SARS-CoV-2-induced impaired immune response by Prostaglandin E2 is accelerated by age, male sex and air pollution

The SARS-CoV-2 coronavirus has led to a pandemic with millions of people affected. The present study nds prostaglandin E2 (PGE2) blood levels elevated in COVID-19 patients with positive correlation with disease severity. SARS-CoV-2 induces PGE2 generation and secretion in infected lung epithelial cells by upregulating cyclo-oxygenase (COX)-2 and reducing the PG-degrading enzyme 15-hydroxyprostaglandin-dehydrogenase. Also living human-lung-precision-slices infected with SARS-CoV-2 display upregulated COX-2. PGE2 in serum of COVID-19 patients lowers the expression of Paired-Box-Protein-Pax-5 (PAX5), a master regulator of B-cell survival, proliferation and differentiation, in both human and mouse pre-B-cells, while the PGE2 inhibitor taxifolin directly reduces SARS-CoV-2-induced PGE2 production and attenuates viral replication. Risk-factors for severe disease courses, i.e. older age, male sex and air pollution are associated with higher PGE2 production and lower PAX5 expression in pre-B-cells. Since PGE2 acts broadly immunosuppressive its elevation might reduce the early anti-viral defense and its inhibition may therefore reduce severe disease courses. lung


Results
Circulating levels of PGE2 in COVID-19 patients and sex-and age-matched healthy controls We analyzed PGE2 levels in individuals with mild/moderate (n=41) and severe (n=48) COVID-19 disease from hospitals in Hanover (Germany), Milan (Italy) and Brussels (Belgium) and age-and sex-matched healthy controls (n=29) ( Table 1, S3). Clinical data and laboratory characteristics of the COVID-19 patients revealed that the more severely affected patients were signi cantly older with a higher proportion of males than females compared with the mildly/moderately affected group (Table 1). BMI and diabetes rate are increased in the entire COVID-19 cohort with no signi cant difference between the mild/moderate and the severe groups (Table 1). In addition, C reactive protein (CRP) was elevated, while the total leukocyte-and neutrophil counts were within the normal range, although some patients displayed markedly increased levels ( Table 1). The mean lymphocyte counts (T-and B-cells) were reduced in the majority of COVID-19 patients and were speci cally low in patients with severe disease courses (Table 1).
Platelets were in the normal range in all COVID-19 patient groups and lactate dehydrogenase (LDH) was increased and highest in the severely affected patients (Table 1). Mortality was 15% for the entire cohort with no patient deaths in the mild/moderate group and 27% of patients dying in the severe disease group (Table 1). Circulating PGE2 levels were increased in COVID-19 patients at the time of hospitalization compared with healthy controls, and PGE2 levels were signi cantly higher in the severely affected patients compared with the mild/moderate COVID-19 patients (Fig. 1A, B, Table 1).
Effect of PGE2 on the expression of pre-B-cell differentiation and survival factor PAX5 in human pre-Bcells PGE2 is known to attenuate the proliferation, differentiation and survival of B-cells 35,36 . Here, we observed that the addition of PGE2 (10 µM, i.e. 3525 pg/ml), in the range measured in COVID-19 patients' sera (1300 to >20.000 pg/ml), to two human B-cell precursor lines, 697 and SUP-B15, signi cantly reduced PAX5 mRNA expression (Fig. 3A, B). This effect was blocked upon co-treatment with the PGE2 receptor 4 (EP4) antagonist, GW627368 but not with the EP2 receptor antagonist, AH6809 (Fig. 3A).
Effect of serum from COVID-19 patients on PAX5 expression in human pre-B-cells Serum from COVID-19 patients with elevated PGE2 levels reduced the expression of PAX5 in SUP-B15 cells compared with serum from healthy controls. Again, this effect was blocked upon co-treatment with the PGE2 receptor 4 (EP4) antagonist, GW627368 (Fig. 3C).
Analyses of B-cells in lungs from patients who died of severe acute COVID-19 disease compared with healthy controls and transplant rejection biopsies In lung biopsies from patients who died of severe acute COVID-19 disease (AC group, con rmed by qRT-PCR for NSP7, Fig. 3D), the signals for CD20 pre-B-cells (qRT-PCR and immunohistochemical quanti cation) and plasma cells (qRT-PCR for CD138) were barely detectable and lower than in control lung tissue (ctrl) and markedly lower than in lung tissue obtained after transplant rejection (TR, Fig. 3E-H).
Lung tissue immunostaining showed increased numbers of CD68 + macrophages and CD4 + T-cells in AC and TR compared with ctrl lung biopsies (Fig. 3E).
Analyses of IL-17A, a marker for pro-in ammatory TH17 cells in lung tissue and in serum from patients with COVID-19 disease Previous reports suggest that PGE2 upregulates IL-17A and thereby promotes the formation of proin ammatory TH17 T-cells 37,38 . However, IL-17A measured by ELISA was below detection level in the serum from all controls and the majority of COVID-19 patients (sFig. 2A).
ffect of diesel exhaust particles as air pollutants on COX-2 expression and PGE2 secretion in human endothelial cells To investigate whether air pollution, which has been suggested as an additional risk factor for severe COVID-19 disease 14,15 , affects circulating PGE2 levels, we stimulated primary human cardiac endothelial cells (hCEC) with diesel exhaust particles (DEP). DEP stimulation of hCECs increased PGE2 secretion into the cell culture supernatants of hCECs (Fig. 4A, B). This was associated with the upregulation of COX-2 expression (Fig. 4B). In turn, the DEP-induced PGE2 secretion in hCECs was prevented by treatment with taxifolin 33,34 (Fig. 4A, B).

PGE2 levels in healthy individuals in relation to sex and age
In healthy control individuals aged <50, circulating PGE2 levels were higher (P>0.01) in men than in women (Fig. 5A). Circulating PGE2 levels were markedly higher in older (>60 years) healthy males and females than in respective sex-matched younger (<50 years) individuals (Fig. 5B). Both males and females showed a signi cant positive correlation of circulating PGE2 levels with age ( Fig. 5D, E), while no correlation with BMI, body weight (BW) or body fat content was observed (Table S4, Table S4). PAX5 expression was higher in 697 and SUP-B15 pre-B-cells incubated with serum from elderly individuals collected after 12 months of controlled physical exercise compared with the BL serum of the same individuals (Fig. 5H, I, Table S4). In addition, the EP4 antagonist, GW627368 increased PAX5 in 697 and SUP-B15 pre-B-cells exposed to BL serum, indicating that the suppressive effect is mediated by PGE2-EP4 (Fig. 5J, K).
Cardiac PGE2 secretion and cardiac PAX5 expression in male mice with impaired androgen receptor signaling in cardiomyocytes We showed previously that cardiac STAT3 de ciency in mice (STAT3-CKO) leads to the upregulation of COX-2 in both sexes, while HPGD is reduced only in male STAT3-CKO mice due to impaired androgen receptor (AR) signaling 39 . Here, we observed that PGE2 levels secreted from cardiomyocytes of male STAT3-CKO were signi cantly higher compared with cardiomyocytes from WT mice, while no such differences were observed between female STAT3-CKO and WT-cardiomyocytes (Fig. 6A, B). The treatment of male STAT3-CKO mice with the COX inhibitor ibuprofen (10 mg/kg bodyweight) for two weeks markedly reduced PGE2 secretion in isolated cardiomyocytes (Fig. 6C). Resident Sca-1 positive cardiac progenitor cells (CPC), of which pre-B-cells are a subpopulation (0.38% are Sca-1 + /CD19 + ; Fig. 6D-G) display lower PAX5 mRNA levels when isolated from STAT3-CKO compared with CPC isolated from WT male hearts; conversely, no difference in PAX5 mRNA levels was observed in CPC from young female STAT3-CKO and WT hearts (Fig. 6H, K). Stimulation of isolated WT-CPC with PGE2 in culture reduced PAX5 expression (Fig. 6L, M, -7.5x10 3 -fold, P<0.01).

Discussion
The key nding of this study is that PGE2 is elevated in patients with COVID-19 disease, with the highest blood levels observed in those severely affected by COVID-19. Since PGE2 exerts immunosuppressive effects on T-and B-cells and monocytes, its elevation might critically reduce the initial anti-viral defense in the early phase of infection and thereby lead to more severe disease courses. Interestingly, our data show that the SARS-CoV-2 virus, not only hijacks the host cell gene expression machinery in order to replicate, but also forces infected host cells to produce PGE2 by upregulating the PGE-generating enzyme COX-2, and at least in part by reducing the expression of the PGE2-degrading enzyme HPGD (Fig. 7). In addition, we provide evidence that reported risk factors for more severe COVID-19 disease courses, i.e. male sex, age and a sedentary life style 13,40 as well as air pollution 14,15 , are either associated with higher PGE2 levels or directly increase the production of PGE2 (Fig. 7). Our study shows that PGE2 serum levels are higher in men than women. In elderly (>60 years) patients, PGE2 levels are higher than in younger individuals and were reduced by regular exercise. We provide evidence that age-related ARdysfunction may lead to a higher PGE2 production ( Fig. 7) and show that DEP, a component of air pollution, increases the production of PGE2 in endothelial cells, an aspect that has also been reported for airway epithelial cells 41,42 (Fig. 7). Mechanistically, we demonstrate that PGE2 in serum from COVID-19 patients via its EP4 receptor reduces the expression of PAX5 in human pre-B-cell lines (Fig. 7). PAX5 is a B-cell speci c transcription factor responsible for pre-B-cell survival, proliferation, and differentiation 43,44,45 and reduced PAX5 therefore reduces the B-cell population, an observation we made in the lung tissue of patients who died of COVID-19 and one that ts well with the recently reported absence of germinal centers and reduction in Bcl-6 + germinal center B-cells in the lymphatic system of patients who died of COVID-19 19 . PGE2 impacts not only on B-cells, also targets the innate (monocytes/macrophages) and adaptive (T-cells) immune systems in response to viral infections, evidenced by the lowering of the immune response and the formation of immunity 25,26,27 . Based on the hypothesized key role of PGE2 in COVID-19 disease, we provide evidence supporting the bene t of lowering PGE2 levels during SARS-CoV-2 infection either by the use of COX-inhibitors such as ibuprofen or by more speci c PGE2 inhibitors such as taxifolin. Furthermore, our data also suggest that regular exercise and reducing air pollution would lower PGE2 levels and, with this, reduce the risk for severe COVID-19 disease (Fig. 7).
PGE2 directly alters B-and T-cells by negatively affecting their proliferation and survival and altering their differentiation 35,36,37,38 . PGE2 suppresses B-cells' continued growth and differentiation and thereby regulates their B-cell responses against pathogens 46 , as well as their proliferation and survival 47 . Here, we report that PGE2 reduces the expression of PAX5 in human pre-B-cells via its EP4 receptor. PAX5 is a master regulator of most aspects of the life cycle of B-cells by controlling genes that are required for early development, antigen-receptor recombination, signaling and adhesion. PAX5 represses the transcription of genes required for the development of other hematopoietic lineages and plasma cells 43,44,45 . In fact, the reduction of PAX5 is important for the nal differentiation of short-lived plasma cells and their antibody (AB) production. In this regard, a PGE2-mediated reduction of PAX5 not only reduces the number of pre-B-cells, but also boosts the differentiation of B-cells towards plasma cells. This initially leads to high SARS-CoV-2-directed AB titers, but at the price of the depletion of B-cell pools in the long run. Indeed, a severe disease course in patients who died of severe COVID-19 is associated with a reduction in germinal centers 19 and with low CD20 + B-cells counts in lung tissue. Postmortem analyses in patients who died of COVID-19 disease showed no signi cant lymphocyte invasion in cardiac tissue despite the presence of SARS-CoV-2 particles 10,11 . Accordingly, a lower percentage and count in CD3 + , CD4 + , and CD8 + lymphocytes populations have strong predictive values for in-hospital mortality, organ injury, and severe pneumonia 16 . Additional studies suggest higher risks for severe disease courses in COVID-19 patients with dysfunctional B-cells due to common variable immune de ciencies (CVIDs) 18 ; correspondingly, patients with larger pools of naïve B-cells seem to build a more effective immune response to SARS-CoV-2 48 .
PAX5 expression is also necessary for the development of memory B-cells after follicular B-cells have encountered antigens 45,49 . In this regard, elevated PGE2 would also reduce the ability of an organism to develop longstanding immunity after COVID-19 infection. Indeed, there are reports on reinfection in individuals with SARS-CoV-2 50, 51 including a recent case report of a patient with a CD20 + B-cell acute lymphoblastic leukemia who developed high AB titers against COVID-19 after an initial recovery. However, the patient experienced a viral reactivation after she lost her COVID-19 AB following the administration of rituximab, cytarabine, and dasatinib for her leukemia, and experienced severe COVID-19 pneumonia with lymphopenia and high in ammatory markers 52 . PGE2 not only affects B-cells, but also promotes T-cell exhaustion and viral expansion through EP2 and EP4, as revealed by recent studies 53 . Moreover, immunosuppression caused by T-cell depletion and exhaustion have been suggested as contributing to viral persistence and mortality in COVID-19 patients 2 . PGE2 also impacts on the innate immune system, i.e. monocytes/macrophages, where it exerts anti-in ammatory effects and lowers the release of proin ammatory cytokines 54 55 . Using our mouse model of premature age-related heart failure (STAT3-CKO), we also addressed the question of whether PGE2 may be involved in the predisposition of males for severe COVID-19 disease and the frequently observed cardiac complications in COVID-19 patients. In this model, we showed that impaired AR-signaling promotes PGE2 production in the heart of male mice but not in females due to an insu cient expression of HPGD 39 . Of note, impaired AR signaling in humans due to an age-related decrease in circulating testosterone is present in around 20% of men >60 years, 30% of men > 70 years and 50% of men >80 years 56 . The resulting increased PGE2 abundance in our mouse cardiac tissues was associated with a lower expression of PAX5 in cardiac pre-B-cells. The causal role of PGE2 was con rmed via the direct addition of PGE2-reduced PAX5 in WT-CPC in vitro. These observations highlight the in uence of local PGE2 production on the heart's immune response and its contribution to the high prevalence of cardiac complications and heart failure in COVID-19 patients 57 .
Finally, our study shows for the rst time, that PCLS can be used as a disease model to study SARS-CoV-2 infection in human tissues. In addition, it unveils potential therapeutic avenues to reverse the adverse effects on the immune system by reducing PGE2 production. First, we show that the PGE2 inhibitor taxifolin, also known as dihydroquercetin, e ciently reduces SARS-CoV-2-induced PGE2 production in lung cells, e.g. in endothelial cells exposed to DEP (Fig. 7A). Additionally, taxifolin slightly reduced the SARS-CoV-2 virus production in lung cells, an observation supported by a recent publication on screening for natural inhibitors for SARS-CoV-2 in silico, which identi ed taxifolin as a direct inhibitor of the SARS-CoV-2 main protease 58 and suggested that taxifolin may also have antiviral potential (Fig. 7A). Taxifolin is a potent avonoid with anti-in ammatory activity, which is present as a natural compound in vegetables and fruits and the Siberian larch, Larix sibirica, 33,34 . It is readily available in foodstuffs and could be tested directly in COVID-19 patients. Inhibition of the microsomal prostaglandin E synthase-1 (mPGES-1) by sonlicromanol (Khondrion; a drug currently in phase 2b studies for mitochondrial disease), may also be bene cial in COVID-19 patients (Fig. 7A). Furthermore, using the STAT3-CKO mouse model, we demonstrated that the COX-inhibitor, ibuprofen, reduced PGE2 production in male cardiomyocytes (Fig. 7A). We also provide evidence that enhanced physical activity lowers PGE2 in the elderly and may thereby support their immune systems in ghting SARS-CoV-2 infection (Fig. 7A).
In conclusion, our data suggest that PGE2 plays an important role in severe COVID-19 disease courses, either induced by SARS-CoV-2 or produced by endogenous and exogenous risk factors (Fig. 7A). Mechanistically, we show that PGE2 in COVID-19 disease speci cally targets B-cells by reducing PAX5, a key factor for B-cell proliferation and differentiation (Fig. 7A+B). Reducing PGE2 preventively and/or during COVID-19 disease may therefore provide a valuable therapeutic strategy to prevent and ght SARS-CoV-2 infection and to enhance and prolong immunity.

Limitations of the study
Limitations of our study include the limited numbers of blood samples from COVID-19 patients and that clinical data on COVID-19 patients, i.e. as C-reactive protein (CRP), lactate dehydrogenase (LDH), leukocytes normal count, neutrophils normal count, and lymphocytes were not available for all patients.

Materials And Methods
Unless otherwise stated, chemicals and reagents were all purchased from Sigma-Aldrich.

Study design
The aim of the study is to determine the secretion of circulating PGE2 levels in dependence of several risk factors in mild and severe diseased COVID-19 patients compared to healthy individuals, and how PGE2 modulates the host's immune response.
COVID-19 study: In this study of 89 patients diagnosed with COVID-19, 41 presented with mild/moderate symptoms and 48 were hospitalized with severe disease. Blood samples were also obtained from male (n=18) and female subjects (n=28) (age 18-50 years) from a healthy population established by Hannover Uni ed Biobank (HUB).
The local ethics committees at Hannover Medical School, Comité d'Ethique Hospitalo-Facultaire of UCLouvain, and the Ethical Committee of IEO has been obtained (IEO1271) approved this study. All patients and healthy control subjects provided written informed consent. The study conforms to the principles outlined in the Declaration of Helsinki.
Physical assessment and exercise program in healthy elderly individuals (Rebirth 60plus cohort, DRKS00013885) All subjects in the Rebirth 60plus cohort (DRKS00013885) were initially tested for maximum power output on a cycle ergometer with graded exercise test (GXT). Based on their activities, physical tness and pathologies, each subject was given an aerobic exercise training program. Once a month, the subjects were contacted by phone to assess training progress and adjust the exercise program, if necessary. All subjects of the Rebirth 60plus study were informed about bene ts and risks regarding all study procedures. Height and weight were measured using a scale (seca gmbh & co. kg, Hamburg, Germany). Body fat was measured with a medical Body Composition Analyzer mBCA (seca gmbh & co. kg, Hamburg, Germany). The physical activity was tracked using a GPS watch Forerunner 30 (Garmin Deutschland GmbH, Munich, Germany) and a daily diary where all physical activities were additionally documented. All study procedures were approved by the local ethics committee of Hannover Medical School (Vote #7617) and all subjects provided informed written consent prior to the commencement of the study procedures.
Mice study: Mice with a cardiomyocyte-restricted knockout of STAT3 (CKO: aMHC-Cre tg/+ ; STAT3 ox/ ox ) and wildtype mice (WT: STAT3 ox/ ox ) were used to analyze the effect of an altered androgen receptor signaling in male mice for cardiac PGE2 secretion and modulation of the immune response. Application of the COX inhibitor ibuprofen as therapeutic strategy was tested.
Mice were housed in a speci c pathogen-free barrier facility and fed standard chow. All animal studies were conducted in accordance with the German animal protection legislation and with the European Communities Council Directives 86/609/EEC and 2010/63/EU for the protection of animals used for experimental purposes. All experiments were approved by the Local Institutional Animal Care and Research Advisory Committee and permitted by the relevant local authority for animal protection "Niedersächisisches Landesamt für Verbrauerschutz und Lebensmittelsicherheit" (LAVES).

Blood sampling and blood tests
Blood samples were collected in S-Monovette® tubes containing ethylenediaminetetraacetic acid (EDTA, for plasma) or clot activator (for serum) at the time of hospitalization or at study inclusion (baseline, BL) and at the follow-up (FU) visits after 12 months for the Rebirth 60Plus male and female subjects (age >60 years). Blood samples were also obtained from young male and female subjects (age 18-50 years) from a healthy population established by Hannover Uni ed Biobank (HUB). Plasma or serum was separated by centrifugation at 1500 rpm for 10 min and aliquots were stored at -80 °C. Laboratory workup was performed as part of routine analysis by hospital laboratories for leukocytes, neutrophils, lymphocytes, platelets, CRP and LDH. PGE2 serum and plasma levels were measured using the prostaglandin E2 ELISA kit (abcam ab133021) according to the manufacturer's protocol. Since the PGE2 ELISA system used in the study is suitable for research use only and is not intended for diagnostic use, we showed relative PGE2 expression in %. IL-17A serum levels were measured using the IL-17A ELISA kit (abcam ab216167). PGE2 detection in supernatants of Calu-3, hCEC and murine adult cardiomyocytes PGE2 levels in the supernatants of the primary hCEC (normalized to cell density), murine adult cardiomyocytes (normalized to aMHC mRNA expression), and the cell lines Calu-3 (normalized to total RNA content) were measured using the prostaglandin E2 ELISA kit (abcam ab133021) respectively, according to the manufacturer's protocols.

Isolation of RNA and qRT-PCR
Total RNA was isolated with TRIzol (Thermo Fisher Scienti c) and cDNA synthesis was performed as described previously 62 . Real-time PCR with the SYBR green dye method (Brilliant SYBR Green Mastermix-Kit, Thermo Fisher Scienti c) was performed with the AriaMx Real-Time PCR System (Agilent Technologies) as described 62 . Expression of mRNA levels was normalized using the 2-ΔΔCT method relative to 18S, B2M and GAPDH. A list of qRT-PCR primers used in this study is provided in the supplements le Table S1, S2. RNA Isolation from formalin xed and para n embedded tissue RNA isolation from formalin-xed and para n embedded tissue was performed using the Maxwell ® RSC RNA FFPE Puri cation Kit (Promega Corporation, Madison, WI). RNA content was measured by using the Qubit RNA IQ Assay (Thermo Fisher Scienti c, Waltham, MA).
Isolation, characterization and culture of Sca-1 + cardiac progenitor cells Mice with a cardiomyocyte-restricted knockout of STAT3 (CKO: aMHC-Cre tg/+ ; STAT3 ox/ ox ) and wildtype mice (WT: STAT3 ox/ ox ) were generated and isolation and cultivation of Sca-1 cardiac progenitor cells from hearts of 3 month-old mice was performed as described 63 39 . Isolated WT-CPC were incubated with PGE2 (1 µM) for 48 h and were harvested in TRIzol.
Male CKO mice at the age of 9-10 weeks were treated with ibuprofen (10 mg/kg bodyweight, ibuprofen sodium salt, Sigma-Aldrich, dissolved in drinking water) for two weeks 39 . CKO control animals received drinking water.
Freshly isolated Sca-1 positive CPC were stained with CD19-PE (BD Pharmingen, 557399) for 15 min at room temperature. Flow cytometry was performed using the FACSCalibur (BD Biosciences).

Immunostaining
For immunostainings using antibodies recognizing CD19 (CD19-PE, BD Pharmingen, 557399) and Sca-1 (Sca-1-FITC microbead kit, Miltenyi Biotec), cryosections were xed in acetone, washed 3 times with PBS and blocked with 10 % donkey serum and 0.3 % Triton in PBS for 1 h at room temperature. Cryosections were stained with the antibodies for 2 h at room temperature. Nuclei were stained with DAPI Hoechst 33342 (Sigma-Aldrich). Images were acquired with Axio Observer 7 and Zen 2.6 pro software (Carl Zeiss).

Statistical Analyses
Statistical analysis was performed using GraphPad Prism version 5.0a, 7.0 and 8.1.2 for Mac OS X (GraphPad Software, San Diego, CA, USA).
Normal distribution was tested using the D'Agostino normality test. Continuous data were expressed as mean ± SD or median and interquartile range (IQR), according to the normality of distribution. Comparison between two groups was performed using one sample t-test or unpaired two-tailed t-test for Gaussian distributed data and the Mann-Whitney-U test where at least one column was not normally distributed. When comparing more than two groups, ANOVA and Bonferroni's post hoc test or Dunnett's post hoc test were used according to the normality of distribution. Categorical variables are presented as frequencies (percentages) and compared using Fisher's exact test. A two-tailed P value of <0.05 was considered statistically signi cant. Correlation for BMI, BW, body fat content and age was analyzed via ozone correlation analysis by using Pearson correlation coe cients for Gaussian distributions or for nonparametric Spearman correlation coe cients for non-normal distribution.    HCEC ctrl PBS, unpaired two-tailed t-test. Circulating PGE2 levels in healthy individuals in relation to sex and age. (A) The dot plots summarize relative circulating serum PGE2 levels (in %) of males (n=18) and females (n=28) below the age of 50 years; the median for males was set at 100 %. (B) The dot plots summarize relative circulating serum PGE2 levels (in %) of males (n=18) <50y and males (n=28) >60y; the median for males <50y was set at 100 %. (C) Dot plots summarize relative circulating serum PGE2 levels (in %) of females (n=28) <50y and Modulators of PGE2 synthesis and degradation are SARS-CoV-2 infection, but also air pollution, male sex, physical inactivity, and older age, which are all risk factors for more severe COVID-19 disease courses <sup>7, 13, 14, 15</sup>. These modulators upregulate the expression of the PGE2-generating enzyme