Competition between clonal plasma cells and normal cells for ... - Nature

Leukemia (2011) 25, 697–706 & 2011 Macmillan Publishers Limited All rights reserved 0887-6924/11 www.nature.com/leu

ORIGINAL ARTICLE Competition between clonal plasma cells and normal cells for potentially overlapping bone marrow niches is associated with a progressively altered cellular distribution in MGUS vs myeloma B Paiva1,2, M Pe´rez-Andre´s2,3, M-B Vı´driales1,2, J Almeida2,3, N de las Heras4, M-V Mateos1,2, L Lo´pez-Corral1,2, NC Gutie´rrez1,2, J Blanco1, A Oriol5, MT Hernańdez6, F de Arriba7, AG de Coca8, M-J Terol9, J de la Rubia10, Y Gonza´lez11, A Martıń12, A Sureda13, M Schmidt-Hieber2,3, A Schmitz14, HE Johnsen14, J-J Lahuerta15, J Blade´16, JF San-Miguel1,2 and A Orfao2,3 on behalf of the GEM (Grupo Espanõl de MM)/PETHEMA (Programa para el Estudio de la Terapeútica en Hemopatıás Malignas) cooperative study groups and the Myeloma Stem Cell Network (MSCNET) 1

Servicio de Hematologıá, Hospital Universitario de Salamanca, Salamanca, Spain; 2Servicio de Hematologia, Centro de Investigacioń del Cańcer (CIC, IBMCC USAL-CSIC), Salamanca, Spain; 3Servicio General de Citometrıá and Departamento de Medicina, Universidad de Salamanca, Salamanca, Spain; 4Servicio de Hematologıá, Complejo Hospitalario de Leoń, Leoń, Spain; 5 Servicio de Hematologıá, Hospital Universitari Germans Trias i Pujol, Badalona, Spain; 6Servicio de Hematologıá, Hospital Universitario de Canarias, Tenerife, Spain; 7Servicio de Hematologıá, Hospital Morales Meseguer, Murcia, Spain; 8Servicio de Hematologıá, Hospital Clıńico Universitario de Valladolid, Valladolid, Spain; 9Servicio de Hematologıá, Hospital Clinico Universitario de Valencia, Valencia, Spain; 10Servicio de Hematologıá, Hospital La Fe, Valencia, Spain; 11Servicio de Hematologıá, Hospital Josep Trueta, Girona, Spain; 12Servicio de Hematologıá, Hospital Virgen de la Concha, Zamora, Spain; 13 Servicio de Hematologıá, Hospital Santa Creu I Sant Pau, Barcelona, Spain; 14Servicio de Hematologia, Service of Hematology, Aalborg Hospital, Aarhus University Hospital, Aalborg, Denmark; 15Servicio de Hematologıá, Hospital 12 de Octubre, Madrid, Spain and 16Servicio de Hematologıá, Hospital Clıńic, IDIBAPS, Barcelona, Spain

Disappearance of normal bone marrow (BM) plasma cells (PC) predicts malignant transformation of monoclonal gammopathy of undetermined significance (MGUS) and smoldering myeloma (SMM) into symptomatic multiple myeloma (MM). The homing, behavior and survival of normal PC, but also CD34 þ hematopoietic stem cells (HSC), B-cell precursors, and clonal PC largely depends on their interaction with stromal cell-derived factor-1 (SDF-1) expressing, potentially overlapping BM stromal cell niches. Here, we investigate the distribution, phenotypic characteristics and competitive migration capacity of these cell populations in patients with MGUS, SMM and MM vs healthy adults (HA) aged 460 years. Our results show that BM and peripheral blood (PB) clonal PC progressively increase from MGUS to MM, the latter showing a slightly more immature immunophenotype. Of note, such increased number of clonal PC is associated with progressive depletion of normal PC, B-cell precursors and CD34 þ HSC in the BM, also with a parallel increase in PB. In an ex vivo model, normal PC, B-cell precursors and CD34 þ HSC from MGUS and SMM, but not MM patients, were able to abrogate the migration of clonal PC into serial concentrations of SDF-1. Overall, our results show that progressive competition and replacement of normal BM cells by clonal PC is associated with more advanced disease in patients with MGUS, SMM and MM. Leukemia (2011) 25, 697–706; doi:10.1038/leu.2010.320; published online 21 January 2011 Keywords: monoclonal gammopathies; malignant transformation; bone marrow niche competition; plasma cells

Introduction Plasma cell (PC) disorders include an heterogeneous spectrum of diseases, from pre-malignant conditions with variable rates of Correspondence: Professor A Orfao, Centro de Investigacioń del Cańcer, Avda. Universidad de Coimbra S/N (Campos Miguel de Unamuno), Salamanca 37007, Spain. E-mail: [email protected] Received 21 October 2010; revised 30 November 2010; accepted 10 December 2010; published online 21 January 2011

progressionFE1 and 10% per year for monoclonal gammopathy of undetermined significance (MGUS) and smoldering myeloma (SMM), respectively1–4Fto symptomatic multiple myeloma (MM) with median overall survival rates of 3–7 years.5,6 Despite the different tumor mass7–14 and clinical behavior of the disease, clonal PC from MGUS, SMM and MM patients show highly similar and largely overlapping genetic profiles.15–17 In addition, no clear phenotypic differences have been reported so far among clonal PC from MGUS, SMM and MM,18 except for a few molecules involved in the interaction between PC and their microenvironment.19 We have previously shown that the proportion of normal PC within the bone marrow (BM) PC compartment (normal PC/BM PC) is an efficient single parameter for discrimination between MGUS and MM;20 at diagnosis, most MGUS cases (480%) display 45% normal PC/BM PC,21 whereas the great majority (485%) of symptomatic MM show o5% normal PC/BM PC.22 Additionally, the presence of 45% normal PC/BM PC at diagnosis is associated with both a lower risk of progression of MGUS and SMM,21 and a favorable outcome in MM.22 In turn, a progressive decrease in the serum levels of uninvolved immunoglobulins (Ig) is observed from MGUS to SMM and MM.21,22 Altogether, these results suggest that disappearance of normal BM PC followed by impaired secretion of normal Igs is associated with malignant transformation of MGUS and SMM as well as disease progression in MM. Serum Ig levels are tightly regulated in humans and they largely depend on Ig production and secretion by long-living BM and mucosa-associated lymphoid tissue-derived PC.23,24 Recently produced normal PC, which have left secondary lymphoid tissues through peripheral blood (PB), usually need to migrate into BM niches and adhere to CXCL12 (SDF-1)expressing BM stromal cells to become long-living BM PC.23,24 SDF-1-expressing BM stromal cells are also crucial for early pro-B25 andFtogether with interleukin-7-expressing cells25Fpotentially also for later pre-B-cell development26,27 into immature/transitional and naive B-cells that will exit the BM

Altered bone marrow homing to PC niches in MG B Paiva et al

698

through PB.28 These findings suggest that due to the limited number of BM PC niches, normal PC as well as normal B-cell precursors25 could compete with clonal PC for BM PC niches; in turn, CXCL12-CXCR4 signaling appears to be also a key component of the CD34 þ hematopoietic stem cell (HSC) niche.29,30 Interestingly, recent results show that in normal individuals, both CD34 þ HSC and PC recirculate between the BM and the PB.23,31,32 Accordingly, it can be hypothesized that BM B-cell precursors, CD34 þ HSC and PC are under a tight and continuous regulation by the SDF-1/CXCR4 axis, with a certain degree of competition among them, for the same BM niches. In this competition, an advantage could be expected for the malignant clone based on its numerical expansion. An additional factor to be considered is the advanced age of these patients as SDF-1-expressing BM niches undergo age-induced impairment in their ability to support normal hematopoiesis,33 which could be partly explained by a progressive increase in the number of adipocytic deposits in the BM.34,35 Thus, progressive replacement of normal precursors and normal PC by the malignant clone could contribute to explain the occurrence of cytopenias and hypogammaglobulinemia in MM patients. In this study, we analyze the distribution and competitive migrating capacity of B-cell precursors, CD34 þ HSC, normal PC and clonal PC in the BM and PB of patients with PC disorders vs healthy adults (HA) aged 460 years. Taken together, our results show that progressive competition and replacement of normal BM cells by clonal PC is associated with the malignant transformation of MGUS and SMM and progression of symptomatic MM.

Materials and methods

Patients, controls and samples Overall a total of 206 patients with clonal PC disorders were prospectively studied. These included newly diagnosed MGUS (n ¼ 60), SMM (n ¼ 47) and MM (n ¼ 87) patients, plus 12 MM cases studied at day þ 100 after high-dose therapy/autologous stem cell transplantation (MM POST-HDT/ASCT). In all cases, diagnosis was established according to the International Myeloma Working Group criteria.36 Patients with SMM had a high-risk of progression to symptomatic disease with X2 of the following three criteria at diagnosis: BM PCX10%; M-component IgGX30 g/l or IgAX20 g/l or Bence–Jones protein 410 g/l, and X95% clonal PC from all BM PC and immune paresis. Patients’ relatively short follow-up precluded survival analysis in the present study. The study was conducted on PB and BM samples collected from MGUSFn ¼ 44 and 27 samples from patients with a median age of 72 (range: 49–90 years) and 66 years (range: 41–83 years), respectivelyFSMMFn ¼ 32 and 21 samples, median age of 64 (range: 39–89 years) and 69 years (range: 39–83 years), respectivelyFand MMFn ¼ 74 and 30 samples from patients with a median age of 64 (range: 41–83 years) and 66 years (range: 41–84 years), respectivelyFpatients studied at diagnosis, as well as 12 MM patients POST-HDT/ASCT (median age of 58 years; range: 52–65 years). Paired BM and PB samples were available in 11 of the 60 patients with MGUS, 6 of 47 SMM and 17 of 87 MM studied at diagnosis, as well as in all 12 MM POST-HDT/ASCT cases. As controls, PB samples from 16 HA older than 60 years (median age of 75 years; range: 62–82 years) as well as 6 PB and 20 BM samples obtained from a group of HA undergoing orthopedic surgery (median age of 50 years; range: 38–81 years) Leukemia

were studied, with paired BM and PB samples available in 6 of these latter 20 HA analyzed. For BM control samples, no significant differences in their compositions were observed between subjects who were younger and those older than 55 years (Supplementary Table 1). All control and patient samples were collected after informed consent was given by each individual, according to the local ethical committees and the Helsinki Declaration.

Multiparameter flow cytometry immunophenotypic studies Approximately 1 ml of EDTA-anti-coagulated PB/case was immunophenotyped using a direct eight-color immunofluorescence stain-and-then-lyse technique,32,37 with the following combination of monoclonal antibodies (Pacific blue (PB)/ anemonia majano cyan (AmCyan)/fluorescein isothiocyanate/ phycoerythrin (PE)/peridinin chlorophyll protein–cyanin 5.5 (PerCP–Cy5.5)/PE–cyanin 7 (PE–Cy7)/allophycocyanin/alexafluor 700 (AF700)): CD20/CD45/surface IgM (sIgM) or sIgl/ sIgG, sIgA or sIgk/CD19/CD10/CD27/CD38. In all PB samples analyzed, the following maturation-associated B-cell subsets were identified32 in addition to circulating plasmablasts/PC (CD10/CD27 þ þ /CD38 þ þ ): (i) immature (CD10 þ /CD27/ CD38 þ ), (ii) naive (CD10/CD27/CD38), (iii) memory nonswitched (CD10/CD27 þ /CD38/IgM þ ) and (iv) memory switched (CD10/CD27 þ /CD38/IgM) B-cells. Because of their different patterns of expression of CD38, CD19 and CD45, and their unique light scatter characteristics,22,38 this multiparameter eight-color staining also allowed detection of circulating PB clonal PC, as clearly different from normal circulating PB plasmablasts/PC. In a subset of 13 cases (2 MGUS, 2 SMM and 9 MM), a more detailed phenotypic characterization of PB vs BM clonal PC was performed using the following antibody combinations: (i) CD19/CD45/sIgl/sIgG plus sIgA/CD138/CD27/sIgk/CD38; (ii) CD19/CD45/CD20/CD22/ CD138/CD27/CD56/CD38; (iii) CD19/CD45/FMC7/CD24/CD138/ CD27/CD43/CD38 and (iv) CD19/CD45/CD40/CXCR4/CD138/ CD27/CD28/CD38. Erythrocyte-lysed whole-BM samples (and also their paired PB samples) were immunophenotyped using a single eight-color antibody combinationFCD20/CD45/CD34/CXCR4/CD19/CD10/ CD27/CD38Faimed at the identification and enumeration of the following BM cell populations: (i) CD34 þ precursors and their CD34 þ /CD38/ þ dim (immature), CD34 þ /CD38 þ /CD19 (myeloid) and CD34 þ /CD38 þ /CD19 þ (lymphoid) subsets; (ii) B cells and their CD19 þ /CD34 þ /CD38 þ (Pro-B), CD19 þ / CD34/CD38 þ (Pre-B) and CD19 þ /CD20 þ /CD38 (mature) precursor B-cell subsets and (iii) BM PC, including normal PC (CD38 þ þ /CD19 þ /CD45/ þ ) and clonal PC.22,39,40 In PB samples, the following cell populations were identified with the same eight-color combination: (i) CD34 þ HSC, normal PC and clonal PC. Data acquisition was performed in a FACSCanto II flow cytometer (Becton Dickinson Biosciences (BDB), San Jose, CA, USA) using the FACSDiva software (version 6.1; BDB), and a two-step acquisition procedure: first, information on 5 104 events corresponding to the whole-sample cellularity was stored; second, data were stored only about CD19 þ and/or CD38 þ and/or CD34 þ gated events, for a minimum of 106 leukocytes per tube.

SDF-1 enzyme-linked immunosorbent assay Plasma was collected from PB samples obtained from 6 HA, 10 MGUS, 5 SMM and 14 MM patients at diagnosis and 5 MM


699 POST-HDT/ASCT cases. Plasma levels of SDF-1 were measured using a quantitative enzyme-linked immunosorbent assay (Quantikine, R&D systems Inc., Minneapolis, MN, USA) following the manufacturer’s recommendations.

Transwell migration assay Heparinized BM samples were obtained from HA (n ¼ 4) and MGUS (n ¼ 4), SMM (n ¼ 4) and MM (n ¼ 4) patients at diagnosis. Mononuclear cells were separated by ficoll-hypaque density gradient centrifugation and re-suspended in RPMI-1640 with 1% L-glutamine and 0.5% bovine serum albumin. Cells were placed in the upper chambers of 6.5 mm transwell plates (Costar, Corning, Acton, MA, USA) separated by an 8 mm pore size filter. SDF-1 (R&D systems) was added to the lower chambers at concentrations of 0, 30 and 70 nM. After 4 h at 37 1C, cells in the lower compartments were collected in Trucount tubes (BDB) and quantified by multiparameter flow cytometry, following the manufacturer’s instructions. The different cell subsets were identified by multiparameter flow cytometry using the following five-color antibody combination (fluorescein isothiocyanate/PE/ PE-Cy7/PerCP-Cy5.5/allophycocyanin): CD38/CXCR4/CD19/CD45/ CD34. Specific migration for SDF-1 of potentially competing cell populationsFBM normal PC, B-cell precursors, CD34 þ HSC and clonal PCFwas determined for each cell population using the following formula:

No: of cells=ml ð30 or 70 nM SDF-1Þ No: of cells=ml ð0 nM SDF-1Þ 100 No: of BM normal PC; B-cell precursors; CD34þ HSC and clonal PC=ml

Of note, migration of cells from HA into the lower chambers in the presence of SDF-1 was significantly higher than when SDF-1 was absent with two to five fold increased migration rates.

Statistical analyses

200.000

Amount of light scatter / clonal PC

150.000

100.000

50.000

***

10.000

2.000

Amount of antigen MFI (mean fluorescence intensity) - expression / clonal PC

The Mann–Whitney U and the Kruskal–Wallis tests were used to evaluate the statistical significance of differences observed

between two or more groups, respectively. Antigenic profiles of BM vs PB clonal PC were compared using the Wilcoxon signed rank test, and correlation studies were performed using the Pearson’s test. For all statistical analyses, the SPSS software (version 15.0; SPSS Inc., Chicago, IL, USA) was used and corrections were applied for multiple comparisons (Bonferroni’s test).

Results

Distribution and immunophenotype of PB vs BM clonal PC in MGUS, SMM and MM As expected, the overall percentage and median number of BM and PB circulating clonal PC found at diagnosis significantly increased from MGUS to SMM and MM patients (Supplementary Table 2). Upon comparing in detail the immunophenotypic features of clonal PC from paired BM vs PB samples in a subset of 13 MGUS, SMM and MM cases, highly similar patterns were found, with a few differences (Figure 1): PB clonal PC showed significantly lower amounts of sideward light scatter (Po0.001), together with lower levels of CD38 (P ¼ 0.001), CD40 (P ¼ 0.006), CD56 (P ¼ 0.03) and CD138 (P ¼ 0.02; Figure 1); by contrast, no significant differences were found regarding forward light scatter and other B-cell markers (for example, CD19, CD20, CD22, CD24, CD27, CD28, CD43, CD45, FMC7 and sIg). Of note, CXCR4 (CD184) was also detected at similar levels in paired BMFmean fluorescence intensity of 1145Fvs PB (mean fluorescence intensity: 1084) clonal PC (r2 ¼ 0.86; Po0.001).

Cell surface expression of CXCR4 on BM clonal PC and PB SDF-1 levels in MGUS, SMM and MM Overall, no significant differences were found between normal BM PC from HA (42%) vs clonal BM PC from neither MGUS (39%), SMM (37%) and MM (35%) patients at diagnosis nor MM POST-HDT/ASCT cases (28%) as regards the percentage of CXCR4+ cells; (Figure 2a). By contrast, mean plasma levels of SDF-1 (Figure 2c) were significantly increased in MM patients

BM clonal PC

200.000

PB clonal PC 150.000

100.000

**

50.000

* * 10.000

* 2.000

0

Immunophenotypic markers

Figure 1 Detailed immunophenotypic features (expressed as MFI; relative mean fluorescence intensity; arbitrary units scaled from 0 to 262, 144) of paired BM vs circulating PB clonal PC from patients with monoclonal gammopathy of undetermined significance (n ¼ 2), smoldering multiple myeloma (n ¼ 2) and symptomatic multiple myeloma (n ¼ 9) studied at diagnosis. *Po0.05, **Pp0.005, ***Po0.001 (Wilcoxon signed rank test). Notched boxes represent 25th and 75th percentile values; the line in the middle and vertical lines correspond to the median value and both the 10th and 90th percentiles, respectively. Leukemia


700

Figure 2 Surface expression of CXCR4 (CD184) by normal BM and normal PB PC from HA aged 460 years vs BM clonal PC and PB normal PC from patients with MGUS, SMM and symptomatic MM studied at diagnosis, and MM POST-HDT/ASCT is shown in (a) and (b), respectively. In (c), SDF-1a levels measured by enzyme-linked immunosorbent assay in the plasma of HA vs patients with MGUS, SMM and MM studied at diagnosis as well as MM POST-HDT/ASCT are displayed. Notched boxes represent 25th and 75th percentile values; the line in the middle and vertical lines correspond to the median value and both the 10th and 90th percentiles, respectively (a). For (c), mean values±s.d. are shown. *P ¼ 0.04 vs HA.

(1778 pg/ml; P ¼ 0.04 vs HA)Fbut not MGUS (1352 pg/ml) or SMM (1656 pg/ml)Freturning to normal levels in MM POST-HDT/ASCT (1331 pg/ml).

Distribution of normal BM B-cell compartments in MGUS, SMM and MM patients The overall median proportion of normal B-cells from the whole BM cellularity was significantly decreased in MM (1.6%; Po0.001) and SMM (2.1%; P ¼ 0.02) patients vs HA (3.7%), whereas it was normal in MGUS (P40.05), even when excluding clonal PC. This was mainly because of the presence of decreased percentages of both the pro-B (1.0 vs 5.0%; P ¼ 0.001) and pre-B (10 vs 24%; P ¼ 0.001) subsets within the overall B-cells in MM, whereas mature B-lymphocytes (87 vs 70%; P ¼ 0.008) were proportionally increased within the overall BM B-cell compartment (Figure 3). Interestingly, the percentage of total B-cells in MM POST-HDT/ASCT BM returned to normal values, but with a clear predominance of pro-B (11 vs 5%; P ¼ 0.01) and pre-B (72 vs 24%; Po0.001) precursors over more mature B-cells (17 vs 70%; Po0.001) within the BM B-cell compartment.

Distribution of normal B-cell compartments and plasmablasts/PC in the PB of MGUS, SMM and MM patients In addition to circulating normal plasmablasts/PC, four other maturation-associated subsets of B cells were systematically identified in the PB of MGUS (n ¼ 33), SMM (n ¼ 26) and MM Leukemia

(n ¼ 57) patients at diagnosis (Table 1). Upon considering both the absolute and relative numbers of all PB B cells, no significant differences were found between HA vs MGUS, SMM and MM patients (Table 1). However, a more detailed analysis of the composition of the overall PB B-cell compartment showed significantly lower percentages of naive B-lymphocytes (Pp0.03) in both MM and SMM vs HA (Table 1). Conversely, non-switched memory B cells were increased (P ¼ 0.04) within the PB B-cell compartment in MM and SMM vs HA cases. Despite the overall percentage of normal PC in the BM was significantly lower among MM and SMM vs MGUS cases (Po0.001; Supplementary Figure 1), the distribution of circulating PB normal PC was similar among groups (P40.05 vs HA). Interestingly, however, although the median proportion of CXCR4 þ normal PB PC (Figure 2b) was similar between HA (11%), MGUS (14%) and SMM (15%), it was increased in MM (21%; P ¼ 0.05); MM POST-HDT/ASCT cases showed the lowest median percentage of CXCR4 þ PB normal PC (5%).

BM and circulating PB CD34 þ HSC from patients with MGUS, SMM and MM

Overall, the median percentage of CD34 þ HSC from the whole BM cellularity was significantly decreased in MM (0.3%; P ¼ 0.001) and SMM (0.4%; P ¼ 0.002) patients vs HA (0.9%) and MGUS (0.8%), even upon excluding clonal PC (Figure 4a). Interestingly, however, using the overall compartment of BM CD34 þ HSC as denominator, the median percentage of immature CD34 þ HSC was significantly increased in MM (39%; Po0.001) and SMM (28%; Po0.001) vs HA (20%);


701 Total B-cells

BM B-cell subsets +

Pro-B / CD19 BM cells

+

Pre-B / CD19 BM cells

+

Mature / CD19 BM cells

*

100%

80%

60%

+

10%

% of cells / CD19 BM B-cells

% of CD19+ BM cells / all BM cells

***

5%

*

***

40%

*

20%

***

* **

0%

0% HA (n = 20) MGUS (n = 27) SMM (n = 21) MM (n = 30) MM POST-HDT/ASCT (n = 12)

HA (n = 20) MGUS (n = 27) SMM (n = 21) MM (n = 30) MM POST-HDT/ASCT (n = 12)



Figure 3 Distribution of B cells and their CD19 þ /CD34 þ /CD38 þ (pro-B), CD19 þ /CD34/CD38 þ (pre-B) and CD19 þ /CD20 þ /CD38 (mature) subsets (within the B-cell compartment) in the BM of HA aged 460 years vs patients with MGUS, SMM and symptomatic MM at diagnosis, as well as MM POST-HDT/ASCT. In this figure, CD38hi/CD19 þ plasma cells were excluded from the analysis. *Po0.05, **Pp0.005, ***Po0.001 vs HA (Mann–Whitney U-test).

Table 1 Distribution of normal PB B-cell and plasmablast/PC subsets in HA aged 460 years (HA460 years) vs patients with MGUS, SMM and symptomatic MM studied at diagnosis HA460 years (N ¼ 16) No. of PB B-cells/ml Immaturea Naivea Memory non-switcheda Memory switcheda No. of PB normal PC/ml Percentage of PB B-cells Immatureb Naiveb Memory non-switchedb Memory switchedb Percentage of PB normal PC

82 2.6 60 12 13 0.6 1.8 2.7 74 11 11 0.01

MGUS (N ¼ 33)

SMM (N ¼ 26)

MM (N ¼ 57)

P-value

(29–307) (0.2–12) (23–226) (1.1–94) (3.0–69) (0.1–1.5)

82 1.5 56 5.4 13 0.6

(9.5–250) (0–22) (4.1–170) (1.3–109) (2.1–69) (0.1–2.3)

60 1.2 34 13 10 0.4

(3.2–200) (0–8.3) (0.5–150) (0.6–74) (1.7–52) (o0.06–4.0)

84 1.0 44 12 13 0.5

(6.6–485) (0–26) (1.8–340) (0.4–100) (2.3–89) (o0.06–11)

NS 0.02 0.05 0.07 NS NS

(0.6–6.4) (0.3–5.5) (43–88) (2.5–31) (4.0–31) (33–0.02)

1.7 2.0 66 10 19 93

(0.2–5.0) (0–11) (28–95) (1.5–48) (2.8–46) (13–0.05)

1.5 1.9 51 21 23 93

(0.2–4.5) (0–6.0) (17–75)** (8.0–56)* (6.3–61) (o13–0.08)

1.6 1.3 57 19 18 0.01

(0.2–6.0) (0–9.8) (14–87)* (2.0–64)* (3.8–52) (o13–0.2)

NS 0.04 0.003 0.005 NS NS

Abbreviations: HA, healthy adults aged 460 years; MGUS, monoclonal gammopathy of undetermined significance; MM, multiple myeloma; M-PC, (mono)clonal/aberrant plasma cells; NS, statistically no significant differences were detected; PB, peripheral blood; PC, plasma cells; SMM, smoldering multiple myeloma. Results expressed as median and range between brackets. *Po0.05, **Pp0.005 vs HA460 years (Bonferroni’s test) a Absolute count (no. of cells/ml). b Percentage of cells using all PB B-cells as the denominator.

by contrast, the proportion of myeloid precursors was similar among all groups, whereas the median percentage of lymphoid precursors was lower in MM (3%; Po0.001) and SMM (13%; P ¼ 0.04) vs HA (16%). Interestingly, MM POST-HDT/ASCT patients showed normal numbers of CD34 þ HSC in their BM (1.1 vs 0.9% in HA; P40.05); however, a more detailed analysis of the different subsets of CD34 þ HSC in these latter patients showed a marked reduction of the more immature precursors (Po0.001) at the expense of an increased number of lymphoid precursors (Po0.001) within the overall compartment of CD34 þ HSC. In a subgroup of cases, we further investigated the distribution of circulating PB CD34 þ HSC in paired BM and PB samples

(Figure 4b). Our results show that both the relative and absolute number of PB CD34 þ HSC are normal in MGUS and SMM, but increased in MM (P ¼ 0.008 and P ¼ 0.06, respectively) when compared with HA. By contrast, in MM POST-HDT/ASCT patients the number of PB CD34 þ HSC returned to normal/ lower levels. Of note, these results suggest that the distribution of BM and PB CD34 þ HSC are progressively altered from MGUS to MM at diagnosis with an inverse behavior; lower levels of BM CD34 þ HSC are paralleled by increased counts of circulating PB CD34 þ HSC. Accordingly, the median percentage of BM non-lymphoid CD34 þ HSC (Supplementary Figure 2) in MM patients with anemia was significantly decreased vs HA (0.24 vs 0.63%; P ¼ 0.04), whereas MM patients with Leukemia


702

Figure 4 Distribution of CD34 þ HSC and their CD34 þ /CD38/ þ dim (immature), CD34 þ /CD38 þ /CD19 (myeloid), and CD34 þ /CD38 þ / CD19 þ (lymphoid) subsets (within the CD34 þ HSC compartment) in the BM (a) and PB (b) of HA aged 460 years vs patients with MGUS, SMM and symptomatic MM at diagnosis, as well as MM POST-HDT/ASCT. *Po0.05, **Pp0.005, ***Po0.001 vs HA (Mann–Whitney U-test).

hemoglobin levels X120 g/l showed a normal percentage of non-lymphoid CD34 þ HSC (0.77 vs 0.63%).

Relationship between different subsets of PB and BM cells competing for potentially overlapping BM niches Upon comparing the distribution of different cell populationsFCD34 þ HSC, pre-B precursors and PC (both normal and clonal)Fpotentially competing for SDF-1-associated BM niches in paired BM and PB samples from HA (n ¼ 6) and patients with MGUS (n ¼ 12), SMM (n ¼ 6) and MM (n ¼ 17) studied at diagnosis, as well as MM POST-HDT/ASCT (n ¼ 12; Supplementary Figure 3), a significant correlation (r2 ¼ 0.42; Po0.001) was found between the percentage of these cell populations in the BM (CD34 þ HSC plus pre-B cells and PC) and the corresponding circulating populations in PB (CD34 þ HSC and PC). Interestingly, HA grouped together with MGUS cases, whereas SMM scattered between HA/MGUS and MM patients; MM POST-HDT/ASCT showed features similar to HA/MGUS cases (Supplementary Figure 3). Despite these Leukemia

findings, it should be noted that while MM patients with detectable clonal PC in PB (Supplementary Table 3) showed significantly higher numbers of BM clonal PC vs cases who had no circulating PB clonal PC (median of 2.2 vs 11% clonal PC from the whole BM cellularity, respectively; P ¼ 0.001), among MGUS and SMM patients no differences were found or they were less pronounced.

‘Ex vivo’ competition between BM B cells, PC and CD34 þ HSC for SDF-1-induced migration

In HA, both the immature and myeloid subsets of CD34 þ HSC showed the highest migration potential in the presence of SDF-1, followed by pre-B cell precursors; conversely, normal PC barely migrated (for example, medians for an SDF-1 concentration of 30 nM of 6, 22, 4 and 0.01%, respectively; Figure 5a). No significant differences were found for the migration of all cell populations analyzed between MGUS and SMM patients vs HA, except for clonal PC that showed an impaired migration in the presence of SDF-1 for the two concentrations used (Figure 5).


703

Figure 5 Transwell migration assays performed on isolated BM mononuclear cells from HA (n ¼ 4) vs patients with MGUS (n ¼ 4), SMM (n ¼ 4) and symptomatic MM (n ¼ 4) studied at diagnosis. The migration ofFCD34 þ HSC CD34 þ /CD38/ þ dim (immature), CD34 þ /CD38 þ /CD19 (myeloid) and CD34 þ /CD38 þ /CD19 þ (lymphoid) subsetsFB-cell precursors CD19 þ /CD34 þ /CD38 þ (pro-B) and CD19 þ /CD34/CD38 þ (pre-B)Fnormal PC and clonal PC was triggered with SDF-1 concentrations of 30 and 70 nM (a and b, respectively). Results are expressed in terms of migration index (percentage of cells that migrate in the presence of SDF-1 after subtracting the percentage of cells that migrate in the absence of SDF-1); bars represent median values and vertical lines the upper bound of the 95% confidence intervals. *Po0.05 vs HA (Mann–Whitney U-test).

In turn, the migration potential of both the myeloid (P ¼ 0.04) and lymphoid (P ¼ 0.02) subsets of CD34 þ HSC, and pre-B cell precursors (P ¼ 0.02) was markedly reduced in symptomatic MM, particularly at lower SDF-1 concentrations (Figure 5a). Most interestingly, the migration of clonal PC from symptomatic MM was markedly increased at both concentrations of SDF-1 used: median of 2.9 and 1.0% for SDF-1 concentrations of 30 and 70 nM, respectively.

Discussion It has been shown that the proportion of residual normal/ polyclonal BM PC (from all BM PC) can be of value for the discrimination between MGUS and MM.20 More recently, this parameter has also proved to be predictive for the progression of MGUS and SMM to symptomatic myeloma,21 as well as for disease outcome in MM.22 Despite its clinical impact, Leukemia


704 the potential mechanisms leading to such clinical associations remain largely unknown and only an association with immune paresis has been reported so far.21,22 Serum levels of polyclonal Igs are typically depressed in MM 5,22 and, less frequently, also in MGUS and SMM;21,41 this has been related to the depletion of normal PC in the patients’ BM. However, it should be noted that in MM, both parametersFimmune paresis and the proportion of normal PC from all BM PCFappear to be independently associated with patient outcome.22 In this study, we confirm the existence of progressively decreasing numbers of normal BM PC from MGUS to SMM and MM patients, in parallel to an increase in both BM and PB clonal PC. In order to become long-living PC, recently produced normal PC (and also clonal PC) need to be in close contact with SDF-1secreting stromal cell niches in the BM. However, the number of such BM niches is finite23 and functionally impaired in the elderly.28,33 Thus, the progressively decreased numbers of normal PC in the BM of MM patients could be because of an increased competition with the malignant clone for the BM PC niches. Interestingly, it has been suggested that adipocytes (whose numbers progressively increase in the BM microenvironment of elderly individuals) have equivalent functions to BM stromal cells in supporting clonal PC.35 In such case, the decreased number of normal BM PC could be due, not to a decreased PC production, but to limited homing of recently produced PB PC into the BM because BM niches are overloaded with clonal PC. In line with this hypothesis, our results show that MM patients display normal numbers of normal PC in PB, despite undetectable in the BM of most patients. Of note, normal PB PC from patients with MM (but not MGUS and SMM) showed increased expression of the CXCR4 chemokine (SDF-1) receptor. As expression of CXCR4 by PC is considered to be a marker for predominant BM homing,23,42 these findings further support the notion that in symptomatic MM, the fraction of normal PB PC that should home to the BM remains in the circulation and accumulates in the PB of these patients. Usually, PB clonal PC have been viewed as cells that leave the BM because of unique migration properties acquired at relatively advanced disease stages, particularly when clonal PC decrease their BM microenvironment dependency.10 Alternatively, our results could also indicate that in order to spread the diseaseFlike normal BM PC23,32,42Fclonal PC could also leave the BM, recirculate into PB and home again into the BM at a different localization, in a kind of ‘metastatic’/dissemination process. In line with this hypothesis, in this and other studies7–14,43 circulating clonal PC were detected in a significant proportion of MGUS and SMM cases in addition to MM. A potential explanation for the recirculation of clonal PC in the early stages of the disease (for example, MGUS) could rely on the advanced age of MGUS patients, which could lead to lower numbers of available BM niches.44 However, we failed to show any correlation between the presence of clonal PC in PB and patients’ age. In turn, clonal PC from paired samples showed a slightly more immature immunophenotypic profile for PB vs BM clonal PC (for example, lower internal complexity and lower expression of CD38, CD138 and CD40). Clonotypic cells with a more immature phenotype have been repeatedly found in the PB of patients with monoclonal gammopathies,12,45–47 and decreased CD138 expression by circulating PB clonal PC could also indicate they are enriched in clonogenic cells, in line with the previously reported immunophenotypic properties of clonal PC showing stem cell properties.46,48 In the early steps of B-cell maturation, B-cell precursors also adhere to SDF-1-expressing reticular cells25 and SDF-1 also has an important role in the colonization of the BM by circulating Leukemia

CD34 þ HSC. In addition, it has been demonstrated in mice that pre-pro-B cells and PC compete for the same BM niche.25 Therefore, it could be hypothesized that such PC competition for BM niches could also extend to B-cell precursors and CD34 þ HSC which use the SDF-1-CXCR4 axis in common. In line with this hypothesis we found here that in parallel to normal PC, the distribution of different B-cell subsets and CD34 þ HSC is also dysregulated in patients with monoclonal gammopathies. Interestingly, the frequency of BM pro-B and pre-B cell precursors and CD34 þ HSC in MGUS was similar to that of HA, whereas MM showed a progressively lower number of these cell populations from smoldering to symptomatic disease. These findings would further support the model of cell competition,25 where clonal PC progressively replace B-cell precursors from their BM niches, in addition to normal PC. Whether the decrease in pre-B cells is due to an intrinsically impaired B-cell production or to direct competition with clonal PC (similarly to what has been suggested for normal PC26,27), remains to be elucidated. In addition to normal PC and pre-B cells, total CD34 þ HSC were also decreased in the BM of SMM and MM patients, at the expense of a decreased number of lymphoid progenitors (which could be further responsible for the lower number of PB immature and naive B-lymphocytes found among MM patients). By contrast, the distribution of immature and myeloid CD34 þ HPC was not significantly altered in MGUS. Based on these data, it could be speculated that during progression from premalignant conditions (for example, MGUS and SMM) to symptomatic MM, there is a derailment in the tightly regulated and balanced competition between clonal PC and different normal cell populations (normal PC, B-cell precursors and CD34 þ HSC) for a limited number of BM niches. In line with this hypothesis, MM patients also showed increased numbers of circulating PB CD34 þ HSC and a significant correlation was observed between PB and BM cell populations (including clonal PC) potentially competing for the same BM niches; both values (and thus also the competition between them) increased from HA to MGUS, SMM and MM; this was followed by a decrease in MM POST-HDT/ASCT cases, which could be explained by niche emptiness after therapy. It would be interesting to further investigate whether this model of cell competition correlates with the malignant transformation of MGUS and progression of ‘evolving’ MGUS cases as well as MM patients’ response after therapy; herein, MM patients achieving complete response after HDT/ASCT showed slightly (non-significant) increased numbers of normal PC, total B-cells and CD34 þ HSC vs cases in partial response (data not shown). Our results in a new model of competitive ex vivo migration assay between normal BM cells vs clonal PC in the presence of variable concentrations of SDF-1 would further support the hypothesis of cell competition, in which normal PC, B-cell precursors and CD34 þ HSC from MGUS and SMM, but not MM patients, were able to abrogate the migration of clonal PC, particularly at the lower concentrations evaluated. Despite all associations described above, it should be noted that a relatively wide variation was found within MGUS, SMM and MM patients for the distribution of some cell populations, with overlapping numbers between HA and MGUS, MGUS and SMM and between SMM and MM. In such cases, the specific genetic background and other relevant features of tumor cells may determinate more closely the behavior of the disease. Further studies are necessary to confirm this hypothesis. In summary, our results suggest that progressive competition and replacement of normal BM PC, B-cell precursors and CD34 þ HSC by clonal PC in BM niches from patients with


705 monoclonal gammopathies is associated with more advanced forms of the disease. The exact composition and distribution of such BM niches and the causes leading to the potential competing advantage of clonal PC vs normal cells (for example, specific cytogenetic or epigenetic abnormalities) for such niches deserve further investigations.

14 15 16

Conflict of interest The authors declare no conflict of interest.

Acknowledgements This work was supported by the Cooperative Research Thematic Cancer Network (RTICs; RD06/0020/0006, RD06/0020/0035 and G03/136), MM Jevitt, SL firm, Instituto de Salud Carlos III/ Subdireccioń General de Investigacioń Sanitaria (FIS: PI060339; 02/0905; 01/0089/01-02; PS09/01897), Consejerıá de Sanidad and Conserjerıá de Educacion (GR37), Junta de Castilla y Leoń, Valladolid, Spain (557/A/10) and from MSCNET European strep (N1E06005FF).

References 1 Kyle R, Therneau T, Rajkumar S, Larson D, Plevak M, Offord J et al. Prevalence of monoclonal gammopathy of undetermined significance. N Engl J Med 2006; 354: 1362–1369. 2 Weber D, Dimopoulos M, Moulopoulos L, Delasalle K, Smith T, Alexanian R. Prognostic features of asymptomatic multiple myeloma. Br J Haematol 1997; 97: 810–814. 3 Kyle R, Therneau T, Rajkumar S, Offord J, Larson D, Plevak M et al. A long-term study of prognosis in monoclonal gammopathy of undetermined significance. N Engl J Med 2002; 346: 564–569. 4 Cesana C, Klersy C, Barbarano L, Nosari A, Crugnola M, Pungolino E et al. Prognostic factors for malignant transformation in monoclonal gammopathy of undetermined significance and smoldering multiple myeloma. J Clin Oncol 2002; 20: 1625–1634. 5 San Miguel JF, Gutierrez NC, Mateo G, Orfao A. Conventional diagnostics in multiple myeloma. Eur J Cancer 2006; 42: 1510–1519. 6 Kyle RA, Rajkumar SV. Multiple myeloma. N Engl J Med 2004; 351: 1860–1873. 7 Luque R, Brieva J, Moreno A, Manzanal A, Escribano L, Villarrubia J et al. Normal and clonal B lineage cells can be distinguished by their differential expression of B cell antigens and adhesion molecules in peripheral blood from multiple myeloma (MM) patients–diagnostic and clinical implications. Clin Exp Immunol 1998; 112: 410–418. 8 Chandesris MO, Soulier J, Labaume S, Crinquette A, Repellini L, Chemin K et al. Detection and follow-up of fibroblast growth factor receptor 3 expression on bone marrow and circulating plasma cells by flow cytometry in patients with t(4;14) multiple myeloma. Br J Haematol 2007; 136: 609–614. 9 Rawstron AC, Owen RG, Davies FE, Johnson RJ, Jones RA, Richards SJ et al. Circulating plasma cells in multiple myeloma: characterization and correlation with disease stage. Br J Haematol 1997; 97: 46–55. 10 Nowakowski GS, Witzig TE, Dingli D, Tracz MJ, Gertz MA, Lacy MQ et al. Circulating plasma cells detected by flow cytometry as a predictor of survival in 302 patients with newly diagnosed multiple myeloma. Blood 2005; 106: 2276–2279. 11 Schneider U, van LA, Huhn D, Serke S. Two subsets of peripheral blood plasma cells defined by differential expression of CD45 antigen. Br J Haematol 1997; 97: 56–64. 12 Billadeau D, Van NB, Kimlinger T, Kyle RA, Therneau TM, Greipp PR et al. Clonal circulating cells are common in plasma cell proliferative disorders: a comparison of monoclonal gammopathy of undetermined significance, smoldering multiple myeloma, and active myeloma. Blood 1996; 88: 289–296. 13 Kumar S, Rajkumar S, Kyle R, Lacy M, Dispenzieri A, Fonseca R et al. Prognostic value of circulating plasma cells in monoclonal

17 18

19

20

21

22

23

24 25 26

27 28

29

30

31

32

gammopathy of undetermined significance. J Clin Oncol 2005; 23: 5668–5674. Witzig T, Kyle R, O’Fallon W, Greipp P. Detection of peripheral blood plasma cells as a predictor of disease course in patients with smouldering multiple myeloma. Br J Haematol 1994; 87: 266–272. Kuehl WM, Bergsagel PL. Multiple myeloma: evolving genetic events and host interactions. Nat Rev Cancer 2002; 2: 175–187. Fonseca R, Barlogie B, Bataille R, Bastard C, Bergsagel PL, Chesi M et al. Genetics and cytogenetics of multiple myeloma: a workshop report. Cancer Res 2004; 64: 1546–1558. Fonseca R, Bailey R, Ahmann G, Rajkumar S, Hoyer J, Lust J et al. Genomic abnormalities in monoclonal gammopathy of undetermined significance. Blood 2002; 100: 1417–1424. Paiva B, Almeida J, Perez-Andres M, Mateo G, Lopez A, Rasillo A et al. Utility of flow cytometry immunophenotyping in multiple myeloma and other clonal plasma cell-related disorders. Cytometry B Clin Cytom 2010; 78: 239–252. Perez-Andres M, Almeida J, Martin-Ayuso M, Moro MJ, MartinNunez G, Galende J et al. Clonal plasma cells from monoclonal gammopathy of undetermined significance, multiple myeloma and plasma cell leukemia show different expression profiles of molecules involved in the interaction with the immunological bone marrow microenvironment. Leukemia 2005; 19: 449–455. Ocqueteau M, Orfao A, Almeida J, Blade J, Gonzalez M, GarciaSanz R et al. Immunophenotypic characterization of plasma cells from monoclonal gammopathy of undetermined significance patients. Implications for the differential diagnosis between MGUS and multiple myeloma. Am J Pathol 1998; 152: 1655–1665. Perez-Persona E, Vidriales MB, Mateo G, Garcia-Sanz R, Mateos MV, de Coca AG et al. New criteria to identify risk of progression in monoclonal gammopathy of uncertain significance and smoldering multiple myeloma based on multiparameter flow cytometry analysis of bone marrow plasma cells. Blood 2007; 110: 2586–2592. Paiva B, Vidriales MB, Mateo G, Perez JJ, Montalban MA, Sureda A et al. The persistence of immunophenotypically normal residual bone marrow plasma cells at diagnosis identifies a good prognostic subgroup of symptomatic multiple myeloma patients. Blood 2009; 114: 4369–4372. Radbruch A, Muehlinghaus G, Luger EO, Inamine A, Smith KG, Dorner T et al. Competence and competition: the challenge of becoming a long-lived plasma cell. Nat Rev Immunol 2006; 6: 741–750. Shapiro-Shelef M, Calame K. Regulation of plasma-cell development. Nat Rev Immunol 2005; 5: 230–242. Tokoyoda K, Egawa T, Sugiyama T, Choi BI, Nagasawa T. Cellular niches controlling B lymphocyte behavior within bone marrow during development. Immunity 2004; 20: 707–718. Ma Q, Jones D, Springer TA. The chemokine receptor CXCR4 is required for the retention of B lineage and granulocytic precursors within the bone marrow microenvironment. Immunity 1999; 10: 463–471. Nie Y, Waite J, Brewer F, Sunshine MJ, Littman DR, Zou YR. The role of CXCR4 in maintaining peripheral B cell compartments and humoral immunity. J Exp Med 2004; 200: 1145–1156. Perez-Andres M, Paiva B, Nieto WG, Caraux A, Schmitz A, Almeida J et al. Human peripheral blood B-cell compartments: a crossroad in B-cell traffic. Cytometry B Clin Cytom 2010; 78 (Suppl 1): S47–S60. Foudi A, Jarrier P, Zhang Y, Wittner M, Geay JF, Lecluse Y et al. Reduced retention of radioprotective hematopoietic cells within the bone marrow microenvironment in CXCR4/ chimeric mice. Blood 2006; 107: 2243–2251. Sugiyama T, Kohara H, Noda M, Nagasawa T. Maintenance of the hematopoietic stem cell pool by CXCL12-CXCR4 chemokine signaling in bone marrow stromal cell niches. Immunity 2006; 25: 977–988. Serke S, Sauberlich S, Huhn D. Multiparameter flow-cytometrical quantitation of circulating CD34(+)-cells: correlation to the quantitation of circulating haemopoietic progenitor cells by in vitro colony-assay. Br J Haematol 1991; 77: 453–459. Caraux A, Klein B, Paiva B, Bret C, Schmitz A, Fuhler GM et al. Circulating human B and plasma cells. Age-associated changes in counts and detailed characterization of circulating normal CD138 and CD138+ plasma cells. Haematologica 2010; 95: 1016–1020. Leukemia


706 33 Mayack SR, Shadrach JL, Kim FS, Wagers AJ. Systemic signals regulate ageing and rejuvenation of blood stem cell niches. Nature 2010; 463: 495–500. 34 Corre J, Planat-Benard V, Corberand JX, Peńicaud L, Casteilla L, Laharrague P. Human bone marrow adipocytes support complete myeloid and lymphoid differentiation from human CD34 cells. Br J Haematol 2004; 127: 344–347. 35 Caers J, Deleu S, Belaid Z, De Raeve H, Van Valckenborgh E, De Bruyne E et al. Neighboring adipocytes participate in the bone marrow microenvironment of multiple myeloma cells. Leukemia 2007; 21: 1580–1584. 36 International Myeloma Working Group. Criteria for the classification of monoclonal gammopathies, multiple myeloma and related disorders: a report of the International Myeloma Working Group. Br J Haematol 2003; 121: 749–757. 37 Nieto WG, Almeida J, Romero A, Teodosio C, Lopez A, Henriques AF et al. Increased frequency (12%) of circulating chronic lymphocytic leukemia-like B-cell clones in healthy subjects using a highly sensitive multicolor flow cytometry approach. Blood 2009; 114: 33–37. 38 Paiva B, Vidriales MB, Cervero J, Mateo G, Perez JJ, Montalban MA et al. Multiparameter flow cytometric remission is the most relevant prognostic factor for multiple myeloma patients who undergo autologous stem cell transplantation. Blood 2008; 112: 4017–4023. 39 Matarraz S, Lopez A, Barrena S, Fernandez C, Jensen E, Flores J et al. The immunophenotype of different immature, myeloid and B-cell lineage-committed CD34+ hematopoietic cells allows discrimination between normal/reactive and myelodysplastic syndrome precursors. Leukemia 2008; 22: 1175–1183. 40 van Lochem EG, van der Velden VH, Wind HK, te Marvelde JG, Westerdaal NA, van Dongen JJ. Immunophenotypic differentiation

41 42

43

44 45

46 47

48

patterns of normal hematopoiesis in human bone marrow: reference patterns for age-related changes and disease-induced shifts. Cytometry B Clin Cytom 2004; 60: 1–13. Kyle RA, Rajkumar SV. Epidemiology of the plasma-cell disorders. Best Pract Res Clin Haematol 2007; 20: 637–664. Mei HE, Yoshida T, Sime W, Hiepe F, Thiele K, Manz RA et al. Blood-borne human plasma cells in steady state are derived from mucosal immune responses. Blood 2009; 113: 2461–2469. Rawstron A, Barrans S, Blythe D, Davies F, English A, Pratt G et al. Distribution of myeloma plasma cells in peripheral blood and bone marrow correlates with CD56 expression. Br J Haematol 1999; 104: 138–143. Frasca D, Landin A, Riley R, Blomberg B. Mechanisms for decreased function of B cells in aged mice and humans. J Immunol 2008; 180: 2741–2746. Isaksson E, Bjorkholm M, Holm G, Johansson B, Nilsson B, Mellstedt H et al. Blood clonal B-cell excess in patients with monoclonal gammopathy of undetermined significance (MGUS): association with malignant transformation. Br J Haematol 1996; 92: 71–76. Matsui W, Huff CA, Wang Q, Malehorn MT, Barber J, Tanhehco Y et al. Characterization of clonogenic multiple myeloma cells. Blood 2004; 103: 2332–2336. Pilarski LM, Hipperson G, Seeberger K, Pruski E, Coupland RW, Belch AR. Myeloma progenitors in the blood of patients with aggressive or minimal disease: engraftment and self-renewal of primary human myeloma in the bone marrow of NOD SCID mice. Blood 2000; 95: 1056–1065. Matsui W, Wang Q, Barber JP, Brennan S, Smith BD, Borrello I et al. Clonogenic multiple myeloma progenitors, stem cell properties, and drug resistance. Cancer Res 2008; 68: 190–197.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Leukemia