Dr R. Appelberg, IBMC, Rua do Campo Alegre 823, Porto, Portugal. E-mail: email@example.com
Knockout mice for CXC receptor 2 (CXCR2) chemokine receptor were used to study the recruitment of neutrophils during acute and chronic inflammatory responses. When treated with lipopolysaccharide (LPS), either intraperitoneally or intratracheally, these animals had a significant reduction in the neutrophil recruitment in the first 24–48 h as compared with control mice. During 15 days of intraperitoneal infection by Mycobacterium avium, the knockout mice showed significantly reduced numbers of neutrophils in the peritoneal cavity as compared with the control mice. In contrast, the recruitment of neutrophils to the lungs during an aerogenic M. avium infection was not affected by the CXCR2 mutation throughout the 60 days of the study. Finally, we could not find any impact of the mutation on the mycobacterial growth of the infected animals. These findings indicate that CXCR2 may be essentially involved in acute inflammatory responses where an early and rapid recruitment of neutrophils is observed.
Neutrophil recruitment is observed in most acute inflammatory reactions but is generally not found in chronic inflammation. In contrast, mycobacterial infections induce, in addition to these acute neutrophilic responses, a protracted response characterized by the immune-mediated accumulation of neutrophils at the sites of infection [1–7]. Antigen-specific T cells mediate this persistent influx of neutrophils [2, 4–6], and chemokines produced by the immune cells are involved in the recruitment of the neutrophils . In the mouse models of mycobacterial infection, where such neutrophil accumulation was found, depletion of neutrophils has shown that these cells play a small but consistent role in resistance to infection [7–9]. Although the way the neutrophils protect against mycobacteria is not yet clearly understood, both a role in the induction of the immune response [7, 9] and the participation in the effector mechanisms that involve the macrophage through the cooperation between the two phagocytes  have been postulated.
In contrast, it is well documented that the local or systemic exposure to lipopolysaccharide (LPS) triggers an acute inflammatory response and induces a rapid recruitment of neutrophils into the host tissues. Cytokines expressed by the LPS-stimulated cells are believed to control the nature and magnitude of the inflammatory infiltration [10, 11]. LPS appears to trigger the release of such cytokines by engaging Toll-like receptors, namely TLR4 , and signalling through the NF-κB transcription factor .
The accumulation of neutrophils in the inflamed tissues depends on a series of events that lead to the migration of this predominantly blood-borne leucocyte into those tissues . Adhesion molecules are required to allow the marginalization of the circulating leucocytes and their transmigration across the endothelial barrier, and chemotactic factors signal both the leucocyte and the endothelium in order for this migration to occur. This recruitment is mediated by different chemoattractants like the complement component C5a , bioactive lipids such as platelet-activating factor and leukotriene B4 , and the chemotactic cytokines [17–19]. These chemokines are expressed by several cell types including epithelial cells, macrophages, T cells and to a lesser extent neutrophils, in response to various stimuli including pro-inflammatory cytokines such as interleukin 1 (IL-1) and tumour necrosis factor (TNF)-α. Chemokines are small proteins with four cysteines forming two essential disulphide bonds. The subfamilies are distinguished according to the position of the first two highly conserved cysteines, which are either adjacent (CC) or separated by one amino acid (CXC) in most molecules, with the exception of lymphotactin (a C-type chemokine) and fractalkine (a CX3C chemokine) [17–19]. Members of the CXC subfamily are predominantly chemotactic for neutrophils, while the CC elements are chemotactic for mononuclear cells [17–19]. IL-8, a CXC member, is the major chemotactic element for human neutrophils, and to this date, a murine homologue has not yet been described. Two types of CXC receptors are known in humans: CXC receptor 1 (CXCR1) is described as having high affinity only for IL-8, but CXCR2 has equal affinity for all CXC chemokines [20, 21]. As type 2 CXC receptor binds a series of different chemokines , it may therefore play a major role in neutrophil recruitment during inflammation. The recent availability of CXCR2 gene-disrupted mice  prompted us to analyse the role of this receptor on the recruitment of neutrophils taking place during mycobacterial infections. The choice for this model was made given the broad range of chemokines to which CXCR2 responds and therefore the likelihood that a deficiency at this level would overcome the functional redundancy observed within the neutrophil-specific chemokines. The study was controlled for the expression of the natural resistance-associated macrophage protein 1 (Nramp1) gene, namely by selecting as control strain the one which harbours the allele expressed in the mutant strain, because this gene plays a major role in determining resistance to Mycobacterium avium infection [23, 24].
Materials and methods
Animals CXCR2 gene-disrupted (CXCR2–/–) mice  were obtained from the Jackson Laboratories (Bar Harbor, ME, USA). Even though the breeders had been back-crossed to BALB/c mice, they were genotyped for the Nramp1 allele as both this gene and the mutated CXCR2 gene locate very closely in chromosome 1. The polymerase chain reaction (PCR) screening for the Nramp1 allele was done as previously described [24, 25]. Briefly, genomic DNA samples were obtained from each mouse by treating a portion of a ear with proteinase K. The amplification of the Nramp1 gene was performed using Taq DNA polymerase (Ampligene-Oncor, Gaithersburg, MD, USA) and primers from the Nramp1 gene, one oligonucleotide being common to both alleles and the other being specific for either the R or S allele as described elsewhere . The amplification was done in a Gene Amp PCR System 9600 (Perkin Elmer-Roche, Branchblurg, NJ, USA). As expected from the physical proximity, the CXCR2–/– mice retained the resistant (G169) allele from 129 mice (from which the embryonic stem cell derives) rather than acquiring the susceptible (D169) allele from the BALB/c (Fig. 1). Therefore, for most experiments, we used as wild-type control mice the BALB/c congenic strain, C.D2, which has the resistant allele of Nramp1 from DBA/2 mice . In some experiments, heterozygous (CXCR2+/–) mice were used, and no differences in the results were found between these and the C.D2 animals.
BacteriaM. avium strain ATCC 25291 was grown for 2 weeks until mid-log phase in Middlebrook 7H9 medium plus 0.04% Tween 80 at 37 °C. The bacteria were harvested by centrifugation, suspended in a small volume of saline and sonicated with Branson (Danbury, CT, USA) sonifier for 15 s at 50 W to disrupt bacterial clumps. This suspension was then diluted, frozen in aliquots, and kept at −70 °C until use. Before inoculation, bacterial aliquots were thawed at 37 °C and diluted in saline to the desired concentration.
Evaluation of bacterial growth Infected mice were sacrificed at different time-points of infection, and the organs were aseptically collected and homogenized in a 0.04% Tween 80 solution in distilled water. The number of colony-forming units (CFU) of M. avium in the livers, spleens, and lungs of the infected mice was determined by serial dilution of the tissue homogenates and plating onto 7H10 agar medium supplemented with oleic acid/albumin/dextrose. Plates were incubated for 2 weeks at 37 °C, and the number of colonies was counted. Results are expressed as the log10 values of CFU per organ.
Study of cellular response to the peritoneal infection The protocol was followed as described elsewhere [1, 2]. Briefly, mice were intraperitoneally (i.p.) infected with 106 CFU of M. avium 25291 or given 30 µg of LPS i.p., which were previously shown to lead to inflammatory reactions resulting in the accumulation of neutrophils [1–6]. At different time intervals, mice were sacrificed by ether anaesthesia, and the inflammatory peritoneal cells were collected with 4 ml of cold phosphate-buffered saline (PBS). Differential cell counts were made from stained cytospin preparations. A total of 400 or more cells in several fields were counted in each slide.
Study of cellular response to the intratracheal infection Protocols used to study lung recruitment were adapted from Saunders and Cheers . Anaesthetized mice received an intratracheal instillation of 5 µg of LPS/animal  or 5 × 104 CFU of M. avium 25291 . At each time-point, animals were sacrificed and pulmonary blood vessels were washed with PBS with 0.02% ethylenediaminetetraacetic acid (EDTA) until the tissue turned white. Lungs were then canulated in the trachea and lavaged with 5 ml of cold PBS/0.02% EDTA. Differential alveolar cell counts were made from stained cytospins.
To obtain the cells from the deeper lung tissues, lungs were sliced with the aid of surgical blades and incubated for 30 min at 37 °C in 5 ml of Dulbecco's modified Eagle's medium (DMEM) with 0.25% Dispase (Boehringer Mannheim, Mannheim, Germany). Tissue remains were removed by passing through metal sieves. Lung cell suspensions were then centrifuged at 800×g for 10 min. Erythrocytes were lysed using a haemolytic solution (155 mm NH4Cl, 10 mm KHCO3, pH 7.2). Differential cell counts were made from stained cytospins as described above.
Statistical analysis Data were compared using the Student's t-test.
CXCR2 –/– mice show deficient recruitment of neutrophils after challenge with LPS
We initially characterized the inflammatory response to LPS in the CXCR2-deficient animals as compared with the wild-type controls (C.D2 mice). LPS was either injected i.p. or was instilled into the trachea of anaesthetized mice. At different time intervals, inflammatory cells were obtained by peritoneal or bronchoalveolar lavage (BAL), and counted and characterized. As shown in Fig. 2, LPS induced the accumulation of neutrophils when injected i.p. or given intratracheally. The recruitment of neutrophils to the peritoneal cavities of the CXCR2–/– mice was significantly reduced as compared with the control animals (Fig. 2A). The recruitment of neutrophils into the alveoli of LPS-treated mice was almost completely abrogated in the CXCR2–/– mice as compared with the control animals (Fig. 2B). The cellular composition of the peritoneal cavity and of the bronchoalveolar space was markedly different with a higher cellularity in the peritoneum. As most of the cells recruited into the lung following administration of LPS were neutrophils (80–95% of the cells), the total number of cells in the BAL was drastically reduced in the mutant strain, whereas no such effect was observed in the peritoneal cavity where a significant number of mononuclear cells was always detected. In either case, no statistically significant differences were found between the two strains regarding the number of macrophages and lymphocytes present in the exudates induced by LPS (data not shown).
Neutrophil recruitment in M. avium-infected CXCR2–/– versus control mice
To characterize the neutrophil accumulation during mycobacterial infections, we infected mice with strain 25291 of M. avium either by the intraperitoneal route or by the intratracheal instillation. The accumulation of neutrophils in the peritoneal cavity of mice infected i.p. had a biphasic pattern in the control mice as initially reported by us  and was significantly reduced to very low levels by the mutation of CXCR2 (Fig. 3). Again, no statistically significant differences were found in the number of macrophages and lymphocytes present in the mycobacteria-induced exudates (data not shown). In contrast, in the aerogenic infection, no statistically significant differences were found in the neutrophil influx into the alveolar space between the two groups of mice (Fig. 4A) or among the macrophage and lymphocyte populations (data not shown). Neutrophils associated with the solid lung tissue, i.e. after performing the BAL and the in vitro digestion of the tissue, were found in smaller numbers in the control mice as compared with the CXCR2–/– animals on day 15, but statistically significant differences were not found at other time-points (Fig. 4B). No statistically significant differences were found for the number of macrophages and lymphocytes present in the solid lung tissue (data not shown). Given the lack of effect of the mutation, the analysis of the lung infection was extended up to 60 days, but no significant differences were found even at later time-points.
Mycobacterial growth is not affected by the absence of CXCR2
No significant differences in mycobacterial loads in the CXCR2–/– mice as compared with the controls were found in the organs of the animals from the previous experiments. After the intratracheal infection, the mycobacterial loads in the lungs at day 30 were 5.93 ± 0.29 log10 CFU in the control mice and 5.59 ± 0.45 log10 CFU in the CXCR2–/– mice, and at day 60 were 6.89 ± 0.49 log10 CFU in the control mice and 6.56 ± 0.42 log10 CFU in the CXCR2–/– mice. After the i.p. infection, there was a rapid dissemination of the inoculum to the liver and spleen, and at day 15 of infection, mycobacterial counts in the liver were 5.52 ± 0.72 log10 CFU for the control mice and 4.93 ± 0.10 log10 CFU for the CXCR2–/– mice, and in the spleen were 5.04 ± 0.12 log10 CFU in the control mice and 5.10 ± 0.37 log10 CFU in the CXCR2–/– mice. Given the difficulties in interpreting the latter experiment owing to the simultaneous occurrence of bacterial removal from the peritoneum and concomitant multiplication in the tissues, an additional experiment was performed where the proliferation of M. avium in the organs of the CXCR2–/– and control mice was followed after intravenous infection. An additional group of BALB/c mice was included. As shown in Fig. 5, the proliferation of M. avium was dependent on the allele of Nramp1 but not on the presence of a functional CXCR2 gene. Thus, bacterial growth was greater in BALB/c mice than in the other two strains (P < 0.01 for all time-points and for both strains). The mycobacterial numbers in the CXCR2–/– mice were not statistically different from those in the C.D2 mice.
In this study, we have compared the participation of CXCR2 during acute and chronic inflammatory responses, namely as regards the recruitment of neutrophils. In the first part of this study, we looked at acute inflammatory responses using as a model the intraperitoneal or the intratracheal administration of LPS, a well-studied phlogistic agent. We observed that in both the experimental settings, CXCR2–/– mice showed a deficient recruitment of neutrophils after LPS inoculation, suggesting a major role of CXCR2 in the acute inflammatory responses. These data are in agreement with previous work which used CXC chemokine antagonists as tools to study the involvement of CXCR2 in acute inflammation. A macrophage inflammatory protein (MIP)2 antagonist was found to inhibit the neutrophil recruitment induced by subcutaneous or peritoneal injection of several pro-inflammatory products (TNF-α, LPS and IL-1α) . Other authors  demonstrated that the direct instillation of LPS into the lungs of rats induced neutrophil accumulation, enhanced the production of CXC chemokines and increased the expression of intercellular adhesion molecule (ICAM)-1. Thus, it is clear from these studies that CXC chemokines and their receptor, CXCR2, play a central role in the inflammatory recruitment of neutrophils in LPS-induced responses.
The participation of CXCR2 during mycobacterial infections is less clear. We have already demonstrated that mycobacteria-specific T cells mediate the accumulation of neutrophils in the peritoneal cavity during mycobacterial infections [2, 4–6]. In addition, macrophages are also involved in the recruitment of neutrophils  as are the chemokines MIP1 and MIP2 . The expression of these chemokines was also found to be reduced in beige mice which have an increased susceptibility to M. avium infection . Here we show that the recruitment of neutrophils to the peritoneal cavity also involves CXCR2 as the CXCR2–/– mice had an almost complete absence of neutrophils in the first 2 weeks of infection. Despite this, the growth of M. avium in the target organs (liver and spleen) of the animals was not affected. As most of the intraperitoneal inoculum is rapidly drained out of the peritoneal cavity and reaches the target organs, the lack of peritoneal influx of neutrophils is expected to have no major effect on mycobacterial proliferation if no deficient recruitment of these cells is found in the solid tissues, as discussed below.
In contrast with the previous results, we could not find any differences in leucocyte accumulation during aerogenic M. avium infection in the two groups of mice studied. One possible explanation may relate to the delayed cellular response observed in the lungs. A greater cell recruitment was only observed at 2 months of infection. This is in agreement with the chemokine expression data reported by Flórido et al. . However, even then no statistically significant differences between neutrophil accumulation in the control and mutant mice were observed. A second explanation may relate to the tissues analysed. Thus different organs may have different pathways of cell recruitment. In fact, the presence of neutrophils in solid tissues was never found to be affected by the mutation (histological data not shown). These cells were present in similar numbers in the liver, spleen and septa of the lungs in the control and mutant mice or were even found at slightly higher numbers in the deficient animals, although statistically significant differences were only found at one time-point. The latter observation may relate to the already-described accumulation of neutrophils in the lymphoid organs of gene-disrupted mice even in the absence of any infection, as well as to an increase in the number of circulating neutrophils that may find their way into the infectious foci by CXCR2-independent pathways . Thus, it was not surprising to find that resistance to infection by mycobacteria was not reduced in the CXCR2–/– mice. In this respect, it is important to stress that the correct control mice for such experiments were the BALB/c congenics expressing the resistant allele (G169) of the Nramp1 gene. Genomic analysis of the CXCR2–/– mice showed that these mice, albeit having been back-crossed to BALB/c mice, still retained the Nramp1 allele of the manipulated embryonic stem cell (the 129 strain with the D169 allele of Nramp1). This is not surprising given the close proximity of the two loci (roughly 0.5 cM). These data stress the importance of carefully taking genetics into consideration when using gene-disrupted mice to study infection. Had we used BALB/c mice as the controls for our experiments, marked differences between these mice and the mutated animals would have been found and wrongly attributed to the induced mutation. Previous work on the role of neutrophils in M. avium infection has used BALB/c or C57Bl/6 mice which have the susceptible D169 allele of Nramp1. Thus, it is still possible that the lack of an effect of the mutation in the resistance of the mice used here may relate to their Nramp1 phenotype inasmuch as neutrophils may be more important when mice are naturally susceptible to infection.
Other models have shown that CXCR2 plays a role in protective immunity to infection. Moore et al.  showed that an antibody specific for CXCR2 both blocked neutrophil influx to the lungs of mice aerogenically infected with Nocardia asteroides and increased susceptibility to infection. Balish et al.  studied the CXCR2–/– mice used in this work and showed that the genetic deficiency in this receptor led to an increased susceptibility to Candida albicans infection. However, a word of caution regarding the use of the CXCR2–/– mice should be made as phenotypic manifestations of these mice include splenomegaly and lymphadenopathy resulting from the extensive B-cell hyperplasia and an increase in circulating neutrophils .
In conclusion, we have shown that CXCR2 participates in neutrophil recruitment in some acute inflammatory situations such as that induced by LPS in the peritoneal cavity and in the alveolar space of the lung, and in M. avium-induced inflammation in the peritoneal cavity. However, neutrophil recruitment to other sites such as the liver, spleen and lung septa may not be affected, consistent with the fact that we could not see any impact of the mutation in the control of M. avium.