The Beta Subunit of Hemoglobin (HBB2/HBB) Suppresses Neuroblastoma Growth and Metastasis

Metastasis is the major killer of patients with cancer. Bidirectional interaction between cancer cells and their microenvironment is a critical determinant of tumor progression and metastasis (1–5). Numerous reports in the last decade deal with mechanisms by which the microenvironment promotes tumor progression (6–9). In contrast, relatively little is known regarding inhibitory microenvironmental cells and molecules (10) with a few notable exceptions such as immunocytes and their products (11) or granulocytes (12).

Neuroblastoma is the most common extracranial solid tumor in children. Lung metastasis is a rare event (3%–4%) but its presence is clinically important because it signals a poor prognosis (13). Sixty percent to 70% of children with high-risk disease will ultimately experience relapse due to the presence of micrometastasis (14). As cure after relapse is extremely rare, novel modalities for the inhibition and elimination of neuroblastoma metastases are needed.

In a previous study (15), we demonstrated that the microenvironment of the normal lung possesses the capacity to restrain lung-metastasizing neuroblastoma cells and block their metastatic potential. Factors derived from normal mouse lungs significantly inhibited the viability of neuroblastoma lung-metastasizing cells by inducing cell-cycle arrest and apoptosis. Micrometastatic neuroblastoma cells (MicroNB), generated as described previously (16), were significantly more susceptible to this growth-restraining function than cells derived from frank neuroblastoma metastasis (MacroNB). The difference in susceptibility between micro- and macrometastatic cells raised the hypothesis that factors in the lung microenvironment exert antimetastatic functions including the maintenance of micrometastatic tumor cells in a state of growth arrest thereby blocking progression to overt lung metastasis. In the current study, we set out to isolate and characterize the lung-derived metastasis-inhibitory factor and probe its metastasis-restraining activity.

The human neuroblastoma lung micrometastatic (MicroNB) and macrometastatic (MacroNB) variants were generated using a mouse model for human neuroblastoma metastasis (17) from the parental cell lines MHH-NB11 (18) and SH-SY5Y (19) as detailed here (16), and were maintained in culture as described previously (17). Primary human pulmonary fibroblasts (HPF) were purchased from Promo-cell. Primary foreskin fibroblasts were generated from discarded foreskin tissue. Human pulmonary endothelial cells (HPEC; ref. 20) were kindly provided by Dr. V. Krump-Konvalinkova (Institute of Pathology, Johannes-Gutenberg University, Mainz, Germany). All other mentioned cell lines were purchased from the ATCC. Cells were authenticated every three months according to the ATCC guidelines, as detailed here (15). All cultures were periodically examined for mycoplasma contamination.

Lungs of 100 BALB/c athymic nude mice were used to prepare the lung-derived factors as described previously (15). Lyophilized lung-derived factors were reconstituted in Milli-Q purified water (EMD Millipore) to a concentration of 1 mg/mL, filtered (0.45 μm), and subjected to separation by Alliance reversed-phase high-performance liquid chromatograph (HPLC; Waters) system using Waters XBridge C18 column (30 × 150 mm, 5 μm) running a gradient of 35% to 50% acetonitrile (Thermo Fisher Scientific) in water containing 0.1% trifluoroacetic acid (Halocarbon, Inc.) at a flow rate of 15 mL per minute. The HPLC-separated fraction found to inhibit the viability of neuroblastoma cells was HPLC-purified running the same conditions.

The purified inhibitory factor was digested in Coomassie-stained polyacrylamide gel. Protein spots were excised from the gel and digested with trypsin according to the published procedures (21). The digested inhibitory factor was injected to a Thermo Electron Orbitrap Velos ETD mass spectrometer using a 8 cm × 75 μm Phenomenex Jupiter 10 μm C18 capillary column, and the peptides eluted from the column by an acetonitrile-0.1 mol/L acetic acid gradient at a flow rate of 0.5 μL per minute over 30 minutes. The digest was analyzed using the double play function acquiring full mass spectra followed by ion spectra to determine molecular mass and amino acid sequence in sequential scans. The data were analyzed using the Sequest search algorithm against the Mouse International Protein Index (IPI).

Native human hemoglobin was dissolved in Milli-Q purified water (EMD Millipore) to a concentration of 1 mg/mL and filtered (0.45 μm). Human hemoglobin was then chromatographed by Alliance RP-HPLC system (Waters) using Waters XBridge C18 column (50 × 250 mm, 10 μm) running a gradient of 35% to 50% acetonitrile (Thermo Fisher Scientific) in water containing 0.1% trifluoroacetic acid (Halocarbon) at a flow rate of 40 mL per minute. The separated alpha and beta subunits of hemoglobin (HBA and HBB, respectively) were collected and purified by reverse-phase HPLC and their molecular masses were verified by electrospray ionization mass spectrometry (ESI-MS).

The inhibitory human HBB peptide (ENFRLLGNVLVCVLA) designated Metox, and a control peptide of a scrambled amino acid sequence (ANVLNECVFVGRLLL) designated scrambled-Metox, were chemically synthesized. The synthesis was performed on appropriate PAM resins (Applied Biosystems) on an 433A peptide synthesizer (Applied Biosystems) using an optimized HBTU (Oakwood Chemical) activation/DIEA (Sigma-Aldrich) in situ neutralization protocol for Boc solid-phase peptide synthesis (22). After chain assembly, the peptides were cleaved by anhydrous hydrogen fluoride (Airgas) in the presence of 5% p-cresol (Sigma-Aldrich) at 0°C for 1 hour, followed by precipitation with cold ether. The Metox and scrambled-Metox peptides were purified by reverse-phase HPLC, and their molecular masses were ascertained by ESI-MS.

Mice were orthotopically inoculated with MicroNB cells to the adrenal gland to generate local adrenal tumors and lung and bone marrow micrometastasis as described previously (16). Fourteen days after tumor cell inoculation, mice were treated intranasally with 15 mg/kg of the human HBB peptide, Metox, once a week for 8 weeks. The lyophilized peptide was dissolved prior to each administration in dimethyl sulfoxide (Sigma-Aldrich), diluted in sterile PBS, and filtered (0.2 μm). Mice were forced to inhale 20 μL of Metox (0.3 mg/mouse) or of the control scrambled peptide, scrambled-Metox (control group).

Paired or unpaired Student t test was used to compare in vitro and in vivo results.

For more details on Materials and Methods, see Supplementary Data

Soluble lung-derived factors induce growth arrest and apoptosis of lung-metastasizing human neuroblastoma cells (15). Here we isolated the inhibitory factor from mouse lungs. Soluble factors derived from the lungs of 100 athymic nude mice were generated as described previously (15). Dialysis of the lung-derived factors suggested that the molecular weight of the inhibitory factor(s) is higher than 3,500 Da (Supplementary Fig. S1A). The biologically active dialyzed lung-derived factors were separated by reverse-phase HPLC to numerous fractions (Fig. 1A). These fractions were incubated with micrometastatic (MicroNB) and macrometastatic (MacroNB) neuroblastoma cells for 72 hours (Supplementary Fig. S1B). An MTS-based viability assay indicated that one of the distinct separated peaks inhibited the viability of the cells by 25%–50% (P < 0.05; Fig. 1B) to the same extent as unseparated lung-derived factors (15).

The active inhibitory fraction was subjected to high-resolution purification using reverse-phase HPLC. This resulted in a single, narrow, and symmetric peak (Fig. 1C) representing, most probably, a single factor. ESI-MS analysis confirmed that the purified fraction was indeed a single factor of a molecular mass of 15,824.5 Da (Fig. 1C). This fraction reduced the viability of MicroNB and MacroNB cells by 65% (P < 0.01) and 35% (P < 0.05), respectively (Fig. 1D). Sequence analysis by LC/MS-MS coupled with tryptic digestion followed by a database search in the International Protein Index (IPI), identified the fraction as mouse hemoglobin subunit beta-2 (HBB2), a protein with 147 amino acid residues (Fig. 1E). N-terminal Edman degradation (23) of the isolated factor further verified its sequence identity.

The inhibitory activity of the isolated factors against MicroNB and MacroNB cells could be effectively neutralized by anti-mouse HBB2 antibodies, validating HBB2 as the inhibitory factor in mouse lungs (Fig. 1F). The inhibitory activity was not affected by an isotype control (Fig. 1F).

MicroNB cells were orthotopically inoculated to the adrenal gland of athymic nude mice generating local tumors. The control group was injected with PBS. Quantitative real-time PCR (qRT-PCR) analyses performed 8 weeks after the intra-adrenal inoculation of the tumor cells revealed the presence of micrometastatic human neuroblastoma cells in lungs, bone marrow, and liver of the inoculated mice (Supplementary Fig. S2). At this point, there was no evidence of overt metastasis.

qRT-PCR analyses indicated that the level of mouse HBB2 mRNA (Supplementary Table S1) was 25 times higher in lungs of micrometastasis-bearing mice than in lungs of normal mice (P < 0.001; Fig. 2A). Levels of the alpha subunit of hemoglobin (HBA) mRNA were low and similar in the two groups of mice (Fig. 2A). Immunofluorescence analyses of frozen lung sections stained with anti-mouse HBB2 antibody revealed a higher expression of HBB2 in lungs harboring neuroblastoma micrometastases than in normal lungs (Fig. 2B). In these lungs, an intracellular expression of HBB2 in cells lining the alveoli was observed (Fig. 2B). The higher expression of HBB2 in lungs harboring micrometastases was verified by Western blot analysis of lung tissue lysates (P < 0.005; Fig. 2C). HBB2 expression was also significantly higher in liver (P < 0.01) and bone marrow (P < 0.05) of micrometastasis-bearing mice than in liver and bone marrow of control normal mice (Fig. 2C). HBB2 concentration in the serum of micrometastasis-bearing mice was 7 times higher (P < 0.005) than its concentration in the serum of normal mice (Fig. 2D). The serum concentration of the alpha subunit, HBA, was very low and was similar in normal control mice and in micrometastasis-bearing mice (Fig. 2D). Similarly there was no significant difference between the concentration of the whole, intact hemoglobin protein in the serum of normal and micrometastasis-bearing mice (Fig. 2D). This result excludes the possibility that the high concentrations of HBB2 in serum of tumor bearers were due to hemolysis. It is not unlikely that the lungs are an important contributor to the elevated HBB2 serum levels, as the most significant elevation in the expression of HBB2 in micrometastasis-bearing organs was in the lungs (Fig. 2C; Supplementary Fig. S2). However, other micrometastasis-bearing organs such as bone marrow and liver (Fig. 2C; Supplementary Fig. S2) do contribute as well, to the increased HBB2 protein levels in the serum.

We next set out to identify the HBB2-producing lung cells and in particular those cells in whom transcription is upregulated in micrometastasis-bearing mice (Fig. 2A). It was previously reported that HBB2 is synthesized by pulmonary epithelial cells (24). Confirming these results, we found that such cells do indeed express HBB2. A higher expression of HBB2 was seen in epithelial cells lining the alveoli in lung sections of micrometastasis-bearing mice than in alveolar epithelial cells from normal mice (Fig. 3A; Supplementary Fig. S3). A higher expression of HBB2 was also seen in endothelial cells lining blood vessels of micrometastasis-bearing mice than in endothelial cells of control mice (Fig. 3A; Supplementary Figs. S3 and S4).

To verify that epithelial and endothelial cells from micrometastasis-bearing mice indeed produce higher levels of HBB2 than similar cells from control, normal mice, single-cell suspensions were prepared from lung tissues of control and micrometastasis-bearing mice using the GentleMACS dissociator (25). Lung cells were separated using magnetic-activated cell sorting (MACS) to isolate immune, epithelial, and endothelial cells using correspondingly the lineage cell markers CD45, CD326, and CD31 (26). Flow cytometry analysis was performed to verify the efficacy of the separation procedure (Fig. 3B). Confocal microscopy confirmed the expression of HBB2 protein in the sorted epithelial and endothelial cells (Fig. 3C). qRT-PCR analyses of the separated cell populations indicated that indeed HBB2 mRNA is produced by pulmonary epithelial and endothelial cells and not by pulmonary immunocytes (Fig. 3D). HBB2 mRNA expression was 30 times higher in pulmonary epithelial cells of micrometastasis-bearing mice than in control mice (P < 0.001). HBB2 mRNA expression was 5 times higher in pulmonary endothelial cells of micrometastasis-bearing mice than in pulmonary endothelial cells of normal mice (P < 0.05; Fig. 3D). An additional HBB2 mRNA–expressing pulmonary cell population (CD45/CD31/CD326-negative cells, possibly hematopoietic stem cells) was identified to express HBB2 mRNA. However, HBB2 expression by these cells was only 2 times higher in cells derived from micrometastasis-bearing mice compared with control mice (data not shown).

Expression of the alpha subunit of hemoglobin, HBA, was also seen in pulmonary epithelial and endothelial cells; however, there was no significant difference in HBA expression in the pulmonary cell populations derived from control and micrometastasis-bearing mice (Fig. 3D).

We next asked whether the elevated expression of HBB2 in pulmonary cells of micrometastasis-bearing mice is mediated by a direct interaction between these cells or their soluble products and micrometastatic neuroblastoma cells residing in the lungs. To answer this question, we cocultured MicroNB cells with human pulmonary endothelial cells or with human pulmonary fibroblasts in a Transwell system. Following the coincubation, total RNA was isolated from the endothelial cells and fibroblasts and a qRT-PCR was performed to examine human HBB expression in these cells. Expression levels of HBB mRNA were 8 times higher (P < 0.005) in the endothelial cells cocultured with soluble factors from MicroNB cells than in control cells (Fig. 3E). The expression of HBB mRNA was not altered in human pulmonary fibroblasts cocultured with MicroNB cells (Fig. 3E). Expression levels of the mRNA of human hemoglobin alpha chain, HBA, remained unchanged in endothelial cells cocultured with MicroNB cells (Fig. 3E).

The data summarized in this section clearly demonstrate that pulmonary epithelial cells and to a lesser degree pulmonary endothelial cells are an important source for the levels of HBB2 in the lungs of nude mice bearing human neuroblastoma micrometastases. Moreover, the results of the in vitro experiments demonstrate that neuroblastoma-derived soluble factors are capable to stimulate pulmonary endothelial cells to selectively upregulate the transcription of the hemoglobin beta chain.

The next set of experiments was aimed to find out if similarly to mouse HBB2, the beta subunit of human hemoglobin (HBB) would also inhibit the viability of human neuroblastoma cells. Native human hemoglobin was subjected to separation by reverse-phase HPLC, during which the alpha and beta subunits of hemoglobin were fully separated (Supplementary Fig. S5B) and purified. The separation was verified by mass spectrometry analysis (Fig. 4A; Supplementary Fig. S5C).

Whereas the alpha subunit (HBA) did not influence neuroblastoma cell viability, incubation with 100 μg/mL human HBB decreased the viability of MicroNB cells by 62% (P < 0.005) and of MacroNB cells by 39% (P < 0.01; Fig. 4B). These results were strikingly similar to the viability-inhibitory functions mediated by mouse lung-derived factors (15) and by lung-derived mouse HBB2 (Fig. 1B and D). The HBB-mediated inhibitory activity was dose dependent (Supplementary Fig. S5D). Intact human hemoglobin inhibited the viability of neuroblastoma cells, but not to the same extent as its beta subunit (Fig. 4B).

Human HBB was found to be inhibitory against several additional cancer cell lines (Table 1). The lung carcinoma cell line A549 and the melanoma cell line RALL were inhibited by all three doses of HBB used (1, 10, and 100 μg). The highest amount of HBB (100 μg), reduced the viability of the breast cancer cell lines T47D and MCF-7, the prostate cancer cell line 22RVi, the cervical cancer cell line Hela, and the melanoma cell line RKTJ. These results demonstrate that the HBB-mediated inhibition of viability is not restricted to neuroblastoma cells.

Human HBB did not influence the viability of the normal (transformed) cell lines HEK293T and human pulmonary endothelial cells, nor did it influence the viability of the normal (nontransformed) human foreskin and pulmonary fibroblasts (Table 1). HBB did not cause hemolysis of human erythrocytes (Supplementary Fig. S6).

Flow cytometry analysis of Annexin-V and PI apoptotic test indicated that in HBB-treated MicroNB and MacroNB cells, the percentage of cells in early apoptosis and late apoptosis was increased by 17% (P < 0.01) and 13% (P < 0.05), respectively, and that the percentage of necrotic cells was very low and did not change after treatment with HBB (Fig. 4C). However, the viability of HBB-treated MicroNB and MacroNB cells decreased by 62% and 39%, respectively (Fig. 4B). Apoptosis and/or necrosis are thus not the only mechanisms responsible for the decline in cell viability.

Flow cytometry analysis for cell-cycle progression revealed that HBB increased the fraction of MicroNB and MacroNB cells in the G₀–G₁ phase by 59% (P < 0.01) and 23% (P < 0.05), respectively (Fig. 4D). The increased percentage of cells in the G₀–G₁ phase was accompanied by a decrease in cyclin D1 protein level (P < 0.005; Fig. 4E).

In addition to apoptosis and growth arrest, HBB decreased ERK phosphorylation (P < 0.005) and increased p38 phosphorylation (P < 0.005), creating a low ERK/p38 signaling ratio in the tumor cells (Fig. 4F and G). The phosphorylation of TAK1 in HBB-treated cells was also increased (P < 0.01; Fig. 4H).

Taken together, these results indicate that human HBB induces apoptosis and cell-cycle arrest in neuroblastoma cells, leading to growth arrest of the tumor.

To identify the functional region of human HBB responsible for the viability-inhibitory activity on tumor cells, we cleaved the protein to N- and C-terminal fragments by cyanogen bromide (Fig. 5A). MTS-based viability assays showed that the C-terminal fragment is responsible for the growth arrest activity (Fig. 5B). The N-terminal fragment also exhibited an inhibitory effect, however, to a lower extent (Fig. 5B).

HBB was then synthesized in 14 peptide segments of 15 amino acids each (Supplementary table S2). Each segment was composed of 5 amino acids overlapping those of the preceding segment and 5 amino acids overlapping those of the following segment (Fig. 5C). Each of these segments was assayed for its ability to block the viability of MicroNB cells. While peptides 2, 3, and 8 stimulated tumor viability, peptide 11 with the amino acid sequence ENFRLLGNVLVCVLA (designated hereafter “Metox”), significantly inhibited the viability of the cells by 23%–70% depending on its concentration (Fig. 5D). It is not unlikely that the viability-inhibitory function of intact HBB is a net balance between the growth-promoting functions mediated by peptides 2, 3, and 8 and the growth-inhibitory functions of peptide 11.

Confocal microscopy suggested that FITC-labeled Metox is internalized into MicroNB and MacroNB cells (Fig. 5E). Flow cytometry indeed indicated membrane binding and cellular uptake of FITC-labeled Metox after 30 minutes of incubation (Fig. 5F). The uptake was inhibited at 4°C, implying that the process depends on endocytosis (Fig. 5F).

Athymic nude mice were orthotopically inoculated with MicroNB cells to the adrenal gland. Fourteen days after inoculation, mice were treated by the intranasal route (27) once a week for 8 weeks with 15 mg/kg Metox or with the same amounts of a control peptide having the identical amino acid composition as Metox but in a scrambled sequence (Fig. 6A).

Twenty days after tumor cell inoculation, a difference (P < 0.05) was apparent in the local tumor volume between mice treated with Metox and mice treated with scrambled-Metox (Fig. 6B). This difference became more significant with time (Fig. 6B). Seventy days after tumor cell inoculation, the weight of local tumors resected from Metox-treated mice was 12 times lower (P < 0.001) than that of local adrenal tumors resected from mice treated similarly but with scrambled Metox. (Fig. 6C and D).

qRT-PCR analyses indicated that the metastatic load of MicroNB cells was significantly lower (P < 0.001) in lungs and bone marrow of mice treated with Metox compared with that found in organs derived from mice treated with scrambled-Metox (Fig. 6E and F).

The results reported above showed that endogenous mouse HBB2 expression is upregulated in micrometastasis-bearing mice (Fig. 2C; Supplementary Fig. S2). In accordance with these results, the endogenous HBB2 mRNA expression levels in organs of mice treated with scrambled-Metox (a treatment that does not reduce tumor and metastasis load) were significantly higher (P < 0.01) than its expression levels in organs of mice treated with Metox (which reduces tumor and metastasis load; Fig. 6G and H).

Similar results were obtained in an additional in vivo experiment in which Metox was administered either by the intranasal route or intravenously to MicroNB cell–inoculated mice (Supplementary Fig. S7). Both forms of Metox administration inhibited the growth of local adrenal tumors and of lung and bone marrow metastasis; however, intranasal administration of Metox was more effective.

The current study is the first to report that the β subunit of hemoglobin belongs to the group of moonlighting proteins that are capable of performing multiple physiologic functions (24). In addition to its oxygen transport functions, HBB also exhibits an antitumor reactivity and as such joins the arsenal of innate resistance factors such as defensins (28, 29) that regulate cancer progression. Interestingly, a novel function was also recently reported for the α subunit of hemoglobin in regulation of the effects of nitric oxide in non-erythroid cells (30).

In patients with neuroblastoma, lung metastasis is a relatively late event (31). This delay may be caused by the inhibitory function of the lung-derived HBB restraining the further progression of lung-residing micrometastases. Overt neuroblastoma lung metastasis may develop if a subset of such lung-residing cells develops resistance to HBB or if the expression of HBB is downregulated. The latter possibility is supported by an Oncomine meta-analysis (32) of gene expression profiling accumulated by several research groups (33–37), demonstrating that a 17- to 40-fold decrease in HBB expression occurred in overt lung metastatic lesions as compared with its expression in normal lung tissues.

In addition to neuroblastoma cells, other tumor cells are sensitive to the proliferation-restraining function of HBB. Breast cancer, lung cancer, and melanoma are among the sensitive cancer types. However, different tumors belonging to a certain cancer type display a heterogeneous response to the growth-retaining function of HBB; only 2 of the 5 breast cancer cell lines tested were sensitive. Ongoing experiments are aimed to identity the common factor that confers HBB sensitivity to the growth-restraining function of this protein upon different types of tumor cells.

The bearing of a local neuroblastoma tumor and of micrometastasis triggered an adaptive enhanced synthesis of HBB2 by pulmonary epithelial cells and to a lesser degree by pulmonary endothelial cells. An upregulated synthesis of HHB2 was also detected in bone marrow and liver cells. Assuming that endothelial cells in these organs are another source for circulating HBB2 and the fact that endothelial cells constitute a very large overall mass in the body, these cells could play a significant role in resisting the propagation of neuroblastoma (and other tumors). Whereas the synthesis of hemoglobin or its subunits by nonerythroid cells such as pulmonary epithelial cells, mesangial cells in the kidney and neurons in the brain was reported (24, 38–42), we are not aware of studies reporting the synthesis of the beta subunit by pulmonary endothelial cells. Interestingly, cathepsin proteases capable of proteolytic degradation of both α- and β-globin are also expressed by pulmonary epithelial cells, where these proteases are involved in post-translational processing of surfactant proteins (43–45).

The upregulated synthesis of HBB2 is apparently mediated by a direct contact between these host cells and soluble factors derived from the tumor cells; in vitro experiments demonstrated that coculturing pulmonary endothelial cells with culture supernatants of tumor cells stimulated HBB synthesis by the former cells but not by pulmonary fibroblasts.

What drives the upregulation of HBB2 in micrometastasis bearing mice? First we demonstrated that the upregulation of HBB2 is transcriptional and occurs at the mRNA level as well as at the protein level. We then experimentally excluded the possibility that hemolysis occurs in nude mice bearing human neuroblastoma xenografts. Free hemoglobin is therefore not the source for the upregulated expression of HBB2 in these mice. By demonstrating that the concentrations of nitric oxide metabolites, nitrite, and nitrate were similar in the lungs and serum of normal and micrometastasis-bearing mice (Supplementary Fig. S8), we also excluded the possibility that free hemoglobin sequestered nitric oxide, depleting its amounts and causing endothelial dysfunction (41). On the other hand and as indicated above, we provided evidence that the upregulation of HBB2 is triggered in pulmonary cells by tumor-derived factors.

The adaptive upregulated expression of HBB2 in tumor-bearing mice suggests that this protein may alert for danger signals delivered by invading foreign cells such as microorganisms or cancer cells sharing patterns that are recognized by HBB2 (46). The in vivo experiments performed in this study indicate that the upregulated levels of HBB2 in tumor and metastasis bearers are insufficient to eradicate micrometastatic tumor cells and that an exogenous administration of HBB2 or of its derivative Metox is needed to efficiently halt metastasis formation.

The elucidation of the mechanism underlying the growth-restraining activity of HBB and its derived Metox peptide is outside the scope of the current study. We do speculate, however, that the presence of HBB2 at the apical surface of endothelial and epithelial cells may indicate that HBB2 is secreted from these cells. The fact that proliferation inhibitory and proapoptotic signaling were activated in tumor cells by HBB2 also supports the suggestion that this HBB2/HBB-mediated signaling is activated by secreted forms of these proteins. The findings that HBB induces both apoptosis and cell-cycle arrest of tumor cells, that the HBB-derived Metox binds the outer membrane and is internalized into the tumor cells, and that HBB-activated TAK1 and P38, downregulated ERK phosphorylation, and Cyclin D1 stability serve as basis for a working hypothesis as to its mode of action. We hypothesize that these growth arrest–inducing activities are mediated by binding of soluble HBB2/HBB to a yet unidentified HBB receptor. Such a receptor could facilitate the internalization of HBB and Metox into the target cells.

A low ERK/P38 phosphorylation ratio may lead to tumor dormancy (47). Although other signaling mechanisms that do not induce dormancy may also act in conjunction with a low ERK/p38 signaling ratio (48, 49), we hypothesize that in addition to apoptosis, HBB induces tumor dormancy. Future work will confirm or negate this hypothesis.

The current study provides proof of concept that microenvironmental control, in the form of a naturally occurring protein, HBB, exerts proliferation-restraining functions (apoptosis and cell-cycle arrest) on tumor cells; neuroblastoma being the case in point. Similar microenvironmental control mechanisms that block the proliferation of incipient cancer cells are postulated to operate in healthy people (10, 50).

The bioactive tumor-restraining region of HBB (ENFRLLGNVLVCVLA) was found to exert significant antitumor and antimetastasis activities both in vivo as well as in vitro. This peptide offers promising opportunities for the development of novel therapies for the treatment of both primary as well as residual disease. In addition, the inducible expression of HBB in the serum and organs of individuals harboring clinically undetectable metastasis could be exploited for the early detection of minimal residual disease preceding relapse.

Y. Keisari is a consultant/advisory board member for SURI Technologies. No potential conflicts of interest were disclosed by the other authors.

Conception and design: S. Maman, O. Sagi-Assif, Wuyuan Lu

Development of methodology: S. Maman, O. Sagi-Assif

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Maman, W. Yuan, T. Meshel, Weiyue Lu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Maman, O. Sagi-Assif, W. Yuan, R. Ginat, T. Meshel, Y. Keisari, W. Lu

Writing, review, and/or revision of the manuscript: S. Maman, O. Sagi-Assif, T. Meshel, Y. Keisari, Weiyue Lu, Wuyuan Lu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Maman, W. Yuan, I. Zubrilov

Study supervision: O. Sagi-Assif, W. Lu, Wuyuan Lu

The authors thank Dr. Mickey Harlev and Dr. Maya Levin Arama (Animal Care Facilities, Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel), Dr. Joe Bryant, and Dr. Eugene Ateh (Animal Core Facility, Institute of Human Virology, University of Maryland School of Medicine, Baltimore, MD) for the help with animal experiments. We also thank the W.M. Keck Biomedical Mass Spectrometry Laboratory at the University of Virginia Biomedical Research Facility for the MS and MS/MS analyses.

This work was supported by the NIH grant AI087423 (W. Lu), by the National Basic Research Program of China (973 Program) 2013CB932500 (W-Y. Lu), by the German Research Foundation (Deutche Forschungsgemeinschaft DFG) grant BA4027/6-1 (I.P. Witz), by the James & Rita Leibman Endowment Fundfor Cancer Research (I.P. Witz), by the Fred August and Adele Wolpers Charitable Fund (I.P. Witz), and by the Sara and Natan Blutinger Foundation (I.P. Witz).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.