Hypoxia-mediated down-regulation of Bid and Bax in tumors occurs via hypoxia-inducible factor 1-dependent and -independent mechanisms and contributes to drug resistance

Abstract

Solid tumors with disorganized, insufficient blood supply contain hypoxic cells that are resistant to radiotherapy and chemotherapy. Drug resistance, an obstacle to curative treatment of solid tumors, can occur via suppression of apoptosis, a process controlled by pro- and antiapoptotic members of the Bcl-2 protein family. Oxygen deprivation of human colon cancer cells in vitro provoked decreased mRNA and protein levels of proapoptotic Bid and Bad. Hypoxia-inducible factor 1 (HIF-1) was dispensable for the down-regulation of Bad but required for that of Bid, consistent with the binding of HIF-1alpha to a hypoxia-responsive element (positions -8484 to -8475) in the bid promoter. Oxygen deprivation resulted in proteosome-independent decreased expression of Bax in vitro, consistent with a reduction in global translation efficiency. The physiological relevance of Bid and Bax down-regulation was confirmed in tumors in vivo. Oxygen deprivation resulted in decreased drug-induced apoptosis and clonogenic resistance to agents with different mechanisms of action. The contribution of Bid and/or Bax down-regulation to drug responsiveness was demonstrated by the relative resistance of normoxic cells that had no or reduced expression of Bid and/or Bax and by the finding that forced expression of Bid in hypoxic cells resulted in increased sensitivity to the topoisomerase II inhibitor etoposide.

Figures

FIG. 1.

Effect of anoxia on Bcl-2 family protein levels. (A) HCT116, HT29, and SW480 cells were incubated for 16 h under normoxic (−) or anoxic (+) conditions, and lysates were immunoblotted for Bcl-2 family proteins. Forty micrograms of protein per sample was loaded, and equal loading was confirmed by using actin as a control. (B) Immunohistochemical analysis of Bax in SW480 cells that were pelleted, fixed, embedded, and stained after incubation of cells in normoxic and anoxic conditions as described above. Results are representative of at least three independent repeat experiments.

FIG. 2.

Effect of anoxia on protein levels of BH-3-only members of the Bcl-2 family. (A) HCT116, HT29, and SW480 cells were incubated for 16 h under normoxic (−) or anoxic (+) conditions, and lysates were immunoblotted for Bad, Bid, and porin (as a mitochondrial protein control). (B) Immunohistochemical analysis of Bid in SW480 cells that were pelleted, fixed, embedded, and stained after incubation under the same conditions of normoxia and anoxia. (C) V79 Chinese hamster lung fibroblasts cells were incubated for 16 h under normoxic or anoxic conditions. Lysates were prepared and immunoblotted for Bid, Bax, and Bad. (D) SW480 cells were incubated for 16 h under normoxic, anoxic (2), and hypoxic (1 to 2% O2) conditions, and lysates were immunoblotted for Bax, Bad, and Bid. For Western blotting, 40 μg of protein per sample was loaded (equal loading was confirmed by using actin as a control [data not shown]). Results are representative of at least three independent repeat experiments.

FIG. 3.

Down-regulation of Bax and Bid in HCT116 tumors in vivo. HCT116 cells were grown as xenograft tumors in CD1 nude mice, and tumors were excised, fixed in formalin, and paraffin embedded, and 5-μm serial sections were prepared. Data show a typical tumor chord with hypoxic penumbra marked by pimonidazole (Pimo) immunostaining with reciprocal immunostaining for Bid or Bax.

FIG. 4.

Effect of anoxia on mRNA levels of proapoptotic Bcl-2 family members and assessment of HIF-1 proficiency in HCT116, HT29, and SW480 cell lines. (A) cDNA was synthesized from total RNA extracted from 16-h incubations of SW480 cells under normoxic conditions (white bars) or anoxic conditions (black bars). Changes in levels of actin, GAPDH, Bid, Bad, and Bax copy numbers were assessed by quantitative RT-PCR (Taqman) analysis. The mean (± SE) changes in gene copy number in the SW480 cell line (normoxia/anoxia, n = 3) are represented. (B) Data representative of at least three experimental repeats for the change in Bid mRNA levels in HCT116 and HT29 cell lines with actin serving as a housekeeping gene control. (C) Cells were transiently transfected with vectors containing HRE sequences from either control (empty vector) (white bars), PGK (grey bars), or LDH (black bars) cloned upstream of a luciferase transcriptional unit and exposed to anoxia or normoxia (see Materials and Methods). Luciferase activity was normalized to the measured activity of controls under normoxia. Data shown are means ± SEs (n = 3). (D) Lysates were prepared from HCT116, HT29, and SW480 cells incubated for 16 h under normoxic or anoxic conditions and immunoblotted for HIF-1α. (E) Lysates were prepared from HCT116, HT29, and SW480 cells incubated for 16 h under normoxic or anoxic conditions and immunoblotted for CA-IX. (F) HT29 cells were incubated for 16 h under normoxic or anoxic conditions and then pelleted and subjected to immunohistochemical analysis of CA-IX. For Western blotting, 40 to 80 μg of protein per sample was loaded (equal loading was confirmed by using actin as a control [data not shown]). Results are representative of at least three independent repeat experiments.

FIG. 5.

HIF-1-dependent down-regulation of Bid by oxygen deprivation. (A) HT29 cells and MEFs were transiently transfected with either a vector containing a HIF-1α scrambled sequence (−) or a vector containing a HIF-1α targeting sequence (+) 32 h prior to incubation under normoxic (−) or anoxic conditions (+) for 16 h. Lysates were then prepared and immunoblotted for Bid, Bax, Bad, and HIF-1α. (B) CHO wt (C4.5), HIF-1α-deficient (Ka13.5), and C15 (a pooled population control [see Materials and Methods]) cells and HEPA-1 wt, HIF-1β-deficient (c4), and mutant revertant (Rc4) cells were incubated for 16 h under normoxic or anoxic conditions, and the levels of Bid were assessed. (C) HEPA-1 wt, HIF-1β-deficient (c4), and mutant revertant (Rc4) cells were incubated for 16 h under normoxic or anoxic conditions and assessed for Bax and Bad levels. For all Western blots, 20 to 40 μg of protein per sample was loaded, and equal loading was confirmed by using actin as a control. (D) SW480 cells were incubated for 16 h under normoxic or anoxic conditions. Nuclear extracts were prepared and incubated with radiolabeled wt or mutant bid probes in the presence and absence of HIF-1α antibody and subjected to EMSAs (see Materials and Methods). Unlabeled oligonucleotides were used in competitive binding experiments to determine the specificity of the HIF-1α bid promoter complex. Arrows indicate the shift in complex size caused by the addition of the HIF-1α antibody (Ab). All results are representative of at least three independent repeat experiments.

FIG. 6.

Oxygen deprivation-induced, HIF-1-independent decreased expression of Bax and Bad. (A) SW480 cells were treated with the proteosome inhibitor MG115 (10 μM) or dimethyl sulfoxide (DMSO) as a solvent control and immediately incubated for 16 h under normoxic or anoxic conditions prior to analysis of HIF-1α, Bad, and Bax by Western blotting. Twenty to 40 μg of protein per sample was loaded, and equal loading was confirmed by using actin as a control. Data are representative of three independent repeat experiments. (B) SW480 cells were incubated for 16 h under aerobic or anoxic conditions, and the amount of rRNA actively involved in translation (associated with polysomes) was measured (see Materials and Methods). The data show a representative polysome profile from normoxic and anoxic cells, and the average amounts of rRNA in the polysomal fraction are depicted in the bar chart (n = 3).

FIG. 7.

Effect of oxygen deprivation on drug response. (A) Colon cancer cells were incubated for 8 h under normoxic (open diamonds) or anoxic conditions (black squares) prior to 8 h of treatment with a range of concentrations of etoposide (with continued normoxic or anoxic incubation). (B) V79 fibroblasts were deprived of oxygen for 16 h prior to treatment with 5-FU for 4 h. Colonies were counted after 10 to 14 days. Data shown are means ± SEs of three independent repeat experiments.

FIG. 8.

Impact of decreased Bid and/or Bax expression on response to etoposide. (A) SV40-transformed wt and Bid KO cells were incubated for 8 h under normoxic or anoxic conditions prior to 8 h of treatment with 50 μM etoposide (with continued normoxic or anoxic incubation), and apoptosis was assessed. Inset: Cells were incubated for 16 h under normoxic (−) or anoxic (+) conditions, and lysates were immunoblotted for Bid, Bax, Bad, and actin (as a loading control). Statistical analysis using Student's two-tailed t test showed significance (P ≤ 0.05) when comparing normoxic wt and Bid KO cells. (B) Normoxic HCT116 wt (black bar) and Bax null (white bar) cells were treated for 8 h with 1 mM etoposide, and apoptosis was assessed 2 days later. HCT116 Bax null cells were transiently transfected with Bid RNAi (grey bar) and then left for 24 h prior to 8-h treatment with etoposide, and apoptosis was assessed 2 days later. Inset: Western blot analysis to confirm that Bid was down-regulated during the course of the experiment. The apoptotic response was compared to that in HCT116 Bax null cells that were oxygen deprived 8 h prior to and after treatment with etoposide (hatched bar). For Western blots, 40 μg of protein per sample was loaded, and equal loading was confirmed by using actin as a control. (C) SW480 cells were transfected with GFP empty vector and immediately incubated for 8 h under normoxic (black bar) or anoxic (white bar) conditions prior to treatment for 8 h with 500 μM etoposide with continued incubation. GFP-expressing cells were isolated by FACS analysis and replated, and apoptosis was assessed 2 days later. SW480 cells transfected with GFP Bid were immediately incubated for 8 h under anoxia prior to 8-h etoposide treatment with continued anoxic incubation. GFP Bid-positive (grey bar) and -negative (hatched bar) cells were sorted by FACS analysis and replated, and apoptosis was assessed 2 days later. Statistical analysis using Student's two-tailed t test showed significance (P ≤ 0.05) when comparing GFP Bid-positive (grey bar) to GFP Bid-negative (hatched bar) cells, as well as to GFP vector anoxic cells (white bar). All data are plotted as means ± SEs (n ≥ 3). +ve, positive; −ve, negative.