`Demonstrated by Equilibrium Dialysis
`
`Wei)", Li, M O, Joh n II. Rockey, M D, PhD
`
`• The bindil'lg of "uorlleein to normal
`human •• rlJm proteins In 8 physiologic
`.olvent a' 37 ' C Vl' at me .. ured by equllib·
`rlum dla'y. I,. Human ,.r um contained
`3.28 X 10 - ~ M concentration Iluorascein(cid:173)
`binding .. II ••. w llh an average association
`con. tant a' 37 "C 01 0 .54 X IO'M" . The
`percentage of lotal flu orescein bound by
`human • • rum pro teins ranged Irom 83""
`10 53% when the total fluorescein con(cid:173)
`cenlraUo n ra nged hom 6.9 x 10-· to
`6 .2 x 10- ~ M .
`(Arch Opllthalmo/1982 ;100:4M-487)
`
`F luorescein has been widely used as
`a t racer to test the permeability of
`the blood-ocular barrier
`for
`two
`decades. U Although seve ral studies
`have been carried out on the binding
`of fluorescein to isolated serum albu(cid:173)
`min, whether fluorescein is substan(cid:173)
`tially bou nd to any plasma protein
`after intravenous injection in man
`remains controversial.z,:
`Laurenee5
`initially demonstrated
`isolated bovine serum albu(cid:173)
`that
`min showed concentration-dependent
`bi nding of fl uorescein by fluorescence
`polarization. He also reported that
`the intengity of fluorescein fluores(cid:173)
`cence dropped markedly concomitant
`with binding to bovine a lbumin (flu o-
`
`A~pl l'd for publicalion Apri l 29. 19111.
`~'rom the &heie Foye In ~l i lute, Department 0(
`l: nh·tMli ty of Pennsy lv~ni~
`OphlhalmoIOlO·.
`Medical &-hool. Philadtlphia.
`t:ye Institute,
`Repri nt rfquHu to Sche ie
`M>'ri n Cire-It. 51 ~ 39lh 51, Philatkolphia, J>A
`HII04 (Or Roekt>·).
`
`rescence quenching), and t he fluores(cid:173)
`cein absor ption lIpectrum was red
`s hifted .~ Bot h observations are fur(cid:173)
`ther evidence that nuoreseein does
`bind to bovine albumin. Andersson et
`al6 subtrequentiy demonstrated fluo(cid:173)
`rt:scein binding
`to
`isolated bovine
`albumin by equilibrium dialysis at 5
`to 30 °C.
`Recently. however. Ianacone et aP
`have studied the binding of radioac(cid:173)
`tive fluorescein using polyacrylamide
`gel electrophoresis and gel filtrat ion
`and have qUl!stioned whether fluores(cid:173)
`ctlin is bound in significant quantity
`to a ny plasma protein under physio(cid:173)
`logic conditions.
`A knowledge of the character of
`fluorescein bindinlt to blood proteins
`under phySiologic conditions is impor(cid:173)
`tant for an understanding of quanti(cid:173)
`fluorescence measurements,
`tative
`both in clinical diagnosis and in exper(cid:173)
`imental reSt!arch. We therefore have
`reexamined thi s question using eq ui(cid:173)
`librium dialysis, a classic method
`firmly established on thermodynamic
`theory t hat is well suited to study
`small ligand binding to protei n s.I. I~
`Fluorescein binding to human serum
`proteins at 37 °C in a physiologic
`solvent was measu red ins tead of bind(cid:173)
`ing to isolated bovine albu min at low
`temperatures. Substantial concentra(cid:173)
`tion-dependent binding of low affinity
`was observed t hroughout a wide range
`of fluorescein concentrations.
`MATERIALS AND METHODS
`Fre!lh blood was obtained (rom normal
`male \'oluntecn. Serum was used imme·
`
`cliateh' or stored at 4 'C until use. Serum
`prol('(ns were dilut('d tenfold wilh T.\·rode·,
`solut.ic," (1)11 7.4 ) for equil ibri um d ialysis.
`Fluorescein sodium IFunduscein) ..... as
`diluted with Tvro-de's );{)Iul ion to Io(h'e con·
`centralions of 12.5
`to 2,500
`mJ,!"/ 1.
`(0.3.1 x 10-' to 0.66 x to- 1M).
`F,(juilibrium d ialrsis was performed
`with specially dcsi~noo ~Ils. ('ach of which
`comprised two identiC<lI d,,~d (."Uml)art(cid:173)
`ments st'parated by a 23-mm (diameter)
`cellulose·disk dialvsis memhrane.'o.,) The
`cells "'ert immer~d in a constant·temper·
`atuN.' walcr bath at 31 ·C and rotated a t a
`s pel!Cl of 5 rpm. Thrlre J,l:roups of experi(cid:173)
`ments were carried oul. In ~roup 1. a J,!"h'cn
`concenlration of liuortscein !!Ulution (e)(.
`12.:) mjl:/ L I was dialyzed agai nst Tyrode's
`solution (or variations of dialYl>is
`lime
`from 15 minutes to 50 hours to determine
`the time required to attain equilibrium of
`liloland diffusion across the dialy~ i::l m~rn
`brane. In J,!"roup 2. a scri('s of roncentra·
`tio n ~ of fluoresc. .. ei n were dialp.ed a.gainst
`Trrooe's solution alone to dct('rminc the
`amount of dye hound b~' thel;ellulose mem(cid:173)
`brane as a function of the free fluorescein
`eon~n l rati on at equ ilibrium. In I(roup 3,
`the series of concentrations of fluorescein
`solutions w('re dia!ywd a)(ainst the 1:10
`liilutl!Cl serum. The dial.\·sis was terminated
`after 50 hours in experimental groups 2
`and 3. All cells in the t hree Krou))!; were
`run in duplicate. T he optical density of the
`fluorescein in the nonprotein compartment
`was measured at }o. 490 nm with a speoetro(cid:173)
`photomeler (Zeiss PMQ II) usin)( a 10-mm
`path-Ienjl;th eell. The molar extinction
`cot'fJicient (cl of lI uorescei n at 490 nm was
`determinE'd to be 8.6 x 10". The data of the
`group 3 experiments (dialysis of Ruores(cid:173)
`eein a)(ainst serum proteins) were cor·
`reeled for lijl;lInd binding by t he (t'lI ul~
`dialysis membranes with
`the data of
`eXlW"rimenlal Kroup 2. Protein-bou nd
`
`4a.
`
`Ar ch Ophihaimol-Vol 100, Mal ch 1982
`
`Equilibrium DialysiB-Li & Rockey
`
`Apotex v. Abraxis -IPR2018-00152, Ex. 1007, p.OI of4
`
`
`
`'00
`
`•
`" ~
`~ 50
`
`~
`"
`
`•
`
`~ (
`
`-,
`
`-.
`
`-,
`
`-.
`
`-,
`
`,.
`
`..
`
`.<2
`
`"
`
`Time, I\(
`Fig 1.-Time roquired to att ain equilibrium 01 tluorescein dilfu ~iQf1
`across dialysis membrane in ab scnce of sorum proteins. Perc enl of
`tota i lluorescein present in two compartments of equilibrium dialysis
`cell (upper curve, initial fluorescein solution side : lower curve, initial
`solvent side) is plotted against dialysis time
`
`-.
`
`l og F,
`Fig 2. - Concenlralion-dependent billdillg of fluorescein by equilib(cid:173)
`rium dialysis cell membrana. datermined in absence of serum
`protein s_ Log8rithm,o of membr8ne-bo • .md fluorescein (log F ... ) is
`plotted as IUnl;tion of the l ogarilhm,~ of free fluorescein CanCenlr8-
`tion (log F,) at equilibrium. Data were litted to straight line by method
`of teast squares.
`
`F ig 3. - Bindinll of tluoroscoin by human serum proteins al 37 · C,
`measured by equilibrium dialysis. l ogarithm,o of p rotein·bound
`Uoorescein concentra tion (corrected l or lIuoroscei" bound to the
`dietys;a m&mbranel, too F •. is plotted as function of tooarithm,o of
`free lIuarescein concer'ltra1ion, log F" at equilibrium. E)(perimen18'
`curve (solid line), measured with se'l,Im protein diluted 1: 10 in
`Tyrode's soluHon. has been cOtrected lor dilution (binding by
`undiluted serum prOleins) in upper curve (broken line).
`
`Fig 4.-Equilibrium dialysis data of Fig 3 presen ted in terms of
`Sca tc"ard plot where r is amoun t 01 fluorescein bound by protein
`and c is free lIuorescein concentration Method of least squares
`was usad to obtain basHit straight line.
`
`-,
`
`-.
`
`-,
`
`-,
`
`-.
`
`log F,
`
`-,
`
`•
`
`b
`X
`0
`
`•
`
`•
`
`•
`
`2
`
`0
`
`•
`
`,
`
`Arch Dphthaimol-Yol 100. March 1982
`
`Equilibrium Dialysis -Li & Rockey
`
`485
`
`Apotex v. Abraxis - lPR20 18-00 152, Ex. 1007, p.02 of 4
`
`
`
`iluore!l('('in IF.) was obtained from the
`fol1owin)l: rellilions hip:
`
`P. -F, -Fr-P • •
`where F, is the total r:on~nt ra.tion or Huo_
`reo<cein, Fr is the concentration of free fluo'
`rescein at equilibrium, and p ... is t he Huo,
`r~!S(."t'in bound by the dialysis cell memo
`brane.
`
`RESULTS
`
`The time required to obtain e(juilib(cid:173)
`rium of fluoresce in diffusion across
`the dialysis membrane separating t he
`two compartments of the equ ilibrium
`dialysis cell, determined in the ):.l:roup
`1 experiments, was 50 hours (Fig I),
`Therefore, all equilibrium dialysis
`experiments of fl uorescein against
`serum proteins were continued for 50
`hours or longer. The amount of fluo(cid:173)
`rescein hound by the dialysis mem(cid:173)
`brane. plotted against the free fluo(cid:173)
`rescein concentration at e(juilibrium.
`determined
`in the gr(XJP 2 experi(cid:173)
`ments. is sho ... ·n in Fig 2. The binding
`to the cellulose membrane was con(cid:173)
`centration dependent. This cu rve was
`uSt.'tI
`in calculating protein-bound
`fluorescei n (F~) from the experimen(cid:173)
`tal group 3 results, where t he total
`fluorescein (F,) a nd free fluorescein
`(Fr) eoncentrationll at Ctluilibri um
`were known.
`In Fig 3, the logarithm of the pro(cid:173)
`tein-bou nd fl uoresctin concentration
`(loglt F ,), corrected for the dye bound
`to the artificial membrane. is plotted
`against that of the free fluorescein
`concentration. The flatness of the
`binding cune at the highest free
`fluorescein concentrations indicated
`that protein fluo rescein-bi nding site
`saturation had been obtained (Pig 3).
`The experimental data were deter(cid:173)
`mined using iflnfotd-di luted human
`serum proteins. To approxi mate t he in
`vivo situation,
`the :second (upper)
`curve of Fig :.I shows t he data replot(cid:173)
`led for binding to undiluted serum
`proteins, This
`curve
`is
`simply
`obta ined because the free fluorescein
`concentration (P,) determines only
`the fraction of available binding sites
`occupied by fl uorescein; the amount of
`fluor escein bound at a given free
`fl uorescein concentration is a simple
`linear fun ction of the concentration of
`available binding sites (eg. proteins)
`(see below, I<'ig 4).
`The maximal and mi nimal fractio ns
`of the lotal
`fl uorescein bound by
`undiluted serum proteins at 37 °C
`throughout the range of frce fluores(cid:173)
`cein concentrations examined are giv~
`en in the Table, The molar concentra(cid:173)
`tion of fluorescein-binding !!lites in
`undiluted hu man serum, determined
`
`Pereenl age 01 TOl al Fluorescein Bound b~ Undiluted $arom P,oteins at t oweSI
`and Higheat Free Fluoffilseein Concentrations Measured
`
`Total
`
`'"
`15.9 X iO-'
`15.2 X 10- '
`
`frn
`'"
`
`t.2 X 10 •
`2.~ X 10"
`
`.~'"
`IF,)
`5.T x io-'
`3.3 X 10-'
`
`Rallo F,I " .
`X 100. ""
`
`" "
`
`from Fig 3, was 3.28 X 1O-'M,
`Binding data of Fig 3 were replotted
`in Fig 4 in the form of a Scatchard
`plot,II.lt" ric vs r. where r is the
`amount of protein-bound fl uorescein
`and c is the concentration of free
`Ouorescein. The maximum value for
`r (n) was arbitrarily chosen as 2, so
`that the association constant (KeJ is
`given hy the reciprocal or the freu-dye
`concentration at r - I (half of the
`binding sites of the proteins were
`occupied by fl uorescein), as obtained
`from the relationship K." r( n - r )-'
`c- I . " " The average intrinsic associa(cid:173)
`tion oonstant (K.) of human serum
`proteins in a physiologic solvent at 37
`°C for fl uorescei n was 0.54 X I ~M -'.
`
`COMMENT
`
`The present results of equilibrium
`dialysis demonstrate that in a physio(cid:173)
`logie solvent at 37 °C, substantial
`binding of
`fl uorescein
`to human
`serum proteins aceurs throughout a
`wide range of fluorescein concentra(cid:173)
`tions, even thou~h the association is
`of low affinity. The appropriateness of
`equ ilibrium dialysis for evaluating
`fluorescein bi nding under physiologic
`conditions hall been questioned by
`lanaeone et aiT because of the long
`time requi red to obtain equilibrium.
`This misu nderstanding has occurred
`because the time required fo r free
`ligand to reach equi librium across the
`artificial dialysis membrane (eg. 50
`hours) has not been d ifferentiated
`from the forwa rd rale constant for
`fluo rescein-protein association, ob(cid:173)
`tained when fluoresce i n
`is added
`dir~t1y to a protein solution, which is
`limited on ly by the rate of diffusion of
`fluoresce in in free sol u tion.~lf The
`artificial
`semipermeable
`dialysis
`membrane is only used so that an
`unambi~uou s measu re of the free
`ligand concentration may be obtained.
`The technique is en tirely appropriate
`for measurements of equilibrium
`thermodynam ic parameters of asso(cid:173)
`ciation of rapidly
`interacting sys(cid:173)
`tems." "
`Gel filtration and polyacrylamide
`gel electrophoresis,! however, may not
`be appropriate techniques to measure
`associations of
`fluorescein -protein
`
`low affinity. In both techniques, f1uo(cid:173)
`rescein'protein complexes are contin(cid:173)
`uously bei ng separated from
`free
`fluorescein because of their differen(cid:173)
`tial migration rates. This factor is
`particularly
`true
`in gel filt ration,
`where fluorescein-protein complexes,
`because of their larger size, arc con (cid:173)
`tinuously being transferred into sol·
`ven t that contains no free fluorescei n .
`This process is equivalent to dialyting
`the
`fl uorescei n.protein
`com plexes
`against an infini te volume of soillent
`and is an effectille way to dissociate
`ligand-protein com plexes of low affin·
`ity. These tcehniques will not measure
`the extent of binding of fiuo res«in t<l
`blood proteins that oc<:urs in the pres(cid:173)
`ence of high concentrations of free
`fl uorescein.
`The concentration of fluorescein(cid:173)
`binding lIites in serum was deter(cid:173)
`mined to be 3.28 x 1O-3M. The normal
`concentration of albumin in human
`serum
`is 0.65 x 1O-3M.u
`Isolated
`bovine serum albumin has been
`reported to have three binding sites
`per molecu l e.~ The ratio for bound
`l1uores~in to human serum albumin
`of 5 (3.28:0.65) indicates either that
`human serum albumin has five fl uo(cid:173)
`rescein·binding sites per molecule or,
`more likely. that other serum proteins
`(eg, low-affinity IgG immunoglobu(cid:173)
`Ii ns) also bind fl uorescein under phys(cid:173)
`iologic conditions. A knowledge 01 the
`fluorescein bi nding by blood protein s
`of substantially different size and dif·
`fusion rates wou ld be of importance
`fo r a oomplete picture in quantitative
`st udies of blood-ocular permeability
`changes.
`A knowledge of the extent of flu()(cid:173)
`rescei n binding to proteins at differ (cid:173)
`en t free fluorescein concentrations
`also is necCMary when blood flu ores(cid:173)
`cein levels are measured ftuorometri (cid:173)
`cally, since t hese data must be cor(cid:173)
`rected for the fluor escence quenching
`fl uorescein if a true
`of the bound
`measure of total fluorescein content is
`to be obtained by this method.
`The find ing that a substantial frac (cid:173)
`tion of the total injected fl ooresce in is
`bound at low affi nity by human blood
`proteins in a physiologic solvent at
`3'7 "C is eonsistent with clinical obser (cid:173)
`vations. The laminar flow seen in reti-
`
`486
`
`Arc h Ophihaimol-Vol 100. March 1982
`
`Equilibrium Dialysis-li & Rocke~
`
`Apotex v. Abrax is -IPR20 18-00 152, Ex. 1007, p.03 of4
`
`
`
`nal vessels!.l' in all likelihood reflects
`fluorescein bound to proteins, as the
`protein-bound
`fluorescein diffusion
`rate will be much slower than that of
`free fluorescein. The rapid changes in
`the volumes of distribution of fl uores(cid:173)
`cein and 13lI_labeled albumin, ob(cid:173)
`served when both were injected simul(cid:173)
`taneously,!f would reflect
`the low
`association constant of fluorescein
`
`binding.
`A complete understanding of quan(cid:173)
`titative fluorescein angiography of
`the retina and vitreous fluorometry
`will require a detailed knowledge of
`the contribution of fluo rescein bind(cid:173)
`ing by blood constituents during the
`changing concentration ratios occur(cid:173)
`ring intravascularly and extravascu(cid:173)
`larly after the initial injection of the
`
`bolus of fluorescein. The present
`experimental results may be used to
`calculate the contribution of fluores(cid:173)
`cei n binding by human serum proteins
`in such studies.
`
`This study wall sup"".ted by an un~stritl-ed
`Kranl f'om Rnl-areh 10 P re_ent Bl indn .. "... Inc,
`New York. and by
`th e Hatry and Edith
`Hu~hman Re-search ~'und
`
`Raferences
`
`I. Novolmy HR. Alvi. 01..; A method of photo(cid:173)
`graphin.(t Huorescence in cireulating blood in the
`human utina. Circu/ali(l'll 1961;24:S2·SG.
`2. Wes.inp; A: tlioc~ .mical properties of Huo(cid:173)
`r."""in. in Fllmrescrin A ",,;ovraph~ 0/ the Reh·
`nil " 5 t l.oui8, CV MO$by Co. 1969. "" 13-14.
`3. LauRe K. Boyd W : Use of Huo.-.e"""in
`method in ""t .. bli.hment of di8j(n (>Si. and pr<>g(cid:173)
`nO$is of peripheral vascular disease$. A"",h
`hlp", Moo 1944;14:175-18<1 .
`4. H~ JY. Dollcry C1': Retinal iOrt txu(cid:173)
`da le.: A d inica l study by eol""r and nuor<"5(en""
`phoUllluphy. Q J M<id 19&1:33:117-1110
`5. Lauren"" DJR: A study of Ihe aboo,ption of
`dyu on bovine se rum albumin by the mdhod of
`""laTi'alion uf Hoo'h«n"". Biochem J im.
`:>1:168_180.
`6. Andersson LO, Rebnnr(lm A. F,.aker OI~
`StudicM on 'non-specific' bindillJ(: The nature of
`
`the bindinl/ or fluoresuin to bovine ""rum albu·
`min. Eur J BiocMm 1971;20:37hlSO.
`7. lanacone DC. ~'eI Mrg NT. ~·ed~.man JI..;
`Tritiated fluore!l<:ein bindinl/ to normal haman
`plasma proteins. A. rch Ophlhall'rloi 191>(1;98:1643-
`1645.
`8. Karush F. SOnenbe .... ),!: Interaction or
`homolOKOuS alkyl aulfau-s with bovine "".um
`.. lbum in. J Am Cite",&<; 1949;71:1369-1376.
`9. Hascheme)'er RH_ I-Iaschem .. yer AE: Met h(cid:173)
`od balIM on char~ or chal")i:e di~ltibution. in
`l1aschemeyu RI!. lIaschemeyer AE (edsl; I'ro(cid:173)
`tei_ A Guide toSttu/lIbJi Pltlln eol a.w Chemical
`Me1/tods. New York_John Wiley &- !;on. Inc, 197"J.
`pp 2.;4.2(;8.
`10. Karu . 1I F: The interaction of purified ant;(cid:173)
`bnd)" witb npticall)' i""merie hapten~. J Am Cite",
`$x 1%6;18;~ 19-:N26.
`II. Klotz I: Thc a pplicatio n of the law of mass
`
`action to hinding b)' pr<>teins: Interactions with
`calcium. Arch BiQ<;hem 1946;9:1Q9-111.
`12 Sea1.Chrd C. ScheinM'lI lH, A.mstr<>na(cid:173)
`SH: Physi cal chemi stry of prOl ein !(Ilution " IV.
`Thc eombinatio n or human SC rum albumin with
`ohlorid~ i"n. JAm Che", Soc 19liO;12:535-MO.
`1<1. Rockey JH: Equine anlihapt.. n antibody:
`The ~ubunih an d fragm ents of anti-/l-la<:tOl<ide
`antibody. J I::;r:p Med 1967:125:249-2'1:>.
`14. Wu WI-I , Rockey JI-I: Antivasopr~.in anti(cid:173)
`body: Chara<:teri .... tion o f high _a lfinity rabbit
`antibnd)' wilh li mited association constant hete.(cid:173)
`OI«'neil)·. BiochfmW'l' 1969;8,2719-2728.
`15. Gu}·ton AC (ed): Tnrbook oj Medical Phll~
`idOl/II. Philadtlphia. WB Saunders <A. 1976. pp
`."193-394.
`11;. Dollery CT. 1I<>d¥" JV. En~ 1 M: Studi"", of
`the .-.elinal ci",ulalion with Huoreseein. BT M,"'- J
`1962:2:1210-1215.
`
`Arch Ophthaimol-Vol 100. March 1982
`
`E.quilit)rium Di"lys is -Li & Rockey
`
`487
`
`Apotex v. Abraxis - lPR20 18-00 152, Ex. 1007, p.04 of 4
`
`