Abstract
The aim of the study was to determine possible factors related to the risk of developing recurrent bacterial respiratory tract infections in HIV-1-infected patients, regardless of the degree of immune cellular impairment. Thirty-three HIV-1 seropositive patients with previous repetitive bacterial respiratory tract infections (case group), 33 HIV-1 seropositive controls (matched by CD4-cell counts) without these antecedents and 27 healthy controls were studied before and after administration of pneumococcal and Haemophilus influenzae type b vaccines. Clinical or toxicological variables, cutaneous tests, complement factors, beta2-microglobulin, serum IgM, IgA, IgG and subclasses, specific antibodies (IgG, IgG2, IgA) against pneumococcal vaccine and polyribosylribitol phosphate (PRP), their avidity, opsonophagocytosis and IgG2m and FcγRIIa allotypes were determined. A history of drug abuse (P = 0·001), less likelihood of receiving high activity antiretroviral treatment high activity antiretroviral treatment (HAART) (P = 0·01), higher levels of HIV-1 viral load (P < 0·05), serum IgG (P < 0·01) and beta2-microglobulin (P < 0·01) were observed in the case group. Also, a lower increase in specific antibodies to pneumococcal vaccine and PRP was demonstrated in the cases in comparison with the two control groups. No differences were observed in the avidity of antibodies, opsonophagocytic capacity or IgG2m and FcγRIIa allotypes between the three groups. These data indicate that vaccination strategies against encapsulated bacteria can be unsuccessful in the HIV-1-infected patients presenting repetitive bacterial respiratory tract infections if behavioural aspects or measures to improve adherence to HAART therapies are not considered.
Keywords: AIDS, bacterial pneumonia, Haemophilus influenzae vaccine, FcγRIIa, HIV infection, opportunistic infection, phagocytosis, pneumococcal vaccine
INTRODUCTION
Encapsulated bacteria, especially Streptococcus pneumoniae and Haemophilus influenzae, are mainly responsible for upper and lower respiratory tract infections in patients with HIV-1 infection [1–6]. These bacteria are not susceptible to complement-mediated lysis, so the immunological response against them is based on the synthesis of specific immunoglobulins, which favour opsonization and later phagocytosis by polymorphonuclear leucocytes (PMN). For convenience, the protective capacity of serum antibodies is evaluated by the measurement of antibody titration. It is generally true that among vaccinated individuals there is a correlation between polysaccharide specific antibody concentrations, as measured by ELISA and serum opsonophagocytic activity. However, exceptions to this pattern have been observed [7–10]. An important functional characteristic of antibody efficacy is its avidity for the antigen. In healthy adults the IgG2 subclass is the main specific polysaccharide antibody. The link between IgG2 and phagocytosis of encapsulated bacteria is constituted by FcγRIIa receptors on effector cells. The existence of different allotypes in IgG2 and FcγRIIa has suggested that the genetic background also determines the defence of a given individual to encapsulated bacteria [11,12].
During HIV-1 infection, disorders in B lymphocytes and PMN function have been described; in addition, lower vaccination responses have also been observed [13–23]. All these dysfunctions are usually found in all HIV-1 seropositive patients, increasing in measure as the CD4 lymphocyte cell count decreases.
In the era of high activity antiretroviral treatment (HAART), a significant decline in the incidence of all opportunistic infections has been observed [24–26], but bacterial respiratory tract infections continue to be a frequent consultation motive and a subgroup of more predisposed HIV-1-infected patients seems to exist [26–28]. Some external factors, related to certain risk practices such as drug abuse, have been shown to favour the development of recurrent bacterial infections [29–32]. However, an immunological substratum related to a more pronounced deterioration in some humoral immune mechanisms referred to previously could not be rejected.
Our aim was to study jointly the clinical, toxicological and main immunological mechanisms involved in the defence against encapsulated bacteria to assess why some HIV-1-infected patients develop recurrent bacterial respiratory tract infections regardless of their immunological cellular status.
SUBJECTS AND METHODS
Study subjects and study design
Thirty-three HIV-1-infected patients, over 18 years old, with recurrent bacterial respiratory tract infections were studied. All the patients had suffered from at least two bacterial sinusitis or bacterial pneumonia episodes during the previous year or three or more episodes since the diagnosis of HIV-1 infection. The diagnoses of bacterial sinusitis or pneumonia were based on microbiological, clinical or therapeutic response criteria. Patients with previous nosocomial infections, a Karnofsky index of less than 70% and those that did not grant consent in writing were not included. To neutralize the effect of cellular immunological state, a seropositive control without previous bacterial sinusitis or pneumonia episodes and with an equivalent CD4 cell count (±50 cells/mm3) was selected for each case. All HIV-1-infected patients (cases and controls) were attended successively in the Departments of Infectious Diseases of the Hospital Universitari Son Dureta and Fundació Hospital Son Llatzer. A second control group made up of 27 healthy adults, without antecedents of recent acute infection, history of serious infection, recurrent infections, allergy and IgE levels <100 IU/ml, was studied. These individuals were recruited through the Preventive Medicine Service from among the staff of both hospitals. The demographic characteristics (gender, age, race and geographical origin) of this last group were comparable with the two HIV-1-infected groups.
In addition to demographic data (gender and age), history of drug abuse, antiretroviral treatment and level of HIV-1 viral load were collected in both HIV-1-infected groups.
Serum samples were taken before vaccination (basal visit) and 4 weeks after antipneumococcal (PNEUMO-23 Pasteur-Mérieux®, Lyon, France) and anti-H. influenzae type b (HIBTITER Lederle®, New York, NY, USA) vaccination. In the basal visit the following parameters were determined: lymphocyte subpopulations, HIV-1 viral load, cutaneous tests, complement factors C3, C4, CH50, beta2-microglobulin, serum IgM, IgA, IgG (including IgG subclasses), IgG2m (n) and FcγRIIa (CD32) allotypes, specific IgG, IgG2 and IgA antibodies against polyribosylribitol phosphate (PRP) and pneumococcal whole vaccine. Also, specific IgG and IgA antibody avidity and PMN opsonophagocytic capacity were measured. Four weeks after vaccination, specific antibodies, avidity and PMN opsonophagocytic capacity were determined again.
This study received local ethical committee approval and all the participants signed a consent form.
Lymphocyte subpopulations were determined by flow cytometry with monoclonal antibodies (FACSCalibur, Becton Dickinson Immunocytometry Systems®, California, USA). HIV-1 viral load was determined by PCR (Amplicor-Roche Diagnostics Corp®, Indianapolis, USA).
Cutaneous tests
Intradermic inoculations of three antigens (tuberculin, candida and tetanus) in the forearm were carried out and read 48–72 h later. Positive-induration diameters or a lack of response to the three inoculated antigens (anergy) were collected.
Complement factors
C3, C4, serum IgM, IgA, IgG and subclasses were measured by nephelometry (BNAII, Dade Behring®, Marburg, Germany) and CH50 by means of serum haemolytic activity (results expressed as haemolytic units).
Beta2-microglobulin was determined by a microparticle immunoassay (MEIA, Abbott Diagnostics®, Illinois, USA).
Specific antibodies to pneumococcal and H. influenzae type b polysaccharides
Total IgG and IgG2 levels to S. pneumoniae and H. influenzae type b were determined by an ELISA test based on the method described by Metzger et al. [33], but modified by using the pneumococcal vaccine and PRP (Pasteur Mèrieux®, Lyon, France; lot. Hi 107B) as antigen, as described by Siber et al. [34]. This technique has been described extensively elsewhere [35]. All sera were absorbed with cell wall polysaccharide before test.
The results of total antipneumococcal IgG were expressed as arbitrary units using a reference serum of 46 UA/ml calibrated against a pneumococcal reference preparation, labelled EQC 1/96 sample B, and with the assigned value of 14·5 µg/ml anti-Hib IgG antibody from the European Quality Scheme for specific antibodies (Immunology Department [Oxford] Radcliffe Hospital, Oxford, UK).
The amounts of specific antipneumococcal and anti-H. influenzae type b IgG2 and IgA were determined by the same ELISA method used for specific total IgG, using HRP-labelled antihuman IgG2 (clone HP6014) and antihuman IgA (Sigma A3062), and the results were expressed as absorbance (A450) values.
Antibody avidity determination
Antibody avidity of human IgG and IgA antibodies antipneumococcal and anti-H. influenzae type b were determined using the chaotropic thiocyanate ion in ELISA assays based on the method described by Pullen et al. [36]. The results are expressed as affinity index, representing the molar concentration of thiocyanate required to reduce the initial optical density, in the absence of thiocyanate, by 50%.
Opsonophagocytosis assay
Flow cytometric method for quantifying phagocytosis was performed according to the method described by Martin et al. [37], with some modifications. FITC-labelled H. influenzae type b and S. pneumoniae serotype 3 were representative of the most frequent isolates (9·7%) in our population [38].
Allotyping methods
G2m(n) typing was carried out by double diffusion assay according to Sarvas et al. [39]. Each serum sample was simultaneously tested against anti-IgG2 (clone HP6014) and antiG2m(n) (clone SH-21).
Determination of FcγRIIa allotypes was carried out by nested polymerase chain reaction (PCR-SSP) as described by Carlsson et al. [40]. Briefly, DNA was isolated from EDTA-anticoagulated peripheral blood using the proteinase K method in a nested PCR with IIa-R131 and IIa-H131 allele-specific primers. After the first PCR amplification, using primer pair P52 and P63, the product obtained was reamplified using sequence-specific primers. The sequence-specific sense primers P5G (specific for G507 = IIaR131) and P4A (specific for A507 = IIaH131) are located at the polymorphic site on exon 4, and a common antisense primer, P13, is located on intron 4. To increase the specificity, two mismatch bases (T instead of C) were introduced at position 502 in both primers. A 440-bp fragment from the C-reactive protein (CRP) gene was used as an internal positive control.
Statistical analysis
Statistical analyses were performed using SPSS 6·1 (SPSS, Chicago) and Epi Info 6·04a (CDC, USA) software. First, a descriptive analysis of the subjects constituting the sample was carried out, to check the initial comparability of the cases and control groups, and the confidence intervals of each independent variable were calculated. Mean geometric concentrations of total immunoglobulins, complement factors, CH50 and beta-2microglobulin were calculated. Also, pre- and post-vaccination mean geometric concentration of specific antibodies, avidity and PMN opsonophagocytic capacity were determined. In accordance with previous studies in HIV-1-infected patients the increase of specific antibodies was calculated as the quotient between basal (denominator) and post-vaccination (numerator). An adequate humoral immune response was arbitrarily defined as a twofold or greater increase in concentration of specific antibodies for pneumococcal vaccine and a threefold or greater increase for H. influenzae type b [23,41,42]. The Kolmogorov–Smirnov test was used as a normality proof, with a logarithmic transformation carried out on those variables that did not follow a normal distribution (specific antibodies and avidity). In the bivariate analysis χ2 or Fisher's exact test were used to compare non-numerical variables and Student's t-test or a non-parametric test (Mann–Whitney U-test) for numerical variables, according to the results of the normality proof. Multivariate analysis was conducted using a logistic regression method to control for potential confounding variables. Variables showing a statistical significance in bivariate analysis were included in a regression model and the adjusted odds ratios, with 95% confidence limits, as well as P-values were calculated.
RESULTS
Demographic and immunological variables
Gender and age of the two HIV-1-infected groups and their history of drug-abuse, antiretroviral treatment, viral load as well as variables related with cellular and humoral immunity are shown in Table 1. The patients in the cases group had presented, before inclusion, 22 episodes of bacterial sinusitis, mean 1·6 [1–4] per patient, and 107 bacterial pneumonias, mean 3·6 [1–11] per patient. All the HIV-1-infected patients (cases and controls) and healthy adults were of local origin. Thirty-four HIV-1-infected patients had a history of parenteral drug-abuse, nine homosexual relations, one had these two risk factors, 18 heterosexual relations, two acquired the infection after a blood transfusion and in two patients the route of infection was unknown. It is remarkable that 24 patients in the cases group had a history of drug abuse (seven remaining active) in contrast with 11 HIV-1-infected control subjects (no one active) (P = 0·001). No statistical differences were observed for the other risk factors. Also, in the HIV-1-infected control group, more patients were receiving antiretroviral treatment at inclusion (P = 0·01) and had better virological control (P < 0·05).
Table 1.
Demographic characteristics, history of drug-abuse, variables related with HIV-1 infection state and main immunological parameters in the two HIV-1-infected patient groups (cases and controls)
| Seropositive patients’ characteristics | Case (n = 33) | Controls (n = 33) | P |
|---|---|---|---|
| Males | 24 (72·7%) | 25 (75·7%) | n.s. |
| Women | 9 (27·3%) | 8 (24·3%) | n.s. |
| Age | 34·7 (24–46) | 36·8 (28–57) | n.s. |
| Tobacco smokers | 30 (90·9%) | 21 (63·6%) | <0·01 |
| Cannabis smokers | 22 (66·6%) | 14 (42·4%) | <0·05 |
| Parenteral drug-abuse | 24 (72·7%) | 11 (33·3%) | 0·001 |
| Inhaled cocaine | 4 (12·1%) | 1 (3%) | n.s. |
| HAART | 25 (75·7%) | 32 (96·9%) | 0·01 |
| PCR-HIV > 5 × 103 copies/ml | 18 (54·5%) | 10 (30·3%) | <0·05 |
| CD4 lymphocytes (cells/mm3) | 242·6 (±213·1) | 247·57 (±234·4) | n.s. |
| CD4 lymphocytes (%) | 14·78 (±10·3) | 16·57 (±16·5) | n.s. |
| CD4 lymphocytes <200 (cells/mm3) | 18 (54·5%) | 15 (45·5%) | n.s. |
| CD8 lymphocytes (cells/mm3) | 944·3 (±690·2) | 796 (±456·4) | n.s. |
| CD8 lymphocytes (%) | 58·78 (±14·1) | 52·18 (±15·5) | n.s. |
| Cutaneous anergy | 14 (46·6%) | 8 (24·2%) | n.s. |
| IgG (mg/dL) | 2137·3 (±821·3) | 1575·3 (±464·1) | <0·01 |
| IgG1 (mg/dL) | 1710·4 (±754·3) | 1223·2 (±418·7) | <0·01 |
| IgG2 (mg/dL) | 278·8 (±188·7) | 208·3 (±853) | n.s. |
| IgG3 (mg/dL) | 86 (±68·6) | 60·5 (±49·9) | n.s. |
| IgG4 (mg/dL) | 40·7 (±34·7) | 36·3 (±21·7) | n.s. |
| IgM (mg/dL) | 222·8 (±192·8) | 159·5 (±124·9) | n.s. |
| IgA (mg/dL) | 290·8 (±195·7) | 340·3 (±157·9) | n.s. |
| Beta2-microglobulin (µg/l) | 3237·9 (±1857·8) | 2123·1 (±612) | <0·01 |
| C3 (mg/dL) | 125·6 (±35·4) | 126·2 (±50·5) | n.s. |
| C4 (mg/dL) | 32 (±13·5) | 32·3 (±20·4) | n.s. |
| CH50 (arbitrary haemolytic units) | 129·9 (±73·2) | 142·7 (±80·8) | n.s. |
The serum IgG basal value was significantly higher in the cases group compared with HIV-1-infected control patients (P < 0·01) (mainly at the expense of IgG1 subclass), as well as mean beta2-microglobulin levels (P < 0·01). The other two serum immunoglobulins (IgM and IgA), complement factors and CH50 mean levels did not show significant differences between the HIV-1 seropositive groups.
Antibody response to H. influenzae type b vaccine
The basal specific IgG mean value was greater in the cases group compared with the HIV-1 seropositive and healthy controls, although only on comparing between both HIV-1 seropositive groups did the differences show statistical significance (P < 0·05). After vaccination, the relative increase of specific IgG antibodies was greater in the healthy group (P < 0·001) and in HIV-1-infected controls (P < 0·001). Furthermore, eight HIV-1 seropositive controls had a threefold increase in their basal specific IgG value, while in only one of the cases, odds ratio 0·11 (IC95% 0·00–0·11) (P < 0·05) (see Table 2).
Table 2.
Pre- and post-vaccine determinations: mean specific antibodies increases and avidities of IgG and IgA against the PRP. Also shown are the percentage of bacteria (Haemophilus influenzae type b) opsonophagocitosed by PMN in the three groups studied
| Variable | Cases (n = 33) | HIV-1 positive controls (n = 33) | P* | Healthy controls (n = 27) | P† |
|---|---|---|---|---|---|
| Antibodies | |||||
| IgG (µg/ml) | |||||
| Prevaccine | 1·88 (±4·52) | 1·22 (±3·99) | 1·46 (±4·39) | ||
| Post-vaccine | 2·82 (±6·71) | 2·88 (±6) | 5·86 (±6·87) | ||
| Mean increase | 1·59 (0·56–14·99) | 3·94 (0·83–17·88) | <0·001 | 13·30 (0·90–88·7) | <0·001 |
| IgG2 (U/ml) | |||||
| Prevaccine | 39·9 (±84·5) | 36·5 (±99·9) | 107·9 (±105·9) | ||
| Post-vaccine | 46·2 (±85·1) | 67·3 (±140·5) | 166·9 (±113·6) | ||
| Mean increase | 1·43 (0·5–5) | 2·76 (0·5–13·25) | 0·05 | 2·64 (0·68–15) | 0·05 |
| IgA (U/ml) | |||||
| Prevaccine | 67·7 (±90·7) | 85·7 (±122·7) | 108·2 (±151·5) | ||
| Post-vaccine | 92·8 (±121·4) | 169·1 (±185·3) | 476·4 (±281·8) | ||
| Mean increase | 1·62 (0·5–8·17) | 8·75 (0·46–154) | <0·01 | 51·11 (0·73–623) | <0·001 |
| IgG avidity | |||||
| Prevaccine | 1·47 (1–3) | 1·24 (1–3) | n.s. | 1·44 (1–2) | n.s. |
| Post-vaccine | 1·52 (1–3) | 1·33 (1–2) | n.s. | 1·93 (1–4) | n.s. |
| IgA avidity | |||||
| Prevaccine | 1·87 (1–4) | 1·82 (1–4) | n.s. | 1·18 (1–2) | n.s. |
| Post-vaccine | 1·69 (1–3) | 1·67 (1–4) | n.s. | 1·41 (1–2) | n.s. |
| Opsonophagocytosis | |||||
| H. influenzae b | |||||
| Prevaccine | 27·8 (±11·6)% | 34·3 (±13·8)% | n.s. | 27·6 (±12·1)% | n.s. |
| Post-vaccine | 25·5 (±15·4)% | 31 (±11·8)% | n.s. | 34·2 (±8·9)% | n.s.‡ |
Cases versus HIV-1 seropositive controls.
Seronegative controls versus all HIV-1-infected patients considered as a whole group.
Statistical significance (P < 0·05) only between HIV-1 seropositive controls vs. seronegative subjects.
Specific IgG2 basal mean levels and relative increase after vaccination were higher in healthy subjects compared with both seropositive groups (P = 0·05). On comparing both HIV-1-infected groups, a tendency towards a larger specific IgG2 antibody increase in control subjects was observed (P = 0·05) (Table 2).
No statistical differences were observed between the three groups in specific IgA basal mean levels. The relative increase in the level of IgA antibodies against H. influenzae type b vaccine was larger in healthy subjects (P < 0·001), and that increase was also higher in the control group when compared to HIV-1 seropositive patients (P < 0·01) (Table 2). Ten HIV-1 seropositive control subjects showed a threefold increase in their basal specific IgA value compared with three in the cases group, odds ratio 0·27 (IC95% 0·05–1·24) (P = 0·05).
The values of IgG and IgA avidity in the cases, HIV-1 seropositive and healthy controls are described in Table 2. A tendency to a higher increase of avidity for both specific antibodies, IgG and IgA, was only observed in healthy controls.
Antibody response to pneumococcal vaccine
There were no statistical differences between the cases, HIV-1 seropositive and healthy controls in the mean basal specific IgG titres. However, IgG increase after vaccination was larger in the two control groups, both HIV-1 seropositive (P < 0·05) and healthy (P < 0·01) (Table 3). Fifteen patients in the cases group doubled specific IgG basal mean titres against 21 in the HIV-1 seropositive control group (no statistical differences).
Table 3.
Prevaccine and post-vaccine determinations: specific antibodies mean relative increases and avidities of IgG and IgA against the pneumococcal vaccine. Also shown is the percentage of bacteria (Streptococcus pneumoniae serotype 3) opsonophagocited by PMN in the three groups of individuals studied
| Variable | Cases (n = 33) | HIV-1 positive controls (n = 33) | P* | Healthy controls (n = 27) | P† |
|---|---|---|---|---|---|
| Antibodies | |||||
| IgG (U/ml) | |||||
| Prevaccine | 409 (±445·2) | 490·3 (±1002·3) | 785·2 (±792) | ||
| Post-vaccine | 873·7 (±989·8) | 1265·1 (±1516·1) | 2501·5 (±1363·7) | ||
| Mean increase | 2·29 (0·73–8·73) | 4·71 (0·25–22·55) | <0·05 | 5·72 (1·02–26·73) | <0·01 |
| IgG2 (U/ml) | |||||
| Prevaccine | 89 (±210·4) | 30·6 (±37·8) | 81·5 (±90·7) | ||
| Post-vaccine | 229·8 (±353·9) | 193·1 (±351·1) | 399·6 (±307·9) | ||
| Mean increase | 7·68 (0·71–95·86) | 7·94 (0·60–86·09) | n.s. | 11·61 (1·15–67·29) | <0·001 |
| IgA (U/ml) | |||||
| Prevaccine | 143·3 (±375·4) | 62·6 (±54·9) | 55·9 (±46·4) | ||
| Post-vaccine | 197 (±430·9) | 211·5 (±142·3) | 521·9 (±402·3) | ||
| Mean increase | 5·12 (0·61–74·5) | 4·71 (0·62–19·8) | <0·05 | 18·27 (1·09–153·7) | <0·001 |
| IgG avidity | |||||
| Prevaccine | 1·4 (1–2) | 1·36 (1–2) | n.s. | 1·25 (1–2) | n.s. |
| Post-vaccine | 1·51 (1–3) | 1·51 (1–3) | n.s. | 1·66 (1–3) | n.s. |
| IgA avidity | |||||
| Prevaccine | 1·9 (1–3) | 1·69 (1–4) | n.s. | 1·85 (1–3) | n.s. |
| Post-vaccine | 1·79 (1–3) | 1·63 (1–3) | n.s. | 1·88 (1–3) | n.s. |
| Opsonophagocytosis | |||||
| S. pneumoniae 3 | |||||
| Prevaccine | 21·5 (±13·8)% | 18·4 (±13·9)% | n.s. | 19·8 (±14·2)% | n.s. |
| Post-vaccine | 23·7 (±15·4)% | 25·4 (±11·8)% | n.s. | 32·4 (±14·6)% | n.s. |
Cases versus HIV-1 seropositive controls.
Seronegative controls versus all HIV-1-infected patients considered as a whole group.
With regard to the specific IgG2 post-vaccination relative increase, this was higher in the healthy control group (P < 0·001), and no differences were observed between HIV-1 seropositive patients (Table 3).
The relative increase of IgA was also greater in HIV-1 seropositive (P < 0·05) and healthy controls (P < 0·001), with no differences observed in basal mean values between the three groups (Table 3). Between both HIV-1 seropositive patients, 9 cases and 23 controls doubled their initial IgA titre, odds ratio 0·21 (IC95% 0·06–0·69) (P < 0·005).
As observed for the H. influenzae type b vaccine, a tendency to a larger increase in postvaccine avidity for two tested antibodies, IgG and IgA, in healthy subjects compared with the two HIV-1 seropositive groups was observed (Table 3).
Opsonophagocytosis
Statistical differences were not observed in the percentage of opsonophagocytosed bacteria between the three groups in the prevaccine determination. A larger increase in the percentage of opsonophagocytos bacteria after H. influenzae type b and pneumococcal vaccination was observed in the healthy individuals, although this reached statistical significance only on comparing HIV-1 seropositive controls with seronegative ones for H. influenzae (P < 0·05) (Table 3). Healthy subjects showed a positive linear correlation between their post-vaccination specific IgG2 mean levels against H. influenzae type b (r = 0·39, P < 0·05) and pneumococcal (r = 0·46, P < 0·01) vaccines and the increase of opsonophagocytosed capacity. On the other hand, the cases and HIV-1 seropositive controls showed a positive linear correlation with post-vaccination level of antipneumococcal specific IgA and increase of opsonophagocytic capacity (r = 0·45, P < 0·05). Table 4 shows the distribution of IgG2m and FcγRIIa allotypes in the three groups of patients studied, with no statistical differences observed on comparing each of the HIV-1 seropositive patient groups (cases and controls) independently or if considered as a whole group versus healthy controls.
Table 4.
IgG2m (n) and FcγRIIa (CD32) allotypes among the three studied groups (no statistical differences were observed)
| Allotypes | Cases n = 33 (%) | HIV-1 positive controls n = 33 (%) | Healthy controls n = 27 (%) | P |
|---|---|---|---|---|
| IgG2m | ||||
| n+/n+ | 12 (40) | 7 (21·2) | 8 (29·6) | n.s. |
| n+/n- | 11 (36·7) | 15 (45·5) | 15 (55·6) | n.s. |
| n-/n- | 7 (23·3) | 11 (33·3) | 4 (14·8) | n.s. |
| FcγRIIa (CD32) | ||||
| H/H | 8 (26·7) | 9 (27·3) | 9 (33·3) | n.s. |
| H/R | 18 (60) | 20 (60·6) | 11 (40·7) | n.s. |
| R/R | 4 (13·3) | 4 (12·1) | 7 (25·9) | n.s. |
Multivariate analysis
The history of injection-drug use was the only variable related to the risk of developing previous bacterial sinusitis or pneumonia episodes in HIV-1 seropositive patients, odds ratio 4·88 (IC95% 1·6–14·1) (P < 0·01). The logistic regression model selected none of the immunological variables.
DISCUSSION
Since the generalization of HAART, many reports have disclosed an important decline in opportunistic infections in HIV-1-infected patients [24–27]. However, bacterial respiratory tract infections, principally bacterial pneumonia, continue to be one of the first reasons for consultation and hospital admission in this group of immunosuppressed patients [26–28]. Moreover, as observed in clinical practice, a subgroup of especially predisposed patients seems to exist and sometimes these infections acquire a recurrent nature. The reasons for HIV-1-infected patients’ predisposition to bacterial respiratory tract infections are not completely understood. Studies conducted in the pre-HAART era analysed some of the risk factors related to the acquisition of these infections in HIV-1-infected individuals and concluded that cellular immunodepression, the main characteristic of the illness, plays a decisive role. In a multi-centre study, incidences of 2·3 and 10·8 episodes of bacterial pneumonia per person/year were observed in patient groups with greater than 500 and less than 200 CD4-cells/mm3, respectively [43]. Seropositive injection-drug users with a previous episode of bacterial pneumonia and a CD4-cell count of less than 200 cells/mm3 had almost a seven times higher risk of developing a new episode [44]. Two sinusitis studies in HIV-1-infected patients corroborated that a CD4-cell count of less than 200/mm3 was related with the recurrence, persistence and chronic nature of the infection [45,46]. In our series, 50% of HIV-1-infected patients had a CD4-cell count of less than 200/mm3 in agreement with the cited reports. However, the CD4-cell count alone does not explain why some patients with a well-preserved cellular immunity can suffer these infections and, on the other hand, others with a high degree of cellular impairment have never developed previous bacterial sinusitis or pneumonia episodes. Therefore, there must be other associated factors influencing the risk of developing these infections.
The present study, in concordance with other observations during the HIV-1 epidemic, discloses how drug abuse is the main factor related to the risk of presenting with recurrent bacterial respiratory infections [29–32]. Tobacco and cannabis smoking were also associated in univariate analysis with previous bacterial respiratory tract infections. Although these two substances can induce mucociliary clearance abnormalities and probably systemic immunological disturbances, especially related to antibody secretion, their individual contribution to the respiratory infections in the case group is difficult to determine, because these habits normally co-exist in parenteral drug-users. However, we think that the most remarkable finding in our study is the impairment in the antibody response against pneumococcal and H. influenzae type b antigens in the group of patients with previous respiratory bacterial infections, with no differences observed in avidity and opsonophagocytic capacity. To our knowledge, this has not been described before and poses some interesting questions. First, are these humoral disorders primary (genetically based) or secondary? The latter hypothesis seems the more probable due to the fact that the allotype distribution observed in both HIV-1 seropositive groups is in concordance with healthy controls. Secondly, are the impaired antibody responses demonstrated here a direct effect on the immune system of the illicit substances used by these patients or a consequence of the poor virological control shown by this group? Thirdly, are these disorders reversible with effective antiretroviral treatment? Some studies have reported a direct effect of opiates on the immune system but with inconclusive and sometimes conflicting results [47–49]. There are reports indicating that oligoclonality appears to be a general characteristic of human pneumococcal polysaccharide-specific antibody repertoires, and we suggest that a deletion of polysaccharide responsive B cell clones may be a direct harmful effect of pneumococcal polysaccharides leading to a poor response in the cases group of patients [50]. It is possible that specific B cell clones may have been deleted by recurrent infections, leading to a poor vaccine response in the cases group. This fact would be in accordance with a recent study paradoxically showing increased rates of pneumococcal disease in vaccine HIV-1-infected recipients [51]. On the other hand, monotherapy with zidovudine has been shown to have a positive effect on specific antibody production against pneumococcal vaccine [52], which supports the importance of virological control on the boosting of humoral immune responses in these patients, which would probably be stronger with HAART due to their higher potency.
The chronic immune stimulation promoted by repetitive infections and multiple intravenous antigen exposure existing in this group may be the cause of the diminished antibody response observed in this study. The latter is admissible if the rise in serum IgG and beta2-microglobulin displayed by the patients with previous recurrent bacterial infections, in comparison with the HIV-1 control group, is taken into account.
Another important matter to clarify is whether the recurrent nature of these bacterial infections is a consequence of the observed impaired specific IgA responses (main immunoglobulin present on the mucous membrane surfaces). It is known that the molecular structure of the serum and secretory IgA are different, and that a linear correlation between the levels of both immunoglobulins does not necessarily exist. However, our results could express the impossibility of erradicating nasal and oropharingeal colonization by certain microorganisms, which is generally considered to be the source of bacterial respiratory tract infections. It would be interesting to study antigen-specific IgA reponses on mucous membrane surfaces of these patients. The role of IgA in the control of respiratory pathogens is poorly understood and its opsonophagocytic capacity discussed, although a recent report supports the relevance of this immunoglobulin [53]. Interestingly, the opsonophagocytic capacity of all our HIV-1-infected patients correlated only with post-vaccination-specific IgA levels to pneumococcal vaccine in contrast with the healthy individuals that correlated, as predicted, with IgG2 specific levels.
Finally, the avidity of specific antibodies was equal in both HIV-1-infected patient groups compared with healthy controls, and therefore appears not to have any influence on the predisposition of developing recurrent bacterial respiratory infections.
Among the limitations of this study it is necessary to emphasize the low number of patients included. However, our main objective was to detect the factors related to the risk of developing bacterial respiratory infections in a limited group of HIV-1-infected subjects that present these infections with a recurrent course and not to achieve a vaccination efficacy study. Therefore, the inclusion criteria were especially restrictive, as shown by the high bacterial respiratory morbidity presented in the cases group before they were studied. Furthermore, pneumococcal and H. influenzae type b vaccinations were used only as a method for measuring the humoral immunological response against polysaccharide antigens. Finally, we have tested the opsonophagocytic capacity only for S. pneumoniae serotype 3, one of the most frequent agents responsible for infections in adult patients in Spain [38], and H. influenzae type b, although non-encapsulated bacteria are more frequent in this population of HIV-1-infected adults. Therefore, the observed results are not necessarily comparable to the responses induced by other pneumococcal serotypes or non-encapsulated H. influenzae isolations.
More studies are needed to answer some of the questions that have been discussed previously because of the clinical practice consequences arising from our results. Although, as commented above, it is not a vaccination efficacy study, it does suggest the possibility that vaccination strategies, mainly the recommended pneumococcal vaccine, could not obtain the expected results in certain patients affected by recurrent bacterial respiratory morbidity. This seems particularly true if they continue taking illicit drugs and a good HIV-1 virological control is not achieved. It also discloses the importance of improving the adherence to antiretroviral therapies in this population, in order to effectively prevent these bacterial infections. Any other policy would be unsuccessful.
REFERENCES
- 1.Stover DE, White DA, Romano PA, Gellene RA, Robeson WA. Spectrum of pulmonary diseases associated with the acquired immune deficiency syndrome. Am J Med. 1985;78:429–37. doi: 10.1016/0002-9343(85)90334-1. [DOI] [PubMed] [Google Scholar]
- 2.Polsky B, Gold JWM, Whimbey E, et al. Bacterial pneumonia in patients with the acquired immunodeficiency syndrome. Ann Intern Med. 1986;104:38–41. doi: 10.7326/0003-4819-104-1-38. [DOI] [PubMed] [Google Scholar]
- 3.Schlamm HT, Yancovitz SR. Haemophilus influenzae pneumonia in young adults with AIDS, ARC, or risk of AIDS. Am J Med. 1989;86:11–4. doi: 10.1016/0002-9343(89)90222-2. [DOI] [PubMed] [Google Scholar]
- 4.Janoff EN, Breiman RF, Daley CL, Hopewell PC. Pneumococcal disease during HIV infection. Ann Intern Med. 1992;117:314–24. doi: 10.7326/0003-4819-117-4-314. [DOI] [PubMed] [Google Scholar]
- 5.Falcó V, Fernandez de Sevilla T, Alegre J, et al. Bacterial pneumonia in HIV-infected patients: a prospective study of 68 episodes. Eur Respir J. 1994;7:235–9. doi: 10.1183/09031936.94.07020235. [DOI] [PubMed] [Google Scholar]
- 6.Baril L, Astagneau P, Nguyen J, et al. Pyogenic bacterial pneumonia in human immunodeficiency virus-infected inpatients: a clinical, radiological microbiological and epidemiological study. Clin Infect Dis. 1998;26:964–71. doi: 10.1086/513944. [DOI] [PubMed] [Google Scholar]
- 7.Jansen WTM, Väkeväinen-Anttila M, Käyhty H, et al. Comparison of a classical phagocytosis assay and a flow cytometry assay for assessment of the phagocytic capacity of sera from adults vaccinated with pneumococcal conjugate vaccine. Clin Diagn Laboratory Immunol. 2001;8:245–50. doi: 10.1128/CDLI.8.2.245-250.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Väkeväinen M, Jansen WTM, Saeland E, et al. Are the opsonophagocytic activities of antibodies in infant sera measured by different pneumococcal phagocytosis assays comparable? Clin Diagn Laboratory Immunol. 2001;8:363–9. doi: 10.1128/CDLI.8.2.363-369.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Anttila M, Voutilainen M, Jäntti V, Eskola J, Käyhty H. Contribution of serotype specific IgG concentration, IgG subclasses and relative antibody avidity to opsonophagocytic activity against Streptococcus pneumoniae. Clin Exp Immunol. 1999;118:402–7. doi: 10.1046/j.1365-2249.1999.01077.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Vernacchio L, Romero-Steiner S, Martinez JE, et al. Comparison of an opsonophagocytic assay and IgG ELISA to assess responses to pneumococcal polysaccharide and pneumococcal conjugate vaccines in children and young adults with sickle cell disease. J Infect Dis. 2000;181:1162–6. doi: 10.1086/315307. [DOI] [PubMed] [Google Scholar]
- 11.Rodriguez ME, van der Pol VL, Sanders LAM, van de Winkel JGJ. Crucial role of FcγRIIa (CD32) in assessment of functional anti-Streptococcus pneumoniae antibody activity in human serum. J Infect Dis. 1999;179:423–33. doi: 10.1086/314603. [DOI] [PubMed] [Google Scholar]
- 12.Jansen WTM, Breukels MA, Snippe H, Sanders AM, Verheul AFM, Rijkers GT. Fcγ receptor polymorphisms determine the magnitude of in vitro phagocytosis of Streptococcus pneumoniae mediated by pneumococcal conjugate sera. J Infect Dis. 1999;180:888–91. doi: 10.1086/314920. [DOI] [PubMed] [Google Scholar]
- 13.Lane HC, Masur H, Edgar LC, Whalen G, Rook AH, Fauci AS. Abnormalities of B-cell activation and immuno-regulation in patients with the acquired immunodeficiency syndrome. N Engl J Med. 1983;309:453–8. doi: 10.1056/NEJM198308253090803. [DOI] [PubMed] [Google Scholar]
- 14.Aucouturier P, Couderc LJ, Gouet D, et al. Serum immunoglobulin G subclass dysbalances in the lymphadenopathy syndrome and acquired immune deficiency syndrome. Clin Exp Immunol. 1986;63:234–40. [PMC free article] [PubMed] [Google Scholar]
- 15.Parkin JM, Helbert M, Hughes CL, Pinching AJ. Immunoglobulin G subclass deficiency and susceptibility to pyogenic infections in patients with AIDS-related complex and AIDS. AIDS. 1989;3:37–9. [PubMed] [Google Scholar]
- 16.Janoff EN, Douglas JMJ, Gabriel M, et al. Class-specific antibody response to pneumococcal capsular polysaccharides in men infected with human immunodeficiency virus type 1. J Infect Dis. 1988;158:983–90. doi: 10.1093/infdis/158.5.983. [DOI] [PubMed] [Google Scholar]
- 17.Pos O, Stevenhagen A, Meenhorst PL, Kroon FP, Van Furth R. Impaired phagocytosis of Staphylococcus aureus by granulocytes and monocytes of AIDS patients. Clin Exp Immunol. 1992;88:23–8. doi: 10.1111/j.1365-2249.1992.tb03033.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Bandres JC, Trial J, Musher DM, Rossen RD. Increased phagocytosis and generation of reactive oxygen products by neutrophils and monocytes of men with stage 1 human immunodeficiency virus infection. J Infect Dis. 1993;168:75–83. doi: 10.1093/infdis/168.1.75. [DOI] [PubMed] [Google Scholar]
- 19.Gabrilovich DI, Ivanova LA, Serebrovskaya LV, Shepeleva G, Pokrovsky V. Clinical significance of neutrophil functional activity in HIV infection. Scand J Infect Dis. 1994;26:41–7. doi: 10.3109/00365549409008589. [DOI] [PubMed] [Google Scholar]
- 20.Lane HC, Depper JM, Green WC. Qualitative analysis of immune function in patients with the acquired immunodeficiency syndrome: evidence for a selective defect in soluble antigen recognition. N Engl J Med. 1985;313:79–84. doi: 10.1056/NEJM198507113130204. [DOI] [PubMed] [Google Scholar]
- 21.Ballet JJ, Sulcebe G, Couderc LJ, et al. Impaired anti-pneumococcal antibody response in patients with AIDS-related persistent generalized lymphadenopathy. Clin Exp Immunol. 1987;68:479–87. [PMC free article] [PubMed] [Google Scholar]
- 22.Simberkoff MS, El Sadr W, Schiffman G, Rahal JJ., Jr Streptococcus pneumoniae infections and bacteremia in patients with acquired immune deficiency syndrome, with report of a pneumococcal vaccine failure. Am Rev Respir Dis. 1984;130:1174–6. doi: 10.1164/arrd.1984.130.6.1174. [DOI] [PubMed] [Google Scholar]
- 23.French N, Gilks CF, Mujugira A, Fasching C, O'Brien J, Janoff EN. Pneumococcal vaccination in HIV-1-infected adults in Uganda. Humoral response and two vaccine failures. AIDS. 1998;12:1683–9. doi: 10.1097/00002030-199813000-00017. [DOI] [PubMed] [Google Scholar]
- 24.Palella F, Delaney K, Moorman A, et al. Declining morbidity and mortality among HIV-infected patients with advanced HIV-infection. HIV outpatient study investigators. N Engl J Med. 1998;338:853–60. doi: 10.1056/NEJM199803263381301. [DOI] [PubMed] [Google Scholar]
- 25.Brodt H, Kamps B, Gute P, Knupp B, Staszewski S, Helm E. Changing incidence of AIDS-defining illnesses in the era of antiretroviral combination therapy. AIDS. 1997;11:1731–8. doi: 10.1097/00002030-199714000-00010. [DOI] [PubMed] [Google Scholar]
- 26.Paul S, Gilbert HM, Ziecheck W, Jacobs J, Sepkowitz KA. The impact of potent antiretroviral therapy on the characteristics of hospitalised patients with HIV infection. AIDS. 1999;13:414–8. doi: 10.1097/00002030-199902250-00015. [DOI] [PubMed] [Google Scholar]
- 27.Selik RM. Trends in fatal infectious diseases in persons dying of HIV-infection in the era of HAART; 7th Conference on Retroviruses and Opportunistic Infections, San Francisco; 30 January–2 February 2000 [Abstract 462] [Google Scholar]
- 28.Sullivan JH, Moore RD, Keruly JC, Chaisson RE. Effect of antiretroviral therapy on the incidence of bacterial pneumonia in patients with advanced HIV infection. Am J Respir Crit Care Med. 2000;162:64–7. doi: 10.1164/ajrccm.162.1.9904101. [DOI] [PubMed] [Google Scholar]
- 29.Selwyn PA, Feingold AR, Hartel D, et al. Increased risk of bacterial pneumonia in HIV-infected intravenous drug users without AIDS. AIDS. 1988;2:267–72. doi: 10.1097/00002030-198808000-00005. [DOI] [PubMed] [Google Scholar]
- 30.Tumbarello M, Tacconelli E, de Gaetano K, et al. Bacterial pneumonia in HIV-infected patients. Analysis of risk factors and prognostic indicators. J Acquir Immune Defic Syndr Hum Retrovirol. 1998;18:39–45. doi: 10.1097/00042560-199805010-00006. [DOI] [PubMed] [Google Scholar]
- 31.Mientjes GH, van Ameijden EJ, van den Hoek AJ, Coutinho RA. Increasing morbidity without rise in non-AIDS mortality among HIV-infected intravenous drug users in Amsterdam. AIDS. 1992;6:207–12. doi: 10.1097/00002030-199202000-00012. [DOI] [PubMed] [Google Scholar]
- 32.Boschini A, Smacchia C, Di Fine M, et al. Community-acquired pneumonia in a cohort of former injection drug users with and without human immunodeficiency virus infection: incidence, etiologies, and clinical aspects. Clin Infect Dis. 1996;23:107–13. doi: 10.1093/clinids/23.1.107. [DOI] [PubMed] [Google Scholar]
- 33.Metzger WJ, Butler JE, Swanson P, Reinders E, Richerson HB. Amplification of the enzyme-linked immunosorbent assay for measuring allergen-specific IgE and IgG antibody. Clin Allergy. 1981;11:523–31. doi: 10.1111/j.1365-2222.1981.tb02170.x. [DOI] [PubMed] [Google Scholar]
- 34.Siber GR, Priehs C, Madore DW. Standardisation of antibody assays for measuring response to pneumococcal infection and immunisation. Pediatr Infect Dis J. 1989;8(suppl.):84–91. [PubMed] [Google Scholar]
- 35.Rodrigo MJ, Miravitlles M, Cruz MJ, et al. Characterization of specific immunoglobulin G (IgG) and its subclasses (IgG1 and IgG2) against the 23-valent pneumococcal vaccine in a healthy adult population: proposal for response criteria. Clin Diagn Laboratory Immunol. 1997;4:168–72. doi: 10.1128/cdli.4.2.168-172.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Pullen GR, Fitzgerald MG, Hosking CS. Antibody avidity determination by ELISA using thiocyanate elution. J Immunol Meth. 1986;86:83–7. doi: 10.1016/0022-1759(86)90268-1. [DOI] [PubMed] [Google Scholar]
- 37.Martin E, Bhakdi S. Flow cytometric assay for quantifying opsonophagocytosis and killing of Staphylococcus aureus by peripheral blood leukocytes. J Clin Microbiol. 1992;30:2246–55. doi: 10.1128/jcm.30.9.2246-2255.1992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Fenoll A, Jado I, Vicioso D, Pérez A, Casal J. Evolution of Streptococcus pneumoniae serotypes and antibiotic resistance in Spain: update (1990–96) J Clin Microbiol. 1998;36:3447–54. doi: 10.1128/jcm.36.12.3447-3454.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Sarvas H, Rautonen N, Sipinen S, Mäkelä O. IgG subclasses of pneumococcal antibodies. Effect of allotype G2m(n) Scand J Immunol. 1989;29:229–37. doi: 10.1111/j.1365-3083.1989.tb01120.x. [DOI] [PubMed] [Google Scholar]
- 40.Carlsson LE, Santoso S, Baurichter G, et al. Heparin-induced thrombocytopenia. new insights into the impact of the FcγRIIa-R-H131 polymorphism. Blood. 1998;92:1526–31. [PubMed] [Google Scholar]
- 41.Mascart-Lemone F, Gerard M, Libin M, et al. Differential effect of human immunodeficiency virus infection on the IgA and IgG antibody responses to pneumococcal vaccine. J Infect Dis. 1995;172:1253–60. doi: 10.1093/infdis/172.5.1253. [DOI] [PubMed] [Google Scholar]
- 42.Kroon FP, van Dissel JT, Rijkers GT, Labadie J, van Furth R. Antibody response to Haemophilus influenzae type b vaccine in relation to the number of CD4+ T lymphocytes in adults infected with human immunodeficiency virus. Clin Infect Dis. 1997;25:600–6. doi: 10.1086/513750. [DOI] [PubMed] [Google Scholar]
- 43.Hirschtick RE, Glassroth J, Jordan MC, et al. Bacterial pneumonia in persons infected with the human immunodeficiency virus. N Engl J Med. 1995;333:845–51. doi: 10.1056/NEJM199509283331305. [DOI] [PubMed] [Google Scholar]
- 44.Caiaffa WT, Vlahov D, Graham NM, et al. Drug smoking, Pneumocystis carinii pneumonia, and immunosupression increase risk of bacterial pneumonia in human immunodeficiency virus-seropositive injection drug users. Am J Respir Crit Care Med. 1994;150:1493–8. doi: 10.1164/ajrccm.150.6.7952605. [DOI] [PubMed] [Google Scholar]
- 45.Godofsky EW, Zinreich J, Armstrong M, Leslie JM, Weikel CS. Sinusitis in HIV-infected patients: a clinical and radiographic review. Am J Med. 1992;93:163–70. doi: 10.1016/0002-9343(92)90046-e. [DOI] [PubMed] [Google Scholar]
- 46.Zurlo JJ, Feuerstein IM, Lebovics R, Lane C. Sinusitis in HIV-1 infection. Am J Med. 1992;93:157–62. doi: 10.1016/0002-9343(92)90045-d. [DOI] [PubMed] [Google Scholar]
- 47.Eisenstein TK, Hilburger ME. Opioid modulation of immune responses: effects on phagocyte and lymphoid cell populations. J Neuroimmunol. 1998;83:36–44. doi: 10.1016/s0165-5728(97)00219-1. [DOI] [PubMed] [Google Scholar]
- 48.De Waal EJ, Van Der Laan JW, Van Loveren H. Effects of prolonged exposure to morphine and methadone on in vivo parameters of immune function in rats. Toxicology. 1998;129:201–10. doi: 10.1016/s0300-483x(98)00077-8. [DOI] [PubMed] [Google Scholar]
- 49.McLachlan C, Crofts N, Wodak A, Crowe S. The effects of methadone on immune function among injecting drug users: a review. Addiction. 1993;88:257–63. doi: 10.1111/j.1360-0443.1993.tb00809.x. [DOI] [PubMed] [Google Scholar]
- 50.Lucas AH, Granoff DM, Mandrell RE, Connolly CC, Shan AS, Powers DC. Oligoclonality of serum immunoglobulin G antibody responses to Streptococcus pneumoniae capsular polysaccharide serotypes 6B, 14 and 23F. Infect Immun. 1997;65:5103–9. doi: 10.1128/iai.65.12.5103-5109.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.French N, Nakiyingi J, Carpenter LM, et al. 23-valent pneumococcal polysaccharide vaccine in HIV-1 infected Ugandan adults: double-blind, randomised and placebo controlled trial. Lancet. 2000;355:2106–11. doi: 10.1016/s0140-6736(00)02377-1. [DOI] [PubMed] [Google Scholar]
- 52.Glaser JB, Volpe S, Aguirre A, Simpkins H, Schiffman G. Zidovudine improves response to pneumococcal vaccine among persons with AIDS and AIDS-related complex. J Infect Dis. 1991;164:761–4. doi: 10.1093/infdis/164.4.761. [DOI] [PubMed] [Google Scholar]
- 53.Janoff EN, Fasching C, Orenstein JM, Rubins JB, Opstad NL, Dalmasso AP. Killing of Streptococcus pneumoniae by capsular polysaccharide-specific polymeric IgA, complement, and phagocytes. J Clin Invest. 1999;104:1139–47. doi: 10.1172/JCI6310. [DOI] [PMC free article] [PubMed] [Google Scholar]