Original Article doi: 10.1111/joim.12375
Reduced expression of TRIM21/Ro52 predicts poor prognosis in diffuse large B-cell lymphoma patients with and without rheumatic disease S. Brauner1, W. Zhou1, C. Backlin2, T. M. Green3,4, L. Folkersen1, M. Ivanchenko1, B. L€ofstr€om5, Z. Y. Xu-Monette6, K. H. Young6, L. Møller Pedersen4, M. Boe Møller3,4, C. Sundstr€om7, G. Enblad8, E. Baecklund2 & M. Wahren-Herlenius1* From the 1Unit of Experimental Rheumatology, Department of Medicine, Karolinska University Hospital, Karolinska Institutet, Stockholm, Sweden; 2Unit of Rheumatology, Department of Medical Sciences, Uppsala University, Uppsala, Sweden; 3Institute of Clinical Research, University of Southern Denmark; 4Odense University Hospital, Odense, Denmark; 5Rheumatology Clinic, Malar Hospital, Eskilstuna, Sweden; 6Department of Hematopathology, University of Texas, MD Anderson Cancer Center, Houston, TX, USA; 7Department of Immunology, Genetics and Pathology, Rudbeck Laboratory, Uppsala University; and ; 8Unit of Oncology, Department of Radiology, Oncology and Radiation Sciences, Uppsala University, Uppsala, Sweden *Current address: Center for Biological Sequence Analysis, Technical University of Denmark, Copenhagen, Denmark.
Abstract. Brauner S, Zhou W, Backlin C, Green TM, Folkersen L, Ivanchenko M, L€ ofstr€ om B, XuMonette ZY, Young KH, Møller Pedersen L, Boe Møller M, Sundstr€ om C, Enblad G, Baecklund E, Wahren-Herlenius M (Karolinska University Hospital, Karolinska Institutet, Stockholm; Uppsala University, Uppsala, Sweden; Institute of Clinical Research, University of Southern Denmark, Odense; Odense University Hospital, Odense, Denmark; Malar Hospital, Eskilstuna, Sweden; University of Texas, MD Anderson Cancer Center, Houston, TX, USA; Uppsala University, Uppsala; Uppsala University, Uppsala, Sweden). Reduced expression of TRIM21/Ro52 predicts poor prognosis in diffuse large B-cell lymphoma patients with and without rheumatic disease. J Intern Med 2015; 278: 323–332. Objective. TRIM21 (also known as Ro52) is an autoantigen in rheumatic disease and is predominantly expressed in leucocytes. Overexpression is associated with decreased proliferation, and the TRIM21 gene maps to a tumour suppressor locus. We therefore investigated the expression of TRIM21 in patients with diffuse large B-cell lymphoma (DLBCL) and its potential usefulness as a prognostic biomarker. Materials and methods. TRIM21 expression levels were assessed by immunohistochemistry in lymphoma biopsies from three cohorts of patients with DLBCL: 42 patients with rheumatic disease treated with a cyclophosphamide, vincristine, doxorubicin and
prednisone (CHOP)-like regimen, 76 CHOP-treated and 196 rituximab-CHOP-treated nonrheumatic patients. Expression was correlated with clinical and biomedical parameters. TRIM21 expression was assessed in relation to lymphocyte proliferation by quantitative PCR and correlated with 3H-thymidine incorporation and propidium iodine staining. Results. TRIM21 expression levels differed in the lymphomas compared to normal lymphoid tissue, with reduced expression correlating with shorter overall survival in all three cohorts. In the two larger cohorts, progression-free survival was assessed and was also found to correlate with TRIM21 expression. The association was independent of commonly used clinical prognostic scores, lymphoma subtype and several previously reported prognostic biomarkers. In agreement with this clinical observation, we noted an inverse correlation between TRIM21 expression and proliferation of leucocytes in vitro. Conclusions. We show that loss of TRIM21 expression is associated with more aggressive lymphoma and increased proliferation, whereas maintenance of TRIM21 expression is associated with better prognosis in patients with DLBCL. Based on our findings, we suggest that TRIM21 should be considered as a novel biomarker for lymphoma characterization and for predicting patient survival. Keywords: biomarker, diffuse large B-cell lymphoma, prognosis, Ro52, TRIM21.
ª 2015 The Association for the Publication of the Journal of Internal Medicine
323
S. Brauner et al.
Introduction Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoma type, accounting for 30–40% of adult lymphomas. The prognosis of patients with DLBCL is highly variable, and despite recent therapeutic advances with the addition of rituximab (anti-CD20) to the anthracycline-based treatment, approximately 40% of patients relapse [1]. Gene expression profiling studies have led to the division of DLBCL into two major subtypes based on the putative cell of origin: germinal centre B-cell-like (GCB) and activated B-cell-like (ABC) subtypes [2, 3]. Survival is better in the GCB compared to the ABC group, but division into subtype has not proven specific enough to identify patients with particularly poor prognosis. In addition, several biomarkers have emerged to better classify and predict outcome at diagnosis, but are not yet routinely used in clinical practice [4]. There is thus an urgent need for more precise and clinically reliable prognostic markers. TRIM21 (Ro52) was first described as an autoantigen in rheumatic diseases. Interestingly, most patients with rheumatic disease have an elevated risk of developing lymphomas, ranging from twofold to 16-fold increased risk compared to the general population, depending on the underlying disease [5, 6]. Our attention was drawn to the possible role of TRIM21 in lymphoma and its use as a biomarker for several reasons. First, the TRIM21 gene is located in a tumour suppressor locus on the short arm of chromosome 11 (11p15.5), a locus that has been associated with malignancies of many organs [7]. Secondly, TRIM21 is mainly expressed in cells of haematopoietic origin at steady state [8, 9]. Thirdly, we have previously demonstrated that increased expression of TRIM21 leads to decreased proliferative capacity in a mouse B-cell lymphoma cell line [9]. Finally, TRIM21 is an E3 ligase involved in ubiquitination, and there are emerging data to suggest that it regulates multiple proteins important in the cell cycle and apoptosis, such as BCL2 and interferon regulatory factor (IRF)-8 as well as FADD and c-FLIP [8, 10–12]. Based on these observations, we investigated the role of TRIM21 in lymphoma and its potential as a prognostic biomarker in patients with DLBCL. Materials and methods Study population and patient characteristics Three cohorts of patients with DLBCL were included in the study. Cohort 1 consisted of patients with 324
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
TRIM21: a novel biomarker in DLBCL
rheumatic disease treated with an anthracyclinebased regimen. The other cohorts consisted of patients without rheumatic disease treated with either cyclophosphamide, vincristine, doxorubicin and prednisone (CHOP) (Cohort 2) or rituximabCHOP (R-CHOP) (Cohort 3). Patients with primary CNS-, post-transplant- or HIV-related lymphomas were excluded from all cohorts. All lymphomas were re-assessed according to the 2001 World Health Organization classification of tumours by at least two independent pathologists prior to inclusion [4, 13–16]. The lymphomas of patients in cohorts 1 and 2 were classified as GCB and ABC (non-GCB) by an immunostaining algorithm [3]; Cohort 3 was classified by a gene expression profiling and/or immunostaining algorithm [14, 16]. The study was approved by the local ethics committees. Cohort 1 was identified by linking the Swedish Inpatient Register and the Swedish Cancer Register from the period 1964–1995, as previously described [13, 17]. In total, 97 patients with rheumatoid arthritis (RA) or systemic lupus erythematosus (SLE) diagnosed with DLBCL were identified. Due to the long recruitment period, treatment regimens varied substantially, and therefore, only the 42 patients who received anthracycline-based treatment were included in this study (33 with RA and nine with SLE). Patients were followed until July 2010. Mean age at lymphoma diagnosis was 62 years (range 33– 88 years). Mean follow-up was 31 months (range 1–267 months), and 78% of patients were classified as having Ann Arbor stage III or IV, indicating an aggressive lymphoma subtype. Cohort 2 comprised 76 patients with DLBCL treated with anthracycline-based CHOP, diagnosed at Uppsala University Hospital between 1984 and 2002 [4]. Patients were followed until June 2012. The mean age at lymphoma diagnosis was 66 years and mean follow-up time was 64 months (range 4 days to 255 months). In this cohort, 58% of patients were classified with Ann Arbor stage III or IV and 11 (16%) had an International Prognostic Index (IPI; a novel and more accurate predictive score of lymphoma severity) of 3–4. Four patients (10%) in Cohort 1 and 12 (16%) in Cohort 2 were alive at the end of follow-up. All surviving patients were diagnosed with lymphomas expressing high levels of TRIM21. Cohort 3 included 196 R-CHOP-treated patients with DLBCL, as previously described [14, 16].
N/A
N/A, not available; CHOP, cyclophosphamide, vincristine, doxorubicin and prednisone; R-CHOP, rituximab-CHOP; IPI, International Prognostic Index. GC, Germinal centre like B-cell subtype. a According to the Hans algorithm [3].
0.45
0.21 54.8
32.5 40
63.2 55.4
34 0.091
1 46
10.4 33 11
40 33.3
N/A N/A
24.4 GC positivitya
(%)
IPI >2 (%)
17.4
0.29
44.1
0.48 46.5 52.5 47.4 0.726 56.7 66.7 57.9 0.43 86.4 72.2 78 Ann Arbor stage III/IV
(years)
0.92
0.65 62.7 63.9 62.9 0.521 66.7 71.3 66.2 0.048 72.6 69.9 Mean age at diagnosis
66.6
41.7
(n = 156) (n = 40)
42.5 42.1
(n = 196) value
0.499 47.1
(n = 67) (n = 9)
33 45.5
(n = 76) value
0.07 58.3
(n = 24) (n = 18) (n = 42)
66.7
>40% ≤40% >40% ≤40%
Cohort 2: CHOP-treated patients
All P>40% ≤40% All
Cohort 1: patients with rheumatic disease
Table 1 Clinical characteristics of patient cohorts
Sections were analysed using a Polyvar II light microscope (Reichert-Jung, Vienna, Austria) at 910 magnification. Most biopsies displayed homogenous lymphoma tissue and were therefore scored as one entity. Biopsies containing non-lymphoma tissue were only scored in lymphoma-rich areas. A semi-quantitative scoring system was developed using a scale ranging from 0 (no expression) to 3 (ubiquitous expression). Scoring was performed in a blinded manner by two independent investigators (SB and MWH). In addition, computer-assisted image analysis of whole core biopsies was performed to define the percentage of stained lymphoma tissue area using the Quantimet 600 image analyser (Leica, Wetzlar, Germany). The average values were used for statistical analysis. If two biopsies of the same tissue differed by more than 10 percentage points in a stained area, they were reanalysed. Staining and scoring of the included biopsies for BCL2, BCL6 and IRF-4 (MUM1) expression have been described previously [4, 14, 20].
P-
Tissue microarray blocks were constructed as previously described [14, 17]. Briefly, two representative core biopsies from each obtained lymphoma biopsy were arranged into microarray paraffin blocks and sectioned into 4-lm sections. Normal lymphoid tonsillar tissue was used as staining control [17, 18]. Deparaffinized sections were subjected to antigen retrieval and subsequently stained with mouse anti-human TRIM21 antibodies, clones 7.8C7 and 7.12E11 at 2 lg mL1, using an automated Ventana XT module (Ventana Medical Systems, Tucson, AZ, USA) [19]. The antibodies target two different epitopes of the protein: clone 7.8C7 recognizes the centrally located coiled-coil domain and 7.12E11 the Cterminal RING domain of the TRIM21 protein. Binding was visualized with the iVIEW DAB detection kit (Ventana Medical Systems).
61.9
Immunohistochemistry
All
Cohort 3: R-CHOP-treated patients
P-
Patients were diagnosed between June 2001 and October 2009 and followed until July 2011. The mean age at lymphoma diagnosis was 63 years (range 16 to 91 years). In this cohort, 67 patients (34%) were classified as having an IPI >2 and 94 (48%) with an Ann Arbor stage III or IV. Follow-up times ranged from 14 days to almost 99 months. The majority (65%) of included patients were alive at the end of the study. Characteristics of the patients in all three cohorts are summarized in Table 1.
value
TRIM21: a novel biomarker in DLBCL
Female (%)
S. Brauner et al.
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
325
S. Brauner et al.
Cell culture, in vitro proliferation assay and real-time PCR Peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using Ficoll-Paque plus (GE Healthcare, Little Chalfont, UK). Freshly isolated cells were seeded in 96-well plates (1.5 9 105 cells per well) and stimulated with pokeweed mitogen at 1 and 5 lg mL1 for 5 days. Proliferation was evaluated by 3H-thymidine incorporation, as previously described [9]. Cell cycle phase was assessed by flow cytometry using propidium iodide. Briefly, cells were stained with antiCD19 (clone 4G7, BD Bioscience, Franklin Lakes, NJ, USA), fixed on ice in 80% ethanol/20% in phosphate-buffered saline (PBS) for 1 h and then treated with propidium iodide and RNAse for 5 min, prior to analysis (FACSCalibur, BD Biosciences). Total mRNA was extracted as previously described [9], using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA) and was reverse transcribed into cDNA using random hexamers (Life Technologies, Carlsbad, CA, USA). TRIM21 expression was analysed with quantitative SYBR Green PCR (Qiagen) using GAPDH as the internal control with primers: TRIM21-F, 50 -GAACTGCTGCAGGAGGTGATAA-30 and TRIM21-R, 50 -AGTTCTGGAGAGGTAATATCCA GGTC-30 ; GAPDH-F, 50 -AGGGCTGCTTTTAACTCT GGTAAA-30 and GAPDH-R, 50 -CATATTGGAACATG TAAACCATGTAGTTG-30 . Statistical analysis Overall survival was defined as the time from diagnosis to death or last day of follow-up; progression-free survival was defined as the time from diagnosis to progression, death or last day of followup. Prism graph pad 5 (La Jolla, CA, USA) was used to compare two groups with the Mann–Whitney Utest, to perform logistic regression, to analyse correlations between variables and to create Kaplan–Meier plots for survival analysis. Cox proportional hazards regression models were calculated using R-3.1.0 software (R foundation for Statistical Computing, Vienna, Austria). Because of non-normality of data distribution, predefined cut-off values were used as indicated, and the resulting binary variables were used in the primary analysis. The optimal cut-off value for low and high expression of TRIM21 in lymphomas was determined by receiver operator characteristic (ROC) curve analysis, and values for all other variables were based on clinical standards. Hazard ratio (HR) was reported as either exp(coef) or exp(-coef) to allow comparable 326
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
TRIM21: a novel biomarker in DLBCL
HR in presentation, independent of how expression changes associates with risk. In all analyses, Pvalues 40% of the cells expressing TRIM21 was considered high expression), based on ROC curve analysis in Cohort 1, yielding an assay sensitivity of 100% (data not shown). This cut-off value was used for further analysis in all cohorts. Expression of TRIM21 is a prognostic indicator of survival in DLBCL patients with rheumatic disease As TRIM21 is an autoantigen in rheumatic disease, we first investigated its potential prognostic value in DLBCL patients with RA or SLE (Cohort 1). To our surprise, we observed a highly significant correlation between high TRIM21 expression and longer overall survival. Median overall survival was 6 months in patients with low TRIM21 expression compared to 11 months in those with high TRIM21 expression (Fig. 1c, P = 0.0073). Cox regression analysis indicated an HR of 1.41 [95% confidence
S. Brauner et al.
TRIM21: a novel biomarker in DLBCL
(b)
(a)
0
1
2
3
Rheumatic anthracycline treated (Overall survival)
(c) 1.0
TRIM21 high (n = 24) 0.8
TRIM21 low (n = 18) P = 0.0073
0.6 0.4 0.2 0.0 0
50
100 150 200 Time (Months)
(e) 60
Rheumatic anthracycline treated (Overall survival)
(d) % Ro52/TRIM21 expression
Fig. 1 Expression of the autoantigen TRIM21 is prognostic for survival in patients with rheumatic disease and lymphoma, independent of autoantibody status. All lymphoma biopsies were stained with two different anti-TRIM21 antibodies (7.8C7 and 7.12E11) binding different epitopes of the protein; antibody 7.8C7 staining was used in the analysis throughout the study. (a) Example of lymphomas scored manually (score 0–3). (b) Both manual and computerized scores were used, and results were compared with linear regression (r2 = 0.67, P < 0.0001). (c) Kaplan–Meier curves of overall survival in 42 lymphoma patients with rheumatic disease treated with anthracycline-based drugs show a correlation between high (>40%) TRIM21 expression and longer survival (P < 0.01). Antinuclear antibody (ANA) status was tested in 30 of the patients. (d) TRIM21 expression in lymphomas stratified by ANA status shows no association between autoantibodies and protein expression (P = 0.39). (e) Survival is unaffected by ANA status, as shown by Kaplan–Meier curves (P = 0.65).
n.s.
40
20
0
ANA pos
interval (CI) 0.97–2.05, P = 0.076]. As this was a retrospectively collected cohort, specific anti-SSA/ Ro52 autoantibody titres were not measured and serum samples were not available for analysis. However, antinuclear antibody status was recorded in 70% of patients. No correlation was observed between TRIM21 expression level and autoantibody positivity (Fig. 1d, P = 0.39), or overall survival (Fig. 1e, P = 0.65), indicating that the antinuclear autoantibodies do not affect the expression pattern of TRIM21. Predictive significance of TRIM21 protein expression in CHOP- and RCHOP-treated patients To investigate whether TRIM21 expression is a predictive factor for survival also in patients with-
ANA neg
250
300
1.0
ANA pos (n = 16) ANA neg (n = 14)
0.8 0.6
P = 0.65 0.4 0.2 0.0 0
50
100
150
200
250
300
Time (Months)
out rheumatic disease, we examined a cohort of 74 CHOP-treated patients with DLBCL (Cohort 2). Also here, we observed that low TRIM21 expression was associated with shorter overall survival (median survival 10 vs. 48 months; Fig. 2a, P < 0.0001), with an HR of 1.56 (95% CI 1.10–1.99, P < 0.002). In addition, the time from remission to disease progression, the median progression-free survival, was 1 month in the group with lymphomas with low TRIM21 expression compared to 60 months in the high expression group (Fig. 2b, P < 0.0001). Today, most patients with DLBCL are treated with a CHOP-like regimen in combination with rituximab. To verify the role of TRIM21 as a prognostic biomarker in DLBCL, we investigated an additional cohort of 196 R-CHOP-treated patients (Cohort 3). ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
327
S. Brauner et al.
TRIM21: a novel biomarker in DLBCL
1.0
TRIM21 high (n = 67)
0.8
TRIM21 low (n = 9)
0.6
P < 0.0001
0.4 0.2 0.0 0
50
100
150
200
250
(b)
CHOP treated (Progression-free survival)
CHOP treated (Overall survival)
(a)
TRIM21 high (n = 67)
1.0
TRIM21 low (n = 9)
0.8 0.6
P < 0.0001
0.4 0.2 0.0 0
300
50
TRIM21 high (n = 154)
1.0
TRIM21 low (n = 42)
0.8 0.6 0.4 P < 0.0001
0.2 0.0 0
20
40
60
80
(d)
R-CHOP treated (Progression-free survival)
R-CHOP treated (Overall survival)
(c)
1.0
300
TRIM21 low (n = 42)
0.4 P = 0.0005
0.2 0.0 0
0.8 0.6 0.4 P = 0.0006
(f)
20
40
60
80
100
TRIM21 high (n = 70)
1.0
ABC R-CHOP treated (Overall survival)
GCB R-CHOP treated (Overall survival)
250
Time (Months)
TRIM21 low (n = 25)
0.2
200
0.6
100
TRIM21 high (n = 84)
1.0
150
TRIM21 high (n = 154)
0.8
Time (Months)
(e)
100
Time (Months)
Time (Months)
TRIM21 low (n = 17)
0.8 0.6 0.4 P = 0.0018
0.2 0.0
0.0 0
20
40
60
80
100
Time (Months)
0
20
40
60
80
100
Time (Months)
Fig. 2 Expression of TRIM21 correlates with overall survival (OS) and progression-free survival (PFS) in patients without rheumatic disease, independent of treatment and disease subtype. A cohort of 76 cyclophosphamide, vincristine, doxorubicin and prednisone (CHOP)-treated diffuse large B-cell lymphoma (DLBCL) patients without rheumatic disease was investigated. Kaplan–Meier curves are shown for OS (a) (P < 0.0001) and PFS (b) (P < 0.0001) comparing high (>40%) and low ( 30%, IRF-4 > 30%, BCL6 > 30% staining, IPI >2 and Ann Arbor stage >2. CI, confidence interval; IPI, international prognostic index. ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
329
(a)
7
(b)
7
6
TRIM21 expression (relative to GAPDH)
TRIM21: a novel biomarker in DLBCL
TRIM21 expression (relative to GAPDH)
S. Brauner et al.
6
5
P < 0.0001 r2 = 0.5032
4 3 2 1 0
5
P = 0.0003 r2 = 0.4022
4 3 2 1 0
0
25 000
50 000
75 000
100 000
cpm
0
5
10
15
20
25
30
(S+G2)/M
Fig. 3 TRIM21 mRNA expression is downregulated in proliferating cells. Freshly isolated peripheral blood mononuclear cells from healthy donors were stimulated with pokeweed mitogen (1 and 5 lg mL1) for 5 days. TRIM21 mRNA expression levels quantified by real-time PCR were by correlated by linear regression to 3H-thymidine incorporation (a) (r2 = 0.50, P < 0.0001) and cell cycle phase measured by flow cytometry (b) (r2 = 0.40, P < 0.001).
against TRIM21 as well as an increased risk of lymphoma development. We proceeded to analyse the prognostic relevance also in DLBCL patients without rheumatic disease and, interestingly, found that it was equally significant in both groups, indicating that TRIM21 expression is an independent marker of lymphoma outcome. Recently published data have demonstrated that TRIM21 is an E3 ligase involved in ubiquitination, and several IRFs have been identified as targets of TRIM21-mediated ubiquitination [8, 21, 22]. Notably, IRF-4 (MUM1) has been proposed to be involved in lymphoma development and is used as a biomarker in DLBCL [3]. Further, Jauharoh and colleagues have shown that BCL2 protein levels are regulated by TRIM21-mediated ubiquitination [10]. In addition, the apoptosis regulatory protein BCL2 has been associated with lymphoma development [23]. GCB lymphomas are associated with a specific translocation, t(14:18), where the BCL2 gene is translocated to the IgH promoter, leading to protein overexpression and apoptotic resistance [24]. TRIM21 also acts on the cell cycle regulator p27 [25]. Of note, it ha been proposed that TRIM21 monoubiquitinates the upstream regulator of inhibitor kappa B (I-kB), the inhibitory kinase of kappa-beta (IKK-beta), and thereby alters its function [26]. Inhibition of the degradation of IkB has been suggested as the main mode of action of the 26S proteasome inhibitor bortezomib [27], which has shown efficacy in the treatment of both myeloma and lymphoma [28, 29]. Increased levels of I-kB lead to inhibition of the NF-kB pathway, 330
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
subsequently decreasing cellular resistance to apoptosis. This pathway is constitutively active in ABC lymphomas [30], and bortezomib has shown promise as second-line treatment in patients with this subgroup [31]. Together, these findings suggest that TRIM21 controls several pathways involved in cell proliferation and apoptosis, and the role of TRIM21 as a central regulator of cellular homoeostasis is further emphasized by the observation that TRIM21-null mice develop lymphoproliferation with lymphadenopathy and splenomegaly [8]. Hence, there is a biological and cellular basis for the observed low expression of TRIM21 in patients with the poorest survival. It is possible that restoring TRIM21 expression could be clinically beneficial. Both type 1 and type 2 interferons (IFNs) upregulate TRIM21 expression, which can be detected as induced under proinflammatory conditions in several tissues including lymphoid tissues [8, 18, 32, 33]. Interestingly, treatment with type 1 IFNs (IFN-alpha) is a secondline consideration for therapy in specific lymphoma subsets [34–36]. It is possible that patients with no or a low level of TRIM21 expression in their tumours might benefit the most from IFN-alpha treatment, based on the theory that this cytokine would re-establish TRIM21 expression and the control of proliferation and apoptosis. A reliable biomarker for selecting patients for IFN-alpha treatment is important not only to identify patients who would benefit most from the therapy but also to avoid unnecessary treatment of others, especially because of the adverse effects of the drug.
S. Brauner et al.
In summary, our data show for the first time that the level of expression of TRIM21 in lymphoma tissue is a strong prognostic marker in patients with DLBCL. The role of TRIM21 as a biomarker for lymphoma outcome may have a biological basis as TRIM21 regulates cellular proliferation and apoptosis. Further, here we show that the prognostic value of this marker is independent of Ann Arbor and IPI classification as well as subtype, suggesting that TRIM21 expression adds prognostic value to current tools in patients with DLBCL.
TRIM21: a novel biomarker in DLBCL
9
10
11
12
Acknowledgements
13
This study was supported by grants from the Swedish Research Council, the Heart-Lung Foundation, Stockholm County Council, Karolinska Institutet, the Swedish Rheumatism Association, the King Gustaf Vth 80-year Foundation and the Torsten and Ragnar S€oderberg Foundation.
14
15
Conflict of interest statement The authors have no competing interests to declare.
References 1 Fu K, Weisenburger DD, Choi WW et al. Addition of rituximab to standard chemotherapy improves the survival of both the germinal center B-cell-like and non-germinal center B-celllike subtypes of diffuse large B-cell lymphoma. J Clin Oncol 2008; 26: 4587–94. 2 Alizadeh AA, Eisen MB, Davis RE et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 2000; 403: 503–11. 3 Hans CP, Weisenburger DD, Greiner TC et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood 2004; 103: 275–82. 4 Berglund M, Thunberg U, Amini RM et al. Evaluation of immunophenotype in diffuse large B-cell lymphoma and its impact on prognosis. Mod Pathol 2005; 18: 1113–20. 5 Dias C, Isenberg DA. Susceptibility of patients with rheumatic diseases to B-cell non-Hodgkin lymphoma. Nat Rev Rheumatol 2011; 7: 360–8. 6 Theander E, Henriksson G, Ljungberg O, Mandl T, Manthorpe R, Jacobsson LT. Lymphoma and other malignancies in primary Sjogren’s syndrome: A cohort study on cancer incidence and lymphoma predictors. Ann Rheum Dis 2006; 65: 796–803. 7 Feinberg AP. Imprinting of a genomic domain of 11p15 and loss of imprinting in cancer: an introduction. Cancer Res 1999; 59: 1743s–6s. 8 Espinosa A, Dardalhon V, Brauner S et al. Loss of the lupus autoantigen Ro52/Trim21 induces tissue inflammation and
16
17
18
19
20
21
22
23
24
systemic autoimmunity by disregulating the IL-23-Th17 pathway. J Exp Med 2009; 206: 1661–71. Espinosa A, Zhou W, Ek M et al. The Sjogren’s syndromeassociated autoantigen Ro52 is an E3 ligase that regulates proliferation and cell death. J Immunol 2006; 176: 6277–85. Jauharoh SN, Saegusa J, Sugimoto T et al. SS-A/Ro52 promotes apoptosis by regulating Bcl-2 production. Biochem Biophys Res Commun 2012; 417: 582–7. Young JA, Sermwittayawong D, Kim HJ et al. Fas-associated death domain (FADD) and the E3 ubiquitin-protein ligase TRIM21 interact to negatively regulate virus-induced interferon production. J Biol Chem 2011; 286: 6521–31. Zhang J, Fang L, Zhu X et al. Ro52/SSA sensitizes cells to death receptor-induced apoptosis by down-regulating c-FLIP (L). Cell Biol Int 2012; 36: 463–8. Baecklund E, Sundstrom C, Ekbom A et al. Lymphoma subtypes in patients with rheumatoid arthritis: increased proportion of diffuse large B cell lymphoma. Arthritis Rheum 2003; 48: 1543–50. Green TM, Young KH, Visco C et al. Immunohistochemical double-hit score is a strong predictor of outcome in patients with diffuse large B-cell lymphoma treated with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone. J Clin Oncol 2012; 30: 3460–7. Organization WH. World Health Organization classification of tumours: pathology and genetics of tumours of haematopoietic and lymphoid tissue. IARC Press, 2001. Visco C, Li Y, Xu-Monette ZY et al. Comprehensive gene expression profiling and immunohistochemical studies support application of immunophenotypic algorithm for molecular subtype classification in diffuse large B-cell lymphoma: a report from the International DLBCL Rituximab-CHOP Consortium Program Study. Leukemia 2012; 26: 2103–13. Lofstrom B, Backlin C, Pettersson T, Lundberg IE, Baecklund E. Expression of APRIL in diffuse large B cell lymphomas from patients with systemic lupus erythematosus and rheumatoid arthritis. J Rheumatol 2011; 38: 1891–7. Oke V, Vassilaki I, Espinosa A et al. High Ro52 expression in spontaneous and UV-induced cutaneous inflammation. J Invest Dermatol 2009; 129: 2000–10. Strandberg L, Ambrosi A, Espinosa A et al. Interferon-alpha induces up-regulation and nuclear translocation of the Ro52 autoantigen as detected by a panel of novel Ro52-specific monoclonal antibodies. J Clin Immunol 2008; 28: 220–31. Baecklund E, Backlin C, Mansouri M et al. LMO2 protein expression predicts survival in patients with rheumatoid arthritis and diffuse large B-cell lymphoma. Leuk Lymphoma 2011; 52: 1146–9. Higgs R, Ni Gabhann J, Ben Larbi N, Breen EP, Fitzgerald KA, Jefferies CA. The E3 ubiquitin ligase Ro52 negatively regulates IFN-beta production post-pathogen recognition by polyubiquitin-mediated degradation of IRF3. J Immunol, 2008; 181: 1780–6. Yoshimi R, Chang TH, Wang H, Atsumi T, Morse HC 3rd, Ozato K. Gene disruption study reveals a nonredundant role for TRIM21/Ro52 in NF-kappaB-dependent cytokine expression in fibroblasts. J Immunol, 2009; 182: 7527–38. Iqbal J, Neppalli VT, Wright G et al. BCL2 expression is a prognostic marker for the activated B-cell-like type of diffuse large B-cell lymphoma. J Clin Oncol 2006; 24: 961–8. Huang JZ, Sanger WG, Greiner TC et al. The t(14;18) defines a unique subset of diffuse large B-cell lymphoma with a
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
331
S. Brauner et al.
25
26
27
28
29
30
31
332
germinal center B-cell gene expression profile. Blood 2002; 99: 2285–90. Sabile A, Meyer AM, Wirbelauer C et al. Regulation of p27 degradation and S-phase progression by Ro52 RING finger protein. Mol Cell Biol 2006; 26: 5994–6004. Wada K, Niida M, Tanaka M, Kamitani T. Ro52-mediated monoubiquitination of IKK{beta} down-regulates NF-{kappa} B signalling. J Biochem 2009; 146: 821–32. Hideshima T, Anderson KC. Molecular mechanisms of novel therapeutic approaches for multiple myeloma. Nat Rev Cancer 2002; 2: 927–37. Goy A, Younes A, McLaughlin P et al. Phase II study of proteasome inhibitor bortezomib in relapsed or refractory Bcell non-Hodgkin’s lymphoma. J Clin Oncol 2005; 23: 667–75. Richardson PG, Barlogie B, Berenson J et al. A phase 2 study of bortezomib in relapsed, refractory myeloma. N Engl J Med 2003; 348: 2609–17. Davis RE, Brown KD, Siebenlist U, Staudt LM. Constitutive nuclear factor kappaB activity is required for survival of activated B cell-like diffuse large B cell lymphoma cells. J Exp Med 2001; 194: 1861–74. Dunleavy K, Pittaluga S, Czuczman MS et al. Differential efficacy of bortezomib plus chemotherapy within molecular subtypes of diffuse large B-cell lymphoma. Blood 2009; 113: 6069–76.
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
TRIM21: a novel biomarker in DLBCL
32 Sjostrand M, Ambrosi A, Brauner S et al. Expression of the immune regulator tripartite-motif 21 is controlled by IFN regulatory factors. J Immunol 2013; 191: 3753–63. 33 Aqrawi LA, Kvarnstrom M, Brokstad KA, Jonsson R, Skarstein K, Wahren-Herlenius M. Ductal epithelial expression of Ro52 correlates with inflammation in salivary glands of patients with primary Sjogren’s syndrome. Clin Exp Immunol 2014; 177: 244–52. 34 Bazarbachi A, Plumelle Y, Carlos Ramos J et al. Metaanalysis on the use of zidovudine and interferon-alfa in adult T-cell leukemia/lymphoma showing improved survival in the leukemic subtypes. J Clin Oncol, 2010; 28: 4177–83. 35 Smalley RV, Andersen JW, Hawkins MJ et al. Interferon alfa combined with cytotoxic chemotherapy for patients with nonHodgkin’s lymphoma. N Engl J Med 1992; 327: 1336–41. 36 Solal-Celigny P, Lepage E, Brousse N et al. Recombinant interferon alfa-2b combined with a regimen containing doxorubicin in patients with advanced follicular lymphoma. Groupe d’Etude des Lymphomes de l’Adulte. N Engl J Med 1993; 329: 1608–14. Correspondence: Marie Wahren-Herlenius, Unit of Experimental Rheumatology, CMM L8:04, Karolinska Institutet, 171 76 Stockholm, Sweden. (fax: +46-8-51775562; e-mail: [email protected]).
Reduced expression of TRIM21/Ro52 predicts poor prognosis in diffuse large B-cell lymphoma patients with and without rheumatic disease S. Brauner1, W. Zhou1, C. Backlin2, T. M. Green3,4, L. Folkersen1, M. Ivanchenko1, B. L€ofstr€om5, Z. Y. Xu-Monette6, K. H. Young6, L. Møller Pedersen4, M. Boe Møller3,4, C. Sundstr€om7, G. Enblad8, E. Baecklund2 & M. Wahren-Herlenius1* From the 1Unit of Experimental Rheumatology, Department of Medicine, Karolinska University Hospital, Karolinska Institutet, Stockholm, Sweden; 2Unit of Rheumatology, Department of Medical Sciences, Uppsala University, Uppsala, Sweden; 3Institute of Clinical Research, University of Southern Denmark; 4Odense University Hospital, Odense, Denmark; 5Rheumatology Clinic, Malar Hospital, Eskilstuna, Sweden; 6Department of Hematopathology, University of Texas, MD Anderson Cancer Center, Houston, TX, USA; 7Department of Immunology, Genetics and Pathology, Rudbeck Laboratory, Uppsala University; and ; 8Unit of Oncology, Department of Radiology, Oncology and Radiation Sciences, Uppsala University, Uppsala, Sweden *Current address: Center for Biological Sequence Analysis, Technical University of Denmark, Copenhagen, Denmark.
Abstract. Brauner S, Zhou W, Backlin C, Green TM, Folkersen L, Ivanchenko M, L€ ofstr€ om B, XuMonette ZY, Young KH, Møller Pedersen L, Boe Møller M, Sundstr€ om C, Enblad G, Baecklund E, Wahren-Herlenius M (Karolinska University Hospital, Karolinska Institutet, Stockholm; Uppsala University, Uppsala, Sweden; Institute of Clinical Research, University of Southern Denmark, Odense; Odense University Hospital, Odense, Denmark; Malar Hospital, Eskilstuna, Sweden; University of Texas, MD Anderson Cancer Center, Houston, TX, USA; Uppsala University, Uppsala; Uppsala University, Uppsala, Sweden). Reduced expression of TRIM21/Ro52 predicts poor prognosis in diffuse large B-cell lymphoma patients with and without rheumatic disease. J Intern Med 2015; 278: 323–332. Objective. TRIM21 (also known as Ro52) is an autoantigen in rheumatic disease and is predominantly expressed in leucocytes. Overexpression is associated with decreased proliferation, and the TRIM21 gene maps to a tumour suppressor locus. We therefore investigated the expression of TRIM21 in patients with diffuse large B-cell lymphoma (DLBCL) and its potential usefulness as a prognostic biomarker. Materials and methods. TRIM21 expression levels were assessed by immunohistochemistry in lymphoma biopsies from three cohorts of patients with DLBCL: 42 patients with rheumatic disease treated with a cyclophosphamide, vincristine, doxorubicin and
prednisone (CHOP)-like regimen, 76 CHOP-treated and 196 rituximab-CHOP-treated nonrheumatic patients. Expression was correlated with clinical and biomedical parameters. TRIM21 expression was assessed in relation to lymphocyte proliferation by quantitative PCR and correlated with 3H-thymidine incorporation and propidium iodine staining. Results. TRIM21 expression levels differed in the lymphomas compared to normal lymphoid tissue, with reduced expression correlating with shorter overall survival in all three cohorts. In the two larger cohorts, progression-free survival was assessed and was also found to correlate with TRIM21 expression. The association was independent of commonly used clinical prognostic scores, lymphoma subtype and several previously reported prognostic biomarkers. In agreement with this clinical observation, we noted an inverse correlation between TRIM21 expression and proliferation of leucocytes in vitro. Conclusions. We show that loss of TRIM21 expression is associated with more aggressive lymphoma and increased proliferation, whereas maintenance of TRIM21 expression is associated with better prognosis in patients with DLBCL. Based on our findings, we suggest that TRIM21 should be considered as a novel biomarker for lymphoma characterization and for predicting patient survival. Keywords: biomarker, diffuse large B-cell lymphoma, prognosis, Ro52, TRIM21.
ª 2015 The Association for the Publication of the Journal of Internal Medicine
323
S. Brauner et al.
Introduction Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoma type, accounting for 30–40% of adult lymphomas. The prognosis of patients with DLBCL is highly variable, and despite recent therapeutic advances with the addition of rituximab (anti-CD20) to the anthracycline-based treatment, approximately 40% of patients relapse [1]. Gene expression profiling studies have led to the division of DLBCL into two major subtypes based on the putative cell of origin: germinal centre B-cell-like (GCB) and activated B-cell-like (ABC) subtypes [2, 3]. Survival is better in the GCB compared to the ABC group, but division into subtype has not proven specific enough to identify patients with particularly poor prognosis. In addition, several biomarkers have emerged to better classify and predict outcome at diagnosis, but are not yet routinely used in clinical practice [4]. There is thus an urgent need for more precise and clinically reliable prognostic markers. TRIM21 (Ro52) was first described as an autoantigen in rheumatic diseases. Interestingly, most patients with rheumatic disease have an elevated risk of developing lymphomas, ranging from twofold to 16-fold increased risk compared to the general population, depending on the underlying disease [5, 6]. Our attention was drawn to the possible role of TRIM21 in lymphoma and its use as a biomarker for several reasons. First, the TRIM21 gene is located in a tumour suppressor locus on the short arm of chromosome 11 (11p15.5), a locus that has been associated with malignancies of many organs [7]. Secondly, TRIM21 is mainly expressed in cells of haematopoietic origin at steady state [8, 9]. Thirdly, we have previously demonstrated that increased expression of TRIM21 leads to decreased proliferative capacity in a mouse B-cell lymphoma cell line [9]. Finally, TRIM21 is an E3 ligase involved in ubiquitination, and there are emerging data to suggest that it regulates multiple proteins important in the cell cycle and apoptosis, such as BCL2 and interferon regulatory factor (IRF)-8 as well as FADD and c-FLIP [8, 10–12]. Based on these observations, we investigated the role of TRIM21 in lymphoma and its potential as a prognostic biomarker in patients with DLBCL. Materials and methods Study population and patient characteristics Three cohorts of patients with DLBCL were included in the study. Cohort 1 consisted of patients with 324
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
TRIM21: a novel biomarker in DLBCL
rheumatic disease treated with an anthracyclinebased regimen. The other cohorts consisted of patients without rheumatic disease treated with either cyclophosphamide, vincristine, doxorubicin and prednisone (CHOP) (Cohort 2) or rituximabCHOP (R-CHOP) (Cohort 3). Patients with primary CNS-, post-transplant- or HIV-related lymphomas were excluded from all cohorts. All lymphomas were re-assessed according to the 2001 World Health Organization classification of tumours by at least two independent pathologists prior to inclusion [4, 13–16]. The lymphomas of patients in cohorts 1 and 2 were classified as GCB and ABC (non-GCB) by an immunostaining algorithm [3]; Cohort 3 was classified by a gene expression profiling and/or immunostaining algorithm [14, 16]. The study was approved by the local ethics committees. Cohort 1 was identified by linking the Swedish Inpatient Register and the Swedish Cancer Register from the period 1964–1995, as previously described [13, 17]. In total, 97 patients with rheumatoid arthritis (RA) or systemic lupus erythematosus (SLE) diagnosed with DLBCL were identified. Due to the long recruitment period, treatment regimens varied substantially, and therefore, only the 42 patients who received anthracycline-based treatment were included in this study (33 with RA and nine with SLE). Patients were followed until July 2010. Mean age at lymphoma diagnosis was 62 years (range 33– 88 years). Mean follow-up was 31 months (range 1–267 months), and 78% of patients were classified as having Ann Arbor stage III or IV, indicating an aggressive lymphoma subtype. Cohort 2 comprised 76 patients with DLBCL treated with anthracycline-based CHOP, diagnosed at Uppsala University Hospital between 1984 and 2002 [4]. Patients were followed until June 2012. The mean age at lymphoma diagnosis was 66 years and mean follow-up time was 64 months (range 4 days to 255 months). In this cohort, 58% of patients were classified with Ann Arbor stage III or IV and 11 (16%) had an International Prognostic Index (IPI; a novel and more accurate predictive score of lymphoma severity) of 3–4. Four patients (10%) in Cohort 1 and 12 (16%) in Cohort 2 were alive at the end of follow-up. All surviving patients were diagnosed with lymphomas expressing high levels of TRIM21. Cohort 3 included 196 R-CHOP-treated patients with DLBCL, as previously described [14, 16].
N/A
N/A, not available; CHOP, cyclophosphamide, vincristine, doxorubicin and prednisone; R-CHOP, rituximab-CHOP; IPI, International Prognostic Index. GC, Germinal centre like B-cell subtype. a According to the Hans algorithm [3].
0.45
0.21 54.8
32.5 40
63.2 55.4
34 0.091
1 46
10.4 33 11
40 33.3
N/A N/A
24.4 GC positivitya
(%)
IPI >2 (%)
17.4
0.29
44.1
0.48 46.5 52.5 47.4 0.726 56.7 66.7 57.9 0.43 86.4 72.2 78 Ann Arbor stage III/IV
(years)
0.92
0.65 62.7 63.9 62.9 0.521 66.7 71.3 66.2 0.048 72.6 69.9 Mean age at diagnosis
66.6
41.7
(n = 156) (n = 40)
42.5 42.1
(n = 196) value
0.499 47.1
(n = 67) (n = 9)
33 45.5
(n = 76) value
0.07 58.3
(n = 24) (n = 18) (n = 42)
66.7
>40% ≤40% >40% ≤40%
Cohort 2: CHOP-treated patients
All P>40% ≤40% All
Cohort 1: patients with rheumatic disease
Table 1 Clinical characteristics of patient cohorts
Sections were analysed using a Polyvar II light microscope (Reichert-Jung, Vienna, Austria) at 910 magnification. Most biopsies displayed homogenous lymphoma tissue and were therefore scored as one entity. Biopsies containing non-lymphoma tissue were only scored in lymphoma-rich areas. A semi-quantitative scoring system was developed using a scale ranging from 0 (no expression) to 3 (ubiquitous expression). Scoring was performed in a blinded manner by two independent investigators (SB and MWH). In addition, computer-assisted image analysis of whole core biopsies was performed to define the percentage of stained lymphoma tissue area using the Quantimet 600 image analyser (Leica, Wetzlar, Germany). The average values were used for statistical analysis. If two biopsies of the same tissue differed by more than 10 percentage points in a stained area, they were reanalysed. Staining and scoring of the included biopsies for BCL2, BCL6 and IRF-4 (MUM1) expression have been described previously [4, 14, 20].
P-
Tissue microarray blocks were constructed as previously described [14, 17]. Briefly, two representative core biopsies from each obtained lymphoma biopsy were arranged into microarray paraffin blocks and sectioned into 4-lm sections. Normal lymphoid tonsillar tissue was used as staining control [17, 18]. Deparaffinized sections were subjected to antigen retrieval and subsequently stained with mouse anti-human TRIM21 antibodies, clones 7.8C7 and 7.12E11 at 2 lg mL1, using an automated Ventana XT module (Ventana Medical Systems, Tucson, AZ, USA) [19]. The antibodies target two different epitopes of the protein: clone 7.8C7 recognizes the centrally located coiled-coil domain and 7.12E11 the Cterminal RING domain of the TRIM21 protein. Binding was visualized with the iVIEW DAB detection kit (Ventana Medical Systems).
61.9
Immunohistochemistry
All
Cohort 3: R-CHOP-treated patients
P-
Patients were diagnosed between June 2001 and October 2009 and followed until July 2011. The mean age at lymphoma diagnosis was 63 years (range 16 to 91 years). In this cohort, 67 patients (34%) were classified as having an IPI >2 and 94 (48%) with an Ann Arbor stage III or IV. Follow-up times ranged from 14 days to almost 99 months. The majority (65%) of included patients were alive at the end of the study. Characteristics of the patients in all three cohorts are summarized in Table 1.
value
TRIM21: a novel biomarker in DLBCL
Female (%)
S. Brauner et al.
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
325
S. Brauner et al.
Cell culture, in vitro proliferation assay and real-time PCR Peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using Ficoll-Paque plus (GE Healthcare, Little Chalfont, UK). Freshly isolated cells were seeded in 96-well plates (1.5 9 105 cells per well) and stimulated with pokeweed mitogen at 1 and 5 lg mL1 for 5 days. Proliferation was evaluated by 3H-thymidine incorporation, as previously described [9]. Cell cycle phase was assessed by flow cytometry using propidium iodide. Briefly, cells were stained with antiCD19 (clone 4G7, BD Bioscience, Franklin Lakes, NJ, USA), fixed on ice in 80% ethanol/20% in phosphate-buffered saline (PBS) for 1 h and then treated with propidium iodide and RNAse for 5 min, prior to analysis (FACSCalibur, BD Biosciences). Total mRNA was extracted as previously described [9], using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA) and was reverse transcribed into cDNA using random hexamers (Life Technologies, Carlsbad, CA, USA). TRIM21 expression was analysed with quantitative SYBR Green PCR (Qiagen) using GAPDH as the internal control with primers: TRIM21-F, 50 -GAACTGCTGCAGGAGGTGATAA-30 and TRIM21-R, 50 -AGTTCTGGAGAGGTAATATCCA GGTC-30 ; GAPDH-F, 50 -AGGGCTGCTTTTAACTCT GGTAAA-30 and GAPDH-R, 50 -CATATTGGAACATG TAAACCATGTAGTTG-30 . Statistical analysis Overall survival was defined as the time from diagnosis to death or last day of follow-up; progression-free survival was defined as the time from diagnosis to progression, death or last day of followup. Prism graph pad 5 (La Jolla, CA, USA) was used to compare two groups with the Mann–Whitney Utest, to perform logistic regression, to analyse correlations between variables and to create Kaplan–Meier plots for survival analysis. Cox proportional hazards regression models were calculated using R-3.1.0 software (R foundation for Statistical Computing, Vienna, Austria). Because of non-normality of data distribution, predefined cut-off values were used as indicated, and the resulting binary variables were used in the primary analysis. The optimal cut-off value for low and high expression of TRIM21 in lymphomas was determined by receiver operator characteristic (ROC) curve analysis, and values for all other variables were based on clinical standards. Hazard ratio (HR) was reported as either exp(coef) or exp(-coef) to allow comparable 326
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
TRIM21: a novel biomarker in DLBCL
HR in presentation, independent of how expression changes associates with risk. In all analyses, Pvalues 40% of the cells expressing TRIM21 was considered high expression), based on ROC curve analysis in Cohort 1, yielding an assay sensitivity of 100% (data not shown). This cut-off value was used for further analysis in all cohorts. Expression of TRIM21 is a prognostic indicator of survival in DLBCL patients with rheumatic disease As TRIM21 is an autoantigen in rheumatic disease, we first investigated its potential prognostic value in DLBCL patients with RA or SLE (Cohort 1). To our surprise, we observed a highly significant correlation between high TRIM21 expression and longer overall survival. Median overall survival was 6 months in patients with low TRIM21 expression compared to 11 months in those with high TRIM21 expression (Fig. 1c, P = 0.0073). Cox regression analysis indicated an HR of 1.41 [95% confidence
S. Brauner et al.
TRIM21: a novel biomarker in DLBCL
(b)
(a)
0
1
2
3
Rheumatic anthracycline treated (Overall survival)
(c) 1.0
TRIM21 high (n = 24) 0.8
TRIM21 low (n = 18) P = 0.0073
0.6 0.4 0.2 0.0 0
50
100 150 200 Time (Months)
(e) 60
Rheumatic anthracycline treated (Overall survival)
(d) % Ro52/TRIM21 expression
Fig. 1 Expression of the autoantigen TRIM21 is prognostic for survival in patients with rheumatic disease and lymphoma, independent of autoantibody status. All lymphoma biopsies were stained with two different anti-TRIM21 antibodies (7.8C7 and 7.12E11) binding different epitopes of the protein; antibody 7.8C7 staining was used in the analysis throughout the study. (a) Example of lymphomas scored manually (score 0–3). (b) Both manual and computerized scores were used, and results were compared with linear regression (r2 = 0.67, P < 0.0001). (c) Kaplan–Meier curves of overall survival in 42 lymphoma patients with rheumatic disease treated with anthracycline-based drugs show a correlation between high (>40%) TRIM21 expression and longer survival (P < 0.01). Antinuclear antibody (ANA) status was tested in 30 of the patients. (d) TRIM21 expression in lymphomas stratified by ANA status shows no association between autoantibodies and protein expression (P = 0.39). (e) Survival is unaffected by ANA status, as shown by Kaplan–Meier curves (P = 0.65).
n.s.
40
20
0
ANA pos
interval (CI) 0.97–2.05, P = 0.076]. As this was a retrospectively collected cohort, specific anti-SSA/ Ro52 autoantibody titres were not measured and serum samples were not available for analysis. However, antinuclear antibody status was recorded in 70% of patients. No correlation was observed between TRIM21 expression level and autoantibody positivity (Fig. 1d, P = 0.39), or overall survival (Fig. 1e, P = 0.65), indicating that the antinuclear autoantibodies do not affect the expression pattern of TRIM21. Predictive significance of TRIM21 protein expression in CHOP- and RCHOP-treated patients To investigate whether TRIM21 expression is a predictive factor for survival also in patients with-
ANA neg
250
300
1.0
ANA pos (n = 16) ANA neg (n = 14)
0.8 0.6
P = 0.65 0.4 0.2 0.0 0
50
100
150
200
250
300
Time (Months)
out rheumatic disease, we examined a cohort of 74 CHOP-treated patients with DLBCL (Cohort 2). Also here, we observed that low TRIM21 expression was associated with shorter overall survival (median survival 10 vs. 48 months; Fig. 2a, P < 0.0001), with an HR of 1.56 (95% CI 1.10–1.99, P < 0.002). In addition, the time from remission to disease progression, the median progression-free survival, was 1 month in the group with lymphomas with low TRIM21 expression compared to 60 months in the high expression group (Fig. 2b, P < 0.0001). Today, most patients with DLBCL are treated with a CHOP-like regimen in combination with rituximab. To verify the role of TRIM21 as a prognostic biomarker in DLBCL, we investigated an additional cohort of 196 R-CHOP-treated patients (Cohort 3). ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
327
S. Brauner et al.
TRIM21: a novel biomarker in DLBCL
1.0
TRIM21 high (n = 67)
0.8
TRIM21 low (n = 9)
0.6
P < 0.0001
0.4 0.2 0.0 0
50
100
150
200
250
(b)
CHOP treated (Progression-free survival)
CHOP treated (Overall survival)
(a)
TRIM21 high (n = 67)
1.0
TRIM21 low (n = 9)
0.8 0.6
P < 0.0001
0.4 0.2 0.0 0
300
50
TRIM21 high (n = 154)
1.0
TRIM21 low (n = 42)
0.8 0.6 0.4 P < 0.0001
0.2 0.0 0
20
40
60
80
(d)
R-CHOP treated (Progression-free survival)
R-CHOP treated (Overall survival)
(c)
1.0
300
TRIM21 low (n = 42)
0.4 P = 0.0005
0.2 0.0 0
0.8 0.6 0.4 P = 0.0006
(f)
20
40
60
80
100
TRIM21 high (n = 70)
1.0
ABC R-CHOP treated (Overall survival)
GCB R-CHOP treated (Overall survival)
250
Time (Months)
TRIM21 low (n = 25)
0.2
200
0.6
100
TRIM21 high (n = 84)
1.0
150
TRIM21 high (n = 154)
0.8
Time (Months)
(e)
100
Time (Months)
Time (Months)
TRIM21 low (n = 17)
0.8 0.6 0.4 P = 0.0018
0.2 0.0
0.0 0
20
40
60
80
100
Time (Months)
0
20
40
60
80
100
Time (Months)
Fig. 2 Expression of TRIM21 correlates with overall survival (OS) and progression-free survival (PFS) in patients without rheumatic disease, independent of treatment and disease subtype. A cohort of 76 cyclophosphamide, vincristine, doxorubicin and prednisone (CHOP)-treated diffuse large B-cell lymphoma (DLBCL) patients without rheumatic disease was investigated. Kaplan–Meier curves are shown for OS (a) (P < 0.0001) and PFS (b) (P < 0.0001) comparing high (>40%) and low ( 30%, IRF-4 > 30%, BCL6 > 30% staining, IPI >2 and Ann Arbor stage >2. CI, confidence interval; IPI, international prognostic index. ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
329
(a)
7
(b)
7
6
TRIM21 expression (relative to GAPDH)
TRIM21: a novel biomarker in DLBCL
TRIM21 expression (relative to GAPDH)
S. Brauner et al.
6
5
P < 0.0001 r2 = 0.5032
4 3 2 1 0
5
P = 0.0003 r2 = 0.4022
4 3 2 1 0
0
25 000
50 000
75 000
100 000
cpm
0
5
10
15
20
25
30
(S+G2)/M
Fig. 3 TRIM21 mRNA expression is downregulated in proliferating cells. Freshly isolated peripheral blood mononuclear cells from healthy donors were stimulated with pokeweed mitogen (1 and 5 lg mL1) for 5 days. TRIM21 mRNA expression levels quantified by real-time PCR were by correlated by linear regression to 3H-thymidine incorporation (a) (r2 = 0.50, P < 0.0001) and cell cycle phase measured by flow cytometry (b) (r2 = 0.40, P < 0.001).
against TRIM21 as well as an increased risk of lymphoma development. We proceeded to analyse the prognostic relevance also in DLBCL patients without rheumatic disease and, interestingly, found that it was equally significant in both groups, indicating that TRIM21 expression is an independent marker of lymphoma outcome. Recently published data have demonstrated that TRIM21 is an E3 ligase involved in ubiquitination, and several IRFs have been identified as targets of TRIM21-mediated ubiquitination [8, 21, 22]. Notably, IRF-4 (MUM1) has been proposed to be involved in lymphoma development and is used as a biomarker in DLBCL [3]. Further, Jauharoh and colleagues have shown that BCL2 protein levels are regulated by TRIM21-mediated ubiquitination [10]. In addition, the apoptosis regulatory protein BCL2 has been associated with lymphoma development [23]. GCB lymphomas are associated with a specific translocation, t(14:18), where the BCL2 gene is translocated to the IgH promoter, leading to protein overexpression and apoptotic resistance [24]. TRIM21 also acts on the cell cycle regulator p27 [25]. Of note, it ha been proposed that TRIM21 monoubiquitinates the upstream regulator of inhibitor kappa B (I-kB), the inhibitory kinase of kappa-beta (IKK-beta), and thereby alters its function [26]. Inhibition of the degradation of IkB has been suggested as the main mode of action of the 26S proteasome inhibitor bortezomib [27], which has shown efficacy in the treatment of both myeloma and lymphoma [28, 29]. Increased levels of I-kB lead to inhibition of the NF-kB pathway, 330
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
subsequently decreasing cellular resistance to apoptosis. This pathway is constitutively active in ABC lymphomas [30], and bortezomib has shown promise as second-line treatment in patients with this subgroup [31]. Together, these findings suggest that TRIM21 controls several pathways involved in cell proliferation and apoptosis, and the role of TRIM21 as a central regulator of cellular homoeostasis is further emphasized by the observation that TRIM21-null mice develop lymphoproliferation with lymphadenopathy and splenomegaly [8]. Hence, there is a biological and cellular basis for the observed low expression of TRIM21 in patients with the poorest survival. It is possible that restoring TRIM21 expression could be clinically beneficial. Both type 1 and type 2 interferons (IFNs) upregulate TRIM21 expression, which can be detected as induced under proinflammatory conditions in several tissues including lymphoid tissues [8, 18, 32, 33]. Interestingly, treatment with type 1 IFNs (IFN-alpha) is a secondline consideration for therapy in specific lymphoma subsets [34–36]. It is possible that patients with no or a low level of TRIM21 expression in their tumours might benefit the most from IFN-alpha treatment, based on the theory that this cytokine would re-establish TRIM21 expression and the control of proliferation and apoptosis. A reliable biomarker for selecting patients for IFN-alpha treatment is important not only to identify patients who would benefit most from the therapy but also to avoid unnecessary treatment of others, especially because of the adverse effects of the drug.
S. Brauner et al.
In summary, our data show for the first time that the level of expression of TRIM21 in lymphoma tissue is a strong prognostic marker in patients with DLBCL. The role of TRIM21 as a biomarker for lymphoma outcome may have a biological basis as TRIM21 regulates cellular proliferation and apoptosis. Further, here we show that the prognostic value of this marker is independent of Ann Arbor and IPI classification as well as subtype, suggesting that TRIM21 expression adds prognostic value to current tools in patients with DLBCL.
TRIM21: a novel biomarker in DLBCL
9
10
11
12
Acknowledgements
13
This study was supported by grants from the Swedish Research Council, the Heart-Lung Foundation, Stockholm County Council, Karolinska Institutet, the Swedish Rheumatism Association, the King Gustaf Vth 80-year Foundation and the Torsten and Ragnar S€oderberg Foundation.
14
15
Conflict of interest statement The authors have no competing interests to declare.
References 1 Fu K, Weisenburger DD, Choi WW et al. Addition of rituximab to standard chemotherapy improves the survival of both the germinal center B-cell-like and non-germinal center B-celllike subtypes of diffuse large B-cell lymphoma. J Clin Oncol 2008; 26: 4587–94. 2 Alizadeh AA, Eisen MB, Davis RE et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 2000; 403: 503–11. 3 Hans CP, Weisenburger DD, Greiner TC et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood 2004; 103: 275–82. 4 Berglund M, Thunberg U, Amini RM et al. Evaluation of immunophenotype in diffuse large B-cell lymphoma and its impact on prognosis. Mod Pathol 2005; 18: 1113–20. 5 Dias C, Isenberg DA. Susceptibility of patients with rheumatic diseases to B-cell non-Hodgkin lymphoma. Nat Rev Rheumatol 2011; 7: 360–8. 6 Theander E, Henriksson G, Ljungberg O, Mandl T, Manthorpe R, Jacobsson LT. Lymphoma and other malignancies in primary Sjogren’s syndrome: A cohort study on cancer incidence and lymphoma predictors. Ann Rheum Dis 2006; 65: 796–803. 7 Feinberg AP. Imprinting of a genomic domain of 11p15 and loss of imprinting in cancer: an introduction. Cancer Res 1999; 59: 1743s–6s. 8 Espinosa A, Dardalhon V, Brauner S et al. Loss of the lupus autoantigen Ro52/Trim21 induces tissue inflammation and
16
17
18
19
20
21
22
23
24
systemic autoimmunity by disregulating the IL-23-Th17 pathway. J Exp Med 2009; 206: 1661–71. Espinosa A, Zhou W, Ek M et al. The Sjogren’s syndromeassociated autoantigen Ro52 is an E3 ligase that regulates proliferation and cell death. J Immunol 2006; 176: 6277–85. Jauharoh SN, Saegusa J, Sugimoto T et al. SS-A/Ro52 promotes apoptosis by regulating Bcl-2 production. Biochem Biophys Res Commun 2012; 417: 582–7. Young JA, Sermwittayawong D, Kim HJ et al. Fas-associated death domain (FADD) and the E3 ubiquitin-protein ligase TRIM21 interact to negatively regulate virus-induced interferon production. J Biol Chem 2011; 286: 6521–31. Zhang J, Fang L, Zhu X et al. Ro52/SSA sensitizes cells to death receptor-induced apoptosis by down-regulating c-FLIP (L). Cell Biol Int 2012; 36: 463–8. Baecklund E, Sundstrom C, Ekbom A et al. Lymphoma subtypes in patients with rheumatoid arthritis: increased proportion of diffuse large B cell lymphoma. Arthritis Rheum 2003; 48: 1543–50. Green TM, Young KH, Visco C et al. Immunohistochemical double-hit score is a strong predictor of outcome in patients with diffuse large B-cell lymphoma treated with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone. J Clin Oncol 2012; 30: 3460–7. Organization WH. World Health Organization classification of tumours: pathology and genetics of tumours of haematopoietic and lymphoid tissue. IARC Press, 2001. Visco C, Li Y, Xu-Monette ZY et al. Comprehensive gene expression profiling and immunohistochemical studies support application of immunophenotypic algorithm for molecular subtype classification in diffuse large B-cell lymphoma: a report from the International DLBCL Rituximab-CHOP Consortium Program Study. Leukemia 2012; 26: 2103–13. Lofstrom B, Backlin C, Pettersson T, Lundberg IE, Baecklund E. Expression of APRIL in diffuse large B cell lymphomas from patients with systemic lupus erythematosus and rheumatoid arthritis. J Rheumatol 2011; 38: 1891–7. Oke V, Vassilaki I, Espinosa A et al. High Ro52 expression in spontaneous and UV-induced cutaneous inflammation. J Invest Dermatol 2009; 129: 2000–10. Strandberg L, Ambrosi A, Espinosa A et al. Interferon-alpha induces up-regulation and nuclear translocation of the Ro52 autoantigen as detected by a panel of novel Ro52-specific monoclonal antibodies. J Clin Immunol 2008; 28: 220–31. Baecklund E, Backlin C, Mansouri M et al. LMO2 protein expression predicts survival in patients with rheumatoid arthritis and diffuse large B-cell lymphoma. Leuk Lymphoma 2011; 52: 1146–9. Higgs R, Ni Gabhann J, Ben Larbi N, Breen EP, Fitzgerald KA, Jefferies CA. The E3 ubiquitin ligase Ro52 negatively regulates IFN-beta production post-pathogen recognition by polyubiquitin-mediated degradation of IRF3. J Immunol, 2008; 181: 1780–6. Yoshimi R, Chang TH, Wang H, Atsumi T, Morse HC 3rd, Ozato K. Gene disruption study reveals a nonredundant role for TRIM21/Ro52 in NF-kappaB-dependent cytokine expression in fibroblasts. J Immunol, 2009; 182: 7527–38. Iqbal J, Neppalli VT, Wright G et al. BCL2 expression is a prognostic marker for the activated B-cell-like type of diffuse large B-cell lymphoma. J Clin Oncol 2006; 24: 961–8. Huang JZ, Sanger WG, Greiner TC et al. The t(14;18) defines a unique subset of diffuse large B-cell lymphoma with a
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
331
S. Brauner et al.
25
26
27
28
29
30
31
332
germinal center B-cell gene expression profile. Blood 2002; 99: 2285–90. Sabile A, Meyer AM, Wirbelauer C et al. Regulation of p27 degradation and S-phase progression by Ro52 RING finger protein. Mol Cell Biol 2006; 26: 5994–6004. Wada K, Niida M, Tanaka M, Kamitani T. Ro52-mediated monoubiquitination of IKK{beta} down-regulates NF-{kappa} B signalling. J Biochem 2009; 146: 821–32. Hideshima T, Anderson KC. Molecular mechanisms of novel therapeutic approaches for multiple myeloma. Nat Rev Cancer 2002; 2: 927–37. Goy A, Younes A, McLaughlin P et al. Phase II study of proteasome inhibitor bortezomib in relapsed or refractory Bcell non-Hodgkin’s lymphoma. J Clin Oncol 2005; 23: 667–75. Richardson PG, Barlogie B, Berenson J et al. A phase 2 study of bortezomib in relapsed, refractory myeloma. N Engl J Med 2003; 348: 2609–17. Davis RE, Brown KD, Siebenlist U, Staudt LM. Constitutive nuclear factor kappaB activity is required for survival of activated B cell-like diffuse large B cell lymphoma cells. J Exp Med 2001; 194: 1861–74. Dunleavy K, Pittaluga S, Czuczman MS et al. Differential efficacy of bortezomib plus chemotherapy within molecular subtypes of diffuse large B-cell lymphoma. Blood 2009; 113: 6069–76.
ª 2015 The Association for the Publication of the Journal of Internal Medicine Journal of Internal Medicine, 2015, 278; 323–332
TRIM21: a novel biomarker in DLBCL
32 Sjostrand M, Ambrosi A, Brauner S et al. Expression of the immune regulator tripartite-motif 21 is controlled by IFN regulatory factors. J Immunol 2013; 191: 3753–63. 33 Aqrawi LA, Kvarnstrom M, Brokstad KA, Jonsson R, Skarstein K, Wahren-Herlenius M. Ductal epithelial expression of Ro52 correlates with inflammation in salivary glands of patients with primary Sjogren’s syndrome. Clin Exp Immunol 2014; 177: 244–52. 34 Bazarbachi A, Plumelle Y, Carlos Ramos J et al. Metaanalysis on the use of zidovudine and interferon-alfa in adult T-cell leukemia/lymphoma showing improved survival in the leukemic subtypes. J Clin Oncol, 2010; 28: 4177–83. 35 Smalley RV, Andersen JW, Hawkins MJ et al. Interferon alfa combined with cytotoxic chemotherapy for patients with nonHodgkin’s lymphoma. N Engl J Med 1992; 327: 1336–41. 36 Solal-Celigny P, Lepage E, Brousse N et al. Recombinant interferon alfa-2b combined with a regimen containing doxorubicin in patients with advanced follicular lymphoma. Groupe d’Etude des Lymphomes de l’Adulte. N Engl J Med 1993; 329: 1608–14. Correspondence: Marie Wahren-Herlenius, Unit of Experimental Rheumatology, CMM L8:04, Karolinska Institutet, 171 76 Stockholm, Sweden. (fax: +46-8-51775562; e-mail: [email protected]).
