AIDS Research and Therapy BioMed Central Methodology

Background HIV-1 integrase (IN) is an attractive target for the development of drugs to treat AIDS, and inhibitors of this viral enzyme are already in the clinic. Nevertheless, there is a continuing need to devise new approaches to block the activity of this viral protein because of the emergence of resistant strains. To facilitate the biochemical analysis of wild-type IN and its derivatives, and to measure the potency of prospective inhibitory compounds, a rapid, moderate throughput solution assay was developed for IN-catalyzed joining of viral and target DNAs, based on the detection of a fluorescent tag. Results A detailed, step-by-step description of the new joining assay is provided. The reactions are run in solution, the products captured on streptavidin beads, and activity is measured by release of a fluorescent tag. The procedure can be scaled up for the analysis of numerous samples, and is substantially more rapid and sensitive than the standard radioactive gel methods. The new assay is validated and its utility demonstrated via a detailed comparison of the Mg++- and Mn++-dependent activities of the IN proteins from human immunodeficiency virus type 1 (HIV-1) and the avian sarcoma virus (ASV). The results confirm that ASV IN is considerably more active than HIV-1 IN, but with both enzymes the initial rates of joining, and the product yields, are higher in the presence of Mn++ than Mg++. Although the pH optima for these two enzymes are similar with Mn++, they differ significantly in the presence of Mg++, which is likely due to differences in the molecular environment of the binding region of this physiologically relevant divalent cation. This interpretation is strengthened by the observation that a compound that can inhibit HIV-1 IN in the presence of either metal cofactors is only effective against ASV in the presence of Mn++. Conclusion A simplified, assay for measuring the joining activity of retroviral IN in solution is described, which offers several advantages over previous methods and the standard radioactive gel analyses. Based on comparisons of signal to background ratios, the assay is 10–30 times more sensitive than gel analysis, allows more rapid and accurate biochemical analyses of IN catalytic activity, and moderate throughput screening of inhibitory compounds. The assay is validated, and its utility demonstrated in a comparison of the metal-dependent activities of HIV-1 and ASV IN proteins.


Background
Vaccine development has become more complex in the last decades, pursuing new strategies for stimulating immune responses against infectious agents of viral, bacterial or parasitic origin as well as against cancer. A striking example is the long-winded search for an effective HIV-1 vaccine that would be crucial, together with antiretroviral therapy, to limit and possibly stop the worldwide AIDS pandemic. Several candidate HIV-1 vaccines that aim to stimulate cellular immune responses have been tested in phase I and II clinical trials [1][2][3]. An accurate evaluation of the cellular immune response will be key to select vaccine candidates for successive phase III clinical trials. Therefore, methods that qualify and quantify antigen-specific, functional T cells in a precise, sensitive, and robust way will be essential. At present, the standard assays that are commonly used for this purpose are IFN-γ ELISPOT, HLA class I and class II multimer staining and ICS. The ELISPOT assay is currently considered the gold standard in vaccine trials due to its sensitivity and extensive standardization and validation [4][5][6][7]. In fact, several reports demonstrated that the ELISPOT assay is more sensitive in detecting weak responses when compared to the ICS assay [8][9][10][11], a feature that represents an important advantage for the detection and measurement of the immune response in vaccine trials [12]. The most commonly used ELISPOT assay measures IFN-γ secretion by total PBMC stimulated by specific antigens. Albeit ELISPOT assays being able to measure the secretion of two different cytokines have been recently established [13], it is unlikely that future development will increase the simultaneous measurement of cytokines for this kind of assays. On the other hand, the introduction of new reagents, instruments and software, strongly improved the capacity of flow cytometry based assays such ICS and multimer staining to simultaneously measure several parameters in the same sample [14][15][16]. However, between ICS and multimer staining, the former seems to be more suited to be employed in vaccine trials since it does not require previous HLA typing and a priori knowledge of specific epitopes [17,18]. Hence, it is generally accepted that ICS provides more information regarding the quality of the immune response whereas ELISPOT grants a high capacity of detecting low magnitude responses, while multimer staining is the method of choice for a detailed analysis of the immune response in a selected and limited number of samples.
In spite of an intense activity in the development and testing of new vaccines against HIV-1, clear immunological correlates of protection do not still exist although there is strong evidence that CD4 and CD8 T-cells play a role in the control of viral replication [19]. However, neither the magnitude of the immune response (measured as production of IFN-γ) nor the breadth of the recognised epitopes constitute per se valid correlates of protection [20][21][22]. Recently, studies have shown that polyfunctional CD8 Tcell responses are preferentially observed in long term non-progressors (LTNP) when compared to persons with progressive disease [23]. Furthermore, antigen-specific terminally differentiated CD8 T-cells, defined by the lineage markers CCR7 and CD45RA, have been preferentially found in long-term non-progressors [24] and early infections with future control of HIV-1 viremia [25]. These findings highlight the importance of developing assays able to simultaneously measure several parameters in the same sample and strongly suggest the use of flow cytometry to monitor immune responses.
In this regard, we have developed a 9-colour ICS that allows the simultaneous determination of the function and the memory phenotype of antigen specific CD4 and CD8 T-cells. The assay has the capacity to detect the cytokines IFN-γ and IL-2, the chemokine MIP-1β and the activation marker CD154. For the characterization of the memory phenotype, we used CD45RA, an isoform of a membrane phosphatase that is expressed by both naïve and terminally differentiated T-cells [26]. Here, we compared the sensitivity of our recently established 9-colour ICS with ELISPOT assays performed in two different experienced laboratories. In our experimental setting, taking advantage of the simultaneous detection of IFN-γ and MIP-1β producing T-cells, we demonstrated a similar or superior capacity of our ICS assay to detect low magnitude IFN-γ-mediated responses.

The simultaneous evaluation of IFN-γ+ MIP-1β+ T-cells increases the capacity to detect IFN-γ responses in ICS
The 9 colour ICS assay established in our laboratory is routinely used to measure HIV-1 specific immune responses in different clinical settings. The cumulative analysis of 275 samples obtained from 31 HIV-1 positive individuals stimulated with peptides derived from 5 different HIV-1 proteins (Table 1) revealed an interesting feature of IFN-γ-based responses. Upon antigenic stimulation the majority of the IFN-γ producing CD8 Tcells were also producing MIP-1β (IFN-γ+ MIP-1β+ CD8 T-cells in %: mean ± SD, 0.245 ± 0.6341), whereas CD8 Tcells characterized by the sole production of IFN-γ were rarely detected (IFN-γ+ MIP-1β-CD8 T-cells in %: mean ± SD, 0.016 ± 0.0652) ( Figure 1A and 1B). This trend was observed for all the CD8 T-cell responses whereas the few detected CD4 T-cell responses were more heterogeneous, since antigen-specific cells producing IFN-γ but not MIP-1β were detectable (IFN-γ+ MIP-1β+ CD4 T-cells in %: mean ± SD, 0.014 ± 0.0500; IFN-γ+ MIP-1β-CD4 T-cells in %: mean ± SD, 0.006 ± 0.0328; Figure 1C and 1D). The analysis of T-cells positive for both markers is of particular interest, since the simultaneous evaluation of two functions is supposed to decrease the non-specific background [23]. In order to investigate this observation in our experimental setting, we analyzed 52 mock stimulated samples from 31 HIV-1 positive subjects. Mock stimulated samples were run for each analyzed patient to measure spontaneous cytokine production and unspecific antibody staining. They were processed as the other samples but in the absence of antigenic peptides. The measured background was significantly reduced (around 4-fold lower; p < 0.0001, Wilcoxon matched pairs test) in the IFN-γ+ MIP-1β+ CD8 T-cells when compared to the total IFN-γ+ CD8 T-cells (Figure 2A). Similarly, we observed a 7-fold decrease (p < 0.0001, Wilcoxon matched pairs test) of the non-specific background in IFN-γ+ MIP-1β+ CD4 T-cells when compared to total IFN-γ+ CD4 T-cells ( Figure 2B). Representative plots of responding CD8 T-cells and their negative control are shown in Figure 2C.
A linear regression analysis was performed to examine the correlation between percentages of total IFN-γ+ and percentages of IFN-γ+ MIP-1β+ CD8 and CD4 T-cells in samples stimulated with HIV-1-derived peptides. Percentages of total IFN-γ+ and IFN-γ+ MIP-1β+ CD8 T-cells showed a goodness of fit of r 2 = 0.9929 and a slope of 1.052 demonstrating an almost perfect linearity of the two measurements ( Figure 3A). The goodness of fit was slightly lower for CD4 T-cells; although it was still characterized by an r 2 value of 0.7817 ( Figure 3B). The slope was 1.190, confirming the presence of HIV-1 specific CD4 T-cells producing IFN-γ but not MIP-1β, as previously shown ( Figure 1C and 1D).
Since the numbers of IFN-γ+ MIP-1β+ T-cells were essentially equivalent to those of total IFN-γ+ T-cells whereas the background was strongly decreased in the former, we made the assumption that the evaluation of double positive IFN-γ+ MIP-1β+ T-cells could represent an interesting option to increase the sensitivity of the ICS assay in the detection of IFN-γ mediated HIV-1-specific responses.
In order to compare the sensitivity of the two modalities to evaluate the IFN-γ T-cell response, we analyzed the previously described 275 independent samples ( Figure 4A). We calculated the 90 th percentile of the negative values after background subtraction and this value was considered as a threshold. Samples were considered positive when higher than the threshold and at least 2-fold higher than their respective negative control. In the CD8 T-cell population, 187 positive responses were detected using the IFN-γ+ MIP-1β+ data evaluation, while only 146 positive responses were detected using the total IFN-γ+ data evaluation. The difference was significant performing a Fisher's exact test (p = 0.0005). The difference between positive CD4 T-cell responses calculated using the two modalities was not significant. When CD8 and CD4 T-cell responses were considered together, the difference achieved significance with a p value of 0.0058 (Fisher's exact test). The contingency tables in Figure 4B show that the IFN-γ+ MIP-1β+ data evaluation allowed the detection of 41 CD8 responses that were otherwise missed by eval-uation of the total IFN-γ+ T-cells. As expected, the simultaneous detection of IFN-γ+ and MIP-1β+ did not increase the capacity to detect antigen-specific CD4 T-cell responses. In fact, 11 CD4 responses were exclusively detected by the total IFN-γ+ data evaluation whereas 5 were exclusively observed with the simultaneous detection of IFN-γ+ and MIP-1β+.

Evaluation of IFN-γ+ MIP1β+ cells increases the sensitivity of the ICS in comparison to the ELISPOT
ICS is generally considered less sensitive than ELISPOT in detecting low magnitude responses [8][9][10]. Therefore, we tested whether the simultaneous detection of MIP-1β and IFN-γ might increase its sensitivity in comparison to two ELISPOT assays performed in independent laboratories. Each laboratory used its own ELISPOT method, including IFN-γ and MIP1-β expression in CD8 and CD4 T-cells stimulated with HIV-1-derived antigens

&' &'
Comparison between the ICS and two independently performed ELISPOT assays

WRWDO WRWDO WRWDO WRWDO
IFN-γ+ MIP-1β+ data evaluation were also scored positive using ELISPOT demonstrating an improvement of the concordance between the assays. Laboratory 2 analyzed 29 samples obtained from 3 HIV-1 infected subjects stimulated with 18 different peptide formulations derived from 5 HIV-1 proteins using an ELIS-POT assay approved by the Cancer Vaccine Consortium [27]. As observed in the results generated from the first laboratory, the correlation with the ELISPOT results was significant for both ICS methodologies ( Figure 5B). A total of 29 positive responses were detected by ELISPOT, while 27 and 20 positive responses were detected by ICS using the IFN-γ+ MIP-1β+ and the total IFN-γ+ methods, respectively. Only 2 positive responses were lost by the IFN-γ+ MIP-1β+ data evaluation system, whereas 9 responses were lost by the total IFN-γ+ data evaluation system in comparison to the ELISPOT performed in laboratory 2. These combined results of the 2 laboratories demonstrated that the new evaluation method based on the simultaneous detection of IFN-γ and MIP-1β increased the capacity of the ICS to detect low-magnitude responses.

Influence of the variation in cell number input in the ICS assay
Cell counting is a basic technique in use in all cell culture laboratories. Nevertheless, it constitutes an important source of experimental error [27]. The number of cells per sample is a critical parameter in the ELISPOT assay, since results are directly calculated from the total amount of cells seeded in each well. In contrast, in the ICS assay responding cells are calculated as a percentage of CD4 or CD8 T-cells and therefore the results are independent from the total number of cells used in each experimental sample. However, variation in the cell number might still affect the experimental outcome because of changes in the proportion between the amount of cells, growth factors and stimulants. Therefore, we tested the impact of varying the amount of PBMC per experimental sample in our 9colour ICS assay. Stimulation with 2 different peptides representing optimal CD8 epitopes was performed using 0.45, 0.91, 1.82 and 3.66 million of cells/well, while the amount of peptides was kept constant at 2 μg/ml. There was neither a trend nor a high variation between the results for either the total IFN-γ+ response as well as for

UHVSRQVH LQ RI &' 7 FHOOV
the combined IFN-γ and MIP-1β positive cells ( Figure 6). Of note, the background levels were not affected by the number of cells seeded per well.

Discussion
In the present study, we provide experimental evidence in support of a combined 9-colour IFN-γ and MIP-1β ICS method that, unlike commonly used methods based on the flow-cytometric detection of IFN-γ, achieves sensitivity comparable to that typical of ELISPOT assays.
ELISPOT and ICS assays are widely used to measure specific immune responses in different experimental settings. IFN-γ ELISPOT is considered the gold standard for the evaluation of the immune response in vaccination trials even though accumulating evidence demonstrates that the measurement of a single immunological marker does not provide sufficient information about the efficacy of a specific immune response [9,28,29]. In addition, recent disappointing results of phase III efficacy HIV-1 vaccination trials [30] underscored the need for a better evaluation of the immune response in phase I and II clinical trials. A supporting issue in favour of the use of the IFN-γ ELISPOT as primary assay in vaccine trials is its supposed higher sensitivity in comparison to other immune monitoring assays such as ICS [8][9][10]12]. Here, we demonstrated that the use of a combined detection of IFN-γ and MIP-1β could scale-up the sensitivity of ICS assays to levels comparable to those of IFN-γ-based ELISPOT. In this regard, the key observation is that the majority of the IFNγ producing T-cells are simultaneously producing MIP-1β rendering this new modality of evaluation equivalent to the measurement of the total IFN-γ producing T-cells with the relevant advantage of a consistent decrease of the background that in turn increases the sensitivity of the assay.
It is unlikely that the increased sensitivity of the IFN-γ+ MIP-1β+ data evaluation is due to false positive detections, since simultaneous unspecific binding of two antibodies to the same cell is less probable than unspecific binding of a single antibody. In addition, the majority of samples scored as positive with the IFN-γ+ MIP-1β+ data evaluation were scored as positive using the ELISPOT as well, indicating that the increased sensitivity was not due to a higher number of false positive detections but due to a better capacity of the IFN-γ+ MIP-1β+ data evaluation to discriminate positive responses in comparison to the total IFN-γ+ data evaluation.
Our results provide support for an expanded use of polychromatic flow cytometry as primary assay in vaccine trials. The ICS method optimized in our laboratory allows the simultaneous measurement of several fluorescence markers without losing sensitivity in comparison to the gold standard IFN-γ ELISPOT. In our setting, we used a 9 colour ICS; however, the same method can be applied to any staining combination including IFN-γ and MIP-1β in combination with the appropriate lineage markers. Thus, for investigators with no access to sophisticated flow cytometers, a simplified panel can be used for immunemonitoring purpose as alternative to the ELISPOT not losing sensitivity and with the advantage to discriminate CD4 and CD8 mediated responses. In alternative, more complex staining combinations could be designed for laboratory facilities where complex instrumentation is available, provided the inclusion of the simultaneous measurement of IFN-γ and MIP-1β.
Our present study was limited to the analysis of the HIV-1-specific T-cell responses. Nevertheless, this method can be extended to other specific immune responses if T-cells expressing IFN-γ and MIP-1β represent the majority of the total IFN-γ producing T-cells. In this regard, a possible extension of our methodology is the coupling of an activation marker (i.e. CD69, CD154, etc.) to the measurement of cytokines or chemokines (i.e, IFN-γ, IL-2 and MIP-1β). As a general rule, targeting 2 or more molecules on the same cell population should increase the sensitivity of the assay for the selected cell population. Since flow cytometry has been recently advanced by the development of new instrumentation and reagents, the inclusion of more markers in a single sample should aim not only to increase the amount of information per cell but also to increase the sensitivity for populations of special interest.
Finally, in the present study we demonstrate that the number of cells used in each sample does not affect the readout of the ICS. Since the procedure of manual cell counting is a usual source of experimental error and the number of cells directly affects the ELISPOT readout, our data support the concept of a reduced experimental error associated with the use of ICS assays and strengthens the idea to apply ICS as primary assay in vaccine trials.

Conclusion
The simultaneous detection of IFN-γ and MIP-1β provides a clear advantage for the detection of low HIV-1 specific responses compared to the classical way to analyze the total IFN-γ producing T-cells by ICS. The comparison with the results generated by ELISPOT independently by two experienced laboratories demonstrates that the combined IFN-γ+ MIP-1β+ evaluation system allows for the detection of low HIV-1 specific IFN-γ responses to a similar or even higher extent, as they can be detected using ELISPOT assays. The application of the IFN-γ+ MIP-1β+ method in other diseases and immunological fields remains to be assessed. These findings are important to guide the choice for suitable immune assays and to build reagent panels able to accurately characterize the phenotype and function of responding T-cells in a highly sensitive way.

Study samples
PBMC obtained from 31 HIV-1 infected individuals were analyzed in the present study. Their median CD4 T-cell count was 502 cells/μl (range 229 to 1,042). Twenty-one study subjects were under antiretroviral therapy and 16 of them had undetectable viral load. When detectable the median viral load was 2,581 RNA copies/ml (range 151 to 50,577). Six of the study subjects under antiretroviral therapy with undetectable (< 50 copies of RNA/ml) viral load underwent treatment interruption and their range of viremia was then from 600 to 49,600 RNA copies/ml at the time of sampling.

Peptides
As shown in Table 1 (9) variable length overlapping peptides spanning HIV-1 BH10 Tat. Pool 1 to 6 and 7 to 9 were previously described by Cosma et al [32] and Vardas et al. [33], respectively. Several peptides contained in the pools 7, 8 and 9 were used alone in some experiments. The following peptides corresponding to previously described optimal CD8 epitopes [31] were also used in some stimulation experiments: FLKEKGGL (FL8), TPGPGVRYPL (TL10), YPLTFGWCY and RRQDILDLWIY (RY11). All the peptide pools were tested for specificity in healthy subjects in previous studies [32,33].

Intracellular cytokine staining
Cryopreserved PBMC were used for the ICS assay. After thawing, 10 6  . Incubation was carried out on ice for 30 min and after washing, cells were acquired using an LSRII flow cytometer (Becton Dickinson) equipped with a high throughput system. Sample analysis was performed using FlowJo version 8.5.3 (Tree Star, Ashland, OR). The gating strategy is shown in Additional file 1. Lymphocytes were gated on a forward scatter area versus side scatter area pseudo-colour dot plot and dead cells were removed according to EMA staining. CD3+ events were gated versus IFN-γ, IL-2, MIP-1β and CD154 to account for down-regulation. CD3+ events were then combined together using the Boolean operator "Or". The same procedure was used to subsequently gate CD8+ events. CD4+ events were excluded before creating a gate for each function or phenotype. After background subtraction, the 90 percentile of the negative values was calculated and this value was considered as a threshold. Samples were considered positive when higher than the threshold and at least 2 times higher than their respective mock stimulated control.
ELISPOT assay (laboratory 1) Laboratory 1 used the TriSpot™ Human IFN-γ/IL-2 ELIS-POT Kit (Endogen, Rockford, IL/USA) according to the manufacturer instructions. Briefly, PBMC from ACD whole blood were separated on Lymphoprep™ (Axis-Shield PoC, Oslo, Norway), washed in RPMI medium (RPMI 1640, supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin and 2 mM L-glutamine, all from BioWhittaker Europe, Verviers, Belgium) and counted by Trypan Blue exclusion for assessing viability. After resuspension in complete medium (RPMI medium supplemented with 10% heat inactivated fetal bovine serum, BioWhittaker), PBMC were transferred to the ELISPOT plate with a concentration of 0.8 to 2 × 10 5 cells/well in duplicate. Peptides were added at a final concentration of 3 μg/ml each. PBMC in medium alone or stimulated with phytohemagglutinin (PHA-P, Sigma) at 5 μg/ml were used as negative and positive controls, respectively. Incubation was carried out at 37°C in a 5% CO 2 incubator for 18 hours. The resulting spots were counted using the Automated ELISA-Spot Assay Video Analysis System Eli-Scan with the software Eli.Analyse V4.2 (A.EL.VIS, Hannover, Germany). PBMC from each study subject were mock