Summary
`
`Inttvnational Jmrml of Phurmacrutrc~s, X6 ( 1902) XY-Y3 ,f: lYY2 Elsevier Science Publishers B.V. All rights resewed 037X-S 173/Y2/$05.00 1JP 02YO.i Note The adsorption of proteins to pharmaceutical container surfaces Carl J. Burke, Bryan L. Steadman, David B. Volkin, Pei-Kuo Tsai, Mark W. Bruner, and C. Russell Middaugh Kq wor&: Protein adsorption; Surface binding
`The adsorption of a variety of proteins to different pharmaceutical container surfaces was investigated. No correlation wa5 found between the amount adsorbed and molecular maw or isoelectric point. although glass surfaces appeared to hind more protein under the experimental conditions examined. The interaction of proteins with the surfaces of their storage containers is a potentially significant problem in biotechnology. The amphipathic na- ture of protein molecules results in their adsorp- tion to a wide variety of surfaces and can result in both their loss and destabilization (Feigner and Wilson, 1976; Andrade, 1985; Stella, 1986; Van der Oetelaar et al., 1989; Wu and Chen, 1989). This problem can be acute at low protein concen- tration where a substantial portion of what is usually assumed to be solution state protein may actually be adsorbed to container walls. We therefore examined the amount of surface ad- sorption of a number of proteins ranging in molecular mass from 6.5 to 670 kDa and isoelec- Correspondence to: C.J. Burke. Department of Pharmaceutical Research, WP2h-331. Merck Research Laboratories. West Point. PA 19486. U.S.A. tric point (~1) from 4.3 to 10.5 to several com- monly used container surfaces. Protein solutions containing 6.2 mM sodium phosphate, 0.15 M NaCI, pH 7.2. were stored at concentrations of 1. 5, 10, and 20 pg/ml for 24 h at 4°C in 15ml 2-cm diameter cylindrical vials. All proteins were obtained from Sigma Chemical Co., except for acidic fibroblast growth factor and transforming growth factor cu-F’seudomonas exo- toxin conjugate which were obtained from Merck and Co. (West Point, PA). The two latter proteins are highly homogeneous as determined from pre- vious isolation procedures (Heimbrook et al., 1990; Volkin et al., 1992). The glass vials are either untreated, siliconed, sulfur-treated or Purcoat “‘-treated. The plastic vials used are polyester + 0.3%, polyester 5 X 0, polypropylene, and nylon. All vials were supplied by the West Company (Phoenixville. PA) and washed and sterilized prior to use. A 5 ml volume of protein
`
`MYLAN INST. EXHIBIT 1072 PAGE 1
`
`MYLAN INST. EXHIBIT 1072 PAGE 1
`
`

`

`5 IO 15 20 0 Nylon / ” ,’ / i ,, ,’ ,I /” ,’ ,. ” / ,d’ &_._._._.-.--4 ,_..-..-----_~f; _____--- P- L I PobfPwy~~n~ ____--E ,*_--- _._-- protein (ug/mL) Fig. I. Binding isotherms of four container surfaces. Dam tor six proteins arc displayed: alcohol dehydrogenase f + ). fi-am);lase ( n ). lactate dehydrogenase (<.I), tu-chymotrypsinogen A t 0 J, thyroglohulin (+ ). and immunoglohulin G t
`Untreated Glass A ,,’ /’ ,y’ /x P- /,’ 0 5 IO 15 20 Polyester + 0.3% JE // / I I / --_________ ... .f
`Each data point is an average of three determinations. solution was added to each vial, yielding a sur- sure to the solution contact region. Adsorption face/volume ratio of 2.4 cm’ ml-‘. Solutions was allowed to proceed for only 24 h to minimi~c were not mixed or inverted to Iimit surface expo- any effect of changes in protein stability on the Fig. 2. Extent of protein hinding to containers. The maximum amount of protein adsorbed at an initial concentration of 20 @g/ml to ths surfaces of eight containers after 24 h at 4°C is represented. The amount hound was determined by averaging three rn~~isuremel~ts of protein c~~llccntr~lti(~n from each vial and ~lcu~atin~ the alnount bound from the difference in protein in solution before and after incubation. The proteins (with molecular mass and pl as indicated) examined are as follows: A. acidic fihrohla\t growth factor (IS.9 kDa, 7.0): 8. alcohol dehydrogenase (141 kDa. 5.4): c‘. apoferritin (44.1 kDa. 4.3): D. aprotinin (6.5 kDn. 10.5); E, /3-amylase (200 kDa, 4.8); F, bovine serum albumin (66 kDa. 3.9); G, n-chymotrypsinogen A (25 kDa, 9. I ): 14. conalbumin (XI) kDa. 5.6): 1. cytochrome c (12.4 kDa, lf1.3); J, immun~~~lobulin G (150 kDa, 7.5); K, t.-lactic dehydrogenase (2% kDa, X.3): 1,. lysozyme (14.3 kDa, IO); M, thyrt~glubulin (669 kDa, 4.5); N, covalent conjugate ot transforming growth factor o with a JO kDa fragment of Pscwlomonrrs exotoxin (45 kDa, 5.0). Purcoat ” is a trademark of the West Company.
`
`v 1.
`
`0
`
`MYLAN INST. EXHIBIT 1072 PAGE 2
`
`MYLAN INST. EXHIBIT 1072 PAGE 2
`
`

`

`Untreated
`
`Glass
`
`Siliconed Glass
`
`ABCDEFGHIJKLMN
`
`ABCDEFGHIJKLMN
`
`Sulfur-treated
`
`Glass
`
`_
`
`3
`
`Purcoat Glass
`
`T T&
`
`ABCDEFGHIJKLMN
`
`ABCDEFGHIJKLMN
`
`Polyester
`
`+ 0.3%
`
`Polypropylene
`
`2-
`
`1 -
`
`0
`
`ABCDEFGHIJKLMN
`
`ABCDEFGHIJKLMN
`
`Nylon
`
`4-
`
`3-
`
`3-
`
`2-
`
`3
`
`- i-
`
`ABCDEFGHIJKLMN
`
`ABCDEFGHIJKLMN
`
`protein
`
`MYLAN INST. EXHIBIT 1072 PAGE 3
`
`MYLAN INST. EXHIBIT 1072 PAGE 3
`
`5
`

`

`the binding ot
`time dependence of
`The
`results.
`scvcral of the proteins was examined and binding
`was complete within
`h.
`The
`isotherms
`thus
`appear
`to represent
`a steady
`state
`(and probably
`equilibrium)
`process over
`the
`time course of the
`measurements.
`The amount of adsorbed protein
`was determined
`by
`sampling
`three aliquots
`of
`each vial
`through
`the
`tlow cell of a SpectraFO-
`CLJS detector
`(Spectra-Physics.
`Inc.) and
`record-
`ing the absorbance of the peptide bond at 21.5 nm
`with
`the detector and sampling
`lint at ambient
`temperature.
`Bound protein was calculated
`from
`the difference
`between
`initial protein concentra-
`tion and that present
`in solution
`after
`the 24 h
`incubation.
`Protein
`concentrations
`wcrc dctcr-
`mined
`from
`individual
`standard
`curves
`for each
`protein.
`Standards
`wcrc prepared
`immediateI\
`prior
`to analysis at a lower
`surface/volume
`ratio
`of
`I.5 cm’ ml
`in polypropylene
`and measured
`immcdiatcly
`to minimize
`adsorption.
`Standards
`wcrc also
`remeasured
`after all
`the vials were
`analyzed and generally were superimposeable.
`No
`correction was made for protein adsorption
`to the
`inner
`surfaces
`of the
`flow cell. but
`the
`flow cell
`was washed bctwcen proteins
`to ensure no rcsid-
`protein was
`retained
`in
`the
`flow cell.
`Four
`point direct binding
`isotherms were employed
`to
`estimate
`the amount of protein on a surface
`(SW
`Fig. 1).
`in
`six proteins
`isotherms
`binding
`Typical
`The
`in
`Fig.
`I.
`four
`containers
`are
`illustrated
`as
`20 proteins
`for
`complete
`binding
`isotherms
`well as plots of protein molecular mass and pl vs
`various binding parameters
`of surface adsorption
`arc available upon
`request
`from
`the authors. Most
`of the 20 proteins
`examined manifest
`evidence of
`eventual
`saturation
`on all eight containers
`tested
`like
`the examples
`shown
`in Fig.
`I.
`In many cases,
`proteins appear
`to saturate at approx. 5 pg/ml
`of
`protein, a level similar
`to that previously
`observed
`for several proteins
`on different
`surfaces
`(Elgers-
`ma et al..
`1990;
`L,uey ct al..
`IYYl).
`In a
`few
`saturation
`is clearly
`not
`instances,
`however,
`achieved at
`the highest
`protein
`concentration
`tested
`(20 ~g/ml).
`Nevertheless.
`even
`in
`the
`worse cases, only
`lO-15%’
`of
`the protein
`is ad-
`sorbed.
`The
`degree of binding
`appears
`to bc
`primarily
`a property
`of
`the
`individual
`proteins
`
`for
`
`Iactatc dch~
`such as
`since proteins
`themsclvcs
`surface
`intcr-
`increased
`drclgenase which display
`action do so on all surfaces cxaminecl.
`‘I‘hc prop-
`erties
`of
`the
`surface
`itself
`clearly d(~ influence
`binding as
`illustrated
`by the very
`low affinity
`ot
`several proteins
`for untrcatcd
`and the (>‘I)
`the
`hanccd
`interaction
`of HSA
`for
`The maximum amount of protein bound
`proteins
`in eight different
`containers
`is summ;t-
`ri7ed
`in Fig. 3. Inspection
`of the binding
`isothcrnl\
`from which
`the data
`in Fig. 2 was obtained
`;I<
`well as similar
`cxperimcnts
`from
`sc~cn other pro
`tcins
`reveals a number of gcncralitie<;. No L’c)rr-e
`lation was
`found hctwcen molecular- mass ot- rhc
`icoclectric point of the proteins and their
`intcr.;tc
`tions with
`the containers
`under
`these eup~~~im~~tl-
`tal conditions.
`In addition,
`the diffcrcncc
`in 4ur-
`fact
`the proteins w;15
`muc% greater
`in container
`4urtacc
`than
`the variation
`type. Nevertheless,
`some
`types of containers
`(lo.
`on average, appear
`to manifest
`Iowcr protein
`surface adsorption
`than others. Overall.
`siliconcd
`and untreated
`glass appear
`to bind
`less pi-oteiti
`than
`the other materials
`examined, while \ulfllt
`treated glass and polyester
`contsincr4
`bind cct--
`tain proteins
`in somewhat
`incrcascd amount\.
`pro-
`In summary,
`and somewhat
`surprisingly.
`not
`tein adsorption
`to container
`surfaces
`dots
`appear
`to be a major problem
`above
`;I 5 -70
`pg/ml
`protein
`range with many
`if not most pro-
`teins over
`the 3°C‘. 24 h incubation
`pcriotl cyam
`incd.
`Since
`proteins
`sometimes
`manifest
`\Iou
`structural
`changes on surfaces
`over
`longer pcxt-l-
`ods (Andrade,
`lY85).
`further
`studies arc requit-ctl
`to establish
`the utility
`of
`these containcl\
`lot-
`long-term
`storage.
`Ncvcrthcless,
`untreated
`:lnd
`treated glasses gcncrally appear
`to bind
`the
`Ica\t
`protein,
`but all of the surfaces
`tested appeal-
`to
`have acceptable adsorption
`charactcristic4.
`II
`I\
`clear. however.
`that proteins
`riced to bc intlividtl-
`ally evaluated
`in this
`regard. Not cvaluatcd
`in thik
`study was
`the adsorption
`to vial
`stopper-s. which
`could potentially
`contribute
`significantly
`to pro-
`tcin
`loss.
`In situations
`where protein adsorption
`is significant.
`the
`inclusion
`of high concentration\
`of an inert protein
`(e.g..
`strum
`albumin)
`to satri-
`rate
`the container
`surface
`or
`the prcscncc
`01
`compounds
`to reduce surface
`interactions
`\uch :I\
`
`MYLAN INST. EXHIBIT 1072 PAGE 4
`
`MYLAN INST. EXHIBIT 1072 PAGE 4
`
`I
`’
`ual
`glass
`for
`same nlatci i;tI
`I t
`interaction hctwccn
`

`

`Imr~irw
`
`.)‘(.I..
`
`J. Ccdlord
`
`in this study. WC also ac- knowledge Drs K.A. Goklen. Ff.H. Tunp and K.C. Yeh for review of the manuscript. References lleimhrook. D.(‘.. Stirdiwnt. S.M.. Ahrrn. J.D.. 13alichin. N.I... Patrick. D.R.. Edwards. G.M.. Defeo-Jones, D.. Fit~grrald. 11.J.. Pastan. I. anti Oliff. A.. Transforming growth factor. [u-~‘.,c,lcclr,monnr ruotouin fusion protein pro- long\ survival of nude mice hearing tumor ucnografts. I’wc~. ,Yd. .4xtl. .Yu. liS.4. X7 ( l9YO) 4607--470 I I.ucy. J.-K.. McC;uire. J. and Sproull. R.D.. The cffcct of pIl and NaC‘l concentration on adwrption of /J-lactoglobulin at hydrophilic and hydrophobic silicon surtaces.
`surfactants. carbohydrates, or amino acids can be employed to rcducc tho problem (Suelter and Del~ca, 1983: Wang and I-fanson. lY8X). Acknowledgements WC would like to thank Martin Efenley for preparation of treated glass vials and steriliza- tion. and the West Company for providing all of the containers
`I43 ( IOOI ) 1X0%500. Stella. V.J.. (‘hcmical and physical ha% tlctrrmining the iwtahility and incompatibility of formulated injectable drug\. .I. frrwrrl. Sd. Tdrrwl. 40 (10x6) I-12- 16.3. Sueltcr. (‘.I I. ;~nd DeLuca. M.. How to prevent Iosscs of prcltein hy atlsorpt Ic,n 10 claw and plactic. -Irltr/ f~iodrcw~. . 135 (10x3) I I?- I I’1 Van der Oetelaar. P.J.M.. Mentinh, I.M. and Brinks. G.J.. Loss of peptides and proteins upon sterile filtration due to xlsol-ption to mcmhrsne filtrl,. D~rg f)ri cd. /UC/. f11~~wi.. IS (19X9) 97-106 Volkin, D.B.. Tsai. P.-K.. Dnhora. J.M.. Gress, J.O.. Burke, (‘.J.. Linhal-cit. R.J. and Middauph, C.R.. Stabilization of acidic fihrobla\t growth factor hy polyanion\ -Irclz. Riodfcw. B/op/~~~. ( 1 992) in pres. Wang. Y.-C‘.J. and Hanson. M.A., Parental formulations of proteins and peptideh: \tuhility and \tahilixr\.
`
`used
`
`SG.
`
`7~C/l,lOl.. 42 ( I ‘)xX) S3-mr. Wu, (‘.-S.C’. and C‘hen. (i.C‘.. Adsorption of protans onto glaaa surfacc~ antl its effect cm the intensity of circular dichroism spectra. Awl. Riot Iwr?~.. 177 ( IYXY) 17X- IX?.
`
`J. f’trrcwt.
`
`MYLAN INST. EXHIBIT 1072 PAGE 5
`
`MYLAN INST. EXHIBIT 1072 PAGE 5
`
`