Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Neutrophil inhibitory factor (NIF), a 4-kDa antiadhesive glycoprotein has been isolated and cloned from the canine hookworm Ancylostoma caninum (Moyle et al., 1994). NIF at nanomolar concentrations inhibits polymorphonuclear leukocyte (PMN) adhesion to the endothelium (Moyle et al., 1994; Barnard et al., 1995) and PMN adhesion-dependent generation of H₂O₂ (Moyle et al., 1994). In an ex vivo guinea pig lung model, we showed that NIF infusion also prevented PMN-mediated lung vascular injury (Barnard et al., 1995). In addition, NIF release in a transgene mouse model prevented lipopolysaccharide-induced PMN infiltration into lungs, and the resultant lung vascular injury (Zhou et al., 1998). Other studies in a rat model of focal cerebral ischemia showed that NIF blocked PMN infiltration into ischemic brain tissues (Jiang et al., 1995). These effects of NIF were ascribed to inhibition of PMN adhesion to the microvascular endothelium (Moyle et al., 1994; Barnard et al., 1995).

Because CD18 integrins are required for firm and stable PMN adhesion to the endothelium (Smith et al., 1989; Springer 1990; Butcher, 1991; Malik and Lo, 1996), proteins such as NIF that interfere with CD18 function may have applications in inhibiting PMN adhesion to the vascular endothelium and transendothelial PMN migration. Two members of the CD18 integrins on PMNs, CD11a and CD11b, mediate the firm PMN adhesion by binding to different sites on intracellular adhesion molecule-1 (ICAM-1) (Diamond and Springer, 1993). Studies showed that NIF inhibited CD11b function by binding to the I domain, contained in a stretch of ~200-amino acid sequence critical for ICAM-1 binding (Muchowski et al., 1994; Rieu et al., 1994). Because anti-CD11b monoclonal antibodies (mAbs) at saturation concentrations only inhibited PMN adhesion by ~50% (Lo et al., 1989), NIF binding to CD11b cannot fully explain the effects of NIF in abrogating PMN adhesion to endothelial cells in both in vitro and in vivo studies (Moyle et al., 1994; Barnard et al., 1995). In the present study, we show that inhibition of PMN induced by NIF requires the functional inactivation of both CD11a and CD11b ₂ integrins. In addition, we show by the gene transfer approach that stable transfection of endothelial cells with NIF cDNA resulted in NIF release, which in turn prevented PMN adhesion by inhibiting the functions of CD11a and CD11b ₂ integrins on PMNs.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Vector Construction. An 850-bp EcoRI fragment containing the entire NIF cDNA was subcloned into the retroviral plasmid pLNCX (provided by Dr. A. Dusty Miller, Fred Hutchinson Cancer Research Center, Seattle, WA). This 6.2-kilobase plasmid contains a neomycin resistant gene and the immediate early promoter of the human cytomegalovirus (CMV) that drives the NIF cDNA. The full-length NIF cDNA, neomycin resistant gene, and CMV promoter are situated between two long terminal repeats required for DNA integration into the host cell genome. Orientation of NIF cDNA relative to the CMV promoter in pLNCX was verified by restriction enzyme analysis (Sambrook et al., 1989). The ligated DNA (pLNC-NIF) was transformed into Escherichia coli DH5 competent cells (Life Technologies, Gaithersburg, MD).

Cell Culture. PA317 cells (American Type Culture Collection, Rockville, MD) were used to produce amphotropic viral particles (Miller and Buttimore, 1986; Miller and Rosman, 1989). These cells were cultured in Dulbecco's modified Eagle's medium (DMEM) with 4.5 g/l glucose and 10% fetal bovine serum (FBS) (Hyclone, Logan, UT), and were grown to confluency and passaged every 2 to 3 days.

Generation of Retroviral Particles Containing NIF cDNA. Both pLNCX and pLNC-NIF constructs were introduced into packaging PA317 cells with the standard lipofectamine transfection method (Life Technologies). Briefly, the cells were plated at a density of 5 × 10⁵ cells/60-mm dish on day 1. On day 2, the cells were incubated in serum-free and antibiotic-free PA317 medium containing 6 µl/ml lipofectamine and 1 µg/ml pLNC-NIF or the control pLNCX plasmid DNA. On day 3, the cells were washed two times with Hanks' Balanced Salt Solution (HBSS), and covered with serum-containing culture medium. Cells were selected for expression of neomycin phosphotransferase by incubation in the presence of G418 (500 µg/ml) (Life Technologies) starting on day 4 and continuing to day 13. The cell medium was changed every 2 to 4 days during the selection and cells were passaged 1:2 when confluent. PA317 cells surviving G418 selection (i.e., the "producer cells") were then grown to confluency in absence of G418, and the conditioned medium containing the replication-incompetent retroviral particles was used to transduce the endothelial cells as described below.

Transduction of Endothelial Cells with NIF cDNA. Cells were plated at 5 × 10⁵ cells/60-mm dish on day 1. On day 2, fresh conditioned medium from confluent producer PA317 cells was collected and filtered through a 0.45-µm nylon filter, mixed with endothelial cell medium at a 1:1 ratio, and added to the HBSS-washed cells, then the incubation was continued. These steps were repeated on days 3 and 4. On day 5, the endothelial cells were washed with HBSS and fed with endothelial cell medium. The endothelial cells were selected for expression of neomycin phosphotransferase with 500 µg/ml G418 from day 7 to at least day 16. The medium was changed every 3 to 4 days during selection. The NIF-transduced endothelial cells had identical growth characteristics and cell morphology as the nontransduced cells (data not shown).

Northern Blot Analysis. Total RNA was isolated from confluent endothelial cells with TriReagent (Molecular Research Center, Inc., Cincinnati, OH). Twenty micrograms of RNA/sample was size-fractionated by electrophoresis through 1% agarose-2.2 M formaldehyde gel and transferred to nitrocellulose membranes (Bio-Rad, Hercules, CA). The membranes were hybridized with a ³²P (800 µCi/mmol; Amersham Corp., Arlington Heights, IL) labeled 850-bp EcoRI NIF cDNA fragment generated by random priming as described in Rahman et al. (1999). After stringent washes, autoradiography was performed by exposing membranes to X-ray film at 70°C for 1 to 5 days. NIF mRNA was present in NIF-transduced HMEC up to 20 passages (data not shown).

Protein Analysis. Endothelial cells were plated on 60-mm dishes and grown to 80% confluency. The growth medium was removed, and the cell monolayers were washed two times with PBS. Methionine-free DMEM with 4.5 g/l glucose (Life Technologies) was added and the cells were incubated for 2 to 3 h at 37°C in 5% CO₂. The medium was then replaced with fresh methionine-free DMEM containing 150 µCi [³⁵S]methionine (Translabel; ICN Radiochemicals, Irvine, CA), and allowed to incubate for 24 h. NIF was immunoprecipitated by adding 5 µl (1:200 dilution) of rabbit polyclonal antibody raised against the recombinant NIF (Corvas International, San Diego, CA) to the aliquots of conditioned medium from the plates, and incubated overnight at 4°C with gentle rocking. The immunocomplex was pelleted with protein G-Sepharose 4FF (Pharmacia Inc., Piscataway, NJ), rinsed three times with wash buffer (0.05% Tween 20, 120 mM NaC1, 2 mM CaCl₂, 20 mM Tris-HC1, pH 8), solubilized by heating in sample buffer, and subjected to electrophoresis through SDS-polyacrylamide gels (Laemmli, 1970). The gels were soaked in 1% glycerol and dried, and the ³⁵S radiolabelled bands were visualized by autoradiography at 70°C for 1 to 3 days.

PMN Isolation and Labeling. PMNs were isolated from whole human blood donated by healthy volunteers with a single-step separation over a sodium metrizoate/Dextran solution (Polymorphprep, Nycomed Pharma AS, Oslo, Norway) followed by hypotonic lysis of erythrocytes. The viability of the isolated PMNs was >95%. PMNs were labeled with the fluorescent dye calcein/AM as described in Marks et al. (1991).

PMN Adhesion Assay. Endothelial cells were seeded at 5 × 10⁴ cells/well in 96-well plates and grown to confluence over 3 days. PMN adhesion to fibrinogen or endothelial cells was determined as described in Barnard et al. (1995). Phorbol-12-myristate-13-acetate (PMA) (Sigma Chemical Co.) was added to the wells at final concentrations of 10 or 20 nM to activate PMNs, and the cells were incubated for 10 min at 37°C in 5% CO₂. In a control group, the NIF polyclonal Ab (final concentration of 150 nM) was added to endothelial cells 30 min before the adhesion assay. For assays evaluating tumor necrosis factor-alpha (TNF-)-induced increase in endothelial adhesivity, TNF- was added to endothelial cell monolayers at a final concentration of 1000 U/ml and incubated for 3 h at 37°C in 5% CO₂ to activate endothelial cells. The anti-NIF polyclonal (150 nM) or anti-ICAM 1 mAb RR1/1 [10 µg/ml final concentration (supplied by Dr. R. Rothlein, Boehringer-Ingelheim, Ridgefield, CT)] was added to endothelial cell monolayers 30 min before addition of the labeled PMNs.

Flow Cytometry. We used fluorescence-activated cell-sorting analysis (FACS) to determine the expression of cell adhesion molecules on endothelial cells and PMNs with specific mAbs. Endothelial cells or PMNs were harvested and incubated with antiadhesion molecules mAbs (10 µg/ml) at 4°C for 20 min. Cells were washed followed by incubation with fluorescein isothiocyanate-labeled F(ab)'₂ fragments of goat anti-mouse antibody at 10 µg/ml for 20 min at 4°C. Cells were washed before FACS analysis.

JY Lymphoblastoid Cell Aggregation Assay. We adopted an established JY cell aggregation assay to detect NIF's ability to block JY cell aggregation (Rothlein et al., 1985). Briefly, 2 × 10⁵ cells in 100 µl RPMI 1640 medium with 10% FBS were added to a flat-bottomed 96-microtest plate. PMA (200 ng/ml, 18 h) was then added, viewed under an inverted microscope, and JY cell aggregation was scored. Scores ranged from 0 to 5+, where 0 indicated that no cells were in clusters; 1+ indicated <10% of the cells were in aggregates; 2+ indicated that 10 to 50% of the cells were in aggregates; 3+ indicated that 50 to 100% of the cells were in small, loose, aggregated clusters; 4+ indicated that up to 100% of the cells were aggregated in large clusters; and 5+ indicated that 100% of the cells were in large, compact aggregates.

Statistical Analysis. Data are presented as means ± S.E. Comparisions between experimental groups were made by ANOVA and the Wilcoxon test with significance value set at P < .05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

NIF Inhibits CD11a- and CD11b-Dependent PMN Adhesion. We determined the ability of NIF to inhibit CD11b function with the in vitro fibrinogen PMN binding assay, which is dependent solely on CD11b integrin (Diamond and Springer, 1993). NIF prevented binding of PMA-activated PMNs to fibrinogen in a dose-dependent manner with complete blockade of PMN binding occurring at ~10 nM NIF (Fig. 1a). Anti-CD11b mAb (OKM10) also abrogated PMN binding to fibrinogen as expected (Fig. 1a, inset). We also determined the effects of NIF in preventing CD11a- and CD11b-dependent PMN adhesion to endothelial cells. As shown in Fig. 1b, NIF prevented PMA-induced PMN adhesion to endothelial cells in a dose-dependent manner (with full inhibition occurring at ~10 nM); the inhibition induced by NIF was the same as that with an anti-CD 18 mAb (IB4) (Fig. 1b, inset).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   a, NIF prevents PMN binding to fibrinogen (Fg) immobilized onto glass slides with human serum albumin (1 mg/ml). PMNs were layered onto the fibrinogen-coated surface and immediately activated by PMA treatment (10 nM, 10 min) in the absence and presence of indicated concentrations of recombinant NIF. PMN adhered to fibrinogen in a CD11b-dependent manner as indicated by inhibition with anti-CD11b mAb (OKM10, 10 µg/ml) (inset). NIF produced a dose-dependent inhibition of PMN binding to fibrinogen. Data represent means ± S.E. of three separate experiments. b, NIF inhibits PMN adhesion to endothelial cells (EC). PMNs were layered onto confluent HUVEC monolayer and recombinant NIF at different concentrations was added. PMNs were activated by PMA treatment (10 nM, 10 min) and adhesion was allowed to proceed at 37°C. PMNs adhered to HUVEC in a CD18-dependent manner as indicated by inhibition with anti-CD18 mAb (IB4, 10 µg/ml) (inset). NIF inhibited PMN adhesion to HUVEC in a dose-dependent manner. Data represent means ± S.E. of three separate experiments.

8

9

View larger version (16K): [in this window] [in a new window]   Fig. 2.   NIF blocks PMN adhesion to endothelial cells equivalent to anti-CD18 mAb. PMA-treated PMNs were incubated with either anti-CD11a mAb (TS1/22, 10 µg/ml), anti-CD11b mAb (OKM10, 10 µg/ml), or anti-CD18 mAb (IB4, 10 µg/ml). Recombinant NIF (1 µM) was added to the medium containing PMN treated with either anti-CD11a mAb or anti-CD11b mAb to determine the additive effects of NIF with these mAbs. Data represent means ± S.E. of three separate experiments. *, difference from the PMA-treated control (P < .05).

View this table: [in this window] [in a new window]   TABLE 1 Concentration-dependent effects of NIF on PMN adhesion to endothelial cells in combination with anti-CD11a or CD11b mAb PMN adhesion assay was carried out as described in Materials and Methods. PMN were first incubated with either anti-CD11a mAb (TS1/22) or anti-CD11b/CD18 (OKM10) mAbs at a saturating concentration (20 µg/ml). NIF at varying doses from 107 to 1011 M was added to mAb-treated PMN prior to the PMN adhesion assay. PMN were layered onto HUVEC monolayers and immediately activated with phorbol dibutyrate (PDB) (100 ng/ml). Values represent number of adherent PMN ± S.E. from four separate experiments.

NIF Prevents Aggregation of JY Lymphoblastoid Cells Expressing CD11a To address whether NIF binding to CD11a accounts for its potent antiadhesive effect, we used JY lymphoblastoid cells, which aggregate in a CD11a/CD18-dependent manner following PMA stimulation (Rothlein et al., 1985). We first showed by FACS analysis the cell surface expression of CD11a on JY cells. Unstimulated JY cells expressed CD11a (Fig. 3), CD11c (data not shown), CD18 (data not shown), ICAM-1 (Fig. 3) as well as HLA (data not shown); however, CD11b was not detectable (Fig. 3). Upon PMA stimulation (200 ng/ml, 18 h), the expression of CD11a, CD11b, ICAM-1, and HLA remained the same. In two of three experiments, we observed a slight decrease in CD11c as well as CD18 antigens (data not shown). As shown in Table 2, mAbs directed against CD11a, CD18, and ICAM-1 inhibited JY cell aggregation, whereas mAbs directed against CD11b and HLA had no effect; thus, JY cell aggregation was dependent on binding of CD11a to ICAM-1. NIF was shown to inhibit JY cell aggregation following PMA stimulation (Table 2).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   FACS analysis of JY lymphoblastoid cells. JY lymphoblastoid cells were treated with PMA (200 ng/ml, 18 h) and FACS analysis was performed to determine the surface expression of CD11a, CD11b, or ICAM-1. The mAbs used were anti-CD11a (TS2/4), anti-CD11b (CBRM1/20), anti-ICAM-1 (RR1/1), and anti-HLA (W6/32) as control. Control unstimulated JY lymphoblastoid cells were used for comparison. The JY cells did not express CD11b and PMA stimulation also did not alter CD11b expression. Data are representative of three separate experiments.

NIF Release by Transduced Endothelial Cells Prevents PMN Adhesion. We transduced the NIF cDNA into endothelial cells and determined NIF mRNA expression by Northern analysis (Fig. 4a). The synthesis and release of NIF protein by the transduced HMEC were demonstrated by metabolically labeling the cells with [³⁵S]methionine. As shown in Fig. 4b, a 41-kDa protein band following immunoprecipitation was evident on the autoradiographs from SDS-polyacrylamide gel electrophoresis. The amount of NIF protein released from the transduced endothelial cells was 110 ± 5 ng NIF/10⁶ endothelial cells/24 h (n = 3). NIF was undetectable in the vector-transduced or control-nontransduced HMEC. Conditioned medium from NIF-transduced cells prevented PMN adhesion by ~85% (Fig. 5). The inhibitory activity of the conditioned medium was fully neutralized by the anti-NIF antibody, indicating the authenticity of NIF as the sole inhibitor (Fig. 5). In other experiments, PMNs adhered minimally (~5%) to both vector-transduced or NIF-transduced endothelial cells under the basal condition (Fig. 6). However, in coculture experiments, challenge with PMA (10 and 20 nM) increased PMN adhesion to vector-transduced endothelial cells, whereas PMA did not significantly increase PMN adhesion to NIF-transduced endothelial cells (Fig. 6).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   a, Northern blot analysis of pLNCX- and pLNC-NIF-transduced HMEC. NIF mRNA (1.4 kilobases) was visualized by autoradiography after hybridization with 32P labeled NIF cDNA probe. NIF mRNA expression was detected in pLNC-NIF-transduced HMEC but not in the control pLNCX- or nontransduced cells. 28S rRNA expression was performed to demonstrate equal loading of RNA samples in each lane. The results are representative of three separate experiments. b, Protein analysis of conditioned medium obtained from pLNCX- or pLNC-NIF-transduced HMEC. Cells were incubated with [35S]methionine and the secreted NIF protein was immunoprecipitated from conditioned media with an anti-NIF polyclonal antibody. Immunoprecipitated protein was size-fractionated through SDS-denaturing gel by electrophoresis and bands were visualized by autoradiography. NIF band (41 kDa) was present in media from pLNC-NIF HMEC (lane 2) and absent in media from control pLNCX cells (lane 1). Results are representative of three separate experiments.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Effects of conditioned medium (CM) from NIF-transduced HMEC on PMN adhesion to endothelial cells. CM obtained from either pLNCX- or pLNC-NIF-transduced HMEC was added to monolayers of control endothelial cells. PMNs were layered onto endothelial monolayers and activated with PMA treatment (20 nM, 10 min) and the PMN adhesion was determined as described in Materials and Methods. In some experiments, conditioned medium from NIF-transduced HMEC was treated for 30 min with NIF antibody before the PMN adhesion assay. Data represent means ± S.E. and are representative of four separate experiments with eight replicates each. *, difference from PMA-treated control groups (P < .05).

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Effects of PMA on PMN adhesion to NIF-transduced HMEC. PMNs were layered onto to PLNCX- or pLNC-NIF-transduced HMEC and then activated with PMA (10 or 20 nM, 10 min). Adhesion assay was carried out as described in Materials and Methods. Data represent means ± S.E. and are representative of three separate experiments with six replicates each. *, difference from the PMA-treated vector control (P < .05).

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Effect of TNF- on PMN adhesion to NIF-transduced HMEC. pLNCX- and pLNC-NIF-transduced HMEC were treated for 3 h with TNF- (1000 U/ml). Anti-NIF polyclonal or anti-ICAM-1 mAb (RR 1/1) antibodies were added 30 min before the PMN adhesion assay. Data represent means ± S.E. and are representative of three separate experiments with eight replicates each. *, difference from the untreated control (basal) group (P < .05).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

NIF is a 41-kDa glycoprotein released by the canine hookworm (A. caninum) as a part of its defense response against phagocytic cells (Moyle et al., 1994). NIF was shown to bind to the Mac-1 subunit (CD11b) of the CD11b/CD18 integrin (Zhang and Plow, 1996, 1997). This is suggested to be a primary factor contributing to the action of NIF in preventing PMN adhesion to endothelial cells. The specific NIF binding domain may be the I domain because the deletion of the D248 PLGY region of the I domain abolished the interaction with NIF but did not affect the binding to C3bi (Zhang and Plow, 1996). This finding suggests that the binding sites for NIF and C3bi to CD11b are overlapping but not identical. Analysis of these sites with a recombinant glutathione S-transferase fusion protein containing the I domain also showed that NIF binds to this domain in a manner inhibited by the anti-I domain antibody (Muchowski et al., 1994; Rieu et al., 1994). In the present study, we showed that NIF blocked the interaction of PMNs with fibrinogen (a CD11b/CD18 ligand), which occurs as the result of NIF binding to CD11b I domain. Although previous studies have implicated the importance of NIF interactions with the CD11b I domain, NIF also may bind to other sites (Rieu et al., 1994; Zhang and Plow, 1996). Moreover, because the I domain exists in CD11a (Rieu et al., 1994), it is plausible that NIF also can impair the function of CD11a integrin. Binding to both CD11a and CD11b may explain NIF's potent inhibition of PMN adhesion to endothelial cells, equivalent to that of anti-CD18 antibodies but exceeding that of anti-CD11b antibodies (Graf et al., 1996).

We showed that NIF prevented the aggregation of JY lymphocytes that solely expressed CD11a. We also demonstrated that NIF in the presence of anti-CD11b mAb had an additive inhibitory effect in preventing PMN adhesion. In the presence of anti-CD11b mAb, it required 10⁸ M NIF to inhibit PMA-induced PMN binding to endothelial cells, whereas in the presence of anti-CD11a mAb, it required 10⁹ M NIF to inhibit PMN adhesion. Thus, the binding of NIF to CD11a may be one order of magnitude less efficient than binding to CD11b. Because NIF can potentially inactivate both integrins, these findings help to explain the potent action of NIF in preventing adhesion of activated PMNs to vascular endothelial cells (Muchowski et al., 1994; Rieu et al., 1994) and intravascular PMN sequestration and airspace PMN migration in vivo (Barnard et al., 1995; Zhou et al., 1998).

The above-mentioned studies indicate that NIF release from A. caninum and its binding to both CD11a and CD11b ₂ integrins on phagocytes provide a formidable mechanism for evading the immunological response of phagocytic cells. Because a similar strategy can be used by cells releasing NIF to evade PMNs, we stably transfected endothelial cells with NIF cDNA by the retroviral-mediated transduction method, and determined whether the released NIF modified PMN adhesion to endothelial cells. NIF cDNA integration and expression in endothelial cells were evident by the high level of NIF mRNA expression, which persisted up to 20 passages. The NIF-transduced endothelial cells released NIF protein into the cell culture medium. Moreover, the conditioned medium from the NIF-transduced cells inhibited the adhesion of PMNs to endothelial cells activated by 20 nM PMA (used to produce the maximal PMN adhesion response). The results also showed that PMN adhesion to the NIF-releasing endothelial cells activated by TNF- or PMA was inhibited (>90%). The anti-NIF antibody prevented these effects, indicating that the released NIF was responsible for the protective effect.

ICAM-1 binding to members of CD18 integrin family mediates the stable PMN adhesion to endothelial cells. We observed that there was no increase in E-selectin expression in NIF-transduced endothelial cells consistent with the absence of E-selectin expression in HMEC (Chen et al., 1997) used to transduce NIF cDNA. Thus, ICAM-1 expression (rather than E-selectin) was responsible for PMN adhesion to the TNF--activated endothelial cells releasing NIF. This finding is concordant with the observation that the inhibition of PMN adhesion to endothelial cells with anti-ICAM-1 mAb (RR1/1) was similar to the response with anti-CD18 mAb. The finding that mAb RR1/1 blocks domain I of ICAM-1 (Berendt et al., 1992), the site recognized by CD11a (Staunton et al., 1990), coupled with the present observation that RR1/1 prevented PMN adhesion, further supports the contention that the inhibitory action of NIF involves binding of NIF to CD11a I domain.

In summary, the present data suggests that NIF inhibits the function of both CD11a and CD11b integrins on PMNs and that may explain the potent antiadhesive function of NIF. Moreover, the strategy of endothelial cell gene transfer to release NIF suggests the feasibility of using this approach to prevent CD11a- and CD11b-dependent PMN adhesion and migration specifically at sites of expression of endothelial adhesion molecules.