Introduction
Myeloid nuclear differentiation antigen (MNDA) was initially detected in human promyelocytic cell lines [
1,
2] and used as a marker for myeloid neoplasms [
3,
4]. Subsequently, Miranda et al. identified a low level of MNDA expression in a population of normal mantle B lymphocytes and a subset of B-cell lymphomas [
5]. Since being identified as a useful marker for the recognition of marginal zone lymphoma (MZL) by Kanellis et al. in 2009 [
6], MNDA has been considered as a diagnostic marker for MZL recently in routine clinical practice [
7]. MNDA showed a potential power in distinguishing nodal marginal zone lymphoma (NMZL) from follicular lymphoma (FL), given its high frequency in MZL and very limited expression in FL [
6,
8]. However, a relatively higher expression of MNDA in FL was also reported by Wang et al. (21%) [
9] and its frequency in FL of different grades is unclear. In addition, although MNDA expression has been found in various B-cell non-Hodgkin lymphomas (B-NHLs), studies on MNDA expression in B-NHLs other than MZL are limited and inconclusive. High expression of MNDA (78-82%) in mantle cell lymphoma (MCL) was demonstrated [
6,
9], but a dramatically lower positivity (6.4%) was also reported by Metcalf et al [
8]. Moreover, the frequency of MNDA expression in cases of chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL, 12.9-76%), lymphoplasmacytic lymphoma (LPL, 27.7-83%) and diffuse large B-cell lymphoma (DLBCL, 3.3-45%) were even largely varied [
6,
8‐
11]. Therefore, the expression of MNDA in diverse B-NHLs and its clinicopathological significance are needed to be further evaluated in larger study cohorts.
Herein, we comprehensively analyzed the expression of MNDA with the largest series of MZLs and other common types of small B-NHLs by far and evaluated its differential diagnostic value. The expression and clinical relevance of MNDA in DLBCL were also investigated in our study.
Discussion
MNDA has been found to be expressed in a large proportion of MZLs and the existing literature has suggested MNDA as a useful marker for identifying MZL [
6,
8]. However, large series of studies on the expression of MNDA in various subtypes of B-NHL are lacking. Therefore, its potential utility in the differential diagnosis between MZL and other B-NHLs remains poorly understood and inconclusive, and its clinical relevance in DLBCL has never been elucidated by far. In this study, we thoroughly investigated the MDNA expression in a large cohort of diverse B-NHLs as well as RLH cases, which appeared to be the largest case series by far to the best of our knowledge.
Our results revealed frequent expression of MNDA in MZL, including both NMZL and EMZL, which was in line with some previous reports [
8,
9], while slightly lower than the data from Kanellis et al. (75% and 95%) [
6]. Consistent with the results from Kanellis et al. (100%) [
6], MNDA was expressed in all of the SMZL cases in our series, although the sample was quite limited. MNDA staining in tumor cells of MZL cases is usually diffusely distributed, at least partially, which was different from that in RLH cases, mostly showing restricted staining to the mantle/marginal zone of lymphoid follicles. Therefore, MNDA might play an important role in the differential diagnosis of MZL and RLH. However, we noticed that MNDA positivity was less commonly observed in the MZL cases with considerable plasmacytic differentiation. In detail, while the lymphoid component exhibited a certain degree of MNDA expression, the neoplastic cells with plasmacytic differentiation lost MNDA. Our finding indicated limited value of MNDA in the recognization of MZL with extensive plasmacytic differentiation, which had never been documented in the previous literature. Moreover, the histocytes/granulocytes strongly positive for MNDA in the background should not be interpreted as neoplastic cells in clinical practice, which should be kept in mind to avoid mistakes.
Extremely low frequency of MNDA expression in FL cases of our series was consistent with the results of Kanellis et al. (5%) [
6], but lower than those of Kivrak et al. (14%) and Wang et al. (21%) [
9,
10]. Moreover, whereas MNDA expression was slightly higher in grade 3 FL, it was rarely seen in grade 1 ∼ 2 FL cases (3.1%), which was in accord with the results of Metcalf et al. (4.3%). Notably, the expression of MNDA in documented FL cases is mostly restricted to the thin perifollicular areas showing marginal zone differentiation and only in a few scattered centrocyte-like cells in the follicles [
6,
8]. In contrast with the previous reports, MNDA-positive FL cases in our series demonstrated diffuse MNDA staining in neoplastic follicles and scattered in tumor cells in the inter-follicular areas [
6,
8].
In clinical practice, distinguishing between MZL and FL is usually straightforward, given their different cell origin and growth patterns as well as immunophenotypes. In some situations, however, MZLs might present a nodular growth pattern or occasional CD10 aberrant expression, mimicking low-grade FL or sometimes low-grade FL might demonstrate a diffuse pattern, posing challenges for differential diagnosis between the two diseases. Our results in the current study further confirmed the efficacy of MNDA protein in distinguishing MZL from FL, which was in accord with the previous studies [
6,
8]. However, Verdanet et al. demonstrated a similar frequency of MNDA expression in cutaneous MZL (4/13, 30.8%) and cutaneous FL (3/29,10.3%) (
p = 0.18), which made the effectiveness of MNDA in the differential diagnosis of these two unusual lymphoma variants ambiguous [
15].
Whereas Metcalf et al. showed a strikingly low expression of MNDA in MCL (6.4%) [
8], our results revealed a high frequency (79.6%), similar to the majority of studies (from 78 to 82%) [
6,
9]. Given a comparable or even more frequent MNDA expression in MCL compared with MZL, it is useless for MNDA in the distinguishment of these two types of B-NHL. The predilection of MNDA staining in the marginal zone/mantle B-cells and corresponding lymphomas warranted to be investigated further.
MNDA expression in CLL/SLL in the literature varied widely, ranging from 12.9 to 76% [
6,
8‐
10], and we got a rate of 44.8% in the current study. The differences among these results might be partially due to the limited number of cases and the criteria for MNDA positivity varied across each study. For the reason that the thresholds utilized in the existing literature presented minimal differences, with no mention of positive intensity. We have modified the scoring methods based on previous studies and adopted a semi-quantitative method that considered both the positive intensity and the percentage of positive tumor cells. Although the frequency of MNDA in CLL/SLL is statistically lower than that in MZL, it seems that its capability in the differentiation between these two entities is limited in routine surgical pathology practice, given its non-negligible expression in nearly half of CLL/SLL cases. Therefore, a panel of IHC markers is still necessary for their differential diagnosis.
MNDA expression in LPL was 25-83% according to the existing literature [
6,
8‐
11] and 25% in the current study, all with a small number of samples. Recently, Righi et al. showed a much lower frequency in LPLs (4% in the test set and 20% in the validation set) compared with 88% in SMZLs in bone marrow specimens [
11], indicating that MNDA might be helpful in the differentiation between LPL and SMZL, especially when the status of
MYD88 L265P mutation is unavailable. Given the low incidence of LPL and the widely varied frequencies of reported MNDA expression in LPL, more cases need to be accumulated to further explore the features of MNDA expression in this rare type of B-NHL.
Studies focusing on MNDA expression in DLBCL are lacking, and our cohort represented the largest study series investigating MNDA in DLBCL as we know. Our results showed a frequency of 29.6% in DLBCL, which was lower than the data from Kanellis et al. (34/75, 45%) [
6] whereas much higher than that from Metcalf et al. (2/61, 3.3%) [
8]. For the first time, we further investigated the clinicopathological relevance of MNDA in DLBCL, and our results revealed a higher frequency of MNDA in non-GCB subtype and in BCL2/MYC double-expression group. Interestingly, MNDA had a significant correlation with CD5 expression in DLBCL in the current series, which had never been documented in the literature. As we mentioned above, MNDA tended to be positive in the majority of MCLs and a considerable number of CLL/SLL cases, in both of which CD5 is consistently positive as a specific diagnostic marker [
16]. These observations suggested that aberrant MNDA overexpression might correlate with CD5 in B-NHLs, which is worthy of further study and the underlying mechanisms deserve further investigation.
Furthermore, the studies focused on the prognostic value of MNDA are limited, only a few in vitro studies suggested the prognostic potential of MNDA in CLL. For instance, Joshi et al. demonstrated that the expression of MNDA in CLL showed an inverse correlation with disease progression [
17] and Bottardi et al. further revealed the potential role of MNDA in tumor suppression by regulating the apoptosis of neoplastic cells in CLL [
18]. Unfortunately, more than 90% of the small B-NHLs in the current study were diagnosed within 5 years, the time is insufficient to achieve the expected target events necessary for survival analysis if prognostic follow-up has been conducted at this stage. Moreover, the number of MNDA-positive DLBCL cases for prognostic analysis is inadequate. Therefore, we will accumulate more cases and ensure a sufficient follow-up period for analyzing the prognostic value of MNDA in these lymphomas.
In summary, MNDA is highly expressed in MZL except those tumor cells with plasmacytic differentiation. MNDA might be an effective indicator for differentiating MZL from RLH and FL in clinical practice, but cannot be used to identify MZL from MCL and its value for distinguishing MZL from CLL/SLL is limited, thus a panel of markers is always indispensable in these circumstances. In DLBCL, the expression of MNDA has a correlation with immunophenotype and BCL2/MYC double-expression as well as CD5 expression.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.