Histopathology 2014 DOI: 10.1111/his.12330
Expression of Mucin-1 in multiple myeloma and its precursors: correlation with glycosylation and subcellular localization Mindaugas Andrulis,1 Elena Ellert,1 Ulla Mandel,2 Henrik Clausen,2 Nicola Lehners,3,4 Marc-Steffen Raab,3,4 Hartmut Goldschmidt3,4 & Reinhard Schwartz-Albiez5 1
Institute of Pathology, University of Heidelberg, Heidelberg, Germany, 2Faculty of Health Sciences, Copenhagen Center for Glycomics, University of Copenhagen, Copenhagen, Denmark, 3Department of Medicine V, Heidelberg University Hospital, Heidelberg, Germany, 4Department of Medical Oncology, National Center for Tumor Diseases (NCT), Heidelberg, Germany, and 5German Cancer Research Center (DKFZ), Heidelberg, Germany
Date of submission 7 August 2013 Accepted for publication 18 November 2013 Published online Article Accepted 20 November 2013
Andrulis M, Ellert E, Mandel U, Clausen H, Lehners N, Raab M-S, Goldschmidt H & Schwartz-Albiez R (2014) Histopathology
Expression of Mucin-1 in multiple myeloma and its precursors: correlation with glycosylation and subcellular localization Aims: Recent reports suggest a possible role for extracellular (MUC1N) and transmembrane (MUC1C) subunits of Mucin 1 (MUC1) in the pathogenesis of multiple myeloma (MM). Nuclear translocation of MUC1C is involved in activation of various oncogenic signalling pathways and both MUC1 subunits are potential therapeutic targets. We aimed at performing a comprehensive expression analysis of the MUC1 subunits in plasma cell dyscrasias. Methods and results: Immunohistochemistry with monoclonal antibodies against the MUC1N subunit (EMA and 5E10) tumour-associated glycoforms of MUC1N (5E5) and the MUC1C subunit were applied to a series of biopsies from normal controls (n = 10)
and plasma cell dyscrasias (n = 121). Clonal plasma cells showed reduced MUC1N expression, and the 5E5 MUC1N epitope was expressed only in neoplastic plasma cells. Nuclear localization of MUC1C was equally frequent in all disease stages and did not differ from the control cases. Loss of both MUC1 subunits in MM (n = 12) was associated with significantly shorter overall survival and was more frequent in pretreated MM samples. Conclusions: Our findings indicate that aberrant glycosylation of MUC1 is an early event in the pathogenesis of MM. In contrast, MUC1C nuclear localization is not likely to be a driver of tumour progression.
Keywords: glycosylation, MUC1 nuclear expression, mucin 1, multiple myeloma
Introduction Multiple myeloma (MM) is an incurable plasma cell malignancy. MM cells are characterized by an abnormal infiltration of the bone marrow (BM) where growth entails many pathological consequences such Address for correspondence: R Schwartz-Albiez, Department of Translational Immunology D015, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany. e-mail: [email protected] © 2013 John Wiley & Sons Ltd.
as bone destruction, and bone marrow and kidney failure. The clinically evident stages of the disease are frequently preceded by asymptomatic stages that require no treatment: monoclonal gammopathy of undetermined significance (MGUS) and smouldering multiple myeloma (SMM). During disease progression, clonal MM cells can proliferate outside the BM environment, which is manifest as extramedullary plasmacytoma (EP).1 Mucin 1 (MUC1) is a cell membrane glycoprotein that is up-regulated and aberrantly glycosylated in
2 M Andrulis et al.
various types of human tumours, including MM.2–4 Recent reports implicate a possible role for MUC1 in the pathogenesis of MM and, due to this, MUC1 is considered as a potential therapeutic target.5–7 MUC1 is a heterodimer which consists of a large extracellular, Nterminal subunit (further referred to as MUC1N) that in the membrane is attached non-covalently to the Cterminal subunit (MUC1C) containing a short extracellular stem region, a transmembrane domain and a short C-terminal intracellular domain (Figure 1). Both MUC1 units are encoded by a single gene and translated as a single polypeptide which is cleaved posttranslationally, leading to equimolar expression of MUC1N and MUC1C. The extracellular MUC1N subunit contains a variable number of conserved 20 amino acid tandem repeats (n = 40–120) carrying a high density of O-linked glycans attached to the protein backbone.2,8 The O-glycans on MUC1N have a major impact on the functional and antigenic properties of the molecule. In normal cells MUC1N harbours complex elongated and branched glycan structures8 (Figure 1) while MUC1N expressed by tumour cells carries aberrant truncated short O-glycans such as the Tn and STn O-glycan structures. The Tn and STn glycans are among the best-characterized tumour-associated antigens in human carcinomas9,10, and these have diagnostic and therapeutic potential.11,12 MUC1C is involved in activation of oncogenic signalling pathways resulting in Wnt/b-catenin,5 NF-jB13,14 and STAT3 activation15 or p53 repression16,17. MUC1C is expressed aberrantly and translocated to the nucleus in MM cell lines.18,19 In addition, silencing MUC1C in MM cells shows negative effects for cell survival.5,20,21 Taken together, these observations indicate that both MUC1 subunits may have an oncogenic role in MM. However, no systematic investigation of MUC1 subunits in different stages of plasma cell disorders is available as yet. Moreover, the expression of MUC1C has not been studied so far in primary MM samples. In the present study we analysed the differential expression of MUC1 extracellular and transmembrane subunits in bone marrow biopsies from MM patients with different clinical stages of plasma cell neoplasia. Furthermore, we correlate MUC1 expression data with disease outcome.
Materials and Methods PATIENTS AND TISSUE SAMPLES
Formaldehyde-fixed paraffin-embedded tissue samples were obtained from the archives of the Institute of
MUC1-N
Normal
Core 1 Core 2
D O
T R
EMA 5E10
G
Cancer
S
Tn STn
O
T
5E5
A
Cell membrane
MUC1C Gal GalNAc GlcNAc Neu5Ac
MUC1-C
Figure 1. Schematic representation of MUC1 epitopes assayed in this study. EMA and 5E10 mAbs recognize the DTR repeat of MUC1 core protein in its glycosylated and non-glycosylated forms. EMA binding to the DTR peptide sequence is enhanced strongly by glycosylation. 5E10 recognizes the same DTR epitope, but it does not react with MUC1N peptide fully covered by sialylated oligosaccharides. 5E5 reacts with GSTA tandem repeats of the MUC1N core protein covered by truncated, cancer-specific Tn and/or STn oligosaccharides. The oligosaccharide subunits are abbreviated as follows: galactose (Gal), N-acetylglucosamine (GlcNAc), N-acetylgalactosamine (GalNAc), and sialic acid (Neu5Ac). MUC1N and MUC1C are non-covalently linked (dashed line).
Pathology, University of Heidelberg. The study cohort included bone marrow biopsies without plasma cell neoplasia (n = 10) with MGUS (n = 14) SMM (n = 24) MM (n = 60; including biopsies taken before treatment, n = 28 and at relapse/progress, n = 32) and EP (n = 23) (total n = 131). The samples of EP were combined in a tissue microarray. Overall survival (OS) data were available for 61 patients (SMM = 1, MM = 45 and EP = 15). The study was approved by the local ethics committee of the University Hospital Heidelberg (approved on 21 October 2005; approval no. 206/2005 and 207/2005). © 2013 John Wiley & Sons Ltd, Histopathology
MUC1 in multiple myeloma stages 3
ANTIBODIES/IMMUNOHISTOCHEMISTRY
A panel of antibodies was used to assess the glycosylation of MUC1 (summarized in Figure 1). Mouse monoclonal antibody (mAb) EMA (clone E29; Dako, Hamburg, Germany) was employed as a reference antibody because it recognizes both the glycosylated and non-glycosylated MUC1N protein backbone.8 In contrast to EMA, another mAb 5E10 also binds to the MUC1N protein backbone, but this mAb shows reduced reactivity with more dense and elongated glycosylation.22 Detection of the tumour-associated Tn (and STn) glycoform of MUC1N was performed using the mAb 5E5.23 A rabbit polyclonal antibody raised against amino acids 239–255 of MUC1C (cat. no. RB-9222-P; Thermo Scientific, Dreieich, Germany) was used to specifically detect the MUC1C subunit. Proliferating plasma cells were detected by Ki67 immunostaining (clone MIB1; Dako). The tissue sections were deparaffinized, subjected to heat-induced epitope retrieval in target retrieval solution (pH6.1; Dako) and immunostained using standard procedures and a labelled streptavidin-biotin detection system (Dako) as described previously.24 The percentage of the plasma cells was evaluated by CD138 (clone MI15; Dako) immunostaining and was taken into account during the semiquantitative assessment of the investigated MUC1 epitopes. Consecutive serial sections were used to stain for MUC1 epitopes. The plasma cells were identified by morphology and direct comparison with the corresponding CD138-stained slide (Figure 2B). Immunostaining was assessed semiquantitatively and classified as negative (no positive plasma cells detected, score 0) low (50% plasma cells positive, score 3). The cases with a score >1 were classified as positive. Membranebound, cytoplasmic and nuclear MUC1C expression were scored separately. In the series investigated we relied upon morphology to categorize the MUC1-negative and -positive cells. To quantify the proliferation rate, the MIB1-stained slides were digitalized using a slide scanner (Aperio Technologies, Inc., Vista, CA, USA) and the percentage of MIB1-positive plasma cells was quantified by automatic image analysis software using a nuclear V9 algorithm (Aperio). STATISTICAL ANALYSIS
All analyses were conducted in R.25 For analysis of discrete variables the v2 or Fisher’s exact test was employed. For analysis of continuous variables the © 2013 John Wiley & Sons Ltd, Histopathology
Wilcox test was used. Unsupervised cluster analysis of ranked immunohistochemical expression data was performed using the gplots library.26 The survival curves were built according to the Kaplan–Meier method. Survival comparison was performed with a two-sided log-rank test.
Results MUC1N EXPRESSION IN BIOPSIES OF DIFFERENT DISEASE STAGES
Staining of bone marrow and extramedullary specimens of control individuals, MGUS, SMM, MM and EP patients revealed that MUC1N (mAb EMA) was expressed in all plasma cells from normal bone marrow and the majority of patients in preclinical disease stages (MGUS and SMM) (Figure 2A,B). Significant loss of MUC1N was observed in advanced stages (MM and EP) (Figure 2A). In contrast, mAb 5E10 binding was diminished at the preclinical stages of MGUS and SMM and remained at comparable levels in symptomatic MM and EP cases (Figure 2A). Tumour-associated MUC1N glycoform (mAb 5E5) was not detectable in normal plasma cells, but expressed in all stages of plasma cell dyscrasias. The 5E5 expression increased along with disease progression (Figure 2A) but this observation did not reach statistical significance. MUC1C SUBCELLULAR LOCALIZATION IN DIFFERENT DISEASE STAGES
Further analysis of subcellular localization of MUC1C revealed three different patterns: membrane-bound, granular–cytoplasmic and nuclear (Figure 2C). The membrane-bound and nuclear patterns were mutually exclusive (P-value = 0.003, v2 test; Figure 4A). In contrast, the cells with granular–cytoplasmic MUC1C localization frequently coexpressed MUC1C in the nucleus. The membrane-bound MUC1C was found in all bone marrow biopsies of the control group. In contrast, the expression of membrane-associated MUC1C was reduced significantly in neoplastic plasma cells (Figure 2A) but did not differ between samples from preclinical and clinical MM stages. Interestingly, nuclear localization of MUC1C was also observed in a subpopulation of plasma cells in almost half the control cases (Figure 2A,B). In neoplastic plasma cells the frequency of nuclear MUC1C expression did not differ significantly between disease stages. EP cases showed the lowest frequency of MUC1C nuclear expression, but this finding was not statistically significant. The
4 M Andrulis et al.
A
P < 0.01
5E10 % 100 80 60 40 20 0
MGUS
P < 0.01
50
MGUS
Normal
MM
EP
clinically manifest
38
SMM n.s.
n.s.
64
n.s.
preclinical
100
5E5
47
41
MM
EP
n.s. clinically manifest
preclinical
5E5
% 100 80 60 40 20 0
SMM
60
EMA
Normal
clinically manifest
CD138
preclinical 93 79
100
MM
0 Normal
12
17
MGUS
SMM
MM
EP
P < 0.01 n.s. clinically manifest
preclinical
100
50 33
28
SMM
MM
21 Normal
MUC1C-N % 100 80 60 40 20 0
8
P < 0.001
MUC1C-M % 100 80 60 40 20 0
7
MUC1C
% 100 80 60 40 20 0
MGUS
BM
P < 0.05
n.s.
5E10
EMA
B
P < 0.05
MGUS
n.s.
n.s.
64
Normal
MGUS
C
n.s.
preclinical
50
EP
58
SMM
i
ii
iii
clinically manifest
45
MM
32
EP
Figure 2. Expression of MUC1 subunits in normal and neoplastic plasma cells. A, relative expression frequency of EMA, 5E10, 5E5, MUC1CM (membrane-bound MUC1C) and MUC1C-N (nuclear MUC1C) in normal plasma cells and different stages of plasma cell dyscrasia (MGUS, monoclonal gammopathy of undetermined significance; SMM, smouldering multiple myeloma; MM, multiple myeloma; EP, extramedullary plasmacytoma). Depicted are the percentages of positive cases, scored as described in detail in Materials and Methods. Statistical significance was assessed using the v2 test (n.s.: non-significant). B, Representative examples of plasma cells in biopsies from normal bone marrow (BM) preclinical (MGUS) and symptomatic (MM) disease stages expressing EMA, 5E10, 5E5 and nuclear MUC1C. Upper row shows biopsy areas stained for CD138 to visualize plasma cells. C, Observed variants of subcellular MUC1C localization: (i) membrane-bound, (ii) granular–cytoplasmic and (iii) nuclear, in these cases coexisting with granular–cytoplasmic pattern in some plasma cells. © 2013 John Wiley & Sons Ltd, Histopathology
MUC1 in multiple myeloma stages 5
WITH PATIENT SURVIVAL
The expression frequency of the epitopes investigated did not differ in untreated and relapsed/progressing MM cases. Furthermore, none of the investigated MUC1 epitopes correlated with overall survival as a single parameter. Therefore, we next tested whether or not a combination of epitopes was associated with any of the above-mentioned factors. For this we applied an unsupervised cluster analysis of scored immunohistochemical data to a subset of MM cases with available OS data (n = 61). This revealed that EMA clusters together with 5E10, whereas 5E5 clusters with membrane-associated MUC1C (Figure 4A). Interestingly, we found a number of MM cases that expressed either MUC1N (n = 12) or MUC1C (n = 11). Most importantly, the cluster analysis revealed a group of 12 MM cases that were negative for all MUC1 epitopes investigated and showed significantly shorter OS (Figure 4B). Pretreated/relapsed MM cases were significantly more frequent in this group (n = 10, P-value = 0.04, v2 test).
Discussion Although MUC1 is used widely as a marker for MM cells, our knowledge with regard to its specific glycosylation variants in MM is still sparse. In the present study we applied a panel of mAbs to MUC1, which enabled us to demonstrate that aberrant glycoforms of MUC1 defined by mAb 5E5 exhibit cancer-specific expression in MM and its precursor stages. MUC1 was expressed in normal cells and all MM stages (defined by mAb EMA) and characterised by a differ© 2013 John Wiley & Sons Ltd, Histopathology
100
90
90
80
80
Ki67 positive plasma cells (%)
100
70 60 50 40 30 20
B
70 60 50 40 30 20
10
10
0
0 negative positive
CORRELATION OF MUC1 EPITOPE EXPRESSION
MUC1C-N
MUC1C-M
A
Ki67 positive plasma cells (%)
granular–cytoplasmic MUC1C expression pattern was found in only 10 cases in total. Among those were control bone marrow samples and cases from all disease stages. Six cases showed coexistent nuclear MUC1C expression (BM = 1; SMM = 2; MM = 1; EP = 2) whereas four cases exhibited a granular–cytoplasmic pattern exclusively (MGUS = 3; SMM = 1). To test further how MUC1C subcellular localization correlates with the proliferation of MM cells we quantified MIB1-positive plasma cells in EP cases. The loss of nuclear MUC1C was associated significantly with increased proliferation rate (P-value = 0.05, Wilcox test, Figure 3A,B). In contrast, membrane-bound expression of MUC1-C was seen in cases with higher MIB1 score, but this observation did not reach statistical significance (Figure 3A,B).
MUC1C
negative positive Ki67
Figure 3. Correlation of subcellular localization of MUC1C and proliferative activity of MM cells. A, box-and-whisker plots showing the percentage of Ki67-expressing plasma cells in extramedullary plasmacytoma (EP) cases with membrane-bound (MUC1C-M, left) and nuclear (MUC1C-N, right) expression of MUC1C. The median is depicted by a horizontal line, the lower and upper quartiles (25% and 75%) are shown as a box and the minimum and maximum values are depicted by the whiskers. The cases with membranebound MUC1C have a higher proliferation rate, but this finding did not reach statistical significance (left). The loss of nuclear MUC1C expression is associated significantly with an increased proliferation rate (right, P-value = 0.05, Wilcox test). B, Representative examples of EP cases with nuclear (top row) and membrane-bound (bottom row) MUC1C expression (left) as well as corresponding serial sections demonstrating the proliferation rate by MIB1-staining (Ki67, right).
6 M Andrulis et al.
Colour key
0
1 2 3 Score MUC1C-N 5E5 MUC1C-M 5E10
EMA
1.0 Positive Negative
0.8
0.6
0.4
P = 0.02
0.2
0.0 0
50
100
150
200
Figure 4. MUC1 epitope expression: cluster analysis and correlation with overall survival. A, unsupervised cluster analysis of semiquantitatively scored immunohistochemical MUC1 epitope expression in a subset of patients with available survival data (n = 61). Black box marks a cluster of multiple myeloma cases (n = 12) that express none of the MUC1 epitopes investigated. B, MUC1-negative cases (black line) have a significantly shorter OS (months) compared to the cases with at least one MUC1 epitope detectable (red line).
ent reactivity pattern of mAb 5E10; it is likely that the density and sialylation of O-glycans increase during disease progression. The increased expression of Tn and STn in tumour cells may be due to either a reduced activity of distinct glycosyltransferases neces-
sary for the extension of cores 1 and 2 oligosaccharides or to the overexpression of sialyltransferases, which may block the chain elongation.27 In particular, the enzymatic activity of the core 1 b1,3-galactosyltransferase is controlled by the chaperone Cosmc, which may be inactivated by somatic point mutations or epigenetic silencing and lead to increased expression of the Tn antigen in cancer cells.28,29 The major difference in EMA and 5E10 binding properties is that 5E10 does not recognize MUC1N fully covered by sialylated oligosacharides.22 Thus, in MUC1N-expressing tumours with aberrant glycosylation one would expect a reduction of 5E10 binding and/or positivity for 5E5 mAb that binds specifically to Tn-antigen independent of its sialylation. Our data demonstrate this difference of 5E10 and 5E5 expression between control cases and MGUS. This indicates that MUC1N glycosylation changes are early events in the oncogenic transformation of plasma cells. In contrast, we did not observe significant differences in 5E10 and 5E5 expression between preclinical and symptomatic disease stages. Moreover, even EMA reactivity was lost in advanced MM stages. This may imply that complete loss of MUC1N from the cell surface but not aberrant glycosylation itself is associated with tumour progression. In the present work we provide the first detailed immunohistochemical analysis of MUC1C subcellular localization in various stages of plasma cell neoplasia. In accordance with published cell culture-based data, we found a significant loss of membrane-bound MUC1C in neoplastic plasma cells compared to normal plasma cells. Furthermore, the membrane-bound and nuclear MUC1C were mutually exclusive. This finding is in agreement with earlier observations that MUC1C indeed translocates into the nucleus to exert its signalling activity.19 Unexpectedly, we found nuclear MUC1C in 50% of normal controls, indicating that MUC1C translocation into the nucleus is not an aberrant event by itself. It may suggest that MUC1Cassociated signalling is involved in certain stages of normal plasma cell differentiation. Further studies are needed to address this question. We did not observe major differences of MUC1C nuclear expression frequency between controls and patient samples of various clinical disease stages. Nuclear MUC1C expression was increased slightly in MGUS and SMM compared to normal controls. Therefore, MUC1C nuclear translocation may occur early in the course of disease, which is in accordance with a possible oncogenic role of MUC1C. However, based on our observations, it is unlikely that nuclear MUC1C translocation is a major initiating event in the pathogenesis of multiple myeloma. Furthermore, our data reveal a tendency of reduced © 2013 John Wiley & Sons Ltd, Histopathology
MUC1 in multiple myeloma stages 7
nuclear MUC1C expression in the most advanced stages of myeloma, although this did not reach statistical significance. In addition, the EP cases with nuclear MUC1C expression were less proliferative compared to MUC1C negative cases. These findings may question a driver role of nuclear MUC1C in the disease progression, and are surprising because a number of in-vitro studies suggest that MUC1C may be involved in the oncogenic signalling of MM 13–15,20,21 and epithelial tumours.7 Conversely, a decreased expression of MUC1C was reported in some epithelial tumours, e.g. prostate cancer.30 Thus, the MUC1C signalling may be different in diverse types of tumours depending on the respective signalling pathways associated with MUC1C. Both MUC1 units are translated as a single polypeptide leading to equimolar expression of MUC1N and MUC1C. However, we found a number of MM cases that express either MUC1N or MUC1C exclusively. This finding indicates that different mechanisms such as shedding of MUC1N and cytoplasmic degradation of MUC1C may be responsible for loss of one of the MUC1 isoforms dependent upon its posttranslational modifications. Our results are in line with previously published data and confirm the expression of aberrantly glycosylated MUC1N on the malignant plasma cell surface as well as nuclear translocation of MUC1C in primary myeloma cells. Additionally, we demonstrate here that the expression of MUC1 is reduced in the clinically most advanced MM stages for which no further established therapeutic options are available. Moreover, the loss of MUC1 is associated significantly with adverse clinical outcome. These observations imply a pretreatment screening for MUC1 targets in patients for whom an anti-MUC1 therapy may be considered. Summarizing, the data presented here provide a concise analysis of both MUC1 subunits in multiple myeloma and its precursor stages. Our findings indicate that MUC1N glycosylation changes are early events in the pathogenesis of multiple myeloma. We further conclude that MUC1C nuclear localization is not an aberrant event and is unlikely to be a major driver of tumour progression in MM. We also demonstrate that complete loss of MUC1 is observed frequently in relapsed disease and is associated with adverse outcome.
Acknowledgements This study was supported by a grant of the ‘Deutsche Jose Carreras Leuk€ amie-Stiftung e.V. (DJCLS R 08/13) © 2013 John Wiley & Sons Ltd, Histopathology
to R.S.A. H.C. and U.M. were supported by the Danish National Research Foundation (DNRF107). We thank Tina Philipp and the Tissue Bank of the National Centre for Tumour Diseases Heidelberg for excellent technical assistance.
References 1. Morgan GJ, Walker BA, Davies FE. The genetic architecture of multiple myeloma. Nat. Rev. Cancer 2012; 12; 335–348. 2. Baldus SE, Engelmann K, Hanisch FG. MUC1 and the MUCs: a family of human mucins with impact in cancer biology. Crit. Rev. Clin. Lab. Sci. 2004; 41; 189–231. 3. Baldus SE, Monig SP, Huxel S et al. MUC1 and nuclear betacatenin are coexpressed at the invasion front of colorectal carcinomas and are both correlated with tumor prognosis. Clin. Cancer Res. 2004; 10; 2790–2796. 4. Cloosen S, Gratama J, van Leeuwen EB et al. Cancer specific mucin-1 glycoforms are expressed on multiple myeloma. Br. J. Haematol. 2006; 135; 513–516. 5. Kawano T, Ahmad R, Nogi H, Agata N, Anderson K, Kufe D. MUC1 oncoprotein promotes growth and survival of human multiple myeloma cells. Int. J. Oncol. 2008; 33; 153– 159. 6. Kufe DW. Mucins in cancer: function, prognosis and therapy. Nat. Rev. Cancer 2009; 9; 874–885. 7. Uchida Y, Raina D, Kharbanda S, Kufe D. Inhibition of the MUC1-C oncoprotein is synergistic with cytotoxic agents in the treatment of breast cancer cells. Cancer Biol. Ther. 2013; 14; 127–134. 8. Hanisch FG, Muller S. MUC1: the polymorphic appearance of a human mucin. Glycobiology 2000; 10; 439–449. 9. Cao Y, Merling A, Karsten U et al. Expression of CD175 (Tn) CD175s (sialosyl-Tn) and CD176 (Thomsen–Friedenreich antigen) on malignant human hematopoietic cells. Int. J. Cancer 2008; 123; 89–99. 10. Pinto R, Carvalho AS, Conze T et al. Identification of new cancer biomarkers based on aberrant mucin glycoforms by in situ proximity ligation. J. Cell Mol. Med. 2012; 16; 1474–1484. 11. Beatson RE, Taylor-Papadimitriou J, Burchell JM. MUC1 immunotherapy. Immunotherapy 2010; 2; 305–327. 12. Mall AS. Analysis of mucins: role in laboratory diagnosis. J. Clin. Pathol. 2008; 61; 1018–1024. 13. Ahmad R, Raina D, Joshi MD et al. MUC1-C oncoprotein functions as a direct activator of the nuclear factor-kappa b p65 transcription factor. Cancer Res. 2009; 69; 7013–7021. 14. Ahmad R, Raina D, Trivedi V et al. MUC1 oncoprotein activates the Ikappa b kinase beta complex and constitutive NFkappa b signalling. Nat. Cell Biol. 2007; 9; 1419–1427. 15. Ahmad R, Rajabi H, Kosugi M et al. MUC1-C oncoprotein promotes STAT3 activation in an autoinductive regulatory loop. Sci. Signal 2011; 4: ra9. 16. Raina D, Ahmad R, Chen D, Kumar S, Kharbanda S, Kufe D. MUC1 oncoprotein suppresses activation of the ARF-MDM2p53 pathway. Cancer Biol. Ther. 2008; 7; 1959–1967. 17. Wei X, Xu H, Kufe D. Human MUC1 oncoprotein regulates p53-responsive gene transcription in the genotoxic stress response. Cancer Cell 2005; 7; 167–178. 18. Leng Y, Cao C, Ren J et al. Nuclear import of the MUC1-C oncoprotein is mediated by nucleoporin NUP62. J. Biol. Chem. 2007; 282; 19321–19330.
8 M Andrulis et al.
19. Li Y, Chen W, Ren J et al. Df3/MUC1 signaling in multiple myeloma cells is regulated by interleukin-7. Cancer Biol. Ther. 2003; 2; 187–193. 20. Yin L, Ahmad R, Kosugi M et al. Survival of human multiple myeloma cells is dependent on MUC1 c-terminal transmembrane subunit oncoprotein function. Mol. Pharmacol. 2010; 78; 166–174. 21. Yin L, Kosugi M, Kufe D. Inhibition of the MUC1-C oncoprotein induces multiple myeloma cell death by down-regulating tigar expression and depleting NADPH. Blood 2012; 119; 810–816. 22. Tarp MA, Sørensen AL, Mandel U et al. Identification of a novel cancer-specific immunodominant glycopeptide epitope in the MUC1 tandem repeat. Glycobiology 2007; 17; 197–209. 23. Sørensen AL, Reis CA, Tarp MA et al. Chemoenzymatically synthesized multimeric Tn/STn MUC1 glycopeptides elicit cancer-specific anti-MUC1 antibody responses and override tolerance. Glycobiology 2006; 16; 96–107. 24. Andrulis M, Dietrich S, Longerich T et al. Loss of endothelial thrombomodulin predicts response to steroid therapy and sur-
25.
26. 27.
28.
29.
30.
vival in acute intestinal graft-versus-host disease. Haematologica 2012; 97; 1674–1677. R Core Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2012. Warnes GR. Gplots: various R programming tools for plotting data 2012. http://cran.r-project.org/web/packages/gplots/index.html Engelmann K, Baldus SE, Hanisch FG. Identification and topology of variant sequences within individual repeat domains of the human epithelial tumor mucin MUC1. J. Biol. Chem. 2001; 276; 27764–27769. Ju T, Aryal RP, Stowell CJ, Cummings RD. Regulation of protein o-glycosylation by the endoplasmic reticulum-localized molecular chaperone cosmc. J. Cell Biol. 2008; 182; 531–542. Mi R, Song L, Wang Y et al. Epigenetic silencing of the chaperone cosmc in human leukocytes expressing Tn antigen. J. Biol. Chem. 2012; 287; 41523–41533. Singh AP, Chauhan SC, Bafna S et al. Aberrant expression of transmembrane mucins, MUC1 and MUC4, in human prostate carcinomas. Prostate 2006; 66; 421–429.
© 2013 John Wiley & Sons Ltd, Histopathology
Expression of Mucin-1 in multiple myeloma and its precursors: correlation with glycosylation and subcellular localization Mindaugas Andrulis,1 Elena Ellert,1 Ulla Mandel,2 Henrik Clausen,2 Nicola Lehners,3,4 Marc-Steffen Raab,3,4 Hartmut Goldschmidt3,4 & Reinhard Schwartz-Albiez5 1
Institute of Pathology, University of Heidelberg, Heidelberg, Germany, 2Faculty of Health Sciences, Copenhagen Center for Glycomics, University of Copenhagen, Copenhagen, Denmark, 3Department of Medicine V, Heidelberg University Hospital, Heidelberg, Germany, 4Department of Medical Oncology, National Center for Tumor Diseases (NCT), Heidelberg, Germany, and 5German Cancer Research Center (DKFZ), Heidelberg, Germany
Date of submission 7 August 2013 Accepted for publication 18 November 2013 Published online Article Accepted 20 November 2013
Andrulis M, Ellert E, Mandel U, Clausen H, Lehners N, Raab M-S, Goldschmidt H & Schwartz-Albiez R (2014) Histopathology
Expression of Mucin-1 in multiple myeloma and its precursors: correlation with glycosylation and subcellular localization Aims: Recent reports suggest a possible role for extracellular (MUC1N) and transmembrane (MUC1C) subunits of Mucin 1 (MUC1) in the pathogenesis of multiple myeloma (MM). Nuclear translocation of MUC1C is involved in activation of various oncogenic signalling pathways and both MUC1 subunits are potential therapeutic targets. We aimed at performing a comprehensive expression analysis of the MUC1 subunits in plasma cell dyscrasias. Methods and results: Immunohistochemistry with monoclonal antibodies against the MUC1N subunit (EMA and 5E10) tumour-associated glycoforms of MUC1N (5E5) and the MUC1C subunit were applied to a series of biopsies from normal controls (n = 10)
and plasma cell dyscrasias (n = 121). Clonal plasma cells showed reduced MUC1N expression, and the 5E5 MUC1N epitope was expressed only in neoplastic plasma cells. Nuclear localization of MUC1C was equally frequent in all disease stages and did not differ from the control cases. Loss of both MUC1 subunits in MM (n = 12) was associated with significantly shorter overall survival and was more frequent in pretreated MM samples. Conclusions: Our findings indicate that aberrant glycosylation of MUC1 is an early event in the pathogenesis of MM. In contrast, MUC1C nuclear localization is not likely to be a driver of tumour progression.
Keywords: glycosylation, MUC1 nuclear expression, mucin 1, multiple myeloma
Introduction Multiple myeloma (MM) is an incurable plasma cell malignancy. MM cells are characterized by an abnormal infiltration of the bone marrow (BM) where growth entails many pathological consequences such Address for correspondence: R Schwartz-Albiez, Department of Translational Immunology D015, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany. e-mail: [email protected] © 2013 John Wiley & Sons Ltd.
as bone destruction, and bone marrow and kidney failure. The clinically evident stages of the disease are frequently preceded by asymptomatic stages that require no treatment: monoclonal gammopathy of undetermined significance (MGUS) and smouldering multiple myeloma (SMM). During disease progression, clonal MM cells can proliferate outside the BM environment, which is manifest as extramedullary plasmacytoma (EP).1 Mucin 1 (MUC1) is a cell membrane glycoprotein that is up-regulated and aberrantly glycosylated in
2 M Andrulis et al.
various types of human tumours, including MM.2–4 Recent reports implicate a possible role for MUC1 in the pathogenesis of MM and, due to this, MUC1 is considered as a potential therapeutic target.5–7 MUC1 is a heterodimer which consists of a large extracellular, Nterminal subunit (further referred to as MUC1N) that in the membrane is attached non-covalently to the Cterminal subunit (MUC1C) containing a short extracellular stem region, a transmembrane domain and a short C-terminal intracellular domain (Figure 1). Both MUC1 units are encoded by a single gene and translated as a single polypeptide which is cleaved posttranslationally, leading to equimolar expression of MUC1N and MUC1C. The extracellular MUC1N subunit contains a variable number of conserved 20 amino acid tandem repeats (n = 40–120) carrying a high density of O-linked glycans attached to the protein backbone.2,8 The O-glycans on MUC1N have a major impact on the functional and antigenic properties of the molecule. In normal cells MUC1N harbours complex elongated and branched glycan structures8 (Figure 1) while MUC1N expressed by tumour cells carries aberrant truncated short O-glycans such as the Tn and STn O-glycan structures. The Tn and STn glycans are among the best-characterized tumour-associated antigens in human carcinomas9,10, and these have diagnostic and therapeutic potential.11,12 MUC1C is involved in activation of oncogenic signalling pathways resulting in Wnt/b-catenin,5 NF-jB13,14 and STAT3 activation15 or p53 repression16,17. MUC1C is expressed aberrantly and translocated to the nucleus in MM cell lines.18,19 In addition, silencing MUC1C in MM cells shows negative effects for cell survival.5,20,21 Taken together, these observations indicate that both MUC1 subunits may have an oncogenic role in MM. However, no systematic investigation of MUC1 subunits in different stages of plasma cell disorders is available as yet. Moreover, the expression of MUC1C has not been studied so far in primary MM samples. In the present study we analysed the differential expression of MUC1 extracellular and transmembrane subunits in bone marrow biopsies from MM patients with different clinical stages of plasma cell neoplasia. Furthermore, we correlate MUC1 expression data with disease outcome.
Materials and Methods PATIENTS AND TISSUE SAMPLES
Formaldehyde-fixed paraffin-embedded tissue samples were obtained from the archives of the Institute of
MUC1-N
Normal
Core 1 Core 2
D O
T R
EMA 5E10
G
Cancer
S
Tn STn
O
T
5E5
A
Cell membrane
MUC1C Gal GalNAc GlcNAc Neu5Ac
MUC1-C
Figure 1. Schematic representation of MUC1 epitopes assayed in this study. EMA and 5E10 mAbs recognize the DTR repeat of MUC1 core protein in its glycosylated and non-glycosylated forms. EMA binding to the DTR peptide sequence is enhanced strongly by glycosylation. 5E10 recognizes the same DTR epitope, but it does not react with MUC1N peptide fully covered by sialylated oligosaccharides. 5E5 reacts with GSTA tandem repeats of the MUC1N core protein covered by truncated, cancer-specific Tn and/or STn oligosaccharides. The oligosaccharide subunits are abbreviated as follows: galactose (Gal), N-acetylglucosamine (GlcNAc), N-acetylgalactosamine (GalNAc), and sialic acid (Neu5Ac). MUC1N and MUC1C are non-covalently linked (dashed line).
Pathology, University of Heidelberg. The study cohort included bone marrow biopsies without plasma cell neoplasia (n = 10) with MGUS (n = 14) SMM (n = 24) MM (n = 60; including biopsies taken before treatment, n = 28 and at relapse/progress, n = 32) and EP (n = 23) (total n = 131). The samples of EP were combined in a tissue microarray. Overall survival (OS) data were available for 61 patients (SMM = 1, MM = 45 and EP = 15). The study was approved by the local ethics committee of the University Hospital Heidelberg (approved on 21 October 2005; approval no. 206/2005 and 207/2005). © 2013 John Wiley & Sons Ltd, Histopathology
MUC1 in multiple myeloma stages 3
ANTIBODIES/IMMUNOHISTOCHEMISTRY
A panel of antibodies was used to assess the glycosylation of MUC1 (summarized in Figure 1). Mouse monoclonal antibody (mAb) EMA (clone E29; Dako, Hamburg, Germany) was employed as a reference antibody because it recognizes both the glycosylated and non-glycosylated MUC1N protein backbone.8 In contrast to EMA, another mAb 5E10 also binds to the MUC1N protein backbone, but this mAb shows reduced reactivity with more dense and elongated glycosylation.22 Detection of the tumour-associated Tn (and STn) glycoform of MUC1N was performed using the mAb 5E5.23 A rabbit polyclonal antibody raised against amino acids 239–255 of MUC1C (cat. no. RB-9222-P; Thermo Scientific, Dreieich, Germany) was used to specifically detect the MUC1C subunit. Proliferating plasma cells were detected by Ki67 immunostaining (clone MIB1; Dako). The tissue sections were deparaffinized, subjected to heat-induced epitope retrieval in target retrieval solution (pH6.1; Dako) and immunostained using standard procedures and a labelled streptavidin-biotin detection system (Dako) as described previously.24 The percentage of the plasma cells was evaluated by CD138 (clone MI15; Dako) immunostaining and was taken into account during the semiquantitative assessment of the investigated MUC1 epitopes. Consecutive serial sections were used to stain for MUC1 epitopes. The plasma cells were identified by morphology and direct comparison with the corresponding CD138-stained slide (Figure 2B). Immunostaining was assessed semiquantitatively and classified as negative (no positive plasma cells detected, score 0) low (50% plasma cells positive, score 3). The cases with a score >1 were classified as positive. Membranebound, cytoplasmic and nuclear MUC1C expression were scored separately. In the series investigated we relied upon morphology to categorize the MUC1-negative and -positive cells. To quantify the proliferation rate, the MIB1-stained slides were digitalized using a slide scanner (Aperio Technologies, Inc., Vista, CA, USA) and the percentage of MIB1-positive plasma cells was quantified by automatic image analysis software using a nuclear V9 algorithm (Aperio). STATISTICAL ANALYSIS
All analyses were conducted in R.25 For analysis of discrete variables the v2 or Fisher’s exact test was employed. For analysis of continuous variables the © 2013 John Wiley & Sons Ltd, Histopathology
Wilcox test was used. Unsupervised cluster analysis of ranked immunohistochemical expression data was performed using the gplots library.26 The survival curves were built according to the Kaplan–Meier method. Survival comparison was performed with a two-sided log-rank test.
Results MUC1N EXPRESSION IN BIOPSIES OF DIFFERENT DISEASE STAGES
Staining of bone marrow and extramedullary specimens of control individuals, MGUS, SMM, MM and EP patients revealed that MUC1N (mAb EMA) was expressed in all plasma cells from normal bone marrow and the majority of patients in preclinical disease stages (MGUS and SMM) (Figure 2A,B). Significant loss of MUC1N was observed in advanced stages (MM and EP) (Figure 2A). In contrast, mAb 5E10 binding was diminished at the preclinical stages of MGUS and SMM and remained at comparable levels in symptomatic MM and EP cases (Figure 2A). Tumour-associated MUC1N glycoform (mAb 5E5) was not detectable in normal plasma cells, but expressed in all stages of plasma cell dyscrasias. The 5E5 expression increased along with disease progression (Figure 2A) but this observation did not reach statistical significance. MUC1C SUBCELLULAR LOCALIZATION IN DIFFERENT DISEASE STAGES
Further analysis of subcellular localization of MUC1C revealed three different patterns: membrane-bound, granular–cytoplasmic and nuclear (Figure 2C). The membrane-bound and nuclear patterns were mutually exclusive (P-value = 0.003, v2 test; Figure 4A). In contrast, the cells with granular–cytoplasmic MUC1C localization frequently coexpressed MUC1C in the nucleus. The membrane-bound MUC1C was found in all bone marrow biopsies of the control group. In contrast, the expression of membrane-associated MUC1C was reduced significantly in neoplastic plasma cells (Figure 2A) but did not differ between samples from preclinical and clinical MM stages. Interestingly, nuclear localization of MUC1C was also observed in a subpopulation of plasma cells in almost half the control cases (Figure 2A,B). In neoplastic plasma cells the frequency of nuclear MUC1C expression did not differ significantly between disease stages. EP cases showed the lowest frequency of MUC1C nuclear expression, but this finding was not statistically significant. The
4 M Andrulis et al.
A
P < 0.01
5E10 % 100 80 60 40 20 0
MGUS
P < 0.01
50
MGUS
Normal
MM
EP
clinically manifest
38
SMM n.s.
n.s.
64
n.s.
preclinical
100
5E5
47
41
MM
EP
n.s. clinically manifest
preclinical
5E5
% 100 80 60 40 20 0
SMM
60
EMA
Normal
clinically manifest
CD138
preclinical 93 79
100
MM
0 Normal
12
17
MGUS
SMM
MM
EP
P < 0.01 n.s. clinically manifest
preclinical
100
50 33
28
SMM
MM
21 Normal
MUC1C-N % 100 80 60 40 20 0
8
P < 0.001
MUC1C-M % 100 80 60 40 20 0
7
MUC1C
% 100 80 60 40 20 0
MGUS
BM
P < 0.05
n.s.
5E10
EMA
B
P < 0.05
MGUS
n.s.
n.s.
64
Normal
MGUS
C
n.s.
preclinical
50
EP
58
SMM
i
ii
iii
clinically manifest
45
MM
32
EP
Figure 2. Expression of MUC1 subunits in normal and neoplastic plasma cells. A, relative expression frequency of EMA, 5E10, 5E5, MUC1CM (membrane-bound MUC1C) and MUC1C-N (nuclear MUC1C) in normal plasma cells and different stages of plasma cell dyscrasia (MGUS, monoclonal gammopathy of undetermined significance; SMM, smouldering multiple myeloma; MM, multiple myeloma; EP, extramedullary plasmacytoma). Depicted are the percentages of positive cases, scored as described in detail in Materials and Methods. Statistical significance was assessed using the v2 test (n.s.: non-significant). B, Representative examples of plasma cells in biopsies from normal bone marrow (BM) preclinical (MGUS) and symptomatic (MM) disease stages expressing EMA, 5E10, 5E5 and nuclear MUC1C. Upper row shows biopsy areas stained for CD138 to visualize plasma cells. C, Observed variants of subcellular MUC1C localization: (i) membrane-bound, (ii) granular–cytoplasmic and (iii) nuclear, in these cases coexisting with granular–cytoplasmic pattern in some plasma cells. © 2013 John Wiley & Sons Ltd, Histopathology
MUC1 in multiple myeloma stages 5
WITH PATIENT SURVIVAL
The expression frequency of the epitopes investigated did not differ in untreated and relapsed/progressing MM cases. Furthermore, none of the investigated MUC1 epitopes correlated with overall survival as a single parameter. Therefore, we next tested whether or not a combination of epitopes was associated with any of the above-mentioned factors. For this we applied an unsupervised cluster analysis of scored immunohistochemical data to a subset of MM cases with available OS data (n = 61). This revealed that EMA clusters together with 5E10, whereas 5E5 clusters with membrane-associated MUC1C (Figure 4A). Interestingly, we found a number of MM cases that expressed either MUC1N (n = 12) or MUC1C (n = 11). Most importantly, the cluster analysis revealed a group of 12 MM cases that were negative for all MUC1 epitopes investigated and showed significantly shorter OS (Figure 4B). Pretreated/relapsed MM cases were significantly more frequent in this group (n = 10, P-value = 0.04, v2 test).
Discussion Although MUC1 is used widely as a marker for MM cells, our knowledge with regard to its specific glycosylation variants in MM is still sparse. In the present study we applied a panel of mAbs to MUC1, which enabled us to demonstrate that aberrant glycoforms of MUC1 defined by mAb 5E5 exhibit cancer-specific expression in MM and its precursor stages. MUC1 was expressed in normal cells and all MM stages (defined by mAb EMA) and characterised by a differ© 2013 John Wiley & Sons Ltd, Histopathology
100
90
90
80
80
Ki67 positive plasma cells (%)
100
70 60 50 40 30 20
B
70 60 50 40 30 20
10
10
0
0 negative positive
CORRELATION OF MUC1 EPITOPE EXPRESSION
MUC1C-N
MUC1C-M
A
Ki67 positive plasma cells (%)
granular–cytoplasmic MUC1C expression pattern was found in only 10 cases in total. Among those were control bone marrow samples and cases from all disease stages. Six cases showed coexistent nuclear MUC1C expression (BM = 1; SMM = 2; MM = 1; EP = 2) whereas four cases exhibited a granular–cytoplasmic pattern exclusively (MGUS = 3; SMM = 1). To test further how MUC1C subcellular localization correlates with the proliferation of MM cells we quantified MIB1-positive plasma cells in EP cases. The loss of nuclear MUC1C was associated significantly with increased proliferation rate (P-value = 0.05, Wilcox test, Figure 3A,B). In contrast, membrane-bound expression of MUC1-C was seen in cases with higher MIB1 score, but this observation did not reach statistical significance (Figure 3A,B).
MUC1C
negative positive Ki67
Figure 3. Correlation of subcellular localization of MUC1C and proliferative activity of MM cells. A, box-and-whisker plots showing the percentage of Ki67-expressing plasma cells in extramedullary plasmacytoma (EP) cases with membrane-bound (MUC1C-M, left) and nuclear (MUC1C-N, right) expression of MUC1C. The median is depicted by a horizontal line, the lower and upper quartiles (25% and 75%) are shown as a box and the minimum and maximum values are depicted by the whiskers. The cases with membranebound MUC1C have a higher proliferation rate, but this finding did not reach statistical significance (left). The loss of nuclear MUC1C expression is associated significantly with an increased proliferation rate (right, P-value = 0.05, Wilcox test). B, Representative examples of EP cases with nuclear (top row) and membrane-bound (bottom row) MUC1C expression (left) as well as corresponding serial sections demonstrating the proliferation rate by MIB1-staining (Ki67, right).
6 M Andrulis et al.
Colour key
0
1 2 3 Score MUC1C-N 5E5 MUC1C-M 5E10
EMA
1.0 Positive Negative
0.8
0.6
0.4
P = 0.02
0.2
0.0 0
50
100
150
200
Figure 4. MUC1 epitope expression: cluster analysis and correlation with overall survival. A, unsupervised cluster analysis of semiquantitatively scored immunohistochemical MUC1 epitope expression in a subset of patients with available survival data (n = 61). Black box marks a cluster of multiple myeloma cases (n = 12) that express none of the MUC1 epitopes investigated. B, MUC1-negative cases (black line) have a significantly shorter OS (months) compared to the cases with at least one MUC1 epitope detectable (red line).
ent reactivity pattern of mAb 5E10; it is likely that the density and sialylation of O-glycans increase during disease progression. The increased expression of Tn and STn in tumour cells may be due to either a reduced activity of distinct glycosyltransferases neces-
sary for the extension of cores 1 and 2 oligosaccharides or to the overexpression of sialyltransferases, which may block the chain elongation.27 In particular, the enzymatic activity of the core 1 b1,3-galactosyltransferase is controlled by the chaperone Cosmc, which may be inactivated by somatic point mutations or epigenetic silencing and lead to increased expression of the Tn antigen in cancer cells.28,29 The major difference in EMA and 5E10 binding properties is that 5E10 does not recognize MUC1N fully covered by sialylated oligosacharides.22 Thus, in MUC1N-expressing tumours with aberrant glycosylation one would expect a reduction of 5E10 binding and/or positivity for 5E5 mAb that binds specifically to Tn-antigen independent of its sialylation. Our data demonstrate this difference of 5E10 and 5E5 expression between control cases and MGUS. This indicates that MUC1N glycosylation changes are early events in the oncogenic transformation of plasma cells. In contrast, we did not observe significant differences in 5E10 and 5E5 expression between preclinical and symptomatic disease stages. Moreover, even EMA reactivity was lost in advanced MM stages. This may imply that complete loss of MUC1N from the cell surface but not aberrant glycosylation itself is associated with tumour progression. In the present work we provide the first detailed immunohistochemical analysis of MUC1C subcellular localization in various stages of plasma cell neoplasia. In accordance with published cell culture-based data, we found a significant loss of membrane-bound MUC1C in neoplastic plasma cells compared to normal plasma cells. Furthermore, the membrane-bound and nuclear MUC1C were mutually exclusive. This finding is in agreement with earlier observations that MUC1C indeed translocates into the nucleus to exert its signalling activity.19 Unexpectedly, we found nuclear MUC1C in 50% of normal controls, indicating that MUC1C translocation into the nucleus is not an aberrant event by itself. It may suggest that MUC1Cassociated signalling is involved in certain stages of normal plasma cell differentiation. Further studies are needed to address this question. We did not observe major differences of MUC1C nuclear expression frequency between controls and patient samples of various clinical disease stages. Nuclear MUC1C expression was increased slightly in MGUS and SMM compared to normal controls. Therefore, MUC1C nuclear translocation may occur early in the course of disease, which is in accordance with a possible oncogenic role of MUC1C. However, based on our observations, it is unlikely that nuclear MUC1C translocation is a major initiating event in the pathogenesis of multiple myeloma. Furthermore, our data reveal a tendency of reduced © 2013 John Wiley & Sons Ltd, Histopathology
MUC1 in multiple myeloma stages 7
nuclear MUC1C expression in the most advanced stages of myeloma, although this did not reach statistical significance. In addition, the EP cases with nuclear MUC1C expression were less proliferative compared to MUC1C negative cases. These findings may question a driver role of nuclear MUC1C in the disease progression, and are surprising because a number of in-vitro studies suggest that MUC1C may be involved in the oncogenic signalling of MM 13–15,20,21 and epithelial tumours.7 Conversely, a decreased expression of MUC1C was reported in some epithelial tumours, e.g. prostate cancer.30 Thus, the MUC1C signalling may be different in diverse types of tumours depending on the respective signalling pathways associated with MUC1C. Both MUC1 units are translated as a single polypeptide leading to equimolar expression of MUC1N and MUC1C. However, we found a number of MM cases that express either MUC1N or MUC1C exclusively. This finding indicates that different mechanisms such as shedding of MUC1N and cytoplasmic degradation of MUC1C may be responsible for loss of one of the MUC1 isoforms dependent upon its posttranslational modifications. Our results are in line with previously published data and confirm the expression of aberrantly glycosylated MUC1N on the malignant plasma cell surface as well as nuclear translocation of MUC1C in primary myeloma cells. Additionally, we demonstrate here that the expression of MUC1 is reduced in the clinically most advanced MM stages for which no further established therapeutic options are available. Moreover, the loss of MUC1 is associated significantly with adverse clinical outcome. These observations imply a pretreatment screening for MUC1 targets in patients for whom an anti-MUC1 therapy may be considered. Summarizing, the data presented here provide a concise analysis of both MUC1 subunits in multiple myeloma and its precursor stages. Our findings indicate that MUC1N glycosylation changes are early events in the pathogenesis of multiple myeloma. We further conclude that MUC1C nuclear localization is not an aberrant event and is unlikely to be a major driver of tumour progression in MM. We also demonstrate that complete loss of MUC1 is observed frequently in relapsed disease and is associated with adverse outcome.
Acknowledgements This study was supported by a grant of the ‘Deutsche Jose Carreras Leuk€ amie-Stiftung e.V. (DJCLS R 08/13) © 2013 John Wiley & Sons Ltd, Histopathology
to R.S.A. H.C. and U.M. were supported by the Danish National Research Foundation (DNRF107). We thank Tina Philipp and the Tissue Bank of the National Centre for Tumour Diseases Heidelberg for excellent technical assistance.
References 1. Morgan GJ, Walker BA, Davies FE. The genetic architecture of multiple myeloma. Nat. Rev. Cancer 2012; 12; 335–348. 2. Baldus SE, Engelmann K, Hanisch FG. MUC1 and the MUCs: a family of human mucins with impact in cancer biology. Crit. Rev. Clin. Lab. Sci. 2004; 41; 189–231. 3. Baldus SE, Monig SP, Huxel S et al. MUC1 and nuclear betacatenin are coexpressed at the invasion front of colorectal carcinomas and are both correlated with tumor prognosis. Clin. Cancer Res. 2004; 10; 2790–2796. 4. Cloosen S, Gratama J, van Leeuwen EB et al. Cancer specific mucin-1 glycoforms are expressed on multiple myeloma. Br. J. Haematol. 2006; 135; 513–516. 5. Kawano T, Ahmad R, Nogi H, Agata N, Anderson K, Kufe D. MUC1 oncoprotein promotes growth and survival of human multiple myeloma cells. Int. J. Oncol. 2008; 33; 153– 159. 6. Kufe DW. Mucins in cancer: function, prognosis and therapy. Nat. Rev. Cancer 2009; 9; 874–885. 7. Uchida Y, Raina D, Kharbanda S, Kufe D. Inhibition of the MUC1-C oncoprotein is synergistic with cytotoxic agents in the treatment of breast cancer cells. Cancer Biol. Ther. 2013; 14; 127–134. 8. Hanisch FG, Muller S. MUC1: the polymorphic appearance of a human mucin. Glycobiology 2000; 10; 439–449. 9. Cao Y, Merling A, Karsten U et al. Expression of CD175 (Tn) CD175s (sialosyl-Tn) and CD176 (Thomsen–Friedenreich antigen) on malignant human hematopoietic cells. Int. J. Cancer 2008; 123; 89–99. 10. Pinto R, Carvalho AS, Conze T et al. Identification of new cancer biomarkers based on aberrant mucin glycoforms by in situ proximity ligation. J. Cell Mol. Med. 2012; 16; 1474–1484. 11. Beatson RE, Taylor-Papadimitriou J, Burchell JM. MUC1 immunotherapy. Immunotherapy 2010; 2; 305–327. 12. Mall AS. Analysis of mucins: role in laboratory diagnosis. J. Clin. Pathol. 2008; 61; 1018–1024. 13. Ahmad R, Raina D, Joshi MD et al. MUC1-C oncoprotein functions as a direct activator of the nuclear factor-kappa b p65 transcription factor. Cancer Res. 2009; 69; 7013–7021. 14. Ahmad R, Raina D, Trivedi V et al. MUC1 oncoprotein activates the Ikappa b kinase beta complex and constitutive NFkappa b signalling. Nat. Cell Biol. 2007; 9; 1419–1427. 15. Ahmad R, Rajabi H, Kosugi M et al. MUC1-C oncoprotein promotes STAT3 activation in an autoinductive regulatory loop. Sci. Signal 2011; 4: ra9. 16. Raina D, Ahmad R, Chen D, Kumar S, Kharbanda S, Kufe D. MUC1 oncoprotein suppresses activation of the ARF-MDM2p53 pathway. Cancer Biol. Ther. 2008; 7; 1959–1967. 17. Wei X, Xu H, Kufe D. Human MUC1 oncoprotein regulates p53-responsive gene transcription in the genotoxic stress response. Cancer Cell 2005; 7; 167–178. 18. Leng Y, Cao C, Ren J et al. Nuclear import of the MUC1-C oncoprotein is mediated by nucleoporin NUP62. J. Biol. Chem. 2007; 282; 19321–19330.
8 M Andrulis et al.
19. Li Y, Chen W, Ren J et al. Df3/MUC1 signaling in multiple myeloma cells is regulated by interleukin-7. Cancer Biol. Ther. 2003; 2; 187–193. 20. Yin L, Ahmad R, Kosugi M et al. Survival of human multiple myeloma cells is dependent on MUC1 c-terminal transmembrane subunit oncoprotein function. Mol. Pharmacol. 2010; 78; 166–174. 21. Yin L, Kosugi M, Kufe D. Inhibition of the MUC1-C oncoprotein induces multiple myeloma cell death by down-regulating tigar expression and depleting NADPH. Blood 2012; 119; 810–816. 22. Tarp MA, Sørensen AL, Mandel U et al. Identification of a novel cancer-specific immunodominant glycopeptide epitope in the MUC1 tandem repeat. Glycobiology 2007; 17; 197–209. 23. Sørensen AL, Reis CA, Tarp MA et al. Chemoenzymatically synthesized multimeric Tn/STn MUC1 glycopeptides elicit cancer-specific anti-MUC1 antibody responses and override tolerance. Glycobiology 2006; 16; 96–107. 24. Andrulis M, Dietrich S, Longerich T et al. Loss of endothelial thrombomodulin predicts response to steroid therapy and sur-
25.
26. 27.
28.
29.
30.
vival in acute intestinal graft-versus-host disease. Haematologica 2012; 97; 1674–1677. R Core Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2012. Warnes GR. Gplots: various R programming tools for plotting data 2012. http://cran.r-project.org/web/packages/gplots/index.html Engelmann K, Baldus SE, Hanisch FG. Identification and topology of variant sequences within individual repeat domains of the human epithelial tumor mucin MUC1. J. Biol. Chem. 2001; 276; 27764–27769. Ju T, Aryal RP, Stowell CJ, Cummings RD. Regulation of protein o-glycosylation by the endoplasmic reticulum-localized molecular chaperone cosmc. J. Cell Biol. 2008; 182; 531–542. Mi R, Song L, Wang Y et al. Epigenetic silencing of the chaperone cosmc in human leukocytes expressing Tn antigen. J. Biol. Chem. 2012; 287; 41523–41533. Singh AP, Chauhan SC, Bafna S et al. Aberrant expression of transmembrane mucins, MUC1 and MUC4, in human prostate carcinomas. Prostate 2006; 66; 421–429.
© 2013 John Wiley & Sons Ltd, Histopathology
