Prostaglandin D₂-supplemented “functional eicosanoid testing and typing” assay with peripheral blood leukocytes as a new tool in the diagnosis of systemic mast cell activation disease: an explorative diagnostic study

Twenty-two Caucasian individuals, consecutively presenting with MCAD (12 with diagnosis of MCAS and 10 with diagnosis of SM, according to WHO criteria) from February to July 2012 at the St. Franziskus Hospital, Eitorf (Germany) or the Department of Oncology, Haematology and Palliative Care, District Hospital of Waldbröl (Germany), were randomly enrolled in the present prospective explorative study to better reflect disease diversity (for patients’ characteristics, see Table 1 and Additional file 1). The MCAD patients were clinically diagnosed according to the current criteria for MCAD [1]-[4]. Diagnostic criteria of SM and MCAS as well as the respective diagnostic findings of the individual patients are depicted in Additional file 2. The age of the patients ranged from 20 to 71 years (mean: 48.5 years; male to female ratio: 6:16). For comparison, twenty age-matched healthy individuals from Medical Clinic III, Friedrich-Alexander-University Erlangen-Nuremberg (Germany) were included (mean age: 44.3 years; range: 20–60 years; male to female ratio: 10:10). An extensive medical history of all healthy individuals was conducted. The presence of an dysfunctional immune system, immune defects, auto-immune disease, inflammatory disease of skin, airways, gastro-intestinal tract, atopy/allergy, hypersensitivity to drugs, aero- or food-allergens, food-additives, also exhaustive physical stress, immunization or xeno-immunoglobulin transfusion was ruled out. IgE-mediated allergy was ruled out by negative prick and specific IgE testing. Finally, no intake of antihistamines, steroids, or regular drugs other than oral contraceptives was accepted. To estimate the influence of MCAD-specific medication on eicosanoid patterns, subgroup analysis for MCAD patients without medication (MCAD-M; mean age: 40.3) and MCAD patients with MCAD-specific medication (MCAD+M; mean age: 52.7 years) were performed. In particular the majority of MCAS patients were devoid of MCAD-specific medication at the time of FET analysis, because they were included into the study shortly after making the diagnosis, but before starting MCAD-specific drug therapy. For diagnostic purposes, typical mast cell mediator-related symptoms were recorded by a standardized, validated questionnaire [14]-[16] (see also Additional file 1). Data on basal tryptase level, as analyzed by the “ImmunoCAP^® Tryptase” standard assay for quantification of tryptase in serum (Phadia GmbH, Freiburg, Germany; normal range < 11.4 μg/L) and KIT^D816V mutation status, determined by polymerase chain reaction based methods, were available for the included MCAD patients (see Table 1). For differential diagnosis, other conditions that could also be responsible for one or more of the reported symptoms were ruled out by unchanged relevant pathognomonic laboratory parameters, imaging techniques, and/or endoscopic methods. Before participating in the present study, all participants gave written informed consent to analysis of their data for scientific purposes and were free to withdraw from the study at any time. Blood samples were collected by standard routine diagnostic methods without further interventions. Therefore, a special ethical approval was not mandatory for the MCAD patients. Blood collection from healthy individuals, in general, was approved by the local ethical committee of the Friedrich-Alexander-University Erlangen-Nuremberg (Germany). The present study was performed in accordance with the ethical standards of the Helsinki Declaration 1975 (revised 2004).

Peripheral venous blood was collected from patients and healthy individuals and processed as described elsewhere [17]-[19]. Briefly, 4.5 ml of heparinized blood were used for isolation of PBLs. PBLs were resuspended in RPMI-1640 cell culture medium without fetal calf serum and adjusted to a cell density of 1 × 10⁶ cells/ml. PBLs were subsequently exposed to triggering factors, i.e. AA (10⁻⁵ M), ASA (10⁻⁸ M), SP (10⁻¹² M), or diluent for 20 min. In-vitro formation and release of eicosanoids were stopped by freezing at -80 °C. Samples were then stored for up to four weeks at −80 °C before quantification of eicosanoids. After thawing and centrifugation, eicosanoids were quantitatively analyzed in the cell supernatants by specific enzyme immunoassays (EIA) using highly sensitive and specific competitive EIAs for PGE₂, pLT (SIAT, Bad Essen, Germany), and PGD₂ (Cayman, Ann Abor, USA) according to manufacturers’ protocols. Intra- and inter-assay variances for the PGE₂, PGD₂, and pLT EIA were 7.4% and 8.6%, < 20% and < 20.6%, and 5.6% and 8.1%, respectively. Recovery rate for PGE₂, PGD₂, and pLT was ≥ 95%; sensitivities of the EIAs were 3.0, 19.5, and 1.0 pg/ml, respectively ([20]; Cayman, Ann Abor, USA; [21]). The FET validation process demonstrated no interference of the test with ingestion of leukotriene receptor antagonists prior to blood sampling. The personnel performing the eicosanoid analyses were blinded to the source of the specimen (patient, healthy subject).

The FET value represents the integrating endpoint of the standard FET assay. It considers both the metabolic network of the cyclooxygenase and 5-lipoxygenase pathways and summarizes the individual’s metabolic activities of triggered and un-triggered PBLs [13],[19],[22]. For calculation of the FET value, the raw data of released PGE₂ and pLTs (analyzed in pg/ml) were processed according to a specific mathematical algorithm [19]. This is achieved by using both basal and triggered PGE₂ and pLT levels, which are then integrated according to their biochemical interactions. The FET value is finally normalized, resulting in dimensionless categories between 0.0 and 3.0, which reflect the balance/imbalance (i.e. the complex pattern of formation, release and mutual interaction) of in-vitro PBL-derived PGE₂ and pLT and thus constitute the patients’ individual eicosanoid pattern. Higher values are indicative of disturbance of the physiologic PGE₂-pLT balance. Reliability of the calculated FET values was clinically confirmed for patients suffering e.g. from aspirin induced asthma and aspirin tolerant asthma (17, 22), gastric ulcer, intestinal cancer (23), NSAID-triggered hypersensitivity (18, 19), and urticaria (24).

The supplemental PGD₂ ratios R_A3, R_S3, and R_P3 were calculated according to the following equations E1, E2, and E3:

with:

a₃: AA-triggered release of PGD₂ [pg/ml];

b₃: basal release of PGD₂ [pg/ml];

s₃: ASA-triggered release of PGD₂ [pg/ml];

p₃: SP-triggered release of PGD₂ [pg/ml].

All data were tested for normal distribution using the Kolmogorov-Smirnov test with Lilliefors correction. If the assumption of normally distributed data could not be rejected, arithmetic group means ± standard deviations were calculated. The Student’s t-test for unpaired values was used for group comparisons. If the normality test failed in at least one of the study groups compared, all data were expressed as group medians with data ranges given in parentheses and comparison was calculated using the Mann–Whitney U test. P values were considered statistically significant, if P ≤ 0.05. These statistical analyses were performed using the SigmaPlot software (version 12.0, Systat Software GmbH, Erkrath, Germany). Optimal cut-off levels and misclassification rates were calculated by the method of ROC analysis, whereby the optimal cut-off point is derived by maximizing the Youden index [23]. Respective calculations were performed with the BIAS software for Windows (version 10.04; © 1989–2013 epsilon Verlag; http://​www.​bias-online.​de).

All patients demonstrated, to a varying degree, intermittently or chronically, a pattern of characteristic mast cell mediator-related symptoms (see Additional file 1) which was similar in patients with SM and MCAS. All SM patients were KIT^D816V-positive and all MCAS patients were KIT^D816V-negative. Five of 10 SM patients had tryptase blood levels > 20 μg/L, defined as the cut-off level in the corresponding WHO criteria. The remaining five SM patients and all MCAS patients exhibited normal basal tryptase levels (Table 1 and Additional files 1 and 2).

All 10 patients with SM were on treatment with mast cell activity-reducing drugs (i.e. H₁- and H₂-antihistamines and cromoglicic acid) to alleviate symptoms induced by the various mast cell mediators, in particular by histamine (Table 1 and Additional file 1). Among the 12 patients with MCAS, only three were on medication with antihistamines, whereas eight did not take any medication, because they were enrolled in the study after making the diagnosis, but before starting symptomatic therapy. Two patients (one with MCAS, one with SM) ingested NSAIDs, another two patients with SM were treated with glucocorticoids, and three SM patients took the leukotriene receptor antagonist montelukast at the time of blood sampling. Nine of 22 patients (40.9%) did not receive any MCAD-specific medication (MCAD-M) at the time of blood sampling, whereas 13 of 22 patients (59.1%) were treated with MCAD-specific medication (MCAD+M). MCAD-specific as well as other regular medications of the enrolled MCAD patients varied highly, depending on the individual symptom constellation, intolerances, and overlapping pathologies.

The mean FET value (1.75 ± 0.356) of MCAD patients was significantly (P < 0.001) elevated, as compared to the healthy individuals (0.53 ± 0.119) (Figure 1). Hence, standard FET analysis distinguished MCAD patients from healthy individuals with a distinct cut-off FET value of 0.945 with 0.0% misclassification. A maximum FET value of 2.36 was detected for MCAS patient #7 and SM patient #1, whereas a minimum FET value of 1.14 was found for MCAS patient #5. SM patients exhibited a slightly higher mean FET value (1.87 ± 0.284) than MCAS patients (1.65 ± 0.389), but the difference did not reach statistical significance (P = 0.155; Figure 1A). This was similar for the mean FET value of MCAD patients without (MCAD-M: 1.67 ± 0.299) and with MCAD-specific drug treatment (MCAD+M: 1.80 ± 0.393; P = 0.43), indicating that the standard FET value, in the present explorative study, was suitable to distinguish MCAD patients from healthy individuals, even when patients were already treated with MCAD-specific drugs or NSAID (Figure 1B). To get an impression concerning robustness of the FET test and the FET algorithms with respect to NSAID intake raw data of MCAS patient #8, who ingested ASA prior to blood sampling, and SM patient #8, who indicated use of ibuprofen, were modeled with subsequent calculation of the FET value under the assumption that the NSAID medication led to inhibition of cyclooxygenase activity and thus underestimated prostaglandin values and, in addition, increase in 5-lipoxygenase activity with subsequent overestimation of pLT. However, even at the assumption of an unrealistically high under- (PGE₂) and overestimation (pLT) of 50% none of the two patients would have been misdiagnosed by using the FET algorithms (see Additional file 3).

We further analyzed the release of PGD₂, because PGD₂ represents a prominent mast cell eicosanoid, which is related to enhanced mast cell activity [10],[11]. We thus hypothesized that integration of PGD₂ into the FET analysis might reveal more MCAD-specific individual eicosanoid patterns. Mean basal release of PGD₂ from PBLs was indeed markedly enhanced (6.6-fold) in patients with MCAD (946 ± 302.2 pg/ml; P < 0.001), as compared to healthy individuals (142 ± 47.8 pg/ml) (Figure 2A). A distinct cut-off value of 333.31 pg/ml was calculated, however, with a misclassification rate of 2.3%. PGD₂ release from PBLs of MCAS patients (mean: 1045 ± 280.3 pg/ml) was higher than that from PBLs of SM patients (mean: 829 ± 298.0 pg/ml), but the difference did not reach statistical significance (P = 0.096) (Figure 2A). Maximum basal PGD₂ level of 1574 pg/ml was observed for MCAS patient #5 whereas the lowest PGD₂ level was observed for SM patient #1. MCAD-M and MCAD+M patients (mean: 964 ± 250.4 pg/ml and 934 ± 343.0 pg/ml, respectively) had similar mean basal PGD₂ levels (Figure 2B), with a broader variance in the MCAD+M subgroup.

Mean basal release of PGE₂ and pLT from PBLs of MCAD patients was significantly higher (1.8-fold, P = 0.005, and 3.2-fold, P < 0.001, respectively), than that of healthy controls (Additional file 4). However, basal PGE₂ and pLT levels were less discriminative than the FET value (which considers both basal and triggered PGE₂ and pLT levels and their biochemical interactions) or the basal PGD₂ release concerning the discrimination of patients from healthy controls. ROC analysis indicated misclassification rates of 27.0% and 11.4% for basal PGE₂ and pLT, respectively, when calculating the respective optimal cut-off levels of 207.65 pg/ml (PGE₂) and 140.55 pg/ml (pLT). Notably, PBLs from MCAS patients released significantly more PGE₂ (median: 228 pg/ml; range: 108–994 pg/ml; P = 0.018) than PBLs from SM patients (median: 136; range: 40–629 pg/ml) (Additional file 4). Mean basal PGE₂ as well as pLT release was not significantly different in patients without and with MCAD-specific medication (data not shown).

On the basis of the marked basal release of PGD₂ in MCAD-patients, we expanded PGD₂ analysis by investigating AA-, ASA-, and SP-triggered release of PGD₂, since these triggering factors are routinely used in standard FET analysis and are of relevance in MCAD.

AA, the physiological substrate of both the cyclooxygenase and 5-lipoxygenase pathways, seemed to be less active in stimulating PGD₂ release from PBLs of MCAD patients than of healthy individuals, when expressed as PGD₂ ratio (i.e. PGD₂ level after AA stimulation divided by basal PGD₂ release; Table 2): Median AA-triggered PGD₂ ratio was significantly lower for MCAD patients (1.13; range: 0.20-3.26; including six non- or adverse-responder; P = 0.007) than for healthy controls (1.43; range: 0.88-2.46; one non-responder). There was no significant difference (P = 0.100) between MCAS (1.15; range: 0.63-1.63) and SM patients (1.07; range: 0.20-3.26); the drug treatment subgroups MCAD-M (1.17; range: 0.64-1.63) and MCAD+M (1.11; range: 0.20-3.26) also had similar median PGD₂ ratios after AA stimulation (Table 2).

The observed lower median AA-triggered PGD₂ ratio for MCAD patients was perhaps a result of the significantly higher basal PGD₂ levels in this group, as in particular PBLs from some MCAD patients with basal PGD₂ levels > 1000 pg/ml did not respond with enhanced, but with reduced release of PGD₂ upon AA-triggering (Figure 2, Additional file 5). We thus hypothesized that high basal PGD₂ release might limit further physiological AA-triggering. Nevertheless, median AA-triggered PGD₂ level for MCAD patients were significantly higher (median: 964 pg/ml; P < 0.001) than for healthy controls (median: 180 pg/ml; range: 144–473 pg/ml), however with high variance (range: 236–1818 pg/ml). Because of the in part opposite effects of AA on PBLs from MCAD patients and the high range of dispersion of the AA-triggered PGD₂ values, AA-triggered PGD₂ release seemed not to be a sufficiently decisive endpoint for diagnostically differentiating MCAD patients from healthy controls.

PBLs were exposed to ASA, a known trigger in some but not all MCAD patients. The median of the ASA-triggered PGD₂ ratios (i.e. PGD₂ level after incubation with ASA divided by basal PGD₂ release) was not elevated in MCAD patients, when compared with healthy individuals (1.10 and 1.10, respectively; P = 1.000), irrespective of MCAD subclass or medication subgroup (for details see Table 2). However, when evaluating the individual ASA-triggered reactions on PGD₂ release, highly variable effects were observed, in terms of no effect (e.g. SM patient #2), elevation (e.g. 3.3-fold increase for SM patient #1), or reduction in PGD₂ release (e.g. decrease to 21% of the basal PGD₂ level for SM patient #8). Variance of the ASA-triggered PGD₂ ratio was markedly higher in MCAD patients (range: 0.21-3.30) than in healthy controls (range: 0.93-1.45) (Table 2).

Since ASA can be used therapeutically to decrease prostaglandin-related symptoms in MCAD patients [24],[25], but can also trigger anaphylactic reactions, it was of special interest to look more closely into the individual ASA-triggered reactions of PBLs from MCAD patients. We, therefore, used vector plots to investigate individual ASA-triggered dynamics of the PGD₂-pLT eicosanoid pattern, because both eicosanoids can account for profound adverse reactions. These plots summarize and integrate both basal (origin of arrows) and ASA-triggered PGD₂ and pLT release (arrowhead) (Figures 3A and 3B). PBLs from healthy individuals were characterized by no or only marginal ASA-triggered modulation of the dynamics of eicosanoid release. In contrast, PBLs from MCAD patients demonstrated extended and highly individual responses to ASA with all kinds of different PGD₂-pLT dynamics. PBLs from some patients reacted with a markedly increased release, of mainly PGD₂ (e.g. MCAS patients #1 and #4, Figure 3A) or pLT [e.g. MCAS patients #6 (known to be ASA intolerant) and patient #12; Figure 3A] and some with an enhanced release of both PGD₂ and pLT (e.g. MCAS patient #8, Figure 3A, and SM patient #1, not depicted in Figure 3B, because of extraordinarily high pLT values). In other MCAD patients no or only minor effects on the PGD₂-pLT eicosanoid pattern [e.g. SM patients #2 (known to be ASA tolerant), and SM patient #5; Figure 3B] were observed upon in-vitro exposure to ASA. Furthermore, some patients demonstrated down regulation of PGD₂ and/or pLT release (e.g. MCAS patients #3 and #9, Figure 3A, and SM patient #6, Figure 3B), in part with opposed effects on the other eicosanoid (e.g. SM patients #8 and #10, Figure 3B). Taken together, the effects of ASA on the PGD₂-pLT dynamics in MCAD patients thus strongly differed between individuals with regard to the eicosanoid type/s involved as well as the extent and direction of the effect. Susceptibility of MCAD patients to ASA did not seem to be uniformly influenced by the individual MCAD-directed medication, because comparably treated patients did not demonstrate the same modulation pattern (see Additional file 6).

The neuropeptide SP was shown to be elevated in blood of patients suffering from mastocytosis [26]. In our study, SP caused a marked increase in PGD₂ release from PBLs of 19 of 22 MCAD patients (Table 2, Figures 4 and 5). The SP-triggered PGD₂ ratio (i.e. PGD₂ level after SP stimulation divided by basal PGD₂ level) was significantly higher in MCAD patients (median: 2.05; P < 0.001) than in healthy controls (median: 0.88), with wide ranges for both MCAD patients (0.79-6.64) and healthy controls (0.40-1.34) (Table 2). This was observed irrespective of the MCAD subclass or MCAD-specific drug treatment. The SP-triggered PGD₂ ratio was significantly higher for SM patients (median: 2.56, range: 1.61-6.64; P = 0.044) than for MCAS patients (median: 1.71; range: 0.79-4.87). The medication subgroups demonstrated nearly identical SP-triggered PGD₂ ratios (MCAD-M: median: 2.07; range: 0.99-4.87; MCAD+M: median 2.04; range: 0.79-6.64), however with a broader variance in MCAD-M patients. The SP-triggered PGD₂ ratio was maximal for SM patient #1 (6.64), followed by MCAS patient #11 (4.87), and SM patient #3 (3.48). In 3 of 22 MCAD patients (MCAS patients #5, #8, and #12) no increase or even a slight decrease in SP-triggered PGD₂-ratio of PBLs occurred.

In view of the obviously strong effect of SP on PGD₂ release from PBLs of MCAD patients and considering the known biochemical interaction of the COX and 5-LOX pathways, we further analyzed the effect of SP on both PGD₂ and pLT levels in more detail. Again, we used vector plots to investigate individual SP-triggered dynamics of the PGD₂-pLT eicosanoid pattern. These plots summarize and integrate both basal (origin of arrows) and SP-triggered PGD₂ and pLT release (arrowhead). PBLs from healthy individuals were characterized by no or only marginal SP-triggered modulation of the dynamics of eicosanoid release (Figures 4A and 4B). In contrast, PBLs from 19 out of 22 MCAD patients demonstrated marked increase in SP-triggered PGD₂ release revealing a significantly (P < 0.001) enhanced mean PGD₂ level of 1896 ± 389.7 pg/ml (range: 939 to 2303 pg/ml), as compared to 130 ± 53.5 pg/ml (range: 50–294 pg/ml) for healthy individuals (Figures 4 and 5). PBLs from SM patients exhibited significantly higher SP-triggered PGD₂ release (2079 ± 207.2 pg/ml; P = 0.0205; Figure 5A) than PBLs from MCAS patients (1743 ± 446.0 pg/ml; Figure 5A). No significant difference was noted between the two medication groups MCAD-M and MCAD+M (1863 ± 376.8 and 1919 ± 411.9 pg/ml, respectively; P = 0.373; Figure 5B). Notably, two of the SP non-responders, concerning PGD₂ release, instead showed enhanced pLT release upon SP-stimulation. Analysis of the vector plots demonstrated partially pronounced stimulation of SP-triggered pLT release for 18 of 22 MCAD patients (Figure 4). Mean pLT release amounted to 393 ± 137.2 pg/ml (range: 188–687 pg/ml) for MCAD patients, as compared to 72 ± 33.1 pg/ml (range: 23–131 pg/ml) for healthy individuals (Additional file 7). MCAS patients exhibited a higher mean pLT release of 433 ± 153.9 pg/ml than SM patients (mean: 345 ± 101.4 pg/ml) upon SP stimulation, but the difference did not reach statistical significance (P = 0.068; Additional file 7).

As depicted in Figures 4 and 5 and in Additional file 7, there was no overlap of the individual SP-triggered PGD₂ and pLT levels of MCAD patients with that of healthy individuals. Subsequently, distinct cut-off values of 616.38 pg/ml and 159.59 pg/ml could be calculated for SP-triggered PGD₂ and SP-triggered pLT levels, respectively, with each 0.0% misclassification rate. Especially the SP-triggered PGD₂ level was highly decisive concerning distinction of MCAD patients from healthy controls (Figure 5).