Rearrangement and expression of the immunoglobulin μ-chain gene in human myeloid cells

Immunoglobulin (Ig), a characteristic marker of B cells, has been reported to be expressed in epithelial cells, with a suggested role in their growth and survival. We have previously reported that IgG heavy chain is expressed in acute myeloid leukemia (AML), but not in the monocytes or neutrophils from patients with non-hematopoietic neoplasms or healthy controls. In the present study, we assessed IgM heavy chain expression and repertoire in human myeloid cells. We detected VHμDJHμ rearrangement and expression in 7/7 AML cell lines, 7/14 primary myeloblasts from AML patients, and interestingly, 8/20 monocytes and 3/20 neutrophils from patients with non-hematopoietic neoplasms and healthy individuals. We also found evidence of somatic hypermutation of the variable (V) gene segments in AML-derived IgM gene rearrangements but not in IgM from monocytes or neutrophils from patients with non-hematopoietic neoplasms and healthy individuals. Furthermore, IgM VHμDJHμ gene rearrangements in AML cell lines, primary myeloblasts, and monocytes and neutrophils from patients with non-hematopoietic neoplasms showed a restricted V usage and repertoire, whereas the VHμDJHμ gene rearrangements in monocytes and neutrophils from healthy individuals displayed more diversity. Anti-human IgM inhibited cell proliferation, but did not induce apoptosis in AML cell lines. Our findings suggest that AML-derived IgM might be a novel AML-related molecule that is involved in leukemogenesis and AML progression and might serve as a useful molecular marker for designing targeted therapy and monitoring minimal residual disease.


INTRODUCTION
It is well known that immunoglobulin (Ig) plays an essential role as an antibody in blood and body fluids and is a protective protein produced by the immune system in response to foreign antigens. Although it was once presumed that B lymphocytes and plasma cells are the only sources of Ig, it has been recently reported that Ig is expressed in many types of neoplastic epithelial cells 1-6 and several types of non-hematopoietic cells from healthy individuals. [7][8][9] Moreover, non-hematopoietic cell-derived Ig shares some interesting features, including the restricted or biased usage of certain sequences, 9,10 a unique glycosylation profile 11,12 and novel regulatory mechanisms of gene expression. 13,14 Furthermore, non-hematopoietic cellderived Ig has a role in cell growth and survival. 1, 3,15,16 These findings suggest that Ig derived from non-hematopoietic cells might be a novel molecule with a unique structure and function and might play a role in carcinogenesis and/or tumor progression.
It has long been thought that myeloid cells and lymphocytes originate from different hematopoietic precursor cells in the bone marrow and that myeloid cells do not produce Ig. In an earlier study, we showed that IgG is rearranged, expressed and secreted in acute myeloid leukemia (AML) cells but not in monocytes or neutrophils. 15 Furthermore, AML-derived IgG variable (V) segments demonstrated somatic hypermutation and restricted or biased V region usage, and anti-human IgG inhibited cell proliferation and promoted apoptosis in AML cell lines. In the current study, we sought to assess the expression and repertoire of IgM heavy chain in human myeloid cells. We detected V Hm DJ Hm rearrangement and expression in both AML cell lines and primary CD33 1 myeloblasts from AML patients. Interestingly, unlike IgG, IgM expression was also found in CD33 1 monocytes and CD33 1 neutrophils from patients with non-hematopoietic neoplasms and healthy individuals, with a similar frequency as that in primary myeloblasts.
In addition, IgM V Hm DJ Hm gene rearrangements in AML cell lines, primary myeloblasts, and monocytes and neutrophils from patients with non-hematopoietic neoplasms exhibited a restricted V usage and repertoire, whereas the V Hm DJ Hm gene rearrangements in monocytes and neutrophils from healthy individuals showed more diversity. Moreover, we found evidence of somatic hypermutation of the V gene segments in the AML-derived IgM gene rearrangements, but not in IgM derived from monocytes or neutrophils from patients with non-hematopoietic neoplasms and healthy individuals. Antihuman IgM inhibited cell proliferation, but did not induce apoptosis in AML cell lines.
These findings provide further evidence that Ig can be expressed by myeloid cells and that myeloid cell-derived V Hm DJ Hm gene rearrangements may have a restricted usage pattern under certain selection pressure due to neoplastic conditions. All these results suggest that similar to AML-derived IgG, AML-derived IgM might be involved in AML pathogenesis and/or progression and might serve as a useful marker for the development of targeted therapy and the monitoring of minimal residual disease.

Cell lines
The AML cell lines (HL-60, THP-1 and U937), the B-cell lymphoma cell line (Daudi), and the T-cell lymphoma cell line  were provided by the Center for Human Disease Genomics, Peking University. The AML cell lines (KG-1, NB4, HEL and OCI-AML3) were provided by the Department of Hematopathology, MD Anderson Cancer Center. All cell lines were maintained as described previously. 15 Patient samples Peripheral blood samples were obtained from 14 AML patients for the assessment of membrane IgM expression on CD33 1 myeloblasts. In addition, peripheral blood samples were obtained from another 14 AML patients, 12 patients with non-hematopoietic neoplasms, and 8 healthy individuals to assess IgM V Hm DJ Hm gene transcription in CD33 1 myeloblasts, monocytes and neutrophils. The clinical, laboratory and pathological data were collected by review of the medical records. The study was performed under an institutional review board-approved protocol.
The peripheral blood samples from patients with nonhematopoietic neoplasms and healthy individuals were stained with monoclonal antibodies against CD19 (APC), CD33 (FITC) and CD138 (PE) and sorted for monocytes and neutrophils after the lysis of red blood cells, as described previously. 15 The monocytes and neutrophils were gated using forward versus side-scatter dot plots. The CD33 1 CD19 2 CD138 2 cells in both gates were selected. We also collected CD33 2 CD19 1 cells and used them as a positive control.

Flow cytometry immunophenotyping
To assess IgM expression on the cellular membrane of different cell lines, cultured cells were washed with PBS, blocked with 2% fetal bovine serum (FBS) at 4 uC for 30 min, and stained with monoclonal mouse anti-human IgM (m-chain specific, 2 ng/ ml; Mabworks Biotech Co. Ltd, Beijing, China) at 4 uC for 40 min. After two washes, the cells were stained with FITCconjugated goat anti-mouse IgG (1 : 100, Santa Cruz Biotechnology, Inc., Dallas, TX, USA) at 4 uC for 30 min; the cells were then washed and analyzed using a FACSCalibur flow cytometer with CellQuest software (BD Biosciences, San Jose, CA, USA). To detect intracellular IgM, the cells were washed, fixed with 4% paraformaldehyde at room temperature for 30 min, washed, permeabilized with permeabilization buffer (eBioscience, San Diego, CA, USA), and centrifuged at 1500 r.p.m. at 4 uC for 5 min. The cells were then stained with the appropriate primary and secondary antibodies and analyzed as described above.
To assess IgM expression on the myeloblasts from AML patients, PBMCs were double stained with CD33-FITC and IgM-PE at 4 uC for 15 min after the lysis of red blood cells and analyzed using a FACSCalibur flow cytometer. The cells that were double stained with CD19-APC and IgM-PE were used as a positive control.

Immunocytochemistry
Cultured AML cells were centrifuged onto slides by cytospin, fixed, processed and stained with monoclonal mouse antihuman IgM (2 ng/ml) in a 37 uC humidified chamber for 1 h, as described previously. 15 The cells were then incubated with anti-mouse IgG-horseradish peroxidase (Dako, Carpinteria, CA, USA), and the bound antibodies were detected using 3,39diaminobenzidine tetrahydrochloride (Dako). 15 Immunoprecipitation and western blot analysis To extract cytoplasmic proteins, AML cells were lysed with radioimmunoprecipitation assay lysis buffer (Thermo Fisher Scientific Inc., Rockford, IL, USA) on ice for 30 min. After centrifugation at 15 000 r.p.m. for 15 min, the lysate was incubated with monoclonal mouse anti-human IgM or mouse IgG1 (isotype control; Sigma, St Louis, MO, USA) and protein G/A beads (Abcam, Cambridge, MA, USA) for 30 min. The beads were washed, and the antigen-antibody complex was then eluted, followed by sodium dodecyl sulfate polyacrylamide gel electrophoresis and western blot analysis using goat antihuman IgM (1 : 2000; Sigma) as the primary antibody, and IRDye 800-conjugated, affinity-purified rabbit anti-goat IgG (LI-COR Biosciences, Lincoln, NE, USA) as the secondary antibody. The signals were detected using an Odyssey Infrared Imager (LI-COR Biosciences).

Analysis of IgM V Hm DJ Hm transcripts by reverse transcription PCR and sequencing
Total RNA was isolated from AML cell lines using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) or from sorted CD33 1 CD19 2 CD138 2 or CD33 2 CD19 1 CD138 2 cells using the RNeasy Micro Kit (Qiagen Inc., Chatsworth, CA, USA). Reverse transcription (RT) was performed using the Sensiscript RT Kit (Qiagen Inc.). The IgM V Hm DJ Hm transcripts were detected using nested PCR, 10,15,18 and the amplification products were gel-purified, cloned, and sequenced. 15 The rearranged IgM V Hm DJ Hm sequences were analyzed for the V, D and J region usage and evidence of somatic hypermutation, as described previously. 15,19 Cell proliferation assay A cell proliferation assay was performed using the HL-60 and OCI-AML3 cell lines and a Cell Counting Kit-8 (CCK-8; Dojindo Molecular Technologies, Inc., Kumamoto, Japan). The cells were cultured in a 96-well plate at 5310 3 cells/well in the presence of goat anti-human IgM (10 or 20 mg/ml; Sigma) or goat IgG (isotype control, 10 or 20 mg/ml; Sigma). At 24, 48, 72 and 96 h, the cell viability was assessed by adding 10 ml CCK8 solution in 100 ml RPMI-1640 media, incubating at 37 uC for 2 h, and measuring the absorbance with a 450-nm filter. This study was performed in triplicate, and the results were averaged.

Cell apoptosis assay
The HL-60 and OCI-AML3 cell lines were cultured in a 12-well plate at 1310 5 cells/well in the presence of goat anti-human IgM (10 or 20 mg/ml; Sigma) or goat IgG (isotype control, 10 or 20 mg/ml; Sigma). At 24, 48 and 72 h, the cells were harvested, washed and resuspended in 200 ml binding buffer (10 mM HEPES (pH 7.4), 140 mM NaCl, 1 mM MgCl 2 , 5 mM KCl and 2.5 mM CaCl 2 ). The cells were then incubated with FITC-conjugated annexin V (Beijing Biosea, Beijing, China) at room temperature for 30 min, followed by the addition of propidium iodide (1 mg/ml) and apoptosis analysis using the FACSCalibur flow cytometer.

Statistical analyses
Statistical analyses were performed using SAS 8.1 software (SAS Institute Inc., Cary, NC, USA). Differences between the groups were assessed using Student's t-test or the x 2 test. A P value of ,0.05 was regarded as statistically significant.

Patients' characteristics
We assessed membranous IgM expression on myeloblasts from 14 AML patients, the clinicopathological, cytogenetic and molecular features of whom have been previously described. 15 We analyzed IgM V Hm DJ Hm gene transcription in sorted CD33 1 myeloblasts from another 14 AML patients, including acute myelomonocytic leukemia (AML-M4, five cases), AML with myelodysplasia-related changes (AML-MRC, three cases), AML without maturation (AML-M1, two cases), AML with maturation (AML-M2, two cases), acute promyelocytic leukemia (one case) and acute monocytic leukemia (AML-M5, one case) (Table 1). We also evaluated IgM V Hm DJ Hm gene transcription in CD33 1 monocytes and neutrophils from 12 patients with non-hematopoietic neoplasms and 8 healthy individuals. The group of patients with non-hematopoietic neoplasms included those with colon adenocarcinoma (four cases) and one case each of stomach adenocarcinoma, hepatocellular carcinoma, pancreas adenocarcinoma, glioblastoma, high-grade sarcoma, thymoma, squamous carcinoma of the tongue and a benign thyroid nodule (Table 1).

IgM is expressed in AML cell lines
We first assessed IgM expression in AML cell lines by immunocytochemical studies and flow cytometric analyses using a mouse monoclonal antibody against human IgM (m-chain specific). A B-cell lymphoma cell line (Daudi) was used as a positive control, and a T-cell lymphoma cell line (MOLT-4) was used as a negative control. Flow cytometric analyses revealed that IgM was expressed both on the plasma membrane and in the cytoplasm in all four AML cell lines assessed (THP-1, OCI-AML3, HL-60 and U937) and in the Daudi cell line but not in the MOLT-4 cell line ( Figure 1a). Immunocytochemistry further confirmed that the IgM molecule was localized both on the plasma membrane and in the cytoplasm of these cell lines ( Figure 1b). Moreover, we performed immunoprecipitation and western blot analyses and detected the presence of Ig mchain in these AML cell lines (Figure 1c).

IgM is expressed in primary AML cells
To demonstrate whether IgM is also produced in primary AML cells, we assessed IgM expression on PBMCs from 14 AML patients by two-color flow cytometry using monoclonal mouse anti-human IgM-PerCP/Cy5.5 and CD33 1 -FITC antibodies. Our results revealed that IgM was expressed on the cell surface of CD33 1 myeloblasts in 7 of 14 AML patients, including patients with AML-M4 (three cases), AML-M1 (two cases), AML-M5 (one case) and AML-MRC (one case; Table 2). Moreover, we observed that strong IgM staining was more frequently observed in cases of AML-M4 and AML-MRC compared to AML-M1 and AML-M2 ( Figure 2).
IgM V Hm DJ Hm transcripts are detected in AML cell lines, primary myeloblasts, and non-neoplastic monocytes and neutrophils To detect rearrangements and Ig m-chain transcription in AML cell lines, we performed RT-PCR using degenerate primers specific for the FR2 and JH regions of IgM V Hm DJ Hm that can amplify nearly all IgM V Hm DJ Hm rearrangements (Figure 3a). Rearranged IgM V Hm DJ Hm transcripts were successfully amplified in all seven AML cell lines assessed (THP-1, OCI-AML3, HL-60, U937, HEL, KG-1, and NB4) and in the Daudi cell line (positive control), but not in the MOLT-4 cell line (negative control, Figure 3b). As expected, CD19 transcripts were only detected in the Daudi cell line and not in the AML cell lines and MOLT-4 cell line, indicating that no contamination of B cells was present in the AML cell lines utilized.
To identify rearrangements and Ig m-chain transcription in primary AML cells, we conducted flow cytometry cell sorting of PBMCs (Figure 4a, P1) from 14 AML patients using monoclonal antibodies against CD19 (APC), CD33 (FITC) and CD138 (PE). The CD33 1 CD19 2 CD138 2 cells were isolated and were regarded as myeloblasts (Figure 4a, P4); we also isolated CD19 1 CD33 2 cells (Figure 4a, P3) and used them as a positive control. We then conducted RT-PCR using the primers described above on the sorted CD33 1 CD19 2 CD138 2 myeloblasts. We detected IgM V Hm DJ Hm transcripts in the CD33 1 CD19 2 CD138 2 myeloblasts from 7 of 14 AML patients, including AML-M4 (three cases), AML-M2 (two cases), AML-M5 (one case) and AML-MRC (one case; Table 2). Moreover, we performed immunoprecipitation and western blot analyses and detected the expression of Ig m-chain in two cases of primary AML cells (Figure 1c).
To investigate whether IgM rearrangements and transcription also occur in adult monocytes and neutrophils, we sorted  (Figure 4b, P6). We then conducted RT-PCR using the primers described above on the sorted CD33 1 adult monocytes and neutrophils. Surprisingly, IgM V Hm DJ Hm transcripts were also detected in monocytes from 4 of 12 patients with non-hematopoietic neoplasms and 4 of 8 healthy individuals and in neutrophils from 1 of 12 patients with non-hematopoietic neoplasms and 2 of 8 healthy individuals (Tables 3 and 4).
IgM V Hm DJ Hm rearrangements demonstrate differentially restricted or biased usage of V, D and J segments in AML cell lines, primary AML cells, and monocytes and neutrophils from patient with non-hematopoietic neoplasms, but show diversity in healthy individuals We further analyzed the rearranged sequences of AML-derived IgM V Hm DJ Hm and observed a clonal (HEL) or biclonal (THP-1, HL-60, U937, NB4 and KG-1) pattern of V Hm DJ Hm rearrangements (Table 5). Interestingly, we found that the VH3 gene family, particularly VH3-23, was preferentially used in six of the seven cells lines assessed (THP-1, OCI-AML3, HL-60, U937, NB4 and KG-1). This was in contrast to the positive control cell line, Daudi, which showed a preference of the   (Table 5). Furthermore, the D6 gene family, particularly D6-19 and D6-13, and the JH5 gene family were preferentially used in five of the seven cell lines assessed (THP-1, OCI-AML3, HL-60, NB4 and KG-1) ( Table 5). Moreover, several identical patterns of V Hm DJ Hm combinations were observed in these cell lines, with VH3-23/D6-19/JH5 being the most common, suggesting a restricted or biased usage of V, D and J segments (Table 5).
We also studied the rearranged sequences of IgM V Hm DJ Hm in primary AML cells. We found the typical V Hm DJ Hm recombination in the V region that usually occurs with B-cell-derived IgM, such as the inclusion of an N or P insert, and that JH gene recombination occurred primarily at the TG nucleotide sequence. 20 Similar to the occurrence in AML cell lines, the V Hm DJ Hm sequences in primary AML cells demonstrated a unique monoclonal (six of seven) or oligoclonal (one of seven, with three different V Hm DJ Hm rearrangements) V Hm DJ Hm pattern (Table 2). However, unlike V Hm DJ Hm rearrangements in AML cell lines, none of the seven patients showed an identical V Hm DJ Hm pattern. When comparing these V Hm DJ Hm sequences with the best-matching functional germline sequences, there was a preferential use of VH3-48 (three cases) and VH3-23 (two cases). Each of the other two cases exhibited the use of VH6-1 or VH4-61 (Table 2). Among 36 cloned V Hm DJ Hm sequences from the seven AML samples, VH3-48 and VH3-23 accounted for 33.3% (12 clones) and 27.8% (10 clones) of all sequences, respectively, and VH6-1 (6 clones, 16.7%) and VH4-61 (5 clones, 13.9%) were also used at a higher frequency than was expected (Table 2). Furthermore, the D1 and D2 sequences were among the most frequently used D sequences, and the JH sequences showed preferential JH4 and JH5 usage ( Table 2).
Similar to the observations in primary AML patients, the V Hm DJ Hm rearrangements in the CD33 1 monocytes and neutrophils from patients with non-hematopoietic neoplasms exhibited a monoclonal (one of four) or biclonal (three of four) pattern. However, no identical V Hm DJ Hm rearrangement was detected in the monocytes or neutrophils from different individuals (Table 3). Only one case of monocytes from a patient with gastric adenocarcinoma showed a monoclonal pattern. In contrast, the V Hm DJ Hm rearrangements from healthy individuals appear to display more diversity; again, no identical V Hm DJ Hm rearrangement was detected in the monocytes or neutrophils from different individuals (Table 4). Moreover, the usage of V, D, and J segments in the healthy individuals was different from that of the primary CD33 1 myeloblasts from AML patients (Tables 3 and 4). Furthermore, we assessed the mutational status of the IgM V genes from AML patients. Unexpectedly, somatic hypermutations were detected in the AML-derived IgM V Hm DJ Hm genes but not in IgM derived from monocytes or neutrophils from patients with non-hematopoietic neoplasms or healthy individuals.

Anti-human IgM inhibits cell proliferation, but does not induce apoptosis in AML cell lines
We next addressed whether AML-derived IgM was involved in cell proliferation and survival. The HL-60 and OCI-AML3 cell lines were incubated with goat anti-human IgM or goat IgG as an isotype control. We assessed cell proliferation at different time points using a CCK8 kit and apoptosis using flow cytometry after labeling with Annexin-V/propidium iodide. As shown in Figure 5, anti-human IgM slightly inhibited cell proliferation at 48, 72 and 96 h in a dose-dependent manner, but did not induce apoptosis in the HL-60 and OCI-AML3 cell    lines, suggesting that AML-derived IgM is involved in cell proliferation but not survival.

DISCUSSION
It has been thought that Ig can only be produced by B lymphocytes and plasma cells. However, in the past decade, we and other groups have found that Ig can also be expressed in other types of cells, for instance, epithelial cells, germ cells and neurons, and in particular, many cancer cells. 1-9 More recently, we have shown that IgG heavy chain is transcribed and expressed in AML cells but not in non-neoplastic monocytes or neutrophils. Moreover, IgG V H DJ H rearrangements in AML cells display restricted or biased V segments and have experienced somatic hypermutation. In addition, anti-human IgG inhibits cell proliferation and induces apoptosis in AML cell lines. 15 These findings suggest a role of IgG gene expression in AML pathogenesis and progression.
In the present study, we extended our initial findings and demonstrated that IgM heavy chain is also transcribed and expressed in AML cell lines and primary myeloblasts. Surprisingly, unlike IgG, IgM was also transcribed and expressed in non-neoplastic monocytes and neutrophils from patients with non-hematopoietic neoplasms or healthy individuals. However, the observed myeloblast-derived V Hm DJ Hm rearrangements displayed a unique monoclonal or oligoclonal pattern, which was different from the diversity of V Hm DJ Hm  Figure 4 FACS successfully isolated CD33 1 CD19 2 CD138 2 myeloblasts from AML patients and CD33 1 monocytes and neutrophils from patients with non-hematopoietic neoplasms and healthy individuals. (a) CD33 1 CD19 2 CD138 2 myeloblasts from the peripheral blood of AML patients. P1, mononuclear cells; P2, CD138 1 cells; P3, CD19 1 cells; P4, CD33 1 CD19 2 CD138 2 myeloblasts. The cells in P4 were selected and defined as CD33 1 CD19 2 CD138 2 myeloblasts; the cells in P3 were collected and used as a positive control. (b) CD33 1 monocytes and neutrophils from the peripheral blood of patients with non-hematopoietic neoplasms or healthy controls. P1, mononuclear cells; P2, granulocytes; P3, CD138 2 mononuclear cells; P4, CD138 2 granulocytes; P5, CD33 1 CD19 2 monocytes; P6, CD19 1 B cells; P7, CD33 1 neutrophils; P8, CD33 2 or dim 1 granulocytes. Cells in P5, P6 and P7 were selected. AML, acute myeloid leukemia; FACS, fluorescence-activated cell sorting.
rearrangements in the non-neoplastic myeloid cells from healthy individuals. Moreover, myeloblast-derived V Hm DJ Hm revealed somatic hypermutation. Our preliminary functional assay showed that anti-human IgM inhibited AML cell proliferation in culture. These findings may indicate a role of IgM expression in AML pathogenesis and progression.   n a n a 7 n a n a 8 n a n a Abbreviation: na, no V Hm DJ Hm rearrangement detected.  Although it has been reported, albeit rarely, that the Ig gene is expressed in a small subset of AML patients, 21 it has remained unclear whether Ig is truly produced by myeloid cells or is secreted by residual B lymphocytes in these specimens. It is also unknown whether Ig is transcribed and translated or merely derived from residual B lymphocytes and simply bound nonspecifically to myeloid cells. Moreover, it has not been investigated whether Ig is expressed in non-neoplastic myeloid cells, e.g., monocytes and neutrophils. To address these issues, we conducted flow cytometry cell sorting to isolate CD33 1 CD19 2 CD138 2 myeloblasts from AML patients and monocytes and neutrophils from patients with non-hematopoietic neoplasms and healthy individuals. We utilized flow cytometry immunophenotyping to confirm that there were no residual B lymphocytes and/or plasma cells present in the specimens. We then studied IgM gene rearrangement and transcription using RT-PCR with primers specific to the IgM V Hm DJ Hm genes and successfully detected V Hm DJ Hm transcripts in CD33 1 myeloblasts in 7 of 14 AML patients. Surprisingly, unlike our previous findings for IgG, IgM was also detected in CD33 1 monocytes and neutrophils from patients with nonhematopoietic neoplasms and healthy controls. These findings confirm that the expression of IgM in myeloid cells, as detected by flow cytometry and immunohistochemistry, is indeed produced by myeloid cells and is not due to the non-specific membranous binding of IgM that is produced and secreted by residual B lymphocytes in the specimens analyzed.
We compared the diagnostic classification between AML patients with and without V Hm DJ Hm transcript expression and found that IgM was expressed at a higher frequency and level in AML with monocytic differentiation (AML-M4 or -M5). Consistently, IgM was expressed at a higher frequency in monocytes (4 of 12 patients with non-hematopoietic neoplasms, 4 of 8 healthy controls) than neutrophils (1 of 12 patients with non-hematopoietic neoplasms, 2 of 8 healthy controls). Although the number of cases in our study was small, these findings suggest that monocytes may be able to express IgM more frequently than neutrophils under certain physiological and pathological conditions. Moreover, unlike epithelial cancer cell-derived IgG, which tends to be expressed in cancer cells with higher frequency than that in normal cells, 10 we did not observe a significant difference in the frequency of IgM expression between primary AML cells (50%) and nonneoplastic monocytes (50%) in healthy individuals.  noted that two of the seven patients who exhibited IgM expression showed dim partial CD19 expression; however, a complete flow cytometry immunophenotyping proved that the blasts were of myeloid lineage rather than mixed myeloid and B-cell lineage. We also compared the clinical outcome between AML patients with and without IgM expression and did not observe any significant differences. Regardless, the number of cases in this study may be too small to reach any definitive conclusions.
We have previously reported that epithelial cancer cellderived V Hm DJ Hm sequences display high conservation among different tumor types and different individuals. 10 Several sets of V Hm DJ Hm rearrangements, such as VH3-33 and VH6-1 (not VH3-23 and VH3-48, which were used more often in the AML cells in this study), have been reported to be frequently expressed among different types of epithelial cancer cells, sometimes with 100% homology. 10 These data suggest that epithelial cancer cell-derived IgM has a similar function among different types of cancer cells. However, although myeloid cellderived V Hm DJ Hm rearrangements showed an oligoclonal or monoclonal pattern in each AML patient, no conservative usage was observed among the AML-derived V Hm DJ Hm rearrangements. These findings suggest that the structure and/or function of AML-derived IgM may be unique in each individual, a fact that may prove to be useful for implementing this molecule as a marker for monitoring minimal residual disease after chemotherapy.
With 22 functional genes, the VH3 gene family is the largest of the seven families (VH1 to VH7) and comprises approximately one-half of the expressed VH repertoires in adult peripheral B cells. 22 Similar to B cells, among a total of 115 V Hm DJ Hm sequences derived from myeloid cells (25 in AML cell lines, 36 in AML patients, 27 in patients with non-hematopoietic neoplasms and 27 in healthy controls; Tables 2-5), the VH3 gene family was used most frequently (90 clones, 78.2%), particularly VH3-23 (43 clones, 37.4%) and VH3-48 (16 clones, 13.9%). It has been reported that VH3-23 is a promiscuous VH gene segment that contributes to many antigenic specificities in B cells. VH3-23 is also frequently expressed during the fetal development of antibody repertoires. 23 In a single-cell PCR analysis of peripheral blood lymphocytes from a normal adult, VH3-23 was found in approximately 17% of all functional rearrangements, 24 and VH3-48 has been reported as an unusual gene in circulating B-cell repertoires with a high homology to VH3-23. 25 Our results also revealed that VH3-23 accounted for 47.5% (29 of 61 clones; Tables 2 and 5) of all rearrangements in AML cell lines and primary AML cells, which was higher than that in patients with non-hematopoietic neoplasms (26.0%, 7 of 27 clones) and healthy individuals (26.0%, 7 of 27 clones). Furthermore, although our results showed an over-representation of VH3-23 and VH3-48 similar to that in B cells, it was surprising that VH4-61 and VH6-1 occurred at a much higher frequency than expected when compared to CD19 1 B-cell-derived VH. Furthermore, there was preferential usage of JH4 and JH5 that was also remarkably higher than expected compared to CD19 1 B cell-derived JH. Such preferential usage of certain V and J gene segments in myeloid cell-derived V Hm DJ Hm , combined with our findings that epithelial cell-derived IgM displays natural antibody activity (unpublished data), suggest that myeloid cell-derived IgM may have a unique function, such as involvement in innate immunity as a natural antibody.
In summary, we demonstrated that IgM is transcribed and expressed in AML cell lines and primary myeloblasts and in the monocytes and neutrophils from patients with non-hematopoietic neoplasms and healthy controls. Anti-human IgM inhibits AML cell proliferation in culture. Thus, AML-derived IgM might play a role in AML pathogenesis and progression and might serve as a useful marker for monitoring minimal residual AML and as a potential target for AML therapy.
AUTHOR CONTRIBUTIONS JH and XQ designed and performed most of the experiments and wrote the manuscript. XS, XG, ZH and LC performed some of the experiments. CCY and XQ conceptualized the study and edited the manuscript.