Enhanced adhesive capacities of the naturally occurring Ile249-Met280 variant of the chemokine receptor CX3CR1.

It was recently shown that individuals carrying the naturally occurring mutant CX3CR1-Ile(249)-Met(280) (hereafter called CX3CR1-IM) have a lower risk of cardiovascular disease than individuals homozygous for the wild-type CX3CR1-Val(249)-Thr(280) (CX3CR1-VT). We report here that peripheral blood mononuclear cells (PBMC) from individuals with the CX3CR1-IM haplotype adhered more potently to membrane-bound CX3CL1 than did PBMC from homozygous CX3CR1-VT donors. Similar excess adhesion was observed with CX3CR1-IM-transfected human embryonic kidney (HEK) cell lines tested with two different methods: the parallel plate laminar flow chamber and the dual pipette aspiration technique. Suppression of the extra adhesion in the presence of pertussis toxin indicates that G-protein mediated the underlying transduction pathway, in contrast to the G-protein-independent adhesion previously described for CX3CR1-VT. Surprisingly, HEK and PBMC that expressed CX3CR1-IM and -VT were indistinguishable when tested with the soluble form of CX3CL1 for chemotaxis, calcium release, and binding capacity. In conclusion, only the membrane-anchored form of CX3CL1 functionally discriminated between these two allelic isoforms of CX3CR1. These results suggest that each form of this ligand may lead to a different signaling pathway. The extra adhesion of CX3CR1-IM may be related to immune defenses and to atherogenesis, both of which depend substantially on adhesive intercellular events.


INTRODUCTION
Adhesion, a critical stage in cell trafficking and migration (1,2), requires the presence of numerous adhesion molecules, such as integrins, that need divalent ions to function. Recent studies show that two chemokines, namely, CX3CL1 and CXCL16, are not only chemoattractant as soluble molecules, but also function as adhesion molecules since they are membrane-anchored, regardless of the presence of divalent ions (3)(4)(5)(6)(7). The best known of these is CX3CL1, also called fractalkine. It is expressed on the surface of many types of cells, in particular IL-1-and TNF-activated endothelial (4) and dendritic cells (8), as a membrane molecule containing the classic chemokine domain, a mucin-like stalk, and a transmembrane domain tethering it to the cell membrane. CX3CL1 may be cleaved by TACE and released from cells (9,10). In this soluble form, it behaves like a chemoattractant molecule, just as other chemokines do. The transmembrane feature of the native CX3CL1 protein, combining it with its receptor, CX3CR1, produces a strongly adhesive pair (4,5,11) that mediates the rapid capture and firm adhesion of leukocytes. Because this activity persists in the absence of divalent cations, it is thought to be independent of integrins (5,11,12). This adhesive feature is also independent of the Gi pathway, since it is still present after pertussis toxin (PTX 1 ) treatment (4,5,11).
The CX3CR1 molecule is expressed on leukocytes, especially monocytes (4) and cytotoxic cells (13,14), on dendritic cells (15), and on neurons and microglial cells (16,17). Recently, we identified two common polymorphisms in strong linkage disequilibrium in the CX3CR1 gene: V249I and T280M (18). We also found that these mutations are associated with more rapid progression to AIDS (18,19), although two studies have failed to confirm this association (20,21). A recent work indicates that these mutations are

CX3CR1 variant constructs
Open reading frames corresponding to CX3CR1-V249T280 (CX3CR1-VT) and CX3CR1-I249M280 (CX3CR1-IM) were amplified from genomic DNA prepared from PBMC from two healthy donors with the corresponding genotypes. To do that, we used Hind III tailed forward primer (LT5-CX3CR1: GCGCATATAAGCTTGCCACCATGGATCAGTTCCCTGAATCAG) and Xho I tailed reverse primer (LT3-CX3CR1: GCGGATATGTCGACCTCGAGTCACGAGTCAGAGAAGGAGCAA). The PCR cycling conditions were 95°C for 5 min followed by 30 cycles at 95°C for 1 min, 52°C for 1 min, and 72°C for 1 min. The PCR product was then digested with Hind III and Xho I (Promega, Charbonnières, France), subcloned into the mammalian expression vector pCDNA3.1(+) (Invitrogen, Cergy-Pontoise, France), and then sequenced on both strands with the BigDye Terminator kit (Applied Biosystems, Warrington, UK).

Cell culture and transfection
HEK cells were cultured in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal calf serum, 1 mM sodium pyruvate, 2 mM glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin (Invitrogen). To generate stably transfected clones, we used Transfast (Promega) in accordance with the manufacturer's instructions. We transfected 10 6 HEK-293 cells in 3-cm dishes with 2 µg of each pcDNA3.1 construct (Invitrogen) with CX3CR1 inserts encoding the CX3CR1 variant gene. Clones were derived by selection in 1 mg/ml G418 (Life Technologies, Rockville, MD, USA). The clones thus obtained were assayed for CX3CL1 binding, and those expressing high levels of CX3CR1 were selected for further study. The clones were maintained in DMEM media containing G418 and were checked with CX3CL1 binding for CX3CR1 expression before each experiment. To obtain transiently transfected HEK cells, we used JetPei TM (cationic polymer transfection reagent, Qbiogene, Illkirch, France) in accordance with the manufacturer's instructions. We used 3 µg of each pcDNA3.1 construct to transfect 4 10 5 HEK-293 cells. After 2 days, transfected cells were resuspended by incubation with PBS for 30 min at 37°C and were checked with CX3CL1 binding for CX3CR1 expression before use. For CX3CL1 transfection in HEK, we used pBlast plasmid (InvivoGen, Toulouse, France) with a CX3CL1 insert or with no insert. For stable expression, clones were derived by selection in 5 µg/ml Blasticidin (Euromedex, Mundolsheim, France) for at least 1 month. The clones thus obtained were assayed for CX3CL1 staining by flow cytometry, and those expressing high levels of CX3CL1 were selected for further study. The clones were maintained in DMEM media containing Blasticidin and were checked before each experiment for CX3CL1 expression by flow cytometry. PBMC were isolated from heparinized venous blood from healthy volunteers by one-step centrifugation on a Ficoll separating solution (Biochrom KG, Berlin, Germany).

Flow cytometry
The cells (10 5 PBMC or 2.5.10 5 HEK) were tested for CX3CR1 expression by flow cytometry after staining by fluorescein isothiocyanate (FITC) conjugated anti-CX3CR1 monoclonal antibody (MBL, Nagoya, Japan). As a control, the cells were incubated without any antibody. We verified that this produced the same signal as an isotype antibody control. The different PBMC subsets were quantified sequentially, in 2 ways: first, by discriminating lymphocytes and monocytes according to their width and granulometry (SSC versus FSC diagram), and second, by staining with various antibodies as follows. The Fura-2/AM and 2 µM of pluronic acid in 1 ml of HBSS buffer supplemented with 10 mM HEPES, 0.5 mM MgCl 2 , and 1 mM CaCl 2 . After centrifugation, the pellet was resuspended in 2 ml of the same buffer and transferred to a quartz cuvette for reading. CX3CL1 was added to the cell volume at various concentrations. Fluorescence was monitored with a SAFAS spectrofluorometer (SAFAS S.A., Monaco) in cuvettes thermostatically controlled at 37°C and stirred continuously. The cell suspension was excited alternately at 340 and 380 nm and fluorescence measured at 510 nm. Ten-nanometer slit widths were used for both excitation and emission. Graphic representation of intracellular calcium concentrations were computed with the equation: [Ca ++ ] = 225xR/(R max -R)xSf 380 /Sb 380 , previously determined by Grynkiewicz (30), with R the ratio of the fluorescence measured at the 340 nm and 380 nm excitations; R max was evaluated by lysing the cells with 0.5% Triton X-100, and R min determined by adding the excess EGTA. Sb 380 and Sf 380 were the fluorescence levels at 380 nm excitation, both determined in the same conditions.

Chemotactic migration assay
Chemotaxis was assayed in a 96-well chemotaxis chamber with a filter porosity of 10 µm (NeuroProbe,

Parallel-plate laminar flow chamber adhesion assay
Adhesion experiments used the parallel-plate flow technique and the chamber previously described (31).
The coverslips we used were either cultured with adherent HEK cells (HEK-pBlast or HEK-FKN clones) or coated with CX3CL1 ( Figure 1C), as follows: the coverslip was coated with 10 µl of anti-  isotonic sucrose solution (300-330 mOs) and preincubated in BSA. Cells were manipulated with two micropipettes, each held in its own micromanipulator and connected to a combined hydraulic/pneumatic system that provided the necessary control of the aspiration force applied to the cells.
The protocol we used is very similar to that of Chien's group (32). Two cells, collected by gentle aspiration onto the tip of each pipette (cell number 1 in pipette A, cell 2 in pipette B), were brought into contact through the use of the micromanipulators and allowed to remain in contact for different periods of time ( Figure 3C; 2 to 30 min). To separate the cells, aspiration in pipette B was maintained at a level sufficiently high to hold cell number 2 tightly, while the aspiration in pipette A was increased in steps measured with a pressure sensor (Validyne: model DP103-38; ranging from 0 to 50 000 Pa). After each step, the pipettes were moved apart in an effort to detach the adherent cells from one another. A pair pulled intact from pipette A was moved back to the pipette orifice, the aspiration in the pipette was increased, and another attempt was made to detach the cells from each other. The cycle was repeated until the level of aspiration in pipette A was sufficient to pull one cell apart from the other. The aspiration employed in each cycle was monitored continuously. In most cases, cell deformation and contact area variation during the separation process were very limited (less than 20% for the contact area), and the separation took place suddenly, in less than a tenth of a second. The cells appeared to behave more like rigid structures than like two adhering deformable capsules. The usual approach (33,34) of measuring contact angles at the end of the pipette and at the edge of the contact thus did not seem useful. The separation force (F) for rigid structures can be deduced from the data (Chu et al., submitted). The values recorded for each of the last two cycles in the series (P n-1 and P n ) were used to calculate F for the pair tested, with the equation: with d the internal diameter of pipette A. This relation assumes that the pressure inside the cell is the same as that in the chamber, valid in our case since the tension of the cell is essentially zero. The results were expressed as mean ± SEM for 13 or more measurements.

Western blot analysis of p44/42 MAP kinase phosphorylation
HEK cells were starved for 18 h in DMEM without SVF and suspended at 10

Enhanced cell adhesive functioning for PBMC that express CX3CR1-IM
We used the parallel-plate adhesion method to compare the adhesion of PBMC from donors with different CX3CR1 genotypes to a monolayer of CX3CL1-expressing HEK. At low perfusion rates (1.5 dynes.cm -2 ), we repeatedly found that cells from donors heterozygous for the 249 and 280 positions (VI-TM) of CX3CR1 ( Figure 1A, solid squares) were captured in significantly larger numbers than cells from those homozygous for the wild-type genotype (VV-TT) ( Figure 1A, solid triangles). Cells from individuals with both genotypes began to dissociate at a shear stress higher than 15-20 dynes.cm -2 ( Figure 1A, solid symbols). Non-specific adhesion, i.e., capture on CX3CL1-negative HEK cells, was very low and very similar for cells of both genotypes ( Figure 1A, open symbols regularly lower in PBMC with the CX3CR1-VI-TM than in cells expressing the non mutated receptor CX3CR1-VV-TT (Table 1). Similarly, the [ 125 I]CX3CL1 binding assay indicated that, as we noted previously (26), cell suspensions with the CX3CR1-VI-TM genotype had only 67 ± 11% (n=7) as many binding sites (Bmax) as suspensions with CX3CR1-VV-TT cells. Yet, although there were fewer CX3CR1+ cells, there were more cells adhering to membrane CX3CL1. The effect of CX3CR1 mutations was therefore underestimated. Accordingly, we expressed the specific adhesion by calculating the ratio of CX3CR1+ cells specifically adhering to membrane CX3CL1 (see Experimental Procedures). Figure 1B reports PBMC specific adhesion for each CX3CR1 allele. No significant differences in adhesion were observed between the PBMC expressing CX3CR1 that differed only at the 280 position ( Figure 1B, compare VI-TT with VI-TM and II-TM with II-MM). In contrast, the I249 substitution appeared crucial.
Adhesion was already significantly greater with the PBMC from heterozygous VI-TT individuals than from the VV-TT homozygote ( Figure 1B). Moreover, PBMC from carriers homozygous for the 249 position (i.e., II-TM and II-MM, Figure 1B, right) adhered significantly more than PBMC from heterozygous (i.e., VI-TT and VI-TM, Figure 1B, center) donors. The amplitude of extra adhesion therefore appears to be directly correlated with the number of I249 alleles, according to a simple gene dosage effect.
To verify the absence of nonspecific cell-cell interaction, we performed experiments with immobilized CX3CL1. We found that PBMC from CX3CR1-VI-TM donors adhered to coverslip coated with recombinant CX3CL1 at a rate more than twice that of cells from CX3CR1-VV-TT individuals, and this was the case regardless of the presence of divalent cations ( Figure 1C, solid bars). This confirms that the excess adhesion in the presence of the I249-M280 CX3CR1 mutations is due solely to its interaction with the CX3CL1 ligand. More specifically, it demonstrates that the phenomenon is independent of the divalent ions as is the basal adhesion caused by the CX3CR1-VV-TT molecule (5,11,12) (see Figure 1C, open bars).
Finally, we assayed the chemotactic migration of PBMC from individuals with the CX3CR1-VV-TT and CX3CR1-VI-TM genotypes ( Figure 1D). Surprisingly, no difference was detected between these variants in response to soluble CX3CL1, in contrast to the notable differences in their responses to membraneanchored CX3CL1 ( Figure 1A-1C). It thus appears that these CX3CR1 variants can discriminate between the two forms of the ligand.
The enhanced adhesiveness of the mutated CX3CR1 was observed with whole PBMC. We then considered whether this effect was specific for a single leukocyte population. Phenotyping the adherent PBMC revealed that all the CX3CR1+ PBMC subpopulations (i.e., monocytes, NK, CD4+ and CD8+ lymphocytes) contributed to this effect in similar proportions ( Figure 1E). This indicates that the excess adhesion we observed is caused by the intrinsic potency of the mutated CX3CR1. We checked this finding further with purified monocytes from various individuals. We found that CX3CR1-VI-TM monocytes adhered at a rate three times higher than the CX3CR1-VV-TT monocytes (data not shown). In contrast, both monocyte populations were indistinguishable in their chemotactic response to CX3CL1. These findings agree with those obtained with whole PBMC ( Figure 1D). We therefore concluded that the excess adhesion we observed with PBMC bearing the I249-CX3CR1 allele ( Figure 1B) was not specific to one particular leucocyte subpopulation, but occurs once the mutation is present. Moreover, this change can only be observed when CX3CR1 binds the membrane form of CX3CL1 ( Figure 1A-C) and not in response to its soluble form ( Figure 1D).

Enhanced cell adhesive functioning for HEK clones that express CX3CR1-IM
To characterize the adhesive properties of the CX3CR1 variants further, we generated HEK clones Two different clones of each type were tested independently with flow adhesion. As with PBMC, the clones expressing CX3CR1-IM ( Figure 3A, solid circles) adhered more than those expressing CX3CR1-VT ( Figure 4A, solid triangles): the number of CX3CR1+ HEK cells adhering specifically to membrane CX3CL1 at a low perfusion rate (1.5 dynes/cm 2 ) was almost twice as high for the mutated CX3CR1-IM form than for the standard CX3CR1-VT. Both types of clones began to dissociate at shear stress higher than 15-20 dynes.cm -2 , as previously reported (5,11). Nonspecific adhesion of the clones was assessed with a CX3CL1-negative adherent cell layer ( Figure 3A, open triangles and circles). A control HEK clone expressing the chemokine receptor CCR5 adhered equally poorly to the CX3CL1-expressing cells ( Figure   3A, diamonds). Moreover, the adhesion of the HEK clones that expressed CX3CR1 was almost wholly suppressed when the cells were preincubated with soluble CX3CL1 for 45 min at 37°C (data not shown).
Again this indicates that this adhesion is specific to the CX3CR1/CX3CL1 pairing.
We verified that no particular property selected by clone generation caused these adhesion features in the stable HEK clones: they were also found with HEK transiently transfected with either CX3CR1-VT or CX3CR1-IM plasmids ( Figure 3B, compare VT and IM). In addition, we assayed the HEK transfected with the CX3CR1 plasmid carrying only the V249I mutation, i.e., CX3CR1-IT. Surprisingly these cells adhered at rate similar to that of the double mutant CX3CR1-IM ( Figure 3B). Moreover the cells expressing the no naturally occurring variant CX3CR1-VM adhered at a rate similar to that observed with CX3CR1-VT cells ( Figure 3B). These findings provide further support for the hypothesis that the excess adhesion is due only to the mutation at the 249 position.
To confirm and quantify this enhanced adhesion with the CX3CR1-IM variant, we used another cell-cell adhesion assay, the dual pipette aspiration technique, previously used to verify CTL-target adhesion (32).
Briefly, this method consists in determining the force required to dissociate a pair formed by two cells brought into contact by micropipettes. The dissociation force is measured at a given time after pair formation. We found that paired CX3CR1-VT / CX3CL1 HEK cells adhered after only 2 min of contact with a separation force of about 6 nN ( Figure 3C, solid  was independent of time, as expected from previous data (4,5,11), and lasted for 30 min without attenuation ( Figure 3C We tested the chemotactic responses of the transfected HEK cell clones to the CX3CL1 gradient. As with PBMC ( Figure 1D), we found no differences between the clones expressing the two CX3CR1 variants ( Figure 3D).

The excess adhesion due to the mutated CX3CR1 is PTX-dependent
Although most signals triggered by CX3CR1 ligation with the soluble CX3CL1 are G-protein dependent, the adhesive properties of the CX3CR1/CX3CL1 pair are independent of the Gi pathway, i.e., they are still  (4,5,11). We confirmed this finding here with flow chamber dynamic adhesion assays that used PBMC from CX3CR1-VV-TT donors ( Figure 4A, open bars) or HEK cells expressing CX3CR1-VT (data not shown). Surprisingly, the adhesion observed with cells expressing the CX3CR1-IM variant was reduced after PTX treatment to the level observed with the CX3CR1-VT ( Figure 4A, solid bars). The same result was observed with PBMC adhering to immobilized CX3CL1, either in the presence or absence of divalent ions (data not shown) or using the dual pipette assay with HEK cell pairs ( Figure 4B). This result suggests that the adhesive feature of the mutated CX3CR1 is composed of two additive events, one basal adhesion common to both variants and one specific to the CX3CR1-IM conformation. In contrast, the excess adhesion obtained with the CX3CR1-IM haplotype was insensitive to other pharmacological agents, including LY-294002 and PD-98059, which inhibit, respectively, phosphatidylinositol 3-kinase and p44/42 MAP kinase enzymes (data not shown).

Signaling pathways mediated by CX3CR1 variants
Possible differences between the CX3CR1 variants were tested by assaying two other cellular responses.
We first examined the calcium response of HEK cell clones that expressed each of the CX3CR1 variants ( Figure 5A): the dose-response curves were indistinguishable. We also tested the activation of the cellular MAP kinase pathway, which CX3CL1 triggers in neurons (37), intestinal epithelial cells (38), microglia cell lines (16), and monocyte cell lines (39). In both of our HEK cell line clones, the maximum p44/42 MAP kinase stimulation was reached within 2 min of CX3CL1 application ( Figure 5B and 5C). The extent of MAP kinase phosphorylation was slightly higher in the CX3CR1-IM than in the VT HEK clone ( Figure 5B and 5C), but the difference was not statistically significant.

DISCUSSION
In view of its effect on prognosis in AIDS (18,19) and in cardiovascular diseases (26)(27)(28), understanding the molecular modifications caused by the chemokine receptor CX3CR1-IM mutation is an important challenge. We found here that intercellular adhesion mediated by the CX3CR1/CX3CL1 pair was substantially greater with cells that express the CX3CR1-IM variant than with those expressing the CX3CR1-VT variant (Figures 1, 3 and 4). Moreover, the data from both PBMC ( Figure 1B) and HEK ( Figure 3B) indicate that the V249I mutation alone is responsible for the high level of adhesion we observed. Finally, adhesion mediated by CX3CR1-IM was independent of divalent ions and involved only CX3CL1 as counterligand (Figure 1), as was the adhesion mediated by CX3CR1-VT (5,11,12).
This surprisingly enhanced adhesiveness of the CX3CR1 variant was demonstrated with two different techniques that determined distinct indicators. The parallel-plate method furnishes the fraction of adhering cells under shear stress, while the dual pipette procedure directly assesses the force required to dissociate cell pairs under axial stress. Both techniques indicate that the CX3CL1-specific adhesion force generated by the CX3CR1-IM variant is significantly greater than that induced by the CX3CR1-VT genotype. Moreover, the dual pipette procedure indicates that this excess adhesion occurs slowly, after a few minutes, thereby suggesting that the adhesive potency of CX3CR1-IM results from the addition of two phenomena: first, immediate adhesion, as observed for CX3CR1-VT, followed by a time-dependent attachment that seems specific to CX3CR1-IM. This slow time course may point to a signalingdependent mechanism, a hypothesis supported by our experiments with PTX ( Figure 4). Thus the mutated CX3CR1 form may specifically trigger a signal that, added to the basal and instantaneous adhesion due to activation is somewhat higher in CX3CR1-IM cells ( Figure 5, B and C), the testing of specific inhibitors ruled out the involvement of the MAP kinases dependent-and the PI3K pathways in generating this extra adhesion.
The enhanced adhesiveness of the CX3CR1-IM variant was observed in both transfected HEK cells and peripheral blood cells. All the CX3CR1+ PBMC subpopulations adhered to membrane CX3CL1 (5) and showed enhanced adhesiveness when they had the CX3CR1-IM haplotype ( Figure 1B and 1E). The association of the CX3CR1-IM genotype with a reduced risk of cardiovascular disease was previously thought to be due to the receptor's reduced capacity to bind its ligand, and frozen PBMC from HIV patients with a mutated genotype showed less ligand affinity (18). A recent report proposes that the I249 mutation is associated with a promoter mutation that may result in differential CX3CR1 expression (40).
This might explain the significantly lower number of receptors per cell on PBMC from VI compared with VV donors (18,26 and this report). It cannot, however, account for the excess adhesion we observed here.
Our experiments indicate that the differences we observed between CX3CR1 variants are due to intrinsic molecular properties.  Our study implies that CX3CR1 behaves differently when addressing soluble or membrane ligand. A similar difference was recently observed for IFNγ production by NK cells (41). Our work also shows that the specific mechanism triggered by CX3CR1-IM binding to membrane CX3CL1 is dependent on the PTX sensitive G-protein family Gi (Figure 4). This signal-dependent adhesion might be due to more effective oligomerization of CX3CR1-IM at the adhesive interface, possibly related to a differential association with membrane lipid rafts. It has been suggested that the association of membrane protein to To our knowledge, our paper is the first to report a chemokine receptor mutation associated with increased functions. It appears to originate in a single mutation, replacement of a valine residue by an isoleucine at the 249 position. This increase in functioning seems to be mediated by gene dosage ( Figure 1B Recent reports show that inactivating the CX3CR1 gene leads to a decrease in the risk of atherogenesis (49,50). It was therefore paradoxical to find that mutations that appear to protect against cardiovascular diseases (26)(27)(28) actually enhance the molecule's adhesive properties. The monocytes recruited in the intima layer to form atherosclerotic plaque should first adhere and cross the endothelium barrier (51,52).
The reduction of this transmigration step in the presence of CX3CL1 (53) indicates that the adhesion function of CX3CL1 may counteract the migration driven by inflammatory chemoattractants. It is thus conceivable that excess adhesion might further diminish monocyte extravasation and hence weaken atherogenesis.
The additional adhesion we observed may also be involved in NK-or CTL-cell target interactions in ganglia, especially in HIV patients. The lymph nodes of such patients over express CX3CL1 (54), while disease severity is correlated with CX3CR1 expression (14). Hence, the excess adhesion we describe here may profoundly affect both innate and acquired immunity.    B. HEK clones were also assayed for p44/42 MAP kinase phosphorylation during the indicated times in response to 50 nM CX3CL1.  Table   Table 1   The PBMC from individuals with CX3CR1-VV-TT (n=3, left), and VI-TM (n=3, right) were assayed for adhesion as in Figure 1B, except that the layer cell was a smooth muscle aortic cell line pre-treated for 4 h at 37°C by 20 ng/ml TNFα and 500 u/ml INFγ. We report the number of adherent cells after the first flow step (10 min at 1.5 dynes.cm -2 ). Specific adhesion was obtained by subtracting the number adhering to resting smooth muscle aortic cells. The difference in adherence is significant (p<0.05).    The PBMC from individuals with CX3CR1-VV-TT (n=4, left), and VI-TM (n=4, right) were assayed for adhesion as in Figure 1C, with coverslips coated with immobilized FKN-His 6 . The protocol was modified according to McDermott et al. [28]. The cell suspension was injected through the chamber at a wall shear stress of 0.25 dynes.cm -2 . After a 5-min wash at 0.25 dynes.cm -2 , the shear stress was set at 0.5 dynes.cm -2 for 1min, 1 dynes for 1 min, 2 dynes.cm -2 for 1 min, 5 dynes.cm -2 for 1 min, and finally 10 dynes.cm -2 for 5 min. Moreover, the experiment was done using buffer containing divalent cations, as McDermott et al. The adherent cells were then counted. Specific adhesion was obtained by subtracting the number adhering to anti-His 6 alone.
The difference in adherence is significant (p<0.001).