Enhancing the Therapeutic Window for Spectinamide Anti-tuberculosis Agents: Synthesis, Evaluation, and Activation of Phosphate Prodrug 3408

Abstract

Spectinamides are a novel class of narrow-spectrum antitubercular agents with the potential to treat drug-resistant tuberculosis infections. Spectinamide 1810 has shown a good safety record following subcutaneous injection in mice or infusion in rats but exhibits transient acute toxicity following bolus administration in either species. To improve the therapeutic index of 1810, an injectable prodrug strategy was explored. The injectable phosphate prodrug 3408 has a superior maximum tolerated dose compared to 1810 or Gentamicin. Following intravenous administration in rodents, prodrug 3408 was quickly converted to 1810. The resulting 1810 exposure and pharmacokinetic profile after 3408 administration was identical to equivalent molar amounts of 1810 given directly by intravenous administration. 3408 and the parent 1810 exhibited similar overall efficacy in a BALB/c acute tuberculosis efficacy model. Delivery of 1810 in phosphate prodrug form, therefore, holds the potential to improve further the therapeutic index of an already promising tuberculosis antibiotic.

Graphical Abstract

Despite significant progress in the fight against tuberculosis (TB) in the past decades, TB remains a leading cause of death from a single infectious disease, with multidrug-resistant TB (MDR-TB) and extensively drug-resistant (XDR) TB creating major challenges for treatment strategies.¹ Recent advancements include a 4-month regimen with rifapentine and moxifloxacin, non-inferior to the standard 6-month treatment for drug-susceptible TB, and a 6-month regimen with bedaquiline, pretomanid, and linezolid for MDR and XDR TB, offering improved efficacy.² However, Linezolid in shorter MDR/XDR TB regimens is associated with significant toxicity, such as peripheral neuropathy and blood count reductions, limiting its use.³ The ongoing need for shorter, better-tolerated treatments across all patient populations underscores the necessity for new antibiotics and innovative drug combinations.

The spectinamides are narrow-spectrum antibiotics that target protein synthesis in M. tuberculosis at a unique ribosomal binding site. They are derived from the natural product spectinomycin (Figure 1).⁴ The spectinamides have emerged recently as a new class of injectable agents under development to treat MDR/XDR TB owing to a high safety margin, lack of cross-resistance to other anti-tubercular agents, and good efficacy in multiple preclinical murine TB infection models when given alone and when used in combination with other TB therapeutics.^{4, 5} Key to their potent antitubercular properties are structural modifications to avoid the Rv1258c efflux pump in Mycobacterium tuberculosis.⁶ Lead molecule 1810 (MBX4888A) is currently in the late preclinical development phase for delivery by inhalation and intravenous administration.^7–9

Figure 1.

Structures of Spectinamides and the designed phosphate prodrug 3408

Although generally very safe, the spectinamides, as a class, have a safety liability issue when administered as a bolus injection due to their high free fraction leading to very high exposures,¹⁰ which limits the delivery of most spectinamides to 200 mg/kg by the intravenous (i.v.) route (Table 1). In this work, we describe the use of an injectable phosphate prodrug strategy to further improve the therapeutic safety window of lead molecule 1810 when given by injection, while maintaining in vivo efficacy in a murine TB acute efficacy model. This approach offers the potential for the development of a second injectable form of 1810, providing a backup to the inhaled therapy approach.

Table 1.

MTD values for spectinamides in rats compared to Gentamycin by i.v. infusion time

Administration	MTD (mg/kg)
	1599	1810	3408	Gentamicin
5 second i.v. bolus	75	200	>500	100
30 second i.v. bolus	100	200	-	100
5 minute i.v. infusion	100	350	-	200
15 minute i.v. infusion	150	>500	-	300
30 minute i.v. infusion	300	-	-	400

1599 (Figure 1) was the first preclinical candidate to emerge from the spectinamide series and has proven efficacious in multiple murine infection models when given alone or when used in combination with other TB therapeutics.⁵ In vitro safety profiling of 1599 failed to detect any relevant off-target pharmacologic effects.¹¹ Although 1599 exhibits excellent in vivo efficacy and a good in vitro safety profile, its development was ultimately abandoned due to i.v. bolus safety issues in rats (Table 1). The exact basis for the acute toxicities associated with 1599 following i.v. bolus administration, which produces transient high maximum plasma concentration (C_max) values, has thus far proven elusive.

Spectinamide 1810 is a closely related chemical analog of 1599, with a polar phenolic group replacing the chlorine atom present in 1599.⁶ However, unlike 1599, 1810 shows a higher MTD and thus a larger in vivo therapeutic safety margin when administered by i.v. bolus or timed infusion to rats (Table 1). Indeed, 1810 is better tolerated in these assays compared to a conventional control antibiotic Gentamicin, which is associated with known toxicities in human patients.¹² 1810 is currently in preclinical development for delivery by inhalation as a candidate therapy for MDR/XDR TB.¹³

The phosphate prodrug approach has emerged as a viable option for increasing the bioavailability of many drug candidates.¹⁴ In an effort to further expand the therapeutic safety window of 1810 when administered via injection, we explored a slow-release injectable phosphate prodrug strategy. To evaluate this approach for the spectinamides, a phosphate prodrug moiety was designed by attaching phosphate to the phenol group of 1810, resulting in prodrug 3408 (Figure 1).

To generate 3408, we followed the route shown in Scheme 1; spectinomycin was initially protected by Cbz groups (1), followed by reductive amination, by which the 3’ ketone group was stereospecifically converted to an amine 2.⁶ Subsequently, 2 was coupled with 2-(5-hydroxypyridin-2-yl) acetic acid to form bi-Cbz 1810 (3). The regioselective phenolic phosphorylation step was carried out using dibenzyl phosphorochloridate, resulting in the formation of bi-Cbz and bi-benzyl protected 3408 (4). The final step involved hydrogenation, utilizing 10% Pd-C as the catalyst and 10% acetic acid/methanol as solvent, yielding the final product 3408 as a white solid.

Scheme 1.

Synthesis of phosphate prodrug 3408

Reagents and conditions: (1) BnOCOCl, NaHCO₃, acetone/water, room temperature, 10 h, 90%; (2) NH₄NO₃, 2-Methylpyridine borane, 10% acetic acid in methanol, room temperature, 2 h, 40%; (3) HBTU, DIPEA, DMF, room temperature, 40%; (4) DMAP, DIPEA and dibenzyl phosphorochloridate; (5) Acetic acid, MeOH, H₂ and 10%Pd/C.

3408 was then profiled against spectinamides 1810 and 1599, and comparator control aminoglycoside Gentamicin in a maximum tolerated dose i.v. infusion study (Table 1). 3408 showed a marked improvement over 1810 and Gentamicin in the rat 5 second i.v. bolus study indicating an excellent safety margin in vivo. The in vitro activity of 3408 was next compared to spectinomycin and the parent 1810. Unlike the parent molecule 1810, 3408 was inactive in the ranges tested herein in in vitro whole cell MIC assays (Table S1). 3408 was also found to exhibit only weak protein synthesis inhibition with an IC₅₀ of 6.7 μg/mL compared to 0.84 and 0.36 μg/mL for 1810 and spectinomycin, respectively (Table S1). These results show that the appended phosphate moiety negatively impacts antitubercular activity and that prodrug cleavage is required for biological activity.

Prodrug conversion of 3408 to the active drug 1810 was subsequently confirmed in two separate studies. First, 3408 was converted to 1810 with a half-life of 19 min following incubation in fresh rat plasma in vitro (Figure S1). This is unlikely to be due to spontaneous liberation of the phosphate prodrug moiety since 3408 was stable for at least 6 hours at room temperature when incubated in the vehicle alone (Table S2). Second, 1810 liberated from 3408 was readily detected in plasma following i.v. administration of the prodrug 3408 to rats (Table 2). The liberation of parent 1810 from 3408 occurred at levels nearly identical to equivalent molar amounts of 1810 given separately by i.v. administration (Table 2). Taken together, these data indicate that 3408 is stable in vitro, but undergoes rapid and specific hydrolysis in vivo or in fresh rat plasma in vitro to release the active parent molecule 1810.

Table 2.

Summarized plasma pharmacokinetic parameters of Lee 1810 after i.v. administration of 1810 and 3408 (10 mg/kg i.v. bolus) in rats. AUC area under the plasma concentration-time curve, CL clearance, V_SS volume of distribution at steady state, t_1/2 terminal half-life

Compound	AUC (h·mg/L)	Cmax (mg/L)	CL (L/h/kg)	Vss (L/kg)	t1/2 (h)
1810 from 3408	18.5 ± 2.0	38.9 ± 4.8	0.546 ± 0.057	0.322 ± 0.067	0.390 ± 0.039
1810	22.0 ± 3.0	47.0 ± 6.1	0.463 ± 0.074	0.494 ± 0.308	0.392 ± 0.016

1810 and 3408 were next compared for in vivo efficacy in the BALB/c TB acute efficacy model.^{15, 16} Testing in this model employs 12 consecutive days of dosing starting 7 days post low-dose aerosol infection. One day following aerosol exposure with M. tuberculosis Erdman pFCA-LuxAB,¹⁷ mice were found to have an average 2.6 log10 bacilli in lungs. By day 7 (start of treatment), there was an average of 3.7 log10 bacilli in lungs. Treatment of mice with either 1810 or 3408 at 400 mg/kg once daily (QD) promoted significant reductions in lung burdens relative to the untreated control (P < 0.001). 1810 reduced lung burdens by 1.13 logs relative to untreated (Table S3). This is consistent with previous studies in a second acute model using the gamma-interferon knock out mouse model.⁶ 3408 also reduced lung burdens by 1.11 logs relative to untreated (Table S3). There was no statistical difference between these two treatment groups, indicating similar overall efficacy for the parent 1810 or the phosphate prodrug 3408 (Figure 2, Table S3). The rifampin control administered orally at 10 mg/kg QD reduced lung burdens by 1.21 log10 CFU, which is consistent with previous studies in this model¹⁶ (Figure 2, Table S3).

Figure 2.

Efficacy of 3408, 1810, and rifampin in a BALB/c TB acute efficacy model. Each group consisted of six mice.

Plasma PK studies in these infected mice (Table S4) revealed high levels of circulating 1810 for both, the 1810 and 3408 treatment groups at 15 min post dose, and detectable but low levels of 3408 in the 3408 group. At 8 hours post dose, low but detectable concentrations of only 1810 were measured in both treatment groups. In total, these tolerability, efficacy, and PK data indicate that the 3408-phosphate prodrug strategy is effective and holds the potential to further increase the therapeutic safety window for the 1810 lead drug candidate.

In conclusion, introducing a phosphate moiety to the phenol group of preclinical antituberculosis candidate 1810 generated the prodrug 3408, which has similar overall efficacy in vivo, but an improved therapeutic window and a higher MTD when delivered by i.v. bolus administration. 3408 is thus an attractive backup preclinical candidate to 1810, which is under investigation for inhaled delivery as an MDR/XDR-TB therapeutic.

Highlights.

• Spectinamides are a novel class of narrow-spectrum antitubercular agents with the potential to treat drug-resistant tuberculosis infections.

• Spectinamides can exhibit transient acute toxicity following bolus administration due to their high free fraction.

• A phosphate prodrug strategy was developed to expand the therapeutic index of the spectinamides.

• Spectinamide prodrug 3408 demonstrates an improved maximum tolerated dose and equivalent efficacy in a BALB/c acute tuberculosis mouse infection model to the spectinamide parent.

Abbrevations

HBTU

[benzotriazol-1-yloxy(dimethylamino)methylidene]-dimethylazanium

DIPEA

N,N-Diisopropylethylamine

DMF

N,N-Dimethylformamide hexafluorophosphate

DMAP

4-Dimethylaminopyridine

HRMS

High-resolution mass spectrometry