Here we describe an in-house extracellular vesicle tissue factor activity assay. Activity-based assays and antigen-based assays have been used to measure tissue factor in extracellular vesicles from human plasma samples. Activity-based assays have higher sensitivity and specificity than antigen-based assays.
Tissue factor (TF) is a transmembrane receptor for factor (F) VII and FVIIa. The TF/FVIIa complex initiates the coagulation cascade by activating both FIX and FX. TF is released from cells into the circulation in the form of extracellular vesicles (EVs). The level of TF-positive (+) EVs is increased in various diseases, including cancer, bacterial and viral infections, and cirrhosis, and is associated with thrombosis, disseminated intravascular coagulation, disease severity, and mortality. There are two ways to measure TF+ EVs in plasma: antigen- and activity-based assays. Data indicates that activity-based assays have higher sensitivity and specificity than antigen-based assays. This paper describes our in-house EVTF activity assay based on a two-stage FXa generation assay. FVIIa, FX, and calcium are added to the TF+ EV-containing samples to generate FXa in the presence and absence of anti-TF antibody to distinguish TF-dependent FXa generation from TF-independent FXa generation. A chromogenic substrate cleaved by FXa is used to determine the FXa level, while a standard curve generated with a relipidated recombinant TF is used for the determination of the TF concentration. This in-house EVTF activity assay has higher sensitivity and specificity than a commercial TF activity assay.
Blood coagulation is initiated with the binding of factor (F) VII/VIIa to tissue factor (TF)1. The TF/FVIIa complex activates both FIX and FX to activate blood coagulation1. There are two forms of full-length, membrane-bound TF: encrypted and active. In addition, there is an alternatively spliced form of TF (asTF). Sphingomyelin and phosphatidylcholine in the outer leaflet of the cell membrane maintain TF in an encrypted state2,3,4. When cells are activated or damaged, phospholipid scramblase transfers phosphatidylserine and other negatively charged phospholipids to the outer leaflet1. The activation of cells also results in the translocation of acid sphingomyelinase to the outer leaflet where it degrades sphingomyelin to ceramide5. These two mechanisms convert encrypted TF to the active form. It is also proposed that protein disulfide isomerase mediates disulfide bond formation between Cys186 and Cys209 in encrypted TF, which results in the de-encryption of TF6,7,8. asTF is also present in the circulation but lacks the transmembrane domain and is therefore soluble9,10. Importantly, asTF has very low levels of procoagulant activity compared to full-length active TF10,11.
Extracellular vesicles (EVs) are released from resting, activated, and dying host cells, as well as cancer cells12. EVs express proteins from their parental cells12. Active TF-bearing EVs are released from activated monocytes, endothelial cells, and tumor cells into the circulation13,14,15. Levels of TF in plasma can be measured by activity- and antigen-based assays. Antigen-based assays include ELISA and flow cytometry16. There are two different activity-based assays: one and two-stage TF activity assays. The one-stage assay is based on a plasma-based clotting assay. The TF-containing sample is added to plasma and the time to form a clot is measured after re-calcification. The two-stage assay measures FXa generation of samples by adding FVII or FVIIa, FX, and calcium. FXa levels are determined using a substrate that is cleaved by FXa.
In both the one- and two-stage TF activity assays, the TF concentration is determined using a standard curve generated with recombinant TF. Two-stage assays have higher sensitivity and specificity than the one-stage assay. Many studies have confirmed that activity-based assays have higher sensitivity and specificity than antigen-based assays17,18,19,20,21. In addition, our in-house activity assay has higher sensitivity and specificity than a commercial activity assay22. Healthy individuals have very low or undetectable levels of EVTF activity in plasma. In contrast, individuals with pathologic conditions, such as cancer, cirrhosis, sepsis, and viral infection, have detectable levels of EVTF activity and this is associated with thrombosis, disseminated intravascular coagulation, disease severity, and mortality23,24,25,26,27,28. Here, we will describe this in-house two-stage EVTF activity assay.
The research was approved by the Institutional Review Board of the University of North Carolina at Chapel Hill (protocol number: 14-2108).
1. Blood collection from donors
2. Preparation of negative control plasma
NOTE: Samples should not be placed on ice at any time before the final freeze.
3. Preparation of positive control plasma
NOTE: Response to lipopolysaccharide is variable between individuals.
4. Isolation of extracellular vesicles from plasma
NOTE: EV pellets may not be visible. The defrosting time is dependent on sample volume but we usually defrost 100 µL plasma for 30 min at 37 °C.
5. Option: Isolation of small extracellular vesicles using ultracentrifuge
6. Measurement of extracellular vesicle tissue factor activity
NOTE: EV samples should be pipetted and vortexed well to mix before adding to wells. EDTA-disodium will inhibit the ability of FXa to cleave the chromogenic substrate. Therefore, EDTA-disodium cannot be used as a replacement for EDTA-tetrasodium in step 6.7.
A successful result gives a positive control value of ≥0.5 pg/mL and a negative control value of <0.5 pg/mL. It is best to find a high LPS responder with >1.0 pg/mL EVTF activity for a positive control. The representative result shows the EVTF activity of EVs isolated from plasma from the whole blood of 11 healthy donors, with and without LPS activation (Figure 4). Six out of eleven donors (Donors 2, 4, 5, 8, 10, 11) were medium to high LPS responders, whereas five out of eleven donors (Donors 1, 3, 6, 7, 9) were low LPS responders.
Figure 1: Preparation of platelet-depleted plasma. The figure was modified from Hisada and Mackman29. Abbreviation: PDP = platelet-depleted plasma. Please click here to view a larger version of this figure.
Figure 2: Isolation of extracellular vesicles from platelet-depleted plasma. The figure was modified from Hisada and Mackman29. Abbreviations: PDP = platelet-depleted plasma; EVs = extracellular vesicles. Please click here to view a larger version of this figure.
Figure 3: Measurement of extracellular vesicle tissue factor activity. The figure was modified from Hisada and Mackman29. Abbreviations: TF = tissue factor; EVs = extracellular vesicles. Please click here to view a larger version of this figure.
Figure 4: EVTF activity of plasma from whole blood of 11 healthy donors, with and without LPS activation. LPS activation was performed under the same condition as the positive control. The positive control was from a known LPS responder whereas the responses to LPS of 11 healthy donors were not known at the time of the experiment. White and black bars indicate with and without LPS activation, respectively. Abbreviations: EVTF = extracellular vesicle tissue factor; LPS = lipopolysaccharide. Please click here to view a larger version of this figure.
Table 1: A summary of the studies comparing this in-house EVTF activity assay, a commercial TF ELISA, and an activity assay. Abbreviations: EVTF = extracellular vesicle tissue factor; ELISA = enzyme-linked immunosorbent assay. Please click here to download this Table.
Here, the protocol of our in-house EVTF activity assay is presented. There are three critical steps in the protocol. When reconstituting the EV pellet, it is important to pipette up and down at the location of the EV pellet even if it is not visible. Incomplete reconstitution of the EV pellet will result in a false negative or underestimation of EVTF activity values of the samples. Second, using HBSA-Ca(+) is critical in protocol step 6.5 because FXa generation cannot be generated without calcium. Third, using EDTA-tetrasodium but not EDTA-disodium to stop FXa generation is critical because the latter inhibits the ability of FXa to cleave the chromogenic substrate.
We recommend the following storage and use method for each reagent. We store assay buffers and the reconstituted substrate at 4 °C and repeatedly use them until we run out of them. If we do not run out of the substrate within 2-3 weeks, we discard it because it turns yellow. We store recombinant TF, antibodies, FVIIa, and FX at -80 °C and recommend making aliquots for one assay to avoid a freeze-thaw cycle.
In a preliminary study, we left plasma samples for 9 h at room temperature, refroze them at -80 °C overnight, and performed the assay the next day. As a control, we thawed the same plasma and refroze it immediately. We found that the sample left for 9 h at room temperature had a 16% decrease in EVTF activity compared to the one that was thawed and refrozen immediately (Hisada and Mackman, unpublished data 2017).
In addition to the difference in response to LPS, the difference in the concentration of EVs in plasma from individual donors may explain the different EVTF activity of different donors. Indeed, we observed some variations in the concentrations of EVs in plasma from LPS-stimulated whole blood (1.9 ± 0.6 × 1010 particles/mL, n = 6, mean ± standard deviation)30.
This in-house EVTF activity assay has several features that commercial assays do not have. First, the assay uses an anti-TF antibody to distinguish TF-dependent FXa generation from TF-independent FXa generation. Three commercial activity assays are claimed to measure TF activity in plasma but none of them uses anti-TF antibodies16. Therefore, these commercial assays simply measure total FXa generation but not TF-dependent FXa generation. Second, we use 2.4 nM of FVIIa (the final concentration in the reaction) to minimize TF-independent FXa generation by FVIIa. One commercial activity assay uses 12 nM of FVIIa in the reaction, which gives a high background22. Indeed, FVIIa activates FX linearly depending on FVIIa concentration in the absence of TF31. We summarized the studies that compared this in-house EVTF activity assay and commercial TF ELISA and an activity assay in Table 1.
Four response classifications of EVTF activity for citrated platelet-poor plasma are proposed: zero (0 to <0.5 pg/mL); weak (0.5 to <1.0 pg/mL), moderate (1.0 to <2.0 pg/mL), and strong (≥2.0 pg/mL)23. Notably, different anticoagulants affect the results of EVTF activity assays. For instance, higher levels of EVTF activity were observed using heparin after LPS stimulation compared to sodium citrate (Donor A: 4.6 (citrate) vs 28.6 (heparin) pg/mL and Donor B: 3.4 (citrate) vs 26.0 (heparin) pg/mL, respectively, Hisada and Mackman unpublished data 2017). In contrast, lower levels of EVTF activity were observed using EDTA after LPS stimulation compared to sodium citrate (Donor C: 3.9 (citrate) vs 1.4 (EDTA) pg/mL, Donor D: 3.8 (citrate) vs 0.4 (EDTA) pg/mL, respectively, Tatsumi and Mackman unpublished data 2016). These results indicate that EVTF activity data are not comparable when different anticoagulants are used.
There are a few limitations to the method. The coefficient of variation (CV) is relatively high compared to clinical assays. Indeed, the CV for the positive control in seven independent studies was 24%17,29. Isolation of EVs using centrifugation pellets not only EVs but also cellular debris. We previously analyzed the pellet of platelet-rich plasma by transmission electron microscopy and observed platelets, EVs, and cellular debris32. We use a fixed-angle rotor and have not experienced a swing-out rotor for 20,000 × g centrifugation. We recently found that small EVs that can be pelleted with 100,000 × g but not with 20,000 × g have EVTF activity in patients with pancreatic cancer and COVID-1930. This indicates that this protocol does not measure EVTF activity from small EVs, such as exosomes. Notably, exosomes are sphingomyelin-rich EVs and the presence of TF-bearing exosomes has been reported30,33,34,35. We recommend pelleting EVs in plasma using 100,000 × g to more accurately determine the total amount of EVTF activity in the sample. Of note, it requires an ultracentrifuge and a longer assay time because the centrifugation time increases from 15 min to 70 min.
This method can be applied to plasma samples from any kind of disease. EVTF activity is associated with disease severity and survival in patients with different diseases, including cancer, COVID-19, bacterial infection, and cirrhosis23,26,27.
The authors have nothing to disclose.
This work was supported by the NIH NHLBI R35HL155657 (N.M.) and the John C. Parker professorship (N.M.). We would like to thank Ms. Sierra J. Archibald for her helpful comments
1.5 mL tube for 20,000 x g centrifuge | any company | N/A | We use the one from Fisher Scientific (Catalog number: 05-408-129). |
1.5 mL tube for ultracentrifuge | any company | N/A | We use the one from Beckman Coulter (Catalog number: 357448) |
15 mL tube | any company | N/A | We use the one from VWR (Catalog number: 89039-666) |
21 G x .75 in. BD Vacutainer Safety-Lok Blood Collection Set with 12 in. tubing and luer adapter | BD | 367281 | |
96-well plate | any company | N/A | We use the one from Globe Scientific (Catalog number: 120338). |
BD Vacutainer Citrate Tubes | BD | 363083 | |
Bovine serum albumin | Sigma Aldrich | A9418 | |
Calcium chloride | Fisher Scientific | C69-500 | |
Centrifuge for 1.5 mL tube | any company | N/A | We use the Centrifuge 5417R (Eppendorf). |
Centrifuge for 15 mL tube | any company | N/A | We use the Centrifuge 5810R (Eppendorf). |
D-(+)-Glucose | Sigma Aldrich | G7021 | |
Ethylenediaminetetraacetic acid tetrasodium salt dihydrate | Sigma Aldrich | E6511 | |
Hepes | Sigma Aldrich | H4034 | |
Human FVIIa | Enzyme Research Laboratory | HFVIIa | The solution should be diluted with HBSA-Ca(+). |
Human FX | Enzyme Research Laboratory | HFX1010 | The solution should be diluted with HBSA-Ca(+). |
Inhibitory mouse anti-human tissue factor IgG, clone HTF-1 | Fisher Scientific | 550252 | |
Lipopolysaccharide from Escherichia coli O111:B4 | Sigma Aldrich | L2630 | There are several lipopolysaccharide from different E. coli. Different lipopolysaccharide have different potential to activate monocytes. |
Mouse IgG | Sigma Aldrich | I5381 | |
Pefachrome FXa 8595 | Enzyme Research Laboratory | 085-27 | |
Plate reader | any company | N/A | We use the SpectraMax i3x from Molecular Devices |
Re-lipidated recombinant tissue factor, Dade Innovin | Siemens | 10873566 | |
Sodium chloride | Fisher Scientific | S271-500 | |
Ultracentrifuge | Beckman Coulter | Optima TLX | |
Ultracentrifuge rotor | Beckman Coulter | TLA-55 |