NAD(P)H Oxidase 1, a Product of Differentiated Colon Epithelial Cells, Can Partially Replace Glycoprotein 91^phox in the Regulated Production of Superoxide by Phagocytes

Cell culture

HT29 cells (American Type Culture Collection, Manassas, VA) were maintained in McCoy’s 5a medium containing 10% FBS, glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml). Caco2 cells (American Type Culture Collection) and PT67 cells (Clonetics, San Diego, CA) were maintained in DMEM containing 10% FBS (HyClone, Logan, UT), glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml). gp91^phox-deficient PLB-985 cells (X-CGD PLB-985; provided by Dr. M. Dinauer) (11) and K562 cells (12) were grown in suspension cultures in RPMI 1640 medium containing 10% FBS, glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml). Clonally derived p67^phox- and p47^phox-expressing K562 lines were isolated following transduction with MFG-S retroviruses, using methods described previously (13, 14). For induction of a granulocytic phenotype, PLB-985 cells were treated with 0.3 mM dibutyryl cAMP (Sigma-Aldrich, St. Louis, MO) for 3 days. Cell number and viability were determined by trypan blue exclusion.

Transfection of HT29 cells

The coding sequence for human Nox1 was amplified from Caco2 cDNA (AMV reverse transcriptase-derived) using primers containing NotI and EcoRI sites. For antisense studies, the coding sequence of human Nox1 was subcloned in an antisense orientation into pcDNA3.1⁺ (Invitrogen, San Diego, CA). HT29 cells were transfected at 60–70% confluence, with pcDNA3.1-asNox1 or the empty pcDNA3.1 vector using GENEPORTER (Gene Therapy Systems, San Diego, CA), selected with G418 (2 mg/ml) started 48 h later; resistant colonies were isolated at 2 wk.

Northern blotting, multiple tissue expression array blotting, RT-PCR, and in situ hybridization

Total RNA was prepared from 10⁷ cells (15 μg), electrophoretically separated on 1% agarose formaldehyde gels, and transferred to nylon membranes. Membranes were probed at 65°C with a radiolabeled, full-length, human Nox1 cDNA fragment (Amersham Pharmacia Biotech, Arlington Heights, IL), by standard hybridization protocols. The Nox1 cDNA was labeled with Prime-It RmT Random Primer Labeling Kit (Stratagene, La Jolla, CA). This fragment was also used to probe a human RNA Multiple Tissue Expression Array (BD Biosciences, Franklin Lakes, NJ) by the same methods.

For RT-PCR experiments, 5 μg of total RNA was transcribed into cDNA using a First Strand cDNA Synthesis Kit (Roche, Indianapolis, IN). Human Nox1 cDNAs were amplified using primers 5′-GTTGTTATGCACCCATCCAAAGTT-3′ (sense) and 5′-TCAAAAATTTTCTTTGTTGA-3′ (antisense) and were confirmed by sequencing.

For in situ hybridization experiments on mouse colon, Nox1 cDNA was subcloned into pBluescript KS vector (Stratagene). ³⁵S-Labeled transcripts (sense or antisense) were synthesized by SP6 or T7 RNA polymerases using linearized templates. Preparation and probing colon thin section specimens were performed as described previously (15).

Multitumor tissue microarray slides (TARP1 and -2) were obtained from the Cooperative Human Tissue Network under the Tissue Array Research Program (TARP) of National Cancer Institute, National Institutes of Health (http://resresources.nci.nih.gov/tarp). For in situ hybridization experiments on multitumor tissue microarrays, a 564-bp fragment of human Nox1 cDNA (3′-untranslated region) was cloned into PCR 4.0 TOPO TA-cloning vector (Invitrogen). ³⁵S-Labeled RNA transcripts (sense or antisense) were synthesized by T7 or T3 RNA polymerases. Methods for hybridization probing of the microarray slides were described previously (15).

Production of Nox1 MFG-S onco-retrovirus and transduction of PLB-985 and K562 cells

The human Nox1 cDNA was cloned into the NcoI site of MFG-S retroviral vector (16) following PCR amplification with primers containing terminal NcoI sites. Following sequence confirmation of this construct, PT67 packaging cells were cotransfected with the MFG-S-Nox1 plasmid (15 μg) and pcDNA 3.1⁺ plasmid (1.5 μg) using 25 μl of GenePorter (Gene Therapy, Gaithersburg, MD). For controls, MFG-S-enhanced green fluorescence protein (EGFP) was used to transfect PT67 cells using the methodology used to generate MFG-S-EGFP onco-retrovirus. Cells were selected 48 h later (0.8 μg/ml G418) and isolated as individual colonies. Stable, high titer, virus-producing cell lines were identified by RT-PCR screening of culture supernatants for virus-derived Nox1 cDNA or by examination of green fluorescence (MFG-S-EGFP). For target cell transduction experiments, 8- to 14-h virus-producing culture supernatants were harvested from subconfluent (70–80%) PT67 cultures. Fresh supernatants (25% dilutions in RPMI 1640 medium, with 10% FBS and 6 μg/ml protamine) were added to PLB-985 or K562 cells in six-well plates (150,000 cells/well) using previously described “spinoculation” methods (13). Cells were transduced with Nox or EGFP virus repeatedly (six to nine rounds) and monitored in control cultures for green fluorescence.

Mobilized CD34⁺ peripheral blood stem cell (PBSC) culture and transduction procedures

G-CSF-mobilized human CD34⁺ PBSC were obtained from an X-linked chronic granulomatous disease (X-CGD) patient (X-CGD CD34⁺ PBSC) after obtaining informed consent (National Institutes of Health institutional review board protocols 94-I-0073 and 95-I-0134) (13), seeded at 0.5 × 10⁶ cells/well in Retronectin (Takara Shuzo, Kyoto, Japan)-precoated, six-well plates (17), and grown in 4 ml of growth medium (X-VIVO10/1% human serum albumin; 50 ng/ml FLT3 ligand, 50 ng/ml stem cell factor, 20 ng/ml IL-6, 10 ng/ml thrombopoietin, 10 ng/ml IL-3, and 10 ng/ml G-CSF) at 37°C in a 7% CO₂ atmosphere. The X-CGD CD34⁺ PBSCs were transduced 16 h later with either MFGS-Nox1 or MFGS-EGFP virus for 6 h/day for 6 consecutive days using 75% of respective undiluted virus supernatant and 5 μg/ml protamine. At the start of each transduction, the plates were centrifuged at 1200 × g at 32°C for 20 min (spinoculation) (13). After each 6-h transduction period, cells were transferred into fresh medium. CD34⁺ cells were maintained in liquid culture for up to 28 days and were analyzed for PMA-stimulated superoxide production. The gp91^phox expression was determined on a FACSort (argon laser; BD Biosciences) by indirect staining with murine mAb 7D5, followed by an FITC-conjugated goat anti-mouse IgG Ab (18). GFP expression was analyzed on the same instrument. Cell viability (trypan blue exclusion) was assessed every other day and ranged from 78–95%.

Measurement of superoxide production

Superoxide production was determined by chemiluminescence using the DIOGENES reagent (National Diagnostics, Atlanta, GA), a superoxide-specific chemiluminescence reagent (12). Cells were washed once in 1× HBSS lacking Ca²⁺ and Mg²⁺. Measurements were performed in 96-well microtiter chemiluminescence plates (10⁵ cells/well) in HBSS at 37°C for 40 min following stimulation with PMA (2 μg/ml) using a Luminoskan luminometer (Labsystems, Helsinki, Finland). The total luminescence was completely sensitive to superoxide dismutase.

Measurement of cell proliferation

For cell proliferation studies, transfected HT-29 cells were plated at 50,000 cells/well and were treated with the indicated compounds on day 1. Total viable cell counts were determined on day 5 using trypan blue staining. Unless otherwise indicated, counts were averaged from wells plated in triplicate.

Analysis of Nox1 expression in the mouse colon

Although the Nox1 mRNA transcript was originally isolated from the colon epithelial carcinoma cell line, Caco2, it is also detected in vascular smooth muscle cells (4). Therefore, we examined which colon cells express Nox1. In situ hybridization on mouse colon (Fig. 1⇓) revealed that Nox1 is present in colon epithelial cells, and that highest expression occurs in the lower two-thirds of the colon crypt (Fig. 1⇓E); the surface epithelium signals resembled negative control background signals. The results from these in situ hybridization experiments indicate an equal distribution of Nox1 mRNA throughout the mouse large intestine. Hybridization experiments on human multiple tissue array RNA blots also showed high Nox1 expression in the entire colon and rectum, and much lower expression in the small intestine (jejunum, ileum) and the ileocecal region (Fig. 1⇓G). These experiments confirmed the low uterine and prostate Nox1 mRNA levels reported previously (4). In addition, the absence of high Nox1 levels outside the colon suggested a relatively specific function for Nox1. These expression patterns were verified by PCR amplification of several human gastrointestinal cDNAs (data not shown). We also observed Nox1 expression in Caco2, HT29 (Fig. 1⇓H), and T84 human colon carcinoma lines (not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Detection of Nox1 transcript in colon epithelial cells. A–F, In situ hybridization of Nox1 mRNA to transverse sections of mouse colon. A and D, H&E staining. B and E, Hybridization with Nox1 antisense probe detects Nox1 mRNA within the lower two-thirds of colon crypts. C and F, Control hybridization with sense probe reveals background signal only within argentophilic clusters. G, Dot hybridization to RNA from gastrointestinal tissues reveals uniform Nox1 expression within all colon segments (J, jejunum; I, ileum; Ic, ileocecum; Ap, appendix; Ca, ascending colon; Ct, transverse colon; Cd, descending colon; R, rectum). H, Nox1 Northern hybridization to Caco2 and HT29 colon carcinoma lines. Data are representative of three experiments with identical results.

Nox1 expression in human tumor tissues

Initial reports demonstrating the ability of Nox1-expressing NIH-3T3 fibroblasts to form tumors in nude mice suggested that Nox1 functions as a mitogenic oxidase (4). We surveyed Nox1 expression patterns in 600 tumor samples represented on a human multitumor tissue array (http://resresources.nci.nih.gov/tarp; Fig. 2⇓). The tumor samples included cancerous tissues from ovary, breast, prostate, colon, and lung as well as melanoma and lymphoma tumors. We performed in situ hybridization to detect Nox1 mRNA using a probe encompassing the 3′-untranslated region of the human Nox1 transcript to avoid cross-hybridization with other Nox isoforms. This analysis revealed that Nox1 is exclusively expressed in colon cancer samples; other tumors were essentially negative (Fig. 2⇓A). Further histological examination of these individual tumor samples evaluated the grade of differentiation and independently assessed Nox1 expression levels. As shown in Fig. 3⇓, the highest Nox1 expression was observed in more differentiated tumor samples.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Detection of Nox1 mRNA in human colon tumors by in situ hybridization. A, Phosphorimage analysis of TARP1 multitumor tissue microarray probed with Nox1 antisense riboprobe. Hybridization signal strengths: red, strong positive; yellow, intermediate; green, low Nox1 mRNA expression. B, Phosphorimage analysis of a multitumor tissue microarray slide incubated with the sense Nox1 riboprobe (control experiment). C, Positions of tumor tissues represented on this microarray. For more information, visit: http://resresources.nci.nih.gov/tarp. Similar results were obtained with TARP2.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Nox1 mRNA expression in differentiated colon epithelial tumors. A, H&E staining of a representative differentiated colon cancer sample from the tissue array in Fig. 2⇑. B, Darkfield in situ hybridization image demonstrates Nox1 expression within differentiated epithelial cells shown in A (signal appears as white silver grains). Arrowheads indicate positive signals within highly organized epithelial layers. C, H&E staining of a representative poorly differentiated colon cancer tissue sample. D, Darkfield in situ hybridization image demonstrates the comparatively lower Nox1 expression in this poorly differentiated tumor sample.

Induced differentiation of colon adenocarcinoma cell lines enhanced Nox1 expression

Our in situ hybridization experiments on tumor arrays did not support a role for Nox1 in the regulation of mitogenesis in colon cancer. We also investigated whether altered Nox1 levels affect cell proliferation in colon carcinoma cell lines (Caco2 and HT29). We assumed that if Nox1 regulates mitogenesis, changes in expression levels would alter proliferation rates. The active vitamin D₃ metabolite, 1α,25-(OH)₂D₃, is a potent inhibitor of colon cancer cell proliferation (19) that promotes differentiation and apoptosis (20, 21). We were particularly interested in the effect of 1α,25-(OH)₂D₃, because this compound also effectively induces expression of phagocytic oxidase components in myeloid cells (11, 22). Interestingly, 1α,25-(OH)₂D₃ treatment (100 nM) dramatically enhanced Nox1 expression in both Caco2 and HT29 cells (Fig. 4⇓A). The effect was detectable at a concentration as low as 1 nM and was maximal 6–12 h after treatment. The increased Nox1 expression levels were also verified by RT-PCR experiments, in which products corresponding to Nox1 (798 bp) and its alternatively spliced (650 bp) variant, NOH-1Lv (23), were induced (Fig. 4⇓A). In agreement with previous studies, 1α,25-(OH)₂D₃ treatment reduced cell proliferation in HT29 cells (Fig. 4⇓B) and Caco2 cells (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Induced differentiation of colon carcinoma cell lines increases Nox1 expression and decreases cell proliferation. A, Detection of Nox1 mRNA following 0- to 48-h treatments of HT29 (top panel) or Caco2 (middle panel) cells with 100 nM 1α,25-(OH)₂D₃ (D3). Bottom panel, RT-PCR detection of full-length and spliced variants of Nox1 mRNA within untreated and 12-h D3-treated HT29 cell RNA. B, Proliferation assays of untreated and 100 nM D3-treated HT29 cells. Similar results were obtained in five experiments. C, Detection of Nox1 mRNA following 24-h treatments of cells with 1000 U/ml IFN-γ. D, Proliferation assay of untreated and IFN-γ-treated HT29 cells. Data in B and D represent averages of assays preformed in triplicate. Similar results were obtained in three separate experiments.

The 1α,25-(OH)₂D₃-induced differentiation of myelocytic leukemia cell lines, such as HL-60 or PLB-985, induces phagocytic oxidase components (11, 22); IFN-γ also stimulates gp91^phox expression in these cell lines (22), consistent with the presence of IFN-γ-responsive elements within the gp91^phox promotor (24). Since we found that 1α,25-(OH)₂D₃ was an effective stimulator of Nox1 expression, we examined IFN-γ for similar effects on colorectal adenocarcinoma cells. Fig. 4⇑C shows that IFN-γ effectively increases Nox1 expression in Caco2 cells and HT29 cells. We tested other cytokines, TNF-α and IL-1, that do not affect Nox1 mRNA levels (not shown). Consistent with previous reports, IFN-γ inhibits cell proliferation of HT29 and Caco2 cell lines (Fig. 4⇑D).

Effect of antisense suppression of Nox1 expression on HT29 cell proliferation

We also examined cell proliferation when Nox1 expression is suppressed. To inhibit Nox1 expression, we stably transfected HT29 cells with human Nox1 cDNA constructed in an antisense orientation. After G418 selection, we isolated individual colonies and analyzed Nox1 mRNA levels by Northern blotting. One clone (HT29-2) exhibited significantly reduced Nox1 mRNAs levels (Fig. 5⇓A). Compared with empty vector-transfected HT29 clones, we observed similar rates of cell proliferation (Fig. 5⇓B). This result again argues against a direct role for Nox1 in the control of mitogenesis.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Antisense suppression of Nox1 does not affect HT29 cell proliferation. A, Northern blot analysis showing transfection with Nox1 antisense reduces Nox1 mRNA levels in HT29-2 cells compared with control (empty vector-transfected) line. B, Nox1 antisense-transfected (HT29-2) cells proliferate as well as control cells. Assays were preformed in triplicate. Similar results were obtained in three separate experiments.

Functional interaction between Nox1 and components of the phagocytic oxidase

The similarities between the regulation of Nox1 and gp91^phox mRNA expression, described above, prompted us to investigate whether Nox1 has sufficient homology to gp91^phox to enable it to interact with other phox components. To examine this, we created a recombinant MFG-S retrovirus containing the Nox1 cDNA (MFG-S-Nox1) and transduced a PLB-985 cell line in which the gp91^phox gene was disrupted (X-CGD PLB-985) (11). This cell line is a model for X-CGD, capable of differentiating into mature myeloid cells that express all the phox components except gp91^phox. The transduction efficiency was monitored by Northern blot analysis, where the Nox1 mRNA was detected in MFG-S-Nox1-transduced cells (Fig. 6⇓A). After differentiation of these cells with dibutyryl cAMP (72 h), we observed PMA-stimulated superoxide production by the Nox1-transduced PLB-985 cells (Fig. 6⇓B). The activity of these Nox1-reconstituted cells was ∼2% the activity observed in normal PLB-985 cells. Interestingly, there was a significant delay in the onset of this Nox1-mediated oxidative response, unlike a typical PMA-stimulated response of normal PLB-985 cells or neutrophils, which reaches a maximum within 10 min of stimulation. Differentiated, but unstimulated, Nox1-transduced PLB-985 cells did not produce superoxide; their activity was similar to that of the stimulated MFG-S-EGFP-transduced X-CGD PLB-985 cells (both <1% of differentiated Nox1-transduced X-CGD PLB-985 cells). More importantly, PMA-stimulated, but undifferentiated, Nox1-expressing cells did not produce significant superoxide (also <1% of differentiated Nox1-transduced X-CGD PLB-985 cells), suggesting that proteins expressed during differentiation support Nox1 activity. It is well known that phox components are among the proteins produced during differentiation (11, 22). Thus, Nox1 required both the presence of complementary factors and an activation stimulus for superoxide generation, in contrast to the results reported by Suh et al. (4), who observed constitutive, low level superoxide production in Nox1-transfected fibroblasts. We also used the K562 erythroleukemia cell model that constitutively expresses only p22^phox (12) to study the possible interaction between the cytosolic phox components and Nox1. The p47^phox and p67^phox cDNAs were introduced by retroviral transduction, followed by clonal selection for high expression of one or both factors. Compared with control Nox1-transduced K562 cells lacking p47^phox and p67^phox (Nox1⁺, p47⁻, p67⁻), the coexpression of p47^phox (Nox1⁺, p47⁺, p67⁻) demonstrated increased PMA-stimulated superoxide production, while coexpression of p67^phox alone (Nox1⁺, p47⁻, p67⁺) had minimal effects on Nox1 activity (Fig. 6⇓C). Significantly higher (>100-fold) superoxide production was observed when both p47^phox and p67^phox (Nox1⁺, p47⁺, p67⁺) were present. These results clearly indicate that Nox1 interacts with phagocytic oxidase components in a regulated fashion and raise the possibility of a functional homology between the phagocytic and intestinal oxidase enzymes.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

MFG-S-Nox1 retrovirus restores superoxide release in promyelocytic X-CGD PLB-985 and K562 cell lines. A, Detection of Nox1 RNA by Northern blot analysis in MFG-S-Nox1-transduced X-CGD PLB-985 cells (lane a) compared with control MFG-S-EGFP-transduced cells (lane b). B, PMA-stimulated superoxide production in differentiated, MFG-S-Nox1-transduced PLB-985 cells (▪) and MFG-S-EGFP-transduced cells (▴). The total integrated activity of Nox1-transduced X-CGD PLB-985 cells was 95 relative light units, vs 4477 relative light units observed in normal differentiated PLB-985 cells. C, PMA-stimulated superoxide production in MFG-S-Nox1-transduced K562 cells: Nox1⁺, p47⁺, p67⁺ (▪); Nox1⁺, p47⁺, p67⁻ (▴); Nox1⁺, p47⁻, p67⁺ (♦); and Nox1⁺, p47⁻, p67⁻ (X) (total integrated activities for theses curves were 13.36, 3.26, 0.55, and 0.98, respectively). Results are representative of five independent experiments, confirming the same relative activities.

Superoxide production by Nox1-transduced X-linked CGD patient CD34⁺ cells

Our results from experiments with Nox1-transduced X-CGD PLB-985 and K562 cells suggest that heterologous Nox1 expression could correct the superoxide deficiency of X-CGD phagocytes. To examine this possibility in a more relevant oxidase-reconstituted system, we transduced CD34⁺ PBSC mobilized from a gp91^phox-deficient X-CGD patient (X-CGD CD34⁺PBSC) in vitro with Nox1-MFG-S retrovirus. At 3 wk of culture the superoxide produced by the Nox1-transduced X-CGD CD34⁺ PBSC was as much as 3% the superoxide generation measured in normal CD34⁺ PBSC (Fig. 7⇓), while the nontransduced and EGFP mock-transduced X-CGD CD34⁺ PBSC produced insignificant superoxide levels. To determine whether the retroviral transduction procedure in some way caused gp91^phox production in these cells, thereby accounting for this increase in superoxide, flow cytometry was performed to monitor gp91^phox expression. This analysis confirmed that gp91^phox was not detectable in these cells, and that the restored oxidase activity is attributable to transduced Nox1.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Superoxide production by X-CGD CD34⁺ PBSC. PMA-stimulated superoxide production by 21-day differentiating X-CGD C34⁺ PBSC of nontransduced, EGFP-transduced and Nox1-transduced cultures is shown. Data represent averages of triplicate assays; similar results were obtained in separate assays performed 2 days later.