Biochip Array-based Analysis of Plasma Cytokines in HIV Patients with Immunological and Virological Discordance

Authors


Dr Deshratn Asthana, Laboratory for Clinical and Biological Studies, University of Miami-Miller School of Medicine, 1550 NW 10th Avenue, Suite 118, Miami, FL 33136, USA. E-mail: desh@miami.edu

Abstract

Assessment of cytokines in body fluids or cells provides important information in understanding the disease process and designing treatment strategies. Recent introduction of antibody-based protein arrays have provided investigators simultaneous and specific detection of multiple analytes in a single sample using minimum volumes. In this study, we used a biochip array system capable of measuring 12 cytokines and growth factors (IL-2, IL-4, IL-6, IL-8, IL-10, IL-1α, IL-1β, IFN-γ, TNF-α, monocyte chemoattractant protein-1 (MCP-1), vascular endothelial growth factor (VEGF) and epidermal growth factor (EGF)) in HIV patients with immunological and virological discordance (discordant) to find out differences if any, in their plasma cytokine profiles when compared with concordant HIV-infected individuals. A sandwich chemiluminescent assay was performed with plasma specimens of 110 HIV patients (55 discordant, 55 concordant) and 22 normal healthy individuals followed by enzyme-linked immunosorbent assay (ELISA) to the confirm levels of cytokines and growth factors that showed significant differences in the two groups. The discordant HIV patients showed significantly higher levels of plasma VEGF (P = 0.001) and EGF (P = 0.034) levels when compared with concordant patients. Overall, the patients showed significantly higher levels of TNF-α, MCP-1 and VEGF when compared with the normal healthy controls (P < 0.05). ELISA for VEGF (P < 0.001) and EGF (P = 0.004) confirmed the comparison obtained with biochip array, between the discordant and concordant patients. The results of cytokine quantitation by biochip array and ELISA confirmed that this technology is not only comparable but also has a good potential in the future applications involving measurement of multiple cytokines with limiting specimens.

Introduction

Assessment of cytokines in body fluids, tissues or cells provides important information in understanding the disease process and designing treatment strategies. Conventional techniques like enzyme-linked immunosorbent assay (ELISA) can measure only one cytokine at a time. Other techniques such as RT-PCR, ELISPOT, immunohistochemistry though provide information on the types of cells that secret cytokines, but these too have the limitation of measuring one or few cytokines at a time [1]. Simultaneous assessment of many cytokines in a biological sample provides more comprehensive information rather than assessing a single cytokine [2]. Thus, the traditional cytokine technologies that measure one cytokine at a time are being gradually replaced by multiplex-type formats (DNA and protein microarrays), which are sensitive enough to detect multiple analytes even in small quantities of biological samples. Several multiplexed bead systems are currently marketed by different vendors for simultaneous detection of multiple cytokines. For example, a multiplexed particle-based flow cytometric assay available from Luminex FlowmetrixTM system (DeSoto, TX, USA) consists of different bead sets with distinct proportions of red and orange fluorescent dyes that can simultaneously measure up to 25 cytokines in a single sample [3]. Similarly, BD Biosciences (San Jose, CA, USA) offer a Cytometric Bead ArrayTM system, which is capable of distinguishing Th1 and Th2 cytokines and detection of pro- inflammatory cytokines in a given sample [4]. Recent introduction of antibody-based protein arrays have provided investigators simultaneous and specific detection of multiple analytes at a time from a single sample using minimum volumes [5–7]. Antibody-based cytokine microarrays can be successfully employed to analyze many cytokines simultaneously with enhanced sensitivities of chemiluminescence and high specificities of conventional enzyme immunoassays.

There are some HIV-infected patients in whom complete suppression of the viral load does not assure restoration of CD4 T cell counts [8, 9]. Such discordant responses suggest that there are other factors in immune system that influence the restoration of CD4 T cell numbers despite decrease in their plasma HIV-1 levels. It is well established that the combined effect of multiple cytokines and chemokines in the immune response to disease or immune modulation by drugs is often more important than the function of specific cytokines [10, 11]. Therefore, the plasma cytokine milieu might be significant in influencing the immune reconstitution or inflammatory damage during the course of disease. In this study, we used a biochip array system capable of measuring 12 cytokines and growth factors (IL-2, IL-4, IL-6, IL-8, IL-10, IL-1α, IL-1β, IFN-γ, TNF-α, MCP-1, VEGF and EGF) in HIV patients with Immunological and virological discordance (discordant) to find out differences if any, in their plasma cytokine profiles when compared with concordant HIV-infected individuals. The results of the biochip array were compared with ELISA to confirm the levels of those cytokines and growth factors that showed significant differences in the discordant and concordant HIV-infected subjects.

Materials and methods

Study samples.  Ethylenediaminetetraacetic acid (EDTA) plasma specimens were obtained from 55 discordant and 55 concordant HIV patients coming to the laboratory for Clinical and Biological Studies (LCBS), University of Miami-Miller School of Medicine, Miami, FL, USA for routine diagnostic investigations such as plasma viral loads and T- cell subset analysis during 2004–2006. The criteria for selection of discordant patients was; whole blood CD4 T-cells <200/μl or CD4/CD8 ratio <1.0 and undetectable plasma HIV-1 RNA for at least 1 year (I−/V+) with no CD4 T-cell increase >100/μl during 24 weeks following HAART as described previously [8]. All patients in this study had baseline determination of CD4 and CD8 T cell counts, plasma viral load. The patients were excluded for any acute illness and recorded for opportunistic infection, lymphomas, or psychiatric illness. For normal reference ranges plasma specimens were also obtained from 22 healthy HIV negative individuals. All the specimens were stored in aliquots at −80 °C prior to the analysis.

Biochip array analysis.  Cytokine levels in plasma specimens were measured using a biochip array system, Evidence InvestigatorTM (Randox laboratories Ltd, Crumlin, UK). The testing platform consists of biochips secured in the base of a well placed in a carrier holding nine biochips in a 3 × 3 format. Each biochip is coated with the capture antibodies specific for each of the 12 cytokines and growth factors (IL-2, IL-4, IL-6, IL-8, IL-10, IL-1α, IL-1β, IFN-γ, TNF-α, MCP-1, VEGF and EGF) on a particular test region. A sandwich chemiluminescent assay was performed with 100 μl plasma using reagents (including the calibrators and controls) and protocols supplied by the same manufacturer. The light signal generated from each of the test regions on the biochip was detected using a charge-coupled detector camera and imaging system and compared with a calibration curve generated with known standards during the same run. All specimens were run in duplicate and the concentration of each cytokine present in each plasma specimen was calculated from the calibration curve and reported in pg/ml.

ELISA.  Enzyme-linked immunosorbent assay for determination of VEGF and EGF levels was performed in duplicate using the same patient and normal control plasma specimens as per manufacturer's instructions (Quantikine Immunoassay; R&D systems, Minneapolis, MN, USA). Absorbance of the ELISA plates was measured at 450 nm using an automatic microplate reader (Molecular devices, Sunnyvale, CA, USA). The sensitivity of the assay was 5.0 pg/ml for VEGF and 0.7 pg/ml for EGF.

Statistical analysis.  The levels of each cytokine in the plasma of discordant and concordant HIV patients and normal healthy controls were compared by performing Wilcoxon's signed rank test using GraphPad Prism software (version 4.0; GraphPad Software Inc., San Diego, CA, USA). Sample wise correlation between biochip array results and ELISA results were analyzed by calculating Spearman's rank correlation coefficient. P values of <0.05 were considered significant.

Results

Patient characteristics

The characteristics of the patients are summarized in Table 1. Except for the viral load and CD8 T-cell counts, there were no significant differences in the CD4 T-cell counts of the discordant and concordant HIV patients. Both groups of patients underwent similar treatment regimens and were age and sex matched. However, as expected the CD4 and CD8 T-cell counts of normal healthy controls did differ significantly from the HIV-infected patients.

Table 1.   Patient characteristics.
CharacteristicsDiscordant HIV patientsConcordant HIV patientsNormal healthy controls
n555522
Males444315
Females11127
Age, years (mean ± SD)38.7 ± 8.340.36 ± 9.442.3 ± 8.1
CD4/μl (mean ± SD)443 ± 214466 ± 293907 ± 135
CD8/μl (mean ± SD)1185 ± 447913 ± 459675 ± 241
CD4/CD8 ratio (mean ± SD)0.37 ± 0.160.59 ± 0.431.34 ± 0.59
Viral load (mean ± SD)<2517,756 ± 42,185NA
Duration of HIV infection, years (mean ± SD)5.0 ± 4.15.5 ± 6.4NA
Months on HAART (mean ± SD)55 ± 761 ± 11NA

Plasma cytokine levels: biochip array

The sensitivity of detection of various cytokines with biochip array analysis and the percentage of patients showing values above detection limit is described in Table 2. Some of the cytokines such as IL-1α, IL-1β, IL-4, IL-10 and IFN- γ had concentrations below the detection limit in majority of the patients and normal subjects. Overall, there appeared to be highly significant differences in the levels of TNF-α, MCP-1 and VEGF between the patients versus the controls with patients showing higher levels of TNF-α (mean; 5.68 ± 6.60 versus 1.49 ± 2.21 pg/ml), MCP-1 (mean; 186.06 ± 121.79 versus 109.08 ± 72.84 pg/ml) and VEGF (mean; 75.41 ± 84.83 versus 41.36 ± 27.17 pg/ml; P < 0.05). The levels of EGF also seemed to be higher in the patient population when compared with controls though the difference was not statistically significant (mean; 95.64 ± 140.70 versus 63.67 ± 56.71 pg/ml).

Table 2.   Plasma cytokine levels in the study population as determined by biochip array analysis.
AnalyteSensitivity (pg/ml)Percentage detectedMean ± SD (pg/ml)P-value (discordant versus concordant)P-value (patients versus controls)
  1. *Significant.

IL-2
 Discordant3.398.88.36 ± 9.040.1460.431
 Concordant7.83 ± 4.46
 Controls10.13 ± 7.72
IL-4
 Discordant4.832.11.30 ± 5.340.3090.844
 Concordant1.33 ± 2.54
 Controls1.37 ± 2.71
IL-6
 Discordant1.171.47.96 ± 28.610.9730.764
 Concordant6.54 ± 20.47
 Controls4.68 ± 8.85
IL-8
 Discordant2.569.016.83 ± 36.810.4840.472
 Concordant21.87 ± 87.72
 Controls5.21 ± 4.14
IL-10
 Discordant1.419.00.68 ± 2.250.8370.437
 Concordant0.33 ± 0.70
 Controls0.10 ± 0.30
IL-1α
 Discordant2.08.30.53 ± 1.890.034*0.625
 Concordant0.06 ± 0.26
 Controls0.15 ± 0.40
IL-1β
 Discordant3.327.42.99 ± 6.680.014*0.322
 Concordant0.69 ± 1.69
 Controls0.87 ± 2.87
IFN- γ
 Discordant4.022.61.59 ± 4.720.3350.195
 Concordant1.77 ± 3.08
 Controls1.85 ± 2.94
TNF-α
 Discordant6.761.25.93 ± 7.590.9640.015*
 Concordant5.43 ± 1.89
 Controls1.49 ± 2.21
MCP-1
 Discordant3.0100191.17 ± 127.020.9450.032*
 Concordant180.94 ± 117.28
 Controls109.08 ± 72.85
VEGF
 Discordant12.091.7102.75 ± 107.190.001*0.019*
 Concordant48.08 ± 38.92
 Controls41.36 ± 27.17
EGF
 Discordant6.586.9125.39 ± 159.700.034*0.451
 Concordant65.90 ± 112.48
 Controls63.67 ± 56.71

On comparing the discordant and the concordant patients, we observed that the concentrations of most of the cytokines including plasma IL-2, IL-4, IFN- γ and IL-10 did not differ significantly. However, there were significant differences in the concentration of plasma VEGF and EGF in both the groups of patients, with the discordant patients showing higher levels of VEGF (102.75 ± 107.19 versus 48.08 ± 38.92 pg/ml, P = 0.001) and EGF (125.39 ± 159.70 versus 65.90 ± 112.48 pg/ml, P = 0.034). One of the important pro-inflammatory cytokines, TNF- α, had relatively similar levels in both the groups of HIV patients. Other pro-inflammatory cytokines IL-1α and IL-1β that were detected in very few specimens (8.3% and 27.4% respectively), also differed significantly, with discordant patients showing comparatively higher levels (P < 0.05).

VEGF and EGF levels: ELISA

To confirm the findings of biochip array analysis, ELISA was performed on same specimens. Overall ELISA showed higher values of both growth factors in HIV patients and controls when compared with biochip array analysis. Mean VEGF level of all the 132 specimens measured was 100.63 pg/ml by ELISA when compared with 69.74pg/ml by biochip array, while in case of EGF the levels were 112.71 and 90.31 pg/ml, respectively (Fig. 1). More importantly, the results of ELISA showed the same trend of results as obtained with biochip array analysis showing significantly higher levels of VEGF (P < 0.001) and EGF (P = 0.004) in discordant patients when compared with concordant patients (Table 3). The normal controls had lower levels of VEGF and EGF when compared with HIV patients, with the levels of VEGF being significantly lower (P < 0.001) as observed with biochip array analysis. Sample wise comparison of VEGF and EGF levels obtained by ELISA and biochip array too showed a highly significant correlation between both of these techniques (VEGF, r = 0.80; EGF, r = 0.97, P < 0.001).

Figure 1.

 Box and whisker plot showing distribution of plasma (A) VEGF levels and (B) EGF levels in discordant and concordant HIV-infected subjects and normal healthy controls as compared by cytokine biochip array and ELISA. Boxes represent 25th and 75th percentile with the median value (solid line) between boxes.

Table 3.   VEGF and EGF levels in the study population as determined by ELISA.
AnalyteSensitivity (pg/ml)Percentage detectedMean ± SD (pg/ml)P-value (discordant versus concordant)P-value (patients versus controls)
  1. *Significant

VEGF
 Discordant5.095.4135.45 ± 85.520.000*0.001*
 Concordant79.40 ± 57.22
 Controls66.64 ± 44.92
EGF
 Discordant0.790.9147.25 ± 143.640.004*0.234
 Concordant95.27 ± 122.39
 Controls69.95 ± 62.31

Discussion

Cytokine biology is quite complex and many cytokine assays still need correct and uniform standardization. At the same time, improved understanding of cytokine interactions has led to a consensus that simultaneous assessment of many cytokines in a biological specimen provides more relevant information rather than assessing a single cytokine. Newer techniques capable of detecting multiple analytes in small quantities of biological samples use different approaches in detection (multi parameter flow cytometry, protein microarrays) besides, differences in their sensitivities and specificities. For example, the chemiluminescent peroxidase substrate luminol, is approximately 10 times more sensitive than tetramethyl benzidine (TMB) commonly used in ELISA even though they are catalyzed by the same enzyme horseradish peroxidase (HRP) [12]. Hence, amidst the availability of various techniques to measure cytokines in biological specimens, rationale for selecting the right assay becomes critical. In this small study, we have utilized a technology that can quantify up to 12 cytokines and growth factors simultaneously in a single plasma/serum sample to assess whether discordant HIV patients have differing concentrations of cytokines or growth factors expressed in their blood when compared with the concordant patients. Additionally, we wanted to explore the utility of this technology in a routine diagnostic setting in comparison with a conventional and widely accepted immunoassay such as ELISA.

We observed that both groups differed mainly in the expression of two growth factors, VEGF and EGF in the plasma, which was quite surprising. There have been many reports in the past on discordance in viral load and CD4 cell counts in HIV-infected patients [8, 9]. However, there is very little information available on the immune factors which may influence or are associated with discordance in HIV patients. The discordant and concordant HIV patients in our study had similar CD4 cell counts and the only apparent differences were viral load and CD8 cell counts, with the former having undetectable viral loads and higher CD8 cell numbers. To assume that differences in their plasma cytokine profiles in both the groups could be attributed to differences in the viral load and CD8 cells would be difficult to state here without further evidence. So far, the VEGF and EGF have never been reported to be associated with decreased HIV viral load or CD4 cell counts therefore; further studies would be needed to investigate why the discordant patients expressed higher levels of plasma VEGF and EGF than their concordant counterparts. The VEGF is mainly recognized as a specific mitogen for the vascular endothelial cells that plays an important role in revascularization, especially after inflammatory responses. Increases in the levels of plasma VEGF have been reported in various malignancies like Kaposi's sarcoma, which is generally associated with advanced stages of HIV infection [13]. In HIV infection, increase in VEGF levels by the T-cells has been reported earlier [14] and the EGF too has been reported to stimulate promoters of HIV-1 in the isolated villous trophoblast cells [15]. Therefore, it would be interesting to identify the principal cell types in discordant patients which might be producing high levels of plasma VEGF and EGF; the same is currently under investigation in our laboratory. Amongst the other cytokines, the levels of pro-inflammatory cytokines IL-1α and IL-1β were also higher in the discordant group, although they were not detected in majority of the patients and hence were ignored in making comparisons. Interestingly the levels of TNF-α, a major pro-inflammatory cytokine, were relatively same in both the groups suggesting viral replication is similar in both the groups.

Another important factor to consider is increased T-cell activation/apoptosis that has been reported to be associated with increased inflammatory-mediated immune damage. Unabated viral replication in lymphoid tissues, despite undetectable plasma viral load, has been proposed as the underlying mechanism of cellular activation [11]. The differences in expression of VEGF, EGF and also, IL-1α and IL-1β perhaps indicate that inflammatory damage might be more in the discordant HIV patients, which might be due to higher CD8 cell counts and hence lower CD4/CD8 ratios, despite persistent low viral load in these patients over a long duration. However, this ‘paradoxical response’ can probably be associated with other events. We must also consider here that systemic measurement of cytokine levels may not represent their true levels found in selected microenvironments while, at the same time it should be noted that measurement of local cytokine milieu is practically difficult to determine especially when they are released at varying levels at different periods, depending upon the cell stimuli. Hence, systemic measurement of multiple cytokines and growth factors in plasma or serum as described in this study gives an indication of any underlying pro-inflammatory responses and also gives information on the skewness of the T-helper responses (Th1 or Th2) in response to infection.

Overall, this study has demonstrated the reliability of biochip array technology, in the context of clinical diagnostics, where systemic measurement of multiple cytokines could be more informative and useful, such as combination therapy of HIV infection, involving cytokines (e.g. IL-2) or cytokine inhibitors (anti-TNF-α, anti-NFκB, anti-IL-1β etc) besides monitoring disease progression or response to antiretroviral therapy. The results of cytokine quantitation by biochip array and ELISA confirmed that this technology is not only comparable but also has a good potential in the future, in applications involving measurement of multiple cytokines with limiting specimens.

Acknowledgment

The authors are thankful to Angela Wilson, Randox Laboratories Ltd, Crumlin, UK for providing biochip array kits.

Ancillary