Identification of Campath-1 (CD52) as Novel Drug Target in Neoplastic Stem Cells in 5q-Patients with MDS and AML

Myelodysplastic syndromes (MDS) and acute myelogenous leukemia (AML) are stem cell-derived, myeloid neoplasms (1–3). Both conditions share pathogenetic and clinical features; and many patients with MDS transform to overt AML during disease evolution. MDS and AML are classified according to blast cell counts as well as cytogenetic and molecular features (3–6). The prognosis in AML varies, depending on age, burden of neoplastic (stem) cells, and certain cytogenetic and molecular lesions. In patients with MDS, prognostically relevant cytogenetic subgroups of patients have also been identified (7–9). Although in low-risk MDS, an “isolated del(5q)” is indicative of a good prognosis, the prognosis of del(5q) patients with advanced MDS or AML is poor, especially when the clone exhibits additional cytogenetic defects (7–11). A complex karyotype is almost always a bad prognostic sign, independent of age and the category of MDS or AML.

Clonal cells in MDS and AML are organized hierarchically similar to normal hematopoiesis (2, 12–16). In this hierarchy, only the most primitive progenitors, also termed neoplastic stem cells (NSC), or leukemic stem cells (LSC) in AML, have the capacity of long-term self-renewal, and thus are responsible for unlimited proliferation, clonal evolution, and relapse after therapy (12–16). In both MDS and AML, NSC and LSC are considered to reside in the CD34⁺ compartment of the clone (12–16). Depending on the disease variant, CD34⁺/Lin⁻ cells or CD34⁺/CD38⁻ cells may exhibit long-term repopulating capacity in NOD/SCID or NOD/SCID IL-2Rgamma^null (NSG) mice and thus stem cell function. However, at least in AML, NSG-repopulating NSC may also reside in the CD34⁺/CD38⁺ fraction or even in a CD34-negtive subset of the clone (17, 18).

During the past few years, substantial efforts have been made to characterize target expression profiles in NSC/LSC in MDS and AML (15, 16, 19–24). Among these targets are several surface molecules that are recognized by targeted antibodies or antibody-toxin-conjugates. Likewise, it has been described that NSC/LSC-enriched CD34⁺/CD38⁻ cells in AML frequently coexpress CD33, CD44, and CD123 (19, 20, 23, 24).

The CD52 antigen, also known as Campath-1, is expressed broadly in lymphatic cells as well as on blood monocytes (25, 26). On the basis of its expression on lymphoid progenitors and B cells, CD52 has been developed as a drug target in advanced chronic lymphocytic leukemia (CLL) (25, 27–29). Indeed, the CD52-targeting drug alemtuzumab, is capable of eliminating CD52⁺ lymphocytes in most patients with CLL (28, 29).

More recently, alemtuzumab has also been administered in patients with MDS (30, 31). The primary rational of this approach has been the observation that T-cell–targeting immunosuppressive agents, like anti-thymocyte globulin, are effective in a subgroup of patients, especially those with hypoplastic MDS (32, 33). Alemtuzumab was also found to induce remarkable responses and even remission in a few patients with MDS (30, 31). Whether these effects of alemtuzumab resulted from its immunosuppressive activity remains unknown. We explored an alternative mode of action of alemtuzumab in MDS, namely a direct effect on NSC. The results of our study show that CD52 is expressed on NSC/LSC-enriched CD34⁺/CD38⁻ cells in patients with MDS and AML. In addition, our data show that alemtuzumab can attack CD52⁺ NSC/LSC.

Sixty-two patients with AML (35 females, 27 males; median age: 62 years) and 29 with MDS (14 females, 15 males; median age: 70 years) were examined. Diagnoses were established according to French-American-British (FAB) and World Health Organization (WHO) criteria (4–6). The patients' characteristics are shown in Supplementary Tables S1 (MDS) and S2 (AML). Bone marrow (BM) aspirates were obtained from the iliac crest. Control bone marrow cells were obtained from 9 patients with chronic myelomonocytic leukemia (CMML), 7 with chronic myelogenous leukemia (CML), 8 with myeloproliferative neoplasms (MPN), 5 with unclassifiable MDS/MPN (MDS/MPN-U), 10 with acute lymphoblastic leukemia (ALL), 19 with cytopenia of undetermined significance (ICUS), and 25 with normal bone marrow (staging for lymphoma, n = 19, Ewing sarcoma, n = 1; remission marrow, n = 5; Supplementary Tables S3 and S4). Bone marrow mononuclear cells (MNC) were isolated using Ficoll. Karyotyping and molecular analyses were performed according to standard techniques (34, 35). All donors gave written informed consent. The study was approved by the ethics committee of the Medical University of Vienna (Vienna, Austria).

RPMI-1640 medium and fetal calf serum (FCS) were purchased from PAA laboratories, FITC-labeled CD34 monoclonal antibody (mAb) 581, PerCP-labeled CD45 mAb 2D1, APC-labeled CD38 mAb HIT2, phycoerythrin (PE)-labeled anti-CLL1 mAb 50C1 from BD Biosciences, PE-labeled CD52 mAb HI186, PE-labeled CD123 mAb 32703 from R&D Systems, FITC-labeled CD14 mAb TÜK4 from Dako, APC-labeled CD123 mAb AC145 from Miltenyi Biotech, and PerCP/Cy5.5-labeled CD90 mAb 5E10 and Pacific Blue-labeled CD45RA mAb HI100 from Biolegend. Alemtuzumab was purchased from Genzyme and IgG from Abcam. For measuring drug effects on NSC/LSC, the CountBright absolute counting beads (Invitrogen), the PE-labeled CD34 mAb 581 (BD Biosciences), and the APC/Cy7-labeled CD45 mAb HI30 (Biolegend) were used.

The CD52⁻ AML cell line HL60 was maintained in RPMI-1640 medium with 10% FCS (37°C), and the CD52-positive ALL cell line Raji in RPMI-1640 plus 20% FCS. HL60 and Raji cells were obtained from the DSMZ Institute (Braunschweig, Germany). The biologic stability of these cell lines was checked by cell surface phenotyping (flow cytometry) and their identity was confirmed by reauthentication in the DSMZ Institute. HL60 cells were engineered to express mutated RAS (NRAS G12D; NRAS Q61K) by lentiviral transduction essentially as described (36). In brief, the coding sequences of RAS mutants were cloned into lentiviral pWPI vector kindly provided by Dr. D. Trono (University of Geneva, Switzerland). The pWPI vector contains an internal ribosome entry site and a GFP coding region. Recombinant lentiviruses were produced as described previously (37). After one week, transduced cells were examined for expression of CD52 by flow cytometry.

A number of mAb were applied to characterize NSC/LSC, monocytes, and blood basophils. NSC/LSC-rich cells (called NSC/LSC below throughout this manuscript) were defined as CD34⁺/CD45⁺/CD38⁻ cells, and progenitor-enriched cells as CD34⁺/CD45⁺/CD38⁺ cells (20, 24, 38). Monocytes were identified as CD14⁺ cells and basophils by their characteristic side-scatter properties and expression of CD123 (39, 40). In patients with CD34-negative AML, CD34⁻ blasts were identified by their characteristic side-scatter properties. In a subset of patients with MDS (n = 6) and AML (n = 6), we examined CD34⁺/CD38⁻/CD90⁺/CD45RA⁻ cells or CD34⁺/CD38⁻/CD90⁻/CD45RA⁻ cells, respectively. Heparinized bone marrow cells (30–100 μL) were incubated with combinations of mAb for 15 minutes. After erythrocyte lysis with BD lysing solution (BD Biosciences), expression of cell surface antigens was examined by multicolor flow cytometry on a FACSCalibur or on a FACSCantoII (BD Biosciences). Antibody reactivity was controlled by isotype-matched antibodies. The staining index (SI) was calculated from median fluorescence intensities (MFI) obtained with the CD52 antibody and an isotype-matched control antibody (SI = MFI_CD52:MFI_control).

In 11 patients with AML and 6 with del(5q) MDS, the CD34⁺/CD38⁻ NSC/LSC and CD34⁺/CD38⁺ progenitors were purified from bone marrow MNC by cell sorting on a FACSAria (BD Biosciences) using a PE-labeled CD34 mAb and an APC-conjugated CD38 mAb as described (24, 38). After sorting, the purity of CD34⁺/CD38⁻ stem cells and CD34⁺/CD38⁺ was >95% in each case, and cell viability was >80% in all samples.

In 5 patients with MDS and del(5q), sorted CD34⁺/CD38⁻ cells, sorted CD34⁺/CD38⁺ cells, and total MNC were examined by FISH using a dual color probe-set (Cytocell) with a red probe for EGR1 (chromosomal band 5q31.1) and a green (control) probe for TAS2R1 (chromosomal band 5p15.31). The presence of trisomy 8 was confirmed in xenotransplanted cells by FISH using a probe for centromere 8, obtained from Kreatech. FISH was performed according to the manufacturer's instructions.

Total RNA was isolated from MNC of 11 patients with AML (FAB M1, n = 5; M2, n = 1; M4, n = 4; secondary AML, n = 1) and 6 with MDS and del(5q), using RNeasy Micro-Cleanup Kit (Qiagen). cDNA was synthesized using Moloney murine leukemia virus reverse transcriptase (Invitrogen), random primers, first-strand buffer, dNTPs (100 mmol/L), and RNasin (all from Invitrogen) according to the manufacturer's instructions. qPCR was performed as described (24) using iTaq SYBR-Green-Supermix with ROX (Bio-Rad) and primers specific for CD52, EVI1, and CD300a (Supplementary Table S5). mRNA levels were expressed as percentage of ABL transcript levels.

Bone marrow MNC (AML, n = 6; MDS, n = 6; control samples, n = 6) were incubated in various concentrations of alemtuzumab (10–300 μg/mL) in RPMI-1640 medium plus 30% complement-containing human serum at 37°C for 1 hour. Then, 10 μL calibration beads were added. After washing, cells were stained with fluorochrome-conjugated mAb against CD34, CD45, and CD38 for 15 minutes. Cells were then subjected to 4′, 6—diamidino—2—phenylindole (DAPI) staining to count viable cells on a FACSCanto-II (BD Biosciences). HL60 cells, Raji cells, and HL60 cells transfected with empty vector or NRAS Q61K, were incubated with various concentrations of alemtuzumab (0.1–500 μg/mL) in RPMI-1640 medium with 30% serum at 37°C for 1 hour. After incubation, cells were examined for viability by propidium iodide staining. In experiments performed with transfected HL60 cells, 10 μL calibration beads were added and cells were examined for viability by propidium iodide staining. In control experiments, cells were incubated with alemtuzumab in medium containing either IgG (20 μg/mL) or 30% human serum.

Primary AML cells (n = 3 patients) were incubated in control medium or in alemtuzumab (500 μg/mL) with 30% human serum at 37°C for 1 hour. After incubation, AML cells were viable without signs of cell death or apoptosis (alemtuzumab: 75%–80% cells viable; vs. control medium: 80%–90% of cells viable by Trypan blue staining). Drug-exposed cells and control cells were washed, resuspended in 0.15 mL PBS with 2% FCS, and injected into the tail vein of adult female NSG mice (2–5 × 10⁶ per mouse, 4–5 mice per group; The Jackson Laboratory). Twenty-four hours before injection, mice were irradiated (2.4 Gy). After injection, mice were inspected daily and sacrificed after 10 weeks. Bone marrow cells were obtained from flushed femurs, tibias, and humeri. Human AML cells were detected in bone marrow samples by multicolor flow cytometry using mAb against CD19, CD33, and CD45. AML repopulation was measured by determining the percentage of CD45⁺ cells in mouse bone marrow samples by flow cytometry. Animal studies were approved by the ethics committee of the Medical University of Vienna and the University of Veterinary Medicine Vienna (Vienna, Austria), and carried out in accordance with guidelines for animal care and protection. Animal experiment license was granted under no. GZ 66.009/0040-II/10b/2009.

To analyze the significance of differences in expression of CD52 on NSC/LSC in various subgroups of patients with MDS and AML, the Mann–Whitney U test was applied. To determine correlations between surface and mRNA expression levels of CD52 in NSC/LSC, between EVI1 mRNA and CD52 mRNA expression, between CD300a mRNA and CD52 mRNA expression, and between CD52 expression and cytotoxic effects of alemtuzumab on NSC/LSC, a linear regression model was applied. The probability of overall survival (OS) in our patients with MDS (n = 29) and AML (n = 62), and the AML-free survival in our patients with MDS were calculated by the product limit method of Kaplan and Meier. The median follow-up in our patients with MDS and AML was 422 days and 382 days, respectively. Moreover the number of patients at risk was calculated at 0 to 84 months. Statistical significances in differences among patients with or without CD52⁺ stem cells concerning survival and AML-free survival (patients with MDS) were determined by log-rank test. For determining the level of significance in drug inhibition experiments, the Student's t test was applied. Differences were considered significant when P < 0.05.

As assessed by flow cytometry, CD52 was found to be expressed at high levels on CD34⁺/CD38⁻ cells in 8/11 patients with del(5q) MDS, including 6 with isolated del (5q) (Fig. 1A; Table 1). In most other MDS patients examined, CD52 was weakly expressed or not expressed on CD34⁺/CD38⁻ NSC (Fig. 1A; Table 1). We also confirmed that CD52 is expressed on the CD34⁺/CD38⁻/CD90⁺/CD45RA⁻ subset of NSC in our patients with MDS. These CD34⁺/CD38⁻/CD90⁺/CD45RA⁻ cells also coexpressed CD123 but did not express CLL-1 (Supplementary Table S6). For control purpose, we also examined expression of CD52 on monocytes and basophils in our patients with MDS. As expected, CD52 was detected on both cell types, without substantial differences in expression levels among the subgroups of patients examined (Table 1).

In 23/62 patients with AML, LSC were found to express CD52. Of these patients with AML, 4 had a complex karyotype including del(5q), one del(5q), 2 dysplasia and inv(3), 2 t(8;21), one isolated inv(16), one 13q- anomaly, three trisomy 8, one monosomy 7, and 8 patients with AML had a normal karyotype. The highest levels of CD52 on NSC/LSC were recorded in one patient with del(5q), one with monosomy 7, one with inv(16), and one with a normal karyotype (Fig. 1A; Table 1). In 5/7 patients with AML exhibiting del(5q) (71.4%), the CD34⁺/CD38⁻ LSC expressed high levels of CD52, whereas in only 18/55 patients with AML without del(5q) (32.7%), LSC expressed high levels of CD52. This difference was found to be statistically significant (P < 0.05). In a majority of patients in whom CD34⁺/CD38⁻ cells displayed CD52, the CD34⁺/CD38⁺ cells also expressed CD52 homogenously without a negative subfraction (MDS: 9/15 = 60%; AML, 14/23 = 60.9%). Overall, CD52 was found to be expressed at higher levels on CD34⁺/CD38⁻ cells and CD34⁺/CD38⁺ cells in patients with MDS or AML compared with control bone marrow samples (Fig. 1B and Supplementary Fig. S1). In patients with AML, CD52 was also found to be expressed on monocytes and basophils in most donors, although in several patients, expression levels were low or even undetectable, thereby contrasting MDS (Table 1). Recent data suggest that in patients with NPM1-mutated AML and other AML types, LSC may (also) reside in a CD34-negative fraction of the clone (18). Therefore, we were interested to learn whether CD52 is expressed on CD34-negative blasts in our patients with AML. We were able to detect CD34-negative blast cells in 41/62 patients with AML. In 11 of these 41 patients, including 3 carrying a NPM1 mutation, the CD34-negative blasts expressed substantial amounts of CD52 (Table 1). We also found that CD52 is expressed on the CD34⁺/CD38⁻/CD90⁻/CD45RA⁻ subset of LSC in our patients with AML. These CD34⁺/CD38⁻/CD90⁺/CD45RA⁻ cells also expressed CD123, and in a subset of patients (2/6), these cells expressed CLL-1 (Supplementary Table S6).

In a next step, we examined the expression of CD52 on NSC in various control cohorts, CML, CMML, and other MDS/MPN overlap syndromes. In 6/6 patients with chronic phase CML, CD34⁺/CD38⁻ cells expressed CD52, whereas in one patient with accelerated phase CML, CD34⁺/CD38⁻ cells did not express CD52 (Supplementary Table S7). In 4/9 patients with CMML, CD34⁺/CD38⁻ cells expressed substantial amounts of CD52. In the other patients with CMML, CD34⁺/CD38⁻ NSC/LSC displayed low levels of CD52 (Supplementary Table S7). We also examined 7 patients with JAK2 V617F⁺ MPN and one with JAK2 V617F⁻ MPN. In two of these patients, CD34⁺/CD38⁻ cells stained positive for CD52 (Supplementary Table S7). Finally, we examined NSC in 2 patients with acute leukemia with mixed (lymphoid/myeloid) phenotype (acute undifferentiated leukemia; AUL). In one patient with AUL, a complex karyotype with del(5q) was detected. In this patient, CD34⁺/CD38⁻ NSC/LSC expressed CD52 (Supplementary Table S7). In the other patient with AUL, who had a normal karyotype, CD34⁺/CD38⁻ NSC/LSC did not express CD52 (Supplementary Table S7). Finally, we were able to show that in most patients with ALL (8/10 = 80%), the CD34⁺/CD38⁻ stem cells express CD52 (Supplementary Table S7).

In 2/19 patients with ICUS, CD34⁺/CD38⁻ stem cells clearly expressed CD52, whereas in the other ICUS patients tested, stem cells did not express substantial amounts of CD52 (Supplementary Table S8). In control bone marrow samples obtained from controls with normal blood counts (n = 5), CD34⁺/CD38⁻ stem cells did not express any detectable CD52. However, in 6 of 19 lymphoma patients without known bone marrow infiltration, CD34⁺/CD38⁻ cells were found to stain positive for CD52. In 5 of these patients, stem cells expressed low levels of CD52, and in 8 patients, stem cells stained negative for CD52 (Supplementary Table S8). We also examined the expression of CD52 on monocytes and basophils in our control donors. As shown in Supplementary Table S8, CD52 was detected on both cell types in all control cases tested.

Highly enriched CD34⁺/CD38⁻ NSC of patients with del(5q) MDS or AML were found to express CD52 mRNA. As visible in Fig. 2A, there was a good correlation between surface expression of CD52 and CD52 mRNA expression levels detected in NSC/LSC. Recent data suggest that CD52 expression is associated with EVI1 and CD300a expression in myeloid leukemias (41). Therefore, we were interested to learn whether EVI1 and CD300a transcripts are detectable in NSC/LSC. Indeed, in all samples tested, CD34⁺/CD38⁻ NSC/LSC expressed EVI1 and CD300a mRNA (Fig. 2B). Although we found a correlation between EVI1- and CD52 mRNA expression in CD34⁺/CD38⁻ cells (R = 0.66; P < 0.05), no correlation between CD300a and CD52 mRNA levels was found (P > 0.05; Fig. 2B). The clonal nature of the sorted NSC populations was confirmed by FISH (Fig. 2C and Supplementary Table S9). In particular, in all 5 patients with del(5q) MDS examined, FISH analysis revealed that nearly 100% of the CD34⁺/CD38⁻ and CD34⁺/CD38⁺ cell populations tested expressed this chromosome abnormality (Supplementary Table S9). Similarly, in one patient with AML with monosomy 7, FISH confirmed that 100% of the CD34⁺/CD38⁻ and 100% of the CD34⁺/CD38⁺ expressed the −7 anomaly (not shown).

Because RAS activation has been associated with EVI1 expression, we asked whether activated RAS may play a role in CD52 expression in AML cells. As visible in Fig. 2D, two different oncogenic NRAS mutants tested induced the expression of CD52 in HL60 cells. Incubation of these CD52⁺ HL60 cells with alemtuzumab induced rapid cell lysis and a dose-dependent decrease in cell numbers, whereas no drug effect was seen in empty vector-transduced control cells (Supplementary Fig. S2A). These data suggest that CD52 expression on AML cells can be triggered by RAS activation.

To demonstrate functional significance of expression of the target receptor CD52 on NSC/LSC, we performed experiments using alemtuzumab and various MDS and AML samples. Samples were selected on the basis of high-level expression of CD52 (MFI ratio >3.01) or lack of CD52 (negative controls, MFI ratio <1.5) in these experiments. As visible in the top panels of Fig. 3A, short-term exposure to alemtuzumab (10–300 μg/mL; 1 hour) resulted in a significant decrease in the numbers of CD52⁺ NSC/LSC in MDS and AML compared with medium control. When human serum was replaced by IgG in these experiments, no effect of alemtuzumab was seen, suggesting a complement-dependent reaction (Supplementary Fig. S2B). In our patients with MDS and AML in whom only CD52⁻ NSC/LSC was detected, no significant effects of alemtuzumab were seen (Fig. 3A, bottom). In control bone marrow samples containing CD52⁺ NSC, alemtuzumab showed a slight effect on stem cell numbers, whereas in control bone marrow samples containing only CD52⁻ stem cells, alemtuzumab showed no effects (Fig. 3B). Overall, we found a significant correlation between expression of CD52 on stem cells and the cytotoxic effects of alemtuzumab on these cells (R = 0.66, P < 0.05; Supplementary Fig. S2C). We also examined the effects of alemtuzumab on HL60 cells and Raji cells (control experiments). As expected, alemtuzumab (in 30% serum) induced a rapid and dose-dependent decrease in the numbers of CD52⁺ Raji cells but showed no effects on CD52⁻ HL60 cells (Fig. 3C). In all experiments performed, alemtuzumab induced rapid cell lysis rather than apoptosis in CD52⁺ Raji cells (not shown). No effects of alemtuzumab on CD52⁺ cells were seen in the presence of IgG or heat-inactivated serum (Fig. 3C).

To confirm that alemtuzumab acts on primitive disease-initiating cells (NSC/LSC), we determined the effects of the drug on NSG engraftment of AML cells in a xenotransplantation model. As shown in Fig. 4, preincubation of AML cells with alemtuzumab (500 μg/mL in 30% human serum for 1 hour) resulted in a decreased engraftment of AML cells in vivo in NSG mice in all three samples examined. The engraftment was reduced to similar levels by alemtuzumab in the three donors, namely to 66% in patient #73; to 44% in patient #79; and to 47% in patient #49 compared with control. In two of the three samples used in these xenotransplantation experiments (#73, #79), a NPM1 mutation was found. In these two patients, the CD34-negative blasts stained positive for CD52, and in vivo engraftment of AML cells was blocked by alemtuzumab. qPCR confirmed the presence of the NPM1 mutation and the FLT3-ITD mutation in xenotransplanted leukemic cells (#73, #79) grown in NSG mice; and in the sample of patient #49, FISH analysis confirmed the presence of trisomy 8 (not shown).

Recent data suggest that the numbers and phenotype of CD34⁺/CD38⁻ stem cells in MDS and AML are of prognostic significance (42–45). In the present study, we asked whether expression of CD52 on CD34⁺/CD38⁻ stem cells in MDS and AML would be of prognostic significance. In both groups of patients (MDS n = 29 and AML n = 62), expression of CD52 was found to correlate with survival (Fig. 5A and B). In AML, the impact of CD52 expression on survival was found to be significant (P < 0.05), whereas in patients with MDS, the difference was not significant, which may be explained by the relatively small numbers of patients. We also attempted to correlate expression of CD52 on NSC/LSC with the WHO type of the disease. As visible in Supplementary Fig. S3, expression of CD52 showed a good correlation with the WHO classification in AML and MDS. In MDS, an obvious correlation was found between CD52 expression and del(5q). In contrast, no correlation between CD52 expression and Revised International Prognostic Scoring System (IPSS-R) subgroups was found (not shown). In AML, LSC consistently expressed CD52 in several subgroups, including AML with inv16, whereas in other groups, such as acute monocytic leukemia, LSC were consistently CD52 negative. No correlation was found between expression of CD52 and risk groups defined by Southwest Oncology Group (SWOG) criteria (not shown).

The target antigen CD52 (Campath-1) is expressed on B lymphocytes, monocytes, and basophils. During the past 10 years, the CD52-targeting drug alemtuzumab has been used successfully in patients with advanced CLL (25–29). More recently, alemtuzumab was also found to induce hematologic responses in patients with low-risk MDS (31). Initially this effect was considered to be mediated by the immunosuppressive activity of alemtuzumab. We here propose an alternative mode of drug action and show that the target antigen CD52 is expressed on immature stem cells in a group of patients with MDS and AML. In addition, we show that alemtuzumab induces rapid cell lysis in CD52⁺ NSC/LSC in these patients. These data suggest that CD52 is a potential therapeutic target in MDS and AML, and that treatment effects of alemtuzumab in these patients may, in part, be explained by targeting disease-initiating cells via Campath-1 (CD52).

Among patients with low-risk MDS reported to respond clinically to alemtuzumab in a previous study, several patients were found to exhibit the del(5q) anomaly (31). This is of particular interest, because in our study, CD34⁺/CD38⁻ NSC in patients with del(5q) MDS expressed high levels of CD52, thereby contrasting other patients with MDS or controls. In addition, we found that CD34⁺/CD38⁻ stem cells in patients with AML exhibiting del(5q) express detectable (mostly high) levels of CD52. Together, marked expression of CD52 is usually seen in MDS and AML with del(5q), but is not a specific marker for these disease categories. In fact, CD52 was also found to be expressed on NSC/LSC in MDS and AML cases without del(5q). In this regard it is noteworthy, that in patients with other myeloid neoplasms, such as CML or CMML, and also in patients with ALL, the CD34⁺/CD38⁻ (putative) NSC/LSC also expressed CD52. Even in a subset of patients with lymphomas without visible bone marrow involvement, CD34⁺/CD38⁻ cells expressed CD52. In contrast, in normal control bone marrow samples and almost all patients with ICUS, CD34⁺/CD38⁻ stem cells did not express CD52. Overall, the levels of CD52 on CD34⁺/CD38⁻ NSC in patients with del(5q) MDS are significantly higher than that found on stem cells in normal bone marrow.

A number of previous studies have shown that LSC in AML can also reside within a CD34⁺/CD38⁺ fraction of the clone (17). We therefore extended our studies to these cells and were able to show that CD34⁺/CD38⁺ cells in patients with AML or MDS exhibiting del(5q) express higher levels of CD52 compared with normal CD34⁺/CD38⁺ cells in control bone marrow samples. Moreover, we found that the CD34⁺/CD38⁺ cells stained positive for CD52 in all patients with AML in whom CD34⁺/CD38⁻ cells expressed CD52. In patients with NPM1-mutated AML and other AML types, LSC may reside within a CD34-negative subfraction of the clone (18). Therefore, we were also interested to learn whether CD52 is expressed on CD34-negative blast cells in our patients with AML. In these experiments, a CD34-negative blast cell population was detectable in 41 of our 62 patients with AML, and in 11 of these 41 patients, blast cells expressed CD52.

In a next step, we confirmed expression of CD52 in AML and MDS stem cells by qPCR. In these experiments, we were able to show that highly enriched (sorted) CD34⁺/CD38⁻ NSC/LSC in patients with del(5q) MDS and del(5q) AML, express detectable CD52 transcripts, and that CD52 surface expression correlates with CD52 mRNA levels. We also confirmed that the CD34⁺/CD38⁻ cells, sorted from bone marrow samples of our patients with MDS and AML exhibiting del(5q), are clonal cells. In particular, in each case examined, the CD34⁺/CD38⁻ fraction of neoplastic cells expressed the del(5q) anomaly by FISH.

So far, little is known about the regulation of expression of CD52 on NSC/LSC in MDS and AML. Recent data suggest that expression of CD52 in AML cells is associated with EVI1 and CD300a expression in myeloid leukemias (41). Although the exact mechanism remains unknown, this observation suggests that EVI1 may be involved in the expression of CD52 in AML cells. In the present study, we were able to show that CD34⁺/CD38⁻ stem cells in MDS and AML with del(5q) express detectable levels of EVI1 and CD300a mRNA, and that EVI1 transcript expression correlates with expression of CD52 mRNA. Because EVI1 expression has been described in the context of RAS activation, we also examined the effects of mutated NRAS variants on expression of CD52 (46, 47). In these experiments, we found that mutant NRAS induces the expression of CD52 in HL60 cells. These data suggest that RAS-dependent signaling may be involved in abnormal expression of CD52 in AML (stem) cells.

Recent data suggest that the numbers and phenotype of CD34⁺/CD38⁻ stem cells in MDS and AML are of prognostic significance (42–45). A clinically important question in this study was whether expression of CD52 on CD34⁺/CD38⁻ bone marrow stem cells in MDS and AML is associated with prognosis. The results of our study show that expression of CD52 on CD34⁺/CD38⁻ cells in MDS and AML is indicative of poor survival. We also asked whether CD52 expression on CD34⁺/CD38⁻ stem cells in MDS or AML correlates with other prognostic (clinical or laboratory) parameters. However, we were unable to substantiate any correlations between CD52 expression on CD34⁺/CD38⁻ cells and other (potentially prognostic) parameters analyzed such as the number of cytopenias, blast cells, or the IPSS category, which is best explained by the low numbers of patients in each group. Nevertheless, our observations suggest that expression of CD52 on NSC/LSC may be a prognostic feature in MDS and AML. However, prospective studies with more patients are required to confirm that CD52 is an independent risk factor concerning survival in patients with MDS and AML.

On the basis of the intriguing effects of alemtuzumab in patients with MDS and AML (31, 48) and our flow cytometry data, we were interested to learn whether alemtuzumab would attack CD34⁺/CD38⁻ and CD34⁺/CD38⁺ cells in MDS and AML. In these experiments, we found that alemtuzumab induces cell lysis in CD34⁺/CD38⁻ cells in patients with del(5q) MDS. Moreover, alemtuzumab induced rapid cell lysis in CD34⁺/CD38⁻ and CD34⁺/CD38⁺ cells in all patients with AML in whom these cells expressed CD52, whereas no drug effect was seen with CD52⁻ NSC/LSC. Finally, we were able to show that incubation of CD52⁺ NSC/LSC with alemtuzumab reduces their leukemic engraftment in NSG mice. Whether these effects are responsible for the remarkable clinical effects of alemtuzumab seen in patients with MDS, remains unknown. However, it is at least tempting to speculate that some of the effects of the drug are exerted through NSC/LSC elimination in these patients. Clinical trials correlating CD52 expression on NSC/LSC with responses to alemtuzumab are now warranted to clarify this question.

Treatment with alemtuzumab is sometimes accompanied by severe prolonged cytopenia (25–29). We therefore asked whether CD52 is also expressed on normal bone marrow stem cells. However, in normal bone marrow samples and in most patients with ICUS, CD34⁺/CD38⁻ cells did not express CD52. In contrast, in some patients with non-Hodgkin lymphoma without histologically detectable bone marrow involvement, CD52 was found to be expressed on CD34⁺/CD38⁻ stem cells. In these patients, alemtuzumab was found to induce cell lysis in CD34⁺/CD38⁻/CD52⁺ stem cells. In contrast, no effects of alemtuzumab on normal CD52⁻ stem cells were seen.

In summary, our data show that CD34⁺/CD38⁻ NSCs in patients with del(5q) MDS and a group of AML, express the target antigen CD52. Moreover, our data show that alemtuzumab induces rapid, complement-dependent lysis of NSC/LSC in these patients. In addition, CD52 may be a prognostic NSC/LSC marker in patients with MDS and AML. The exact value of CD52 as a NSC/LSC marker and target remains to be determined in future studies.

No potential conflicts of interest were disclosed.

Conception and design: K. Blatt, P. Valent

Development of methodology: G. Hoermann

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Blatt, H. Herrmann, G. Hoermann, M. Willmann, S. Cerny-Reiterer, I. Sadovnik, S. Herndlhofer, B. Streubel, W. Rabitsch, W.R. Sperr, M. Mayerhofer, T. Rülicke

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K. Blatt, H. Herrmann, G. Hoermann, M. Willmann, S. Cerny-Reiterer, I. Sadovnik, S. Herndlhofer, B. Streubel, W.R. Sperr, M. Mayerhofer

Writing, review, and/or revision of the manuscript: K. Blatt, G. Hoermann, M. Willmann, W. Rabitsch, T. Rülicke, P. Valent

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Willmann, S. Herndlhofer

Study supervision: P. Valent

The authors thank Gabriele Stefanzl (Department of Internal Medicine I, Medical University of Vienna) as well as Günther Hofbauer and Andreas Spittler (Cell Sorting Core Unit, Medical University of Vienna) and Tina Bernthaler (University of Veterinary Medicine Vienna) for excellent technical support.

This work was supported by the Austrian Science Fund (FWF) SFB project #F4704-B20 (to P. Valent).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.