Introduction
Class IA Phosphatidylinositol 3-kinases (PI3Ks) are heterodimeric lipid kinases consisting of a catalytic subunit (p110α, p110β, or p110δ) that forms a constitutive association with a regulatory peptide (p85). While p110γ and p110β are widely expressed, leukocytes are the primary source of p110δ expression. Distinct functions for every single p110 isoform have been delineated, mirroring their varying modes of expression and the ways in which they bind to their corresponding receptors. Growth factor, costimulatory, cytokine, and antigen receptors can all activate p110δ, which is highly expressed in both lymphocytes and myeloid cells [
1]. PI3Kδ, which consists of p85α (encoded by PIK3R1) and p110δ (encoded by PIK3CD), operates at the leukocyte plasma membrane. It responds to antigen and cytokine receptor stimulation by phosphorylating phosphatidylinositol 4,5-bisphosphate (PIP2) to produce PIP3. This secondary messenger initiates diverse cascades that involve PDK1, AKT, mTOR and downstream targets such as FOXO1, which then contribute to the metabolic and transcriptional alterations needed for lymphocyte proliferation, differentiation, and effector functions [
2,
3].
PI3K plays a role in both innate and adaptive immunity. In the former it promotes inflammation in human neutrophils after chemotactic peptide challenge, through activation of reactive oxygen species production [
4]. Depending on the context, it can also support or attenuate Toll-like receptor pathways, for example by promoting type I interferon production or by suppressing NFκB-mediated pro-inflammatory transcription [
5]. In B cells PI3Kδ cooperates with the tyrosine kinase BTK, downstream of the B cell receptor, for proper triggering of Ca2 + flux and NFκB activation [
2,
6]. In T cells PI3Kδ is primarily coupled to CD28, and although it is not critical for NFκB signaling, its downstream phosphorylation of AKT facilitates transcription of IL-2 and nuclear localization of NFAT [
7].
PI3Kδ gene mutations are known to cause both gain of function (GOF) and loss of function (LOF) inborn errors of immunity (IEI) in humans. This highlights the critical role that balanced p110δ activity plays in human adaptive immunity [
1,
8]. Dominant-acting germline mutations in PIK3CD result in activated PI3Kδ syndrome (APDS), an immune dysregulatory disease that primarily manifests as childhood recurrent sinopulmonary infections, herpesvirus infections, and CD4 + lymphopenia [
9]. Cells undergo chronic lymphoproliferation, leading to activation-induced cell death and immune deficiency [
2]. For LOF, only 10 patients have been reported to date with germline autosomal recessive deficiency of PIK3CD; however, phenotypic interpretation is complicated by the fact that three of them also had a concurrent mutation [
10‐
16]. Here we report the clinical, genetic and molecular findings from three siblings bearing a novel bi-allelic, truncating mutation in PIK3CD. All three patients exhibited autoimmunity and B cell dysregulation, and in vitro analysis of total and of enriched naïve B cells revealed impaired proliferation and class-switching.
METHODS
Study Approval
The institutional review board at King Faisal Specialist Hospital & Research Centre (KFSHRC) approved all research within this study, under RAC # 2080 025. Informed consent was acquired from the mother for herself as well as on behalf of her children. The father was invited to join the study but refused participation. Clinical and family histories were noted at the time of blood draw. PBMCs were isolated using standard Ficoll density gradient centrifugation methods for cryopreservation till ready for use.
Next-Generation Sequencing (NGS)
Whole exome sequencing (WES) was performed on blood-derived genomic DNA using SureSelectXT Target Enrichment system (Illumina NovaSeq 6000, San Diego, CA, USA). Quantitative measures of raw data quality were computed using FastQC, and the reads were aligned to the human reference genome (hg19) using Burrows-Wheeler Aligner (BWA)21. The resulting bam files were assessed for possible PCR duplication. Local realignment and base-quality recalibration were performed using Picard (
http://broadinstitute.github.io/picard) and GATK22, respectively, to guarantee the production of high-quality base calls. MuTect223 was employed to detect somatic mutations, whereas ANNOVARv21.07 was used to annotate mutations throughout different databases. Single nucleotide variations (SNVs) and indels that met the criteria of the conventional MuTect2 filters were subjected to further screening. We utilized the Saudi Human Genome Database to assess whether variants of interest found in the WES data were low-level unreported SNPs segregating in the local population.
Sanger Sequencing
Genomic DNA was obtained from whole blood. Primers incorporating 5’-tagged M13 sequences were designed as needed. DNA sequencing was performed using a BigDye Terminator (Thermo Fisher, Waltham, MA, USA).
RNA Expression Analysis
Lymphoblastoid cell lines were generated via standard protocols from patients and healthy controls. Subsequent to RNA extraction and cDNA synthesis, real-time reverse-transcriptase PCR (Applied Biosystems 7500 Fast Real-Time PCR System, Thermo Fisher) was performed using exon–exon spanning primers. Primer sequences were 5′- GGGTCTACCTGAACTTCCCTG − 3′ and 5′- GCCACTGAGCATGTGGAAG − 3′. Quantification was performed using the QuantiTect SYBR Green mix (Qiagen, Limburg, Netherlands), employing the deltaCt technique to quantify fold difference. Reactions were conducted in triplicate each time, with at least two independent runs, then normalized to GAPDH as the internal control.
Immunoblotting
Lymphoblastoid-derived cell lysates were prepared with RIPA buffer (Sigma, St. Louis, MO, USA) then quantified on a Smart-Spec Plus spectrophotometer (Bio-Rad, Hercules, CA, USA) using the Protein Assay Dye Reagent (Bio-Rad). Protein samples were electrophoresed on an SDS-PAGE (National Diagnostics, Atlanta, GA, USA), and transferred onto a PVDF membrane (GE Healthcare, Waukesha, WI, USA). Primary antibodies utilized in this study included: anti-GAPDH (#2118) (Cell Signaling Technology, Danvers, MA, USA), anti-PIK3CD (ab109006) and anti-p85 (ab86714) (Abcam, Cambridge, UK).
Immunophenotyping
Frozen PBMCs were thawed and rested for 2 h in complete medium (RPMI 1640 containing 10% FBS and 1% penicillin/streptomycin). Cells were then counted and suspended at 10
6 cells per ml in FACS buffer (2% FBS in PBS) before being stained and analyzed on an LSR II flow cytometer (Becton Dickinson, Mountain View, CA). Gating and antibody details are as described previously [
17].
Lymphocyte Stimulation
For the carboxyfluorescein 6 succinimidyl ester (CFSE) proliferation assay, thawed PBMCs were rested for 2 h in complete medium. Cells were then stained with 1 µM CFSE (BD) at 37 degrees for 15 min, washed extensively in complete medium, then seeded in a U-bottom 96-well plate at a density of 1 × 106 cells/ml. For T cell proliferation, CFSE-labelled PBMCs were incubated in the presence or absence of Dynabeads Human T-Activator CD3/CD28 (Thermo Fisher) for 3 days and analyzed by FACS. Experiments on total CD19 + cells were conducted as part of a mixed PBMC population, while those involving enriched naïve B cells were performed on the CD19 + CD27- IgD + subset after processing through a BD FACSMelody Cell Sorter. Cells were stained with CFSE (for the proliferation assay) or not, then treated with CD40L (MEGACD40L, Enzo Life Sciences, Farmingdale, NY) supplemented with IL-2 (202-IL-010) and IL-21 (8879-IL-050) (both R&D Systems, Minneapolis, MN) for the indicated durations. Fluorescent antibodies used for analysis included anti-CD4-APC, anti-CD27-PE, anti-IgM-APC (all BD), anti-IgD-FITC (348206, Biolegend) and anti-CD19-PE-Cy7 (25-0199-42, Thermo Fisher). After addition of 4′,6-diamidino-2-phenylindole (DAPI) for live/dead screening, cells were analyzed on an LSR II flow cytometer (BD) with lymphocyte proliferation assessed as a measure of CFSE loss. DAPI/Annexin V cellular profiling was performed using Annexin V-Alexa Fluor 488 (V13241, Invitrogen).
DISCUSSION
Here we describe the clinical, molecular and cellular aspects of a family that segregates a null phenotype of PIK3CD. In leukocytes PI3Kδ, composed of a p85α and p110δ heterodimer, acts upon PIP2 to generate PIP3, an essential mediator of cellular growth and survival. Key downstream effectors of PI3Kδ include Akt, which activates mTOR and thereby facilitates glucolysis and cellular growth, and IL-2 inducible T cell kinase (ITK) in T cells and Bruton tyrosine kinase (BTK) in B cells [
1]. In the latter, PIP3 recruits BTK to the cell membrane where activation occurs through autophosphorylation, highlighting the essential role of PI3K in B cell development and function. Indeed, patients with LOF PIK3CD display reduced B cell counts and diminished immunoglobulins [
11,
14,
15]. Similarly, patients with p85α mutations that lead to PI3K underactivation also cause defects in B cell development and function [
19,
20], including an almost total absence of CD19 + cells in blood, and of B lineage (including pro-B) cells in bone marrow. In line with these reports, immune phenotyping of our patients also revealed low B cell ratio as well as abnormally high naïve, and abnormally low class-switched, CD19 + percentages across all three patients. These findings were also reflected in our in vitro work, which revealed that patients’ total CD19 + and enriched naïve B cells exhibited deficient responses to T cell-dependent stimulation, including weak proliferation activity and poor class-switching. CD19 + cells also displayed increased apoptosis.
The centrality of PI3K in B cell function has been shown across species. Previous work on p110δ-deficient mice revealed a significantly impaired or completely abolished B cell response to T cell-dependent and T cell-independent stimulation [
21‐
23], with compromised cell cycle entry following BCR stimulation and the inability of IL-4 to protect B cells from apoptosis [
24]. Interestingly, in mice with B cell-specific mutation, suppression of PIP3 lead to an increase in CSR at the expense of plasmablast differentiation, while elevated levels of PI3K activity suppressed immunoglobulin CSR [
25,
26]. These studies illustrated that tuning the level of PIP3 plays a role in the B cell’s decision of whether to differentiate into plasmablasts, which generate mainly low-affinity IgM antibodies, or to undergo CSR for eventual secretion of higher affinity (IgA, IgG, IgE) immunoglobulins [
2]. Although an attenuated humoral response is indeed observed in
Pik3cd-deficient mice, this appears to be a consequence of poor development of follicular helper T cells (T
FH) due to the absence of p110δ. In knockout mice the levels of two major effector molecules, Cd40l and Il-21, are reduced in T
FH cells. Interestingly, the adoptive transfer of extrinsic, wildtype T cell help into the mice generates a strong class-switched response to immunization, despite the B cells being p110δ-deficient. Hence the impaired humoral response in mice is likely secondary to loss of p110δ in T cells [
27]. In our patients however, both the in vivo class-switched CD19 + memory fraction (CD27 + IgD-) and the plasmablast percentage (CD27 + CD38+) were drastically reduced compared to controls. Despite strong exogenous CD40L and IL-21 in vitro stimulation there was evidence of poor CSR outcome in our patients’ cells. Impaired proliferation and class-switching in total CD19⁺ cells could reflect skewed naïve-to-memory B cell ratios. However, similar defects were seen (albeit less pronounced) when using enriched naïve B cells. This may indicate a potential inter-species difference for the role of PIK3CD in B cell function, an observation that warrants further investigation. We note that the role of p110δ in B cells has been elucidated in mice but remains less well studied in humans.
The observed increase in T cell proliferation contrasts sharply with the marked decrease in B cell proliferation and function. Our data showed significantly enhanced T cell proliferation in response to CD3/CD28 stimulation, as well as to PHA and ConA mitogens. This phenomenon has been observed before, not only in other p110δ-deficient patients, but even in Jurkat cell lines that were engineered for removal of the PIK3CD gene [
12], indicating that the enhanced proliferation is not simply a function of weakened Treg activity. Murine p110δ-deficient T cells exhibit a similar tendency [
23]. This hyperactivation may result from the increased availability of PIP2, which is hydrolyzed by phospholipase Cγ1 (PLCγ1) into crucial secondary messengers that trigger several distal signaling cascades vital for CD3 + activation [
7]. Loss of homeostasis between the PI3K and PLCγ1 pathways in p110δ-deficient T cells may therefore provide a boost to T cell activity [
12]. In light of this, one of the clinical hallmarks of PIK3CD-deficiency in both mice and humans is enteropathy [
1,
23], which was also a consistent finding across all affected members of our family. In PI3Kδ-deficient mice, improper activation of effector T cells by gut microbes leads to colitis, which PI3Kδ-deficient Treg are unable to suppress [
23,
28]. Reduced abundance of mRNAs encoding proteins linked to inflammation and cytotoxicity, such as IFN-γ, granzyme B, and perforin, is a characteristic of PI3Kδ-deficient mouse CD8 + T cells stimulated in vitro [
29,
30]. In humans, patients who are homozygous for truncating mutations exhibit severe infections alongside inflammatory colitis [
11,
14,
15], molecularly characterized as CD8 + expansion into colonic tissue with drastically upregulated expression of TBET and perforin proteins [
15]. However, colitis was not observed in the two reports where patients with LOF PIK3CD had concurrent genetic mutations. In the first report, the nature of the hypomorphic missense mutation in p110δ, which was linked to normal expression and residual PIK3CD enzymatic activity, may explain the absence of colitis and the generally normal immunoglobulin readings [
12]. In the second report the presence of a homozygous truncating mutation in SKAP (a kinetochore-associated protein), alongside the PIK3CD mutation, resulted in defects in T cell spreading and migration which likely prevented colitis onset [
13]. Our data supports the conclusion that CD3 + hyperactivation is a natural consequence of PIK3CD protein loss.
A total of ten p110δ LOF patients have been reported to date. However, two of the patients were not well-characterized [
10], while three other patients had a second concomitant mutation which complicated interpretation of the PIK3CD phenotype. If we remove these individuals, we are left with only five well-characterized patients. Our study raises this number significantly to eight, allowing us to better understand the disease phenotype. The majority of the patients came from consanguineous families in the Middle East and Pakistan. Early clinical manifestation during the first few months is frequent, while some instances appear at a later age of 6–9 years. Inflammatory bowel disease with chronic bloody diarrhea and colitis is seen in 7 of the 8 patients. All reported cases had variable hypogammaglobulinemia and a history of recurrent sinopulmonary infections, of which some were caused by opportunistic organisms. Diseases of the immune system dysregulation, such as juvenile idiopathic arthritis, autoimmune hepatitis, autoimmune thrombocytopenia and inflammatory bowel disease, were also common. Clinical management consisted mainly of prophylactic antibiotics, immunoglobulin replacement and antibiotic treatment for acute infections. Immunosuppressive (steroids and 6-mercaptopurine) or anti-inflammatory (mesalazine) medication were used to control autoimmune manifestations [
12‐
15]. Two of the eight patients died from pneumonia and sepsis. The remaining patients were followed on conservative therapy. One patient underwent hematopoietic stem cell transplantation with a lack of long term follow up [
15], which was not clinically considered in our patients due to their disease course and treatment response.
Our study indicates that loss of PIK3CD in our patient cells coincides with a B cell phenotype that does not adequately respond to T-dependent stimulation. This suggests that the role of PI3Kδ in human B cells differs from what has been observed in mouse models. Further work is warranted to corroborate these findings, and to search for other potential areas of interspecies divergence relating to the PIK3CD pathway in B cells and other lymphocyte populations.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.