(2014) 28, 166–173 & 2014 Macmillan Publishers Limited All rights reserved 0887-6924/14 www.nature.com/leu

ORIGINAL ARTICLE Multiparameter flow cytometry for the identification of the Waldenstro¨m’s clone in IgM-MGUS and Waldenstro¨m’s : new criteria for differential diagnosis and risk stratification

B Paiva1,2, MC Montes1, R Garcı´a-Sanz1,2, EM Ocio1,2, J Alonso1, N de las Heras3, F Escalante3, R Cuello4, AG de Coca4, J Galende5, J Herna´ndez6, M Sierra7, A Martin1, E Pardal8,ABa´rez9, J Alonso10, L Suarez11, TJ Gonza´lez-Lo´ pez12, JJ Perez1, A Orfao2,13, M-B Vidrı´ales1,2 and JF San Miguel1,2

Although multiparameter flow cytometry (MFC) has demonstrated clinical relevance in of undetermined significance (MGUS)/myeloma, immunophenotypic studies on the full spectrum of Waldenstro¨m’s Macroglobulinemia (WM) remain scanty. Herein, a comprehensive MFC analysis on bone marrow samples from 244 newly diagnosed patients with an (IgM) monoclonal protein was performed, including 67 IgM-MGUS, 77 smoldering and 100 symptomatic WM. Our results show a progressive increase on the number and light-chain-isotype-positive B-cells from IgM-MGUS to smoldering and symptomatic WM (Po.001), with only 1% of IgM-MGUS patients showing 410% B cells or 100% light-chain-isotype-positive B-cells (Po.001). Complete light-chain restriction of the B-cell compartment was an independent prognostic factor for time-to progression in smoldering WM (median 26 months; HR: 19.8, P ¼ 0.001) and overall survival in symptomatic WM (median 44 months; HR: 2.6, P ¼ 0.004). The progressive accumulation of light-chain-isotype-positive B-cells accompanied the emergence of a characteristic Waldenstrom’s phenotype (CD22 þ dim/CD25 þ /CD27 þ /IgM þ ) that differed from other B-NHL by negative expression of CD5, CD10, CD11c or CD103. In contrast to myeloma, light-chain-isotype-positive plasma cells in IgM monoclonal gammopathies show otherwise normal antigenic expression. Our results highlight the potential value of MFC immunophenotyping for the characterization of the Waldenstro¨m’s clone, as well as for the differential diagnosis, risk of progression and survival in WM.

Leukemia (2014) 28, 166–173; doi:10.1038/leu.2013.124 Keywords: Waldenstrom’s Macroglobulinemia; MGUS; flow cytometry; phenotype; plasma cells; survival

INTRODUCTION there is some controversy to distinguish between smoldering WM The WHO defines Waldenstro¨m’s Macroglobulinemia (WM) as a and IgM-MGUS; some groups use the amount of serum 6 lymphoplasmacytic associated with a monoclonal concentration of the M-protein and BM infiltration, while a immunoglobulin M (IgM) protein (regardless of size) and bone consensus panel defined BM disease involvement on histological examination as the main criterion.2 It is plausible to assume that marrow (BM) infiltration by small lymphocytes that may exhibit 7 plasma cell (PC) differentiation.1 As lymphoplasmacytic lymphoma similarly to (MM), most (if not all) WM patients have eventually gone through the benign stages of IgM-MGUS infrequently involves the lymph nodes or other extramedullary and smoldering WM before developing clinical symptoms. sites, demonstration of BM infiltration (for example, through 2 Therefore, the identification of new objective criteria for the trephine biopsy) is therefore essential for the diagnosis of WM. differential diagnosis between these conditions, as well as more Many patients who fulfill the criteria for WM do not require accurate estimation of their risk of progression for optimal clinical immediate therapy because they are detected before developing management are mostly welcome, particularly in two settings: 3 disease-associated symptoms. These patients are classified as (i) to decide which patients are candidates for an early treatment smoldering WM, a clinically recognized entity with a cumulative intervention, and (ii) to optimize monitoring of premalignant probability of progression to symptomatic WM, amyloidosis or stages. However, the differential diagnosis among these patient lymphoma of 65% at 10 years.4 groups may be challenging as some asymptomatic cases can Smoldering WM patients have a greater risk of progression to present with markedly elevated IgM levels and BM infiltration, active disease than IgM-monoclonal gammopathy of undeter- while others show minimal tumor burden but require therapy due mined significance (MGUS) cases (18% at 10 years).5 Currently, to disease-related complications. Moreover, although serum IgM

1Hospital Universitario de Salamanca, Department of Hematology, Salamanca, Spain; 2IBSAL IBMCC (USAL-CSIC), Salamanca, Spain; 3Complejo Hospitalario de Leo´n, Leo´ n, Spain; 4Hospital Clı´nico Universitario de Valladolid, Valladolid, Spain; 5Hospital Comarcal del Bierzo, PonferradaLeo´ n, Spain; 6Hospital General de Segovia, Segovia, Spain; 7Hospital Virgen de la Concha, Zamora, Spain; 8Hospital Virgen del Puerto, Plasencia, Spain; 9Hospital Nuestra Sen˜ ora de Sonsoles, A´ vila, Spain; 10Hospital Rio Carrio´ n, Palencia, Spain; 11Hospital Santos Reyes, Aranda de Duero, Spain; 12Complejo Asistencial Universitario de Burgos, Burgos, Spain and 13Servicio General de Citometrı´a and Departamento de Medicina, Universidad de Salamanca, Salamanca, Spain. Correspondence: Professor JF San Miguel, IBSAL-Hospital Universitario de Salamanca, Paseo de San Vicente 58–182, Salamanca 37007, Spain. E-mail: [email protected] Received 12 March 2013; revised 10 April 2013; accepted 12 April 2013; accepted article preview online 24 April 2013; advance online publication, 14 May 2013 Immunophenotyping of IgM monoclonal gammopathies B Paiva et al 167 concentration is increased in WM, it may often do not exceed MFC immunophenotypic studies 2 3 g/dl, thus showing considerable overlap with IgM-MGUS; Erythrocyte-lysed whole-BM samples were immunophenotyped using similar variability has also been reported for the levels of BM 4-color combinations of the following fluorochrome-conjugated—fluores- infiltration.8–13 Finally, the differential diagnosis of WM may also cein isothiocyanate/phycoerythrin/peridinin chlorophyll protein-cyanin 5.5 be challenged by other diseases, which may also show an IgM (PerCPCy5.5)/allophycocyanin—monoclonal : cytoplasmatic Ig paraprotein (for example, IgM-MM, marginal zone, mantle cell or lambda light chain (CyIgl)/cytoplasmatic Ig kappa CyIgk/CD19/CD38, ). CyBcl2/CD103/CD19/CD38, surface membrane IgM (SmIgM)/CD25/CD19/ CD38, SmIgM/CD27/CD19/CD38 CD11c/CD5/CD19/CD38, CD22/CD23/ In recent years, multiparameter flow cytometry (MFC) immuno- CD19/CD38, CD138/CD25/CD19/CD38, HLA-DR/CD11b/CD19/CD38, FMC7/ phenotyping has demonstrated to be of great value in the 14 CD24/CD19/CD38, CD45RA/CD45RO/CD19/CD38, CD38/CD10/C19/CD20, differential diagnosis of B-cell lymphoproliferative disorders, as CD38/CD56/CD19/CD45. Thus, a total of 23 different phenotypic markers, well as for baseline prognostication of patients with (non-IgM) including SmIgM, CyIgl and CyIgk were systematically evaluated in both MGUS and MM.15,16 Its use in WM was strongly recommended in BM mature B cells and PCs identified as CD19 þ /SSClow and CD38hi/SSCint the Second International WM Workshop in 2003,2 but immuno- cells, respectively, after careful exclusion of B-cell precursors (CD19 þ / hi low phenotypic studies on patients with IgM monoclonal CD38 /SSC ). A stain-and-then-lyse direct immunofluorescence techni- gammopathies remain limited, focused only on patients with que was used to evaluate all surface markers; for the staining of CyIgl, the most aggressive forms of the disease, and no consensus has CyIgk and CyBcl2, the Fix & Perm reagent kit (DAKO, Glostrup, Denmark) was used, strictly following the recommendations of the manufacturer. The been reached so far about the phenotypic characteristics of the 17–23 evaluation of SmIgM, CyIgl and CyIgk was performed after washing serum Waldenstro¨m’s clone. Ig. In all cases, one tube stained with CD19-PerCPCy5.5 and CD38-APC was Here, we report on a comprehensive immunophenotypic used for specific evaluation of baseline autofluorescence levels of B cells analysis of a large series of newly diagnosed IgM-MGUS, smolde- and PCs, respectively (negative control). ring and symptomatic WM patients. Our results show that MFC is a Data acquisition was performed in a FACSCalibur (4-color) flow valuable technique to assess the tumor burden and the degree of cytometer (Becton Dickinson Biosciences, BDB, San Jose, CA, USA) using clonality in patients’ BM aspirate, providing invaluable information the FACSDiva 6.1 software (BDB), and a two-step acquisition procedure. In 4 for the differential diagnosis of IgM-MGUS and prognostication of the first step, information about 5 Â 10 events corresponding to the WM. In addition, the specific phenotypic characteristics of clonal B whole-sample cellularity was stored; while in the second step, data about CD19 þ or CD38 þ gated events was selectively stored, for a minimum of cells and PCs are also described in detail. 106 leukocytes/tube. Data analysis was performed using the Infinicyt software (Cytognos, Salamanca, Spain). PATIENTS AND METHODS Patients Statistical analysis A total of 244 newly diagnosed patients with an IgM monoclonal The Mann–Whitney U test was used to estimate the statistical significance gammopathy—classified as IgM-MGUS (n ¼ 67), smoldering WM (n ¼ 77) of differences observed between groups. Smoldering WM patients who and symptomatic WM (n ¼ 100)—who were referred to our institution for had not progressed were censored on the last date they were known to be immunophenotypic investigations are the focus of the present study. alive and without progression. Overall survival (OS) was measured from the Patients were classified according to the Second International Workshop moment of diagnosis to the date of death or last visit. Time-to progression on WM criteria.2 In brief, the absence versus presence of BM infiltration in a (TTP) and OS curves were plotted by the Kaplan–Meier method, and the trephine biopsy (or 420% B cells in the BM aspirate by cytomorphology if log-rank test was used to estimate the statistical significance of differences biopsy was not available) was used for the differential diagnosis between observed between curves. For multivariate analysis, the stepwise Cox IgM-MGUS and WM, respectively; in turn, the distinction between proportional hazard and logistic regressions were used, retaining those smoldering and symptomatic WM was based on the absence versus variables with a statistically significant predictive value (Pp0.05) in the presence of symptoms related to tumor infiltration (typically anemia, predictive model. For all statistical analyses, the SPSS software (version hepatosplenomegaly, hiperviscosity and/or peripheral neuropathy).24 IgM- 15.0; SPSS Inc., Chicago, IL, USA) was used. MGUS patients without trephine biopsy (N ¼ 24) showed non-significantly different clinical-biological parameters and risk of progression into symptomatic WM. Samples were collected before therapy after informed RESULTS consent was given, in accordance with the ethical committee guidelines Quantification of tumor mass and clonality by MFC in IgM-MGUS and the Declaration of Helsinki. Median follow-up for the whole series was versus WM of 36 months (range: 2–126 months). Survival data was missing in four IgM- MGUS (6%), two smoldering WM (3%) and seven symptomatic WM (7%). Overall, the median percentage of mature BM B cells (among total Baseline demographics and disease characteristics are shown in Table 1. nucleated cells) progressively increased (Po0.001) from IgM- Hemoglobin and platelet values below the normal range in IgM-MGUS and MGUS (2.2%) to smoldering (8.7%) and symptomatic WM patients smoldering WM patients were considered as unrelated to the underlying (12.2%) (Table 2). In fact, only 1% of IgM-MGUS patients showed B-cell clone. 410% B cells, in contrast to 34 and 55% of smoldering and

Table 1. Demographic and clinical characteristics of newly diagnosed patients with an IgM monoclonal gammopathy grouped according to their diagnosis

IgM-MGUS (N ¼ 67) Smoldering WM (N ¼ 77) Symptomatic WM (N ¼ 100)

Male/femalea 40 (60%)/27 (40%) 49 (64%)/28 (36%) 69 (69%)/31 (31%) Age (years) 72 (39–86) 72 (44–93) 74 (30–94) Marrow infiltration (%) 14 (2–20) 30 (3–92) 45 (13–100) Serum IgM (g/l) 0.8 (0.2–3.1) 1.7 (0.4–6.0) 4.0 (0.6–11.5) Hemoglobin (g/l) 134 (90–175) 132 (86–159) 104 (56–158) No. of leukocytes ( Â 109/l) 6.7 (2.1–14.3) 7.3 (3.0–22.3) 6.8 (2.0–30.5) No. of platelets ( Â 109/l) 257 (104–739) 249 (92–483) 216 (4–530) Follow-up (months) 33 (3–138) 45 (3–162) 33 (2–121) Abbreviations: MGUS, monoclonal gammopathy of undetermined significance; WM, Waldenstrom’s Macroglobulinemia. aResults expressed as median (range) or as number (percentage) of cases.

& 2014 Macmillan Publishers Limited Leukemia (2014) 166 – 173 Immunophenotyping of IgM monoclonal gammopathies B Paiva et al 168 Table 2. BM tumor burden and degree of clonality as assessed by MFC in newly diagnosed patients with an IgM monoclonal gammopathy distributed according to final diagnosis

IgM-MGUS (N ¼ 67) Smoldering WM (N ¼ 77) Symptomatic WM (N ¼ 100)P-value

% of BM B cells 2.2 (0.2–12) 8.7 (0.9–54) 12.2 (0.3–76) o0.001 % of patients with 410% BM B cellsa 1% 34% 55% o0.001 % of light-chain-isotype-positive B-cells 75 (8–100) 96 (16–100) 99 (45–100) o0.001 % of patients with complete B-cell light-chain restriction 1% 19% 40% o0.001 % of BM PCs 0.32 (0–2.4) 0.28 (0–1.3) 0.30 (0–7.0) 0.66 % of light-chain-isotype-positive PCs 70 (10–100) 85 (24–100) 97 (34–100) o0.001 % of patients with complete PC light-chain restrictiona 8% 14% 37% o0.001 Abbreviations: BM, bone marrow; PC, plasma cell; MGUS, monoclonal gammopathy of undetermined significance; WM, Waldenstrom’s Macroglobulinemia. aResults expressed as median (range) or as number (percentage) of cases.

symptomatic WM patients (Po0.001). As could be expected, this investigated whether the estimation of tumor burden by MFC was associated with a progressively increased (Po0.001) infiltra- could also be prognostically informative in symptomatic WM tion by light-chain-isotype-positive B-cells (as reflected by the patients. In this regard, cases with 410% B cells showed a trend percentage of B cells with cytoplasmatic expression of the patient- for inferior OS (Figure 1c; P ¼ 0.08), which translated into significant specific light-chain isotype within the mature B cell compartment) differences when the OS of patients with 100% versus o100% from IgM-MGUS to smoldering and symptomatic WM patients light-chain-isotype-positive B-cells was compared (Figure 1d; (median of 75%, 96% and 99%, respectively; Table 2). Conse- median OS of 44 versus 78 months; P ¼ 0.001). The later parameter quently, only 1% of IgM-MGUS patients showed full light-chain (full light chain restriction of the B cell compartment; HR: 2.6, restriction of the B cell compartment (as reflected by the absence P ¼ 0.004) together with the WM International Prognostic Scoring of B cells with cytoplasmatic expression of the uninvolved light- System (IPSSWM; HR: 2.2, P ¼ 0.006) represented the best chain), in contrast to 19 and 40% of smoldering and symptomatic combination of variables to predict for OS of symptomatic WM WM patients (Po.001). Conversely, similar percentages of BM PCs patients (Table 4). Of note, patients classified as high-risk according (among total nucleated cells) (P40.05) were found among IgM- to the IPSSWM showing 100% light-chain restricted B cells (20% of MGUS and WM patients for the percentage of BM PCs between all symptomatic WM cases) were particularly associated with a (overall median of 0.3%; Table 2). Despite this, a progressively dismal outcome (median OS of 16 months; P ¼ 0.004) (Figure 2). higher degree of light-chain-isotype-positive PCs (as reflected by Full light chain restriction of the PC compartment had no effect the percentage of PCs with cytoplasmatic expression of the both in TTP and OS of smoldering and symptomatic WM patients, patient-specific light-chain isotype within the PC compartment) respectively (data not shown). was also noted from IgM-MGUS to smoldering to symptomatic WM: median of 70, 85% and 97%, respectively (Table 2). Most Immunophenotypic characteristics of the B cell Waldenstro¨m’s interestingly, a significant correlation was found between the clone across IgM monoclonal gammopathies extent of clonality depicted in the two tumor (B cell and PC) Detailed analysis of the immunophenotypic profile of the B cell compartments (R ¼ 0.685; Pp.001). However, when evaluating the correlation between the serum concentration of the IgM compartment showed significant differences among patients with M-component and the degree of clonality at the two tumor cell IgM-MGUS, smoldering and symptomatic WM (Supplementary compartments, only the degree of clonality of the PC (and not the Table 1), either in the percentage of positive B cells expressing B cell) compartment independently correlated with the amount of individual markers (within the B cell compartment) or in the pattern ± serum IgM paraprotein (P ¼ 0.001). of antigen expression—that is, negative ( À ), bimodal ( ), dim ( þ low) and bright positive ( þ hi)—of the following markers: CD19, CD22, CD23, CD25, CD27, FMC7, HLA-DR and SmIgM. Accordingly, expression of the CD19 and CD22 pan B-cell markers was found in Impact of tumor mass and clonality by MFC on progression and lo lo survival of WM patients all cases, but the frequency of patients with CD19 and/or CD22 progressively increased from IgM-MGUS to WM. Similarly, the The above mentioned thresholds of BM B-cell infiltration (410%), percentage of B cells, which expressed SmIgM also progressively as well as the degree of clonality (100%) were highly specific to increased, whereas the proportion of FMC7 þ and HLA-DR þ cells exclude the diagnosis of IgM-MGUS, but not to identify WM decreased. In addition, the transition from IgM-MGUS to WM (either patients. This led us to investigate if the same thresholds could be smoldering or symptomatic), was associated with a progressive loss used for risk stratification. Accordingly, among smoldering WM of reactivity for CD23, accompanied by a progressively increased patients (Figure 1a), those who showed 410% B cells at baseline staining for CD25 and CD27. Overall, our results show that when the were at higher risk of progression into symptomatic (treatment majority of B cells (within the B cell compartment) are clonal (for requiring) disease as compared with cases with p10% B cells example, in WM patients), these show a characteristic Walden- (median TTP of 45 versus 145 months, respectively; P ¼ 0.01); stro¨m’s phenotype (Figure 3): CD22lo/CD23 À /CD25 þ /CD27 þ / noteworthy, an increased risk of disease progression was SmIgM þ . This Waldenstro¨m’s B cell clone is also typically particularly evident in those smoldering WM patients who showed characterized (X89% of cases) by absence of expression of CD5, full light-chain restriction of the B cell compartment versus those in CD10, CD11c and CD103 in contrast with most other mature which normal residual B cells were still detected (Figure 1b; median lymphoid malignancies except marginal zone lymphoma (Figure 2). TTP of 26 versus 145 months, respectively; Pp0.001). Multivariate analysis including both cytometric features and other variables with significant prognostic value for TTP (that is, M-component The phenotypic profile of PCs in patients with IgM monoclonal X3 g/dl and X50% lymphoplasmacytic BM cells by morphology), gammopathies showed that the presence of 100% light-chain-isotype-positive Detailed analysis of the antigenic profile of BM PCs from patients B-cells at baseline was the only parameter that retained its with IgM-MGUS, smoldering and symptomatic WM (Supplementary prognostic value—P ¼ 0.001; HR: 19.8 (3.3–118) (Table 3). We then Table 2) consistently showed a phenotype profile similar to that of

Leukemia (2014) 166 – 173 & 2014 Macmillan Publishers Limited Immunophenotyping of IgM monoclonal gammopathies B Paiva et al 169

Figure 1. TTP of smoldering WM patients and OS of symptomatic WM patients according to the BM tumor burden and degree of clonality assessed by MFC at diagnosis. In panels (a and c) TTP and OS from the moment of diagnosis of smoldering and symptomatic WM patients according to the percentage of BM B cells is shown, respectively. In panels (b and d) TTP and OS of smoldering and symptomatic WM patients according to the extent of light-chain-isotype-positive B-cells in the B-cell compartment is displayed.

Table 3. Univariate and multivariate analysis with predictive markers for TTP of smoldering WM patients into active disease

Univariate analysis Multivariate analysis

Median TTP Risk at 2-years P-value Hazard ratio (95% CI) P-value

M-componentX3 g/dl 67 m 27% 0.02 1.8 (0.5–6.1) 0.34 BM lymphoplasmacytic cellsX50% 45 m 43% 0.007 1.2 (0.2–5.7) 0.82 410% B cells by MFC 45 m 25% 0.01 1.5 (0.4–5.6) 0.55 100% light-chain-isotype-positive B-cells by MFC 26 m 42% o0.001 19.8 (3.3–118) 0.001 Abbreviations: BM, bone marrow; CI, confidence interval; MFC, multiparameter flow cytometry; TTP, time to progression. normal BM PCs, and clearly different from that of myeloma CD20 þ and CD45 þ and SmIgM þ cells, together with a higher patients (for example, CD56 þ hi,CD19À ,CD27À and/or CD45 À ), proportion (68% versus 32%) of symptomatic WM versus even if increasing percentages of light-chain-isotype-positive PCs IgM-MGUS patients showing complete lack of reactivity for CD56 were found from IgM-MGUS to WM. However, significant differ- on BM PCs. Overall, these results suggest that during the transition ences (either in the percentage of antigen-expressing positive PCs from IgM-MGUS to smoldering and symptomatic WM (in which the within the PC compartment or in the pattern of phenotypic majority of PCs are light-chain restricted), the PC compartment is expression) were noted between the three groups of patients for 5 enriched in clonal PCs displaying an immature/plasmablastic of the 23 antigens analysed. Thus, from IgM-MGUS to WM, patients phenotype with restricted isotype of both the Ig heavy and light BM PCs showed a progressively higher frequency of CD19 þ , chains (Figure 4).

& 2014 Macmillan Publishers Limited Leukemia (2014) 166 – 173 Immunophenotyping of IgM monoclonal gammopathies B Paiva et al 170 BM evaluation in patients suspicious of having WM, even in the Table 4. Univariate and multivariate analysis with predictive markers era of genomics.2,3,25 Of note, this does not apply in MM, where for OS of symptomatic WM patients there is flexibility between BM aspirate or biopsy for diagnostic 26 Univariate Multivariate purposes; in fact, in recent years a growing body of evidence has analysis analysis emerged, which supports the added value of MFC for the evaluation of the BM aspirate samples to discriminate between 27 Median OS P-value Hazard ratio P-value (non-IgM) MGUS and MM, and to identify patients with MGUS or (95% CI) smoldering MM at different risk of progression into symptomatic disease.15 IPSSWM 2.2 (1.3–3.9) 0.006 In the present study, we performed an extensive phenotypic Low Not reached 0.017 Intermediate 78 m characterization of BM B cells and PCs from a large series of newly High 46 m diagnosed patients with an IgM M-component, covering the full spectrum of the disease: IgM-MGUS, smoldering and symptomatic % of B cells by MFC 1.5 (0.8–3.1) 0.22 WM. MFC evaluation of the BM aspirate revealed progressively o10 78 m 0.08 X10 46 m increasing numbers of B cells from IgM-MGUS to WM: while only 1% of IgM-MGUS patients showed at baseline 410% B cells, one % of light-chain- 2.6 (1.4–5.0) 0.004 third of smoldering WM and half of symptomatic WM patients isotype-positive displayed 410% BM B cells. Of note, the increase in the overall B-cells by MFC o100 78 m 0.001 percentage of BM B cells was associated with a progressively 100 44 m higher number of light-chain-isotype-positive B-cells from IgM- MGUS to smoldering and symptomatic WM; as a consequence, Abbreviations: CI, confidence interval; IPSSWM: International prognostic complete full light-chain restriction of the mature B cell scoring system for Waldenstrom’s Macroglobulinemia; MFC: multipara- meter flow cytometry; OS, overall survival. compartment (absence of residual normal B cells) was virtually undetectable in IgM-MGUS patients, whereas it was present in around one-fifth and one-half of smoldering and symptomatic WM patients, respectively. Thus, although the presence of higher tumor burden and a more prominent clonal excess could virtually exclude the diagnosis of IgM-MGUS, no MFC-based criteria could completely differentiate between IgM-MGUS and WM. These results are in line with previous observations, which suggest a substantial overlap between the three stages of the disease,2,3 including also its genetic features. In this regard, the very low frequency of MYD88 L265P initially reported in IgM-MGUS (10%)25 suggested to be of potential utility for the differential diagnosis with WM, but this could not be confirmed in more recent studies using sensitive ASO-PCR, which show that the frequency of mutated MYD88 can be typically found in at least half of IgM-MGUS patients.28–30 Overall, these results support that similarly to MM,7 also in WM there may be a continuum between IgM-MGUS and WM, rather than these entities being considered as separate. We have further investigated the potential role of MFC for prognostication in WM patients. Thus, smoldering WM patients with 410% BM B cells and complete light-chain restricted B cells Figure 2. OS of symptomatic WM patients with IPSSWM high-risk were at higher risk for progression then other smoldering WM disease according to degree of clonality in the BM B cell cases. Conversely, symptomatic WM patients showing o10% B compartment assessed by MFC at diagnosis. cells in association with persistent residual normal B cells were characterized by less aggressive disease and superior OS. Notably, complete light-chain restriction of the B-cell compartment DISCUSSION identified a subset of high-risk smoldering WM patients that WM is a relatively rare and heterogeneous disease, both at the evolve into symptomatic disease in B2 years after clinical cellular level (tumor cells range from small mature-appearing B recognition. Accordingly, we have identified a new prognostic cells to lymphoplasmocytes and PCs), and at the clinical level marker in smoldering WM that adds up to a rather short list, (intermediate features between non-Hodgkin and namely the degree of BM infiltration, hemoglobin, b2 micro- monoclonal gammopathies with both indolent and aggressive globuin and IgM paraprotein values.4,13,31 In a multivariate model forms). This is partially reflected on the distinct criteria currently including the later risk factors, full light-chain restriction of the B used or proposed by different groups to define its diagnosis. In cell compartment showed independent predictive value; the same fact, WM is defined by the WHO as a subtype of lymphoplasma- applied in symptomatic WM patients where this marker predicted cytic lymphoma with BM involvement and an IgM monoclonal for inferior OS independently of the IPSSWM.32 Moreover, by protein of any concentration.1 In turn, the Second International integrating the IPSSWM with the clonal extent found in B cells we Workshop on WM recommended that the disease should be were able to identify a subset (20%) of patients with a dismal regarded as a distinct clinicalpathological entity and not a clinical outcome and a median OS of only 16 months; those with high-risk syndrome secondary to IgM secretion.2 Because of this, the disease by the IPSSWM plus 100% light-chain-isotype-positive concentration of serum IgM becomes irrelevant for diagnostic B-cells. These results suggest that similarly to MGUS, MM or AL purposes if there is evidence of BM infiltration by amyloidosis,5,15,16,33–36 the degree of clonal excess (for example, lymphoplasmacytic lymphoma;2 this slightly differs from the balance between malignant and residual normal cells) is at least as Mayo Clinic criteria that requires X10% BM involvement.7 In effective as the overall evaluation of the total tumor burden in any case, it should be highlighted the relevance of trephine biopsy predicting outcome.

Leukemia (2014) 166 – 173 & 2014 Macmillan Publishers Limited Immunophenotyping of IgM monoclonal gammopathies B Paiva et al 171

Figure 3. Immunophenotypic characterization of the Waldenstro¨ m’s B cell clone. In panels (a and c) a representative dot-plot for surface IgM versus CD27 and CD22 versus CD25 of the B cell clone from a patient with WM is shown, respectively. In panels (b and d) the WM B cell clone is represented by one and two s.d. (white) curves together with residual normal BM B cells and PCs. Panels (e and f) reflect the immunophenotype of the B cell clone found in X89% of WM patient for CD5, CD10, CD11c and CD103 versus a schematic representation of the typical phenotype present in (HCL), Follicular lymphoma (FL), Burkitt’s lymphoma (BL), B-cell chronic lymphocytic leukemia (B-CLL) and (MCL).

Already 410 years ago the Second International Workshop on regulation of CD23 together with increased CD27) up to the stage WM agreed in that the diagnosis of WM should be supported by of a lymphoplasmacytic cell (for example CD19 þ lo and HLA- immunophenotypic studies.2 However, a clear consensus on the DR þ lo), which show a characteristic Waldenstro¨m’s phenotype: immunophenotypic features of the WM clone has not been SmIgM þ /CD22 þ lo/CD25 þ . Moreover, in the vast majority of WM reached so far. This may be partially explained by the low number patients clonal B cells were CD5 À (94%) and CD10 À (98%). This of patients usually included in most studies (typically o50) and phenotype differs from other B-cell lymphoproliferative disorders the different methodological approaches (immunohistochemistry that may show an IgM paraprotein and some degree of overlap for versus MFC), gating strategies (B cells versus lymphoplasmocytes the MYD88 mutation with WM,25,37–39 namely B cell chronic versus PCs) or monoclonal antibodies used.17–20,22,23 To the best of lymphocytic leukemia and mantle cell lymphoma (typically our knowledge, here we report on the largest series of WM CD5 þ hi) or follicular lymphoma (usually CD10 þ ). Moreover, in patients characterized on immunophenotypic grounds by MFC, the small fraction of patients in which CD5 þ hi was noted, other which also includes for the first time IgM-MGUS cases. Overall, our hallmarks of the B-cell chronic lymphocytic leukemia phenotype results show a continuously evolving immunophenotypic pattern (CD20 þ lo, sIg þ lo or FMC7 À ) were absent. By contrast, the from IgM-MGUS to smoldering and symptomatic WM. Such Waldenstro¨m’s clone showed some degree of phenotypic phenotypic profile is consistent with a progressive accumulation overlap with that of marginal zone lymphoma. This entity is, of light-chain restricted mature B cells (for example, down- however, characterized by dimmer expression of surface IgM,

& 2014 Macmillan Publishers Limited Leukemia (2014) 166 – 173 Immunophenotyping of IgM monoclonal gammopathies B Paiva et al 172

Figure 4. Immunophenotypic characterization of the Waldenstro¨ m’s PC clone. In panels (a and b) a representative dot-plot for surface CD19 versus HLA-DR and CD45 versus CD38 of the whole-PC clone (plasmablasts plus mature PCs) and B cell clone (represented by one and two s.d. (white) curves) from a patient with WM is shown. Panels (c and d) reflect the immunophenotype of the PC clone of IgM MUS and WM patients for surface IgM, CD27, CD38 and CD56 versus a schematic representation of the typical phenotype present in MM.

heterogeneous CD25 and bright staining for CD22,18 but these extent of clonal involvement of the BM PC compartment (but subtle differences may be difficult to interpret and mainly depend not that of mature B cells) directly correlated with the serum on the experience and visual criteria of the observer. Further concentration of the IgM paraprotein, in line with previous investigations with the input of novel polychromatic cytometry observations.42 Moreover, as the percentage of BM PCs remained and multidimensional analysis of the flow data are thus stable across the three disease stages, a correlation between these warranted.14 This input should increase its sensitivity to clearly and serum IgM concentration was obviously not found. One can depict the WM clone in IgM-MGUS patients (possible in only 12% speculate whether this is a result from peripheral of the cases using conventional 4-color MFC), as well as to contamination of BM specimens or due to specific PC loss elucidate if the phenotypic differences between IgM-MGUS during sample preparation. Alternatively, this observation could and WM patients as regards the B cell compartment are reflect an intrinsic feature of WM biology; in fact, increasing related to specific-disease-progression events, or rather to an numbers of BM PCs are typically detected by MFC in MM as expansion of the WM clone from one stage to the other compared with MGUS patients.15,16 (as depicted previously by the progressive increment of light- In summary, our results support the use of MFC immunophe- chain-isotype-positive B-cells). notyping to evaluate BM aspirates from patients with suspected Despite the progressive accumulation of light-chain-isotype- WM. This technique affords information for the differential positive B-cells from IgM-MGUS to WM, the overall percentage of diagnosis of IgM-MGUS, risk of transformation in smoldering BM PCs remained stable across the three disease stages. However, disease and survival in symptomatic WM. We have also unraveled within the PC compartment there was a progressive increase in the specific phenotypic characteristics of clonal B cells and PCs, the proportion of light-chain-isotype-positive PCs from IgM-MGUS and this information should now be used in the setting of novel to WM. Such increasingly more pronounced clonal excess was multidimensional flow cytometry. associated by a more immature/plasmablastic phenotype of BM PCs, as reflected by a greater reactivity for CD19, CD20 CD45 and SmIgM40), always in the absence of myeloma-related phenotypic CONFLICT OF INTEREST aberrations (for example, upregulation of CD56 or absence of The authors declare no conflict of interest. CD27).15,16 Consequently, the phenotype of clonal PCs from WM patients is clearly different from that of MM patients. These findings indicate that MFC immunophenotyping can also be of ACKNOWLEDGEMENTS great utility to differentiate between WM and IgM-MM (particularly 41 This work was supported by the Cooperative Research Thematic Network (RTICC) in patients with t(11;14)), as well as to distinguish true WM from RD06/0020/0006, Instituto de Salud Carlos III/Subdireccio´n General de Investigacio´n multiple unrelated diseases (for example, B-cell chronic Sanitaria (FIS: PI060339; 06/1354; 02/0905; 01/0089/01-02; PS09/01897/01370) and lymphocytic leukemia with concurrent MGUS). Notably, the Consejerı´a de Sanidad, Junta de Castilla y Leo´n, Valladolid, Spain (557/A/10).

Leukemia (2014) 166 – 173 & 2014 Macmillan Publishers Limited Immunophenotyping of IgM monoclonal gammopathies B Paiva et al 173 AUTHOR CONTRIBUTIONS 20 Konoplev S, Medeiros LJ, Bueso-Ramos CE, Jorgensen JL, Lin P. Immunopheno- JFSM and BP conceived the idea and together with RGS, EMO, AO and MBV typic profile of lymphoplasmacytic lymphoma/Waldenstrom macroglobulinemia. designed the study protocol; BP, MCM, JJP and MBV analyzed the flow cytometry Am J Clin Pathol 2005; 124: 414–420. data; RGS, EMO, JA, NdH, FE, RC, AGdC, JG, JH, MS, AM, EP, AB, JA, LS, TJGL and 21 Merli M, Arcaini L, Boveri E, Rattotti S, Picone C, Passamonti F et al. Assessment of bone marrow involvement in non-Hodgkin’s lymphomas: comparison between JFSM contributed with provision of study material or patients; BP and JFSM histology and flow cytometry. Eur J Haematol 2010; 85: 405–415. analyzed and interpreted data; BP performed statistical analysis; BP and JFSM 22 Morice WG, Chen D, Kurtin PJ, Hanson CA, McPhail ED. Novel immunophenotypic wrote the manuscript. All authors reviewed and approved the manuscript. features of marrow lymphoplasmacytic lymphoma and correlation with Wal- denstrom’s macroglobulinemia. Mod Pathol 2009; 22: 807–816. 23 Remstein ED, Hanson CA, Kyle RA, Hodnefield JM, Kurtin PJ. Despite apparent REFERENCES morphologic and immunophenotypic heterogeneity, Waldenstrom’s macro- 1 Swerdlow S, Campos E, Harris N, Jaffe E, Pileri S, Stein H et al. WHO Classification of globulinemia is consistently composed of cells along a morphologic continuum Tumors of Hematopoietic and Lymphoid Tissues. IARC: Lyon, 2008. of small lymphocytes, plasmacytoid lymphocytes, and plasma cells. Semin Oncol 2 Owen RG, Treon SP, Al-Katib A, Fonseca R, Greipp PR, McMaster ML et al. Clin- 2003; 30: 182–186. icopathological definition of Waldenstrom’s macroglobulinemia: consensus panel 24 Kyle RA, Treon SP, Alexanian R, Barlogie B, Bjorkholm M, Dhodapkar M et al. recommendations from the Second International Workshop on Waldenstrom’s Prognostic markers and criteria to initiate therapy in Waldenstrom’s macro- Macroglobulinemia. Semin Oncol 2003; 30: 110–115. globulinemia: consensus panel recommendations from the Second International 3 Gertz MA. Waldenstrom macroglobulinemia: 2012 update on diagnosis, risk Workshop on Waldenstrom’s Macroglobulinemia. Semin Oncol 2003; 30: 116–120. stratification, and management. Am J Hematol 2012; 87: 503–510. 25 Treon SP, Xu L, Yang G, Zhou Y, Liu X, Cao Y et al. MYD88 L265P somatic mutation 4 Kyle RA, Benson JT, Larson DR, Therneau TM, Dispenzieri A, Kumar S et al. Pro- in Waldenstrom’s macroglobulinemia. N Engl J Med 2012; 367: 826–833. gression in smoldering Waldenstrom macroglobulinemia: long-term results. Blood 26 Dimopoulos M, Kyle R, Fermand JP, Rajkumar SV, San Miguel J, Chanan-Khan A et al. 2012; 119: 4462–4466. Consensus recommendations for standard investigative workup: report 5 Kyle RA, Therneau TM, Rajkumar SV, Remstein ED, Offord JR, Larson DR et al. Long- of the International Myeloma Workshop Consensus Panel 3. Blood 2011; 117: term follow-up of IgM monoclonal gammopathy of undetermined significance. 4701–4705. Blood 2003; 102: 3759–3764. 27 Ocqueteau M, Orfao A, Almeida J, Blade J, Gonzalez M, Garcia-Sanz R et al. 6 Kyle RA, Rajkumar SV. Criteria for diagnosis, staging, risk stratification and Immunophenotypic characterization of plasma cells from monoclonal response assessment of multiple myeloma. Leukemia 2009; 23: 3–9. gammopathy of undetermined significance patients. Implications for the differ- 7 Weiss BM, Abadie J, Verma P, Howard RS, Kuehl WM. A monoclonal gammopathy ential diagnosis between MGUS and multiple myeloma. Am J Pathol 1998; 152: precedes multiple myeloma in most patients. Blood 2009; 113: 5418–5422. 1655–1665. 8 Kyle RA, Benson J, Larson D, Therneau T, Dispenzieri A, Melton Iii LJ et al. IgM 28 Landgren O, Staudt L. MYD88 L265P somatic mutation in IgM MGUS. N Engl J Med monoclonal gammopathy of undetermined significance and smoldering 2012; 367: 2255–2257. Waldenstrom’s macroglobulinemia. Clin Lymphoma Myeloma Leuk 2009; 9: 17–18. 29 Varettoni M, Arcaini L, Zibellini S, Boveri E, Rattotti S, Riboni R et al. Prevalence 9 Kyle RA, Therneau TM, Rajkumar SV, Offord JR, Larson DR, Plevak MF et al. and clinical significance of the MYD88 (L265P) somatic mutation in Walden- A long-term study of prognosis in monoclonal gammopathy of undetermined stro¨m’s macroglobulinemia and related lymphoid neoplasms. Blood 2013; 121: significance. N Engl J Med 2002; 346: 564–569. 2522–2528. 10 McMaster ML, Caporaso N. Waldenstrom macroglobulinaemia and IgM 30 Xu L, Hunter ZR, Yang G, Zhou Y, Cao Y, Liu X et al. MYD88 L265P in monoclonal gammopathy of undetermined significance: emerging under- Waldenstro¨m’s macroglobulinemia, IgM monoclonal gammopathy, and other standing of a potential precursor condition. Br J Haematol 2007; 139: 663–671. B-cell lymphoproliferative disorders using conventional and quantitative 11 Schop RF, Van Wier SA, Xu R, Ghobrial I, Ahmann GJ, Greipp PR et al. 6q deletion allele-specific PCR. Blood 2013; 121: 2051–2058. discriminates Waldenstrom macroglobulinemia from IgM monoclonal gammo- 31 Garcia-Sanz R, Montoto S, Torrequebrada A, de Coca AG, Petit J, Sureda A et al. pathy of undetermined significance. Cancer Genet Cytogenet 2006; 169: 150–153. Waldenstrom macroglobulinaemia: presenting features and outcome in a series 12 Baldini L, Goldaniga M, Guffanti A, Broglia C, Cortelazzo S, Rossi A et al. Immu- with 217 cases. Br J Haematol 2001; 115: 575–582. noglobulin M monoclonal gammopathies of undetermined significance and 32 Morel P, Duhamel A, Gobbi P, Dimopoulos MA, Dhodapkar MV, McCoy J et al. indolent Waldenstrom’s macroglobulinemia recognize the same determinants of International prognostic scoring system for Waldenstrom macroglobulinemia. evolution into symptomatic lymphoid disorders: proposal for a common prog- Blood 2009; 113: 4163–4170. nostic scoring system. J Clin Oncol 2005; 23: 4662–4668. 33 Paiva B, Vidriales MB, Perez JJ, Lopez-Berges MC, Garcia-Sanz R, Ocio EM et al. 13 Dimopoulos MA, Panayiotidis P, Moulopoulos LA, Sfikakis P, Dalakas M. Walden- The clinical utility and prognostic value of multiparameter flow cytometry strom’s macroglobulinemia: clinical features, complications, and management. immunophenotyping in light-chain amyloidosis. Blood 2011; 117: 3613–3616. J Clin Oncol 2000; 18: 214–226. 34 Zhan F, Barlogie B, Arzoumanian V, Huang Y, Williams DR, Hollmig K et al. 14 van Dongen JJ, Orfao A. EuroFlow: resetting leukemia and lymphoma Gene-expression signature of benign monoclonal gammopathy evident in immunophenotyping. Basis for companion diagnostics and personalized medi- multiple myeloma is linked to good prognosis. Blood 2007; 109: 1692–1700. cine. Leukemia 2012; 26: 1899–1907. 35 Larsen JT, Kumar SK, Dispenzieri A, Kyle RA, Katzmann JA, Rajkumar SV. Serum 15 Perez-Persona E, Vidriales MB, Mateo G, Garcia-Sanz R, Mateos MV, free light chain ratio as a biomarker for high-risk smoldering multiple myeloma. de Coca AG et al. New criteria to identify risk of progression in monoclonal Leukemia 2012; 27: 941–946. gammopathy of uncertain significance and smoldering multiple myeloma based 36 Katzmann JA, Clark R, Kyle RA, Larson DR, Therneau TM, Melton 3rd LJ et al. on multiparameter flow cytometry analysis of bone marrow plasma cells. Blood Suppression of uninvolved immunoglobulins defined by heavy/light 2007; 110: 2586–2592. chain pair suppression is a risk factor for progression of MGUS. Leukemia 2012; 27: 16 Paiva B, Vidriales MB, Mateo G, Perez JJ, Montalban MA, Sureda A et al. The 208–212. persistence of immunophenotypically normal residual bone marrow plasma cells 37 Ngo VN, Young RM, Schmitz R, Jhavar S, Xiao W, Lim KH et al. Oncogenically active at diagnosis identifies a good prognostic subgroup of symptomatic multiple MYD88 mutations in human lymphoma. Nature 2011; 470: 115–119. myeloma patients. Blood 2009; 114: 4369–4372. 38 Pasqualucci L, Trifonov V, Fabbri G. Analysis of the coding genome of diffuse large 17 San Miguel JF, Vidriales MB, Ocio E, Mateo G, Sanchez-Guijo F, Sanchez ML et al. B-cell lymphoma. Nat Genet 2011; 43: 830–837. Immunophenotypic analysis of Waldenstrom’s macroglobulinemia. Semin Oncol 39 Puente X, Pinol M, Quesada V. Whole-genome sequencing identifies recurrent 2003; 30: 187–195. mutations in chronic lymphocytic leukaemia. Nature 2011; 470: 101–105. 18 Ocio EM, Hernandez JM, Mateo G, Sanchez ML, Gonzalez B, Vidriales B et al. 40 Perez-Andres M, Paiva B, Nieto WG, Caraux A, Schmitz A, Almeida J et al. Human Immunophenotypic and cytogenetic comparison of Waldenstrom’s peripheral blood B-cell compartments: a crossroad in B-cell traffic. Cytometry B macroglobulinemia with splenic marginal zone lymphoma. Clin Lymphoma 2005; Clin Cytom 2010; 78(Suppl 1): S47–S60. 5: 241–245. 41 Owen R, Feyler S, O’Connor S, Bond L, De Tute R, Rawstron A. Defining IgM 19 Howard MT, Hodnefield J, Morice WG. Immunohistochemical phenotyping of multiple myeloma. Am J Hematol 2011 717. plasma cells in lymphoplasmacytic lymphoma/Waldenstrom’s macroglobulinemia 42 Pasricha SR, Juneja SK, Westerman DA, Came NA. Bone-marrow plasma cell is comparable to flow cytometric techniques. Clin Lymphoma Myeloma Leuk 2011; burden correlates with IgM paraprotein concentration in Waldenstrom macro- 11: 96–98. globulinaemia. J Clin Pathol 2011; 64: 520–523.

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

& 2014 Macmillan Publishers Limited Leukemia (2014) 166 – 173