Identification of a Novel Structure in Heparin Generated by Sequential

May 23, 2012 - Copyright © 2012 American Chemical Society. *Phone: 1-617-395-5100. Fax: 1-617-621-0431. E-mail: [email protected]...
3 downloads 0 Views 405KB Size
Article pubs.acs.org/ac

Identification of a Novel Structure in Heparin Generated by Sequential Oxidative−Reductive Treatment Daniela Beccati, Sucharita Roy, Miroslaw Lech, Jennifer Ozug, John Schaeck, Nur S. Gunay, Radouane Zouaoui, Ishan Capila, and Ganesh V. Kaundinya*,† Momenta Pharmaceuticals Inc., 675 West Kendall Street, Cambridge, Massachusetts 02142, United States S Supporting Information *

ABSTRACT: Unfractionated heparin is isolated from animal organs, predominantly porcine intestinal mucosa, and goes through an extensive process of purification before it can be used for pharmaceutical purposes. While the structural microheterogeneity of heparin is predominantly biosynthetically imprinted in the Golgi, subsequent steps involved in the purification and manufacture of commercial heparin can lead to the introduction of additional modifications. Postheparin crisis of 2008, it has become increasingly important to identify what additional structural diversity is introduced as a function of the purification process and thus can be determined as being heparinrelated, as opposed to being an adulterant or contaminant, e.g., oversulfated chondroitin sulfate. Our study focuses on the identification of a previously unreported structure in heparin that arises due to specific steps used in the manufacturing process. This structure was initially observed as a disaccharide peak in a complete enzymatic digest of heparin, but its presence was later identified in the NMR spectra of intact heparin as well. Structural elucidation experiments involved isolation of this structure and analysis based on multidimensional NMR and liquid chromatography coupled with mass spectrometry (LC-MS). Heparin was also subjected to specific chemical reactions to determine which steps in the manufacturing process are responsible for this novel structure. Our results allowed for the definitive assignment of the structure of this novel process-related modification and enabled an identification of the putative steps in the process that give rise to the structure.

H

revise existing guidelines/monographs and implement high resolution analytical tests that could allow for the detection of possible contamination in heparin and effectively ensure drug safety in the future. In addition, efforts were made to minimize the amount of potential impurities like dermatan sulfate, residual proteins, and nucleotides in the large scale production of pharmaceutical-grade heparin. These attempts mainly involved changes in the isolation and purification processes used in the manufacture of heparin and have been reported to affect the composition of pharmaceutical heparin as well.9 As a consequence, it became essential to have a good understanding of the manufacturing process-related heterogeneity introduced in the heparin backbone to be able to assess pharmaceutical heparin in terms of its purity, as well as its structural composition. While performing composition analysis of unfractionated heparin samples by ion-pairing reverse phase HPLC (IPRPHPLC),10 we observed that some lots presented an unidentified peak (referred to as Peak X henceforth) at approximately 34−35 min in the elution profile. In order to ensure that this peak was not due to a contaminant in heparin, a comprehensive set of experiments for identification and structural analyses was conducted. An enzymatic digest of heparin was used to isolate the peak of interest, which was then

eparin is a polydisperse, highly sulfated, linear polysaccharide comprising repeating 1→4 linked hexuronic acid and D-glucosamine units. The major disaccharide sequence of heparin is the trisulfated L-IdoA (2S)-D-GlcNS (6S), which constitutes roughly 70% of heparin from porcine intestinal mucosa. The amino group of the glucosamine residue can be substituted with an acetyl or a sulfate group. Additionally, the 3O and 6-O positions of the glucosamine residues can be either nonsulfated or substituted with a sulfate group. Finally, the uronic acid of each disaccharide repeating unit can be either Liduronic or D-glucuronic acid and may contain a 2-O sulfate group.1−3 The inherent structural microheterogeneity of heparin arises due to the variable substitution of the disaccharide units and is encoded in a nontemplate driven manner during the biosynthesis of heparin.4,5 Pharmaceutical heparin is predominantly derived from porcine intestinal mucosa, initially as crude heparin that is subsequently subjected to numerous isolation and purification steps. These purification steps often use strong basic and oxidative conditions that can give rise to additional structural heterogeneity in the heparin backbone, as has been previously reported.6 In 2008, a series of adverse events, including numerous fatalities, were reported following administration of pharmaceutical heparin. Subsequent investigations confirmed the presence of a contaminant in pharmaceutical heparin that was identified as oversulfated chondroitin sulfate (OSCS).7,8 This event prompted regulatory and standard setting authorities to © 2012 American Chemical Society

Received: March 26, 2012 Accepted: May 11, 2012 Published: May 23, 2012 5091

dx.doi.org/10.1021/ac3007494 | Anal. Chem. 2012, 84, 5091−5096

Analytical Chemistry

Article

analyzed by multidimensional NMR11 and liquid chromatography coupled with mass spectrometry (LC-MS)12 to enable structural elucidation. In parallel, various process conditions used in the manufacture of heparin were evaluated to determine whether any of the performed reactions could generate the unidentified peak. The results of our study demonstrate that this peak is the consequence of a process-related modification of heparin and occurs during a sequential oxidation−reduction process using hydrogen peroxide and sodium sulfite, respectively. The structure of the heparin disaccharide unit that gives rise to this peak has been completely elucidated as a result of our study. The newly identified structure presents characteristic NMR signals that can easily be identified in HSQC spectra of intact heparin samples.

rotating-frame Overhauser enhancement spectroscopy (ROESY) spectra were measured in phase-sensitive mode, for 16−24 scans and 240−256 transients. TOCSY spectra were acquired with 80 ms of MLEV-17 mixing. All experiments were zero filled and apodized prior to Fourier transformation. LC-MS/MS Analyses. An Ultimate 3000 capillary HPLC workstation (Dionex) was used for microseparation in front of the LTQ XL instrument (Thermo Fisher Scientific). A step gradient elution was performed using a binary solvent system composed of water (eluent A) and 70% aqueous methanol (eluent B), both containing 8 mM di-n-dibutylamine acetate as an ion-pairing reagent. A PROTO 200, 5 μm, 200 oA, C18 polymeric silica column (250 mm × 0.3 mm, Higgins Analytical, CA, USA) at 4 μL/min flow rate was used for the separation. To improve the ionization, a postcolumn addition of methanol at 6 μL/min flow rate was performed using a secondary solvent system and a NanoMixer (Upchurch Scientific). The negative-ion mode mass spectra were acquired using an electrospray ionization source (ESI) on the quadrupole ion trap LTQ XL mass spectrometer (Thermo Fisher Scientific) using Xcaliber version 2.0.7 data acquisition software. The nitrogen gas (N2) was used as a desolvation/nebulizer gas. ESI-MS and collision-induced dissociation (CID) MS2 experiments were performed using data dependent triple play acquisition. The MS to MS2 switch criteria was based on the peak intensity. The first scan event was MS scan of full scan type generated by scanning the range of m/z 120−1400, followed by the data dependent ZoomScan and MS2 scan. For MS2 experiments, an isolation window was set at 2 Da and the normalized collision energy at 25. All mass spectra consisted of multiple scans.



EXPERIMENTAL SECTION Materials. Unfractionated heparin samples (UFH, porcine intestinal) were obtained from various heparin suppliers. 3(Trimethylsilyl) propionic acid-d4 sodium salt (TSP) was obtained from Sigma-Aldrich (St. Louis, MO, USA). Deuterium oxide (D2O, 99.9% D or 99.96% D) was purchased from Cambridge Isotope laboratories, Inc. (Andover, MA, USA). Enzymatic Digestion of UFH Samples. Primary Digestion. UFH (100 mg) was digested using heparinase enzyme cocktail at 30 °C for 16 h. Secondary Digestion. The product of the primary digestion was incubated for 6 h at 30 °C simultaneously with 2-Osulfatase and Δ4,5-glycuronidase. On completion of digestion, the sample was frozen and lyophilized. Peak Isolation by Strong Anion Exchange-HPLC. UFH Lot 1 was subjected to secondary enzymatic digestion, lyophilized, reconstituted in Ultra-Pure water, and injected onto a preparative CarboPac PA1 SAX (250 × 22 mm, Dionex) column at ambient temperature. Optimal peak separation was obtained using a NaCl gradient, pH 7.0, at a flow rate of 5 mL/ min. The profile was monitored at 232 nm on an Agilent 1100 HPLC equipped with a diode array detector and fraction collector. Multiple runs were carried out, and Peak X was collected on the basis of a set minimum threshold, desalted on a G10 Sephadex resin (GE Healthcare), lyophilized, and subsequently submitted for LC-MS and NMR analyses. Ion-Pairing Reverse Phase HPLC Analyses. Digested heparin samples were analyzed by IP-RPHPLC using tetra-nbutyl ammonium chloride (TBA) as the ion-pairing reagent in 15% acetonitrile (ACN) and 10 mM Tris buffer at pH 7.0, with a gradient ranging from 0 to 1 M NaCl. The digested samples were separated using an analytical C18 Discovery column (4.6 × 250 mm, Supelco) maintained at 25 °C, at a flow rate of 0.7 mL/min over 130 min of total run time. The profile was monitored by UV absorption at 232 nm. NMR Analyses. Samples for NMR analysis were dissolved in 99.9% or 99.96% D2O and sonicated for 30 s to remove air bubbles. NMR spectra were obtained on a 600 MHz Bruker Avance spectrometer with a 5 mm TXI cryo probe. All spectra were obtained at 298 or 303 K. 1D 1H spectra were acquired for 8−64 scans, with presaturation of the residual water signal and a recycle delay of 10 s. Two-dimensional HSQC spectra were recorded with carbon decoupling during acquisition, for 8−64 scans and 256−320 transients. The polarization transfer delay was set with a 1JC−H coupling value of 155 Hz. For HSQC spectra, the matrix was zero filled to 2 K × 1 K prior to Fourier transformation. 2D total correlation spectroscopy (TOCSY), double-quantum filtered correlation spectroscopy (COSY), and



RESULTS Pharmaceutical grade (USP) heparin lots (Table 1) from different suppliers were digested with a cocktail of heparin Table 1. Description of Samples Used in This Study sample ID UFH lot 1 UFH lot 2 PIH PIHox PIHox‑red

sample description pharmaceutical grade unfractionated heparin sample with a significant amount of Peak X, used for peak isolation pharmaceutical grade unfractionated heparin sample that contains minimal amount of Peak X porcine intestine mucosal heparin sample with a minimal amount of Peak X porcine intestine mucosal heparin oxidized with hydrogen peroxide porcine intestine mucosal heparin oxidized with hydrogen peroxide and reduced with sodium sulfite

lyases (primary digestion), followed by 2-O-sulfatase and Δ4,5glycuronidase treatment (secondary digestion) as described in the Experimental Section. Both digestion mixtures were subsequently analyzed by IP-RPHPLC to monitor and quantify the presence of the peak eluting at 34−35 min, henceforth referred to as Peak X. IP-RPHPLC analyses showed that Peak X was present in varying intensity among different lots (Figure 1). LC-MS analysis of the digested UFH Lot 1 assigned a mass of 561 Da to Peak X. The observed mass could not be assigned to any previously observed disaccharides in the enzymatic depolymerization of UFH, thus suggesting the presence of an unusual structure. 5092

dx.doi.org/10.1021/ac3007494 | Anal. Chem. 2012, 84, 5091−5096

Analytical Chemistry

Article

Peak X sample isolated by SAX-HPLC was further analyzed by 1D and 2D NMR experiments. 1D 1H NMR identified two structures: a trisaccharide of formula HNAc-G-HNS3S6S and a disaccharide containing a N-,6-O sulfated glucosamine residue and an unsaturated residue (Figure 2A). The respective percentage molar ratio between these structures was calculated as approximately 1:3.

Figure 1. IP-RPHPLC analysis of digested UFH lots containing different amounts of Peak X (indicated with ∗). (A) Primary and (B) secondary digestion of UFH Lot 1; (C) primary and (D) secondary digestion of UFH Lot 2.

In order to isolate Peak X for structural characterization, UFH Lot 1 was digested and submitted to secondary digestion before being subjected to fractionation by SAX-HPLC (Figure S1, Supporting Information). Peak X was collected and desalted. Peak collection by SAX-HPLC was preferred over IP-RPHPLC since it did not require extensive clean up steps to remove the residual ion-pairing reagent present in the latter. Fragmentation (MS2) was done on the most abundant doubly charged [M − 2H]2− molecular ion at m/z 279.6. Tandem mass spectrum (MS2) of this doubly charged ion was identical to the tandem mass spectrum of the doubly charged ion observed in UFH Lot 1 (in terms of the observed number of product ions, their masses and relative abundances), thus showing structural equivalence between the peak collected by SAX-HPLC and Peak X present in the UFH Lot 1 (Figure S2, Supporting Information). A singly charged product ion at m/z 301.92 indicated that at least two sulfate groups were present on the glucosamine residue. Additionally, the same peak in combination with the two other singly charged product ions at m/z 203.92 and 222.00 hinted at an unmodified glucosamine residue, with a modification located on the unsaturated uronic acid residue. The sample isolated by SAX-HPLC also contained a secondary species (Mass: 798 Da). This species was assigned a composition corresponding to a trisaccharide of molecular formula H−U−H (where H = hexosamine and U = uronic acid) containing one acetyl and three sulfate groups. It was inferred that this structure did not contain unsaturated residues and thus did not exhibit a UV absorbance at 232 nm.

Figure 2. (A) 1H NMR spectrum of Peak X isolated by SAX-HPLC. Signals assigned to the disaccharide of formula ΔU2X−HNS,6S are indicated. Peaks due to the trisaccharide HNAc−G−HNS3S6S are labeled with ∗. (B) HSQC spectrum of Peak X isolated by SAX-HPLC. Assignments for ΔU2SO3−HNS,6S are indicated close to the relative contours. Additional signals belong to the trisaccharide HNAc−G− HNS3S6S..

2D NMR experiments, i.e., HSQC, COSY, and TOCSY, were performed in order to assign the structure of the disaccharide containing an unsaturated residue. The unsaturated moiety was identified in the HSQC spectrum by two peaks at 5.80/98.2 ppm (H1/C1) and 6.09/110.3 ppm (H4/ C4). H1 showed a COSY correlation with a H2 signal at 3.60 ppm. The strong shift of the H2/C2 signal to 3.60/66.3 ppm compared to H2/C2 of ΔU (3.86/73.2 ppm) and ΔU2S (4.48/ 77.1 ppm) typical of digested heparin samples was consistent with the presence of a carbon−sulfonate (C−SO3−) bond. Structure A shown in Scheme 1 was thus proposed on the basis of both the NMR and MS2 data. The NMR assignment of the disaccharide detected in Peak X is reported in Figure 2B and Table 2. To identify the conditions that would generate a modified uronic acid residue substituted with a SO3− at the H2/C2 position, a heparin sample (PIH) was subjected to oxidation with hydrogen peroxide, under basic conditions. IP-RPHPLC 5093

dx.doi.org/10.1021/ac3007494 | Anal. Chem. 2012, 84, 5091−5096

Analytical Chemistry

Article

Scheme 1. Mechanism and Structure of Heparin Modification Resulting in the Peak X

Table 2. NMR Assignments of the Disaccharide Detected in Peak X (ΔU2SO3-HNS,6S) and of the Modified Uronic Acid Structure Identified in the PIH ox‑red Sample (I2SO3-) H1/C1 H2/C2 H3/C3 H4/C4 H5/C5 H6, H6′/C6

ΔU2SO3-

HNS,6S

I2SO3-

5.80/98.2 3.60/66.3 4.59/62.6 6.09/110.3 N/A N/A

5.46/93.7 3.28/60.2 3.76/71.6 3.86/81.6 4.16/70.5 4.33−4.22/69.5

5.37/n.d. 3.13/66.4 4.32/72.3 4.15/78.7 4.80/72.9

analysis of the oxidized material showed no increase in the intensity of Peak X. 1H NMR and HSQC analysis indicated that the most significant effect of H2O2 treatment under basic conditions was the conversion of I2S into epoxide and galacturonic acid (GalA) residues13 (Figure S3, Supporting Information). The peroxide-oxidized samples were subsequently subjected to reduction with varying concentrations of sodium sulfite. Treatment with 10% (w/w) sodium sulfite caused no significant change in the intensity of Peak X, as determined by IP-RPHPLC analysis. However, treatment with 40% (w/w) sodium sulfite led to a substantial increase in Peak X (Figure S4, Supporting Information). Composition analysis by LC-MS of fully digested PIHox‑red showed a substantial increase in the amount of the peak with mass of 561 Da (Figure S5c, Supporting Information). Tandem mass spectrometry was performed on the precursor ion at m/z 186.98 [M − 3H]3‑ in negative ion mode using LTQ XL instrument (Figure 3B). Similar tandem mass spectrometry was performed on the precursor ion at m/z 191.33 [M-3H]3− of a trisulfated disaccharide with formula ΔU2S−HNS6S (577 Da) (Figure 3A). Precursor ions with negative charges equal to the number of sulfate groups were preferred for the MS analysis, as they produce larger number of fragments (glycosidic-bond and cross-ring cleavages) and the sulfate losses are minimized.14 The doubly charged product ions at m/z 126.92 and 217.92 present in the MS2 spectrum of the triply charged precursor at m/z 191.33 showed the unmodified uronic acid residue. However, both ions were absent in the MS2 spectrum of the triply charged precursor at m/z 186.98. Moreover, the presence of a new doubly charged product ion at m/z 209.92 in this spectrum, in combination with singly charged product ions at m/z 114.92 and 137.92 and a doubly charged product ion at m/ z 168.42 clearly implied a modification of the uronic acid. Product ions were assigned on the basis of cleavage patterns previously observed in HS disaccharides.15,16 Fragmentation

Figure 3. Negative tandem mass spectra of (A) trisulfated disaccharide of formula ΔU2S−HNS6S (577 Da) using the precursor ion at m/z 191.33 [M − 3H]3− and (B) Peak X (561 Da) using the precursor ion at m/z 186.98 [M − 3H]3−.

results of the precursor ion at m/z 186.98 [M − 3H]3− were consistent with a disaccharide containing a SO3− group with a C−S bond at the 2 position of an unsaturated residue and N-,6O disulfated glucosamine. The modification at the 2 position of the unsaturated residue was observed to result in a loss of 16 mass units compared to the trisulfated disaccharide standard, equivalent to one oxygen atom, thus giving more credence to our hypothesis of Structure A. In the 1H NMR and HSQC spectrum of PIHox‑red, a new peak at 3.13/66.4 ppm was detected (Figure 4A). By analysis of COSY and TOCSY experiments, a correlation between the peak at 3.13 ppm and peaks at 5.37 ppm (H1), 4.32 ppm (H3), and 4.15 ppm (H4) was established, indicating that it belonged to an uronic acid residue. The shift of the H2/C2 signal as compared to I2S and I2OH heparin residues (4.37/78.3 ppm and 3.81/71.4 ppm, respectively) suggested a different substitution at the position 2 of the uronic acid. The chemical shift of the H4 proton indicated the presence of an iduronic rather than a glucuronic acid residue. A cross-peak between H2 and a H5 at 4.80 ppm was observed in the ROESY spectrum, allowing assignment of the H5 signals. This ROESY correlation indicated that both H2 and H5 were both on the same side of the molecule, while the substituent at the position 2 was on the opposite side with respect to H5, thus confirming the presence of Structure A (Scheme 1). The NMR assignment performed by COSY, TOCSY, ROESY, and HSQC analysis is reported in Table 2. 5094

dx.doi.org/10.1021/ac3007494 | Anal. Chem. 2012, 84, 5091−5096

Analytical Chemistry

Article

Figure 4. (A) NMR spectra of PIHox‑red and (B) 1H NMR and HSQC spectra of UFH Lot 1. Peaks associated with I2SO3- are indicated with a circle.

units compared to a regular trisulfated disaccharide of formula ΔU2S−HNS,6S. To establish the complete structure of Peak X, the peak was isolated using preparative chromatography as described in the Experimental Section and characterized by tandem mass spectrometry and NMR spectroscopy. Multidimensional NMR analyses (1D and 2D NMR) coupled with MS2 analysis confirmed the presence of an intact 2,6-disulfated glucosamine and assigned the modified residue to a Δ4,5-unsaturated uronic acid substituted at the 2 position with a sulfonic acid. The presence of an unusual carbon−sulfur (C−S) bond explained the resistance of the Δ4,5-unsaturated uronic acid to digestion with 2-O-sulfatase and Δ4,5-glycuronidase. Nevertheless, this residue remained a suitable substrate for heparinases. Concomitantly, to understand the chemical origin of Peak X, a set of reactions were performed in an effort to generate the modification of heparin resulting in Peak X. Since Peak X was detected in commercial UFH lots, it was decided to mimic conditions and reagents currently employed in the process for manufacturing pharmaceutical heparin sodium. On the basis of our previous experience with heparin modifications introduced under potassium permanganate oxidation conditions,6 we hypothesized that other strong oxidizing/reducing conditions employed during the process could result in modified structures leading to generation of Peak X. To test our hypothesis, heparin samples were subjected to oxidative conditions with basic hydrogen peroxide followed by treatment with excess sodium sulfite. These conditions were observed to yield products which exhibited a substantial increase in the area percent of the Peak X as compared to untreated samples. The LC-MS analyses performed on one such compound showed significant enrichment of the peak with mass of 561 Da while the LC-MS/MS analyses confirmed the structure observed in the isolated material. Interestingly, neither basic peroxide nor sulfite treatment independently gave this result. The multidimensional NMR analyses performed on obtained UFH lots detected a peak at 3.13/66.4 ppm in the above-mentioned compound

Commercial samples of unfractionated heparin, whose IPRPHPLC analysis confirmed the presence of Peak X, were analyzed by 1H NMR and HSQC in an attempt to identify the peak associated with the uronic acid structure modified at the 2 position. As indicated in Figure 4B, a signal was detected at 3.13 ppm in the 1H NMR spectrum and at 3.13/66.4 ppm in the HSQC spectrum and was assigned to the latter.



DISCUSSION Characterization of atypical peaks/signals in the analyses of USP grade heparin has increased in importance since the 2008 heparin crisis wherein contaminated heparin samples exhibited distinct signals in 1H NMR spectra and CE/SAX-HPLC profiles due to oversulfated chondroitin sulfate.17 The present study summarizes our study on the isolation and characterization of a unique structure responsible for an unusually intense signal (Peak X) observed in the IP-RPHPLC analyses of multiple heparin samples. A set of experiments were designed, focused on the three main pieces of evidence provided by the IP-RPHPLC-analyses: (a) Peak X was detected in the primary digest; (b) Peak X detected in primary digest exhibited UV absorption at 232 nm; (c) Peak X detected in primary digest was refractory to the secondary digest. Digestion of heparin with heparinase enzymes generates saccharide fragments containing a Δ4,5-unsaturated uronic acid at the nonreducing end, which is responsible for their UV absorbance. Since Peak X was observed to possess UV absorbance, it clearly contained a Δ4,5-unsaturated uronic acid generated by the action of heparinases. The lack of reactivity to both 2-O-sulfatase and Δ4,5-glycuronidase indicated that the Δ4,5-unsaturated uronic acid was unusually modified and therefore was not a good substrate for these enzymes.18,19 Analyses by LC-MS of Peak X exhibited a molecular ion at m/z 560 (M-H)−1, consistent with a molecular formula of C12H19NO18S3. Further tandem mass spectrometry analyses assigned this formula to a disaccharide containing a N,6-Odisulfated glucosamine linked to a monosulfated-Δ4,5-uronic acid with a modification responsible for the loss of 16 mass 5095

dx.doi.org/10.1021/ac3007494 | Anal. Chem. 2012, 84, 5091−5096

Analytical Chemistry



which could be assigned to the H2/C2 of an uronic acid residue with a sulfonic acid substitution. On the basis of the obtained experimental evidence, the mechanism of formation of the residue with a sulfonic acid at the H2/C2 position is proposed (Scheme 1). In fact, the use of sodium sulfite in the analyses and quantification of epoxides has been previously reported.20 The mechanism hypothesizes an initial formation of uronic acid epoxides under basic oxidative conditions (confirmed in Figure S3, Supporting Information), which then undergoes ring-opening when treated with sodium sulfite to generate a modified uronic acid with a hydroxysulfonic acid substitution. In this case, the newly formed signal at 3.13/66.4 ppm observed in the HSQC spectrum can be assigned to the proton at the C-2 position of this modified uronic acid moiety. Finally, it was attempted to identify this modified uronic acid moiety in commercial samples of unfractionated heparin by 1H NMR and HSQC, since these methods do not require enzymatic digestion and can constitute a useful alternative to analysis by IP-RPHPLC. As indicated in Figure 4, a signal assigned to H2 of the modified uronic acid was clearly detected at 3.13 ppm in the 1H NMR spectrum and at 3.13/66.35 ppm (1H/13C) in the HSQC spectrum. These signals can be used as fingerprints to screen for the presence of I2SO3- residues in commercial heparin samples. These results support the proposal that Peak X results from a structural modification of heparin caused by the conditions used in the process for heparin sodium manufacture, specifically base-promoted formation of heparin epoxide, followed by reaction with sodium sulfite to generate a β-hydroxy-sulfonic acid substituted uronic acid residue.

Article

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone: 1-617-395-5100. Fax: 1-617-621-0431. E-mail: [email protected]. Notes

The authors declare no competing financial interest. † Note that Ganesh V. Kaundinya was previously known as Ganesh Venkataraman.



ACKNOWLEDGMENTS D.B. and S.R. contributed equally to this work. The authors would like to acknowledge and thank Dr. Richard Sachleben for valuable input and discussion, Drs. Fei Yu and Jing Wang for assistance with NMR/MS procedures, and Andre Jones and Danica Meine for help with the HPLC analyses. This study was funded by Momenta Pharmaceuticals Inc.



REFERENCES

(1) Linhardt, R. J. J. Med. Chem. 2003, 46, 2551−2564. (2) Venkataraman, G.; Shriver, Z.; Raman, R.; Sasisekharan, R. Science 1999, 286, 537−542. (3) Vives, R. R.; Pye, D. A.; Salmivitra, M.; Hopwood, J. J.; Lindahl, U.; Gallagher, J. T. Biochem. J. 1999, 339, 767−773. (4) Prydz, K.; Dalen, K. T. J. Cell Sci. 2000, 113, 193−205. (5) Sugahara, K.; Kitagawa, H. IUBMB Life 2002, 54, 163−175. (6) Beccati, D.; Roy, S.; Yu, F.; Gunay, N. S.; Capila, I.; Lech, M.; Linhardt, R. J.; Venkataraman, G. Carbohydr. Polym. 2010, 82, 699− 705. (7) Guerrini, M.; Beccati, D.; Shriver, Z.; Naggi, A.; Viswanathan, K.; Bisio, A.; Capila, I.; Lansing, J. C.; Guglieri, S.; et al. Nat. Biotechnol. 2008, 26, 669−675. (8) Kishimoto, T. K.; Viswanathan, K.; Ganguly, T.; Elankumaran, S.; Smith, S.; Pelzer, K.; Lansing, J. C.; Sriranganathan, N.; Zhao, G.; et al. New Eng. J. Med. 2008, 358, 2457−2467. (9) Liverani, L.; Mascellani, G.; Spelta, F. Thromb. Haemostasis 2009, 102, 846−853. (10) Karamanos, N.; Vanky, P.; Tzanakakis, G. N.; Tsegenidis, T.; Hjerpe, A. J. Chromatogr., A 1997, 765, 169−179. (11) Guerrini, M.; Guglieri, S.; Naggi, A.; Sasisekharan, R.; Torri, G. Semin. Thromb. Hemostasis 2007, 33, 478−487. (12) Kuberan, B.; Lech, M.; Zhang, L.; Wu, Z. L.; Beeler, D. L.; Rosenberg, R. D. J. Am. Chem. Soc. 2002, 124, 8707−8718. (13) Rej, R. N.; Perlin, A. S. Carbohydr. Res. 1990, 200, 437−447. (14) Zaia, J.; Costello, C. E. Anal. Chem. 2003, 75, 2445−55. (15) Saad, O. M.; Leary, J. A. J. Am. Soc. Mass Spectrom. 2004, 15, 1274−86. (16) Domon, B.; Costello, C. E. Biochemistry 1988, 27, 1534−43. (17) Keire, D.; Mans, D. J.; Ye, H.; Kolinski, R. E.; Buhse, L. F. J. Pharm. Biomed. Anal. 2010, 52, 656−64. (18) Myette, J.; Shriver, Z.; Kiziltepe, T.; Mclean, M. W.; Venkataraman, G.; Sasisekharan, R. Biochemistry 2002, 41, 7424−7434. (19) Raman, R.; Myette, J. R.; Shriver, Z.; Pojasek, K.; Venkataraman, G.; Sasisekharan, R. J. Biol. Chem. 2003, 278, 12167−12174. (20) Swan, J. D. Anal. Chem. 1954, 26, 878−880.



CONCLUSION On the basis of experimental data, we conclude that Peak X observed in the IP-RPHPLC analysis of heparin sodium USP lots is a disaccharide containing a N,6-O-disulfated glucosamine at the reducing position linked to a modified Δ-4,5-uronic acid with a 2-sulfonic acid substitution. Structural elucidation of Peak X was performed by a combination of orthogonal techniques, ion-pairing RPHPLC, LC-MS/MS, and 1D and 2D NMR. This disaccharide was the result of process-related chemical modification of heparin that occurs upon sequential treatment with an oxidizing agent and sodium sulfite. The presence of the 2-sulfonic acid substitution can be detected in digested heparin by IP-RPHPLC or SAX-HPLC and in intact unfractionated heparin by NMR (1H NMR or HSQC) analyses due to the presence of a peak at 3.13/66.4 ppm. This report represents the first instance of identification of this novel structure in heparin and also elaborates the sequential steps in the manufacturing process of heparin that give rise to this structure. This study provides a good example on how to detect, characterize, and monitor for the presence and amount of process-related signatures in pharmaceutical heparin using high resolution analytical methods. It also demonstrates that compositional analysis techniques are useful tools not only in monitoring the composition of pharmaceutical heparin but also in discriminating between contaminants and process-related chemical modifications occurring in the heparin backbone. Furthermore, this study emphasizes the importance of the interplay between good process understanding and application of high resolution analytical techniques. 5096

dx.doi.org/10.1021/ac3007494 | Anal. Chem. 2012, 84, 5091−5096