Competition Between Clonal Plasma Cells and Normal Cells for Potentially

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,MPe´rez-Andre´s2,3, M-B Vı´driales1,2, J Almeida2,3, N de las Heras4, M-V Mateos1,2,LLo´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)

1Servicio 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; 5Servicio 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; 13Servicio 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) progressionFE1 and 10% per year for monoclonal gammo- predicts malignant transformation of monoclonal gammopathy pathy of undetermined significance (MGUS) and smoldering of undetermined significance (MGUS) and smoldering myeloma myeloma (SMM), respectively1–4Fto symptomatic multiple (SMM) into symptomatic multiple myeloma (MM). The homing, þ myeloma (MM) with median overall survival rates of 3–7 behavior and survival of normal PC, but also CD34 hemato- 5,6 7–14 poietic stem cells (HSC), B-cell precursors, and clonal PC years. Despite the different tumor mass and clinical largely depends on their interaction with stromal cell-derived behavior of the disease, clonal PC from MGUS, SMM and MM factor-1 (SDF-1) expressing, potentially overlapping BM stromal patients show highly similar and largely overlapping genetic cell niches. Here, we investigate the distribution, phenotypic profiles.15–17 In addition, no clear phenotypic differences have characteristics and competitive migration capacity of these cell been reported so far among clonal PC from MGUS, SMM and populations in patients with MGUS, SMM and MM vs healthy MM,18 except for a few molecules involved in the interaction adults (HA) aged 460 years. Our results show that BM and 19 peripheral blood (PB) clonal PC progressively increase from between PC and their microenvironment. MGUS to MM, the latter showing a slightly more immature We have previously shown that the proportion of normal PC immunophenotype. Of note, such increased number of clonal within the bone marrow (BM) PC compartment (normal PC/BM PC is associated with progressive depletion of normal PC, þ PC) is an efficient single parameter for discrimination between B-cell precursors and CD34 HSC in the BM, also with a parallel MGUS and MM;20 at diagnosis, most MGUS cases (480%) increase in PB. In an ex vivo model, normal PC, B-cell 21 þ display 45% normal PC/BM PC, whereas the great majority precursors and CD34 HSC from MGUS and SMM, but not 22 MM patients, were able to abrogate the migration of clonal PC (485%) of symptomatic MM show o5% normal PC/BM PC. into serial concentrations of SDF-1. Overall, our results show Additionally, the presence of 45% normal PC/BM PC at that progressive competition and replacement of normal BM diagnosis is associated with both a lower risk of progression of cells by clonal PC is associated with more advanced disease in MGUS and SMM,21 and a favorable outcome in MM.22 In turn, a patients with MGUS, SMM and MM. progressive decrease in the serum levels of uninvolved Leukemia (2011) 25, 697–706; doi:10.1038/leu.2010.320; immunoglobulins (Ig) is observed from MGUS to SMM and published online 21 January 2011 21,22 Keywords: monoclonal gammopathies; malignant transformation; MM. Altogether, these results suggest that disappearance of bone marrow niche competition; plasma cells 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 Introduction largely depend on Ig production and secretion by long-living BM and mucosa-associated lymphoid tissue-derived PC.23,24 Plasma cell (PC) disorders include an heterogeneous spectrum Recently produced normal PC, which have left secondary of diseases, from pre-malignant conditions with variable rates of lymphoid tissues through peripheral blood (PB), usually need to migrate into BM niches and adhere to CXCL12 (SDF-1)- Correspondence: Professor A Orfao, Centro de Investigacioń del expressing BM stromal cells to become long-living BM ´ Cancer, Avda. Universidad de Coimbra S/N (Campos Miguel de PC.23,24 SDF-1-expressing BM stromal cells are also crucial for Unamuno), Salamanca 37007, Spain. 25 F E-mail: [email protected] early pro-B and together with interleukin-7-expressing 25 26,27 Received 21 October 2010; revised 30 November 2010; accepted cells Fpotentially also for later pre-B-cell development 10 December 2010; published online 21 January 2011 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 were studied, with paired BM and PB samples available in number of BM PC niches, normal PC as well as normal B-cell 6 of these latter 20 HA analyzed. For BM control samples, no precursors25 could compete with clonal PC for BM PC niches; in significant differences in their compositions were observed turn, CXCL12-CXCR4 signaling appears to be also a key between subjects who were younger and those older than component of the CD34 þ hematopoietic stem cell (HSC) 55 years (Supplementary Table 1). niche.29,30 Interestingly, recent results show that in normal All control and patient samples were collected after informed individuals, both CD34 þ HSC and PC recirculate between consent was given by each individual, according to the local the BM and the PB.23,31,32 Accordingly, it can be hypothesized ethical committees and the Helsinki Declaration. that BM B-cell precursors, CD34 þ HSC and PC are under a tight and continuous regulation by the SDF-1/CXCR4 axis, Multiparameter flow cytometry immunophenotypic with a certain degree of competition among them, for the studies same BM niches. In this competition, an advantage could be Approximately 1 ml of EDTA-anti-coagulated PB/case was expected for the malignant clone based on its numerical immunophenotyped using a direct eight-color immunofluores- expansion. An additional factor to be considered is the cence stain-and-then-lyse technique,32,37 with the following advanced age of these patients as SDF-1-expressing BM niches combination of monoclonal antibodies (Pacific blue (PB)/ undergo age-induced impairment in their ability to support anemonia majano cyan (AmCyan)/fluorescein isothiocyanate/ 33 normal hematopoiesis, which could be partly explained by a phycoerythrin (PE)/peridinin chlorophyll protein–cyanin 5.5 progressive increase in the number of adipocytic deposits in (PerCP–Cy5.5)/PE–cyanin 7 (PE–Cy7)/allophycocyanin/alexa- 34,35 the BM. Thus, progressive replacement of normal precursors fluor 700 (AF700)): CD20/CD45/surface IgM (sIgM) or sIgl/ and normal PC by the malignant clone could contribute to sIgG, sIgA or sIgk/CD19/CD10/CD27/CD38. In all PB samples explain the occurrence of cytopenias and hypogammaglobuli- analyzed, the following maturation-associated B-cell subsets nemia in MM patients. were identified32 in addition to circulating plasmablasts/PC In this study, we analyze the distribution and competitive (CD10À/CD27 þþ/CD38 þþ): (i) immature (CD10 þ /CD27À/ þ migrating capacity of B-cell precursors, CD34 HSC, normal CD38 þ ), (ii) naive (CD10À/CD27À/CD38À), (iii) memory non- PC and clonal PC in the BM and PB of patients with PC disorders switched (CD10À/CD27 þ /CD38À/IgM þ ) and (iv) memory vs healthy adults (HA) aged 460 years. Taken together, our switched (CD10À/CD27 þ /CD38À/IgMÀ) B-cells. Because of results show that progressive competition and replacement of their different patterns of expression of CD38, CD19 and normal BM cells by clonal PC is associated with the malignant CD45, and their unique light scatter characteristics,22,38 this transformation of MGUS and SMM and progression of sympto- multiparameter eight-color staining also allowed detection of matic MM. 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 Materials and methods characterization of PB vs BM clonal PC was performed using the following antibody combinations: (i) CD19/CD45/sIgl/sIgG Patients, controls and samples plus sIgA/CD138/CD27/sIgk/CD38; (ii) CD19/CD45/CD20/CD22/ Overall a total of 206 patients with clonal PC disorders were CD138/CD27/CD56/CD38; (iii) CD19/CD45/FMC7/CD24/CD138/ prospectively studied. These included newly diagnosed MGUS CD27/CD43/CD38 and (iv) CD19/CD45/CD40/CXCR4/CD138/ (n ¼ 60), SMM (n ¼ 47) and MM (n ¼ 87) patients, plus 12 MM CD27/CD28/CD38. cases studied at day þ 100 after high-dose therapy/autologous Erythrocyte-lysed whole-BM samples (and also their paired PB stem cell transplantation (MM POST-HDT/ASCT). In all cases, samples) were immunophenotyped using a single eight-color diagnosis was established according to the International antibody combinationFCD20/CD45/CD34/CXCR4/CD19/CD10/ 36 Myeloma Working Group criteria. Patients with SMM had CD27/CD38Faimed at the identification and enumeration of a high-risk of progression to symptomatic disease with X2 the following BM cell populations: (i) CD34 þ precursors and of the following three criteria at diagnosis: BM PCX10%; their CD34 þ /CD38À/ þ dim (immature), CD34 þ /CD38 þ /CD19À M-component IgGX30 g/l or IgAX20 g/l or Bence–Jones protein (myeloid) and CD34 þ /CD38 þ /CD19 þ (lymphoid) subsets; 410 g/l, and X95% clonal PC from all BM PC and immune (ii) B cells and their CD19 þ /CD34 þ /CD38 þ (Pro-B), CD19 þ / paresis. Patients’ relatively short follow-up precluded survival CD34À/CD38 þ (Pre-B) and CD19 þ /CD20 þ /CD38À (mature) analysis in the present study. precursor B-cell subsets and (iii) BM PC, including normal The study was conducted on PB and BM samples collected PC (CD38 þþ/CD19 þ /CD45À/ þ ) and clonal PC.22,39,40 In PB from MGUSFn ¼ 44 and 27 samples from patients with a samples, the following cell populations were identified with the median age of 72 (range: 49–90 years) and 66 years (range: same eight-color combination: (i) CD34 þ HSC, normal PC and 41–83 years), respectivelyFSMMFn ¼ 32 and 21 samples, clonal PC. median age of 64 (range: 39–89 years) and 69 years (range: Data acquisition was performed in a FACSCanto II flow 39–83 years), respectivelyFand MMFn ¼ 74 and 30 samples cytometer (Becton Dickinson Biosciences (BDB), San Jose, CA, from patients with a median age of 64 (range: 41–83 years) and USA) using the FACSDiva software (version 6.1; BDB), and a 66 years (range: 41–84 years), respectivelyFpatients studied at two-step acquisition procedure: first, information on 5 Â 104 diagnosis, as well as 12 MM patients POST-HDT/ASCT (median events corresponding to the whole-sample cellularity was age of 58 years; range: 52–65 years). Paired BM and PB samples stored; second, data were stored only about CD19 þ and/or were available in 11 of the 60 patients with MGUS, 6 of 47 CD38 þ and/or CD34 þ gated events, for a minimum of 106 SMM and 17 of 87 MM studied at diagnosis, as well as in all 12 leukocytes per tube. 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 SDF-1 enzyme-linked immunosorbent assay 20 BM samples obtained from a group of HA undergoing Plasma was collected from PB samples obtained from 6 HA, 10 orthopedic surgery (median age of 50 years; range: 38–81 years) MGUS, 5 SMM and 14 MM patients at diagnosis and 5 MM

Leukemia Altered bone marrow homing to PC niches in MG B Paiva et al 699 POST-HDT/ASCT cases. Plasma levels of SDF-1 were measured between two or more groups, respectively. Antigenic profiles using a quantitative enzyme-linked immunosorbent assay of BM vs PB clonal PC were compared using the Wilcoxon (Quantikine, R&D systems Inc., Minneapolis, MN, USA) follow- signed rank test, and correlation studies were performed using ing the manufacturer’s recommendations. the Pearson’s test. For all statistical analyses, the SPSS software (version 15.0; SPSS Inc., Chicago, IL, USA) was used and correc- tions were applied for multiple comparisons (Bonferroni’s test). 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 Results diagnosis. Mononuclear cells were separated by ficoll-hypaque density gradient centrifugation and re-suspended in RPMI-1640 Distribution and immunophenotype of PB vs BM clonal with 1% L-glutamine and 0.5% bovine serum albumin. Cells PC in MGUS, SMM and MM were placed in the upper chambers of 6.5 mm transwell plates As expected, the overall percentage and median number of BM and (Costar, Corning, Acton, MA, USA) separated by an 8 mm pore PB circulating clonal PC found at diagnosis significantly increased size filter. SDF-1 (R&D systems) was added to the lower chambers from MGUS to SMM and MM patients (Supplementary Table 2). at concentrations of 0, 30 and 70 nM. After 4 h at 37 1C, cells in the Upon comparing in detail the immunophenotypic features of lower compartments were collected in Trucount tubes (BDB) and clonal PC from paired BM vs PB samples in a subset of 13 quantified by multiparameter flow cytometry, following the MGUS, SMM and MM cases, highly similar patterns were found, manufacturer’s instructions. The different cell subsets were with a few differences (Figure 1): PB clonal PC showed significantly identified by multiparameter flow cytometry using the following lower amounts of sideward light scatter (Po0.001), together five-color antibody combination (fluorescein isothiocyanate/PE/ with lower levels of CD38 (P ¼ 0.001), CD40 (P ¼ 0.006), CD56 PE-Cy7/PerCP-Cy5.5/allophycocyanin): CD38/CXCR4/CD19/CD45/ (P ¼ 0.03) and CD138 (P ¼ 0.02; Figure 1); by contrast, no CD34. Specific migration for SDF-1 of potentially competing significant differences were found regarding forward light scatter cell populationsFBM normal PC, B-cell precursors, CD34 þ and other B-cell markers (for example, CD19, CD20, CD22, HSC and clonal PCFwas determined for each cell population CD24, CD27, CD28, CD43, CD45, FMC7 and sIg). Of note, using the following formula: CXCR4 (CD184) was also detected at similar levels in paired BMFmean fluorescence intensity of 1145Fvs PB (mean 2 No: of cells=ml ð30 or 70 nM SDF-1ÞÀNo: of cells=ml ð0 nM SDF-1Þ fluorescence intensity: 1084) clonal PC (r ¼ 0.86; Po0.001). Â100 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 Cell surface expression of CXCR4 on BM clonal PC and the presence of SDF-1 was significantly higher than when SDF-1 PB SDF-1 levels in MGUS, SMM and MM was absent with two to five fold increased migration rates. 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 Statistical analyses POST-HDT/ASCT cases (28%) as regards the percentage of The Mann–Whitney U and the Kruskal–Wallis tests were used to CXCR4+ cells; (Figure 2a). By contrast, mean plasma levels of evaluate the statistical significance of differences observed SDF-1 (Figure 2c) were significantly increased in MM patients

200.000 200.000 BM clonal PC PB clonal PC

150.000 150.000

100.000 100.000 **

50.000 50.000 *** * Amount Amount of antigen MFI - *

10.000 10.000 Amount of light scatter / clonal PC *

2.000 (mean fluorescence intensity) - expression / clonal PC 2.000

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 Altered bone marrow homing to PC niches in MG B Paiva et al 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) (n ¼ 57) patients at diagnosis (Table 1). Upon considering both or SMM (1656 pg/ml)Freturning to normal levels in MM the absolute and relative numbers of all PB B cells, no significant POST-HDT/ASCT (1331 pg/ml). 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 Distribution of normal BM B-cell compartments in significantly lower percentages of naive B-lymphocytes MGUS, SMM and MM patients (Pp0.03) in both MM and SMM vs HA (Table 1). Conversely, The overall median proportion of normal B-cells from the whole non-switched memory B cells were increased (P ¼ 0.04) within BM cellularity was significantly decreased in MM (1.6%; the PB B-cell compartment in MM and SMM vs HA cases. Po0.001) and SMM (2.1%; P ¼ 0.02) patients vs HA (3.7%), Despite the overall percentage of normal PC in the BM was whereas it was normal in MGUS (P40.05), even when significantly lower among MM and SMM vs MGUS cases excluding clonal PC. This was mainly because of the presence (Po0.001; Supplementary Figure 1), the distribution of circulat- of decreased percentages of both the pro-B (1.0 vs 5.0%; ing PB normal PC was similar among groups (P40.05 vs HA). P ¼ 0.001) and pre-B (10 vs 24%; P ¼ 0.001) subsets within Interestingly, however, although the median proportion of the overall B-cells in MM, whereas mature B-lymphocytes CXCR4 þ normal PB PC (Figure 2b) was similar between HA (87 vs 70%; P ¼ 0.008) were proportionally increased within (11%), MGUS (14%) and SMM (15%), it was increased in MM the overall BM B-cell compartment (Figure 3). Interestingly, (21%; P ¼ 0.05); MM POST-HDT/ASCT cases showed the lowest the percentage of total B-cells in MM POST-HDT/ASCT BM median percentage of CXCR4 þ PB normal PC (5%). 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) þ precursors over more mature B-cells (17 vs 70%; Po0.001) BM and circulating PB CD34 HSC from patients with within the BM B-cell compartment. MGUS, SMM and MM Overall, the median percentage of CD34 þ HSC from the whole BM cellularity was significantly decreased in MM (0.3%; Distribution of normal B-cell compartments and P ¼ 0.001) and SMM (0.4%; P ¼ 0.002) patients vs HA (0.9%) plasmablasts/PC in the PB of MGUS, SMM and and MGUS (0.8%), even upon excluding clonal PC (Figure 4a). MM patients Interestingly, however, using the overall compartment of BM In addition to circulating normal plasmablasts/PC, four other CD34 þ HSC as denominator, the median percentage of maturation-associated subsets of B cells were systematically immature CD34 þ HSC was significantly increased in MM identified in the PB of MGUS (n ¼ 33), SMM (n ¼ 26) and MM (39%; Po0.001) and SMM (28%; Po0.001) vs HA (20%);

Leukemia Altered bone marrow homing to PC niches in MG B Paiva et al 701 Total B-cells BM B-cell subsets + + + Pro-B / CD19 BM cells Pre-B / CD19 BM cells Mature / CD19 BM cells 100% * ***

80%

10%

60% BM B-cells + BM cells / all + 40% * 5% * *** *** % of cells / CD19

% of CD19 20% * **

0% 0%

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

Figure 3 Distribution of B cells and their CD19 þ /CD34 þ /CD38 þ (pro-B), CD19 þ /CD34À/CD38 þ (pre-B) and CD19 þ /CD20 þ /CD38À (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) MGUS (N ¼ 33) SMM (N ¼ 26) MM (N ¼ 57) P-value

No. of PB B-cells/ml 82 (29–307) 82 (9.5–250) 60 (3.2–200) 84 (6.6–485) NS Immaturea 2.6 (0.2–12) 1.5 (0–22) 1.2 (0–8.3) 1.0 (0–26) 0.02 Naivea 60 (23–226) 56 (4.1–170) 34 (0.5–150) 44 (1.8–340) 0.05 Memory non-switcheda 12 (1.1–94) 5.4 (1.3–109) 13 (0.6–74) 12 (0.4–100) 0.07 Memory switcheda 13 (3.0–69) 13 (2.1–69) 10 (1.7–52) 13 (2.3–89) NS No. of PB normal PC/ml 0.6 (0.1–1.5) 0.6 (0.1–2.3) 0.4 (o0.06–4.0) 0.5 (o0.06–11) NS

Percentage of PB B-cells 1.8 (0.6–6.4) 1.7 (0.2–5.0) 1.5 (0.2–4.5) 1.6 (0.2–6.0) NS Immatureb 2.7 (0.3–5.5) 2.0 (0–11) 1.9 (0–6.0) 1.3 (0–9.8) 0.04 Naiveb 74 (43–88) 66 (28–95) 51 (17–75)** 57 (14–87)* 0.003 Memory non-switchedb 11 (2.5–31) 10 (1.5–48) 21 (8.0–56)* 19 (2.0–64)* 0.005 Memory switchedb 11 (4.0–31) 19 (2.8–46) 23 (6.3–61) 18 (3.8–52) NS Percentage of PB normal PC 0.01 (3À3–0.02) 9À3 (1À3–0.05) 9À3 (o1À3–0.08) 0.01 (o1À3–0.2) 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) aAbsolute count (no. of cells/ml). bPercentage of cells using all PB B-cells as the denominator. by contrast, the proportion of myeloid precursors was similar (Figure 4b). Our results show that both the relative and absolute among all groups, whereas the median percentage of lymphoid number of PB CD34 þ HSC are normal in MGUS and SMM, but precursors was lower in MM (3%; Po0.001) and SMM (13%; increased in MM (P ¼ 0.008 and P ¼ 0.06, respectively) when P ¼ 0.04) vs HA (16%). Interestingly, MM POST-HDT/ASCT compared with HA. By contrast, in MM POST-HDT/ASCT patients showed normal numbers of CD34 þ HSC in their BM patients the number of PB CD34 þ HSC returned to normal/ (1.1 vs 0.9% in HA; P40.05); however, a more detailed analysis lower levels. Of note, these results suggest that the distribution of the different subsets of CD34 þ HSC in these latter patients of BM and PB CD34 þ HSC are progressively altered from showed a marked reduction of the more immature precursors MGUS to MM at diagnosis with an inverse behavior; lower (Po0.001) at the expense of an increased number of lymphoid levels of BM CD34 þ HSC are paralleled by increased counts of precursors (Po0.001) within the overall compartment of circulating PB CD34 þ HSC. Accordingly, the median percen- CD34 þ HSC. tage of BM non-lymphoid CD34 þ HSC (Supplementary Figure 2) In a subgroup of cases, we further investigated the distribution in MM patients with anemia was significantly decreased vs of circulating PB CD34 þ HSC in paired BM and PB samples HA (0.24 vs 0.63%; P ¼ 0.04), whereas MM patients with

Leukemia Altered bone marrow homing to PC niches in MG B Paiva et al 702

Figure 4 Distribution of CD34 þ HSC and their CD34 þ /CD38À/ þ dim (immature), CD34 þ /CD38 þ /CD19À (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 findings, it should be noted that while MM patients with detect- non-lymphoid CD34 þ HSC (0.77 vs 0.63%). able 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 Relationship between different subsets of PB and BM the whole BM cellularity, respectively; P ¼ 0.001), among cells competing for potentially overlapping BM niches MGUS and SMM patients no differences were found or they Upon comparing the distribution of different cell were less pronounced. 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) ‘Ex vivo’ competition between BM B cells, PC and and patients with MGUS (n ¼ 12), SMM (n ¼ 6) and MM (n ¼ 17) CD34 þ HSC for SDF-1-induced migration studied at diagnosis, as well as MM POST-HDT/ASCT (n ¼ 12; In HA, both the immature and myeloid subsets of CD34 þ HSC Supplementary Figure 3), a significant correlation (r2 ¼ 0.42; showed the highest migration potential in the presence of Po0.001) was found between the percentage of these cell SDF-1, followed by pre-B cell precursors; conversely, normal populations in the BM (CD34 þ HSC plus pre-B cells and PC) PC barely migrated (for example, medians for an SDF-1 concen- þ and the corresponding circulating populations in PB (CD34 tration of 30 nM of 6, 22, 4 and 0.01%, respectively; Figure 5a). HSC and PC). Interestingly, HA grouped together with MGUS No significant differences were found for the migration of all cell cases, whereas SMM scattered between HA/MGUS and MM populations analyzed between MGUS and SMM patients vs HA, patients; MM POST-HDT/ASCT showed features similar to except for clonal PC that showed an impaired migration in the HA/MGUS cases (Supplementary Figure 3). Despite these presence of SDF-1 for the two concentrations used (Figure 5).

Leukemia Altered bone marrow homing to PC niches in MG B Paiva et al 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 /CD38À/ þ dim (immature), CD34 þ /CD38 þ /CD19À (myeloid) and CD34 þ /CD38 þ /CD19 þ (lymphoid) subsetsFB-cell precursors CD19 þ /CD34 þ /CD38 þ (pro-B) and CD19 þ /CD34À/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) Discussion 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 It has been shown that the proportion of residual normal/ MM, particularly at lower SDF-1 concentrations (Figure 5a). polyclonal BM PC (from all BM PC) can be of value for the Most interestingly, the migration of clonal PC from symptomatic discrimination between MGUS and MM.20 More recently, this MM was markedly increased at both concentrations of SDF-1 parameter has also proved to be predictive for the progression used: median of 2.9 and 1.0% for SDF-1 concentrations of 30 of MGUS and SMM to symptomatic myeloma,21 as well as 22 and 70 nM, respectively. for disease outcome in MM. Despite its clinical impact,

Leukemia Altered bone marrow homing to PC niches in MG B Paiva et al 704 the potential mechanisms leading to such clinical associations CD34 þ HSC. In addition, it has been demonstrated in mice that remain largely unknown and only an association with immune pre-pro-B cells and PC compete for the same BM niche.25 paresis has been reported so far.21,22 Serum levels of polyclonal Therefore, it could be hypothesized that such PC competition for Igs are typically depressed in MM 5,22 and, less frequently, also BM niches could also extend to B-cell precursors and CD34 þ in MGUS and SMM;21,41 this has been related to the depletion of HSC which use the SDF-1-CXCR4 axis in common. In line with normal PC in the patients’ BM. However, it should be noted that this hypothesis we found here that in parallel to normal PC, the in MM, both parametersFimmune paresis and the proportion of distribution of different B-cell subsets and CD34 þ HSC is also normal PC from all BM PCFappear to be independently dysregulated in patients with monoclonal gammopathies. associated with patient outcome.22 In this study, we confirm the Interestingly, the frequency of BM pro-B and pre-B cell existence of progressively decreasing numbers of normal BM PC precursors and CD34 þ HSC in MGUS was similar to that of from MGUS to SMM and MM patients, in parallel to an increase HA, whereas MM showed a progressively lower number of these in both BM and PB clonal PC. cell populations from smoldering to symptomatic disease. These In order to become long-living PC, recently produced normal findings would further support the model of cell competition,25 PC (and also clonal PC) need to be in close contact with SDF-1- where clonal PC progressively replace B-cell precursors from secreting stromal cell niches in the BM. However, the number of their BM niches, in addition to normal PC. Whether the decrease such BM niches is finite23 and functionally impaired in the in pre-B cells is due to an intrinsically impaired B-cell elderly.28,33 Thus, the progressively decreased numbers of production or to direct competition with clonal PC (similarly normal PC in the BM of MM patients could be because of an to what has been suggested for normal PC26,27), remains to be increased competition with the malignant clone for the BM PC elucidated. niches. Interestingly, it has been suggested that adipocytes In addition to normal PC and pre-B cells, total CD34 þ HSC (whose numbers progressively increase in the BM microenvir- were also decreased in the BM of SMM and MM patients, at the onment of elderly individuals) have equivalent functions to BM expense of a decreased number of lymphoid progenitors (which stromal cells in supporting clonal PC.35 In such case, the could be further responsible for the lower number of PB decreased number of normal BM PC could be due, not to a immature and naive B-lymphocytes found among MM patients). decreased PC production, but to limited homing of recently By contrast, the distribution of immature and myeloid CD34 þ produced PB PC into the BM because BM niches are overloaded HPC was not significantly altered in MGUS. Based on these with clonal PC. In line with this hypothesis, our results show that data, it could be speculated that during progression from pre- MM patients display normal numbers of normal PC in PB, malignant conditions (for example, MGUS and SMM) to despite undetectable in the BM of most patients. Of note, symptomatic MM, there is a derailment in the tightly regulated normal PB PC from patients with MM (but not MGUS and SMM) and balanced competition between clonal PC and different showed increased expression of the CXCR4 chemokine (SDF-1) normal cell populations (normal PC, B-cell precursors and receptor. As expression of CXCR4 by PC is considered to be a CD34 þ HSC) for a limited number of BM niches. In line with marker for predominant BM homing,23,42 these findings further this hypothesis, MM patients also showed increased numbers of support the notion that in symptomatic MM, the fraction of circulating PB CD34 þ HSC and a significant correlation was normal PB PC that should home to the BM remains in the observed between PB and BM cell populations (including clonal circulation and accumulates in the PB of these patients. PC) potentially competing for the same BM niches; both values Usually, PB clonal PC have been viewed as cells that leave (and thus also the competition between them) increased from the BM because of unique migration properties acquired at HA to MGUS, SMM and MM; this was followed by a decrease in relatively advanced disease stages, particularly when clonal PC MM POST-HDT/ASCT cases, which could be explained by decrease their BM microenvironment dependency.10 Alterna- niche emptiness after therapy. It would be interesting to further tively, our results could also indicate that in order to spread the investigate whether this model of cell competition correlates diseaseFlike normal BM PC23,32,42Fclonal PC could also with the malignant transformation of MGUS and progression of leave the BM, recirculate into PB and home again into the BM at ‘evolving’ MGUS cases as well as MM patients’ response after a different localization, in a kind of ‘metastatic’/dissemination therapy; herein, MM patients achieving complete response after process. In line with this hypothesis, in this and other studies7–14,43 HDT/ASCT showed slightly (non-significant) increased numbers circulating clonal PC were detected in a significant proportion of normal PC, total B-cells and CD34 þ HSC vs cases in partial of MGUS and SMM cases in addition to MM. A potential response (data not shown). Our results in a new model of explanation for the recirculation of clonal PC in the early competitive ex vivo migration assay between normal BM cells stages of the disease (for example, MGUS) could rely on the vs clonal PC in the presence of variable concentrations of SDF-1 advanced age of MGUS patients, which could lead to lower would further support the hypothesis of cell competition, in numbers of available BM niches.44 However, we failed to show which normal PC, B-cell precursors and CD34 þ HSC from any correlation between the presence of clonal PC in PB and MGUS and SMM, but not MM patients, were able to abrogate patients’ age. In turn, clonal PC from paired samples showed a the migration of clonal PC, particularly at the lower concentra- slightly more immature immunophenotypic profile for PB vs BM tions evaluated. clonal PC (for example, lower internal complexity and lower Despite all associations described above, it should be noted expression of CD38, CD138 and CD40). Clonotypic cells with a that a relatively wide variation was found within MGUS, SMM more immature phenotype have been repeatedly found in the and MM patients for the distribution of some cell populations, PB of patients with monoclonal gammopathies,12,45–47 and with overlapping numbers between HA and MGUS, MGUS and decreased CD138 expression by circulating PB clonal PC could SMM and between SMM and MM. In such cases, the specific also indicate they are enriched in clonogenic cells, in line with genetic background and other relevant features of tumor cells the previously reported immunophenotypic properties of clonal may determinate more closely the behavior of the disease. PC showing stem cell properties.46,48 Further studies are necessary to confirm this hypothesis. In the early steps of B-cell maturation, B-cell precursors also In summary, our results suggest that progressive competition adhere to SDF-1-expressing reticular cells25 and SDF-1 also has and replacement of normal BM PC, B-cell precursors and an important role in the colonization of the BM by circulating CD34 þ HSC by clonal PC in BM niches from patients with

Leukemia Altered bone marrow homing to PC niches in MG B Paiva et al 705 monoclonal gammopathies is associated with more advanced gammopathy of undetermined significance. J Clin Oncol 2005; 23: forms of the disease. The exact composition and distribution of 5668–5674. such BM niches and the causes leading to the potential 14 Witzig T, Kyle R, O’Fallon W, Greipp P. Detection of peripheral competing advantage of clonal PC vs normal cells (for example, blood plasma cells as a predictor of disease course in patients with smouldering multiple myeloma. Br J Haematol 1994; 87: 266–272. specific cytogenetic or epigenetic abnormalities) for such niches 15 Kuehl WM, Bergsagel PL. Multiple myeloma: evolving genetic deserve further investigations. events and host interactions. Nat Rev Cancer 2002; 2: 175–187. 16 Fonseca R, Barlogie B, Bataille R, Bastard C, Bergsagel PL, Chesi M et al. Genetics and cytogenetics of multiple myeloma: a workshop Conflict of interest report. Cancer Res 2004; 64: 1546–1558. 17 Fonseca R, Bailey R, Ahmann G, Rajkumar S, Hoyer J, Lust J et al. The authors declare no conflict of interest. Genomic abnormalities in monoclonal gammopathy of undetermined significance. Blood 2002; 100: 1417–1424. 18 Paiva B, Almeida J, Perez-Andres M, Mateo G, Lopez A, Rasillo A Acknowledgements et al. Utility of flow cytometry immunophenotyping in multiple myeloma and other clonal plasma cell-related disorders. Cytometry This work was supported by the Cooperative Research Thematic B Clin Cytom 2010; 78: 239–252. Cancer Network (RTICs; RD06/0020/0006, RD06/0020/0035 and 19 Perez-Andres M, Almeida J, Martin-Ayuso M, Moro MJ, Martin- G03/136), MM Jevitt, SL firm, Instituto de Salud Carlos III/ Nunez G, Galende J et al. Clonal plasma cells from monoclonal Subdireccioń General de Investigacioń Sanitaria (FIS: PI060339; gammopathy of undetermined significance, multiple myeloma and plasma cell leukemia show different expression profiles of 02/0905; 01/0089/01-02; PS09/01897), Consejerıá de Sanidad molecules involved in the interaction with the immunological and Conserjerıá de Educacion (GR37), Junta de Castilla y Leoń, bone marrow microenvironment. Leukemia 2005; 19: 449–455. Valladolid, Spain (557/A/10) and from MSCNET European strep 20 Ocqueteau M, Orfao A, Almeida J, Blade J, Gonzalez M, Garcia- (N1E06005FF). Sanz R et al. Immunophenotypic characterization of plasma cells from monoclonal gammopathy of undetermined significance patients. Implications for the differential diagnosis between MGUS References and multiple myeloma. Am J Pathol 1998; 152: 1655–1665. 21 Perez-Persona E, Vidriales MB, Mateo G, Garcia-Sanz R, Mateos 1 Kyle R, Therneau T, Rajkumar S, Larson D, Plevak M, Offord J et al. MV, de Coca AG et al. New criteria to identify risk of progression Prevalence of monoclonal gammopathy of undetermined signifi- in monoclonal gammopathy of uncertain significance and smoldering cance. N Engl J Med 2006; 354: 1362–1369. multiple myeloma based on multiparameter flow cytometry analysis 2 Weber D, Dimopoulos M, Moulopoulos L, Delasalle K, Smith T, of bone marrow plasma cells. Blood 2007; 110: 2586–2592. Alexanian R. Prognostic features of asymptomatic multiple 22 Paiva B, Vidriales MB, Mateo G, Perez JJ, Montalban MA, Sureda A myeloma. Br J Haematol 1997; 97: 810–814. et al. The persistence of immunophenotypically normal residual 3 Kyle R, Therneau T, Rajkumar S, Offord J, Larson D, Plevak M et al. bone marrow plasma cells at diagnosis identifies a good prognostic A long-term study of prognosis in monoclonal gammopathy of subgroup of symptomatic multiple myeloma patients. Blood 2009; undetermined significance. N Engl J Med 2002; 346: 564–569. 114: 4369–4372. 4 Cesana C, Klersy C, Barbarano L, Nosari A, Crugnola M, Pungolino E 23 Radbruch A, Muehlinghaus G, Luger EO, Inamine A, Smith KG, et al. Prognostic factors for malignant transformation in monoclonal Dorner T et al. Competence and competition: the challenge of gammopathy of undetermined significance and smoldering multiple becoming a long-lived plasma cell. Nat Rev Immunol 2006; 6: myeloma. JClinOncol2002; 20: 1625–1634. 741–750. 5 San Miguel JF, Gutierrez NC, Mateo G, Orfao A. Conventional 24 Shapiro-Shelef M, Calame K. Regulation of plasma-cell develop- diagnostics in multiple myeloma. Eur J Cancer 2006; 42: 1510–1519. ment. Nat Rev Immunol 2005; 5: 230–242. 6 Kyle RA, Rajkumar SV. Multiple myeloma. N Engl J Med 2004; 25 Tokoyoda K, Egawa T, Sugiyama T, Choi BI, Nagasawa T. Cellular 351: 1860–1873. niches controlling B lymphocyte behavior within bone marrow 7 Luque R, Brieva J, Moreno A, Manzanal A, Escribano L, Villarrubia J during development. Immunity 2004; 20: 707–718. et al. Normal and clonal B lineage cells can be distinguished by 26 Ma Q, Jones D, Springer TA. The chemokine receptor CXCR4 is their differential expression of B cell antigens and adhesion required for the retention of B lineage and granulocytic precursors molecules in peripheral blood from multiple myeloma (MM) within the bone marrow microenvironment. Immunity 1999; 10: patients–diagnostic and clinical implications. Clin Exp Immunol 463–471. 1998; 112: 410–418. 27 Nie Y, Waite J, Brewer F, Sunshine MJ, Littman DR, Zou YR. The 8 Chandesris MO, Soulier J, Labaume S, Crinquette A, Repellini L, role of CXCR4 in maintaining peripheral B cell compartments and Chemin K et al. Detection and follow-up of fibroblast growth factor humoral immunity. J Exp Med 2004; 200: 1145–1156. receptor 3 expression on bone marrow and circulating plasma cells 28 Perez-Andres M, Paiva B, Nieto WG, Caraux A, Schmitz A, by flow cytometry in patients with t(4;14) multiple myeloma. Br J Almeida J et al. Human peripheral blood B-cell compartments: a Haematol 2007; 136: 609–614. crossroad in B-cell traffic. Cytometry B Clin Cytom 2010; 78 9 Rawstron AC, Owen RG, Davies FE, Johnson RJ, Jones RA, (Suppl 1): S47–S60. Richards SJ et al. Circulating plasma cells in multiple myeloma: 29 Foudi A, Jarrier P, Zhang Y, Wittner M, Geay JF, Lecluse Y et al. characterization and correlation with disease stage. Br J Haematol Reduced retention of radioprotective hematopoietic cells within 1997; 97: 46–55. the bone marrow microenvironment in CXCR4À/À chimeric mice. 10 Nowakowski GS, Witzig TE, Dingli D, Tracz MJ, Gertz MA, Lacy Blood 2006; 107: 2243–2251. MQ et al. Circulating plasma cells detected by flow cytometry as a 30 Sugiyama T, Kohara H, Noda M, Nagasawa T. Maintenance of the predictor of survival in 302 patients with newly diagnosed multiple hematopoietic stem cell pool by CXCL12-CXCR4 chemokine myeloma. Blood 2005; 106: 2276–2279. signaling in bone marrow stromal cell niches. Immunity 2006; 25: 11 Schneider U, van LA, Huhn D, Serke S. Two subsets of peripheral 977–988. blood plasma cells defined by differential expression of CD45 31 Serke S, Sauberlich S, Huhn D. Multiparameter flow-cytometrical antigen. Br J Haematol 1997; 97: 56–64. quantitation of circulating CD34(+)-cells: correlation to the 12 Billadeau D, Van NB, Kimlinger T, Kyle RA, Therneau TM, Greipp quantitation of circulating haemopoietic progenitor cells by PR et al. Clonal circulating cells are common in plasma cell in vitro colony-assay. Br J Haematol 1991; 77: 453–459. proliferative disorders: a comparison of monoclonal gammopathy 32 Caraux A, Klein B, Paiva B, Bret C, Schmitz A, Fuhler GM et al. of undetermined significance, smoldering multiple myeloma, and Circulating human B and plasma cells. Age-associated changes active myeloma. Blood 1996; 88: 289–296. in counts and detailed characterization of circulating normal 13 Kumar S, Rajkumar S, Kyle R, Lacy M, Dispenzieri A, Fonseca R CD138À and CD138+ plasma cells. Haematologica 2010; 95: et al. Prognostic value of circulating plasma cells in monoclonal 1016–1020.

Leukemia Altered bone marrow homing to PC niches in MG B Paiva et al 706 33 Mayack SR, Shadrach JL, Kim FS, Wagers AJ. Systemic signals patterns of normal hematopoiesis in human bone marrow: regulate ageing and rejuvenation of blood stem cell niches. Nature reference patterns for age-related changes and disease-induced 2010; 463: 495–500. shifts. Cytometry B Clin Cytom 2004; 60: 1–13. 34 Corre J, Planat-Benard V, Corberand JX, Peńicaud L, Casteilla L, 41 Kyle RA, Rajkumar SV. Epidemiology of the plasma-cell disorders. Laharrague P. Human bone marrow adipocytes support complete Best Pract Res Clin Haematol 2007; 20: 637–664. myeloid and lymphoid differentiation from human CD34 cells. 42 Mei HE, Yoshida T, Sime W, Hiepe F, Thiele K, Manz RA Br J Haematol 2004; 127: 344–347. et al. Blood-borne human plasma cells in steady state are 35 Caers J, Deleu S, Belaid Z, De Raeve H, Van Valckenborgh E, derived from mucosal immune responses. Blood 2009; 113: De Bruyne E et al. Neighboring adipocytes participate in the bone 2461–2469. marrow microenvironment of multiple myeloma cells. Leukemia 43 Rawstron A, Barrans S, Blythe D, Davies F, English A, Pratt G et al. 2007; 21: 1580–1584. Distribution of myeloma plasma cells in peripheral blood and 36 International Myeloma Working Group. Criteria for the classifi- bone marrow correlates with CD56 expression. Br J Haematol cation of monoclonal gammopathies, multiple myeloma and 1999; 104: 138–143. related disorders: a report of the International Myeloma Working 44 Frasca D, Landin A, Riley R, Blomberg B. Mechanisms for decreased Group. Br J Haematol 2003; 121: 749–757. function of B cells in aged mice and humans. J Immunol 2008; 37 Nieto WG, Almeida J, Romero A, Teodosio C, Lopez A, 180: 2741–2746. Henriques AF et al. Increased frequency (12%) of circulating 45 Isaksson E, Bjorkholm M, Holm G, Johansson B, Nilsson B, chronic lymphocytic leukemia-like B-cell clones in healthy Mellstedt H et al. Blood clonal B-cell excess in patients with subjects using a highly sensitive multicolor flow cytometry monoclonal gammopathy of undetermined significance (MGUS): approach. Blood 2009; 114: 33–37. association with malignant transformation. Br J Haematol 1996; 38 Paiva B, Vidriales MB, Cervero J, Mateo G, Perez JJ, Montalban 92: 71–76. MA et al. Multiparameter flow cytometric remission is the most 46 Matsui W, Huff CA, Wang Q, Malehorn MT, Barber J, Tanhehco Y relevant prognostic factor for multiple myeloma patients who et al. Characterization of clonogenic multiple myeloma cells. undergo autologous stem cell transplantation. Blood 2008; 112: Blood 2004; 103: 2332–2336. 4017–4023. 47 Pilarski LM, Hipperson G, Seeberger K, Pruski E, Coupland RW, 39 Matarraz S, Lopez A, Barrena S, Fernandez C, Jensen E, Flores J Belch AR. Myeloma progenitors in the blood of patients with et al. The immunophenotype of different immature, myeloid and aggressive or minimal disease: engraftment and self-renewal of B-cell lineage-committed CD34+ hematopoietic cells allows primary human myeloma in the bone marrow of NOD SCID mice. discrimination between normal/reactive and myelodysplastic Blood 2000; 95: 1056–1065. syndrome precursors. Leukemia 2008; 22: 1175–1183. 48 Matsui W, Wang Q, Barber JP, Brennan S, Smith BD, Borrello I 40 van Lochem EG, van der Velden VH, Wind HK, te Marvelde JG, et al. Clonogenic multiple myeloma progenitors, stem cell Westerdaal NA, van Dongen JJ. Immunophenotypic differentiation 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