HLA-B and -C leader peptides which again are missing together with HLA-E (12). Immunohistochemistry suggests that the extramedullary AML tissue was clearly accessible to T cells but not to NK cells. Besides the potentially impaired recognition of tumor cells and insufficient activation of NK cells, the tumor microenvironment might have suppressed NK-cell tropism and invasion.

Further experimental studies are required to determine whether quiescent leukemic blasts may persist at extramedullary sites after allogeneic HSCT, acquire further genetic aberrations, making them less susceptible to allorecognition, or whether subclones with altered MHC expression preferentially move toward and persist in extramedullary sites (sanctuary sites). In AML, it is unknown whether these sites represent tissues with reduced immunologic surveillance or whether formation of a “metastatic niche” with accumulation of BM-derived stromal cells is necessary, as suggested for solid malignancies (13, 14). As information about extramedullary disease in AML is scarce, animal models are desired in which these conditions are simulated to delineate the precise mechanisms involved. Furthermore, EM tissues need to be analyzed to characterize whether partial 6p deletion is a common mechanism of relapse in AML patients after MRD allogeneic HSCT. This is the first report on partial HLA gene loss in extramedullary AML relapse after allogeneic HSCT. Because of the complete donor chimerism in the BM and the fact that there was no del5q observed with FISH analysis in the BM at relapse and the aberrant subclones at the time of initial diagnosis 9 years earlier had a del5q, we exclude the possibility of the existence of an occult relapse subclone in the BM with a partial loss of chromosome 6p. Therefore, our data indicate that 6p deletion including class I HLA genes may provide a mechanism whereby leukemic cells can escape the effects of GVL after matched related allogeneic HSCT. During the preparation of this manuscript, another group reported on the loss of HLA haplotype during AML relapse shortly after matched unrelated allogeneic HSCT (15). It was speculated in their report and by others in the context of haplotransplantation that selective pressure mediated by donor T cells through donor lymphocyte infusions led to the appearance of loss of HLA haplotype in AML relapse (7, 16). Our report plants evidence that AML relapse due to HLA loss also occurs in extramedullary AML and after matched related allogeneic HSCT (both of which have not been reported so far) and also occurs without the proposed triggering mechanism of donor lymphocyte infusions. In contrast to the above-mentioned reports, the deleted region in our patient did not span the whole HLA locus on chromosome 6p indicating that not all HLA genes are necessary for triggering relapse. Furthermore, occurrence of relapse after a long time provides evidence for persistence of quiescent leukemic cells in either extramedullary sites or in BM after HSCT. The finding of a NPM1 wild type in the EM relapse suggests a clonal evolution of AML blasts with partial 6p deletion which have an assertive survival advantage when compared with other clones that were identified initially. Alternative treatment strategies for reinducing immunogenicity in these leukemic subsets or targeting AML blasts at the signaling level are warranted. Detailed molecular characterization of extramedullary AML will help to better understand the mechanisms of the disease to which many patients succumb to, despite tremendous therapeutic efforts.


The patient was treated in the AML96 trial of the Study Alliance Leukemia which is in agreement with the Helsinki declaration and was approved by the institutional review boards of all participating centers. The study was registered with the NCT number 00180115. DNA isolation, chimerism, and analyses for FLT3-internal tandem duplication and NPM1 mutations were performed as described previously (17). Conventional cytogenetic analyses were performed on short-term cultured BM cells (24 and 48 hr). Chromosome preparation and GTG banding were performed using standard techniques. FISH was performed on freshly prepared slides using methanol/acetic acid-fixed cells. Hybridization and posthybridization washing followed the manufacturer's recommendations (LSI EGR1/D5S23, D5S721 Dual Color Probe Set, CEP 6 [D6Z1] SpectrumOrange; Abbott, Wiesbaden, Germany). During relapse, 27 and 23 metaphases and 200 interphases were investigated. FFPE tissue was cytogenetically analyzed using the above-mentioned probe in breast tissue biopsy samples. Array CGH analyses were performed using the Human Genome CGH Microarray kit 244A (Agilent, Santa Clara, CA) according to the manufacturer's instructions. The hybridized arrays were scanned using an Agilent microarray scanner. Raw data were processed using Feature Extraction software (Agilent). Deleted or amplified regions were determined using the DNA Analytics 4.0.76 program (Agilent). DNA from FFPE BM and breast tissue was extracted according to the protocol: “Isolation of Genomic DNA from FFPE Tissue Sections,” Qiagen's QIAmp DNA FFPE Tissue Handbook 10/2007 (Qiagen, Hilden, Germany).

Xylene treatment of the sample in step 3 of the protocol was modified to 2×5 min at 60°C. Buffer AW2 was replaced with 80% ethanol in step 18. DNA from BM blasts was extracted using the DNeasy Tissue kit from Qiagen. FFPE tissues were immunohistochemically stained using monoclonal antibody (mAb) anti-CD33 (Abnova, Taipei City, Taiwan) 1:100; mAb anti-CD34 (Dako, Glostrup, Denmark) 1:200; mAb TAP2, as described previously (3); MICA polyclonal antibody (LifeSpan BioSciences, Seattle, WA) 1:200; mAb CD3 (ThermoFisher Scientific, Waltham, MA) 1:500; and mAb CD56 (Novocastra, Wetzlar, Germany) 1:50. HLA typing was molecularly performed for HLA-A, -B, -C, -DRB1, -DQB1, and -DPB1 at the allele level based on a combination of sequencing PCR-SBT (Life Technologies, Carlsbad, CA) and PCR-SSP (Olerup, Vienna, Austria). KIR genotyping was performed by PCR-SSP (Olerup). Flow cytometric evaluation of cytotoxic T cells and helper T cells (CD8 FITC/CD4 PE/CD3 APC; DAK-TC660-Biozol, Eching, Germany) was performed using a standard lyse-wash procedure. CellquestPro software was used for measurement on a FACS Calibur flow cytometer (both BD Biosciences, San Jose, CA) equipped with two lasers (488 and 633 nm).


The authors thank Norbert Grunow, M.D., from the Institut für Pathologie, Klinikum Görlitz, Germany, for providing FFPE material, and also thank Kenan Onel, M.D., Ph.D., from the Pediatric Familial Cancer Clinic, University of Chicago, for critically reviewing the manuscript.


1. Thakar MS, Forman SJ. ASH evidence-based guidelines: Is there a role for second allogeneic transplant after relapse? Hematology Am Soc Hematol Educ Program 2009; 414.
2. Schmid C, Schleuning M, Ledderose G, et al.. Sequential regimen of chemotherapy, reduced-intensity conditioning for allogeneic stem-cell transplantation, and prophylactic donor lymphocyte transfusion in high-risk acute myeloid leukemia and myelodysplastic syndrome. J Clin Oncol 2005; 23: 5675.
3. Seliger B, Stoehr R, Handke D, et al.. Association of HLA class I antigen abnormalities with disease progression and early recurrence in prostate cancer. Cancer Immunol Immunother 2010; 59: 529.
4. Isoda T, Ford AM, Tomizawa D, et al.. Immunologically silent cancer clone transmission from mother to offspring. Proc Natl Acad Sci USA 2009; 106: 17882.
5. Murgia C, Pritchard JK, Kim SY, et al.. Clonal origin and evolution of a transmissible cancer. Cell 2006; 126: 477.
6. Siddle HV, Kreiss A, Eldridge MD, et al.. Transmission of a fatal clonal tumor by biting occurs due to depleted MHC diversity in a threatened carnivorous marsupial. Proc Natl Acad Sci USA 2007; 104: 16221.
7. Vago L, Perna SK, Zanussi M, et al.. Loss of mismatched HLA in leukemia after stem-cell transplantation. N Engl J Med 2009; 361: 478.
8. Mortimer J, Blinder MA, Schulman S, et al.. Relapse of acute leukemia after marrow transplantation: Natural history and results of subsequent therapy. J Clin Oncol 1989; 7: 50.
9. Platzbecker U, Thiede C, Fussel M, et al.. Reduced intensity conditioning allows for up-front allogeneic hematopoietic stem cell transplantation after cytoreductive induction therapy in newly-diagnosed high-risk acute myeloid leukemia. Leukemia 2006; 20: 707.
10. Cunningham I. Extramedullary sites of leukemia relapse after transplant. Leuk Lymphoma 2006; 47: 1754.
11. Bonnet D, Warren EH, Greenberg PD, et al.. CD8(+) minor histocompatibility antigen-specific cytotoxic T lymphocyte clones eliminate human acute myeloid leukemia stem cells. Proc Natl Acad Sci USA 1999; 96: 8639.
12. Farag SS, Fehniger TA, Ruggeri L, et al.. Natural killer cell receptors: New biology and insights into the graft-versus-leukemia effect. Blood 2002; 100: 1935.
13. Erler JT, Bennewith KL, Cox TR, et al.. Hypoxia-induced lysyl oxidase is a critical mediator of bone marrow cell recruitment to form the premetastatic niche. Cancer Cell 2009; 15: 35.
14. Kaplan RN, Riba RD, Zacharoulis S, et al.. VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic niche. Nature 2005; 438: 820.
15. Waterhouse M, Pfeifer D, Pantic M, et al.. Genome-wide profiling in AML patients relapsing after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant 2011; 17:1450.
16. Villalobos IB, Takahashi Y, Akatsuka Y, et al.. Relapse of leukemia with loss of mismatched HLA resulting from uniparental disomy after haploidentical hematopoietic stem cell transplantation. Blood 2010; 115: 3158.
17. Thiede C, Koch S, Creutzig E, et al.. Prevalence and prognostic impact of NPM1 mutations in 1485 adult patients with acute myeloid leukemia (AML). Blood 2006; 107: 4011.

HLA loss; Allogeneic; Extramedullary; AML

Supplemental Digital Content

© 2012 Lippincott Williams & Wilkins, Inc.