- Email: [email protected]
Validation of a miniaturized assay based on IFNg secretion for assessment of specific T cell immunity
Journal of Immunological Methods 355 (2010) 68–75
Contents lists available at ScienceDirect
Journal of Immunological Methods j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j i m
Research paper
Validation of a miniaturized assay based on IFNg secretion for assessment of specific T cell immunity Giusi Li Pira a,⁎, Federico Ivaldi b, Paolo Moretti b, Marco Risso b, Gino Tripodi b, Fabrizio Manca b a b
Advanced Biotechnology Center, Largo Benzi 10, 16132, Genoa, Italy G. Gaslini Institute, Largo Gaslini 5, 16148, Genoa, Italy
a r t i c l e
i n f o
Article history: Received 14 January 2010 Received in revised form 22 February 2010 Accepted 23 February 2010 Available online 1 March 2010 Keywords: T cell assay Cytokine secretion Recall antigens Cell-ELISA Assay validation
a b s t r a c t A miniaturized method for detection of antigen induced secretion of IFNg by specific T cells cultured in 384 well plates has been recently reported. In order to confidently apply this assay to clinical investigations for monitoring of specific T cell immunity, an intralaboratory validation study has been undertaken. High reproducibility and linearity of reference curves was demonstrated. Consecutive replicate experiments handled by different operators using broad panels of recall antigens were reproducible when tested on individual biological samples. Kinetics of IFNg secretion with different antigens showed a plateau after 24 h culture. Similar trends were observed with secretion of TNFa, GM-CSF and IL17, suggesting that the same kinetics can be applied if other cytokines are tested with this assay. It was demonstrated that frozen-thawed cells can be tested by cell-ELISA and that when PBMC are replaced by whole blood similar reactivity profiles were observed even though cytokine concentration was lower. T cell responses were higher in round bottom than in flat bottom wells, but these plates could not be applied to cell-ELISA as clear plates are not available for scanning. In conclusion, the assay proved flexible, since plates can be frozen at different times during the process, fresh or frozen PBMC and PBMC or whole blood could be used, and robust, since reproducibility was remarkable even when different operators performed the procedures. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Analysis of pathogen specific T cell immunity is a relevant parameter for monitoring infected or immunocompromised patients, vaccinees or vaccine candidates. Numerous methods are currently available, as recently reviewed (Thiel et al., 2004; Kern et al., 2005; Li Pira et al., 2007a,b,c). Since the number of PBMC required for testing is a limiting factor with these assays, thereby limiting the number of antigens that can be screened, we developed a miniaturized assay in 384 and in 1536 well plates that may overcome some of these limitations (Li Pira et al., 2007a,b,c, 2008). The assay, named cell-ELISA, is based on PBMC cultures in wells precoated with an antic⁎ Corresponding author. Cellular Immunology Laboratory, Advanced Biotechnology Center, Largo Benzi 10, 16132 Genoa, Italy. Tel./fax: + 39 010 5737370. E-mail address: [email protected] (G. Li Pira). 0022-1759/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2010.02.010
ytokine antibody and containing predispensed antigens. After incubation, cytokines released by antigen activated specific T cells are captured by the solid phase antibody and eventually revealed by a conventional miniaturized ELISA procedure. After reporting the methodological details and potential applications (Li Pira et al., 2007a,b,c, 2008), we tackled the validation issue using the 384 well plate format with the aim of proposing cell-ELISA as a routine and reliable high throughput functional assay for specific T cell immunity. 2. Materials and methods 2.1. Instrumentation The 96-channel Hydra II (Matrix Technologies, Hudson, NH) was used for distribution of antigens from 96 well master plates to 384 well plates (NUNC Roskilde, Denmark, #164688). Round bottom 96 well plates were from NUNC (#163320) and
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
round bottom 384 well plates were from Corning (Lowell, MA, #3671). Small volume 384 well plates (25 ul working volume instead of 50 ul as for standard 384 well plates) were from Greiner (Frickenhausen, Germany, #691161). The 8-channel dispensers MultiWell (Matrix) and MultiDrop Combi (Thermo) were used to distribute reagents and cells in plates. The ELx800 (Biotek, Winooski, VE) or the Victor3 (Perkin Elmer, Waltham MA) scanners were used for plate readout. 2.2. Reagents Phosphate buffered saline (PBS) and RPMI 1640 were purchased from BioWhittaker, Verviers, Belgium. RPMI was enriched with L-glutamine 10 mM and with 5% fetal calf serum selected for low background in the cell-ELISA assay (complete medium). Antibody pairs for detection of IFNg and other cytokines were from Mabtech (Stockholm, Sweden). An IFNg containing supernatant used as reference standard was obtained by culturing PBMC at 2 × 106/ml in complete medium with PHA at 1 ug/ml for 48 h. After spinning, the supernatant was aliquoted and frozen at −20 °C. The IFNg concentration of the supernatant used as secondary standard (24 ng/ml) was determined by ELISA using the IFNg kit (Mabtech) and the primary standard provided by the manufacturer. Recall antigens included CMV lysate, pools of immunodominant CD4 and C8 peptides from CMV pp65, tetanus toxoid TT, PPD, rd-ESAT6 and Ag85 of M. tuberculosis, heat inactivated bodies of C. albicans, A. niger, A. fumigatus, C. neoformans, lysates of P. carinii and T. gondii. These antigens were obtained from commercial sources or prepared in house as described in detail (Li Pira et al., 2007a,b, 2004, 2005). Synthetic peptides were produced by INBIOS
69
(Naples, Italy) and by JPT (Berlin, Germany). Automatic 8 channel and 96 channel dispensers were used for plate preparation, antigen dispensing, cell seeding and plate development (Li Pira et al., 2007a,b,c, 2008). 2.3. PBMC preparation Blood was obtained from healthy donors who gave their informed consent. PBMC were separated either from leukopacks or from heparinized venous blood by conventional Ficoll gradient and brought to 2 × 106/ml in complete medium. PBMC were frozen in 10% DMSO in FCS and stored in liquid nitrogen. PBMC were thawed and washed twice in cold medium before counting. 2.4. Cell-ELISA The cell-ELISA method, based on an assay originally designed for mouse splenocytes in 96 well plates (McKinney et al., 2000) has been described in detail for 384 well plates and 1536 well plates (Li Pira et al., 2007a,b,c, 2008). Briefly, wells were coated with the first anti IFNg antibody and after washing each well received 20 ul complete medium and 5 ul antigen as 10×. Plates with predispensed antigens were sealed and stored at −20 °C. PBMC were automatically dispensed in thawed plates at 25 ul per well. Volumes were reduced by 50% when small volume 384 well plates were used. The plates were incubated for 36 h in a 5% CO2 atmosphere and frozen at −20 °C. For cell-ELISA development, plates were thawed, washed and developed using a second biotinylated antibody, followed by streptavidin-alkaline phosphatase and
Fig. 1. Reproducibility of titration curves in cell-ELISA. Panel A: four representative titration curves out of 25 generated over a twelve month period. Panel B: mean and SD of the R2 values for the 25 titration curves. Panel C: mean OD values and SD for the four IFNg dilutions determined on the 25 titration curved.
70
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
Fig. 2. Reproducibility of PBMC cultures in cell-ELISA. Cultures were performed using 384 replicates for each condition. Panel A: results obtained with PBMC stimulated with PHA using a manual (multichannel pipette) or automated (WellMate) cell dispensing procedure. Panel B: Panel B: PBMC from a PPD responsive subject were stimulated with PPD using two automated dispensers (MultiDrop and WellMate). Panel C: as in Panel B, with cultures run in small volume 384 well plates instead of standard 384 well plates.
3. Results
at 37 °C for 36 h, followed by freezing and thawing for development. Four representative curves out of 25 are shown in Fig. 1, panel A. The plots demonstrate remarkable linearity, with an average R2 for the 25 curves of 0.95 (SD +/−0.04), as shown in panel B. Panel C shows the mean OD405 for each of the four IFNg concentrations, as determined in the 25 curves.
3.1. Reproducibility of cell-ELISA titration curves
3.2. Reproducibility of culture replicates
Titration curves were developed over a twelve month period using a single batch of plates that already contained the predispensed standard dilutions. Titration curves were handled as cell-ELISA samples, with a preliminary incubation
Automatic and manual 8 channel dispensers were used to seed 384 replicates of PBMC. Fig. 2, panel A, shows the mean OD value of the control replicates (PBMC without stimulation) and of the positive replicates (PHA stimulated PBMC).
para-nitro-phenyl-phosphate PNPP as substrate. Plates were scanned after 1 h incubation at room temperature. Results are shown as OD405 × 1000 or as cytokine concentration referred to a titration curve.
Fig. 3. Comparison of fresh vs. frozen PBMC. Fresh and frozen/thawed PBMC from four donors were compared in response to stimulation with recall antigens. The upper panels show the results obtained with two responders (IFNg N5 ng/ml to all tested antigens). The middle panels show the results obtained with two non responders (IFNg b 1 ng/ml to all tested antigens). The bottom panels show the comparison of fresh PBMC and PBMC stored in liquid nitrogen for one month.
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
71
Fig. 4. Reproducibility of the assay performed by different operators. Operators 1–4 set up cell-ELISA assays using two donors. The blood samples were divided in four aliquots and all phases of the protocol were independently carried out by the four operators, including the density gradient preparation of PBMC. The plates were also developed independently at the end of the culture period, after freezing and thawing. Negative controls in the absence of antigen were b 1 ng/ml IFNg in all cases and thus they do not appear on the log scale that ranges from 1 to 1000 ng/ml.
The SD corresponding to a SE b 0.1% shows the high reproducibility of the replicates dispensed either automatically or manually. For a more accurate evaluation, we also tested antigen stimulated PBMC, as shown in panel B. PBMC were stimulated with PPD and seeded using two different automated 8 channel dispensers. Also in this case a remarkable interwell reproducibility was observed, with identical results using the two instruments. Small volume 384 well plates were also tested by using 25 ul culture volume instead of 50 ul, as used for conventional 384 well plates. In this case interwell reproducibility was less accurate, as expected, with 5% and 9% SE for antigen stimulated PBMC seeded using the two automated dispensers (Panel C). It should be noted that reduction by about 50% in OD in comparison to standard 384 well plates was due to the smaller volume of substrate in these wells, that corresponds to a 50% reduction of the light path during scanning. With this consideration in mind, the overall OD is comparable in the two culture systems.
3.3. Fresh vs frozen PBMC Fresh PBMC were compared in parallel cultures with the same PBMC frozen and stored in liquid nitrogen for 4 h before thawing. Previous data (not shown) suggested that frozen PBMC are less efficient than fresh PBMC on a per cell basis, therefore we used a concentration of frozen PBMC 1.5 higher than that of fresh PBMC to balance for this defect, based on previous titrations (not shown). The upper panels in Fig. 3 show two high responders defined as producers of N5 ng/ml IFNg after stimulation with all of the tested recall antigens. The middle panels show the profiles of two low responders. In all cases positive or negative responses are clearly concordant with both fresh and frozen PBMC. The bottom panels show the profiles of one low and one high responder, whose PBMC stored in liquid nitrogen for one month were compared with fresh PBMC. Also in this case, profiles are similar. Results are given here as ng/ml IFNg.
Fig. 5. Role of vessel shape. A subject with low PPD response was selected for this experiment. Cultures were performed in flat and round bottom wells using 96 and 384 well plates. Three different cell concentrations were tested, using serial 1/2 dilutions of the standard cell concentration. After development in the round bottom cell-ELISA wells, the colored product was transferred with the 96 channel Hydra liquid handler to flat bottom plates for scanning.
72
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
Fig. 6. Kinetics of IFNg secretion. The kinetics of IFNg secretion in response to a panel of different antigens was tested on two subjects (upper and lower panels). Left panels: CD4 and CD8 peptide pools derived from CMV pp65 protein. Middle panels: TT, mycobacterial antigens (PPD, rd-ESAT6, Ag85), PHA. Right panels: Candida, A. fumigatus, A. niger, C. neoformans, P. carinii, and T. gondii. IFNg concentrations were measured after culture for different time intervals. The baseline value of 0.01 ng/ml corresponds to lowest cytokine concentration in the titration curve and it overlaps with the background.
3.4. Reproducibility of PBMC cultures
3.5. Role of vessel geometry
In order to test reproducibility of the assay performed by different operators, two fresh blood samples were divided in four aliquots and processed by four technicians. Processing included PBMC separation and counting, plate seeding and plate development. Fig. 4 demonstrates a remarkable reproducibility even when different operators performed the assay. This was the case for all antigens tested. The SE of the mean cytokine concentrations obtained by the four replicate assays (not shown) never exceeded 4.3%, indicating that trained technicians can produce reliable cell-ELISA results.
Round bottom wells facilitate cell to cell contacts, thereby increasing the chances for a specific T cell to encounter an APC. Therefore we compared IFNg secretion induced in flat or in round bottom well plates. For the experiment shown in Fig. 5 we selected a donor known for low response to PPD. His PBMC were stimulated with PPD in flat and in round bottom wells in 96 and in 384 well plates. Since 384 well plates with round bottoms are not commercially available as clear plates, after cell-ELISA development the colored substrate was transferred to clear flat bottom plates for reading. It can be seen that with
Fig. 7. Kinetics of secretion of different cytokines. The kinetics of secretion of IL2, IFNg, TNFa, GM-CSF and IL17 were tested in response to a panel of different antigens using three subjects (A, B, and C). Control responses in the absence of antigen are shown in the upper left panel for each donor. The stimulatory antigens are indicated in each panel. Cytokine concentrations were tested after 16, 24 and 36 h.
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
73
74
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
Fig. 8. Comparison of PBMC vs whole blood. Two blood samples obtained one week apart from two subjects were tested using fractionated PBMC and whole blood. Whole blood was diluted 1/4 in complete medium before dispensing in antigen containing wells. Four different antigen preparations (CMV, TT, PPD, CD4/CD8 peptide pools) were tested. Results are given as OD405.
both 96 and 384 well plates, the round bottoms revealed a positive responses that were hardly detectable when cultures were performed in flat bottom plates. 3.6. Kinetics of cell-ELISA for production of IFNg and other cytokines Monitoring of IFNg secretion was performed by testing two donors with a library of recall antigens. The kinetics of responses to different groups of antigens are illustrated in Panels A–C and D–F of Fig. 6. While most, but not all of the responses were already detectable after a 12 h incubation period, all positive responses reached a plateau after 24 h, with no further increase after 36 h. This was independent of the antigenic stimulation for both donors. Similar kinetics were observed for secretion of other cytokines induced by antigen stimulation using three donors, as shown in Fig. 7, panels A,B,C. Also in this case, cytokines like IL2, TNFa and GM-CSF, IL17, in addition to IFNg, reached a plateau after 36 h most of the times, suggesting a synchronous expression of their genes. IFNg, TNFa and GM-CSF were more consistently detected then IL2 and IL7. Cytokines like IL4, IL5 and granzyme B were not informative, since they were produced at very low levels (not shown). 3.7. Use of whole blood versus PBMC The possibility of using whole blood in place of purified PBMC is an appealing feature of T cell assays. Fig. 8 demonstrates that if similar numbers of mononuclear cells are cultured, based on the assumption that a healthy donor contains about 2 × 106 PBMC per ml blood, whole blood can
be used to provide functional information similar to that obtained with PBMC. In fact, the two donors tested here twice at one week interval (left and right panels) exhibited very similar reactivity profiles using either PBMC or whole blood. The examples given here represent one donor (FM) with higher responses with whole blood and one donor (FI) with higher responses with purified PBMC. Additional experiments (not shown) suggest that, on average, higher responses are seen with purified PBMC. Nevertheless, whole blood proved a convenient and reliable analyte if purification of PBMC is difficult because of special conditions. 4. Discussion Numerous assays are currently available to evaluate antigen specific T cell immunity (Thiel et al., 2004; Kern et al., 2005; Li Pira et al., 2007a,b,c). Their pros and cons have been evaluated and at the present time no assay can be considered as ideal. In fact assays that provide functional and phenotypic information, like intracytoplasmic cytokine staining, are technically demanding, expensive and cell consuming. Several assays can enumerate specific T cells by using functional properties (ICS, ELISPOT) or by exploiting the specificity and restriction of their TCR (multimers), but none of these assays can quantitate the amount of secreted cytokines. The cell-ELISA assay we developed, thanks to its simplicity, miniaturization, suitability to automation and ability to quantitate secreted cytokines, can be added to the arsenal of assays for T cell immunity. In addition, its sensitivity to low frequency of specific T cells overlaps with sensitivity of other assays (Li Pira et al., 2007a,b,c), as determined by using antigen specific T-cell lines or PBMC containing known frequencies of specific T cells defined by multimer staining.
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
IFNg release assays, collectively referred to as IGRA, include ELISPOT and ELISA measurement of secreted IFNg (Pai et al., 2004; Mori, 2009). These two methods are also commercially available as diagnostic kits. Intracytoplasmic staining for IFNg should be included in the same category. Likewise, cell-ELISA can be defined as an IGRA assay, with the advantage of being miniaturized, automated and capable of detecting more than one cytokine. T cell assays are useful for clinical studies to monitor immunocompromised or immunoreconstituted patients (Li Pira et al., 2009), infected patients or vaccinees and vaccine candidates (Kern et al., 2005). Therefore in this study we approached the intralaboratory validation issue of cell-ELISA. We demonstrated a high intra and interassay reproducibility and the optimal kinetics of secreted cytokines. It was also shown that frozen PBMC can be used in place of fresh cells. The possibility of using whole blood in place of purified PBMC can further simplify the assays, even though it was observed a lower response with whole blood. It is likely that by optimizing culture conditions (e.g. well geometry, medium, blood to medium ratio) we can improve whole blood responses and make them closer to purified PBMC. In conclusion, our data suggest that cell-ELISA is a reliable test, that requires few PBMC and thus can be used to test numerous antigens with clinically acceptable blood draws or to test relevant recall antigens in lymphopenic subjects. In addition, cell-ELISA is a robust test that can be performed in different steps with plate freezing in between. Therefore it can also be applied to field studies thanks to its simplicity and to the fact that batches of plates can be prepared and frozen in a central laboratory, distributed to peripheral laboratories for PBMC seeding and culture, frozen again and shipped back to the central laboratory for final development. Finally, if conditions for use of whole blood are made more efficient and comparable to the use of fractionated PBMC, a further progress in assay simplification can be achieved. Acknowledgements This work was supported by Ministry of Health, Rome (Finalized Project 2008 Establishment of a GMP validated
75
Biobank); Ministry of University, Rome (FIRB RBNEOI-RB9B.003, FIRB RBIP064CRT-006); Italian Health Institute, Rome (AIDS 40G.33, 45G.17, 50G.25); CIPE (Rome, 2007), Regione Liguria, Genoa (2007); EU, Bruxelles (grant LSHP-CT-2005-018680).
References Kern, F., LiPira, G., Gratama, J.W., Manca, F., Roederer, M., 2005. Measuring Ag-specific immune responses: understanding immunopathogenesis and improving diagnostics in infectious disease, autoimmunity and cancer. Trends Immunol. 26, 477. Li Pira, G., Bottone, L., Ivaldi, F., Pelizzoli, R., Del Galdo, F., Lozzi, L., Bracci, L., Loregian, A., Palu, G., De Palma, R., Einsele, H., Manca, F., 2004. Identification of new Th peptides from the cytomegalovirus protein pp 65 to design a peptide library for generation of CD4 T cell lines for cellular immunoreconstitution. Int. Immunol. 16, 635. Li Pira, G., Bottone, L., Ivaldi, F., Tagliamacco, A., Fiordoro, S., Ricciardi, A., Barbano, G., Manca, F., 2005. Generation of cytomegalovirus (CMV)specific CD4 T cell lines devoid of alloreactivity, by use of a mixture of CMV-phosphoprotein 65 peptides for reconstitution of the T helper repertoire. J. Infect Dis. 191, 215. Li Pira, G., Ivaldi, F., Bottone, L., Koopman, G., Manca, F., 2007a. Helper function of cytolytic lymphocytes: switching roles in the immune response. Eur. J. Immunol. 37, 66. Li Pira, G., Ivaldi, F., Bottone, L., Manca, F., 2007b. High throughput functional microdissection of pathogen-specific T-cell immunity using antigen and lymphocyte arrays. J. Immunol. Methods 326, 22. Li Pira, G., Kern, F., Gratama, J., Roederer, M., Manca, F., 2007c. Measurement of antigen specific immune responses: 2006 update. Cytom. B Clin. Cytom. 72, 77. Li Pira, G., Ivaldi, F., Dentone, C., Righi, E., Del Bono, V., Viscoli, C., Koopman, G., Manca, F., 2008. Evaluation of antigen-specific T-cell responses with a miniaturized and automated method. Clin. Vaccine Immunol. 15, 1811. Li Pira, G., Kapp, M., Manca, F., Einsele, H., 2009. Pathogen specific T-lymphocytes for the reconstitution of the immunocompromised host. Curr. Opin. Immunol. 21, 549. McKinney, D.M., Skvoretz, R., Qin, M., Ishioka, G., Sette, A., 2000. Characterization of an in situ IFN-gamma ELISA assay which is able to detect specific peptide responses from freshly isolated splenocytes induced by DNA minigene immunization. J. Immunol. Methods 237, 105. Mori, T., 2009. Usefulness of interferon-gamma release assays for diagnosing TB infection and problems with these assays. J. Infect. Chemother. 15, 143. Pai, M., Riley, L.W., Colford Jr., J.M., 2004. Interferon-gamma assays in the immunodiagnosis of tuberculosis: a systematic review. Lancet Infect. Dis. 4, 761. Thiel, A., Scheffold, A., Radbruch, A., 2004. Antigen-specific cytometry—new tools arrived! Clin. Immunol. 111, 155.
Contents lists available at ScienceDirect
Journal of Immunological Methods j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j i m
Research paper
Validation of a miniaturized assay based on IFNg secretion for assessment of specific T cell immunity Giusi Li Pira a,⁎, Federico Ivaldi b, Paolo Moretti b, Marco Risso b, Gino Tripodi b, Fabrizio Manca b a b
Advanced Biotechnology Center, Largo Benzi 10, 16132, Genoa, Italy G. Gaslini Institute, Largo Gaslini 5, 16148, Genoa, Italy
a r t i c l e
i n f o
Article history: Received 14 January 2010 Received in revised form 22 February 2010 Accepted 23 February 2010 Available online 1 March 2010 Keywords: T cell assay Cytokine secretion Recall antigens Cell-ELISA Assay validation
a b s t r a c t A miniaturized method for detection of antigen induced secretion of IFNg by specific T cells cultured in 384 well plates has been recently reported. In order to confidently apply this assay to clinical investigations for monitoring of specific T cell immunity, an intralaboratory validation study has been undertaken. High reproducibility and linearity of reference curves was demonstrated. Consecutive replicate experiments handled by different operators using broad panels of recall antigens were reproducible when tested on individual biological samples. Kinetics of IFNg secretion with different antigens showed a plateau after 24 h culture. Similar trends were observed with secretion of TNFa, GM-CSF and IL17, suggesting that the same kinetics can be applied if other cytokines are tested with this assay. It was demonstrated that frozen-thawed cells can be tested by cell-ELISA and that when PBMC are replaced by whole blood similar reactivity profiles were observed even though cytokine concentration was lower. T cell responses were higher in round bottom than in flat bottom wells, but these plates could not be applied to cell-ELISA as clear plates are not available for scanning. In conclusion, the assay proved flexible, since plates can be frozen at different times during the process, fresh or frozen PBMC and PBMC or whole blood could be used, and robust, since reproducibility was remarkable even when different operators performed the procedures. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Analysis of pathogen specific T cell immunity is a relevant parameter for monitoring infected or immunocompromised patients, vaccinees or vaccine candidates. Numerous methods are currently available, as recently reviewed (Thiel et al., 2004; Kern et al., 2005; Li Pira et al., 2007a,b,c). Since the number of PBMC required for testing is a limiting factor with these assays, thereby limiting the number of antigens that can be screened, we developed a miniaturized assay in 384 and in 1536 well plates that may overcome some of these limitations (Li Pira et al., 2007a,b,c, 2008). The assay, named cell-ELISA, is based on PBMC cultures in wells precoated with an antic⁎ Corresponding author. Cellular Immunology Laboratory, Advanced Biotechnology Center, Largo Benzi 10, 16132 Genoa, Italy. Tel./fax: + 39 010 5737370. E-mail address: [email protected] (G. Li Pira). 0022-1759/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2010.02.010
ytokine antibody and containing predispensed antigens. After incubation, cytokines released by antigen activated specific T cells are captured by the solid phase antibody and eventually revealed by a conventional miniaturized ELISA procedure. After reporting the methodological details and potential applications (Li Pira et al., 2007a,b,c, 2008), we tackled the validation issue using the 384 well plate format with the aim of proposing cell-ELISA as a routine and reliable high throughput functional assay for specific T cell immunity. 2. Materials and methods 2.1. Instrumentation The 96-channel Hydra II (Matrix Technologies, Hudson, NH) was used for distribution of antigens from 96 well master plates to 384 well plates (NUNC Roskilde, Denmark, #164688). Round bottom 96 well plates were from NUNC (#163320) and
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
round bottom 384 well plates were from Corning (Lowell, MA, #3671). Small volume 384 well plates (25 ul working volume instead of 50 ul as for standard 384 well plates) were from Greiner (Frickenhausen, Germany, #691161). The 8-channel dispensers MultiWell (Matrix) and MultiDrop Combi (Thermo) were used to distribute reagents and cells in plates. The ELx800 (Biotek, Winooski, VE) or the Victor3 (Perkin Elmer, Waltham MA) scanners were used for plate readout. 2.2. Reagents Phosphate buffered saline (PBS) and RPMI 1640 were purchased from BioWhittaker, Verviers, Belgium. RPMI was enriched with L-glutamine 10 mM and with 5% fetal calf serum selected for low background in the cell-ELISA assay (complete medium). Antibody pairs for detection of IFNg and other cytokines were from Mabtech (Stockholm, Sweden). An IFNg containing supernatant used as reference standard was obtained by culturing PBMC at 2 × 106/ml in complete medium with PHA at 1 ug/ml for 48 h. After spinning, the supernatant was aliquoted and frozen at −20 °C. The IFNg concentration of the supernatant used as secondary standard (24 ng/ml) was determined by ELISA using the IFNg kit (Mabtech) and the primary standard provided by the manufacturer. Recall antigens included CMV lysate, pools of immunodominant CD4 and C8 peptides from CMV pp65, tetanus toxoid TT, PPD, rd-ESAT6 and Ag85 of M. tuberculosis, heat inactivated bodies of C. albicans, A. niger, A. fumigatus, C. neoformans, lysates of P. carinii and T. gondii. These antigens were obtained from commercial sources or prepared in house as described in detail (Li Pira et al., 2007a,b, 2004, 2005). Synthetic peptides were produced by INBIOS
69
(Naples, Italy) and by JPT (Berlin, Germany). Automatic 8 channel and 96 channel dispensers were used for plate preparation, antigen dispensing, cell seeding and plate development (Li Pira et al., 2007a,b,c, 2008). 2.3. PBMC preparation Blood was obtained from healthy donors who gave their informed consent. PBMC were separated either from leukopacks or from heparinized venous blood by conventional Ficoll gradient and brought to 2 × 106/ml in complete medium. PBMC were frozen in 10% DMSO in FCS and stored in liquid nitrogen. PBMC were thawed and washed twice in cold medium before counting. 2.4. Cell-ELISA The cell-ELISA method, based on an assay originally designed for mouse splenocytes in 96 well plates (McKinney et al., 2000) has been described in detail for 384 well plates and 1536 well plates (Li Pira et al., 2007a,b,c, 2008). Briefly, wells were coated with the first anti IFNg antibody and after washing each well received 20 ul complete medium and 5 ul antigen as 10×. Plates with predispensed antigens were sealed and stored at −20 °C. PBMC were automatically dispensed in thawed plates at 25 ul per well. Volumes were reduced by 50% when small volume 384 well plates were used. The plates were incubated for 36 h in a 5% CO2 atmosphere and frozen at −20 °C. For cell-ELISA development, plates were thawed, washed and developed using a second biotinylated antibody, followed by streptavidin-alkaline phosphatase and
Fig. 1. Reproducibility of titration curves in cell-ELISA. Panel A: four representative titration curves out of 25 generated over a twelve month period. Panel B: mean and SD of the R2 values for the 25 titration curves. Panel C: mean OD values and SD for the four IFNg dilutions determined on the 25 titration curved.
70
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
Fig. 2. Reproducibility of PBMC cultures in cell-ELISA. Cultures were performed using 384 replicates for each condition. Panel A: results obtained with PBMC stimulated with PHA using a manual (multichannel pipette) or automated (WellMate) cell dispensing procedure. Panel B: Panel B: PBMC from a PPD responsive subject were stimulated with PPD using two automated dispensers (MultiDrop and WellMate). Panel C: as in Panel B, with cultures run in small volume 384 well plates instead of standard 384 well plates.
3. Results
at 37 °C for 36 h, followed by freezing and thawing for development. Four representative curves out of 25 are shown in Fig. 1, panel A. The plots demonstrate remarkable linearity, with an average R2 for the 25 curves of 0.95 (SD +/−0.04), as shown in panel B. Panel C shows the mean OD405 for each of the four IFNg concentrations, as determined in the 25 curves.
3.1. Reproducibility of cell-ELISA titration curves
3.2. Reproducibility of culture replicates
Titration curves were developed over a twelve month period using a single batch of plates that already contained the predispensed standard dilutions. Titration curves were handled as cell-ELISA samples, with a preliminary incubation
Automatic and manual 8 channel dispensers were used to seed 384 replicates of PBMC. Fig. 2, panel A, shows the mean OD value of the control replicates (PBMC without stimulation) and of the positive replicates (PHA stimulated PBMC).
para-nitro-phenyl-phosphate PNPP as substrate. Plates were scanned after 1 h incubation at room temperature. Results are shown as OD405 × 1000 or as cytokine concentration referred to a titration curve.
Fig. 3. Comparison of fresh vs. frozen PBMC. Fresh and frozen/thawed PBMC from four donors were compared in response to stimulation with recall antigens. The upper panels show the results obtained with two responders (IFNg N5 ng/ml to all tested antigens). The middle panels show the results obtained with two non responders (IFNg b 1 ng/ml to all tested antigens). The bottom panels show the comparison of fresh PBMC and PBMC stored in liquid nitrogen for one month.
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
71
Fig. 4. Reproducibility of the assay performed by different operators. Operators 1–4 set up cell-ELISA assays using two donors. The blood samples were divided in four aliquots and all phases of the protocol were independently carried out by the four operators, including the density gradient preparation of PBMC. The plates were also developed independently at the end of the culture period, after freezing and thawing. Negative controls in the absence of antigen were b 1 ng/ml IFNg in all cases and thus they do not appear on the log scale that ranges from 1 to 1000 ng/ml.
The SD corresponding to a SE b 0.1% shows the high reproducibility of the replicates dispensed either automatically or manually. For a more accurate evaluation, we also tested antigen stimulated PBMC, as shown in panel B. PBMC were stimulated with PPD and seeded using two different automated 8 channel dispensers. Also in this case a remarkable interwell reproducibility was observed, with identical results using the two instruments. Small volume 384 well plates were also tested by using 25 ul culture volume instead of 50 ul, as used for conventional 384 well plates. In this case interwell reproducibility was less accurate, as expected, with 5% and 9% SE for antigen stimulated PBMC seeded using the two automated dispensers (Panel C). It should be noted that reduction by about 50% in OD in comparison to standard 384 well plates was due to the smaller volume of substrate in these wells, that corresponds to a 50% reduction of the light path during scanning. With this consideration in mind, the overall OD is comparable in the two culture systems.
3.3. Fresh vs frozen PBMC Fresh PBMC were compared in parallel cultures with the same PBMC frozen and stored in liquid nitrogen for 4 h before thawing. Previous data (not shown) suggested that frozen PBMC are less efficient than fresh PBMC on a per cell basis, therefore we used a concentration of frozen PBMC 1.5 higher than that of fresh PBMC to balance for this defect, based on previous titrations (not shown). The upper panels in Fig. 3 show two high responders defined as producers of N5 ng/ml IFNg after stimulation with all of the tested recall antigens. The middle panels show the profiles of two low responders. In all cases positive or negative responses are clearly concordant with both fresh and frozen PBMC. The bottom panels show the profiles of one low and one high responder, whose PBMC stored in liquid nitrogen for one month were compared with fresh PBMC. Also in this case, profiles are similar. Results are given here as ng/ml IFNg.
Fig. 5. Role of vessel shape. A subject with low PPD response was selected for this experiment. Cultures were performed in flat and round bottom wells using 96 and 384 well plates. Three different cell concentrations were tested, using serial 1/2 dilutions of the standard cell concentration. After development in the round bottom cell-ELISA wells, the colored product was transferred with the 96 channel Hydra liquid handler to flat bottom plates for scanning.
72
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
Fig. 6. Kinetics of IFNg secretion. The kinetics of IFNg secretion in response to a panel of different antigens was tested on two subjects (upper and lower panels). Left panels: CD4 and CD8 peptide pools derived from CMV pp65 protein. Middle panels: TT, mycobacterial antigens (PPD, rd-ESAT6, Ag85), PHA. Right panels: Candida, A. fumigatus, A. niger, C. neoformans, P. carinii, and T. gondii. IFNg concentrations were measured after culture for different time intervals. The baseline value of 0.01 ng/ml corresponds to lowest cytokine concentration in the titration curve and it overlaps with the background.
3.4. Reproducibility of PBMC cultures
3.5. Role of vessel geometry
In order to test reproducibility of the assay performed by different operators, two fresh blood samples were divided in four aliquots and processed by four technicians. Processing included PBMC separation and counting, plate seeding and plate development. Fig. 4 demonstrates a remarkable reproducibility even when different operators performed the assay. This was the case for all antigens tested. The SE of the mean cytokine concentrations obtained by the four replicate assays (not shown) never exceeded 4.3%, indicating that trained technicians can produce reliable cell-ELISA results.
Round bottom wells facilitate cell to cell contacts, thereby increasing the chances for a specific T cell to encounter an APC. Therefore we compared IFNg secretion induced in flat or in round bottom well plates. For the experiment shown in Fig. 5 we selected a donor known for low response to PPD. His PBMC were stimulated with PPD in flat and in round bottom wells in 96 and in 384 well plates. Since 384 well plates with round bottoms are not commercially available as clear plates, after cell-ELISA development the colored substrate was transferred to clear flat bottom plates for reading. It can be seen that with
Fig. 7. Kinetics of secretion of different cytokines. The kinetics of secretion of IL2, IFNg, TNFa, GM-CSF and IL17 were tested in response to a panel of different antigens using three subjects (A, B, and C). Control responses in the absence of antigen are shown in the upper left panel for each donor. The stimulatory antigens are indicated in each panel. Cytokine concentrations were tested after 16, 24 and 36 h.
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
73
74
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
Fig. 8. Comparison of PBMC vs whole blood. Two blood samples obtained one week apart from two subjects were tested using fractionated PBMC and whole blood. Whole blood was diluted 1/4 in complete medium before dispensing in antigen containing wells. Four different antigen preparations (CMV, TT, PPD, CD4/CD8 peptide pools) were tested. Results are given as OD405.
both 96 and 384 well plates, the round bottoms revealed a positive responses that were hardly detectable when cultures were performed in flat bottom plates. 3.6. Kinetics of cell-ELISA for production of IFNg and other cytokines Monitoring of IFNg secretion was performed by testing two donors with a library of recall antigens. The kinetics of responses to different groups of antigens are illustrated in Panels A–C and D–F of Fig. 6. While most, but not all of the responses were already detectable after a 12 h incubation period, all positive responses reached a plateau after 24 h, with no further increase after 36 h. This was independent of the antigenic stimulation for both donors. Similar kinetics were observed for secretion of other cytokines induced by antigen stimulation using three donors, as shown in Fig. 7, panels A,B,C. Also in this case, cytokines like IL2, TNFa and GM-CSF, IL17, in addition to IFNg, reached a plateau after 36 h most of the times, suggesting a synchronous expression of their genes. IFNg, TNFa and GM-CSF were more consistently detected then IL2 and IL7. Cytokines like IL4, IL5 and granzyme B were not informative, since they were produced at very low levels (not shown). 3.7. Use of whole blood versus PBMC The possibility of using whole blood in place of purified PBMC is an appealing feature of T cell assays. Fig. 8 demonstrates that if similar numbers of mononuclear cells are cultured, based on the assumption that a healthy donor contains about 2 × 106 PBMC per ml blood, whole blood can
be used to provide functional information similar to that obtained with PBMC. In fact, the two donors tested here twice at one week interval (left and right panels) exhibited very similar reactivity profiles using either PBMC or whole blood. The examples given here represent one donor (FM) with higher responses with whole blood and one donor (FI) with higher responses with purified PBMC. Additional experiments (not shown) suggest that, on average, higher responses are seen with purified PBMC. Nevertheless, whole blood proved a convenient and reliable analyte if purification of PBMC is difficult because of special conditions. 4. Discussion Numerous assays are currently available to evaluate antigen specific T cell immunity (Thiel et al., 2004; Kern et al., 2005; Li Pira et al., 2007a,b,c). Their pros and cons have been evaluated and at the present time no assay can be considered as ideal. In fact assays that provide functional and phenotypic information, like intracytoplasmic cytokine staining, are technically demanding, expensive and cell consuming. Several assays can enumerate specific T cells by using functional properties (ICS, ELISPOT) or by exploiting the specificity and restriction of their TCR (multimers), but none of these assays can quantitate the amount of secreted cytokines. The cell-ELISA assay we developed, thanks to its simplicity, miniaturization, suitability to automation and ability to quantitate secreted cytokines, can be added to the arsenal of assays for T cell immunity. In addition, its sensitivity to low frequency of specific T cells overlaps with sensitivity of other assays (Li Pira et al., 2007a,b,c), as determined by using antigen specific T-cell lines or PBMC containing known frequencies of specific T cells defined by multimer staining.
G. Li Pira et al. / Journal of Immunological Methods 355 (2010) 68–75
IFNg release assays, collectively referred to as IGRA, include ELISPOT and ELISA measurement of secreted IFNg (Pai et al., 2004; Mori, 2009). These two methods are also commercially available as diagnostic kits. Intracytoplasmic staining for IFNg should be included in the same category. Likewise, cell-ELISA can be defined as an IGRA assay, with the advantage of being miniaturized, automated and capable of detecting more than one cytokine. T cell assays are useful for clinical studies to monitor immunocompromised or immunoreconstituted patients (Li Pira et al., 2009), infected patients or vaccinees and vaccine candidates (Kern et al., 2005). Therefore in this study we approached the intralaboratory validation issue of cell-ELISA. We demonstrated a high intra and interassay reproducibility and the optimal kinetics of secreted cytokines. It was also shown that frozen PBMC can be used in place of fresh cells. The possibility of using whole blood in place of purified PBMC can further simplify the assays, even though it was observed a lower response with whole blood. It is likely that by optimizing culture conditions (e.g. well geometry, medium, blood to medium ratio) we can improve whole blood responses and make them closer to purified PBMC. In conclusion, our data suggest that cell-ELISA is a reliable test, that requires few PBMC and thus can be used to test numerous antigens with clinically acceptable blood draws or to test relevant recall antigens in lymphopenic subjects. In addition, cell-ELISA is a robust test that can be performed in different steps with plate freezing in between. Therefore it can also be applied to field studies thanks to its simplicity and to the fact that batches of plates can be prepared and frozen in a central laboratory, distributed to peripheral laboratories for PBMC seeding and culture, frozen again and shipped back to the central laboratory for final development. Finally, if conditions for use of whole blood are made more efficient and comparable to the use of fractionated PBMC, a further progress in assay simplification can be achieved. Acknowledgements This work was supported by Ministry of Health, Rome (Finalized Project 2008 Establishment of a GMP validated
75
Biobank); Ministry of University, Rome (FIRB RBNEOI-RB9B.003, FIRB RBIP064CRT-006); Italian Health Institute, Rome (AIDS 40G.33, 45G.17, 50G.25); CIPE (Rome, 2007), Regione Liguria, Genoa (2007); EU, Bruxelles (grant LSHP-CT-2005-018680).
References Kern, F., LiPira, G., Gratama, J.W., Manca, F., Roederer, M., 2005. Measuring Ag-specific immune responses: understanding immunopathogenesis and improving diagnostics in infectious disease, autoimmunity and cancer. Trends Immunol. 26, 477. Li Pira, G., Bottone, L., Ivaldi, F., Pelizzoli, R., Del Galdo, F., Lozzi, L., Bracci, L., Loregian, A., Palu, G., De Palma, R., Einsele, H., Manca, F., 2004. Identification of new Th peptides from the cytomegalovirus protein pp 65 to design a peptide library for generation of CD4 T cell lines for cellular immunoreconstitution. Int. Immunol. 16, 635. Li Pira, G., Bottone, L., Ivaldi, F., Tagliamacco, A., Fiordoro, S., Ricciardi, A., Barbano, G., Manca, F., 2005. Generation of cytomegalovirus (CMV)specific CD4 T cell lines devoid of alloreactivity, by use of a mixture of CMV-phosphoprotein 65 peptides for reconstitution of the T helper repertoire. J. Infect Dis. 191, 215. Li Pira, G., Ivaldi, F., Bottone, L., Koopman, G., Manca, F., 2007a. Helper function of cytolytic lymphocytes: switching roles in the immune response. Eur. J. Immunol. 37, 66. Li Pira, G., Ivaldi, F., Bottone, L., Manca, F., 2007b. High throughput functional microdissection of pathogen-specific T-cell immunity using antigen and lymphocyte arrays. J. Immunol. Methods 326, 22. Li Pira, G., Kern, F., Gratama, J., Roederer, M., Manca, F., 2007c. Measurement of antigen specific immune responses: 2006 update. Cytom. B Clin. Cytom. 72, 77. Li Pira, G., Ivaldi, F., Dentone, C., Righi, E., Del Bono, V., Viscoli, C., Koopman, G., Manca, F., 2008. Evaluation of antigen-specific T-cell responses with a miniaturized and automated method. Clin. Vaccine Immunol. 15, 1811. Li Pira, G., Kapp, M., Manca, F., Einsele, H., 2009. Pathogen specific T-lymphocytes for the reconstitution of the immunocompromised host. Curr. Opin. Immunol. 21, 549. McKinney, D.M., Skvoretz, R., Qin, M., Ishioka, G., Sette, A., 2000. Characterization of an in situ IFN-gamma ELISA assay which is able to detect specific peptide responses from freshly isolated splenocytes induced by DNA minigene immunization. J. Immunol. Methods 237, 105. Mori, T., 2009. Usefulness of interferon-gamma release assays for diagnosing TB infection and problems with these assays. J. Infect. Chemother. 15, 143. Pai, M., Riley, L.W., Colford Jr., J.M., 2004. Interferon-gamma assays in the immunodiagnosis of tuberculosis: a systematic review. Lancet Infect. Dis. 4, 761. Thiel, A., Scheffold, A., Radbruch, A., 2004. Antigen-specific cytometry—new tools arrived! Clin. Immunol. 111, 155.