Abstract
INTRODUCTION
Uncontrolled HIV infection is known to activate the complement system, leading to an increase in chronic inflammation. Whether or not this activation of complement persists and contributes to chronic inflammation in subjects with HIV infection that is well-controlled through use of antiretroviral therapy has not been studied.
METHODS
We conducted an observational, cross-sectional study using sera from 305 adults with well-controlled HIV infection and 30 healthy controls. Sera was tested for markers of complement activation (C3a and C5a levels), complement function (CH50 assay), and immunoglobulin levels (IgG1-IgG4) as IgG can activate complement. We evaluated the association of well-controlled HIV infection with C3a, C5a, CH50, IgG1-IgG4, and total IgG levels using both univariate and multivariate analyses, controlling for factors such as age, gender race, comorbidities (including hepatitis C co-infection), smoking status, and statin use.
RESULTS
Well-controlled HIV infection was associated with a 54% increase in complement activation as measured by C3a levels compared with healthy controls (p<0.0001). Hepatitis C co-infection was associated with a further 52% increase in complement activation, as measured by C3a levels, over HIV alone (p = 0.003).
CONCLUSION
These results suggest that complement activation may contribute to a pro-inflammatory state even in well-controlled HIV infection. Further, HCV-co-infection may be even more pro-inflammatory, in terms of complement activation, compared with HIV infection alone.
Keywords: Complement, HIV, Hepatitis C, C3a, C5a, CH50
INTRODUCTION
HIV infection creates a chronic inflammatory state that improves but does not resolve when HIV is well-controlled through the use of antiretroviral therapy (ART)1–3. This chronic inflammation has been linked to a 2–4 fold higher risk of mortality from non-AIDS defining events such as cardiovascular disease3. It has also been linked to metabolic disorders, bone disease, kidney disease, and neurocognitive dysfunction1.
Untreated HIV infection is known to lead to complement activation through envelope proteins on the HIV virion4. In addition, untreated HIV infection produces elevated levels of IgG subclasses 1 and 3, the two IgG subclasses which are strongly complement activating, in contrast with subclasses 2 and 4, which are minimal activators of complement5–7. The complement system is a critical regulator of inflammation with multiple effects including opsonization, elaboration of inflammatory signaling molecules, and cell lysis. Complement activation leads to formation of the anaphylatoxins C3a and C5a, which can stimulate numerous immune cells. C3 is the central component of the complement system and is present in high concentrations in human plasma (1 g/l)8. C3a results from cleavage of C3, leading to the generation of opsonins, C3b/iC3b, and terminal complement cascade activation. C3a can have pro- or anti-inflammatory functions in different situations9. C5a is an anaphylatoxin with strong pro-inflammatory effects, including an association with cardiovascular events10. To date, the role of complement activation in well-controlled HIV infection has not been studied.
In order to better understand the potential roles of immunoglobulin and complement in the chronic inflammation seen in well-controlled HIV infection, we compared the levels of C5a, C3a, IgG subclasses, and classical pathway complement function via the CH50 assay in serum from adults with well-controlled HIV infection versus uninfected controls.
METHODS
This was an observational, cross-sectional study conducted at the primary HIV clinic (C3ID) at Eastern Virginia Medical School (EVMS) in Norfolk, VA. We used stored serum and data from 305 adults with well-controlled HIV infection who participated in one of two prior studies conducted 2012–2014, and who consented that their excess serum and data could be used in future studies. Inclusion criteria for both prior studies included documented HIV infection, age ≥18 years, and HIV viral load <400 copies/ml on the most recent test11. We also prospectively enrolled and collected serum and questionnaire data from 30 healthy controls who were 22–63 years old, not HIV-infected, not taking immunosuppressive medications, and not acutely ill. All healthy controls underwent informed consent. The EVMS Institutional Review Board approved the study.
On the day of collection, blood samples were centrifuged, and the serum aliquotted and stored at −80°C until laboratory analysis. All serum was extracted and stored by the same person using identical procedures and equipment to ensure continuity between samples. Serum was tested by six unique ELISA assays for the specific quantitative measurement of IgG 1- IgG 4 antibody levels (eBiosciences, San Diego, CA), C3a levels (eBiosciences, San Diego, CA), and C5a levels (R&D systems, Minneapolis, MN). Whole Complement Titration (CH50) was performed with the sera using Ab-sensitized sheep erythrocytes at 5×108 (CompTech, Tyler, TX) to test classical and terminal pathway complement function12. Total IgG levels were calculated by adding the levels of IgG1-4.
We estimated that with a sample size of 305 subjects with well-controlled HIV infection, we would need to enroll 30 healthy controls to detect a 25% change in C5a levels between HIV-infected subjects and controls to achieve a statistical power of 80% at α = 0.05.
Categorical variables were presented as percentages, and continuous variables were presented as means and standard deviations (SD). The primary outcomes included C3a levels, C5a levels, CH50 levels, IgG1 levels, IgG2 levels, IgG3 levels, IgG4 levels, and total IgG levels. The primary independent variable of interest was HIV infection. Covariates included in the analysis were age, gender, race, diabetes diagnosis, current smoking status, hepatitis C virus (HCV) coinfection, hepatitis B virus (HBV) coinfection, and current statin use. Among HIV-infected subjects only, covariates also included protease inhibitor use, atazanavir use, darunavir use, non-nucleoside reverse transcription inhibitor (NNRTI) use, integrase inhibitor use, tenofovir use, abacavir use, current CD4 count, history of AIDS diagnosis (defined as ever having a CD4 count<200 cells/mm3 or ever having an AIDS-defining illness), and duration of HIV infection. In addition, for the outcomes of C3a levels, C5a levels, and CH50 levels, we analyzed the association of IgG1, IgG2, IgG3, IgG4, and total IgG levels as variables. To examine the association between the primary outcomes and variables of interest, we used linear regression for continuous variables, t-test for binary variables, and ANOVA for categorical variables with more than two choices. For multivariate analyses, we included the primary variable of interest (HIV infection) as well as variables that were significant on univariate analysis. A subanalysis was done excluding the HCV-coinfected subjects once it was determined that HCV co-infection was associated with a number of the outcomes. Two-sided statistical tests were conducted at α = 0.05. Statistical analysis was performed using SAS version 9.3 (SAS institute, Cary, North Carolina).
RESULTS
Study Subjects
The demographics of the 305 adults with well-controlled HIV infection and the 30 healthy controls are shown in Table 1. A number of variables were significantly different between the two groups, including mean age, racial composition, smoking status, HCV infection status, and statin use.
Table 1.
Demographics of the Study Subjects
| Variables | Legend | HIV + | HIV − | p Value |
|---|---|---|---|---|
| Study Population | (freq, %) | 305 (91) | 30 (9) | - |
| Age in years | (mean ± SD) | 45.2 ± 11.1 | 34.8 ± 10.9 | <0.0001 |
| Gender | (freq, %) | 0.07 | ||
| Male | 212 (70) | 16 (53) | ||
| Female | 93 (30) | 14 (47) | ||
| Race | (freq, %) | <0.0001 | ||
| White | 95 (31) | 26 (86) | ||
| Black | 208 (68) | 2 (7) | ||
| Other | 2 (0.7) | 2 (7) | ||
| Diabetic | (freq, %) | 0.06 | ||
| No | 273 (90) | 30 (100) | ||
| Yes | 32 (10) | 0 | ||
| Current smoker | (freq, %) | <0.0001 | ||
| No | 176 (58) | 29 (97) | ||
| Yes | 129 (42) | 1 (3) | ||
| Hepatitis C | (freq, %) | 0.03 | ||
| No | 263 (86) | 30 (100) | ||
| Yes | 42 (14) | 0 | ||
| Hepatitis B | (freq, %) | 0.2 | ||
| No | 289 (95) | 30 (100) | ||
| Yes | 16 (5) | 0 | ||
| Statin Use | (freq, %) | 0.03 | ||
| No | 233 (76) | 28 (93) | ||
| Yes | 72 (24) | 2 (7) | ||
| Current CD4 Count in cells/mm3 | (mean ± SD) | 638 ± 326 | NA | |
| Duration of HIV Infection | (mean ± SD) | 11 ± 8 | NA | |
| History of AIDs Diagnosis | (freq, %) | NA | ||
| No | 128 (42) | |||
| Yes | 176 (58) | |||
| On Protease Inhibitor | (freq, %) | NA | ||
| No | 153 (50) | |||
| Yes | 149 (49) | |||
| On Integrase Inhibitor | (freq, %) | NA | ||
| No | 237 (78) | |||
| Yes | 65 (22) | |||
| On NNRTI | (freq, %) | NA | ||
| No | 138 (45) | |||
| Yes | 164 (55) | |||
HIV = Human immunodeficiency Virus;Freq = frequency; SD = standard deviation; NNRTI=Non-Nucleoside Reverse Transcriptase Inhibitors
Complement Activation as Measured by C3a Levels
HIV infection was associated with a 64% increase in C3a levels on univariate analysis (p=0.001), and showed a trend towards being increased on multivariate analysis (p=0.06). HCV co-infection appeared to strongly impact the association of C3a and HIV (Table 2, Figure 1). In the subanalysis excluding the 42 HCV co-infected subjects, HIV infection remained significantly associated with a 54% increase in C3a levels (7524 versus 4901 ng/ml, p<0.0001, in HIV-infected versus uninfected subjects, Figure 1).
Table 2.
Association of Well-Controlled HIV Infection with Complement Activation and Function and IgG Levels
| HIV + (mean ± SD) | HIV − (mean ± SD) | Unadjusted p value | Adjusted p value | |
|---|---|---|---|---|
| C5a (ng/mL) | 73 ± 31 | 87 ± 33 | 0.02 | 0.09 |
| C3a (ng/mL) | 8059 ± 7632 | 4901 ± 2240 | <0.0001 | 0.06 |
| CH50 (U/mL) | 46 ± 8 | 42 ± 6 | 0.0006 | 0.2 |
| IgG1 (µg/mL) | 8811 ± 5942 | 6760 ± 3268 | 0.004 | 0.5 |
| IgG2 (µg/mL) | 1960 ± 1278 | 2361 ± 1330 | 0.10 | |
| IgG3 (µg/mL) | 1955 ± 984 | 1433 ± 660 | 0.0003 | 0.04 |
| IgG4 (µg/mL) | 357 ± 356 | 481 ± 296 | 0.07 | 0.2 |
| Total IgG (µg/mL) | 13082 ± 6891 | 11035 ± 3906 | 0.02 | 0.6 |
Covariates were included in the multivariate model if they were significant on univariate analysis. Multivariate models included age and HCV infection for C5a; HCV infection for C3a; age, race, HCV infection, and statin use for CH50; race for IgG1, nothing for IgG2 (no covariates were significant on univariate analysis); age and HCV infection for IgG3; smoking status for IgG4; and race for IgG total.
Figure 1.
Mean Levels of C3a, C5a, IgG1, and IgG3 by HIV and HCV status; HIV=human immunodeficiency virus; HCV=hepatitis C virus. Lines represent mean and bars represent standard deviation. Of note, all HIV-infected subjects had viral loads <400 copies/ml.
Among the 305 subjects with well-controlled HIV infection (Table 3), HCV infection was the only covariate significantly associated with C3a levels, with HCV-infected subjects having a 52% increase in C3a levels (p=0.03). These results suggest that both HCV infection and well-controlled HIV infection are associated with increased complement C3 activation.
Table 3.
Other Variables Associated with Complement Activation and Function and IgG Levels among Subjects with Well-Controlled HIV Infection
| Variable | Mean ± SD | Unadjusted p value | Adjusted p value |
|---|---|---|---|
| C5a (ng/mL) | |||
| Hepatitis C | †0.03 | †0.03 | |
| No | 75 ± 30 | ||
| Yes | 64 ± 31 | ||
| C3a (ng/mL) | |||
| Hepatitis C | †0.03 | †0.03 | |
| No | 7524 ± 6819 | ||
| Yes | 11407 ± 11022 | ||
| CH50 (U/mL) | |||
| Race | 0.03 | 0.005 | |
| Black | 47± 8 | ||
| White | 44 ± 8 | ||
| Other | 45 ± 6 | ||
| Hepatitis C | 0.01 | 0.0009 | |
| No | 47± 7 | ||
| Yes | 42 ± 12 | ||
| Statin Use | <0.0001 | 0.0003 | |
| No | 45 ± 8 | ||
| Yes | 49 ± 7 | ||
| IgG1 (µg/mL) | |||
| Race | 0.005 | 0.0007 | |
| Black | 9566 ± 6212 | ||
| White | 7178 ± 5012 | ||
| Other | 7790 ± 3063 | ||
| Current CD4 count (continuous variable) | 0.0009 | 0.01 | |
| On Darunavir | 0.02 | 0.01 | |
| No | 8220 ± 4733 | ||
| Yes | 12164 ± 10214 | ||
| History of AIDs Diagnosis | <0.0001 | 0.003 | |
| No | 7222 ± 3688 | ||
| Yes | 9959 ± 6929 | ||
| IgG2 (µg/mL) | |||
| IgG3 (µg/mL) | |||
| Age (continuous variable) | 0.02 | 0.008 | |
| Statin Use | 0.04 | 0.002 | |
| No | 2020 ± 980 | ||
| Yes | 1744 ± 975 | ||
| History of AIDs Diagnosis | 0.01 | 0.02 | |
| No | 1793 ± 868 | ||
| Yes | 2071 ± 1047 | ||
| IgG4 (µg/mL) | |||
| Current Smoker | 0.03 | 0.04 | |
| No | 393 ± 391 | ||
| Yes | 309 ± 295 | ||
| History of AIDs Diagnosis | 0.03 | 0.04 | |
| No | 307 ± 278 | ||
| Yes | 394 ± 399 | ||
| Total IgG (µg/mL) | |||
| Race | 0.02 | 0.004 | |
| Black | 13852 ± 7031 | ||
| White | 11437 ± 6361 | ||
| Other | 11175 ± 3381 | ||
| On Darunavir | 0.03 | 0.01 | |
| No | 12476 ± 5822 | ||
| Yes | 16435 ± 10938 | ||
| Current CD4 count (continuous variable) | 0.003 | 0.04 | |
| History of AIDs Diagnosis | <0.0001 | 0.004 | |
| No | 11322 ± 4651 | ||
| Yes | 14356 ± 7908 | ||
only variable significant on univariate analysis.
Current CD4 count in cells/mm3 inversely correlated with IgG1 levels, and age in years directly correlated with IgG3 levels. Of note, smoking status, atazanavir use, and hepatitis C infection were significantly associated on univariate analysis with CH50, IgG1, and IgG3 levels respectively, but the significance did not remain on multivariate analysis.
As all but one of the 30 healthy controls had C3a levels <10,000 ng/ml, we further investigated the 305 subjects with well-controlled HIV infection to assess whether the C3a level of 10,000 ng/ml is a natural breakpoint, and what levels ≥ or <10,000 ng/ml might correlate with. We found that C5a levels were significantly higher (83 ±31 versus 71 ±31 ng/ml, p=0.01), and IgG4 levels were significantly lower (252 ±239 versus 382 ±373 µg/ml, p=0.001), in the 58 subjects with C3a ≥10,000 versus the 247 subjects with C3a <10,000 ng/ml respectively. We also found that the proportions of subjects with HCV infection (24% versus 11%, p=0.01), protease inhibitor use (67% versus 45%, p=0.002), NNRTI use (41% versus 57%, p=0.03), integrase inhibitor use (33% versus 19%, p=0.02), and abacavir use (19% versus 9%, p=0.03) were significantly different by C3a levels ≥10,000 versus <10,000 ng/ml respectively.
C5a Levels
A small (16%) decrease in C5a levels noticed on univariate analysis between HIV-infected versus uninfected subjects did not remain significant on multivariate analysis when age and HCV infection were included in the model (Table 2, Figure 1). In the subanalysis excluding the 42 HCV co-infected subjects, HIV status was the only variable significantly associated with C5a levels, with a similar small (14%) decrease in C5a levels in the HIV-infected subjects (p=0.04). Thus, as the change was very small and the association was in the opposite direction, the anaphylatoxin C5a does not appear to contribute to inflammation in well-controlled HIV infection.
Among the 305 subjects with well-controlled HIV infection (Table 3), HCV infection was the only covariate significantly associated with C5a levels, with HCV-infected subjects demonstrating a 15% lower C5a level (p=0.03). This small difference in C5a concentration is unlikely to be clinically relevant.
Complement Function as Measured by CH50 Levels
HIV infection was significantly associated with increased complement function as measured by CH50 levels on univariate analysis, but this did not remain significant on multivariate analysis when age, race, HCV infection, and statin use were included in the model (Table 2). Similarly, in the subanalysis excluding HCV-infected subjects, HIV infection was significantly associated with increased CH50 levels on univariate analysis, but not on multivariate analysis when age, race, diabetes status, and statin use were included in the model.
Among the 305 subjects with well-controlled HIV infection (Table 3), black race and statin use were significantly associated with increased CH50 levels, and HCV infection was significantly associated with decreased CH50 levels. The clinical relevance of increased CH50 levels is unknown. However, decreased CH50 levels are likely due to decreased production or increased consumption of complement components. In agreement with HCV co-infection being associated with increased C3a levels, the CH50 results provide supportive evidence for increased complement activation leading to the consumption of complement components in the setting of HCV co-infection.
IgG Levels
As expected, HIV infection was significantly associated with higher IgG1, IgG3, and total IgG levels on univariate analysis, but this only remained significant for IgG3 levels on multivariate analysis (Table 2, Figure 1). This same pattern was seen when HCV co-infected subjects were excluded from the analysis.
Among the 305 subjects with well-controlled HIV infection (Table 3), history of AIDS diagnosis was significantly correlated with higher IgG1, IgG3, IgG4, and total IgG levels. Taking darunavir and black race were also significantly correlated with higher IgG1 and total IgG levels. CD4 count, statin use, and currently smoking were inversely correlated with IgG1, IgG3, and IgG4 levels respectively.
Correlation Between IgG Levels and Complement Tests
When the associations between IgG levels and C5a, C3a, or CH50 levels were examined, we found that IgG1 levels were significantly inversely related to CH50 levels on both univariate and multivariate analysis (r coefficient −0.12, p=0.03 and 0.008 respectively). IgG4 was significantly related to C5a levels on univariate analysis, but the association did not remain significant on multivariate analysis (r coefficient 0.12, p=0.04 and 0.06 respectively). This is consistent with IgG1 being a strong complement activator via the classical pathway and potentially leading to some depletion of complement components and hence lower CH50 levels.
DISCUSSION
We measured complement activation and function in 305 subjects with well-controlled HIV infection and 30 healthy controls, to assess whether well-controlled HIV infection is associated with complement activation, and by extension whether this might contribute to the chronic inflammatory state seen in subjects with well-controlled HIV infection. C3a data, indicative of activation of the central complement component and most important in assessing overall complement activation, was increased 1.5-fold for HIV-infected individuals. In contrast, C5a levels showed minimal difference (−16%) with HIV infection (p=0.09 in multivariate analysis), suggesting that this anaphylatoxin has little or no clinical impact on inflammation in well-controlled HIV infection. CH50 values are calculated as a z-score, limiting the sensitivity of this measure unless there are very large amounts of complement component consumption or decreased component production (e.g. C6 deficiency). Other variables that were associated with complement activation and function included HCV infection, statin use, and race.
A number of studies have demonstrated an association between untreated HIV infection and complement activation, along with a decrease in available uncleaved complement factors and thus complement function13–17. Untreated HIV infection is believed to activate the classical pathway both through HIV-specific antibodies binding the HIV virions, and by the HIV surface protein gp41 binding C1q18. In addition, the HIV surface protein gp120 can bind mannose-binding lectin and activate the lectin pathway18. However, to our knowledge, only one prior study on complement activation in HIV-infected subjects included any subjects on ART19. This study took place in Gabon, and included 40 HIV-infected subjects on and 46 subjects off ART, as well as uninfected controls. They found that HIV infection was associated with complement activation both in asymptomatic and septic subjects, but that complement activation correlated with HIV viral load19. Our study included much larger numbers of HIV-infected subjects on ART, all with HIV viral loads <400 copies/ml, and suggests that even in the setting of well-controlled HIV infection, there is significant activation of the central complement component C3. This suggests that even well-controlled HIV infection is associated with complement activation, which may be contributing to a pro-inflammatory state.
Our data show significant elevation of C3a levels without a corresponding increase in C5a levels in the HIV-infected cohort. However, HIV-infected subjects with C3a levels ≥10,000 ng/ml had significantly elevated C5a levels compared to HIV-infected subjects with C3a levels <10,000 ng/ml. Despite the cascade nature of complement, it has been shown that C3a and C5a levels do not correlate for a variety of diseases including intestinal ischemia reperfusion20, sepsis21 and cystic fibrosis22. C3a and C5a have described receptors including, C3aR, C5aR1 and C5aR2 that modulate different effects, although there can be some crosstalk where C3a-desArg may also bind to C5aR2 and mediate downstream pro-inflammatory effects without generation of C5a23. It has been shown that C5a may increase the susceptibility of monocyte derived macrophages to HIV infection24. Our data suggest this may be attenuated in well-controlled HIV infection. We speculate that in well-controlled HIV, C3b may be rapidly converted to its inactivated isoform iC3b, preventing further propagation of the terminal complement cascade and its downstream effects25.
In addition to HIV infection by itself being associated with C3 activation, HCV co-infection was associated with a further increase in C3 activation, as evidenced by a 1.5-fold increase in C3a which was significant on both univariate and multivariate analysis. This was supported by decreased CH50 levels in HCV coinfection, suggesting complement component consumption as a result of increased complement activation. To our knowledge, this is the first report of HCV co-infection with HIV being associated with increased complement activation. A previous report showed lower serum C3 and C4 levels in HCV-infected subjects compared with uninfected controls26, potentially reflecting complement activation and consumption or decreased component synthesis in the liver. Additionally, it has been shown that elevated C3a in HCV-infected patients is associated with the development of hepatocellular carcinoma27. Thus, the association of HCV co-infection in subjects with well-controlled HIV infection with elevated complement activation appears to suggest a pro-inflammatory state.
Well-controlled HIV infection was associated with increased levels of IgG1, IgG3, and total IgG, but this only remained significant for IgG3 on multivariate analysis. Our findings are consistent with a number of studies that have demonstrated that uncontrolled HIV infection is associated with increased IgG1, IgG3, and total IgG levels5,6,28. IgG1 and total IgG levels have been shown to markedly decrease in HIV-infected subjects after ART initiation28. As our subjects were all on ART with well-controlled HIV, it makes sense that the increased levels of IgG1, IgG3, and total IgG were not as marked as in studies of subjects with uncontrolled HIV. Our finding that a diagnosis of AIDS was significantly associated with higher levels of three of the four IgG subtypes is consistent with a prior study showing that advanced HIV infection is associated with elevated levels of IgG2 and IgG4 in addition to IgG1 and IgG35, and suggests that these elevated levels may persist even after the HIV is well-controlled.
We found that being on a lipid-lowering drug of the statin class was associated with significantly increased CH50 levels and decreased levels of IgG3. Statin drugs are known to have anti-inflammatory properties, and have been shown to decrease complement activation in a mouse model29. One could hypothesize that the increased complement function could be a result of less complement activation and thus more available uncleaved complement factors. However, our study did not see an association between statin use and C5a or C3a levels, our measures of complement activation.
We found race was associated with increased CH50 levels, and increased levels of IgG1 and total IgG. There is very little data on the effect of race on complement activation and function. We found one study from 1975 in 163 infants and children that noted no difference by race of levels of CH50, C3, C4, or C530. However, consistent with our findings, black race has been linked to increased IgG levels in prior studies31.
Our finding that currently smoking was associated with decreased levels of IgG4 is consistent with prior literature that showed an association of smoking with decreased IgG levels31. Our finding that subjects taking darunavir versus other antiretroviral agents had higher levels of IgG1 and total IgG was surprising. To our knowledge, there is no prior data in the literature on the effect of darunavir on IgG levels.
Our study has several limitations. Demographics of our HIV-infected subjects and uninfected subjects were significantly different, which reflects real life, but makes teasing out the contribution of HIV infection more difficult. However, to account for this, we conducted multivariate analysis to control for demographic differences that influenced our outcomes, and we did a subanalysis excluding the HCV co-infected subjects. In addition, it would have been ideal to also include subjects with uncontrolled HIV-infection in the study, but we were unable to do this because of budget constraints.
Despite these limitations, we are the first to evaluate complement activation in the setting of well-controlled HIV infection. Well-controlled HIV infection was associated with increased complement activation as measured by C3a levels in our study. HCV co-infection was associated with a further increase in complement activation, suggesting a pro-inflammatory state for patients with both HCV and HIV infection, even when their HIV infection is well-controlled. The clinical importance of this finding deserves further investigation.
Acknowledgments
We would like to thank the providers, staff, and especially patients at the EVMS C3ID clinic for their generous support and cooperation. We would also like to thank Julia Siik RN for enrolling the initial subjects and collecting the samples.
Source Funding:
Samples used for this study were taken from studies funded by the Doris Duke Charitable Foundation: (Clinical Scientist Development Award 2012061 [principal investigator, S.Troy]). S.Troy and A. Rossheim also received salary support while working on this project from both the Doris Duke Charitable Foundation and the US National Institutes of Health: (Career Development Award 5K23AI093678 [principal investigator, S. Troy.]). However, neither funding agency was involved in the decision to conduct this study or in the preparation of the manuscript.
Footnotes
Conflicts of Interest: Authors have no conflict of interest to declare.
REFERENCES
- 1.Lederman MM, Funderburg NT, Sekaly RP, Klatt NR, Hunt PW. Residual immune dysregulation syndrome in treated HIV infection. Adv Immunol. 2013;119:51–83. doi: 10.1016/B978-0-12-407707-2.00002-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hearps AC, Maisa A, Cheng WJ, et al. HIV infection induces age-related changes to monocytes and innate immune activation in young men that persist despite combination antiretroviral therapy. AIDS. 2012;26(7):843–853. doi: 10.1097/QAD.0b013e328351f756. [DOI] [PubMed] [Google Scholar]
- 3.Tenorio AR, Zheng Y, Bosch RJ, et al. Soluble markers of inflammation and coagulation but not T-cell activation predict non-AIDS-defining morbid events during suppressive antiretroviral treatment. J Infect Dis. 2014;210(8):1248–1259. doi: 10.1093/infdis/jiu254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Huber G, Bánki Z, Lengauer S, Stoiber H. Emerging role for complement in HIV infection. Curr Opin HIV AIDS. 2011;6(5):419–426. doi: 10.1097/COH.0b013e3283495a26. [DOI] [PubMed] [Google Scholar]
- 5.McGowan JP, Shah SS, Small CB, et al. Relationship of serum immunoglobulin and IgG subclass levels to race, ethnicity and behavioral characteristics in HIV infection. Med Sci Monit. 2006;12(1):CR11–CR16. [PubMed] [Google Scholar]
- 6.Raux M, Finkielsztejn L, Salmon-Céron D, et al. IgG subclass distribution in serum and various mucosal fluids of HIV type 1-infected subjects. AIDS Res Hum Retroviruses. 2000;16(6):583–594. doi: 10.1089/088922200309007. [DOI] [PubMed] [Google Scholar]
- 7.Bindon CI, Hale G, Brüggemann M, Waldmann H. Human monoclonal IgG isotypes differ in complement activating function at the level of C4 as well as C1q. J Exp Med. 1988;168(1):127–142. doi: 10.1084/jem.168.1.127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Engström G, Hedblad B, Berglund G, Janzon L, Lindgärde F. Plasma levels of complement C3 is associated with development of hypertension: a longitudinal cohort study. J Hum Hypertens. 2007;21(4):276–282. doi: 10.1038/sj.jhh.1002129. [DOI] [PubMed] [Google Scholar]
- 9.Coulthard LG, Woodruff TM. Is the complement activation product C3a a proinflammatory molecule? Re-evaluating the evidence and the myth. J Immunol. 2015;194(8):3542–3548. doi: 10.4049/jimmunol.1403068. [DOI] [PubMed] [Google Scholar]
- 10.Speidl WS, Exner M, Amighi J, et al. Complement component C5a predicts future cardiovascular events in patients with advanced atherosclerosis. Eur Heart J. 2005;26(21):2294–2299. doi: 10.1093/eurheartj/ehi339. [DOI] [PubMed] [Google Scholar]
- 11.Troy SB, Kouiavskaia D, Siik J, et al. Comparison of the Immunogenicity of Various Booster Doses of Inactivated Polio Vaccine Delivered Intradermally Versus Intramuscularly to HIV-Infected Adults. J Infect Dis. 2015;211(12):1969–1976. doi: 10.1093/infdis/jiu841. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Todd JC. Clinical diagnosis and management by laboratory methods. In: Gaither TAFM, editor. Complement in clinical pathology. Philadelphia London: Saunders; 1979. pp. 1245–1261. [Google Scholar]
- 13.Jarvis JN, Taylor H, Iobidze M, Dejonge J, Chang S, Cohen F. Rheumatoid factor expression and complement activation in children congenitally infected with human immunodeficiency virus. Clin Immunol Immunopathol. 1993;67(1):50–54. doi: 10.1006/clin.1993.1044. [DOI] [PubMed] [Google Scholar]
- 14.Jarvis JN, Taylor H, Iobidze M. Complement activation and immune complexes in early congenital HIV infection. J Acquir Immune Defic Syndr Hum Retrovirol. 1995;8(5):480–485. doi: 10.1097/00042560-199504120-00008. [DOI] [PubMed] [Google Scholar]
- 15.Carini C, Perricone R, Fratazzi C, et al. Complement activation is associated with the presence of specific human immunodeficiency virus (HIV)-anti-HIV immune complexes in patients with acquired immunodeficiency syndrome-related complex or lymphoadenopathy syndrome. Scand J Immunol. 1989;30(3):347–353. doi: 10.1111/j.1365-3083.1989.tb01220.x. [DOI] [PubMed] [Google Scholar]
- 16.Perricone R, Fontana L, de Carolis C, Carini C, Sirianni MC, Aiuti F. Evidence for activation of complement in patients with AIDS related complex (ARC) and/or lymphoadenopathy syndrome (LAS) Clin Exp Immunol. 1987;70(3):500–507. [PMC free article] [PubMed] [Google Scholar]
- 17.Lin RY, Wildfeuer O, Franklin MM, Candido K. Hypocomplementemia and human immunodeficiency virus infection. Clinical correlates and relationships to circulating immune complex and immunoglobulin G levels. Int Arch Allergy Appl Immunol. 1988;87(1):40–46. doi: 10.1159/000234646. [DOI] [PubMed] [Google Scholar]
- 18.Yu Q, Yu R, Qin X. The good and evil of complement activation in HIV-1 infection. Cell Mol Immunol. 2010;7(5):334–340. doi: 10.1038/cmi.2010.8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Huson MA, Wouters D, van Mierlo G, Grobusch MP, Zeerleder SS, van der Poll T. HIV Coinfection Enhances Complement Activation During Sepsis. J Infect Dis. 2015;212(3):474–483. doi: 10.1093/infdis/jiv074. [DOI] [PubMed] [Google Scholar]
- 20.Wu MC, Brennan FH, Lynch JP, et al. The receptor for complement component C3a mediates protection from intestinal ischemia-reperfusion injuries by inhibiting neutrophil mobilization. Proc Natl Acad Sci U S A. 2013;110(23):9439–9444. doi: 10.1073/pnas.1218815110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hollmann TJ, Mueller-Ortiz SL, Braun MC, Wetsel RA. Disruption of the C5a receptor gene increases resistance to acute Gram-negative bacteremia and endotoxic shock: Opposing roles of C3a and C5a. Mol Immunol. 2008;45(7):1907–1915. doi: 10.1016/j.molimm.2007.10.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sass LA, Hair PS, Perkins AM, Shah TA, Krishna NK, Cunnion KM. Complement Effectors of Inflammation in Cystic Fibrosis Lung Fluid Correlate with Clinical Measures of Disease. PLoS One. 2015;10(12):e0144723. doi: 10.1371/journal.pone.0144723. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Coulthard LG, Woodruff TM. Is the Complement Activation Product C3a a Proinflammatory Molecule? Re-evaluating the Evidence and the Myth. J Immunol. 2015;194(8):3542–3548. doi: 10.4049/jimmunol.1403068. [DOI] [PubMed] [Google Scholar]
- 24.Kacani L, Banki Z, Zwirner J, et al. C5a and C5a(desArg) enhance the susceptibility of monocyte-derived macrophages to HIV infection. J Immunol. 2001;166(5):3410–3415. doi: 10.4049/jimmunol.166.5.3410. [DOI] [PubMed] [Google Scholar]
- 25.Cunnion KM, Hair PS, Buescher ES. Cleavage of Complement C3b to iC3b on the Surface of Staphylococcus aureus Is Mediated by Serum Complement Factor I. Infect Immun. 722004:2858–2863. doi: 10.1128/IAI.72.5.2858-2863.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.El-Fatah Fahmy Hanno A, Mohiedeen KM, Deghedy A, Sayed R. Serum complements C3 and C4 in chronic HCV infection and their correlation with response to pegylated interferon and ribavirin treatment. Arab J Gastroenterol. 2014;15(2):58–62. doi: 10.1016/j.ajg.2014.04.005. [DOI] [PubMed] [Google Scholar]
- 27.Kanmura S, Uto H, Sato Y, et al. The complement component C3a fragment is a potential biomarker for hepatitis C virus-related hepatocellular carcinoma. J Gastroenterol. 2010;45(4):459–467. doi: 10.1007/s00535-009-0160-5. [DOI] [PubMed] [Google Scholar]
- 28.Voltersvik P, Albrektsen G, Ulvestad E, Dyrhol-Riise AM, Sørensen B, Asjö B. Changes in immunoglobulin isotypes and immunoglobulin G (IgG) subclasses during highly active antiretroviral therapy: anti-p24 IgG1 closely parallels the biphasic decline in plasma viremia. J Acquir Immune Defic Syndr. 2003;34(4):358–367. doi: 10.1097/00126334-200312010-00002. [DOI] [PubMed] [Google Scholar]
- 29.Gonzalez JM, Pedroni SM, Girardi G. Statins prevent cervical remodeling, myometrial contractions and preterm labor through a mechanism that involves hemoxygenase-1 and complement inhibition. Mol Hum Reprod. 2014;20(6):579–589. doi: 10.1093/molehr/gau019. [DOI] [PubMed] [Google Scholar]
- 30.Norman ME, Gall EP, Taylor A, Laster L, Nilsson UR. Serum complement profiles in infants and children. J Pediatr. 1975;87(6 Pt 1):912–916. doi: 10.1016/s0022-3476(75)80904-8. [DOI] [PubMed] [Google Scholar]
- 31.Tollerud DJ, Brown LM, Blattner WA, et al. Racial differences in serum immunoglobulin levels: relationship to cigarette smoking, T-cell subsets, and soluble interleukin-2 receptors. J Clin Lab Anal. 1995;9(1):37–41. doi: 10.1002/jcla.1860090107. [DOI] [PubMed] [Google Scholar]