Prevention of hypoxia by myoglobin expression in human tumor cells promotes differentiation and inhibits metastasis

Abstract

As a tumor grows, it requires increased amounts of oxygen. However, the tumor blood vessels that form to meet this demand are functionally impaired, leading to regions of hypoxia within the tumor. Such hypoxia is one of the hallmarks of malignancy and is thought to promote a number of tumorigenic properties. Here, we sought to determine how tumors without hypoxia would progress by engineering A549 human lung carcinoma cells to ectopically express myoglobin (Mb), a multifunctional heme protein that specializes in oxygen transport, storage, and buffering. Mb expression prevented the hypoxic response in vitro and delayed tumor engraftment and reduced tumor growth following xenotransplantation into mice. Experimental tumors expressing Mb displayed reduced or no hypoxia, minimal HIF-1alpha levels, and a homogeneously low vessel density. Mb-mediated tumor oxygenation promoted differentiation of cancer cells and suppressed both local and distal metastatic spreading. These effects were primarily due to reduced tumor hypoxia, because they were not observed using point-mutated forms of myoglobin unable to bind oxygen and they were abrogated by expression of a constitutively active form of HIF-1alpha. Although limited to xenograft models, these data provide experimental proof of the concept that hypoxia is not just a side effect of deregulated growth but a key factor on which the tumor relies in order to promote its own expansion.

Figure 1. Mb expression attenuates the hypoxic response of cancer cells.

(A) Lentiviral vector–transduced cells were incubated in normoxia (21% O₂) or hypoxia (1% O₂) for 24 hours, and intracellular ATP levels were determined on total cell lysates. Statistical significance was calculated between each experimental sample and the empty vector (EV) control at 21% O₂. *P < 0.05; **P < 0.01. (B) Cells were incubated in normoxia or hypoxia, and lactate concentration was determined in the conditioned medium. (C) Cells were incubated in normoxia or hypoxia and then resuspended in normoxic medium. Oxygen consumption was measured in a sealed chamber using a Clark-type electrode. (D) Cells were incubated at different pO₂ values for 3 hours, and HIF-1α levels were determined by Western blotting using anti–HIF-1α antibodies. (E) Cells were subjected to repeated cycles of hypoxia and reoxygenation (1 hour at 21%, 3%, 1%, or 0.1% O₂ followed by 30 minutes at 21% O₂, repeated 3 times), and HIF-1α levels were determined by Western blotting. Cells incubated at 21% O₂ were used as controls.

Figure 2. Mb expression inhibits tumor growth.

Lentiviral vector–transduced A549 lung carcinoma cells were injected subcutaneously into CD1 nu/nu mice, and the formation of experimental tumors was monitored over time. (A) Kaplan-Meier–like analysis of tumor latency. Statistical significance was calculated comparing the median values. (B) Analysis of tumor volume. Statistical significance refers to the average values after day 20.

Figure 3. Mb expression promotes tumor oxygenation.

(A) Approximately 5 weeks after tumor cell injection, intratumoral pO₂ was measured directly using a ruthenium probe. At least 6 distinct measurements in different tumor locations were performed for each mouse. Red bars indicate average values. M1, mouse 1. (B) Following i.p. injection of pimonidazole hydrochloride, mice were euthanized and tumors extracted for analysis. The extent of tumor hypoxia was determined by immunohistochemical analysis of tumor sections using anti-pimonidazole (PIMO) antibodies. Hypoxia was quantified by calculating the percentage of pimonidazole-positive area (identified by the dotted outlines) over the total area (mean ± SD; in parentheses). Original magnification, ×100.

Figure 4. Mb expression inhibits tumor angiogenesis.

(A) Immunofluorescence analysis of tumor vessels using anti-CD31 antibodies (red). Nuclei are stained with DAPI (blue). Original magnification, ×100. (B) Analysis of MVD. Statistical significance refers to the difference between the empty vector and Mb groups. **P < 0.01. (C) Analysis of intercapillary distance (ICD). (D) MLA of tumor vessels. The mean value is indicated by the inner (pink) circle. The outer circle indicates SD. The linear scale refers to mean vessel diameter.

Figure 5. Mb expression promotes tumor differentiation.

Rows 1–3, tumor sections were stained with H&E. Original magnification, ×100 (rows 1 and 2); ×50 (row 3). Inset shows histological preparation from a normal mouse lung. Original magnification, ×200. Row 4, tumor sections were stained with PAS–Alcian blue. Original magnification, ×50.

Figure 6. Mb expression inhibits metastatic spreading of tumor cells.

(A) Inguinal lymph node invasion was determined by immunohistochemical analysis of lymph node sections using anti-GFP antibodies. Values in parentheses refer to the number of GFP-positive cells per field analyzed (mean ± SD). Original magnification, ×200. (B) The amount of gfp DNA present in lungs was determined by TaqMan analysis of lung genomic DNA. Values are expressed as 2^–ΔΔCT using lung genomic DNA from a tumor-free mouse as reference sample. (C) Immunohistochemical analysis of lung sections using anti-GFP antibodies. A representative image for each group is shown. Original magnification, ×200. (D) Lung metastasis was determined as in B in mice bearing tumors of equal size (approximately 0.2 cm³). **P < 0.01.

Figure 7. Engineering mutant forms of Mb unable to bind oxygen.

(A) Alignment of helices E and F of mouse Mb, human Mb, human α-Hb (α-Hb), and human β-Hb (β-Hb). Iwate α-Hb and Hyde Park β-Hb are naturally occurring human Hb M mutants unable to transport O₂. Mouse HF8Y Mb is one of the Mb mutants used in this study. Numbers refer to the amino acid positions occupied by the first and last residues shown within the respective protein sequence. (B) Western blot analysis of A549 cells expressing wild-type or mutant Mb using anti-FLAG antibodies. Cells were also engineered to express GFP to track metastatic dissemination. HF8Y Mb, His(F8)Tyr Mb mutant; HE7A/HF8A Mb, His(F8)Ala/His(E7)Ala double mutant. (C) Growth rate analysis of lentiviral vector–transduced cells in the presence of 2% FBS.