Second Generation Triple-Helical Peptide Inhibitors of Matrix Metalloproteinases

Abstract

The design of selective matrix metalloproteinase (MMP) inhibitors that also possess favorable solubility properties has proved to be especially challenging. A prior approach using collagen-model templates combined with transition state analogs produced a first generation of triple-helical peptide inhibitors (THPIs) that were effective in vitro against discrete members of the MMP family. These THPI constructs were also highly water-soluble. The present study sought improvements in the first generation THPIs by enhancing thermal stability and selectivity. A THPI selective for MMP-2 and MMP-9 was redesigned to incorporate non-native amino acids (Flp and mep), resulting in an increase of 18 °C in thermal stability. This THPI was effective in vivo in a mouse model of multiple sclerosis, reducing clinical severity and weight loss. Two other THPIs were developed to be more selective within the collagenolytic members of the MMP family. One of these THPIs was serendipitously more effective against MMP-8 than MT1-MMP and was utilized successfully in a mouse model of sepsis. The THPI targeting MMP-8 minimized lung damage, increased production of the anti-inflammatory cytokine IL-10, and vastly improved mouse survival.

Introduction

Collagenolysis is a well-known physiological process involved in normal tissue and organ development, morphogenesis, and wound healing.¹ Conversely, pathological conditions resulting from abnormal collagen catabolism include primary and metastatic tumor growth, arthritis, arteriosclerosis, and periodontitis.^1–4 The matrix metalloproteinases (MMPs) are the most extensively studied proteases that catalyze the hydrolysis of the collagen triple-helix.⁵

MMP inhibitor programs began in the early 1980s, using extracellular matrix substrates as models for inhibitor design.⁶ The first generation of MMP inhibitors were peptidic whereas the second generation were nonpeptidic small molecules.^2,7 However, clinical trials with both generations of MMP inhibitors were unsuccessful due to, among several reasons, a lack of inhibitor selectivity and metabolic stability, suboptimal dosing of inhibitor, and often considerable side effects.^2,8–11

A well-established strategy to develop protease inhibitors is modification of a bioactive peptide such that the hydrolytically susceptible peptide bond is replaced. The use of phosphinic pseudo-dipeptides can be a very effective approach to develop highly selective and potent inhibitors of a variety of Zn²⁺ metalloproteases.^12,13 Phosphinic dipeptides contain a hydrolytically stable, tetrahedral phosphinic moiety that mimics the tetrahedral intermediate formed during enzymatic hydrolysis. Previously, we reported that phosphinic triple-helical peptides (THPs) behave as effective transition state analog inhibitors of collagenolytic MMPs.^12,14,15 More specifically, two homotrimeric triple-helical peptide inhibitors (THPIs), C₆-(Gly-Pro-Hyp)₄-Gly-Pro-Pro-GlyΨ{PO₂H-CH₂}Val-Val-Gly-Glu-Gln-Gly-Glu-Gln-Gly-Pro-Pro-(Gly-Pro-Hyp)₄-NH₂ [designated C₆-α1(V)GlyΨ{PO₂H-CH₂}Val THPI] and (Gly-Pro-Hyp)₄-Gly-mep-Flp-Gly-Pro-Gln-GlyΨ{PO₂H-CH₂}Leu-Ala-Gly-Gln-Arg-Gly-Ile-Arg-Gly-mep-Flp-(Gly-Pro-Hyp)₄-Tyr-NH₂ [designated α1(I-III)GlyΨ{PO₂H-CH₂}Leu THPI], were shown to inhibit MMPs in the low nanomolar range.^12,14 The THPI sequences were based on either the α1(V)436−450 collagen region, which is hydrolyzed at the Gly-Val bond by MMP-2 and MMP-9, or the interstitial collagen consensus sequence α1(I−III)769−783, which is hydrolyzed by MMP-1, MMP-2, MMP-8, MMP-9, MMP-13, membrane-type 1 MMP (MT1-MMP), and MT2-MMP.^16,17

THPIs represent small, well-folded peptidic structures. There has been a recent surge in the advancement of peptide drugs, as they tend to have lower toxicities and greater efficacies than existing drugs.^18–21 The size of THPIs makes immunogenicity unlikely; in fact, with the exclusion of specific glycosylated sequences, collagen peptides are nonimmunogenic. The triple-helical nature of the inhibitor renders it reasonably stable to proteolysis, as observed in vitro in mouse, rat, and human serum and/or plasma and in vivo in rats.^22–26 The stability of THPs has allowed for their oral administration.²⁷

The first generation THPIs were limited for potential in vivo applications, due to (a) thermal instability or (b) lack of selectivity. The melting temperature (T_m) of C₆-α1(V)GlyΨ-{PO₂H-CH₂}Val THPI was 25 °C, below that desired for a stable triple-helix under physiological conditions.¹² The α1(I−III)GlyΨ{PO₂H-CH₂}Leu THPI was selective toward collagenolytic MMPs, but that group included seven enzymes. The present experiments were designed to overcome these structural and functional shortcomings by using “second” generation THPIs. First, we wished to increase the thermostability of the MMP-2/MMP-9 selective THPI for in vivo use. We utilized the non-native amino acids (2S,4R)-4-fluoroproline (Flp) and (2S,4R)-4-methylproline (mep) to create a more thermally stable α1(V)GlyΨ{PO₂H-CH₂}Val THPI. Moreover, we streamlined the synthetic routes for the synthesis of 9-fluorenylmethoxycarbonyl (Fmoc)-protected Flp (Figure 1, 4) and mep (Figure 1, 5), resulting in a more practical THPI synthesis. Second, we wished to develop THPIs that offered selectivity among the MMP collagenolytic enzymes. We applied two different approaches to obtain selective THPIs for MT1-MMP, utilizing data obtained from phage display generation of MT1-MMP selective substrates and combinatorial chemistry generation of MT1-MMP selective inhibitors. The resulting THPIs were tested in in vivo models of multiple sclerosis (MS) and sepsis.

Figure 1.

Structures of Fmoc-phosphinate dipeptide building blocks (3), Fmoc-Flp (4), and Fmoc-mep (5). The phosphinate dipeptides are analogs of Gly-Val (3a), Gly-Leu (3b), and Gly-Ile (3c).

Results

Improved Thermal Stability of α1(V)Glyψ{PO₂H-CH₂}Val THPI

Our prior study established that a THPI, modeled on the type V collagen sequence Gly-Pro-Pro-Gly₄₃₉-Val₄₄₀-Val-Val-Gly-Glu-Gln, efficiently inhibited MMP-2 and MMP-9.¹² However, that THPI had a T_m value of 25 °C, which was too low for in vivo applications. To produce a more stable analog, our strategy involved incorporating non-native amino acids that favor the preorganization of triple-helical structure.²⁸ Flp offers greater stability than Hyp in the Yaa position of the triple-helix, while mep offers greater stability than Pro in the Xaa position.^28,29 Fmoc-Flp (Figure 1, 4) and Fmoc-mep (Figure 1, 5) were synthesized in large scale. The Fmoc-GlyΨ{PO₂H-CH₂}Val phosphinate dipeptide building block [(R,S)-2-isopropyl-3-((1-(N-(Fmoc)amino)methyl)adamantyl-oxyphosphinyl)propanoic acid], Fmoc-Flp, and Fmoc-mep were then incorporated by solid phase methodology to assemble (Gly-Pro-Hyp)₄-Gly-mep-Flp-Gly-Pro-Pro-GlyΨ-{PO₂H-CH₂}Val-Val-Gly-Glu-Gln-Gly-Glu-Gln-Gly-Pro-Pro-Gly-mep-Flp-(Gly-Pro-Hyp)₄-NH₂ [designated α1(V)GlyΨ-{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI]. Both the Flp and the mep residues were positioned to minimize potential interactions with MMP active sites (see Discussion).

The circular dichroism (CD) spectrum of α1(V)GlyΨ-{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI was indicative of a collagen-like triple-helix with strong negative molar ellipticity ([Θ]) at λ = 212 nm and a positive [Θ] at λ = 225 nm (Figure 2). To examine the THPI thermal stability, [Θ] at λ = 225 nm was monitored as a function of increasing temperature. α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI exhibited a cooperative transition indicative of the melting of a triple-helix to single strands (Figure 2). The T_m was determined to be 43.2 °C (Figure 2), which represents appropriate stability for in vivo studies.

Figure 2.

(Top) CD spectrum of α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI. (Middle) Thermal transition of α1(V)GlyΨ-{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI, measured at λ = 225 nm. (Bottom) First derivative of the thermal transition of α1(V)GlyΨ-{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI, indicating T_m = 43.2 °C.

α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI was examined for inhibition of MMP-2 and MMP-9 hydrolysis of the Knight substrate [Mca-Lys-Pro-Leu-Gly-Leu-Lys-(Dnp)-Ala-Arg-NH₂].³⁰ The THPI exhibited K_i values of 189 and 91 nM for MMP-2 and MMP-9, respectively, at 25 °C (Table 1). It is noteworthy that the inhibitory potency was less compared to the first generation THPI (Table 1). At 25 °C, α1(V)-GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI is ~87% folded (Figure 2, middle panel), while the first generation THPI is ~50% folded at this temperature.¹² The decrease in inhibitory effect may be due to an enhancement in stability of the triple-helical structure, which could slow its destabilization/unwinding by MMP-2 and MMP-9. Alternatively, the decrease in activity could be due to the mep and/or Flp residues interacting unfavorably with the respective enzyme. At 37 °C, the K_i values for MMP-2 and MMP-9 were improved to 2 and 1 nM, respectively (Table 1). At this temperature the triple-helix is more “unwound” than at 25 °C (~64% folded), and the destabilized structure more favorably interacts with MMP-2 and MMP-9 (see Discussion). For MMP-9, the activity is comparable to the first generation THPI at 37 °C (Table 1). For MMP-2, the activity is better compared with the first generation THPI at 37 °C (Table 1). At 37 °C, the first generation THPI is ~17% folded.¹² Thus, inhibitor interaction with MMP-2 and MMP-9 appears to require a balance of folded and unfolded states, although MMP-9 interaction with the triple-helix is less critical for activity than MMP-2.

Table 1. Inhibition of MMP-2 and MMP-9.

enzyme	inhibitor	temperature (°C)	Ki(app) (nM)
MMP-2	C6-α1(V)GlyΨ{PO2H-CH2}Val THPI	10	4.14 ± 0.47a
MMP-2	C6-α1(V)GlyΨ{PO2H-CH2}Val THPI	37	19.23 ± 0.64a
MMP-2	α1(V)GlyΨ{PO2H-CH2}Val [mep14,32,Flp15,33] THPI	25	189.1 ± 26.54
	α1(V)GlyΨ{PO2H-CH2}Val [mep14,32,Flp15,33] THPI	37	2.24 ± 0.11
MMP-9	C6-α1(V)GlyΨ{PO2H-CH2}Val THPI	10	1.76 ± 0.05a
	C6-α1(V)GlyΨ{PO2H-CH2}Val THPI	37	1.29 ± 0.00a
MMP-9	α1(V)GlyΨ{PO2H-CH2}Val [mep14,32,Flp15,33] THPI	25	90.6 ± 6.67
MMP-9	α1(V)GlyΨ{PO2H-CH2}Val [mep14,32,Flp15,33] THPI	37	0.98 ± 0.092

^aAs previously published.¹²

MS Model

In the active stage of MS, CD4+ T-cells activated in the periphery penetrate the blood−brain barrier (BBB), recruit immune cells, initiate destruction of the myelin sheath, and cause axonal loss. Experimental autoimmune encephalomyelitis (EAE) is a well-established murine model of MS. During MS and in the EAE animal model, MMP-2 and MMP-9 are required at key stages of disease progression. They are assumed to act as effector molecules in the (a) generation of neuroantigens, (b) disruption of the BBB, and (c) invasion of inflammatory cells into the brain parenchyma.^31–33

There are various versions of the EAE model that differ based on the applied neuroantigen and/or mouse strain. The most widely used EAE model is the chronic progressive model, in which C57BL/6J mice are immunized with myelin oligodendrocyte glycoprotein (MOG). In this model, days 7−10 approximate the time of T-cell activation and migration across the BBB which manifests as disease onset. Approximately 3−5 days after initial onset, maximum clinical severity is observed prior to a slight reduction and stabilization. In addition to the clinical severity, the concurrent weight loss represents a key indicator of disease progression. Weight loss is typically observed right before disease onset.^34,35

EAE was induced herein with the MOG₃₅₋₅₅ peptide. EAE mice were treated daily from day 7 with an intraperitoneal dosage of 12.43 μg of α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI/animal. With a molecular weight of 12.4 kDa, one 12.43 μg dosage of α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI corresponded to 1 μmol. Mice have ~58.5 mL of blood per kg of body weight. Hence, a mouse weighing 25 g would have a total blood volume (TBV) of ~58.5 mL/kg × 0.025 kg = 1.46 mL. Consequently, the blood concentration of α 1(V)Gly Ψ {PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI would reach (1 × 10⁻⁶ mol)/(1.46 × 10⁻³ L) = 690 nM. This concentration is considerably higher than the α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI K_i for MMP-9 (Table 1).

THPI treatment reduced the clinical severity of EAE from day 12 on (Figure 3A, Figure 3C). The THPI treated mice on average had less weight loss per day and increased their weight before the control group (Figure 3B, Figure 3D).

Figure 3.

Results of 12.43 μg of α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI/day treatment of EAE mice. (A, top left) Clinical score of treated (gray) and nontreated (black) animals. (B, top right) Percent weight change of treated (black) and nontreated (gray) animals. (C, bottom left) Average clinical score of treated and nontreated animals over the course of the experiment. (D, bottom right) Average percent weight change of treated and nontreated animals from induction to sacrifice. n = 5 per group, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Statistical analysis was performed using a two-tailed Student t test.

MT1-MMP Selectivity of THPIs

Of the two prior homotrimeric THPIs, C₆-α1(V)GlyΨ{PO₂H-CH₂}Val THPI did not inhibit MT1-MMP,¹² while α1(I−III)GlyΨ{PO₂H-CH₂}Leu THPI inhibited MT1-MMP with K_i = 120 nM.¹⁵ Unfortunately, α1(I−III)GlyΨ{PO₂H-CH₂}Leu THPI is not selective for MT1-MMP, as it also effectively inhibits MMP-1, MMP-2, and MMP-9.^14,15

Two screening approaches were used to obtain THPI sequences with greater selectivity for MT1-MMP. The first was based on a phage display data utilized to identify MT1-MMP substrates,³⁶ and the second was based on a hydroxamic acid peptide combinatorial library utilized to obtain selective MMP inhibitors.³⁷ Motifs from each of these studies were incorporated into THPs, starting from the general collagenolytic MMP substrate template (Gly-Pro-Hyp)₅-Gly-Pro-Gln-Gly~Leu-Arg-Gly-Gln-Arg-Gly-Val-Arg-(Gly-Pro-Hyp)₅-NH₂. The THP sequences, designations, and sources were as follows:

• (Gly-Pro-Hyp)₅-Gly-Arg-Ile-Gly-His~Leu-Arg-Thr-Gln-Gly-Val-Arg-(Gly-Pro-Hyp)₅-Tyr-NH₂ (designated HLRT, phage display motif bolded);

• (Gly-Pro-Hyp)₅-Gly-Arg-Ile-Gly-Phe~Leu-Arg-Thr-Gln-Gly-Val-Arg-(Gly-Pro-Hyp)₅-Tyr-NH₂ (designated FLRT, phage display motif bolded);

• (Gly-Pro-Hyp)₆-Gly-Pro-Gln-Gly~Val-Gln-His-Gln-Arg-Gly-Val-Arg-(Gly-Pro-Hyp)₅-NH₂ (designated VQH, inhibitor motif bolded);

• (Gly-Pro-Hyp)₆-Gly-Pro-Gln-Gly~Val-Ser-Trp-Gln-Arg-Gly-Val-Arg-(Gly-Pro-Hyp)₅-NH₂ (designated VSW, inhibitor motif bolded);

• (Gly-Pro-Hyp)₆-Gly-Pro-Gln-Gly~Ile-His-Lys-Gln-Arg-Gly-Val-Arg-(Gly-Pro-Hyp)₅-NH₂ (designated IHK; inhibitor motif bolded)

• (Gly-Pro-Hyp)₆-Gly-Pro-Gln-Gly~Ile-Tyr-Phe-Gln-Arg-Gly-Val-Arg-(Gly-Pro-Hyp)₅-NH₂ (designated IYF, inhibitor motif bolded);

• (Gly-Pro-Hyp)₅-Gly-Arg-Ile-Gly-Phe~Leu-Asp-Pro-Gln-Gly-Val-Arg-(Gly-Pro-Hyp)₅-Tyr-NH₂ (designated FLDP, combined phage display and inhibitor motif bolded).

All THPs exhibited triple-helical CD spectra and cooperative melting curves (Figure S1 in Supporting Information). The THP melting temperatures were 27, 32, 42, 43, 39, 41, and 28 °C for HLRT, FLRT, VQH, VSW, IHK, IYF, and FLDP, respectively (Figure S1). The % folded at 25 °C was 72, 79, 87, 92, 83, 88, and 72 for HLRT, FLRT, VQH, VSW, IHK, IYF, and FLDP, respectively. Thus, all THPs were examined for their inhibitory potential of MT1-MMP at 25 °C, where the majority of the THP existed as a folded triple-helix. Only IHK and IYF showed good inhibitory activity at both concentrations tested (50 and 300 μM) (Figure 4).

Figure 4.

Inhibition of MT1-MMP hydrolysis of fTHP-15 by THPs. Codes for the THPs are given in the text. THP concentrations were 50 or 350 μM.

The sequences of IHK and IYF were used to design two homotrimeric THPIs containing a Gly-Ile phosphinic dipeptide. The THPI sequences were (Gly-Pro-Hyp)₄-Gly-mep-Flp-Gly-Pro-Gln-{GlyΨ(PO₂H-CH₂)Ile}-Tyr-Phe-Gln-Arg-Gly-Val-Arg-Gly-mep-Flp-(Gly-Pro-Hyp)₄-Tyr-NH₂ (designated GlyΨ-{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI) and (Gly-Pro-Hyp)₄-Gly-mep-Flp-Gly-Pro-Gln-{GlyΨ(PO₂H-CH₂)Ile}-His-Lys-Gln-Arg-Gly-Val-Arg-Gly-mep-Flp-(Gly-Pro-Hyp)₄-Tyr-NH₂ (designated GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI). Both THPIs exhibited triple-helical CD spectra and cooperative melting curves (Figure S2). GlyΨ{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI and GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI had T_m values of 25.4 and 25.9 °C, respectively. These lower T_m values are not surprising when one considers the interruption of the Gly-Xaa-Yaa repeat in the sequences.

GlyΨ{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI and GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI were initially tested for the inhibitory activity against MT1-MMP and MMP-8, as these two enzymes have similar THP sequence specificity profiles^17,38 which complicates the development of selective MT1-MMP inhibitors. GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI was a poor MT1-MMP inhibitor (Table 2). Surprisingly, GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI effectively inhibited MMP-8 and MMP-1 with K_i values of 120 and 170 nM, respectively (Table 2), despite the design principles that, based on prior studies, would have predicted the opposite outcome.³⁷ GlyΨ{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI was an excellent MT1-MMP inhibitor (Table 2). However, GlyΨ{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI inhibited all collagenolytic MMPs tested (Table 2) In fact, GlyΨ{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI was an outstanding MMP-9 inhibitor and demonstrated almost 1000-fold better inhibition of MMP-9 compared with MMP-2 (Table 2). While the thermal stabilities of GlyΨ{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI and GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI were lower than desired for in vivo experiments, the low levels of MMP-3 inhibition (Table 2) indicated selectivity against noncollagenolytic proteases.

Table 2. Inhibition of MMPs by GlyΨ{PO₂H-CH₂}Ile THPIs.

enzyme	inhibitor	Ki (nM)
MT1-MMP	GlyΨ{PO2H-CH2}Ile-His-Lys-Gln THPI	4704 ± 708.4
MMP-8	GlyΨ{PO2H-CH2}Ile-His-Lys-Gln THPI	124.6 ± 6.9
MMP-1	GlyΨ{PO2H-CH2}Ile-His-Lys-Gln THPI	169.2 ± 28.4
MMP-3	GlyΨ{PO2H-CH2}Ile-His-Lys-Gln THPI	NIa
MT1-MMP	GlyΨ{PO2H-CH2}Ile-Tyr-Phe-Gln THPI	46.15 ± 4.7
MMP-1	GlyΨ{PO2H-CH2}Ile-Tyr-Phe-Gln THPI	110.59 ± 29.8
MMP-2	GlyΨ{PO2H-CH2}Ile-Tyr-Phe-Gln THPI	17.82 ± 1.9
MMP-3	GlyΨ{PO2H-CH2}Ile-Tyr-Phe-Gln THPI	13600.33 ± 5160.7
MMP-8	GlyΨ{PO2H-CH2}Ile-Tyr-Phe-Gln THPI	62.1 ± 2.5
MMP-9	GlyΨ{PO2H-CH2}Ile-Tyr-Phe-Gln THPI	0.03 ± 0.02
MMP-13		77.13 ± 14.4

^aNI = no inhibition at a THPI concentration of 1000 nM.

Sepsis Model

MMP-8 gene expression and activity are increased in human sepsis, and the degree of increase is associated with worsening clinical outcome.^39,40 Genetic ablation or pharmacological inhibition of MMP-8 conferred a survival advantage in an animal model of sepsis.⁴⁰ In turn, ablation of MT1-MMP in an animal model of endotoxemia and decreased MT1-MMP in human sepsis resulted in an increased proinflammatory response, more severe lung injury, and increased mortality.⁴¹ GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI was an effective MMP-8 inhibitor and offered 40-fold selectivity between MMP-8 and MT1-MMP (Table 2). GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI was thus applied to a murine model of sepsis. We used the cecal ligation and puncture (CLP) sepsis model, which entails opening the abdomen, ligating the cecum, and puncturing the cecum with a needle to induce bacterial peritonitis. The severity can be controlled depending on the size of the needle and the number of punctures.

A dosage of 10 mg of GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI per kg was chosen for the sepsis model. With a molecular weight of 12.6 kDa, a delivery of 10 mg of GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI per kg corresponded to 0.79 mmol/kg, assuming an average mouse weight of 25 g results in delivery of 19.75 μmol per mouse. By use of the earlier calculation of TBV of ~1.46 mL, the concentration of GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI would be (19.75 × 10⁻⁶ mol)/(1.46 × 10⁻³ L) = 13.5 μM. This concentration is considerably higher than the GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI K_i for MMP-8 (Table 2).

Following CLP, none of the nontreated wild-type mice survived after 7 days, while 70% of the wild-type mice treated with GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI survived after 7 days (Figure 5). Thus, treatment of wild-type mice with GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI after CLP conferred a significant survival advantage. To examine the specificity of the THPI in vivo, Mmp8 null mice underwent CLP and were then either treated with the THPI or not treated. A survival advantage during sepsis was observed for Mmp8 null mice compared with wild-type mice (Figure 5 gray line versus Figure 6 gray line). Survival of inhibitor treated wild-type mice mirrored that of nontreated Mmp8 null mice (Figure 5 black line versus Figure 6 gray line), while survival of Mmp8 null mice was not augmented by inhibitor treatment (Figure 6 gray versus black line). Thus, the THPI appears to be acting specifically toward MMP-8 in vivo.

Figure 5.

Seven-day survival curves for wild-type mice subjected to CLP. One group of wild-type mice (n = 10) was treated with the GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI (10 mg/kg) immediately after CLP and then every 12 h for an additional 5 doses (black line). The other group of wild-type mice (n = 10) was treated with saline vehicle using the identical schedule (gray line). None of the nontreated wild-type mice survived CLP after 7 days, while 70% of the wild-type mice treated with the THPI survived after 7 days. p < 0.05; Kaplan−Meier survival analysis.

Figure 6.

Seven-day survival curves for Mmp8 null mice subjected to CLP. One group of mice (n = 10) was treated with the GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI (10 mg/kg) immediately after CLP and then every 12 h for an additional 5 doses (black line). The other group of mice (n = 10) was treated with saline vehicle using the identical schedule (gray line). p < 0.918.

We next examined potential modes of action of the THPI. Bacterial colony counts in blood 24 h after CLP were not different between the wild-type treated and nontreated groups (Figure S4), suggesting that the THPI was not acting through an antibacterial mechanism. The lung is an important distal organ of injury in clinical sepsis and is recapitulated in the CLP model.⁴⁰ Neutrophil invasion into the lungs was quantified by determining lung myeloperoxidase (MPO) activity. The improvement in survival was found to be correlated with significantly decreased MPO activity (Figure 7) and hence neutrophil infiltration into the lungs. The inflammatory cytokine profiles in these mice were next studied. There was no effect on interleukin 1β (IL-1β), IL-6, keratinocyte chemoattractant (KC, also known as CXCL1), or LPS-induced CXC chemokine (LIX) levels upon GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI treatment (Figure S3). Increased levels were observed for TNF-α, macrophage inflammatory protein 1 (MIP-1, also known as CCl3 + CCL4), and IL-10 following GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI treatment (Figure 8).

Figure 7.

Wild-type mice lung neutrophil infiltration 24 h after CLP, as measured by MPO activity. One group of wild-type mice (n = 10) was treated with the GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI (10 mg/kg) immediately after CLP and then 12 h later. The other group of wild-type mice (n = 10) was treated with saline vehicle using the identical schedule. THPI treatment significantly reduced neutrophil inflltration into the lungs. *p < 0.05; Student’s t test.

Figure 8.

Wild-type mice serum cytokine profiles 24 h after CLP. One group of mice (n = 10) was treated with the GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI (10 mg/kg) immediately after CLP and then 12 h later. The other group of mice (n = 10) was treated with saline vehicle using the identical schedule. p < 0.05 versus vehicle.

Discussion

THPI Treatment of MS

The design of MMP inhibitors has been significantly improved by the integration of secondary binding sites (exosites). Exosite binding peptides and small molecules have been described as selective MT1-MMP inhibitors.⁴² Selective antibodies have also been developed for MMP-9 and MT1-MMP based on binding to exosites.^43–49 Finally, allosteric, small molecule inhibitors of MMP-13 are well documented.^50–61

The development of collagen helical model THPIs is based on the concept that collagenolytic MMPs will have distinct exosites from noncollagenolytic MMPs and, thus, that this class of inhibitors will exhibit selectivity within the MMP family. Further fine-tuning of the sequence, based on studies with triple-helical peptide substrates,^{17,38,62–64} would enhance selectivity. However, two key problems were associated with the first generation of THPIs. First, the incorporation of the phosphinate significantly decreased the triple-helical stability of the constructs compared with the analogous substrates.^12,14 Second, triple-helical peptide studies did not always provide selective sequences.

Our initial MMP-2/MMP-9 THPI offered excellent selectivity toward those two MMPs but inadequate thermal stability.¹² To overcome this weakness, the present studies incorporated non-native amino acids in an attempt to enhance inhibitor stability. Flp and mep, in the Yaa and Xaa positions of the triple-helix, respectively, favor the preorganization of triple-helical structure.²⁸ Because the phosphinate moiety destabilizes the triple-helix, the Flp and mep residues needed to be incorporated near the phosphinate but not in subsites that might result in detrimental interactions with the enzymes (MMP-2 and MMP-9). Flp was incorporated at the P₅ and P₁₄′ subsites, while mep was incorporated at the P₆ and P₁₃′ subsites. The first generation THPI had Hyp and Pro at these subsites, respectively.¹² Due to the structural similarity of Flp to Hyp and mep to Pro, these substitutions were not anticipated to produce significant unfavorable interactions with MMP-2 and MMP-9. Ultimately, the judicious use of two Flp and two mep residues (per strand) greatly improved THPI thermal stability (ΔT_m = 18 °C). The resulting inhibitor (α1(V)GlyΨ-{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI) was less effective than the parent inhibitor at 25 °C, most likely due to less mobility of the individual peptide strands. Models of MMP-2/MMP-9 collagenolysis suggest that the triple-helix needs to be substantially unwound locally to facilitate hydrolysis of the strands.⁶⁵ However, the activity of the inhibitor was excellent at 37 °C, the appropriate temperature for in vivo studies.

Treatment with the α1(V)GlyΨ{P O₂ H-CH₂ }Val [mep_14,32,Flp_15,33] THPI in the EAE mouse model resulted in reduction of clinical severity and weight loss per day, both of which are hallmarks of EAE. There are numerous indications for the role of MMP-9, and perhaps MMP-2, in MS.^66,67 MMP-9 levels are elevated in MS serum and cerebrospinal fluid, particularly in patients with active disease.^32,68–70 Injection of MMP-9 into the brain causes disruption of the basement membrane and is associated with breakdown of the BBB.^71,72 In MS patients, leukocyte-derived MMP-9 facilitates leukocyte penetration of the BBB.⁶⁷ Concurrent knockout of MMP-2 and MMP-9 results in the inability to induce EAE in mice,⁷³ while loss of MMP-9 alone leaves young (but not older) mice significantly less susceptible to EAE.^31,74 Hydrolysis of dystroglycan by MMP-2 and MMP-9 promotes leukocyte penetration of the parenchymal basement membrane during EAE.⁷³ Inhibitory activities for C₆-α1(V)GlyΨ{PO₂H-CH₂}Val THPI were similar for full-length (82 kDa) MMP-9 or TIMP-1 resistant (65 kDa) MMP-9¹⁴ and coincide with the finding that TIMP-1 resistant MMP-9 appears to be the predominant enzyme form contributing to MS.⁷⁵

THPI Treatment of Sepsis

Our initial THPI that was designed based on interstitial collagen hydrolysis offered no selectivity among the collagenolytic MMPs.^14,15,76 Creating heterotrimeric analogs resulted in THPIs that did not inhibit MMP-1 but were still not selective within the remaining collagenolytic MMPs.¹⁵ Expression of MT1-MMP has been associated with a variety of cellular and developmental processes, as well as multiple physiopathological conditions. MT1-MMP is also known to convert pro-MMP-2 to active protease and to cleave a number of matrix proteins including collagen, fibronectin, and vitronectin. In an attempt to obtain an MT1-MMP selective THPI, we utilized prior studies that described MT1-MMP selective single-stranded peptides and inhibitors. Unfortunately, those selectivities did not translate to triple-helical constructs. A similar observation was made previously.³⁸ Fortuitously, GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI, which was not particularly effective toward MT1-MMP but was a good MMP-8 inhibitor, could be applied in an in vivo sepsis model.

High serum or plasma levels of MMP-8 correlate to sepsis severity and fatality.^39,40 Protection from endotoxin shock following LPS challenge was observed in mice treated with peptidic inhibitors of MMP-8 and MMP-9, and an MMP-8 selective nanobody.^77–79 MMP-8 has been indicated to be a key mediator in the regulation of innate immunity.⁸⁰ The three inflammatory cytokines constituting the “cytokine storm” during septic shock are IL-1β, IL-6, and TNF-α, and recent studies have implicated IL-3.⁸¹ Comparison of CLP in wild type and Mmp8 knockout mice showed no difference in increases of KC, MIP-1α and TNF-α, while LIX did not increase in either model.⁴⁰ The increase in wild-type mice IL-6 was reduced and the increase in wild-type mice IL-1β was completely eliminated in the knockout mice.⁴⁰ Thus, the septic shock increase of IL-6 and IL-1β observed in the wild-type mouse CLP model was positively correlated to MMP-8 concentration.

In contrast to the observations in the Mmp8 knockout mice 6 h after CLP,⁴⁰ the cytokine storm was only partially affected by GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI treatment. Differences observed between THPI treatment and Mmp8 knockout may have several origins. First, Mmp8 deletion removed all MMP-8 activity, while the THPI presumably reduced, but did not eliminate, MMP-8 activity. Second, global Mmp8 deletion may induce a background environment in the mouse that had not been previously appreciated. Third, the THPI may interact with alternative targets, resulting in pleiotropic effects. Dose–response testing of the THPI may yield insights into the origin(s) of the differences.

Prior treatment of CLP mice with a small molecule inhibitor designed for MMP-8 ((3R)-(+) - [2 -(4-methoxy-benzenesulfonyl)-1,2,3,4-tetrahydroisoquinoline-3-hydroxamate], whose true selectivity was not studied),⁸² reduced whole lung MPO activity and improved survival,⁴⁰ consistent with observations here. However, this inhibitor significantly reduced IL-6, IL-1β, MIP-1α, and TNF-α while increasing IL-10.⁴⁰ GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI is potentially more selective than the previously used hydroxamate, which may explain the observed differences in cytokine regulation.

IL-10 is an anti-inflammatory cytokine.⁸³ It was significantly elevated in the THPI treated CLP mice, compared to the vehicle treated CLP mice. This suggested that treatment with the THPI altered the balance of pro-/anti-inflammatory cytokines toward a “less inflamed” response to sepsis and is consistent with the paradigm of overwhelming inflammation being deleterious in sepsis. Whether this is a direct effect of MMP-8 inhibition or some nonspecific effect of the THPI is unknown.

Several additional roles for MMP-8 in sepsis have been suggested. MMP-8 efficiently cleaves IL-13, inactivating it.⁸⁴ MMP-8 cleaves the N-terminus of LIX, CXCL8, and CXCL5, increasing their activity including enhancement of polymorphonuclear neutrophil chemotaxis.⁸⁰ However, we observed no effect on LIX using GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI. MMP-8 directly activates NF-kB, which may lead to regulation of the inflammatory state.⁴⁰ MMP-8 may also be acting indirectly by cleaving and inactivating α1-protease inhibitor, which in turn increases the presence of active neutrophil elastase.⁸⁵ Neutrophil elastase can directly damage lung tissue.⁸⁶

Conclusions

THPIs hold great promise as probes of enzyme activity and imaging agents. THPs have excellent stability in vivo.^22,24,27 Phosphinic peptides designed as MMP inhibitors also exhibit favorable in vivo properties.⁸⁷ The α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI described herein is sufficiently stable to adopt primarily a triple-helical structure in vivo. The triple-helical structure facilitates binding of the inhibitor to the targeted MMPs while minimizing proteolytic degradation. Although MMP-9 selective, inhibitory antibodies and antibody fragments have been described,^{47,48,88–91} their application for neurological disorders is limited based on the inability of antibodies to cross the BBB. The effectiveness of the MMP-2/MMP-9 α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI in the EAE mouse model indicated that this inhibitor may cross the BBB. Subsequent studies will address the efficiency of THPI delivery across the BBB and the integrity of the BBB during EAE.

While our attempt to design a selective MT1-MMP THPI was not successful, we did obtain a THPI that was effective in the CLP animal model of sepsis, most likely through targeting of MMP-8. Although the GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI had low triple-helical thermal stability, it retained selectivity as indicated by its lack of inhibition of MMP-3 and most likely did not undergo general proteolysis in vivo. Gly-Pro-Hyp repeats themselves are stable in human serum²⁶ and thus prevent exopeptidase and possibly inhibit endopeptidase action toward the THPI. The intestine has been found to be the source of MMP-8 in septic peritonitis.⁹² The application of MMP-8 inhibitors such as GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI could be considered in abdominal sepsis associated with intestinal injury⁹² and in adult, rather than juvenile, patients.⁹³

Experimental Section

Chemistry, General

All reactions were carried out under an inert atmosphere and anhydrous conditions with dry solvents unless otherwise stated. Analytical thin layer chromatography was performed on 0.25 mm silica gel 60-F plates. Visualization was accomplished with UV light and aqueous potassium permanganate solution staining followed by air drying. Purification of reaction products was carried out using flash silica gel 40–63 μm. All the standard chemicals and reagents are commercially available and were used without further purification.

All standard peptide synthesis chemicals were peptide synthesis grade. N,N-Dimethylformamide (DMF), trifluoroacetic acid (TFA), dichloromethane (DCM), and ethyl acetate (EtOAc) were purchased from Fisher Scientific (Pittsburgh, PA). N-Methylmorpholine (NMM), 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), and triisopropylsilane were purchased from Acros Organics. 2-(6-Chloro-1H-benzotriazole-1-yl)-1,1,3,3-tetramethylaminium hexafluorophosphate (HCTU) was purchased from Anaspec. All Fmoc-amino acid derivatives and Nova PEG Rink amide resin were purchased from Novabiochem. Amino acids are of l-configuration (except for Gly).

¹H NMR spectra were recorded on a Varian Mercury 400 (400 MHz) spectrometer and are reported in ppm using solvent as an internal standard (CDCl₃ at 7.26 ppm). Data are reported as (b = broad, s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, m = multiplet; coupling constant(s) in Hz, integration). ¹³C NMR were recorded on Varian Mercury 400 (100 MHz) spectrometer. Chemical shifts are reported in ppm, with solvent resonance employed as the internal standard (CDCl₃ at 77.0 ppm). High-resolution mass spectra were obtained from the University of Florida Mass Spectrometry Laboratory.

THPIs were obtained in ≥95% purity as determined by RP-HPLC on a Vydac C₁₈ column (5 μm, 300 Å, 150 mm × 4.6 mm) at a flow rate of 1 mL/min. The elution gradient was 2–98% of B in 20 min (where A is 0.1% TFA in water and B is 0.1% TFA in CH₃CN), and detection was at λ = 220 nm.

The synthesis of Fmoc-Gly-Val phosphinate dipeptide [(R,S)-2-isopropyl-3-((1-(N-(Fmoc)amino)methyl)adamantyloxyphosphinyl)-propanoic acid] (Figure 1, 3a) started from commercially available materials and followed an efficient bis-deprotection strategy as the key step.⁹⁴ Fmoc-Gly-Ile phosphinic dipeptide [(R,S)-2-(2-butyl)-3-((1-(N-(Fmoc)amino)methyl)adamantyloxyphosphinyl)propanoic acid] (Figure 1, 3c) was prepared as previously described.⁹⁵

To efficiently incorporate Flp and mep into a peptide sequence, convenient synthetic routes for the preparation of Fmoc-Flp (Figure 1, 4) and Fmoc-mep (Figure 1, 5) were required. We recently reported a convenient synthetic route for the preparation of Fmoc-Flp from readily available, inexpensive trans-(4R)-hydroxy-l-proline.⁹⁶ The fluorination step was modified herein. Morpholinosulfur trifluoride (3.8 mL, 31.2 mmol) was added dropwise to a solution of N-benzyloxycarbonyl (Cbz)-trans-(4S)-hydroxy-l-proline-O-benzyl (Bzl) (3.7 g, 10.4 mmol) in dry DCM (50 mL) at −78 °C under argon atmosphere. The reaction mixture was stirred for additional 1 h at −78 °C and allowed to warm to room temperature. After 24 h, the reaction mixture was carefully quenched with saturated aqueous NaHCO₃ solution. The organic layer was separated and washed with water (2 × 75 mL) and dried over anhydrous Na₂SO₄. Solvent evaporation under reduced pressure followed by flash chromatography using hexanes/EtOAc (4:1) yielded pure N-Cbz-(4R)-Flp-OBzl as light yellow gummy liquid (1.87 g, 50%). ¹H NMR (500 MHz, CDCl₃): δ 2.0–2.25 (m, 1H), 2.40–2.70 (m, 1H), 3.55–3.80 (m, 1H), 3.85–4.15 (m, 1H), 4.45–4.70 (m, 1H), 4.95–5.40 (m, 4H), 7.20–7.40 (m, 10H). A one-pot three-step procedure involving a bis-deprotection of the N- and C-termini of N-Cbz-4(R)-Flp-OBzl under catalytic hydrogenation conditions followed by selective protection of the N-terminus with an Fmoc group was employed to yield Fmoc-Flp.⁹⁶

To obtain Fmoc-mep, N-Boc-trans-mep was initially synthesized as reported by Del Valle and Goodman.⁹⁷ Fmoc group installation was achieved according to the procedure described by Shoulders et al.⁹⁸

Peptide Synthesis and Purification

The solid-phase synthesis of THPs was performed using Fmoc chemistry on either a Protein Technologies PS3 or CEM Liberty automated microwave peptide synthesizer as described⁹⁹ with Nova PEG Rink amide resin (loading of 0.44 mmol/g except for FLDP, where loading was 0.5 mmol/g). For assembly of FLRT and IHK on the Liberty, coupling was achieved with 5.0 equiv of each amino acid, 4.9 equiv of HCTU (0.5 M in DMF), and 8.0 equiv of NMM (2 M in DMF) at 25 W, 50 °C for 300 s (double coupling for all Arg residues first at 0 W, 25 °C for 25 min followed by 25 W, 50 °C for 300 s and FLRT Pro11 and IHK Pro32 at 25 W, 50 °C for 300 s twice), while deprotection was achieved by treatment with 20% piperidine in 0.1 M HOBt/DMF initially at 35 W, 75 °C for 30 s followed by 35 W, 75 °C for 180 s. For assembly of HLRT and FLDP on the Liberty, coupling was achieved with 5.0 equiv of each amino acid, 4.9 equiv of HCTU (0.5 M in DMF), and 8.0 equiv of NMM (2 M in DMF) at 25 W, 50 °C for 300 s (double coupling for all Arg residues first at 0 W, 25 °C for 25 min followed by 25 W, 50 °C for 300 s and Pro11 at 25 W, 50 °C for 300 s twice), while deprotection was achieved by treatment with 10% piperazine in DMF initially at 35 W, 75 °C for 30 s followed by 35 W, 75 °C for 180 s. For assembly of VQH, VSW, and IYF on the PS3, coupling was achieved with 4.0 equiv of each amino acid, 3.9 equiv of HCTU, and 8.0 equiv of NMM (0.4 M in DMF) for 30 min (double coupling for VQH Pro8, VSW and IYF Pro14), while deprotection was achieved by treatment with 1% DBU, 5% piperidine in DMF 3 times for 6 min each.

Side chain deprotection and cleavage of all THPs from the resin were achieved by two treatments with H₂O–thioanisole–EDT–TFA (5:5:2.5:87.5) for 3 h each in Ar. THPs were purified by RP-HPLC (see below) and characterized by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry as described⁹⁹ using an Applied Biosystems Voyager DE-PRO. [M + H]⁺ values for THPs were 4183.4 Da for HLRT (theoretical 4184.64 Da), 4195.6 for FLRT (theoretical 4194.67 Da), 4261.1 Da for VQH (theoretical 4257.63 Da), 4267.0 Da for VSW (theoretical 4265.65 Da), 4272.8 Da for IHK (theoretical 4271.72), 4318.9 Da for IYF (theoretical 4316.75 Da), and 4150.8 Da for FLDP (theoretical 4149.59 Da).

The solid-phase synthesis of GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI and GlyΨ{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI was performed on a CEM Liberty automated microwave peptide synthesizer using Nova PEG Rink amide resin (loading 0.5 mmol/g) utilizing Fmoc chemistry. The coupling of usual Fmoc-amino acids was achieved with 5 equiv of each amino acid, 4.9 equiv of HCTU (0.5 M in DMF), and 8.0 equiv NMM (2 M in DMF) using microwave power of 25 W at 50 °C with the reaction time of 300 s (double coupling for all Arg residues first at 0 W, 25 °C for 25 min followed by 25 W, 50 °C for 300 s). The removal of Fmoc-protecting group was accomplished using 10% piperazine in DMF (microwave power of 35W at 75 °C and the reaction time of 30 s, followed by a second treatment under the same conditions for 180 s). The unusual amino acids, Fmoc-Flp-OH and Fmoc-mep-OH, were coupled using 3 equiv of amino acid, 2.9 equiv of HCTU, and 6 equiv of NMM with microwave power of 25 W at 50 °C for 300 s. The coupling of Fmoc-mep-OH was repeated. The Fmoc-protected phosphinic dipeptides were attached using 3 equiv of dipeptide, 3 equiv of DIPCDI, 6 equiv of HOBt, and DIPEA to adjust pH to ~8 at 40 °C for 40 min and room temperature overnight. Following dipeptide addition the peptide-resin was capped using 0.5 M Ac₂O, 0.125 M DIPEA, and 0.015 M HOBt in DMF at room temperature for 1 h. The deprotection of Fmoc group, immediately after incorporation of the phosphinic dipeptide into the sequence, was performed without microwave power but with increased reaction times of 500 and 800 s. The amino acid immediately after the dipeptide was attached using the same conditions as for the dipeptide; Pro11 was double coupled.

The solid-phase synthesis of α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI was performed on a CEM Liberty Blue automated microwave peptide synthesizer using Nova PEG Rink amide resin (loading 0.48 mmol/g) utilizing Fmoc chemistry. The coupling of usual Fmoc-amino acids was achieved with 5 equiv of each amino acid (0.2 M), 5 equiv of DIPCDI (0.5 M in DMF), and 5 equiv of oxyma (1 M in DMF). Fmoc-amino acid coupling cycles were carried out for 15 s at 75 °C and 110 s at 90 °C. Fmoc deprotection was achieved with 10% piperazine (w/v) in ethanol/N-methyl-2-pyrrolidone (10:90) for 15 s at 75 °C and 50 s at 90 °C. Fmoc-Flp-OH was incorporated using 2.5 equiv of amino acid for 600 s at 25 °C and 300 s at 75 °C. Fmoc-mep-OH and Fmoc-GlyΨ{PO₂H-CH₂}Val were incorporated using 2.5 equiv of each compound for 1500 s at 25 °C and 600 s at 75 °C. Coupling of Fmoc-Pro-OH following GlyΨ{PO₂H-CH₂}Val was performed for 1500 s at 25 °C and 600 s at 75 °C for using 5 equiv of amino acid. For all THPIs, the final cleavage of peptide from the resin and side chain deprotection were carried out using 7 mL of triisopropylsilane–phenol–H₂O–TFA (2:5:5:88) under inert atmosphere (Ar) for 3 h.

THPs and phosphinic peptides were purified using RP-HPLC on a Vydac C₁₈ column (15–20 μm, 300 Å, 250 mm × 22 mm) at a flow rate of 10 mL/min. The elution gradients were 7–30% of B in 40 min [for GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI], 10–35% of B in 50 min [for GlyΨ{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI], and 20–50% of B in 30 min [for α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI] (where A is 0.1% TFA in water and B is 0.1% TFA in CH₃CN) and detection was at λ = 220 nm. Homogenous fractions were combined and lyophilized. Purified THPIs were characterized by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry as described^12,15 using an Applied Biosystems Voyager DE-PRO or Bruker microflex workstation. [M + H]⁺ values for THPIs were 4145.3 Da for α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI (theoretical 4143.36 Da), 4235.5 Da for GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI (theoretical 4234.64 Da), and 4280.5 Da for GlyΨ{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPI (theoretical 4279.00 Da).

CD Spectroscopy

CD spectra were recorded over the range λ = 190–250 nm on a JASCO J-810 spectropolarimeter using a 10 mm path length quartz cell. The thermal transition curves were obtained by recording the molar ellipticity ([Θ]) at 225 nm, while the temperature was continuously increased in the range of 5–85 °C at a rate of 15 °C/h. The temperature was controlled using a JASCO PTC-348WI temperature control unit. For samples exhibiting sigmoidal melting curves, the inflection point in the transition region (first derivative) is defined as the T_m. Alternatively, T_m was evaluated from the midpoint of the transition.

MMPs

Human recombinant, full-length proMMP-1, proMMP-3, proMMP-8, proMMP-13, and proMT1-MMP were purchased from R&D Systems (Minneapolis, MN) and activated prior to use. Trypsin-3 and 4-(2-aminoethyl)benzensulfonyl fluoride hydrochloride (AEBSF) were also purchased from R&D Systems. p-Aminophenylmercuric acetate was purchased from EMD Biosciences (San Diego, CA), and trypsin (TPCK-treated) was purchased from Worthington Biochemical Corporation (Lakewood, NJ). Recombinant MMP-2 and MMP-9 catalytic domains were purchased from R&D Systems and used directly without further activation in the desired concentration.

ProMMP-1, proMMP-3, proMMP-8, and proMMP-13 were activated by mixing an equal volume of stock enzyme solution and activator (p-aminomercuric acetate to a final concentration of 1 mM) followed by a 2–3 h incubation in a 37 °C water bath. ProMMP-2 was activated by incubating with 2 mM p-aminomercuric acetate for 1 h at 37 °C. ProMT1-MMP was activated by using trypsin-3 at a final concentration of 0.1 μg/mL and incubating for 1 h at 37 °C. The reaction was stopped by addition of AEBSF (at a final concentration of 1 mM) and incubation for 15 min at room temperature. Activated MMPs were diluted to 20–100 nM in ice cold TSB*Zn (50 mM Tris, 100 mM NaCl, 10 mM CaCl₂, 0.05% Brij-35, 0.02% NaN₃, 1 μM ZnCl₂, pH 7.5) to prevent autoproteolysis. Enzyme aliquots were kept on wet ice and used the same day. MMP activity was initially evaluated by using the Knight substrate and compared with prior data.^14,76 In this way, activity toward the substrate was used as an indicator of enzyme integrity, rather than TIMP titration, as performed previously.¹⁷

Inhibition Kinetic Studies

Peptide substrate and inhibitor solutions were prepared using TSB*Zn. For testing THPs as inhibitors, 1–2 nM enzyme was incubated with varying concentrations of THPs for 2 h at room temperature. A 2 h incubation was utilized based on our prior studies using peptidic inhibitors of MMPs.¹⁰⁰ Residual enzyme activity was monitored by adding fTHP-15 solution in TSB*Zn to produce a final concentration of <0.1K_M. Initial velocity rates were determined from the first 20 min of hydrolysis when product release is linear with time. Fluorescence was measured on a Bio-Tek Synergy H1 reader or H4 hybrid reader using λ_excitation = 324 nm and λ_emission = 393 nm. Apparent K_i values were calculated using SigmaPlot by fitting data to the equation v = v₀/(1 + I/K_i), where v₀ is the activity in the absence of inhibitor and K_i is the apparent inhibition constant. Because the substrate concentration is less than K_M/10, apparent K_i values are insignificantly different from true K_i values.

For THPIs, 1−2.5 nM enzyme was incubated with varying concentration of inhibitor for 2 h at room temperature or 37 °C. A 2 h incubation was utilized based on the generally observed behavior of slow on and off rates for tight-binding inhibitors¹⁰¹ and studies demonstrating that high affinity phosphinate inhibitors of Zn²⁺ metalloproteinases are slow binding.¹⁰² Residual enzyme activity was monitored by adding Knight substrate solution in TSB*Zn to produce a final concentration of <0.1K_M. Initial velocity rates were determined from the first 10 min of hydrolysis when product release is linear with time. Fluorescence was measured on a Bio-Tek Synergy H1 reader or H4 hybrid reader using λ_excitation = 324 nm and λ_emission = 393 nm. Apparent K_i values were calculated from the following formulas:¹²

$$\frac{v_{\text{i}}}{v_0}=\frac{E_{\text{t}}-I_{\text{t}}-K_{\text{i}}^{\text{(app)}}+{({(E_{\text{t}}-I_{\text{t}}-K_{\text{i}}^{\text{(app)}})}^2+4E_{\text{t}}{K}_{\text{i}}^{\text{(app)}})}^{0.5}}{2E_{\text{t}}}$$ (1)

$$K_{\text{i}}^{\text{(app)}}=K_{\text{i}}\left(\frac{A_{\text{t}}+K_{\text{M}}}{K_{\text{M}}}\right)$$ (2)

where I_t is the total inhibitor concentration, E_t is the total enzyme concentration, A_t is the total substrate concentration, v₀ is the activity in the absence of inhibitor, and K_M is the Michaelis constant. In our assays the value of E_t/K_i^(app) does not exceed 100 so that the inhibitor is distributed in both free and bound forms, and K_i^(app) can be calculated by fitting inhibition data to eq 1. Because the substrate concentration is less than K_M/10, K_i^(app) values are insignificantly different from true K_i values. In cases where weak inhibition occurred, K_i^(app) values were calculated using v_i = v₀/(1 + I_t/K_i^(app)).

Murine Model of MS

All procedures involving animals were performed in accordance to the guidelines of the Institutional Animal Care and Use Committee. Complete Freund’s adjuvant (CFA) was produced by emulsifying incomplete Freund’s adjuvant (Thermo Scientific) and M. tuberculosis H37Ra (Becton Dickinson, Franklin Lakes, NJ). MOG_35–55 peptide (Met-Glu-Val-Gly-Trp-Tyr-Arg-Ser-Pro-Phe-Ser-Arg-Val-Val-His-Leu-Tyr-Arg-Asn-Gly-Lys-OH) was synthesized using standard Fmoc solid-phase chemistry¹⁰³ and RP-HPLC purified to >90%. EAE was induced in female C57BL/6J mice (Jackson Laboratory; 10–12 weeks of age) by subcutaneously inoculating an emulsification of CFA and the MOG_35–55 peptide (200 μg/mouse) on day 0. Pertussis toxin (Hooke’s Laboratory, 320 ng) was introduced via intraperitoneal injection 3 and 24 h later. The control group received phosphate buffered saline (PBS)/animal/day, while the treated group received 12.43 μg of α1(V)GlyΨ{PO₂H-CH₂}Val [mep_14,32,Flp_15,33] THPI/animal per day by intraperitoneal injection. Disease severity (scored based on the Hooke’s Laboratory scoring guide (www.hookelabs.com/protocols/eaeAI_C57BL6.html)) and weight change were recorded from day 5 until sacrifice.

Murine Model of Sepsis

All experimental procedures involving mice complied with the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (NIH Publication No. 85-23, revised 1996) and met approval of the Cincinnati Children’s Research Foundation Institutional Animal Care and Use Committee. Mmp8–/– mice on a C57BL/6J background were provided by Dr. Steven Shapiro, University of Pittsburgh. Loss of the Mmp8 gene was confirmed by PCR using Mmp8 specific primers (data not shown). Wild type C57BL/6J mice were obtained from Harlan Laboratories (Indianapolis, IN). All mice were fed standard rodent chow and maintained on 12 h light–dark cycles.

Male mice aged 5–9 weeks underwent CLP as previously described.⁴⁰ Briefly, mice were anesthetized and a midline laparotomy was performed. The cecum was isolated and ligated to 30% original diameter using 3–0 silk suture (Ethicon, Cincinnati, OH). Two through and through cecal perforations were made distal to the ligation using a 21-gauge needle, and a small amount of fecal content was expressed from each site. The cecum was then returned to the abdominal cavity, and the peritoneum was closed using 7–0 silk suture (Ethicon), followed by skin closure with surgical glue (Abbott Laboratories, Abbott Park, IL). Mice treated with vehicle received an intraperitoneal injection of PBS at a dose of 10 mL/kg after abdominal closure. The GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln THPI was dissolved in PBS to a final inhibitor concentration of 10 mg/mL. Animals received a 10 mg/kg dose of inhibitor immediately following abdominal closure. All mice received 0.6 mL of normal saline subcutaneously in the nape of the neck at the conclusion of the operation. Animals were redosed with the THPI or PBS vehicle at 12 h intervals, for up to 5 additional doses. Following CLP, animals were monitored for survival up to 7 days or were sacrificed at 24 h to procure lung and blood samples.

Measurement of Myeloperoxidase Activity

MPO was measured as an indication of neutrophil infiltration in lung tissue, as previously described.¹⁰⁴ Briefly, whole lung tissue was homogenized and myeloperoxidase activity was assessed using spectrophotometry and defined as the quantity of enzyme degrading 1 μmol of hydrogen peroxide/min at 37 °C, expressed in units per 10 mg of tissue.

Measurement of Plasma Cytokines and Chemokines

Serum concentrations of IL-10, IL-6, KC, IL-1β, MIP-1α, TNFα, and LIX were analyzed using a Luminex multiplex system (Luminex Corporation, Austin, TX) according to instructions from the manufacturer.

In Vivo Bacterial Load

Twenty-four hours after CLP mice were sacrificed with carbon dioxide asphyxiation, serum samples were obtained via cardiac puncture and subjected to serial log-fold dilutions using sterile normal saline.¹⁰⁵ The dilutions from all samples were then plated onto blood agar plates (Becton Dickinson) and incubated for 24 h in a bacterial culture incubator. After the 24 h incubation period, bacterial culture plates were manually counted by a technician blinded to the experimental conditions.

Supplementary Material

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jmedchem.7b00018.

• CD melting curves for THPs, CD spectra and melting curves for GlyΨ{PO₂H-CH₂}Ile-His-Lys-Gln and GlyΨ-{PO₂H-CH₂}Ile-Tyr-Phe-Gln THPIs, serum profiles of IL-1β, IL-6, KC and LIX 24 h after CLP, and bacterial colony counts 24 h after CLP (PDF)

Abbreviations Used

MMP

matrix metalloproteinase

THP

triple-helical peptide

THPI

triple-helical peptide inhibitor

MT1-MMP

membrane type 1 matrix metalloproteinase

Flp

(2S,4R)-4-fluoroproline

mep

(2S,4R)-4-methylproline

Fmoc

9-fluorenylmethoxy-carbonyl

MS

multiple sclerosis

BBB

blood–brain barrier

CNS

central nervous system

MOG

myelin oligodendrocyte glycoprotein

EAE

experimental autoimmune encephalomyelitis

TBV

total blood volume

CD

circular dichroism

CLP

cecal ligation and puncture

KC

keratinocyte chemoattractant

LIX

LPS-induced CXC chemokine

MPO

myeloperoxidase

IL

interleukin

MIP

macrophage inflammatory protein

TNF

tumor necrosis factor

TIMP

tissue inhibitor of metal-loproteinase

DMF

N,N-dimethylformamide

TFA

trifluoroacetic acid

DCM

dichloromethane

EtOAc

ethyl acetate

NMM

N-methylmorpholine

DBU

1,8-diazabicyclo[5.4.0]-undec-7-ene

HCTU

2-(6-chloro-1H-benzotriazole-1-yl)-1,1,3,3-tetramethylaminium hexafluorophosphate

Cbz

benzyloxycarbonyl

Bzl

benzyl

EDT

1,2-ethanedithiol

DIPCDI

N,N′-diisopropylcarbodiimide

HOBt

1-hydroxybenzotriazole

DIPEA

N,N-diisopropylethylamine

AEBSF

4-(2-aminoethyl)-benzensulfonyl fluoride hydrochloride

PBS

phosphate buffered saline

CFA

complete Freund’s adjuvant