|Home | About | Journals | Submit | Contact Us | Français|
Deleterious mutations in genes involved in the Fas apoptosis pathway lead to Autoimmune Lymphoproliferative Syndrome (ALPS). Demonstration of an apoptosis defect is critical for the diagnosis and study of ALPS. The traditional in vitro apoptosis assay, however, requires a week of experimental procedures. Here, we show that defects in Fas-induced apoptosis in PBMCs can be evaluated directly ex vivo using multicolor flow cytometry to analyze the apoptosis of effector memory T cells, a Fas-sensitive subset of PBMCs. This method allowed us to sensitively quantify defective apoptosis in ALPS patients within a few hours. Some ALPS patients (ALPS-sFAS) without germline mutations have somatic mutations in Fas specifically in double-negative αβ T cells (DNTs), an unusual lymphocyte population that is characteristically expanded in ALPS. Since DNTs have been notoriously difficult to culture, defective apoptosis has not been previously demonstrated for ALPS-sFAS patients. Using our novel ex vivo apoptosis assay, we measured Fas-induced apoptosis of DNTs for the first time and found that ALPS-sFAS patients had significant apoptosis defects in these cells compared to healthy controls. Hence, this rapid apoptosis assay can expedite the diagnosis of new ALPS patients, including those with somatic mutations, and facilitate clinical and molecular investigation of these diseases.
Autoimmune Lymphoproliferative Syndrome (ALPS) is a disorder of lymphocyte apoptosis, resulting in chronic lymphoproliferation and autoimmunity. The disease is characterized by chronic lymphadenopathy and/or splenomegaly, autoantibodies, autoimmune cytopenias of hematopoietic cells (i.e. neutropenia, thrombocytopenia, and anemia), and an expansion of a normally rare population of TCRαβ+ lymphocytes that do not express the CD4 or CD8 co-receptors (double-negative T cells, DNTs) [1, 2]. ALPS patients have defects in the Fas death receptor pathway of apoptosis, which has been shown to be critical for maintaining lymphocyte homeostasis and immune tolerance. Discovery of the molecular defects in cell death pathways in different forms of ALPS has been instrumental in the understanding of normal immune homeostasis [1, 4].
The current criteria for a definitive diagnosis of ALPS requires that the patient have chronic, non-malignant, non-infectious lymphadenopathy and/or splenomegaly, elevated TCRαβ+ DNTs (greater than or equal to 1.5% of total lymphocytes), and either a mutation in one of the known ALPS causative genes (FAS, FASLG, or CASP10) or defective lymphocyte apoptosis . Therefore, the demonstration of defective apoptosis in vitro is essential for a conclusive diagnosis of ALPS, especially for patients in which the genetic defect is unknown. Apoptosis testing has been particularly critical for the proper identification of ALPS patients with overlapping symptoms of other disorders such as Evans syndrome, familial hemophagocytic lymphohistiocytosis (fHLH), and common variable immunodeficiency (CVID) [6, 7]. However, the conventional procedure for measuring lymphocyte apoptosis in patient specimens requires activation of T lymphocytes and in vitro culture for at least one week (Fig. 1). Activation and extended culture in IL-2 is thought to be required for the expansion of activated lymphocytes to enhance susceptibility to apoptosis. The expanded cells are stimulated through Fas and assessed for cell death by propidium iodide staining and flow cytometry analysis. However, with this method, inconsistencies have been observed in results between laboratories and even between different phlebotomy samples from the same patient . These problems have been attributed to differences in activation or proliferation of the lymphocytes and length of time in culture. Thus a faster, more standardized apoptosis assay that is highly reproducible between different laboratories and investigators is needed. In freshly isolated peripheral blood mononuclear cells (PBMCs), there are multiple T cell subpopulations with different levels of sensitivity to Fas-induced apoptosis [9, 10]. Previous studies have shown that among PBMCs, the effector memory T cells express Fas and are the most sensitive T cell subset to Fas-mediated apoptosis [10, 9]. Potentially, multicolor flow cytometry could be used to selectively assess cell death of effector memory T cells in PBMCs, eliminating the need for activation and in vitro propagation. Here, we describe a novel ex vivo assay for assessing defects in apoptosis that is performed directly on freshly isolated PBMCs and can yield highly reproducible results within a few hours.
The following antibodies were used: FITC-conjugated anti-CCR7 (R&D Systems), PE-conjugated anti-CD27 (BD Pharmingen), eFluor™450-conjugated anti-CD45RA (eBioscience), PerCP-Cy5.5 conjugated anti-CD3 (eBioscience), APC-Cy™7 conjugated anti-CD4 (BD Pharmingen), FITC-conjugated anti-TCRαβ (BD Pharmingen), PE-conjugated anti-B220 (eBioscience), eFluor 605NC-conjugated anti-CD8a (eBioscience), eFluor 450-conjugated anti-CD3 (eBioscience), APC-conjugated anti-CD95 (BD Pharmingen), and APC-conjugated mouse IgG1, k isotype control (BD Pharmingen). APC-conjugated annexin V was purchased from Invitrogen. Agonistic anti-Fas antibody (APO-1-3) was obtained from Enzo Life Sciences. Rat anti-mouse IgG3 (Southern Biotech) was used to crosslink anti-Fas antibody.
Blood samples were obtained from patients and healthy individuals with informed consent under protocols approved by the Institutional Review Boards of the National Institute of Allergy and Infectious Diseases. PBMCs were enriched from whole blood collected in acid citrate dextrose (ACD) by Ficoll-Hypaque density gradient centrifugation. Cells were washed and resuspended at 2 × 106 cells/mL in RPMI 1640 supplemented with 10% FBS, 2 mM glutamine, and antibiotics, with or without added IL-2 (100 U/mL).
The following lists the ALPS patients and healthy mutation-positive relatives (HMPRs), included in the study, along with their corresponding Fas mutations: 121.11 (c.1073_1074delAT, p.L278DfsX2), 137.6 (c.973A>G, p.D244G), 143.16 (c.970T>C, p.I243T), 143.17 (c.970T>C, p.I243T), 320.8 (c.845+1G>T, p.V174GfsX9 or p.0), 330.1 (c.870+4A>G, p.E202MfsX4? or p.0?), 243.1 (c.498A>G,p.R86G), 161.1 (c.535A>T, p.0), 232.1 (c.535A>T, p.0), 324.11 (c.669_683del15insA, p.L143KfsX10? or p.0?), 43.1 (c.825G>T,p.E195X), 360.1 (c.913delTinsGA, p.M224RfsX7), and 276.1 (c.876 del 5; D212fs). ? means effect unknown; p.0 indicates no protein produced; p.0? means effect unknown, probably no protein produced. The mutations are listed using the classic notation for Fas mutations, in which nucleotides are numbered in relation to cDNA nucleotide positions per GenBank M67454.1 and amino acids are numbered starting at Arginine 1 of the mature Fas protein per GenBank AAA63174.1. Mutations for these individuals are also listed using the Human Genome Variation Society (HGVS) recommended nomenclature in Tables 1 and and22.
PBMCs (2 × 105 cells/well) from healthy blood donors and ALPS patients were plated in triplicate (unless otherwise noted) in 96-well flat bottom plates in media supplemented with 100 U/mL IL-2 and treated with varying concentrations of agonistic Fas antibody plus anti-mouse IgG3 at one-tenth the concentration of anti-Fas for crosslinking of the Fas antibody. Cells were incubated for 8 hours at 37°C 5% CO2, then washed with Annexin-binding buffer (ABB; 10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl2, 0.5% BSA, pH 7.4), and stained with fluorochrome-conjugated annexin V and antibodies to CCR7, CD27, CD3, CD45RA, and CD4 to identify apoptotic TEM cells. After staining, the cells were washed and resuspended in ABB then fixed by adding equal volume of 10% neutral buffered formalin (NBF). Cells were analyzed by FACS using an LSR II or LSRFortessa (BD Biosciences). At least 50,000 events per sample were acquired. Using FlowJo software (Tree Star, Inc.), we gated on CD3+CD4+CD45RA-CCR7-CD27- cells (TEM) and the percent annexin V+ fraction of TEM cells was assessed. Percentage of cell loss was calculated as 100 × (1-[(100-% of annexin V+ TEM with treatment)/(100-% of annexin V+ TEM without treatment)]).
For analysis of percentage cell loss of other T cell subsets, the percent annexin V+ fraction of that particular subset was used with the same formula. Naive, TCM, and TTM CD4+ T cells were defined as CD3+CD4+CD45RA+CCR7+CD27+, CD3+CD4+CD45RA-CCR7+CD27+, and CD3+CD4+CD45RA-CCR7-CD27+, respectively.
For the FasT Kill assay experiments in which different incubation time points (4, 6, 8, and 12 hours) were assessed, IL-2 was not added to the media as later discussed. All subsequent FasT Kill assays performed with the shorter 4 hour incubation time do not include IL-2 in the media.
For DNT apoptosis experiments, after cells were treated with Fas antibody and anti-IgG3 for 4 hours in media without IL-2, cells were washed and stained with annexin V and antibodies to CD3, TCRαβ, CD4, CD8, and B220 to identify apoptotic DNTs. DNTs were defined as CD3+TCRαβ+ CD4-CD8-B220+. The percentage of DNT cell loss was calculated with the above formula using the percent annexin V+ fraction of DNTs.
Data in FasT Kill assays represent mean ± standard deviation for triplicate wells, unless otherwise noted. The FasT Kill assay normal control range was calculated for each dose of agonistic Fas antibody as the 10th to 90th percentiles based on the FasT Kill results obtained from 20 healthy blood donors.
To develop an improved apoptosis assay, we took advantage of the inherent sensitivity of effector memory T cells to Fas-mediated apoptosis. We used multicolor flow cytometry to identify effector memory T cells among total PBMCs and assay apoptosis. We prepared whole PBMCs from a fresh blood sample by a one-step Ficoll gradient separation. We then stimulated the cells for 8 hours with APO-1-3, an agonistic anti-Fas antibody of the IgG3 isotype that was then cross-linked with anti-mouse IgG3. No in vitro activation or proliferation was induced. Subsequently, the cells were stained to examine the viability of different T cell subsets. We gated using CD3 and CD4 to identify CD4+ T cells and CD45RA to eliminate naïve cells. We also used CCR7 and CD27 to identify CD4+ effector memory T (TEM) cells, which express CD45RO (no CD45RA) and are CCR7- and CD27- (Fig. 2a). As previously observed, most CD4+CD45RA+ T cells were CCR7+ and CD27+ [9, 10]. Concomitantly, we used annexin V staining to mark apoptotic cells and compared TEM with or without Fas stimulation. With this method, we observed that the TEM cells are clearly the most sensitive T cell subset to Fas-induced apoptosis. There was a dose-response of apoptosis to log increases in Fas stimulation and at 1 μg/mL, >95% of TEM underwent apoptosis within eight hours despite the fact that the PBMCs from which they were derived had not been stimulated with TCR agonists or IL-2 in vitro as in conventional assays (Fig. 2b). These results are in accordance with previous observations where a stabilized form of Fas ligand could induce the death of TEM cells . The sensitivity of the other populations to Fas-induced apoptosis was lower compared to TEM cells, with transitional memory (TTM) appearing partially sensitive and central memory (TCM) cells showing a very mild induction of apoptosis. No specific killing was observed in naïve cells (Fig. 2b).
In the conventional Fas kill assay, death is quantitated in total CD3+ T cells and so we compared this unseparated population to the TEM in the ex vivo kill assay and found that the percent cell death of the former was much lower, approximately 20-30%, compared to the TEM cells. The latter subset showed nearly 100% cell loss at the highest dose of anti-Fas tested (Fig. 2c). Thus, using multiparameter flow cytometry to specifically quantitate death in the TEM population provided a greater range in which to study cell death in patient specimens. We also observed that there was a variation in the fraction of TEM cells among different individuals, which ranged between 0.5-10 percent (data not shown). A variation of this magnitude could result in major differences in the percentage of quantified cell loss when examining the total population of CD3+ T cells. Focusing the analysis only on the TEM cells eliminates this adventitious variation in apparent cell death sensitivity between individuals. Because of the rapidity with which Fas apoptosis can be assessed in effector memory T cells we call the new assay the “FasT Kill” assay.
To validate the FasT Kill assay as a potentially useful clinical tool for ALPS diagnosis we examined the normal control range and reproducibility of the assay. In preliminary experiments on normal control donors using the FasT Kill assay, we found a modest variation in the degree of cell death of TEM cells. Therefore, we compared 20 healthy individuals to define a normal control range (Fig. 3a). We also examined assay variability by testing the PBMCs from four different individuals on three separate occasions, each representing a separate phlebotomy from the individual. We found that the FasT Kill assay is highly reproducible with minimal day-to-day variation when different phlebotomy specimens from the same individual were assayed (Fig. 3b). We hypothesized that the control range that we established would allow us to categorize patients with defective Fas-mediated apoptosis.
We defined the normal range for the FasT Kill assay as the 10th to 90th percentiles based on the 20 normal controls. We then tested five individuals with intracellular Fas mutations (three with ALPS and two healthy mutation-positive relatives (HMPRs) of ALPS patients) and four ALPS patients with extracellular Fas mutations (Table 1). Previous studies have shown that HMPRs typically have abnormal in vitro apoptosis responses in the traditional Fas kill assay [11, 12]. Compared to the controls, the ALPS patients and HMPRs were all demonstrably defective in Fas-mediated apoptosis, as defined by a fractional cell loss after Fas stimulation that fell below the 10th percentile of normal controls (Fig. 4a and b). Consistent with previous findings, HMPRs with deleterious Fas mutations but no clinical disease also exhibited apoptosis defects similar to ALPS patients using the new assay. Interestingly, the majority of individuals with intracellular Fas mutations displayed quantitatively greater apoptosis defects in this assay (Fig. 4a) than the patients with extracellular Fas mutations (Fig. 4b). Since previous work established that intracellular Fas mutations typically affect the “death domain” and dominantly interfere with Fas function more severely than extracellular mutations, our results indicate that this assay is sensitive enough to distinguish the severity of apoptosis impairment by different molecular variants of Fas [13-15].
We originally used an 8 hour stimulation in the FasT Kill assay since high levels of apoptosis were observed at this time point across a broad range of concentrations of the anti-Fas antibody used in the conventional Fas kill assay. However, to make the assay more amenable to clinical laboratory use as a diagnostic test, we tested shortening the treatment time to 4 hours. We found that 4 hours of stimulation yielded less cell loss using the lower doses of anti-Fas antibody but at 1000 ng/mL the cell loss was comparable to that observed with longer periods of stimulation (Fig. 5a). A range of higher concentrations of agonist antibody with the 4 hour stimulation resulted in a clear dose response with greater percentages of cell death (data not shown). We had previously added IL-2 to the media during the FasT Kill assay to help reduce background death (presumably due to cytokine withdrawal-driven apoptosis), however we found that in the absence of exogenously added IL-2, there was slightly more Fas-induced apoptosis (data not shown). Since the reduction in background death by IL-2 addition was negligible and IL-2 appeared to slightly suppress Fas-mediated death, we eliminated its use in all subsequent 4 hour stimulation assays. We tested ALPS patients with the FasT Kill assay using the 4 hour stimulation time and found clear differences between ALPS patients with intracellular (330.1) or extracellular (232.1) mutations and the normal controls (Fig. 5b and c and Table 1). Therefore, the shortened FasT Kill assay appears to be suitable as a clinical test that can detect a functional Fas apoptosis defect within a single working day.
ALPS patients with somatic Fas mutations (ALPS-sFAS) have all the clinical features of ALPS (Table 2) but represent a diagnostic challenge since they have no germline Fas mutations and undergo cell death normally in PBMC apoptosis assays [16, 17]. The somatic Fas mutations are found in the DNT cell population of these patients. We therefore asked whether we could detect impaired apoptosis in DNT cells directly ex vivo using a multicolor flow cytometry-based method similar to the FasT Kill assay. DNTs in ALPS patients are defined as TCRαβ-positive T cells that lack the CD4 and CD8 co-receptors. However, unlike the expanded DNTs in ALPS patients, the small numbers of DNTs from normal individuals are composed of a more heterogeneous population of cells . Therefore, we used the B220 marker, which appears to be uniquely expressed on the expanded population of DNTs in ALPS patients, to selectively gate on the DNT sub-population in normal controls that most likely represents the same class of DNTs found in ALPS . By gating on CD3+TCRαβ+CD4-CD8-B220+ cells within PBMCs, we found that DNTs clearly express Fas (Fig. 6a). We first tested PBMCs from ALPS-sFAS patients along with normal control individuals for Fas-mediated apoptosis of effector memory CD4+ T cells using the FasT Kill assay. Consistent with our previous experience with conventional apoptosis assays, the TEM cells of ALPS-sFAS patients appeared to undergo apoptosis normally (Fig. 6b). This is due to the fact that the TEM cells do not harbor a germline genetic mutation in Fas and express normal Fas protein. However, when we examined the CD3+TCRαβ+CD4-CD8-B220+ DNT cells, we found that the DNTs from ALPS-sFAS patients were severely impaired in Fas-mediated apoptosis compared to the normal controls (Fig. 6c). A dose response in percent cell loss was observed in the DNTs from normal individuals whereas little, if any, Fas-induced cell death was detected in DNTs of ALPS-sFAS patients at any dose tested (Fig. 6c). Thus, the combined use of the FasT Kill assay and a similar apoptosis assay for DNTs is a new functional approach for the rapid diagnosis and identification of ALPS patients with somatic Fas mutations.
The study of human disorders of apoptosis has yielded important insights about immunological disease pathogenesis and new approaches to therapy [8, 20]. In the case of ALPS, the demonstration of a lymphocyte apoptosis defect is a critical diagnostic criterion [21, 5]. Also, it has become clear that lymphocyte apoptosis defects contribute to the pathogenesis of other severe immunological disorders such as X-Linked Lymphoproliferative Syndrome . A persistent problem has been the utilization of assays that require activation and extensive culture of lymphocytes, which makes it impractical for all except a few specialized centers to carry out in vitro tests of lymphocyte apoptosis. Furthermore, in our experience there is a great deal of variability in currently utilized protocols which diminishes confidence in assessing an abnormality even for genetic mutants that cause profound apoptosis defects . Therefore, we have devised a new test in which multiparameter flow cytometry is used to assess the apoptosis response of TEM and other lymphocyte subpopulations after a relatively short (~ 4 hours) stimulation through the Fas receptor.
Comparing apoptosis sensitivities between different individuals by analyzing CD4+ TEM cells rather than whole PBMCs is advantageous for several reasons. First, different individuals will have different proportions of lymphocyte subsets with variable apoptosis susceptibility in their PBMC pool. We have observed significant differences in the proportion of memory populations in different individuals when comparing whole PBMCs or isolated pan-T cells. To some degree this correlates with age; T cells from younger subjects generally have more naïve, Fas-insensitive cells (data not shown). Gating on the CD4+ effector memory population and analyzing apoptosis solely in this population eliminates this variability, allowing us to better compare apoptosis sensitivity between different individuals. Second, the Fas sensitive population represents a very small fraction of total PBMCs, such that even among T cells, only a minor percentage of the cells will undergo apoptosis. By contrast, the majority of CD4+ effector memory cells die after stimulation with high concentrations of anti-Fas (Fig. 2c). Third, we also observed that CCR7-CD27+ “transitional memory” T cells may be as sensitive as TEM to Fas-mediated apoptosis but because CCR7 staining does not give clear separation between the central (TCM) memory and transitional (TTM) memory T cells, some of the TCM cells in the transitional T cell gate may reduce the apoptosis measured. Thus, for apoptosis assessment, it is simpler to gate on CCR7-CD27- effector memory cells since this provides clear separation from the other two memory T cell populations. Flow cytometry is usually available in immunological testing laboratories and medical centers and the new test was easy to perform and gave robust results even in individuals with subtle apoptosis abnormalities. Thus, this assay should be widely accessible to clinicians and researchers and provides a quick quantitative assessment of functional defects.
Based on the current criteria, apoptosis testing is essential for a definitive diagnosis of individuals displaying clinical features consistent with ALPS but lacking mutations in any of the known causative genes . Those individuals that meet all ALPS diagnostic criteria but have no known genetic defects are classified as ALPS-U. Given the importance of apoptosis assays for the diagnosis of ALPS-U, we have been testing our FasT Kill assay on potential ALPS-U patients. Currently, we have found our FasT Kill assay results corroborate well with the results of the traditional Fas apoptosis assay in diagnosing ALPS-U or ALPS-like patients (Supplementary Fig. 1). With the FasT Kill method, we confirmed apoptosis defects in ALPS-U patients previously shown to have impaired death with the conventional Fas kill assay (Supplementary Fig. 1a and b). Patients with clinical features of ALPS but no defect in cell death evaluated by the traditional Fas apoptosis test, correspondingly, had no defect in apoptosis as assessed by our FasT Kill assay (Supplementary Fig. 1c and d). Thus, the FasT Kill assay may be a faster method for the diagnosis of ALPS-U patients.
In each of the FasT Kill assays we have performed, we used several different concentrations of anti-Fas antibody. Our data suggests that a single anti-Fas antibody dose at 1000 ng/mL may be sufficient to detect an apoptosis defect in most ALPS patients, especially ALPS-Fas patients with intracellular mutations. However, we recommend using at least 2-3 concentrations of anti-Fas between 100-1000 ng/mL because having more than one data point below the normal controls can generate more confidence when declaring an apoptosis defect for patients with milder defects in cell death, such as ALPS-Fas patients with extracellular mutations or ALPS-U patients.
Although a lymphocyte apoptosis defect is critical for a definitive diagnosis of ALPS-U, it is important to note that patients with clinical features resembling ALPS may still be diagnosed with “probable ALPS” without being tested for defective apoptosis . For a “probable ALPS” diagnosis, the patients must have chronic lymphadenopathy and/or splenomegaly, elevated TCRαβ+ DNTs, and have either elevated ALPS biomarkers (i.e. high plasma levels of soluble FasL, IL-10, IL-18, or vitamin B12), immuno-histological findings characteristic of ALPS, autoimmune cytopenias along with elevated IgG levels, or a family history of ALPS-like lymphoproliferative disease. It is recommended that patients with “probable ALPS” be “treated and monitored” in the same manner as those with a definitive ALPS diagnosis . However, patients with “probable ALPS” should still be tested for apoptosis defects or mutations in FAS, FASLG, or CASP10 in order to obtain a definitive diagnosis of ALPS .
ALPS-sFAS patients present the clinician with a difficult diagnosis. They have the same clinical phenotype and biomarker profile as ALPS-FAS patients with germline FAS mutations, but they do not display defects in apoptosis using a traditional Fas kill assay nor genetic mutations in unfractionated peripheral blood samples . This is because DNTs do not survive well in culture and are eliminated during the prolonged culture prior to the apoptosis assay [16, 17]. The remaining conventional T cells generally do not possess the Fas mutation and therefore exhibit normal apoptosis responses. It is also difficult to establish a molecular diagnosis in such patients because they lack germline mutations in ALPS-related genes. At present, it is necessary to isolate DNT populations from suspected ALPS-sFAS patients and search for mutations in known ALPS-associated genes. Establishing a functional defect has heretofore not been possible. Using a multiparameter flow cytometric direct ex vivo assay, we find that B220+ DNTs from normal individuals, closely resembling those from ALPS patients, can undergo Fas-mediated apoptosis. However, the B220+ DNTs from ALPS patients were insensitive to Fas-induced apoptosis. Thus, a comparison of the results from ex vivo Fas killing assays between peripheral blood T cells and DNTs can yield the presumptive diagnosis of ALPS-sFAS in just a few hours. Taken together, our data show that the functional evaluation of selected lymphocyte subsets using current multiparameter flow cytometry may be more widely utilized to yield robust diagnostic assays even in immune diseases due to somatic mutations.
We thank the patients and healthy blood donors. We also thank Josh Milner for use of his LSRFortessa. We thank Julie Niemela for assistance in mutation nomenclature. We also thank Claire Liu for assay name recommendations, and we thank Helen Su and Chryssa Kanellopoulou for critical reading of the manuscript. This research was supported by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.
Conflict of interest: The authors declare that they have no conflict of interest.