COX-2/PGE2 upregulation contributes to the chromosome 17p-deleted lymphoma

Prostaglandin E2 is specifically enriched in del(17p) and Alox8-deficient cells

Our previous work has shown that arachidonic acid (AA) was accumulated in murine B-cell lymphomas with Alox8-knockdown or chromosome 11B3 deletion, which is syntenic to human chromosome 17p13 [2]. Here, we collected peripheral white blood cells from B-cell chronic lymphocytic leukemia (B-CLL) patients at West China Hospital, and measured AA amount through liquid chromatography-mass spectrometry (LC-MS). We found that cellular AA levels were significantly higher in del(17p) CLL (n = 3) than no del(17p) ones (n = 5) (Fig. 1A). Given that AA can be oxygenated into diverse bioactive lipid metabolites, we wondered how these elevated AA levels would alter the AA-derived metabolites in del(17p) tumors. Therefore, we profiled the AA metabolites, produced through either lipoxygenase or cyclooxygenase pathway, in primary mouse B-cell lymphomas with or without 11B3 loss by LC-MS. As Fig. 1B shown, two AA derivatives oxygenated by the cyclooxygenase pathway, prostaglandin E2 (PGE2) and thromboxane B2 (TXB2), were increased in 11B3-deleted lymphoma cells, while multiple metabolites produced from arachidonate lipoxygenases, including hydroxyeicosatetraenoates (HETEs) were reduced in 11B3-deleted cells compared to Trp53 loss only lymphoma cells. Interestingly, the levels of several other ALOX products, such as several hydroxy docosahexaenoic acids (HDoHEs) derived from docosahexaenoic acid (DHA) were decreased. Altogether, the AA metabolism is largely disrupted in del(17p) tumor cells, with a downregulated lipoxygenase pathway and upregulated cyclooxygenase pathway.

Fig. 1: Prostaglandin E2 was upregulated in mouse Alox8- or chromosome 11B3 deleted lymphomas.figure 1

A AA levels in fresh patients’ white blood cells enriched with chromosome 17p-deleted CLL cells, compared to those without 17p deletion, measured by liquid chromatography-mass spectrometry (LC-MS). n = 3 for no del(17p) group and n = 5 for del(17p) group. Error bar represents “Mean with SEM”, *p < 0.05 (unpaired two-tailed t-test). B Profiling AA metabolites in mouse primary lymphoma cells with chromosome 11B3 deletion (11B3+/-), compared to Trp53-/- lymphoma cells, by LC-MS (n = 3). C AA levels in shAlox8 vs shRen Ba/F3 cells, measured by LC-MS and ELISA, respectively. n = 4 for shRen group and n = 8 for shAlox8 group. Error bar represents “Mean with SEM”, *p < 0.05 (unpaired two-tailed t-test). D AA levels in Alox8+/+ vs Alox8-/- Ba/F3 cells measured by LC-MS and ELISA, respectively. n = 6 for Alox8+/+ group and n = 4 for Alox8-/- group. Error bar represents “Mean with SEM”, ***p < 0.001 (unpaired two-tailed t-test). E Profiling AA metabolites in primary shAlox8 lymphoma cells, compared to shTrp53 lymphoma cells, by LC-MS (n = 3). F PGE2 levels in Myc; shTrp53-shRen (shRen) or Myc; shTrp53-shAlox8 (shAlox8) lymphoma cells, measured by ELISA. n = 4 for shRen group and n = 6 for shAlox8 group. Error bar represents “Mean with SEM”, *p < 0.05 (unpaired two-tailed t-test).

Given that there are five ALOX genes on chromosome 17p, next we tested whether ALOX15B loss would be able to mimic del(17p) and directly led to the AA metabolism dysregulation. In a previous study, we showed that AA was increased in mouse NIH3T3 fibroblast cells with Alox8 knockdown, homologous to human ALOX15B [2]. Here, we confirmed this observation in Ba/F3 cells, a murine pro-B cell line that is more relative to B cell lymphoma. Ba/F3 cells infected with shAlox8 had significantly higher levels of AA than that in cells with shRen, measured by LC-MS (Fig. 1C). Similar results of higher AA levels were observed in the corresponding medium of Ba/F3 cells with shAlox8 (Supplementary Fig. 1A). Further, we generated Ba/F3 cells with Alox8 loss-of-function mutations by CRISPR/Cas9 technology, verified by T7 endonuclease I assay and Sanger sequencing (Supplementary Fig. 1C and D). The cellular and medium AA levels were significantly higher in Alox8-mutated cells (Alox8-/-) than in control cells (Fig. 1D and Supplementary Fig. 1B), consistently with what we observed in shAlox8 cells. Then we profiled the AA metabolites by LC-MS in primary B-cell lymphoma cells with or without shAlox8. In consistent with the finding in del(11B3) lymphoma cells, multiple AA downstream metabolites produced from cyclooxygenase pathway, such as PGE2, PGD2 (Prostaglandin D2) and TXB2, were upregulated in Alox8-deficient lymphoma cells (shAlox8;Myc), compared with control lymphoma cells (shTrp53;Myc) (Fig. 1E). The accumulation of PGE2 in tumor cells with Alox8 loss was further confirmed by enzyme-linked immunosorbent assay (ELISA) (Fig. 1F). These data suggested that Alox8 loss contributed to the enrichment of AA and it derives like PGE2 via cyclooxygenase pathway in tumor cells with del(17p).

RNA-seq analyses reveal a transcriptomic similarity between Alox8-deficient cells and those treated with PGE2

To decipher the molecular consequences of Alox8 loss, we analyzed the transcriptomes of pre-B cells with Alox8 or control Renilla shRNAs by RNA sequencing (RNA-seq). Both the unsupervised hierarchical clustering and principal component analysis (PCA) plot showed that the RNA-seq results of pre-B cells with shAlox8 were grouped together, distinguished from shRen cells (Fig. 2A and Supplementary Fig. S2A). Gene ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enrichment analysis show that 154 GO terms and 26 KEGG pathways were significantly downregulated in shAlox8 pre-B cells, compared to control shRen cells. Among them were immune response-related pathways such as cytokine-cytokine receptor interaction, chemokine signaling pathway, cellular response to interleukine-1, response to lipopolysaccharide, and the apoptosis pathway significantly downregulated in shAlox8 cells (Fig. 2B and Supplementary Fig. S2B). Gene set enrichment analysis (GSEA) results revealed that the upregulated genes by PGE2 treatment were positively enriched in shAlox8 pre-B cells, compared to shRen cells (NES = 1.135). In contrast, the downregulated genes by PGE2 treatment were negatively enriched (NES = −1.119) (Fig. 2C) [16]. The transcriptomic similarity further confirmed the altered cyclooxygenase pathway by Alox8 loss.

Fig. 2: Transcriptome analysis revealed that inhibition of Alox8 has similar gene expression as PGE2-treated cells.figure 2

A Unsupervised clustering of RNA-seq data of shAlox8 or shRen pre-B cells. n = 4, B KEGG enrichment analysis of genes downregulated in shAlox8 cells comparing to shRen cells. CE Gene set enrichment analysis of genes belonging to the top 100 differentially expressed genes with PGE2 treatment (C), HALLMARK_P53_PATHWAY and HALLMARK_APOPTOSIS gene sets (D) or GO_LYMPHOCYTE_DIFFERENTIATION and GO_B_CELL_DIFFERENTIATION gene sets (E) in shAlox8 cells comparing to shRen cells. NES, normalized enrichment score; FDR, false discovery rate. F The percentage of Annexin V + 7-AAD- apoptotic cells of pre-B cells treated with different concentrations of PGE2, measured by flow cytometry. n = 3. *p < 0.05 (unpaired two-tailed t-test). G, H The percentage of IgM+ in B220 + B cells as measured by flow cytometry analysis. n = 3. Error bar represents “Mean with SEM”, **p < 0.01(unpaired two-tailed t-test).

PGE2 increase prevents cellular apoptosis and B-cell differentiation

Gene set enrichment analysis (GSEA) results also showed that HALLMARK_P53 PATHWAY (NES = −1.546, FDR q = 0.013) and HALLMARK_APOPTOSIS pathway gene sets (NES = −1.560, FDR q = 0.011) were significantly negatively enriched in the transcriptome of shAlox8 cells compared to that of shRen cells (Fig. 2D). Multiple lymphocyte differentiation and B cell differentiation gene signatures were negatively enriched in Alox8-deficient pre-B cells (Fig. 2E). The enrichment of these pathways might underlie the anti-apoptosis and differentiation block effects of Alox8 deficiency.

To test whether upregulated PGE2 by Alox8 loss contributed to apoptosis and B-cell differentiation, we treated pre-B cells, the assumable cell-of-origin for Myc-driven B lymphoma in our model, with PGE2. The results showed that ectopic PGE2 prevented the apoptosis of pre-B cells, measured by Annexin V and 7-AAD staining, in a dosage-dependent manner (Fig. 2F). Further, we found that the differentiation of pre-B cells, indicated by IgM expression, was significantly repressed by PGE2 treatment (Fig. 2G, H). The survival advantage and differentiation block suggested that PGE2 might promote the transformation of B progenitor cells.

Cox-2 (Ptgs2) but not Cox-1 (Ptgs1) is induced by Alox8 deficiency

Cyclooxygenase is the key enzyme to catalyze AA into prostaglandins, including PGE2. There are two cyclooxygenases COX-1 and COX-2, encoded by PTGS1 and PTGS2, respectively. It is known that while PTGS1 is constitutively-expressed, PTGS2 gene transcription could be rapidly induced in response to stimuli [17]. Consistent with the increased levels of the cyclooxygenase pathway metabolites, we found that Ptgs2 but not Ptgs1 gene expression was dramatically induced up to 250-fold in Alox8-deficient lymphoma cells, compared to that in control shRen lymphoma cells (Fig. 3A). The synergistically increased expressions of PTGS2, PTGES and its metabolite PGE2 indicated a shift of the AA metabolism from the lipoxygenase pathway to the cyclooxygenase pathway.

Fig. 3: Ptgs2 was upregulated in Alox8-deficient lymphomas, which contributed to lymphomagenesis.figure 3

A Relative mRNA levels of Ptgs1(Cox-1) and Ptgs2(Cox-2) in shRen or shAlox8 lymphoma cells, measured by RT-qPCR. n = 3 for shRen group and n = 6 for shAlox8 group. *p < 0.05 (unpaired two-tailed t-test). B Knockdown efficiency of Ptgs2 by shPtgs2.562 and shPtgs2.3383 in shTrp53 lymphoma cells, measured by RT-qPCR. n = 3. Error bar represents “Mean with SEM”, *p < 0.05 (unpaired two-tailed t-test). C Kaplan–Meier tumor-free survival of recipient mice transplanted with tandem shRNAs for both Alox8 and Ptgs2. n = 6. *p < 0.05 (log-rank test). D Relative cell counts of Myc; shTrp53-shRen (shRen) or Myc; shTrp53-shAlox8 (shAlox8) lymphoma cells treated with 50 μM celecoxib or vehicle for 48 h. n = 3 for shRen group and n = 5 for shAlox8 group. Error bar represents “Mean with SEM”, *p < 0.05 (unpaired two-tailed t-test). E Relative cell counts of Trp53-/- or 11B3+/- lymphoma cells treated with 50μM celecoxib or vehicle for 48 h. n = 3 for shRen group and n = 4 for shAlox8 group.

Upregulation of cyclooxygenase pathway is required for Alox8 loss-driven lymphomagenesis

Because of the disrupted AA metabolism balance between the lipoxygenase pathway and the cyclooxygenase pathway in Alox8-deficient tumors and the known oncogenic functions of PGE2 in various cancers, we hypothesized that the enhanced cyclooxygenase pathway might mediate the tumor-promoting function of Alox8 loss. Two independent shRNA against Ptgs2 were cloned and validated by qRT-PCR (Fig. 3B). Ptgs2 shRNAs were cloned into a retrovirus-based vector carrying Myc cDNA and tandem shRNAs (shAlox8.2865-shPtgs2.562 or shAlox8.2865-shPtgs2.3383), with which Alox8 and Ptgs2 would be simultaneously repressed in the same cell with Myc overexpression. These tandem shRNA constructs were introduced into B progenitor pre-B cells, followed by transplantation into sub-lethally irradiated recipient C57BL/6 mice. The recipient mice were monitored daily by palpation, and the tumor-free survivals were recorded. Kaplan–Meier curve showed that recipient mice transplanted with shAlox8.2865-shRen transduced pre-B cells developed lymphomas with a median latency of 17 days, similar to those with shAlox8.2865 only [2]. However, compared to control shAlox8-shRen recipient mice, mice carrying shAlox8-shPtgs2 pre-B cells developed lymphomas with significantly longer tumor latencies (Fig. 3C; Median onset 27 and 39 days, respectively). Intriguingly, although the resulting lymphoma cells containing shAlox8-shPtgs2 as enriched GFP + cells in enlarged lymph nodes, these cells have upregulated Ptgs2 gene expression to a similar level as that in control shAlox8-shRen lymphoma cells (data not shown) and displayed similar histological features (Supplementary Fig. 3A, B), suggesting that shPtgs2 suppressive effect on Ptgs2 mRNA level had been bypassed or compensated during lymphomagenesis. Altogether, these results strongly indicated that the cyclooxygenase pathway upregulation, following Alox8 loss, was required for lymphomagenesis driven by Alox8 deficiency.

The AA metabolism alteration with the lipoxygenase-cyclooxygenase pathway unbalances implied that a gain of vulnerability for cancers with del(17p) or ALOX15B loss. As a proof-of-concept, we treated lymphoma cells with shRen or shAlox8 with celecoxib, a commonly used anti-inflammation drug inhibiting COX-2. The results showed that shAlox8 tumor cells were significantly more sensitive to COX-2 repression than the control cells (Fig. 3D). Further, we tested the inhibitory effect of celecoxib on mouse del(11B3) lymphoma cells. Tumor cells with del(11B3) also displayed a better response to celecoxib treatment than those with only p53 loss (Fig. 3E). The susceptibility of these Alox8 deficient tumor cells to celecoxib suggested that targeting the unbalanced AA metabolism might be a new treatment for cancers with del(17p) in patients.

The unbalanced lipoxygenase pathway and cyclooxygenase pathway in lymphoma patients with del(17p) suggest new susceptibility

Given that there are multiple tumor suppressor genes like p53 and PHF23 on chromosome 17p, we wondered how much ALOX15B repression and the resulted aberrant AA metabolism would contribute to del(17p) cancer. GSEA results showed that the genes upregulated in shAlox8 cells were significantly positively enriched in human chromosome 17p-deleted DLBCL cells (NES = 1.74, FDR q = 0.03). And in contrast, the genes downregulated in shAlox8 cells were significantly negatively enriched in del(17p) DLBCL (NES = −1.48, FDR q = 0.03) (Fig. 4A). These results suggest that Alox8 loss partially mimicked the molecular characteristics of del(17p). The cyclooxygenase pathway was upregulated in both Alox8 deficient and 11B3-deleted mouse lymphoma cells, indicated by increased levels of Ptgs2 expression and its product PGE2 (Figs. 1E and 1F). Here, we further tested the expressions of PTGS2 in human del(17p) tumors. Since the expression levels of genes on chromosome 17p are reduced in 17p-deleted tumors, we used the corrected expression levels of chromosome 17p13 genes to indicate 17p deletions. By analyzing the transcriptomes of 48 DLBCL samples in the TCGA cohort, we found that the expression levels of PTGS2 were significantly negatively correlated with those of chromosome 17p13 genes (Pearson = −0.46, p < 0.01) (Fig. 4B). A similar correlation was also observed in the TCGA-AML cohort (Pearson = −0.33, p < 0.001) (Fig. 4C). The negative correlation between 17p genes and PTGS2 was consistent with our hypothesis that the cyclooxygenase pathway of AA metabolism was upregulated in del(17p) cancers.

Fig. 4: Upregulation of cyclooxygenase pathway in lymphoma patients with del(17p) suggest new susceptibility.figure 4

A GSEA plots showing that genes were chosen from top 200 upregulated (shAlox8_UP) or downregulated (shAlox8_DOWN) genes in shAlox8 cells comparing to shRen cells significantly positively or negatively enriched in human chromosome 17p-deleted DLBCL cells. B, C Correlations between expression levels of PTGS2 and averages of chromosome 17p genes in TCGA-DLBCL or TCGA-AML. Data analyzed from TCGA-DLBCL (n = 48) or TCGA-AML (n = 187).

Taken together, our studies revealed the AA metabolism alterations in del(17p) cancers. The deficiency of the co-deleted 17p TSG ALOX15B led to a repressed lipoxygenase pathway and an enhanced cyclooxygenase pathway for AA metabolism (Fig. 5). Moreover, this metabolism abnormality implied a new susceptibility to this recalcitrant malignancy.

Fig. 5: Unbalanced arachidonate lipoxygenase and cyclooxygenase pathway in 17p-deleted lymphomas.figure 5

Schematic representation showing that arachidonate lipoxygenase and cyclooxygenase pathway in normal cell (left) or lymphoma cells with chromosome 17p deletions (right).

留言 (0)

沒有登入
gif