CD106 and activated-CD29 are expressed on myelomatous bone marrow plasma cells and their downregulation is associated with tumour progression

CD106 and activated-CD29 are expressed on myelomatous

bone marrow plasma cells and their downregulation

is associated with tumour progression

Rosario Luque,1

Jose Antonio Garcıa-Trujillo,1

Carmen Camara,1

Ana Moreno,1

Pablo Eiras,1

Garbine Roy,1

Luisa Marıa Villar,1

Manuel Lombardıa,1

Jose Antonio Brieva,2

Alfredo Bootello1

and Ernesto Roldan1 1Servicio de Inmunologıa, Hospital Ramon y Cajal, Madrid, and 2Servicio de Inmunologıa,

Hospital Puerta del Mar, Cadiz, Spain

Received 19 March 2002; accepted for publication 1 May 2002

Summary. Malignant plasma cells (PC) from multiplemyeloma (MM) patients characteristically home to the bonemarrow (BM). High numbers of tumour cells are found inthe peripheral blood (PB) only at end-stage disease (secon-dary plasma cell leukaemia, PCL) in a minority of patients.Using flow cytometric and fluorescence in situ hybridization(FISH) analysis, a high percentage of tumoral BM PC fromuntreated patients was found to express CD106. In addition,these cells also expressed an activated form of CD29, asdetermined using the CD29 activation reporter monoclonalantibody HUTS-21. Adhesion-binding experiments showedthat CD106+-activated CD29+ BM PC from these patientsadhered to fibronectin (FN) in a CD29/CD49d-dependentmanner. In contrast, marrow PC from progressive patients

and BM or circulating malignant cells from secondary PCLpatients expressed lower levels or were negative for CD106and activated CD29, respectively, with a decreased or zeroability to adhere to FN. The expression of constitutive CD29and CD49d, however, was similar during disease progres-sion. We conclude that BM myelomatous cells co-expressCD106 and a functionally active form of CD29. Moreover,our results suggest that the loss of expression and/or func-tion of these antigens are associated with the progressionof MM and may explain the exit of tumoral cells from theBM.

Keywords: myeloma, CD106, activated CD29, progression,adhesion.

In human multiple myeloma (MM), malignant plasma cells(PC) are retained within the bone marrow (BM) microen-vironment. This preferential localization of myelomatouscells, which defines the MM stage of this disease, has beenpartially explained by their adhesion molecule expressionpattern (Van Camp et al, 1990; Witzig, 1999). Among thesemolecules, CD49d/CD29 (a4/b1 integrin) is alwaysexpressed on all malignant PC (Kim et al, 1994; Pellat-Deceunynck et al, 1995; Luque et al, 1998). This finding iscompatible with the fact that human myeloma-derived celllines specifically adhere to fibronectin (FN) through thisintegrin (Uchiyama et al, 1992), which may contribute tolocalization of malignant cells in the BM. However, thisadhesive profile does not explain by itself this phenomenon,because other tumoral cells express either similar or

identical adhesion structures and are not preferentiallylocalized within the BM (Reuss-Borst et al, 1992). In thisregard, the expression of an unusual adhesion molecule onmyeloma cells could explain their selective retention atthis site. Moreover, a massive exit of malignant cells fromBM to PB occurs during disease progression toward secon-dary plasma cell leukaemia (PCL) (Kosmo & Gale, 1987).Although it has been reported recently that the absenceof CD56 on malignant PC is a characteristic feature ofsecondary PCL (Pellat-Deceunynck et al, 1995, 1998), themechanisms responsible for the leukaemic phase in MM arepoorly understood (Kalasz et al, 1989). For the aforemen-tioned reasons, it is supposed that the CD29/CD49d integrinplays an essential role in the BM retention of myelomatouscells. However, the expression of this integrin on marrow PCfrom MM and PCL patients is very similar (Pellat-Deceunynck et al, 1995). In spite of this fact, because thefunction of integrins is activation dependent (Hynes, 1992)and their expression per se does not reflect their function,CD29/CD49d involvement in the BM retention of PC cannot

Correspondence: Ernesto Roldan, Servicio de Inmunologıa, HospitalRamon y Cajal, Ctra. Colmenar Km. 9.100, 28034 Madrid, Spain.

E-mail: [email protected]

British Journal of Haematology, 2002, 119, 70–78

70 � 2002 Blackwell Publishing Ltd

be discarded. Therefore, we have extensively studied theunexpected expression of certain adhesion molecules in MMand/or the activation state of integrins in order to under-stand and predict the biological behaviour of this disease.Here, it is reported that CD106 and an active form of CD29are expressed on malignant BM cells in the MM stage of thedisease, but not on circulating or marrow PC from PCLpatients. In this regard, CD106 and activated CD29 could beconsidered to be excellent markers to study the progressionof MM.

PATIENTS AND METHODS

Patients. BM and PB samples from 75 patients with MMwere analysed. The diagnosis and clinical staging of thesepatients were made on the basis of clinical features, haema-tological characteristics and immunophenotyping (Durie &Salmon, 1975; Pellat-Deceunynck et al, 1994). Patients instages II and III received either melphalan at 8 mg/m2 for5 d and prednisone at 60 mg/m2 for 5 d or cyclophos-phamide at 600 mg/m2, adriamycin at 30 mg/m2, 1Æ3-bis(2-chloroethyl)-1-nitrosourea (BCNU) at 30 mg/m2,and prednisone at 60 mg/m2 with or without a-inter-feron. Eighteen MM patients underwent PB stem celltransplants. Patients were grouped as either untreated oraccording to whether they had progressive disease (thiscategory includes patients in relapse or with refractorydisease). Four patients with secondary plasma cell leuk-aemia (PCL) and 15 healthy donors were also studied.Cytomorphological studies of PC in these PCL patients didnot reveal any differences with respect to those of classicMM. The patients’ median follow-up was 31 months(range, 2–108 months). All patients and controls wereinformed about the objectives and methods of the study,and gave their consent.

Staining, detection and isolation of BM and PB PC. Thefollowing monoclonal antibodies (mAbs) were used for flowcytometric analyses: phycoerythrin (PE)- and fluoresceinisothiocyanate (FITC)-labelled anti-CD106 (vascular celladhesion molecule 1,VCAM-1) (clone 1G11) (CymbusBioscience, Southampton, UK); anti-CD106-PE (51–10C9),purified anti-CD106 (E1/6), anti-CD49d-PE (L25), anti-CD38-CyChrome and anti-CD29-PE mAbs (clones MAR4and HUTS-21) (Becton Dickinson Biosciences, San Jose, CA,USA); anti-CD38-FITC and purified anti-CD106 (F(ab¢)2

fragment, clone 1G11) (Caltag, San Francisco, CA, USA);purified anti-CD138 (ImmunoQuality Products, Groningen,The Netherlands); purified anti-CD49d (HP2/1), purifiedanti-CD29 (4B4) and mouse IgG (679Æ1Mc7) (Immunotech,Marseille, France). Malignant (or normal) BM or PB PC weredefined as CD38++ (Pellat-Deceunynck et al, 1994). PC wereidentified on a dot-plot display with side scatter (SSC) vsgreen fluorescence as the two parameters. Additional flowcytometry analysis with the CD138 mAb confirmed thatalmost all BM or PB PC were positive for this specific PCantigen (Luque et al, 1998). The expression of adhesionmolecules were analysed by two- or three-colour flowcytometry and quantified both in terms of frequency(percentage) and of mean fluorescence intensity (MFI), as

previously described (Luque et al, 1998). BM monocyteswere identified as CD14+ cells with intermediate SSC signal.

BM or PB myelomatous cells were isolated with the DynalRAM IgG1 CollectionTM kit using Dynabeads (M-280)(Dynal, Oslo, Norway) coated with a rat mAb againstmouse IgG1, according with the manufacturers instruc-tions. Briefly, anti-CD138 mAb was incubated with thebeads (5 lg/107 beads). Then, the mAb anti-CD138-coatedbeads were incubated with unseparated BM or PB samples.The positive fraction (CD138+ cells) was obtained byrepeated washes on a magnet and was subsequentlyreleased from the beads with a releasing buffer containingDNase. Purity of PC was > 95% in all cases.

Fluorescence in situ hybridization (FISH). Cells were fixedto poly l-lysine coated slides with 4% paraformaldehyde(PFA) for 30 min and washed twice (1 · phosphate-buffered saline, PBS) for 5 min. Non-specific binding ofprobes to positively charged amino groups was preventedby acetylation of these residues with freshly mixed 0Æ25%acetic anhydride in 0Æ1 mol/l triethanolamine HCl, pH 8 for10 min at room temperature (RT). After that, the slideswere washed briefly in 2 · saline-sodium citrate (SSC) anddehydrated by passing them for 3 min each through agraded series of ethanol concentrations (50, 70, 90 and100%). CD106 mRNA hybridization was performed using acocktail of six commercially available biotinylated 3¢- and5¢-antisense oligodeoxynucleotides (R & D Systems, Abing-don, UK) which were previously mixed with yeast tRNA(0Æ5 mg) and the appropriate amount of the hybridizationmixture (50% formamide, 10% dextran sulphate, 1 ·Denhardt’s solution, 0Æ3 mol/l NaCl, 0Æ02 mol/l Tris-HCl,50 mmol/l EDTA, pH 8) and incubated at 37�C overnight.Slides were washed at high stringency. Probes weredetected with PE-labelled streptavidin (Becton Dickinson)diluted 1:100 in Tris-buffered saline (TBS) at 37�C for30 min. Cells were visualized and photographed using aLeitz microscope.

Adhesion assay. Binding of PC to human fibronectin (FN)(Sigma Chemicals, St Louis, MO, USA) was assessed asdescribed elsewhere (Goodwin & Pauli, 1995). Ninety-sixwell flat-bottomed plates were incubated with 2 lg/well ofFN overnight at 4�C. Wells were subsequently blocked withCa+2/Mg2 ± free PBS/2% human serum albumin (HSA) for2 h at RT to reduce non-specific attachment. PC from bloodor marrow samples, which were biotinylated at RT for20 min by adding 0Æ5 mg/ml of NHS-LC-biotin (Pierce,Rockford, IL, USA) in HEPES/NaCl buffer containing1 mmol/l Mg2+ and 1 mmol/l Ca+2, were added to wells(105 cells/100 ll) and allowed to sediment onto the bottomof well for 10 min at 4�C. Plates were then transferred to aCO2 incubator and incubated for 45 min at 37�C. Afterthat, 200 ll of Percoll floatation medium (density 1Æ10 g/ml)was slowly added. Next, 50 ll of fixative (final concentrationof glutaraldehyde 1Æ4%) was added for 30 min at RT.Unbound cells were removed by three washing steps. Forblocking experiments, biotin-labelled PC were pretreated for30 min at 4�C with HP2/1 (anti-CD49d), 4B4 (anti-CD29)mAb or with an irrelevant mouse IgG mAb. After thesequential addition of peroxidase-labelled streptavidin

Myelomatous Cells Express CD106 and Activated CD29 71

� 2002 Blackwell Publishing Ltd, British Journal of Haematology 119: 70–78

(Pierce) and OPD, cell adhesion was spectrophotometricallyquantified using an enzyme-linked immunosorbent assayreader.

Statistical analysis. The relationships between the expres-sion of adhesion molecules and quantitative characteristicsof the patients were studied by the Spearman’s rankcorrelation test. The relationships between expression ofadhesion proteins and qualitative parameters were studiedby the Mann–Whitney test. All P-values indicated wereconsidered as statistically significant if P < 0Æ05.

RESULTS

Malignant BM cells from MM patients express CD106CD106 expression on malignant PC was determined in 60cases of untreated MM using the mAb 1G11. As can be seenin Fig 1(A and B), the majority of malignant BM PC wereCD106+ (84 ± 8%; MFI: 310 ± 132). Moreover, the stain-ing pattern was independent of FcR binding as thepercentage of CD106+ PC was similar using either complete1G11 mAb or its F(ab¢)2 fragments (75 ± 11%; n ¼ 19). Tosubstantiate the finding of CD106 expression on myeloma-tous BM cells, two additional anti-CD106 mAbs, 51–10C9and E1/6 (Bevilacqua et al, 1995), were used: both mAbsalso reacted with malignant BM PC in similar proportions as

found with 1G11 (74 ± 7% and 69 ± 11% respectively;n ¼ 31), although the MFI tended to be lower than thatobtained with 1G11 (data not shown). As a reliable internalpositive control, a high percentage of BM monocytes wereCD106+ (77 ± 13%; MFI: 739 ± 186; n ¼ 5) (Fig 1C andD). To confirm these findings, a FISH study was performed.As shown in Fig 1E, malignant BM PC contained detectablelevels of CD106 mRNA transcripts, data consistent withthat shown in Fig 1B.

Although MM has long been regarded as a neoplasialocalized to the BM, a very small number of circulating PC inthe majority of MM patients has been described (Billadeauet al, 1996). Based on these findings, we investigated as towhether this commonly observed minor circulating PCsubset was related to quantitative changes in the expressionof CD106. As can be seen in Fig 1F, these circulating PCwere CD106– (6 ± 3%; MFI: 74 ± 27; n ¼ 40).

On the one hand, we next addressed the questionwhether the CD106 expression could be helpful in distin-guishing between normal and malignant BM PC. As shownin Fig 1G, CD106 was not restricted to malignant cellsbecause this antigen was similarly detected on a largeproportion of normal BM (71 ± 12%; MFI: 301 ± 129;n ¼ 15). Therefore, CD106 expression on human PC maynot be associated with neoplasia.

Fig 1. CD106 is differentially expressed on malignant PC from marrow and PB compartments. Gating protocols (A and C) have been previously

described in Materials and methods. Biparametric fluorescence histograms show respective FL1 (CD38-FITC or CD14-FITC signals; y-axes) and

FL2 (CD106-PE signal; x-axes) data through gate R1 (B, F and G) or R2 (D). (A, B) CD106 expression on malignant marrow PC from MMpatients. (C and D) Expression of CD106 on BM monocytes. (E) FISH analysis: the detection of CD106 mRNA on malignant BM PC using a

cocktail of six antisense oligodeoxynucleotides. Note the striking cellular resolution of the mRNA signal. (F) The very minor subset of malignant

PB PC usually detected in the majority of MM patients was stained negatively for CD106. (G) Expression of CD106 on normal marrow PC.

Numbers in each dot-plot represent the percentage of positivity and, in brackets, the MFI for CD106 on CD38++ (or CD14+) cells. Quadrantmarkers were set based on identical isotype controls.

72 R. Luque et al


Myeloma BM cells from untreated patients also expressan activated form of CD29It has been reported that the co-expression of CD106 andCD49d/CD29 is associated with the activation of thisintegrin (Liu et al, 1998, 1999). As most marrow PCco-express both CD106 (present report) and CD49d/CD29(Pellat-Deceunynck et al, 1995; Luque et al, 1998) and animportant feature of integrins is that they exist in active andinactive states (Hynes, 1992), we examined the expressionof the activated form of CD29 on PC from untreated MMpatients. To this end, we explored the binding of mAbHUTS-21 (Luque et al, 1996), which recognizes an activa-tion epitope of the CD29 integrin: in all the MM patientstested (n ¼ 25), CD29 was constitutively active on BMmalignant cells (87 ± 9%; MFI: 179 ± 62) (Fig 2A). Thesefindings suggest that the CD49d/CD29 integrin on CD106+

marrow PC in untreated MM patients may be present in aconstitutive, in vivo activated state. In contrast, the vastmajority of malignant BM B cells from patients withdifferent chronic lymphoproliferative diseases were negativefor the activated form of CD29 (data not shown). As shownwith CD106, activated CD29 expression was not signifi-cantly different between myelomatous and normal(85 ± 12%; MFI: 169 ± 78; n ¼ 10; P > 0Æ05) marrowPC (Fig 2B). Moreover, the minor circulating PC subsetcommonly observed in these patients expressed very low orundetectable levels of the activated form of CD29 (9 ± 8%;MFI: 42 ± 17; n ¼ 19) (Fig 2C).

CD106 and activated CD29 levels correlate with diseaseprogressionIn order to explore whether the expression of these twoadhesion molecules was associated with myeloma progres-sion, untreated patients who developed, after treatment andduring the time of this study, progressive disease (relapse ortreatment failure) were studied. As secondary PCL isconsidered the terminal phase of the disease, four patientswho developed it were classified as an independent group.As shown in Fig 3, marrow PC from patients in relapse or

with refractory disease showed lower expression of CD106and activated CD29, both in terms of percentage and MFI,than in myelomatous cells from these same patients before

Fig 2. An activated form of CD29 is expressed on malignant BM, but not circulating, PC from untreated MM patients. BM or PB cells were

incubated with the antiactivated CD29 HUTS-21 mAb and analysed by flow cytometry. Gating protocols have been previously described in

Materials and methods. The expression of the activated form of CD29 on BM PC from a representative untreated MM patient (A), normal BM PCfrom a healthy subject (B) and circulating PC from the same patient (C) is shown. Numbers in each dot-plot represent the percentage of

positivity and, in brackets, the MFI for activated CD29 on CD38++ cells. Quadrant markers were set based on identical isotype controls.

Fig 3. CD106 and activated CD29 are downregulated during MM

progression. The expression of CD106 and activated CD29 on PC

from BM aspirates and PB samples of patients at diagnosis (n ¼ 15),with progressive disease (relapse or refractory disease; R/RD)

(n ¼ 15) or with PCL (n ¼ 4) were analysed by flow cytometry. The

percentage of cells that stained positively (A) and the MFI (B) foreach antigen are shown.



treatment (P ¼ 0Æ01; n ¼ 15). Similar differences wereobserved between patients with progressive disease incomparison with patients in complete remission or inplateau phase (not shown). These differences were moresignificant between marrow and circulating PC fromsecondary PCL patients (n ¼ 4), which were all virtuallynegative for both antigens, in comparison with BM PC frompatients at diagnosis (P < 0Æ0001).

Moreover, the absolute numbers of circulating PC weremeasured over time: 14/15 patients at relapse or withrefractory disease showed significantly increased numbers ofcirculating PC, which were also negative for CD106 andactivated CD29 (data not shown), in comparison with thenumbers prior to treatment (mean: 112 · 106/l and4Æ7 · 106/l respectively; P < 0Æ001). Therefore, the down-modulation of these two adhesion molecules on marrow PCwas concomitant not only with progressive disease, but alsowith an increased number of circulating PC. These findingsare in agreement with a previous report that demonstrateda close relationship between progressive disease andincreased numbers of malignant circulating PC (Rawstronet al, 1997).

Given that the percentages of the marrow PC lackingCD106 and activated CD29 similarly decreased during thedifferent phases of the disease, we next investigated whetherthere was a correlation between these two parameters.Figure 4A shows a scattergram illustrating a linear regression

analysis: when comparing the expression of both antigenson myelomatous cells in terms of the percentage of positivecells, a correlation r2 value of 0Æ703 (P < 0Æ0001) wasfound. This positive correlation was further demonstratedby using multicolour immunofluorescence and multipara-meter flow cytometric analysis with anti-CD106, anti-CD29(activated) and anti-CD38 mAbs. As can be seen in Fig 4Band C, most of the BM PC were either positive or negativefor both CD106 and activated CD29, demonstrating a closerelationship in the expression of these two adhesionmolecules on PC.

We next examined whether the differences in activatedCD29 expression during the disease progression were due toa change in constitutive CD49d or CD29 integrin expres-sion. Although lower quantities of constitutive CD49d andCD29 (clone MAR4) levels were detected on marrow PC fromprogressive myeloma patients (CD49d: 96 ± 3%; MFI:380 ± 91; CD29: 78 ± 17%; MFI: 458 ± 101; n ¼ 15) orPCL patients (CD49d: 94 ± 3%; MFI: 372 ± 74; CD29:84 ± 11%; MFI: 401 ± 120; n ¼ 4) than in PC fromuntreated MM patients (CD49d: 98 ± 1%; MFI: 410 ± 59;CD29: 93 ± 5%; MFI: 489 ± 120; n ¼ 17), these differenceswere not statistically significant (P > 0Æ05). In Fig 5, anuntreated MM patient who relapsed and, to a further extent,developed a secondary PCL illustrates these findings. Othervery late antigens were not usually expressed (data notshown).

Fig 4. Relationship between the expression of CD106 and activated CD29 on malignant BM PC during MM progression. (A) Linear regression

analysis shows highly significant correlation between CD106 and activated CD29 expression in MM (P < 0Æ0001). The abscissa and ordinatein the graph show the CD106 and activated CD29 expression respectively. (B and C) BM cells were incubated with the anti-CD106-FITC,

antiactivated CD29-PE and anti-CD38-CyChrome mAbs, and analysed by flow cytometry. The gate was drawn on a SSC vs CD38 dot-plot

around CD38++ cells. Results from an untreated (B) and a progressive (C) patient expressing high or low percentages of CD106+/activated

CD29+ marrow PC are shown.

74 R. Luque et al


Therefore, myeloma progression is associated with adownregulation of CD106 and HUTS-21 epitope withoutaltering the level of CD49d/CD29 integrin expression.

Myeloma cell adhesion to FN decreaseswith disease progressionTo investigate whether the expression of HUTS-21 activa-tion epitope correlated with a significant adherence to FN,the binding of BM and circulating PC to FN-coated wellsfrom untreated, progressive and secondary PCL patients wasstudied. Myeloma BM cells from three untreated patientssignificantly adhered to FN without any exogenous stimulus(Fig 6). Moreover, the binding of tumoral cells to FN wasinhibited approximately 70–80% by anti-CD29 (4B4) oranti-CD49d (HP2/1) mAbs, confirming an essential role forthese integrins in this process. Phenotypic and functionalstudies demonstrated, in contrast, that CD49e, CD51 andCD41a were not involved in binding of myeloma cells to FN(not shown). As expected, malignant BM cells from patientswith progressive disease (n ¼ 3), which expressed lowerlevels of activated CD29, displayed a weaker adhesion to FNin vitro than those from MM patients at presentation(P ¼ 0Æ01). In close agreement, moreover, with the fact thatalmost all BM and circulating malignant cells from secondaryPCL patients were negative for activated CD29; their bindingto FN was very poor: significant decreases were observed inthe adhesion to FN when comparing PC from untreated

Fig 5. Activated CD29 is downregulated during myeloma progression, whereas constitutive CD29 and CD49d remain unchanged. Flow

cytometric analysis of the differential expression of epitopes HUTS-21 (activated CD29), MAR4 (an epitope constitutively expressed in the CD29

subunit, regardless of the state of integrin activation) and L25 (CD49d) on malignant PC from a MM patient at diagnosis, relapse (1 year later),and after developing a secondary PCL (3 years later). M1 markers were set based on identical isotype controls. Numbers at the top of each

histogram represent the percentages of positivity and the MFI values (in brackets) for each antigen.

Fig 6. Differences in the adhesion to FN between PC from untreated

or progressive MM and secondary PCL patients. Adhesion capacities

of purified BM (patients at presentation, n ¼ 3; patients at relapse or

with refractory disease, R/RD, n ¼ 3; PCL patients, n ¼ 3) and PB(PCL patients, n ¼ 3) malignant cells were quantified as indicated

in Materials and methods. Adhesion assays were performed either in

the absence of any mAb or in the presence of mAbs (10 lg/ml)

against CD29 (4B4), CD49d (HP2/1) or control mouse IgG. Mn+2-stimulated normal T-cell population was used as positive control

and its adhesion to FN, expressed in optical density (O.D.) units, was

1Æ05 ± 0Æ2. Background binding to albumin was < 5% and has not

been subtracted from the data presented.



patients with BM (p)0Æ001; n ¼ 3) or with circulating PC(p)0Æ001; n ¼ 4) of patients in the leukaemic phase.

Therefore, the myeloma cell adhesion to FN decreasedwith disease progression in parallel to the downmodulationof the activated CD29 expression.

DISCUSSION

Unlike most other haematopoietic malignancies, malignantPC in MM preferentially home to the BM. In this regard, it issupposed that the expression of certain critical adhesionmolecules on malignant PC (Witzig, 1999) can explain thispredilection to growth in the BM environment. We foundthat CD106 was expressed on malignant marrow PC fromuntreated MM patients. CD106 is a cytokine-inducedendothelial adhesion molecule that mediates the bindingof various leucocytes (Bevilacqua et al, 1995). In normaltissues, CD106 is restricted to few cell types. It is wellknown, however, that in normal BM several cell typesexpress this molecule (fibroblasts, macrophages, endothelialcells) (Ryan et al, 1991; Simmons et al, 1992), suggestingthat the BM environment can provide a unique milieu forCD106 expression. Therefore, our finding that malignantBM PC express CD106 was unexpected, but consistent withthe fact that the BM microenvironment can have profoundregulative effects to induce the expression of CD106.Moreover, our results are in agreement with a recent study,which describes the expression of CD106 on myeloma celllines (Yi-Feng et al, 2001). No difference was found inCD106 expression on normal, also considered as sessilecells, and malignant BM PC. Consequently, it cannot beassociated with neoplasia but rather with an enhancedretention of PC within the BM.

Several reports have described the constant presence ofthe integrin CD29/CD49d on malignant PC in MM (Pellat-Deceunynck et al, 1995; Luque et al, 1998). FN and CD106are recognized by this adhesion receptor, although itsconversion to an active state is necessary for efficientbinding to their ligands (Hynes, 1992). Using HUTS-21mAb, which recognizes an activation epitope of the CD29integrin (Luque et al, 1996), we demonstrated that malig-nant BM cells from untreated MM patients expressed anactivated form of CD29. This is related to the fact that theadhesion of the myeloma cells to FN is almost totally CD29/CD49d dependent, results which were in agreement withprevious reports (Sanz-Rodriguez et al, 1999). The activa-tion state of CD29 in this disease had been previouslystudied only in MM cell lines (Garcia-Gila et al, 1997). Wealso describe that the attachment of these cells to immobi-lized FN decreases during disease progression, correlatingwith the downmodulation of activated CD29. In contrast,constitutive CD29 and CD49d display conserved levels ofexpression during the course of the disease. On the basis ofthese results, we suggest that, not only in vivo myeloma,adhesiveness to FN may change during myeloma progres-sion, but also that the deactivation of CD29 integrin, asfound in malignant BM cells from PCL patients, is correlatedwith the tendency of the myeloma to leukaemic dissemin-ation through a deficient adhesion to FN. Currently, there is

no clear understanding of the cellular mechanisms bywhich CD29 integrin is produced with different activationstates. It is well known, however, that tumour cells expressCD29 with various levels of constitutive activity. Incontrast, CD29 integrin on normal cells is mostly inactiveunless activated by several stimuli (Masumoto & Hemler,1993). As previously described for CD106 expression,however, the fact that normal marrow PC also express theactivated form of CD29 may reflect that the BM is anoptimal microenvironment to induce the expression of theactive form of this integrin. Indeed, BM immature cells fromsome lineages were also HUTS-21+ (data not shown).

In contrast to activated CD29, the consequences ofCD106 expression and its loss on the adhesive propertiesof the PC are still unclear. Several lines of evidence suggest,however, that CD106 could be involved in homotypicadhesion of these BM myelomatous cells. First, it is wellknown that the finding of clusters of PC is generallysufficient to diagnose MM (Sukpanichnant et al, 1994).Second, it has been suggested that the expression of CD106on melanoma cells, which are also CD49d+ cells, isimportant in preventing metastasis by mediating melanomacell–cell adhesion (Denton et al, 1992). Third, it has beenrecently published that the co-expression of CD106 andCD49d/CD29 is associated with the activation of thisintegrin (Liu et al, 1998, 1999), thereby mediating amarked homotypic cell aggregation. In relation to this, anintriguing result not shown in the present work was that,despite that malignant BM PC expressed intermediate tohigh levels of activated CD49d/CD29, they displayed a poorbinding to CD106 expressed on stromal or activatedendothelial cells, in contrast to their strong adhesion toFN. CD49d has distinct binding sites for the two ligands(Elices et al, 1990) and it has also been reported that CD29can assume at least three stages of activation states(Masumoto & Hemler, 1993). Depending on these states,the extent of cell adhesion to CD106 and FN is different. Asa consequence, malignant BM cells from untreated MMpatients that express intermediate and, sometimes, highlevels of activated CD29 should be capable of binding toboth ligands. Although the reason why these cells adheredpoorly to CD106 is unknown, we hypothesize that, onCD106+ marrow PC, homotypic adhesion of PC could beanalogous to their attachment to activated endothelium inthat both processes are supported by CD106. As a conse-quence, the CD49d/CD106 binding site will be occupied(homotypic contacts), avoiding additional heterotypic inter-actions. In contrast, the CD49d/FN binding site will beavailable. In this regard, we are currently testing whetherhomotypic adhesion may influence the adhesion of thesecells to endothelial (or stromal) cells.

Of note, the minor circulating PC population detected inthe majority of MM patients (Billadeau et al, 1996) showedphenotypic (CD106–/activated CD29–) and functional (lackof adhesion to FN) (data not shown) characteristics identicalto those of PC from PCL patients. Whether myelomaprogression to PCL results from the expansion of the formerpopulation of cells that pre-existed in the initial phase of thedisease is not yet clear.

76 R. Luque et al


An interesting observation in this work was the simul-taneous downmodulation of CD106 and activated CD29during myeloma progression. Although it is supposed thatCD106 expression and integrin activation are regulated bydifferent mechanisms because the cellular distributions ofCD29/CD49d and CD106 are quite different, several reportsdemonstrated the simultaneous expression of these twoadhesion molecules in the same cells (Rosen et al, 1992). Itremains to be determined, however, to what extent thisobserved pattern of decreased expression of CD106 andactivated CD29 during disease progression reflectsco-ordinated regulation via shared regulatory elements. Inthis regard, several BM microenvironmental changes duringMM progression have been described (Merico et al, 1993).Whether or not these changes play a relevant role in thepathophysiology of the disease, as well as in the expressionof CD106 and activated CD29, is yet to be determined.

ACKNOWLEDGMENTS

We acknowledge the clinical staff of the HaematologyDepartment for their diligence in acquiring specimens foranalysis. We also thank Francine Doat and Eva Gallo forexcellent technical assistance. The authors gratefullyacknowledge the generous economic supports of Vitro S.A.(Madrid, Spain). This study was supported by Grants94/0529 and 96/1555 from Fondo de InvestigacionesSanitarias de la Seguridad Social (FIS) of Spain. R.L. wassupported by a fellowship grant from FIS of Spain. J.A.G.-T.was supported by a fellowship grant from the LAIRFoundation.

REFERENCES

Bevilacqua, M.P., Schwartz, B.R. & Harlan, J.M. (1995) CD106

cluster report. In: Leukocyte Typing V. White cell differentiation

antigens (ed. by S.F. Schlosmann) Vol. 2, pp. 1764–1766. OxfordUniversity Press, Oxford.

Billadeau, D., Van Ness, B., Kimlinger, T., Kyle, R.A., Therneau,

T.M., Greipp, P.R. & Witzig, T.E. (1996) Clonal circulating cells

are common in plasma cell proliferative disorders: a comparison ofmonoclonal gammopathy of undetermined significance, smolder-

ing multiple myeloma, and active myeloma. Blood, 88, 289–296.

Denton, K.J., Stretch, J.R., Gatter, K.C. & Harris, A.L. (1992) A

study of adhesion molecules as markers of progression inmalignant melanoma. Journal of Pathology, 167, 187–191.

Durie, B.G.M. & Salmon, S.E. (1975) A clinical staging system for

multiple myeloma. Cancer, 36, 842–854.Elices, M.J., Osborn, L., Takada, Y., Crouse, C., Luhowskyj, S., Hemler,

M.E. & Lobb, R.R. (1990) VCAM-1 on activated endothelium

interacts with the leukocyte integrin VLA-4 at a site distinct

from the VLA-4/fibronectin binding site. Cell, 60, 577–584.Garcia-Gila, M., Cabanas, C. & Garcıa-Pardo, A. (1997) Analysis of

the activation state of a4b1 integrin in human B cell lines derived

from myeloma, leukemia or lymphoma. FEBS Letters, 418, 337–

340.Goodwin, A.E. & Pauli, B.U. (1995) A new adhesion assay using

buoyancy to remove non-adherent cells. Journal of Immunological

Methods, 187, 213–219.Hynes, R.O. (1992) Integrins: versatility, modulation, and signal-

ling in cell adhesion. Cell, 69, 11–25.

Kalasz, V., Pap, T., Csanaky, G.Y. & Kelenyi, G. (1989) Endothelial

cell receptors on leukemic plasma cells. Leukemia Reseach, 13,

863–868.

Kim, I., Uchiyama, H., Chauhan, D. & Anderson, K.C. (1994) Cellsurface expression and functional significance of adhesion

molecules on human myeloma-derived cell lines. British Journal of

Haematology, 87, 483–493.Kosmo, M.A. & Gale, R.P. (1987) Plasma cell leukemia. Seminars in

Hematology, 24, 202–208.

Liu, Z.J., Tanaka, Y., Mine, S., Morinobu, A., Yagita, H., Okumuara,

K., Taniguchi, T., Yamamura, H. & Minami, Y. (1998) Func-tional cooperation of cyclin C and c-myc in mediating homotypic

cell adhesion via very late antigen-4 activation and vascular cell

adhesion molecule-1 induction. Blood, 92, 4700–4711.

Liu, Z.J., Tanaka, Y., Fujimoto, H., Mine, S., Morinobu, A., Yagita,H., Okumura, K., Oishi, I., Udagawa, J., Yamamura, H. &

Minami, Y. (1999) A novel role for H-ras in the regulation of very

late antigen-4 integrin and VCAM-1 via c-myc-dependent and-independent mechanisms. Journal of Immunology, 163, 4901–

4908.

Luque, A., Gomez, M., Puzon, W., Takada, Y., Sanchez-Madrid, F. &

Cabanas, C. (1996) Activated conformations of very late activa-tion integrins detected by a group of antibodies (HUTS) specific

for a novel regulatory region (355–425) of the common b1

chain. Journal of Biological Chemistry, 271, 11067–11075.

Luque, R., Brieva, J.A., Moreno, A., Manzanal, A., Escribano, L.,Villarrubia, J., Velasco, J.L., Lopez-Jimenez, J., Cervero, C., Otero,

M.J., Martınez, J., Bellas, C. & Roldan, E. (1998) Normal and

clonal B lineage cells can be distinguished by their differentialexpression of B cell antigens and adhesion molecules in periph-

eral blood from multiple myeloma (MM) patients. Diagnostic and

clinical implications. Clinical and Experimental Immunology, 112,

410–418.Masumoto, A. & Hemler, M.E. (1993) Multiple activation states of

VLA-4. Mechanistic differences between adhesion to CS1/fibro-

nectin and to vascular cell adhesion molecule-1. Journal of Bio-

logical Chemistry, 268, 228–234.Merico, F., Bergui, L., Gregoretti, M.G., Ghia, P., Aimo, G., Lindley,

I.J. & Caligaris-Capio, F. (1993) Cytokines involved in the pro-

gression of multiple myeloma. Clinical and Experimental

Immunology, 92, 27–31.Pellat-Deceunynck, C., Bataille, R., Robillard, N., Harousseau, J.L.,

Rapp, M.J., Juge-Morineau, N., Wijdenes, J. & Amiot, M. (1994)

Expression of CD28 and CD40 in human myeloma cells: a com-parative study with normal plasma cells. Blood, 84, 2597–2603.

Pellat-Deceunynck, C., Barille, S., Puthier, D., Rapp, M.J.,

Harousseau, J.L., Bataille, R. & Amiot, M. (1995) Adhesion

molecules on human myeloma cells: significant changes inexpression related to malignancy, tumor spreading, and

immortalization. Cancer Reseach, 55, 3647–3653.

Pellat-Deceunynck, C., Barille, S., Jego, G., Puthier, D., Robillard, N.,

Pineau, D., Rapp, M.J., Harousseau, J.L., Amiot, M. & Bataille, R.(1998) The absence of CD56 (NCAM) on malignant plasma cells

is a hallmark of plasma cell leukemia and of a special subset of

multiple myeloma. Leukemia, 12, 1977–1982.Rawstron, A.C., Owen, R.G., Davies, F.E., Johnson, R.J., Jones, R.A.,

Richards, S.J., Evans, P.A., Child, J.A., Smith, G.M., Jack, A.S. &

Morgan, G.J. (1997) Circulating plasma cells in multiple myelo-

ma: characterization and correlation with disease stage. BritishJournal of Haematology, 97, 46–55.

Reuss-Borst, M.A., Steinke, B., Waller, H.D., Buhring, J.J. & Muller,

C.A. (1992) Phenotypic and clinical heterogeneity of CD56-

positive acute nonlymphoblastic leukemia. Annals of Hematology,64, 78–82.



Rosen, G.D., Sanes, J.R., LaChance, R., Cunningham, J.M., Roman, J.

& Dean, D.C. (1992) Roles for the integrin VLA-4 and its counter

receptor VCAM-1 in myogenesis. Cell, 69, 1107–1119.Ryan, D.H., Nuccie, B.L., Abboud, C.N. & Winslow, J.M. (1991)

Vascular cell adhesion molecule-1 and the integrin VLA-4 med-

iate adhesion of human B cell precursors to cultured bone mar-

row adherent cells. Journal of Clinical Investigation, 88, 995–1004.Sanz-Rodriguez, F., Ruiz-Velasco, N., Pascual-Salcedo, D. & Teixido,

J. (1999) Characterization of VLA-4-dependent myeloma cell

adhesion to fibronectin and VCAM-1. British Journal of Haema-

tology, 107, 825–834.Simmons, P.J., Masinovsky, B., Longenecker, B.M., Berenson, R.,

Torok-Storb, B. & Gallatin, W.M. (1992) Vascular cell adhesion-1

expressed by bone marrow stromal cells mediates the binding of

hematopoietic progenitor cells. Blood, 80, 388–395.Sukpanichnant, S., Cousar, J.B., Leelasiri, A., Graber, S.E., Greer,

J.P. & Collins, R.D. (1994) Diagnostic criteria and histologic

grading in multiple myeloma: histologic and immunohistologic

analysis of 176 cases with clinical correlation. Human Pathology,

25, 308–318.Uchiyama, H., Barut, B.A., Chauhan, D., Cannistra, S.A. &

Anderson, K.C. (1992) Characterization of adhesion molecules

on human myeloma cell lines. Blood, 80, 2306–2314.

Van Camp, B., Durie, B.G., Spier, C., De-Waele, M., Van-Riet, I.,Vela, E., Frutiger, Y., Richter, L. & Grogan, T.M. (1990) Plasma

cells in multiple myeloma express a natural killer cell-associated

antigen: CD56 (NKH-1; Leu-19). Blood, 76, 377–382.

Witzig, T.E. (1999) The role of adhesion receptors in the patho-genesis of multiple myeloma. Hematology Oncology Clinics of North

America, 13, 1127–1143.

Yi-Feng, X., Li-Ping, L., Cui, Z. & Geng, J.G. (2001) NF-jB activation

for constitutive expression of VCAM-1 and ICAM-1 on B-lym-phocytes and plasma cells. Biochemical and Biophysical Research

Communications, 289, 851–856.

78 R. Luque et al


Documents

CD106 and activated-CD29 are expressed on myelomatous bone marrow plasma cells and their downregulation is associated with tumour progression