Pyruvate Kinase M2 in Blood Circulation Facilitates Tumor Growth by Promoting Angiogenesis*

Background: It is known that PKM2 is present in cancer patient blood. It is not known if PKM2 functions in cancer progression. Results: PKM2 in blood facilitates tumor growth by promoting tumor angiogenesis via increasing angiogenic endothelial cell migration and ECM attachment. Conclusion: Extracellular PKM2 promotes tumor angiogenesis. Significance: We reveal a novel mechanism of cancer angiogenesis that could potentially be a new target for anti-angiogenesis. It is long known that pyruvate kinase isoform M2 (PKM2) is released into the circulation of cancer patients. The PKM2 levels in patients have been suggested as a diagnostic marker for many types of cancers. However, it is not known how PKM2 is released in the blood, and whether the circulating PKM2 has any physiological function(s) in tumor progression. In this report, we demonstrate that PKM2 in the blood facilitates tumor growth by promoting tumor angiogenesis. Our experiments show that PKM2 promotes tumor angiogenesis by increasing endothelial cell proliferation, migration, and cell-ECM adhesion. Only the dimeric PKM2 possess the activity in promoting tumor angiogenesis, which is consistent with the observations that PKM2 in circulation of cancer patients is a dimer form.

An important molecular signature of tumor development is that a shift in expression of isoenzymes of pyruvate kinases occurs in tumors of almost all types. The tissue-specific isoform (L, R, or M1) disappears in many tissue types. In replacement, PKM2 2 is expressed in cancer cells (1)(2)(3). In tumor cells, growth stimulations convert the pyruvate kinase active tetramer PKM2 to a pyruvate kinase inactive dimer form (4). It is believed that during tumor development the demands for biosyntheses, especially syntheses of nucleotides and amino acids, are high. One main source of the carbon frame and reducing power of NADPH for the biosynthesis comes from metabolites of glycolysis. Thus, the inactive dimeric PKM2 actually provides a metabolic advantage to supply precursors for biosynthesis. Activation of pyruvate kinase activity is actually unfavorable for tumor growth (5,6). Christofk et al. (7) demonstrate that PKM2 is important for cancer cell metabolism and tumor growth. Interestingly, a number of recent studies show that PKM2 is functionally involved in multiple cellular processes in different subcellular locations, including metabolism control, transcription regulation, and chromatin package (8 -12).
High serum levels of PKM2 have long been observed in cancer patients of many types, including gastrointestinal cancer, pancreatic cancer, renal cell carcinoma, lung cancer, and ovarian cancer (13)(14)(15)(16). Studies show that there is a strong correlation between the serum levels of PKM2 and tumor progression. Thus, it is proposed that serum levels of PKM2 can be used as an important molecular marker for cancer diagnosis/prognosis. PKM2 is a glycolytic enzyme. The forms and the mechanism of its release into the circulation of cancer patients are not known. It is also not known whether the circulating PKM2 has any physiological function in tumor progression. In the present study we provide evidence showing that the circulating PKM2 facilitates tumor growth by promoting angiogenesis. PKM2 promotes tumor angiogenesis by increasing endothelial cell proliferation, migration, and cell-ECM adhesion.

Reagents, Cell Lines, Antibodies, and Protein Expression/
Purifications-Antibodies against ␤-actin, mouse CD31, and Ki-67 were purchased from Cell Signaling, Santa Cruz, and Abcam, respectively. The antibody against PKM2 was raised using recombinant PKM2 expressed/purified from Escherichia coli as an antigen. IgGs were purified from the rabbit anti-serum over a protein G column. Cell lines SW620 and PC-3 were purchased from ATCC, and HUVECs were purchased from Invitrogen. The cells were cultured by following the vendor's instructions. The cDNAs that encode human PKM2 and PKM1 were purchased from Addgene. The cDNAs were subcloned into bacterial expression vector pET-32a. The recombinant proteins were purified from bacterial lysates by a two-column procedure.
Mice Xenografts and Treatments-All animal experiments were carried out in accordance with the guidelines of Institutional Animal Care and Use Committee of Georgia State University. Nude mice (athymic nude, 5-6 weeks of age) were subcutaneously injected with 5 ϫ 10 6 of SW620 or PC-3 cells. Tumor formation and volumes were assessed every 2 days. Tumor volumes were measured by two perpendicular diameters of the tumors with the formula 4/3 ϫ (width/2) 2 ϫ (length/2). The tumor-bearing mice were subjected to the intraperitoneal injections of appropriate agents once every other day for 8 days. The treatments started 5 days post tumor inoculations. The tumors were collected and weighed at the end of the experiments. Tissue sections were prepared from harvested tumors and stained using commercially available antibodies against Ki-67 or mouse CD31. Statistical analyses were done in comparison to the control group with Student's t test.
Boyden Chamber and Cell Proliferation Assays-QCM TM 24-Well Fluorimetric Cell Migration Assay kit was used to measure the migration of different cells. The cells were first treated under the different conditions (indicated in figure legends) in regular cell culture plates. The treated cells were resuspended into optimum medium (without serum) and seeded into the inner chamber of the migration assay kit. The culture medium with 10% FBS was added to the outer chambers. After overnight incubation, medium in the inner chamber was removed, and the cells attached to the outer bottom side were detached using the cell detachment buffer (included in the kit). The detached cells were then lysed using the cell lysis buffer (included in the kit). The amounts of the migrated cells were determined by measuring the fluorescence using ex ϭ 485 nm and em ϭ 535 nm.
For analyses of cell proliferation, a cell proliferation ELISA kit that measures BrdU incorporation was used. Briefly, cells were incubated for appropriate time in the presence of 10 M BrdU under different conditions (indicated in figures). The cells were fixed after incubation and washed 3 times. The fixed cells were detected by an anti-BrdU-POD antibody and a secondary antibody. The nuclei incorporations of BrdU were measured by chemiluminescence emission (Victor 3TM, PerkinElmer Life Sciences). Cell proliferation was also measured by cell number counting. Cells were incubated for the appropriate time under FIGURE 1. Antibody against PKM2 and recombinant rPKM2 affect tumor growth. A, upper, the levels of PKM2 in the serum of mouse blood collected from tumor-bearing nude mouse (SW620) and non-tumor nude mouse (Con) were examined by an immunoblot of PKM2 (IB: PKM2). Coomassie Blue staining (CBS: M-Albumin) of serum albumin was a loading control. Bottom, the PKM2 levels in serum of nude mice with and without SW620 tumor implantation and with and without intraperitoneal administration of rPKM2 (48 h after injection) were determined by sandwich ELISA using an in-house-developed rabbit monoclonal antibody against PKM2 and a commercially available goat polyclonal antibody against PKM2. The PKM2 concentration was presented as the range of high and low levels (ng/ml) in serum samples from five mice per group tested. B, the levels of PKM2 in the cell extracts or medium (CM, 10 -20-fold concentration) of SW620 or PC-3 cells (indicated) were analyzed by an immunoblot of PKM2 (IB: PKM2). The PKM2 level in culture medium without cell culturing (Con) is a negative control. The immunoblots of ␤-actin (IB: Actin) in the culture medium and cell lysate are controls indicating that no ␤-actin presents in cell culture medium. C-E, growth of SW620 tumor under the treatment of purified IgGs of IgGPK or pre-immune serum (IgGCon) was monitored by growth curve by measuring tumor volumes every 2 days (C), end point pictures (D), and end point weights of the harvested tumors with 8 days treatment (treatment started 5 days post tumor inoculation) (E). F-H, growth of SW620 tumor under treatment of different recombinant proteins (indicated) was monitored by a growth curve measuring tumor volumes every 2 days (F) and end point pictures (G) and end point weights (H) of the harvested tumors with 8 days of treatment (treatment started 5 days post tumor inoculation). The p values are presented by * (p Յ 0.05) or NS (statistically insignificant; p Ͼ 0.05) and were calculated using unpaired two-tailed Student's t test. The error bars in C and F are S.D. from the measurements of five mice. appropriate conditions. Cell numbers were counted before and after the indicated time of culture by five independent cell counting.
Endothelial Tube Formation and Cell Attachment Assays-Endothelial tube formations were carried out with the endothelial tube kit. Briefly, HUVECs were seeded in culture plates coated with Matrigel. After 30-min incubations, agents, e.g. FBS, proteins, or cancer cell culture medium, were added to the HUVECs. The cells were further cultured for an additional 16 h. The formed endothelial tubes were analyzed under light microscope. For tube formations with supplement of SW620 culture medium, no FBS was added to the HUVEC cell culture.
For cell attachment, cells were cultured overnight under standard conditions. The next day different cells (with appropriate cell numbers) were transferred to new plates with wells that were coated with different proteins (indicated in the figures) and with fresh medium containing appropriate agents (indicated in the figures). The cells were further cultured for 2 h and washed gently. The attached cells were either directly counted or lysed. The cell lysates were then measured to determine the amounts of attached cells.
Size-exclusion Chromatography-Size exclusion chromatography was performed with a Superdex 200 10/300GL column. The samples of mouse serum (2-8 mg/ml total protein), rPKM2 (ϳ15 M), and rPKM1 (ϳ15 M) were prepared in Tris-HCl buffer with and without FBP. 100 l of the sample was loaded into the column and eluted with elution buffer (50 mM phosphate, 0.15 M NaCl, pH 7.2). The fraction of 300 l was collected, and 20 l of each fraction was analyzed by immunoblot. The elution profiles were compared with that of a size exclusion chromatography calibration kits (GE Healthcare) under identical conditions. The elution profile was plotted against log M r according to vendor's instructions.
Pyruvate Kinase Activity-Pyruvate kinase activity was analyzed by following the experimental procedure similar to that was described by Christofk et al. (17).

PKM2 in Blood Circulation
Facilitates Tumor Growth-We sought to investigate the physiological function(s) of circulating PKM2 in tumor progression. We first tested whether PKM2 was also released to the blood circulation of a xenograft model of human colon cancer cell SW620. After 4 weeks of tumor growth in nude mice, blood samples from the mice with or without tumor implantation were collected. PKM2 levels in the blood samples were analyzed by immunoblot and ELISA of the serum. It was evident that the PKM2 levels in blood of the SW620 tumor mice were very high. As a control, negligible PKM2 levels were detected in the blood of mice without tumor inoculation (Fig. 1A). We also examined the PKM2 levels in the cell culture medium of SW620 cells. Consistent with these data, we observed high levels of PKM2 in the medium (Fig. 1B). We questioned what would be the effects if the PKM2 in the mouse blood circulation was neutralized using an antibody against PKM2. We used an in-house developed rabbit polyclonal antibody raised against full-length recombinant PKM2 (referred to as PabPKM2). Antibody screening indicated specific recognition of PKM2 in the cell extracts ( Fig. 2A). The recognition of cellular PKM2 was completely abolished by the bacterially expressed PKM2 (Fig. 2B). The antibody did not recognize any protein in serum of nude mouse (Fig. 2C). IgGs were purified from the FIGURE 2. Antibody against PKM2 specifically recognizes PKM2. A, immunoblot of PKM2 in the whole cell lysate (WCL) of H146 and H460 cells using the anti-serum of the antibody PabPKM2. M, molecular weight markers. BSA was used as a control. B, immunoblot of PKM2 in the whole cell lysate of SW620 cells using the anti-serum of the antibody PabPKM2 (IB: PKM2). The anti-serum was preincubated with 5 M rPKM2 (ϩrPKM2) or BSA (Non) before adding the extracts. The immunoblot of ␤-actin (IB: Actin) is the loading control. C, immunoblot analyses of PKM2 in the serum collected from nude mice carrying the SW620 tumor (SW620) or no tumor (Con) using the antibody IgGPK (IB: IgGPK) or pre-immune serum (IB: IgGCon). D, Coomassie Blue staining (CBS) of the SDS-PAGE of purified IgGs from the anti-serum of the antibody PabPKM2 (IgGPK) and the preimmune serum of the same rabbit (IgGCon). Numbers on the left are the molecular weight markers. E, the intraperitoneally injected (48 h post injection) antibody IgGPK was analyzed by immuno blot of the serum sample collected from the injected mice using antibody against rabbit IgG (IB: Rabbit IgG). The control is the blot of serum from mouse without IgGPK injection.
antiserum of the PabPKM2 (referred to as IgGPK) by an affinity column of rPKM2 (Fig. 2D). The purified IgGPK or the IgGs purified from preimmune serum by protein A/G beads (referred to as IgGCon) were intraperitoneally injected (3 mg/kg in 100 l) into nude mice that carried the SW620 xenograft tumor every 2 days for 8 days. The administered IgGPK was detectable in mouse serum (Fig. 2E). It was clear that the IgGPK inhibited the tumor growth, whereas administration of the IgGCon did not exhibit any significant effect on the growth of the same tumor ( Fig. 1,  C-E). The results suggest that PKM2 in the blood circulation is critically important for the tumor growth.
We next tested whether the addition of PKM2 to blood circulation would promote tumor growth. We employed the bacterially expressed recombinant PKM2 (referred to as rPKM2) and its isoenzyme PKM1 (referred to as rPKM1) as a control. Because PKM2 is secreted from cancer cells, presumably the protein should be present in the extracellular space of tumors. Thus, the purified rPKM2 and rPKM1 were premixed with cancer cells at concentrations of 2 M. The mixtures were then subcutaneously implanted into nude mice. The purified recombinant proteins were also subsequently intraperitoneally injected (5 mg/kg in 100 l) in the tumor-bearing nude mice every other day for 8 days. The first injection started 5 days post tumor inoculation. Clearly, the SW620 tumors that were treated with the rPKM2 experienced over 2-fold higher growth rates compared with the tumors that were treated with the rPKM1 and buffer. The tumors treated with the rPKM1 and buffer had almost similar growth rates (Fig. 1, F-H). Fructose 1,6-bisphosphate (FBP) is an allosteric activator of PKM2 (18). We, therefore, examined whether the addition of FBP (3 mM) would affect the activity of the rPKM2 in promoting tumor growth. There was an apparent decrease in the activity of the rPKM2 in promoting tumor growth with the addition of FBP. However, the results were statistically insignificant ( Fig. 1, F-H). To test whether the observed effects of rPKM2 is specific to the SW620 tumor, we employed another xenograft model, human prostate cancer PC-3 cells, by the same treatment schedule. PKM2 was detected in the cell culture medium of PC-3 cells. Evidently, administration of the rPKM2 facilitated PC-3 tumor growth (Fig. 3, A-C).
PKM2 Promotes Angiogenesis-It is intriguing that cancer cells release PKM2 to the blood circulation and the circulating PKM2 promotes cancer growth. We questioned what the functional role of the circulating PKM2 is in promoting tumor growth. One possibility is that extracellular PKM2 promotes cancer cell proliferation. Tissue section stained using an antibody against Ki-67 indeed indicated that the tumors treated with the IgGPK had reduced proliferation rates, whereas the tumors treated with rPKM2 had higher proliferation rates (Fig.  4, A-D). However, the addition of the IgGPK, IgGCon, rPKM2, rPKM1, and rPKM2 ϩ FBP into SW620 and PC-3 cell culture medium did not lead to any significant change in cell proliferation (Fig. 4, E and F). Thus, the tumor growth promotion by the rPKM2 and inhibition by the IgGPK were unlikely due to their actions on cancer cells. The other possibility is that the PKM2 in circulation promotes angiogenesis to facilitate tumor growth. To test this conjecture, we carried out histology analy-ses with the tumor tissue sections using antibody against mouse CD31, a marker for endothelial cells. It was clear that treatment of mice with the IgGPK dramatically reduced blood vessels in the SW620 tumor (Fig. 5, A and B). Reversely, treatment of mice with the rPKM2 led to substantial increases in blood vessels in both PC-3 and SW620 tumors. The rPKM2 ϩ FBP had reduced effects, whereas the rPKM1 had no significant effects (Fig. 5, C-G). Our analyses strongly suggested that the PKM2 in blood circulation promotes angiogenesis.
To further verify the role of PKM2 in angiogenesis, we employed the in vitro tube formation assay using HUVECs. The rPKM2, rPKM2ϩFBP, rPKM1, and buffer alone were added to the culture medium of the cells. Formation of endothelial tubes was analyzed. The rPKM2 strongly promoted endothelial tube formation (Fig. 6, A and B). The time required for the formation of the tubes was substantially shortened, and the tubes were maintained for a much longer time (data not shown). The rPKM2 ϩ FBP had less effects compared with that of the rPKM2. The rPKM1 had only marginal effects, whereas saline had no effects (Fig. 6, A and B). We subsequently tested the effects of the PKM2 antibody on the tube formation by co-culturing the medium collected from SW620 cell cultures in which PKM2 was immunodepleted by the IgGPK with HUVECs. Immunoblots indicated that PKM2 in SW620 cell culture medium was completely removed by the IgGPK (Fig. 7A). Clearly, the antibody depletion greatly reduced the endothelial tube formation in the co-culture of SW620 medium with HUVECs (Fig. 7, B  and C). These in vitro tests supported our notion that PKM2 promotes angiogenesis. We further tested the effects of PKM2 on angiogenesis by Matrigel plug assay using Balb/c mouse. The presence of rPKM2 (3 M) in the Matrigel enhanced the vessels growth into the Matrigel compared with that in the presence of rPKM1 (3 M) (Fig. 7D).
PKM2 Facilitates Endothelial Cell Migration-How does circulating PKM2 promote tumor angiogenesis? To address this question, we analyzed cell proliferation, migration, and adhesion of HUVEC cells. The addition of rPKM2 to the cell culture medium led to a marginal increase in cell proliferation (Fig. 6C). It is unlikely this small effect would be the sole factor that confers the in vivo effects on tumor growth. On the other hand, Boyden chamber assays showed that the addition of the rPKM2 led to a strong increase in cell migration, and the effects were substantially reduced by the addition of FBP. The rPKM1 had almost no effects (Fig. 6D). The increases in cell migration were not observed with SW620 and PC-3 cells (Fig. 8A). We noted a strong attachment of HUVEC cells to the culture plates with rPKM2 coating (Fig. 6E). In addition, the attachment of HUVEC cells to the ECM-coated plate was substantially strengthened by the addition of SW620 cell culture medium, whereas the enhancement was abolished by the addition of the antibody PabPKM2 (Fig. 8B). Analyses of the actin filament by fluorescence staining suggested formation of strong stress microfilament fibers upon rPKM2 treatment (Fig. 8C). We suspected that the rPKM2 might affect endothelial cell adhesion to ECM. Indeed, the cell attachment assays showed that HUVEC adhesion to vitronectin and fibronectin, matrix molecules, was strongly enhanced upon the addition of the rPKM2 to the culture medium (Fig. 6F). The effects were not observed with rPKM1. In consistent with these data, we observed a strong effect on the HUVEC spreading by the addition of rPKM2 into the cell culture medium, whereas this effect was not observed with rPKM1 (Fig. 8D). The effects of PKM2 on cell adhesion to ECM and cell migration were not observed with epithelial cancer cells (Fig. 8, A, E, and F), indicating that the effects were endothelial cell-specific. Thus, we conclude that PKM2 promotes angiogenesis by facilitating endothelial cell migration and cell adhesion to ECM.
Dimer PKM2 Promotes Angiogenesis-It is believed that the PKM2 in the cancer patient blood circulation exists as a dimer (19,20), whereas the protein in cancer cells exists as a mixture of tetramer and dimer (18,21) (8). Thus, an interesting issue is whether the dimer and tetramer status of PKM2 have different effects in promoting tumor growth. We first examined the tetramer and dimer status of the administered rPKM2/rPKM1 in the mouse serum. Chromatography analyses followed by immunoblots indicated that the rPKM2 existed mostly as a dimer in the circulation, whereas the rPKM1 was mostly a tetramer (Fig. 9A). Using an ELISA analysis, we estimated that the concentration of the intraperitoneally injected rPKM2 and rPKM1 (at dose of 5 mg/kg) in the mouse blood was around 400 -800 nM 4 h after the administration. We also analyzed the tetramer and dimer status of the rPKM2/rPKM1 in vitro. Chromatography profiles indicated that the rPKM2 existed as a mixture of tetramer and dimer (with tetramer to dimer ratio at around 80% to 20%) at concentrations of 12 M, whereas the rPKM1 was almost completely tetramer at the same concentration (Fig. 9B). The purified proteins possessed pyruvate kinase activity (Fig. 9C). Interestingly, dilution of the rPKM2 led to conversion of tetramer to dimer with the rPKM2 almost completely dimer at around 1 M (Fig. 9, D and G). This concentration is very close to the concentration of the administered rPKM2 in mouse blood. Most of the rPKM1 still existed as tetramer at this concentration (Fig. 9, F and G). The addition of 3 mM FBP converted the rPKM2 to the tetramer even at concentrations as low as 1 M (Fig. 9E). Consistently, a large portion of rPKM2 was tetramer in mouse blood circulation when the protein was co-administered with 3 mM FBP (Fig. 9A). Clearly, the observed activity of FBP in facilitating the conversion of the rPKM2 from dimer to tetramer is consistent with the results that the addition of FBP to the rPKM2 reduced the effects of the protein on promoting tumor growth (see in Fig. 1,  F-H, and in Fig. 3, A-C). Both support a conclusion that introduction of the dimeric PKM2 to blood circulation promoted tumor angiogenesis. Can FBP also convert the circulating dimer PKM2 to a tetramer, therefore, inhibit tumor angiogenesis? To answer this question, we collected the serum samples from the mice that had been implanted tumors. FBP was added to the serum samples. The PKM2 in the serum was analyzed by the same chromatography procedure. Clearly, the addition of FBP did not convert the circulating PKM2 to tetramer (Fig. 10A). The observation is consistent with previous report that FBP could not convert the nuclear dimer PKM2 to a tetramer (8). . PKM2 affects endothelial tube formation, migration, and extracellular matrix attachment. Shown are representative light microscopic images of endothelial tubers formed by HUVECs in the presence of rPKM2, rPKM2 ϩ FBP, rPKM1, and buffer saline (A) or quantitative analyses of the branch points in the formed HUVEC tubes (B). The quantification was the mean of branch point counting of four randomly selected fields from each slide and with five repeating experiments (slides). C and D, cell proliferation (C) and migration (D) of HUVECs in the presence of rPKM2, rPKM2 ϩ FBP, rPKM1, and buffer saline were analyzed by a commercial BrdU proliferation kit and Boyden chamber assay, respectively. The cell proliferation and migration assays are presented as relative proliferation (C) or relative migration (D) by defining the proliferation of buffer saline-treated cells as 100% and cell migration of rPKM2-treated cells as 100%. E and F, cell attachment of HUVECs to cell culture plates on which rPKM2, rPKM1, or BSA was coated (E) or fibronectin or vitronectin was coated, and the indicated proteins were added to the culture medium (F). The cell attachments are presented as relative attachment by defining the cell attachment in the plate on which rPKM2 was coated as 100% in E or buffer saline was added to the culture medium as 100% in F. In A-D, buffer saline is a control. The p values are presented by * (p Յ 0.05), *** (p Յ 0.005), or NS (statistically insignificant; p Ͼ0.05) and were calculated using unpaired two-tailed Student's t test. The error bars in B-F represent the S.D. from five repeating experiments.

DISCUSSION
A very high percentage of cancer patients of different cancer types have elevated levels of PKM2 in their blood circulation (22). We show that the circulative PKM2 plays a critical role in facilitating tumor growth by promoting tumor angiogenesis. An obvious question is why tumors release this glycolysis enzyme to the circulation to promote angiogenesis. One plausible explanation is that dimeric PKM2 is a sensor for glycolysis status and glucose demands in cancer cells for high proliferation. It is demonstrated by many laboratories that PKM2 is converted to dimer in high proliferation cells with low glycolysis activity (17,21). Thus, it is possible that dimeric PKM2 is a signal reflecting a nutrition requirement and therefore a signal for angiogenesis. Most recently, Luo et al. (10) reported that PKM2 co-activates the transcription factor HIF1␣. It is conceivable that PKM2 may also be an important molecular factor in response to hypoxia conditions for angiogenesis stimulation.
Interestingly, PKM2 is the main pyruvate kinase isoform during fetus development. It is well known that angiogenesis is active during organ development and maturation. Is PKM2 also secreted into blood circulation of fetus for promoting angiogenesis?
Another open question is why PKM2 but not PKM1 promotes tumor angiogenesis. PKM2 and PKM1 differ by only 23 amino acids resulting from alternative pre-mRNA splicing. These two isoforms share an almost identical structure. However, structural analyses revealed that the segment with a different amino acid sequence is mostly localized at the dimerdimer interface (4). Thus, it is conceivable that the dimer interface is engaged in interaction with the target molecule on the endothelial cell surface and consequently promotes angiogenesis. This notion is consistent with the results that the dimer PKM2 is effective in promoting angiogenesis, whereas tetramer is inactive. We previously demonstrated a PKM2 mutant R399E B and E, cell attachment of HUVECs (B) or SW620 cells (E) to the cell culture plate on which ECM was coated. B, SW620 cell culture medium (620CM) or the medium without SW620 cell culturing (ConCM) with the addition of IgGPK (IgGPK) or the IgGCon (IgGCon) added to the HUVEC culture before the attachment analyses. E, rPKM2, rPKM2 ϩ FBP, rPKM1, or buffer saline was added to the SW620 culture. The cell attachments are presented as relative attachment by defining the cell attachment in the plate in which ConCM was added to HUVEC culture as 100% (B), or buffer saline was added to the SW620 culture medium as 100% (E). C, representative microscopic images of F-actin staining of HUVECs (red). The cells were seeded on microscopic chamber slides with ECM coating. The indicated proteins (at 5 M) were added to the slides. The cells were fixed and stained using Rhodamine phalloidin. D, representative phase contrast images of cell spreading of HUVECs on microscopic chamber slides coated with ECM and in the presence of rPKM1, rPKM2, and BSA (indicated). Quantification of the spreading cells (defining by obvious membrane protrusions and visible lamellipodia) by manually counting the cells per field is shown. Results are represented as the percentage of spreading cells counted in five fields. F, HUVEC migration in the presence of IgGPK, IgGCon, and buffer saline were analyzed by Boyden chamber assay. The cell migrations are presented as relative migration by defining the cell migration of buffer saline-treated cell as 100%. The p values are presented by * (p Յ 0.05) or NS (statistically insignificant; p Ͼ 0.05) and were calculated using unpaired two-tailed Student's t test. The error bars in A, B, D, E, and F are S.D. from five repeating experiments.
that is a dimer and less active in pyruvate kinase activity (8). The R399E mutant is fully functional (slightly more active) in promoting HUVEC attachment to ECM (Fig. 10B), suggesting that pyruvate kinase activity is not required and the exposed dimerdimer interface is required, which also supports the notion that the dimer interface may be engaged in interaction with the target(s) and consequently promotes angiogenesis. Curiously, the effects of FBP on PKM2 for tumor growth in vivo were statistically insignificant, whereas the effects of FBP on PKM2 activities in vitro were statistically significant. The discrepancy may be due to differences of effective FBP concentrations in both in vivo and in vitro conditions. Likely, the input FBP (premixed with rPKM2) was largely diluted once administered into mouse blood circulation. The question of the molecular target(s) of PKM2 in mediating the endothelial cell migration and ECM attachment is intriguing. It was suggested that PKM2 might interact with TEM8, tumor endothelial marker 8 (23). To probe whether TEM8 is a target of PKM2, we probed the potential interaction between TEM8 and PKM2 by coimmunoprecipitation. We did not observe a strong interaction between TEM8 and PKM2 in our experiments (data not shown). Thus, it is unlikely that PKM2 targets TEM8. Clearly, to fully understand the molecular mechanism by which extracellular PKM2 facilitates endothelial cell migration and attachment to ECM, it will FIGURE 9. Tetramer or dimer of PKM2 and PKM1. A, upper, immunoblot analyses of the chromatography fractions of the mouse serum collected from mice that were administered rPKM2, rPKM2 ϩ FBP, and rPKM1 using the antibody PabPKM2. Bottom, the chromatography profiles of the standard molecular weight calibration kit. mAU, milliabsorbance units. B, chromatography profiles of rPKM2 and rPKM1 at concentrations of 12 M. Elution volumes equivalent to tetramer and dimer are indicated by arrows. C, pyruvate kinase activity of the rPKM2, rPKM2 ϩ FBP, and rPKM1 (5 g/ml) was analyzed by the method described by Christofk et al. (17). The pyruvate kinase activity was expressed as relative pyruvate kinase activity by defining the activity in the rPKM1 as 100%. D-F and G, the chromatography profiles of different concentrations (indicated) of rPKM2 (D), rPKM2 ϩ FBP (E), and rPKM1 (F) as well as 1 M rPKM2/rPKM1 (G). The dimer and tetramer ratios (T/D Ratio) in B-F were calculated by the areas under the dimer and tetramer peaks. In A, B, D, E, F, and G the y axis is the absorbance of the elution at 280 nm. The numbers in the x axis are the elution volumes. The p values are presented by * (p Յ 0.05), ** (p Յ 0.01), or NS (statistically insignificant; p Ͼ 0.05) and were calculated using unpaired two-tailed Student's t test. The error bars in C are S.D. from five repeating experiments. be a necessary first step to find out the PKM2 interacting target on endothelial cells. Apparently, inhibition of tumor growth by the PKM2 antibody reveals a potential new anti-angiogenesis target for cancer patients with high circulative PKM2. One great advantage of targeting PKM2 is that the levels of circulative PKM2 would be a prognosis marker to predict the possible outcomes.