Updates in Immunohistochemistry for Hematopoietic and Lymphoid Neoplasms

As stated above, Zhang and Aguilera¹  previously reviewed the traditional IHC workup for small B-cell lymphomas (Table 1) and introduced some novel markers to assist in that workup (Table 2), especially when immunophenotypes present with unclassical patterns. These new stains may also be helpful in the ever-increasing use of minimally invasive biopsy techniques challenging the practicing pathologist with scant tissue that may be distorted by sampling and crush artifact.

A tissue review⁸  of 517 small B-cell lymphomas used a panel of markers that included LEF1, MNDA, CD27, IRTA1, and Stathmin/STMN1. LEF1 was consistently positive in cases of chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) but did show expression in 2 of 22 marginal zone lymphoma (MZL) cases. MNDA demonstrated expression most significantly in MZL cases. However, significant expression was also seen in CLL/SLL (76%), lymphoplasmacytic lymphoma (LPL) (27.7%), and follicular lymphoma (FL) (14%), possibly limiting its usefulness in differentiating these entities in small biopsy samples. Stathmin/STMN1 expression was significant in the evaluation of FL compared to splenic MZL, LPL, and other types of MZL; however, there was significant expression of Stathmin/STMN1 (57.1%) in cases of CLL/SLL, which was restricted to paraimmunoblastic cells. IRTA1 expression was observed in 35.7% of nodal MZL cases and in 41.2% of extranodal MZL cases, with limited expression in other small B-cell lymphomas. From these findings, the authors suggest that using LEF1, MNDA, and IRTA1 may be useful in the workup of atypical immunophenotypes of diffuse small B-cell infiltrates, and suggest using Stathmin/STMN1, MNDA, and IRTA1 in the workup of atypical immunophenotypes of nodular small B-cell infiltrates.

Another recent article⁹  evaluated expression of IRTA1 and MNDA in 127 lymphomas comprising 80 MZLs and 47 other small B-cell lymphomas; however, IRTA1 was evaluated by using an RNA ISH assay. This study demonstrated IRTA1 positivity in 42% (31 of 74) of all MZLs and only 1 FL. All cases of LPL, mantle cell lymphoma (MCL), and CLL were negative for IRTA1. MNDA was positive in 64% (51 of 79) of all MZLs, 37% (3 of 8) of LPLs, 21% (3 of 14) of FLs, 53% (8 of 15) of CLLs, and 78% of (7 of 9) MCLs. These findings further advocate for the use of IRTA1 and MNDA, especially in the context of a diagnosis of MZL.

LEF1 and SOX11 expression were evaluated in 354 cases of small B-cell lymphomas.¹⁰  In this study, 98% (126 of 129) of CLL/SLL cases were positive for LEF1, while only 6% (2 of 33) of MCLs were positive. All examined cases of MZL (142) and LPL (50) were negative for LEF1. SOX11 expression was positive in 82% (27 of 33) of cases of MCL but only in 5% (2 of 42) of MZLs. Overall, these findings support the use of LEF1 and SOX11 in the workup of small B-cell lymphomas.

LEF1 expression was also recently reviewed¹¹  in a series of 117 Hodgkin lymphomas, including 24 cases of Richter transformation of CLL/SLL to classic Hodgkin lymphoma (CHL), 66 cases of CHL, and 27 cases of nodular lymphocyte-predominant Hodgkin lymphoma (NLPHL). LEF1 was expressed in most cases that involved both Richter transformation of CLL/SLL to CHL (80%, 19 of 24) and CHLs (88%, 58 of 66), suggesting that LEF1 expression should not be used for evidence for Richter transformation of CLL/SLL. Only 44% (12 of 27) of NLPHL cases showed expression of LEF1.

However, several previously unnoted potential pitfalls of these new markers have also been recently described. MNDA is expressed in primary follicles, and this morphology may overlap with MZLs,¹²  especially in limited biopsies or atypical presentations. Atypical CLL/SLL and/or cases with increased prolymphocytes tend to show downregulation of LEF1 expression,¹³  further limiting its usefulness in difficult-to-classify cases. Finally, in a review¹⁴  of the expression of LEF1 in non-CLL/SLL CD5⁺ B-cell lymphomas with a total of 18 cases, including MZL, LPL, and FL, 50% (1 of 2) of CD5⁺ FLs were positive for LEF1, suggesting possible expression of LEF1 in a subset of FLs; however, 100% (13 of 13) of CD5⁺ MZLs were negative for LEF1, and 100% (3 of 3) of CD5⁺ LPLs were negative for LEF1.

Accumulating genetic evidence suggests that not only cyclin D1 but also other cyclin D family members (CCND2 and CCND3) might play an oncogenic role in the pathobiology of MCL. Approximately 10% to 15% of MCLs lack the typical t(11;14) but have similar clinical, molecular, and morphologic features. The so-called cyclin D1–negative MCLs display increased expression of either cyclin D2 or cyclin D3. Recurrent translocations involving CCND2 and the immunoglobulin heavy and light chains have been identified in such cases and are associated with high cyclin D2 expression levels. In addition, cryptic insertions that result in the integration of immunoglobulin light-chain enhancers near CCND2 or CCND3 have also been observed as another genetic mechanism for cyclin D2 or cyclin D3 overexpression.¹⁵ 

Cyclin D3 is a D-type cyclin that functions in the transition of the cell cycle from G1- to S-phase and is important in early B-cell development in the bone marrow and the proliferation of B cells in germinal center (GC) dark zones.¹⁶  Mutations in the CCND3 gene have been reported in multiple B-cell lymphomas. Cyclin D3 expression was evaluated in a review¹⁷  of splenic B-cell lymphoma and showed expression in 24 of 33 splenic diffuse red pulp small B-cell lymphomas (SDRPBLs). Cyclin D3 was only rarely expressed in the other splenic B-cell lymphomas evaluated (0%, 0 of 40 CLLs; 14%, 1 of 7 hairy cell leukemias [HCLs]; 6%, 2 of 35 MCLs; and 1%, 1 of 74 splenic marginal zone lymphomas [SMZLs]).

Cortactin is a cytoskeletal protein and functions in the processes of cell motility, adhesion, membrane trafficking, and extracellular matrix degradation. Overexpression of cortactin has been described in carcinomas, sarcomas, and astrocytic tumors. A recent study¹⁸  further evaluated cortactin expression in 131 cases of non-Hodgkin B-cell lymphomas and leukemias. Positive expression was observed in 82% (14 of 17) of CLL cases, 0% (0 of 16) of MCL cases, 8% (2 of 25) of FL cases, 83% (5 of 6) of MZL cases, 41% (7 of 17) of extranodal marginal zone lymphoma cases, 14% (1 of 7) of SMZL cases, 0% (0 of 3) of SDRPBL cases, 100% (10 of 10) of HCL cases, 75% (12 of 16) of GC B-cell type diffuse large B-cell lymphoma (DLBCL) cases, and 93% (13 of 14) of non–GC B-cell type DLBCL cases. Overall, cortactin appears to be commonly expressed in CLL, DLBCL, and HCL, and may be another marker useful in the differential diagnosis of CLL versus MCL.¹⁸ 

Aberrant MAL expression in primary mediastinal large B-cell lymphoma (PMBL) was initially reported in 1999¹⁹; however, evaluation of MAL expression by commercially available IHC stain is suboptimal and underused. Recently, a study²⁰  of 15 cases of PMBL using both MAL IHC stain and MAL RNA ISH showed 67% (10 of 15) of PMBL cases with expression of MAL by both IHC stain and RNA ISH, with an additional 3 cases detected by RNA ISH only and 1 case detected by IHC stain only. These findings suggest incorporating MAL RNA ISH in the workup of large B-cell lymphomas of the mediastinum.²⁰ 

J-chain is a small protein that facilitates multimerization of the immunoglobulin (Ig) A and IgM antibodies into dimers and pentads, respectively.²¹  In reactive tissues, weak to variably strong expression of J-chain is seen in reactive GCs as well as in reactive plasma cells. In a study evaluating J-chain in the differential diagnosis of Hodgkin lymphoma,²²  J-chain showed strong expression in lymphocyte-predominant (LP) cells of NLPHL in 100% (20 of 20) of cases and showed variable staining in 50% (2 of 4) of T-cell–rich large B-cell lymphomas (THRLBCLs) and 67% (10 of 15) of PMBLs. J-chain was negative in all 43 evaluated cases of CHL, including 23 cases of nodular sclerosis CHL, 9 cases of lymphocyte-rich CHL, 7 cases of mixed cellularity CHL, and 4 cases of lymphocyte-depleted CHL. In addition to positivity in LP cells, J-chain is expressed in nonneoplastic GC B cells and plasma cells and because of that reason, they cannot be used in isolation in the discrimination between the progressive transformation of GC and NLPHL.²² 

MEF2B is an activating transcription factor that participates in the regulation of the BCL6 proto-oncogene and shows a similar pattern of expression to BCL6 in GC B cells. MEF2B expression was also evaluated in the differential diagnosis of Hodgkin lymphoma and, like J-chain, showed strong expression in LP cells in 100% (20 of 20) of NLPHL cases but was more commonly expressed in neoplastic lymphocytes of PMBL (93%, 14 of 15 cases) and THRLBCL (100%, 4 of 4 cases).²²  MEF2B was negative in all 43 evaluated cases of CHL (see subsets in the preceeding paragraph).

GATA3, a transcription factor gene in T-cell maturation, may also be useful in the workup of large B-cell lymphomas. In normal lymph nodes, GATA3 shows staining in paracortical and GC T cells. In the workup of large B-cell lymphomas, GATA3 showed positive staining in 80% (39 of 49) of CHL variants, 0% (0 of 17) of NLPHLs, 0% (0 of 72) of DLBCLs, 75% (3 of 4) of PMBLs, 0% (0 of 2) of Epstein-Barr virus–positive DLBCLs, and 100% (1 of 1) of gray zone lymphomas (GZLs).²³ 

Another review²⁴  of mediastinal lymphomas evaluated the expression of p63 and GATA3 in the differential diagnosis of CHL and PMBL. Expression of p63 was observed in 94% (15 of 16) of PMBL cases and 15% (2 of 13) of CHL cases. Expression of GATA3 was observed in 77% (10 of 13) of CHL cases and 0% (0 of 16) of PMBL cases. These findings are similar to those of Kezlarian et al,²³  further supporting the use of GATA3 in challenging mediastinal lymphoma diagnoses and presenting another potential use for p63.

IRF8 is a transcription factor in the determination of B-cell differentiation and induction of tolerance with expression in B cells as well as monocytic leukemias. IRF8 expression was evaluated in a study²⁵  of 74 cases of CHL, 7 cases of NLPHL, 15 cases of ALK-negative anaplastic large cell lymphoma (ALCL), and 4 cases of ALK-positive ALCL to evaluate whether IRF8 expression may be helpful in lymphoma with loss of B-cell– or T-cell–defining lineage markers. One hundred percent (7 of 7) of NLPHLs expressed IRF8, and 85% (61 of 72) of CHLs expressed IRF8. All cases of ALCL were negative. IRF8 showed expression in 60% (6 of 10) of PAX5-negative CHLs. These findings support the use of IRF8 as a B-cell marker in the differential diagnosis of CHL and ALK-negative ALCL.²⁵ 

A panel consisting of CD83, fascin, and CD23 may also be useful in the differential diagnosis of mediastinal lymphomas.²⁶  CD83 is a glycoprotein expressed on the membrane of dendritic cells, activated monocytes or macrophages, and subsets of B cells. The intermediate filament protein, fascin, is typically expressed in neurons, dendritic cells, macrophages, and endothelial cells and is known to be expressed in Hodgkin Reed-Sternberg cells of CHL, ALCL, and a few cases of DLBCL. CD23 is a receptor for IgE and is known to show expression in PMBL. When this panel of 3 stains was used in the differential diagnosis of CHL, GZL, and PMBL, CHL showed expression of CD83 in 100% (22 of 22) of cases, fascin in 100% (22 of 22) of cases, and CD23 in 9% (2 of 22) of cases. PMBL showed expression of CD83 in 41% (9 of 22) of cases, fascin in 32% (7 of 22) of cases, and CD23 in 95% (21 of 22) of cases. GZL showed mixed expression results of CD83 in 89% (16 of 18) of cases, fascin in 86% (18 of 21) of cases, and CD23 in 67% (12 of 18) of cases.²⁶  These findings support the usage of a 3-marker panel as a useful additional tool in the workup of mediastinal lymphomas.

EZH2 is an epigenetic regulator protein that is expressed in nonneoplastic proliferating cells and may be irregularly expressed in neoplasia. Small molecular inhibitors of EZH2 have potential therapeutic benefits. A recent study²⁷  showed positive expression for EZH2 in most cases of large B-cell lymphomas, including NLPHL, CHL, THRLBCL, GZL, DLBCL, Burkitt lymphoma, and double-hit lymphomas, suggesting that most aggressive B-cell lymphomas have expression of EZH2 and could potentially benefit from small-molecule inhibitors of the EZH2 oncogenic pathway.

Table 3 summarizes the above-described new IHC markers.

Diagnostic evaluation of T-cell lymphomas is difficult and generally requires an extensive list of IHC stains (Table 4). A review of recent literature shows evaluation of several interesting new markers to assist in this evaluation.

CD28 has a pivotal role in T-cell activation, and its expression is strictly regulated in normal T cells. Gain-of-function genetic alterations involving CD28 have been frequently observed in adult T-cell leukemia/lymphoma (ATLL). A recent study²⁸  examined semiquantitative expression of CD28 and its ligands CD80 and CD86 by immunohistochemistry in 120 ATLL cases. The study found that CD28 was overexpressed in 55 cases (46%), and CD80 or CD86 (CD80/CD86) was infrequently overexpressed in 12 (11%). The most striking result of the study was that CD28 overexpressers showed a significantly poorer overall survival than non-overexpressers (P = .001), which suggested that patients with ATLL and CD28 overexpression have a more aggressive clinical course, with tumors that are more refractory to treatment with multidrug chemotherapy.

Chemokines and their receptors orchestrate cell migration and homing in the body and are subclassified by the N-terminal conserved cysteine motifs.²⁹  Mogamulizumab targets the extracellular N-terminal domain of CCR4, which is expressed in most ATLL cases. A recent study³⁰  showed that CCR4 C-terminal gain-of-function mutations were frequent in ATLL cases, and a subgroup with these mutations who were treated without allogenic hematopoietic stem cell transplant (HSCT) and with mogamulizumab-containing [HSCT⁻ and mogamulizumab⁺] regimens had a superior survival rate. Another study³¹  examined the immunohistochemical expression of both CCR4 N-terminus (CCR4-N-IHC) and CCR4 C-terminus (CCR4-C-IHC) and analyzed the clinicopathologic significance of CCR4 expression of the respective termini in 92 ATLL cases. They found that CCR4-C-IHC, but not CCR4-N-IHC, showed a negative correlation with the presence of CCR4 mutations. In addition, CCR4-C-IHC was found to be a useful marker for high-sensitivity screening of the CCR4 mutational status (87%). Overall, negative staining for CCR4-C-IHC may predict CCR4 mutations and mogamulizumab efficacy in ATLL.³¹ 

Extranodal natural killer/T-cell lymphoma (ENKTCL), nasal type is a high-grade malignancy and is prevalent in Asia and South and Central America. A recent study³²  aimed to investigate the expression profiles of cell cycle–related proteins in ENKTCL by using a cell cycle antibody array and showed a significant negative correlation between p-ATM expression and survival in 23 patients with high expression compared with 26 patients with low expression (P = .011). Kaplan-Meier survival analysis also showed a significant negative correlation between CHK2 expression and survival in 24 patients with high expression compared with 25 patients with low expression (P = .025). This study demonstrated that abnormalities in the ATM/CHK2 pathway may play a crucial role in the oncogenesis and chemoradiotherapy resistance of nasal ENKTCL.

BCL11B plays an essential role in T-cell lineage commitment and maturation. A recent study³³  evaluated the expression of BCL11B protein in a cohort of 115 T-cell acute lymphoblastic leukemia/lymphoblastic lymphoma (T-ALL/LBL) cases, including early T-cell precursor acute lymphoblastic leukemia/lymphoma (ETP-ALL/LBL) (n = 29; 25%), early T-ALL/LBL (n = 26; 23%), thymic T-ALL/LBL (n = 42; 37%), and mature T-ALL/LBL (n = 18; 16%). The study found that most ETP-ALL cases (83%) showed negative BCL11B expression, whereas most (84%) non–ETP-ALL/LBL cases (84%) were positive for BCL11B. Additionally, patients with ETP-ALL and positive BCL11B expression had a better overall survival than those with negative BCL11B expression (P = .009). Overall, the BCL11B expression status may help distinguish ETP-ALL/LBL from other subtypes of T-ALL/LBL, and BCL11B may serve as a potential prognostic marker in patients with ETP-ALL.³³ 

LIM Domain Only 2 (LMO2) is a protein involved in scaffolding of transcription factors (including the GATA family) necessary for hematopoiesis and angiogenesis.³⁴  T-lymphoblastic lymphoma (T-LBL) is an aggressive neoplasm of T-lymphoid precursors, which can rarely occur in the setting of myeloid/lymphoid neoplasms with eosinophilia (M/LNs-Eo),³⁵  but can be difficult to identify. LMO2 is overexpressed in most T-LBLs, but not in immature TdT-positive T cells in the thymus or indolent T-lymphoblastic proliferations. In a study³⁶  of 11 cases of T-LBLs associated with M/LNs-Eo, 9 of 11 cases were negative for LMO2, suggesting a different molecular pathogenesis of T-LBLs associated with M/LNs-Eo. This finding suggests that T-LBLs that are negative for LMO2 may warrant investigation for genetics associated with M/LNs-Eo.³⁶  LMO2 expression is also found in B-cell lymphomas derived from GC lymphocytes, including follicular, Burkitt, and DLBCLs, as well as in LP Hodgkin lymphoma.²² Table 5 summarizes the above-described new IHC markers.

The differential diagnosis of normal monocytes, abnormal monocytes, promonocytes, and blasts is essential for the diagnosis of myelomonocytic neoplasms; however, this task is reliant on accurate and experienced morphologists and is therefore often subjective. The common lack of CD34 expression in blasts of acute monocytic/monoblastic leukemias further contributes to diagnostic difficulty in this group of myeloid neoplasms. Recently, IRF8 was evaluated as a marker for identification of monoblasts.³⁷  IRF8 is a specific monocyte and dendritic cell progenitor transcription factor. IRF8 expression in bone marrow cores of patients with known acute monocytic/monoblastic leukemia showed high correlation (R = 0.94). Additionally, expression of IRF8 was not observed significantly in other subtypes of acute myeloid leukemia (AML) (R = 0.56). IRF8 expression may also be useful in the evaluation for residual disease after therapy; however, less concordance was identified in the evaluation of blast counts in chronic myelomonocytic leukemia.³⁷  IRF8 may also be expressed in blastic plasmacytoid dendritic cell neoplasms, mature plasmacytoid dendritic cell proliferations, and reactive plasmacytoid dendritic cell proliferations.³⁸ 

Glucose transporter 1 (GLUT1) is a membrane transporter for cellular glucose uptake that is highly expressed in red blood cells. Expression of GLUT1 was evaluated in both benign and neoplastic erythroid populations in bone marrow biopsies.³⁹  GLUT1 expression was sensitive and specific for benign and reactive erythroid precursors, with only some cytoplasmic weak expression in megakaryocytes in 10 cases of nonneoplastic bone marrow biopsies. GLUT1 showed strong to intermediate cytoplasmic and membranous staining in proerythroblasts of erythroid malignancies in 100% (6 of 6) of cases of pure erythroid leukemia and 100% (2 of 2) of cases of therapy-related AML with erythroid differentiation. GLUT1 expression was also demonstrated in 80% (4 of 5) of evaluated cases of B-lymphoblastic leukemia/lymphoma, but staining was less intense than that of proerythroblasts.³⁹ 

In the 5th edition of the WHO classification of hematolymphoid tumors,³  the diagnostic criteria for systemic mastocytosis (SM) have been modified to include CD30 expression—in addition to CD25 and CD2—by immunohistochemistry as one of the minor diagnostic criteria. Morgado et al⁴⁰  reported that CD30 is expressed in most SM cases and would improve diagnostic accuracy when combined with CD25 expression compared to using CD25 alone. Except for mast cell leukemia, all other subtypes of SM, including indolent, smoldering, aggressive, and SM with an associated hematologic neoplasm (SM-AHN), show a similar level of CD30 expression. Currently, there is no significant association of CD30 expression with clinical presentation or disease prognosis in SM.

CD30 is a very useful diagnostic marker for well-differentiated SM. In this rare variant, bone marrow–infiltrating neoplastic mast cells are mature appearing and usually lack CD25 and/or CD2 expression but are positive for CD30.⁴¹ 

SM is traditionally treated with cytoreductive therapy to control severe refractory symptoms or organ dysfunction. For patients with CD30 expressing SM, an anti-CD30 antibody drug, such as brentuximab vedotin, provides a novel treatment solution.⁴² 

In addition to including CD30 as a minor criterion, another refinement in SM diagnosis is to include any activating KIT mutation in addition to codon 816 only as a minor diagnostic criterion. Although CD117 (KIT) expression by IHC stain is commonly used to identify immature and atypical mast cells in tissue sections, there is not enough evidence to support its correlation with the detection of KIT mutations by molecular technology.⁴³ Table 6 summarizes the above-described new IHC markers.

Recently, our institution instituted a proven care work group for rapid diagnosis and treatment of AMLs. In their 2022 update, European LeukemiaNet recommend that, for best treatment decisions, results for FLT3, IDH1, IDH2, and NPM1 be received in 3 to 5 days.⁴⁴  Generally, our molecular laboratory has full hematology gene panel results within 5 to 7 days, but, depending on the day of the biopsy and batched runs, some molecular results may not be available for 10 days. Rapid FLT3 mutation analysis is performed separately, and results are usually received within the recommended time frame. As part of this work group, we performed a literature review for recent articles describing evidence of correlation between IHC expression (a rapid and comparably more economical test) and molecular mutations in AML, with limited results.

A study⁴⁵  of 71 Chinese adult de novo AML cases evaluated with IHC stain for cytoplasmic delocalization of NPM1 showed 86.7% sensitivity and 96.4% specificity for NPM1 mutation. A similar study⁴⁶  showed 19 of 21 patients with NPM1 gene mutations with positive aberrant cytoplasmic localization of NPM1 staining.

Two recent studies^47,48  evaluated the correlation between p53 expression and TP53 mutations. One study⁴⁷  showed concordant expression and mutation in 67% (24 of 36) of cases with a cutoff of 7% p53 expression or greater. However, the other small study⁴⁸  showed p53 staining had poor correlation with type of disease and mutational status except in instances of clonal cytopenia of undetermined significance, in which staining was never greater than 1%.

One study⁴⁹  evaluated the use of IDH1 p.R132H mutation-specific antibody IHC stain in diagnosis and follow-up of myeloid neoplasms. At diagnosis, all neoplasms with IDH1 p.R132H mutations showed positive staining. These cases included AML (30 cases), myelodysplastic syndrome (10 cases), myelodysplastic syndrome/myeloproliferative neoplasm (4 cases), and myeloproliferative neoplasm (5 cases), with positive expression in immature and mature myeloid cells and negative expression in erythroid precursors, lymphocytes, endothelial cells, and osteoblasts. All evaluated IDH1 wild-type and mutant non-p.R132H samples were also negative by IHC stain. The study also evaluated follow-up biopsies with prior positive IDH1 mutations. Interestingly, 40% (8 of 20) of polymerase chain reaction–negative cases in the follow-up samples were positive by IHC stain.⁴⁹ 

Additional recent studies evaluating the usefulness of IDH1/2 staining in myeloid neoplasms were not identified, despite the availability of monoclonal antibodies and use of stains for diagnosis in other neoplasms. Similarly, recent studies for IHC expression of FLT3 are also lacking.

The practicing hematopathologist requires knowledge of the ever-expanding repertoire of immunohistochemistry for the diagnosis and treatment of hematolymphoid neoplasia. New markers presented in this article help to complete our understanding of disease, diagnosis, and management.