Forty-six patients, who were diagnosed with DLBCL at Osaka University Hospital during the period from April 2006 to December 2009 and whose follow-up data could be accessed, were examined. Clinicopathological findings in these 46 patients are summarized in Table I. There were 19 men and 27 women with ages ranging from 29 to 78 years (median age, 63). Location of DLBCL was nodal in 25 cases and extranodal in 21 cases. Immunohistochemically, the large lymphoid cells were CD20⁺ and CD3⁻. According to the criteria proposed by Hans et al (12), DLBCL is categorized into germinal center-type B cell lymphoma (GCB) (CD10⁺ or CD10⁻/bcl-6⁺/MUM1⁻) and non-GCB (CD10⁻/bcl-6⁻/MUM1⁺) type. The subtype of the studied DLBCL cases was GCB in 11 cases and non-GCB in 35 cases. Although the number of females and non-GCB cases was higher than that of males and GCB cases, there was no collection bias. Histological specimens were fixed in 10% formalin and routinely processed for paraffin-embedding. Paraffin-embedded specimens were stored in a dark room at room temperature at the Department of Pathology of Osaka University Hospital, and sectioned (4 μm) at the time of staining and stained with H&E and immunoperoxidase procedure.

Based on physical examination records, surgical notes and pathological assessments of the specimens, the Ann Arbor staging system was applied. Eighteen cases were at stage I, 7 at stage II, 4 at stage III and 17 at stage IV. Ten patients received a combination of radiotherapy and chemotherapy, 28 received chemotherapy only, 2 received chemotherapy after surgery, 4 transplantation (3 auto- and 1 allo-transplantation), 1 received radiation therapy only and 1 received no therapy. The clinical outcome was evaluated according to the Guidelines of the International Workshop to Standardize Response Criteria for non-Hodgkin Lymphoma (13). The characteristics of the patients, such as age, tumor location and stage, in the present study were similar to those in previous reports (14,15). The proportion of male DLBCL patients in general was ∼55%, indicating that the proportion of female patients was slightly high in the present study. The study was approved by the Ethics Review Board of the Graduate School of Medicine, Osaka University.

ALDH1 expression was immunohistochemically examined using the anti-ALDH1 antibody (BD Biosciences, Franklin Lakes, NJ, USA). The sections were incubated with anti-ALDH1 diluted at x100, and subsequently processed using the ChemMate EnVision kit (Dako A/S, Glostrup, Denmark). DAB (Dako) was used as a chromogen. As the negative control, staining was carried out in the absence of the primary antibody. For immunophenotyping of DLBCL, the following monoclonal antibodies were used: CD20, CD3, Bcl-6, CD10 and MUM1 (Dako; dilution at 1:400, 1:50, 1:50, 1:100 and 1:100, respectively). Stained sections were evaluated independently by two pathologists (S.F. and E.M.).

Double staining of ALDH1 with CD20 (a marker for DLBCL cells), fascin (a marker for DCs; Dako) and CD68 (a marker for macrophages; Dako) was carried out using the EnVision G/2 doublestain system (Dako) according to the manufacturer’s protocol. For double staining with CD20, sections were first incubated with the anti-CD20 antibody (1:100), colored with DAB and, subsequently, ALDH1 expression was detected with Permanent Red (Dako). For double staining with CD68 or fascin, sections were initially incubated with ALDH1, colored with DAB and, subsequently, antigen retrieval was carried out using a Pascal pressurized heating chamber (Dako). After incubation with the anti-CD68 (1:100) or anti-fascin (1:100) antibody, sections were colored with Permanent Red. Since red fluorescence is released from Permanent Red, the red-color signal was detected with a fluorescence microscope (Biozero, Keyence, Osaka, Japan).

The Chi-square test was used to analyze the correlation between ALDH1 expression and clinicopathological characteristics of the lymphoma cases. Kaplan-Meier methods were used to calculate the overall survival (OS) rate, and differences in survival curves were evaluated with the log-rank test. p-values of <0.05 were considered statistically significant.

The expression of ALDH1 was examined in 46 DLBCL cases. The signals for ALDH1 were detected in two types of cells: cells with ample cytoplasm and fine chromatin and those with reticular structure (Fig. 1A). To confirm the absence of ALDH1 signals in DLBCL cells, double staining of ALDH1 and CD20 was performed; CD20⁺ large cells (brown-colored cells) did not express ALDH1 (Fig. 1B and C).

Since ALDH1⁺ cells with ample cytoplasm and fine chromatin morphologically resembled macrophages, the double staining of ALDH1 and CD68, a marker for macrophages, was carried out. ALDH1⁺ cells with ample cytoplasm and fine chromatin were positive for CD68 (Fig. 2A and B). By contrast, ALDH1⁺ cells with reticular morphology were negative for CD68 (Fig. 2A and B).

Reticular morphology in ALDH1⁺ cells suggested that a portion of the DCs was ALDH1⁺. To confirm this, double staining of ALDH1 and fascin was performed. ALDH1⁺ DC-like cells were positive for fascin (Fig. 2C and D).

Clinical significance of the presence of ALDH1⁺ cells in DLBCL was examined. Seventeen and 7 of 46 cases possessed ALDH1⁺ macrophages and ALDH1⁺ DCs, respectively. The presence of ALDH1⁺ macrophages did not show any correlation with gender, location, subtype, stage of diseases and response to therapies (Table II). By contrast, the presence of ALDH1⁺ DCs was correlated with location (p=0.038) and response to therapies (p=0.039), although no correlation was detected in gender, subtype and stage of disease (Table III). There was a statistically significant difference in OS between patients with and without ALDH1⁺ DCs (p=0.006), although no difference was detected between patients with and without ALDH1⁺ macrophages (Fig. 3).