Amplification of tumor hypoxic responses by macrophage migration inhibitory factor-dependent hypoxia-inducible factor stabilization

Abstract

Low oxygen tension-mediated transcription by hypoxia-inducible factors (HIF) has been reported to facilitate tumor progression, therapeutic resistance, and metastatic adaptation. One previously described target of hypoxia-mediated transcription is the cytokine/growth factor macrophage migration inhibitory factor (MIF). In studies designed to better understand hypoxia-stimulated MIF function, we have discovered that not only is MIF induced by hypoxia in pancreatic adenocarcinoma but MIF is also necessary for maximal hypoxia-induced HIF-1alpha expression. Cells lacking MIF are defective in hypoxia- and prolyl hydroxylase inhibitor-induced HIF-1alpha stabilization and subsequent transcription of glycolytic and angiogenic gene products. Moreover, COP9 signalosome subunit 5 (CSN5), a component of the COP9 signalosome previously reported to functionally interact with MIF, has recently been shown to interact with and stabilize HIF-1alpha. Our results indicate that MIF interacts with CSN5 in pancreatic cancer cells and that MIF-depleted cells display marked defects in hypoxia-induced CSN5/HIF-1alpha interactions. This functional interdependence between HIF-1alpha and MIF may represent an important and previously unrecognized pro-tumorigenic axis.

Figure 1

MIF transcription and secretion are enhanced by hypoxia, and patients with pancreatic adenocarcinoma have elevated levels of plasma MIF. A, MIA-PaCa-2 cells were transfected with nonsense (NS) or MIF siRNA. After 48 h, cells were placed in 0.5% FCS–containing media and placed under either normoxic or hypoxic conditions. After 24 h, cell supernatants were collected and concentrated. MIF protein levels were analyzed by Western blot. B, MIA-PaCa-2 cells were transfected with nonsense or MIF siRNA. After 48 h, cells were incubated for 16 h in either hypoxia or normoxia. RNA was isolated and cDNA was synthesized. Primers for MIF and β-actin were used in real-time PCR to determine levels of MIF transcript in the cells. Columns, ΔC_t for each condition between MIF and β-actin; representative of three independent experiments. C, HIF-1α^+/+ and HIF-1α^−/− MEFs were challenged with 1% hypoxia for the indicated times and MIF mRNA levels were assessed by Northern blotting. D, plasma from healthy donors and pancreatic adenocarcinoma patients collected before chemotherapeutic treatment were assessed for MIF by ELISA. *, P < 0.05, Student’s t test (two tailed).

Figure 2

A functional requirement for MIF in hypoxia-induced gene expression. A, MIA-PaCa-2 cells were transfected with nonsense or MIF siRNA. Whole-cell lysates were prepared from cells after a 24-h exposure to either normoxia or hypoxia and immunoblotted for LDH-A and inducible nitric oxide synthase (iNOS). B, MIA-PaCa-2 cells were transfected with nonsense or MIF siRNA for 48 h, at which time cells were incubated for the indicated amount of time under hypoxia. RNA was harvested, cDNA synthesis was done, and real-time PCR for β-actin and VEGF was done as described in Materials and Methods. Columns, ΔC_t of the average of duplicate reactions for each condition between VEGF and β-actin; representative of two independent experiments.

Figure 3

A requirement for MIF in hypoxia-induced HIF-1α expression. MIA-PaCa-2 cells were transfected with nonsense or MIF siRNA for 48 h. Cells were then incubated for 16 h in anoxia (A) or the indicated amount of time in hypoxia (B). Cytoplasmic (A) and nuclear (A and B) extracts were prepared and assessed for HIF-1α followed by stripping and reprobing for β-actin and/or MIF. C, MIF^+/+ and MIF^−/− murine embryonic fibroblasts were cultured in duplicate under nearly confluent conditions with or without 150 μmol/L CoCl₂ for 6 h. Nuclear extracts were immunoblotted for HIF-1α (top band, arrow) and then stripped and reprobed for β-actin. D, MIA-PaCa-2 cells were transfected with nonsense or MIF siRNA for 48 h. Conditioned medium from confluent MIA-PaCa-2 cells was mixed 1:1 with fresh medium, pretreated with isotype control mAb, anti-MIF mAb, or 100 μmol/L final concentration of the small-molecule inhibitor ISO-1, and used to replace media from nonsense- and MIF-transfected cells. After challenge for 4 h with 150 μmol/L CoCl₂, nuclear extracts were assessed for HIF-1α and β-actin expression. A to D, representative of at least two independent experiments.

Figure 4

HIF-1α destabilization by loss of MIF is 26S proteasome dependent. A, for RNA analysis of HIF-1α, MIA-PaCa-2 cells were transfected with nonsense or MIF siRNA oligos for 48 h. Cells were placed in either 21% or 1% O₂ for 16 h and RNA was prepared, reverse transcribed, and analyzed by real-time PCR for HIF-1α and β-actin. Columns, ΔC_t of the average of duplicate reactions for each condition between HIF-1α and β-actin; representative of two independent experiments. For proteasome inhibitor studies, MIA-PaCa-2 cells were transfected with nonsense or MIF siRNA. After 48 h, 10 μmol/L MG-132, a 26S proteasome inhibitor, was added to the indicated cells for 30 min before challenging for 6 h with either 150 μmol/L CoCl₂ (B) or hypoxia (C). Equal amounts of nuclear fractions were analyzed by Western blot analysis. HIF-1α and β-actin antibodies were used for immunoblotting.

Figure 5

MIF interacts with CSN5 and promotes CSN5/HIF-1α interaction. A, MIA-PaCa-2 cells were incubated with indicated amounts of CoCl₂ for 4 h. Cells were lysed and whole-cell extracts were analyzed by Western blotting for MIF CSN5 and β-actin or immunoprecipitated with a CSN5 antibody. CSN5 interaction with MIF was evaluated by Western blotting of coimmunoprecipitates with anti-MIF antibody. B, MIA-PaCa-2 cells were transfected with nonsense or MIF siRNA. Forty-eight hours later, cells were incubated with or without 150 μmol/L CoCl₂ for 4 h. Whole-cell extracts and nuclear extracts were obtained in parallel. Whole-cell extracts were immunoprecipitated with a CSN5 antibody. Nuclear extracts and coimmunoprecipitates were analyzed for HIF-1α by immunoblotting. C, forty-eight hours after MIA-PaCa-2 cells were transfected with nonsense or MIF siRNA, cells were incubated with or without 10 μmol/L MG-132 for 6 h. Both whole-cell extracts and nuclear extracts were obtained. Whole-cell extracts were immunoprecipitated with a CSN5 antibody. Nuclear extracts and coimmunoprecipitates were analyzed by Western blotting for HIF-1α whereas a small fraction of the whole-cell lysate was assessed for total CSN5. A to C, representative of at least three independent experiments.

Figure 6

Schematic of proposed mechanism of the coregulatory axis between MIF and HIF-1α in pancreatic adenocarcinoma. MIF expression is coordinately regulated by several intracellular and extracellular stimuli. Consistent with prior observations, at least some intracellular MIF exists in a bound state with the COP9 signalosome subunit CSN5 (19). During hypoxia or prolyl hydroxylase inhibitor challenge, MIF is necessary for the previously described binding and stabilization of HIF-1α by CSN5 (6). Once stabilized, HIF-1α directs the transcription of many protumorigenic enzymes and growth factors, among these MIF (23). We propose that the net result of this coregulation between MIF and HIF-1α is an amplification of hypoxia-initiated responses and subsequent hypoxia-independent autocrine and paracrine effects of MIF (7).