Main

The well documented tendency of neonates to acquire serious bacterial infections has been explained as partly due to an impaired stimulus-induced adhesion and migration of NN. This has been demonstrated both in animalin vivo experiments(1, 2) and in studiesin vitro with adhesion of isolated human NN to biologic surfaces in the presence or absence of flow and shear stress(35). In neutrophil endothelial interaction, L-selectin on the neutrophil surface participates in the initial rolling contact of the neutrophil with carbohydrate ligands expressed on the endothelium(69), whereas the firm and irreversible contact and subsequent migration mainly is exerted by theβ2-integrin Mac-1 interacting with ICAM-1 on the endothelium(1012). Flow cytometric analysis has clearly demonstrated a reduced surface expression of L-selectin and a diminished stimulus-induced up-regulation of Mac-1 in NN(4, 5, 13, 14). The latter has been suggested to be caused by impaired mobilization of Mac-1 from gelatinase-containing granules(15), albeit the existence of separate gelatinase granules has been a matter of debate. We have recently confirmed the existence of gelatinase granules as an easily mobilizable subpopulation of peroxidase-negative granules and demonstrated this subset to be a reservoir of Mac-1 in AN(1618). Yet, the up-regulation of Mac-1 observed upon stimulation of neutrophils with weak secretagogues mainly originates from a newly discovered mobilizable compartment, the so-called secretory vesicles, rather than from gelatinase granules(19). Secretory vesicles are unique in their mobilization characteristics, because unlike granules, they are almost completely mobilized at low concentrations of inflammatory mediators likely to be operable during neutrophil endothelial interaction(1921). Besides Mac-1, secretory vesicles have been shown to contain CR1, the b-cytochrome component of the NADPH-oxidase, the receptor for formyl peptides, Fcγ receptor III, and urokinase type plasminogen activator receptor(2228), and are therefore likely to serve their function by rapidly supplementing the plasma membrane with these functional proteins upon interaction of the neutrophil with inflamed endothelium. Secretory vesicles were originally identified by their content of latent alkaline phosphatase, i.e. alkaline phosphatase activity measurable only in the presence of a detergent(20). It is well known that the content of alkaline phosphatase of NN is severalfold higher compared with adult cells(15, 29), and although one report demonstrated the presence of latent alkaline phosphatase in NN(30), it remains to be established whether the increased amount of alkaline phosphatase indicates an increased number of secretory vesicles in these cells or merely a higher concentration of the enzyme per vesicle. Moreover, because mobilization of secretory vesicles may supplement the plasma membrane with as yet unidentified functional proteins of importance for neutrophil adherence and migration, we found it of particular interest to investigate the existence in NN of secretory vesicles and to evaluate their mobilization characteristics. This is performed by subcellular fractionation of unperturbed and FMLP-stimulated, nitrogen-cavitated cord blood neutrophils on three-layer Percoll density gradients. Furthermore, mobilization of granules and secretory vesicles is investigated by measurement of release of matrix markers and by flow cytometric analysis of surface up-regulation of alkaline phosphatase, CR1, and Mac-1 upon stimulation of isolated cells with an array of inflammatory mediators.

METHODS

Isolation of neutrophils. Cord blood was drawn before delivery of the placenta from uncomplicated deliveries of normal full-term pregnancies and immediately anticoagulated in 25 mM sodium citrate, 126 mM glucose. Adult cells were obtained from whole blood donated by healthy volunteers. Cells were isolated by dextran sedimentation [2% (wt/vol) dextran (Pharmacia, Uppsala, Sweden) in saline] followed by centrifugation of the leukocyte-rich supernatant on Lymphoprep (Nygaard, Oslo, Norway) to remove mononuclear cells. The remaining erythrocytes were lysed for 30 s in ice-cold water and reconstituted in an equal volume of 1.8% NaCl. The cells were washed once in saline before resuspension in the desired buffer. Isolation of cells was carried out at 4 °C, except for dextran sedimentation, which was performed at room temperature. Isolated neutrophils were more than 93% pure, as judged from flow cytometry (forward side scatter plots).

Subcellular fractionation. Unperturbed or FMLP-stimulated (see below) neutrophils were incubated at 0.5-1 × 107/mL for 5 min in diisopropylfluorophosphate (5 mM, Aldrich Chemical Co., Milwaukee, WI), followed by centrifugation and resuspension in the initial volume of disruption buffer (100 mM KCl, 3 mM NaCl, 1 mM ATPNa2, 3.5 mM MgCl2, 10 mM 1,4-piperazinediethanesulfonic acid, pH 7.2) containing 0.5 mM phenylmethylsulfonyl fluoride (Sigma Chemical Co.). Cells were disrupted by nitrogen cavitation (pressurized for 5 min) as described(31). Nuclei and intact cells were pelleted by centrifugation at 400 g for 15 min (P1). Seven to 10 milliliters of the postnuclear supernatant (S1) were applied on top of a 3 × 9 mL three-layer Percoll (Pharmacia) gradient (1.05/1.09/1.12 g/mL) containing 0.5 mM phenylmethylsulfonyl fluoride and centrifuged at 37,000 × g for 30 min. This resulted in a gradient with four visible bands, from the bottom designated the α-band, the β1-band, theβ2-band, and the γ-band. The cytosol was present above theγ-band on top of the Percoll. The gradient was collected in 35 equally sized fractions by aspiration from the bottom of the tube. All fractions were assayed for markers as described below. Recovery was calculated as total amount of marker present in the fractions in percentage of the amount applied in the S1. NN and AN were always fractionated in parallel at equal cell concentrations.

Stimulation of neutrophils. For stimulation, the AN and NN were resuspended in KRP (130 mM NaCl, 5 mM KCl, 1.27 mM MgSO4, 0.95 mM CaCl2, 5 mM glucose, 10 mM NaH2PO4/Na2HPO4, pH 7.4) at 1 × 107 cells/mL and preincubated for 5 min at 37°C. After addition of the stimulus, cells were incubated for 15 min. One vial of cells was kept on ice, another at 37 °C, both without addition of stimulus. The incubation was stopped by addition of 1 volume of ice-cold KRP followed by centrifugation at 200 × g for 6 min. The supernatant (S) was aspirated and the pellet (P) resuspended to the initial volume in KRP. The content of marker enzymes (myeloperoxidase, lactoferrin, NGAL, gelatinase, and albumin; see below) was measured in supernatant and pellet. Release of granule markers was calculated as S/(S + P), and expressed in percent.

For evaluation of stimulus-induced changes in surface expression of various membrane-associated proteins, unperturbed or stimulated cells were processed for flow cytometric analysis as described below.

Cells were stimulated with the following secretagogues: FMLP (10 nM, Sigma Chemical Co.), LTB4 (10 nM, Sigma Chemical Co.), PAF (10 nM, Sigma Chemical Co.), ZAS (5%, vol/vol), GM-CSF (200 U/mL, Sandoz Pharmaceuticals, East Hanover, NJ), IL-8 (200 pg/mL, a kind gift from dr. K. Thestrup-Pedersen, Department of Dermatology, Marselisborg Hospital, Aarhus, Denmark), TNF-α (500 U/mL, Amersham International, Amersham, UK), and ionomycin (1μM, Calbiochem, La Jolla, CA). AN and NN were always investigated in parallel.

Marker assays. Myeloperoxidase (azurophil granules), lactoferrin, and NGAL (specific granules)(32), gelatinase (gelatinase granules), albumin (secretory vesicles)(33), and HLA (plasma membranes) were all measured by ELISA as described elsewhere(32). Subcellular fractions were also analyzed for their content of lysozyme, measured by a recently developed ELISA(34). Secretory vesicles were also identified by latent alkaline phosphatase, i.e. alkaline phosphatase measurable only in the presence of detergent (0.2% Triton X-100)(20, 35). The percentage of latent alkaline phosphatase is calculated as the content of latent alkaline phosphatase in S1 and P1 divided by the total content in S1 and P1.

Flow cytometric analysis. A total of 0.5 × 106 neutrophils were incubated for 30 min in saturating concentrations of primary MAb in PBS containing 0.5% BSA. Neutrophils were washed twice in PBS/BSA followed by 30′ incubation in FITC-conjugated Fab fragments of rabbit anti-mouse Ig (DAKO, diluted 20-fold in PBS/BSA). Subsequently cells were washed three times in PBS/BSA and resuspended in PBS, containing 1% formaldehyde and analyzed in a FACSort (Becton Dickinson, Mountain View, CA) flow cytometer. Results are expressed as median fluorescence intensity of 5000 cells. The following primary antibodies were used: anti-CR1 (CD35, DAKO, Glostrup, Denmark, diluted 50-fold), anti-CD11b (α-chain of Mac-1, DAKO, diluted 20-fold), anti-HLA class 1 (DAKO, diluted 20-fold), anti-alkaline phosphatase(36) (diluted 100-fold), anti-L-selectin(Becton Dickinson, diluted 20-fold). Cells were incubated with preimmune mouse IgG1 (Becton Dickinson, diluted 50-fold), to estimate the degree of nonspecific binding.

Statistical analysis. Differences between AN and NN were tested by the paired t test. p values < 0.05 were considered significant.

RESULTS

Subcellular fractionation of AN on three-layer Percoll density gradients results in four bands: the α-band containing azurophil granules, theβ1-band containing the specific granules (lactoferrin-rich, peroxidase-negative granules), the β2-band containing the gelatinase granules (defined as gelatinase-rich, lactoferrin-poor, peroxidase-negative granules)(17, 18), and theγ-band containing plasma membranes and secretory vesicles. When the distribution profile of membrane and matrix markers in fractions from NN is compared with adult cells (Fig. 1) the overall picture is the same, i.e. the azurophil granule marker myeloperoxidase is located largely to the α-band, the specific granule markers lactoferrin and NGAL colocalize in the β1-band, the majority of gelatinase is localized distinct from NGAL and lactoferrin in the β2-band, and secretory vesicles (latent alkaline phosphatase and albumin) are localized with the plasma membranes (HLA) in the γ-band. However, marked differences in peroxidase-negative granules (specific and gelatinase granules) of NN and AN were consistently observed in all of eight experiments. Notably the amount of lactoferrin and NGAL was considerably lower in NN, and a larger part is localized to granules in the β2-band(Table 1 and Fig. 1). Thus, the specific granules are lighter in NN than in AN. This indicates a structural difference between granules from NN and AN. Whether this is simply a matter of there being less protein in each granule as indicated by the lower total amount of NGAL and lactoferrin in NN (Table 1) or due to less efficient concentration of proteins in the granules is unknown. To some extent, the same difference is observed regarding gelatinase, because the amount of gelatinase is lower in NN (Table 1), and especially the shoulder of gelatinase extending into the β1-band in AN (representing gelatinase present in specific granules)(18) is greatly diminished in NN (Fig. 1). The distribution of lysozyme, mainly a consistent of azurophil and specific granules(37, 38) is consistent with these observations,i.e. that the lysozyme present in specific granules is shifted to a lesser density (Fig. 1). As judged from the profiles of myeloperoxidase and from the lysozyme present in the α-band, azurophil granules of NN do not differ from AN with regard to their isopycnic density.

Figure 1
figure 1

Subcellular fractionation of AN and NN on three-layer Percoll density gradients. Subcellular fractionation of isolated NN and AN was performed as described in “Methods.” Subcellular fractions and postnuclear supernatant (S1) were assayed for the following markers: latent alkaline phosphatase, albumin, gelatinase, lactoferrin, NGAL, myeloperoxidase, lysozyme, and HLA. The graphs represent the average of eight experiments (normalized to 1 × 107 cells/mL), except for lysozyme(average of five experiments). The average recovery of marker proteins(determined as total content in fractions 1 through 35 in percent of the content in the postnuclear supernatant (S1)) was between 80 and 120%. Please note the different y axis for neonatal and adult cells(except for myeloperoxidase and HLA). For comparison of total amount of marker proteins in NN and AN, please refer to Table 1.

Table 1 Content of marker proteins in neonatal vsadult neutrophils

Marked differences were also observed between NN and AN regarding the content of secretory vesicles. NN contains 10 times as much alkaline phosphatase as AN (Table 1), but the fraction which is latent, i.e. localized to secretory vesicles, was the same in NN and AN (74 versus 72%). Surprisingly, the amount of matrix protein of secretory vesicles, albumin was reduced in NN compared with AN(Fig. 1 andTable 1), probably reflecting the lower concentration of albumin in neonates compared with adults. These findings indicate that the amount of secretory vesicle membrane is not increased in NN compared with AN (which is supported by the observations on CR1, a marker of secretory vesicles, described below).

To investigate the mobilization characteristics of the different granule and vesicle subsets, unperturbed and FMLP-stimulated NN and AN were subjected to subcellular fractionation (Fig. 2). The hierarchy in mobilization known from AN(18, 39) also holds true in neonates, because secretory vesicles are the most easily mobilized organelle followed by gelatinase granules, specific granules, and azurophil granules in that order. The mobilization of secretory vesicles in NN is readily observed by the almost complete disappearence of both latent alkaline phosphatase and albumin upon FMLP stimulation (Fig. 2). Also there is a significant stimulus-induced loss of gelatinase (42% reduction after stimulation), especially from the lightest gelatinase granules, whereas very little NGAL and lactoferrin (Fig. 2) and no myeloperoxidase disappear after stimulation. This is in agreement with the findings in adult cells (not shown; please refer toRef. 18).

Figure 2
figure 2

Subcellular fractionation of unperturbed or FMLP-stimulated NN. NN at 1 × 107/mL in KRP were either kept on ice (control) or stimulated with FMLP at 10 nM for 15 min at 37 °C after a 5-min preincubation period at 37 °C. The cells were subsequently subjected to subcellular fractionation as described in “Methods.” Fractions were assayed for their content of latent alkaline phosphatase, albumin, gelatinase, lactoferrin, myeloperoxidase, and HLA. Data represent the average of three experiments.

To further characterize the mobilization of the different exocytosable subsets, NN and AN in suspension were stimulated with either one of an array of inflammatory mediators, that in part employ different signal transduction pathways in addition to the calcium ionophore, ionomycin. The release of the matrix markers albumin, gelatinase, lactoferrin, NGAL, and myeloperoxidase was determined, and simultaneously the change in surface content of HLA class 1, CD11b (αm subunit of Mac-1), alkaline phosphatase, CR1, and L-selectin was evaluated by flow cytometry. Table 2 andFigures 3 and4 sum up the findings. The overall observation is that NN respond to all secretagogues tested, and regardless of the stimulus, the hierarchy in mobilization of the various subsets described above holds true, because the secretory vesicle marker albumin is mobilized to the largest extent followed by gelatinase, lactoferrin/NGAL, and myeloperoxidase in that order. Generally, the release of the different granule and vesicle markers is augmented in NN compared with adult cells, but not in all cases significantly (Fig. 3,Table 2). There is a higher release of NGAL in comparison to lactoferrin in NN, although the subcellular fractionation data indicate these to colocalize in specific granules. This is also the case in AN(32), and may be caused by lactoferrin sticking to the plasma membrane after exocytosis, thus resulting in underestimation of lactoferrin release.

Table 2 Release of secretory vesicle and granule markers upon stimulation of AN and NN
Figure 3
figure 3

Granule and secretory vesicle mobilization upon stimulation of NN and AN. Isolated NN or AN at 1 × 107 cells/mL in KRP were incubated either on ice (4 °C), at 37 °C without addition of stimulus, or at 37 °C in the presence of FMLP, PAF, IL-8, ZAS, GM-CSF, LTB4, TNF-α, or ionomycin at the concentrations indicated in“Methods.” Release of marker proteins is calculated as their content in supernatant in percentage of the total content in supernatant and pellet. Open bars are AN, hatched bars are NN. Neonatal and adult cells were always stimulated in parallel. Bars represent the average+ SEM of seven experiments (for ionomycin, six experiments). For statistical analysis, see Table 2.

Figure 4
figure 4

Expression of surface antigens upon stimulation of NN and AN. Isolated NN or AN at 1 × 107 cells/mL in KRP were incubated either on ice (4 °C), at 37 °C without addition of stimulus, or at 37 °C in the presence of FMLP, PAF, IL-8, ZAS, GM-CSF, LTB4, or ionomycin at the concentrations indicated in “Methods.” One million control or stimulated cells were subjected to flow cytometric analysis using primary MAb against alkaline phosphatase (AP), CR1, CD11b(αm subunit of Mac-1), L-selectin, and HLA. The bars show median fluorescent intensity (MFI, average + SEM of five to seven experiments). Open bars are AN, hatched bars are NN. Unspecific fluorescense using preimmune mouse IgG1 was always below 15. Please note the different y axis for neonatal and adult cells in the diagrams showing AP and CD11b.

Secretory vesicle mobilization can be evaluated by release of albumin and loss of latent alkaline phosphatase(20, 33) as shown for NN above. With a MAb against alkaline phosphatase, we have measured the surface up-regulation of alkaline phosphatase by flow cytometric analysis(Fig. 4). It is obvious that there is a similar pattern of stimulus induced surface up-regulation in NN and AN, and that no further up-regulation occurs after stimulation with ionomycin which in addition to mobilizing secretory vesicles also leads to substantial exocytosis of gelatinase and specific granules (Fig. 3 andTable 2), confirming the lack of alkaline phosphatase in these subsets. Furthermore, the very high content of alkaline phosphatase in NN is confirmed, although the difference between neonates and adults is not as striking as when evaluated by enzymatic activity. By combined subcellular fractionation and free flow electrophoresis and by flow cytometry we have demonstrated that secretory vesicle mobilization is responsible for the up-regulation of CR1 and Mac-1 observed after stimulation of AN with inflammatory mediators likely to be operable during early neutrophil activation(19, 21, 22). The parallel up-regulation of neonatal and adult CR1 and Mac-1, respectively, indicates that this is also likely to be the case in NN, although these cells are deficient in Mac-1 relative to adult cells, with a surface expression of 60 to 70% of that observed in adults (Fig. 4). The further up-regulation of Mac-1 observed in NN after ionomycin stimulation indicates that a substantial proportion of Mac-1 is localized in gelatinase and specific granules as is the case in adult cells(18, 21, 40). There is a reduction in CR1 levels in ionomycin-stimulated cells due to proteolytic degradation of the protein, which could not be completely inhibited in the presence of the serine protease inhibitor phenylmethylsulfonyl fluoride.

L-Selectin participates in the initial rolling contact of neutrophils with E- and P-selectin-presenting endothelium(69). Generally, the surface expression of L-selectin is lower on NN compared with AN, and the stimulus-induced shedding in NN is less than that observed in AN in the presence of most agonists (Table 3, Fig. 4). This confirms findings obtained by others(4, 13).

Table 3 Relative stimulus induced shedding of L-selectin in AN and NN

DISCUSSION

We have presented data suggesting the importance of secretory vesicle mobilization in the conversion of the neutrophil from a serene, circulating cell to a cell expressing increased amounts of a number of functional proteins including Mac-1 and thus being capable of adhering firmly to inflamed endothelium(19, 21). Because stimulus-induced adhesion is impaired in NN, we investigated the presence of secretory vesicles in NN and their mobilization characteristics. Indeed, secretory vesicles are present in NN, as demonstrated by subcellular fractionation which revealed the presence of a pool of latent alkaline phosphatase in the light membrane region colocalizing with albumin, another secretory vesicle marker. The mobilization kinetics of secretory vesicles of NN are similar to those of adult cells, because they are almost completely mobilized upon stimulation by weak secretagogues. This was demonstrated by the disappearance of both latent alkaline phosphatase and albumin in subcellular fractions upon stimulation of the cells by 10 nM FMLP, and by release of albumin and surface up-regulation of alkaline phosphatase upon stimulation of NN with an array of inflammatory mediators employing different signal transduction pathways. The release of albumin upon stimulation was very similar in AN and NN, although there was a surprisingly large release of albumin in neonatal control cells kept at 4°C, where one would expect exocytosis to be completely inhibited. This albumin is unlikely to originate from secretory vesicles, because it is not accompanied by a relative up-regulation of alkaline phosphatase in NN compared with AN, or by a loss of latent alkaline phosphatase in these control cells, in which 74% of the alkaline phosphatase activity is latent (compared with 72% in AN).

Previous investigations have demonstrated increased alkaline phosphatase activity in NN(15, 29). We could confirm this finding, and the flow cytometric demonstration of a 4-5-fold higher surface expression of alkaline phosphatase in neonatal cells compared with adults indicates that the increased activity is at least partly caused by an increased number of molecules, rather than merely by a higherVmax of the neonatal enzyme. The increased amount of latent alkaline phosphatase in NN could indicate an increased number of secretory vesicles in these cells as proposed by others(30), but in that case one would expect an increase in the albumin content of NN, because albumin is taken up by an endocytic process during the formation of secretory vesicles(33). This increase was not observed, but on the contrary the albumin content of NN was reduced, probably reflecting the lower plasma albumin concentration of neonates. Patients with myeloproliferative diseases with either elevated (myelofibrosis, polycythemia vera) or decreased (chronic myelogenous leukemia) alkaline phosphatase content also had a normal number of secretory vesicles(41) as we have found to be the case in neutrophils from third trimester pregnant women (L. Kjeldsen and N. Borregaard, unpublished observation), who like neonates are known to have elevated alkaline phosphatase levels.

The pattern of stimulus-induced surface up-regulation of CR1 and Mac-1(CD11b) in NN (Fig. 4) strongly indicates these to be located in secretory vesicles as is the case in AN(19, 22). In accordance with this it is conceivable that secretory vesicles of NN also are an intracellular store of the receptor for formyl peptides, the Fcγ receptor III, the receptor for urokinase-type plasminogen activator, and the b-cytochrome component of the NADPH-oxidase(22, 24, 25, 27, 28). So far, alkaline phosphatase is the only known membrane constituent which is elevated in secretory vesicles because CR1 expression is normal(14, 42), whereas the expression of both Mac-1 and the receptor for FMLP is reduced(5, 14) in resting and activated NN. The reduced surface up-regulation of Mac-1 in response to FMLP has been claimed to be caused by a diminished mobilization from gelatinase-rich, peroxidase-negative granules (gelatinase granules), as evaluated by subcellular fractionation, but that study did not pay attention to the existence of secretory vesicles(15). Furthermore, a recent paper demonstrated an overall reduction of total cell Mac-1 content to approximately 66% of the adult level(43), which is in complete agreement with our flow cytometric data showing that Mac-1 surface expression of NN is approximately 60 to 70% of the adult level regardless of the induced activation state of the cells (Fig. 4). It is thus likely that the Mac-1 content is evenly reduced in all Mac-1 containing subsets of NN.

We have confirmed the existence of gelatinase granules as a gelatinase-rich, lactoferrin-poor subpopulation of peroxidase-negative granules by combined subcellular fractionation and double-labeling immunogold electron microscopy(1618). This granule subset was found to be more easily mobilized than the remaining peroxidase-negative granules, which are defined as specific granules by their content of lactoferrin. Our data clearly demonstrate that this granule heterogeneity and the mobilization characteristics of the subpopulations also holds true in NN, as previously shown by Jones et al.(15) who also found a lower overall density of specific granules in cord blood neutrophils. Our finding regarding the reduced total gelatinase content of NN is in accordance with the data obtained by Joneset al. but we found the stimulus induced release of gelatinase to be higher (not in all cases significantly) in NN in comparison to AN, whereas Jones found similar gelatinase release upon stimulation (with FMLP). This difference may be explained as due to their use of a functional assay for gelatinase, which is sensitive to the activity of neutrophil proteases present during the assay, thus precluding precise determination of gelatinase release in percentage of the total content. This is not a problem with our immunologic assay, which in contrast to the enzymatic assay does not involve incubation of neutrophil samples at 37 °C(44). Our demonstration of augmented release of both gelatinase and specific granules in NN indicates a primed state of the cord blood neutrophils, as previously described by others(30, 45).

By pulse chase experiments and immunohistochemical staining of bone marrow cells separated according to their maturational stage, we have demonstrated that the peroxidase-negative granule heterogeneity can be explained by a differential but overlapping synthesis of peroxidase-negative granule proteins(46). Gelatinase granules are thus a marker of terminal neutrophil differentiation and are formed as the gelatinase synthesis continues at the band stage where the synthesis of lactoferrin has ceased. The presence of gelatinase granules in NN indicates a normal timing of the biosynthesis of peroxidase negative granule proteins, but because the content of many granule proteins including lactoferrin, NGAL, gelatinase, and Mac-1 is diminished in NN (Table 1)(15, 43, 45), the biosynthetic capacity of neutrophil precursors is likely to be reduced. It has been suggested that spontaneous release of granule proteins during isolation of cord blood neutrophils could explain the reduced content of the various granule proteins(15), but the preservation in NN of secretory vesicles, the most readily mobilized subset in the neutrophil, precludes that significant granule mobilization has occurred during the isolation procedures. Contamination of the isolated NN with immature myeloid cells or nucleated red cells could possibly explain the observed difference, but this is unlikely, because NN were more than 93% pure as judged from flow cytometry (forward side scatter plots).

In conclusion, we have confirmed the presence of secretory vesicles in NN and demonstrated their profound sensitivity to mobilization by inflammatory mediators. Abnormalities in secretory vesicles are thus unlikely to cause the well known functional defects of NN, which are more likely explained by the combined effects of reduced surface expression of L-selectin and Mac-1 in combination with diminished liberation of gelatinase during neutrophil activation. Furthermore, our data indicate a reduced capability of NN precursors to synthesize a number of peroxidase-negative granule proteins, which may also be the case for as yet unidentified proteins of importance for neutrophil adhesion, diapedesis, and migration.