Main

Diffuse large B-cell lymphoma is the most common subtype of non-Hodgkin's lymphoma and accounts for 30–40% of new diagnoses.1 Prognosis of diffuse large B-cell lymphoma patients is poor. Despite multiagent chemotherapy, durable remissions are achieved in only 40–50% of patients. Current attempts to determine prognosis in diffuse large B-cell lymphoma rely on clinical parameters, but are still not reliable enough to predict the course of the disease in individual patients.2 It is hoped that a better understanding of the molecular basis of the disease will eventually lead to better prognostic markers. Indeed, several new proteins or groups of genes playing a role in prognosis or that may potentially serve as therapeutic targets have recently been discovered.3, 4, 5, 6

The fragile histidine triad (FHIT) gene located on chromosome 3p14.2 at fragile site, FRA3B, belongs to these genes that have recently been linked to diffuse large B-cell lymphoma.7, 8 FHIT is known to be inactivated in various human malignancies.9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 FHIT inactivation by point mutation is a rare event,21, 22, 23 but significant loss or reduction of expression can be caused by other mechanisms, including loss of heterozygosity (LOH) and/or promoter hypermethylation.24 For example, FHIT hypermethylation with consequent transcriptional inactivation has been shown in breast, lung, esophageal, cervical, prostate and bladder cancer.25, 26, 27, 28, 29 For breast cancer, it was demonstrated that hypermethylation of one allele can occur in conjunction with LOH, and that these two events can constitute the ‘two hits’ required for the complete gene silencing.30

A recent immunohistochemistry study on 31 diffuse large B-cell lymphoma patients had suggested that decreased or absent FHIT protein expression may herald poor prognosis in diffuse large B-cell lymphoma.7 More recently, it was shown that microdeletions within the FHIT gene result in the selective loss of certain exons, which can cause aberrant RNA expression in diffuse large B-cell lymphoma. Other mechanisms of reduced FHIT expression have not been analyzed in diffuse large B-cell lymphoma. The aims of this study were therefore two-fold. First, we aimed at a confirmation of the prognostic relevance of reduced FHIT expression in a series of >100 diffuse large B-cell lymphoma. Second, we investigated the role of promotor methylation status for FHIT inactivation. Overall, our data confirm a major role of FHIT alteration in the pathogenesis of diffuse large B-cell lymphoma.

Materials and methods

Tissue Samples

Formalin-fixed, paraffin-embedded samples from 190 newly presenting and previously untreated patients with diffuse large B-cell lymphoma were investigated. Diagnosis was confirmed by pathologic review using the diagnostic criteria defined in the revised European-American Classification Lymphoid Neoplasms/WHO Classification.31 Clinical follow-up information was available from all patients. Study approval was obtained from the Research Advisory Council (RAC #2030 019) at King Faisal Specialist Hospital and Research Centre. Tissue microarry construction was as described.32 Briefly, tissue cylinders with a diameter of 0.6 mm were then punched from representative tumor regions of each ‘donor’ tissue block and brought into a recipient paraffin block using a home made semiautomated precision instrument.

Methylation-Specific Polymerase Chain Recation Analysis

For methylation-specific polymerase chain reaction analysis, genomic DNA was either extracted with a Puregene kit (Gentra, Minneapolis, MN, USA) or was available from previous studies.29, 33 One microgram of genomic DNA was denatured in 0.4 M NaOH and modified with 3 M sodium bisulfite and 10 mM hydroquinone at 55°C for 16 h. After purification with a GeneCleanIII kit (Bio 101, Vista, CA, USA), the DNA was desulfonated in 0.4 M NaOH, precipitated in ethanol, and resuspended in dH2O. Then 200 ng was used as a template in methylation-specific polymerase chain reactions with 1.5 mM MgCl2 and 20 pmol of primers specific for methylated (M) and unmethylated (U) forms.25 The methylated FHIT reaction consisted of 32 cycles of touchdown PCR at annealing ranging from 71 to 63°C with primers TGGGGCGCGGGTTTGGGTTTTTACGC and CGTAAACGACGCCGACCCCACTA. The unmethylated FHIT reaction was done at 64°C for 33 cycles with primers TTGGGGTGTGGGTTTGGGTTTTTATG and CATAAACAACACCAACCCCACTA, corresponding to nucleotides 189–301 (GenBank Accession Number U76263). Each reaction was tested with untreated DNA to ensure lack of amplification, and three controls were included to ensure specificity: (1) normal human DNA previously treated with the CpG methylase SSS1 in the presence of S-adenosylmethionine (in vitro methylated DNA); (2) DNA from peripheral lymphocytes from a healthy individual (normal control); and (3) no template (blank). PCR products were analyzed after electrophoresis on 4% agarose gels containing ethidium bromide.

Immunohistochemical Staining for FHIT Protein

Paraffin-embedded 5 μm sections from the tissue microarry block were stained for FHIT protein, according to the method described by Yang et al.34 Briefly, paraffin embedded sections on polylysine coated slides were dewaxed with xylene and rehydrated through a graded alcohol series. Endogenous peroxidase activity was blocked in 3% hydrogen peroxidase in methanol for 10 min. Antigen retrieval was performed by placing the sides in a Citrate buffer (pH 6.0) and microwaving them for 5 min at 750 W and for 15 min at 250 W. The sections were incubated for 90 min in 1:900 dilutions of polyclonal rabbit antibodies reacting against FHIT protein (ZR44 Zymed, USA). Bound antibody was detected with biotinylated link antibody (Dako, Glostrup, Danmark) and horse radish peroxidase labeled streptavidin (Dako). The reaction was developed in 3,3′-diamino benzidine with H2O2 as substrate (Dako). The sections were then counterstained with Gills hematoxylin. The primary antibody was omitted in negative control sections.

Expression was scored on a four tiered scale for both intensity (grade 0, no staining; grade 1, weak; grade 2, moderate; grade 3, strong) and extent (grade 1, percentage of positive cells is <10%; grade 2, 10–50%; grade 3, >50%). The intensity and extent scores were multiplied to give a composite score (1–9) for each tumor. Score 0 was defined as absent or lost expression, scores 1–3 were defined as markedly reduced FHIT expression and scored 4–9 were considered as normal expression.35, 36, 37

Statistical Analysis

Statistical analysis was performed using SAS's (SAS Institute Inc.) JMP 5.1 software (Cary, NC, USA), and all P-values reported are two-tailed. Univariate analysis of categorical variables was conducted using contingency analysis and χ2 tests. Surviving curves were plotted according to the Kaplan–Meier method. Survival differences between groups were analyzed by log-rank test.

Results

FHIT Immunohistochemistry

Immunohistochemical staining for FHIT protein expression was successful in 114 of 190 diffuse large B-cell lymphoma. The absence of tissue or lack of clearly discernible tumor cells were the cause of noninformative results in 76 additional cases. Out of 114 informative cases, 39 (34%) showed strong (Figure 1c), 57 (50%) weak (Figure 1b), and 18 (16%) absent FHIT staining (Figure 1a), according to our definition. No survival difference was seen between diffuse large B-cell lymphoma with different FHIT expression level.

Figure 1
figure 1

Diffuse large B-cell lymphoma tissue microarry. (a) absent fhit expression, × 10; (b) reduced fhit expression, × 10; (c) normal fhit expression, × 10; (d) H&E stained tissue micro array of diffuse large B-cell lymphoma; (e) normal fhit expression, × 40; and (f) absent fhit expression, × 40.

Full size image

FHIT Methylation

FHIT promoter hypermethylation analysis was successful in 93 of 114 diffuse large B-cell lymphoma with available immunohistochemistry data (82%). Unsuccessful analyses were due to insufficient DNA quality in 21 cases. FHIT hypermethylation was found in 29 (23%) of 93 interpretable samples (Figure 2). FHIT methylation was unrelated to lymphoma stage (Table 1), but was significantly associated with short patient survival P=0.023 (Figure 3). A comparison of methylation and immunohistochemistry data revealed methylation in 15 of 59 (25%) cases with absent or reduced FHIT expression by immunohistochemistry and in six of 34 (17%) tumors with normal FHIT expression. This association was statistically not significant.

Figure 2
figure 2

Methylation analysis. Methylation-specific PCR analyses of seven representative NHL samples (labeled 1–7 on the top) including normal PBL as positive control for unmethylated reacation and in vitro methylase treated (IVM) DNA as positive control for methylated reaction. Both methylated (M) and unmethylated (U) reactions were amplified for each bisulfite-treated DNA and run in a 4% agarose gel.

Full size image
Table 1 FHIT expression by IHC, methylation status of FHIT protein and characteristics of patient with diffuse large B-cell lymphoma
Full size table
Figure 3
figure 3

Prognostic significance of FHIT promoter methylation in diffuse large B-cell lymphoma.

Full size image

Relationship to Diffuse Large B-Cell Lymphoma Subtype

CD10 and bcl6 immunohistochemistry to define germinal center (CD10/bcl6 positive) and nongerminal center (CD10/bcl6 negative) diffuse large B-cell lymphoma subtypes had previously been performed in our tumors.38 This analysis had unequivocally identified eight germinal center (CD10/bcl6 positive) and 45 nongerminal center (CD10/bcl6 negative) in our 114 interpretable diffuse large B-cell lymphoma. Remarkably, our comparison of FHIT results and diffuse large B-cell lymphoma phenotype revealed discrepant results for methylation and immunohistochemistry data. Methylation results showed a tendency towards more FHIT methylation in nongerminal center phenotype (12 of 45; 27%) than in germinal center phenotype (1 of 8; 13%). At the same time, the immunohistochemistry data suggested expression loss to be more frequent in germinal center (reduced in nine of 10 cases, 90%) than in nongerminal center (reduced in 35 of 57, 61%) phenotype (P=0.05).

Discussion

Our data suggest that promotor hypermethylation contributes to FHIT downregulation in diffuse large B-cell lymphoma. This is of potential clinical importance as new treatment regimens targeting and reversing hypermethylation of FHIT are now in clinical trials. FHIT belongs to the most commonly altered genes in all human cancers, and is believed to be inactivated in 20–100% (depending on the tumor type) of neoplasias (reviewed in Pekarsky et al39). A potential efficacy of such drugs is supported by findings from clinical trials in various solid tumors,40 for example, non-small-cell lung cancers41 and squamous cell carcinomas of the cervix.42 If such drugs should proof to be efficient in humans, about one-third of diffuse large B-cell lymphoma patients could potentially benefit from such treatments.

The exact molecular mechanism or functional pathway mediating FHIT's tumor suppressor action is still not fully understood. It is known that FHIT hydrolases diadenosine nucleotides into ADP and AMP, but since this enzymatic activity does not seem to be required for its tumor-suppressor function there must be other relevant features of the protein,12 that is regulation of apoptosis. It has been demonstrated that restoration of FHIT expression in lung and cervical cancer cell lines resulted in efficient induction of apoptosis and suppression of tumorigenicity, and that the apoptotic mechanism seems to be FADD (Fas associated via death domain) dependent, caspase-8 mediated and independent from regulation through Bcl-2 or Bcl-xl.43, 44 Most recent, it has bee shown that FHIT modulates the Pi3k/AKT pathway by downregulation of the antiapoptotic survivin, an inhibitor of apoptosis protein (IAP) family member.45

Our data indicate that inactivation of FHIT might be due to different reasons. A comparison of expression data as observed by immunohistochemistry and promotor hypermethylation did not show a significant association. The much higher frequency of expression loss (66%) as compared to hypermethylation (23%) raises the possibility that other mechanisms than hypermethylation may reduce FHIT expression in most diffuse large B-cell lymphoma. Small deletions that selectively eliminate individual FHIT exons have recently been found in about 30% of diffuse large B-cell lymphoma.8 As the epitope where the antibody binds is not known, it cannot be excluded that microdeletions may constitute another main reason for reduced expression as detected by our antibody. Discrepancies between immunohistochemistry and methylation analysis also included a small number of cases (n=6) with normal FHIT expression but hypermethylation. Tumor heterogeneity for methylation is a possible explanation for these cases. In addition, it is presumed that methylation usually occurs monoallelic, and complete loss of expression is a consequence of a combination of methylation and allelic loss.25 Finally, technical immunohistochemistry problems, including variable immunoreactivity because of different fixation conditions, might have contributed to the discrepant findings. For example, we used the same antibodies as described by Chen et al,7 but found a slightly higher fraction (65%) of cases with reduced expression as compared to the 58% in Chen's description.

Inherent limitations of the tissue microarry approach could also have contributed to the relatively high number of FHIT negative cases in this study. Focal reduction of immunoreactivity or biologic heterogeneity can lead to false negative immunostainings on tissue microarrys. It has been shown, however, that some of the disadvantages caused by the small size of samples analyzed on a tissue microarry will be compensated by the maximal standardization of tissue microarry analysis and interpretation.46 For example, in one previous study, the prognostic significance of p53 positivity in breast cancer was identified on several different tissue microarrys manufactured from a series of >500 cancers but not on corresponding large sections.47 This study shows that at least in some instances, tissue microarrys composed of one 0.6 mm sample per tumor can be superior over traditional large sections for identification of prognostic biomarkers.

A true prognostic role of FHIT inactivation in diffuse large B-cell lymphoma could be supported by the significant association observed between FHIT hypermethylation and short survival. PCR based hypermetylation analysis clearly is a more robust and reproducible method than immunohistochemistry, which is prone to numerous technical shortcomings.48, 49 However, our data also raise the possibility that certain FHIT inactivation mechanisms could be linked to different diffuse large B-cell lymphoma subtypes. With the exception of one case, FHIT hypermethylation was only seen in the nongerminal center diffuse large B-cell lymphoma subtype. Thus, the poor prognosis observed for FHIT methylated diffuse large B-cell lymphoma could be explained by the generally poor prognosis of nongerminal center diffuse large B-cell lymphoma previously reported in both Western2 and Saudi patients.38 Remarkably, such a tendency to an association with nongerminal center subtype was not found for reduced FHIT protein expression. In contrary, there was even a clear tendency towards a lower frequency of reduced FHIT expression in nongerminal center (60%) as compared to germinal center (88%) diffuse large B-cell lymphoma subtype. Although the respective P-values did not reach significance (P=0.1) and the reliability of immunohistochemistry analysis is to some extent limited, this result raises the possibility that hypermethylation is primarily inactivating FHIT in nongerminal center diffuse large B-cell lymphoma while other mechanisms apply for FHIT inactivation in the germinal center diffuse large B-cell lymphoma subtype.

In summary, our data show that hypermethylation is a relevant mechanism for FHIT inactivation in diffuse large B-cell lymphoma and suggest a link of hypermethylation to nongerminal center subtype and poor prognosis. If methylated FHIT should indeed constitute a suitable therapeutic target, diffuse large B-cell lymphoma patients could substantially benefit from such new drugs. Overall, the accumulating data on DNA level FHIT alterations provide strong evidence for an important role of FHIT in development or progression in diffuse large B-cell lymphoma.