Search

The associations of cytogenetics with complete remission (CR) rates, overall survival (OS), and outcomes after CR were studied in 609 previously untreated AML patients younger than 56 years old in a clinical trial comparing 3 intensive postremission therapies: intensive chemotherapy, autologous transplantation (ABMT), or allogeneic bone marrow transplantation (alloBMT) from matched related donors. Patients were categorized into favorable, intermediate, unfavorable, and unknown cytogenetic risk groups based on pretreatment karyotypes. CR rates varied significantly (P

Pulmonary arterial hypertension (PAH) is a proliferative vasculopathy characterized by high circulating CD34+CD133+ proangiogenic progenitors, and endothelial cells that have pathologic expression of hypoxia-inducible factor 1 α (HIF-1α). Here, CD34+CD133+ progenitor cell numbers are shown to be higher in PAH bone marrow, blood, and pulmonary arteries than in healthy controls. The HIF-inducible myeloid-activating factors erythropoietin, stem cell factor (SCF), and hepatocyte growth factor (HGF) are also present at higher than normal levels in PAH blood, and related to disease severity. Primary endothelial cells harvested from human PAH lungs produce greater HGF and progenitor recruitment factor stromal-derived factor 1 α (SDF-1α) than control lung endothelial cells, and thus may contribute to bone marrow activation. Even though PAH patients had normal numbers of circulating blood elements, hematopoietic alterations in myeloid and erythroid lineages and reticulin fibrosis identified a subclinical myeloproliferative process. Unexpectedly, evaluation of bone marrow progenitors and reticulin in nonaffected family members of patients with familial PAH revealed similar myeloid abnormalities. Altogether, the results show that PAH is linked to myeloid abnormalities, some of which may be related to increased production of HIF-inducible factors by diseased pulmonary vasculature, but findings in nonaffected family suggest myeloid abnormalities may be intrinsic to the disease process.

The 8;21 translocation is one of the most common chromosomal translocations in acute myelogenous leukemia (AML), accounting for 40% of pediatric AML with French-American-British (FAB)-M2 morphology. The chromosomal breakpoints have recently been identified at the molecular level and shown to involve the AML1 gene on chromosome 21 and the ETO gene on chromosome 8. Translocation results in the consistent fusion of these genes on the der(8) chromosome, resulting in the production of a novel chimeric gene and message. Using oligonucleotide primers derived from the AML1 and ETO cDNAs, we were able to amplify a specific fusion transcript from 26 of 26 patients with t(8;21) by a reverse transcriptase polymerase chain reaction (PCR) approach. DNA fragments of identical size were generated from each case including two with complex translocations. Studies on the sensitivity and specificity of this approach show that PCR analysis can be used as a rapid, accurate, and sensitive means for detecting this chromosomal abnormality, and for following the patients' response to therapy.

Using metaphase cytogenetics (MC), chromosomal abnormalities are found in only a proportion of patients with myelodysplastic syndrome (MDS). We hypothesized that with new precise methods more cryptic karyotypic lesions can be uncovered that may show important clinical implications. We have applied 250K single nucleotide polymorphisms (SNP) arrays (SNP-A) to study chromosomal lesions in samples from 174 patients (94 MDS, 33 secondary acute myeloid leukemia [sAML], and 47 myelodysplastic/myeloproliferative disease [MDS/MPD]) and 76 controls. Using SNP-A, aberrations were found in around three-fourths of MDS, MDS/MPD, and sAML (vs 59%, 37%, 53% by MC; in 8% of patients MC was unsuccessful). Previously unrecognized lesions were detected in patients with normal MC and in those with known lesions. Moreover, segmental uniparental disomy (UPD) was found in 20% of MDS, 23% of sAML, and 35% of MDS/MPD patients, a lesion resulting in copy-neutral loss of heterozygosity undetectable by MC. The potential clinical significance of abnormalities detected by SNP-A, but not seen on MC, was demonstrated by their impact on overall survival. UPD involving chromosomes frequently affected by deletions may have prognostic implications similar to the deletions visible by MC. SNP-A–based karyotyping shows superior resolution for chromosomal defects, including UPD. This technique further complements MC to improve clinical prognosis and targeted therapies.

Abstract 2555 High hyperdiploidy is present in 30% of children with acute lymphocytic leukemia (ALL), and is associated with a favorable prognosis. We evaluated pts with newly diagnosed ALL treated on SWOG trials S9400 (1995–2000) and S0333 (2005–2010) to determine the prevalence and prognostic impact of hyperdiploidy in adults with ALL. Additionally, we examined the prognostic impact of hypodiploidy, a feature typically associated with a poor prognosis in children. Methods: One-hundred and eighty-five pts treated on S9400 and S0333 with successful cytogenetic (CG) analysis were included. The treatment regimens were: S9400 [Induction: Daunorubicin (D), vincristine (V), prednisone (P), PEG-asparaginase (PEG); Consolidation: Cytoxan (Cy), cytarabine (AraC), 6-mercaptopurine (6MP), intrathecal methotrexate (IT Mtx). Consolidation was followed by allogeneic stem cell transplant or maintenance chemotherapy] and S0333: Double Induction Chemotherapy [Induction 1: D, V, P, PEG; Induction 2: high dose AraC, mitoxantrone, decadron. Consolidation: Cy, AraC, 6MP, Mtx; consolidation was followed by maintenance therapy]. Karyotypes were centrally reviewed and clonal abnormalities described according to ISCN (2009). Hyperdiploidy was defined as: low hyperdiploidy [47–49 chromosomes (cs)], high hyperdiploidy (51–65 cs), near triploidy (66–79 cs), and near tetraploidy (84–100 cs). Hypodiploidy was defined as: near haploidy (25–29 cs), low hypodiploidy (31–39 cs), and high hypodiploidy (42–45 cs). When more than one cell line was present, ploidy was assigned by the most complex clonal karyotype. Hypodiploidy and hyperdiploidy were analyzed as prognostic factors for complete response (CR) rate and residual disease (RD) by logistic regression and chi-square tests; and for overall survival (OS) and relapse-free survival (RFS) by proportional hazards. Multivariable analyses were stratified by study and using the baseline variables: age, WBC, lineage, and CG risk. Results: The median age was 32 yrs (range 17–64), and median WBC at diagnosis 17.2 K/uL (range 0.6–396.6). CG risk was ascribed by (Pullarket V et al. Blood 2008; 111: 2563). Forty-five pts (24%) had normal CG, and 73 (39%) had poor risk CG. Fourteen pts (8%) had hypodiploidy (2: low hypodiploidy; 12: high hypodiploidy). Fifty-three pts (29%) had hyperdiploidy [40: low hyperdiploidy, 10: high hyperdiploidy (5%), 3: near tetraploidy or tetraploidy (2%)]. The CR rate for all pts was 72%; with a median RFS of 15 mos (95% CI: 12–29 mos) and median OS of 28 mos (95% CI: 21–36 mos). There was no significant association with ploidy status and age, WBC, or lineage. However, there was an increased prevalence of the t(9;22) in the high hypodiploidy group compared to the normal/pseudo diploidy group (p=0.049). Neither hypodiploidy nor hyperdiploidy were predictive of CR or RD; although pts with hypodiploidy had a higher rate of RD (p=0.062). The 2 pts with low hypodiploidy had very poor outcomes (1 had RD and died after 11 mos; the other relapsed after 3 mos from CR and died 4 mos after study registration). There were no statistically significant differences in OS, CR rate, or RFS between the ploidy groups even after adjusting for baseline characteristics in multivariate analysis. Surprisingly, when excluding pts with poor risk CG there was still a trend towards a worse RFS (29 vs. 32 months, p=0.20) and OS (40 vs. 68 mos, p=0.29) in pts with hyperdiploidy compared to normal/pseudodiploidy. In addition, the 3 pts in the high hyperdiploidy group without poor risk CG had poor OS (median 23 mos). Conclusions: The prevalence of high hyperdiploidy is much lower in adults with ALL, compared to children. The prevalence of hypodiploidy and near tetraploidy/tetraploidy is comparable to that seen in children with ALL. Hypodiploidy and high hyperdiploidy were not prognostic factors for outcome in this group of patients. Given the low prevalence of these abnormalities, it is possible that larger numbers of pts may be needed to detect such a difference. The poor outcomes of pts with low hypodiploidy are consistent with findings by Moorman et al. (Blood 2006; 109: 3189). However, in contrast to Moorman's results, there was no evidence of an association of hyperdiploidy with age/WBC, and there was a trend towards a worse prognosis in this subset of patients. This suggests that the biology and prognosis of high hyperdiploidy may be affected more by WBC and age in the adult population. Disclosures: No relevant conflicts of interest to declare.

Abstract 1086 Poster Board I-108 T-cell large granular lymphocyte (LGL) leukemia is a rare clonal lymphoproliferative disorder derived from cytotoxic T-lymphocytes (CTL) associated mostly with lineage-restricted cytopenias. While the clinical course is often indolent, some patients exhibit severe morbidities due to transfusion dependence, infections or thrombocytopenia. Intuitively, the level of the aberrant CTL clone as expressed by high LGL count or TCR Vβ clonal expansion by flow cytometry, the severity of neutropenia, pancytopenia, or the association with clonal myeloid disorders such as myelodysplastic syndromes (MDS) should herald a worse clinical outcome. Moreover, expression of CD56 has been established as an adverse marker for morbidity and mortality in number of hematological malignancies. In addition, rare cases of aggressive T-cell LGL leukemia were reported to display a CD3+/CD56+ immunophenotype. Based on the availability of a large, well characterized cohort of patients with T-cell LGL leukemia (n=86), we studied features associated with a poor clinical outcome and perceived need for aggressive therapy, including histomorphologic parameters, immunophenotype (CD56 expression). Both clinical response and overall survival (OS) were analyzed by means of categorical and survival statistical methods. The median patient age was 64 years (range, 15–80), and 55% were males. Rheumatoid arthritis was present in 14% of patients, 47% of patients had splenomegaly and 43% of those underwent splenectomy. Concomitant hematologic malignancies (5 cases of plasma cell dyscrasia, 8 with B-cell malignancies, and 3 with MDS/sAML) were found in 18% of patients, whereas solid cancers accounted for 15%. Neutropenia, anemia, and thrombocytopenia were seen in 63%, 50%, and 24%, respectively. Overall clinical response to a variety of therapies given was observed in 60% of patients. CD56 expression on the LGL clone was found in 15 patients (21%). Cases expressing CD56 were enriched among females (p=.005). CD56-positive cases were less likely to have neutropenia (p=.03) or thrombocytopenia (p=.03). In contrast to aggressive NK-cell lymphoma, the CD56 phenotype in T-cell LGL leukemia did not negatively impact OS (p=.85). However, increasing number of cytopenias (p=.006) were associated with poor survival. A dose-response pattern of association with OS was detected for pancytopenia (hazard ratio [HR]=9.9, p=.002) vs. bicytopenia (HR=4, p=.06) vs. single/none (reference) cytopenia. Similar results were obtained from logistic regression of factors associated with clinical response to therapeutic intervention. Neutropenia (p=.004) and thrombocytopenia (p=.02), but not anemia were associated with poorer clinical responses. All 5 patients with plasma cell dyscrasia had a complete response to the therapies targeting LGL (p=.01). Similarly, patients who underwent splenectomy tended to have a more favorable clinical response (p=.045). Multiple lineage cytopenias adversely affect both clinical outcomes and OS of T-cell LGL leukemia. In contrast to other diseases (T-cell acute lymphoblastic leukemia, AML, and multiple myeloma) a CD56+ immunophenotype was not associated with poor outcome in our cohort. Thus, as opposed to other studies, we would not suggest aggressive systemic chemotherapy in management of patients with T-cell LGL leukemia based purely on CD56 expression. Disclosures No relevant conflicts of interest to declare.

We have evaluated various types of single nucleotide polymorphism arrays (SNP-A) as a karyotyping platform in over 600 cases of bone marrow failure and various myeloid disorders including MDS, MDS/MPD, and AML and in over 360 controls. We have shown that SNP-A not only reliably confirms chromosome gains and losses identified by metaphase cytogenetics (MC) but also allows for detection of previously cryptic chromosomal lesions. Moreover, through the ability of combining copy number measurement with genotyping, SNP-A also enables detection of copy-neutral loss of heterozygosity (LOH), a lesion that cannot be recognized using traditional MC. We have previously shown that this type of lesion occurs frequently in myeloid disorders. Recently, ultra-high density SNP-A were introduced (Affymetrix 6.0 array platform) with modified chemistry and a combination of SNP (>906,600) and copy number variant probes (CNV >946,000) to allow for precise delineation of the location and size of submicroscopic chromosomal defects, copy-neutral LOH and germline CNVs. This technology will likely supplant previous SNP-A platforms in performance and opens a new era in clinical cytogenetic diagnostics. In this pilot trial, we applied the Affymetrix SNP 6.0 array to samples obtained at diagnosis from 193 patients (59 AML, 115 MDS, 19 MDS/MPD). For 95 samples, both Affymetrix 250K and 6.0 arrays were applied; 6.0 SNP-A allowed for more identification of 33 regions of copy number aberrations and 11 copy-neutral LOH, suggesting a higher resolution of this technology over 250K arrays. By MC, 43% (71/167) of patients had chromosomal aberrations including a complex karyotype; 48% (81/167) had a normal karyotype and 9% (15/167) showed no growth. By 6.0 SNP-A, aberrations were detected in 78% (150/193) of patients. Significantly, among the samples with either a normal karyotype or non-informative results, aberrations were found by SNP-A in 57% and 73%, respectively. In an illustrative case with a balanced translocation by MC, copy number changes were found at the junctions, indicating that the translocation breakpoints contained small deletions. Comparison of 6.0 SNP-A results with MC showed concordance for detection of MC-defined lesions. However, 9% (15/167) of defects were only detected by MC and could not be detected by SNP 6.0 array in patients with a low proportion of abnormal metaphases. Gains and losses detected by SNP-A 6.0 required a sensitivity threshold of 20% clonal cells. We identified recurrent lesions not previously reported as being CNV. These comprised 238 regions of loss, 257 regions of gain, and 97 regions of copy-neutral LOH. When feasible (or if indicated to resolve a discordance with MC or to exclude CNV), DNA from sorted CD3+ cells was used as a reference. Recurrent losses were located in 37%(71/193) patients at 5q, 7p, 7q, 13q, 17p, 20q, 1p, 12p, 3q and 4q. Recurrent gains in 28%(55/193) patients included 13q, 11p, 11q, 21q, 3p, 8p, 8q, 9q, 19p and 5p. Somatic copy-neutral LOH was identified in 20%(39/193) patients at 13q, 1p, 5q, 6p, 7q, 11q, 17q, 3p, 4q, and 12q. Homozygous deletion was observed in one patient with a 6 Mb loss at 5q34 (not visible on MC) within a heterozygous 5q11.2–5q35.3 deletion Copy neutral LOH was found in 29% (56/193) patients overall. AML patients appeared to have more frequent copy neutral LOH(34%) than MDS (24%), but not significant (P=0.2). The 5%(9/193) of patients had more than 2 chromosomes with copy-neutral LOH. Among the samples showing a normal karyotype, 20% (16/81) of patients showed copy -neutral LOH as the sole aberration without additional copy number aberrations. To our knowledge, this is the first systematic application of SNP-A 6.0 in a large cohort of patients with myeloid malignancies. Our results demonstrate that this technology complements, but does not replace conventional MC for identification of abnormalities in myeloid neoplasia.

Large granular lymphocyte leukemia (LGL leukemia) is a semiautonomous clonal lymphoproliferation of cytotoxic T cells associated with various immune cytopenias. Within the spectrum of acquired immune-mediated bone marrow failure states, LGL leukemia can serve as monoclonal model of usually polyclonal T cell-mediated pathology. Mechanisms of unopposed clonal expansion of LGL cells in many aspects resemble true lymphoma and are not well understood. In addition to its reactive character, intrinsic clonal defects may be present in some patient with LGL, in particular those with more pronounced lymphoproliferative features. However, unlike in B-cell lymphomas, recurrent chromosomal abnormalities have not been frequently identified in LGL leukemia using traditional metaphase karyotyping techniques. We have applied Affymetrix 250K single nucleotide polymorphism-arrays (SNP-A) and 6.0 SNP-A in 28 patients with LGL leukemia, to elicit a far higher resolution of chromosomal content through genomic mapping of individual SNP and respective copy number analysis in LGL leukemia. SNP-A based cytogenetics have been applied successfully to MDS patients and has increased prognostic reliability. Blood mononuclear cells containing high proportion of clonal cells as determined by TCR Vβ flow cytometry were used as a source of DNA. For comparison, a large number of control blood and marrow specimens (N=119 for 6.0 and 124 for 250K SNP arrays) were analyzed. Data were processed using Genotyping Console v2.1 software (Affymetrix, Santa Clara, CA). After exclusion of known copy number variants (CNV) referenced in public databases and our own set of 178 normal controls, we found distinct chromosomal changes in 16/28 (57%) of LGL leukemia patients. Consensus regions of deletion/gain or uniparental disomy (UPD) were identified. The most common abnormalities included either UPD or copy number loss of chromosome 3q21.2–q21.31, which was identified in 6 (21%) patients. This region harbors CD86, the gene encoding the B7.2 protein responsible for T cell activation and regulation through costimulatory mechanisms. Copy number loss/UPD was also identified at chromosome 1p31.1–p32.3 in 4 (18%) patients. Interestingly, copy number gain in this same region was identified in 2 (9%) patients suggesting that this region may correspond to either a new, infrequent germ line encoded CNV or represents a somatic microdeletion. Recent studies indicated a role for SIL/SCL, which resides at this locus, in V(D)J recombination, lymphocyte development, and maturation. Additional conserved chromosomal abnormalities included copy number gains in 14q (11%) and copy number loss at 11p15 (14%), all possibly representing germ line or somatic CNV physiologically acquired during lymphocyte ontogenesis. We compared chromosomal lesions identified in our LGL cohort with clinical features including age, presence of neutropenia, anemia, thrombocytopenia, splenomegaly, immunophenotype, and degree of clonal expansion. No difference in clinical course was found between patients with and without cytogenetic abnormalities with regard to type and severity of cytopenias or size of the LGL clone. However, patients with 1p31.1–p32.3 deletions were found to have lesser degree of neutropenia compared to patients without 1p31.1–p32.3 deletions (p

Refractory anemia with ringed sideroblasts and thrombocytosis (RARS-T) has been considered a provisional subtype within the diagnostic entity of myelodysplastic/myeloproliferative diseases (MDS/MPD). Since JAK2 V617F mutation is present in a significant proportion of RARS-T patients (Szpurka et al. Blood, 2006), many investigators consider this entity to be more closely related to classical MPD. However, a significant minority of patients with RARS-T do not display a JAK2 V617F mutation. We have studied a cohort of patients with RARS-T (N=18) characterized by the presence of ringed sideroblasts, reticulin fibrosis and thrombocytosis (>600×109/L), that lack obvious causes of secondary thrombocytosis. While 8/18 patients harbored a JAK2 V617F mutation, a molecular pathogenesis for the remaining patients was unexplained. The successful application of SNP-A to characterize the genomic lesions in MDS prompted us to use this technology to study RARS-T. SNP-A allows detection of copy neutral loss of heterozygosity such as UPD9p which is associated with JAK2 V617F mutation. SNP-A facilitated detection of previously cryptic lesions; 9/18 patients showed an abnormal SNP-A-based karyotype often involving multiple lesions (only 5 of these defects were detected by metaphase cytogenetics). The new lesions seen by SNP-A included gains of chromosome 11p, 20q and 21q; deletion of 2p and various areas of UPD including 1p, 9p, 6p, 2p and 8p. SNP-A allowed identification of seemingly invariant UPD1p in 4/18 patients. As this region includes the Mpl gene, we analyzed patients for the presence of MPL W515L/K mutations which have been described in MPD. We did not find any patients with MPL W515K, however MPL W515L mutation was present in 2/4 RARS-T patients with UPD1p; another patient showed monoallelic MPL W515L variant. In addition, 1 patient with UPD1p harbored both JAK2 V617F and MPL W515L mutations. To further delineate the molecular lesion we analyzed all patients for the presence of abnormal STAT5 activation. An aberrant phospho-STAT5 staining pattern was present in all cases that were positive for either JAK2 V617F or MPL W515L mutations (N=10); unexplained STAT5 activation was found in only 4 cases, pointing towards a molecular defect involving this pathway. In these 4 patients, and in 1 additional with UPD1p who did not harbor MPL W515L mutation, we searched for other genes which might explain the pathogenesis of this disease by potentially causing aberrant activation of STAT5. We sequenced Jak1T478S, Jak1V623A and Ntrk1S677N as well as the transmembrane, juxtamembrane and kinase domain of Tie1, Epha2 and Ephb2 genes, but no mutation was found. In addition, we found a group of phospho-STAT5-negative patients (N=4) that showed typical genetic features of myelodysplasia e.g. del(5), +8 and partial loss of chromosome X; these cases are probably best considered to be of MDS origin rather than MPD. To our knowledge, our work is the first description of biallelic MPL W515L mutation and UPD1p found in RARS-T patients. This data is important for understanding the clonal selection process and pathophysiology of activating mutations in MDS/MPD. Overall, our studies demonstrate that somatic UPD1p is associated with homozygous MPL W515L mutation in MDS/MPD cases. Localizing areas of somatic UPD by SNP-A may help identify candidate genes within the shared regions that are likely targets for mutations.

Areas of loss of heterozygosity (LOH) can be precisely delineated using single nucleotide polymorphism arrays (SNP-A) allowing for detection of submicroscopic chromosomal defects and segmental somatic uniparental disomy (UPD) not revealed by metaphase cytogenetic analysis (MC). This study focused on aberration of chromosome 17. We analyzed marrow specimens in 1162 MDS/AML by MC and found 39 patients whose karyotype involved monosomy 17. All had a complex karyotype, aggressive histomorphologic features (1/39 low risk, 26/39 advanced MDS/sAML, 2/39 MDS/MPD, and 10/39 pAML) with a median survival of 3 months. In addition to loss of chromosome 17, 35/39 patients also showed either −5/del(5q) (N=10), −7/del(7q) (N=2), or both (N=23). To better delineate the boundaries of LOH on 17th chromosome, we analyzed a subset of 532 patients by SNP-A and identified 43/532 samples with an abnormal chromosome 17; 28 had interstitial deletions and 15 had somatic UPD. In 17/19 samples with monosomy 17 by MC, SNP-A revealed a deletion in 17p or 17q, indicating incomplete loss of chromosome 17 material. SNP-A yielded a total of 11 additional lesions on 17q not detected by MC. We were able to define two commonly deleted regions (CDR1 and CDR2). CDR1 (bp 6,828,482 to 8,075,871; 1.25 Mb) encompassed around 90 genes, including TP53, and was present in 11/14 samples with del17p. CDR2 (bp 25,320,435 to 27,355,332; 2Mb) was detected in 7/14 patients and encompassed approximately 33 genes, including NF1. Overall, the frequency of UPD17 was high: 17p UPD was detected in 7 and 17q in 8 samples analyzed. In all cases with 17p UPD, the region of UPD overlapped with CDR1. CDR2 overlapped with the region of 17q UPD in 4/8 samples. In analogy to monosomy 17, 18/21patients with LOH 17p (7 UPD, 14 losses) had a complex karyotype, 21/21 had aggressive histomorphologic features (1/21 RCMD, 3/21 RCMD-RS, 6/21 RAEB-1/2, 3/21 pAML, 8/21sAML) and a poor prognosis with a median survival of 2.6 months. Moreover, in 13/14 patients with del(17p) by SNP-A, −5/del(5q) (N=1) or both −5/del(5q) and −7/del(7q) (N=12) were present. One patient did not show deletions of chromosomes 5 or 7, but had del (4)(q26) and del(6)(q23.2). No patient had del(17p) as the sole abnormality. Strong association between 17p UPD and abnormalities of chromosomes 5 and/or 7 was also noted: of 7 patients with 17p UPD, 3 had 5/del(5q), 1 showed −7/del(7q), and 3 had −5/del(5q) and −7/del(7q). We hypothesized that LOH within the 17p CDR1 that includes TP53 might be associated with a distinct clinical phenotype and point toward pathogenic TP53 mutations. Overall, 18 instances of 17p LOH included the TP53 locus. When TP53 exons 5–9 were screened for mutations in patients with 17p LOH, we found biallelic TP53 mutations in 5/6 patients with somatic 17p UPD and in 6/8 patients with 17p deletions. We detected 10 missense mutations and 1 insertion. All missense mutations were located in the DNA-binding domain of TP53 (4/10 in exon 5: C141Y, V172F, C176Y, H179Q; 2/10 exon 6: H193N, H193R; 1/10 exon 7: R249G and 3/10 exon 8: V272L, V272M, R273H). Our study demonstrates that LOH of 17p in myeloid malignancies should prompt consideration of TP53 mutation. TP53 mutation is linked with an aggressive clinical phenotype and is highly associated with partial or complete loss of chromosomes 5 and/or 7. To our knowledge this is the first report of biallelic TP53 mutations due to UPD17p in myeloid malignancies, and indicates that both heterozygous and homozygous mutations can be encountered and comprise part of the pathologic continuum of the selection process of malignant myeloid clones.

Deletion of the long arm of chromosome 20 (20q) is one of the more common recurring abnormalities associated with myeloid malignancies. The previous studies utilizing FISH and microsatellite PCR attempted to define the commonly deleted regions (CDR) of chromosome 20, but due to the limited resolution of conventional technologies, the precise delineation of such a region was difficult. Among our cohort of 1162 patients with myeloid malignancies, 3% of the patients (34/1162) showed 20q deletion (27 MDS, 2 AML, 5 MDS/MPD). The incidence of 20q deletion as a sole cytogenetic abnormality was 44% (15/34). To decrease the possible size of CDR and better delineate the boundaries of areas of copy-neutral LOH, we analyzed a subcohort of 532 of the patients using high density single nucleotide polymorphism microarray (SNP-A) with 250K and SNP 6.0 including 1.8x106 markers. As a result, we were able to identify chromosome 20q deletion in 21 patients (2 RARS, 3 RCMD, 6 RAEB, 4 MDS-U, 2 AML, 2 MPD, 2 MDS/MPD), while metaphase cytogenetics in the same cohort of patient revealed only 14 cases with 20q deletion. Although the size and breakpoints of the deleted regions varied, they showed considerable overlap. We have defined 2 CDRs on 20q shared by all patients with deletion of chromosome 20q. CDR1 (present 18/21 patients) spanned 2.5Mb and is located between bands 20q11.23 and 20q12 (34,968,632-37,417,814), while CDR2 (present in 21/21 patients) encompassed 1.8Mb within bands 20q13.12 (43,636,105-45,417,399). We found 3 patients (1 RAEB and 2 MDS-U) whose lesions included CDR2 only and without CDR1. Two patients with MDS-U had 20q deletion as a sole abnormality and belong to the low and intermediate-1 risk group, respectively. One RAEB patient who only had CDR2 had duplication of the short arm of chromosome 20 and progressed to AML. As most of the lesions were interstitial, the terminal regions of chromosome 20q and the short arms of chromosome 20 were retained or even duplicated in 8 cases. For RARS and MPD, MDS/MPD patients, there were no patients that have duplicated regions. CDR1 and CDR2 spanned 31 and 33 genes known to be expressed in hematopoietic tissues. We selected 3 candidate genes (retinoblastoma like protein 1), CTNNBL1 (beta-catenine like protein 1) and TP53RK (TP53 regulating kinase) and we did not find any inactivating mutations when the genes were sequenced. Theoretically, deletions result in haploinsufficiency but underexpression may be due to methylation silencing which can result in events similar to a functional knockout if it affects genes in the deleted area. Using 27K methylation arrays, we studied methylation of genes corresponding to CDR1 and CDR2 (10 patients with 20q deletion and 53 patients without 20q deletion). CpG islands that were hypermethylated within these CDR regions relative to controls were studied and there was no hypermethylated CpG site in >25% of patients for CDR1. There were 4 hypermethylated CpG sites (MMP9, SLC12A5, SLC13A3 and CHD22) in >25% of patients for CDR2. It is still unclear if SLC12A5, SLC13A3 and CDH22 are associated with cancer. Among 10 patients with 20q CDR2 deletions, we found 3 patients with hypermethylated CpG sites within MMP9 gene (1 AML, 2 MDS-U). Based on the assumption that methylation can results in deficient gene expression analogous to haploinsufficiency and hypermethylation of genes corresponding to the CDRs in patients without deletion could result in a similar phenotype. Among 53 patients without deletion of 20q, we found hypermethylated MMP9 genes in 42% (22/53) of patients (13 AML, 1 RA, 4 RCMD, 1 RAEB, 1 5q- syndrome, 1 MDS/MPD). Among 10 patients with 20q deletion, there was no evidence of hypermethylation in our candidate genes of RBL1, CTNNBL1 and TP53RK genes. For CTNNBL1 gene, 25% (13/53) of patients without 20q deletion showed hypermethylation (8 AML, 1 RAEB, 3 RCMD, 1 MDS-U). In conclusion, our results suggest that CDR1 and CDR2 regions of chromosome 20 which were delineated by SNP-A may be associated with pathogenesis of myeloid malignancies and candidate genes of RBL1, CTNNBL1, TP53RK and MMP9 need to be further studied.

Cytogenetic aberrations identified by conventional metaphase cytogenetics (MC) have an important diagnostic and prognostic role in evaluating patients with myelodysplastic syndromes (MDS), and results can affect the choice of therapy. Fluorescence in situ hybridization (FISH) can complement MC by providing information derived from both interphase and metaphase nuclei. However, clinically practical FISH strategies are limited to detection of the most common lesions in MDS, including −5/del(5q), −7/del(7q), del(20q), and trisomy 8. The ability to obtain informative results from interphase nuclei and the relative ease of scoring greater numbers of cells are advantages of FISH as compared to MC. Still, the clinical relevance of small numbers of abnormal cells, apart from detection of residual disease, remains unclear. Single nucleotide polymorphism array (SNP-A)-based karyotyping can reveal genetically unbalanced defects with superior resolution compared to MC and FISH, as well as identify segmental uniparental disomy (UPD) that cannot be detected by either method. We sought to determine whether the overall diagnostic yield for detecting common recurring genetic defects associated with MDS could be improved using a strategy incorporating MC, FISH and SNP-A. Using a standardized approach, we focused our investigation on detection of −5/del(5q), −7/del(7q), trisomy 8, and del(20q). We studied 62 patients, including 42 MDS, 5 MDS/MPD, and 15 secondary AML, with standard MC, FISH probes for chromosomes 5, 7, 8, and 20, and SNP-A karyotyping using Affymetrix 250K and/or 6.0 SNP array platform. The detection rate for del(5q) was 35%, 35%, and 37% by MC, FISH, and SNP-A, respectively. No single method detected all of the defects, and detection rates improved when results of all methods were combined. For example, the rate for detection of del(5q) increased incrementally to 39% (MC+FISH), 44% (MC+SNP-A), 42% (FISH+SNP-A), and 44% when all 3 methods were applied. Similar findings were observed for −7/del(7q), trisomy 8, and −20/del(20q): after combining all methods the detection rates improved from 8% to 17%, from 10% to 17%, and from 8% to 10%, respectively, as compared with MC alone. Discrepant results among these methods were related to poor growth (N=2) and low percentage of positive metaphases (small clonal size; N=2). In addition, small somatic deletions (N=6) and UPD (N=2) were not detected by MC or FISH. Larger defects that were detected by SNP-A (e.g., from 5q14.2 to 5q23.1) did not overlap with either loci 5p15.2 (D5S630) or 5q31 (EGR1) used in the FISH probes. We conclude that metaphase cytogenetics, interphase FISH, and SNP-A are complementary techniques that, when applied and interpreted together, can improve the diagnostic yield for identifying genetic lesions in MDS. SNP-A allows for identification of topographically smaller defects and copy-neutral loss of heterozygosity without a requirement for successful cytogenetic analysis. While FISH affords the ability to quantitate the number of affected cells, it is only useful to screen for specific, known defects of a certain size. Whether novel defects as identified by FISH or SNP-A karyotyping will have prognostic impact or affect the results of therapy is the subject of ongoing investigation.

Among WHO low-risk categories of MDS, refractory anemia with ringed sideroblasts (RARS) can be more accurately diagnosed by characteristic pathomorphology. Clonal hematopoiesis and chromosomal abnormalities exemplify a close pathogenetic relationship to other forms of MDS. RARS shows considerable clinical variability even for patients (pts) with identical cytogenetic defects. Due to the low resolution of metaphase cytogenetics (MC) and its dependence on cell growth in vitro, this test is often non-informative in MDS. High-density SNP arrays (SNP-A) allow for a precise identification of unbalanced genomic lesions and copy-neutral loss of heterozygozity. We hypothesize that cryptic chromosomal (chr) aberrations exist in most, if not all, pts with RARS. Their detection may help to improve prognostication, distinguish distinct phenotypes and point towards unifying pathogenic defects. Initially, we analyzed the results of MC in pts with MDS and MDS/MPD (N=455) and in a sub-cohort of RARS, RCMD-RS, RARSt and other MDS subtypes with >15% RS. When we compared pts with/without RS, chr defects were found at comparable frequencies (∼50%). The most commonly occurring defects associated with RS, compared to other forms of MDS, included those of chr 5 (9% vs. 16%, 7 (8% vs. 12%) and 20 (3% vs. 8%). DNA was available for 36 pts with RS and was subjected to 250K SNP-A karyotyping. Pathologic lesions were defined upon exclusion of normal copy number polymorphisms identified in 81 controls (O’Keefe at al ASH 2007), as well as the Database of Genomic Variants (http://projects.tcag.ca/variation). By MC, a defective karyotype was present in 16/36 pts (44%). Deletions involving chr 5, 7 and complex MC were found in 3, 5, and 2pts, respectively. However, when SNP-A was applied as a karyotyping tool (copy number and LOH analysis), all aberrations found by MC were confirmed, but also new lesions were detected so that an abnormal karyotype was established in 62% of pts. Several previously cryptic/recurrent lesions included losses of a portion of chr. 2 (N=2; 2p16.2, 2p16.3), and deletions (N=4; 7p11.1–14.1, 7p21.3, 7q11.23–21.11, 7q21.12-qter) as well as gains (N=1; 7q33) on chr 7. We have also detected segmental uniparental disomy (UPD) in chr 1 (N=2; 1p21.3–22.2, 1p). This type of lesion cannot be detected using MC and provides an additional mechanism leading to LOH. When both bone marrow and blood of 5 RARS patient were tested using SNP-A, blood analysis had 100% accuracy rate as compared to marrow; all defects seen in the marrow were also found in blood. We conclude that chromosomal defects are present in a majority of RARS patients and arrays with higher resolution will identify defects in most, if not all of the patients. Our study also demonstrates testing of peripheral blood by SNP-A can complement marrow MC, especially in cases in which marrow is not available. Detection of clonal marker aberrations in blood of RARS patients suggests that mostly clonal dysplastic progenitor cells contribute to blood production rather than residual “normal” progenitors.

Fludarabine(FLU) and 200cGy TBI is commonly used for RIC AHSCT, but we observed 12% graft rejections and a 43% incidence of disease relapse when used at our institution. We hypothesized that this might be improved with dose escalation of TBI to 400cGy. From 12/03–4/07 40 pts with hematologic malignancies received RIC AHSCT using FLU 30 mg/m2/d on days -5, -4 and -3 and then TBI 200cGy on days -1 and 0. Our analysis compared outcomes with 42 historical control pts who received 200cGy TBI from 1/00–11/03. Matched sibling donors (MSD) were used for 32(76%) pts in the 200cGy group and 26 (65%) in the 400cGy group (p=0.27); other pts had 8/8 HLA matched unrelated donors (MUD). MSD pts received cyclosporine/mycophenolate and MUD pts received tacrolimus/mycophenolate. There were no differences in diagnostic categories between the 200cGy and 400cGy groups, which included 19(45%) and 22(56%) pts with myeloid (MY) diseases, respectively, (AML most common for both) while the remaining pts with lymphoid (LY) diseases had NHL most commonly. No other baseline characteristics differed between the groups. 200cGy pts received a higher median CD34+ cell dose (6.77 vs 4.93 × 106/kg, p Figure Figure

In MDS, cytogenetics has a major prognostic influence on the phenotype of the malignant clone and specific defects may point towards potential therapeutic targets. However, traditional metaphase cytogenetics (MC) has limited resolution and does not allow for detection of uniparental disomy (UPD). Defects on chromosome 5q have been studied using various methods to identify a minimal commonly deleted region (CDR). SNP-array karyoptyping (SNP-A) allows for precise detection of copy number changes as well as UPD. We hypothesized that SNP-A may reveal new lesions on chromosome 5 and allow for better definition of CDRs and pathogenic genes. Of 512 patients, 15% showed a 5q abnormality as a sole or associated aberration by MC. DNA was available in 189 patients and was subjected to 250K SNP-A. In 7 patients with normal/non-informative MC, a deletion on 5q was clearly detectable by SNP-A; in total, SNP-A identified 5q abnormalities in 14% patients in this group (vs. 11% by MC). UPD 5q was found in one patient with CMML. By SNP-A, 6/27 patients showed an isolated 5q deletion. SNP-A can also be used to construct precise cytogenetic maps. The commonly deleted region (CDR1,5q31.2, 137,472,900–139,451,900) was present in 24/27 patients. Significant overlap occurs with the CDR previously defined by Fairman, Zhao, Horrigan et al. This region includes important genes such as Cdc25C and EGR1. Of 24 patients with a deletion in CDR1, 21 had multilineage dysplasia predominantly in the megakaryocytic line (92%). While elevated platelet counts occurred in 3 patients, increased levels of megakaryocytes were common (83%). Previous studies by Bouldwood/Jaju suggested that the minimal CDR among patients with 5q- syndrome (CDR2, 5q33.1-33.2) differs slightly from that associated with secondary AML/MDS (CDR1). However, when patients (5/27) with classical 5q- syndrome were analyzed, all displayed single deletions spanning both CDR1 and CDR2. Other areas of partial overlaps were also identified (5q12.1; 5q13.3) more centromeric to CDR1 and present in 7/27 patients. 2 cases were particularly interesting: 1 with segmental UPD involving the CDR, the other showing a small deletion defining the CDR itself. In the latter patient, marked thrombocytosis was present and SNP-A demonstrated a complex chromosomal rearrangement. While MC revealed a segmental deletion of 5q and a concomitant duplication of this abnormal homolog, SNP-A showed that while the p arm portion had been duplicated, the q arm, with the exception of two small deletions (1.35 and 1.98Mb in length, confirmed by FISH), had a normal diploid set. SKY clarified that chr. 5 material had indeed been displaced to both chr. 3 and 7 with a reciprocal translocation of chr. 3 material occurring on the abnormal chr. 5. In sum, our studies demonstrate the utility of SNP-A as a karyotyping tool that can detect previously cryptic areas of LOH on chr. 5 and facilitate definition of shared 5q defects. We also show that our patients with 5q- syndrome had lesions spanning both 5q33 and the more proximal 5q31.2 area, making pathogenic distinction based on cytogenetics difficult.

The reactivity of NK cells and some T cell populations is regulated by KIR interactions with HLA class I molecules. Such interactions have been suggested to influence outcomes after myeloablative allogeneic HSCT. However, in NMHSCT the effect of KIR interactions on outcomes including the development of CDC has not been well described. We analyzed 51 pts who received related donor NMHSCT at our institution from 1/10/00–10/25/05. All pts received fludarabine 30 mg/m2/d x 3 days followed by total body irradiation 200 cGy (n=35) or 400 cGy (n=16) for conditioning. The median age was 54 (range, 21–64). Short tandem repeat analysis for T cell (CD3+) chimerism was performed on peripheral blood and CDC was defined as achievement of >95% DNA of donor origin in the CD3+ T cells. 37 (73%) of patients achieved CDC at a median time of 3.3 mos (range, 0.4–11.2). KIR genotypes were determined for recipients by PCR-rSSOP analysis. Donor HLA KIR ligands were categorized as: HLA-Cw groups C1 (+ or −); C2 (+ or −); HLA-Bw4 (+ or −); and HLA-A3 or -A11 (+ or −) [as reviewed by Farag et al. Blood2002; 100:1935–47]. Recipient KIR genotype and donor HLA KIR ligands were used to generate an inhibitory KIR score for pts from 1 to 4 corresponding to the potential number of inhibitory KIRs engaged. 7 pts had a score of 1, 27 had a score of 2, 14 had a score of 3 and 3 had a score of 4. Figure Figure The Kaplan-Meier method was used to estimate the achievement of CDC by inhibitory KIR score (figure, p=0.09). Pts with a score of 1 were less likely to achieve CDC compared to those with a score of 2 (p=0.02), while those with a score of 2 tended to be less likely to develop CDC than those with a score of 4 (p=0.07). There were no differences in CD34+ or CD3+ cell doses between any of the groups. When combined with the inhibitory KIR score data the presence of single or multiple activating KIR’s was not found to influence the development of CDC. Thus, pts with lower inhibitory KIR scores may have more active anti-donor effector cells (NK cells and T cell subsets) that may reduce donor cell chimerism. Conversely, those with higher inhibitory KIR scores may have less active populations and be more likely to achieve CDC. Given the genotypic potential to inhibit all NK cells KIR expression may be variable among different clones, and may affect the development of CDC. Further investigation of KIR expression at the cellular level rather than by genotyping alone should be pursued.

Assessment of chimerism after AHSCT has been important to monitor donor hematopoietic engraftment and to assess for residual disease or relapse. Achievement of T cell (CD3+) complete donor chimerism (CDC), particularly after NM AHSCT, has been considered important to achieve a graft-vs.-malignancy effect. Formal comparisons of the onset and frequency of T cell CDC in pts treated with MY vs. NM conditioning regimens have not been well described. The current analysis compared rates of achieving T cell CDC for 116 pts transplanted from 1/10/00–5/19/06 who were categorized into the following 4 groups based upon type of transplant conditioning: 1) 200 cGy total body irradiation (TBI) + fludarabine 30 mg/m2/d × 3 days (NM200; n=47); 2) 400 cGy TBI + fludarabine (NM400; n=23); 3) MY AHSCT with busulfan/cyclophosphamide without TBI (MY-noTBI; n=31); 4) MY AHSCT with TBI (MY-TBI; n=15). All NM AHSCT pts received peripheral blood stem cells and all MY AHSCT patients received bone marrow as their stem cell source. The total nucleated cell (TNC) and CD34+ cell doses were higher in the NM AHSCT patients. 36 (31%) pts had matched unrelated donors, all with at least an 8/8 match (HLA-A, -B, -Cw, -DR) by HLA class I and II DNA-based typing [10 (21%) NM200, 7 (30%) NM400, 9 (29%) MY-noTBI, 10 (67%) MY-TBI; p=0.011]. Graft-vs-host disease prophylaxis consisted of mycophenolate mofetil and cyclosporine or tacrolimus for most pts. Short tandem repeat analysis for T cell (CD3+) chimerism was performed on peripheral blood post transplant and CDC was defined as achievement of >95% DNA of donor origin isolated from CD3+ T cells. Post-transplant T cell chimerism was found in both the MY and NM AHSCT groups. The number of pts who achieved CDC and the median time to its occurrence for each group was as follows: NM200 - 34 (72%) at 4 mos; NM400 - 18 (78%) at 2.7 mos; MY-noTBI - 20 (65%) at 3.3 mos; and MY-TBI - 13 (87%) at 1.3 mos. The Kaplan-Meier curves for achievement of CDC are shown above (p=0.23). The group of pts who received MY-TBI developed T cell CDC more rapidly than the NM200 pts (p=0.05). No significant differences were observed between the NM400, MY-noTBI and MY-TBI groups with regards to achieving CDC. NM conditioning with either 200 cGy or 400 cGy TBI did not show a significant difference in rate of achieving T cell CDC as compared to MY conditioning with busulfan/cyclophosphamide without TBI. This may be related in part to the higher TNC and CD34+ cell doses in the NM AHSCT pts. However, the increased TBI doses utilized for MY conditioning may more effectively suppress anti-donor immune effector cells from the recipient, which resulted in the increased CDC compared to the NM200 group. In conclusion, post-transplant monitoring for T cell CDC is important in both MY and NM AHSCT to allow for immune manipulation to maintain a state of donor-host tolerance in order to prevent graft rejection. Figure Figure

Fludarabine and TBI (200 cGy) is a common NMHSCT preparative regimen. However, this regimen resulted in a 14% graft rejection rate and 42% incidence of disease relapse at our institution. We hypothesized that escalation of the TBI dose to 400 cGy may improve post-transplant outcomes. From 12/10/03 to 4/19/05, 17 patients (pts) with hematologic malignancies underwent NMHSCT using a preparative regimen of fludarabine 30 mg/m2/d IV on days −5, −4 and −3 and TBI 200 cGy on days −1 and 0 (total dose 400 cGy). Immunosuppressant therapy consisted of cyclosporine and mycophenolate for matched sibling donor pts (n = 10) or Tacrolimus and mycophenolate for matched unrelated donor pts (MUD) (n = 7), which was started day −1 and discontinued day +56 in the absence of GVHD. Diagnoses included 4 NHL, 3 AML, 2 MDS, 2 myelofibrosis (MFB), 1 Hodgkin lymphoma (HL), 1 ALL, 1 multiple myeloma (MM), 1 CLL, 1 chronic myeloproliferative disorder (CMD) and 1 bilineal acute leukemia (BAL). Only 2 (12%) pts were in complete remission at the time of transplant (both AML). The median time from diagnosis to transplant was 11 months (range, 3–246 months). All transplants were performed as an outpatient, but 16 (94%) pts required hospitalization and the most common reason was for fever (9 pts– 56%). The median CD34+ and CD3+ cell doses infused were 4.92 x 106/kg and 4.44 x 108/kg, respectively. The median time to absolute neutrophil count recovery of ≥500/μL was 10 days (range, 8–13 days) while time to platelet recovery ≥20K/μL was 12 days (range, 11–16 days). T-cell (CD3+) chimerism was monitored by short tandem repeat analysis and complete donor chimerism (CDC) was defined as ≥ 95% donor DNA in CD3+ T-cells. CDC was achieved in 14 (82%) pts at a median of 28 days (range, 21–130 days), whereas, in our prior analysis with pts receiving 200 cGy TBI, 75% of pts achieved durable CDC at a median of 77 days (range, 14–310 days). Ten (59%) pts developed acute GVHD at a median of 61 days (range, 13–85 days) with 4 grade I, 4 grade II, 1 grade III, and 1 grade IV. Chronic GVHD developed in 5 (29%) pts at a median of 10 months (range, 3–13 months) and only 1 (6%) developed extensive chronic GVHD. Responses included 4 CR (1 NHL, 1 AML, 1 CLL, 1 MFB), 4 PR (2 NHL, 1 HL, 1 MM), 2 stable disease (1 CMD, 1 NHL), and 3 not evaluable (2 MDS, 1 MFB). Graft rejection occurred in only 1 AML patient with a MUD and HLA-B and -Cw disparities. Three (18%) pts (1 ALL, 1 AML, 1 BAL) had disease relapse at a median of 6 months post-transplant (range, 2–12 months). The Kaplan-Meier method reported a median relapse-free survival of 11.6 months. Seven pts died, 3 within 100 days of NMHSCT; estimated median survival was 12.7 months. Causes of death included 2 acute GVHD, 1 chronic GVHD, 1 relapse, 1 sepsis, 1 ARDS, 1 cardiac arrest. We conclude that escalation of the TBI dose to 400 cGy in combination with fludarabine for NMHSCT is an effective approach which when compared to the 200 cGy regimen has resutled in less graft rejection (6% vs. 14%, respectively) and a lower relapse rate (18% vs. 42%, respectively). The more rapid achievement of T-cell CDC may be responsible for these differences. Further investigation and follow-up with this regimen is warranted.

The evolution of abnormal hematopoietic clones characterized by acquired chromosomal abnormalities is the central event in the pathogenesis of MDS. Defective chromosomes have significant clinical implications in the management of MDS and suggest the presence of an inherent chromosomal instability. As karyotypic lesions are not found in all MDS patients, it is possible that in some the dysplastic clone may evolve without a chromosomal defect or, more likely, the resolution of routine metaphase cytogenetics is not sufficient to detect smaller lesions; in many instances lack of growth precludes the analysis. Array-based comparative genomic hybridization (A-CGH) allows for a high-resolution genomic scan that circumvents some of the limitations associated with the use of conventional cytogenetics. We hypothesized that high-resolution genomic analysis of genetic gains and losses by A-CGH may detect cryptic lesions, particularly in patients with negative/non-informative cytogenetics that may be of clinical/scientific significance. We examined bone marrow cells from 39 MDS patients (18 RA/RARS, 11 RAEB-t, 6 CMML and 4 secondary AML) and 11 controls using a 2632 BAC microarray and CGH. Dye swapping on duplicate arrays assured reproducibility of the CGH results, confirmed globally by a high resolution 50K SNP microarray in 4 patients and by microsatellite analysis in others. By traditional cytogenetics 19 patients had chromosomal lesions, 18 were normal and 2 tests non-informative. When A-CGH was applied, a normal karyotype was found in only 15% of patients in comparison to 46% by metaphase cytogenetics. Of note is that both cases with uninformative cytogenetics showed an abnormal CGH result and in several patients (N=11) with an abnormal karyotype additional lesions were found. Karyotypic results were confirmed in 7 cases; discordant analysis may be due to a lower proportion of dysplastic cells in marrow. Irrespective of the genomic area affected, when we studied the raw number of lesions more advanced forms of MDS (RAEB-t/AML) were evenly distributed between patients subdivided on sheer number of lesions (0, 1–17, >17). Many hotspots of genomic instability shared between patients were identified. For example, 1p26.3, 10q26 and 4p16 lesions were found in 2 or more patients. Interestingly, these regions contain genes of potential pathologic significance, including tubulin gamma complex associated protein 2 (TUBGCR2) and histone stem-loop binding protein (SLBP). Cryptic lesions on chromosome 7 (e.g. 7p21, 7q31) were identified in 5 patients with normal cytogenetics. These patients suffered from severe cytopenias, consistent with the prognosis of monosomy 7 and highlighting a consensus defect on chromosome 7. Certain chromosomes were rarely or never affected, implying that a more targeted array might be designed for clinical use. A-CGH Cytogenetics Unsuccessful Normal Abnormal Unsuccessful (N=2) 0 0 2 Normal (N=18) 0 3 15 Abnormal (n=19) 0 3 16 In summary, our study highlights the superior level of resolution of A-CGH as compared to metaphase analysis in the diagnosis of MDS. A prospective analysis is underway to determine the prognostic value of CGH-detected lesions and their pathophysiologic significance.

The WHO classification of myeloid neoplasms recognizes a category of myeloid disorders that overlaps traditional myelodysplastic syndromes (MDS) and myeloproliferative disorders (MPD) as the entity MDS/MPD disease. This disease category includes CMML, JMML, atypical CML and MDS/MPD unclassifiable (MDS/MPD-u). The provisional entity termed RARS associated with marked thrombocytosis (RARS-t) is currently best classified as MDS/MPD-u until further information is available regarding its pathogenesis. Recently a JAK2 mutation V617F (G→T) was identified as a pathogenetic lesion in typical myeloproliferative disorders (Kralovics et al, NEJM, 2005). Subsequent studies of the JAK2 mutation concentrated on more proliferative forms of MDS or MDS/MPD such as CMML and CNL (Steensma et al, Blood, 2005). These studies demonstrated that homo- and heterozygous JAK2 mutants are present in a rather small proportion of these patients. Based on similar clinical features, we theorized that JAK2 mutants might be also found in patients with MDS/MPD-u. We have collected a cohort of these patients (N=202) and analyzed them for the presence of JAK2 mutation using a molecular allele-specific PCR assay. Positive cases were confirmed by sequencing with a sensitivity of about 20% of mutated cells. A group of patients with PV, MF and ET served as positive controls (N=66). In agreement with previous reports, the detection rate of JAK2 mutants for PV, MF and ET was 92%, 55% and 55%, respectively. Our experimental group included 104 patients with MDS or MDS/MPD (53 RA/RS, 14 RAEB, 22 RAEB-t/sAML, and 15 CMML). Most significantly, the RA/RS group contained 13 patients with WHO-defined MDS/MPD overlap and 3 with RARS-t (among these patients, 3 were JAK2 mutants; all of them were heterozygous - 18.8%). Within 16 patients with CMML1 or 2 we found only 2 heterozygous for JAK2 mutation (12.5%). The remaining cohort of 73 patients in MDS categories revealed only 1 patient to be heterozygous for JAK2 mutation (RARS, 1.4%). As expected MDS/MPD patients with JAK2 mutation showed various degrees of BM fibrosis, splenomegaly and less pronounced cytopenias. Except for one patient, JAK2 mutants had normal cytogenetics. All had normal MCV and ANC. In 4/5 increased megakaryocytes with/or without atypia was seen. One CMML patient with abnormal cytogenetics showed an unusual translocation t(8;9)(q22;p24). JAK2 is located at 9p24 so it is possible that the JAK2 gene was involved in the translocation generating a novel fusion protein in addition to an activating JAK2 mutant. Our results showed that JAK2 mutations are rarely found in typical cases of MDS or CMML. However, further analysis of JAK2 mutational status in patients with MDS/MPD-u is warranted. The reported detection rate may suggest that the pathogenesis of these entities is more akin to myeloproliferative than myelodysplastic syndromes.

Pre and post transplant dendritic cell (DC) levels may correlate with survival, GVHD, and relapse in allogeneic BMT. Whether they have prognostic importance in autologous peripheral blood progenitor cell (PBPC) is unknown. We prospectively collected pre and post dendritic cell levels, including DC1 and DC2 levels in 53 patients with diffuse large B cell lymphoma (DLBCL) to investigate their potential prognostic importance. Pre-transplant analysis of DC1 and DC2 by flow cytometry was obtained 24 hours prior to the initiation of VP-16 (2 gm/m2) or G-CSF (10 mcg/kg/d) (n=4) for PBPC mobilization; post transplant samples were obtained 6 weeks after transplant. 51 of 53 (96%) were sensitive to chemotherapy at the time of transplant; 34% received prior radiation therapy; 15% had an elevated LDH at transplant. 49 of 53 received VP-16 + G-CSF for stem cell mobilization and 4 patients received G-CSF alone. 98% received a preparative regimen of Bu/Cy/VP and one patient received Busulfan alone. In a univariate analysis, higher numbers of pre-transplant DC1 cells and pre-transplant total DC (DC1 + DC2) cells correlated with improved survival, as shown graphically below: Figure Figure Patients received a median CD34+ cell dose of 8.7 x 106/kg. DC values did not correlate with CD34+ cell yield. Post-transplant DC1 (p=0.03), DC2 (p=0.035), and post-transplant total DCs (p=0.04) also correlated with improved survival. In all cases, more DCs was associated with a better outcome. In a multivariable model, pre-transplant total DC number (p=0.028), and a normal LDH at the time of transplant (p=0.017) correlated with improved outcome. Median follow up of surviving patients is 15 months. Of 26 patients with pre-transplant total DC numbers greater than 9.10 x 106/mL, 77% of patients are alive, as compared to 56% of patients with pre-transplant total DC numbers less than 9.10 x 106/mL (n= 27)(p=0.022). The high DC group had fewer relapses and fewer late infections. We conclude that higher pre-transplant total DC numbers are associated with improved outcome in ABMT for DLBCL.

We hypothesized that after allogeneic hematopoietic stem cell transplant (HSCT), GvHD-affected tissues harbor expanded “immunodominant” T-cell clones, and characterization of these clones can be used to develop markers of disease. A multiplex PCR was used to detect T-cell receptor variable beta (VB) chain rearrangements in target tissue. Molecular analysis of the amplified VB CDR3 sequences allowed for identification and quantitation of putative disease-associated “clonotypes” and for the development of clone-specific PCR. We studied 5 HSCT patients for the presence of signature clonotypes in 10 skin biopsies taken during diagnostic GvHD work-up. Size distribution analysis of VB PCR products showed a skewed peak pattern in 9 biopsies; immunodominant clones (per definition frequency ≥30%) were detected in 6/7 biopsies with histologically confirmed GvHD, consistent with the presence of expanded clonotypes and the oligoclonal nature of the tissue-specific alloresponse. Immunodominant clones were also found in 2 of 3 biopsies not diagnostic for GvHD but obtained based on strong clinical suspicion, raising the possibility that they were associated with early evolving GvHD not distinguishable by histology. For example, when serial skin biopsies were analyzed, a GvHD-positive post-transplant d63 biopsy contained an immunodominant clone (frequency 60%), which was also detected in a subsequent biopsy positive for GvHD (frequency 33%). Similar results were seen in another patient, in whom serial biopsies taken on d214 (not diagnostic) and d217 (GvHD-positive) showed an identical immunodominant clone, not present in a d13 GVHD-negative biopsy. This finding suggests that the d214 biopsy might have contained early GvHD that was not detectable morphologically. In a patient who rejected an initial MUD graft (Tx 1) and then received a MUD SCT (Tx 2) from a different, unrelated donor, immunodominant clones were identified in GvHD-positive biopsies following each transplant, that were distinct for each graft. To examine whether immunodominant clonotypes derived from biopsies could be used as markers of disease, clonotypic PCR was developed for an immunodominant biopsy-derived clonotype for each transplant. The Tx 1 clonotype was detected in blood and skin following Tx 1, but not in tissue or blood taken after Tx 2. Specificity and correct size of the clonotypic PCR product were confirmed by both Genescan analysis and sequencing. Clonotypic PCR designed for an immunodominant clonotype from the Tx 2 donor detected the putative allospecific clonotype in serial samples after the second engraftment. Neither clonotype could be found in either donor, indicating that the disease-associated clones expanded to detectable levels following transplant. These results indicate that clonotypic PCR can distinguish distinct GvHD-associated clonotypes from different donors in both blood and tissue following transplant. Monitoring of the relative frequency of disease-associated clones in recipient blood indicated a significant peripheral expansion of disease-associated clones at the time of active GvHD. Our results demonstrate an efficient method for identification of disease-associated clonotypic markers, which can be used to aid diagnosis and monitoring of GvHD.