Prostaglandin E2 Increases Bovine Leukemia Virustax and pol mRNA Levels via Cyclooxygenase 2: Regulation by Interleukin-2, Interleukin-10, and Bovine Leukemia Virus

ABSTRACT Prostaglandin E2 (PGE2), produced by macrophages, has important immune regulatory functions, suppressing a type 1 immune response and stimulating a type 2 immune response. Type 1 cytokines (interleukin-2 [IL-2], IL-12, and gamma interferon) increase in freshly isolated peripheral blood mononuclear cells (PBMCs) of animals with an early disease stage of bovine leukemia virus (BLV) infection, while IL-10 increases in animals with a late disease stage. Although IL-10 has an immunosuppressive role in the host immune system, IL-10 also inhibits BLV tax and polmRNA levels in vitro. In contrast, IL-2 stimulates BLVtax and pol mRNA and p24 protein expression in cultured PBMCs. The inhibitory effect of IL-10 on BLV expression depends on soluble factors secreted by macrophages. Thus, we hypothesized that PGE2, a cyclooxygenase 2 (COX-2) product of macrophages, may regulate BLV expression. Here, we show that the level of COX-2 mRNA was decreased in PBMCs treated with IL-10, while IL-2 enhanced the level of COX-2 mRNA. Addition of PGE2 stimulated BLV tax and polmRNA levels and reversed the IL-10 inhibition of BLV mRNA. In addition, the specific COX-2 inhibitor, NS-398, inhibited the amount of BLV mRNA detected. Addition of PGE2 increased BLV tax mRNA regardless of NS-398 addition. PGE2 inhibited antigen-specific PBMC stimulation, suggesting that stimulation of BLVtax and pol mRNA levels by PGE2 is independent of cell proliferation. These findings suggest that macrophage-derived COX-2 products, such as PGE2, regulate virus expression and disease progression in BLV infection.

Bovine leukemia virus (BLV), closely related to human Tcell leukemia virus type 1 (HTLV-1), is a type C retrovirus that infects bovine B cells and leads to enzootic bovine leukosis (16). The genomes of BLV and HTLV-1 are similarly arranged. In particular, the 3Ј region that contains the tax, rex, R3, and G4 genes is unique to BLV and HTLV-1 (2). In addition, there are several features of pathogenesis that are shared by BLV and HTLV-1. For both viruses, many infected individuals develop antibodies, but clinical symptoms are relatively rare. Disease progression in BLV-infected animals is divided into three stages: serologically positive, persistent lymphocytosis negative (alymphocytotic [AL]); serologically positive, persistent lymphocytosis positive (persistently lymphocytotic [PL]); and tumor-bearing stages (usually lymphosarcoma). Most infected animals never display outward signs of disease and are referred to as asymptomatic or aleukemic. Fewer than 5% of infected animals develop malignant lymphosarcoma (11), while 30% of infected animals progress to persistent lymphocytosis, in which nonneoplastic B cells proliferate and leukocyte counts may exceed 10,000 cells/mm 3 (17). Usually there is a long duration between these disease stages. The mechanism of disease progression from AL to PL or tumor-bearing stage is not clear.
Recent investigation has revealed that cytokine production plays a critical role in the progression of many different diseases (9,32). In previous studies, we found cytokine polariza-tion in BLV-infected animals at different disease stages (27,29). Type 1 cytokines, such as interleukin-2 (IL-2), IL-12, and gamma interferon (IFN-␥), were expressed in high amounts in AL animals, while IL-10 was increased in PL and tumor-bearing animals. Other studies showed that antigen-specific lymphocyte proliferation (25) and cytotoxic ␥␦ T-cell activity (P. S. Lundberg and G. A. Splitter, submitted for publication) were reduced in PL and tumor-bearing animals. Because IL-10 had an inhibitory effect on type 1 immune responses, increased IL-10 expression in late disease stages may result in suppressed cell-mediated immune responses. Nevertheless, we found that IL-10 inhibited BLV tax and pol mRNA levels, while IL-2 stimulated detection of BLV in peripheral blood mononuclear cells (PBMCs) in vitro (28). These data suggest that increasing levels of IL-10 during BLV infection may have a beneficial effect to suppress BLV expression in infected B cells. In addition, we found that macrophages have a critical role in BLV expression and secrete certain factor(s) to activate BLV tax and pol mRNA.
We hypothesized that one of the soluble factor(s) may be prostaglandin E 2 (PGE 2 ) that is produced by cyclooxygenase 2 (COX-2). PGE 2 is the only prostaglandin produced by macrophages in response to inflammatory cytokines, such as IL-1 and tumor necrosis factor alpha (TNF-␣), that is known to inhibit cell-mediated immune response and stimulate type 2 cytokine production (12,14,18,26,36). There are two COX enzymes, referred to as COX-1 and COX-2, that produce prostaglandins from arachidonic acid (10). While COX-1 is a constitutive enzyme that is expressed at its highest concentrations in the kidney, stomach, platelets, and vascular endothelium, COX-2 is an inducible enzyme whose expression is regulated by growth factors, tumor promoters, or cytokines. COX-2 is expressed in a few specialized tissues and cells, including macrophages and follicular dendritic cells. The functions of COX-2 and prostaglandins are very important in regulating normal physiological processes (8,19), as well as the immune response. Here, we demonstrate that IL-10 decreased detection of COX-2 mRNA by PBMCs, while conversely, IL-2 increased COX-2 mRNA. Although PGE 2 reduced antigen-specific PBMC proliferation, PGE 2 increased detection of BLV and reversed the IL-10 inhibition of BLV tax and pol mRNA levels. In addition, BLV acts as an autocrine stimulator to increase the levels of BLV tax and pol mRNA and COX-2 mRNA.
COX-2 mRNA from PBMCs was inhibited by IL-10 and enhanced by IL-2. In a previous study, we found that human recombinant IL-10 (hrIL-10) inhibited BLV tax and pol mRNA, while IL-2 enhanced the detection of BLV tax and pol mRNA and BLV p24 protein in PBMCs in vitro (27). Also, IL-10-mediated regulation of BLV expression was macrophage dependent. Because COX-2, an important enzyme for prostaglandin synthesis, is mainly expressed by macrophages (18,26,36), we hypothesized that COX-2 and its products may affect BLV expression regulated by IL-2 and IL-10. Quantitative competitive PCR (QC-PCR) was performed to determine COX-2 mRNA levels in PBMCs cultured with or without hrIL-10. Transcripts of COX-2 mRNA were quantified by a competitive reverse transcriptase PCR (RT-PCR) assay using standard curve methodology. Validation of this assay and synthesis of native and competitor standards has been published previously (37,38). A standard curve was created by RT-PCR using a constant amount of competitor RNA (2 aM) together with twofold serial dilutions of native RNA (15 to 0.11 aM). Unknown mRNA samples were diluted as needed, reverse transcribed and amplified with the same amount of competitor RNA, and compared to the standard curve. Reverse transcription was carried out in 1ϫ RT buffer (Promega), 0.2 mM deoxynucleoside triphosphates, 100 M random hexamer, and 40 U of Moloney murine leukemia virus RT for 1.5 h at 37°C, followed by 95°C for 10 min in a final volume of 20 l. Four microliters of RT reaction mixture were then PCR amplified (30 s, 95°C; 30 s, 57°C; 30 s, 72°C) for 30 cycles, followed by 72°C for 5 min in 20 l of 1ϫ PCR buffer (Promega, Madison, Wis.), 1.5 mM MgCl 2 , 0.2 mM deoxynucleoside triphosphates, 0.4 M upstream and downstream primers, and 0.5 U of Taq DNA polymerase. PCR products were separated on a 5% polyacrylamide gel electrophoresis (PAGE) gel and stained with ethidium bromide. Bands observed in gels were quantified using Collage a software (Fotodyne, Heartland, Wis.). Densitometric values for the standard curve were plotted as follows: log([native RNA]/[competitor RNA]) versus log [native RNA (aM)]. Addition of hrIL-10 reduced the detection of COX-2 mRNA in PBMCs from both AL and PL animals (Fig. 1a). These results paralleled the effects of IL-10 inhibition of BLV tax and pol mRNA levels (27). In contrast, IL-2 dramatically increased detection of COX-2 mRNA (Fig. 1b). These results suggest that cytokines, such as IL-2 and IL-10, can have opposing effects on the regulation of COX-2 mRNA. PGE 2 , a COX-2 product, enhances detection of BLV tax and pol mRNA and reverses IL-10 inhibition of BLV tax and pol mRNA. PGE 2 is the only COX-2 product expressed by macrophages (14). To determine whether PGE 2 affects BLV expression, BLV-infected PBMCs were cultured with or without PGE 2 in the presence of hrIL-10. Subsequently, the tax and pol mRNA produced by PBMCs from BLV-infected animals was measured by QC-PCR (27). For standardization, PCR was performed using this serially twofold diluted standard plasmid, with concentrations ranging from 8,192 to 16 fg/l for tax and from 2,048 to 4 fg/l for pol and with a fixed amount of mimic (10 fg/l). Synthesized cDNA from each sample and fixed amounts of mimic were added to the same tube and amplified simultaneously with tubes for a standard reaction. Gel photographs were scanned, and the amplified DNA bands were analyzed by densitometry using the NIH Image program, version 1.61, with standard curves constructed with Cricket graph. The amount of cytokine produced was determined by comparing the density ratios of sample and standard reaction mixtures. As expected, we observed that different animals have different viral loads. Usually, PL animals have more viral load than AL animals. While IL-10 reduced detection of BLV tax and pol mRNA, PGE 2 reversed this IL-10 inhibitory effect (Fig. 2a). Addition of PGE 2 alone also enhanced BLV tax and pol mRNA levels (Fig. 2b). These results suggest that macrophage-derived PGE 2 may stimulate BLV expression from infected B cells and that IL-10 may inhibit BLV expression by reducing PGE 2 production. At the higher concentration of PGE 2 , BLV tax and pol mRNA levels were slightly diminished, indicating that the stimulatory effect of PGE 2 was maximal at 1 M. To confirm the effects of PGE 2 to increase BLV mRNA levels, a selective COX-2 inhibitor, NS-398 (21), was added to PBMC cultures. As expected, the level of BLV tax mRNA was dramatically suppressed by the COX-2 inhibitor (Fig. 2c). In contrast, addition of PGE 2 increased BLV tax mRNA detection regardless of NS-398 addition (Fig. 2c). These data suggest that a selective COX-2 inhibitor may reduce the level of BLV mRNA in infected B cells and that PGE 2 addition may bypass the NS-398 inhibition of COX-2 activity.
Purified BLV increases the level of COX-2 mRNA and has an autocrine effect that increases the levels of BLV tax and pol mRNA. Numerous factors can stimulate COX-2 and PGE 2 expression by macrophages. To investigate how BLV infection might affect COX-2 and BLV expression, BLV proteins were purified from the supernatant of the BL3* cell line using metrizamide density gradient centrifugation. Purified BLV antigens were confirmed by sodium dodecyl sulfate-PAGE and immunoblotting with anti-gp51 and -p24 antibodies (data not shown). When purified BLV was added to PBMC cultures, COX-2 mRNA detection rapidly increased (Fig. 3a). Also, detection of BLV tax and pol mRNA increased by addition of purified BLV (Fig. 3b), suggesting that BLV antigens regulate BLV expression by an autocrine mechanism. To remove the possibility that other factors isolated during BLV antigen purification could affect COX-2 and BLV expression, a BLVnegative reagent was prepared from the BL3 cell line, which does not produce any BLV antigens. BLV gp51 and p24 were not detected by immunoblotting in the purified BLV-negative material. The BL3 purified material could not enhance COX-2 mRNA or BLV tax and pol mRNA (Fig. 3).

PGE 2 suppresses antigen-specific PBMC proliferation.
To determine if PBMC proliferation and BLV expression were correlated, different concentrations of PGE 2 were added to the PBMC cultures with or without BL3* supernatant as a viral antigen source (23). Following incubation of PBMCs for 3 to 5 days, cells were harvested for further experiments. In cell pro- liferation assays, [ 3 H]thymidine was added 8 to 12 h before harvest and the radioactivity of the harvested cells was measured by a ␤-counter (MATRIX 9600; Packard, Meriden, Conn.). Medium alone was used as a control in proliferation assay. Addition of BL3* supernatant increased PBMC proliferation more than 10-fold. However, antigen-specific PBMC proliferation was dramatically suppressed with increasing concentrations of PGE 2 (Fig. 4). Spontaneous cell proliferation in the absence of antigens was also slightly reduced by addition of PGE 2 (Fig. 4). Thus, PGE 2 enhanced BLV detection but suppressed antigen-specific PBMC proliferation. The concentration of cells differed between transcription and proliferation assays, as fewer cells were optimal in the U-bottom wells for the proliferation assay.
The results presented here demonstrate that IL-10 inhibited detection of BLV tax and pol mRNA and reduced COX-2 transcription from macrophages, while PGE 2 activated BLV tax and pol mRNA. These data indicate that regulation of BLV is closely related to signals induced by PGE 2 . Both AL and PL animals, but not noninfected animals, produced similar findings when tested with IL-2, IL-10, and PGE 2 , supporting a common mechanism of pathogenesis in infected animals. Also, in spite of the relatively low IL-10 levels, AL animals produced more IL-10 than uninfected animals, and AL and PL animals responded similarly to IL-10 and PGE 2 . These findings suggest that macrophages produce PGE 2 and have a central role in regulating BLV expression in infected B cells (Fig. 5).
The E series of prostaglandins are widely known as immunosuppressive products produced by macrophages, follicular dendritic cells, and fibroblasts (10,17). These prostaglandins, especially PGE 2 , can downregulate many aspects of B-and T-cell functions. PGE 2 production is triggered by inflammatory cytokines, such as IL-1 and TNF-␣, that are produced in viral and bacterial infections (4,20). Increasing PGE 2 negatively regulates type 1 cytokines, such as IL-2, IFN-␥, and TNF-␣, by increasing production of type 2 cytokines, such as IL-10 (13). Thus, PGE 2 may have a central role in regulating production of type 1 and type 2 cytokines. PGE 2 activates a humoral immune response, stimulating B-cell differentiation and immunoglobulin class switching (26). In this paper, we show that PGE 2 stimulated detection of BLV tax and pol mRNA and inhibited PBMC proliferation. These findings suggest that enhancement of BLV expression by PGE 2 may not depend on cell proliferation. Ironically, B cells are the only PBMCs that are significantly infected with BLV (22), while macrophages are the only source of PGE 2 in PBMCs (14). However, B cells FIG. 3. BLV proteins stimulate COX-2 mRNA (a) and BLV tax and pol mRNA (b). BLV proteins were purified from BL3* supernatant, and BLVnegative reagent was prepared from BL3 supernatant using metrizamide density gradient centrifugation. BLV-infected PBMCs were cultured for 3 days with similar volumes of purified BLV (10 g/ml) and BLV-negative (BLVϪ) materials. COX-2 and BLV tax and pol mRNA were quantified as described above. have a number of PGE 2 receptors that regulate B-cell activation (7). Signal transduction by PGE 2 receptors mediates increased cyclic AMP (cAMP) production (3). BLV long terminal repeats (LTRs) contain a cAMP-response element (CRE) that facilitates BLV gene transcription (1,39). Tax stabilizes CRE-binding protein (CREB) to bind CRE in LTRs (5). Thus, CREB and Tax may activate BLV expression of infected B cells. In addition, protein kinase C (PKC) increases BLV expression with increased Ca 2ϩ influx (15). Therefore, increased PGE 2 production by macrophages may stimulate BLV tax and pol mRNA expression through cAMP-dependent PKA and/or PKC signal transduction pathway. BLV LTRs also contain NF-B binding sites that facilitate BLV transcription (6). Recently, antiinflammatory agents, such as aspirin and salicylate, reportedly inhibit the activity of IB kinase-␤, which facilitates the degradation of IB and activates NF-B (40). Antiinflammatory agents that inhibit prostaglandin synthesis may suppress BLV expression via NF-B inhibition.
We demonstrate that BLV functions as a stimulant of COX-2 expression. Although PGE 2 enhances IL-10 expression (13) to inhibit COX-2 and BLV expression, a synergistic effect of BLV expression, opportunistic infections (35), pregnancy (19), and/or stress (8) could induce disease progression in BLV infection. COX-2 and PGE 2 also inhibit programmed cell death and facilitate tumor formation (33,34), and thus these activities might promote lymphosarcoma and leukemia with other carcinogenic factors, such as Bcl-2 and BLV Tax protein, in BLV infection. Thus, in other retrovirus infections, the inhibitory function of IL-10 on human immunodeficiency virus expression has been reported (24,30,31). Most studies have utilized macrophage cell lines and primary macrophages, while the studies with T-cell lines or primary T cells failed to demonstrate the IL-10 inhibition of human immunodeficiency virus expression. Therefore, macrophages may have a direct role in regulating retrovirus expression responding to IL-10 (32). Our preliminary data showed that bovine herpes virus type 1 and Brucella abortus, two common opportunistic infections in cattle, activated COX-2 mRNA expression (D. Pyeon and G. A. Splitter, unpublished data). We anticipate that further research regarding PGE 2 and opportunistic infections will reveal additional clues to solve the complicated mechanisms of disease progression in retrovirus infections.
To examine whether inhibitors of PGE 2 or COX-2 would be efficacious for treatment of BLV infection, in vivo studies are necessary. Infected animals could be treated with a PGE 2 inhibitor, such as indomethacin, and viral load and BLV expression could be measured. Treatment with a PGE 2 inhibitor may reduce BLV load in BLV-infected animals and would support a role for PGE 2 to stimulate BLV replication and disease progression in vivo. This study provides an additional strategy to treat retrovirus infection combined with currently available antiretroviral treatment. This work was supported by National Cancer Institute grant R01 CA59127, BARD 95-34339-2556, and the College of Agricultural and Life Sciences.