Conventional MHC class I-restricted CD8+T cells (TCD8) play a pivotal role in immune surveillance against spontaneously arising neoplastic cells. They are also targeted in many immunotherapeutic protocols for established and advanced malignancies. Therefore, understanding TCD8 activation cascades and the cellular and molecular mechanisms that govern or modulateTCD8 functions is of prime importance. 

Naïve TCD8 activation is initiated by professional antigen (Ag)-presenting cells (pAPCs), particularly dendritic cells (DCs), which process and load 8-11-amino acid residue-long antigenic peptides onto their MHC I molecules for presentation to naïve TCD8 bearing a T cell receptor (TCR) of unique specificity. This highly specific interaction delivers “signal 1” to TCD8. Signal 1 is necessary but not sufficient for optimal T cell activation, which also requires a second or costimulatory signal provided by other molecular interactions such as those established between CD28 and CD80/CD86. In fact, TCR triggering in the absence of appropriate costimulatory signaling may lead to T cell anergy or even death by apoptosis. 

Naïve anticancer TCD8 are activated through direct priming or cross-priming and also by cross-dressed DCs. Direct priming of TCD8 occurs when signals 1 and 2 can be co-delivered, for instance by tumors of hematopoietic origin. However, many tumors downregulate MHC I as a way to avoid CTL detection or do not express costimulatory molecules by nature. Under such circumstances, naïve TCD8 activation can still occur through cross-priming. Accordingly, DCs capture certain exogenous materials from tumor cells and degrade them to generate peptides, which can then be complexed with their own MHC I and presented in trans (i.e., along with costimulatory molecules) to naïve TCD8. Finally, DCs may acquire “pre-made” peptide:MHC I complexes from dying or dead tumor cells for presentation to TCD8. We coined the term “cross-dressing” to describe this phenomenon (Yewdell and Haeryfar: Annu Rev Immunol 2005). “Cross-dressed” APCs have since been implicated in generation of TCD8 memory and in TCD8 responses to tumor vaccines. Once optimally activated, naïveTCD8 proliferate and differentiate into armed CTLs harboring a potent arsenal of cytotoxic effector molecules (e.g., perforin, granzymes and Fas ligand), which will be employed to destroy tumor targets displaying cognate peptide:MHC I complexes. 

A puzzling feature of TCD8 responses is immunodominance, which dictates that out of thousands of peptides present in complex Ags, only a selected few elicit measurableTCD8 responses of varying magnitude. This establishes a hierarchy among Ag-specific TCD8 clones. Accordingly, immunodominant epitopes provoke robust TCD8 responses, whereas subdominant epitopes activate TCD8 clones that occupy modest ranks in the hierarchy. The reason for immunodominance is unknown. Several factors are known to shape TCD8 hierarchies. These include the abundance of foreign gene products and the efficiency of their degradation by proteasomes, the rate and degenerate specificity of peptide transport into the endoplasmic reticulum (ER), the binding affinity of peptides for MHC I molecules within the ER, and the existence of epitope-specific TCD8 within one’s T cell repertoire. Using multiple models, our team has revealed that immunodominance hierarchies of TCD8 are shaped by the immunosuppressive function of naturally occurring CD4+CD25+ regulatory T (nTreg) cells (Haeryfar et al: J Immunol 2005) and indoleamine 2,3-dioxygenase (IDO) (Rytelewski et al: PLoS ONE 2014), by the template-independent DNA polymerase terminal deoxynucleotidyl transferase (TdT) (Haeryfar et al: J Immunol 2008; Leon-Ponte et al: Immunol Invest 2008), by a serine-threonine protein kinase called the mammalian target of rapamycin (mTOR) [Vareki et al: Am J Transplant 2012) and by cell surface interactions between programmed death-1 (PD-1) and one of its ligands, PD-L1 (Memarnejadian et al: J Immunol 2017). 

mTOR inhibitors are immunosuppressive agents commonly used in the clinic to prevent or delay allograft rejection but also paradoxically exert anticancer activities and are indeed approved for the treatment of certain cancers. Using a mouse model, we discovered that the prototypic mTOR inhibitor rapamycin augments the TCD8 response to a clinically relevant tumor Ag while at the same time attenuating TCD8 alloreactivity [Vareki et al: Am J Transplant 2012). Accordingly, we have proposed that the anticancer properties of mTOR inhibitors are owed, at least in part, to their adjuvant effects on tumor-specific TCD8. We are currently investigating the adjuvanticity of rapamycin analogs in cancer patients. 

PD-1 is a co-inhibitory or “checkpoint” molecule known to mediate TCD8 exhaustion. TCR triggering drives the expression of PD-1 at both transcriptional and translational levels, which subsides once the Ag source is eliminated. However, when the immune system fails to eradicate cancer, for instance in patients with high tumor burden, prolonged antigenic stimulation can lead toTCD8 functional impairments, including exhaustion and anergy. Exhausted and anergic TCD8 are often unable to secrete effector cytokines or launch optimal proliferative and oncolytic responses, which may compromise anticancer immunity and worsen clinical outcomes. Immune Checkpoint Inhibitors (ICIs) that block PD-1-PD-L1 interactions have shown significant therapeutic benefits in numerous malignancies. Unfortunately however, favorable clinical responses to PD-1-based ICIs are not universally achieved in patients with different cancers or even in all patients with the same type of cancer. This highlights the need to better understand how the PD-1-PD-L1 pathway operates to incapacitate antitumor responses. We recently discovered that PD-1 blockade can selectively reinvigorate subdominant anti-tumor TCD8 responses, thus inducing “epitope spreading” (Memarnejadian et al: J Immunol 2017). This is important because: i) subdominant TCD8 are more likely than immunodominant clones to escape tolerance mechanisms and may contribute to protective anticancer immunity; ii) the quality of anticancer host defense is thought to depend not only on the magnitude but also on the breadth of TCD8 responses. In fact, too narrowly focused T cell responses may favor the outgrowth of neoplastic cells that do not display bona fide Ags within heterogeneic tumors. We are now studying the breadth of TCD8 responses in cancer patients receiving ICIs.

Anticancer TCD8 clones recognizing T Ag-derived immunodominant (site IV) and subdominant (site I) epitopes express varying levels of PD-1 (Memarnejadian et al: J Immunol 2017). 


Invariant T cells are unconventional, innate-like T lymphocytes that uniquely recognize non-MHC-restricted Ags and rapidly secrete a wide array of cytokines with remarkable immunomodulatory properties. Consequently, they can shape the nature and course of ensuing immune responses. A significant component of our research is focused on invariant natural killer T (iNKT) cells and more recently on mucosa-associated invariant T (MAIT) cells. 

iNKT cells are rare but potent T cells that recognize glycolipid Ags in the context of a monomorphic MHC I-like molecule called CD1d. In the past few years, we have been investigating iNKT cells’ means and modes of activation and their therapeutic properties. Just to give examples, we identified a costimulatory function for glycosylphosphatidylinositol (GPI)-anchored proteins, typified by mouse Thy-1 and human CD55, during iNKT cell responses to glycolipid Ags (Mannik et al: Immunology 2011), and also demonstrated for the first time that both mouse and human iNKT cells can be directly activated by group II bacterial superantigens (SAgs) (Hayworth et al: Immunol Cell Biol 2012). Furthermore, using wild-type and humanized mouse models, we reported the therapeutic benefits of skewing iNKT cell functions towards an anti-inflammatory T helper (TH)-2-type phenotype in delaying cardiac allograft rejection (Haeryfar et al: Transplantation 2008), in preventing and curing citrulline-induced autoimmune arthritis (Walker et al: Immunol Cell Biol 2012), and in reducing the morbidity of acute intraabdominal sepsis (Anantha et al: Clin Exp Immunol 2014) and SAg-provoked cytokine storm (Szabo et al: J Infect Dis 2017). It is noteworthy that mouse and human iNKT cells are functionally homologous to the extent that mouse iNKT cells can recognize human CD1d and vice versa. In addition, the same glycolipid Ags that activate mouse iNKT cells in experimental settings have shown promise in clinical trials for cancer. Therefore, at least some of the observations made in mouse models of iNKT cell-based therapies may be translatable to humans. Our ongoing investigations include a CIHR-funded project on the therapeutic efficacy of iNKT cell agonists in hyperinflammatory and immunosuppressive phases of sepsis. In a separate set of projects, which is supported by NSERC, we are exploring the impact of psychological stress on iNKT cells, which could in turn compromise the anti-metastatic ability of these cells. 

Costimulatory and co-inhibitory interactions that control iNKT cell responses (van den Heuvel et al: Trends Mol Med 2011). 

Treatment with rapamycin and OCH to skew iNKT cell responses towards a TH2-type phenotype attenuates allografted cardiac tissue damage (Haeryfar et al: Transplantation 2008). 

OCH administration prevents the progression of citrulline-induced autoimmune arthritis in HLA-DR4-transgenic mice. Dashed arrows show areas of synovial hyperplasia and arrowheads point to infiltrating leukocytes. The solid arrow identifies an area of bone reformation (Walker et al: Immunol Cell Biol 2012).

OCH administration reduces SAg-induced morbidity and mortality (Szabo et al: J Infect Dis 2017)

iNKT cell activation pathways in sepsis (Szabo et al: Front Immunol 2015)

More recently, we have extended our studies to MAIT cells, an invariant T cell type that is abundant in humans, but not in mice. MAIT cells see a distinct array of Ags, including but not limited to vitamin B metabolites of bacterial origin, in the context of the monomorphic MHC I-like molecule MR1. Although MAIT cells were discovered in the 1990s, they have come under scrutiny only recently. Our team recently discovered that MAIT cells are a predominant source of interferon (IFN)-γ in toxic shock (Shaler et al: PLOS Biol 2017). We found MAIT cells to be quickly hyperactivated by SAgs in a largely IL-12/IL-18-dependent fashion, followed immediately by their exhaustion, which in turn impeded their cognate antimicrobial functions. We have proposed that this phenomenon may partially explain why systemic exposure to SAgs induces immunosuppression. We are currently evaluating MAIT cell functions in clinical sepsis. Although MAIT cells are commonly viewed as antimicrobial entities, they can be detected within tumor microenvironments. We recently demonstrated that MAIT cells efficiently infiltrate hepatic metastases of colorectal carcinoma but are rendered dysfunctional within patients’ tumor masses (Shaler et al: Cancer Immunol Immunother 2017). We are now exploring the role(s) of MAIT cells in multiple human malignancies and also testing the efficacy of several modalities aimed at restoring or boosting MAIT cells’ pro-inflammatory and direct/indirect tumoricidal functions.

Tumor-inflicted liver (segments II and III) of a patient undergoing two-stage hepatectomy for hepatic metastasis from colorectal cancer (Selby and Hernandez-Alejandro: CMAJ 2014)

Potential MAIT cell roles in tumor promotion/progression and in antitumor immune surveillance (Haeryfar et al: Cancer Immunol Immunother 2018)