PI3Kδ and primary immunodeficiencies

Abstract

Primary immunodeficiencies are inherited disorders of the immune system, often caused by the mutation of genes required for lymphocyte development and activation. Recently, several studies have identified gain-of-function mutations in the phosphoinositide 3-kinase (PI3K) genes PIK3CD (which encodes p110δ) and PIK3R1 (which encodes p85α) that cause a combined immunodeficiency syndrome, referred to as activated PI3Kδ syndrome (APDS; also known as p110δ-activating mutation causing senescent T cells, lymphadenopathy and immunodeficiency (PASLI)). Paradoxically, both loss-of-function and gain-of-function mutations that affect these genes lead to immunosuppression, albeit via different mechanisms. Here, we review the roles of PI3Kδ in adaptive immunity, describe the clinical manifestations and mechanisms of disease in APDS and highlight new insights into PI3Kδ gleaned from these patients, as well as implications of these findings for clinical therapy.

Figure 1. BCR signaling

PI3Kδ is a heterodimeric enzyme, typically composed of a p85α regulatory subunit and a p110δ catalytic subunit. In B cells, PI3Kδ is activated upon cross-linking of the BCR, after stimulation with IL-4 or by the chemokine CXCL13 via CXCR5. The BCR co-opts the co-receptor CD19 or the adapter protein BCAP, both of which have YXXM motifs to which the p85α SH2 domains can bind. The IL-4R co-opts IRS1, which also has YXXM motifs. The mechanism whereby CXCR5 is coupled to PI3Kδ remains to be defined (indicated by a dotted line). PI3Kδ signalling through AKT promotes the activation of mTOR and suppresses FOXO1 function (via phosphorylation-dependent nuclear export). FOXO1 is a transcription factor that activates the genes encoding RAG proteins involved in V(D)J recombination, IKAROS which is required for early B cell development, CD62L which is required for homing to lymph nodes and AID, which is required for CSR and SHM. The amino acid sensor mTOR contributes to the growth and proliferation of B cells. All proteins coloured in green have been affected by LOF mutations causing PID. Of these, only p85α and p110δ have also been affected by GOF mutations causing APDS.

Figure 2. TCR signaling

PI3Kδ is a heterodimeric enzyme, typically composed of a p85α regulatory subunit and a p110δ catalytic subunit. In T cells, the TCR, the costimulatory receptor ICOS and the IL-2R can activate PI3Kδ. ICOS contains a YXXM motif in the cytoplasmic domain which is essential for ICOS-mediated co-stimulation. Precisely how the TCR activates PI3Kδ remains incompletely understood, though TCR ligation is known to induce ZAP70-mediated phosphorylation of LAT. Whether PI3K binds LAT directly or via other adapter proteins remains to be established. Mechanisms of PI3Kδ activation downstream of IL-2R are even less clear, but a role for JAK3 has been implicated. PI3Kδ contributes to the downregulation of the expression of IL-7Rα and CD62L,via the AKT-dependent inactivation and nuclear export of FOXO1, preparing the T cell to exit the lymph nodes and circulate through the vascular systems and organs. PI3Kδ also increases metabolism and contributes to T cell effector-associated phenotypes by promoting activation of mTOR.

Figure 3. Dynamic regulation of PI3Kδ signaling in the immune system

PI3Kδ activity needs to be dynamically regulated for normal immune cell function, as some cell types and processes require high PI3Kδ activity, while other depend on low PI3Kδ activity (e.g., if they require FOXO1-dependent gene transcription). Problems arise if cells cannot increase or suppress PI3Kδ due to mutations, and have chronically low or high PI3Kδ activity. Immunosuppression is associated with loss-of function and gain-of-function in PIK3CD, which encoded the PI3Kδ subunit p110δ. Illustrated are some key cell types and processes affected by high or low PI3Kδ activity, and the consequences of being locked in one state or the other. In the healthy state (top), PI3Kδ signalling is low in naïve and memory T cells, which are characterised by low mTOR and metabolic activity and high expression of FOXO1-dependent lymph node homing receptors. In activated effector T cells, PI3Kδ activity is high as a consequence of TCR, IL2R and ICOS signalling. Effector T cells are also characterised by high mTOR and metabolic activity, whereas FOXO1-dependent expression of lymph node homing receptors is reduced. Inhibition of PI3K signalling during thymic development is thought to favour the development of Treg. However, PI3Kδ activity is required to maintain normal numbers of Treg cells in the peripheral lymphoid tissues and for Treg to adopt an effector phenotype, especially peripheral tissues. Maintainenance of low-level signaling (also referred to as tonic signalling) via PI3Kδ (and to a lesser extent PI3Kα) maintains survival of naïve follicular B cells. Upon activation, PI3K is increased and this contributes to B cell proliferation. However, for B cells to undergo CSR in the GC, PI3Kδ signalling needs to be tuned back down to allow higher FOXO1 transcription and proper AID targeting. In disease states (bottom) caused by gain-of-function or loss-of-function in PI3Kδ, the proper dynamics of signalling result in cellular defects associated with immunodeficiency. Chronically high PI3Kδ activity leads to T cell (more senescence, death, and Tregs) and B cell (more transitional B cells and less CSR and SHM) abnormalities with increased susceptibility to B cell lymphoma, infections, and lymphoproliferative disease. Chronically low PI3Kδ activity leads to a different set of T cell (poor responses and low Tregs) and B cell (low numbers) abnormalities resulting in prevalent infections and colitis.

Figure 4. APDS mutations lower the threshold of PI3Kδ activation

a | Schematic diagram of the protein domains in the p85α regulatory and p110δ catalytic subunits with mapped interactions shown with lines, where the black line indicates the binding interaction mediated constitutive interaction and the red lines indicate inhibitory contacts. The locations of the described amino acid substitutions caused by APDS mutations are indicated. ABD: adaptor-binding domain, RBD: RAS-binding domain, SH3: SRC-homology 3 domain, P: proline-rich region, BH: breakpoint-cluster region homology domain, SH2: SRC-homology 2 domain, N-: N-terminal, C-: C-terminal, i-: inter-. b | Class IA PI3Ks are activated by their recruitment to tyrosine kinase-associated receptors at the plasma membrane. The p85α regulatory subunit (p50 fragment containing the N-SH2–i-SH2–C-SH2 domains, shown here in blue) stabilizes the p110δ catalytic subunit (orange) through constitutive binding of the p85α i-SH2 domain (coiled portion) to the p110δ adaptor-binding domain (ABD). Binding of the p85α SH2 domains to tyrosine-phosphorylated residues on an activated receptor releases the inhibitory contacts between the p85α SH2 domains and the p110δ C2, helical and kinase domains (shown in red in part a). It is possible that the ΔEx11 mutation (red) that truncates the p85α inter-SH2 domain affects p110δ more than it affects p110α, hence the lack of more dramatic pleiotropic effects on growth and metabolism in individuals with this deletion. Ras–GTP further tethers p110δ to the membrane by binding to the Ras-binding domain (RBD) of p110δ. GOF mutations in PIK3R1 and PIK3CD increase kinase activity by interfering with inhibitory interactions between the p85α regulatory and p110δ catalytic subunit (ΔEx11, N334K, C416R and E525K), or by increasing the affinity of p110δ for the plasma membrane (E1021K). The E1021K mutations may also interfere with inhibitory contacts from the p85α C-SH2 domain. See ref (17) and references therein for further details of the structures and mechanisms of regulation of PI3Kδ.