IM14 I· I :1:1'i /AA/41Uf•·Ul°1i:!$O:!JMl/O
`TKANS�l.ANTATION
`( '11pyrii:h1 c:• l�!\l'i h\' Tfw \\'illiam< & Wilkin< ('o
`MONOCLONAL ANTIBODY THERAPY
`
`Vol.41. No. a
`/'rmto•d 111 l l.S.A.
`
`
`
`ANTIBODIES TO OKT:� ARISINC: DESPITE 1NTEN3E
`ANTl-II>IOTYPIC ANO NON-ANTl-IOIOTYPIC
`IMMUNOSUPPRESSION1
`
`GREGORY J. ,JAFFERS, THOMAS C. FULLER, A. BENEDICT COSIMI, PAUL S. RUSSELL,
`HENRY J. WINN, ANO ROBERT B. Cor.vrN'.!
`
`Tronspla11tati1111-lm1111mri/11;:y 11ncl lmm11n11path11/11;:y I /nits, lJepnrlm1•11/,, 11/ S11r;:1•ry n11d Path11/11�\'. Ma.<.<achusells G1•rwrnl H11spital
`and Harvard Ml!d1cal Schwl. H11.,t1m, Ma.,,;arhu.w?ll.� 02114
`
`nomenon has acquired new relevance with the application of
`
`The frequency, timing, and specificity of the humoral
`
`antibodies to clinical medicine. De­
`
`antibody response to a murine monoclonal antibody
`murine hybridom11-derived
`(0KT3, lgGi.) were measured in 21 consecutive renal
`spite some init inl optimism, the injection of the1;e murine
`
`
`allograft recipients. These patients received i.v. OKT3,
`
`
`protein1; into man. a1; expected, reh'ularly provokes an immune
`response hy the host ( / - / / ) .
`
`1-5 mg/day for J 0-20 days as treatment for acute graft
`
`
`rejection. Maintenance immunosuppression consisted of
`
`We have had extensive experience with the use of one mono­
`
`and corticosteroids. Using three different
`azathioprine
`
`clonal antibody, OKT:\ ClgG" .. �. BALH/c), used as an adjuvant
`assays, an antibody response was detected in 75% of the
`
`to azathioprine and prednisone for treatment of acute renal
`20 patients with adequate samples. The ELISA assay of
`rejection (I, 2, //). This antibody, reactive with an
`allograft
`the overall JgM and lgG reactivity to OKT3 revealed
`invariant component (T;W of t he T cell antigen receptor com­
`that lgM anti-OK1'3 appeared in 65% and IgG anti­
`plex ( /:ll, lrns proved to he dram atically effective in reversing
`OKT3 in 50% of the patients, reaching a peak 20-33
`acute cPllular rejection when ndministered at doses of 1-!i mg/
`days after the last dose of OKT3. The lgM preceeded the
`day ovt'r Ill 10 days. This report describes
`lgG in most cases (P<0.02) and in 8 cases was detected
`in detail the speci­
`ficity of the antibody response to OKT:l in 21 renal allograft
`during therapy. One patient had high levels of IgM anti­
`
`OKT3 before therapy, yet responded normally to OKT3.
`
`rel·ipients. and the implications for therapy. A preliminary
`
`
`Interference with the therapeutic effectiveness was ev­
`(6, 7).
`report of the lirst 11 patients has been presented
`ident in one patient who developed IgG antibodies dur­
`
`ing therapy. His serum blocked the binding of F-OKT3
`MATEHIALS AND METHODS
`to normal lymphocytes in the presence of normal
`BALB/c serum. The blocking assay, done by flow cytom­
`Patient.�. Thl• 11 con:=;ecut ivr OKT:l-treated patients studied
`
`etry, measured anti-idiotypic (Id) reactivity since the
`were initially treated with prednisone And azathioprine after
`sera did not affect the binding of OKTS (another lgG2_)
`a cnclavl'r clonor renal alloi:raft (I. 2). They were given
`rt•ct'iving
`or anti-Leu'& (another anti-T3), und the blocking activ­
`Raritan, N.J) when they had
`OKT:{ <Ort ho Pharmaceuticals,
`ity remained after affinity absorption with normal
`their first cpi!mde nf renal allograft rejection. The ant ihody was
`
`
`mouse IgG. Using this assay, 60"/,. of the pateints made
`given once dnily as a 1-f>-mh' i.v. bolus. No patient manifested
`an anti-Id response. One made only anti-Id, and several
`a skin react ion to intradermal OKT:I (O. l µg) prior to treatment.
`had anti-Id at times when other reactivities were unde­
`
`In an attempt to modify thl' immune response to OKT3, 6
`
`
`tectable. Antibodies to non-idiotypic, presumably iso­
`typic, determinants represented on OKT8 occurred in
`
`patients ahm rerrived rydophosphamide (patients N()s. IO and
`only 44%, while other reactivity (0KT4; lgG2bK) was
`17-�ll.
`less common (12%) and weaker.
`.'frrum snmpll's. Hloo<l samples were collected in acid-cit rate­
`
`While no adverse allergic reactions occurred in this
`dextrose (ACD) at lt>ast twicE' weekly beginninl{ a t the time of
`�roup of patients, the anit-Id antibodies, which are a
`
`transplantation and continuing for up to ten months after
`prominent feature of the immune response to this and
`OKT:I therapy. An averagE' of' 19 samples per patient were
`probably other monoclonal antibodies, can block their
`analyzed in :W patients (range l:l-14). One patient (No.;{) had
`
`therapeutic effectiveness and can arise despite intense
`
`only :1 posttreatment samples (all negAtive) and therefore was
`
`immunosuppression. This response may require the use
`
`not induded in further analysis. The samples were centrifuged
`of different idiotypes for prolon�ed or repeated courses
`at I !"100 rpm and the plasma was stored at -:lD°C. Peripheral
`of therapy and may be the major obstacle to the use of
`human monoclonal antibodies.
`hlood lympl,ocytes from these samples were used for immuno­
`as described previously (I).
`logir moni111rin1'(
`1�·11zy11w-link1•d immur10.<11rbn11 assay (f;LJSA). This assay,
`serum proteins has The parenteral administration of li>reil{n
`
`
`performed by a sli�ht modiliC'at ion of puhlisbed met hods ( /3),
`het>n known since the early days of scrot herapy to stimulate an
`was used to cll'tect hoth anti-Id and non-anti-Id anti-OKT:l
`im nn1ne response. The typical allergic manifest at ions include
`skin rash. anaphylaxis, and serum sickness. This phe-
`f'l'ver.
`
`
`antibodies. Polyvinyl vinyl plastic microtiter plates (imulon I,
`• Ahhrt>\'ialion,. lht>d: Al'D. al'id 1·itralt• dexJro.-;t>: ('AF,. CHAI.HJ
`' This work was supported h�· lJSl'HS (;rant H L-18">41i and by funds
`pru\'idt•cl hy llw Ori ho l'harma«C'Ut ic-al C'urpurnt ion.
`cxA/.JJF,: El.ISA. t>niynw·linkl'd 1111munosorhen1 assay: F·. lluores­
`t·l'in conjui:at('(I (t'.J:., F-OKT:IJ: Id. idiulype: MFC. median lluorese!'nl'!'
`,..f1011lcl ht• adrlres,..ed 111 H.B. Colvin, M.D .. Im·
`
`: lh•print rl'qllt'"'"
`nn111u11a1holu),[y lJnil. llt'partnwnl of Pmhnlui.ty. Massal'huse11s (;<'11·
`<'hanilt'I; l'Bs. phosphau .. hufft•n•d ,.alirw: l'HA. phytolwmai:i:lutinin:
`t•rnl Hos11ital. Hoslon. MA 0:!11-1.
`Tl. 1 lw :!II :.l!"> kilodailon surfan· anl ii:C'n of millure T <·ells .
`...... ,
`., ,_
`
`1 of 7
`
`BI Exhibit 1027
`
`

`

`May, 19�
`
`JAFFERS ET AL.
`
`sn
`
`Dynatech, Alexandria, VA) were coated with 1 µg/ml of purified
`e.g .. anti-HLA. Positive values were defined as 2 SD units
`parenteral grade OKT3, 0.05 M NaHC0:1 buffer (pH 9.6). The
`above the mean of 21 normals.
`columns. OKT3 and normal mouse IgG (Miles) were
`plates were incubated overnight at 4 °C then washed with phos­
`Affinity
`coupled to CNBr-activated Sepharose 4B (Pharmacia) by
`phate-buffered saline containing .05% (v/v) Tween-20 (0.01 Na
`standard techniques, with approximately 10 mg of protein/ml
`phosphate, 0.14 M NaCl, (pH 7.4) (PBS-Tween). Each serum
`swollen gel. Columns were 12X0.9 cm (mouse lgG) and 4.5x0.9
`sample was diluted 1:240 in PBS-Tween (5 µl/1.2 ml buffer).
`cm (0KT3). Fractions were eluted in PBS and then 3M KSCN,
`200 µI of this mixture was added to the wells. The plates were
`dialyzed, and reconcentrated to the starting volume by positive
`incubated for 2 hr at room temperature, washed with PBS­
`pressure filtration (Am icon XM-50).
`Tween, and 200 µI of a 1:500 dilution of alkaline-phosphatase­
`(Id) as.�ay. The basis of this assay is that certain
`conjugated goat antihuman IgG (Kirkegaard-Perry, Gaithers­
`Anli-idiotype
`anti-Id block the binding of the Id-bearing antibody with an­
`berg, MD)-or !{Oat antihuman IgM (both heavy chain specific)
`tigen. The test was done by flow cytometry using fluorescein­
`was added at a similar concentration. Anti-IgG with light chain
`ated OKT:l (F-OKT3) as the Id and T3+ T cell blasts (or
`reactivity was used in early studies as a screening test. These
`normal peripheral blood lymphocytes) as tbe antigen in the
`solutions were incubated for 2 hr at room temperature and
`presence of the test serum. Normal BALB/c mouse serum was
`washed again using PBS-Tween. Finally, 200 µl of I mg/ml p­
`added to block non-anti-Id antibodies. The controls included
`nitrophenyl phosphate in 10% diethanolamine was added. The
`normal human serum and fluoresceinated OKTS (F-OKT8).
`resultant color change of substrate was read on an ELISA plate
`The amount of F-OKT3 required to saturate PHA blasts or
`spectrophotometer (Bio-Tek Instruments) at 40fi nM. Sera
`from 15 healthy volunteers and assays without OKT3, serum,
`peripheral lymphocytes was determined by titration with a
`constant number of cells, plotting the MFC-vs.-concentration
`or anti-lg were used as controls. All assays were done in
`of F-OKT:l used (Fig. 1). Saturation was achieved at 8 µg/ml
`triplicate. Each daily run was standardized by running aliquots
`of F -OKT:l: beyond this point no further shift in the MFC was
`of a known postive control. This serum had a net OD .. i:. (test­
`achieved. A point of about 30% saturation was selected (2.5 µg/
`blank) of about 650 and 70 for the lgG and IgM assays,
`ml) for use in the assays to maximize sensitivity. A similar
`respectively. Standardized values for IgG were obtained by
`point was determined for F-OKT8.
`multiplying the net OD.111. of the test sample by 650/(00,,,, of
`The assay was performed by adding 25 µI of test or normal
`standard) on that day (or 70/0D.11., of standard for lgM). A
`serum to 20 µl of BALB/c serum and incubating for 30 min at
`sample was considered positive if the corrected OD., .... was 2 SD
`room temperature. Subsaturating amounts of F-OKT3 or F­
`above the mean of the controls (95th percentile) in the case of
`the pretreatment samples, or if the on ....... was more than double
`OKT8 were added and these were incubated for 30 min on ice
`
`
`Normal buffy coat peripheral blood lymphocytes (25 µI), pre·
`
`the pretreatment oo •. �. in the case of subsequent samples
`(provided the OD,,,., was >100).
`viously prepared from ACD blood, were added to each tube
`fl HA blast pr<'paration.
`with further incubation for ao min at 4°C. Residual erythro­
`The preparation of phytohemagglu­
`cytes were lysed with NH.Cl and the cells were washed twice
`tinin (PHA)-stimulated T cell blasts was accomplished hy
`methods previously described ( /4). Briefly, normal peripheral
`in PBS. Samples were analyzed using the Spectrum III flow
`mononuclear cells were isolated on Ficoll-Hypaque gradients
`cytometer gated on the lymphocyte or lymphoblast cluster. The
`and stimu!ated with 1 µg of PHA per JO" lymphocytes in 1 ml
`MFC of the positive cells above the positive threshold channel
`of RPMI 1640 medium containing 5% normal human serum,
`was taken as the index of staining intensity. Pretreatment
`antibiotics, and 5 µM mercaptoethanol. Cells were maintained
`samples had inhibition of 7±7%. Therefore a decrease in the
`MFC of greater than 20% of the normal serum control, provided
`by dilution to 101'/ml in T cell growth factor medium containing
`there was not a concommitant inhibition of F-OKT8, was taken
`interleukin-2 about every third day.
`as evidence for the presence of anti-Id antibody. Samples
`Using flow cy­
`Enhancin� assay for rnJn-Anti-ld antibodies.
`tometry, non-anti-Id human antibodies that included isotype­
`specific reactivity to murine IgG". were detected by the binding
`to OKT8-coatE"d (lgGt.K,CAF,) PHA blasts. Antibodies not
`specific for Id or isotype were similarly measured with OKT4-
`coated (lgG""A,CAF,) targets. OKT3-coated targets were used
`as a positive control. PHA blasts (which contained both T:3•T4+
`and T:rTs+ cells) were suspended at a concentration of 10°/ml
`in RPM! 1640. Nonfluoresceinated OKT8, OKT4, or OKT3
`(0.5 µg) or PBS was added to each tube containing 50 µl of
`PHA blasts. The suspensions were incubated at 4°C for :JO min
`and washed twice with PBS. 50 µI of test or normal serum was
`added and the mixture incubated and washed as before. Flu­
`oresceinated goat antihuman lgG (heavy and light chain reac-
`1ive) was then 11dded and the tubes were incubated and washed
`as before. The median intensity of the positive staining was
`evaluated using the Spectrum Ill flow cytometer (Ortho Diag­
`nostics System, Westwood, MA) ( /.'}). The percentage of in­
`crease in the median fluorescence channel (MFC, linear units)
`above uncoated blasts (no monoclonal antibody) for each sam­
`ple was calculated. This corrects for any contribution by anti·
`T-cell antihodies that may be present in the patient's serum,
`
`75
`
`�
`
`� � -
`� t1
`t3 � :::i �
`
`50
`
`25
`
`10 20 40 eo 160 320 640 12eo 2!J60
`DILUTION
`I. 1'101 of t1uore!>rence intensity (MFC, lineur unils) vs.
`F1<a11u:
`dilution or F-OKT:l on l'HA lym1>hohlasts. A point of uhout :!0%
`�aturntion wns >'eler1ed for thl' anti-Id inhibition assay.
`
`2 of 7
`
`BI Exhibit 1027
`
`

`

`574
`
`TRANSPLANTATION
`
`Vol. 41. No . • 5
`
`OKT8 and OKT3 enhancing reactivity were disparate in 2
`patients. The serum from patient No. I reacted with OKT8
`coated, hut not OKT:l coated cells. She had blocking anti-Id
`antibodies and it is possible that these displaced the OKT3
`from the target cell, giving a false negative. A similar phenom­
`enon may explain the negative OKT3-enhancing results in
`patients Nos. 2 und 17 who also had blocking anti-Id. A i:;econd
`patient !No. 14) had the opposite pattern, reacting with OKT3-
`coated, but not. OKT8-coated cells. The restricted reactivity in
`this case may have been due to anti-Id of the nonblocking type.
`Hloc:kin;: anti-Id assay. None of the pretreatment sera specif­
`ically blocked the binding of F-OKT:-1 or F-OKT8 to normal
`lymphocytes (Table I). However, after the course of OKT:l
`therapy. 12 patients (60'};,) developed antihodiei:; that selectively
`blocked the antigen-binding of F-OKT:l. The sera blocked F­
`OKT:l in the presen('e of normal BALH/c serum and did not
`block F-OKT8 or F-anti-Leu 4. arguing that the specificity was
`restricted to idiotypic-and not allotypic or isot.ypic--deter­
`minants on OKT:1 IFiJ!. 2).
`
`TAHU: :L Onset of anti-OKT:I tintihodies and effect on OKT:I
`response"
`
`r:LISA- Hirn··
`OKT:I
`kini:
`
`Enhnnl'inl(
`
`Pa1it·nt
`
`Response lo
`OKT:I («ells/
`mm3>d
`
`containing no BALB/c serum were analyzed in parallel. F-anti­
`Leu 4 <Becton Dickinson Monoclonal Center, Mountain View,
`CA) was used for comparison.
`Statistical analysis. Values given are mean ± SD unless
`otherwise noted. Student's t and Fisher's exact tests were used
`for evaluating statistical significance. An antibody response
`was considered positive if two or more sample values were
`above the defined threshold, either a doubling of the pretreat­
`ment value or above the 95th percentile of normal controls.
`
`RESULTS
`ELISA: total Anti-OKT3 response. Using the ELISA with
`OKT:l as the antigen, the lgG and lgM antibodies of all
`specilicities to OKT:3 were determined. Except for patient No.
`18, the ELISA reactivity of the patient's pretreatment i:;erum
`was similar to that of t he If> controls. hoth for IgG and for lgM
`(Table l) and the anti-lg with light chain reactivity gave
`essentially the same result as anti-')' chain.
`Patient No. 18 had the highest pretreatment reactivity, which
`was exclusively lgM. This hinding could be completely elimi­
`nated by preincubation of the tei:;t serum with BALB/c serum,
`and partially blocked with preincubation in normal goat serum.
`A rheumatoid factor assay with human lgG was within normal
`limits (<60 l.lJ.). The level of the lgM antimouse lgG antibody
`did not change hy more than !)";, during treatment, and the
`patient responded normally to the OKT:-1 therapy (Table 2).
`An anti-OKT:{ antibody response, defined as doubling of the
`pretreatment oo ... ! .. wns deteC'tahle at various times post treat­
`ment in 14 (7()';;,) of t hose individuals (Tables 2 and :H. Four
`h11d only lgM and one had only lgG. The lgG was detected
`significantly later than the lgM by an average of 10 days,
`although the time of the peak level was the same (20-2:l days)
`(Tables 2-4). Antihorlies were detected during OKT:S adminis­
`tration in 8 patients. All 8 had lgM antibodiei:; and 2 also had
`lgG. Only one patient (No. 4), who had lgG as well as lgM
`anti-OKT:I, showed resistance to the pharma<·ologic cfft•cts of
`OKT:l. The other 7 cleared their T C'ells normally despite the
`presence of lgM anti-OKT:l. Only JO':;, (fgG) and :l:l% <lgM)
`of the responders had persistent antibodies when checked 6
`weeks or longer after the last dose. Two had detectable levels
`14 weeks posttherapy (Nos. 7, l!J).
`Enhandn;: assay. This assay was used to detect ii:;otype and
`light chain reactivity by ass11ying T8• and T4• PHA blasts
`coated with OKT8, OKT4, or OKT:S antibodieR. Three patients
`(Nos. 10, 12. and 14) had pretreatment reactivity above the
`9f1th percentile (Tables I and 2). Of the 16 tested, 7 (44<;;,)
`showed a two-fold or greater increase in reactivity to OKT8
`{lg(;" .. �)-conted cells. Only two (12%) hound (both weakly) t.o
`OKT4 (lg(;,.,,)-coated cells. suggesting that the reactivity de­
`teC'ted was primarily to isotype specificities on the heavy chain.
`Tht• timing of the anti-isotype was closer to that of the lgM
`than the lg(; (Table 4).
`
`:!
`:I''
`·I
`;,
`Ii
`I
`8
`!I
`Ill
`II
`l:!
`1:1
`l·I
`I�>
`Iii
`J'i
`Ill
`l!I
`:.!O
`
`I�-<; li:M
`
`I!) .j
`-·I
`1:1
`
`Anli· 01\T:I 01\TX OK'l't T:l THTI\
`l<l
`ll
`�:,
`
`ll
`
`()
`II
`
`-:!
`7
`x
`
`:, -:t
`fi
`-ll
`
`:.!x
`
`()
`:!
`;,
`
`. ,
`
`-;
`
`"
`
`"
`
`10
`
`·I
`p-
`
`:.!
`·I
`p-
`
`p-
`•)
`
`p:I
`
`p-
`:!
`p-
`
`:.!
`:11 -JO
`-:.
`-I
`
`:!
`x
`
`Ill
`
`l ·I
`ti
`
`p-
`-IO
`-ti .,,,
`
`p-
`4;,
`
`72
`p-
`
`:1
`1r,:1
`I:"> :!16
`40:.!
`7
`.177
`!">:!·1
`n
`·II
`() (!);\
`I
`!IX
`:1
`191
`()
`1:10
`;, f1(i
`;,
`80
`() :1.1
`:l!):l
`l
`:r;9
`16
`:n
`24:?
`129
`·I
`4:16
`l i
`JI
`:?i-1
`:IO
`2:11
`MIX
`:.!·I
`I )Of,
`:!I
`"Thi· clay wh1·n antibody wa" first d1•tt•ctrd (0 = last clay of OKT:l
`t h1•rap�·: rwt.:at iH• numlwrs-da�·s lwfon • I h..rap_v end('(!).
`'Onlv :1 post! rl';tt ment sampl1•s.
`'Abhrt>viations: p = prt>treot1111•111 �ampll' positiv!'. >!lf>th per<'entile
`of normal control�: - = m·1wtiv1•: blank= not donl'.
`"Last blood values on OKT:I.
`
`TAHU: I. Pr1•treatment rPil<.'livity
`
`Enhnndng As�HY
`4(·,· increa:wl
`
`Hlu.-kin,: assn�·
`( i·; inhihi1 ion l
`
`Ii:
`
`li:C
`
`ll(M
`
`OKT:I
`
`OKTX
`
`OKT4
`
`Pat ii·nts
`J.l·l±:.!!-.17
`:lll±XO
`l>O±!l:.!
`:!i±:l-1
`('0111rofs
`Wx±lf>I
`IH±:.!l
`·-·--·
`"Excludes patit>nt No. 18 whos1• n1l1ws were t:l!q !Ii:>. ! Ill (lg(;J. and 171/l llgM) on tht> :1 ELISA assays. lg wns a scr1>ening test usinl{ goat
`anti human inununoglobulin n•m ·t ivt' to ) !wavy 1·hain and hoth lit.:ht 1·hains. whill' lg(; and h:M wt-re dt'tt'l'ted hy hl•1wy-diain-specifir ant isera.
`
`:16!1±:.!l!I
`:t:.t!:t I :!f,
`
`;1:1!1±141
`
`!i9±7(i
`:.!9±:10
`
`OKT:I
`7±7
`
`OKTI\
`
`8±11
`
`3 of 7
`
`BI Exhibit 1027
`
`

`

`May, 191:i6
`
`.JAFFERS ET AL.
`
`575
`
`TAHU: :1. Maximum antihody level"
`. -·-----·----------
`Enhancinj.! (% increase
`Hlocking:
`ELISA-OK'J':I: OD..,.
`MFC>
`. ·-- - -- - --------------
`Anti-Id t% In·
`lgM
`OKTA
`OKT:I
`OKT4
`hihition)
`- ----------------------
`- 1122(19)
`(;:14( 19)
`7fi( 8)
`.59fi(.59l
`176Jllfi) 14:.!(7.'))
`l:i9(2.S)
`
`1:114(14) 460( f>)
`172:�(1fi) 221)(10)
`-
`2921 l:i)
`
`1971(12) 5583(12)
`7/'i:NIOJ AAl(lOl
`
`f>99( ;39 l
`
`60H(-l.5)
`
`- 869< 1:n
`!IOO(fif>l
`;,:l8(29)
`6f'>f>( -I )
`.'J/'i.51-J)
`
`81if>(:IO)
`
`.5971 II)
`
`8:!(14)
`91(/2)
`92(1-1)
`
`94( l'i)
`:l�I( 1 Ol
`
`f)69( 2)
`:158( 1)
`489( 7) 178( 7)
`1116(14) 118:1(19)
`
`640(22) 200(22)
`400113) 263(1 :1)
`
`- 94(52)
`
`89( 2)
`
`.')f)(l4)
`
`1'200(76) 1500(76) 67(76)
`1067( :l)
`
`AA!22l
`92(1G)
`
`f)8(4!i)
`
`Pt1tien1
`
`•)
`:1
`4
`f)
`I)
`7
`8
`9
`10
`II
`I:!
`1:l
`14
`lfi
`Hi
`17
`18
`I 9
`
`:w
`21
`
`• Uay of maximum !after OKT:l therapy ended) in parentheses.
`•Assays in iwlks were positive in the last sample 1.ested (more than
`6 weeks from the last OKT:I dose).
`
`absence of RALB/c serum. The analogous eluate from patient
`No. 4 had no blocking activity. In contrast, the effluent blocked
`F-OKT:l completely but had little activity against F-OKT8,
`even in the absence of added BALB/c serum (Fig. 4). The
`effluent (patient No. 4) and the starting serum were then
`absorbed on an OKT3 column. This removed the blocking
`activity, which was undetectable even after reconcentration.
`Taken together, these data indicate that the blocking activity
`was due to antibodies that reacted selectively with OKT3 and
`not to other mouse immunoglobulins, satisfying the criteria for
`anti-Id specificity.
`The time course of the anti-Id antibody production was
`variable, but generally paralleled that of the IgG (Table 4). The
`anti-Id antibodies first became detectable at an average of 1 1
`days (range 0-45) after therapy had ended. The most rapid
`antibody response was manifested at 12 days after the start of
`therapy in patient No. 4 (Fig. 3). At this time he became
`refractory to the T lymphopenic effect of OKT3. The only
`clinical signs of an allergic reaction were an eosinophilia of
`12% and a fever upon administration of the OKT3. No serum
`sickness lesions (glomerulonephritis, vasculitis, mouse lg de­
`posits by immunof1uorescence) were seen in renal biopsy taken
`during the OKT3 treatment. The patient suffered no adverse
`systemic effects from this response, although the rejection
`process became reactivated and the allograft was eventually
`rejected. This was the only patient whose therapy was stopped
`hecause of anti-OKT:� antibodies. Three other patients had
`anti-Id antihody in the
`first sample analyzed, 2 days after
`therapy ended. Anti-Id antibody persisted in 33% of patients
`after 6 weeks; the longest documented was 6 months (No. 8).
`Of the 12 patients that made a blocking anti-Id antibody, 3
`(25%) had no IgG and 2 ( 17%) had no JgM reactivity detectable
`by ELISA. Similarly, 30% of those with blocking anti-Id had
`no isotype/light chain reactivity detectable by the enhancing
`assay. Patient No. 17 made exclusively an anti-Id response,
`detectable only in the blocking assay. Furthermore, the anti-Id
`assay detected antibody on one or more occasions in 7 other
`patients when the other assays were negative. This can be seen,
`for example, in the time course of the antibody response in
`patient No. 8 (Fig 4), in which the anti-Id response was persis­
`tently positive when the ELISA and enhancing assays were
`negative. Two patients (Nos. 11 and 15) made anti-OKT3
`antibody detected on ELISA that did n'ot block antigen binding
`or have reactivity to OKT8. These antibodiel-l may include
`nonblocking anti-Id.
`To estimate the portion of the antibody response that was
`anti-Id, serum from patient No. 8 at the peak of the antibody
`response (day 12, Fig. 4) was analyzed by ELISA using normal
`BALB/c serum as a competitive inhibitor for non-anti·Id anti­
`bodies. Normal mouse serum (100 µI) was able to block 69% of
`the lgG binding to OKT:l, while 100% could he blocked by 10
`µg OKT3, so that anti-Id constituted about 3 1 % of the reactiv­
`ity.
`Correlations with clinical cour.�e. With the exception of the
`one patient (No. 4) who developed anti-Id antibodies that
`neutralized the in vivo effectiveness of OKT3, no adverse
`effects were associated with the antimouse l g or anti-OKT3
`idiotype response. The response to therapy and outcome of the
`graft were not correlated with the presence or absence of an
`antibody response. However, of the 6 patients who received
`cyclophosphamide (Nos. 10 and 17-21) only 173 made an lgG
`anti-OKT3 response detectable in ELISA, whereas 9 of 14 of
`
`TAHl.F. 4. Summary of time <'ourse of antihody respon!<e
`Numher -- - -·-·-·
`1:l
`10
`12
`7
`
`Pcuk•
`
`2:1±7
`20±5
`16±:!
`2:1±9
`
`Onset•
`
`l!!M
`li:C
`ldiotype
`lsotype
`•Day after last OKT:l dose. mean± SE.
`b Preceeds Ii:<:. /'<.0'2.
`
`O±:lb
`10±:1
`11±4
`4±1
`
`Further experiments were performed to document whether
`the effect was indeed due to anti-Id antibodies. The presence
`of residual free OKT3 in the serum was ruled out by the lack
`of detectable mouse lg on normal lymphocytes after incubation
`in the blocking serum. In all patients OKT3 antibody was
`undetectable in the serum by 2 days after therapy had ended
`(data not shown}. For this reason only sera taken at least 48
`hr after the end of therapy were analyzed (except for patient
`No. 4 who was demonstrated to clear the OKT3 within 7 hr).
`Serum that blocked OKT3 did not block F-anti-Leu 4, an IgG1
`antibody t.o a different epitope on the same molecule (Fig. 2).
`This argues strongly against free or complexed T3 antigen itself
`as a significant source of antigenic competition in the blocking
`assay.
`Sera from two of the patients were analyzed by affinity
`chromatography using columns with immobilized OKT3 or
`normal mouse IgG. The sera from patients Nos. 4 (Fig. 2 and
`3) and 8 (Figs. 4 and 5) were absorbed with mouse IgG on
`Sepharose. The effluent and the dialyzed 3M KSCN eluate
`were tested for blocking activity. The eluate from normal mouse
`lgG (patient No. 8, day 5) showed equivalent blocking of F­
`OKT3 and F-OKT8 of 88% and 83%, respectively, in the
`
`4 of 7
`
`BI Exhibit 1027
`
`

`

`NORMAL
`OKT3
`
`OKT8
`
`576
`
`10
`
`Leu4
`
`TRANSPLANTATION
`
`Vol . .//, N<>. ,5
`
`C4/13
`
`OKT3
`
`1 0
`
`1·1·1 1-·, ,-,-,. , 1·1-n
`1a
`�·�
`
`'"° �
`
`1 �0
`
`104
`1 '.""r'rl
`:?O
`
`I�
`
`OKT8
`
`Leu4
`
`FH:t11u: '2. Fluorescence histograms ofhlocking
`anti· Id assay. Normal lympho(·ytes stuined with
`fluores\'einated monoclonal antibodies in the
`prese1we of normal HALH/c serum and either
`normal human serum {left t•olumnl or serum
`from patient No.� !right column). The patient's
`serum hlocb F'·OKT:I. hut not F·OKTR (both
`are lg(;,,.,1 or F·anti·Leu 4 {li:<:l. reactive to the
`T:I antigen}. ('ell numher is on the y·axis and
`lluores«em·t· intensit�' is on the x·axis. The num·
`her in the lower right of ea(·h histogram is the
`median tluort'seence channel oft he positive cells.
`
`.JIM!����
`"--11��,.:..:.....,;:.........��,�
`
`..
`
`..
`
`ll't
`
`1.0 ,..
`
`l ..
`
`10
`
`..
`
`•
`
`llO
`
`I.. - l'ot
`
`Poti.nt 4
`
`Skin '"'e
`Ba: R•i•t!lon
`'
`
`Ba. Rei.tllon
`
`8Ml9G i
`
`ELISA aOK 13 (0040Y
`to.,. _ 'o
`lgO
`'114
`�Qf�!,.�G i;i� 11
`' 1nhol;Hhon
`
`]�omQ/d
`I ]�!>()mQ/d
`
`11� J�e
`,�, !>fJ
`
`8l
`
`4GO
`181
`
`.,
`
`PREOHISONE
`
`AZATHIOPRINE
`,.
`
`l!.tt
`
`OKT3
`
`T3+
`
`25
`20
`15
`10
`5
`OAYS POST TRANSPLANT
`FH;l't<�: :1. ('liniral ('ourse of pati<·nt No. 4. who developed blocking
`anti·ld during therapy and IX'came resistent to the T lymphopenic
`effects of OKT:I. Ii:<; and lgM ant i·OKT:I was detected during therapy.
`Hlo\'king anti·ld appearance was accompanied hy resistance to the T
`lymphopenic response to OKT:I (day:!:!}.
`
`the non-cyclophosphamide group made such antibodies
`(/'<0.076). Cyclophosphamide had little effect on the blocking
`anti-Id response. which still occurred in !JO'/(, of patients­
`although somewhat later than the others, peaking at 16-45
`days. No correlation with the recipient's HLA phenotype and
`the antibody response was found.
`
`0000
`
`SO<IO
`
`4000
`
`a � JOOO
`� iil
`.., 2000
`....
`� ..
`
`1000
`
`lgM a0KT3
`
`0
`
`"'
`�
`
`�iii
`
`� � 300 L'.:I l!; � � ; 100
`a�KA IC
`. " �
`100 ,.: '("'<
`l_� ·""""""· 1.&?Ki4 i I ,
`I � I:: r � � 40
`
`.
`:, .
`
`iC-:
`8" 10
`�
`01'_-,.---.. - ..
`'
`.
`.
`•
`-20 ·•O o 10 20 :ao co so 60 10 ao
`(0KT3l DAY AFTER tAST OKT3
`FHaai�; ·I. Time course of antibocly response in patient No. R, as
`dete\'ted hy each assay in this study. The horizontal bar is the thresh·
`hold of positivity <:! SD ahovt' tht• mean of normals). Note the early
`rise of li:M and the presence of both blocking anti· Id and anti·isotype
`<OKTH enhancini: assay). No antibodies to OKT4 were detectable.
`
`OISC'USSION
`In this series 60'};, of the patients treated with OKT3 devel·
`oped anti-Id antibodies and 44'7/, anti-isotypic antibodies, de·
`spite intense immunosuppressive treatment (Table !J). The
`
`5 of 7
`
`BI Exhibit 1027
`
`

`

`May, 1986
`
`JAFFERS ET AL.
`
`577
`
`The nonidiotypic reactivity detected by the binding of OKT8
`was probably isotypic, because little reactivity was found to
`OKT4. We cannot exclude the possibility that a portion of the
`OKT8 reactivity may be to allotypic determinants not ex·
`pressed on OKT4. OKT3 was derived from a BALB/c mouse,
`and both OKT8 and OKT4 were obtained from (BALB/cxA/
`J) F, mice ( /6). The strains of origin of the heavy chains in
`OKT4 and OKT8 are unknown. Several lgGi. allotypic deter­
`minants of BALB/c are expressed by A/J but none are appar­
`ently expressed on lgG2i. (17). Anti-isotype specificities have
`been detected in other patients treated with OKT3 ( /0).
`The presence of some degree of reactivity to mouse immu­
`noglobulin in the pretreatment sera and in normal controls was
`expected. Antibodies to equine A TG have been detected by
`passive hemagglutination in up to 50% of patients awaiting
`renal transplantation with no previous exposure to ATG (18) .
`These antihorse immunoglobulin antibodies were found to
`crossreact with goat and rabbit immunoglobulin. The specificity
`of these antibodies has not been established. They may be
`crossreactive antibodies to immunoglobulin antigens absorbed
`orally (beef, pork, etc.) or rheumatoid factors. We found no
`correlation between preexisting antibodies and the severity of
`side effects from monoclonal therapy. The one patient with a
`high titer prior to treatment responded fully to OKT3 therapy
`for rejection. A further analysis of these antibody crossreactiv­
`ities will be necessary to determine the importance of these
`interactions.
`Although early studies were optimistic that monoclonal anti­
`bodies might have low immunogenicity ( /, 3, 19), following our
`initial report (6), several other studies have documented the
`presence of both non-anti-Id and anti-Id responses (4, 5, 7, 9,
`10, 14, 15). The production of such antibodies may be influ­
`enred by the particular monoclonal antibody used and the
`immune reactivity of the individuals treated. More than half
`the patients treated for lymphoreticular malignancies with a
`nonmitogenic antibody (anti-Leu l) produce antimouse anti­
`body, similar to that of our transplant recipients. Although
`these antibodies do not cause immune complex disease, the
`response effectively neutralized the monoclonal antibody, as in
`our patients (15).
`OKT3 does not 1;eem particularly more or less irnmunogenic
`than other anti-T-cell or antitumor antibodies, despite its
`known mitogenic activity. The one distinctive feature of the
`immune response to OKT3 may be the predominance of the
`anti-Id response. Only 5% of the antibodies to the murine
`immunoglobulin have been to idiotypic determinants in lym­
`phoma patients treated with anti-Leu 1 (20), whereas in one of
`our patients only anti-Id antibodies were detected, and in
`another they accounted for 31 % of the antibody response.
`Overall, 75% of our patients who had any antibody response
`made anti-Id that was able to block antigen binding. Why
`OKT3 should be particularly effective at promoting an anti-Id
`response is unknown, but one could speculate that this might
`be related to the target molecule, a part of the T cell antigen
`receptor complex. However, the strong anti-Id response is not
`unique to OKT3, as anti-Id constituted about one-third of the
`antibody response of goats immunized with myeloma proteins
`(2/). Furthermore, we have observed blocking anti-Id in mon­
`keys treated with anti-Leu 2a (Cosimi et al., manuscript in
`preparation).
`That the recipients of murine monoclonal antibodies should
`develop an immune response to these foreign proteins is, of
`
`�
`,: 1.\..� .�, . ,.,., , ., ... ' . ' . . ..... .,
`
`u
`
`..
`• )
`••
`I�
`u•
`:01
`:•O
`Fu:URE 5. Rlocking assay of the mouse lgG absorbed serum from
`patient No. 8, day f>. Fluorescence histograms of T cell blasts stained
`with F-OKT:l (T�) or F-OKT8 (T8) in the presence of the effluent (E)
`that passed through the mouse IgG column or the control (C) which
`was the ultrafiltrate of the effluent. (passed through XM-50). (U)
`unstained cells. No BALB/c serum was added. The effluent completely
`hlocks F-OKT:l, in contrast to its slight inhibitory effect on F-OKT8.
`
`TABLE f>. Summary of response pattern
`
`Anti-OKT:J
`by ELISA
`
`Block in�
`anti-Id
`
`+
`+
`+
`
`+
`+
`
`+
`
`N
`
`7
`4
`:l
`
`!)
`
`Ant.iisotype•
`
`+
`
`° Four patients were not tested for antiisotype.
`
`response was transient and, with one exception, did not affect
`therapeutic outcome even though lgM antibodies frequently
`appeared before the completion of therapy.
`A number of experiments indicated that these patients did
`make antibody with anti-Id specificity. Th