Neutrophil adhesion molecules in HIV disease

Abstract

Neutrophil dysfunction in HIV disease is well described. We examined the expression of neutrophil adhesion molecules amongst 72 HIV-infected subjects using a whole blood flow cytometric assay with FITC- and R-PE-labelled isotype-specific MoAbs. We report lesser expression of CD11a (LFA-1) and l-selectin (CD62L) on the circulating neutrophils of HIV⁺ subjects compared with HIV⁻ controls. Expression of CD11b (Mac-1) was unchanged. Shedding of l-selectin and up-regulation of CD11b in response to in vitro stimulation with N-formyl-methionyl-leucyl-phenylalanine (fMLP) were less in HIV⁺ compared with HIV⁻ subjects, most markedly in subjects with CD4 cell counts < 100 cells/mm³. These results suggest that neutrophil dysfunction in HIV disease, which increases with disease progression, may be attributable to dysregulated adhesion molecule expression.

INTRODUCTION

The incidence of non-neutropenic bacterial sepsis is markedly increased in HIV disease [1]. Although B cell abnormalities are recognized [2], neutrophil dysfunction is also well described [3]. Published data on neutrophil function in HIV infection are difficult to interpret, as results are inconsistent and often conflicting due to differences in method, in study populations and in concomitant medication [3–7]. Abnormal neutrophil chemotaxis is the most commonly described defect for most HIV⁺ groups.

Adherence of the circulating neutrophil to endothelium, followed by extravasation and chemotactic movement to the site of inflammation, relies on the sequential expression of adhesion molecules. For this reason we have investigated the expression of the principal neutrophil adhesion molecules in HIV-infected patients.

MATERIALS AND METHODS

Reagents

The following IgG1 mouse anti-human MoAbs were used: FITC-conjugated anti-CD11a (Serotec Ltd, Oxford, UK), FITC-conjugated anti-CD18 (Dako Ltd, High Wycombe, UK), FITC-conjugated anti-l-selectin/CD62L (R&D Systems Europe Ltd, Abingdon, UK), R-PE-conjugated anti-CD11b (Ancell; Alexis Corp. (UK) Ltd, Nottingham, UK), FITC-conjugated anti-CD15 (Dako), FITC-conjugated isotype control (Dako) and R-PE-conjugated isotype control (Sigma Chemical Co., Poole, UK).

The other reagents used were PBS (Unipath Ltd, Basingstoke, UK), N-formyl-methionyl-leucyl-phenylalanine (fMLP), and Orthomune lysing reagent (Ortho Diagnostic Systems Ltd, High Wycombe, UK).

Subjects

Seventy-two HIV-infected adults were studied, 66 men and six women. Patients with ongoing bacterial or protozoal infections were excluded. Subjects were reviewed either through case notes or in person 2–4 weeks later to confirm there had been no subsequent emergence of infection. Details of all therapies received at the time of investigation were recorded and patients taking non-steroidal anti-inflammatory drugs (NSAIDs) or opiates were excluded. HIV⁻ control samples were taken from clinic and laboratory staff. Whole blood samples were drawn into EDTA tubes and transferred immediately onto ice. Samples were processed within 1 h of venesection in all cases. All samples were processed by the same investigator (D.A.J.M.).

Determination of neutrophil adhesion molecule expression

Experiments were performed in duplicate on three subjects at a time, two patients and one HIV⁻ control. Utilizing a whole blood flow cytometric method the expression of the adhesion molecules CD11a (LFA-1), CD11b (Mac-1), CD18 and l-selectin (CD62L) was analysed. Further, the response of neutrophil l-selectin and CD11b expression to in vitro stimulation with the bacterial peptide fMLP was assessed. Briefly, from EDTA samples on ice, 50 μl of whole blood were either (i) kept on ice and incubated with 50 μl PBS, or (ii) incubated in a water bath at 37°C with 50 μl PBS, or (iii) incubated in a water bath at 37°C with 50 μl fMLP (10⁻⁷m). After 30 min all samples were then returned to ice and remained so or at 4°C until flow cytometric analysis was performed. Following incubation, labelling of the 100-μl samples with R-PE- and FITC-conjugated MoAbs was performed. R-PE-conjugated anti-CD11b and FITC-conjugated anti-l-selectin were combined in the same tube, as were the R-PE- and FITC-conjugated isotype control antibodies. FITC-labelled anti-CD18 and anti-CD11a were added to separate tubes. All samples were stained on ice for 30 min, after which 2 ml ice-cold Orthomune lysing solution were added to each tube. Erythrocyte lysis for 30 min was followed by centrifugation at 200 g for 10 min at 4°C and 1.5 ml of cell-free supernatant was discarded. The cells were resuspended in the residual liquid for flow cytometry. Centrifuging markedly speeded up the processing. Earlier experiments had demonstrated that alteration of l-selectin and CD11b expression only occurred when centrifuging was performed at temperatures above 15°C (data not shown).

Flow cytometry analysis

We used an Ortho Diagnostic Systems' Ortho Cytoron Absolute flow cytometer with Immunocount II software. Forward and side scatter identified the granulocyte population and the validity of this gate was confirmed using FITC-conjugated anti-CD15 antibody. R-PE- and FITC-conjugated isotype controls were run initially so that > 98% of cells appeared in the lower left quadrant (non-specific staining). The percentage of specifically stained cells was calculated by subtracting the percentage of isotype control stained cells found beyond the statistical area (generally 0.2–2%) from that seen with the specific antibody. Further, because of occasional variability in non-specific staining, mean cell fluorescence (MCF) of the whole neutrophil population, i.e. not just those with fluorescence in excess of isotype background, was calculated.

This method proved robust with low inter- and intra-assay variability. As an example, the coefficient of variability (CV%) for CD11a MCF was 4.48 for all HIV⁻ controls and 3.09 for multiple repeat assays on one individual.

Statistical analysis

Reference ranges were established from the data for the HIV⁻ control group and were defined as values lying within 2 s.d. of the mean (ref. range = mean ± 2 s.d.). The two-tailed heteroscedastic (for two samples of unequal variance) Student's t-test (P value) was used to determine differences between HIV⁺ and HIV⁻ groups.

The study was approved by the Riverside Hospitals Ethics Committee.

RESULTS

The percentage of polymorphs expressing CD11a, CD18 and l-selectin in HIV⁺ patients was significantly reduced compared with HIV-1⁻ controls (Tables 1 and 2), with the most marked effect being seen in those with a CD4 count < 100 cells/mm³. There was great variability in the percentage of neutrophils expressing these adhesion molecules in this latter group (Fig. 1). Many patients had a percentage of neutrophils expressing these markers below the reference range for HIV⁻ controls.

Table 1.

Neutrophil l-selectin and CD11b expression

Table 2.

Neutrophil CD11a and CD18 expression

Fig. 1.

Increased variability of neutrophil adhesion molecule expression in HIV⁺ subjects with CD4 counts < 100.

In contrast, the unstimulated expression of CD11b was much lower than for the other adhesion molecules and there was no significant difference between HIV⁺ and HIV⁻ subjects (Table 1).

For CD11a and l-selectin the median and mean MCF were significantly lower than HIV⁻ controls, with again the greatest reduction in those with a CD4 count < 100 cells/mm³. For CD11b and CD18 the median and mean MCF were significantly increased in the HIV⁺ group.

Response of adhesion molecules to stimulation with fMLP

The mean percentage of neutrophils expressing residual l-selectin after fMLP stimulation was similar in HIV⁺ and HIV⁻ controls, although 24% of those with CD4 counts < 100 cells/mm³ exceeded the reference range. Residual MCF was significantly increased amongst HIV-infected subjects, with a significant trend as CD4 cell count fell (Table 1).

As expected, CD11b expression increased markedly following stimulation, but was significantly greater in HIV⁻ controls than in HIV⁺ subjects. The ability to up-regulate CD11b in response to stimulation was lowest in the CD4 < 100 cells/mm³ subgroup, where almost two thirds were below the lower limit of normal for HIV⁻ controls (85.4%). MCF values gave the same picture (Table 1).

DISCUSSION

We believe this to be the first report of neutrophil adhesion molecules measured in HIV⁺ patients by a whole blood flow cytometric assay. The technique avoids procedures for the isolation of neutrophils from whole blood known to alter their function and cell surface molecule expression [8–11]. Further advantages include reduced sample handling and the fact that low neutrophil counts, which can limit the number of traditional tests of neutrophil function performed, pose no problem.

The reproducibility of the technique is good, with minimal inter- and intrasubject and inter- and intra-assay variability.

An important problem in HIV⁺ patients is the careful definition of clinical groups, as it is recognized that many concomitant drugs, particularly opiates and non-steroidal anti-inflammatory agents, as well as occult bacterial infection, may alter neutrophil adhesion molecules [12,13]. Therefore all our patients were carefully questioned and were followed closely for a minimum of 1 month following the laboratory assay.

We have demonstrated that expression of the principal neutrophil adhesion molecules in subjects with HIV infection exhibits a broader range of expression and altered neutrophil responses to in vitro stimulation compared with controls.

Expression of CD11a was observed on a reduced proportion of neutrophils in HIV-infected subjects and at reduced intensity, most markedly in those with CD4 counts < 100 cells/mm³.

CD11b expression on circulating neutrophils was negligible in both HIV⁺ and HIV⁻ groups, but up-regulation of cell surface expression in response to in vitro stimulation was impaired amongst HIV-infected subjects, in terms of both the proportion of neutrophils which responded and the intensity to which the up-regulated CD11b was expressed. These findings were again most marked amongst those with CD4 counts < 100 cells/mm³. The bacterial peptide fMLP activates neutrophils in vitro in a manner which provides a model for what happens at the interface with activated endothelium in vivo. Up-regulation of CD11b expression facilitates firm adhesion to the endothelial ligand intercellular adhesion molecule-1 (ICAM-1) and facilitates diapedesis to sites of infection. Reduced up-regulation would impair these processes.

CD18 forms the common β subunit for at least three other cell surface molecules (CD11a, CD11b, CD11c) and overall expression will therefore be the net result of several different influences. Nevertheless, expression of CD18 was seen on a lower proportion of HIV⁺ neutrophils in the CD4 < 100 cells/mm³ subgroup. It is unlikely that CD18 itself has any direct role in adhesion to any endothelial ligand, so its adhesive role is likely to be as part of the complexes with CD11a (LFA-1) and CD11b (Mac-1).

The intensity of neutrophil l-selectin expression was reduced in HIV-infected subjects. Amongst HIV-infected subjects, particularly the CD4 < 100 cells/mm³ subgroup, a lower percentage of neutrophils expressed this molecule, and shedding of l-selectin in response to in vitro stimulation was reduced in the latter group. l-selectin controls neutrophil rolling by forming brief weak adhesive interactions with endothelial ligands. Alterations in its expression could clearly have an adverse effect, impairing ability to respond to chemotactic stimuli by ‘rolling’. Moreover, their failure to shed l-selectin appropriately might interfere with subsequent adherence and extravasation.

Whilst the physiological relevance of these statistically significant but numerically modest differences is unclear, it should be noted that all the differences detected are in the direction we would have predicted given our knowledge of neutrophil dysfunction in HIV disease.

Three previous reports of leucocyte adhesion molecules in HIV-infected individuals have appeared [14–16] and only two of these studied neutrophils. In one, of 42 HIV⁺ subjects, there were no data concerning risk group, prescribed and unprescribed drug (opiate and NSAID) use, disease stage or presence or absence of intercurrent infections [14]. However, their finding of impaired l-selectin shedding concurs with our own. Their finding of CD11b up-regulation on circulating neutrophils might be explained by artefactual activation from the isolation methods employed in their study. Another study reporting elevated CD11b expression in intravenous drug users [15] is difficult to interpret because of the confounding effects of opiates [9].

Previous data on HIV⁺ individuals have mainly concentrated on functional tests of neutrophils and are confusing. Phagocytic function has been reported as subnormal [17], supranormal [18], oxidative function as excessive [19] and impaired [20], and bacterial killing as reduced [21] and unchanged [22]. Two studies have suggested that heat-labile soluble serum factors may play an important part in HIV-associated neutrophil dysfunction [23,24]. The one functional test which appears to be impaired in all studies is chemotaxis. Chemotactic responses rely on appropriate sequential expression of cell surface adhesion molecules. Dysregulation of this sequence would lead to impaired neutrophil chemotaxis, and we have therefore used these molecules as a marker for neutrophil dysfunction, an approach which has been validated previously in other patient groups [25,26]. A significant limitation of this and other studies is the observation that conformational change may be at least as important as expression intensity for the in vivo activity of many adhesion molecules. However, we believe that results demonstrating impaired chemotaxis could be explained by abnormalities in leucocyte adhesion molecule expression and shedding shown in this and other studies.

The clinical importance of these results is difficult to assess. Bacterial sepsis in HIV infection is sometimes due to neutropenia or to indwelling central venous catheters, but in other individuals occurs without obvious precipitating cause and it is in these that neutrophil dysfunction may have a role. It might be that administration of cytokines such as granulocyte colony-stimulating factor (G-CSF) or granulocyte-macrophage colony-stimulating factor (GM-CSF) may have a role in increasing adhesion molecule expression in this subgroup of patients, which would explain reports of augmented in vitro adhesion [27].

The cause of these changes in neutrophil adhesion molecules in HIV-infected individuals is unclear. Although one report has suggested that HIV is able to infect polymorphs [28], this is not the conventional view. It seems more likely that complex changes in the cytokine milieu associated with advancing HIV infection are responsible. A wide range of cytokines is known to influence adhesion molecules in vitro and such changes are likely to be of central importance in many homeostatic mechanisms within the body. Neutrophil ‘priming’ or ‘preactivation’ may occur as a result of specific cytokine influences, and this is one possible explanation for our findings. Such ‘primed’ neutrophils may thus be less functionally capable. Changes of adhesion molecules in HIV⁺ individuals may have a more generalized importance than simply in relation to neutrophil function. For instance, very similar adhesion molecules are present on lymphocytes, and the homing of these to the gut mucosa is controlled by similar mechanisms to those involved in neutrophil diapedesis. It is striking that lymphocyte depletion of the lamina propria occurs at an early stage of HIV infection and is out of proportion to the circulating CD4 cell count [29]. It is equally striking that with the advent of highly effective anti-retroviral therapy, marked improvements in diarrhoea, secondary to protozoal infection, often occur rapidly, and may be related to recovery of mucosal immune function secondary to changes in lymphocyte adhesion molecules.