Depolymerization of Heparin Benzyl Ester in a Flow System †

: The preparation of enoxaparin via the base-promoted depolymerization of heparin is well known. In this paper, we present our efforts to convert this preparation into a ﬂow method. Streams of base and ester solutions are combined and heated in a ﬂow reactor, and the variations in ﬂow rate, temperature, and residence time were examined. NMR analysis demonstrated that careful control of the reaction parameters affords enoxaparin-like materials. An analysis was carried out by NMR spectroscopy and the optimum conditions were determined


Introduction
Heparin is a highly charged, highly anionic glycosaminoglycan valued primarily for its anticoagulant properties [1]. In general, it is a linear polymer of alternating glucosamine and iduronic acid saccharides. In addition, heparin may include charged groups including sulfonates and carboxylates. The amino group in the glucosamine is generally observed as either a bare amino (NH 2 ), an acetamide (NHCOCH 3 ), or an aminosulfate (NHSO 3 − ). The molecular weight is distributed at about an average of around 15,000 g/mol. Certain crucial pentasaccharide sequences along the chain are responsible for the anticoagulant activity of heparin as it binds to antithrombin. See Figure 1 for a generalized structure of heparin.

Introduction
Heparin is a highly charged, highly anionic glycosaminoglycan valued primarily for its anticoagulant properties [1]. In general, it is a linear polymer of alternating glucosamine and iduronic acid saccharides. In addition, heparin may include charged groups including sulfonates and carboxylates. The amino group in the glucosamine is generally observed as either a bare amino (NH2), an acetamide (NHCOCH3), or an aminosulfate (NHSO3 − ). The molecular weight is distributed at about an average of around 15,000 g/mol. Certain crucial pentasaccharide sequences along the chain are responsible for the anticoagulant activity of heparin as it binds to antithrombin. See Figure 1 for a generalized structure of heparin. While unfractionated heparin remains a very important drug, fractionated or low molecular weight heparins (LMWH) were introduced as substitutes. These heparin derivatives are important in terms of clinical use due to reduced hemorrhagic side effects, more predictable pharmacological action, and improved bioavailability. Additionally, delivery of LMWHs is more convenient through subcutaneous administration [2].
One of the most commonly used LMWHs is enoxaparin with an average molecular weight about 4500 [3]. Presently, enoxaparin is prepared in a batch process [4]. In a typical process, a benzyl ester of heparin is treated with a base and moderate heat (55-65 °C) for   While unfractionated heparin remains a very important drug, fractionated or low molecular weight heparins (LMWH) were introduced as substitutes. These heparin derivatives are important in terms of clinical use due to reduced hemorrhagic side effects, more predictable pharmacological action, and improved bioavailability. Additionally, delivery of LMWHs is more convenient through subcutaneous administration [2].
One of the most commonly used LMWHs is enoxaparin with an average molecular weight about 4500 [3]. Presently, enoxaparin is prepared in a batch process [4]. In a typical process, a benzyl ester of heparin is treated with a base and moderate heat (55-65 • C) for from 30 min to 2 h with some post-reaction processing to afford the LMWH. The structure of enoxaparin is shown in Figure 2, and can be characterized by the carboncarbon double bond that forms at the non-reducing end. When produced, about 15-25% of the polysaccharide chains in enoxaparin have the 1,6 anhydro ring structure at the reducing end.
from 30 min to 2 h with some post-reaction processing to afford the LMWH. The structure of enoxaparin is shown in Figure 2, and can be characterized by the carbon-carbon double bond that forms at the non-reducing end. When produced, about 15-25% of the polysaccharide chains in enoxaparin have the 1,6 anhydro ring structure at the reducing end. The objective of our study was to investigate the base promoted depolymerization of heparin benzyl ester using a flow system. The potential advantages of a flow process have been well articulated and include reproducibility, safety, and quality control [5].

Methods and Materials
Heparin benzyl ester was provided by Smithfield Biosciences. All other reagents were used as received. NMR spectra were recorded on a Bruker Avance 500 MHz NMR instrument at 298K and using D2O as the solvent.

Procedure for Flow Reactions
Continuous flow experiments were carried out on the Syrris FRX flow system with a 12.05 mL reactor coil ( 1 /16" i.d. tubing). Note: The continuous flow system was flushed with water before the reaction was run. An aqueous solution of heparin benzyl ester and an aqueous solution of sodium hydroxide were merged through a t-mixer. The merged solution was pumped through the coil submerged in an oil bath at a regulated temperature. The reaction stream was collected in a flask and neutralized with dilute HCl. Subsequently, the product was diluted with 3X volume of methanol, the solution was evaporated, washed with acetone, purified through cassette dialysis (slide-A-Lyzer G2, 2000 MWCO, Thermo-Fisher, Waltham, MA, USA), and freeze dried. HNMR was recorded on the recovered material.
The variations in reaction conditions are presented in Table 1.  The objective of our study was to investigate the base promoted depolymerization of heparin benzyl ester using a flow system. The potential advantages of a flow process have been well articulated and include reproducibility, safety, and quality control [5].

Methods and Materials
Heparin benzyl ester was provided by Smithfield Biosciences. All other reagents were used as received. NMR spectra were recorded on a Bruker Avance 500 MHz NMR instrument at 298 K and using D 2 O as the solvent.

Procedure for Flow Reactions
Continuous flow experiments were carried out on the Syrris FRX flow system with a 12.05 mL reactor coil ( 1 / 16 " i.d. tubing). Note: The continuous flow system was flushed with water before the reaction was run. An aqueous solution of heparin benzyl ester and an aqueous solution of sodium hydroxide were merged through a t-mixer. The merged solution was pumped through the coil submerged in an oil bath at a regulated temperature. The reaction stream was collected in a flask and neutralized with dilute HCl. Subsequently, the product was diluted with 3X volume of methanol, the solution was evaporated, washed with acetone, purified through cassette dialysis (slide-A-Lyzer G2, 2000 MWCO, Thermo-Fisher, Waltham, MA, USA), and freeze dried. HNMR was recorded on the recovered material.
The variations in reaction conditions are presented in Table 1.

Results
Our objective in this study was to derive flow conditions that gave rise to the most 'enoxaparin-like' product by HNMR. Our inspection of the results in the various entries concentrated on two peaks: the vinyl resonance at about 6.2 ppm and the acetamide peak at about 2.2. As will be discussed below, the best match for USP enoxaparin as compared to a reaction product was for entry 12 with 25.1 min residence time and 55 • C temperature. The stacked spectra are presented below in Figure 3.

Results
Our objective in this study was to derive flow conditions that gave rise to the most 'enoxaparin-like' product by HNMR. Our inspection of the results in the various entries concentrated on two peaks: the vinyl resonance at about 6.2 ppm and the acetamide peak at about 2.2. As will be discussed below, the best match for USP enoxaparin as compared to a reaction product was for entry 12 with 25.1 min residence time and 55°C temperature. The stacked spectra are presented below in Figure 3. From this comparison, we have established that the crude reaction product using these conditions are similar to enoxaparin when refined and purified. The supplemental information includes the NMR spectra of the other entries. The variations produced features similar to enoxaparin but the conditions described here are the optimal one that we observed.

Discussion
The parameters we chose for our investigation were informed by the batch reaction data that was previously reported. There are two simultaneous reactions that must be balanced: saponification of the ester and elimination. If the saponification predominates, the molecular weight would be higher than desired. If the elimination predominates, the molecular weight would be low. Figure 4 shows the HNMR spectrum of USP enoxaparin. The vinyl peak at 6.2 is new to the spectrum (compared to heparin sodium) and is a marker for the depolymerization extent while the NAc peak at 2.2 acts as a standard for the starting enoxaparin ester [6,7]. The balance of the signals is a good indicator of the average molecular weight. The spectra of the other entries in Table 1 are included in the Supplementary Materials. A smaller peak in the vinyl region is diagnostic for too high a molecular weight, while those at a higher temperature showed substantial formation of extraneous peaks [8]. From this comparison, we have established that the crude reaction product using these conditions are similar to enoxaparin when refined and purified. The supplemental information includes the NMR spectra of the other entries. The variations produced features similar to enoxaparin but the conditions described here are the optimal one that we observed.

Discussion
The parameters we chose for our investigation were informed by the batch reaction data that was previously reported. There are two simultaneous reactions that must be balanced: saponification of the ester and elimination. If the saponification predominates, the molecular weight would be higher than desired. If the elimination predominates, the molecular weight would be low. Figure 4 shows the HNMR spectrum of USP enoxaparin. The vinyl peak at 6.2 is new to the spectrum (compared to heparin sodium) and is a marker for the depolymerization extent while the NAc peak at 2.2 acts as a standard for the starting enoxaparin ester [6,7]. The balance of the signals is a good indicator of the average molecular weight. The spectra of the other entries in Table 1 are included in the Supplementary Materials. A smaller peak in the vinyl region is diagnostic for too high a molecular weight, while those at a higher temperature showed substantial formation of extraneous peaks [8].

Conclusions
The base-promoted depolymerization of heparin benzyl ester can be conducted using a flow reactor. The conditions have been developed to produce materials that are similar to enoxaparin by NMR. Further work should be directed at elaborating the flow scheme to include refining of the crude material by a flow method for desalting and isolating the product.

Patents
The authors have filed a provisional patent application based on this work.

Supplementary Materials:
The following supporting information can be downloaded at: www.mdpi.com/xxx/s1, NMR spectra of enoxaparin USP and selected entries from Table 1.