T‐cell activation–induced marker assays in health and disease

Cell surface receptors validated for AIM assays with human blood/PBMCs

Numerous cell surface receptors have been identified and authenticated as being able to identify human antigen–specific CD4⁺ and CD8⁺ T cells. However, a variety of marker combinations and stimulation conditions have been published, making reproducibility and standardization difficult. Here we provide a summary and discussion of the advantages and disadvantages of these approaches.

CD25 (IL‐2Rα) is one of the most commonly used cell surface receptors to identify antigen‐specific CD4⁺ T cells. It is important to note that CD25 must always be used in tandem with another marker to define antigen‐specific cells, as it is constitutively present on 0.5–6% of CD4⁺ T cells, being the Tregs. ⁸ , ⁹ Initial upregulation of CD25 is a result of TCR signaling via the nuclear factor‐kappa B pathway, with frequency directly correlated with signaling strength. ⁹ Maintenance of CD25 is mediated by IL‐2–induced STAT5 (signal transducer and activator of transcription 5) signaling, which peaks at 48 h following activation. ⁹ CD25 is also upregulated on CD8⁺ T cells following activation; however, it has not been validated to define antigen‐specific CD8⁺ T cells and, of note, increased levels of CD8⁺CD25⁺ T cells have been reported in some elderly individuals (mean age 75 years). ¹⁰ Instead, the type II C‐lectin receptor, CD69, is more commonly used to identify antigen‐specific CD8⁺ T cells and was the original marker in this field; indeed, an alternate name for CD69 is activation inducer molecule. ¹¹ , ¹² CD69 is upregulated via TCR signaling, and its surface expression peaks at 16 h after activation on CD8⁺ T cells. ¹³ However, signaling through α₄β₁ integrin or via IL‐2, IL‐4, IL‐12, interferon‐(IFN)α and IFN‐γ cytokines through their receptors will increase surface expression of CD69, meaning it also needs to be used in tandem with at least one other marker. ¹³ , ¹⁴ Of note, Tregs are more sensitive to TCR‐independent, CD69 induction signals than non‐Tregs, which may lead to false identification of antigen‐specific Tregs. ¹⁵

CD40 ligand (CD40L; CD154), a member of the tumor necrosis factor (TNF) superfamily, was one of the first surface receptors to be used and validated for AIM assays. CD40L was originally described for CD4⁺ T cells and later expanded for use with CD8⁺ T cells. ⁶ , ⁷ , ¹⁶ CD40L is expressed transiently by CD4⁺ and CD8⁺ T cells in response to antigen stimulation, and CD40L⁺ T cells produce large amounts of cytokines, IL‐2, IL‐4, IL‐5, IL‐10, IFN‐γ and TNF depending on the stimulus. ⁶ , ⁷ However, following binding to its cognate receptor, CD40, CD40L is rapidly downregulated within 6 h of activation. ¹⁷ Activation probably causes CD40L to cycle between the cell surface and cytoplasm as it has been detected intracellularly within HIV‐antigen‐stimulated IFN‐γ‐producing cells in HIV‐infected patients. ¹⁸ To combat this downregulation, a fluorescent anti‐CD40L antibody can be added for the duration of in vitro stimulation. This method was first described by Chattopadhyay et al. where, although CD40L surface expression peaked at 6 h, the protein–antibody complex could be detected up to 24 h post stimulation. ⁶ This technique has been demonstrated on both fresh and previously frozen PBMCs. Of note, monensin, an inhibitor of protein export from the Golgi complex, is required to stabilize the internalized CD40L protein–antibody complexes while use of brefeldin abrogates detection. ⁶ Alternatively, anti‐CD40 blocking antibodies can be used to prevent internalization of CD40L, resulting in similar expression kinetics as the Chattopadhyay method. ⁷ In addition, costaining for additional activation markers, such as CD69, can increase assay sensitivity. ¹⁹

More recently, OX40 (CD134), another member of the TNF receptor superfamily, has been validated in combination with CD25 to identify antigen‐specific CD4⁺ T cells in AIM assays. ²⁰ Unlike other activation‐induced markers, OX40 is not expressed on nonactivated T cells in humans, but similarly its surface expression is driven by TCR signaling, with expression levels peaking at about 48 h after activation. ²¹ OX40 expression levels are tightly controlled by TCR signaling, with additional fine‐tuning via CD28, IL‐2 and IL‐4 signaling. ²² OX40 signaling plays a vital role in T‐cell expansion and survival, as well as T‐helper cell differentiation. Unlike CD40L, which is biased toward expression on T‐helper type 1 (Th1) cells, OX40 is expressed equally on all Th cells, including Tregs, potentially reducing bias when analyzing antigen‐specific T cells. ¹⁶ , ²³ , ²⁴ It is important to note that OX40 is highly expressed on murine FOXP3⁺ Tregs, making it unsuitable for analysis of peripheral murine CD4⁺ T cells. ²⁵

The CD25/OX40 AIM assay has been validated clinically using fresh, heparinized whole blood within 24 h of collection, with analysis following 40–50 h of antigen stimulation. ²⁰ , ²⁶ It is important to note that this whole blood assay is validated for blood collected in tubes containing heparin (sodium or lithium), and although a comprehensive testing of all anticoagulants has not been performed, it is known that ethylenediaminetetraacetic acid and acid–citrate–dextrose inhibit marker upregulation due to inhibiting Ca²⁺ influx; however, PBMCs can be used after isolation from any of these blood collection tubes. ²⁷ This assay correlates with cell proliferation, intracellular cytokine and tetramer staining measures of antigen‐specific T cells, as well as serology. ²⁶ , ²⁸ In comparison to CD69/CD40L and IFN‐γ intracellular cytokine staining assays, the CD25/OX40 AIM assay detects a larger population of antigen‐specific CD4⁺ T cells, including rare, noncytokine‐producing T‐cell subsets such as T follicular helper (Tfh) cells. ²⁶ , ²⁹ This suggests that CD25/OX40 is currently the best published method to capture the full breath of antigen‐specific CD4⁺ T‐cell responses.

Another validated marker for AIM assays is the TNF receptor superfamily member, 4‐1BB (CD137). Activation of T cells via TCR drives upregulation of 4‐1BB, which through binding its ligand 4‐1BBL (CD137L) on antigen‐presenting cells drives T‐cell proliferation and cytokine production. ³⁰ , ³¹ Notably, 4‐1BB can be paired with CD69 to detect antigen‐specific CD8⁺ T cells ³² , with optimal surface expression at 24 h after activation. ³³ Although best validated for detecting CD8⁺ T cells, 4‐1BB is also expressed on CD4⁺ T cells as early as 6 h following activation, and combined with a lack of CD40L expression, could detect stable human Tregs in long‐term culture. ³⁴ , ³⁵ Therefore, within 6 h of activation, when CD40L is only upregulated on nonregulatory cells, the combination of 4‐1BB and CD40L might be able to differentiate Treg and T‐effector antigen‐specific CD4⁺ T‐cell populations.

Programmed death‐ligand 1 (PD‐L1 or CD274) is another early activation marker that can be used in AIM assays. PD‐L1 functions to suppress autoimmunity and, importantly, plays a pivotal role in suppressing antitumor immunity by signaling to programmed cell death protein 1 (PD‐1) on antitumor T cells. As such, PD‐L1/PD‐1 signaling is an immunotherapy drug target in various cancers. ³⁶ When combined with CD25, PD‐L1 was found to be equivalent to OX40 in detecting antigen‐specific Tfh cells. ²⁹ In addition, while CD25/OX40 staining captures the full range of T‐cell subtypes, OX40/PD‐L1 staining has been described to exclude Tregs, thus providing a way to discriminate between Treg and non‐Tregs. ³⁷ It is important to note that PD‐L1 signaling prevents Th1 polarization, directs Th17 differentiation and induces an IFN‐γ‐anergic phenotype. ³⁸ As such, PD‐L1 may identify an anergic and nonfunctional antigen‐specific T‐cell subset, especially when characterizing antitumor responses.

AIM assays using tissue and nonhuman samples

AIM assays have also been adapted for use with nonhuman samples, primarily mouse and nonhuman primates, and with tissue samples. Notably, CD40L has been validated as a marker of murine antigen‐specific circulating CD4⁺ T cells, with CD25 and OX40 also validated for splenocytes and Tfh cells, with detection after 18 h of stimulation. ²⁰ , ³⁹ Similar to human T cells, murine T cells rapidly downregulate CD40L, necessitating the use of CD40 blocking antibody or the addition of anti‐CD40L antibody to culture. ⁴⁰ The combination of CD25 and OX40 to detect antigen‐specific CD4⁺ T cells seems to have the widest validation, including macaque blood and for lymph node and tonsil samples from humans and macaques. ²⁰ , ²⁹ , ³⁹ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ Although AIM assays for murine CD8⁺ T cells are not as well characterized, it is known that CD69 and 4‐1BB surface expression are induced upon activation of murine T cells. Following stimulation, CD69 expression peaks between 24 and 48 h, although expression levels decrease with 2‐month‐old mice compared with 24‐month‐old mice, ⁴⁵ and up to 5% of non‐antigen‐specific T cells express CD69 after stimulation. ⁴⁰ 4‐1BB can be detected on murine CD8⁺ T cells 12 h after stimulation, with expression peaking at about 24 h. ⁴⁶ Of note, AIM assays using human bone marrow, spleen, lung, lung‐associated and gut‐associated lymph node samples have defined antigen‐specific CD4⁺ T cells as CD40L⁺OX40⁺ or 4–1BB⁺OX40⁺ or 4–1BB⁺CD40L⁺ and antigen‐specific CD8⁺ T cells as 4–1BB⁺CD25⁺. ⁴⁷ In addition, while circulating human T cells in blood are CD69^neg, tissue‐resident memory T cells coexpress CD69 and CD103, therefore CD69 should be avoided when using tissue samples (reviewed by Thome and Farber ⁴⁸ ). Because of species‐ and tissue‐specific differences in activation‐induced marker utility and kinetics, assays with nonhuman and tissue samples require careful validation.

Characterization of the antigen‐specific T‐cell response

With the rapid advancement of high‐dimensional flow cytometry, it is possible to pair identification of antigen‐specific T cells with phenotypic analysis and proliferation assays. ⁴⁹ This approach utilizes previously identified surrogate cell surface markers of interest to characterize or sort the antigen‐specific T‐cell response, allowing for deeper insights into adaptive immune responses.

Detection of chemokine receptors on the cell surface has been widely used to identify subsets of T cells ex vivo. When coupled with AIM assays, phenotyping via chemokine receptors becomes a powerful tool in identifying the cell subsets that constitute an antigen‐specific response. Notably, naïve, effector memory, effector memory re‐expressing CD45RA (TEMRA) and central memory T cells can be easily identified via their expression of CD45RA, CD45RO and CCR7 (naïve: CCR7⁺CD45RA⁺CD45RO^neg; effector memory: CCR7^negCD45RA^negCD45RO⁺; TEMRA:CCR7^negCD45RA⁺CD45RO^neg; central memory: CCR7⁺CD45RA^negCD45RO⁺). ⁵⁰

CD4⁺ T cells can be further broken down into T helper (Th) subsets by their surface expression of chemokine receptors. Notably, CXCR3 expression can be used to identify Th1 cells, with the addition of CCR4 and CCR6 used to identify IFN‐γ‐ and IL‐17A‐producing Th17.1 cells (Th1 cells are CCR4^negCCR6^neg). ⁵¹ , ⁵² , ⁵³ In addition, the alternate combinations of these markers can be used to identify Th2 (CXCR3^negCCR4⁺ CCR6^neg) and Th17 cells (CXCR3^negCCR4⁺ CCR6⁺), with Th9 cells proposed to be CXCR3^+/− (losing CXCR3 during acute reaction) CCR4^negCCR6⁺ and Th22 cells to be CCR4⁺CCR6⁺CCL10⁺. ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ When coupled with the CD25/OX40 assay, expression of the ecto‐nucleoside triphosphate diphosphohydrolase CD39 identifies a cell population highly enriched (> 80%) for Tregs, ⁵⁹ noting that a single‐nucleotide polymorphism (rs10748643) determines surface expression levels of CD39. ⁶⁰ However, CD39 alone also identifies a second population that is FOXP3^neg and produces high levels of IFN‐γ and IL‐17A. ⁶¹ CD39⁺ Treg identification may prove to be a useful alternative to intracellularly staining for FOXP3, especially if CD25 is included in the assay. ²⁴ In addition, it is possible to identify circulating Tfh (cTfh) based on CXCR5 and PD‐1 expression. ²⁹ Of note, CXCR5‐expressing CD8⁺ T cells important for the control of autoantibody generation in the germinal center and for shaping the antibody response to viral infections. ⁶² , ⁶³ , ⁶⁴

It is important to note that while studies have shown chemokine receptors to be a useful alternative to cytokine detection, proper validation should be carried out to verify findings. This is especially important in disease cohorts, where alterations to chemokine receptor expression may occur in vivo. Care must also be taken to validate in vitro stimulation–induced changes in lineage‐defining surface markers during the time course of the assay, with one method for this being stimulation of sorted ex vivo cell populations. ⁴⁹ Prolonged stimulation should also be carefully optimized as this may cause changes in surface receptor expression, particularly through inducing cell proliferation and/or death. Validation can be achieved through various methods, including intracellular staining or cell sorting and qualitative multiplex single‐cell RT‐PCR of transcription factors. ⁶⁵ AIM assays are also particularly useful for repertoire analysis of TCR sequences as viable antigen‐specific T cells can be isolated prior to cell division occurring. ⁶⁶

Clinical applications of AIM assays

Detecting SARS‐CoV‐2–specific T cells following infection and vaccination

In response to the COVID‐19 pandemic, there was a rapid development of vaccines, which created an immediate need to evaluate and track adaptive immune responses. especially with the continued emergence of new SARS‐CoV‐2 variants. AIM assays quickly dominated the COVID‐19 T‐cell literature, providing immediate insights into SARS‐CoV‐2–specific responses. These findings have been summarized in Table 1.

The Sette and Crotty groups were among the first to publish data from AIM assay characterization of COVID‐19 T‐cell immunity, discovering that natural infection–induced CD4⁺ and CD8⁺ T cells have a half‐life of 3–5 months and that the magnitude of CD4⁺ and CD8⁺ T‐cell responses negatively correlates with disease severity. ³² , ⁹⁸ These groups also identified SARS‐CoV‐2–specific T cells (from stimulation with HLA‐I– and HLA‐II–predicted peptide megapools) in 40–60% of uninfected individuals, hypothesizing that this was due to cross‐reactivity between SARS‐CoV‐2 and previously circulating coronaviruses. ⁹⁹ They subsequently used AIM assays to map the immunodominant T‐cell–activating epitopes, generating megapools that were provided to the research community and used as a stimulus in AIM assays. ¹⁰⁰ Peptide pools spanning the entirety of the matrix (M), nucleocapsid (N) and spike (S) proteins are also commercially available. The readily available peptide products have allowed for rapid and reproducible evaluation of COVID‐19 vaccine–induced T‐cell responses to both the ancestral and variant viral proteins.

When coupled with phenotypic analysis, AIM assays can reveal T‐cell subsets that have predictive capability. This is especially important when making decisions about redosing populations, especially those who are immunocompromised, as these T‐cell subsets can act as early warning signs of waning antibody levels. AIM assays using CD40L have identified that, following COVID‐19 vaccination, Th1 and Tfh cells are expanded, with the latter correlating with protective humoral immunity. ¹⁰¹ A study of long‐term immunity to SARS‐CoV‐2 found that increased levels of S receptor–binding domain–specific CD25⁺OX40⁺ CD4⁺ T cells associated with higher neutralizing antibody titers, with single‐cell transcriptomics revealing a heterogenous population. ²⁴

More recently, AIM assays have been used to determine vaccine‐induced protection against omicron and other emerging SARS‐CoV‐2 variants. The CD25/OX40 assay indicated that individuals receiving ChAdOx‐1S, Ad26.COV2.S, mRNA‐1273 or BNT162b2 vaccines had no significant differences in the level of antigen‐specific T cells when stimulated with ancestral spike protein as compared with delta, beta and omicron variant spike proteins, while antibody levels against omicron were significantly decreased. ¹⁰² Further, 77% of vaccinated individuals were found to have omicron cross‐reactive CD4⁺ T cells, using CD69 and CD40L AIM assays. ¹⁰³ Most notably, when analyzed independently, T‐cell immunity induced by all vaccine types was highly conserved against all variants, with greater than 80% of CD4⁺ and CD8⁺ memory responses conserved against omicron. ¹⁰⁴

Challenges and future directions

Although most studies show robust vaccine‐induced T‐cell immunity, the lack of standardization and variety of AIM assay used makes data interpretation challenging. A diverse combination of cell surface markers in AIM assays have been used to track COVID‐19 vaccine–induced T‐cell responses. Most commonly, CD25, CD40L and OX40 are used for CD4⁺ T cells and CD69 and 4‐1BB are used for CD8⁺ T cells. However, the differing combinations seem to yield varying results when evaluating the phenotypes comprising the response. For example, studies using CD69, CD40L and CD200 as activation markers for CD4⁺ T cells identify Th1 cells comprising the majority of the T‐cell response. ¹⁰² , ¹⁰⁵ Interestingly, the use of CD40L may be skewing these results. In fact, CD40L⁺ cells have been shown to be enriched for IFN‐γ‐producing Th1 cells. ¹⁸ , ¹⁰⁶ Alternatively, Th17‐like cells, identified using the activation marker CD69 alone, were found to be predominant in the nasal mucosa following SARS‐CoV‐2 mRNA vaccination. ¹⁰⁷ This highlights the heterogeneous nature of antigen‐specific T cells when using different combinations of activation markers. These data indicate that the best practice is to include multiple activation markers when evaluating vaccine‐induced T‐cell responses.