HIF-1 Alpha Regulates the Response of Primary Sarcomas to Radiation Therapy through a Cell Autonomous Mechanism

Abstract

Hypoxia is a major cause of radiation resistance, which may predispose to local recurrence after radiation therapy. While hypoxia increases tumor cell survival after radiation exposure because there is less oxygen to oxidize damaged DNA, it remains unclear whether signaling pathways triggered by hypoxia contribute to radiation resistance. For example, intratumoral hypoxia can increase hypoxia inducible factor 1 alpha (HIF-1α), which may regulate pathways that contribute to radiation sensitization or radiation resistance. To clarify the role of HIF-1α in regulating tumor response to radiation, we generated a novel genetically engineered mouse model of soft tissue sarcoma with an intact or deleted HIF-1α. Deletion of HIF-1α sensitized primary sarcomas to radiation exposure in vivo. Moreover, cell lines derived from primary sarcomas lacking HIF-1α, or in which HIF-1α was knocked down, had decreased clonogenic survival in vitro, demonstrating that HIF-1α can promote radiation resistance in a cell autonomous manner. In HIF-1α-intact and -deleted sarcoma cells, radiation-induced reactive oxygen species, DNA damage repair and activation of autophagy were similar. However, sarcoma cells lacking HIF-1α had impaired mitochondrial biogenesis and metabolic response after irradiation, which might contribute to radiation resistance. These results show that HIF-1α promotes radiation resistance in a cell autonomous manner.

Conflict of interest statement

CONFLICT OF INTEREST

The authors declare that there are no conflicts of interest.

Figures

Figure 1. Human soft tissue sarcomas accumulated nuclear HIF-1α after interdigitated chemo-radiation therapy

(A–C) Formalin-fixed paraffin embedded (FFPE) tissue microarray from human soft tissue sarcoma biopsies prior to chemo-radiation therapy and after chemo-radiation therapy was stained with antibody against HIF-1α. The HIF-1α stained tissue microarray was subsequently read by a pathologist blinded to the samples and scored based on HIF-1α staining intensity. A) HIF-1α nuclear accumulation score increased significantly with interdigitated chemo-radiation therapy using the Fisher’s exact test. B–C) Paired pre- and post-therapy human soft tissue sarcoma biopsies stained with an antibody to HIF-1α. Scalebar = 50μm. * p<0.05.

Figure 2. P7NP sarcomas were either UPS or RMS by histology, and had regions of HIF-1α accumulation and tumor hypoxia

A) Schematic of novel Pax7Cre-ER-T2/+; NrasLSL-G12D/+; p53FL/FL; ROSA26mTmG/mTmG (P7NP) mouse model of soft tissue sarcoma generated by intramuscular 4-hydroxytamoxifen injection. (B, C) FFPE sections of primary soft tissue sarcomas from P7NP mice were stained with hematoxylin and eosin. B) P7NP sarcoma with histology resembling undifferentiated pleomorphic sarcoma (UPS). Scalebar = 100μm. C) P7NP sarcoma with histology resembling rhabdomysarcoma (RMS). Scalebar = 100μm. D) Representative FFPE sarcoma section stained with antibody against HIF-1α and signal-amplified with DAB. HIF-1α staining of sarcomas from P7NP mice revealed regional heavy nuclear accumulation of HIF-1α. Scalebar = 200μm. (E–H) Representative frozen sarcoma whole tumor cross-section in a P7NP mouse that received intraperitoneal injection of EF5 and intravenous injection of Hoechst 33342 prior to tumor harvest. E) EF5 distribution in a whole cross-section of P7NP sarcoma showed areas of tumor hypoxia. F) Hoechst 33342 perfusion showed well-perfused tumor periphery and surrounding normal tissue, and poorly perfused areas in the tumor core. G) eGFP and tdTomato visualization showed eGFP positive tumor with surrounding tdTomato positive normal tissue as well as tdTomato positive stromal infiltration within the tumor. H) Overlay of E-G shows distribution of tissue hypoxia and perfusion relative to tumor and surrounding normal tissue. Scalebar = 2mm.

Figure 3. HIF-1α deletion in sarcomas from P7NPH1 mice was efficient and did not change tumor subtype or degree of tumor hypoxia

(A–C) Primary cells from P7NP and P7NPH1 tumors were dissociated into single cell suspensions with tissue proteases, and cultured deplete stromal cells. After 2–4 serial passages, primary tumor cells were collected to verify HIF-1α status by genomic DNA, RNA expression, and protein expression. A) Using primers specific for the HIF-1α conditional allele, the loxP flanked exon 2 of the Hif1a conditional allele was efficiently recombined (1-loxP) at the genomic level in sarcomas from P7NPH1 mice. B) qRT-PCR specific for exon 2-containing Hif1a mRNA transcript showed significant loss of expression in tumor cells from P7NPH1 mice when normalized to 18S by a two-tailed student t-test. C) Western Blot using antibody against HIF-1α in denatured cell lysates from tumors in P7NPH1 mice showed no detectable HIF-1α protein in the nuclear protein extracts after culture under 0.5% oxygen for 16 hours. Antibody probing against histone H3 served as nuclear protein loading control. (D–H) FFPE sections of primary P7NP and P7NPH1 tumors were stained with hematoxylin and eosin and scored as UPS or RMS by a pathologist blinded to the genotypes of the tumors. D) Histologic examination of P7NP and P7NPH1 tumors indicated deletion of HIF-1α did not alter the distribution of tumor histologic subtype. E–H) Representative hematoxylin and eosin stained FFPE sections of UPS and RMS from P7NP and P7NPH1 tumors. Scalebar = 50μm. (I–M) After primary tumors reached 600mm3, P7NP and P7NPH1 mice received intraperitoneal injection of EF5 and intravenous injection of Hoechst 33342 prior to tumor harvest. Tumors were evaluated for the presence of hypoxia by EF5 staining and for vascular perfusion by Hoechst 33342. (I–L) Representative frozen sarcoma whole tumor cross-sections with low levels of hypoxia and high levels of hypoxia in P7NP and P7NPH1 mice. Scalebar = 2mm. M) EF5 staining intensity of tumors using whole tumor cross-sections, normalized to tumor area, showed no significant variations between degree of tumor hypoxia from P7NP and P7NPH1 mice by a two-tailed student t-test. n.s. = not significant. **** p<0.0001.

Figure 4. Deletion of HIF-1α in the P7NP mouse model of soft tissue sarcoma did not affect in vivo tumor initiation but did sensitize primary tumors to radiation therapy both in vivo and in vitro

(A–C) P7NP and P7NPH1 mice were injected with intramuscular 4-hydroxytamoxifen in the hindlimb and followed for tumor formation and tumor growth kinetics. After the tumors reached 200mm3, a cohort of nonirradiated tumors was measured for tumor growth. A separate cohort underwent daily irradiation to 10Gy for 5 days. These tumors were followed for local control. A) Deletion of HIF-1α (P7NPH1) did not affect primary tumor onset as compared to tumors from P7NP mice by a log-rank test. B) Deletion of HIF-1α (P7NPH1) did not affect nonirradiated tumor growth kinetics compared to tumors from P7NP mice as measured by time to tumor tripling by a two-tailed student t-test. C) Deletion of HIF-1α (P7NPH1) sensitized tumors to fractionated radiation therapy as compared to tumors from P7NP mice by a log-rank test. (D–F) Primary cells from P7NP and P7NPH1 tumors were isolated as described in Figure 3. The growth rate and radiosensitivity as measured by clonogenic survival and apoptosis after radiation for primary tumor cells were compared in vitro under 21% oxygen. D) The growth rates were measured by cell counting at indicated times. Deletion of HIF-1α (P7NPH1) did not affect sarcoma cell proliferation as compared to sarcoma cells from P7NP mice in vitro by a two-tailed student t-test. E) Deletion of HIF-1α (P7NPH1) increased sarcoma cell radiosensitivity as compared to sarcoma cells from P7NP mice in vitro by a two-tailed student t-test. Inset: P7NP cells irradiated with 6Gy accumulate nuclear HIF-1α at 24 and 48 hours after irradiation. Histone H3 served as a control for the nuclear extract. F) The percentages of apoptotic cells after 6Gy were measured by incubating P7NP and P7NPH1 cells at indicated times with Annexin V antibody and evaluated with flow cytometry. Deletion of HIF-1α (P7NPH1) sarcoma cells resulted increased apoptosis 72 hours following 6Gy irradiation as compared to sarcoma cells from P7NP mice by a two-tailed student t-test. n.s. = not significant. * p<0.05. ** p<0.01. **** p<0.0001.

Figure 5. Deletion of HIF-1α did not affect DNA damage repair or cellular ROS clearance after irradiation

Primary cells from P7NP and P7NPH1 tumors were isolated as described in Figure 3. All experiments were performed with in vitro cultures under 21% oxygen. A) DNA damage repair after 10Gy irradiation was measured with the Comet assay. Cells were irradiated and collected at indicated times, and embedded in agarose. DNA damage was measured by the percentage of visible tail after electrophoresis. Deletion of HIF-1α (P7NPH1) did not affect non-homologous end-joining repair as compared to sarcoma cells from P7NP mice by a two-tailed student t-test. Representative images of the DNA tail were shown in the right panels. B) Cells were irradiated with 6Gy. Accumulation of reactive oxygen species (ROS) was measured first by incubating nonirradiated and irradiated cells with CellRox DeepRed reagent at 2, 24, and 72 hours, then by performing flow cytometry. Relative ratios of ROS were calculated by normalizing the irradiated samples to nonirradiated controls at each time point. Deletion of HIF-1α (P7NPH1) did not affect ROS accumulation as compared to sarcoma cells from P7NP mice by a two-tailed student t-test. C) Cells were irradiated with 6Gy. Accumulation of mitochondrial superoxide was measured first by incubating nonirradiated and irradiated cells with MitoSOX reagent at 2, 24, and 72 hours, then by performing flow cytometry. Relative ratios of mitochondrial superoxide were calculated by normalizing the irradiated samples to nonirradiated controls at each time point. Deletion of HIF-1α (P7NPH1) did not affect mitochondrial superoxide accumulation as compared to sarcoma cells from P7NP mice by a two-tailed student t-test. D) Two P7NP and P7NPH1 primary sarcoma cell lines were irradiated with 6Gy. Damage to cellular structures by ROS was measured by protein carbonylation at indicated times after irradiation using OxyBlot per the manufacturer’s instructions. Deletion of HIF-1α (P7NPH1) did not affect the rate of accumulation or resolution of ROS-induced protein carbonylation. ImageJ quantification of cabonylated proteins for each time point was calculated as a ratio to the 0h time point, and listed below each lane.

Figure 6. Deletion of HIF-1α did not affect changes in autophagy after irradiation

Primary cells from P7NP and P7NPH1 tumors were isolated as described in Figure 3. All experiments were performed with in vitro cultures under 21% oxygen. Cells were irradiated with 6Gy with and without adding 25μM chloroquine 2 hours prior to protein collection at the indicated times. Lysates were probed with an antibody against LC3A/B-I (top band) and LC3A/B-II (bottom band) to measure changes in autophagic flux, and BECLIN1 and ATG3 to measure other components of the autophagy pathway. Deletion of HIF-1α (P7NPH1) did not affect the changes in autophagic flux after irradiation.

Figure 7. Loss of HIF-1α was correlated with metabolic changes in sarcoma cells in response to irradiation

Primary cells from P7NP and P7NPH1 tumors were isolated as described in Figure 3. All experiments were performed with in vitro cultures under 21% oxygen. A) Cells were irradiated with 6Gy and lysates were collected at the indicated times after irradiation, and intracellular lactate was measured with mass spectrometry. The results showed P7NP cells had significantly increased intracellular lactate content at 24 hours after irradiation, both compared to nonirradiated P7NP cells and compared to P7NPH1 cells by a two-tailed student t-test. B) Cells were irradiated with 6Gy and lysates were collected at the indicated times after irradiation, and extracellular lactate was measured with the lactate colorimetric kit. The results showed P7NP cells had significantly increased extracellular lactate content at 0 and 2 hours after irradiation as compared to P7NPH1 cells by a two-tailed student t-test. (C–G) Oxygen consumption rates (OCR) were measured in cells that were irradiated at 6Gy at indicated times after irradiation. DMSO, oligomycin (Oligo), FCCP, antimycin A (Anti A), rotenone (Rot), and 2-deoxyglucose (2-DG) were added at indicated times. C) P7NP and P7NPH1 cells had similar OCR at baseline. D) P7NP and P7NPH1 cells both had reduced OCR at 8 hours following irradiation. E) P7NPH1 cells restored maximal electron transport chain (ETC) activity at 24 hours following irradiation and has more reserve capacity as compared to P7NP cells by two-tailed student t-test. F) P7NP cells restored maximal ETC activity at 72 hours following irradiation. G) Respiration in P7NP or P7NPH1 cells at 24 hours after irradiation was not inhibited by 2-DG. * p<0.05.

Figure 8. Loss of HIF-1α was correlated with decreased mitochondrial biogenesis in sarcoma cells in response to radiation

Primary cells from P7NP and P7NPH1 tumors were isolated as described in Figure 3. All experiments were performed with in vitro cultures under 21% oxygen. A) The mitochondrial density in P7NP and P7NPH1 cells in response to 6Gy irradiation was measured first by incubating nonirradiated and irradiated cells with the MitoTracker DeepRed reagent after 72 hours, then by performing flow cytometry. Signal from irradiated cells were normalized to nonirradiated controls. The P7NPH1 cells had significantly decreased mitochondrial density at 72 hours following irradiation when compared to P7NP cells by a two-tailed student t-test. B) The mitochondrial DNA (mtDNA) content was measured via real-time PCR using DNA extracted from purified P7NP and P7NPH1 sarcoma cells at 72 hours after 6Gy irradiation, and normalized to 18S and nonirradiated controls. P7NP cells had more mtDNA content at 72 hours following irradiation as compared to P7NPH1 cells by a two-tailed student t-test. * p<0.05. ** p<0.01.

Figure 9. Knockdown of HIF-1α sensitizes P7NP cells to radiation and leads to a radiation-induced mitochondrial biogenesis defect

Primary cells from P7NP tumors were isolated as described in Figure 3. All experiments were performed with in vitro cultures under 21% oxygen. A) Efficiency of Hif-1α knockdown by shRNA was measured by measuring HIF-1α protein level in nuclear extracts of parental P7NP sarcoma cells and P7NP cells stably expressing Hif-1α shRNA cultured under 0.5% oxygen for 24 hours. This showed that the efficiency of Hif-1α knockdown in P7NP sarcoma cells obtained from 3 different tumors was beyond the level of detection by Western Blot. B) Knockdown of Hif-1α by shRNA resulted in significantly increased radiation sensitivity compared to the parental controls after exposure to 6Gy under 21% oxygen as measured by clonogenic survival, by a two-tailed student t-test. Each symbol shape represents a parental control P7NP cells and its corresponding knockdown daughter sarcoma cells. C) The mitochondrial DNA (mtDNA) content was measured via real-time PCR using DNA extracted from parental P7NP and P7NP cells with Hif-1α knockdown at 72 hours after 6Gy irradiation, and normalized to 18S. P7NP cells had more mtDNA content at 72 hours following irradiation as compared to P7NP cells with Hif-1α shRNA, by a two-tailed student t-test. Each symbol shape represents a parental control P7NP and its corresponding knockdown daughter sarcoma cells. * p<0.05. ** p<0.01.