Heart and Circulatory Physiology

Exogenous stromal cell-derived factor-1 induces modest leukocyte recruitment in vivo

Steven M. Kerfoot, Graciela Andonegui, Claudine S. Bonder, Lixin Liu


Stromal cell-derived factor-1 (SDF-1; CXCL12), a CXC chemokine, has been found to be involved in inflammation models in vivo and in cell adhesion, migration, and chemotaxis in vitro. This study aimed to determine whether exogenous SDF-1 induces leukocyte recruitment in mice. After systemic administration of SDF-1α, expression of the adhesion molecules P-selectin and VCAM-1 in mice was measured using a quantitative dual-radiolabeled Ab assay and leukocyte recruitment in various tissues was evaluated using intravital microscopy. The effect of local SDF-1α on leukocyte recruitment was also determined in cremaster muscle and compared with the effect of the cytokine TNFα and the CXC chemokine keratinocyte-derived chemokine (KC; CXCL1). Systemic administration of SDF-1α (10 μg, 4–5 h) induced upregulation of P-selectin, but not VCAM-1, in most tissues in mice. It caused modest leukocyte recruitment responses in microvasculature of cremaster muscle, intestine, and brain, i.e., an increase in flux of rolling leukocytes in cremaster muscle and intestines, leukocyte adhesion in all three tissues, and emigration in cremaster muscle. Local treatment with SDF-1α (1 μg, 4–5 h) reduced leukocyte rolling velocity and increased leukocyte adhesion and emigration in cremasteric venules, but the responses were much less profound than those elicited by KC or TNFα. SDF-1α-induced recruitment was dependent on endothelial P-selectin, but not P-selectin on platelets. We conclude that the exogenous SDF-1α enhances leukocyte-endothelial cell interactions and induces modest and endothelial P-selectin-dependent leukocyte recruitment.

  • inflammation
  • emigration
  • adhesion

stromal cell-derived factor-1 (SDF-1; CXCL12), a CXC chemokine, acts through its sole receptor CXCR4 to mediate its diverse functions. SDF-1 has been shown to have six splice variants, among which SDF-1α is the predominant and the most studied isoform (21, 55). It was first identified as pre-B-cell growth-stimulating factor (34), but its functions have been demonstrated to be numerous and far beyond B cell lymphopoiesis, including T cell activation and migration, hematopoiesis, neuronal development, vasculogenesis, cardiogenesis, immune cell homing and trafficking, human immunodeficiency virus pathogenesis, and tumorigenesis and metastasis (6, 21, 33). It is interesting to note that, apart from its role in homeostasis in vivo, there is a large body of evidence suggesting that SDF-1 may be involved in inflammation (13, 14, 37, 46).

SDF-1 has been found to be constitutively and highly expressed in most normal tissues, but not leukocytes (43). However, the SDF-1 receptor CXCR4 is highly expressed in many subsets of leukocytes, including B cells, naïve T cells, monocytes, mast cells, eosinophils, and neutrophils (11, 26, 30, 35, 36). Notably, endothelial cells express SDF-1 and CXCR4 (19, 38, 42, 51). The SDF-1-CXCR4 interaction triggers multiple intracellular signaling pathways downstream of CXCR4. In various cell types, the subsequent events include G protein-dependent and -independent signaling events, prolonged activation of PKB and ERK-2, CXCR4 dimerization, and receptor tyrosine phosphorylation, internalization, and degradation (6, 12, 41, 50). Some of these signaling events are similar to those activated by prototypic proinflammatory cytokines.

Research performed in animal models has suggested that SDF-1 has a direct proinflammatory role in vivo. Subcutaneous injections of SDF-1α induced local vessel formation that was accompanied by leukocyte infiltration (42). In an allergic airway model, the involvement of SDF-1 and CXCR4 was found to be critical in the development of inflammation (13). In an arthritis model, injection of exogenous SDF-1 elicited an inflammatory response that was blocked by AMD-3100, a CXCR4 antagonist (31). It has also been suggested that the proinflammatory role of SDF-1 may be indirect. In human astroglioma cells, Oh et al. (37) showed that SDF-1α induced expression of a number of chemokines, including monocyte chemoattractant protein-1, IL-8, and interferon-inducible protein-10, all of which are known to be proinflammatory. In models of various skin inflammatory diseases, CXCR4-expressing lymphocytes were found infiltrating the tissue and were in close contact with a large number of mononuclear cells and fibroblasts expressing SDF-1 (38). Clearly, although SDF-1 has a role in many physiological and pathological situations, it probably also plays a role in inducing or maintaining inflammation in vivo. In the present study, we used intravital microscopy to investigate the role of exogenous SDF-1α in leukocyte-endothelial cell interactions. We evaluated changes in recruitment parameters (leukocyte rolling, adhesion, and emigration) in microvasculatures after systemic treatment with SDF-1α and also quantified the endothelial activation state by determining P-selectin or VCAM-1 expression in vivo. We also compared the effects of local SDF-1α treatment with the effects of the cytokine TNFα and an inflammatory CXC chemokine, keratinocyte-derived chemokine (KC; CXCL1), in mice.



Male C57BL/6 mice (Charles River Laboratories, Wilmington, MA) weighed 20–30 g and were used at 6–12 wk of age. All animal protocols met the standards of the Canadian Association of Animal Care and were approved by the Animal Care Committee of the University of Calgary.

For the following experimental procedures, except intravital microscopy of the small intestine, mice were anesthetized with an intraperitoneal injection of a mixture of 10 mg/kg xylazine (Bayer Animal Health, Toronto, ON, Canada) and 200 mg/kg ketamine hydrochloride (Rogar/STB, Montreal, PQ, Canada). For intravital microscopy on the small intestine, mice were anesthetized with isoflurane, an inhalation anesthetic, as previously described (4). For cremaster muscle and intestine protocols, the left jugular vein and, for the brain protocol, the tail vein were cannulated for administration of additional anesthetic, fluorescent dyes, or drugs when necessary.

Quantification of P-Selectin and VCAM-1 Expression

A modified dual-radiolabeled Ab assay was used to determine expression of P-selectin and VCAM-1 on the endothelium as a measure of endothelial activation (1, 17, 18). All MAbs were purchased from Pharmingen (San Diego, CA). The IodoGen method was used to label Abs RB40.34 (against P-selectin) and 429 (MVCAM.A, against VCAM-1) with 125I and Abs A110-1 (a rat IgG1, λ-isotype standard) and R35-95 (a rat IgG2a, κ-isotype control Ig) with 131I, as previously described (17, 18). A110-1 and R35-95 were used as control for nonspecific binding in the murine system.

For study of P-selectin expression, mice were injected intravenously with a mixture of 10 μg of 125I-labeled Ab RB40.34 and a variable dose of 131I-labeled nonbinding Ab A110-1. For measurement of VCAM-1, mice were injected with 10 μg of 125I-labeled anti-VCAM-1 [429 (MVCAM.A)], 25 μg of unlabeled anti-VCAM-1 [429 (MVCAM.A)], and a variable dose of 131I-labeled nonbinding Ab (R35-95) calculated to achieve a total injected 131I activity of 400,000–600,000 cpm (200 μl total volume). This Ab combination was chosen after pilot experiments, conducted over a range of doses of unlabeled 429 (MVCAM.A), showed that this protocol ensured receptor saturation under stimulated conditions. In both cases, the Abs were allowed to circulate for 5 min; then the animals were heparinized. A blood sample was obtained from a carotid artery catheter; then the mice were exsanguinated through the carotid artery catheter and simultaneously infused intravenously with bicarbonate-buffered saline. Lungs, muscle, heart, brain, spinal cord, stomach, small bowel, large bowel, skin, pancreas, kidneys, bone marrow, and liver were harvested and weighed. 131I and 125I activities were measured in plasma and tissue samples.

P-selectin or VCAM-1 expression was calculated per gram of tissue by subtraction of the accumulated activity of the nonspecific (131I-labeled) Ab from the accumulated activity of the binding (125I-labeled) Ab. The expression data for P-selectin and VCAM-1 are represented as the percentage of the injected dose of Abs per gram of tissue. It has been previously demonstrated that this approach provides reliable quantitative values of adhesion molecule expression and that radiolabeled binding Abs can be displaced specifically with sufficient amounts of unlabeled Abs (10, 39). The technique is sufficiently sensitive that very small, basal levels of P-selectin can be detected in wild-type mice relative to P-selectin-deficient mice, where values are zero (10).

Intravital Microscopy

Cremaster muscle.

An incision was made in the scrotal skin to expose the left cremaster muscle, which was then carefully removed from the associated fascia. A cautery was used to make a lengthwise incision on the ventral surface of the cremaster muscle. The testicle and the epididymis were separated from the underlying muscle and reintroduced into the abdominal cavity. The muscle was then spread out over an optically clear viewing pedestal and secured along the edges with 4-0 suture. The exposed tissue was superfused with 37°C warmed bicarbonate-buffered saline (pH 7.4). An intravital microscope (Axioskop, Carl Zeiss Canada, Don Mills, ON, Canada) with a ×25 objective lens (Weltzlar L25/0.35, Leitz, Munich, Germany) and a ×10 eyepiece was used to examine the cremasteric microcirculation. A video camera (model 5100 HS, Panasonic, Osaka, Japan) was used to project the images onto a monitor, and the images were recorded for playback analysis using a videocassette recorder.

Single unbranched cremasteric venules (25–40 μm diameter) were selected; to minimize variability, the same section of cremasteric venule was observed throughout the experiment. The number of rolling, adherent, and emigrated leukocytes was determined offline during video playback analysis. Rolling leukocytes were defined as cells moving at a velocity less than that of erythrocytes within a given vessel. The flux of rolling cells was measured as the number of rolling cells passing a given point in the venule per minute. A leukocyte was considered to be adherent if it remained stationary for ≥30 s, and total leukocyte adhesion was quantified as the number of adherent cells within a 100-μm length of venule. Leukocyte emigration was defined as the number of cells in the extravascular space within a 260 × 370 μm area (field of view). Only cells adjacent to and clearly outside the vessel under study were counted as emigrated.

Small intestine.

A segment of small intestine was exteriorized through an abdominal incision and draped over a viewing pedestal, and the exposed tissue was suffused with bicarbonate-buffered saline. Mice were allowed to stabilize for 15 min after surgery. The intestinal microcirculation was visualized by autofluorescence before experiments for identification of postcapillary venules. Leukocytes were visualized by intravenous injection of rhodamine 6G (0.3 mg/kg body wt; Sigma-Aldrich, St. Louis, MO). Rhodamine 6G-associated fluorescence is visualized by epi-illumination at 510–560 nm using a 590-nm-emission filter. A microscope (Optiphot-2, Nikon, Mississauga, ON, Canada) with a ×25 water-immersion lens (Leitz Wetzlar L25/0.35) and a ×10 eyepiece was used. A silicon-intensified fluorescence camera (model C-2400-08, Hamamatsu Photonics, Hamamatsu City, Japan) mounted on the microscope projected the images onto a monitor, and the images were recorded using a video recorder (model NV8950, Panasonic). Five randomly selected 20- to 30-μm-diameter postcapillary venules were observed throughout the 30- to 45-min period during each experiment to obtain a true representation of an overall response. All experiments were recorded for later analysis. Leukocyte rolling flux, rolling velocity, and adhesion were defined as described for the cremaster muscle protocol.


Intravital microscopy of mouse cerebromicrovasculature was performed as previously described (9). Briefly, a craniotomy was performed using a high-speed drill (Fine Science Tools, North Vancouver, BC, Canada), and the dura mater was removed to expose the underlying pial vasculature. Throughout the experiment, the mouse was maintained at 37°C, and the exposed brain was kept moist with an artificial cerebrospinal fluid buffer. A microscope was used to observe leukocyte-endothelial interactions in leukocytes that were fluorescently labeled by intravenous administration of rhodamine 6G as described for the intestine preparation. Three different 30- to 70-μm-diameter postcapillary venules were chosen for observation. All experiments were recorded for later analysis. Flux of rolling leukocytes and leukocyte adhesion were measured as described for the cremaster muscle protocol.

Experimental Protocol

In the first series of experiments, leukocyte kinetics in the microcirculation of cremaster muscle, intestine, and brain were determined at 4–5 h (and, in some experiments, at 2 and 24 h) after mice were treated with an intraperitoneal injection of saline or mouse recombinant SDF-1α (10 μg in 100 μl of saline; R & D Systems, Minneapolis, MN). In the second series of experiments, mice were treated locally by intrascrotal injection of saline (control), 0.5 μg of mouse recombinant TNFα (R & D Systems), 1 μg of SDF-1α, or 1 μg of KC (CXCL1, R & D Systems), each in 100 μl of saline, and leukocyte kinetics in the microcirculation of cremaster muscle were quantified 4–5 h later. In some experiments, 0.5 μg of IL-1β (BD Biosciences, San Diego, CA) or TNFα was injected intraperitoneally for 20 h before intrascrotal injection of SDF-1α or simultaneously with SDF-1α to test whether IL-1β or TNFα potentiates the local or systemic effects of SDF-1α on leukocyte recruitment. In some experiments, mice were intravenously infused with 20 μg of rat anti-mouse P-selectin MAb (RB40.34 in 200 μl saline/mouse) and control mice were infused with A110-1 isotype control Ab after 4 h of SDF-1α local or systemic treatment, and leukocyte recruitment was determined before and after 5 min of anti-P-selectin treatment. In some experiments, platelet depletion in the circulation was accomplished by intraperitoneal administration of 50 μl of anti-thrombocyte serum (adsorbed rabbit anti-mouse thrombocyte serum, Accurate Chemical & Scientific, Westbury, NY) simultaneously with intrascrotal administration of SDF-1α for 4 h or at 2 h of intraperitoneal infusion of SDF-1α; then leukocyte recruitment was determined at the 4-h SDF-1α treatment. This protocol of anti-thrombocyte treatment has been shown to deplete circulating platelets by 96% (2).

Statistical Analysis

Values are means ± SE. Student's t-test was applied to compare the statistical difference within two groups, and ANOVA was used for comparison of the differences in more than two groups. P < 0.05 was considered statistically significant.


Effect of Systemic Administration of SDF-1α on Leukocyte Recruitment in Microvasculatures of Cremaster Muscle, Small Intestine, and Brain

To measure SDF-1α-induced leukocyte recruitment, we treated mice systemically by injecting 10 μg of SDF-1α into the peritoneal cavity or locally by injecting 1 μg of SDF-1α into the scrotum. Such treatment did not change mouse peripheral leukocyte counts and hemodynamic parameters, such as shear rate, central line red blood cell velocity, and venular diameters of the cremasteric microvasculature (Table 1; P > 0.05). Therefore, changes in leukocyte recruitment described below are not due to the changes in hemodynamic parameters after SDF-1α treatment.

View this table:
Table 1.

Hemodynamic parameters in mice

Using intravital microscopy to measure the parameters of leukocyte recruitment in cremaster muscle, we tested the effect of systemic and local SDF-1α treatments in mice 2–24 h after injection of SDF-1α. No effect on leukocyte recruitment was observed at 2 or 24 h after injection, regardless of the method of administration. We then focused on the effect of 4–5 h of local or systemic SDF-1α treatment. We tested the dose responses of leukocyte recruitment after systemic (0, 1, 10, and 20 μg ip) or local (0, 0.1, and 1 μg intrascrotal) SDF-1α treatment. The dose-response studies suggest that the most effective doses for systemic (intraperitoneal) and local (intrascotal) SDF-1α treatment were 10 and 1 μg per injection, respectively. Figure 1 shows the leukocyte recruitment in mouse cremaster muscle after systemic administration of 10 μg of SDF-1α at 4–5 h. Leukocyte rolling flux was very significantly increased approximately twofold after SDF-1α treatment (Fig. 1A; P < 0.01). Leukocyte rolling velocity remained unchanged in SDF-1α-treated groups (P > 0.05; Fig. 1B). Leukocyte adhesion was subtly and significantly increased (from 2–3 cells in controls to 7–8 cells in SDF-1α treated mice, P < 0.05 or P < 0.01) and leukocyte emigration was subtly increased (from 3 cells in saline controls to ∼6 cells in SDF-1α treated mice, P < 0.05) by SDF-1α.

Fig. 1.

Effect of systemic administration of stromal cell-derived factor-1α (SDF-1α) on leukocyte recruitment in cremasteric postcapillary venules. SDF-1α (10 μg in 100 μl of saline) was injected intraperitoneally, and mouse cremaster muscle was prepared for intravital microscopy. Leukocyte rolling flux (A), leukocyte rolling velocity (B), number of adherent leukocytes (C), and number of emigrated leukocytes (D) were determined at 4–5 h after SDF-1α administration. Values are means ± SE (n = 4). *P < 0.05; **P < 0.01 vs. saline control.

Although the cremaster muscle remains the gold standard for the examination of leukocyte recruitment in microvasculature in mice, other tissues may have more clinical relevance. Previous studies showed that large amounts of CXCR4 were expressed in brain and intestine vasculature and that when CXCR4 was deleted, the animals exhibited dramatic developmental problems, including defects in vascularization (47, 56). We thus examined leukocyte recruitment in small intestine and brain in mice. Leukocyte rolling flux and adhesion were modestly but significantly increased in microvasculature of the small intestine after 4 h in mice treated systemically with 10 μg of SDF-1α compared with saline control mice (P < 0.05; Fig. 2, A and C). Again, there was no statistical difference in leukocyte rolling velocity in the intestine between saline- and SDF-1α-treated mice (P > 0.05; Fig. 2B).

Fig. 2.

Effect of systemic administration of SDF-1α on leukocyte recruitment in postcapillary venules of mouse small intestine. SDF-1α (10 μg) was injected intraperitoneally, and a segment of mouse small intestine was prepared for intravital microscopy. Leukocyte rolling flux (A), leukocyte velocity (B), and number of adherent leukocytes (C) were determined at 4 h after SDF-1α administration. Values are means ± SE (n = 3). *P < 0.05 vs. saline.

In contrast to cremaster muscle and small intestine, leukocyte rolling flux in the mouse brain did not change after systemic administration of SDF-1α (Fig. 3A). In cremaster muscle and small intestine, there is significant basal rolling, which may be required for enhancement of leukocyte rolling flux by SDF-1α treatment; in brain microcirculation, however, basal rolling flux is not evident, and enhancement may not be possible (Fig. 3A). In the brain, the baseline level of leukocyte adhesion in the control mice was also very low (<1 cell/100-μm length of venule). After SDF-1α treatment, leukocyte adhesion was modestly, but significantly, increased to ∼1–2 cells/100-μm length of venule (Fig. 3B; P < 0.01).

Fig. 3.

Effect of systemic administration of SDF-1α on leukocyte recruitment in mouse brain pial venules. SDF-1α (10 μg) was injected intraperitoneally, and leukocyte rolling flux (A) and number of adherent leukocytes (B) were determined 4 h later. Values are means ± SE (n = 3–4). **P < 0.01 vs. control.

Effect of Systemic Administration of SDF-1α on P-Selectin and VCAM-1 Expression

Because leukocyte recruitment in various organs after systemic SDF-1α administration was increased and it is known that endothelial cells express CXCR4 (15, 42, 56), we hypothesized that systemic SDF-1α treatment activates endothelial cells and that this activation of the endothelium, through the expression of adhesion molecules, supports the increase in leukocyte recruitment. We chose to measure the expression of adhesion molecules in vivo using the dual-radiolabeling technique, inasmuch as it is both quantitative and sensitive (10, 16). Although systemic treatment with SDF-1α induced a modest increase in leukocyte recruitment, we found that in vivo VCAM-1 expression in mice systemically treated with SDF-1α at 5 or 10 μg for 4 h was not upregulated in any tissue except the spleen (Table 2). It is unlikely that E-selectin is important in this model, because systemic SDF-1α treatment did not change leukocyte rolling velocity in the microvasculature (Figs. 1B and 2B), and E-selectin is known to be important in reducing rolling velocity and is also believed to have only a minor role in leukocyte flux (18, 25). We therefore determined intravascular P-selectin expression in vivo after systemic SDF-1α administration.

View this table:
Table 2.

Effect of systemic administration of SDF-1α on VCAM-1 expression in vivo

Table 3 demonstrates the level of P-selectin expression in various organs in mice systemically treated with 10 μg of SDF-1α for 4 h. With only a few exceptions (lung, heart, and bone marrow), this treatment significantly increased P-selectin expression in the majority of tissues, including cremaster muscle, small intestine, and brain, where an increase in leukocyte recruitment was observed. These data suggest that systemic treatment with exogenous SDF-1α activates endothelial cells in vivo to upregulate P-selectin expression in a majority of tissues, and this upregulation may be an underlying mechanism that supports the increase in leukocyte recruitment.

View this table:
Table 3.

Effect of systemic administration of SDF-1α on P-selectin expression in vivo

Effect of Local Administration of SDF-1α on Leukocyte Recruitment in Cremaster Muscle

Next we examined local SDF-1α delivery to determine whether a more robust leukocyte recruitment response occurs. SDF-1α has been shown to be chemotactic for lymphocytes, neutrophils, mast cells, CD34+ hematopoietic stem cells, and tumor cells in various model systems (3, 5, 7, 24, 27). We recently showed that superfusion of the CXC chemokine macrophage inflammatory protein-2 (CXCL2) or KC (CXCL1) or a gradual release of KC from a 1-mm3 piece of agarose gel on mouse cremaster muscle induced substantial neutrophil recruitment in the tissue (28, 29). We performed similar experiments using 5 nM SDF-1α superfusion or 0.5 μM SDF-1α in a piece of 1-mm3 agarose gel on cremaster muscle to induce leukocyte recruitment. However, neither treatment induced leukocyte recruitment in cremasteric microvasculature (unpublished observation). It has also been shown that intrascrotal injection of macrophage inflammatory protein-2 or KC (1 μg) induced massive neutrophil emigration in cremaster muscle at 4–5 h (29). Preliminary observations suggested that more leukocyte recruitment was induced by intrascrotal injection of 1 than 0.1 μg of SDF-1α. We then determined leukocyte recruitment in cremaster muscle at 4–5 h after intrascrotal injection of 1 μg of SDF-1α. Figure 4A shows that leukocyte rolling flux remained at ∼40–60 cells/min in saline-injected and SDF-1α-treated mice. The decrease in rolling flux induced by TNFα compared with saline controls (P < 0.05) did not differ from that induced by SDF-1α (P > 0.05). Leukocyte rolling flux at 4–5 h was profoundly increased by KC to ∼80–100 cells/min, which was significantly higher than leukocyte rolling flux in saline- and SDF-1α-treated groups (P < 0.05). Leukocyte rolling velocity decreased ∼50% after local treatment with SDF-1α for 4–5 h (P < 0.05 or P < 0.01; Fig. 4B). However, KC did not change rolling velocity (P > 0.05), whereas TNFα very significantly lowered leukocyte rolling velocity (to ∼3 μm/s) compared with saline or SDF-1α treatment (P < 0.01). Although the number of adherent leukocytes was increased modestly from the baseline of 2–4 cells to 8–10 cells after treatment with SDF-1α and KC (P < 0.05 or P < 0.01; Fig. 4C), local treatment with TNFα profoundly increased the number of adherent cells to ∼17, which was significantly higher than the number of adherent cells in the groups treated with SDF-1α and KC (P < 0.05 or P < 0.01). Intrascrotal administration of SDF-1α modestly increased the number of emigrated cells at 4 h: approximately three cells per field of view in the baseline control group and nearly nine cells in the group injected with 1 μg of SDF-1α (P < 0.05; Fig. 4D). Locally administered TNFα increased the number of emigrated cells to ∼21, a value significantly higher than in SDF-1α-injected group (P < 0.01). As expected, local injection of 1 μg of KC induced very significantly greater numbers of emigrated cells (mean 109) in the tissue than local treatment with TNFα or SDF-1α (P < 0.05). These results clearly demonstrate that, similar to the systemic SDF-1α effect, local SDF-1α treatment also modestly increased leukocyte adhesion in venules and emigration into tissue.

Fig. 4.

Effect of local administration of SDF-1α, KC, and TNFα on leukocyte recruitment in cremasteric postcapillary venules. After intrascrotal injection of 1 μg of SDF-1α or KC or 0.5 μg of TNFα (each in 100 μl of saline), mouse cremaster muscle was prepared for intravital microscopy. Leukocyte rolling flux (A), leukocyte rolling velocity (B), number of adherent leukocytes (C), and number of emigrated leukocytes (D) were determined at 4–5 h after administration of the reagents. Values are means ± SE (n = 3–5). *P < 0.05; **P < 0.01 vs. saline control. #P < 0.05; ##P < 0.01 vs. SDF-1α.

We examined the tissue histology of cremaster muscles after SDF-1α local injection for 4 h. It has been shown that almost all the cells recruited by TNFα and KC are neutrophils (8, 9, 29, 44, 48). After SDF-1α local treatment, we found that all the cells infiltrated in cremaster muscle were neutrophils. There was no difference in the number of mononuclear cells between the SDF-1α-treated tissue and the non-SDF-1α-treated tissue.

Cytokines and GM-6001 Do Not Potentiate the Effect of SDF-1α on Leukocyte Recruitment

It is known that, during inflammation, matrix metalloproteinases are able to inactivate chemokines in tissues and reduce their activity on leukocyte recruitment (32). To determine whether the low leukocyte recruitment response we observed was due to the inactivation of locally injected SDF-1α by matrix metalloproteinases, we injected mice intrascrotally with a combination of 2 nmol of GM-6001 (Calbiochem, San Diego, CA; 10 μM in 200 μl of saline), the dose that is known to effectively inhibit the activity of these proteases (52), and 1 μg of SDF-1α. We then examined leukocyte recruitment at 4–5 h and compared it with the response to SDF-1α injected alone. There was no difference in leukocyte rolling, adhesion, and emigration between these two groups (not shown). This suggests that the moderate effect of SDF-1α on leukocyte recruitment is unlikely due to the degradation of this chemokine in the tissue.

In tissues or cells, treatment with TNFα or IL-1β may enhance functional responses to another stimulus. Treatment of endothelial cells with TNFα or IL-1β was shown to substantially increase the CXCR4 mRNA transcription in endothelial cells (15) and the mRNA and protein expression in astroglioma cells at different time points (37). We treated mice systemically with IL-1β or TNFα (0.5 μg ip) for 20 h, injected SDF-1α (1 μg) intrascrotally, and examined leukocyte recruitment parameters 4–5 h later. This 24-h IL-1β or TNFα systemic treatment did not potentiate any SDF-1α local effects (not shown). Alternatively, we simultaneously injected SDF-1α (10 μg) and TNFα (0.5 μg) intraperitoneally and examined the recruitment parameters in cremaster muscle at 3–5 h. The response to this treatment was not different from the response to treatment with TNFα alone. These results suggest that inflammatory cytokines may not potentiate the effects of local or systemic SDF-1α in vivo.

Endothelial P-Selectin, but not P-Selectin on Platelets, Is Responsible for SDF-1α-Induced Leukocyte Recruitment

To confirm that P-selectin expression in response to SDF-1α stimulation is responsible for the increased leukocyte recruitment, we treated the mice intravenously with anti-P-selectin MAb R40.34. In systemically and locally SDF-1α-treated groups, MAb R40.34 completely blocked all leukocyte rolling and reduced the number of adherent leukocytes to the level in saline controls (Fig. 5, A and B). These results suggest that the upregulation of P-selectin expression is responsible for the increased leukocyte rolling and adhesion induced by systemic or local SDF-1α treatment. Using the protocol that we have previously shown to reduce circulating platelets by 96% (2), we also treated the mice with a thrombocyte-depleting Ab. In response to systemic and local SDF-1α stimulation, mice treated with anti-thrombocyte demonstrated the same level of leukocyte recruitment as the untreated mice (Fig. 5, C and D; P > 0.05). These results indicate that the upregulation of endothelial P-selectin mediates the increase of leukocyte recruitment induced by exogenous SDF-1α stimulation and that the P-selectin on platelets is not involved in the SDF-1-induced recruitment response.

Fig. 5.

Effect of anti-P-selectin MAb and anti-thrombocyte serum on leukocyte recruitment in cremasteric postcapillary venules in response to systemic and local stimulation with SDF-1α. Mice were treated with SDF-1α [10 μg ip or 1 μg intrascrotally (isc) in 100 μl of saline], and leukocyte recruitment in cremaster muscle was measured at 4 h using intravital microscopy. A and B: after baseline measurement at 4 h of SDF-1α treatment, 20 μg of anti-P-selectin (RB40.34) were infused intravenously (arrow), and measurement was repeated after 5 min. C and D: 50 μl of anti-mouse thrombocyte serum was injected intraperitoneally for 2–4 h, and leukocyte recruitment was measured at 4 h of SDF-1α treatment. There is no significant statistical difference between untreated (−) and anti-thrombocyte-treated (+) mice (P > 0.05). No change in SDF-1α-induced leukocyte emigration was observed after anti-P-selectin or anti-thrombocyte treatment (not shown). Values are means ± SE (n = 3–4). *P < 0.05; **P < 0.01 vs. before anti-P-selectin.


In the present study, using intravital microscopy and quantifying adhesion molecule expression in vivo, we show that administration of SDF-1 induces the upregulation of P-selectin expression on vascular endothelial cells and leukocyte-endothelial cell interactions in the microvasculature of various tissues. This inflammatory response is short-lived (at 4–5 h) and does not last to 24 h. Local or systemic administration of SDF-1 resulted in leukocyte recruitment, which is dependent on endothelial P-selectin and does not involve P-selectin on platelets. However, in contrast to the pronounced leukocyte recruitment responses elicited by systemic TNFα (9) or local TNFα or local KC (present study) treatment, the recruitment response induced by systemic or local SDF-1 is modest.

SDF-1 is a CXC chemokine that is widely and constitutively expressed in vivo but not by leukocytes (43). A well-defined role for chemokines in leukocyte recruitment is induction of leukocyte adhesion through activation of leukocyte integrins. Adherent leukocytes can migrate across the endothelial cellular layer and move toward the source of chemokines or other chemoattractants. Here, we demonstrate that exogenous chemokine SDF-1 acts to induce P-selectin expression on the endothelial cells. A bolus of SDF-1 functions similarly to a cytokine, rather than as a “classical” chemokine. Therefore, exogenous SDF-1 appears to influence all stages of the recruitment process: activation of the endothelium to express P-selectin to mediate leukocyte rolling and induction of adhesion and emigration. Interestingly, KC also induced increased rolling when locally administered to cremaster muscle (Fig. 4). Further study is required to determine whether it can similarly induce adhesion molecule expression or whether any other process is responsible.

Our results suggest that exogenous SDF-1 acts, at least on endothelial cells, to induce P-selectin expression and leukocyte rolling, adhesion, and emigration in the microvasculature. It has been documented that endothelial cells express functional CXCR4, which is the predominant chemokine receptor, and that SDF-1-CXCR4 interactions activate multiple intracellular signaling events in endothelial cells and mediate numerous activities, e.g., endothelial cell activation and angiogenesis (24, 41, 42). In addition, SDF-1- and CXCR4-deficient mice have profound and identical defects in vascular development, suggesting that CXCR4 is the sole receptor for SDF-1 (47). Therefore, among chemokines, SDF-1 is unique to the functionalities of endothelium.

Similar to the effect of the cytokine TNFα (9), systemic administration of SDF-1α resulted in endothelial activation to express P-selectin in most tissues, but, in contrast to TNFα, SDF-1 did not upregulate VCAM-1 expression in nearly all tissues. In the microvasculature of cremaster muscle and small intestine, upregulation of P-selectin would result in a greater numbers of rolling and adherent cells. In the brain vasculature, however, leukocyte adhesion increased, but no rolling was observed, despite the presence of low-level upregulation of P-selectin expression. VCAM-1 is constitutively expressed in the brain vasculature (45) (Table 2), in which blood cells are circulating under high shear conditions. Activated T cells have been shown to tether and immediately adhere in brain vessels via interactions between leukocyte α4-integrin and endothelial VCAM-1 (49). We recently demonstrated that α4-integrin-VCAM-1 interactions in the brain are dependent on or enhanced by P-selectin, presumably by facilitating the initial capture or tethering of circulating leukocytes under high shear conditions (22). SDF-1 has been known to enhance adherent activities of αLβ2-integrin (LFA-1), α4β1-integrin, and α4β7-integrin to their ligands (53, 54). Therefore, the low level of upregulation of P-selectin expression induced by SDF-1 may not increase P-selectin-mediated rolling under high shear conditions but may facilitate integrin-mediated adhesion in the brain.

Although SDF-1 is expressed constitutively in many tissues (43), its activity is likely to be local. Intrascrotal administration of SDF-1 resulted in activation of leukocyte adhesion and emigration in cremaster muscle similar to our observations with systemic administration of SDF-1, except for some minor differences in leukocyte rolling flux and rolling velocity. However, the quality of the local response was very different from the responses to local TNFα and KC. TNFα-induced local inflammation resulted in very slow leukocyte rolling, so slow, in fact, that the number of rolling cells per minute was significantly lower than baseline, although many cells were interacting with the vessel wall. On the other hand, SDF-1 administration resulted in rolling that was slower than baseline, but not as slow as in response to TNFα. Similarly, adhesion and emigration were significantly increased by exogenous SDF-1, but only to about half the levels observed with TNFα. The effects of KC, a chemokine with a well-defined role in inflammation and neutrophil recruitment, were principally on leukocyte emigration. Indeed, the number of emigrated leukocytes observed in the tissue far exceeded that induced by TNFα or SDF-1. Therefore, these three mediators produced three markedly different responses: 1) KC administration resulted in large numbers of quickly rolling cells, moderate adhesion, and tremendous emigration, 2) TNFα resulted in accumulation of many very slow-rolling and adherent cells in the vasculature and significant emigration, and 3) SDF-1 resulted in a very mild activation of all stages of leukocyte recruitment, including slightly slow rolling, modest adhesion, and very subtle emigration.

Because CXCR4 is broadly expressed by many leukocyte subsets (11, 30, 35, 36) and SDF-1 has been shown to be a potent chemoattractant to various subsets of leukocytes, including neutrophils (3, 24, 36), and to hematopoietic stem cells (20, 24), the selective accumulation of only neutrophils in the cremaster muscle tissue in response to 4 h of local SDF-1 stimulation is somewhat unexpected. In one report, subcutaneous injections of SDF-1 over a number of days resulted in accumulation of neutrophils, as well as T cells and other mononuclear leukocytes, in the skin (42). This suggests that longer-term exposure to SDF-1 may permit the recruitment of additional subsets besides neutrophils. Matthys et al. (31) demonstrated that, in mice with collagen-induced arthritis, a 1-μg exogenous SDF-1 injection in periarthritic tissue for 4 h elicited infiltration of mainly polymorphonuclear cells, an inflammation response that could be completely inhibited by AMD-3100, a selective CXCR4 antagonist. This observation suggests that, under some pathological conditions, SDF-1 is able to selectively recruit neutrophils in the inflamed tissue. In the human system, SDF-1 has been shown to stimulate mast cells to secrete IL-8 (26) and induce astroglioma cells to produce monocyte chemoattractant protein-1, IL-8, and interferon-inducible protein-10 (37). These chemokines are chemotactic to certain subsets of leukocytes, including neutrophils. It is unclear whether, in our mouse system, exogenous SDF-1 stimulated the production of exclusive neutrophil chemoattractants in the tissue and, thereby, selectively recruited neutrophils. The specific recruitment of a particular subset of leukocytes can be not only the function of the specific chemokines present on the endothelial surface or in the tissue, but also selective capturing of leukocytes by particular adhesion molecules. It is thus possible that adhesion molecule expression induced by SDF-1 could also contribute to the selective neutrophil recruitment in the present study. This is supported by the fact that a substrate of P-selectin or endothelium expressing P-selectin selectively recruits neutrophils (23, 40). Obviously, this may explain why SDF-1 combined with P-selectin expression elicited exclusive neutrophil recruitment in our cremaster muscle system.

In conclusion, the present study describes a novel role for SDF-1 in the induction of adhesion molecule expression by endothelial cells and in leukocyte-endothelial cell interactions. The effect of exogenous SDF-1 in leukocyte recruitment is modest compared with the inflammatory effect of the cytokine TNFα or the chemokine KC. Our findings may uncover mechanisms that explain the observations of an inflammatory role of SDF-1 in various inflammatory models. We propose that, under physiological conditions, constitutively expressed SDF-1 is a “housekeeping” chemokine, fulfilling its numerous physiological functions, including mediating homing of hematopoietic stem cells, neutrophil homeostasis, and lymphocyte recirculation, and that, under pathological conditions when an extra large amount of SDF-1 is produced or exogenous SDF-1 is introduced, it becomes inflammatory and elicits P-selectin expression and leukocyte recruitment.


This study was supported in part by research funds from the University of Saskatchewan (to L. Liu).


We thank Lori Zbytnuik for expert technical assistance and Dr. Adil I. Khan for help with histological analysis.

Present addresses: S. M. Kerfoot, Department of Laboratory Medicine, Yale University School of Medicine, New Haven, CT 06510; G. Andonegui, Immunology Research Group, Department of Physiology and Biophysics, University of Calgary, Calgary, AB, Canada; C. S. Bonder, Division of Human Immunology, Institute for Medical and Veterinary Science, Hanson Institute, Adelaide, South Australia, Australia.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract