`prOdUCtioO of functional
`chimaeric mouse/human antibody
`
`• IL-2
`...
`Spleen SN /""_.... ...
`
`·-•;::...-•=• anti-Tac
`
`1·
`I
`
`Q>
`
`....
`,.;
`§'
`0::
`""
`E-;'
`.;i:
`<,...
`0
`;::
`0
`]
`:a
`.=
`
`80
`
`60
`
`40
`
`20
`
`~
`
`0
`
`., , , .• 4F2
`
`'•
`
`·-~:,
`I 11(
`1_,
`1,,
`//
`
`... ==_ ......
`
`..,
`a::::.•
`
`Gabrielle L. Boulianne*t:I:, Nobumichi Hozumi*t
`& Marc J. Shulman*:!:
`
`* Department of Medical Biophysics, University of Toronto, and
`t Ontario Cancer Institute, 500 Sherbourne Street, Toronto, Ontario,
`Canada M4X IK.9
`:j: Rheumatic Disease Unit, Wellesley Hospital, Toronto, Ontario,
`Canada M4Y 1J3
`
`The availability of monoclonal antibodies has revived interest in
`immunotherapy. The ability to influence an individual's immune
`state by administering immunoglobulin of the appropriate specific(cid:173)
`ity may provide a powerful approach to disease control and preven(cid:173)
`tion. Compared with immunoglobulin from other species, human
`immunoglohulin (lg) might be best for such therapeutic interven(cid:173)
`tion; it might function better with the recipient's effector cells and
`should itself be less immunogenic. The success of the mouse
`hybridoma system suggests that immunoglobulin of virtually any
`specificity can be obtained from a properly immunized animal. In
`the human system, however, immunization protocols are restricted
`by ethical considerations, and it is not yet clear whether human
`antibody-producing cell lines of the required specificity can be
`obtained from adventitiously immunized individuals or from
`in vitro immunized cells. A method which might circumvent these
`difficulties is to produce antibodies consisting of mouse variable
`regions joined to human constant regions. Therefore, we have
`constructed immunoglobulin genes in which the DNA segments
`encoding mouse variable regions specific for the hapten trinitro(cid:173)
`phenyl (fNP) are joined to segments encoding human µ and 1e
`constant regions. These 'chimaeric' genes are expressed as func(cid:173)
`tional TNP-binding chimaeric lgM. We report here some of the
`properties of this novel lgM.
`The variable regions used in the experiments described here
`are derived from the hybridoma cell line Sp6 which secretes
`• The specific K gene2 and µ gene3
`IgM(K) specific for TNP1
`have been cloned and were used as a source of TNP-specific
`variable regions which were joined to cloned human µ, and K
`constant regions4
`5 in the vector pSVrneo6 (Fig. I).
`•
`To assay the chimaeric light- and heavy-chain genes indepen(cid:173)
`dently of each other, we transferred these genes into appropriate
`mutant hybridoma cell lines derived from Sp6. The vector pN · x(cid:173)
`KTNP bearing the chimaeric K light-chain gene, was transferred
`9 to the mutant cell line, igk14 (ref. 7).
`as described elsewhere1
`-
`This cell line has lost the ability to produce the TNP-specific K
`light chain (KTNp) but continues to synthesize the TNP-specific
`µ, heavy chain (Jl.TNp). In a similar manner, the vector pN- x(cid:173)
`µ, TNP bearing the chimaeric heavy-chain gene, was transferred
`to another mutant cell line, igmlO (ref. 3), which produces KTNP
`but no ILTNP- To produce the totally chimaeric IgM, we transfer(cid:173)
`red the vectors p N · x-K TNP and p N · x-µ TNP together into the
`cell line, Sp2/0 (ref. 10), which produces neither heavy nor light
`immunoglobulin chains. Transformants were selected for resist(cid:173)
`ance to the drug G418, then tested for their production of
`TNP-specific lgM, as measured by their ability to agglutinate
`TNP-coupled sheep red blood cells
`(TNP-SRBC). The
`frequency at which stable 0418-resistant transformants were
`generated was found to be 10-3
`_ In experiments involving a
`single vector-transfer, approximately 30% of the resistant
`transformants produced detectable lgM. In the co-transfer
`experiment, where the chimaeric K and µ, were introduced into
`Sp2/ 0, we estimate that 0.1- l % of the transformants produced
`enough of both ILTNP and KTNP to make detectable IgM. The
`selective advantage to co-transfer seems here to be significantly
`less than that reported by others using calcium phosphate co(cid:173)
`precipitation 11 or protoplast fusion 12
`• We have also transferred
`the chimaeric heavy- and light-chain genes sequentially: the
`heavy-chain vector, pN · x-µ, TNP, was first introduced into
`
`
`
`(cid:141)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:50)(cid:69)(cid:88)(cid:89)(cid:86)(cid:73)(cid:4)(cid:52)(cid:89)(cid:70)(cid:80)(cid:77)(cid:87)(cid:76)(cid:77)(cid:82)(cid:75)(cid:4)(cid:43)(cid:86)(cid:83)(cid:89)(cid:84)(cid:21)(cid:29)(cid:28)(cid:24)
`
`10- 0
`
`10-•
`Dilutions of ascites added
`
`10-•
`
`Fig. 4 Anti-Tac inhibits the IL-2-induced proliferation of S.
`aureus-activated B cells. Serial dilutions of anti-Tac or anti-4F2
`were added to 5 x I 04 spleen B cells that had been cultured for 3
`days with optimal concentrations of S. aureus in round-bottomed
`96-well microtitre plates. Spleen supernatant or recombinant IL-2
`were added subsequently at a final concentration of 7 U m1- 1 IL-2.
`Triplicate cultures were incubated for 72 h at 37 °C and 0.5 mCi
`of 3H-TdR was added 18 h before collection (see also Fig. I). Data
`are expressed as % inhibition of the response obtained in cultures
`receiving spleen supernatant or recombinant IL-2 alone (53,500±
`3,400 c.p.m. and 46,800± 1,400 c.p.m., respectively).
`
`We next investigated whether the IL-2-induced B-cell prolife(cid:173)
`rative response could be inhibited by anti-Tac (in the absence
`of complement). The B-cell proliferative response to recom(cid:173)
`binant IL-2 or spleen supernatant was stron~ly inhibited by the
`addition of anti-Tac at a final dilution of 10- of anti-Tac ascitic
`fluid whereas anti-4F2 ascites, used as control, had minimal
`inhibitory effect even when added at final dilutions as high as
`10-3 (Fig. 4). According to previous studies, supernatant of
`lectin-stimulated lymphocytes should contain lymphokines with
`BCGF activity in addition to IL-2. However, in our study, up
`to 90% inhibition of B-cell proliferation was achieved with the
`anti-Tac. Taken together, the present results suggest that IL-2
`may be responsible for a large part of the BCGF activity gener(cid:173)
`ated by PHA-stimulated human spleen cells. Our data do not
`rule out the possibility that molecules other than IL-2 display
`BCGF activity. For example, supematants of lectin-stimulated
`72-h lymphocyte cultures have been reported to contain strong
`BCGF activity but little TCGF activity2; in addition, T-T-cell
`hybridomas have been described which released BCGF and not
`14
`IL-2 13
`• In any case, limiting dilution analysis of the precursors
`•
`of IL-2-producing cells has previously indicated that as many
`as 60% of peripheral blood T cells have this functional poten(cid:173)
`tial 15, and it is therefore evident that a large fraction of human
`T cells can also influence B-cell responses via IL-2.
`We thank Dr J.C. Cerottini for helpful discussion, Dr A. S.
`Fauci for the 4F2 antibody and Dr B. Malissen for the B9-12
`(anti-HLA) antibody, J. Hosking for technical help and M. van
`Overloop for secretarial assistance.
`
`Received 13 August; accepted 8 October 1984.
`
`I. Smith. K. A. Immun. Rev. 51, 337-357 (1980).
`2. Fauci, A. S., Muraguchi, A., Butler, J. L. & Kehr!, J. H. in Prag. Immun. 5, 1053-1067 (1984).
`3. Howard, M. <I al J. exp. Med. !SS, 914-923 (1982).
`4. Leonard, W. J. et al. Nature 300, 267-269 (1982).
`5. Leonard, W. J .• Deppert J. M., Robb, R. J .• Waldmann~ T. A. & Greene, W. C. Proc. natn.
`Acad. Sci. U.S.A 80, 6957-6961 (1983).
`6. Korsmeyer, S. J, et al Proc. natn. Acad. Sci. U.S.A 80, 4522-4527 (1983).
`7. Zubler, H. R. et al. J. exp. Med. (in the press).
`8. Devos, R. et al Nucleic Acids Res. 11, 4307-4322 (1983).
`9. Muraguchi, A. & Fauci, A. S. J. Immun. 129, 1104-1108 (1982).
`IO. Mingari, M. C. et al Eur. J. Immun. (in the press).
`11. Kehrl, J. H., Muraguchi, A. & Fauci, A. S. J. lmmun. 132, 2857-2861 (1984).
`12. Hubbard, A. L. & Cohn, Z. A. J. Cell Biol 64, 438-460 (1975).
`13. Butler, J. L., Muraguchi, A., Lane, H. C. & Fauci, A. S. J. exp. Med. 157, 60-68 (1983).
`14. Okada, M. et al. J. exp. Med. 157, 583-590 (1983).
`15. Moretta, A. Eur. J. lmmun. (in the press).
`16. Moretta, L., Webb, S. R., Grossi, C. E., Lydyard, P. M. & Cooper, M. D. J. exp. Med. 146,
`184-200 (1977).
`17. Moretta, A., Pantaleo, G., Moretta, L., Cerottini, J.-C. & Mingari, M. C. J. exp. Med. 157,
`743-754 (1983).
`18. Malissen, B., Rebai, N., Llabeuf, A. & Mawas, C. Eur. J. lmmun. 12, 739-747 (1982).
`19. Laemmli, U. K. Nature 227, 680-685 (1970).
`
`1 of 4
`
`Celltrion, Inc., Exhibit 1117
`
`
`
`~644'-'--------------------L£TTER$TQNATLJRE-------~N_A_TU_R~E"--v~oL_._3_12_1~3-o_EC~EM_B_E_R_1_9_84
`
`a
`
`L l/H TNP
`
`SP6-718 ·~ """'ll11IH1-1 I-~ --1111~
`
`b
`
`B
`
`A
`
`a b c d e
`
`B
`Q b c d
`
`----"""'011111!111f+-41c
`~c1s 11-
`8
`
`h-CJJ.
`
`8
`
`1..Y"°TNP m-CK
`
`I I
`
`I l
`BH H
`
`, ......
`
`M
`
`8
`
`T111:I
`
`......
`'"
`
`-
`
`-1~1"'tl1
`
`sv•e
`
`ori
`M
`
`l
`VHTNP
`
`~ ,N·X·l-'TNP H
`
`n-Cµ.
`
`h-CK
`
`Fia. 1 a, Construction of the chimaeric heavy-chain gene. The
`3.5-ldlobase (kb) EcoRI-HindIII fragment containing the mouse
`TNP-specific heavy-chain variable region was obtained from the
`cloned TNP-specific heavy chain, Sp6-718 (ref. 3). This fragment,
`together with the 8.0-kb BamHI-Hindlll fragment containing the
`human constant region from the cloned segment AC75 (ref. 4), was
`introduced into pSV2-neo by ligation at the EcoRI-BamHI site.
`b, Construction of the chimaeric light-chain gene. The 5.4-kb
`BamHl-Hindlll fragment containing the human C, gen.e was
`introduced into pBR322 by ligation at the BamHI-HindllI site.
`The 4.2-kb HindIII-Hindlll fragment containing the mouse TNP(cid:173)
`specific light-chain variable region was obtained from the cloned
`TNP-specific light chain T. 1 (ref. 2) and introduced into the
`Hindlll site of the pBR322 vector containing the human c .. From
`this plasmid, pVCK, the 9.6-kb BamHI-Clal fragment containing
`the chimaeric light chain was obtained and introduced at the
`BamHl-Clal site of a modified pSV2-neo vector containing the
`BamHl-EcoRI segment derived from pBR322. The chimaeric
`light-chain gene was also introduced into the vector pSV2-gpt 13 in
`the manner described above. Restriction enzymes: B, BamHI; C,
`Clal; E, EcoRl ; H, Hindlll .
`
`Sp2/0; in a subsequent step, transfer of the light-chain gene,
`carried in this case on the vector pSV2-gpt 13
`, was selected by
`resistance to mycophenolic acid. Stable transformants which
`produced the highest levels (-5 µg ml- 1
`) of IgM were selected
`for further study and cloned by limiting dilution. The vector
`pR-HLTNP, which bears in their entirety the mouse genes for
`TNP-specific µ, and K chains, was also transferred to the cell
`line Sp2/0, (refs 3, 14) and the IgM made by one such transfor(cid:173)
`mant, TSp2/mlgM12, is compared here with the chimaeric IgM
`made by the transformant TSp2/ x- IgM 1. By using various anti(cid:173)
`sera specific for antigenic determinants of the mouse and human
`µ, and K constant regions, we confirmed that the chimaeric genes
`encode chimaeric proteins, that is, the TNP-binding capacity of
`the mouse IgM is linked to antigenic determinants of human µ,
`and K chains (results not shown).
`To determine whether the chimaeric IgM is pentameric, we
`analysed the biosynthetically labelled lgM secreted by these
`transformants using SDS-polyacrylamide gel electrophoresis
`(SDS-PAGE). The SOS-denatured chimaeric IgM from TSp2/ x(cid:173)
`IgM 1 migrates at nearly the same rate as pentameric mouse lgM
`produced by the hybridoma Sp6 and
`the
`transformant
`TSp2/mlgMl2 (Fig. 1A), indicating that the chimaeric genes
`produce µ, and K chains which combine to form pentameric
`lgM. After reduction of the disulphide bonds, chimaeric µ, and
`K chains migrate slightly slower than the corresponding mouse
`chains (Fig. 2B) . The molecular weights of the human µ, and
`
`-"
`
`Fig. 2 Production of immunoglobulin heavy and light chains in
`transformants. The transformants expressing the chimaeric µ, and
`K genes are compared with two cell lines making TNP-specific
`murine IgM: the parental Sp6 hybridoma (Sp603 subclone) from
`which the µ,TNP and KTNP genes were cloned, and a transformant
`(TSp2/ mlgM 12) expressing the transferred murine IA-TNP and KTNP
`genes (denoted as cell line Sp2/Tl 2 in ref. 14). The secreted IgM( K)
`of the indicated cell lines was biosynthetically labelled with 14C(cid:173)
`leucine
`described 23
`and
`subjected
`to SDS-PAGE
`as
`•
`lmmunoglobulin from cell
`lines Sp2/ 0 (a), Sp6 (c) and
`TSp2/ mlgMl2 (d) was precipitated with anti-mouse
`lgM .
`Immunoglobulin from the cell lines Sp2/ 0 {b) and TSp2/ x-IgM
`(e) was precipitated with anti-human IgM. A, Material was not
`reduced so that the immunoglobulin disulphide bonds are intact.
`B, Material was treated with 2-mercaptoethanol to reduce disul-
`phide bonds.
`
`K constant regions are nearly the same as the corresponding
`murine amino acid sequences 15
`; therefore the difference in
`mobility does not reflect simply a difference in molecular weight.
`Work is in progress to determine the reason for these differences
`in mobility.
`We used two methods to compare the binding sites of mouse
`and chimaeric IgM. First, we determined the affinity of these
`IgMs for TNP by measuring the ability of free hapten to inhibit
`the inactivation by these IgMs of TNP-coupled phage T4 (Fig.
`3); the association constant of mouse and chimaeric lgM for
`the hapten TNP-cap
`(2,4,6-trinitrophenyl-E-aminocaproic
`acid) was l.3±0.5Xl04 M- 1 and 1.5±0.5xl04 M- 1
`, respec(cid:173)
`tively. Within experimental error, we can detect no significant
`difference in affinity for TNP between the mouse and chimaeric
`lgMs.
`We also compared the affinity of these lgMs for each of several
`trinitrophenyl-like compounds by measuring the ability of these
`compounds to block the agglutination of TNP-SRBC by IgM.
`Figure 4 illustrates the results obtained for TNP-cap and DNPy(cid:173)
`Ala (3,5-dinitropyridine-13-alanine). The displacement of the
`inhibition curves indicates that the affinity of each IgM for
`DNPy-Ala is about threefold less than for TNP-cap. Table I
`summarizes the results obtained for other compounds. Again,
`these results suggest that the hapten binding sites are comparable
`in the mouse and chimaeric IgMs. The direct analysis of
`immunoglobulin structure predicted that immunoglobulin speci(cid:173)
`ficity should be unchanged when the same variable region is
`joined to different constant regions 16
`• This prediction has been
`verified both in studies of IgM and IgD in normal cells 17 and
`in analyses of the binding specificity of the immunoglobulin
`made by hybridoma cell lines which have switched in vitro 18
`19
`•
`•
`In addition, Sharon et al. 20 have shown that the substitution of
`a light-chain constant region for a heavy-chain constant region
`does not affect the affinity of the resulting protein. On the other
`hand, we can distinguish between the chimaeric and mouse
`lgMs using hapten inhibition of TNP-SRBC agglutination as a
`binding assay (Fig. 4). For each hapten, the curve for chimaeric
`IgM is displaced from the corresponding curve for mouse IgM.
`The phage inactivation analysis described above indicated that
`
`
`
`(cid:141)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:50)(cid:69)(cid:88)(cid:89)(cid:86)(cid:73)(cid:4)(cid:52)(cid:89)(cid:70)(cid:80)(cid:77)(cid:87)(cid:76)(cid:77)(cid:82)(cid:75)(cid:4)(cid:43)(cid:86)(cid:83)(cid:89)(cid:84)(cid:21)(cid:29)(cid:28)(cid:24)
`
`2 of 4
`
`Celltrion, Inc., Exhibit 1117
`
`
`
`Table 1 Relative affinity of mouse and chimaeric lgMs for trinitro(cid:173)
`phenyl-like compounds
`
`Ha pt en
`TNP-cap
`TNP-Lys
`DNP-Lys
`DNPy-Ala
`NIP
`NP
`
`mlgM
`1.00
`1.00
`1.00
`3.00
`>10
`>10
`
`x-IgM
`1.00
`1.00
`1.00
`3.00
`>25
`>25
`
`As described in Fig. 4 legend, we measured the ability of each hapten
`to inhibit agglutination of TNP-SRBC. TNP-cap was obtained from
`Dr G. D'Agostura; TNP-Lys (2,4,6-trinitrophenyl-L-lysine) and DNPy(cid:173)
`Ala were obtained from Research Plus; NIP (4-hydroxy-3-iodo-5-
`nitrophenylacetic acid) and NP (4-hydroxy-3-nitrophenylacetic acid)
`were obtained from Sigma and Aldrich, respectively. The affinity of
`each lgM for the indicated haptens, relative to TNP-cap, was determined
`by the displacement of the inhibition curves. mlgM, murine lgM; x- lgM,
`chimaeric lgM.
`
`these lgMs have the same affinity for TNP-cap. In this context,
`the displacement of the agglutination curves suggests that the
`binding of the chimaeric lgM to TNP-SRBC is different in some
`way from that of the mouse lgM. The significance of this
`difference in molecular terms is unclear. Crystallographic analy(cid:173)
`sis has suggested that there are several sites at which the variable
`region interacts with the first domain of the constant region 16
`;
`such interactions may affect the binding site differently in the
`chimaeric and mouse lgMs. On the other hand, agglutination
`is a complex process involving the binding of multiple lgMs at
`each of the multiple sites. Subtle differences in aspects such as
`flexibility of the constant region may affect binding and thus
`the sensitivity to inhibition by free hapten. The molecular stress
`necessary to effect agglutination may distort the variable region
`in ways which would not usually occur while
`the
`immunoglobulin is free in solution.
`We have also compared the chimaeric and mouse IgMs for
`their ability to activate complement. Culture supematants were
`titred for their ability to promote lysis of TNP-SRBC in the
`presence of a source of complement, that is, guinea pig serum.
`Compared with the haemagglutination titre, the haemolysis titre
`for chimaeric IgM was about fourfold less than that for mouse
`lgM; we do not know whether this reflects a difference in the
`intrinsic ability of these lgMs to activate complement or the
`difference in TNP-SRBC binding discussed above.
`Several therapeutic uses for specific antibodies have been
`proposed. The transfer of passive immunity by injecting specific
`immunoglobulins is a long-standing treatment. Othe~ uses are
`more speculative. In animal models, anti-idiotype antibodies
`have been used both to elicit and to suppress the production of
`specific antibodies by the recipient animal 21
`• These results might
`be extended to humans so that anti-idiotype antibodies could
`be used in some cases as a vaccine to enhance antibody produc(cid:173)
`tion and in other cases to depress the destructive immune
`responses which cause autoimmune diseases. The identification
`of tumour-associated antigens might lead to the production of
`monoclonal antibodies that selectively destroy tumour cells
`in vivo 22
`• However, as mentioned above, it may prove difficult
`to obtain human monoclonal antibodies having the specificities
`required for immunotherapy. Chimaeric immunoglobulins may
`provide a good compromise. For both monoclonal human
`immunoglobulin and chimaeric immunoglobulin, the constant
`region is expected to be non-immunogenic. We have no reason
`to expect the mouse variable region would in this form be more
`immunogenic than a human variable region of the same speci(cid:173)
`ficity.
`In terms of DNA and protein, the chimaeric antibody system
`works well. The regulatory signals for RNA transcription, initi(cid:173)
`ation, termination and splicing function, so that these genes
`express a high level of chimaeric µ and K chains. The specificity
`of the mouse variable region is preserved in the chimaeric IgM,
`and the human constant regions function so that the µ and K
`
`1.0
`
`0.9
`
`0.6 5
`
`0.4
`
`0.2
`
`· - · - • mtoM
`·--o X-IoM
`
`0
`
`0.01
`
`0.02
`
`0.5
`0.2
`0.1
`0.05
`Haplen concentration (M w; 10-4)
`
`1.0
`
`2.0
`
`5.0
`
`Fig. 3 Affinity of chimaeric and mouse lgMs for TNP. TNP was
`coupled to phage T4 as described elsewhere24
`• The affinity for
`hapten was calculated as follows. The rate-limiting step for the
`inactivation of phage T4
`is
`the attachment of the first
`immunoglobulin binding site25
`, that is, the concentration of surviv(cid:173)
`ing phage, <I>, is given by<I> = P exp (-aB) where P =initial hapten(cid:173)
`ated phage concentration before reacting with the anti-TNP, B =
`concentration of free binding sites and a = proportionality con(cid:173)
`stant. By incubating free hapten in the reaction mix, the concentra(cid:173)
`tion of free binding sites can be manipulated according to the
`formula K. = [hB]/[h][B], where K. is the association constant,
`[h] the hapten concentration, [B] the concentration of free binding
`sites and [hB] the concentration of binding sites bound to hapten.
`The inactivation index, Ih, is defined as log <l>(h)/ P where <l>(h)
`is the phage concentration after incubation with anti-TNP in the
`presence of hapten at concentration h. The figure plots the ratio
`Ih/ I 0 as a function ofhapten concentration where I0 is the inactiva(cid:173)
`tion index obtained in the absence of free hapten (h =0). Each
`point was determined in triplicate. The affinity constant for the
`hapten can be calculated from any point as K. = 1/ h[I0 / Ih -1].
`The values of K. given in the text represent the average of all points.
`
`>91
`9
`
`-~
`;;
`._§ 7
`~
`~
`~ 5
`-~
`~
`~ 3
`
`I 1
`I
`I /
`
`;
`
`. 1
`t
`, , ,
`J ,
`,
`,
`I ,
`, ,
`,
`i
`,J
`,. ... -"
`
`I
`I
`I
`
`0 8 - i - i
`0.02
`0.05
`
`1.0
`0.5
`0.2
`0.1
`Hapten concentration (M x 10-•)
`
`2.0
`
`5.0
`
`Fig. 4 Relative affinity of mlgM (e) and x-IgM (0) for TNP
`(--)and DNPy (- - - -) haptens. Culture supematants from the
`transformants TSp2/mlgMl2 and TSp2/ x-IgMI were serially
`diluted in twofold steps and incubated with either TNP-cap or
`DNPy-Ala at the indicated concentration in V-bottomed 96-well
`trays. TNP-SRBC were then added and the wells were scored for
`agglutination. The inhibition of agglutination, that is, the reduction
`in the number of wells with positive haemagglutination, is shown
`here for the indicated concentrations of each hapten. The displace(cid:173)
`ment of the curves for the different haptens indicates the hapten
`concentrations required to yield the equivalent number of free
`binding sites, and thus measures the relative affinities of each lgM
`for the two haptens26
`•
`
`chains are covalently bound to form pentameric lgM which can
`activate complement. We have, nevertheless, detected differen(cid:173)
`ces in the binding to TNP-SRBC of the chimaeric and mouse
`lgMs. Further work is needed to assess whether these differences
`have significance for immunotherapy.
`We acknowledge important contributions by Dr C. Heusser
`in the conception of this work. This work was supported by
`
`
`
`(cid:141)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:50)(cid:69)(cid:88)(cid:89)(cid:86)(cid:73)(cid:4)(cid:52)(cid:89)(cid:70)(cid:80)(cid:77)(cid:87)(cid:76)(cid:77)(cid:82)(cid:75)(cid:4)(cid:43)(cid:86)(cid:83)(cid:89)(cid:84)(cid:21)(cid:29)(cid:28)(cid:24)
`
`3 of 4
`
`Celltrion, Inc., Exhibit 1117
`
`
`
`~646=--~~~~~~~~~~~~~~~~~~-LETfER$TQ NATLJRf~~~~~~~~-N_AT_U_R_E_v_o_L_._3_12_1_J_D_E_C_E_M_B_ER~19_8_4
`
`grants from the MRC of Canada, the NCI of Canada, the
`Arthritis Society of Canada, the Wellesley Hospital Research
`Institute and the Allstate Foundation. G.L.B. is supported by a
`studentship of the NCI of Canada.
`ed 5 September 1984.
`
`Re.:eived 2 July; accepted 5 September 1984.
`
`I. Kohler, G. & Milstein, C. Eur. J. lmmun. 6, 511-519 (1976).
`2. Hawley. R. G., Shulman, M. J., Murialdo, H., Gibson, D. M. & Hozumi, N. Proc. natn.
`Acad. ScL U.S.A. 79, 7425-7429 (1982).
`3. Ochi, A. et al. Proc. natn. Acad. ScL U.S.A 80, 6351-6355 (1983).
`4. Rabbitts, T. H., Forster, A. & Milstein, C. P. Nucleic Acids Res. 9, 4509-4524 (1981).
`5. Hieter, P.A., Max, E. E., Seidman, J. G., Maize!, J. V. & Leder, P. Cel/ 22, 197-207 (1980).
`6. Southern, P. J. & Berg, P. J. molec. appl. Genet. 1, 327-341 (1982).
`7. Ochi, A., Hawley, R. G., Shulman, M. J. & Hozumi, N. Nature 301, 340-342 (1983).
`8. Schaffner, W. Proc. natn. Acad. Sc< U.S.A 77, 2163-2167 (1980).
`9. Sandri-Goldin, R. M., Goldin, A. L., Levine, M. & Glorioso, J. C. Molec. cell Biol. I,
`743-752 (1981 ).
`IO. Shulman, M. J., Wilde, C. & Kohler, G. Nature 276, 269-270 (1978).
`11. Wigler, M. et al Cell 16, 777-785 (1979).
`12. Rassoulzadegan, M. et al Nature 300, 713-718 (1982).
`13. Mulligan, R. C. & Berg, P. Proc. natn. Acad. Sci. U.S.A 78, 2072-2076 (1981).
`14. Shulman, M. J. et al. Can. J. Biochem Cell Biol. 62, 217-224 (1984).
`15. Kabat, E. A., Wu, T. T., Bilofsky, H., Reid-Miller, M. & Perry, H. Sequence of Proteins of
`Immunological Interest (NIH, Bethesda, Maryland, 1983).
`16. Amsel, L. M. & Poljak, R. J. A Rev. Biochem. 48, 961-997 (1979).
`17. Goding, J. & Layton, J. J. exp. Med. 144, 852-857 (1976).
`18. Neuberger, M. S. & Rajewsky, K. Proc. natn. Acad. Sci. U.S.A 78, 1138-1142 (1981).
`19. Oi, V. T. et al. Nature 307, 136-140 (1984).
`20. Sharon, J. et al. Nature 309, 364-367 ( 1984).
`21. Rajewsky, K. & Takemori, T. A Rev. Immun. l, 569 (1983).
`22. Levy, R. & Miller, R. A. A Rev. Med. 34, 107-116 (1983).
`23. Shulman, M. J., Heusser, C., Filkin, C. & Kohler, G. Molec. cell. Biol. l, 1033-1044 (1982).
`24. Haimovich, J. & Sela, M. J. lmmun. 103, 45-55 (1969).
`25. Adams, M. H. Bacteriophages (lntcrscience. New York, 1959).
`26. Williams, C. A. & Chase, M. W. Methods in Immunology and Immunochemistry Vol. 4
`(Academic, New York, 1977).
`
`Participation of pS3 cellular tumour
`antigen in transformation
`of normal embryonic cells
`
`Daniel Eliyahu, Avraham Raz*, Peter Grusst,
`David Givol & Moshe Oren
`
`Departments of Chemical Immunology and *Cell Biology,
`The Weizmann Institute of Science, Rehovot 76100, Israel
`tZMBH, Universitat Heidelberg, D-6900 Heidelberg I, FRG
`
`Table 1 Transformation of rat and Chinese hamster embryo fibroblasts
`by various gene combinations
`
`Transfected DNA
`
`Carrier
`(BALB/c DNA)
`pEJ6.6
`pMSVp53G
`pPyp53c
`pMSVp53G + pEJ6.6
`pPyp53c + pEJ6.6
`pLSVmyc + pEJ6.6
`pLA8+pEJ6.6
`pMSVE + pEJ6.6
`None
`
`Foci per I 06 cells
`
`REF CHEF
`REF
`expt 1 expt 2 expt 3 Tumorigenicity
`
`0
`0
`0
`ND
`13
`ND
`ND
`68*
`ND
`
`0
`0
`0
`ND
`5
`ND
`21
`ND
`0
`
`0
`0
`ND
`0
`ND
`20
`ND
`165*
`ND
`
`15/15
`
`9/9
`
`0/8
`
`Primary Fisher rat or Chinese hamster embryo fibroblasts were pre(cid:173)
`pared30 and maintained in Dulbecco's modified Eagle's medium supple(cid:173)
`mented with I 0% fetal calf serum and 4 mM L-glutamine (maintenance
`medium). 106 cells were seeded per 90-mm dish, and were transfected
`with the indicated DNA combination after I day (10 µg of each plasmid,
`made up to a total of25 µgwith sheared BALB/c liver DNA). pLSVmyc
`was constructed linking the 5.6 kbBamHI fragment of the murine c-myc
`gene (see ref. 14) to the SV40 early promoter. pLA8 contains the left-end
`9.1 % of the adenovirusl2 genome, including the EIA and part of the
`EIB region 16
`• pMSVE is a derivative of pMSVp53G, containing only
`the MSV enhancer inserted in the BamHI site but no p53-specific
`sequences (compare with Fig. I). The cells were transfected by the
`calcium phosphate procedure31 for16 h, glycerol-shocked (10% glycerol
`in maintenance medium for 90 s), replenished with maintenance
`medium, allowed to recover for an additional day, then split and
`reseeded at a density of 4 x I 05 (rat fibroblasts) or 2 x I 05 (Chinese
`hamster cells) per 90-mm dish. Medium was changed every 6-7 days.
`Distinct foci overgrowing the monolayer were scored 16 days (expt I),
`14 days (expt 2) or 24 days (expt 3) after replating ofttansfected cultures.
`REF, rat embryo fibroblasts; CHEF, Chinese hamster embryo fibro(cid:173)
`blasts. The tumorigenicity of cell lines established from corresponding
`foci was determined by injecting subcutaneously 5x106 cells info 5-8
`day-old Fisher rats whole-body irradiated with 200 rads. As a non(cid:173)
`transfected control (bottom line) we used REF propagated in culture
`to obtain a sufficient number of cells. Data are derived from two different
`p53 + Ha-ras lines and one myc + Ha-ras line. ND, not determined.
`* Foci possessing a distinctly different morphology from that induced
`by pLA8 alone.
`
`The cellular tumour antigen p53 is found at elevated levels in a
`wide variety of transformed cells (for reviews see refs 1, 2). Very
`little is yet known about the precise relationship of p53 to malig(cid:173)
`nant transformation. Although the increase in p53 levels could be
`a secondary by-product of the transformed state, It is equally
`possible that pS3 is actively involved in altering cellular growth
`properties, especially as it has been Implicated in the regulation
`of normal cell proliferation~. We sought to test whether p53
`could behave in a manner similar to known genes In a biological
`test system, and we demonstrate here that pS3 can cooperate with
`the activated Ha-ras oncogene to transform normal embryonic
`cells. The resultant foci contain cells of a markedly altered mor(cid:173)
`phology which produce high levels of pS3. Cell lines established
`from such foci elicit tumours in syngeneic animals.
`Recent findings have suggested certain similarities between
`p53 and the product of the oncogene myc. Both are DNA binding
`proteins (ref. 7 and D. Lane, personal communication) that
`accumulate in the nuclei of transformed cells 7
`8
`• Both are regu(cid:173)
`•
`10 and are induced at an early stage
`9
`lated with the cell cycle6
`•
`•
`following the treatment of resting cells with mitogens3
`5
`9
`10
`•
`•
`•
`•
`Cycloheximide-treated cells accumulate p53 11 and show
`increased myc messenger RNA levels9
`• Detailed analysis of the
`amino acid sequences predicted for the two proteins shows weak
`similarities in both the overall molecular organization and the
`positioning of charged residces within distinct domains of the
`13
`molecules 12
`• Hence, ifp53 can function like known oncogenes,
`•
`it is likely to do so in a manner similar to myc.
`
`A biological test system demonstrating the involvement of the
`myc product in malignant transformation has been established
`15
`recently 14
`; primary rat embryo fibroblasts are transformed
`•
`stably by the joint action of myc and another oncogene such as
`Ha-ras. Similar results are obtained when another nuclear
`oncogene, the adenovirus-2 EIA region, is assayed by co(cid:173)
`transfection with Ha-ras in an analogous system 16
`• In both cases,
`the transformation is visualized by the appearance of dense
`foci capable of overgrowing the monolayer of normal cells and
`are dependent on the presence of both oncogenes. This system
`is therefore a suitable test of the oncogenic properties of p53.
`Two types of recombinant DNA constructs were used as
`templates for efficient p53 expression in transfected cells (Fig.
`I). The plasmid pMSVp53G contains the 16 kilobase (kb) Eco RI
`17 jux(cid:173)
`fragment encompassing the functional murine p53 gene 12
`•
`taposed to the enhancer portion of the Moloney murine sarcoma
`virus (MoMSV) long terminal repeat. This approach utilizes the
`presence of a functional pr.omoter in the 16 kb fragment (B.
`Bienz, unpublished results). The second construct, pPyp53c,
`contains a stretch of p53 cDNA linked to the polyoma virus
`early promoter; this cDNA contains the intact coding region
`for p53 17 and directs the synthesis of authentic p53 in a
`heterologous system 18
`•
`Secondary Fisher rat embryo fibroblasts were co-transfected
`with pMSVp53G and pEJ6.6 (ref. 19), carrying an activated
`human c-Ha-ras l gene. As a positive control, we performed a
`
`
`
`(cid:141)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:4)(cid:50)(cid:69)(cid:88)(cid:89)(cid:86)(cid:73)(cid:4)(cid:52)(cid:89)(cid:70)(cid:80)(cid:77)(cid:87)(cid:76)(cid:77)(cid:82)(cid:75)(cid:4)(cid:43)(cid:86)(cid:83)(cid:89)(cid:84)(cid:21)(cid:29)(cid:28)(cid:24)
`
`4 of 4
`
`Celltrion, Inc., Exhibit 1117
`
`