CD28 was identified in early 1980s by Martin and colleagues by monoclonal antibodies recognizing a 44 kDa protein on the surface of human T lymphocytes (Martin et al. 1980). Further experiments from Gmunder and Lessener evidenced the crucial role of this molecule in costimulating T cell responses by synergizing with PHA in inducing T cell proliferation (Gmunder and Lesslauer 1984). Later in 1987, Aruffo and Seed cloned the cDNA of human CD28 and showed that it encodes for a 23 kDa glycosylated type I transmembrane protein expressed as disulfide-linked homodimer on the surface of 80% of human CD4+ T cells and 50% of human CD8+ T cells (Aruffo and Seed 1987). In mouse, CD28 was cloned in 1990 by Gross and colleagues and found to be expressed 100% on both CD4+ and CD8+ T cells (Gross et al. 1990). In the same year, the natural ligands of CD28 were also identified. In 1989, Freeman and colleagues identified B7.1, also known as CD80, on the surface of activated B lymphocytes (Freeman et al. 1989), and further studies from Azuma and colleagues demonstrated that B7.1 binds to CD28 (Azuma et al. 1993). Later, the same group also cloned and identified B7.2 or CD86 as a CD28 binding partner (Freeman et al. 1993).
Since its discovery, it has become clear that CD28 was the most prominent costimulatory molecule able to deliver the signal two necessary for full T lymphocyte activation. The two signal model of T lymphocyte activation predicts that optimal T cell response to antigen is achieved following the recognition of peptide-major histocompatibility complex (MHC) by TCR (signal one) together with a subset of costimuli (signal two), generally provided by counterreceptors expressed on the surface of APCs. Extensive in vitro and animal model studies demonstrated that CD28 delivers signals that complement TCR in both qualitative and quantitative manners, thus promoting/enhancing cell proliferation, high levels of cytokines, cell survival, and T cell differentiation. More recent studies also evidenced the ability of CD28 to function in a TCR-independent manner (Porciello and Tuosto 2016).
The short cytoplasmic tail of CD28 (41 aa in human and 38 aa in mouse) is highly conserved and has no enzymatic activity. However, it contains several tyrosine and proline-based motifs, which bind the Src homology (SH)2 and SH3 domain of intracellular signalling molecules. It contains an N-terminal YMNM motif that following phosphorylation binds the SH2 domain of the p85 subunit of class 1A PI3K and the adaptor proteins Grb2 and Grb2-related adaptor downstream of Shc (GADS). The YMNM motif is followed by two proline-rich regions, the N-terminal PRRP that binds the SH3 domain of the IL-2 inducible kinase (Itk) and a C-terminal motif PYAP that binds important signalling molecules, including Lck, filamin-A, and Vav1 (Porciello and Tuosto 2016) (Fig. 1).
CD28 and the Regulation of IS Formation
CD28 engagement by either agonistic antibodies or its natural ligands CD80/CD86 lowers the T cell activation threshold and leads to the augmentation of TCR signalling events necessary for efficient cytokine production (via augmented transcriptional activity and messenger RNA stabilization), cell cycle progression, survival, and T cell differentiation. Activation of T cells by APCs bearing the appropriate peptide-MHC complexes requires rapid cytoskeletal reorganization events leading to the polarization of membrane receptors and signalling molecules within the contact site between T cell and APC, a process that is referred as the immunological synapse (IS). TCR CD28 and coreceptors segregate into central supramolecular activation clusters (cSMACs), whereas LFA-1 adhesion molecule and CD45 are enriched into the peripheral SMACs. CD28 plays a critical role in immunological synapse (IS). CD28 mediates the rearrangement of membrane lipid rafts, thus generating a dynamic platform at the IS where many signalling proteins are concentrated and protected from phosphatases (Viola et al. 2010). CD28 also enhances the close contact between T cells and APCs and triggers the actin cytoskeleton rearrangement events, which are necessary for the recruitment and the organization of molecular signalling complexes (Acuto and Cantrell 2000). CD28 capability to promote cytoskeleton rearrangement events relies on its ability to recruit Vav1 and filamin-A (Tavano et al. 2006; Muscolini et al. 2015).
Vav1 is a guanine nucleotide exchange factor of small Rho GTPases, Rac1, and Cdc42 and is required for CD28-dependent signals and actin nucleation (Acuto and Michel 2003; Rudd and Schneider 2003). CD28 stimulation promotes Vav1 tyrosine phosphorylation and activation and Vav1 binding through C-terminal PYAP motif (Muscolini et al. 2015), thus bringing to the membrane the N-Wiskott Aldrich syndrome protein (WASp)/actin related protein (Arp)2/3 complex and filamin-A. All these proteins in turn cooperate in inducing cortical actin polymerization.
CD28-dependent regulation of actin dynamic and cytoskeleton reorganization is fundamental for its role of general amplifier of TCR signalling functions (Boomer and Green 2010).
CD28 Contribution to TCR-Signalling Functions
One of the earliest events initiated by TCR recognition of peptide-MHC complexes is the activation of Src family tyrosine kinases p56lck and p59fyn that phosphorylate tyrosine residues of the immunoreceptor tyrosine-based activation motifs (ITAMs) of CD3 and ζ chains. Tyrosine phosphorylated ITAMs bind the Syk family tyrosine kinase Zap-70 that following activation by p56lck and/or p59fyn phosphorylates the linker for activation of T cells (LAT). Tyrosine phosphorylated LAT binds and recruits to the membrane PLC-γ1 and the growth-factor receptor-bound protein (Grb2). PLCγ1 Hydrolyzes phosphatidylinositol 4,5-biphosphate (PIP2) into inositol-1,4,5-trisphosphate (IP3) and diacylglycerol (DAG). Soluble IP3 triggers intracellular Ca2+ mobilization, thus leading to the activation of calcineurin and nuclear factor of activated T cells (NF-AT). TCR ligation in absence of CD28 costimulation results in a strong reduction of both PLCγ1 phosphorylation and release of endoplasmic reticulum Ca2+ into the cytoplasm.
The membrane lipid DAG activates the protein kinase C (PKC)θ and nuclear factor-κB (NF-κB), and CD28 stimulation is essential to trigger the activation of the NF-ĸB pathway by favoring the recruitment of PKCθ to the IS.
The first readout of CD28 costimulatory signals in TCR-mediated T cell activation is the enhancement of IL-2 expression and secretion at both transcriptional and posttranscriptional levels. The IL-2 promoter is, indeed, specifically regulated by NF-AT, NF-κB, and AP-1 transcription factors that, as described above, are all amplified by CD28. Moreover, the IL-2 promoter contains a specific enhancer region, known as CD28-responsive element (CD28RE), which is composed of NF-κB binding sites with adjacent NF-IL-2B AP-1 sites (Verweij et al. 1991; Shapiro et al. 1997). This regulative unit controls CD28 responsiveness in the IL-2 promoter and results in a site for signal integration and thus mutations of RE/AP strongly impair the transcriptional increase of IL-2 induced by CD28 costimulation.
CD28 and NF-κB Activation
NF-κB family consists of transcriptionally active heterodimers containing RelA, c-Rel, or RelB in association with p50 (NF-κB1) and/or p52 (NF-κB2). In most unstimulated cells, inhibitory proteins belonging to the inhibitor of NF-κB (IκB) family, which include IκBα, IκBβ, and IκBε, bind and sequester the active heterodimers in the cytosol. A protein kinase complex containing two serine kinases, IκB kinase (IKK)α and IKKβ, and a third subunit, IKKγ/NEMO, with regulatory functions, phosphorylates IκBs, thus leading to their proteolytic degradation and release of NF-κB into the nucleus. TCR stimulation induces the activation of the canonical RelA/p50 or c-Rel/p50 heterodimers by recruiting PKCθ and the ternary complex caspase recruitment domain membrane associated guanylate kinase protein 1 (CARMA1), Bcl10, and mucosa-associated lymphoid tissue lymphoma translocation protein 1 (MALT1) that links TCR to the IKK complex.
CD28 contributes to the activation of the canonical NF-κB1 pathway by binding the SH2 domain of p85 subunit of class 1A PI3K through the tyrosine phosphorylated YMNM motif. Class 1A PI3K generates the phosphatydilinositol 3,4,5 triphosphates (PIP3) lipids that bind the pleckstrin homology (PH) domains of phosphoinositide-dependent protein kinase 1 (PDK1). PDK1 leads to the activation and membrane recruitment of PKCθ that cosegregates with CD28 to a spatially unique subregion within the IS, thus favoring the activation of CARMA1/Bcl10/MALT1 complex and IKKs. Moreover, PDK1 also contributes to the phosphorylation and activation of Akt that both cooperates with PKCθ in stimulating the canonical NF-κB1 pathway and synergize with CD28 to activate noncanonical NF-κB2 pathway (Tuosto 2011).
The noncanonical NF-κB2 pathway is generally activated by the IKKα activator NF-κB-inducing kinase (NIK) that, by phosphorylating IKKα homodimers, leads to the processing of NF-κB2 and the release of p52-containing heterodimers. CD28 stimulation is able to recruit and activate NIK and IKKα, thus leading to the nuclear translocation and activation of noncanonical NF-κB2 dimers. This CD28 unique capability to activate NF-κB converges to the selective regulation of the expression of several genes, including antiapoptotic and proapoptotic genes of the Bcl-2 family and cytokine/chemokine genes (Tuosto 2011; Porciello and Tuosto 2016).
Since its discovery and on the basis of the high homology between rodent (mouse and rat) and human CD28, for several years, in vivo mouse models have been used for understanding the mechanisms of T lymphocyte activation and differentiation and the role of CD28 costimulation in health and immune diseases. When Thomas Hunig’s group discovered that CD28 superagonistic antibodies (CD28SAbs) were able to preferentially activate and expand immunosuppressive regulatory T cells (Lin and Hunig 2003), an enormous amount of preclinical experiments have been performed to evaluate the potential use of these CD28SAbs to ameliorate the onset, progression, and clinical course of human autoimmune diseases. However, on March 2006, the phase I clinical trial of a humanized CD28SAb (TGN1412) turned in a catastrophe, because this antibody induced a rapid and massive cytokine production (i.e., IFN-γ, IL-1, IL-6, TNF-α), thus causing a severe systemic inflammatory response syndrome in all healthy volunteers (Suntharalingam et al. 2006). These data evidence that despite the enormous progresses made in identifying the mechanisms and molecules involved in CD28 signalling, much remains to be elucidated, especially in the light of the differences in CD28 signalling capabilities between human and mouse.