Cathepsins play an important role in cellular protein degradation through the lysosomal pathway,1
and as such, represent a significant class of drug targets, having been implicated in a variety of degenerative and invasive processes including: asthma, Alzheimer’s disease,2
diabetes, inflammatory diseases,5
liver disorders, multiple sclerosis, muscular dystrophy, and pancreatitis.6
The large number of disease states associated with the biological effects of the cathepsin protease family, and specifically cathepsin B, demand an understanding of their biological function. The availability of small molecule probes of the cathepsins thus holds the promise both for characterization of this ubiquitous enzyme class and for the discovery of new cysteine protease inhibitors of considerable biomedical value.3,7,8
Recently, the Penn Center for Molecular Discovery (PCMD),9
completed a high throughput screen (HTS) to identify small molecule probes (i.e., inhibitors) for the papain-like cysteine protease family, including cathepsins B, L, and S. While a number of both potent and selective inhibitors have been described previously,2-8
this project presented an opportunity to annotate the NIH Molecular Libraries Small Molecule Repository (MLSMR) through deposition of data in PubChem.10
As such, this work represents one of the first efforts to create a comprehensive, publicly available profile of small-molecule inhibitors of the cysteine protease class. Screening 63,332 members of the MLSMR against human liver cathepsin B resulted in a number of hits.11
Further confirmatory assays included IC50
determination and elimination of false positives resulting from non-specific redox chemistry.12
Based on these results, a family of substituted pyrazole esters was identified that displayed promising activity as inhibitors of cathepsin B ().13
Pyrazole HTS hits in the cathepsin B assay
The pyrazole series (1-6) displayed suitable physicochemical properties (MW 347-379 and XlogP = 2.7-3.3) with structurally related analogs revealing a range of activities in our HTS analysis. Based on these attributes, members of this series were viewed as potential small molecule lead compounds for the inhibition of cathepsin B, and as such were selected for further study to understand and improve their biological profile.
Structurally related pyrazoles that were inactive in the HTS assay were compared to compounds 1-6 in a preliminary effort to understand the requirements for activity. Based on this analysis, we pinpointed the ester group as a functional moiety important for activity; we reasoned that the ester could act as an electrophilic site that would react with the active site cysteine. Alternatively, the compound could behave as an alternate substrate of the enzyme resulting in transesterification or hydrolysis of the ester functionality. To differentiate these two scenarios, a series of analogs were constructed and subjected to biochemical and analytical studies to define the mechanism of action. Herein we report the results of this effort.
To confirm both the activity and identity of the original hits, we resynthesized compounds 1-6
This exercise permitted both the development of an effective method to construct the pyrazole scaffold and the preparation of analogs. The synthesis of l,3-disubstituted-5-aminopyrazoles (1-6
) began with commercially available cyanoacetohydrazide (7
) and the appropriate aryl-substituted sulfonyl chlorides, according to the procedure of Elgemeie and co-workers.16-17
Reaction of 7
with the aryl-substituted sulfonyl chlorides in ethanol furnished the sulfonamides 8-11
which conveniently precipitated from solution ().
Synthetic procedure to prepare pyrazole esters 1-6
As expected, the major by-products, aryl ethyl sulfonates, remained soluble and were easily removed from 8-11
via filtration. Thus, on a 50 mmol scale, gram quantities of the required sulfonamide hydrazides could be readily prepared in high purity. Efficient cyclizations of sulfonamide hydrazides (8-11
) to pyrazolones (12-15
) occurred upon treatment with an aqueous solution of NaOH, followed by acidification (10% HC1) to precipitate the pyrazolones.18
The pyrazolones were used in the subsequent acylation step without purification. For example, treatment of 12-15
with 2-thiophenecarbonyl chloride or 2-furancarbonyl chloride in chloroform resulted in chemoselective19
acylation at the C3-oxygen versus the C5-amino group to furnish pyrazoles 1-6
in moderate to good yields after silica gel chromatography.
The availability of newly synthesized pyrazoles 1-6 allowed confirmation of the structures and observed biological activity of the initial hits. Pleasingly, both the structures and re-assay IC50 values of synthetic pyrazoles 1-6 were in agreement with the results obtained from the original screening experiments ().
The ester carbonyl carbon in 1
, activated by the adjacent pyrazole, is a likely site for nucleophilic attack by the active site cysteine-25 sulfur in cathepsin B. In order to assess the electrophilicity of this carbon atom, equilibrium geometries and electrostatic potential surfaces were calculated for 1
using PC Spartan software.20
The known cysteine protease inhibitor E6421
was also analyzed for comparison to pyrazole ester 1
Compounds for which electrostatic potential surfaces were calculated: pyrazole ester 1 and E64
As expected, the most electrophilic center for pyrazole ester 1 was the carbonyl carbon of the ester, with an electrostatic potential value of 128.0 kJ/mol. The carbons of the epoxide ring in E64 had values of 104.3 and 104.7 for their electrostatic potential surface energies. The greater positive value of 128.0 kJ/mol suggests that the carbonyl carbon of 1 is more prone to nucleophilic attack than the epoxide carbons in the irreversible cathepsin B inhibitor E64. This difference in electrophilicity was confirmed experimentally (vide infra). Presumably, the stability of the pyrazolone by-product of 1 provides the driving force for attack of this atom center.
We reasoned that, modulating the electronic parameters of the electrophilic ester functionality in this series of compounds would quickly validate or eliminate the importance of this structural feature. To test this hypothesis, pyrazole 14 was employed as a common intermediate to construct ether 16, sulfonate 17, and isonicotinoyl ester 18 ().
Reagents and conditions: (a) 2-bromomethyl-thiophene, K2CO3, DMF, 0 °C to 23 °C, 3 h, 50%; (b) thiophene-2-sulfonyl chloride, Et3N, CHCl3, 23 °C, 2 h, 76%; (c) isonicotinoyl chloride, Et3N, CHCl3, 23 °C, 2 h, 43%.
was formed via an alkylation of 14
in moderate yield. Sulfonate 17
and isonicotinoyl ester 18
were constructed under chemoselective acylation conditions similar to those employed to prepare pyrazoles 1-6
In support of the requirement of the ester functionality for biological activity, ether 16
and sulfonate 17
were found to be inactive. Surprisingly, however, related ester analog pyridine 18
was devoid of activity, despite its steric and electronic similarity to the thiophene moiety found in 1
. This unexpected result led us to question the stability of the pyrazoles hits (e.g., hydrolysis or transesterification) under the assay conditions. To probe these issues, we undertook further biological characterization of 1
, as well as a detailed analysis of assay by-products. Kinetic analysis of the inhibition of cathepsin B by 1
revealed a competitive inhibition pattern, in which 1
increased the Km
of the substrate but had no effect on the Vmax
. The Ki
was determined to be 0.9 μM (±) 0.25 μM ().23
Figure 3 Kinetic characterization of inhibition of cathepsin B by 1. cathepsin B was assayed with combinations of 20-350 μM Z-Arg-Arg-AMC substrate and 0.17-1.3 μM inhibitor (1). Initial rates of product release were determined by measuring increase (more ...)
Reversibility of inhibition was next demonstrated by incubating cathepsin B with 1
at a concentration that gave complete inhibition followed by determination of enzyme activity after 100-fold dilution. As expected for reversible inhibition, the enzyme regained full activity upon dilution.24
These experiments ruled out the possibility that 1
acts similar to E64
, an irreversible inhibitor. The possibility that 1
acted as a substrate of cathepsin B, however, remained. That is, under the conditions employed for the kinetic analyses, a competitive inhibitor is indistinguishable from an alternate substrate.23
Thus, experiments were carried out to evaluate the reactivity of 1
under the assay conditions.
As a control, a biochemical time course experiment was designed. Pyrazole 1
was combined with the assay buffer (phosphate, pH 6.8) over a series of time points (0 h, 2 h, 17 h, and 24 h). Samples at each time point were then incubated with cathepsin B (0.1 nM) in the buffer (100 mM) containing dithiothreitol (DTT) (2 mM) under the standard assay conditions, and IC50
data were measured.13
Results from these experiments indicated a significant loss of inhibitory activity which correlated with the time of pre-incubation ().
Time dependent pre-incubation of 1 with and without DTT: An analysis of 1 in the cathepsin B assay
The IC50 values of pyrazole 1 increased from 1.42 μM at a pre-incubation time of 0 h to >50 μM at an incubation time of 24 h. Two additional experiments were conducted to pinpoint the cause of the time dependent loss of activity: (1) pyrazole 1 was added to an assay buffer that did not contain DTT (Column 2, ); and (2) pyrazole 1 was added to a non-buffered solution also without DTT (Column 3, ). Both experiments revealed that the IC50 values changed slightly from the 0 h to 24 h incubation time. Thus, the presence of DTT was deemed largely responsible for the significant loss in activity.
With these results in hand, we set out to define more precisely the role of DTT in the inactivation of 1. Analytical experiments were designed to mimic the biological pre-incubation experiment, wherein the aqueous solutions of 1 with and without DTT would be analyzed via LC-MS for hydrolysis products (e.g., pyrazolone 12). The LC-MS analysis corroborated the previous study and also provided an additional detail, the loss of 1 was the result of its conversion to pyrazolone 12 ().
Time dependent pre-incubation of 1 with and without DTT: An analysis of 1 by LC-MSa
For example, when 1
was incubated in the DTT containing assay buffer, conversion to 12
occurred quickly with 28% of 1
remaining after 1 h, and only 12% after 24 hours.25
Hydrolysis of 1
also occurred, albeit slowly, in both buffer without DTT and non-buffered water. However, on the time-scale of the HTS assay, the contribution of this degradation pathway to the activity was negligible. Additional experiments indicated that when E64
was exposed to DTT (under the same conditions as listed in .) it was found to be stable with no E64
decomposition and no DTT-E64
adduct formed, as monitored by LC-MS.
Anticipating that 12 could arise via nucleophilic displacement of the thiophene ester with DTT, we examined the LC-MS data for the anticipated transesterified DTT-thiophene ester product 19 ().
Synthesis of DTT-thiophene adduct 19.
The molecular ion corresponding to 19 (M+1 = 265) was observed in the LC-MS analysis of the assay buffer with DTT. As further verification, adduct 19 was synthesized independently using DTT and 2-thiophenecarbonyl chloride, and used as a standard in the LC-MS analysis. Bioassay of 19 with cathepsin B revealed no activity. Taken together, these data reveal the basis for the diminished IC50 values following extended pre-incubation times.
The inherent redox chemistry associated with DTT is well known, both in connection with the identification of false positives in cysteine protease assays and with enzymes such as phosphatases.12,26-28
To avoid the redox issues of DTT, cysteine is often used in place of DTT due to its lower oxidation potential. Less well appreciated, however, is the propensity of DTT to act as a nucleophile, thereby confounding bioassay results.29,30
Unfortunately, in our case, replacement of DTT with cysteine does not alleviate the problem. Indeed, the same fast conversion of 1
was observed, via LC-MS, in the presence of cysteine. Thus, the series of pyrazole esters is susceptible to both the hydrolysis and transesterfication pathways. We were, however, intrigued by the apparent lack of any activity of 18
, despite the similar size and electronic attributes of the ester moiety. To explore this observation, we analyzed the rate of hydrolysis and transesterification of 18
. Surprisingly, incubation of 18
in assay buffer, with and without DTT, and a non-buffered aqueous solution, revealed rapid conversion (ca. 1 hr) to pyrazolone 14
, regardless of the presence of DTT ().25
Time dependent pre-incubation of 18 with and without DTT: An analysis of 18 by LC-MSa
The ease of hydrolysis and esterification of 18
is likely a reflection of the fact that the pyridine ring is partially protonated under the conditions.31
Thus, within the timescale of the cathepsin B assay (1 h incubation with the enzyme), a negligible amount 18
would be present.
Taken together, the results support the hypothesis that 1
acts as an alternate cathepsin B substrate rather than an inhibitor. Presumably, the active site cysteine acts much like DTT to transesterify 1
, and thereby form a transient thiophenoyl-enzyme intermediate.7,32,33
To test this scenario, a stoichiometric reaction was devised to incubate pyrazole ester 1
(1 equivalent) with and without cathepsin B (2 equivalents) ().
Standard assay conditions with and without cathepsin B: A stoichiometric reaction analyzed by LC-MSa
Under these stoichiometric reaction conditions, in the presence of cathepsin B, pyrazole 1
was fully converted to 12
after only 15 minutes. However, under identical conditions, in the absence of cathepsin B, 50% of pyrazole 1
We conclude that pyrazole esters such as 1
are competitive substrates for the enzyme cathepsin B.
In summary, we have demonstrated that pyrazole 1 acts as an alternate substrate for the cysteine protease, cathepsin B. Synthesis and evaluation of related analogs revealed the potential reactivity of the ester functionality with the nucleophilic enzyme active site cysteine to form a transient thiphenoyl-enzyme intermediate. Initially, the similar reactivity of DTT and cysteine, in the bioassay, confounded the HTS and subsequent assay results due to the nucleophilic properties of the thiol sulfur. Thus, it is important for the biological and chemical communities to consider the potential of DTT and cysteine to act as nucleophiles in assay systems where substrates contain electrophilic functionality.