Plasminogen activator inhibitor (PAI-1) is an anticancer agent that inhibits plasmin driven proteolysis, limiting angiogenesis and metastasis. In low concentrations it could induce cancer cell motility by interacting with urokinase (uPA), its receptor (uPAR), vitronectin and integrins. Active PAI-1 binds to uPA forming a complex with uPAR, while the latent form of PAI-1 does not. PAI-1 is found in both forms in the circulation. It is not clear which form acts as an anticancer agent and how it interacts with malignant cells. To investigate how these forms reduce angiogenesis or metastasis, we have created PAI-1 cysteine mutants in the active conformation (VLHL PAI-1) with an extended half-life that reaches ~700 h and its R369A mutant, which has an active conformation but cannot bind to uPA (VLHL NS PAI-1). Both VLHL PAI-1s convert into the latent form when treated with a reducing agent (DTT) that breaks disulfide bridges. Unexpectedly, during routine investigation of LnCAP cell proliferation, we have found that cells detach from the culture vessels regardless of PAI-1 conformation or activity. Further investigation showed that treatment of cancer cells with VLHL PAI-1 downregulated nucleophosmin, while all forms of PAI-1 downregulated fortilin. These two proteins are implicated in important cellular processes (cell growth, cell cycle, malignant transformation). This suggests that PAI-1, in addition to its well-known anticancer properties, plays an important role in cell signaling. We hope that by exploring PAI-1's structure and function we might be able to understand and separate the different effects of PAI-1 on cancer cells and develop more effective therapeutic strategies in cancer treatment.

Plasminogen activator inhibitor-1 (PAI-1), a member of the serpin super-family, forms a covalent complex with its target proteinases, such as tissue and urokinase plasminogen activators. Thus, PAI-1 controls the physiological and pathological proteolysis. An abnormal expression of PAI-1 has been observed in different diseases, which can be treated by returning the proteolysis back to normal physiological levels. It has been reported that some PAI-1 inhibitors neutralize its activity by accelerating the conversion of PAI-1 into a latent form. We have found small organic chemicals that also neutralize PAI-1 activity, but by a different mechanism. Using the NBD fluorescent probe [N,N'-dimethyl-N-(acetyl)-N'-(7nitrobenz-2-oxa-1,3-diazol-4-yl)] incorporated into the reactive center loop (RCL) of PAI-1, we measured the kinetics of conversion from an active to a latent form. Unexpectedly, we found that some inhibitors of PAI-1 arrest this serpin in its active form instead of increasing the speed of conversion. Using docking calculations, we located two possible binding sites for these chemicals. The sites are in proximity of the P1/P1' amino acids of the RCL of PAI-1. Binding in this area can inactivate PAI-1 and additionally create a steric obstacle on the RCL making insertion of this loop between the A3 and A5 strands more difficult; hence abolishing a necessary step in the conversion of this protein into the latent form. Additionally, PAI-1 inhibitors link the RCL of one PAI-1 molecule with the strand 3C and strand 4C or helix A and strand 1B regions of the other PAI-1 molecule aiding polymerization or stabilizing the junction of the two. The polymerization of PAI-1 reduces PAI-1 activity by encapsulating the critical RCL fragment inside the formed PAI-1/ PAI-1 polymers.

Proteolytic activity initiated by uPA is commonly recognized as a critical factor in angiogenesis and metastasis. Many cancers overexpress uPA and the reduction of proteolytic activity has been proposed as a cancer treatment option. Indeed, uPA inhibitors have been shown to reduce angiogenesis and tumor growth. Thus, we want to identify novel inhibitors of uPA suitable for cancer treatment. We have chosen PAI-1, which inhibits the urokinase plasminogen activator. However, PAI-1 is not a stable molecule and converts itself into the latent form with a half-life in the range of t1/2~2 hours. Based on the known structure of active PAI-1, we have identified amino acids that can be substituted with a cysteine residue to produce disulfide bridges linking the top and bottom parts of strands A3 and A5 as well as sites within the helix D region in hopes of preventing a conversion into the latent PAI-1. We have created a total of seven cysteine mutants via point mutation (two to six point mutations) generating possible sites for a disulfide bridge formation at the top and bottom parts of A3 and A5, within the helix D region, or by a combination thereof. Desired mutations were introduced by PCR using appropriate primers. The mutant forms of PAI-1 containing the chitin binding intein tag were then purified using affinity chromatography wherein the intein tag is cleaved leaving the mutant PAI-1 protein. Cysteine mutations resulted in proteins with an extended half-life of PAI-1 from 2 to over 700 hours depending on the mutant. Novel PAI-1s were fully functional against uPA and showed activity in the in vitro model of angiogenesis e.g. inhibition of sprout formation. The mutant with the longest half-life (one that produced a disulfide bridge linking the top part of strands A3 and A5) was chosen for further study. VLHL PAI-1 expressed in the E. Coli vector produced a modest yield of 1 mg-purified protein from 1L of cell culture. Thus, we expressed it in the baculovirus vector, with an 6His purification tag that produced ~18 mg of PAI-1/1L. Two different forms were made: fully active VLHL PAI-1 and VLHLNS PAI-1 with an Arg369?Ala mutation in the P1 position, which will be used in future anticancer study as a negative control. VLHLNS PAI-1 mutation (Arg369?Ala) was introduced by PCR and gene was transferred into the baculovirus vector in the same way as VLHL PAI-1. We assume that the VLHLNS construct will remain in an active conformation as VLHL PAI-1 does, but will not have any inhibitory activity toward uPA. Our study suggests that both proteins are in an active conformation, however they convert into the latent form after treatment with reducing agents, which break the disulfide bridge. Fully active VLHL PAI-1 inhibits uPA as demonstrated by chromogenic assay (SPEC-TROZYME®) and forms the uPA/PAI-1 complex as shown on PAGE gel while VLHLNS PAI-1 does not. We believe that PAI-1s with extended half-lives are therapeutically desired in cancer treatment and Cys mutated PAI-1s could launch a new class of novel anti-cancer agents.

The mutation of amino acids, Gln197, Gly355?Cys, produce a PAI-1 with a very long half-life of over 700 h as reported previously. However, VLHL PAI-1 was expressed in E. Coli with a modest yield of 1 mg-purified protein in 1L of cell culture. To increase the yield, we expressed VLHL PAI-1 in baculovirus. The cDNA encoding of PAI-1 was excised from the VLHL PAI-1 plasmid as a NdeI/Xho fragment. Afterwards, the PCR product of the VLHL PAI-1 NdeI/Xho I fragment was ligated into the pFastbac plasmid containing a 6His purification tag. VLHLNS PAI-1 mutation (at P1, Arg369?Ala) was introduced by PCR and the gene was transferred into the baculovirus vector in the same way as VLHL PAI-1. We assumed that the VLHLNS construct would remain in an active conformation as VLHL PAI-1 does, but would not have any inhibitory activity toward uPA. We used the baculovirus expression system, which promises ~20 mg of protein from 1L of cell culture. In the pFastbac vector, the expression of the gene is controlled by the Autographa californica multiple nuclear polyhedrosis virus (AcMNPV), polyhedrin (PH), or the p10 promoter for high-level expression in insect cells. The plasmids were transposed into a recombinant bacmid with the help of DH10Bac E. coli cells (Invitrogen), which contain a baculovirus shuttle vector (Bacmid) with a min-attTn7 target site and a helper plasmid. The recombinant bacmid DNA was isolated from the white colonies grown for 48 h at 37°C on a LB agar plate containing 50mg/L kanamycin, 7mg/L gentamycin, 10mg/L tetracycline, 100mg/L X-gal and 40mg/L, IPTG and which was used to transfect Sf9 cells derived from Spodoptera Frugiperda (Fall Armyworm) by cellfectin reagent (Invitrogen). The virus was amplified to ~ 2x107 plaque forming units (pfu)/mL and was added to Sf9 cells (~2x106/mL) in 6 well culture plates. The plates were incubated at 27°C for different time intervals. The virus was subsequently amplified to ~ 2x108 plaque forming units (pfu)/mL and was used to infect Sf9 cells (~2x106/mL) on a large scale (1L cell culture). The cells were harvested and lysed by two freeze-thaw cycles. The lysate was then centrifuged at 3,000xg for 20 minutes to pellet the cellular debris. The supernatant was transferred to a fresh tube for purification. The supernatant was loaded onto a column packed with nickel resin (Invitrogen) at a flow rate of 0.4 ml/min and the column was then washed with a buffer containing 40 mM imidazole in 20mM Hepes buffer pH 8.0, and protease inhibitors until no washed proteins were detected. The protein was then eluted using a gradient of 40 -250 mM imidazole in native buffer. The peak fractions were dialyzed to remove imidazole and concentrated to a desired concentration. VLHL PAI-1 was active against uPA, however VLHLNS PAI-1, as expected, was not. Most likely, both PAI-1s are in an active conformation. Purity was determined as +95%.