Reproducing the molecular subclassification of peripheral T-cell lymphoma-NOS by immunohistochemistry

Abstract

Peripheral T-cell lymphoma (PTCL) is a heterogeneous group of mature T-cell malignancies; approximately one-third of cases are designated as PTCL-not otherwise specified (PTCL-NOS). Using gene-expression profiling (GEP), we have previously defined 2 major molecular subtypes of PTCL-NOS, PTCL-GATA3 and PTCL-TBX21, which have distinct biological differences in oncogenic pathways and prognosis. In the current study, we generated an immunohistochemistry (IHC) algorithm to identify the 2 subtypes in paraffin tissue using antibodies to key transcriptional factors (GATA3 and TBX21) and their target proteins (CCR4 and CXCR3). In a training cohort of 49 cases of PTCL-NOS with corresponding GEP data, the 2 subtypes identified by the IHC algorithm matched the GEP results with high sensitivity (85%) and showed a significant difference in overall survival (OS) (P = .03). The IHC algorithm classification showed high interobserver reproducibility among pathologists and was validated in a second PTCL-NOS cohort (n = 124), where a significant difference in OS between the PTCL-GATA3 and PTCL-TBX21 subtypes was confirmed (P = .003). In multivariate analysis, a high International Prognostic Index score (3-5) and the PTCL-GATA3 subtype identified by IHC were independent adverse predictors of OS (P = .0015). Additionally, the 2 IHC-defined subtypes were significantly associated with distinct morphological features (P < .001), and there was a significant enrichment of an activated CD8+ cytotoxic phenotype in the PTCL-TBX21 subtype (P = .03). The IHC algorithm will aid in identifying the 2 subtypes in clinical practice, which will aid the future clinical management of patients and facilitate risk stratification in clinical trials.

Graphical abstract

Figure 1.

Selected IHC panel included in the diagnostic algorithm. (A) Positive immunostains for GATA3 (nuclear) and its corresponding target protein CCR4 (membranous) in a PTCL-GATA3 case. This case is negative for TBX21 and CXCR3. (B) Positive immunostains for TBX21 (nuclear) and its target protein CXCR3 (membranous) in a PTCL-TBX21 case. This case is negative for GATA3 and CCR4. (A-B) Original magnification ×600.

Figure 2.

Antibody selection for the IHC algorithm. (A) Correlation between mRNA and IHC expression of GATA3, CCR4, TBX21, and CXCR3 in the training cohort. Positive correlation was observed between mRNA and the corresponding protein expression by IHC. The percentage of positivity was estimated in discrete increments of 10% and correlated with the mRNA gene expression obtained using HG U133 Plus 2 arrays. (B) Distribution of immunostain positivity in the corresponding GEP-defined molecular subtypes. Top panels, The PTCL-GATA3 subtype showed higher positivity for GATA3 and CCR4 immunostains compared with the PTCL-TBX21 subtype, with 50% being the best cutoff to separate the 2 subtypes. Bottom panels, The PTCL-TBX21 subtype showed higher positivity for TBX21 and CXCR3 compared with the PTCL-GATA3 subtype, with most PTCL-TBX21 cases showing ≥20% positivity for TBX21 and CXCR3, and most PTCL-GATA3 cases showing <20% positivity. These cutoffs were subsequently used for the IHC algorithm. (C) Heatmap representation of the GATA3, CCR4, TBX21, and CXCR3 immunostains in GEP-defined subtypes and IHC algorithm classification in training cohort cases. The frequency of the individual immunostain positivity in the training cohort in the PTCL subtypes is shown on the right.

Figure 3.

Decision tree for IHC subclassification. (A) The IHC algorithm was generated using the 40 PTCL-NOS cases that were classified by the GEP as PTCL-GATA3 and PTCL-TBX21. The 9 cases unclassifiable by GEP were not included in the training of the algorithm. (B) Comparison of the IHC and GEP-defined PTCL subtypes resulted in an accurate classification of 85% of cases, with 5% remaining unclassifiable.

Figure 4.

OS of the PTCL subtypes using IHC classification. (A) Significant differences in OS were observed between the PTCL-GATA3 and PTCL-TBX21 subtypes using the IHC algorithm and are comparable to the GEP classification (dotted line) in the training cohort. (B) OS curves in the validation cohort using the IHC algorithm. PTCL-GATA3 was significantly associated with worse prognosis in both cohorts. (C) Distribution (percentage) of IHC-defined subtypes was similar in both cohorts.

Figure 5.

Representative morphologic patterns in PTCL subtypes. PTCL-GATA3 was significantly associated with a monomorphic, monotonous, tumor cell–rich morphology with a minimal inflammatory background characterized by: (A) small-intermediate cells with abundant cytoplasm (pattern 1) or (B) clusters or sheets of large cells (pattern 2). In contrast, PTCL-TBX21 was significantly associated with a polymorphous morphology, characterized by (C) neoplastic cells interspersed in a mixed inflammatory background (pattern 3) or (D) the classic lymphohistiocytic (Lennert) lymphoma morphology composed of small tumor cells within clusters of epithelioid histiocytes (pattern 4). (E) Distribution (percentage) of morphological patterns in the 2 PTCL subtypes with monomorphic patterns associated with PTCL-GATA3, whereas polymorphic patterns associated with PTCL-TBX21 (P < .001). (F) The monomorphic patterns (1 and 2) and the Lennert lymphoma pattern (pattern 4) were associated with worse OS, when both cohorts were combined. (A-D) H&E stain; original magnification ×200; inset magnification ×400.

Figure 6.

Association of morphologic patterns and additional immunophenotypic features in PTCL subtypes classified by the IHC algorithm. (A) Heatmap representation of the morphologic patterns (1-4) and CD4, CD8, TIA-1, granzyme B, and CD30 immunostain and EBER in situ hybridization positivity in PTCL subtypes in training and validation cohort studied on TMA (n = 57). Each case is represented by a column, with CD4/CD8 reported as positive and negative and cytotoxic markers/CD30 positivity graded as 0-3. EBERs was evaluated semiquantitatively calculating the average number of positive cells per field (f): 0/f = 0; <1/f = 1+; 1 to 9/f = 2+; 10 to 50/f = 3+; and >50/f = 4+. Distribution of (B) CD4 and CD8 single-positive, double-positive, or double-negative immunophenotype; (C) cytotoxic markers; (D) CD30; and (E) EBER in the 2 PTCL subtypes. *There is a significant association of cytotoxic marker expression with PTCL-TBX21 subtype (P < .001).