Immunoglobulin G subclass levels and antibody responses to the 2009 influenza A (H1N1) monovalent vaccine among human immunodeficiency virus (HIV)-infected and HIV-uninfected adults

Authors


N. Crum-Cianflone, c/o Clinical Investigation Department (KCA), Naval Medical CenterSan Diego, 34800 Bob Wilson Drive, Ste. 5, San Diego, CA 92134-1005, USA. E-mail: nancy.crum@med.navy.mil

Summary

Immunoglobulin (Ig)G levels are important for antibody vaccine responses and IgG subclass deficiencies have been associated with severe 2009 influenza A (H1N1) infections. Studies have demonstrated variations in immune responses to the H1N1 vaccine, but the aetiology of this is unknown. We determined the associations between pre-vaccination overall and influenza-specific IgG subclass levels and 2009 H1N1-specific antibody responses post-vaccination (robust versus poor at day 28) stratified by human immunodeficiency virus (HIV) status. Logistic regression models were utilized to evaluate whether pre-vaccination IgG subclass levels were associated with the antibody response generated post-vaccination. We evaluated 48 participants as part of a clinical study who were stratified by robust versus poor post-vaccination immune responses. Participants had a median age of 35 years; 92% were male and 44% were Caucasian. HIV-infected adults had a median CD4 count of 669 cells/mm3, and 79% were receiving highly active anti-retroviral therapy. HIV-infected participants were more likely to have IgG2 deficiency (<240 mg/dl) than HIV-uninfected individuals (62% versus 4%, P < 0·001). No association of pre-vaccination IgG subclass levels (total or influenza-specific) and the antibody response generated by HIN1 vaccination in either group was found. In summary, pre-vaccination IgG subclass levels did not correlate with the ability to develop robust antibody responses to the 2009 influenza A (H1N1) monovalent vaccine. IgG2 deficiencies were common among HIV-infected individuals but did not correlate with poor influenza vaccine responses. Further investigations into the aetiology of disparate vaccine responses are needed.

Introduction

Low immunoglobulin (Ig)G subclass levels, in particular IgG2 deficiency, have been associated with poor antibody responses to vaccination, especially among those with a history of severe or recurrent respiratory infections [1]. Patients with low IgG2 levels may have poor responses to Haemophilus influenzae type B (Hib), Streptococcus pneumoniae conjugate and polysaccharide vaccines, as well as tetanus toxoid; these findings have been noted among human immunodeficiency virus (HIV)-infected and HIV-uninfected people [1–6]. In addition to IgG2, other IgG subclasses (e.g. IgG1) may be important in mounting antibody responses to vaccination. Consequently, some have suggested that poor immune responses post-vaccination (e.g. Hib) may warrant assessment of IgG subclasses levels [2]. In addition, a recent study associating severe 2009 pandemic influenza disease with IgG2 subclass deficiency [7] lends further credence to a potential link between the magnitude of an immune response to infection or vaccination and immunoglobulin subclass levels.

We recently conducted a clinical study of the immunogenicity of the 2009 influenza A (H1N1) monovalent vaccine among HIV-infected and HIV-uninfected adults [8] and found large variations in the post-vaccination influenza antigen-specific antibody concentrations (measured by a standard haemagglutinin-inhibition test) among both study arms not explained fully by demographic data, influenza or vaccination history, or HIV-related factors. In light of these observations, we postulated that IgG subclass levels may be important in mounting immune responses to H1N1 vaccination. Accordingly, we evaluated the relationship between pre-vaccination IgG subclass levels (overall and influenza-specific) and the magnitude of influenza antigen-specific antibody responses generated to this novel vaccine.

Methods

We evaluated stored pre-vaccination serum specimens from participants who received the monovalent 2009 influenza A (H1N1) vaccine [strain A/California/7/2009(H1N1), Novartis, Liverpool, UK]. The vaccine manufacturer was not involved in the study in any capacity. Both HIV-infected and HIV-uninfected groups were enrolled simultaneously, and all participants were 18–50 years of age and without serious medical conditions, except for the diagnosis of HIV among the former group. The main study and this substudy were approved by a central military institutional review board, and the vaccine study was registered with the Clinical Trials network (registration NCT00996970).

The initial vaccine study enrolled 132 participants and demonstrated wide variations in antibody responses to the 2009 influenza A (H1N1) vaccine among both study arms [8]. Antibody levels to the 2009 influenza A (H1N1) virus were measured by haemagglutination-inhibition assay (HAI), as described previously [8]. Sera were tested in duplicate in two independent assays, with the geometric mean titre (GMT) reported as the final titre. For computational purposes, titres of <1:10 were assigned a value of 1:5 and those >1:1280 a value of 1:1280.

For this substudy, we evaluated pre-vaccination IgG subclass levels (overall and influenza-specific) among participants with the highest and lowest changes, from baseline (day 0) to day 28 post-vaccination, in antibody GMT for 2009 influenza A (H1N1). We examined four groups (each with n = 12) in this substudy: HIV-infected participants with poor antibody response (those with the smallest changes in GMT), HIV-infected with robust antibody response (those with the largest changes in GMT), HIV-uninfected with poor antibody response and HIV-uninfected with robust antibody response.

IgG total and subclass (IgG1, IgG2, IgG3 and IgG4) levels were performed using nephelometry on a Food and Drug Administration (FDA)-approved platform (Dade Behring Siemens BNII system) at Quest Diagnostics Nichols Institute (San Juan Capistrano, CA, USA). All tests were conducted simultaneously utilizing the same testing plates, and laboratory personnel were blinded to the participants' clinical information. The reference ranges for normal adults according to the manufacturer were 694–1618 for the total IgG, 382–929 for IgG1, 241–700 for IgG2, 22–178 for IgG3 and 4–86 mg/dl for IgG4. For this report, an IgG subclass deficiency was defined as <700 for total, <380 for IgG1, <240 for IgG2, <22 for IgG3 and <4 for IgG4, similar to prior reports [7].

Additionally, influenza-specific IgG subclass levels were measured pre-vaccination and at day 28 post-vaccination using a quantitative subclass enzyme-linked immunosorbent assay (ELISA) [9]. Briefly, serum samples, diluted 1:3000 in blocking buffer, were bound in duplicate (in two independent experiments) to 96-well microtitre plates (Immulon 2HB) precoated with recombinant HA (ΔTM) (H1N1)-A/California/06/2009 (Immune Technology Corp., NY, USA) in ×1 phosphate-buffered saline (PBS). Captured anti-influenza antibody was detected using horseradish peroxidase-labelled anti-human IgG subclass antibody (The Binding Site, San Diego, CA, USA) at a 1:3000 dilution. Standard curves for subclasses were generated by capturing known quantities of IgG1, IgG2, IgG3 or IgG4 (Invitrogen, Grand Island, NY, USA) on plates coated with anti-human IgG (Invitrogen) and detecting them with horseradish peroxidase-conjugated anti-human IgG1, IgG2, IgG3 or IgG4 (The Binding Site).

Clinical data collected in the original H1N1 vaccine study included demographics, body mass index (BMI), number of seasonal influenza vaccinations within the 3 years prior to study enrolment, number of household members and history of self-reported influenza. For the HIV-infected group, receipt of highly active anti-retroviral therapy (HAART), current CD4 count and plasma HIV RNA level were recorded. Clinical events including influenza-like illnesses (ILI) and confirmed influenza infections were documented during a 6-month follow-up period.

Statistical methods included descriptive statistics shown as medians [interquartile ranges (IQR)] or counts (proportions). Unadjusted group comparisons (HIV-infected versus HIV-uninfected, and poor responders versus robust responders within each HIV group) utilized median tests and Fisher's exact tests. Logistic regression models on robust versus poor responders were performed to examine associations with IgG levels and participant characteristics. All P-values are two-sided. Analyses were conducted using r (version 2·10·1; R Development Core Team, Vienna, Austria) and sas version 9·2 (SAS Institute, Inc., Cary, NC, USA).

Results

We evaluated 48 participants with a median age of 35 years; 92% were male and 44% Caucasian (Table 1). The HIV-infected and HIV-uninfected groups were similar, except that the latter were more likely to have received all three seasonal influenza vaccinations during the past 3 years (P = 0·001). The HIV-infected arm had a median CD4 count of 669 cells/mm3 and 79% were receiving HAART. Pre-vaccination overall and influenza-specific IgG subclass levels are shown in Table 2. HIV-infected participants had lower median IgG2 levels compared to HIV-uninfected participants (207 versus 342 mg/dl, P < 0·001) and were more likely to have IgG2 deficiency (<240 mg/dl) detected (62% versus 4%, P < 0·001). There were no differences by HIV-status in baseline influenza-specific IgG levels (all P > 0·05).

Table 1.  Baseline demographic and clinical data, stratified according to human immunodeficiency virus (HIV) status and robust versus poor antibody responses to the 2009 influenza A (H1N1) vaccine.
Factor*HIV-uninfected adultsHIV-infected adults
Total group n = 24Robust antibody response n = 12Poor antibody response n = 12P-valueTotal group n = 24Robust antibody response n = 12Poor antibody response n = 12P-value
  1. *Reported as medians with interquartile ranges (IQR) or counts with proportions. P-values shown compare participants with a poor versus robust response in each arm. Seasonal influenza vaccinations in the past 3 years. HAART: highly active anti-retroviral therapy.

Demographics
Age, years34 (26–38)34 (26–38)28 (25–35)1·036 (25–45)32 (25–38)42 (25–47)0·11
Sex, male23 (96%)12 (100%)11 (92%)1·021 (88%)10 (83%)11 (92%)1·0
Ethnicity        
 Caucasian13 (54%)6 (50%)7 (58%)0·738 (33%)4 (33%)4 (33%)1·0
 African American4 (17%)3 (25%)1 (8%)9 (38%)4 (33%)5 (42%)
 Other7 (29%)3 (25%)4 (33%)7 (29%)4 (33%)3 (25%)
Clinical data
Body mass index, kg/m226 (23–29)27 (24–30)24 (22–28)0·2127 (23–29)24 (22–29)27 (26–29)0·39
Self-reported history of influenza3 (13%)1 (8%)2 (17%)1·05 (21%)4 (33%)1 (8%)0·32
Seasonal influenza vaccinations        
 10 (0%)0 (0%)0 (0%)0·483 (12%)2 (17%)1 (8%)0·74
 22 (8%)2 (17%)0 (0%)11 (46%)6 (50%)5 (42%)
 322 (92%)10 (83%)12 (100%)10 (42%)4 (33%)6 (50%)
Number of household members        
 05 (21%)4 (33%)1 (8%)0·3811 (46%)6 (50%)5 (42%)1·0
 14 (17%)2 (17%)2 (17%)5 (21%)2 (17%)3 (25%)
 ≥215 (62%)6 (50%)9 (75%)8 (33%)4 (33%)4 (33%)
HIV-specific data  
HIV infection duration, years8 (2–16)3 (1–9)12 (2–20)0·11
Current CD4 cell count, cells/mm3669 (480–876)605 (536–819)671 (403–876)0·42
Plasma HIV RNA level, <50 copies/ml12 (50%)5 (42%)7 (58%)0·68
Receipt of HAART19 (79%)10 (83%)9 (75%)1·0
Table 2.  Overall and influenza-specific serum immunoglobulin (Ig)G subclass levels, stratified according to human immunodeficiency virus (HIV) status and robust versus poor antibody responses to the 2009 influenza A (H1N1) vaccine.
Factor*HIV-uninfected adultsHIV-infected adults
Total group n = 24Robust antibody response n = 12Poor antibody response n = 12P-valueTotal group n = 24Robust antibody response n = 12Poor antibody response n = 12P-value
  1. *Reported as medians with interquartile ranges (IQR) or counts with proportions. P-values shown compare participants with a poor versus robust response in each arm.

Baseline overall IgG levels, mg/dl
Total IgG        
 Median1050 (890–1150)1060 (966–1130)933 (865–1180)0·421140 (926–1480)1210 (825–1480)1110 (946–1250)0·42
 No. with deficiency0 (0%)0 (0%)0 (0%)1·03 (12%)2 (17%)1 (8%)1·0
IgG1        
 Median609 (475–632)609 (543–623)550 (433–710)1·0781 (502–998)781 (483–998)781 (674–866)1·0
 No. with deficiency1 (4%)0 (0%)1 (8%)1·02 (8%)1 (8%)1 (8%)1·0
IgG2        
 Median342 (299–454)399 (286–507)337 (306–378)0·42207 (119–257)207 (116–267)193 (119–244)1·0
 No. with deficiency1 (4%)1 (8%)0 (0%)1·015 (62%)7 (58%)8 (67%)1·0
IgG3        
 Median69 (46–92)74 (54–100)50 (31–88)0·4286 (55–124)73 (55–112)86 (48–134)1·0
 No. with deficiency1 (4%)0 (0%)1 (8%)1·02 (8%)1 (8%)1 (8%)1·0
IgG4        
 Median34 (14–50)28 (9–62)34 (23–34)1·016 (7–39)15 (7–40)16 (4–38)1·0
 No. with deficiency0 (0%)0 (0%)0 (0%)1·03 (12%)1 (8%)2 (17%)1·0
Baseline influenza-specific IgG levels, median levels, mg/dl
Total IgG24 (15, 39)21 (7, 30)24 (15, 40)0·4229 (15, 60)26 (10, 57)37 (17, 63)0·42
IgG117 (13, 35)15 (5, 27)22 (14, 35)0·4226 (9, 53)24 (9, 52)26 (4, 60)1·0
IgG21 (0, 2)1 (0, 2)1 (0, 2)1·01 (0, 2)0 (0, 1)1 (0, 2)0·11
IgG33 (0, 6)1 (0, 5)3 (1, 6)0·422 (0, 4)2 (1, 4)2 (0, 5)1·0
IgG40 (0, 0)0 (0, 1)0 (0, 0)0·690 (0, 0)0 (0, 0)0 (0, 0)0·23
Change in influenza-specific IgG levels from baseline to day 28 post-vaccination, median levels, mg/dl
Total IgG40 (105, 31)531 (276, 794)10 (1, 21)<0·0018 (−1, 91)91 (35, 186)3 (−7, 6)0·001
IgG 137 (8, 529)529 (275, 794)8 (0, 16)<0·0018 (−1, 60)60 (35, 186)3 (−7, 6)0·001
IgG 20 (0, 1)0 (0, 1)0 (0, 1)0·420 (0, 0)0 (0, 0)0 (−2, 0)0·39
IgG 31 (0, 2)0 (0, 2)1 (0, 1)0·420 (0, 1)0 (0, 1)0 (0, 0)0·42
IgG 40 (0, 0)0 (0, 0)0 (0, 0)0·620 (0, 0)0 (0, 0)0 (0, 0)0·49

HIV-uninfected participants with a robust antibody response (median GMT change of 1246, IQR 1120–1274) were compared to those generating a poor response (median GMT change of 9, IQR 0–29). No demographic or clinical factors were associated with a more robust response, nor were pre-vaccination overall IgG subclass levels (Tables 1 and 2, Fig. 1a). We also examined pre-vaccination influenza-specific IgG among HIV-uninfected individuals and found no associations between any subclass levels with vaccine robust versus poor antibody responses (Table 2, Fig. 1b). The change (from days 0 to 28) in the total and the IgG1 influenza-specific levels were correlated with the change in 2009 H1N1 antibody GMT, although no relationships were noted for IgG2, IgG3 or IgG4.

Figure 1.

Serum immunoglobulin (Ig)G subclasses for (a) overall and (b) influenza-specific levels stratified according to HIV status and robust versus poor antibody response to the 2009 influenza A (H1N1) vaccine.

HIV-infected participants with robust (median GMT change 419, IQR 339–741) versus poor vaccine responses (median GMT change 0, IQR 0-0) were also compared. There was a marginal association for younger age (median 32 versus 42 years, P = 0·11) and shorter duration of HIV infection (median 3 versus 12 years, P = 0·11) being associated with more robust vaccine responses (Table 1). Overall IgG subclass levels were not associated with robust versus poor vaccine responses (Table 2, Fig. 1a). Regarding influenza-specific IgG levels among HIV-infected participants, there were no associations between any pre-vaccination subclass IgG level with post-vaccination robust versus poor antibody responses. Similar to the HIV-uninfected arm, larger increases in influenza-specific total and IgG1 levels were associated with a larger change in 2009 H1N1 antibody GMT, but there was no such association of IgG2 or other subclass levels with change in antibody level (Table 2, Fig. 1b).

Given prior research regarding the association of IgG2 deficiency with poor vaccine responses, and that the most common IgG subclass deficiency in our HIV-infected cohort was IgG2, we performed additional analyses of this group. Of note, of all HIV-infected patients who had at least one IgG subclass deficiency, all had an IgG2 deficiency. HIV-infected participants with an IgG2 deficiency (n = 15, 63%) compared to those without (n = 9, 37%) were similar, except that the former group had HIV infection for a longer duration (9 versus 4 years, P = 0·68) and were more likely to have CDC stage C disease (27% versus 0%, P = 0·26), although neither was statistically significant. Fifty-eight per cent of those with a robust post-vaccination response had a low overall IgG2 level compared to 67% of those with a poor response (P = 1·0). A poor vaccine response was not associated with a low overall IgG2 level (<240 mg/dl).

Discussion

We conducted this substudy to evaluate if pre-vaccination IgG subclass levels could explain the variability of antibody responses to influenza A (H1N1) vaccination. Furthermore, our study was prompted after IgG2 subclass deficiencies were found to be associated with severe 2009 influenza A (H1N1) virus infections [7], and data suggesting that patients with respiratory infections may generate poor vaccine responses [1]. Overall, our study did not find a significant relationship between pre-vaccination IgG subclass levels (either overall or influenza-specific) and antibody responses to the 2009 influenza A (H1N1) vaccine.

Our study represents the first study, to our knowledge, examining the potential role of IgG subclass levels and immune responses to the 2009 influenza A (H1N1) vaccine. Regarding IgG subclass deficiencies and influenza, most studies have focused on its role on the clinical disease course. Data from animal studies showed that IgG levels are important in virus control, particularly in preventing the development of pneumonia [10,11]. A single human study found a significant relationship between low IgG2 levels and increased severity of 2009 influenza A (H1N1) infections [7]. Although immune responses to natural infections may differ from those elicited by vaccinations, IgG subclass deficiencies have been linked to poor antibody responses to some vaccines [1–6]. From our study results, it appears that pre-vaccination subclass IgG levels do not explain the wide variations in immune responses to H1N1 influenza vaccination. Further, although our study cohort had a high prevalence of IgG2 subclass deficiency, no relationship was observed between low IgG2 levels and poor vaccine antibody responses.

We found that 62% of our HIV-infected cohort had low IgG2 levels. The prevalence of IgG2 deficiency among HIV patients in our study was significantly higher than that seen in the HIV-uninfected arm (4%) and in the general population (2–20%) [12]. Although HIV patients often have hypergammaglobulinaemia, IgG subclass deficiencies may be present [13]. Prior studies have also noted IgG2 deficiencies, most often among those with acquired immune deficiency syndrome (AIDS) [13,14]. Interestingly, we found a high prevalence of IgG2 deficiency in our HIV cohort despite robust CD4 counts (median 669 cells/mm3) and high anti-retroviral coverage. We examined if low IgG2 levels were associated with either prior self-reported influenza or influenza (H1N1) vaccine responses, but found no associations. In a prior study among AIDS patients, low IgG2 levels were associated with pyogenic infections, but this study did not evaluate influenza events [13]. The lack of association with IgG2 levels is due probably to the fact that the change in IgG1 (not IgG2) levels was most predictive of robust post-vaccination immune responses.

IgG subclass deficiencies in our study did not explain overall poor influenza vaccine responses among HIV-infected or HIV-uninfected individuals. The poor vaccine responses in the HIV arm may be due to persistent cellular and humoral immune dysfunction beyond that measured by immunoglobulins, the absolute CD4 cell count or HIV viraemia [15]. Similarly, despite studying a young, healthy HIV-uninfected population, some individuals did not mount appropriate antibody responses to vaccination. The exact mechanism of this impairment is unclear, but may be related in part to host genetics. Because poor vaccine responses may result in continued susceptibility to infection and potential spread within the community, we advocate for future studies to confirm our study findings and evaluate the causes of poor vaccine responses among individuals within various population groups.

Our study had potential limitations. First, it had a small sample size. We attempted to optimize statistical power by focusing on subsets with the greatest and smallest antibody responses to the vaccination. In addition, a power calculation (using an alpha of 0·05 and an estimated variance of 155 mg/dl [7]) demonstrated that our study could detect a true difference of −180 to 180 mg/dl for IgG2 with a power of 0·80; hence, our study would probably have detected a difference similar to that found in the prior publication [7], if such a difference existed. A second limitation is that we did not measure cellular immune responses; however, our study's objective was to evaluate antibody responses, and we examined both total pre-vaccination IgG levels as well as influenza-specific IgG levels pre- and post-vaccination. Thirdly, we attempted to determine the impact of IgG levels on clinical influenza events; however, none occurred during the 6-month follow-up period, due probably to the rapid decline in infections within the community. We evaluated the associations between IgG levels and history of influenza illness and subsequent ILI events, but found no associations in HIV-infected or HIV-uninfected groups.

In summary, pre-vaccination IgG subclass levels (overall or influenza-specific) were not associated with antibody responses to the 2009 influenza A (H1N1) vaccine in our study. Although IgG2 deficiency was common among HIV-infected individuals, the level of IgG2 did not predict immune responses to influenza vaccination. Given the significant morbidity and mortality from influenza infections each year, further investigations regarding the factors associated with severe influenza infections and poor vaccine responses are advocated.

Acknowledgements

Support for this work (IDCRP-053) was provided by the Infectious Disease Clinical Research Program (IDCRP), a Department of Defense (DoD) program executed through the Uniformed Services University of the Health Sciences. This project has been funded in whole, or in part, with federal funds from the National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH), under Inter-Agency Agreement Y1-AI-5072. In addition, funding was provided by the Armed Forces Health Surveillance Center's Global Emerging Infections System via project I204_10. The content of this publication is the sole responsibility of the authors and does not necessarily reflect the views or policies of the NIH or the Department of Health and Human Services, the US Government, the DoD or the Departments of the Army, Navy or Air Force. Mention of trade names, commercial products or organizations does not imply endorsement by the US Government. The authors acknowledge that research protocol (IDCRP-053, ‘Immunogenicity of Novel H1N1 Vaccination among HIV-Infected Compared to HIV-Uninfected Persons’) received applicable Institutional Review Board review and approval. We certify that all individuals who qualify as authors have been listed; each has participated in the conception and design of this work, the writing of the document, and the approval of the submission of this version; that the document represents valid work; that if we used information derived from another source, we obtained all necessary approvals to use it and made appropriate acknowledgements in the document; and that each takes public responsibility for it. Nothing in the presentation implies any Federal/DOD/DON endorsement.

Disclosure

The authors have no financial interest in this work or the vaccine evaluated in this clinical study. All authors contributed to the content of the manuscript and concurred with the decision to submit it for publication.

Ancillary