Mixed-mode and reversed-phase liquid chromatography–tandem mass spectrometry methodologies to study composition and base hydrolysis of polysorbate 20 and 80

Description
Mixed-mode and reversed-phase liquid chromatography–tandem mass spectrometry methodologies to study composition and base hydrolysis of polysorbate 20 and 80

Please download to get full document.

View again

of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information
Category:

Pathophysiology

Publish on:

Views: 2 | Pages: 8

Extension: PDF | Download: 0

Share
Tags
Transcript
   Journal of Chromatography A, 1218 (2011) 2138–2145 Contents lists available at ScienceDirect  JournalofChromatographyA  journal homepage: www.elsevier.com/locate/chroma Mixed-mode and reversed-phase liquid chromatography–tandem massspectrometry methodologies to study composition and base hydrolysis of polysorbate 20 and 80 Daniel Hewitt a , ∗ , Melissa Alvarez a , Kathryn Robinson a , Junyan Ji b , Y. John Wang b ,Yung-Hsiang Kao a , Taylor Zhang a a Department of Protein Analytical Chemistry, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080-4990, USA b Department of Late Stage Pharmaceutical & Process Development, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080-4990, USA a r t i c l e i n f o  Article history: Available online 29 September 2010 Keywords: PolysorbateEvaporative light scatteringMass spectrometryHydrolysis a b s t r a c t Polysorbate 20 (polyoxyethylenesorbitan monolaurate) and polysorbate 80 (polyoxyethylenesorbitanmonooleate) used in protein drug formulations are complex mixtures that have been difficult to char-acterize. Here, two HPLC methods are used with evaporative light scattering detection (ELSD) and massspectrometry (MS) to characterize polysorbate from commercial vendors. The first HPLC method used amixed-modestationaryphase(WatersOasisMAX,mixed-modeanionexchangeandreversed-phasesor-bent)withastepgradienttoquantifyboththetotalpolyoxyethylenesorbitanesterandpolyoxyethylenesorbitan (POE sorbitan, a non-surfactant) in polysorbate. The results indicated POE sorbitan was presentfrom16.0to27.6and11.1to14.5%(w/w)inpolysorbate20and80,respectively.ThesecondHPLCmethodusedareversed-phasestationaryphase(ZorbaxSB-300C 8 )withashallowgradienttoseparate,identify,and quantify the multiple ester species present in polysorbate. For all lots of polysorbate 20 analyzed,only 18–23% of the material was the expected structure, polyoxyethylenesorbitan monolaurate. Up to40%and70%(w/w)di-andtriesterswerefoundinpolysorbate20andpolysorbate80respectively.Like-wise, polyoxyethylenesorbitan monooleate accounted for only 20% of polysorbate 80. A variability of 3–5% was observed for each ester species between multiple lots of polysorbate 20. The reversed-phasemethod was then used to determine the rate of hydrolysis for each polyoxyethylene sorbitan ester of polysorbate20inbasicsolutionatroomtemperature.Increasingratesofhydrolysiswereobservedwithdecreasing aliphatic chain lengths in polysorbate 20. © 2010 Elsevier B.V. All rights reserved. 1. Introduction Polysorbate 20 (polyoxyethylene sorbitan monolaurate) andpolysorbate 80 (polyoxyethylene sorbitan monooleate) are non-ionic surfactants commonly used in the formulation of proteinpharmaceuticals. The role of polysorbate in protein formulationsistopreventtheformationofaggregates[1,2]andprotectthepro- tein from denaturation at liquid-vial and liquid-air interfaces [3].The polysorbate molecule consists of two parts; the polar headgroup,polyoxyethylenesorbitan(POEsorbitan),andthehydropho-bic ester tail (Fig. 1). Structural variability can occur in both the POE sorbitan and the ester tail. The POE sorbitan in polysorbate20 and 80 contains approximately twenty ethylene oxide subunitsthat vary in both total number and positional isomers. Based onthe European Pharmacopeia specification, for the content of fatty ∗ Corresponding author. Tel.: +1 6502256259; fax: +1 6502253554. E-mail address:  dhewitt@gene.com (D. Hewitt). acids, the laurate ester accounts for 40–60% of esters in polysor-bate20,theremainingestersrangefromC 8  toC 18 .Theoleateesteraccounts for 58–85% of esters in polysorbate 80, the remainderrange from C 14  to C 18  including stearate, linoleate, and linoleneateesters[4].Thisvariabilityistheprimaryreasoncharacterizationof  polysorbates has been difficult.Studieshavebeenperformedcharacterizingpolysorbatesusingmatrix assisted laser desorption ionization mass spectrometry[5,6], LC–MS [7], and reversed-phase chromatography [8]. There have also been many assays developed to quantitate polysorbatein various matrices such as plasma [9,10] or drug formulations[11–14]. These studies focused on the identification of the manyspecies in polysorbate or the quantitation of the total polysor-bate in a sample, not the identity and abundance of each of thesub-species present in polysorbate. Here, we discuss two HPLCmethods developed for the characterization and quantitation of the sub-species present in polysorbate 20 and polysorbate 80. Amixed-mode method previously reported separates POE sorbitanfrom the POE sorbitan esters in polysorbate [13]. The mixed-mode 0021-9673/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2010.09.057  D. Hewitt et al. / J. Chromatogr. A 1218 (2011) 2138–2145 2139 Fig. 1.  Theoretical structure of polysorbate 20. Approximately 40–60% of thehydrophobic tail is a laurate ester. method was used to quantify the total percent mass of POE sor-bitan in polysorbate 20 and polysorbate 80, the amount of whichhas not been reported, and its specification has neither been setby vendor certificate of analysis nor by compendial method. Thisis important because POE sorbitan is not a surfactant. Second, areversed-phase (RP) HPLC method used a linear gradient to sep-arate, identify, and quantitate the multiple POE sorbitan esters of polysorbate20and80.TheHPLCmethodswereusedinconjunctionwith evaporative light scattering detection (ELSD) and mass spec-trometry (MS) to characterize and quantitate sub-species presentin polysorbate obtained from commercial vendors.The ester linked hydrophobic tail is critical to the surfactantproperty of polysorbate. Since esters are susceptible to hydrolysis,it is important to understand what effect hydrophobic tail hetero-geneity will have on polysorbate hydrolytic stability. Both HPLCmethodsdescribedinthisreportareabletomonitorthehydrolysisof polysorbate. The mixed-mode method can measure the overallhydrolysis of polysorbate which was previously reported. The RPmethod can measure the hydrolysis of each POE sorbitan ester inpolysorbate.Thisinformationwillallowscientiststogaininsightonthehydrolyticstabilityofpolysorbatelotsbyobtainingareversed-phase profile. 2. Experimental  2.1. Reagents and materials HPLC grade isopropyl alcohol (IPA) (Burdick & Jackson,Muskegon, MI, USA), formic acid (Alfa Aesar, Ward Hill, MA, USA),and water purified using a Milli-Q filtration system (Millipore, Bil-lerica,MA,USA)wereusedintheHPLCmobilephase.Threegradesof polysorbate 20 were evaluated from both standard and “super”refinement processes (Croda, Rancho Cucamonga, CA, USA) andNOF (Tokyo, Japan). Polysorbate 80 samples were obtained fromCroda.POEsorbitanstandardswereobtainedfromNOFandCroda.These materials were used as received.30mg/mLpolysorbateorPOEsorbitanstocksolutionswerepre-pared by accurately weighing 150mg into a 5mL volumetric flask,then filled with MiliQ water. Serial dilutions were carried out toreach concentrations outlined later in the communication.  2.2. Chromatographic system Chromatographic analysis was performed on an Agilent 1200HPLC system (Palo Alto, CA, USA) equipped with a binary gradientpump, autosampler, a temperature-controlled column compart-ment, and an evaporative light scattering detector (ELSD, 380-LS,Varian, Palo Alto, CA, USA). The ELSD settings were as follows; gasflow rate 1.0 SLM, nebulizer temperature 45 ◦ C, and evaporationtubetemperature100 ◦ C.Nitrogengaswasprovidedbyanin-housenitrogen gas generator system at 65psi. The injection volume was20  L, and the column temperature was 30 ◦ C.  2.3. Mixed-mode chromatography The Mixed-mode method has been previously reported [13].AnalyteswereseparatedusinganOasisMAX(20 × 2.1mm,30  m,Waters, Milford, MA, USA) column. Initial conditions were set at90% solvent A (2% formic acid) and 10% solvent B (2% formic acidinIPA).SolventBwasincreasedto20%inthefirstminuteandheldfor 2.4min. POE sorbitan esters were eluted using a step gradientof 20% to 100% B over 0.1min, followed by an equilibration step of 10% B for 0.9min. The flow rate was kept at 1mL/min.  2.4. Reversed-phase chromatography Analytes were separated using a Zorbax SB C8 column(50 × 4.6mm, 5  m, Agilent, Santa Clara, CA, USA). Initial condi-tions were set at 90% solvent A (2% formic acid) and 10% solvent B(2% formic acid in IPA). Solvent B was increased to 20% in the firstminute and held for 2.4min. Separation of polysorbate esters wasachieved using a linear gradient of 20% to 100% B over 19.6min,followed by an equilibration step of 10% B for 5min. The flow ratewas kept at 900  L/min.  2.5. Mass spectrometric analysis Mass spectrometric analysis was carried out on a MicromassQTOF-1(Beverly,MA)massspectrometeroperatinginapositiveionmode via electrospray ionization (ESI). The instrument was set torunwithacapillaryvoltageof3500V,sampleconevoltageof55V,sourceblocktemperatureof125 ◦ Canddesolvationtemperatureof 200 ◦ C. The mass spectrometer was triggered via a contact closure.Calibration was performed in the  m /  z   range of 100–3500 using asolution of sodium cesium iodide (NaCsI). Instrument control anddata analysis were achieved using Waters MassLynx version 4.0softwarepackage.TheMaxEnt3programwasappliedtodeconvo-lute the multiply-charged ions.  2.6. NMR analysis Data were recorded on a Bruker 500MHz spectrometerequipped with a 5mm BBI probe and a Bruker BACS-60 autosam-pler.TheH 2 Osignalwassuppressedbyirradiatingthesamplewitha low power saturation pulse at H 2 O frequency during the relax-ation delay. Prior to the NMR measurement, D 2 O was added to allsamples to a final concentration of 10%.  2.7. Base hydrolysis of polysorbate 20 Polysorbate 20 (1mg/mL) was incubated with ammoniumhydroxide (200mM) for 22h at room temperature. Every 30minthe autosampler injected an aliquot of the reaction mixture intotheHPLCreversed-phasemethod(Section2.4).Foreachtimepoint, theconcentrationofeachPOEsorbitanesterwascalculatedfromastandard curve generated using the mixed mode method (Section2.3). 3. Results and discussion  3.1. Characterization of polysorbate using the mixed-mode HPLC method The mixed-mode HPLC method uses a step gradient to elute allPOE sorbitan esters in a single peak. Two peaks are observed in atypical polysorbate chromatogram (Fig. 2). The flow through peak at 0.3min accounts for approximately 15–20% (w/w) of the totalpeak area. Both peaks were analyzed using liquid chromatograph  2140  D. Hewitt et al. / J. Chromatogr. A 1218 (2011) 2138–2145 Fig.2.  Typicalmixed-modeELSDchromatogramofpolysorbate20.Peaksat0.3minand 4.6min correspond to POE sorbitan and POE sorbitan ester respectively. mass spectrometry (LC–MS) and collected for nuclear magneticresonance spectroscopy (NMR) analysis.Raw mass data were deconvoluted using the Waters MassLynxMaxEnt3softwarepackage,yieldingthemonoisotopicmassoftheanalytes. Both sodiated and non-sodiated forms of polysorbateswere observed in the deconvoluted mass spectra. The deconvo-luted mass spectrum for each peak contains two mass envelopes(Fig. 3B). The more abundant higher molecular weight envelope (1100–2200Da) contains masses consistent with POE sorbitanesters. The masses in each envelope are separated by 44Da, themass of one ethylene oxide residue. The most abundant massesobserved are 1309.74Da (Fig. 3A) and 1491.91Da (Fig. 3B) in the 0.3min and 4.6min peaks respectively, a difference of 182.17Da.The observed monoisotopic mass of 1491.91Da [M+H] + correlateswellwiththetheoreticalmonoisotopicmass(1491.92Da,[M+H] + )of the POE sorbitan laurate ester containing 26 ethylene oxideresidues (within mass error). The expected loss due to hydrolysisof the laurate ester is 183.32Da. Therefore, the higher molecularweight envelope of the 0.3min peak was assigned to POE sorbi-tan and the 4.6min peak was assigned to POE sorbitan esters. Themass spectrum for the 4.6min peak is more complex than that of the 0.3min peak owing to multiple POE sorbitan esters present inpolysorbate 20.The lower molecular weight envelope (400–1000Da) corre-sponds to byproducts of polysorbate synthesis, mainly isosorbidepolyethoxylates (IPE) [5–7]. The observed mass of 813.53Da (Fig. 3B) correlates well with the [M+H] + IPE laurate ester ioncontaining11ethyleneoxidesubunits,whosetheoreticalmonoiso-topic mass is 813.52Da. Similar to what was found for the highmolecular weight envelope, the 587.33Da mass observed in the0.3min peak correlates well with the [M+H] + IPE ion contain-ing 10 ethylene oxide subunits theoretical monoisotopic mass is587.32Da.Thelossofaliphaticresonancesbetween0.5and2.5ppmin the NMR spectrum of the 0.3min peak confirms these assign-ments (Fig. 4).ThepresenceofPOEsorbitaninpolysorbateislikelyabyproductof polysorbate synthesis, either due to incomplete esterification of sorbitan or as a result of hydrolysis during ethoxylation, the con-densationofthesorbitanesterwithethyleneoxide,inthepresenceof an alkaline catalyst [15]. The mixed-mode method was used to determine the amount and variability of POE sorbitan present incommercialpolysorbate20and80.Atotalofnineteenpolysorbate20lotsandfourpolysorbate80lotswereanalyzed.ThePOEsorbitan Fig.3.  Deconvolutedmassspectraofthe0.3min,POEsorbitanpeak(A)andthe4.6min,POEsorbitanesterpeak(B)forpolysorbate20Lot#1acquiredusingthemixed-modeHPLC method.  D. Hewitt et al. / J. Chromatogr. A 1218 (2011) 2138–2145 2141 Fig.4.  NMRspectraoffractioncollectedmixed-mode0.3minpeak(A),4.6minpeak(B), and polysorbate 20 control (C). concentration was calculated from a POE sorbitan standard curveranging from 10 to 100  g/mL prepared from material receivedfrom Croda. LC–MS data confirmed the POE sorbitan masses foundin polysorbate 20 (Fig. 3A) are consistent with those in the stan- dardPOEsorbitan(Fig.5A),andNMRconfirmstheabsenceofester species in the standard (Fig. 5B). Since the ELSD has a logarith- mic not linear response [16], a log–log linear standard curve was obtained by plotting the log(peak area) against the log(mg/mL) of POE sorbitan yielding a Pearson correlation coefficient ( R ) >0.999.The resulting concentration of POE sorbitan ranged from 16.0 to27.6% (w/w) in polysorbate 20 and 11.1 to 14.5% (w/w) in polysor-bate 80 samples (Table 1).  3.2. Characterization of polysorbate using the reversed-phaseHPLC method Polysorbate20ismadeupof40–60%laurateesters,theremain-der is a mixture of esters with chain lengths from C 8  to C 18  [4]. Areversed-phase HPLC method was used to separate, identify, andquantitate the different POE sorbitan ester species in polysorbate.A similar method was previously reported, but made no attemptto identify or quantitate the species observed [17]. Representa- tive chromatograms acquired for polysorbate 20 and polysorbate80 using the RP method are shown in Fig. 6. The 2.4min peak contains non-esterified POE sorbitan, IPE, and polyethylene glycol(PEG). Later in the polysorbate 20 chromatograph, there is a seriesof eight peaks, followed by a broad tailing peak. The peaks wereassigned using LC–MS and are summarized in Table 2. An exam- ple deconvoluted mass spectrum used for the assignment of the10.4min peak is shown in Fig. 7. The IPE laurate [M+K] + species isidentified in the 719, 763, and 807 series of ions and PEG laurate[M+K] + species is identified in the 679, 723, and 767 series of ions,both present in relatively low abundance. The POE sorbitan mono- Fig. 5.  Deconvoluted mass spectrum (A) and NMR spectrum (B) of the POE sorbitan standard obtained from Croda.  2142  D. Hewitt et al. / J. Chromatogr. A 1218 (2011) 2138–2145  Table 1 POE sorbitan mass percentages in commercial polysorbate lots. Lot numbers were arbitrarily assigned.Sample Vendor Lot number Refinement process POE sorbitan (% mass)Polysorbate 20 Croda 1 Standard 26.4Polysorbate 20 Croda 2 Standard 18.9Polysorbate 20 Croda 3 Standard 24.9Polysorbate 20 Croda 4 Standard 23.8Polysorbate 20 Croda 5 Standard 21.9Polysorbate 20 Croda 6 Standard 24.0Polysorbate 20 Croda 7 Standard 21.2Polysorbate 20 Croda 8 Standard 26.5Polysorbate 20 Croda 9 Standard 24.2Polysorbate 20 Croda 10 Standard 21.9Polysorbate 20 Croda 11 Standard 23.1Polysorbate 20 Croda 12 Standard 23.2Polysorbate 20 Croda 13 Standard 23.2Polysorbate 20 Croda 14 Standard 24.7Polysorbate 20 Croda 15 Super refined 27.6Polysorbate 20 Croda 16 Super refined 17.7Polysorbate 20 Croda 17 Super refined 18.0Polysorbate 20 Croda 18 Super refined 16.0Polysorbate 20 NOF 19 Standard 27.6Polysorbate 80 Croda 20 Standard 14.2Polysorbate 80 Croda 21 Standard 14.5Polysorbate 80 Croda 22 Standard 13.4Polysorbate 80 Croda 23 Standard 11.1 laurate [M+Na] + species is identified by the 1381, 1425, and 1469series of ions. The [M+H] + adduct is also present in Fig. 6 identi- fiedbythe1535,1579,and1623seriesofions.Thepresenceofthemassenvelopesisreflectiveofthepolydispersityofethyleneoxidesubunits in polysorbate. It was interesting to note that the aver-age number of ethylene oxide subunits was twenty-six for eachPOE sorbitan ester, similar to that observed for the POE sorbitanin the mixed mode method, which is consistent with previouslyreported results [18]. POE sorbitan esters in polysorbate 20 follow the expected trend where longer hydrophobic esters exhibit laterelution times. The broad peak late in the chromatogram is madeup of several di- and tri-ester polysorbate species, some of thoseidentified are listed in Table 2. Since true standards could not be readilyobtainedforthePOEsorbitanesterspecies,afivepointcal-ibration curve was constructed from the POE sorbitan ester peak(4.6min)inthemixed-modemethodusingthesameinjectionvol-ume,flowrate,andELSDsettingsasthereversedphasemethod.Thecalibration curve (log(peak area) vs. log(mg/mL)) was corrected toaccountforthepresenceofPOEsorbitan.Thiscalibrationcurvewasused to calculate the concentration of each ester species observedin the reversed-phase method. The total mass of ester species inthe reversed-phase method recovered using this calibration wasbetween90and110%oftheexpectedvalueforallpolysorbatelots Fig.6.  Representativechromatogramsforpolysorbate20Lot#1(A)andpolysorbate80 Lot #20 (B) using the reversed-phase HPLC method with ELS detection. Greytraces are water injections demonstrating the baseline of the chromatograms.  Table 2 Reversed-phase method peak assignments for Croda polysorbate 20 Lot #1. The monoisotopic theoretical mass was calculated using the “ethylene oxide subunits reported”column, this value corresponds to the most abundant mass observed. The stearate ester is represented by C 18  and the oleate ester is represented by C 18:1 . Peak numberscorrelate to those designated in Fig. 6.Peak number Retention time(min)Assignment Ethylene oxidesubunit rangeEthylene oxidesubunitreportedMonoisotopictheoreticalmass [M+Na] + Observed mass[M+Na] + ppm1 8.2 C 8  ester 19–32 26 1457.84 1457.80 312 8.8 C 8  IPE 7–16 11 779.44 779.43 133 9.7 C 10  ester 16–34 26 1485.87 1485.82 334 10.4 C 12  ester 17–31 23 1381.83 1381.79 285 11.5 C 14  ester 19–33 26 1541.94 1541.91 196 12.2 C 16  ester 19–29 23 1437.89 1437.87 147 13.0 C 18:1  ester 22–30 24 1507.93 1507.84 59C 18  ester 22–30 24 1509.94 1509.92 138 14.2 C 12  diester 18–33 25 1652.05 1651.99 369 15.0 C 12 /C 14  diester 21–29 26 1724.11 1724.10 610 16.0 C 12 /C 16  diester 19–30 24 1664.08 1664.06 1211 16.9 C 12 /C 12 /C 12  tri-ester 24–32 27 1922.27 1922.15 6212 17.9 C 12 /C 12 /C 16  tri-ester 22–30 26 1934.30 1934.23 3613 18.4 C 12 /C 12 /C 18  tri-ester 22–27 25 1918.31 1918.24 36
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks