The majority of patients with acute promyelocytic leukemia (APL) possess either a bcr1 or a bcr3 type fusion between PML and RARα genes. The junction sequences may possibly be a target for immune response and influence susceptibility to the disease. In this case, HLA class I allele frequencies would be different between bcr1 and bcr3 patients. To test this hypothesis, we typed 102 APL patients for HLA-A, -B and -Cw alleles. The A*1, A*30, B*51, B*41, Cw*0602, and Cw*1701 alleles showed a different distribution between bcr1 and bcr3 patients, but in no case was this statistically significant after correction for the number of comparisons or was confirmed in an independent panel. Moreover, no difference was detected between bcr1 and bcr3 when HLA alleles were grouped according to their peptide binding specificities. Comparing HLA frequencies, clinical features at diagnosis and clinical outcome of the 64 patients homogeneously treated with all-trans retinoic acid and idarubicin (AIDA protocol) we observed a statistically significant association between HLA-B*13 and risk of relapse by univariate and multivariate regression analysis. Should this finding be confirmed in larger future studies, this observation would be of outmost importance in identifying patients at high risk of relapse in which more aggressive consolidation therapies should be used.
APL is characterized by a t(15;17) translocation, involving the retinoic acid receptor α (RARα) gene on chromosome 17 and the PML (promyelocytic) gene on chromosome 15. RARα belongs to the superfamily of nuclear hormone receptors while PML encodes a member of the zinc RING finger family of nuclear proteins. The translocation generates the PML/RARα chimeric gene which is expressed as a fusion product containing large portions of both PML and RAR proteins. The fusion protein acts as a transcriptional repressor that interferes with the normal pathway of gene expression involving the RARs/RXR (retinoid X receptor) complex.4
Within the PML gene, DNA breaks cluster in three different regions named bcr1, bcr2 and bcr3. Bcr1 and bcr3 are the most frequently involved in APL (92–95% of cases) and produce a unique PML/RARα fusion product, whereas bcr2 leads to several different products because of alternative splicing.5
The relevance of an HLA association study in APL stems from the following considerations: the junction regions of bcr1 and bcr3 regions code for distinct amino acid sequences. They can be considered as immunologically ‘non-self’ in that the peptides crossing the fusion point cannot be generated from the normal gene products (Figure 1). Therefore they could act as bcr1- and bcr3-specific tumor-associated antigens and could induce protective anti-APL immunity. The possibility that fusion proteins deriving from chromosomal translocations represent a candidate target for an HLA restricted T cell-mediated antitumor response has been already suggested for CML patients.67
Since APL cells do not express HLA class II molecules,89 it is more likely that any direct killing of such cells by cytotoxic T cells involves class I molecules and CD8+ T cell recognition. Thus, HLA alleles coding for molecules that bind bcr1 or bcr3 fusion peptides should be under-represented among the patients of either bcr1 or bcr3 type. If the immune response does not prevent the disease onset but it influences its clinical features, HLA class I allele frequencies could vary according to clinical parameters.
Accordingly, the aims of the present study were three-fold: (1) to collect the first data on HLA association in APL, by comparing the HLA frequencies in Italian APL patients with those of an ethnically matched control population; (2) to check whether there is any differential HLA profile in the two types of rearrangements that could underlie a different immune-modulated susceptibility; and (3) to test for any correlation between HLA, type of rearrangement and selected clinical parameters.
Materials and methods
One hundred and two APL patients were included in this study. The patient samples were diagnosed and provided by the clinical centers belonging to the Italian Cooperative Group GIMEMA over a period of time ranging from 1988 to 1997. All the patients were Italian. Approximately 43% of the patients were contributed by centers located in northern Italy, 38% in the center, 16% in the south and 3% in Sardinia. Diagnosis of APL was based on French–American–British guidelines.10 All patients presented PML/RARα rearrangements with or without a cytogenetically detectable t(15;17). At diagnosis, patients mean age was 39.4 (±15.7) years and white blood cell (WBC) count was 12.5 ± 22.3 × 109/l. Disseminate intravascular coagulopathy (DIC) and hemorrhages were present in 76.1% and 60.5% of cases. Bcr1 and bcr3 PML/RARα isoforms were in 66 and 36 patients, respectively. Sixty-four out of 102 patients were treated by the AIDA protocol in which ATRA is combined with idarubicin chemotherapy.11 In this latter group of patients, after a minimum follow-up of 1.5 years, 46/64 patients persisted in continuous complete remission (CCR) while 18 patients presented an overt relapse of disease.
Controls included 101 cadaver kidney donors and 200 cord blood units for which genomic HLA-A, B typing was performed in the Transplant Center of Turin1213 in Piedmont (north-west Italy). This population absorbed several migration waves from other Italian regions. The control panel, therefore, though composed predominantly (60%) by people of northern Italian origin, also includes a large (40%) component from the center and the south. HLA-A, B frequencies in this panel were very similar to those reported for 159 311 serologically typed Italian bone marrow donors whose regional composition was 59% from northern Italian regions, 35% from the center and south, and 5% from Sardinia.14 Thus the control panel can be considered as representative of the whole Italian population and reasonably corresponds to the regional distribution of the patients’ panel.
The control HLA-C frequencies were taken from those genomically typed in a limited sample from the Italian population (n = 101) during the XII Histocompatibility Workshop.15
In older samples (n = 55), the fusion type was determined on genomic DNA by Southern blot analysis.16 In more recently collected patient samples (n = 47), the hybrid transcripts were evaluated on cDNA by RT-PCR with a nested polymease chain reaction specific for bcr1, bcr2 and bcr3 fusion types.17 High molecular weight DNA was obtained from isolated mononuclear cells following proteinase-K digestion and phenol-chloroform extraction. Total RNA was extracted from cells cryopreserved in guanidine isothiocyanate according to the method of Chomczynsky and Sacchi.18 cDNA was obtained by retrotranscription of 1 μg of RNA by the Moloney leukemia virus derived reverse transcriptase (Gibco BRL, Italia), according to the manufacturer's instructions.
HLA class I typing
Class I alleles were typed either on genomic DNA (55 patients) or on cDNA (47 patients).
Genomic DNA from 39 bcr1 and 16 bcr3 patients was typed for HLA-A, -B and -Cw alleles by the SSP-ARMS-PCR (sequence specific primer-amplification refractory mutations system-polymerase chain reaction), using the 12th International Histocompatibility Workshop typing kit. It allows a medium to low resolution for all loci.19
cDNA of a further 27 bcr1 and 20 bcr3 patients was typed for HLA-A and -B alleles by a PCR-SSOP method,2021 that allows assignment of all the HLA class I serological specificities and a nearly complete assignment of molecular subtypes. 21 bcr1 and 20 bcr3 patients, included in the same cDNA panel, were also typed for Cw*0602 and Cw*1701. Cw*1701 was typed by the same SSP-ARMS-PCR approach utilized for genomic DNA typing: a 229 bp fragment was obtained after cDNA amplification. A specific approach was designed for Cw*0602: cDNA was amplified by primers 5′TACGGTCGACGATTCTTCCCAGACGCCGAGATGCGG3′ (forward) and 5′CATGGGATCCCAGGCAGCTGTCTC- AGGCTTTACAAG3′ (reverse) that coamplify HLA-B and C cDNA fragments.22 PCR products were then hybridized with the SSO AGCCCCGGGCGCCGT corresponding to the amino acid position 46–51.23
Reported frequencies for the various alleles refer to the number of individuals carrying a given allele (phenotypic frequency).
For categorical variables, significance of association was evaluated by X2 (Yates correction) or, when this is requested from the small number of expected cases, by two-tailed Fisher's exact test. For continuous variables (age and WBC count at presentation), the association was analyzed by the non-parametric Mann–Whitney test. The P values shown in the Tables were not corrected for multiple comparisons.
A multivariate analysis was performed to evaluate the influence of all variables simultaneously upon the clinical outcome (CCR or relapse) using the logistic regression model; in the full model all the variables are considered together, in the reduced model non-significantly-associated variables are excluded by a stepwise procedure.
HLA-A, -B, -Cw frequencies in APL patients
Considering the whole patient panel, regardless of the fusion type, the frequency of HLA-A, -B, -Cw tested alleles was similar to that reported for the Italian population (Tables 1 and 2). The only statistically significant difference concerned B*52 (P = 0.045) but it did not withstand correction for the number of comparisons. Genotype frequencies in the patients were in accordance with Hardy Weinberg equilibrium.
When considering the patients subdivided according to the bcr1 and bcr3 fusion type, borderline deviations between the two types were seen (Table 1) for A*1, A*30, B*51 that were decreased in bcr3 vs bcr1 patients (A*1: 12.5% vs 30%, P = 0.057; A*30: 2.8% vs 17.2%, P = 0.051; B*51: 5.6% vs 21.9%, P = 0.064) and for B*4101 that was decreased in bcr1 (1.6% vs 11.1%, P = 0.055). The observed differences were far from significant when P was corrected for the number of comparisons.
Of the 15 HLA Cw alleles typed on 55 APL genomic DNA samples (Table 2), Cw*0602 (P = 0.0274) and Cw*1701 (P = 0.0116) showed a statistically significant difference between bcr1 and bcr3. To confirm these results, an independent set of 41 APL cDNAs was typed for these two alleles. No significant difference in bcr1 and bcr3 types was seen either in this second panel (Table 2, bottom) or in the combined data set.
HLA-A and -B typing data were also evaluated according to so-called HLA-supertype groups (Table 3). Each HLA-supertype group consists of HLA class I alleles that share peptide binding motifs and thus possibly represents a functionally differentiated family.24 Four supertypes have been described: HLA-A2 including nine HLA-A allelic molecules, HLA-A3 including five -A alleles, HLA-B7 with 17 -B alleles and HLA-B44 including eight -B alleles. Also in this analysis no significant difference between bcr1 and bcr3 was found.
Correlation with clinical features
HLA frequencies were compared in patients subdivided according to age and WCB count at diagnosis and to the presence or absence of DIC or hemorrhagic syndrome. Some correlations were found, with generally modest statistical significance (Table 4). None of them was dependent on the fusion type. Correlation of HLA with response to therapy was tested in the 64 patients homogeneously treated with the AIDA protocol (all-trans retinoic acid combined with anthracycline-based chemotherapy). A significant increase of HLA-B*13 (33.3% vs 4.3%, P = 0.0046) with a relative risk of 11.0 (confidence limits 1.96–61.61) was found in relapsed patients (n = 18) as compared to patients maintaining a CCR (n = 46). Also this correlation was independent of the type of rearrangement (Table 5). A multivariate analysis including B*13, type of rearrangement, age and WBC count at onset, DIC and hemorrhage was performed using the logistic regression model on the 54 patients for which information for all the considered variables was available. The only significantly associated variable was B*13 with P = 0.008 in the full model and P = 0.006 in the reduced model.
The present study failed to detect any evidence of an HLA-class I modulated mechanism in susceptibility to APL. The most likely interpretation is that none of the APL fusion peptides binds to HLA class I with an affinity sufficient to be presented to cytotoxic T cells. This interpretation is in line with the observation that no known HLA binding motifs are present in the bcr1 and bcr3 PML-RARα fusion peptides.25 In fact, no direct binding to HLA-A*0201, B*5101 and B*4402 molecules was found for any of the nonapeptides possibly generated from the APL bcr1 and bcr3 fusion region.2627 Another possible explanation comes from the recent observation by Zheng et al28 that mutations involving PML down-regulate MHC expression with a dominant-negative mechanism acting on transcription of the genes involved in antigen processing. According to these authors, malfunction of PML consequent to its fusion with RARα might be the reason why TAP-1, TAP-2, LMP-2, LMP-7 and cell surface MHC class I are barely detectable on the NB4 APL cell line. If this is a common feature of APL cells, it may hinder host immune surveillance.
The present study also failed to show any bcr1- or bcr3-specific HLA correlation of several clinical parameters such as age and WBC counts at diagnosis and vascular complications.
Concerning correlations among HLA allele frequencies and patients’ clinical features and outcome, the most interesting observation is that HLA-B*13 was significantly increased in relapsed APL patients. This finding was further confirmed by a multivariate analysis which demonstrated that B*13 was the only parameter influencing the risk of relapse (P = 0.006). In our opinion, although derived from a retrospective evaluation, this finding is clinically relevant in order to identify poor prognostic factors with the currently adopted therapy programs for APL patients. In fact, although prognosis of APL patients has dramatically changed with the introduction of therapeutic programs in which all-trans retinoic acid (ATRA) is combined with anthracycline-based chemotherapy, early death during induction and occurrence of relapse still represent the major obstacles to final cure of this disease.
The biological meaning of this finding is not clear. Likely, HLA-B*13 may be a marker in linkage disequilibrium with another genetic factor that can influence response to therapy.
In conclusion, our present data did not show any evidence suggesting a possible HLA-class I related mechanism in suceptibility to APL. However, the statistically significant correlation between the presence of HLA-B*13 and relapse risk is a novel and clinically relevant observation. This finding, if confirmed in larger future prospective studies, might be useful to identify patients at high risk of relapse that should be treated with more aggressive consolidation therapies.
Hester JP, Rossen R, Trujillo J, McCredie KB, Freireich EJ . Frequency of HLA antigens in chronic myelocitic leukemia S Med J 1977 70: 691–693
Navarrete C, Alonso A, Awad J, McCloskey D, Ganesan TS, Amess J, Lister TA, Festenstein H . HLA class I and class II antigens association in acute leukemias J Immunogenet 1986 13: 77–84
Annino L, Ferrari A, Laurenti L, Perrone MP, Romani C, Pacchiarotti A, Girelli G, Mandelli F . HLA typing in hairy cell leukemia Leuk Lymphoma 1994 14: 63–65
Perez A, Kastner P, Sethi S, Lutz Y, Reibel C, Chambon P . PMLRAR homodimers: distinct DNA properties and heterodimeric interactions with RXR EMBO J 1993 12: 3171–3182
Pandolfi PP, Alcalay M, Zangrilli D, Mencarelli A, Diverio D, Biondi A, Lo Coco F, Rambaldi A, Grignani F, Rochette-Egly C, Gaub MP, Chambon P, Pelicci PG . Genomic variability and alternative splicings generate multiple PML/RARα transcripts that encode aberrant PML proteins and PML/RARα isoforms in acute promyelocytic leukemias EMBO J 1992 11: 1397–1407
Bocchia M, Korontsvit T, Xu Q, Mackinnon S, Yang SY, Sette A, Scheinberg DA . Specific human cellular immunity to bcr-abl oncogene-derived peptides Blood 1996 87: 3587–3592
Greco G, Fruci D, Accapezzato D, Barnaba V, Nisini R, Alimena G, Montefusco E, Vigneti E, Butler R, Tanigaki N, Tosi R . Two bcr-abl junction peptides bind HLA-A3 molecules and allow specific induction of human cytotoxic T lymphocytes Leukemia 1996 10: 639–699
Lecchi M, Lovisone E, Genetta C, Peruccio D, Resegotti L, Richiardi P . Gamma-IFN induces a differential expression of HLA-DR, DQ and DP antigens on peripheral blood myeloid leukemic blasts at various stages of differentiation Leukemia Res 1989 13: 221–226
Jinnai I . In vitro growth response to G-CSF and GM-CSF by bone marrow cells of patients with acute myeloyd leukemia Leukemia Res 1990 14: 227–240
Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, Gralnick HR, Sultan C . Proposal for the classification of the acute leukemias Br J Haematol 1976 33: 451–458
Mandelli F, Diverio D, Avvisati G, Luciano A, Barbui T, Bernasconi C, Broccia G, Cerri R, Falda M, Fioritoni G, Leoni F, Liso V, Petti MC, Rodeghiero F, Saglio G, Vegna ML, Visani G, Jehn U, Villemze R, Muus P, Pelicci PG, Biondi A, Lo Coco F . Molecular remission in the PML/RARa positive acute promyelocytic leukemia by combined all-trans retinoic acid and idarubicin AIDA therapy Blood 1997 90: 1014–1021
Fasano ME, Praticò L, Brancatello F, Mazzola G, Cravero T, Bersanti M, Curtoni ES . Caucasian Italian normal. In: Terasaki P, Gjertson D (eds) HLA 1998 American Society for Histocompatibility and Immunogenetics: Lenexa, KS 1999 p 162
Francia di Celle P, Gay E, Marin F, Maderno P, Berrino M, Dall'Omo A, Curtoni ES . Caucasian Italian normal. In: Terasaki P, Gjertson D (eds) HLA 1998 American Society for Histocompatibility and Immunogenetics: Lenexa, KS 1999 p 163
Rendine S, Borelli I, Barbanti M, Sacchi N, Roggero S, Curtoni ES . HLA polymorphisms in Italian bone marrow donors: a regional analysis Tissue Antigens 1998 52: 135–146
Clayton J, Lonjou C . Allele and haplotype frequencies for HLA loci in various ethnic groups. In: D Charron (ed) Genetic Diversity of HLA. Functional and Medical Implication EDK: Paris 1997 1: p 675
Diverio D, Lo Coco F, D'Adamo F, Biondi A, Fagioli M, Grignani F, Rambaldi A, Rossi V, Avvisati G, Petti MC, Testi AM, Liso V, Specchia G, Fioritoni G, Recchia A, Frassoni F, Ciolli S, Pelicci PG for the Italian Cooperative Study group GIMEMA . Identification of DNA rearrangements at the retinoic acid receptor alfa RARa locus in all patients with acute promyelocytic leukemia APL and mapping of APL breakpoints within the RARa second intron Blood 1992 79: 3331–3336
Diverio D, Riccioni R, Pistilli A, Buffolino S, Avvisati G, Mandelli F, Lo Coco F . Improved rapid detection of the PML/RARa fusion gene in acute promyelocytic leukemia Leukemia 1996 10: 1214–1216
Chomczynsky P, Sacchi N . Single step method of RNA isolation by acid guanidium thiocyanate-phenol-chloroform extraction Anal Biochem 1987 162: 156–158
Tonks S, Marsh SGE, Bunce M, Moses JH, Krausa P, Sadler AM, Petronzelli F, Bodmer JG . HLA class I DNA typing study. In: D Charron (ed). Genetic Diversity of HLA. Functional and Medical Implication EDK: Paris 1997; 1: pp188–193
Oh S, Fleischhauer K, Yang SY . Isoelectric focusing subtypes of HLA-A can be defined by oligonucleotide typing Tissue Antigens 1993 41: 135–142
Fleischhauer K, Zino E, Bordignon C, Benazzi E . Complete, generic and extensive fine-specificity typing of the HLA-B locus by the PCR-SSOP method Tissue Antigens 1995 46: 281–292
Zino E, Severini GM, Mazzi B, Bordignon C, Benazzi E, Fleischhauer K . Sequencing of a new HLA-A*32 subtype A*3202 Immunogenetics 1996 45: 76–77
Grundschober C, Rufer N, Sanchez-Mazas A, Madrigal A, Jeannet M, Roosnek E, Tiercy JM . Molecular characterization of HLA-C incompatibilities in HLA-ABDR-matched unrelated bone marrow donor-recipient pairs. Sequence of two new Cw alleles Cw*02023 and Cw*0707 and recognition by cytotoxic T lymphocytes Tissue Antigens 1997 49: 612–623
Sidney J, Grey HM, Kubo RT, Sette A . Practical, biochemical and evolutionary implications of the discovery of HLA class supermotifs Immunol Today 1996 17: 261–266
Gambacorti-Passerini C, Bertazzoli C, Dermime S, Scardino A, Schendel D, Parmiani G . Mapping of HLA class I binding motifs in 44 fusion proteins involved in human cancers Clin Cancer Res 1997 3: 675–683
Bocchia M, Wentworth PA, Southwood S, Sidney J, McGraw K, Scheinberg DA, Sette A . Specific binding of leukemia oncogene fusion protein peptides to HLA class I molecules Blood 1995 85: 2680–2684
Bolognesi E, D'Alfonso S, Greco G, Tanigaki N, Fleischhauer C, Cimino G, Rapanotti C, Diverio D, Momigliano-Richiardi P . HLA class I and susceptibility to acute promyelocytic leukemia Eur J Immunogenet 1998 25: S56
Zheng P, Guo Y, Niu Q, Levy DE, Dyck JA, Lu S, Sheiman LA, Liu Y . Proto-oncogene PML controls genes devoted to MHC class I antigen presentation Nature 1998 396: 373–376
This work was supported by grants from FIRC, the Italian Foundation for Cancer Research, and from Associazione Italiana per la Leucemia, sezione di Roma (ROMAIL). E Bolognesi is a recipient of a fellowship from Gruppo di Cooperazione in Cancerologia. MC Rapanotti is a recipient of a fellowship from FIRC. We are grateful to Dr R Tosi for critically reviewing the manuscript.
About this article
Cite this article
Bolognesi, E., Cimino, G., Diverio, D. et al. HLA class I in acute promyelocytic leukemia (APL): possible correlation with clinical outcome. Leukemia 14, 393–398 (2000). https://doi.org/10.1038/sj.leu.2401691
Incidence of extramedullary disease in patients with acute promyelocytic leukemia: a single-institution experience
International Journal of Hematology (2009)
Critical Reviews in Oncology/Hematology (2005)
CD34 expression is associated with poor clinical outcome in patients with acute promyelocytic leukemia
American Journal of Hematology (2003)
Current Oncology Reports (2002)