Because Syk expression was found in a greater proportion of PTCLs than previously reported,¹⁰ we corroborated the IHC results using western blotting. Reactive splenic lymphocytes were sorted by flow cytometry into B-cell, αβ T-cell and γδ T-cell populations. B-cell lysates demonstrated a 72 kDa band corresponding to Syk, whereas T-cell lysates were negative (Figure 2a). Analysis of frozen tumor tissue lysates (Figure 2b) showed cases that were Syk-negative by IHC to be negative by western blot (PTCL-Us, two cases) as well. All four cases that were Syk-positive by IHC were positive by western blot (two AITLs, one ALK-negative ALCL and one PTCL-U). To evaluate the activation status of Syk in PTCLs, we probed western blots with phospho-specific anti-Syk (Tyr525/526); these tyrosine residues reside in the catalytic domain of Syk kinase and their phosphorylation is necessary for Syk activity.¹⁸ Syk was phosphorylated at these residues in 4/4 Syk-positive PTCLs tested (Figure 2b).

Because the lysates used for western blot might contain Syk derived from non-tumor cells as well as PTCLs, we also evaluated Syk expression by flow cytometry. Reactive T cells from peripheral blood, lymph node and spleen were negative for Syk, whereas reactive B cells were positive (not shown). By using appropriate gating strategies in PTCLs with an aberrant T-cell phenotype, we could assess Syk expression specifically in the neoplastic T cells in four cases. The tumor cells demonstrated Syk expression in three cases (Figure 2c; see also Figure 1c). One case of hepatosplenic T-cell lymphoma was Syk-negative by flow cytometry (Figure 2d) as well as IHC (Figure 1f).

To determine the relationship between Syk overexpression in our series and the t(5;9)(q33;q22) SYK/ITK translocation, we evaluated cases using D-FISH probes for SYK and ITK. Despite appropriate fusion signals in control tissue with the translocation (not shown), no evidence of t(5;9)(q33;q22) was identified in 86 informative study cases of PTCL (84 of which were positive for Syk by IHC). These 86 cases included 23 AITLs, 38 PTCL-Us, 13 ALCLs and 12 other cases, a distribution similar to that in the overall study set. Additional copies of SYK (3–6 signals) were identified in only four cases, including two ALK-negative ALCLs (both Syk protein-positive by IHC) and two PTCL-Us (one Syk-positive and one Syk-negative). The FISH probes used did not allow distinction between gene amplification and polysomy as the cause for additional SYK signals. None of the cases studied had the characteristic features of PTCLs with follicular involvement described by de Leval et al.,³ which were seen in 3/5 previously reported cases with SYK/ITK translocation.² Based on our findings, translocations or additional copies of SYK do not appear to be the mechanisms leading to Syk protein overexpression in most Syk-positive PTCLs.

Syk has been suggested as a potential therapeutic target for PTCL by Mahadevan et al.,¹⁹ but previous data on Syk expression in T-cell lymphomas are limited and somewhat conflicting. A small study found Syk in only 2/19 PTCLs by IHC, including 1/8 PTCL-U and 1/1 mycosis fungoid (weak staining).¹⁰ The higher positivity rate found by us might be due to differences in the antibodies used, or due to unknown differences in the patient populations studied. Syk expression and Syk kinase activity have been reported to be decreased in lysates of cutaneous T-cell lymphoma cells isolated from peripheral blood (n=4),¹⁸ a source not evaluated in our study. This difference in site might account for our finding that Syk was expressed in 4/4 cases of lymph node involvement by mycosis fungoides. Other previous studies have shown Syk overexpression in SYK-translocated cases,² Syk upregulation in adult T-cell leukemia/lymphoma cell lines²⁰ and a relative increase in SYK expression in ALK-positive ALCLs.²¹

Several comments regarding the interpretation of our findings are warranted. First, IHC of reactive lymphoid tissue showed Syk-positive cells to outnumber CD20-positive cells in the paracortex (Figure 1a). Most lymphocytes appeared negative for Syk. By morphology and distribution, many of the positive cells appeared to be histiocytes and/or dendritic cells, which are among the hematopoietic cell types that express Syk.²² Without double immunostaining, the presence of a minimal population of normal Syk-positive T cells cannot be entirely excluded. However, such a population was not identified by flow cytometry, which is a highly sensitive method of detection. Second, unlike most normal T cells, NK cells have been reported to express Syk.²³ Although we did not include tumors of known NK-cell origin in our series, we did include cases of NKTLs, which may be of either NK- or T-cell origin.¹⁵ Four out of five NKTLs were Syk-positive by IHC. If these four positive cases were of NK-cell origin, the observed Syk positivity might reflect constitutive expression in this cell type rather than lymphoma-associated overexpression. Finally, as mentioned above, tumor lysates subjected to western blot would be expected to contain some protein from admixed non-neoplastic cells. The phosphorylation status of Syk in the admixed B cells present in PTCLs such as AITLs is unknown. In B-cell lymphomas such as follicular lymphoma, tumor-infiltrating non-neoplastic B cells appear to demonstrate lesser Syk phosphorylation than the tumor cells on stimulation.²⁴ However, as phosphorylation of Syk is a physiologic event in B-cell receptor-mediated signaling,⁸ we cannot exclude the possibility that some of the phospho-Syk detected by western blot of PTCL samples (Figure 2b) was derived from admixed B cells.