AID	Name	Type	Comment
680	Factor XIa Mixture HTS	screening	Factor XIa mixture screen; hits were confirmed in this assay

Molecular Library Screening Center Network (MLSCN)
Penn Center for Molecular Discovery (PCMD)
Assay Provider: Scott L. Diamond, University of Pennsylvania
MLSCN Grant: X01-MH076406-01

Target

Factor XI (FXI) circulates as a complex with high molecular weight kininogen (HK) in the plasma at a concentration of 5 ug/ml (equivalent to 31.3 nM, dimeric concentration) as a homodimeric glycosylated blood plasma zymogen of approximately 160 kDa, containing monomeric subunits of 80 kDa each (1). Thrombin (2, 3), factor XIa (FXIa) (3), and factor XIIa alpha (FXIIa) (4), all cleave an internal R369-I370 bond in each monomer of FXI, yielding the enzyme FXIa.

After activation from FXI to FXIa, FXIa possesses a heavy chain of 369 residues and a light chain of 238 residues. The heavy chain consists of four apple domains (A1-A4) and the light chain represents a trypsin-like serine protease domain with active site residues at H413, D464, and S557 (1, 5-7). FXIa catalyzes FIX to FIXa activation by cleaving two scissile bonds at R145 and R180 (8). The FIXa generated can catalyze FXa formation on the platelet surface with the active cofactor factor VIIIa (FVIIIa), with FVIIIa increasing the Vmax of FX activation by FIXa by 100,000 fold (9). After activation of sufficient levels of FXa by the consolidation pathway, FXa can go on to form a ternary complex with FVa and prothrombin on the platelet surface, to give sufficient levels of thrombin for activation of fibrinogen to fibrin. Formation of this ternary prothrombinase complex in the presence of phospholipids has been shown to increase the rate of prothrombin to thrombin activation by 300,000-fold more than with FXa and prothrombin alone (10).

Thrombin catalyzes activation of fibrinogen to fibrin, cleaving peptides 14-16 amino acids in length, called fibrinopeptides, from the A alpha and B beta subunits of fibrinogen (11, 12). The fibrin monomers produced by thrombin form a noncovalent meshwork with other fibrin molecules to produce a fibrin clot that stabilizes and retracts the initial platelet plug described above. This stable fibrin clot formation stabilizes the primary platelet plug, and reduces the volume of the plug in order to arrest blood loss.

The role of FXI in thrombosis was investigated using a mouse double knockout of FXI and protein C (PC). First, the single FXI knockout in the mouse did not cause a significant bleeding problem in mice, shown by comparable tail bleeding times between FXI knockout mice and wt mice, as opposed to the FIX knockout mouse which had a tail bleeding time increase of > 5.8-fold over wt mice (13). Additionally, using a FXI deficient mouse model, mice were partially protected from arterial occlusion in a FeCl3 model of thrombosis (13). Thus, by this model, controlling or limiting FXI activity can have a protective effect on hemostasis by preventing pathologic thrombus formation. Further studies of FXI under rapid arterial flow conditions simulating rapid thrombus growth with an anti-FXI antibody that reduced intraluminal thrombus growth did suggest that inhibiting FXI can have a mild, yet controlled antithrombotic effect that may be of significant use clinically (14). A recent report found that increased risk of cerebrovascular events is associated with elevated FXI activity and antigenic levels, implicating FXI as a risk factor in deep venous thrombosis (15). When double knockout mice lacking both FXI and activated protein C (aPC) were examined, the mice possessed greater control over thrombosis than the aPC knockouts, suggesting that FXI deficiency helped these mice to survive past the neonatal period. Thus, the absence of FXI prevented the occurrence of fatal disseminated intravascular coagulation (DIC) in this model (16).

In vivo findings in rodent models of coagulation in addition to clinical data from patients at risk for pathologic cerebrovascular events have given rise to the recognition of FXI as a fine modulator and focal point in blood coagulation whose inhibition does not cause a severe bleeding diathesis, and suggest that drug design using the tools of high throughput chemical inhibitor screening may give rise to novel anticoagulant therapies (17).

HTS was performed using 62,107 compounds of the MLSCN library combined as mixtures of 10 orthogonally pooled compounds totaling 2.5 mM, which were diluted 50-fold into 10 ul 384 well assay plates (final concentration 5 uM each compound). The assay used to test for percent inhibition was a fluorescence assay utilizing hydrolysis of Boc-Glu-Ala-Arg-AMC, as first described by Kawabata et al. (18). Hits with >20% inhibition in both wells of the mixture plates were retested in complete IC50 curves.

1. K. Fujikawa, D. W. Chung, L. E. Hendrickson, E. W. Davie, Biochemistry 25, 2417 (1986).
2. D. Gailani, G. J. Broze, Jr., Science 253, 909 (Aug 23, 1991).
3. K. Naito, K. Fujikawa, Journal of Biological Chemistry 266, 7353 (1991).
4. B. N. Bouma, J. H. Griffin, Journal of Biological Chemistry 252, 6432 (1977).
5. T. Koide, M. A. Hermodson, E. W. Davie, Nature 266, 729 (Apr 21, 1977).
6. F. van der Graaf, J. S. Greengard, B. N. Bouma, D. M. Kerbiriou, J. H. Griffin, Journal of Biological Chemistry 258, 9669 (1983).
7. B. A. McMullen, K. Fujikawa, E. W. Davie, Biochemistry 30, 2056 (Feb 26, 1991).
8. K. Fujikawa, M. E. Legaz, H. Kato, E. W. Davie, Biochemistry 13, 4508 (1974).
9. G. van Dieijen, G. Tans, J. Rosing, H. C. Hemker, J Biol Chem 256, 3433 (Apr 10, 1981).
10. J. P. Miletich, C. M. Jackson, P. W. Majerus, Journal of Biological Chemistry 253, 6908 (1978).
11. K. Bailey, F. R. Bettelheim, L. Lorand, W. R. Middlebrook, Nature 167, 233 (Feb 10, 1951).
12. B. Blomback, M. Blomback, Ann N Y Acad Sci 202, 77 (Dec 8, 1972).
13. X. Wang et al., J Thromb Haemost 3, 695 (Apr, 2005).
14. A. Gruber, S. R. Hanson, Blood 102, 953 (Aug 1, 2003).
15. D. T. Yang, Flanders, M.M., Rodgers, G.M., Blood 106, 2626 (November 16, 2005, 2005).
16. J. C. Chan et al., American Journal of Pathology 158, 469 (Feb, 2001).
17. A. Gruber, S. R. Hanson, Curr Pharm Des 9, 2367 (2003).
18. S. Kawabata et al., European Journal of Biochemistry 172, 17 (Feb 15, 1988).

Materials

Human plasma factor XIa was purchased from Enzyme Research Laboratories (Cat # HFXIa 1111a). Substrate Boc-Glu-Ala-Arg-AMC was from Bachem (Cat #I-1575.0050). Assay buffer consisted of 50 mM Tris, pH 7.4, 150 mM sodium chloride, 0.02% Tween 20. Low-volume 384-well black plates were from Corning (Item #3676).

Assay

Factor XIa (0.23 ug/mL) was incubated with Boc-Glu-Ala-Arg-AMC substrate (15 uM) in 10 uL of assay buffer (see above) for 2 hr at room temperature. Hits from mixture HTS were confirmed on single compounds by IC50 determination as described below.

IC50 protocol

1.Serial dilute single compounds at 50x concentration in DMSO (16 two-fold dilutions from 2.5 mM to 75 nM)
2.Fill low-volume plate with 4 uL water using Multidrop-micro
3.Add 5 uL assay buffer to columns 1 and 23 using Multidrop-384
4.Add 200 nL of compound (in DMSO from step 1) using Evolution pintool
5.Add 1 uL of Boc-Glu-Ala-Arg-AMC substrate (150 uM in 5x assay buffer) using Multidrop-micro
6.Add 5 uL enzyme (0.46 ug/mL in assay buffer) using Multidrop-384
7.Incubate for 2 hr at room temperature
8.Read fluorescence (excitation 355, emission 460) on Envision reader

Data analysis

IC50 plates contained compounds in columns 3-22, controls (enzyme, no compound) in columns 2 and 24, and blanks (no enzyme) in columns 1 and 23. Each column 3-22 contained 16 two-fold dilutions of a single compound, ranging in concentration from 50 uM to 1.5 nM. Percent activity was calculated for each dilution of each compound from the signal in fluorescence units (FU) and the mean of the plate controls and the mean of the plate blanks using the following equation:

% Activity = 100*((signal-blank mean)/(control mean-blank mean))

Dose response curves of percent activity were fit using XLfit equation 205 (four parameter logistic fit with maximum percent activity and minimum percent activity fixed at 100 and 0, respectively).

Activity scoring

The activity score reported here is based on follow-up IC50 testing on compounds that showed >20% inhibition in both locations in the primary HTS:

IC50 score = sum (IC50 score #1, IC50 score #2, IC50 score #3).

IC50 scores were calculated as follows:

(1) Score = 5.75 x (pIC50-3), where pIC50 = -log(10) of IC50 in mol/L
(2) For IC50 >50 uM, the score was calculated from percent activity at maximum concentration tested in assay (50 uM):
Score = [5.75 x (0-3)] + [(100-percent activity at max concentration)/1.75]

Activity Outcome

Compounds that gave percent inhibition >20 in both locations in the primary HTS were judged to be hits and these compounds were selected for follow-up IC50 testing. IC50 values were determined as described in protocol above. The percent activity at the maximum concentration is reported and can be used to estimate the potency of compounds for which the IC50 values were >50 uM.

Activity outcome is reported as follows:

(1) IC50 <50 uM in all three IC50 determinations = active
(2) IC50 <50 uM in only 1 or 2 out of 3 determinations = inconclusive
(3) IC50 >50 uM, percent inhibition 30-50% at 50 uM = inconclusive
(4) IC50 >50 uM, percent inhibition <30% at 50 uM = inactive

Contributors

This assay was submitted to the PCMD by Scott Diamond, assay development and HTS were conducted by Paul Riley, and data were submitted by Andrew Napper and Paul Riley, all of the University of Pennsylvania.