GTP cyclohydrolase (GCH1) is the key-enzyme to produce the essential enzyme cofactor, tetrahydrobiopterin. The byproduct, neopterin is increased in advanced human cancer and used as cancer-biomarker, suggesting that pathologically increased GCH1 activity may promote tumor growth. We found that inhibition or silencing of GCH1 reduced tumor cell proliferation and survival and the tube formation of human umbilical vein endothelial cells, which upon hypoxia increased GCH1 and endothelial NOS expression, the latter prevented by inhibition of GCH1. In nude mice xenografted with HT29-Luc colon cancer cells GCH1 inhibition reduced tumor growth and angiogenesis, determined by in vivo luciferase and near-infrared imaging of newly formed blood vessels. The treatment with the GCH1 inhibitor shifted the phenotype of tumor associated macrophages from the proangiogenic M2 towards M1, accompanied with a shift of plasma chemokine profiles towards tumor-attacking chemokines including CXCL10 and RANTES. GCH1 expression was increased in mouse AOM/DSS-induced colon tumors and in high grade human colon and skin cancer and oppositely, the growth of GCH1-deficient HT29-Luc tumor cells in mice was strongly reduced. The data suggest that GCH1 inhibition reduces tumor growth by (i) direct killing of tumor cells, (ii) by inhibiting angiogenesis, and (iii) by enhancing the antitumoral immune response.
Neopterin is a biomarker for viral infection and cancer and high plasma concentrations have been associated with a poor prognosis.1, 2 It is produced as a by-product in tetrahydrobiopterin (BH4) de novo synthesis and mirrors the activity of the rate limiting enzyme in the BH4 synthesis cascade, GTP cyclohydrolase 1 (GCH1).3 GCH1 produces dihydroneopterin phosphate, which is further processed to 6-pyruvoyl-tetrahydropterin by the downstream enzyme PTP-synthase and finally via sepiapterin reductase to BH4. In the case of high GCH1 activity, and hence substrate overload for the PTP-synthase, the intermediate dihydroneopterin phosphate is metabolized to neopterin, which is a stable “waste” product excreted in urine.3 Neopterin concentrations in plasma are indicative of GCH1 activity in immune cells,4 platelets,5 and endothelial cells,6 that upon proinflammatory stimuli may considerably raise GCH1 expression and activity.7
However, it is not known whether GCH1 activity and BH4 production are just associated with cancer or whether they play a role in the pathophysiology of tumor growth, angiogenesis and tumor-evoked immune responses.8 We have previously reported that specific GCH1 polymorphisms in humans prevent the upregulation of GCH1 upon stimulation9–11 and are associated with slower progression of cancer pain and longer survival times.12 This suggested that GCH1 activity may play a causative role in cancer progression. The biochemical link might be BH4, which is essential as an electron acceptor for the coupling of oxidation and reduction at the catalytic site of nitric oxide synthases (NOS).13–15 The stoichiometric balance between NOS and BH4 is tightly controlled at various levels and both relative deficiency and overload of BH4 are associated with pathological conditions such as arteriosclerosis and neuropathic pain, respectively.9, 15, 16 An imbalance increases the formation of reactive oxygen and nitrogen species.14, 17 Within the tumor environment BH4 and NO are likely to regulate the blood flow, the formation of new blood vessels18, 19 and functions of immune cells.20, 21
Considering the functional associations of GCH1 polymorphisms with neurological and cardiovascular diseases and with pain in humans10 and in light of the association of plasma neopterin concentrations with cancer1, 2 we pursued the hypothesis whether reduction of BH4 activity may suppress tumor progression. We show that GCH1 inhibition or downregulation reduce tumor growth and angiogenesis in vitro and in tumor bearing mice. The results were translated into humans addressing GCH1 upregulation in tumor tissue. The results suggest that inhibition of GCH1 activity might play a future role in cancer therapy.
Material and Methods
Mouse melanoma B16 and B16-F10-luc-G5 cells (ATCC and Bioware cell lines, respectively) originated from C57BL6 mice and human breast cancer cells (MCF-7, ATCC). They were grown in Dulbecco's modified Eagle medium (DMEM) with 10% supplemental fetal bovine serum and 1% penicillin/streptomycin. HT29 and HT29-luc-D6 human colon cancer cells (DSMZ, Heidelberg, Germany and Bioware cell lines, respectively) were cultured in McCoy's medium supplemented with 10% FCS and 1% Penicillin/Streptomycin. All cells were kept in an incubator at 37°C, 95% humidity and 5% CO2 atmosphere. For analysis of biopterin and neopterin 1 × 106 cells were seeded in 15 cm dishes, cultured for 3 days and subsequently exposed to 1% oxygen for 24 hr. For human umbilical vein cell cultures (HUVEC, PromoCell) we used CS-C medium (Sigma) with 10% iron supplemented bovine calf serum, L-glutamine, sodium bicarbonate, and HEPES. Tumor cells were harvested by trysinization, resuspended in the respective media without supplements and kept on ice for less than 30 min until injection in mice.
WST assays and colony forming
WST assays (water soluble tetrazolium) were used according to the manufacturer's instructions (Sigma, Steinheim, Germany) to assess the viability and metabolic activity. In viable cells the mitochondrial succinate tetrazolium dehydrogenase system catalyses the conversion of the red tetrazolium salt WST-1 (4-[3-(4-Iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]-1,3-benzole-disulfonate) to dark red formazan, which is analyzed photometrically.
To assess colony formation cells (500 cells/3 cm dish) were treated with DAHP or vehicle for 10–20 days depending on the cell line, were fixed with 100% methanol and stained with 0.5% crystal violet, 5% Giemsa solution. Single cell clones were counted using QuantityOne (Biorad, München, Germany). The relative survival rate was calculated in relation to vehicle treated control cells set at 100%. The IC50 was calculated using a standard sigmoidal Emax model.
FACS analysis of cell cycle and cell death
For cell cycle distribution analyses, cells (5 × 105) were seeded in 10 cm dishes. They were starved for 72 hr in medium containing 1% FCS to synchronize the cell cycles. Cells were then treated for 24 hr with DAHP or vehicle or were transduced with the respective shGCH1 or control lentivirus particles at 1 MOI (multiplicity of infection) in full medium for 48 or 72 hr. Cells were harvested by trypsinization and fixed with 80% ethanol. After fixation cells were washed two times with phosphate buffered saline (PBS), incubated for 5 min with 0.125% Triton X-100, washed again and stained with propidium iodide in PBS containing 0.2 mg/ml RNAse A. Stained cells were analyzed by flow cytometry (FACSCanto II, Becton Dickinson). Cells were counted until a total of 7,000 cells in a predefined G1-gate were recorded. The cell cycle distribution i.e., the percentage of cells in subG1, G0/G1, S, and G2/M phase, was assessed using VenturiOne software (Applied Cytometry).
For analysis of apoptosis, cells (1 × 106) were treated with DAHP or transduced with shGCH1 or control lentivirus particles at 1 MOI for 48–72 hr. Cells were harvested by trypsinization. After washing with PBS cell pellets were resuspended in binding buffer [10 mM HEPES, pH 7.4; 140 mM NaCl; 2.5 mM CaCl2; 0.1% BSA(bovine serum albumin)] and adjusted to a cell density of 1 × 106 cells/ml. Hundred microliters of the cell suspension was incubated on ice with 5 μl phycoerythrin (PE)-labeled Annexin V and 5 μl 7-amino-actinomycin D (7-AAD) (Annexin V-PE apoptosis kit I, Pharmingen, Hamburg, Germany) for 15 min at room temperature in the dark. Four-hundred microliters of binding buffer was added and fluorescence was measured by flow cytometry. The percentage of live and dead cells was analyzed with VenturiOne software (Applied Cytometry).
FACS analysis of tumor associated immune cells
To ensure normal in vivo immune functions, a syngeneic tumor model was used. B16 melanoma cells were injected in C57BL6 mice. Starting after tumor injection, mice were treated with 2 × 100 mg/kg/day DAHP or vehicle for 12 days. Tumors were then dissected, and single cell suspensions were prepared by treatment with tumor lysis buffer [DMEM/Accutase (PAA) 1:1, collagenase (3 mg/ml, Sigma), DNAse I (1 U/ml, Promega)] for 45 min at 37°C and subsequent mechanical disruption via forcing the tissue through a nylon mesh with 70 μm pore size (Cell Strainer, BD). Cells were then treated with erythrocyte lysis buffer for 4 min and CD16/32 blocking antibody (Fcγ RII/III receptor blocker, BD) for 15 min on ice. For staining of cell surface antigens cells were incubated for 20 min at 4°C in staining buffer with the respective antibodies and were then counted with a flow cytometer (FACS LSRFortessa™-BD). FACS scans were analyzed with VenturiOne software. The following antibodies were used: CD45-PerCP (BD), CD45-V450 (Miltenyi Biotec), CD11b-APC or CD11b eFluor (eBioscience), CD206-FITC (Biozol) or CD206-PE (Serotec), CD11c-Alexa Fluor 700 (BD), HLA-DR (MHC II)-PE (Miltenyi Biotec). The controls were FITC, PE, or APC-conjugated rat IgG1 or IgG2a.
Tube forming assay for angiogenesis
In vitro angiogenesis was analyzed with an in vitro tube forming assay in human umbilical vein endothelial cells (HUVECs) according to the manufacturer's instructions. Briefly, cells were seeded onto an ECM (extracellular matrix) coated 96-well microplate (Cell Biolabs) for 12 hr and were then treated with DAHP at various concentrations (0.5–5 mM in triplicate) for 72 hr. Images were captured every 24 hr and analyzed with the Axiovision AutMess modul (Zeiss) to assess tube formation, cell viability, and proliferation. We used “area covered by tubular structures/total area analyzed” as the major outcome parameter. Drug effects were assessed by comparison of EC50 levels obtained by curve fitting with a standard sigmoidal Emax model.
Generation of GTP cyclohydrolase shRNA expressing lentiviral constructs
The siRNA (small interfering RNA) sequence was designed as described previously22 employing an mRNA sequence of 19–23 nucleotides complimentary to the target GCH1 cDNA (sequence AAG(N18-22)TT). To clone the lentiviral construct with shRNAs specific for the human GCH1, two complimentary oligonucleotides were designed consisting of the sense strand of GCH1-siRNA of 19 nucleotides followed by a short spacer (TTCAAGAGA), the reverse strand and five T tails as an RNA Polymerase III terminator. The oligomers were annealed and inserted into the HpaI and XhoI sites of lentiviral vector pLentiLox-3.7.22 The oligomer annealing reaction was performed at an established thermal cycle protocol consisting of 95°C for 2 min, 65°C for 10 min, 37°C for 10 min, 20°C for 20 min, and final incubation at 4°C for 10 min. Each step was run once.
Virus amplification and production of high-titer lentivirus stocks
Lentiviral particles were produced by transient cotransfection of HEK293T cells seeded in 6 cm dishes with 10 μg of vector DNA (pLL-shGCH1), together with three helper plasmids (5 μg of pMDL/RRE, 3 μg of RSV-Rev, and pMD2G-VSVG). The calcium phosphate precipitation method was used for transfection according to standard techniques. Cell supernatant containing lentivirus particles was harvested at 48 hr after transfection, and passed through a 0.45 μm filter, and concentrated by centrifugation for 90 min at 25,000 rpm at 4°C. The virus pellet was resuspended in 25 μl cold 0.1M PBS and the titer was assessed by GFP expression analysis in transduced tumor cells. Viral titers were 2–3 × 107 transduction units/ml.
For transduction 5 × 105 HT29-Luc cells were seeded in 10 cm dishes and transduced with the respective lentivirus particles at 10 MOI in complete medium. After 3 days the efficiency of virus infection was determined by flow cytometry on the basis of eGFP expression. To exclude the noninfected cells, GFP-positive cells were isolated by the FACSAria Cell sorter (Becton Dickinson, Germany).
Total RNA was extracted from homogenized tissue or cells according to the protocol provided in the RNAeasy tissue Mini Kit (Qiagen, Hilden, Germany), reverse transcribed using poly-dT as a primer to obtain cDNA fragments. Quantitative RT-PCR was performed using an ABI prism 7700 TaqMan thermal cycler (Applied Biosystems, Germany) using the Sybrgreen detection system with primer sets designed on primer express. Specific PCR product amplification was confirmed with gel electrophoresis. Transcript regulation was determined using the relative standard curve method according to the manufacturer's instructions (Applied Biosystems). Amplification was achieved at 50°C for 40 cycles using a specific primer pair for human GCH1: forward: 5′-CCAGGTGCAGCAATGGGTTC–3′; reverse: 5′-TTCAACCACTACCCCGACTC-3′.
Western blot analysis
Whole cell protein extracts were prepared in RIPA lysis buffer containing a protease inhibitor cocktail and PMSF 10 μg/ml, separated on a 12% SDS–PAGE gel, transferred to nitrocellulose membranes by wet-blotting. Tumor tissue samples were homogenized in PhosphoSafe Buffer (Sigma), and protease inhibitor mixture (Complete™, Roche Diagnostics), proteins separated by SDS–PAGE (30 μg/lane) and transferred to nitrocellulose membranes (Amersham Pharmacia) by wet-blotting. Proteins were detected using the following antibodies: rabbit antiserum against full length rat GCH1, antihuman GCH1 (Abnova), eNOS (Chemicon), iNOS (Chemicon), VEGF (Sigma), CD31/PECAM (BD), MPO (Abcam), F4/80 (MD Bioscience), PARP (Cell Signaling), and Phospho-Histone H3 (Ser28) (cell signaling) and secondary antibodies conjugated with IRDye 680 or 800 (1:10,000; LI-COR Biosciences, Bad Homburg, Germany). Beta-actin and Hsp90 were used as loading controls. Blots were visualized and analyzed on the Odyssey Infrared Imaging System (LI-COR Biosciences). The ratio of the respective protein band to the control band was used for semiquantitative analysis.
The experiments were done in male, 8- to 12-week-old, C57BL6 mice for B16 mouse melanomas and in NMRI-nu/nu mice for HT29-Luc human colon cancer xenografts (Harlan Winkelmann GmbH). Mice had free access to food and water, were maintained in climate controlled rooms at a 12 hr light–dark cycle. The experiments were approved by the local Ethics Committee for animal research (Darmstadt, Germany) and were in line with the European and German regulations for animal research. The measurements and imaging analyses were performed by an observer unaware of the treatments.
Mouse tumor models and treatments
Tumor cells were injected under 1.5–2% isoflurane anesthesia. For human colon cancer xenografts we injected HT29 and HT29-Luc-D6 into the subcutaneous tissue in 6- to 8-week-old NMRI-nu/nu male mice. For experiments involving shGCH1 we injected 2.5 × 106 HT29-Luc-D6 transduced (Lv-shGCH1 or control lentivirus, n = 8 per group). To achieve syngeneic tumor growth we used B16 and B16-F10-Luc-G5 mouse melanoma cells for C57BL6 mice. Cells were suspended in cell culture medium and injected subcutaneously at the right dorsal flank (5 × 105 cells in 100 μl culture medium) or into the plantar mouse pad of the left hindpaw (5 × 104 cells in 50 μl culture medium). Mice were treated with 100 mg/kg diamino hydroxypyrimidine (DAHP, Sigma/Aldrich) twice daily or vehicle. DAHP was dissolved in DMSO/H2O 1:1 and administered orally. The first dose was administered after tumor cell injection. Eight to fifteen mice were treated in each group. The tumor dimensions were assessed with a calliper ruler and the volume estimated [long diameter × short diameter2)/2] and by in vivo luciferase analysis. Two or three weeks after tumor cell implantation mice were killed and blood was sampled for the determination of plasma cytokine and biopterin concentrations. Tumor samples were excised and stored at −80°C until further analysis.
Colon cancer was induced in C57BL6 mice via a single injection of azoxymethane (AOM, 7.4 mg/kg, Sigma Aldrich) followed by three cycles of dextran sodium sulfate (3% DSS) in the drinking water. The first cycle started one day after AOM and each cycle consisted of 7 days DSS and 14 days regular water. Mice were killed 10 weeks after AOM injection, the colon flushed with PBS (0.1M), excised, cut open longitudinally along the main axis, washed with PBS, and the number of tumors was evaluated macroscopically. Tumor and healthy surrounding colon tissue samples were dissected and kept at −80°C until analysis.
In vivo luciferase and fluorescence analysis of tumor growth and angiogenesis
In vivo luciferase imaging was done with an IVIS Lumina II imaging system that employs XENOGEN technology (Caliper LifeSciences). Ten minutes before imaging 100 μl D-luciferin (150 mg/ml) was injected intraperitoneally. In vivo near-infrared (NIR) fluorescence was analyzed on a Maestro imaging system that employs CRi's in vivo spectral imaging technology. For the in vivo analysis of tumor angiogenesis we injected 100 μl IntegriSense 680 (2 nmol dissolved in 100 μl 0.1M PBS; VisEN) intravenously via the tail vein 24 hr before imaging. InteriSense comprises a nonpeptide small molecule integrin αVβ3 antagonist and a NI) fluorochrome. It binds to neovasculature and enables monitoring of tumor growth and tumor angiogenesis. During image capturing mice were kept under 1–1.5% isoflurane anesthesia.
Cytokine/chemokine protein array
A mouse cytokine array from R&D Systems was used according to the manufacturer's instructions. Briefly, the assay employs capture antibodies spotted in duplicate on nitrocellulose membranes. Mouse plasma samples were mixed with biotinylated detection antibodies and then incubated with the array membrane. Complexes of cytokines or chemokines with the respective detection antibodies were bound by the cognate immobilized capture antibodies and detected with streptavidin-horseradish peroxidase (HRP) and chemiluminescent detection reagents. The analysis of the spot pixel density was done with ImageJ software.
Concentrations of biopterin and neopterin
After acidic oxidation of homogenized tissue with iodine pteridines were obtained by solid phase extraction using Oasis MCX extraction cartridges (Waters GmbH, Eschborn, Germany). Concentrations of total biopterin, neopterin, and the internal standard rhamnopterin were determined by liquid chromatography coupled to tandem mass spectrometry. HPLC analysis was done under gradient conditions using a Gemini C18 5 μm, 150 × 2 mm column (Phenomenex, Aschaffenburg, Germany). MS/MS analyses were performed on an API 4000 Q TRAP, a hybrid triple quadrupole/linear ion trap mass spectrometer with a turbo ion spray source. Precursor-to-product ion transitions of m/z 236→192 for biopterin, m/z 252→192 for neopterin and m/z 266→192 for rhamnopterin were used for the MRM. Quantification was done with Analyst software 1.5 (AB Sciex). Linearity of the calibration curve was proven from 0.1 to 100 ng/ml. The coefficient of correlation for all measured sequences was at least 0.99. The intra-day and inter-day variability was <10%.
Terminally anesthetized mice were perfused transcardially with 0.9% saline followed by 4% paraformaldehyde (PFA) in 0.1M PBS (pH 7.4). Paws or subcutaneous tumors were dissected and postfixed for 2 hr and then transferred into 20% sucrose in PBS for overnight cryoprotection at 4°C. The tissue was embedded in Tissue-Tek® O.C.T. Compound (Science Services, Munich, Germany) and cut in transverse sections on a cryotome. Mouse paws were dimineralized for 3–4 weeks in 10% EDTA before sectioning. Tumor cryosections were sequentially incubated with methanol, avidin/biotin (Vector Laboratories), and 1% blocking reagent (Dako). Primary antibodies specific for F4/80 (1/500; BD), myeloperoxidase (1/20; Abcam), PECAM1 (1/1000; BD), and IL-6R (1/1000; Santa Cruz) were dissolved in TBST (1% Triton X-100 in 0.1M Tris buffered saline)/0.5% BSA and incubated overnight at 4°C. Subsequently, slides were incubated with biotinylated secondary antibodies and streptavidin-HRP and detected with a Tyramide-Amplification kit (Cy3/FITC) according to the manufactures's instructions (PerkinElmer). Nuclei were counterstained with DAPI and slides were imbedded with Vectashield mounting medium (Vector Laboratories).
Immunostainings of human paraffin embedded tissue arrays (AccuMax tissue array colon cancer 45 cases (A203(VII), Petagen; and tissue array human skin cancer 12 cases, US Biomax T212) followed the manufacturer's instructions. Briefly, section were deparaffinized in graded ethanol, endogenous peroxidase was blocked with 3% H2O2 and antigen retrieval achieved with 0.01% citrate buffer pH 6.0. Incubation with the primary antihuman GCH1 antibody (Abnova) was done in PBST with 1% blocking reagent overnight at 4°C followed with AlexaFluor-594 (Invitrogen). Slides were rinsed in PBS and cover slipped in antifade medium (fluorescent staining). Sections were analyzed on a fluorescent microscope (AxioImager, Zeiss, Germany).
We used SPSS 19.0 for statistical evaluation. Data are presented as means ± SEM or SD as indicated. In the case of time-courses, data was analyzed using analysis of variance (ANOVA) for repeated measurements. In addition, areas under the observations-versus-time curves, obtained by applying the linear trapezoidal rule, were compared by means of one-way ANOVAs. Post hoc comparisons were done using t-tests. counts of cells, WST and FACS data, QRT-PCR, and western blot results were analyzed using t-tests or one-way ANOVA. The α level was set at 0.05 and correction according to Bonferroni was employed to account for multiple testing. The number of animals used for the experiments is indicated in the figure legends.
GCH1 inhibition kills cancer cells in vitro
We assessed effects of GCH1 inhibition on tumor cell viability and survival at concentrations that inhibit the catalytic GCH1 activity and BH4 production. The EC50 was previously shown to range from 1.5 to 3 mM in cell culture experiments.23, 24 A concentration dependent decline of cell viability was seen in the water soluble tetrazolium (WST) assay (IC50 = 1.1 mM, Fig. 1a) and a reduction of colony formation in the colony forming assay in MCF-7 human breast cancer and B16 mouse melanoma cells [IC50 = 1.5 and 3.5 mM, respectively, Figure 1b, IC50 = 2.3 mM in human HT29 colon cancer cells (not shown)]. The effects of the GCH1 inhibitor DAHP were mimicked by inhibition of sepiapterin reductase, i.e., the last step in BH4 synthesis and were partially reversed by BH4 and by the BH4 precursor, sepiapterin, but not by the NO-donor NOC-12 (Fig. 1c). Neither BH4 nor NOC-12 had any effect alone (Fig. 1c). Flow cytometric analysis of the cell cycle distribution showed a G1-phase arrest at 1 mM DAHP in both MCF-7 human breast cancer and B16 mouse melanoma cells and an increase of apoptotic sub-G1 or dead cells at higher concentrations (Fig. 1d). The results were confirmed by flow cytometry analysis of Annexin IV and 7-AAD DNA labeling (Fig. 1e) in both cell lines, which showed mainly signs of late apoptosis and necrosis whereas indices of early apoptosis were weak. Western Blot analyses showed a reduction of phosphorylated histone H3 with DAHP treatment (1 mM, 24 hr) suggesting a reduction of cell proliferation (Fig. 1f).
GCH1 inhibition reduces angiogenesis
Endothelial NO has been suggested to promote angiogenesis. As its production depends on BH4, we studied the effects of GCH1 inhibition on the formation of capillary-like structures of HUVECs on a basement membrane matrix. DAHP treatment dose-proportionally reduced the tube formation of HUVECs (IC50 = 1.5 mM, Figs. 2a and 2b). HUVECs upregulated GCH1 upon hypoxia (1% O2 for 24 hr), which was not affected by DAHP treatment (Fig. 2c). However, DAHP reduced the expression of endothelial NOS (eNOS) both in normal and hypoxic HUVECs (Fig. 2c). In addition, PECAM/CD31 immunoreactive blood vessel like structures were reduced in B16 tumors of C57BL6 mice treated with DAHP (2 × 100 mg/kg/day perorally) as compared with vehicle treated control mice (Fig. 2d).
GCH1 inhibition and downregulation attenuate in vivo tumor growth and angiogenesis
To assess the relevance of the in vitro results we tested the effects of twice daily DAHP treatment in nude mice on the in vivo tumor growth of HT29-Luc colon cancer cells, which express firefly luciferase. The oral DAHP doses of 2 × 100 mg/kg/day are nontoxic and produce plasma concentrations in the range of the IC50 effectively inhibiting BH4 synthesis.9 Results of in vivo luciferase imaging showed a reduced progression of tumor growth in the DAHP treatment group as compared to vehicle (Figs. 3a and 3b). The same was observed on B16-Luc mouse melanoma growth in C57BL6 mice, which represents a syngeneic tumor model (not shown). The time course of changes in the dimensions of HT29 tumors in nude mice confirmed the results of luciferase imaging (Fig. 3d).
To assess in vivo angiogenesis we injected NIR labeled intergrin αvβ3 (IntegriSense), which specifically binds to newly formed blood vessels and therefore accumulates in tumors and highlights tumor angiogensis (Figs. 3c and 3e). Imaging of IntegriSense showed a strong reduction of tumor angiogensis in DAHP treated mice. Again, the result obtained in HT29 tumors was confirmed in B16 melanomas (not shown). Macroscopic inspection of tumors confirmed the reduction of blood vessels following DAHP treatment (Fig. 3f). DAHP treatment resulted in a reduction of plasma biopterin concentrations by 50% (Fig. 3g). Neopterin plasma concentrations directly indicating the enzymatic activity of GCH1 and not influenced by BH4 recycling were elevated in tumor bearing mice, which was prevented by DAHP (Fig. 3h). Overall, neopterin plasma concentrations were however very low. To assess whether the tumor itself contributed to plasma levels we assessed biopterin and neopterin in HT29 cells and in the cell culture medium under naïve conditions and upon exposure to low oxygen (Fig. 3i). Neopterin was below quantification limit. Biopterin concentrations increased in the cells upon hypoxia whereas its release into the medium was reduced (Fig. 3i).
To detect (i) potential off-target effects of DAHP and (ii) the specific role of GCH1 in the tumor cells versus GCH1 expression in surrounding tissue and blood vessels, we assessed the ability of shGCH1 transduction of tumor cells to mimic the DAHP-mediated inhibition.
HT29 colon cancer cells were transduced with a lentiviral vector with small hairpin RNA specific for the human GCH1. The efficacy of transduction and GCH1 downregulation was confirmed by QRT-PCR and Western blotting (Fig. 4a). It resulted in a decrease of forskolin stimulated production of biopterin and NO (Figs. 4b and 4c). The stimulated biopterin increase was reduced by 50%, which is in line with the assumption that de novo synthesis accounts for half of the BH4 pool in the cells while the rest is recycled from BH2. Neopterin concentrations were below quantification limit. The colony forming capacity of shGCH1-transduced HT29 cells was not significantly reduced (Fig. 4d, n.s.). However, cell cycle analysis showed an increase of apoptotic subG1 cells upon GCH1 silencing (Fig. 4e).
Luciferase imaging (Figs. 4f and 4g) and measuring tumor dimensions (Fig. 4h left, i) showed that the progression of shGCH1-transduced HT29-Luc xenotransplants in nude mice was significantly reduced as compared to HT29-Luc cells transduced with the control lentivirus vector. In addition, imaging of IntegriSense showed reduced blood vessel like structures (Fig. 4h right) suggesting that the release of BH4 from the tumor cells was promoting tumor angiogenesis.
GCH1 inhibition fortifies the antitumoral immune response
Since the increase of plasma neopterin concentrations has been associated with an unfavorable outcome in cancer patients and is likely to reflect tumor evoked effects on the immune response,1, 2 plasma chemokine, and cytokine profiles were assessed in tumor bearing C57BL6 mice using the syngeneic B16 melanoma model. Vehicle treatment was associated with tumor growth and a raise in immunosuppressive factors including CXCL13 and IL-10. This was attenuated by DAHP, which also led to an increase in chemokine IP10/CXCL10 and RANTES/CCL5 concentrations, which have been associated with an antitumoral immune response [Fig. 5a; complete result of cytokine/chemokine array in Fig. 1 (Supporting Information)]. Immunostainings showed a higher number of F4/80 immunoreactive myeloid cells, which are either macrophages or dendritic cells in tumor surrounding tissues. These cells were infiltrating the tumor in mice treated with DAHP (Fig. 5b). FACS analyses of freshly dissociated tumors were done for surface markers of various myeloid cell lineages to dissect out the effects of DAHP on macrophage polarization. This showed an increase of tumor associated macrophages in DAHP treated mice (CD45+/CD11b+, Fig. 5c). Subtype analysis suggested a relative increase in tumor-associated macrophages (TAM) of the “classically activated” M1-like phenotype, which is considered to be cytotoxic to tumor cells (CD45+/CD11b+/CD206−),21, 25 and a relative decrease of the alternatively activated, proangiogenic M2 phenotype (CD45+/CD11b+/CD206+) upon treatment with DAHP (Fig. 5d). The number of antigen presenting tumor-associated macrophages and dendritic cells, identified by MHC-II expression, was increased following DAHP treatment (Fig. 5e). The immune cell analysis would be compatible with an increase of tumor attacking immune cells in DAHP treated mice and a reduction of tumor-associated immune cells promoting tumor growth and angiogenesis.
GCH1 upregulation in mouse and human colon and skin cancer
The relevance of GCH1 for human cancer was assessed by studying GCH1 expression in tumors and healthy surrounding tissues motivated by results in mice, which showed an upregulation of GCH1 in mouse AOM/DSS induced colon cancer (Fig. 6a). In human colon cancer tissue arrays, GCH1 immunoreactivity was weak in both, healthy control tissue and highly differentiated colon cancers (G1, T1-2 tumors) (Fig. 6b). However, in grade 2–3, T3–4 cancer tissues, GCH1 immunoreactivity was strong and abundant. Similarly, GCH1 expression was low in healthy human skin but strongly upregulated in melanoma and skin sarcoma and weakly upregulated in squamous cell carcinoma (Fig. 6c). The immunostainings suggested that tumor growth was positively correlated with GCH1 expression.
BH4 is an essential cofactor in the synthesis of biogenic amines and nitric oxide26 and as such, has beneficial effects in the cardiovascular system particularly in diabetics.15, 17 It is also crucial for dopamine synthesis in neurons and its absence following loss-of function mutations causes hereditary dopamine-responsive dystonia.27 However, excessive amounts of BH4 cause neuronal damage28, 29 and enhance nociceptive sensitivity.9, 30 BH4 overproduction is caused by overexpression of GCH1.9, 31, 32 The present results clearly demonstrate that prevention of excessive BH4 production either by pharmacologic inhibition or by RNA interference reduces tumor growth and angiogenesis in vitro and in vivo and shifts the immune response towards antitumoral defense. This was reflected in a positive correlation of human tumor grading with GCH1 expression in colon and skin cancer tissues.
The exact mechanism of the BH4 mediated survival functions in cancer cells remains nevertheless unclear. Cells might require BH4 for maintenance of cellular functions, probably in part owing to its role in NO synthesis. Some tumor cell lines express inducible NOS upon cytokine stimulation or exposure to cytotoxic agents.33–35 It was observed that iNOS deficient HT29 colon cancer cells were resistant to doxorubicin-evoked cell death36 suggesting that NO production within the tumor cells may increase chemosensitivity. However, in vivo NO-release by tumor cells may also favor tumor growth33 by formation of new capillary-like structures of blood vessels37, 38 and a shift of immune responses towards M2-polarization of macrophages.20, 21, 25 Therefore, the effects of BH4 inhibition on cancer need to be evaluated in the context of the tumor environment and the complex interactions of tumor cells, immune cells, and vascular cells. Moreover, since BH4 is secreted39 and thus available for neighboring cells and also produced by endothelial or immune cells, de novo BH4 synthesis within cancer cells may be dispensable. In the present experiments, hypoxia led to an upregulation of GCH1 expression in endothelial and HT29 colon cancer cells. It is likely that this mechanism also accounts for increased GCH1 expression in ischemia models in mice40 and cardiovascular disease in humans.16 In cancerous tissue this effect would sustain BH4 and nitric oxide production to promote blood flow to the tumor and facilitate endothelial cell migration and tube formation. Effects of nitric oxide in tumor cell proliferation and angiogenesis are mediated in part by upregulation of sphinogsine-1-phosphate (S1P) production,41, 42 which is released by apoptotic tumor cells and contributes to the tumor-evoked shift of the immune response.43 Dying tumor cells may release further substances that act in concert to abolish antitumoral immune responses by polarization of macrophages into tumor-supporting and proangiogenic M2-like cells.44 In addition, tumor-evoked immunosuppressive effects involve a shift of chemokine and cytokine plasma profiles45–47 and increase of myeloid derived mononuclear and polymorphonuclear suppressor cells.48, 49 The here observed relative increase of plasma concentrations of IP10/CXCL10 and CCL5/Rantes and decrease of plasma IL-10 and of M2-like tumor associated macrophages suggest that the inhibition of GCH1 allowed for a readjustment of the immune response towards tumor attack. The exact mechanism remains to be elucidated. It may be hypothesized that immune cell iNOS is uncoupled upon depletion of its essential cofactor BH4, which will result in an increased production of reactive oxygen species instead of NO, and facilitate oxidative tumor cell killing abilities. A similar DAHP-evoked uncoupling effect on NOSs within the tumor cells may lead to tumor cell death.
In summary, the inhibition of GCH1 in vivo in the context of tumor growth depends on a concerted effect in tumor cells, endothelial cells, and immune cells. It is likely that effects are in part mediated by suppression of NO production. However, BH4 also acts as coenzyme for biogenic amine and serotonin production and the metabolism of lipid-ethers by the recently identified protein TMEM195/AGMO.50 Therefore, inhibition of BH4 synthesis may be more efficient than inhibition of nitric oxide synthesis. The effects of GCH1 inhibition need to be further tested in combination with cytotoxic agents or immune targeting therapies to evaluate its efficacy as therapeutic anticancer approach.