Main

Primary central nervous system diffuse large B-cell lymphoma is a rare form of diffuse large B-cell lymphoma with an intracerebral or intraocular location that accounts for 1% of all non-Hodgkin lymphomas.1 The tumor cells of central nervous system diffuse large B-cell lymphoma are relatively homogeneous and are characterized by a centroblastic appearance, a late germinal center B-cell origin (as assessed using gene expression profiling), and a non-germinal center B-cell phenotype (as assessed using immunohistochemistry based on the Hans algorithm).2, 3, 4 The prognosis of central nervous system diffuse large B-cell lymphoma is generally unfavorable, because of the unique tumor cell biology, location itself, or central nervous system-specific microenvironment. However, prognostic markers for this condition have not been identified.

In recent years, the role of the MYC gene in the prognosis of non-central nervous system diffuse large B-cell lymphoma has been elucidated. MYC is a transcription factor that activates many genes involved in diverse cellular processes, including the regulation of cell size, survival, proliferation, metabolism, and angiogenesis. Deregulated MYC expression is involved in the pathogenesis of various solid tumor and lymphoid malignancies.5 MYC rearrangements have been identified in 5–15% of non-central nervous system diffuse large B-cell lymphomas and are associated with a poor prognosis, especially when combined with BCL2 and/or BCL6 rearrangement ('double-hit' and 'triple-hit' lymphomas).6, 7, 8, 9 In addition, high MYC and/or BCL2 protein expression results in an adverse prognosis, regardless of the presence of MYC rearrangements.10, 11, 12, 13, 14

MYC rearrangement in central nervous system diffuse large B-cell lymphoma has been reported recently, at a lower incidence (0–9%) than that in non-central nervous system diffuse large B-cell lymphoma. There are limited data on the prognostic significance of MYC rearrangement in central nervous system diffuse large B-cell lymphoma.15, 16, 17, 18 In contrast, high MYC protein expression (43–92%) in central nervous system diffuse large B-cell lymphoma has been reported in recent studies with an inconsistent association with overall survival.17, 18, 19 The mechanisms underlying the deregulation of MYC expression in central nervous system diffuse large B-cell lymphoma have been uncovered. Although aberrant somatic hypermutation involving the MYC gene has been detected in central nervous system diffuse large B-cell lymphoma, its clinical significance is unknown at present.20 Recent gene expression profiling analyses to identify genes that are differentially expressed between central nervous system diffuse large B-cell lymphoma and non-central nervous system diffuse large B-cell lymphoma did not identify the MYC gene.2, 21

In the current study, we investigated the association between MYC protein expression, MYC gene rearrangement, and MYC mRNA expression, as well as their prognostic value for central nervous system diffuse large B-cell lymphoma.

Materials and methods

Sample Collection and Processing

Seventy-four pretreatment tumor biopsies of patients diagnosed with de novo central nervous system diffuse large B-cell lymphoma according to the WHO classification (2008) criteria were collected.1 Patients were diagnosed between January 1995 and December 2012 at the Department of Pathology of the Samsung Medical Center (Seoul, Korea). Clinical information was collected from the medical records and included age, performance status according to the Eastern Cooperative Oncology Group (ECOG), serum lactate dehydrogenase levels at the time of diagnosis, sex, type of treatment, and date of the last follow-up or death. The study was approved by the research ethics boards of our institutions, according to the Declaration of Helsinki.

Tissue Microarrays and Immunohistochemistry

Formalin-fixed, paraffin-embedded specimens were used to generate a tissue microarray. Representative 1 mm cores of each case, in duplicate, were taken from tissue blocks. Immunohistochemical stains of 4-μm paraffin sections of the tissue microarray blocks were performed using a Bond Max automated immunostainer (Leica Biosystems, Melbourne, Australia). Monoclonal antibodies against CD20 (L26, 1/200; Dako, Glostrup, Denmark), CD3 (polyclonal, 1/200; Dako), CD10 (56C6, 1/250; Novocastra, Newcastle upon Tyne, UK), BCL6 (LN22, 1/80; Novocastra), MUM1 (MUM1p, 1/500; Dako), BCL2 (124, 1/100; Dako), and MYC (Y69, cat:ab32072, 1/100; Abcam, Burlingame, CA, USA) were used. The cell-of-origin subtype was determined using the Hans algorithm.4 MYC immunoreactivity was evaluated manually on an average of 200 cells per case. MYC and BCL2 positivity was analyzed using the X-tile statistical software (http://www.tissuearray.org/rimmlab) to determine the optimum cutoff point for dichotomizing the expression of the MYC protein (≥44%) and BCL2 protein (≥70%) based on patient survival.22 The reproducibility of MYC and BCL2 protein expression measurements by IHC was checked by comparing the results obtained by two pathologists (SMS and YHK).

MYC mRNA In situ Hybridization Assay

The MYC mRNA expression was examined in the 74 patients by RNA in situ hybridization using the RNA scope technology (Advanced Cell Diagnostics, Hayward, CA, USA). All RNA in situ hybridization slides were captured and were printed as hard copies. The RNA signals were evaluated by two independent observers (SMS and YHK). MYC mRNA molecules were detected with single-copy detection sensitivity. Single-molecule signals were quantified on a cell-by-cell basis by manual counting. The signals per cell were divided into five groups (0 dots/cell, 1–3 dots/cell, 4–9 dots/cell, 10–15 dots/cell, and>15 dots/cell with >10% of dots in clusters). Each sample was evaluated for the percentage of cells in each group. The H-score was calculated by adding up the percentage of cells in each group, with a weight assigned to each group, using the formula provided by the manufacturer (Supplementary Table 1). H-scores are given on a scale of 0–400.

Fluorescence In situ Hybridization (FISH) Assay

MYC gene rearrangements were evaluated by FISH using a Vysis LSI MYC dual color, break-apart probe (Abbott Molecular, Abbott Park, IL, USA).23 BCL2 gene rearrangements were investigated using a Vysis LSI BCL2 dual color, break-apart probe (Abbott Molecular). Cases with break-apart signals in >3% of nuclei were considered positive for the presence of a translocation.

Statistical Anlaysis

The baseline characteristics are described as the mean and s.d., or median and range, for quantitative variables; and as frequency and percentage for categorical variables. The analysis of the correlation between clinicopathological features and the results of MYC RNA in situ hybridization and MYC immunohistochemistry was performed using the χ2 and Fisher’s exact test. Student’s t-test and the Mann–Whitney U-test were used when applicable for comparisons among study groups. Pearson’s correlation analysis (r) was used to assess the linear relationship between the MYC immunopositive tumor cell percentage and MYC mRNA expression score based on the H-scoring system. Survival curves were plotted according to the Kaplan–Meier method and were compared using the log-rank test. The Cox proportional hazard model was used for univariate and multivariate survival analyses. A multivariate analysis was performed including the variables that were found to be predictive by univariate analysis (P<0.05). Overall survival was defined as the time between the date of the diagnostic biopsy and the date of death or of the last follow-up. Statistical analyses were carried out using the SPSS statistical package, version 23.0 (IBM, NY, USA). Statistical significance was set at P<0.05 (two sided).

Results

Patient Characteristics

The study cohort consisted of 74 immunocompetent patients with central nervous system diffuse large B-cell lymphoma, whose characteristics are summarized in Table 1. The overall sex composition was 40 (54%) males and 34 (46%) females. The median age of the patients at diagnosis was 58 years (range, 19–82 years). Fifty-seven out of 74 (77%) patients had an ECOG performance status score ≥2. Lactate dehydrogenase levels at diagnosis were elevated in 20 out of the 68 (29%) patients with available data. An International Prognostic Index (IPI) of more than high-intermediate risk was observed in 11 out of 68 (16%) patients. Seventy-two of the 74 (97%) patients had available treatment data. Sixty-nine out of 72 (96%) patients received methotrexate-based chemotherapy: high-dose methotrexate alone (n=5), high-dose methotrexate with cytarabine (n=3), high-dose methotrexate with rituximab and cytarabine (n=1), high-dose methotrexate with vincristine (n=6), high-dose methotrexate with vincristine and cytarabine (n=8), high-dose methotrexate with procarbazine and vincristine (n=4), high-dose methotrexate with procarbazine and vincristine with cytarabine (n=41), and high-dose methotrexate with procarbazine and vincristine with rituximab (n=1). Whole-brain radiation therapy (WBRT) was added to chemotherapy in 56 patients. One patient was treated with WBRT only and two patients received palliative care only.

Table 1 Relationship between clinicopathological features of patients and MYC protein expression
Full size table

Of the treated patients with follow-up information, the median follow-up was 35.2 months (range, 1.8–148.8 months) and the median overall survival time was 61.8 months (95% confidence interval, 40.8–94.2). Lymphoma was the cause of death in all the patients who were followed until death. Nine patients (12%) were lost to the follow-up and 45 patients (61%) had died at the time of data collection.

Morphology and Immunohistochemistry

All 74 cases were positive for CD20 and negative for CD3. Fifty-nine out of 74 cases (80%) were negative for CD10 and positive for both BCL6 and MUM1; these cases were of the non-germinal center type according to the Hans algorithm (Figure 1a). The other 15 cases were of the germinal center type by the Hans algorithm; nine were negative for both CD10 and MUM1 and positive for BCL6, four were positive for all three markers, and two were negative for MUM1 and positive for both CD10 and BCL6. BCL2 was expressed in 54 out of 74 cases (73%).

Figure 1
figure 1

Representative example of primary central nervous system diffuse large B-cell lymphoma. (a) Hematoxylin and eosin staining showing large neoplastic cells with prominent nucleoli (original magnification, × 400). (b) MYC staining showing a high percentage of cells with positive protein expression (original magnification, × 400). (c) RNA in situ hybridization analysis showing positive MYC mRNA expression (original magnification, 400 × ). (d) MYC rearrangement shown by FISH (one split signal, red and green, and one fused signal). FISH, fluorescence in situ hybridization.

Full size image

The mean percentage of MYC immunopositive cells was 49% (95% confidence interval, 44–54%) and the median value was 52% (range, 1–90%). MYC protein expression was exclusively nuclear in all cases (Figure 1b). Using the cutoff value of 44%, as assessed using the X-tile package, 49 out of 74 cases (66%) were considered MYC immunopositive cases. The relationships between MYC protein expression and clinicopathological findings are shown in Table 1. Patients with MYC protein overexpression were more frequently female and more frequently showed high-risk IPI score, non-germinal center phenotype, MUM1 protein expression, and MYC mRNA expression. The proportion of MYC immunopositive cases was higher for the non-germinal center than for the germinal center phenotype (75 vs 33%, respectively, P=0.003) (Pearson’s χ2-test). The mean percentage of MYC immunopositive cells was higher in the non-germinal center than in the germinal center phenotype (52 vs 36%, respectively, P=0.01) (Student’s t-test).

Determination of MYC mRNA Expression via RNA in situ Hybridization

To assess MYC mRNA expression levels, a quantitative RNA in situ hybridization analysis was performed by calculating the H-score according to the ;manufacturer’s guidelines (Supplementary Table 1). The mean MYC mRNA H-score was 45.3 (95% confidence interval, 34.8–55.8) and the median value was 34.2 (range, 1.0–207.5). Using the cutoff value of 63.5, as determined using the X-tile package, 16 out of 74 cases (22%) were classified as high MYC mRNA cases (Figure 1c). The relationships between MYC mRNA level and clinicopathological findings are shown in Table 1. The proportion of high MYC mRNA cases was greater in MYC immunopositive cases than it was in MYC immunonegative cases (29 vs 8%, respectively, P=0.042) (Pearson’s χ2-test). The level of MYC mRNA expression was moderately correlated with the level of MYC protein expression (P<0.001, r=0.544) (Pearson’s correlation analysis) (Figure 2a). The level of MYC protein expression in the high MYC mRNA group was higher than that in the low-MYC mRNA group (Figure 2b) (median, 70 vs 47%, respectively; P<0.001) (Mann–Whitney U-test).

Figure 2
figure 2

(a) XY scatter plots showing a correlation between the percentage of MYC immunopositive tumor cells and MYC mRNA expression, as determined by RNA scope. (b) MYC protein expression was higher in the group with high MYC mRNA expression compared with the group with low-MYC mRNA expression (P<0.001).

Full size image

FISH Analyses

FISH assays to determine MYC aberrations were performed in all cases, and informative FISH analyses were obtained in 68 out of 74 (92%) cases. Five (7%) cases showed MYC translocation (Figure 1d). No other genetic aberrations, including amplification or polysomy, were observed. In addition, BCL2 FISH was performed in the five MYC-FISH-positive cases, and no translocation involving BCL2 was found. All cases of MYC translocation belonged to the MYC immunopositive group. The mean percentage of MYC immunopositive cells in the MYC-FISH-positive group was higher than that in the MYC-FISH-negative group (78 vs 47%, respectively, P=0.001) (Student’s t-test) (Figure 3a). However, MYC mRNA expression levels were not significantly different between the MYC-FISH-positive and MYC-FISH-negative cases (median, 110.6 vs 34.6, respectively; P=0.084) (Mann–Whitney U-test) (Figure 3b). The five MYC-FISH-positive cases were divided into two low MYC RNA cases and three high MYC RNA cases (Table 1).

Figure 3
figure 3

(a) MYC protein expression was higher in the MYC-FISH-positive group than it was in the negative groups (P=0.001). (b) The difference in MYC mRNA expression between MYC-FISH-positive and -negative groups was not statistically significant (P=0.084). FISH, fluorescence in situ hybridization.

Full size image

Factors Associated with Clinical Outcome

Table 2 presents the results of univariate and multivariate survival analyses that included age, sex, ECOG status, cell-of-origin phenotype according to the Hans algorithm, BCL6 protein expression, BCL2 protein expression, MYC protein expression, MYC and BCL2 protein co-expression, and MYC mRNA expression status. A univariate analysis showed that age over 60 years, ECOG≥2, and MYC immunopositivity were significantly associated with an increased risk of death. However, there were no significant differences in survival according to MYC mRNA expression status (Table 2 and Figure 4) (log-rank test). The 3-year overall survival rate was not different between the high-MYC mRNA group and the low-MYC mRNA group (53 vs 66%, P=0.352). On multivariate analysis, MYC protein expression and ECOG retained prognostic significance (P=0.016 and P=0.004, respectively). The prognostic significance of MYC protein expression and MYC mRNA expression was evaluated according the cutoff points determined using the X-tile package; cutoffs at 44% and 63.5 points, respectively, exhibited the best correlation with survival. As informative MYC FISH analysis was possible in 68 out of 74 cases, the prognostic evaluation was applied to these 68 cases only; there were no significant differences in survival according to MYC translocation status (P=0.503) (log-rank test).

Table 2 Univariate and multivariate overall survival analyses
Full size table
Figure 4
figure 4

Overall survival of patients with primary central nervous system diffuse large B-cell lymphoma. Kaplan–Meier curves represent overall survival according to (a) high MYC protein expression and (b) high MYC mRNA expression. Log-rank tests for a and b yielded P=0.016 and P=0.055, respectively. Total evaluable patients for the analyses: (a) n=74 and (b) n=74.

Full size image

Discussion

Primary lymphoma of the central nervous system is a specific type of diffuse large B-cell lymphoma that is confined to the central nervous system. The central nervous system is an immunoprivileged organ without a classical lymphatic drainage system, which creates a specific microenvironment for developing lymphoid malignancy. The three major components of a central nervous system with primary central nervous system lymphoma are malignant B cells, resident brain cells, and reactive inflammatory bystander cells.3 The interaction between the tumor B cells, astrocytes, microglial cells, reactive T cells, and endothelial cells via the expression of chemokines, cytokines, cell-adhesion molecules, osteopontin, and major histocompatibility complex class II antigen may contribute to the characteristic perivascular tumor cell cuffing and diffuse brain infiltration. However, at present, the mechanisms underlying the selective tropism and confinement to the central nervous system are unclear.3, 24, 25, 26, 27

A variety of prognostic markers have been investigated in central nervous system diffuse large B-cell lymphoma. A perivascular tumor cell infiltration pattern was identified as a poor prognostic factor.28 The presence of reactive perivascular T cells was associated with superior overall survival.26, 28 Microvessel density appeared to be an inconsistent prognostic factor in central nervous system diffuse large B-cell lymphoma.29, 30, 31 BCL6 expression has been reported to have an inconsistent association with prognosis.17, 18, 32, 33, 34, 35 Chromosome 6q deletions and homozygous 9p deletions were associated with inferior progression-free survival and overall survival.15, 36

In this work, we evaluated MYC as a prognostic marker for central nervous system diffuse large B-cell lymphoma. MYC translocation has a lower prevalence (0–9%) in central nervous system diffuse large B-cell lymphoma compared with non-central nervous system diffuse large B-cell lymphoma.15, 16, 17, 18 MYC protein overexpression is more frequent (43–92%) than MYC translocation in central nervous system diffuse large B-cell lymphoma.16, 17, 18, 34 The prognostic value of MYC protein expression for central nervous system diffuse large B-cell lymphoma remains controversial. Therefore, we analyzed MYC protein expression by immunohistochemistry, MYC gene status by FISH, and MYC mRNA expression by RNA in situ hybridization, to investigate the correlation between these parameters and to evaluate their prognostic significance.

MYC protein overexpression was more frequent in central nervous system diffuse large B-cell lymphoma (43–92%) than it was in non-central nervous system diffuse large B-cell lymphoma (29–64%).10, 11, 12, 14, 16, 17, 18, 34 MYC protein expression was identified in 66% of our cohort. The high prevalence of MYC protein overexpression observed in central nervous system diffuse large B-cell lymphoma may be related to the predominance of the non-germinal center subtype in central nervous system diffuse large B-cell lymphoma, as MYC expression is more frequent in the non-germinal center subtype of non-central nervous system diffuse large B-cell lymphoma.7, 10 Our data also showed the presence of a significant association between MYC protein overexpression and the non-germinal center subtype (P=0.003, Table 1).

The data regarding the prognostic value of MYC protein expression for central nervous system diffuse large B-cell lymphoma are inconsistent.17, 18, 19 Two studies comprising 14 and 42 central nervous system diffuse large B-cell lymphoma cases performed by Chang et al.19 and Tapia et al.,18 respectively, reported that MYC overexpression was associated with an adverse prognosis; however, in the cohort of 59 patients reported by Gill et al.,17 no association was found with clinical outcome. In our work, MYC protein expression was related to poor overall survival (P=0.016) and appeared as an independent prognostic factor in a multivariate analysis (P=0.016). The impact of MYC and BCL2 co-expression was also evaluated as their unfavorable clinical impact was reported in non-central nervous system diffuse large B-cell lymphoma;7, 11, 13 however, concurrent expression of MYC and BCL2 was not of prognostic value in this work, similar to the study of Tapia et al.18 In addition, a prognostic difference according to germinal center/non-germinal center subtype was not observed in this study, which was in line with the studies of Hattab et al.39 and Raoux et al.40

We also evaluated MYC mRNA expression using RNA scope. The central nervous system diffuse large B-cell lymphoma showed high-MYC mRNA expression, which exceeds that reported for non-central nervous system diffuse large B-cell lymphoma.14 The impact of MYC mRNA expression on patient outcome was not statistically significant (P=0.055).

MYC translocation was observed only in 7% of cases. There was no significant difference in prognosis according to MYC translocation status. Consistent with a previous study, the rarity of MYC translocation found here makes it unsuitable as a prognostic tool for central nervous system diffuse large B-cell lymphoma.15, 16, 17, 18 Furthermore, BCL2 FISH was performed in the five MYC-FISH-positive cases to investigating additional gene alterations, and no translocation involving BCL2 was detected.

Deregulation of the MYC oncogene contributes to tumorigenesis, via its transcriptional and nontranscriptional roles in cellular processes, such as proliferation, differentiation, and metabolism. MYC deregulation is mediated not only by direct alterations, such as amplification or chromosomal translocation, but also by the activation of many receptor signaling pathways. The deregulated expression of MYC results in an increase of MYC protein expression.5 The genetic mechanism underlying MYC protein expression in central nervous system diffuse large B-cell lymphoma has not been elucidated. However, studies aimed at disclosing the genetic and epigenetic mechanisms of MYC regulation are in progress. Aberrant somatic hypermutation of the MYC gene has been identified in nine out of 10 cases of central nervous system diffuse large B-cell lymphoma; however, the impact of MYC mutations on MYC protein expression in central nervous system diffuse large B-cell lymphoma remains unknown.20

We examined the correlation between MYC translocation, MYC mRNA expression, and MYC protein expression. In our study, MYC translocation showed no statistically significant correlation with MYC mRNA expression (P=0.081) (Fisher’s exact test). However, all cases with MYC translocation showed MYC protein overexpression. These results suggest that MYC translocation leads to overexpression of the MYC protein, although many more MYC-translocation-negative cases overexpress MYC protein via other mechanisms. The mechanisms of MYC deregulation are not restricted to translocations or amplifications of the MYC locus, and MYC can be deregulated by any one of several mechanisms that target its expression and/or activity, either directly or indirectly.41

In our study, MYC mRNA expression showed moderate positive correlation with MYC protein expression (P<0.001, r=0.544) (Pearson’s correlation analysis). Several reports have shown more frequent MYC protein overexpression in central nervous system diffuse large B-cell lymphoma than in non-central nervous system diffuse large B-cell lymphoma,10, 11, 12, 14 while the difference of MYC mRNA expression between them was controversial. Rubenstein et al.27 presented increased MYC mRNA expression in central nervous system diffuse large B-cell lymphoma compared to non-central nervous system diffuse large B-cell lymphoma. Brunn et al.16 reported that MYC mRNA level in central nervous system diffuse large B-cell lymphoma was not different from that of non-central nervous system diffuse large B-cell. Our results suggest that MYC protein overexpression cannot be completely accounted for by mRNA overexpression; posttranscriptional or posttranslational modulation may also play a role.

The MYC protein is highly unstable and its destruction is mediated by post-translational modification, ubiquitination, and degradation.42, 43 Mutation of the MYC gene may stabilize the MYC protein and contribute to tumorigenesis, as in Burkitt’s lymphoma.38, 44, 45 Moreover, deregulation of microRNAs (miRNAs) may be associated with MYC protein overexpression, via translational modulation. Numerous miRNAs have been shown to regulate MYC expression.46, 47 MYC-regulated miRNAs, such as miR-17-5p and miR-20a, which belong to the miR-17-92 clusters, are upregulated in central nervous system diffuse large B-cell lymphoma compared with non-central nervous system diffuse large B-cell lymphoma.48 MYC-activated miR-17-92 clusters contribute to lymphomagenesis via amplification of B-cell receptor (BCR) signaling, and stimulated BCR responses result in elevation of MYC itself, thus forming a feed-forward loop.49 Further studies of miRNA-mediated pathways in central nervous system diffuse large B-cell lymphoma may explain the observed MYC overexpression.

We evaluated MYC mRNA expression using the RNA scope method, which quantifies RNA via an in situ detection method. This technique, which was first developed by Wang et al.,37 enables direct counting of mRNA molecules in single cells in routine formalin-fixed, paraffin-embedded tissue specimens using bright-field microscopy. The false-negative results obtained from admixtures of many non-malignant cells with tumor cells in real-time RT–PCR can be overcome using this method; however, specimen quality may be another problematic issue. In our study, the experiment was successfully performed using paraffin archives and manual counting of a single signal, followed by H-scoring, which yields accurate measurement of mRNA expression.

In summary, we describe MYC protein expression as an important prognostic marker for the stratification of patients with central nervous system diffuse large B-cell lymphoma. Moreover, we evaluated MYC mRNA expression using the RNA in situ hybridization method, and assessed its impact on patient survival for the first time. In addition, we assessed the correlation between MYC gene status, MYC mRNA expression, and MYC protein expression in central nervous system diffuse large B-cell lymphoma. The present study had several limitations. Because of its retrospective nature, the patients were not highly selected, as would be done in a prospective study. Further studies of the complete molecular mechanisms underlying MYC protein overexpression would be useful to validate and expand our findings.