A combination of LMO2 negative and CD38 positive is useful for the diagnosis of Burkitt lymphoma

Burkitt lymphoma (BL) is one of the most studied human malignant tumors that originates in the B cells. Although it is relatively simple to diagnosis BL in children, it is a challenge to identify reliable subtypes of aggressive B-cell lymphoma in adults [1, 2]. It is crucial to distinguish BL from other lymphomas because of its rapid progress and the planned improvements in treatment for adult aggressive B-cell lymphomas [1‐4].

BL is a highly aggressive B-cell lymphoma with unique morphologic, immunophenotypic, and molecular features [5]. BL tumor cells are monomorphic, composed of medium-sized cells with round nuclei, multiple deeply stained nucleoli, and basophilic cytoplasm. The cell proliferation rate as well as the apoptotic rate are extremely high. Approximately 100% of the cells are Ki-67 positive (MIB-1 positive) and display the “starry sky” pattern. BL has a typical immunophenotype-strong immunoglobulin (Ig) expression and generally expresses markers of B cell-associated antigens (CD19, CD20, CD22, and CD79a) and a germinal center (CD10). It does not express BCL-2 [6]. In nearly all studies, BL was associated with one of three chromosomal translocations on the c-MYC oncogene locus (8q24) and the Ig gene on the long arm of chromosome 14, also the immunoglobulin light chain genes on chromosomes 2 and 22 [7‐9].

High-grade B-cell lymphoma, NOS includes blastoid-appearing large B-cell lymphomas and cases lacking MYC and BCL2 or BCL6 translocations. HGBL, with MYC and BCL2 and/or BCL6 and HGBL, NOS replaces the 2008 category of B-cell lymphoma, unclassifiable, with features intermediate between DLBCL and Burkitt lymphoma (BCLU) [5]. Most morphologic features are intermediate between those of DLBCL and BL, with a high proliferative index and starry sky pattern, and the immunophenotype is consistent with that of BL.

Burkitt-like lymphoma with an 11q aberration has morphologic and immunophenotypic features similar to those of BL, but lacks MYC rearrangement and has the typical 11q aberration, which appears as a partial amplification and partial deletion in the region at the same time [10]. The tumor is rare, accounting for only 3% of BL, is more common in children and young people and more in males than females, and is more likely to involve lymph nodes than BL [11].

The above lymphomas are difficult to distinguish from their histological morphologies and existing routine immunophenotypes. Our study hopes to discover new immunohistochemical markers and analyze their expressions in these lymphomas so as to better diagnosis of BL.

LMO2 is a transcription factor that plays an important role in embryonic development and angiogenesis. Studies have shown that many tumors have LMO2 expression and that it is associated with the prognosis for patients with certain tumors, such as glioblastoma and pancreatic cancer [12, 13]. In the lymphatic and hematopoietic system, in addition to expression in the normal lymphoid germinal center, LMO2 is expressed in germinal center-derived lymphomas, acute B-lymphoblastic leukemia, and acute myeloid leukemia (AML) [14]. Recent studies have found that LMO2 protein expression is downregulated or negative in BL with abnormal MYC [2]. CD38 is a type II transmembrane glycoprotein that has several complex and unique biological characteristics and functions. It is widely expressed in both hematopoietic and non-hematopoietic cells, including bone marrow precursor cells, germinal center B-cells, plasma cells, prostate epithelial cells, skeletal muscle, and other tissues, and in activated T cells, B cells, monocytes, NK cells, and islet cells [15]. CD38 is strongly expressed in both plasma cells and plasma cell tumors. It is also present in acute lymphoblastic leukemia, AML, chronic lymphocytic leukemia, and non-Hodgkin lymphoma (NHL) [16, 17]; however, no in-depth studies have been conducted to verify the positive expression of CD38 in BL.

Our study analyzed the expression of LMO2 and CD38 proteins in BL, HGBL,NOS and Burkitt-like lymphomas with the 11q aberration and hypothesized that the combination of LMO2-negative and CD38-positive expressions can be used to diagnose auxiliary BL. To test this hypothesis, we analyzed the specificity and sensitivity of LMO2-negative, CD38-positive, and the combination of both expressions, as well as their diagnostic efficiency in BL.

BL is a highly aggressive B-cell NHL characterized by the translocation and dysregulation of c-MYC on chromosome 8 [2]. Researchers have questioned whether c-MYC rearrangement is a necessary condition for the diagnosis of BL and have found that ≤5% of the tumors with typical BL characteristics do not have c-MYC rearrangement [1, 22]. Some researchers have speculated that these cases might have molecular pathogeneses other than the MYC activation mechanism, which is the BL’s iconic pathogenesis. Recently, many studies have reported cases with clinical, morphologic, immunophenotypic, or gene expression characteristics consistent with BL, but lacked FISH-detected positive MYC rearrangement. Additional studies have found that there were 11q aberrations in MYC-negative cases [10, 11, 23]; therefore, the 2016 revision of WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues proposed a new temporary type of lymphoma-Burkitt-like lymphoma with the 11q aberration [5].

BL, HGBL,NOS and Burkitt-like lymphoma with the 11q aberration can be diffusely infiltrated by large, medium-sized lymphocytes, no obvious nodule formation, monotonous and consistent cells, and a starry sky pattern. In addition to the expression of B-cell markers, all tumors showed mostly the expression of CD10 positive, BCL-6 positive, and BCL-2 negative in the immunophenotype; therefore, these types of tumors cannot be fully identified using only their morphology and the immunophenotype.

Previous studies have found that LMO2 has high sensitivity and specificity of expression in normal germinal center B cells and germinal center B cell-derived lymphomas. LMO2 was also expressed in myeloid and erythroid progenitor cells, megakaryocytes, lymphocytes, and acute myeloid leukemia. It was rarely expressed in mature T cells, natural killer (NK) cells, or plasma cell tumors. In addition, with the exception of endothelial cells, it did not express in non-lymphoid hematopoietic tissue. In DLBCL, the expression profile of LMO2 was similar to that of other germinal center-related proteins (HGAL, BCL6, and CD10), but was different from nongerminal center proteins (MUM1/IRF4 and BCL2) [14]. Recent studies have found that LMO2 might be a useful indicator for identifying MYC translocation and might also help identify BL [24].

We observed that the deletion of LMO2 expression might be particularly helpful in diagnosing BL. In this series, we found 74 of the 75 BL cases studied were negative for LMO2 using a cutoff of 30%. This was consistent with the data obtained using the GEP study, which indicated that the expression level of LMO2 was lower in BL [1, 2]. Only three studies analyzed the expression of the LMO2 protein in a small number of BL cases. Natkunam and colleagues and Agostinelli and colleagues defined two different cloned LMO2 proteins and evaluated the specificity and effectiveness of their antibodies. In these two studies, the expression rate of LMO2 in BL was 5/10 (50.0%) and 13/32 (41.0%), respectively, and 1/3 (33.3%) in the BL cell line (Ramos cell line). A third study comprised five cases of BL, and LMO2 was expressed in only one case (20%) [14, 25, 26]. Previous studies have also found that the presence of LMO2 protein can distinguish BL from DLBCL [25], because it was more commonly expressed in the latter [26]. In our study, we found that LMO2 protein was 100% positively expressed in HGBL; however, it was only expressed in one of Seventy-five Burkitt lymphomas. There was a statistically significant difference in the negative expression of LMO2 protein between BL and HGBL,NOS(P < .01). The sensitivity and specificity of the negative expression of LMO2 protein were 98.67 and 100%, respectively, and AUC of diagnostic efficiency was 0.993; therefore, we preliminarily concluded that LMO2 deletion might play a role in BL identification. None of the previous studies found a correlation between LMO2 and MYC rearrangements; however, a recent study not only found low expression of LMO2 in BL, but also 100% detected MYC rearrangement. This study suggested that the loss of LMO2 might be a good predictor of the presence of MYC [24]. In MYC rearrangement in BL, the exact mechanism that leads to LMO2 downregulation was not clear; however, Natkunam et al. [14] found that LMO2 protein is highly expressed at the mRNA level in the Ramos cell line, whereas the expression was indeed low at the immunohistochemical protein level. This suggested that LMO2 might be regulated at the posttranscriptional level in BL. These findings suggested that LMO2 protein can be used as an alternative marker for detecting MYC translocation in BL and might have application value in the differential diagnosis of other high-grade lymphomas.

CD38 is a transmembrane glycoprotein and in addition to marking mature plasma cells and plasma cell tumors, is a marker for germinal center B-cells [15]. Previous studies have found that CD38 is positively expressed in BL, but no in-depth studies have been conducted to verify this [27]. The expression of CD38 in HGBL,NOS and Burkitt-like lymphoma with the 11q aberration was even more limited. In our study, the positive rates of CD38 in BL, HGBL,NOS and Burkitt-like lymphoma with the 11q aberration were 98.67 (74/75), 33.3 (4/12), and 100% (3/3), respectively. There was a statistically significant difference in the positive expression rate of CD38 in BL and HGBL,NOS (P < .01). The sensitivity and specificity of the positive expression of CD38 protein were 98.67 and 66.67%, respectively, and AUC of diagnostic efficiency was 0.827. Previous studies have found that CD38, as LMO2, can be considered as a valuable diagnostic marker for identifying BL/DLBCL [28]. At the immunohistochemical level, it has been found that CD38 and CD44 can be used to distinguish between MYC-positive and MYC-negative lymphomas [29]. In the absence of cytogenetic analysis, it was very difficult to identify MYC-R in high-grade B-cell lymphomas. In practice, classical morphologic features of starry sky with medium-sized lymphocytes, typical Ki-67 hyperproliferation/CD10+/bcl-6+/bcl-2-, and recently identified CD38+/CD44−/TCL-1+ can predict a great possibility of MYC-R [29‐40]. All of these suggest that CD38 has a specific value in the differential diagnosis of BL.

Recent studies have suggested that the expression of MYC protein in aggressive B-cell lymphoma can effectively predict a poor prognosis [33‐40]. MYC protein is significantly correlated with MYC rearrangement, but the expression of MYC protein is not necessarily the result of MYC rearrangement [41]. In our study, the cutoff value of the positive expression of MYC protein was defined as 80% because of the differential diagnosis of BL, which was not consistent with previous studies [19, 35]. There was a significant difference in the expression of MYC protein in BL and HGBL,NOS (P < .01; Table 2) because of the defined MYC protein cutoff value. Because of the impact on the statistics of the defined MYC protein–positive cutoff value, we excluded MYC in subsequent statistical analyses. Finally, the combination of LMO2-negative and CD38-positive was used in the differential diagnosis of BL in our study. The sensitivity and specificity of LMO2 negative and CD38 positive were 97.33 and 100%, respectively, and AUC of diagnostic efficiency was 0.998, which was larger than AUC of those only LMO2 negative (0.993) or only CD38 positive (0.827). Further analysis found that AUC of the combination of LMO2-negative and CD38-positive was statistically different (P = .015) from that of CD38 positive, and there was no statistical difference (P = .328) in AUC of the combination of LMO2-negative and CD38-positive compared with that of LMO2 negative. The same results can be obtained by integrated discrimination improvement index analysis. The reasons for this were that first, the sample size of our study was relatively small, and in a follow-up study we will need to increase the sample size to reduce sampling error. Second, Burkitt-like lymphoma with the 11q aberration was rare; therefore, only three cases were included in our study and the expressions of LMO2 and CD38 in those cases were consistent with that in BL. We did not include these three cases in the statistical analyses shown in Table 2. Further analyses with a larger sample must be conducted to assess whether the expressions of LMO2 and CD38 in Burkitt-like lymphoma with the 11q aberration is completely identical to those in BL.

There was another limitation in our study. The best detection method for the 11q aberration is the chip technology of comparative genomic hybridization using oligonucleotide microarrays. In this study, ATM detected by FISH was located in 11q22. There were eight cases in the literature that reported amplification of this gene region [10, 11, 23], which was similar to the results of our study; therefore, the detection of this gene indirectly proved the 11q aberration.

We believe that the assessment of LMO2 and CD38 protein expression can improve the accuracy of the pathological diagnosis of BL. At the same time, with the use of routine immune indices, such as immunohistochemical markers CD10 and BCL2, the combination of LMO2-negative and CD38-positive results can be directly applied to the routine assessment of BL in clinical practice.