`,!:) 1992 Elst'vier Science Publishers £3. V. All 1igh1s rc~c1Ycd 0378-517.1 / 92/$05.00
`
`IJP 02905
`
`Note
`The adsorption of proteins to pharmaceutical container surfaces
`
`Carl J. Burke, Bryan L. Stead man , David B. Volkin, Pei-Kuo Tsa i, Mark W. Bruner,
`and C. Russe ll Middaugh
`Oe1xmm('/l/ 11{ l'lwrmace111ica/ Research. WP2h-33 l. 1\,frrck Research /,ahomtorifs, West l'oim. l'A /9486 (U.\'.4!
`
`(Recciwd 7 January 1992)
`(M odified version received 2.'i M arch 1992)
`(Accepted 7 May 19'>1I
`
`Ke.v words : Protein adsorption; Surface binding
`
`Summary
`
`The <td:iorption of a variety of protein::i to differen t p hH rmaccuticill con tainer ~urfaccs was invcsligatcJ . No l ·o1 rc la1io11 wa'.\
`found be1ween l he amounl adsorbed and molecular mass or isoelec1ric point. although glass surfaces appe,1red 10 hin d more
`protein under 1he expe rimt:n lal condi1ions ex,lmined.
`
`The interaction of proteins with the surfaces of
`their storage contai ners is a potentially significant
`problem in biotechnology. The amphipathic na(cid:173)
`ture of protein molecules results in their adsorp(cid:173)
`tion to a wide variety of surfaces and can result in
`both their loss and destabilization (Feigner and
`Wilson, 1976; Andrade, 1985; Stella, 1986; Yan
`der Oetelaar et al., 1989; Wu and Chen, 1989).
`T his problem can be acute at low protein concen(cid:173)
`tration where a substantial portion of what is
`usually assumed to be solution state protein may
`actually be adsorbed to con ta iner walls. We
`therefore examined the amount of surface ad(cid:173)
`sorption of a number of proteins ranging in
`molecular mass from 6.5 to 670 kDa and isoelec-
`
`Corres{lo11de11ce 10: C.J. Burke. Department of Pharmaceu1ical
`Research. WP26-33 1, Merck Research Lahorn1ories. Wesl
`Point. PA 19486. U.S.A.
`
`tric point (pl) from 4.3 to 10.5 to several com(cid:173)
`monly used container surfaces.
`Protein solutions containing 6.2 mM sod ium
`phosphate, 0. 15 M NaCl, pH 7.2. were stored at
`concentrations of I, 5, 10, and 20 µ.g/m l for 24 h
`at 4°C in 15-ml 2-cm diameter cylindrical vials.
`All proteins were obtai ned from Sigma Chemical
`Co., except for acidic fibrob last growth facto r and
`transforming growth factor a-Pseudomonas exo(cid:173)
`toxin conjugate which were obtained from Merck
`and Co. (West Point, PA). The two latter proteins
`are highly homogeneous as determined from pre(cid:173)
`vious isolation procedures (Heimbrook et al..
`1990; Vol kin ct al., I 992). The glass vials are
`either untreated, sil iconed, sulfur-treated or
`Purcoat''~-treated. T he plastic vials used are
`polyester + 0.3%, polyester 5 x 0, polypropylene,
`and nylon. All vials were :supplit:u by tht: Wt:st
`Company (Phoenixville, PA) and washed and
`sterilized prior to use. A 5 ml volume of protei n
`
`FRESENIUS EXHIBIT 1072
`Page 1 of 5
`
`
`
`9t)
`
`2 ...... - - - - - - . - - - -~ - - - - , - -- -~
`Untreated Glass
`
`3 n-----,-----.----..----.,,
`Nylon
`
`2
`
`E
`(.)
`
`-('I
`0 :::, --0
`
`C:
`::::,
`0
`.0
`C:
`
`G) -0 ....
`
`C.
`
`0
`
`5
`
`10
`
`15
`
`20
`
`0
`
`5
`
`10
`
`15
`
`20
`
`3 ....... - - - - , - -- ---r-----.----"T'I
`Polyester + 0.3%
`
`2
`
`2 rr------.-----.----...... - - - . , ,
`
`0
`
`5
`
`10
`
`15
`
`20
`
`o,.. ___ _._ ___________ _
`20
`15
`5
`10
`0
`
`protein (ug/ml)
`Fig. I. Binding isotherms of four container s urfaces. Data for six pmte.ins arc displaye d: alcohol dehydrogenasc ( + ), µ- :,mylasc
`("').l actate dehydroge nasc ( o ). a -chymotrypsinogen A ( O ), t hyroglobulin ( ♦ }.and immunoglobulin G ( <J ). E ach da w point i, an
`average of thre e determinations.
`
`solution was added to each vial, yielding a sur(cid:173)
`face/volume ratio of 2.4 cm 2 m1 - 1• Solutions
`were not mixed or inverted to limit surface expo-
`
`s ure to the solution contact region. Adsorption
`was allowed to proceed for only 24 h to minimize
`any effect of changes in protein stability on the
`
`Fig. 2. Exte nt of protein binding to containe rs. The maximum amount of protein adsorbed at an initial concentration of 20 µg / ml
`to the surfaces of eight containers afte.r 24 h at 4°C is represented. The amount bound was determined by averaging three
`mea~urcments o f protein concentration from each vial and calculating the amount hound from the difference in prote in in solu tion
`1,o;;for.: and aft~, inc ubat ion. The p1vt,:;i11, (with 1nvkculr11 111a,s a nc.l vi a~ iuuil:att:<l) cxamint:<l an; a s follows: /\. a,idic fih rnhla st
`growth factor (15.9 kDa, 7.0): 8, alcohol dehydrogenase ( 141 kDa. 5.4): C. apoferritin (44J kDa, 4.3): D . aprutinin (n.5 kDa. !05>:
`E, f:l-amylase (200 kDa . 4.8); F, bovine serum albumin (M kDa. 4.9): G, a -chymotrypsinogen A (25 kDa. 9. 1 ): H , C<>nalbumin (l)tl
`kUa. 5.(l): I, q1ochrome c ( lZ.4 kDa, 10.3): J, immunoglobuli n G (1 50 kDa, 7.5); K, 1.-lactic dehydrogenase (228 kDa. X .. ll: L.
`lysozyme ( 14.3 kDa, JO); M, thyroglobulin (669 kDa, 4.5); N , covalent conjugate of transfonning gro\1<1h factor a with a 40 kDa
`fragment of Pseudomona.t exotoxin (45 kDa, 5.0). Purcoa t"'' is a trademark o f the West Company.
`
`FRESENIUS EXHIBIT 1072
`Page 2 of 5
`
`
`
`3~------------------,
`
`Untreated G lass
`
`4,--------------------,
`
`Siliconed Glass
`
`()J
`
`-
`(\J c
`0
`--....
`Ol
`-2
`"E
`::J
`0
`D
`E
`::J
`E
`X
`(U
`E
`
`2
`
`0
`
`5
`
`4
`
`3
`
`2
`
`0
`
`4
`
`3
`
`2
`
`0
`
`4
`
`3
`
`2
`
`0
`
`A B C O E F G H
`
`I J K L M N
`
`Sulf1X- t reated Glass
`
`A B C D e; F G H
`
`I
`
`J K
`
`L MN
`
`Polyester + 0 .3%
`
`A B C D E F G H
`
`I J K L M N
`
`Nylon
`
`A B C D E F G H
`
`I
`
`J K L M N
`
`3
`
`2
`
`3
`
`2
`
`0
`
`3
`
`2
`
`0
`
`e
`7
`
`6
`
`5
`
`4
`
`3
`2
`
`0
`
`A B C D E F G H
`
`I J K L M N
`
`Purcoat Glass
`
`A B C D E F G H
`
`I
`
`J K L M N
`
`Polypropylene
`
`A B C D E F G H
`
`I
`
`J K L M N
`
`Polyester 5x0
`
`A B C D E F G H
`
`I
`
`.J K L MN
`
`protein
`
`FRESENIUS EXHIBIT 1072
`Page 3 of 5
`
`
`
`9'.!
`
`results. The time dependence of the binding of
`several of t he p rotein~ was e xa mined and binding
`was complete within I h. T he isotherms thus
`appear to represent ,1 steady state (and probably
`equil ibrium) process over the time course of the
`measurements. The amount of adsorbed protein
`was determined by sampling three aliquots of
`each vial throug h the now cell of a Spectra FO(cid:173)
`CUS detector (Spectra Physics. Inc.) and rct·ord(cid:173)
`ing the absorhanee of the peptide blind at 215 nm
`with the de tector c1nd sampling line at ambient
`temperature. Bound protein was calculated from
`the difference between in itial prote in concentra(cid:173)
`tion and that present in solution afte r the 24 h
`incubation. Protein concentra tions were deter(cid:173)
`mined from ind ividual standard curves for each
`protein. Standards were prepared immediately
`prio r to analysis at a lower surface/volume ratio
`of 1.5 cm 2 ml - 1 in polyp ropylene and measured
`immediately to minim ize adsorption. Standards
`were a lso remeasured after all the via ls were
`analyzed and generally were supcrimposeable. No
`correction was made for protein adsorption to the
`inner s urfaces of the now cell , but the flow cell
`was washed between protei ns to ensure no resid(cid:173)
`ual protein was retained in the flow cell. Four
`point direct binding isotherms were employed to
`estimate the amount of protein on a su rface (sec
`Fig. I).
`Typical binding isotherms for s ix proteins in
`fou r containers a rc illustrated in Fig. I. The
`complete binding isotherms for 20 proteins as
`well as plots of protein mo lecular mass and p/ vs
`vario us binding parameters of surface adsorption
`arc available upon request from the authors. Most
`of the 20 prote ins examined manifest evidence of
`c;:vcnlual salurat iu11 u 11 all eight containers tested
`like the examples shown in Fig. I. In many cases.
`pro te ins a ppea r to saturate :11 :-ipprnx. 5 11. g/rnl of
`protein, a level similar to that previously observed
`for several proteins on different surfaces (Elgcrs(cid:173)
`ma ct al., 1990; Lucy ct al.. 199 1 ). In a few
`instances, however. saturation
`is clearly not
`achic:vcu a l
`the highest protei n concentrat i\.ln
`tested (20 µ.g/ m l). Nevertheless. even in the
`worse cases, only Hl-15% of the protein is ad(cid:173)
`sorbed . T he degree of binding appears to be
`primarily a property of the individual proteins
`
`themselves since proteins such as lactate deh y(cid:173)
`dn,gcna~e wh ic h display incrc.1~cd surface inter(cid:173)
`action Jo so on all surfaces exam ined. T he prop(cid:173)
`e r ties of the s urface itself clearly d11 influence
`binding as illustra ted by the very low affinity o f
`several proteins for untrcatl'd glass and the en
`hanccd interaction of BSA for the same rnatcri;d .
`The maximum amount of protein hound for 11
`proteins in eight different containers is s umm:1-
`rizccl in F ig. 2. I nspectio n of the binding isothern1s
`from which the data in fi g. 2 was obtained ;1~
`well as sim ilar expe riments from seven nthn pro
`tcins reveals a number 1lf generalities. No c,)rrc ·
`latit)n was round t1ctwcen molecular mass or the
`isoekctric point nf the proteins and their intcr,K (cid:173)
`tions with t he containers undc\r th,:se c,qwrin1-: 11-
`tal conditions. In additil1n, the diffcrrnce in ,ur (cid:173)
`face interaction between the proteins was much
`greater than the varia tion in cnntainer surfoc,·
`type . Nevertheless, somc types (If Cl)ntainers do.
`on average. appear to manifest lower rr()tcin
`surface adsorptio n than others. Overall. siliconct.l
`a nd u ntreated glass appear tv hi nd kss prnkin
`than the other materials examined, while sulfur(cid:173)
`treaced glass and polyester container, hind c(-r(cid:173)
`tain proteins in somewhat increased amounts.
`In summary. and somewhat surprisingly. pro(cid:173)
`te in adsorrtion to container surfaces docs not
`appear to he a major problem above a 5 -20
`µg / ml prote in range with many if not most pro(cid:173)
`teins over the 4"C. 24 h incuhation period exam(cid:173)
`ined. Since proteins sometimes manifest slow
`st ructural c ha nges on surfaces over longer peri(cid:173)
`ods (Andrade, 1985). further studies a rc required
`to establ ish the uti lity of these containers fpr
`long-term storage . Nevertheless. untreated and
`treated gla:;scs gcncrnlly 11ppc,1r to hind th.: kw,t
`protein, but all of the surfaces tested appear to
`i~
`have acceptable adsorption charactcristic-.;. It
`clear. howeve r, that pwteins need to be inJividu (cid:173)
`ally evaluated in this rega rd. Not evaluated in this
`study was the adsorption to vial s topper~. which
`could potentially contribute significantly to pro(cid:173)
`tei n lo:;:;. I n :-.ituation:-. whe n: prot..:in <1ds(1rpti•>n
`is significant. the inclusion of high concentration~
`nf an inert protein (e.g .. sc ru m alhum in) to satu(cid:173)
`rate the conta iner s urface <) r
`the prt'scncc of
`compounds to reduce surface interac tions such as
`
`FRESENIUS EXHIBIT 1072
`Page 4 of 5
`
`
`
`su rfactants, carhohydratcs. or amino acids can be
`employed to reduce th!:! proble m (Suc her and
`De Luca, I 98>: Wang and Hanson, I 988).
`
`Acknowledgements
`
`We wou ld like to thank M artin H enley for
`preparation of treated glass vials anJ steriliza(cid:173)
`t ion, and the West Company for providing all of
`the containers used in this study. We also ac(cid:173)
`knowledge Ors K.A. Gok len. H.H. Tung and
`K.C. Yeh for review of the manuscript.
`
`References
`
`Andrade. J.D .. Principle of protein ads,irption. Sur/ii<',' 1111d
`fntaj,1ci11f Aswcts "f [liomcdical f'of.\'111ers. Plenum. Nc·w
`York. I lJX:>. pp. I-XO.
`Elgersma, AV .. Zsom. R.L.J .. Nortlc. W. and Lyklema . .I .•
`The adsorption of bovine serum albumin in positiv;.:ly and
`nci:atively charged polystyn:1w latices. J. Colloid l111cr(acc
`Sci .. 138 !1990l 145- 15!,.
`Fdgncr. P.L. and Wilson. J .E..
`to
`I lexokinase binding
`rolypropylcne ll''' tuhes. Anal. Ri"ch,..,n. . 7.1 ( 1976) 6.• 1-
`6'.l.".
`
`lkimhrook. D.C.. Stirdivant. S.M .. i\hern. J.D .. Balishin.
`N.I... Patrirk . D.R .. Edwards. G.M .. De feo -Jones. D ..
`Fit7.j!erald. D.J .. Pastan. I. and Oliff. A .. Transforming
`growth fart1ir n -1'.w•udomona.,· exntoxin fusion prote in pro•
`longi.: ~urviv;1I o f nudt.!' mict" bt.!aring tumor xcnograft~.
`/>roe Natl . . '1cad. Sci. USA . 87 ( 1990J 4697 --4701.
`l.ucy . . 1.-K .. McGuire. J. and Sproull. R.D .. The dkcl of pH
`.iml NaCl conccntratilln nn adsorption of fj -lac·wglohulin
`at hydrophi lic and hydrophobic silicon surfaces. J. Col/"id
`lntcrfiu:c Sci .. 14., ( 1991/-189- StXJ.
`Stella. V . .1.. Chemic.ii and physical hasi, de termining the
`inslahil ity and inrnmpatihility of formul.ited injectabk
`drul(s. J. f'am11. S, ·i. frcl11wl.. -10 (1 986) 142- 1<,3.
`Suche r. C.H . and De l uca. M .. How 111 prcvt'nt loss<.:s ,1f
`rrotcin hy ads1>rp1ion to glass and plastic. .41111/. 1Jiod1cm ..
`13:, ( J(Jl(\ \ I 12- 1 ll.l
`Van der Oetc la ar. P.J.M .. Mentink. 1.M. and Brinks. G.J ..
`Loss o f peptides nnd proteins uron steri le filtration J ue to
`.id, 01v1io11 10 rncmt,n ,nc filtt'rs. l>nig J),,r l'I. 1ml. !'harm ..
`15 l 198\J) 97 - 11)6.
`Volkin. D.13.. Tsai, P.-K .. Dahor:1. J.M .. Gre,s. J.O .. Burke.
`CJ.. Linhard l. R.J. and Middaugh. C.R .. S tabilization 11f
`-4rch.
`acidic-
`fibroblast growth factor hy rnlyan ions.
`Biochem. !Jiophrs. ( 1992) in press.
`Wang. Y.-C.J. and Hanson. M.A.. Paren tal fllrmulations nf
`pmkins and peptides: ~tahility and s1ahilizcrs. J. f'armt.
`Sri. 1,•,·11110/ . . J2 ( 1 llRf<) s .,- S26.
`W u. C -S.( ·. anJ Ch<'n. (i.C.. Adsorption o f proteins on to
`glass surfoc·"s and its t'ffo::ct on lho.> intensi ty ,>f circular
`dichroism .,peclra. A rm/. fliu ,hnfl .. 177 ( 1989) 178- l::'12.
`
`FRESENIUS EXHIBIT 1072
`Page 5 of 5
`
`