The starting point for the peptide mimetics considered herein were structural analogs of small regions of BMP-7 that we identified as likely receptor binding sites. The first criterion is that they have to be accessible for receptor binding. Once the three-dimensional structure of BMP-7 had been determined, the solvent-accessible regions were identified using the “rolling water probe” method of Lee and Richards (Lee and Richards, 1971). It should be noted that inherent in this identification process was the assumption that the crystal structure of BMP-7, as determined by x-ray crystallography (Griffith et al., 1996), is similar to its structure when interacting with its receptor extra-cellular domains. The second criterion is specificity. As noted above, BMP-7 is a member of a protein family that displays significant sequence homology at certain positions, which permits a clear alignment of the member sequences and implies structural homology across the family. The fact that members of the BMP family have different activities suggests that those regions displaying sequence variability across the family could also include receptor binding sites. Therefore, we selected regions of overlap between regions of putative solvent accessibility and variability as initial mimetic targets. These regions included the large loop at the end of the “finger one” anti-parallel beta strand, the tight beta turn at the end of the “finger two” anti-parallel beta strand, and the loops at the C-terminal and N-terminal ends of the “Heel” helix (Figure 1). The peptide analogs of these regions of interest were designed to include disulfide bonds intended to stabilize the conformation found in the native protein.

From this point forward the process for selection and optimization of these mimetics was based not on any assumed structural theory of receptor binding, but via their observed activity using biologically relevant in vitro activity and receptor binding assays. Once one or more lead peptides were identified, the same assays are used to optimize the sequence for desired activities. Binding assays measured the ability of the peptides to compete with BMP-7 in binding to cloned extracellular domains of ALKs 2, 3 and 6 as well as of BMPR-II. For cell-based in vitro assays, the HK2 immortalized human kidney cell line was used to look at the proinflammatory response with and without peptides of interest compared to BMP-7. In vitro assays were also used to look at intracellular signaling and actions such as the ability of candidate peptides, relative to that of BMP-7, to stimulate the phosphorylation of Smad-1,5,8, to block apoptosis, and to control the cell cycle. Of particular interest was the relative ability of the peptides to block the epithelial-to-mesenchymal transition (EMT) which is induced by TGF-beta signaling and is key to fibrosis development.

The peptide optimization process employed a peptide model in which each residue position is represented by a binary vector wherein each bit represents the presence 1) or absence (-1) of a physical-chemical feature such as charge, polarity, size, secondary structure propensity, etc. Based on the results from the above assays each candidate sequence is classified as either active 1) or inactive (-1). Each bit value in each sequence is multiplied by the activity classification value to create a preference value for each bit in each sequence. A statistical method, such as Bayesian analysis (Gelman et al., 1995) was employed at each bit position to create a weight, such as the log of the odds ratio of preference values, reflecting the degree to which the presence or absence of the corresponding feature favors or hinders activity. The result is a structure/activity profile based on physical/chemical features at each residue position in the mimetic, rather than on amino acid residues per se. This profile was then used to score the peptide sequences and to predict other likely active sequences. The optimization process proceeded as cycles of synthesizing new candidates, assaying them, and using the new data to update the structure/activity profile, etc. Once leads had been optimized, they were further evaluated in iv vivo animal models for kidney injury such as the Unilateral Clamp Model for ischemia/reperfusion injury. This is the optimization process used to derive THR-184 from early leads, such as THR-123.

BMPs play important roles in regulating numerous physiological processes including cell proliferation, differentiation, apoptosis, and specification of developmental fate during embryogenesis and in adult tissues (Dudley et al., 1995; Luo et al., 1995). Aberrations in the BMP signaling pathways are associated with the pathophysiology of several diseases, including osteoporosis, arthritis, pulmonary hypertension, idiopathic pulmonary fibrosis, myocardial injury, cardiomyopathies, acute and chronic renal disease, hepatic injury, cerebrovascular diseases, and cancer.

Genetic deletion of BMP-7 in mice leads to severe impairment of kidney development, resulting in perinatal death (Dudley et al., 1995; Luo et al., 1995). Genetic defects in BMPs have also been shown to cause malformations in the heart and lung (Wang L. P. et al., 2014). The kidney is the major site of BMP-7 synthesis during embryogenesis and in the adult (Ozkaynak et al., 1991). Synthesis is confined to distal tubules, collecting ducts and podocytes of glomeruli (Vukicevic et al., 1998). BMP-7 expression decreases in several kidney disease models, including acute ischemic renal injury, tubulointerstitial fibrosis, diabetic nephropathy, and the remnant kidney model (Simon et al., 1999; Wang et al., 2001; Hruska, 2002; Dube et al., 2004). The cellular targets for BMP-7 in the kidney are convoluted tubule epithelium, glomeruli, and collecting ducts (Bosukonda et al., 2000).

In the adult, the primary biological role of BMP-7 is to maintain epithelial/endothelial cell homeostasis. BMP-7 also inhibits the induction of inflammatory cytokine expression (Sutton and Molitoris, 1998; Vukicevic et al., 1998; Gould et al., 2002), attenuates inflammatory cell infiltration, and reduces apoptosis of tubular epithelial cells in renal disease models. Thus, BMP-7 plays critical roles in repairing renal tissues. Damage to these renal tubular tissues results in kidney diseases. Accordingly, BMPs would seem to constitute a family with powerful therapeutic potential for many acute and chronic diseases, were it not for their osteoinductive activity. Since the latter activity appears to be mediated through the ALK6 type I receptor, an ideal therapeutic agent should have no or minimal affinity for the ALK6 type I receptor. The ideal BMP-7 mimetic for treating soft tissue damage necessarily would incorporate the positive properties of BMP-7 without osteoinductive activity. Moreover, these mimetics would be unlikely to interact with the natural binding protein inhibitors of the BMP’s such as follistatin, noggin, gremlin, and chordin, thus avoiding inactivation.

A mimetic peptide of BMP-7 was designed and optimized using the process outlined above. The sequences of the approximately 30 members of the human BMP subfamily of the TGF-beta superfamily have several strictly conserved residues throughout the sequences, which suggests a high degree of structural homology across the subgroup. The x-ray structure of BMP-7, which retains a high degree of three-dimensional structural homology with other superfamily members (Griffith et al., 1996), was used to identify solvent accessible sites. The knuckle region covered by the osteogenic peptides and the beta-turn region covered by the non-osteogenic mimetic peptides described here are separate structural regions (Figure 1C). This separation allows small mimetic peptides to be designed with specific activity.

The receptor binding and in vitro assay activity, indicating half maximal effective concentrations (EC₅₀) of the lead peptide variants, were used for structure/activity analysis to arrive at THR-184 as the prime clinical candidate.

The THR-184 and THR-123 are members of the BMP-7 mimetic family. The family of BMP mimetics are agonists that signal through the activin-like kinase three receptor (ALK3). The BMP-7 mimetics are effective in initiating the phosphorylation of Smads 1, 5 and 8, blocking apoptosis, and have the ability to suppress renal tubule cell inflammation. Also, the mimetic (THR-123) failed to induce ectopic bone formation in a rat implant assay (Figures 2A, B). Preclinical work with THR-184 (Table 2) in multiple animal models showed suppression of inflammation, apoptosis, reversal of the epithelial-to-mesenchymal transition (EMT) program, and attenuation of renal injury, including ischemia/reperfusion injury. Moreover, there was no evidence of bone formation in Pre-clinical Tox experiments. Thus, the mimetic, unlike BMP-7, showed no osteoinduction.

Imaging studies of radio-labeled mimetic injected via the tail vein in rats showed distribution primarily to the kidney cortex (Sugimoto et al., 2012).

A particularly pernicious consequence of ischemic injury is Acute Kidney Injury (AKI). AKI is defined as an abrupt (within hours) decrease in kidney function that encompasses both renal structural damage and loss of kidney function. The term AKI has largely replaced Acute Renal Failure (ARF), reflecting the recognition that smaller decrements in kidney function that do not result in overt organ failure are, nevertheless, of substantial clinical relevance and are associated with increased morbidity and mortality. Recent evidence from both basic medical science and clinical research, is beginning to change our view of AKI from a single-organ failure syndrome to a syndrome where the kidney plays an active role in the progress of multi-organ dysfunctions (Makris and Spanou, 2016).

Injury induced by ischemia can result in damage to both the tubular and the microvascular compartment. Resolution of vasoconstriction appears effective at reducing injury when administered prophylactically, but not following established injury. Resistance may be due to exacerbated inflammation, which may impart reductions in RBF ¹ and GFR insensitive to vasodilator therapies. Of central importance in this process is the activation of inflammatory processes which are influenced by factors released by damaged proximal tubules as well as adhesion of damaged microvascular cells. Infiltrating leukocytes may impinge on RBF either by secreting vasoactive factors, or by contributing to the disruption of flow by physical interference. In addition, exacerbated hypoxia leading to tubular obstruction may contribute to reductions in GFR independent of vasodilator therapy (Basile et al., 2012).

Mortality in patients admitted to the hospital with AKI has been estimated to be 20% in patients not in the ICU and 50% in patients admitted to the ICU. Thus far, AKI has defied attempts to find an effective treatment, let alone a cure (Brown et al., 2010). The pathophysiology of AKI (see Figure 3) involves disruption of arterial and venous flow to the kidney as well as injury to various structures within the kidney such as the glomerulus, the collecting ducts, and tubular interstitium (Basile et al., 2012; Case et al., 2013). The primary targets of injury in AKI are the renal proximal tubule epithelial cells (RPTECs) lining the tubules proximal to the glomerulus. Damaged cells sluff-off and clog the downstream tubules and collecting ducts. This clogging lowers the effective Glomerular Filtration Rate (GFR) and affects other functional aspects of the nephron. The two primary sources of injury to RPTECs are: toxins, such as heavy metals, and ischemia/reperfusion (ischemia and the subsequent immune response to reperfusion).

AKI tends to occur in the sickest patients, often requiring the most complicated treatments. These patients are typically undergoing major surgery and/or in need of critical care (Askenazi et al., 2006; Lo et al., 2009; Brown et al., 2010; Mammen et al., 2012). Current therapies are predominantly supportive. They endeavor to promote urine output, to obtain adequate renal perfusion, to prevent fluid overload, to reduce the buildup of toxic metabolites, and to optimize hemodynamics and promote renal blood flow (Askenazi et al., 2006; Mammen et al., 2012). The intensive investigation of AKI reflects the fact that none of the standard therapeutic strategies employed to treat AKI are able to regulate the multiple pathways involved in AKI pathophysiology. Consequently, to date, there are no established pharmacotherapies for AKI.

The complexities of AKI constitute the kind of pharmacotherapeutic challenge that suggests a more robust approach, such as the use of growth factors, might be more effective. The results of several such trials are mixed. Exogenous IGF-1, which is beneficial in the recovery after kidney injury in mouse models, failed to demonstrate the efficacy in patients with acute renal failure (ARF) (Hirschberg et al., 1999). A double-blind placebo-controlled trial, designed to test whether early treatment with erythropoietin could prevent the development of AKI in ICU patients (Endre et al., 2010), failed to show a difference between the placebo and treatment groups. Currently, a small hepatocyte growth factor/scatter factor mimetic, ANG-3777, that has been shown to improve renal function in patients after kidney transplantation (FDA, 2015) is undergoing a clinical trial in patients who are susceptible to kidney injury (Hulse and Rosner, 2019). This limited success with treatments for AKI strengthens our above-described interest in a mimetic of the BMP pathway.

A successful AKI therapeutic would also likely lead to major improvements in quality of life during aging and to major annual health budget savings. Recent studies have demonstrated a relationship between AKI and the subsequent development of chronic renal insufficiency, a progression to hemodialysis and other severe consequences (Askenazi et al., 2006; Coca et al., 2009; Ishani et al., 2011; Mammen et al., 2012). Even mild renal injury can have long term deleterious effects (James et al., 2020).

Thus, there is a large body of evidence indicating that AKI can lead to chronic kidney disease CKD. AKI is thought to be driven by inflammation and apoptosis. The long-term effects of these processes lead to CKD and are known to involve fibrosis.

The current standards for the diagnosis of AKI rely on the loss of functional markers, such as serum creatinine-based estimates of GFR, but such functional markers do not accurately reflect the degree of actual renal injury. Ultimately, an AKI incident that escapes detection could lead to the development of CKD for which hemodialysis is a maintenance treatment, but not a cure.

A reliable ischemia-reperfusion animal model for testing the protective and therapeutic power of BMP mimetics proved to be the Unilateral Clamp Ischemic Reperfusion rat model developed by Bruce Molitoris (Sutton and Molitoris, 1998). This model was used to determine the clinical-trial dosing of the BMP mimetic THR-184. In this model, animals undergo unilateral nephrectomy and clamping of the renal artery of remaining kidney. A 30-min renal artery occlusion was used to produce ischemia reperfusion and trigger apoptosis and minimize necrosis. The model was used to assess efficacy and to establish dose ranges for different modes of delivery. The two modes of administration tested were IV bolus and IV infusion of up to 1 h, before and after clamping.

In this rat model THR-184 produced a dose-dependent reduction in serum creatinine levels with an ED₅₀ of ∼10–30 µg/kg, with maximal efficacy (65% reduction in serum creatinine levels) measured with a 1 mg/kg dose. That is a ∼100-fold ED₅₀ shift toward greater potency from the bolus to the infusion.

Similarly, a 1 h infusion administered prior to the ischemic insult produced a dose-dependent reduction in serum creatinine levels with an ED₅₀ ∼10 μg/kg. Longer infusion times (4 and 24 h) provided no additional efficacy compared to the 1 h infusion. In a pre-ischemic injury model, the same reduction in sCr at 24 h was obtained via slow infusion using ∼30x less THR-184 than was required via bolus injection (Figure 4A). In the post-ischemic injury mode, the same reduction in sCr at 24 h was obtained via slow infusion using ∼100x less THR-184 than was required via bolus injection (Figure 4B).

One kidney is removed and the blood supply to the other is clamped off for 30–40 min. Figure 4A: The effect of treatment pre-ischemic insult using a slow infusion of THR-184 versus that of a bolus injection. The same reduction in [sCr] at 24 h was obtained via slow infusion using ∼30x less THR-184 than was required via bolus injection. Figure 4B: The effect of treatment post-ischemic insult using a slow infusion THR-184 versus that of a bolus injection. The same reduction in [sCr] at 24 h was obtained via slow infusion using ∼100x less THR-184 than was required via bolus injection. In addition, the maximum degree of reduction is greater using a slow infusion than via a bolus injection.

The 1-h post-ischemic infusion was found to require one percent of the total dose of the IV bolus to produce the same reduction in the rise of serum creatinine. Thus, the doses for the clinical trial were scaled down conservatively and it was decided that an IV infusion of 1 hour would be administered 1 hour prior to initiation of cardiopulmonary bypass and an IV infusion of 1 hour would be administered daily for 3 days post operatively.

Two phase 1 safety studies, each a randomized, adaptive, double-blind, placebo-controlled study at a single center, evaluated safety, tolerability, and pharmacokinetic behavior of ascending doses of THR-184. These were done at Algorithme Pharma of Laval, Quebec, Canada, under FDA Guidelines, and their study numbers for single Ascending Dose Trial and Multiple Ascending Dose Trial were THR-P1-427 and THR-P1-428, respectively.

For the reasons outlined above, AKI is a difficult disease to treat. Clinical trials are confounded by several complexities which are hard to control in a randomized clinical trial. Cardio-vascular surgery has many confounding variables that can affect the occurrence of AKI. These include time on cardiopulmonary bypass, hypotension, anesthetic agents, pressors, inotropes, diuretic use, intravascular volume shifts, and blood loss. Currently, there is no pharmacological intervention that has consistently been associated with renal protection (see Section 2.9.3). This failure is likely related to the following:

• AKI after cardiac surgery is pathophysiologically multifactored and complex, thus simple strategies targeting a single pathophysiological factor and/or single pathway will most likely fail.

Biomarkers such as KIM1 and NGAL, although not yet approved as clinical surrogate endpoints, have the advantage of being early markers of renal injury because their concentration rises within a few hours after injury (Mishra et al., 2005; Han et al., 2008; Khreba et al., 2019). The available evidence from our preclinical studies with THR-184 supports the use of these biomarkers in an AKI clinical trial. In animal models, THR-184 provided nephroprotection from Ischemic Reperfusion Injury and Cisplatin-induced AKI by preventing renal tubular damage, inhibiting tubular apoptosis, and decreasing urinary NGAL (Carlson, 2018). We included KIM-1, NGAL and Albumin as exploratory endpoints in our Clinical trial (NCT01830920). Accordingly, serum and urine samples were obtained pre-, during, and post-treatment with THR184. However, they remain to be assayed. Our understanding is: Sample analysis for these markers is not part of either the primary or secondary outcomes of the clinical trial. However, the availability of data from sample analysis for the KIM-1 and NGAL markers will likely add to our understanding of THR-184 efficacy and the use of these markers in AKI. Importantly, previous studies have found that, as early markers, KIM-1 and NGAL can help to diagnose AKI with high sensitivity and specificity (Bonventre, 2008; Devarajan, 2011; Geng et al., 2021).

AKI is increasingly recognized as a major risk factor for progression to CKD after cardiac surgery (Hsu and Hsu, 2016; Mizuguchi et al., 2018). Multiple signaling pathways including SMAD (Vigolo et al., 2019), p38 MAPK (Terada et al., 2007) and Wnt/β-catenin (Xiao et al., 2016) are implicated as mediators of the progression of AKI into CKD. We have observed that the BMP mimetic alone can regulate these pathways. Importantly, the BMP mimetic has a protective role in the AKI-CKD transition by regulating the WNT/β-catenin signaling pathway. Also, it is capable of inhibiting renal inflammation and blocking/reversing epithelial-mesenchymal transition (EMT), one of the hallmarks of the AKI-CKD transition. In our clinical trial, the effect of THR-184 was more prominent in AKI patients with limited renal capacity due to underlying CKD.

Patient population definition is critical in situations where functional endpoint markers can be insensitive to the targeted clinical condition. Most clinical data come from trials conducted on low-risk patient populations. In the case of this trial, where the functional marker, eGFR, is sensitive to the underlying injury only after there is a significant loss of renal capacity, clinical significance is much easier to achieve when the patient pool is limited to those with pre-existing low renal capacity.

Clinical trial designs are problematic where the injury condition does not affect functional markers in the near term but, if left untreated, does have ultimate long-term implications for function. Trials having long term endpoints require long term follow-up, but the per-protocol patient population decays with time due to the inevitable follow-up protocol violations. This kind of long-term trial will be very large, long, and expensive.

In the report “Perioperative THR-184 and AKI after Cardiac Surgery” on the results of the Phase two clinical trial (Himmelfarb et al., 2018), the trial’s renal advisory board and the chief scientific officer of our company, Thrasos Innovation, state on page 670 that they “found that administration of perioperative THR-184 through a range of dose exposures failed to reduce the incidence, severity, or duration of AKI in patients with high-risk cardiac surgery” (Himmelfarb et al., 2018).

There were two important findings in this publication. The first concerns the safety of the compound, THR-184. THR-184 at all doses tested was well tolerated by the patients and was shown to have no adverse effects in treated patients compared to the Placebo. The second finding concerns the efficacy of THR-184 on AKI after cardiac surgery in high-risk patients. The study pointed out that “a higher postoperative THR-184 dose arm showed lower incidence of post-surgery AKI relative to placebo; the difference, although not statistically different, was more pronounced among patients with preexisting CKD in the per protocol dataset” (Himmelfarb et al., 2018). This is certainly a significant advance since there have been no effective therapeutics to reduce AKI after cardiac surgery.

While the phase two trial did not show statistically significant effects, it did provide data that points to the next steps for clinical development. We understand Phase two Clinical Trials as defined by the Food and Drug Administration:

“In Phase two studies, researchers administer the drug to a group of patients with the disease or condition for which the drug is being developed. Typically involving a few hundred patients, these studies are not large enough to show whether the drug will be beneficial.

“Instead, Phase two studies provide researchers with additional safety data. Researchers use these data to refine research questions, develop research methods, and design new Phase 3 research protocols” (https://www.fda.gov/patients/drug-development-process/step-3-clinical-research).

In the significance statement on p. 671 Himmelfarb et al. note that “The lack of observed clinical benefit (in the trial) is similar to that in other recently completed randomized clinical trials” (Himmelfarb et al., 2018). This, we contend, will be the result for any clinical trial of a therapeutic that addresses renal injury so long as the end point is based solely on functional markers. Failure is a loss of function, which is the result of injury, but there can be injury without detectable loss of function. Therapeutics such as THR-184 treat underlying renal injury. The intent is to prevent loss of function, but as noted above, there is mounting evidence that multiple episodes of even minor AKI ultimately lead to Chronic Kidney Disease (Askenazi et al., 2006; Coca et al., 2009; Ishani et al., 2011; Mammen et al., 2012; James et al., 2020). Unfortunately, the current clinical definition of AKI involves functional parameters such as estimated Glomerular Filtration Rate (eGFR), which are poor surrogate markers of underlying renal injury when there is excess renal capacity. Only in the case of patients entering the trial with reduced renal capacity, where they are closer to the point where loss of function is detectable, will functional markers be reliable in diagnosing actual injury (see Table 3). In this context, we took a closer look at the trail results for the subpopulation of patients entering the trial with preexisting CKD (Table 4 above).

In Table 4, the incidence of AKI, as judged by the sCr KDIGO standard, was 57.0% in the placebo arm and 49.3% in the highest dose arm. This is a reduction of 7.7%. In the Per Protocol subgroup analysis of patients with eGFR <60 the incidence of AKI, as judged by the sCr KDIGO standard, was 75.0% in the placebo arm and 58.1% in the highest dose arm. This is a 16.9% reduction. The incidence of Stage 1 AKI in the Per Protocol population was 47.0% for the placebo arm and 29.0% for the highest dose arm. This is an 18% reduction. In the Per Protocol subgroup analysis of patients with eGFR <60 the incidence of AKI, as judged by the sCr KDIGO standard, was 62.5% in the placebo arm and 35.5% in the highest dose arm. This is a 27.0% reduction. Thus, while not statistically significant due to the small sample size of this subpopulation and the masking of the injury due to the use of eGFR, there is evidence that THR-184 reduced acute renal injury.

Because of the afore-mentioned correlation between incidences of AKI and ultimate development of CKD (Askenazi et al., 2006; Coca et al., 2009; Ishani et al., 2011; Mammen et al., 2012; James et al., 2020), it is important to monitor the level of renal injury during surgical episodes. One way to assess the degree of renal injury independent of functional markers is by monitoring specific renal injury markers such as KIM-1 and NGAL. In fact, the design for the Phase two trial included the collection of samples for assaying the levels of KIM-1, NGAL and Albumin (see Himmelfarb Table 2). Unfortunately, there has been no published report of those data.