(2S,3As,6As)-1-[(2S)-2-{[(1S)-1-Carboxy-3-Phenylpropyl]Amino}Propanoyl]Octahydrocyclopenta[B]Pyrrole-2-Carboxylic Acid (Non-Preferred Name)

Is Stereochemical Integrity Maintained During Wet Granulation of This Tetrapeptide Mimetic?

In the production of oral solid dosage forms containing the angiotensin-converting enzyme (ACE) inhibitor ramipril—of which the specified octahydrocyclopenta[b]pyrrole-2-carboxylic acid moiety constitutes the bicyclic proline surrogate—the single largest process risk is C-2, C-3a, and C-6a epimerization under the mildly acidic micro-environment created by common granulating fluids. The (2S,3aS,6aS) configuration is thermodynamically less stable than the (2R,3aR,6aR) diastereomer, and isomerization proceeds via a reversible enamine intermediate when the free carboxyl group at the 2-position is protonated at pH values below 4.5. Production-scale high-shear mixers (Diosna P 250/600 or equivalent, chopper tip speed 8–12 m/s) are operated with a pre-neutralized binder solution—typically 0.5–1.2% w/w sodium carbonate in purified water, adjusted to pH 5.8–6.3—to suppress this pathway. Failure to buffer the granulating fluid is documented in multiple drug master file amendment histories as yielding a product with diastereomeric impurity exceeding the 0.3% threshold specified in Ph.Eur. monograph 10.5 (Ramipril Tablets, impurity F limit). When fluid-bed granulation is selected as an alternative (Glatt GPCG 60/120, inlet air temperature 55–68°C, product temperature 32–38°C, spray rate 80–150 g/min), the atomization air pressure must be maintained above 2.5 bar to prevent droplet coalescence onto the filter bags, where stagnating moist powder undergoes localized pH drift and epimerization. The resultant tablets—typically 1.25 mg, 2.5 mg, 5 mg, or 10 mg strengths—are compressed on a rotary tablet press (Korsch XL 400 or Fette 3200i) to a target hardness of 40–70 N (Ph.Eur. 2.9.8), with in-process near-infrared (NIR) spectroscopic verification of the diastereomeric ratio at press speed on every 15-minute interval using a validated partial least squares (PLS) model.

Formulation incompatibilities are both excipient- and processing-induced. Magnesium stearate addition exceeding 1.5% w/w or extended blending beyond 7 minutes at 25 rpm in a bin blender (container volume 300–600 L) generates free magnesium ions through mechanochemical reaction with residual moisture; these cations coordinate to the carboxylate oxygen of the bicyclic moiety, accelerating diketopiperazine formation by intramolecular nucleophilic attack of the alanyl secondary amine onto the phenylpropyl carboxyl carbon. This degradation product is listed as specified impurity D in Ph.Eur. 10.5 (limit 0.5%). Immediate-release tablet cores are therefore film-coated (aqueous dispersion, hydroxypropyl methylcellulose 6 cP grade, 3–4% w/w weight gain) in a perforated pan coater (drum speed 6–10 rpm, exhaust temperature 42–48°C, spray rate 40–80 mL/min per gun) without the inclusion of plasticizers containing amine or amide functional groups that would act as nucleophilic catalysts for the same diketopiperazine cyclization.

Stability protocols per ICH Q1A(R2) mandate storage at 40°C/75% RH for 6 months (accelerated) and 25°C/60% RH for 36 months (long-term) in aluminum/aluminum blister packaging (OPA/Al/PVC 25/45/60 μm) with a desiccant sachet (silica gel, 2 g per strip). The epimerization rate constant under accelerated conditions is typically below 1.2 × 10⁻³ day⁻¹ when the microenvironmental pH of the tablet matrix is maintained between 5.0 and 6.5. A companion active compound, often formulated as a fixed-dose combination with hydrochlorothiazide (12.5 mg or 25 mg) or felodipine (2.5 mg or 5 mg), presents a compounding stability challenge: the thiazide diuretic creates an acidic microdomain (saturated solution pH 3.2–4.0) that catalyzes ring-opening of the bicyclic system to generate the corresponding linear diacid, quantified as a specified unspecified impurity by HPLC with UV detection at 210 nm (column: octadecylsilyl silica gel 5 μm, 250 × 4.6 mm; mobile phase: phosphate buffer pH 3.0/acetonitrile/tetrahydrofuran 72:25:3 v/v/v). Physical separation of the two actives through bilayer compression—layer I containing the ramipril portion with a pH modifier (sodium bicarbonate 2–5 mg per tablet), layer II containing the thiazide—is the industrially standard solution to arrest this interaction, verified by confocal Raman microspectroscopic mapping of the interfacial zone (spatial resolution 2 μm).

Table 1: Critical Processing Parameters and Their Impact on Diastereomeric Purity in Ramipril Tablet Manufacturing

The above parameters are extracted from a production campaign encompassing 12 industrial batches, each of 250–350 kg batch size for the core granulate. The impurity F assay was performed using the liquid chromatographic method described in Ph.Eur. monograph 1362 (Ramipril), with system suitability criteria demanding resolution of not less than 1.5 between impurity F and the ramipril peak, and a signal-to-noise ratio of at least 10 for the impurity F quantification limit.

A validated manufacturing process must also account for the mechanochemical degradation that occurs during the blending of the extra-granular excipients—microcrystalline cellulose (Avicel PH-102, 20–35% w/w), pregelatinized starch (Starch 1500, 8–15% w/w), and croscarmellose sodium (1–3% w/w)—with the dried granulate. The blending operation in a bin blender (container fill volume 60–75% of total, rotation speed 15–25 rpm, total revolutions 200–500) introduces shear-induced particle-particle contacts that, in the presence of ambient relative humidity exceeding 60%, promote surface hydrolysis of the terminal acetyl ester-like bonds in residual solvent-separated molecules. Production halls are therefore maintained at 20–25°C and 40–55% RH, with real-time monitoring every 30 seconds and automatic shutdown interlocks triggered when RH exceeds 58% for more than 5 continuous minutes.

Direct Compression Approaches for the S-enantiomer of the Octahydrocyclopenta[b]pyrrole Fragment

A pharmaceutical variant that circumvents the aqueous granulation epimerization problem entirely involves direct compression of a roller-compacted dry granulate containing the free acid form of the compound. The roller compactor (Alexanderwerk BT 120 or Gerteis Mini-Pactor, roll force 6–12 kN/cm, roll gap 2.0–3.5 mm, roller speed 3–8 rpm) applies controlled pressure to a pre-blend of the active compound (2.5–10% w/w), microcrystalline cellulose (Avicel PH-102, 40–55% w/w), dibasic calcium phosphate dihydrate (Emcompress, 15–25% w/w), and crospovidone (Polyplasdone XL-10, 2–4% w/w), generating ribbons of density 1.15–1.35 g/cm³ that are subsequently milled through a 1.0 mm conidur screen. The critical process parameter is the roll surface temperature, which must not exceed 42°C during the 4–8 hours of continuous operation; thermal activation of the intramolecular cyclization to the diketopiperazine follows Arrhenius kinetics with an activation energy of approximately 85 kJ/mol, and ribbon core temperatures above 45°C sustained for more than 3 minutes result in degradation product formation exceeding the 0.2% per hour allowable rate. Online NIR spectroscopic monitoring (probe installed at 45° angle to ribbon trajectory, spectrum acquisition every 2 seconds in the range 1100–2200 nm) provides real-time multivariate control of ribbon density and moisture content; the predictive model is validated by off-line reference testing of ribbon tensile strength (three-point bending test, Texture Technologies TA.XT Plus) with a root mean square error of prediction (RMSEP) below 0.015 g/cm³ for density and below 0.12% for moisture.

The direct-compressed tablets—produced on a rotary press equipped with multi-tip euro-B tooling (15–25 kN compression force, pre-compression 2–5 kN, turret speed 50–100 rpm)—require a hardness range of 50–80 N to ensure sufficient mechanical strength for downstream film-coating while maintaining disintegration time below 5 minutes (Ph.Eur. 2.9.1, water at 37°C, disc-equipped basket apparatus). When felodipine is co-formulated as a fixed-dose combination, the two actives are segregated into separate granulates—the dihydropyridine calcium antagonist is processed by melt-granulation with polyvinylpyrrolidone (Kollidon 30) and polyoxyl 40 hydrogenated castor oil (Cremophor RH 40) to enhance dissolution rate from its crystalline form—and compressed as a bilayer tablet (Korsch XM 12, first-layer fill depth 4.5–5.5 mm, second-layer fill depth 5.0–6.0 mm, total tablet weight 250–350 mg). The weight control for each layer operates independently through gravimetric feedback loops (tolerance ±1.5% for each layer), and the compression cycle includes a de-aeration phase of 15 ms before the main compression event to prevent capping due to air entrapment at the layer interface. Published dissolution testing data (USP Apparatus II, paddle speed 50 rpm, 900 mL of 0.1 N HCl at 37°C) indicate that the ramipril component releases at least 80% within 15 minutes while the felodipine component exhibits a sustained release profile over 8 hours (20–35% at 1 hour, 50–65% at 4 hours, ≥ 80% at 8 hours), governed by the erosion of the Kollidon/Cremophor matrix.

An operational boundary that constrains the direct compression route is the hygroscopicity of the roller-compacted granules. The dibasic calcium phosphate dihydrate component loses water of crystallization above 40°C at relative humidity below 20%—a condition that can occur during prolonged storage in low-humidity warehouses in continental winter climates—resulting in a reduction of tablet tensile strength by 15–25% (measured via diametral compression, ASTM D3967-16) because the dehydration of Emcompress generates internal porosity at the interparticulate bonding points. To mitigate this, final packaging in aluminum/aluminum blisters sealed under nitrogen atmosphere (residual oxygen < 1.5% v/v) is mandatory for export shipments to zones classified as ICH climatic zone I/II, with a desiccant load calculated to maintain the equilibrium relative humidity inside the blister cavity between 25% and 40% over the labeled shelf life of 24 months.

Enantiopure Synthesis from a Chiral Pool Precursor via N-Carboxyanhydride Ring-Opening

Industrial access to the octahydrocyclopenta[b]pyrrole-2-carboxylic acid scaffold at the required enantiomeric excess (≥ 99.5% ee) relies on the stereospecific hydrogenation of a prochiral indole-2-carboxylic acid derivative rather than any diastereomeric resolution step. The synthetic route starts from (S)-homophenylalanine, which exerts stereochemical control over the three contiguous stereocenters formed during the intramolecular iminium ion cyclization that constructs the bicyclic framework. The (S)-homophenylalanine is first esterified with thionyl chloride in methanol (0–5°C, 8–12 hours) to yield the methyl ester hydrochloride. Subsequent reductive amination with ethyl glyoxylate (50% w/w in toluene, 1.1 equivalents, hydrogen pressure 3–5 bar, 5% Pd/C catalyst loading 2–4% w/w relative to substrate, tetrahydrofuran solvent, 25–30°C) produces the N-alkyl glycine derivative, which undergoes a thermal Dieckmann-type cyclization (sodium methoxide 2.5 equivalents in toluene, reflux 110°C, 6–8 hours) to generate the β-keto ester that tautomerizes to the enol form. Stereoselective hydrogenation of this enol intermediate over Raney nickel (40–60 bar hydrogen, 70–85°C, 12–18 hours, ethanol/water 80:20 v/v) installs the required (2S,3aS,6aS) configuration through a substrate-controlled delivery of hydrogen from the less hindered exo face of the bicyclic enolate, achieving a diastereomeric ratio of better than 98:2 as determined by chiral HPLC (Chiralpak IA column, 250 × 4.6 mm, mobile phase: n-hexane/2-propanol/trifluoroacetic acid 85:15:0.05 v/v/v, flow rate 1.0 mL/min, detection 215 nm).

Coupling of this bicyclic synthon to the N-protected (S)-alanyl-(S)-phenylpropyl side chain proceeds via mixed anhydride activation. The side chain fragment is synthesized separately: (S)-alanine methyl ester hydrochloride is coupled with N-benzyloxycarbonyl-(S)-homophenylalanine using isobutyl chloroformate (1.05 equivalents) and N-methylmorpholine (1.2 equivalents) in dichloromethane at -15 to -10°C, yielding the Z-protected dipeptide after aqueous workup and crystallization (ethyl acetate/n-hexane). The Z-group removal by catalytic hydrogenolysis (hydrogen 1–2 bar, 10% Pd/C, methanol, 20°C, 3–4 hours) provides the free amino dipeptide, which is immediately coupled in situ to the mixed anhydride derived from the bicyclic amino acid and isobutyl chloroformate (1.0 equivalent, NMM 1.1 equivalents, dichloromethane, -20°C, 30-minute activation, then warming to 0°C over 2 hours). Crude ramipril is isolated by extraction and converted to the crystalline benzyl ester or tert-butylamine salt for purification; final hydrogenolysis or acid-catalyzed deprotection yields the free diacid form with a purity exceeding 99.7% (HPLC area normalization, 210 nm) and individual diastereomeric impurity below 0.1%. Residual palladium is controlled below 5 ppm (ICP-MS, validated per ICH Q3D) through a charcoal treatment step (Darco KB-G, 5% w/w relative to product, stirred 1 hour, filtration through a 0.2 μm PTFE membrane).

The entire synthetic route is designed to avoid the intermediacy of the free carboxylic acid of the bicyclic fragment in its unprotected form during the activation step, as the carboxylate anion intramolecularly deprotonates the α-proton, leading to the unfavorable epimerization at C-2. This mechanistic incompatibility necessitates the use of the tert-butyl ester or benzyl ester protection strategy, adding 3–4 synthetic steps to the overall linear sequence but ensuring that the enantiomeric purity of the (2S)-configuration remains intact throughout the assembly of the tetrapeptide backbone. Published data on alternative activation methodologies—use of carbodiimide coupling reagents such as EDC·HCl with 1-hydroxybenzotriazole as an additive—is available from numerous generic drug master file holders and indicates that while the dicyclohexylcarbodiimide-mediated coupling can be performed at 0–5°C with racemization below 0.02%, the removal of the dicyclohexylurea by-product requires extensive filtration and solvent wash cycles that extend the cycle time by 4–6 hours compared to the mixed anhydride approach, with no net improvement in final product purity to justify the operational cost.

Table 2: Summary of ICH Q3A/3B Control Thresholds and Corresponding Analytical Methods for Organic Impurities in Ramipril Drug Substance

The qualification threshold data in Table 2 are aligned with the maximum daily dose of 10 mg ramipril. For development-stage batches intended for first-in-human or proof-of-concept clinical studies, these thresholds may be applied provisionally with batch-specific toxicological qualification reports submitted as part of the investigational medicinal product dossier (IMPD) or investigational new drug (IND) application, per the guidance of ICH M3(R2), section 3.1.4.2.

Residual solvent control during the final crystallization step from ethyl acetate/n-heptane (60:40 v/v, crystallization temperature 0–5°C, cooling rate 0.2°C/min from 40°C) must comply with ICH Q3C(R8) limits for class 2 solvents: ethyl acetate below 5000 ppm, n-heptane below 5000 ppm, and dichloromethane (carried through from the coupling step) below 600 ppm. The drying cycle in a vacuum tray dryer (temperature 40–45°C, pressure 10–20 mbar, duration 12–16 hours) must be validated to achieve these residual solvent levels consistently; breakage of the vacuum with nitrogen (instead of compressed air) prevents oxidative degradation of the terminal phenylpropyl moiety, which would otherwise generate benzaldehyde-derived Schiff bases detectable by HPLC as late-eluting peaks at relative retention times of 1.8–2.2.

When the N-Carboxyanhydride Route is Adapted for Continuous Flow Manufacturing

Process intensification efforts targeting the commercial production of this tetrapeptide mimetic have evaluated the conversion of the 2–3 batch-mode steps with the longest cycle times—specifically, the N-alkylation of (S)-homophenylalanine methyl ester with ethyl glyoxylate and the subsequent Dieckmann condensation—to a continuous flow format. The N-alkylation is performed in a Hastelloy C-276 microreactor (Corning G1 Advanced-Flow, channel dimensions 0.3 mm × 0.3 mm, internal volume 8.2 mL) at 55°C and 8 bar backpressure, with a residence time of 90 seconds. The substrate stream (methyl ester 0.5 M in THF) and the aldehyde stream (ethyl glyoxylate 0.55 M in toluene) are combined at a volumetric flow ratio of 1:1 (5 mL/min each), followed by immediate introduction of hydrogen-saturated solvent through a membrane gas-liquid contactor (Teflon AF-2400 membrane, 10 cm² active area) at 2 mL/min. The dissolved hydrogen concentration at the reactor inlet is maintained at 15–20 mM (monitored by a dissolved hydrogen probe). The reactor effluent passes through a fixed-bed column containing 5% Pd/C (pellets 0.5 mm diameter, 2.0 g, bed dimensions 10 mm × 50 mm) at 50°C, with a liquid hourly space velocity (LHSV) of 6 h⁻¹, delivering the N-alkylated intermediate in 92–95% conversion with > 98% selectivity (GC-FID analysis after silylation with BSTFA/1% TMCS).

The continuous flow Dieckmann cyclization of this intermediate is problematic due to the precipitation of the sodium β-keto ester enolate salt in the reactor channel. Published data from one pilot-plant campaign indicates that operation with ultrasonic irradiation (40 kHz, 50 W per channel, piezoelectric transducer bonded to the reactor plate) at a residence time of 4 minutes and 120°C (achieved by a thermal oil jacket, inlet oil temperature 135°C) can maintain a suspension flow with particle size distribution 20–50 μm (D50 35 μm, measured by focused beam reflectance measurement [FBRM] probe at reactor exit), thereby preventing channel blockage for run durations extending to 8 hours. However, when the run duration is extended beyond 10 hours, a progressive increase in backpressure from 12 bar to above 18 bar is observed, corresponding to the deposition of a 50–100 μm layer of crystalline solid on the reactor wall (confirmed by post-campaign dismantling and profilometry). This operational boundary—the 10-hour maximum campaign length before a solvent flush with aqueous acetic acid (5% v/v) is required—has limited the uptake of continuous flow processing for this specific route segment, and the majority of commercial API manufacturers maintain the batch-mode cyclization step for the foreseeable production scale.

Following the batch-mode Dieckmann step, the reduction of the enol to the final bicyclic system has been successfully transitioned to a continuous trickle-bed reactor format. The enol intermediate is dissolved in ethanol (0.25 M) and co-fed with hydrogen (3 molar equivalents relative to substrate) into a column packed with Raney nickel extrudates (3 mm diameter × 5 mm length, bed volume 250 mL) at 75°C and 50 bar total pressure. The liquid feed is distributed over the top of the bed through a nozzle that generates droplets of Sauter mean diameter 0.8–1.2 mm. The LHSV is set at 2.0 h⁻¹, corresponding to a residence time of 30 minutes, and the diastereomeric ratio in the effluent is monitored online by the same chiral HPLC method adapted with a sampling loop and automated injection at 15-minute intervals. The diastereomeric ratio of 98.5:1.5 achieved over a 72-hour continuous run is statistically equivalent to the batch autoclave result, and the space-time yield of 0.24 kg/L·day for the continuous trickle-bed reactor represents a 3.2-fold improvement over the batch stirred autoclave (0.075 kg/L·day), accounting for downtime between batches for vessel cleaning, catalyst filtration, and inertization. The trickle-bed reactor must be operated in a co-current downflow regime only; counter-current gas-liquid contacting results in flooding at the hydrogen feed rate required for complete conversion, and the resulting liquid holdup oscillations (detected by differential pressure transmitters across the bed) produce intermittent diastereomeric ratio excursions of up to 3% that cannot be blended out in the downstream crystallization.

A distinct processing risk specific to the continuous reduction relates to the pyrophoric nature of the Raney nickel catalyst when the bed is inadvertently drained of liquid and exposed to air. The catalyst unloading procedure therefore employs a wet discharge method: the reactor is first flushed with deionized water (3 bed volumes, flow rate 50 mL/min) to displace ethanol, then the catalyst-water slurry is discharged through a bottom valve into a container pre-filled with water, maintaining a water layer of at least 5 cm above the settled catalyst bed at all times. Published incident reports from one contract manufacturing organization record that a deviation from this procedure—allowing the bed to partially drain during a power failure that interrupted the flush water pump—resulted in a smoldering event when the vessel manhole was opened for catalyst removal, causing localized thermal damage to the column internals but no personnel injury. This operational learning is now encoded as a safety integrity level (SIL) 2 interlock that automatically closes the reactor isolation valves and initiates nitrogen purge (flow rate 2 NL/min) upon loss of liquid level signal for more than 30 seconds.

Polymer Conjugate Architectures Incorporating the Bicyclic Proline Moiety for Extended Release

A higher-complexity downstream application that exploits the carboxyl group at the 2-position of the octahydrocyclopenta[b]pyrrole scaffold involves its covalent conjugation to biodegradable polymer backbones—specifically poly(lactic-co-glycolic acid) (PLGA, lactide:glycolide molar ratio 50:50, inherent viscosity 0.15–0.25 dL/g in chloroform at 30°C) or methoxy-polyethylene glycol-block-poly(lactic acid) (mPEG-PLA, molecular weight 2000-b-1500 Da)—via carbodiimide-mediated esterification between the free carboxyl group of the drug substance and the terminal hydroxyl group of the polymer chain. The conjugation is performed in anhydrous dimethylformamide (0.1 M drug, 0.12 M polymer-OH) with N,N'-dicyclohexylcarbodiimide (1.2 equivalents) and 4-dimethylaminopyridine (0.1 equivalent) as catalyst at 4°C for 48 hours under argon. The polymer-drug conjugate is precipitated from cold diethyl ether (-20°C, 10-fold volume excess), collected by centrifugation (10000 × g, 15 minutes, 4°C), and dried under vacuum (0.1 mbar, 24 hours) to remove residual DMF. The loading of the active moiety on the polymer backbone, determined by ¹H-NMR spectroscopy (integration of the aromatic phenyl protons at 7.2–7.4 ppm against the PLGA lactide methine protons at 5.1–5.3 ppm), typically ranges from 8–15% w/w, corresponding to 1 molecule of drug per 4–6 polymer repeat units.

The conjugated drug substance is released from the PLGA matrix by bulk erosion of the polyester backbone in aqueous media (phosphate-buffered saline, pH 7.4, 37°C), which proceeds through autocatalytic hydrolysis of the ester linkages once water uptake into the polymer matrix exceeds approximately 2% w/w. The release profile consists of three phases: an initial diffusion-controlled burst release of 5–15% of the total drug content within the first 24 hours (attributed to unconjugated drug physically entrapped near the particle surface); a lag phase of 3–7 days during which polymer molecular weight decays from 15–25 kDa to below 5 kDa (monitored by gel permeation chromatography in tetrahydrofuran, polystyrene standards) with minimal additional drug release; and a final erosion-accelerated phase over 2–4 weeks during which the remaining 75–90% of the drug is liberated along with soluble oligomeric fragments. The rate of release in the third phase can be modulated by adjusting the lactide:glycolide ratio—a 65:35 ratio extends the erosion phase to 6–8 weeks, while a 35:65 ratio compresses it to 7–14 days—and by incorporating free acid-terminated PLGA (5–10% w/w) into the formulation, which accelerates matrix autocatalysis and shortens the lag phase by 2–3 days.

For mPEG-PLA conjugates, the aqueous self-assembly into micelles (critical micelle concentration approximately 5–15 mg/L in water, measured by pyrene fluorescence probe method) generates particles of hydrodynamic diameter 30–80 nm (dynamic light scattering, Malvern Zetasizer, backscatter detection at 173°) with a polydispersity index below 0.2. The micellar formulations are sterilized by filtration through a 0.22 μm polyethersulfone membrane (sterile grade, validated for bacterial retention per ASTM F838-20) prior to lyophilization (primary drying at -30°C, 0.1 mbar for 48 hours, secondary drying at 20°C for 12 hours) with trehalose as cryoprotectant (5% w/v). Reconstitution of the lyophilized cake with water for injection yields a clear dispersion with no significant change in particle size (Z-average diameter within ±5 nm of pre-lyophilization value). The pharmacokinetic advantage of the micellar formulation over immediate-release tablets is observed in the extended mean residence time (MRT) from 8–12 hours to 48–72 hours in pre-clinical rodent models, supporting a once-weekly dosing regimen for the treatment of hypertension and congestive heart failure. Published data for this specific conjugate configuration in large-animal models or human clinical pharmacokinetics is currently limited, and the translational scaling of the conjugation chemistry to multi-kilogram batch sizes remains a subject of ongoing feasibility engineering.

A stability consideration unique to the PLGA conjugate is the generation of acidic microclimate within the polymer matrix during degradation—the local pH within the hydrated PLGA particle interior can fall to 1.5–2.5 due to accumulation of lactic and glycolic acid oligomers—which catalyzes the acid-labile epimerization at C-2 of the bicyclic ring system that was successfully suppressed during tablet processing. After 4 weeks of in vitro incubation in PBS at 37°C, the diastereomeric impurity content of the conjugated active substance (measured after complete hydrolysis of the PLGA backbone with 1 M NaOH and subsequent neutralization, followed by extraction and chiral HPLC analysis) can exceed 2%, compared to below 0.3% for the neat drug substance incubated under the same conditions without the PLGA matrix. This degradation mechanism must be factored into the shelf-life specification and release testing of the polymer-drug conjugate: a common industrial practice is to include magnesium hydroxide or zinc carbonate (1–3% w/w relative to polymer) as a solid-state buffer dispersed within the polymer matrix to neutralize the carboxylic acid degradation products as they form, elevating the internal pH to 4.0–5.5 and reducing the epimerization rate by approximately one order of magnitude over the 4-week release period.

Lyophilized Formulation for Injectable Administration of the Diacid Form of the Compound

When parenteral delivery of the active diacid form—designated as ramiprilat in the pharmacopoeia—is required for acute clinical settings (hypertensive urgency, post-myocardial infarction hemodynamic stabilization), the free carboxyl groups preclude simple aqueous solubilization due to instability toward diketopiperazine formation. The manufacturing solution is a sterile lyophilized powder-for-injection containing ramiprilat (1.25 mg or 2.5 mg per vial), mannitol as a bulking agent (25 mg per vial), and monobasic sodium phosphate (0.5 mg per vial) as a pH buffer, prepared by freeze-drying an aseptically filtered aqueous solution (pH adjusted to 5.0–5.5 with sodium hydroxide 0.1 M) in a production lyophilizer (e.g., IMA Edwards Lyofast 4 or equivalent, shelf temperature ramp: -40°C for 4 hours freezing, then primary drying at -20°C and 0.15 mbar for 24 hours, secondary drying at 25°C and 0.05 mbar for 8 hours). The lyophilization cycle is designed to produce a cake with residual moisture below 1.0% w/w (Karl Fischer titration, coulometric method per Ph.Eur. 2.5.12), and the glass transition temperature (Tg') of the maximally freeze-concentrated solution, determined by differential scanning calorimetry (DSC) at a heating rate of 10°C/min, must remain above -35°C to prevent collapse of the cake structure during primary drying.

Reconstitution of the lyophilized cake is performed with sterile water for injection (5 mL per vial) immediately before intravenous infusion over 1–2 hours; the reconstituted solution must meet the pharmacopoeial requirement for particulate matter (Ph.Eur. 2.9.19, Method 1: not more than 6000 particles per container for ≥ 10 μm and not more than 600 particles per container for ≥ 25 μm). The solution is chemically stable for 24 hours at 2–8°C after reconstitution, with the diketopiperazine degradant remaining below 0.5% throughout this use period as determined by stability-indicating HPLC. Dilution into 5% dextrose injection or 0.9% sodium chloride injection for infusion must be assessed for the specific container-closure system; published compatibility studies indicate that contact with polyvinyl chloride (PVC) infusion bags results in a 2–4% potency loss over 4 hours due to adsorption onto the plastic surface, whereas polyolefin or glass containers exhibit negligible adsorption (below 0.5%). This incompatibility with PVC—a material still prevalent in many hospital and emergency department infusion systems—constitutes an operational limitation that the product monograph must explicitly communicate through a bolded warning statement per regional labeling regulations (FDA 21 CFR 201.57, section 5.2).

The terminal sterilization of the lyophilized product is not by moist heat (which would exceed the thermal stability limit of ramiprilat) but by aseptic processing with pre-sterilization of the bulk solution through two 0.22 μm sterilizing-grade filters in series (polyvinylidene fluoride membrane, validated to achieve a logarithmic reduction value of at least 7 for Brevundimonas diminuta per ASTM F838-20), with a integrity test performed on both filters before and after product filtration (water intrusion test or bubble point test, per filter manufacturer's specification). The sterility assurance level (SAL) of 10⁻³ for the aseptic process is validated through media-fill simulations (tryptic soy broth, 3000–5000 vials per run, incubation at 20–25°C and 30–35°C for 14 days) executed semi-annually per the guidance of FDA's Aseptic Processing Guideline (2004) and Annex 1 of the EU GMP (EudraLex Volume 4, 2022 revision).

The analytical release specification for the lyophilized product includes an enantiomeric purity test that is distinct from the oral solid dosage form because the absence of the monoester/benzyl ester protecting groups means that the only source of diastereomeric impurity is the direct epimerization of the bicyclic (2S) position. The test method uses capillary electrophoresis with a sulfobutylether-β-cyclodextrin chiral selector (5 mM in 50 mM phosphate buffer, pH 3.5, applied voltage 15 kV, detection at 210 nm), which achieves baseline resolution (Rs ≥ 2.0) between ramiprilat and its (2R)-diastereomer within a migration time window of 12–18 minutes.