Comparison of Low-Molecular-Weight Heparins Prepared FromBovine Lung Heparin and Porcine Intestine Heparin
Yudong Guan 1, Xiaohui Xu 1, Xinyue Liu 1, 2, Anran Sheng 1, Lan Jin 1,Robert J. Linhardt 2, 3, 4, 5, Lianli Chi 1, *1 National Glycoengineering Research Center, Shandong Provincial Key Laboratory of Carbohydrate Chemistry and Glycobiology, Shandong University, Jinan,Shandong, China2 Department of Chemistry and Chemical Biology, Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, New York121803 Department of Chemical and Biological Engineering, Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, NewYork 121804 Department of Biology, Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, New York 121805 Department of Biomedical Engineering, Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, New York 12180
a r t i c l e i n f o
Article history:Received 9 February 2016Revised 12 March 2016Accepted 30 March 2016
Currently porcine intestine is the only approved source for producing pharmaceutical heparin in mostcountries. Enoxaparin, prepared by benzylation and alkaline depolymerization from porcine intestineheparin, is prevalent in the anticoagulant drug market. It is predicted that porcine intestine heparinederived enoxaparin (PIE) will encounter shortage, and expanding its production from heparins obtainedfromother animal tissuesmay, therefore, be inevitable. Bovine lungheparin is a potential alternative sourcefor producing enoxaparin. Critical processing parameters for producing bovine lung heparinederivedenoxaparin (BLE) are discussed. Three batches of BLEs were prepared and their detailed structures werecompared with PIEs using modern analytical techniques, including disaccharide composition, intact chainmapping by liquid chromatography-mass spectrometry and 2-dimensional nuclear magnetic resonancespectroscopy. The results suggested that the differences between PIEs and BLEs mainly result from N-acetylationdifferencesderived fromtheparentheparins. In addition, bioactivities ofBLEswere about 70%ofPIEs based on anti-factor IIa and Xa chromogenic assays. We conclude that BLE has the potential to bedeveloped as an analogue of PIE, although some challenges still remain.
Heparin has been used as an anticoagulant drug to treatthrombosis for about 80 years.1 It was originally isolated from dogliver and demonstrated to possess anticoagulant activity in 1916.2
During the 1930s, heparin was successfully prepared from bovinelung, and heparin from this source was later developed as a phar-maceutical product in the United States.3,4 The bovine lung was the
primary source of heparin until the end of 1950s, when porcineintestinal mucosa became the preferred alternative due to a simplerextraction process and a higher yield. In 1986, the first bovinespongiform encephalopathy case was reported in Europe, andcertain bovine tissues were consequently banned for food andhuman consumption in many countries.5,6 The global productionand utilization of heparin from cattle decreased significantly.Currently, the only approved animal species for manufacturingpharmaceutical heparin in the United States is pig. However, thereare potential risks with using porcine intestinal mucosa as the solesource of heparin. First, if a major outbreak of a pig infectious dis-ease occurs in the future, heparin would be in severe shortage.Second, China accounts for more than 50% of pig livestock andexports most raw heparin around the world. The over-sulfatedchondroitin sulfate contamination of heparin in China caused theheparin crisis in 2008.7 Third, due to the increasing demand fromdeveloping countries, heparin will eventually be in shortage if noother alternative anticoagulant agents emerge. Thus, the
Abbreviations used: ATIII, antithrombin III; BLE, bovine lung heparinederivedenoxaparin; ESI, electrospray ionization; GlcA, glucuronic acid; GlcN, glucosamine;GPC, gel permeation chromatography; HexA, hexuronic acid; HILIC, hydrophilicinteraction chromatography; HSQC, heteronuclear single quantum coherence; IdoA,iduronic acid; LMWH, low-molecular-weight heparin; MALLS, multi-angle laserlight scattering; MWCO, molecular-weight cutoff; TIC, total ion chromatogram; PIE,porcine intestine heparinederived enoxaparin.* Correspondence to: Lianli Chi (Telephone: þ86-531-88363200; Fax: þ86-531-
reintroduction of bovine heparin to the United States market iscurrently under active investigation.8
Heparin is a linear polysaccharide composed of a repeatingdisaccharide building block of alternating b-1,4-linked hexuronicacid (HexA) and glucosamine residue (GlcN). The HexA can beeither b-D-glucuronic acid (GlcA) or a-L-iduronic acid (IdoA), atwhich the C2 position can be substituted by an O-sulfo group. TheGlcN may be modified by an N-acetyl group (GlcNAc), an N-sulfogroup (GlcNS), or can be unsubstituted, whereas O-sulfo groupsubstitution can occur at its C3 and/or C6 positions (Fig. 1a).9,10 Apentasaccharide sequence of GlcNAc/NS(6S)-GlcA-GlcNS(3S,6S)-IdoA(2S)-GlcNS(6S) (Fig. 1b) is the structural motif for heparinthat specifically binds to antithrombin III (ATIII) and inactivates theblood clotting process.11 Heparin is polydisperse with a molecularweight (MW) range from 5000 to 40,000 Da, in which most com-ponents have a MWof 12,000-25,000 Da.12 The heparins producedfrom different animal species and organs usually possess variablestructural characteristics, physicochemical properties, and biolog-ical properties, such as MW, disaccharide composition, oligosac-charide sequence, anti-factor IIa, and anti-factor Xa activities.13
Although unfractionated heparin is directly used as a drug, over60% of heparin is used as the starting material to produce low-molecular-weight heparins (LMWHs), which demonstrateimproved bioavailability, longer half-life, and more controllableanticoagulation.14,15 Enoxaparin, the most widely used LMWH isproduced through esterification followed by alkaline depolymer-ization.16 The weight-averaged MW (Mw) of enoxaparin rangesfrom 3800 to 5000 Da. It inherits some structural characteristicsfrom the parent heparin, including the natural disaccharide build-ing blocks and oligosaccharide sequences. The terminal structure ofenoxaparin is modified by the chemical process used in its prepa-ration. Unsaturated uronate residues are commonly found at thenonreducing end of enoxaparin chains, and 1,6-anhydro groups arefound at 15%-25% of the reducing ends of enoxaparin chains(Fig. 1a). Structural characterization of a LMWH and its parentheparin is fundamentally important for evaluating the efficacy and
safety of these drugs. It also provides critical data for the devel-opment and approval of alternative anticoagulant drugs.17 Modernanalytical techniques, such as top-down and bottom-up massspectrometric chain mapping approaches, 2-dimensional nuclearmagnetic resonance (NMR) methods and novel liquid chromato-graphic methods, provide powerful tools for determining the finestructure of these complicated polysaccharides.18-23
In this study, we prepared enoxaparin-like LWMHs from bovinelung heparin to evaluate as a potential analogue of porcine intestineheparinederived enoxaparin (PIE). Key chemical reaction param-eters were controlled to make bovine enoxaparin with MWmeeting the requirements of the United States Pharmacopeia (USP).The structural and biological properties between bovine- andporcine-derived enoxaparins were then examined using severalstate-of-the-art analytical techniques.
Materials and Methods
Materials
Porcine intestine heparin and bovine lung heparin were giftsfrom Innokare Bio-pharmaceutical Tech Co., Ltd (Suzhou, China).Enoxaparin reference standard (PIE-S) and heparin referencestandard were obtained from the USP Convention (Rockville, MD).Benzethonium chloride, dichloromethane, benzyl chloride, meth-anol, and hydrochloric acid were purchased from J & K ChemicalTechnology (Beijing, China). Sodium acetate and sodium hydroxidewere purchased fromAmersco (Solon, Ohio). Heparinase I, II, III andreagents for bioactivity analysis, including antithrombin, factor IIa,factor Xa, chromogenic substrate S-2238, S-2222, and S-2765 werepurchased from Adhoc International Technologies (Beijing, China).
Preparation of Enoxaparin Using Porcine Intestine Heparin
Two batches of LMWH samples (PIE-1 and PIE-2) were preparedusing porcine intestine heparin as starting materials according to
Figure 1. Structures of heparin and enoxaparin. (a) The structure modification from heparin to enoxaparin. (b) The ATIII-binding pentasaccharide sequence of heparin andenoxaparin.
Y. Guan et al. / Journal of Pharmaceutical Sciences 105 (2016) 1843-18501844
the procedures described by Roger Debrie.24 Briefly, heparin wasadded to benzethonium chloride solution to form the quaternaryammonium salt. The heparinebenzethonium salt was precipitatedout, collected, and dried. Dichloromethane was used to redissolvethe resulting dry powder. Then benzyl chloride was added, and thereaction was incubated at 40"C for 12 h. The heparin benzyl esterwas formed and collected by centrifugation. The depolymerizationwas performed by incubating 1 g of heparin benzyl ester with 25mL of 4.0 g/L sodium hydroxide solution. The reaction temperaturewas controlled at 55"C in a water bath for 2 h. Hydrochloric acidwas used to neutralize the excess sodium hydroxide after the re-action was completed. The final enoxaparin product was precipi-tated using methanol, dialyzed against dialysis membrane withmolecular weight cutoff 1 kDa, and lyophilized.
Preparation of Enoxaparin Using Bovine Lung Heparin
A total of 5 batches of LMWH samples were prepared usingbovine lung heparin as the starting material. One batch (BLE-a) wasprepared using the same process as PIEs. One batch (BLE-b) wasprepared using modified processing conditions. The amount ofalkaline was reduced to 3.5 g/L, and the temperaturewas decreasedto 50"C in the depolymerization step. Three batches (BLE-1, 2, and3) were prepared using the final processing conditions. The amountof alkaline, the depolymerization temperature, and the reactiontime were adjusted to 3.5 g/L, 50"C, and 6 h, respectively. The keyprocessing parameters for preparing all the LMWH samples aresummarized in Table 1.
Gel Permeation Chromatography (GPC) Analysis
Two different types of GPC analysis were used to characterizethe MWs of both starting heparins and all LMWH samples. GPCmultiangle laser light scattering (MALLS) analysis was performedon a Waters 515 HPLC (Milford, MA) equipped with a Wyatt DAWNHELEOS II MALLS detector and a Wyatt Optilab rEX refractive indexdetector (Santa Barbara, CA). The column was a Shodex OHpak SB-803 HQ GPC column (6 mm, 8.0 # 300 mm; Showa Denko, Tokyo,Japan). The isocratic mobile phase was 0.2-M NaNO3 and 0.2-g/LNaN3 in water and the flow rate was 0.5 mL/min. The MWs werecalculated using the ASTRA software (Wyatt Technology, SantaBarbara, CA).
The MW distribution profiles of all samples were also acquiredon a Shimadzu LC-20A HPLC system equipped with a RID-10Arefractive index detector. The column was a SuperdexTM Peptide10/300 GL column (13 mm, 10.0 # 310 mm; GE Healthcare, Uppsala,Sweden) and the mobile phase was 0.2 M NH4HCO3 in water. Theflow rate was set as 0.4 mL/min.
Analysis of Disaccharide Composition and 1,6-Anhydro Derivatives
Quantitative analysis, of the disaccharide composition and the1,6-anhydro derivatives of LMWHs, was performed by strong anionexchange HPLC after exhaustive enzymatic digestion. The digestionconditions were optimized using PIE-S. According to the USPenoxaparin monograph, to reach exhaustive enzymatic digestion ofenoxaparin, the ratio of the peak area of 1,6-anhydro DIS-IS to 1,6-anhydro DIS should not be more than 1.15. Briefly, 5 mL of 20 mg/mLsample solution was mixed with 17.5 mL of digestion buffer con-taining 2-mM monobasic potassium phosphate and 3-mM bovineserum albumin (pH 7.0). A 25-mL cocktail of heparinases I, II, and III(3.5 mIU each) cocktail was then added and the reaction wasincubated at 25"C for 48 h. The disaccharides and 1,6-anhydroderivatives were recovered by using Millipore ultrafiltration filterunits (molecular weight cutoff 3 kDa), lyophilized, reconstitutedusing 50 mL water, and mixed with 10 mL of freshly prepared30 mg/mL sodium borohydride solution. The reaction was incubatedat room temperature for 4 h.
The HPLC analysis was performed on a Shimadzu LC-20A sys-tem. The digested and reduced samples were separated on aSpherisorb® SAX column (5 mm, 4.6 # 250 mm). Mobile phase Awas 2-mM monobasic sodium phosphate (pH 3.0) and mobilephase B was 1 M sodium perchlorate in mobile phase A (pH 3.0).The flow rate was 0.45 mL/min with a step gradient starting from98% A to 92% A in 20 min, 92% A to 80% A from 21 min to 50 min,80% A to 60% A from 51 to 60 min, and back to 98% A from 61min to80 min. The UV detection wavelength was at 234 nm.
The peaks were identified based on their retention times andtheir areas were integrated. The total molar percentage of three 1,6-anhydro derivatives were calculated using,
% 1;6$ anhydro ¼ 100#X
Ai #Mw
.XðMWx # AxÞ
where Ai is the area of the 1,6-anhydro peak i; MWx and Ax are theMW and the area corresponding to the specified peak X,respectively.
NMR Analysis
1H-NMR and heteronuclear single quantum coherence (HSQC)-NMR spectra of LMWH samples were acquired on a Bruker 600MHz UltraShieldTM 600 PLUS spectrometer equipped with a
Table 1Adjustment of Key Processing Parameters for Preparing PIEs and BLEs
Sample Alkaline Depolymerization Parameters Analysis Result
AlkalineConcentration (g/L)
Temperature ("C) Time (h) MW 1,6-AnhydroDerivative
PIE-1 4.0 55 2 Ya YPIE-2 4.0 55 2 Y YBLE-a 4.0 55 2 Nb Not testedBLE-b 3.5 50 2 Y NBLE-1 3.5 50 6 Y YBLE-2 3.5 50 6 Y YBLE-3 3.5 50 6 Y Y
a Y, within the scope of USP enoxaparin monograph.b N, does not meet the definition of USP enoxaparin monograph.
Table 2MWs and MW Distributions of Starting Heparins and LMWH Products
a PIEavg, averaged from PIE-S, PIE-1, and PIE-2.b BLEavg, averaged from BLE-1, BLE-2, and BLE-3.
Y. Guan et al. / Journal of Pharmaceutical Sciences 105 (2016) 1843-1850 1845
cryoprobe at 298 K (Bruker BioSpin, Billerica, MA). The test sampleswere dissolved in D2O (99.9%) and lyophilized once for solventexchange. Then the samples were redissolved at the concentrationof 90mg/600 mL in D2O for NMR analysis. Standard pulse sequencesof 1H and HSQC spectra were used. All data were processed usingBruker TopSpin, version 3.2 software.
Intact Chain Mapping Analysis
This intact chain mapping analysis was performed using hy-drophilic interaction chromatography (HILIC)-electrospray ioniza-tion (ESI)-MS approach.18 A Luna HILIC column (200 A, 2.0 # 150mm; Phenomenex, Torrance, CA) was used. Mobile phase A was 5-mM ammonium acetate in water and mobile phase B was 5-mMammonium acetate in 98% acetonitrile. LMWH samples (8 mL at 1mg/mL) were injected, and the flow rate was 150 mL/min, with alinear gradient from 10% A to 35% A over 40 min. An LTQ-OrbitrapXL FT-MS (Thermo Fisher Scientific, San Jose, CA) was used to re-cord the mass spectra in the negative-ion mode. The parameterswere set as following: spray voltage at 4.2 kV; capillary voltage
at $40 V; capillary temperature at 275"C; tube lens voltage at $50V; sheath flow rate at 20 arbitrary units; and auxiliary gas flow rateat 5 arbitrary units. All mass spectra were acquired at a mass rangebetween 400 and 2000 Da.
Anti-factor IIa and Anti-factor Xa Activities
The bioactivities of starting heparins and LMWH products weredetermined using chromogenic assays.25,26 For porcine intestineand bovine lung heparin samples, the USP heparin referencestandard was used to make standard solutions at 4 concentrationsin buffer containing 50-mM Tris, 7.5-mM EDTA, 175-mM NaCl, and0.1% polyethylene glycol 6000 (pH 8.4). Antithrombin solution wasadded to the standard solution or test solution and incubated for4 min. For anti-factor IIa assay, substrate S-2238 was added and theincubation time was 4 min, whereas for anti-factor Xa assay, sub-strate S-2222was used, and the incubation timewas 12min. All thereagents were prewarmed to 37"C and all reactions were kept atthe same temperature. Acetic acid solution was used to quench thereactions, and UV absorbance was measured at 405 nm using amicroplate reader. For LMWH samples, the buffer containing 50-mM Tris, 150-mM NaCl, and 0.1% polyethylene glycol 6000 (pH7.4) was used to dissolve enoxaparin standard and test samples. Theincubation time of antithrombin solution with test samples was1 min. Substrate S-2238 was used in the anti-factor IIa assay,whereas substrate S-2765was used in the anti-factor Xa assay. Bothincubation times were 4 min. All other experimental procedureswere the same as heparin. The bioactivities of all the samples werecalculated based on the parallel-line models.
Results and Discussion
Process Adjustment for Preparing BLEs
According to the USP monograph, enoxaparin possesses an Mwof 4500 Da, the range being between 3800 and 5000 Da. About 20%
Figure 2. GPC analysis of LMWHs. The smaller oligosaccharides, including degree ofpolymerization (dp)4, dp6, dp8, and dp10, were well separated and marked witharrows.
Table 3Disaccharide Composition and 1,6-Anhydro Derivative Analysis of LMWH Samples
a The abbreviations are DIVA, DUA-GlcNAc; DIVS, DUA-GlcNS; DIIA, DUA-GlcNAc6S; DIIIA, DUA2S-GlcNAc; DIIS, DUA-GlcNS6S; DIIIS, DUA2S-GlcNS; DIA, DUA2S-GlcNAc6Sand DIS, DUA2S-GlcNS6S; DIIA-IVSglu, DUA-GlcNAc6S-GlcA-GlcNS3S; and DIIA-IISglu, DUA-GlcNAc6S-GlcA-GlcNS3S6S.
b PIEavg, averaged from PIE-S, PIE-1, and PIE-2.c BLEavg, averaged from BLE-1, BLE-2, and BLE-3.
Y. Guan et al. / Journal of Pharmaceutical Sciences 105 (2016) 1843-18501846
of the enoxaparin chains contain a 1,6-anhydro derivative on thereducing end, the range being between 15% and 25%. We tried toprepare BLEs using the benzyl esterification and alkaline depoly-merization reactions, with the 2 key structural characteristicsmeeting the USPmonograph description. As shown in Table 1, PIE-1and PIE-2 were prepared from porcine intestine heparin. Threecritical processing parameters, the alkaline concentration, the re-action temperature, and the reaction time, were optimized to4.0 g/L, 55"C, and 2 h, respectively. BLE-a was prepared using thesame processing parameters. Its Mw was below the lower limit ofenoxaparin, which is expected because the Mw of bovine lungheparin is significantly lower than that of porcine intestine heparin.BLE-b was prepared by reducing the alkaline concentration to3.5 g/L and reaction temperature to 50"C, whereas the reactiontime remained unchanged. The Mw of BLE-b met the description ofUSP monograph. However, its content of 1,6-anhydro derivativeswas below 15%. To increase the amount of 1,6-anhydro derivatives,the reaction time was adjusted to 6 h. With the final modifiedprocess, BLE-1, 2, and 3 were prepared. Their Mw and 1,6-anhydroderivatives were within the ranges described in the USP enox-aparin monograph. The structure and bioactivity comparison be-tween BLEs and PIEs were made using BLE-1, 2, and 3 against PIE-S,PIE-1, and PIE-2.
MW and MW Distributions of Heparin and Enoxaparin Samples
Two types of GPC were used to analyze the heparin and LMWHsamples. GPC-MALLS with the Shodex columnwas used to generatetheMw and number-averageMW (Mn). All MWs are summarized inTable 2. Porcine intestine heparin had larger Mw and Mn thanbovine lung heparin, but bovine lung heparin was more poly-disperse. In addition to an Mw range between 3800 to 5000 Da, theUSP monograph also requires 12.0%-20.0% components of enox-aparin to have a MW less than 2000 Da and 68.0%-82.0% compo-nents with a MW between 2000 and 8000 Da. The Mw and MWdistribution of 2 lots of PIEs met the USP monograph requirementsand were very close to PIE-S. For bovine lung heparin, 4 lots ofLWMH products were able to meet the USP requirements whenmilder depolymerization conditions were applied.
GPC with Superdex columnwas used to provide better-resolvedoligosaccharide profiles of LMWH samples. As shown in Figure 2,oligosaccharides with sizes up to decasaccharides were resolved inthe chromatography. The oligosaccharide distribution of PIE-1 wasvery similar to the PIE-S. BLE-a was obviously over depolymerized.The profiles of BLE-b and BLE-1 were very similar to one anotherbut slightly different from that of PIE-S. The centers of tetra-saccharides, hexasaccharides, octasaccharides, and deca-saccharides peaks are marked with arrows in Figure 2. Theoligosaccharide components in BLEs had slightly larger MWs thantheir corresponding components in PIEs, most likely due to thedifferent sulfation patterns between bovine lung heparin andporcine intestine heparin.
Determination of Disaccharide Composition and 1,6-AnhydroDerivatives
The LMWHs were digested to basic building blocks using acocktail of heparinase I, II, and III. Disaccharides with underivatized
Figure 3. The 1H-NMR and 2D-HSQC-NMR spectra of PIE-S and BLE-1. (a) Peaks in the1H-NMR spectra were assigned: a, H4 DU2S; b, H1 ANS6X-(G); c, H1 A3S,3SNS; d, H1 ANS6X-(I2S) and H1 ANAc6X-(G); e, H1 I2S; f, H1 I-(ANS6S); g, H5 I2S; h, H1 G; i, H2 I2S; j, H6 ANS6S;k, H3 I2S; l, H4 I2S; m, H6, ANS; n, H4 ANS6S; o, H3 ANS6X; p, H2 ANS6X; and q, acetyl CH3(abbreviations are: DU, unsaturated uronic acid; A, GlcN; I, IdoA; G, GlcA; 1,6-an.A,
1,6-anhydro glucopyranose and 1,6-an.M, 1,6-anhydro mannopyranose). (b) The HSQCspectrum of PIE-S. (c) The HSQC spectrum of BLE-1. Three regions in the HSQC spectrawere zoomed in: (i) cross peaks region of 1,6-anhydro derivative in PIE-S; (ii) crosspeaks region of 1,6-anhydro derivative in BLE-1; and (iii) region of HexA C1H1 in BLE-1.
Y. Guan et al. / Journal of Pharmaceutical Sciences 105 (2016) 1843-1850 1847
reducing ends were reduced to alditols by sodium borohydride,whereas the 1,6-anhydro derivatives remained unchanged. Thepurpose of the reduction step was to allow the separation of DIIA
and 1,6-anhydro DIIS on the Spherisorb SAX column, which werepoorly resolved before reduction reaction. As summarized inTable 3, 8 natural heparin disaccharides and 3 1,6-anhydro
Figure 4. The HILIC-ESI-MS chain mapping analysis. (a) The TICs of PIE-S, PIE-1, and BLE-1. The TIC is labeled with the oligosaccharides corresponding to the peak time. (b)Comparison of individual oligosaccharides identified in PIEs and BLEs. The structure of oligosaccharides was represented as (DHexA, HexA, GlcN, 1,6-Anhydro, Ac, SO3). The numberswere the composition of each monosaccharide residue or substitution group.
Y. Guan et al. / Journal of Pharmaceutical Sciences 105 (2016) 1843-18501848
disaccharides or tetrasaccharide were identified. The LMWHsshowed similar disaccharide compositions within species, whereasthe differences between PIEs and BLEs were obvious. LMWHsinherited the disaccharide compositions from their parent hepa-rins. Bovine lung heparin contains more N-sulfo groups than doesporcine intestine heparin,13 so the percentage of DIS in BLEs wassignificantly higher than that in PIEs. The total 1,6-anhydro de-rivatives in the 3 batches of PIEs and the 3 batches of BLEs were allin the range of between 15% and 25%, which is within the definitionof the USP monograph. Two 3-O-sulfo group containing tetra-saccharides, DIIA-IVSglu, and DIIA-IISglu were also detected. Theywere derived from the ATIII binding pentasaccharide as the 3-O-sulfated GlcNS residue is resistant to the heparinases. The amountof these 2 tetrasaccharides in PIEs is significantly higher than thatin BLEs (Table 3), suggesting that BLEs contain less ATIII-bindingsites than do PIEs.
NMR Spectroscopy
BLE-1 and PIE-S were analyzed using 1H-NMR and two-dimensional HSQC-NMR to demonstrate the detailed structuralproperties of PIE and BLE. According to the 1H-NMR spectra(Fig. 3a), the 2 samples had similar structures but displayed severalsubtle differences. Clearly, there was more GlcNAc in PIE than inBLE, which was demonstrated by the relative intensity of the peakat 1.96 ppm (peak q) corresponding to the N-acetyl protons of theGlcNAc residue. In the HSQC spectra (Figs. 3b and 3c), BLE-1 againexhibited overall similar structural properties compared with PIE,but with a higher level of sulfation. More specifically, both Glc6OHand GlcA were less in BLE-1 compared to that in PIE-S. In addition,the 1,6-anhydro structures mainly appeared to be 1,6-anhydroglucopyranose in BLE-1, whereas in PIE, the amounts of 1,6-anhydro glucopyranose and 1,6-anhydro mannopyranose wereapproximately equal. The ratios of HexA epimerization were alsodifferent between PIE and BLE. The cross-peak signal correspond-ing to the GlcA was observed in the HSQC spectrum of PIE, but itwas not visible even in the magnified spectrum of BLE.
Fingerprinting Profiles of PIE and BLE by HILIC-ESI-MS
HILIC-ESI-MS provides high-resolution fingerprint profiles forLMWHs. The oligosaccharides were separated on the HPLC andidentified by the on-line ESI-MS. The representative total ion chro-matograms (TICs) of PIE-S, PIE-1, and BLE-1 are shown in Figure 4a.The TIC of PIE-1 is almost identical to the standard, whereas the TICof BLE-1 is clearly different. Over 50 oligosaccharide species wereidentified from the LC-MS analysis, and their structures and relativeintensities are summarized in Figure 4b. The chain length, degree ofsulfation, degree of N-acetylation, and 1,6-anhydro groups at thereducing endwere determined based on their high-resolutionmassspectra. Some oligosaccharide species contained saturated nonre-ducing ends similar to the original nonreducing ends of parentheparins. ThediscrepancybetweenBLEs andPIEswasmainlycausedby the different degrees of N-acetylation. For example, the relativeintensity of hexasaccharide (1, 2, 3, 0, 0, 9) containing no N-acetylgroupwas significantlyhigher inBLEs than that in PIEs,whereas PIEscontained more hexasaccharide (1, 2, 3, 0, 1, 7), which had one N-acetylation. This result is consistent with both the disaccharidecompositional analysis and NMR analysis.
Bioactivity Assays
The bioactivities of starting heparin materials were measuredfirst. The anti-factor IIa and anti-factor Xa activities of porcine in-testine heparin were measured as 198 IU/mg and 185 IU/mg,
respectively. Bovine lung heparin had lower anti-factor IIa and Xaactivities, measured as 131 IU/mg and 121 IU/mg, respectively. Thebioactivities of LMWHs produced from these 2 heparins were thenmeasured. Accordingly, BLEs showed approximately 30% loweraverage anti-factor IIa and anti-factor Xa activities compared toPIEs. The averaged anti-factor IIa and anti-factor Xa activities of PIE-1 and PIE-2 were 25 IU/mg and 103 IU/mg, respectively. The aver-aged anti-factor IIa and anti-factor Xa activities of BLE-1, 2, and 3were 19 IU/mg and 76 IU/mg, respectively. They were unable tomeet the lower limit of USP enoxaparin monograph (20.0 IU/mg foranti-factor IIa activity and 90.0 IU/mg for anti-factor Xa activity).
Conclusions
The demand for LMWHs is increasing each year, and the even-tual shortage of LMWHs produced solely from porcine intestine isforeseeable. It is necessary to explore alternative sources forLMWHs. Bovine lung heparin had been used as a clinical antico-agulant drug for a long period. Its efficacy and safety had beenapproved and would be subject to future approval if the bovinespongiform encephalopathy contamination could be prevented. Byproperly adjusted processing parameters, BLEs with MW and 1,6-anhydro derivative passing the requirements of USP enoxaparinmonograph were successfully prepared. Further structural com-parison using disaccharide composition, chain mapping, and NMRanalysis, showed that the main difference between BLEs and PIEs istheir differing degrees of N-acetylation, inherited from the parentheparin source material. In addition, BLEs contained significantlyless 1,6-anhydro mannopyranose than PIEs. These results suggestthat it may not be possible to prepare enoxaparin exactly matchingthe structure of PIEs, if bovine lung heparin is used as the startingmaterial, making it difficult to meet the sameness criteria by theUnited States Food and Drug Administration.17 Furthermore, BLEscontain less ATIII-binding sites than PIEs. As a result, both anti-factor IIa and anti-factor Xa activities of BLEs are only about 70%of PIEs, a greater mass of drug will be required. The present studyprovides a better understanding of the potential and challenges ofusing bovine lung heparin as an alternative source material forproducing an enoxaparin-like LMWH.
Acknowledgment
This work was supported by grants from the National BasicResearch Program of China (973 Program) (2012CB822102) and theNational Natural Science Foundation of China (21472115,21302113).
References
1. Capila I, Linhardt RJ. Heparin-protein interactions. Angew Chem Int Ed Engl.2002;41:390-412.
2. Mclean J. The thromboplastic action of cephalin. Am J Phys. 1916;41:250-257.3. Charles AF, Scott DA. Studies on heparin I. The preparation of heparin. J Biol
Chem. 1933;102:425-429.4. Charles AF, Scott DA. Studies on heparin II. Heparin in various tissues. J Biol
Chem. 1933;102:431-435.5. Nathanson N, Wilesmith J, Griot C. Bovine spongiform encephalopathy (BSE):
causes and consequences of a common source epidemic. Am J Epidemiol.1997;145:959-969.
6. Schonberger LB. New variant Creutzfeldt-Jakob disease and bovine spongiformencephalopathy. Infect Dis Clin N Am. 1998;12:111-121.
7. Liu H, Zhang Z, Linhardt RJ. Lessons learned from the contamination of heparin.Nat Prod Rep. 2009;26:313-321.
8. Zhang F, Yang B, Ly M, et al. Structural characterization of heparins fromdifferent commercial sources. Anal Bioanal Chem. 2011;401:2793-2803.
9. Rabenstein DL. Heparin and heparan sulfate: structure and function. Nat ProdRep. 2002;19:312-331.
10. Casu B. Structure of heparin and heparin fragments. Ann N Y Acad Sci.1989;556:1-17.
Y. Guan et al. / Journal of Pharmaceutical Sciences 105 (2016) 1843-1850 1849
11. Lindahl U, Thunberg L, Backstrom G, Riesenfeld J, Nordling K, Bjork I. Extensionand structural variability of the antithrombin-binding sequence in heparin.J Biol Chem. 1984;259:12368-12376.
12. Linhardt RJ. 2003 Claude S. Hudson award address in carbohydrate chemistry.Heparin: structure and activity. J Med Chem. 2003;46:2551-2564.
13. Fu L, Li G, Yang B, et al. Structural characterization of pharmaceutical heparinsprepared from different animal tissues. J Pharm Sci. 2013;102:1447-1457.
14. Gray E, Mulloy B, Barrowcliffe TW. Heparin and low-molecular-weight heparin.Thromb Haemost. 2008;99:807-818.
15. Antman EM. The search for replacements for unfractionated heparin. Circula-tion. 2001;103:2310-2314.
16. Linhardt RJ, Gunay NS. Production and chemical processing of low molecularweight heparins. Semin Thromb Hemost. 1999;25:5-16.
17. Lee S, Raw A, Yu L, et al. Scientific considerations in the review and approval ofgeneric enoxaparin in the United States. Nat Biotechnol. 2013;31:220-226.
18. Li L, Zhang F, Zaia J, Linhardt RJ. Top-down approach for the direct character-ization of low molecular weight heparins using LC-FT-MS. Anal Chem. 2012;84:8822-8829.
19. Li G, Steppich J, Wang Z, et al. Bottom-up low molecular weight heparinanalysis using liquid chromatography-Fourier transformmass spectrometry forextensive characterization. Anal Chem. 2014;86:6626-6632.
20. Mourier PAJ, Viskov C. Chromatographic analysis and sequencing approach ofheparin oligosaccharides using cetyltrimethylammonium dynamically coatedstationary phases. Anal Biochem. 2004;332:299-313.
21. Li D, Chi L, Jin L, et al. Mapping of low molecular weight heparins usingreversed phase ion pair liquid chromatographyemass spectrometry. CarbohydrPolym. 2014;99:339-344.
22. Wang Z, Li D, Sun X, Bai X, Jin L, Chi L. Liquid chromatographyediode arraydetectionemass spectrometry for compositional analysis of low molecularweight heparins. Anal Biochem. 2014;451:35-41.
23. Guerrini M, Guglieri S, Naggi A, Sasisekharan R, Torri G. Low molecular weightheparins: structural differentiation by bidimensional nuclear magnetic reso-nance spectroscopy. Semin Thromb Hemost. 2007;33:478-487.
24. Debrie R. Mixtures of Particular LMW Heparinic Polysaccharides for the Prophy-laxis/Treatment of Acute Thrombotic Events. United States patent US 5,389,618 A.1995 Feb 14.
25. Suzuki T, Ishii-Watabe A, Hashii N, et al. The establishment and validation ofefficient assays for anti-IIa and anti-Xa activities of heparin sodium and hep-arin calcium. Biologicals. 2013;41:415-423.
26. Martinez C, Savadogo A, Agut C, Anger P. Reproducibility of the anti-factor Xaand anti-factor IIa assays applied to enoxaparin solution. J Pharm Biomed Anal.2013;81-82:138-145.
Y. Guan et al. / Journal of Pharmaceutical Sciences 105 (2016) 1843-18501850
