High expression of 5-lipoxygenase in normal and malignant mantle zone B lymphocytes

Background Human B lymphocytes can produce leukotriene B4 but the biological function of the 5-lipoxygenase (5-LO) pathway in B cells is unclear. In order to better understand and define the role of 5-LO in B cells, we investigated the expression of 5-LO mRNA and protein in subsets of B cells from human tonsils and different types of B cell lymphoma. Results Based on RT-PCR and western blot/immunohistochemical staining, with a polyclonal antibody raised against 5-LO, high expression of 5-LO was found in mantle zone B cells from tonsils. By contrast, only a weak expression of 5-LO was detected in germinal centre cells and no expression in plasma cells from tonsils. This pattern of 5-LO expression was preserved in malignant lymphoma with high expression in mantle B cell lymphoma (MCL) and weak or no expression in follicular lymphoma. Primary leukemized MCL, so called B-prolymphocytic leukaemia cells, and MCL cell lines also expressed 5-LO and readily produced LTB4 after activation. Conclusion The present report demonstrates the expression of 5-LO mainly in normal and malignant mantle zone B cells while the expression is low or absent in germinal centre B cells and plasma cells, indicating a role of the 5-LO pathway in B cells before the cells finally differentiate to plasma cells.


Background
Arachidonic acid can be converted to leukotrienes which mediate inflammatory and immunological reactions [1].
The key enzyme in leukotriene biosynthesis is 5-lipoxygenase (5-LO), which upon activation and interaction with 5-LO activating protein (FLAP) converts arachidonic acid, via a two step process, to leukotriene (LT) A 4 . This compound can easily be transformed into LTB 4 , through the action of LTA 4 hydrolase, or into LTC 4 , catalyzed by LTC 4 synthase [1]. Leukotriene C 4 can be further converted to LTD 4 and LTE 4 . The biological effects of leukotrienes are dependent on receptor interaction [1][2][3][4][5]. Leukotriene B 4 is a potent chemotactic mediator for granulocytes and T lymphocytes [6][7][8][9]. Several reports have demonstrated a function of LTB 4 in the immune system as a stimulator of monocytes, T lymphocytes and B lymphocytes [10][11][12].
Biosynthesis of leukotrienes is restricted to a few cell types in the human body. Myeloid cells are the main source of leukotriene formation but B lymphocytes have also the capacity to produce LTB 4 . The activation of leukotriene synthesis in B cells is quite different in comparison to myeloid cells. Neutrophils and monocytes readily produce leukotrienes upon stimulation with calcium ionophore A23187. B cells, however, do not produce LTB 4 after challenge with calcium ionophore only but the cells can produce similar amounts of LTB 4 as myeloid cells after changing the cellular oxidative status [13][14][15]. The 5-LO activity in B cells appears to be latent and the mechanism of activation of the enzyme under physiological conditions is not yet known. Endogenously produced LTB 4 , however, plays a pivotal role in CD40-dependent activation of chronic B lymphocytic leukaemia cells (B-CLL) [16].
In resting neutrophils, 5-LO is localized in the cytoplasm but upon cell activation the enzyme translocates to the nucleus and nuclear membranes [17,18]. It has been proposed that this translocation allows for 5-LO to interact with FLAP on the nuclear membrane, thus enabling leukotriene synthesis [1]. The localization of 5-LO seems, however, to differ between different types of myeloid cells [17][18][19]. Phosphorylation of 5-LO appears to influence the nuclear import of 5-LO [20]. In B cell lines and isolated B cells in vitro, both cytoplasmic and nuclear localisation of 5-LO have been reported [14,21].
Mantle cell lymphoma (MCL) constitutes 5% of non-Hodgkin lymphomas. Most MCL carry the t(11;14)(q13;q32) translocation by which cyclin D1 becomes overexpressed [22,23]. Most MCL have unmutated immunoglobulin genes [24] and the current hypothesis is that the tumour cells are derived from the mantle or marginal zone of the B cell follicles. Microarray data of MCL have revealed high expression of 5-LO in these cells in comparison to control lymphoid tissue [25].
The enzyme 5-LO has been reported to be expressed in precursor B cells, B cell populations from the peripheral blood, tonsils and various types of malignant B cells [13,14,26]. However, it is not known which particular subsets of B lymphocytes from the tonsils which can express 5-LO and produce LTB 4 . Therefore, in order to define the function of the leukotriene pathway in B cells, we investigated the cellular expression of 5-LO in different tonsillary subsets of B lymphocytes and the corresponding type of malignant B cell lymphoma.

PCR analysis of genes involved in the biosynthesis of leukotrienes in subsets of B cells
RT-PCR was performed on isolated total RNA from subsets of tonsillary B cells to elucidate the gene expression of enzymes involved in the leukotriene cascade. These analyses demonstrated that 5-LO, FLAP and LTA 4 hydrolase were expressed in various degrees in tonsillary B cells (CD19 + ), memory B cells (CD19 + , CD38 + , IgD -), mantle zone B cells (CD19 + , CD38 + , IgD + ) and germinal centre (GC) B cells (CD19 + , CD38 ++ , IgD -). Highest relative expression of these three genes was found in mantle zone B cells and lowest expression in GC B cells (Figure 1). These results indicate differential expression of these enzymes in subpopulations of B cells. No investigated subpopulation of B cells expressed LTC 4 synthase.

Western blot analysis of 5-LO expression in subsets of B cells
Subpopulations of B cells from tonsils were separated by flow cytometry and analyzed by western blot in order to further characterize the expression of 5-LO. Whole cell lysates of purified cells were submitted to SDS/PAGE followed by western blotting using a polyclonal anti-human 5-LO antibody. A single immunoreactive band at the expected size was observed in the sample consisting of total B cells (CD19 + ) from tonsils, mantle zone B cells (CD19 + , CD38 + , IgD + ) and in memory B cells (CD19 + , CD38 + , IgD -) and a weak band was detected in GC B cells (CD19 + , CD38 ++ , IgD -) ( Figure 2). In contrast, no band was detected in samples from plasma cells (CD19 + , CD38 +++ , IgD -). There was a more marked difference in the expression of the 5-LO protein than in the 5-LO mRNA levels in the various subtypes of B cells (Figures 1  and 2).

Immunofluorescence analysis of 5-LO expression in subsets of tonsillar B lymphocytes
To identify the expression of 5-LO protein in subpopulations of tissue tonsillar B cells, a panel of antibodies were used to identify mantle zone B cells (IgD + ), germinal centre B cells (CD38 ++ ) and plasma cells (CD38 +++ or CD138 + ). Figure 3A shows high expression of 5-LO in mantle zone B cells (IgD + ). In contrast, a very weak expression of 5-LO was observed in GC cells (CD38 ++ ) ( Figure  3B). Figure 3C demonstrates no expression of 5-LO in plasma cells (CD38 +++ ) within the GC. Furthermore, no expression of 5-LO was observed in plasma cells (CD138 + ) outside the GC ( Figure 3D). These results concur with the western blot analysis of 5-LO in different subsets of B cells ( Figure 2).

Immunohistochemical analysis of 5-LO expression in mantle cell lymphoma and follicular lymphoma
In order to further characterize the expression of 5-LO in subsets of B cells, we investigated the expression of 5-LO in malignant cells derived from the mantle zone and the germinal centre i.e. mantle B cell lymphoma and follicular lymphoma. For comparison, lymph node biopsies were obtained from patients with reactive lymph nodes. In agreement with the data shown in figure 3, the expression of 5-LO in a reactive lymph node was mainly found in B cells from the mantle zone rather than from the germinal centre ( Figure 4A). Interestingly, virtually all B cells expressed 5-LO in a biopsy obtained from a patient with MCL ( Figure 4B). In contrast, corresponding analyses of biopsies from patients with follicular lymphoma, demonstrated that very few cells expressed 5-LO. These cells are most likely dendritic cells or tangible body macrophages as described earlier [27] ( Figure 4C). No staining was observed with control antibody ( Figure 4D). Similar results were obtained with another 5-LO antiserum (data not shown). Taken together, all investigated biopsies from patients with mantle zone B cell lymphoma clearly expressed 5-LO, whereas, little or none staining was observed in biopsies from patients with follicular lymphoma ( Table 1).
As a comparison, tonsillar polymorphonuclear granulocytes expressed 5-LO in the nucleus ( Figure 4E), in contrast to tonsillar CD3 positive T lymphocytes which did not express 5-LO ( Figures 4F and 4G). The IHC in figure  4G was stained with human anti-CD3 antibody.

Biosynthesis of LTB 4 in mantle B cell lymphoma
Since it was difficult to isolate large numbers of tonsillary non-malignant mantle zone B cells for chemical analysis, we used MCL cells to study the biosynthesis of LTB 4 in Semiquantitative RT-PCR analysis of enzymes in the leukotriene pathway in subpopulations of B cells from tonsils mantle zone B cells. For this purpose, we used both different MCL cell lines and primary MCL cells isolated from the peripheral blood of patients with leukemized MCL, so called prolymphocytic leukemia (B-PLL) with t (11;14). B-PLL is in fact a heterogenous disease and those cases that are positive for cyclin D1 and carry the t(11:14) translocation are in the current WHO classification of malignant lymphomas [22] considered to be leukemic forms of MCL [28,29]. All three investigated MCL cell lines (Granta 519, JEKO-1 and Rec1) expressed 5-LO protein ( Figure 5).  [30]. This cocktail of compounds is known to induce LTB 4 production in other types of B lymphocytes [14]. Sonicated cells also readily produced LTB 4 ( Figure  6A). Upon calcium ionophore activation only, isolated B-PLL produced low amounts of LTB 4 (about 1 pmol/10 6 cells), showing that there were minimal amounts of myeloid cells contaminating these preparations of B-PLL cells. Figure 6B shows that intact or sonicated Granta 519, JEKO-1 and Rec1 cells also produced significant amounts of LTB 4 as B-PLL cells. The capacity of the different cell lines to produce LTB 4 correlated relatively well with the degree of expression of 5-LO protein ( Figures 5 and 6B). In essence, these results demonstrate that both B-PLL cells and MCL cell lines have the capacity to produce LTB 4 and that the cells contained substantial amounts of 5-LO, which could be activated under certain conditions.

Discussion
The activation mechanisms and biological functions of 5-LO in human B lymphocytes are unclear although the overall capacity of B cells to produce LTB 4 is similar to myeloid cells [30]. We purified various subpopulations of tonsillary B cells and investigated the expression of enzymes involved in the leukotriene pathway. RT-PCR and western blot analysis of B cells from tonsils separated by flow cytometry showed high expression of 5-LO, in particular mantle zone B cells but also memory B cells (Figures 1 and 2), while germinal centre cells and plasma cells showed low or no expression of 5-LO, respectively. This analysis also demonstrated that FLAP and LTA 4 hydrolase were expressed in these various B cell populations ( Figure 1). In order to further investigate the expression of the 5-LO protein, immunofluorescence analyses of tonsillary tissue B cells were performed. A distinct expression of 5-LO was seen in IgD + B cells in the mantle zone, while only a weak expression was found in germinal centre B cells and virtually no expression in plasma cells (Figure 3).
To verify if the expression of 5-LO was preserved in malignant B cells we investigated MCL and FL. Mantle cell lymphomas have a lower somatic mutation rate in immunoglobulin genes than follicular lymphomas and approximately 60-75% of mantle cell lymphomas have immunoglobulin genes in germ-line configuration. Therefore, it has been hypothesized that mantle cell lymphomas are derived from naïve cells, similar to normal mantle zone B cells [31]. Immunohistochemical analysis of biopsies from MCL patients showed high expression of 5-LO in virtually all cells ( Figure 4B). Follicular lymphomas, on the other hand, have a high rate of somatic mutations in immunoglobulin genes and clearly originate from germinal centre B cells [31]. There were no or low expression of 5-LO in follicular lymphoma cells and this concur with the results of normal germinal centre B cells (Figures 2, 3, 4A, 4C and Table 1). There are conflicting data about the expression of 5-LO in T cells. We have previously reported that T cells express FLAP but not 5-LO [13], whereas, one other report indicates the formation of LTB 4 by human T cells [32]. We found no evidence for the expression of 5-LO protein in CD3 positive tonsillar T cells ( Figure 4F).

Western blot analysis of 5-LO expression in B cells from tonsils
Since the numbers of tonsillary mantle zone B cells that can be obtained by flow cytometry cell sorting is limited, we investigated the capacity to produce LTB 4 in primary B-PLL cells and cell lines derived from patients with MCL. Three different MCL cell lines were used in this study and all were found to express 5-LO and to have the capacity to produce LTB 4 ( Figures 5 and 6). Also, primary B-PLL cells    4 have not yet been uncovered. In this regard, the cellular oxidative status of the cells seems to be of importance [13][14][15]. Our finding that the enzyme is highly expressed in mantle zone B cells and less expressed in germinal centre B cells and not expressed in plasma cells could indicate that the enzyme might be activated in the lymph node to attract activated T cells. It is known that activated T cells express BLT1, the LTB 4 receptor, and that LTB 4 is a potent chemotactic agent for T cells [7][8][9]. Therefore, we will now study the effect of T-and B lymphocyte interaction on the activity and expression of 5-LO.

Conclusion
In summary, the present study demonstrates high expression of 5-LO mantle zone B cells but not in germinal centre B cells or plasma cells indicating a role of the 5-LO pathway in B cells before the cells differentiate to plasma cells.
MCL cell lines were washed in PBS (w/o calcium and magnesium) and suspended in PBS (w/o calcium and magnesium) supplemented with complete mini protease inhibitor (Roche, Sweden). Cells were sonicated on ice (3 × 5 seconds) (Sonics Vibra cell, 40% amplitude, CiAB,

Immunohistochemistry on Tissue Microarray (TMA)
Formalin-fixed, paraffin-embedded tumor (lymph nodes) tissue samples and hyperplastic tonsils were obtained from the Department of Pathology, Karolinska University Hospital Huddinge. Before constructing a TMA block, serial 5 μm sections were cut from each donor block. One of these sections was stained with H&E for marking morphologically representative areas of the tumor. One representative area was targeted. Using Beecher Instruments Tissue Arrayer (Silver Springs, MD, USA), tissue cylinders with a diameter of 1.0 mm were punched from the targeted area in each donor block and deposited into a recipient TMA block. The cores of each TMA included duplicate cores from all cases and two cores of normal human heart muscles as a control for orientation of the sections.
Multiple serial 5 μm-thick TMA sections containing samples were cut de-paraffinized in xylene and hydrated in a series of graded alcohols. Enzyme digested antigen retrieval was carried out with Protease (Sigma P-5147, Sweden) at 20°C for 10 min. Anti-5-LO antibody (LO-32, Merck Frosst, Canada) was used at a dilution of 1:300 and anti-CD3 antibody was used at a dilution of 1:100 (DAKO). All staining were semiautomatic and performed on a TechMate 500 plus (DAKO, Glostrup, Denmark) by using the Dako REAL™ Detection System, PeroxidaseDAB, RabbitMouse kit as recommended by the manufacturer.
The results of the staining were evaluated by three persons (XW, BC and BS). Images of immunohistologic staining were acquired using an Olympus BX45 microscope (Olympus, Stockholm, Sweden) and Sony digital camera (DXC-S500, Sony, Stockholm, Sweden). Digitized images were processed using Picsara 8.9 Rev 5 software (Bildanalyssytem AB, Stockholm, Sweden).

Incubation of B-PLL cells and MCL cell lines
Intact cells (10 × 10 6 ) were suspended in 1 mL phosphate buffered saline (PBS) and pre-incubated for two min with azodicarboxylic acid bisdimethylamide (Diamide) (100 μM) prior to stimulation with arachidonic acid (40 μM) and calcium ionophore A23187 (1 μM). The cells were stimulated for five min and the reaction was terminated with 1 mL methanol. The A23187 incubation was performed as described above but without any addition of Diamide and arachidonic acid. Alternatively, cells (10 × 10 6 ) were re-suspended in 1 mL calcium free PBS including EDTA (2 mM) and sonicated for 3 × 5 seconds. The cells were pre-incubated for two min at 37°C in the presence of ATP (1 mM) prior addition of calcium chloride (2 mM) and arachidonic acid (40 μM). The reaction was terminated after five min of incubation with 1 mL methanol.

Analysis of leukotrienes
After addition of PBS (0.5 mL) and the internal standard PGB 1 (50 pmol), the samples were centrifuged (1250 × g, 5 min). The supernatants were subsequently subjected to solid phase extraction on Oasis HLB Extraction Cartridges (10 mg, Waters, Sweden). The methanol fractions were analyzed on a Waters Alliance 2695 reverse phase HPLC and detected with Waters PDA 996. Methanol:water:trifluoroacetic acid (70:30:0.007, v/v) was used as mobile phase at a flow rate of 1.2 mL/min. Qualitative analysis was performed by comparison of retention times of synthetic standards and quantitative determinations were performed by computerized integration of the area of eluted peaks. Mean and standard deviation was calculated.

Ethical perspective
This study was approved by the local ethic committee of Friedrich-Schiller-University of Jena, Germany, and Karolinska University Hospital, Stockholm, Sweden.