Quantifying antigen-specific T cell responses induced by vaccine candidates is a central challenge in immunogenicity monitoring. Among the available methods, the IFN-γ ELISpot assay is widely used for its ability to detect and enumerate individual cytokine-secreting cells, offering sensitivity suited to low-frequency responses. Each secreting cell produces a discrete spot on a coated membrane, and the resulting spot count provides a direct quantitative measure of the frequency of responding cells among PBMC. Its unique characteristic of enabling assessment of immune cell function at a single cell level makes it susceptible to variability arising from multiple sources: sample handling, assay format, and the intrinsic magnitude of the immune responses.
The data presented here comes from IFN-γ ELISpot studies run at KCAS Bio’s Lyon laboratory across three distinct vaccine development contexts, each of which required a tailored approach and generated specific observations on the sources and management of assay variability.
Five vaccine candidates, five indications, three assay formats
Depending on the Context-of-Use, i.e. the type of vaccine, and what is known about the disease, whether it is acute or chronic infectious disease, non-transmissible chronic disease, and anticipated magnitude of immune responses, the assay format will have to be adapted. Across the projects we have been involved in at KCAS Bio, different ELISpot formats were implemented.
For prophylactic vaccines, aiming at protecting the general population from acute viral infections, such as Flu or COVID-19, ex vivo ELISpot successfully quantified circulating T cell effectors, and discriminated pre- from post-vaccination samples, as well as naive from convalescent individuals. This proved also to be true in individuals with an aging immune system. On the other hand, for therapeutic vaccines aiming at restoring immune responses in individuals with either chronic diseases caused by persistent viral infection, such as HPV-induced cervical cancer or chronic hepatitis B, or non-transmissible chronic diseases such as cancer, ELISpot format needs to be adapted. In that specific setting, circulating central memory cells need to be expanded in vitro prior to being quantified by ELISpot.
But getting to that point was no bed of roses.
Vaccine Type
Examples
Format
Prophylactic
SARS-CoV-2
Ex vivo (overnight resting + 24h stimulation)
Prophylactic
Influenza virus
Ex vivo (overnight resting + 24h stimulation)
Therapeutic
HPV16+ cervical cancer
5-day in vitro expansion
Therapeutic
Hepatitis B
10-day in vitro expansion
Personalized cancer vaccine
Solid tumor neoantigens
10-day in vitro expansion
When Positive Samples Are Scarce: Assay Development Off the Beaten Track
Developing and validating these assays can be either straightforward, if you have easy access to antigen-reactive samples, as for influenza virus to which most adults have been exposed during their lifespan, or for SARS-CoV-2 to which most individuals were infected or vaccinated during the COVID-19 pandemic. Or it can get much more challenging in the case of chronic diseases or personalized vaccines in oncology, for which positive samples cannot be easily sourced. In that latter case, one needs to get off the beaten track and think out of the box for the assay development and validation strategy. This can involve using a surrogate antigen for assay development and optimization, and then, confirm that the method is appropriate on in-study samples if anticipated in early discussions with the sponsor. Or it can involve directly developing, optimizing and evaluating the performance of the method on early in-study samples, if this is anticipated with the sponsor, in terms of informed consent and volume of blood. Or using samples from a previous/former clinical study, when available.
Before any of these clinical samples were run, we developed and optimized each assay, evaluating the type of antigen, e.g recombinant protein vs peptide pool for influenza vaccine, peptide concentration, and culture medium (human plasma vs serum-free medium) to identify the best conditions for each context.
What the pre- and in-study variability data showed
As for any cellular assay, the performance of ELISpot needs to be carefully characterized pre-study and monitored in-study to pick up any drift in the method that would impact data interpretation. Naturally, the more complex the method, the greater the variability, e.g. ELISpot after expansion vs ELISpot ex vivo.
Since no reference samples are available for measuring accuracy, we regularly participate in external proficiency panels, as a benchmark against peer laboratories (EQAPOL or CIMT-CIP). Then, pre-study method validation mostly consists of evaluating the precision of the method based on validation samples screened during method development and aims at assuring that clinical samples, that will be analyzed on different days, by different analysts, can still be compared with a maximum of confidence.
These validation samples will then be run in-study, in parallel with clinical samples, included in each analytical plate and monitored in a control chart for assay trending and attest to the assay performance overtime.
But ultimately, even if you have carefully characterized the performance of the method and implemented control charts for the study, if sample integrity cannot be guaranteed, you will still end up with non-interpretable data sets.
Where sample quality made the difference
Across these projects, we experienced different scenarios for sample handling / preparation with notably different outcomes.
The best results came from a PBMC network model, where local laboratories, after being audited, trained and qualified, processed and cryopreserved PBMC within 8 hours of venipuncture using a standardized protocol. Viability and yield were consistently high. When three clinical sites followed the same 8-hour window but without harmonized procedures, viability and yield varied considerably across sites, even though the timing was right. The weakest scenario was a large multi-site study in which samples from 28 US and EU sites were shipped overnight to a central laboratory: despite a standard SOP at the receiving end, shipping time and cold-chain variability degraded cell quality before the assay even began.
The conclusion was clear: standardized local processing outperforms centralized analysis with overnight shipping, and timing alone is not enough without harmonized procedures. Overcoming key challenges, through K.C.A.S.
Deciding when a response is real
Finally, once you’ve secured your samples, confirmed the method performs well, and gathered your sample data, you’re left with THE key question: how do you define a positive response? There are two aspects to this question.
The first aspect is whether reactivity to a specific antigen is higher than background; that can vary a lot between samples. For that, we can rely on empirical rules or statistical tests (Moodie et al., 2010). By default, KCAS Bio applies an empirical rule, and many variations of, after discussion with our sponsors, based on:
- Ratio of specific spots to mock (unstimulated) spots greater than 3 to 4
- Specific minus mock spots exceeding 50 spots per million PBMC
The second aspect is important for experimental antigens that are not expected to generate responses in treatment-naive populations, or to which treatment is supposed to enhance pre-existing responses. In these settings, the aim will be to determine a threshold of positivity above supposedly non-reactive samples. Pre-testing of 20 to up to 100 healthy and / or treatment naive donors is recommended for setting a threshold of positivity (Janetzki, 2024, Patton et al., 2021).
Conclusion
Conducted within the quality management framework of KCAS Bio’s Lyon laboratory, which holds GLP certification and ISO 9001:2015 accreditation, these projects confirm, above all, that sample quality and integrity are what ultimately make or break ELISpot data: no method can rescue a compromised sample. Standardization matters, but it isn’t the whole answer: agility and creativity are just as critical, since no single strategy fits all cases. Each vaccine, indication, and patient population calls for its own tailored approach.
References
To know more about how we can support your vaccine analysis, feel free to contact us here: