Mutations in isocitrate dehydrogenase isoforms 1 and 2 are rare events in primary central nervous system and non-central nervous system diffuse large B cell lymphoma


Dr Se Hoon Kim, MD, PhD, Department of Pathology, Yonsei University College of Medicine, 134 Shinchon-dong, Seodaemun-gu, Seoul 120-752, Korea. Email:


Background and aim: Mutations in isocitrate dehydrogenase enzyme isoforms 1 (IDH1) and 2 (IDH2) have been identified in many cancers, including gliomas and acute myeloid leukemia. The aim of this study was to determine whether these genes are mutated in the primary central nervous system lymphoma (PCNSL). Methods: We analyzed IDH1 and IDH2 mutations in 20 PCNSL and 30 non-central nervous system (CNS) diffuse large B-cell lymphomas (DLBCLs) using routine formalin-fixed paraffin-embedded tissues and a polymerase chain reaction assay focusing on the known mutation hot spots, Arg 132 of IDH1 and Arg 172 of IDH2. Results: We did not detect mutations in IDH1 or IDH2 in PCNSL nor non-CNS DLBCL. Conclusions: Although IDH1 and IDH2 mutations are known to be important in gliomas and in some hematologic malignancies, they appear to be very rare events in PCNS and non-CNS DLBCLs. Further studies on other IDH mutations in larger, non-CNS DLBCL populations are needed.


Isocitrate dehydrogenase enzyme isoforms 1 (IDH1) and 2 (IDH2) catalyze oxidative decarboxylation of isocitrate to α-ketoglutarate and NADP+ to NADPH, and are involved in reverse cellular metabolisms.1–3 IDH1 plays a role in cholesterol synthesis in the peroxisome,4 protects cells against gamma radiation,5 singlet oxygen,6 and UVB radiation,7 and senses glucose and insulin secretion.8 Arg 132 in exon 4 of IDH 1 is aligned with Arg 172 of IDH2. Approximately 90% of IDH1 mutations are R132H amino acid substitutions; other reported mutations are R132C, R132S, R132S, R132G, and R132L. Known amino acid substitutions of IDH2 mutations are R172K, R172M, R172W, and R140Q.9

Recently, IDH1 and IDH2 mutations have been reported in gliomas,10–13 acute myeloid leukemia (AML),14 and myeloproliferative neoplasms.15 Although rare, a mutation of IDH1 has been found in B cell acute lymphoblastic leukemia,16 primitive neuroectodermal tumor,9 prostate cancer,16 and colorectal cancer.17 Interestingly, several studies have demonstrated that IDH1/IDH2 mutations can be a potential prognostic factor for longer survival.11,13,18

Approximately 98% of primary central nervous system (PCNS) lymphomas are of B cell origin, mostly diffuse large B-cell lymphomas (DLBCLs).19 Although it is controversial, PCNS DLBCLs are thought to have a poorer prognosis than non-central nervous system (CNS) DLBCLs.20 The aim of this study was to evaluate and analyze the incidence of IDH1/2 mutations at common mutations sites in PCNS DLBCL and non-CNS DLBCL. Furthermore, we investigated a possible pathogenic relationship between PCNS DLBCLs and IDH mutations, compared with non-CNS cases.


Patients and tissue samples

We retrieved formalin-fixed paraffin-embedded (FFPE) tissues of 20 primary CNS-DLBCLs (14 males and six females; range, 28 to 75 years; mean, 60.15 years) with typical morphologic and immunophenotypical features, which were collected between 2000 and 2009, from the Pathology Department archives at the Yonsei University College of Medicine in Seoul, Korea (Table 1). This study was approved by the Institutional Review Board of Yonsei University College of Medicine (4–2009-0436). We diagnosed primary CNS-DLBCLs, which are defined by the World Health Organization (WHO) as extranodal DLBCLs that arise in the parenchyma of the CNS in the absence of obvious lymphoma outside the nervous system at the time of the diagnosis. We excluded leptomeningeal DLBCLs from this study. All 20 samples were independently reviewed by two pathologists (Noh and Kim). We confirmed the diagnosis by CD20 immunohistochemistry (Fig. 1). Thirty cases of non-CNS DLBCL (18 males and 12 females; range, 34 to 83 years; mean, 60.33 years), which were biopsied between 2000 and 2003 were randomly selected as controls (Table 2). The non-CNS DLBCLs originated from lymph nodes (15 cases), stomach (1), soft tissue (3), tonsil (3), testis (3), small bowel and colon (2), epididymis (1), pancreas (1), and tongue (1) (Table 2).

Table 1.  Clinical information of primary central nervous system diffuse large B-cell lymphoma
Case no.SexAge (years)StageCTxRTxStatusFollow-up (months)GCB/ABC
  1. ABC, activated B-cell-like type; CTx, chemotherapy; GCB, germinal center B-cell-like type; F, female; M, male; RTx, radiotherapy.

Figure 1.

Histologic findings of a case of primary central nervous system diffuse large B-cell lymphomas (case No.15). Diffuse infiltrations of large atypical lymphocytes (a, hematoxylin and eosin, ×200) with membranous CD20 immunoreactivity (b, CD20, ×200).

Table 2.  Clinical information of non-central nervous system diffuse large B-cell lymphoma
Case no.SexAge (years)StageCTxRTxStatusFollow-up (months)GCB/ABCOrgan
  1. ABC, activated B-cell-like type; CTx, chemotherapy; GCB, germinal center B-cell-like type; F, female; M, male; RTx, radiotherapy.

 1F72IYesNoDead18.4GCBLymph node
 2M47IVYesNoDead17.5GCBLymph node
 3F58IYesNoDead15.1ABCLymph node
 4M51IYesNoAlive63.6ABCLymph node
 6M66IIIYesNoDead15.7GCBLymph node
 7M48IIYesNoAlive41.3ABCLymph node
 9F83INoNoDead3.9ABCLymph node
12M36IIIYesNoDead8.9ABCSoft tissue
13F75IIYesNoDead4.4GCBLymph node
14F65IYesYesAlive36.1GCBLymph node
15F62IYesYesAlive35.1ABCLymph node
16M64IYesNoAlive34.6ABCLymph node
19F75IIIYesYesAlive31.5ABCLymph node
23M53IYesYesAlive25.5ABCSoft tissue
24M70IVYesNoAlive35.3GCBSoft tissue
25M67IIIYesYesDead18.8GCBLymph node
28M70IIIYesNoDead17.6ABCLymph node
30M59IIIYesNoDead34.9ABCLymph node

Imuunohistochemical analysis

Immunohistochemical staining was performed as described in the following steps. Serial sections (4 μm) were applied to silane-coated slides (Muto Pure Chemicals, Tokyo, Japan). Deparaffinization and rehydration were performed using xylene and graded alcohols. Endogenous peroxidase was blocked with 3% aqueous hydrogen peroxide for 10 min. For CD3 (1:40; Dako, Glostrup, Denmark), CD20 (1:1,600; Dako) and CD10 (1:100; Novocastra, Newcastle upon Tyne, UK), the slides were pretreated in a microwave oven for 15 min in a pressure cooker filled with 0.01 M citric acid buffer (pH 6.0). On the other hand, we used Tris-HCl buffer (pH 9.0) instead of 0.01 M citric acid buffer (pH 6.0) for Bcl-6 (1:20; Dako) and multiple myeloma-1 (MUM1, 1:150; Dako) immunohistochemical staining. The slides were then incubated at room temperature for 1 h with primary antibodies, followed by overnight incubation at 4°C. For Bcl-6 and CD10 staining, overnight incubation was performed at room temperature. After washing, the signals were detected with an LSAB2 kit (Dako) and diaminobenzidine as a chromogen. The slides were also counterstained with hematoxylin. Appropriate positive and negative controls were included in all stains to ensure the quality and consistency of staining results. Immunohistochemical staining results were interpreted without any clinicopathologic information. In cases showing discrepancy between the two interpreters, the final score of immunohistochemistry results was decided by discussion. The immunohistochemistry results were scored using a semiquantitative scoring system based on the percentage of immunopositive cells (0, negative; 1, < 20% of tumor cells; 2, 20% to approximately 50% of tumor cells; 3, 51% to approximately 80% of tumor cells; and 4, > 80%) as proposed by Chang et al.21 Except for Bcl-2, the expression of each marker was considered “positive” when the score was greater than 2 (> 20% of positive neoplastic lymphocytes). In the case of Bcl-2, staining was considered positive with a score greater than 3 (> 50% of neoplastic lymphocytes positive), in accordance with a previously published method.22 We distinguished two subgroups of DLBCL with the following three markers according to the decision tree presented by Hans et al.23: CD10, Bcl-6, and MUM1. The germinal center B-cell-like (GCB) type includes all CD10+ cases and those with a CD10Bcl-6+MUM1 immunophenotype. Other patients were grouped as activated B-cell-like (ABC) type, which includes MUM1+ cases, regardless of their Bcl-6 status. According to this classification, cases expressing none of the three markers were classified as ABC subgroup.

IDH1/2 sequencing assays

IDH1/2 assays were performed according to the method described by Horbinski et al.24 Briefly, FFPE tumor tissues were manually microdissected to 5 μm in thickness. DNA was isolated from each sample using a QIAamp DNA FFPE tissue kit (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. The quantity of isolated genomic DNA was evaluated using a NanoDrop 1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA). DNA was extracted from samples of PCNS and non-CNS DLBCLs for the detection of IDH1/2 mutations by polymerase chain reaction (PCR) using forward and reverse primers that were designed to amplify exon 4 (codon R132) of the IDH1 gene and exon 4 (codon R172) of the IDH2 gene. IDH1 forward primer (5′-ACC AAA TGG CAC CAT ACG A-3′) and reverse primer (5′-GCA AAA TCA CAT TAT TGC CAA C-3′) generated a 130-bp PCR product; IDH2 forward primer (5′-GCT GCA GTG GGA CCA CTA TT-3′) and reverse primer (5′-TGT GGC CTT GTA CTG CAG AG-3′) generated a 293-bp PCR product. PCR amplification was performed using an AmpliTaq Gold PCR Master Mix (Applied Biosystems, Foster City, CA, USA). The reaction mixture was subjected to an initial denaturation at 95 °C for 10 min, followed by 35 cycles of amplification consisting of denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s, and extension at 72 °C for 60 s. After purification and sequencing amplification, the sequencing products were analyzed by a 3730XL DNA sequencer (Applied Biosystems).


Clinical profiles

Among 20 cases of PCNS-DLBCLs, 19 cases (95%) were stage I and, one case (5%) stage IV case, which directly involved the orbit. Thirty cases of non-CNS DLBCLs included 13 cases (43%) of stage I, seven cases (23%) of stage II, seven cases (23%) of stage III and three cases (10%) of stage IV. Most cases of PCNS-DLBCLs were stage I, compared with non-CNS DLBCLs.

In primary central nervous system lymphomas (PCNSLs), three cases (15%) were GCB type, and 17 cases (85%) were ABC type. In non-CNS DLBCLs, 12 cases (40%) were GCB type, and 18 cases (60%) were ABC type. Although it was not statistically significant (P= 0.059), the percentage of ABC type in PCNSLs was higher.

In PCNS DLBCLs, seven patients received chemotherapy and radiotherapy, and seven patients received chemotherapy only. Six patients did not receive any treatment. Most chemotherapy regimens were the high-dose methotrexate, vincristine sulfate, procarbazine, and dexamethasone (HD MVP) regimen with or without modifications. In non-CNS DLBCLs, 10 patients received chemotherapy and radiotherapy, and 18 patients received chemotherapy only. Two patients did not receive any treatment. Most chemotherapy regimens were the rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP) regimen with or without modifications.

Clinical information of PCNS DLBCLs and non-CNS DLBCLs is summarized in Tables 1 and 2, respectively.

IDH1/2 sequencing assays

All CNS and non-CNS lymphoma had CGT nucleotides in exon 4 of the IDH1 gene and AGG nucleotides in exon 4 of the IDH2 gene (Fig. 2). Eventually, we did not detect IDH1 or IDH2 mutations in any of the PCNS DLBCL and non-CNS DLBCL samples.

Figure 2.

There were no mutations at codon 131 (CGT) of isocitrate dehydrogenase enzyme isoforms (IDH)1 in case No.18 or at codon 172 (AGG) of IDH2 in case No.16.


IDH1 and IDH2 mutations have been found often in gliomas13 and occasionally in AML.14 As mentioned above, most of these mutations were at R132 in IDH1 and at R172 and R140 in IDH2.9 A mutation in only IDH1 has been found in B-cell acute lymphoblastic leukemia,16 primitive neuroectodermal tumor,9 prostate cancer,16 colorectal cancer,17 and rarely in melanoma.25

At first, we hypothesized that IDH1 and IDH2 mutations might be found in PCNS DLBCLs because IDH1 mutation is an unfavorable prognostic factor in AML,14 and PCNS DLBCLs are thought to have a poorer prognosis than non-CNS DLBCLs.20 We raised a question whether an IDH mutation could be an important prognostic factor between PCNS DLBCLs and non-CNS DLBCLs. However, PCNS DLBCLs had no IDH mutations in this study. Furthermore, no IDH mutations were found in non-CNS DLBCLs. As we failed to demonstrate IDH mutations at frequent mutation sites in both PCNS and non-CNS DLBCLS, we concluded that the brain is not a specific environment where IDH mutations play an important role in lymphomatogenesis. In addition, we believe that IDH mutations rarely occur in DLBCLs.

Our findings are consistent with those of Lopez et al.,25 who reported that, in general, IDH mutations may be necessary for the formation of tumors in a cell-lineage dependent manner with a particularly strong selective pressure for mutations in progressive glioma, but not in brain tissue.25 Our findings suggest that IDH mutations are not common in PCNS DLBCLs. Limitations of this study are the small number of PCNSL cases, FFPE tissues, and the fact that we only investigated mutations at well-known hot spot foci. However, we believe that this is a valuable pilot study for more extensive investigations to determine the frequency of a wider range of IDH mutations in a larger study population.


This work was supported by the National Research Foundation of Korea grant funded by the Korean government (MEST) (2010–0021092).