1
`Nature Vol. 256 August 7 I975
`
`
`Page 1
`
`Continuous cultures of fused cells
`secreting antibody of predefined specificity
`THE manufacture of predefined specific antibodies by means of
`permanent tissue culture cell lines is of general interest. There
`are at present a considerable number of permanent cultures of
`myeloma cells” and screening procedures have been used to
`reveal antibody activity in some of them. This, however,
`is
`not a satisfactory source of monoclonal antibodies of predefined
`specificity. We describe here the derivation of a number of
`tissue culture cell
`lines which secrete anti-sheep red blood
`cell (SRBC) antibodies. The cell lines are made by fusion of a
`mouse myeloma and mouse spleen cells from an immunised
`donor. To understand the expression and interactions of the
`Ig chains from the parental lines, fusion experiments between
`two known mouse myeloma lines were carried out.
`Each immunoglobulin chain results from the integrated
`expression of one of several Vand C genes coding respectively
`for its variable and constant sections. Each cell expresses only
`one of the two possible alleles (allelic exclusion; reviewed in
`ref. 3). When two antibody-producing cells are fused,
`the
`products of both parental lines are expressed”, and although
`the light and heavy chains of both parental lines are randomly
`joined, no evidence of scrambling of V and C sections is
`observed‘. These results, obtained in an heterologous system
`involving cells of rat and mouse origin, have now been con-
`firmed by fusing two myeloma cells of the same mouse strain,
`
`A
`
`Chains
`
`H(Pl)
`
`495
`
`The protein secreted (MOPC 21) is an lgG1 (K) which has been
`fully sequenced”. Equal numbers of cells from each parental
`line were fused using inactivated Sendai virus9 and samples
`contining 2><l05 cells were grown in selective medium in
`separate dishes. Four out of ten dishes showed growth in
`selective medium and these were taken as independent hybrid
`lines, probably derived from single fusion events. The karyotype
`of the hybrid cells after 5 months in culture was just under the
`sum of the two parental lines (Table I). Figure 1 shows the
`isoelectric focusing” (IEF) pattern of the secreted products of
`different lines. The hybrid cells (samples c—lz in Fig. 1) give a
`much more complex pattern than either parent (a and b) or a
`mixture of the parental lines (m). The important feature of the
`new pattern is the presence of extra bands (Fig.
`l, arrows).
`These new bands, however, do not seem to be the result of
`differences in primary structure; this is indicated by the IEF
`pattern of the products after reduction to separate the heavy
`and light chains (Fig. 1B). The IEF pattern of chains of the
`hybrid clones (Fig. 1B, g) is equivalent to the sum of the IEF
`pattern (a and b) of chains of the parental clones with no
`evidence of extra products. We conclude that, as previously
`shown with interspecies hybrids”, new lg molecules are
`produced as a result of mixed association between heavy and
`light chains from the two parents. This process is intracellular
`as a mixed cell population does not give rise to such hybrid
`molecules (compare m and g, Fig. IA). The individual cells must
`therefore be able to express both isotypes. This result shows
`that in hybrid cells the expression of one isotype and idiotype
`does not exclude the expression of another: both heavy chain
`
` Fig.
`
`1 Autoradiograph of labelled compo-
`nents secreted by the parental and hybrid cell
`lines analysed by IEF before (A) and after
`reduction (B). Cells were incubated in the
`presence of “C-lysine“ and the supernatant
`applied on polyacrylamide slabs. A, pH range
`6.0 (bottom) to 8.0 (top) in 4 M urea. B, pH
`range 5.0 (bottom) to 90 (top) in 6 M urea;
`the supernatant was incubated for 20 min at
`37 “C in the presence of 8 M urea, 1.5 M
`mercaptoethanol and 0.1 M potassium phos-
`phate pH 8.0 before being applied to the right
`slab. Supernatants from parental cell
`lines
`in: a, P1 Bu] ; b, P3-X67Ag8; and m, mixture
`of equal number of PlBul and P3-X67Ag8
`cells. Supernatants from two independently
`derived hybrid lines are shown: c-f, four
`subclones from Hy-3; g and h, two subclones
`from Hy-B. Fusion was carried out” using
`10"‘ cells of each parental
`line and 4,000
`haemagglutination units
`inactivated Sendai
`virus (Searle). Cells were divided into ten
`equal
`samples and grown separately in
`selective medium (HAT medium,
`ref. 6).
`Medium was changed every 3 d. Successful
`hybrid lines were obtained in four of the cul-
`tures, and all gave similar IEF patterns. Hy-B
`and Hy-3 were further cloned in soft agar“.
`L, Light; H, heavy.
`
`‘
`‘
`‘
`‘
`__ M v__,, __
`'
`
`2"‘
`
`/1
`
`*‘
`
`b
`
`g
`
`a
`
`and provide the background for the derivation and under-
`isotypes (yl and y2a) and both V” and both VL regions
`standing of antibody-secreting hybrid lines in which one of
`(idiotypes) are expressed. There are no allotypic markers for
`the parental cells is an antibody-producing spleen cell.
`the CK region to provide direct proof forthe expression of both
`Two myeloma cell lines of BALB/c origin were used. PlBul
`parental CK regions. But this is indicated by the phenotypic
`link between the Vand C regions.
`is resistant
`to 5-bromo-2’-deoxyuridine“, does not grow in
`Figure IA shows that clones derived from different hybridi-
`selective medium (HAT, ref. 6) and secretes a myeloma protein,
`Adj PCS, which is an lgG2A (K), (ref. 1). Synthesis is not
`sation experiments and from subclones of one line are indistin-
`balanced and free light chains are also secreted. The second
`guishable. This has also been observed in other experiments
`cell line, P3-X63Ag8, prepared from P3 cellsz,
`is resistant to
`(data not shown). Variants were, however, found in a survey of
`I00 subclones. The difference is often associated with changes
`20 ug ml" 8-azaguanine and d_o_e_s__n9_t grow in HAT medium.
`‘ ‘""
`GENENTECH 2003
`GENZYME V. GENENTECH
`©1975 Nature Publishing Group
`IPRZO 16-003 83
`
`GENENTECH 2003
`GENZYME V. GENENTECH
`IPR2016-00383
`
`

`
`496
`
`Nature Vol. 256 August 7 I975
`
`Isolation of an anti-SRBC antibody-
`Fig. 2
`secreting cell clone. Activity was revealed by
`a halo of haemolysed SRBC. Direct plaques
`given by: a, 6,000 hybrid cells Sp-1 ; b, clones
`grown in soft agar from an inoculum of 2,000
`Sp-1 cells; c, recloning of one of the positive
`clones Sp-1/7; a’, higher magnification of a
`positive clone. Myeloma cells (107 P3—X67A
`g8) were fused to 109 spleen cells from an
`immunised BALB/c mouse. Mice were im-
`munised by intraperitoneal injection of 0.2 ml
`packed SRBC diluted 1:10, boosted after 1
`month and the spleens collected 4 d later.
`After fusion, cells (Sp-l) were grown for 8 d
`in HAT medium, changed at 1-3 d intervals.
`Cells were then grown in Dulbecco modified
`Eagle’s medium, supplemented for 2 weeks
`with hypoxanthine and thymidine. Forty days
`after fusion the presence of anti-SRBC act-
`ivity was revealed as shown in a. The ratio of
`plaque forming cells/total number of hybrid
`cells was 1/30. This hybrid cell population
`was cloned in soft agar (50% cloning ef-
`ficiency). A modified plaque assay was used to
`reveal positive clones shown in b—a' as follows.
`When cell clones had reached a suitable size,
`they were overlaid in sterile conditions with
`2 ml 0.6% agarose in phosphate-buffered
`saline containing 25 ul packed SRBC and
`0.2 ml fresh guinea pig serum (absorbed with
`SRBC) as source of complement. b, Taken
`after overnight incubation at 37 “C. The ratio
`of positive/total number ofclones was 1/33. A
`suitable positive clone was picked out and
`grown in suspension. This clone was called
`Sp-I/7, and was recloned as shown in c; over
`90 % of the clones gave positive lysis. A second
`experiment in which 10“ P3- X67Ag8 cells were
`fused with 10*‘ spleen cells was the source of
`a clone giving rise to indirect plaques (clone
`Sp-2/3-3). Indirect plaques were produced by
`the addition of 1:20 sheep anti-MOPC 21
`antibody to the agarose overlay.
`
`in the ratios of the different chains and occasionally with the
`total‘ disappearance of one or other of the chains. Such events
`are best visualised on IEF analysis of the separated chains
`(for example, Fig.
`lh, in which the heavy chain of P3 is no
`longer observed). The important point that no new chains are
`detected by IEF complements a previous study“ of a rat—mouse
`hybrid line in which scrambling of V and C regions from the
`light chains of rat and mouse was not observed. In this study,
`both light chains have identical CK regions and therefore
`scrambled VL—CL molecules would be undetected. On the other
`hand,
`the heavy chains are of different subclasses and we
`expect scrambled VH—CH to be detectable by IEF. They were
`not observed in the clones studied and if they occur must do
`so at a lower frequency. We conclude that in syngeneic cell
`hybrids (as well as in interspecies cell hybrids) V—C integration is
`not the result of cytoplasmic events. Integration as a result of
`DNA translocation or rearrangement during transcription is
`also suggested by the presence of integrated mRNA molecules”
`and by the existence of defective heavy chains in which a
`deletion of V and C sections seems to take place in already
`committed cells”.
`The cell line P3-X63Ag8 described above dies when exposed
`to HAT medium. Spleen cells from an immunised mouse also
`die in growth medium. When both cells are fused by Sendai
`virus and the resulting mixture is grown in HAT medium,
`surviving clones can be observed to grow and become estab-
`lished after a few weeks. We have used SRBC as immunogen,
`which enabled us, after culturing the fused lines, to determine
`the presence of specific antibody-producing cells by a plaque
`assay technique” (Fig. 2a). The hybrid cells were cloned in
`soft agar“ and clones producing antibody were easily detected
`by an overlay of SRBC and complement (Fig. 2b). Individual
`clones were isolated and shown to retain their phenotype as
`almost all the clones of the derived purified line are capable of
`lysing SRBC (Fig. 2c). The clones were visible to the naked
`eye (for example, Fig. 2d). Both direct and indirect plaque
`
`assays” have been used to detect specific clones and representa-
`tive clones of both types have been characterised and studied.
`The derived lines (Sp hybrids) are hybrid cell lines for the
`following reasons. They grow in selective medium. Their
`karyotype after 4 months in culture (Table 1) is a little smaller
`than the sum of the two parental lines but more than twice the
`chromosome number of normal BALB/c cells, indicating that
`the lines are not the result of fusion between spleen cells. In
`addition the lines contain a metacentric chromosome also
`present
`in the parental P3 -X67Ag8. Finally,
`the secreted
`immunoglobulins contain MOPC 21 protein in addition to new,
`unknown components. The latter presumably represent
`the
`chains derived from the specific anti-SRBC antibody. Figure 3A
`shows the lEF pattern of the material secreted by two such
`Sp hybrid clones. The IEF bands derived from the parental P3
`line are visible in the pattern of the hybrid cells, although
`obscured by the presence of a number of new bands. The
`pattern is very complex, but the complexity of hybrids of this
`type is
`likely to result from the random recombination of
`chains (see above, Fig. 1). Indeed, IEF patterns of the reduced
`material secreted by the spleen—P3 hybrid clones gave a simpler
`pattern of lg chains. The heavy and light chains of the P3
`parental line became prominent, and new bands were apparent.
`The hybrid Sp-1 gave direct plaques and this suggested that
`it produces an lgM antibody. This is confirmed in Fig. 4 which
`shows the inhibition of SRBC lysis by a specific anti—IgM
`
`
`Table I Number of chromosomes in parental and hybrid cell lines
`
`Mean
`Number of chromosomes per cell
`Cell line
`65
`66,65,65,65,65
`P3-X67Ag8
`55
`Ref. 4
`Pl Bul
`40
`4
`Mouse spleen cells
`J06
`ll2,1l0,l04,l04,l02
`Hy-B(Pl—P3)
`90
`93,90,89,89,87
`Sp- 1 /7-2
`
`
`97,98,96,96,94,88Sp-2/3-3 95
`
`©1975 Nature Publishing Group
`
`

`
`Nature 1%)]. 256 August 7 I975
`
`Page 3
`
`49.,
`
`IEF techniques usually do not reveal 19S IgM
`antibody.
`IgM is
`therefore unlikely to be present
`in the
`molecules.
`unreduced sample a (Fig. 3B) but I1 chains should contribute
`to the pattern obtained after reduction (sample a, Fig. 3A).
`The above results show that cell fusion techniques are a
`powerful tool to produce specific antibody directed against a
`predetermined antigen. It further shows that it is possible to
`isolate hybrid lines producing different antibodies directed
`against the same antigen and carrying different effector func-
`tions (direct and indirect plaque).
`The uncloned population of P3~spleen hybrid cells seems
`quite heterogeneous. Using suitable detection procedures it
`should be possible to isolate tissue culture cell
`lines making
`different classes of antibody. To facilitate our studies we have
`used a myeloma parental line which itself produced an lg.
`Variants in which one of the parental chains is no longer ex-
`pressed seem fairly common in the case of P1—P3 hybrids
`(Fig. lh). Therefore selection oflines in which only the specific
`antibody chains are expressed seems
`reasonably simple.
`Alternatively, non-producing variants of myeloma lines could
`be used for fusion.
`We used SRBC as antigen. Three different fusion experiments
`were successful
`in producing a large number of antibody-
`producing cells. Three weeks after the initial fusion, 33/1,086
`
`'1
`
`
`
` ab
`
`b
`
`C
`
`
`
`Inhibition of haernolysis by antibody secreted by hybrid
`Fig. 4
`clone Sp-l/7-2. The reaction was in a 9-cm Petri dish with a layer
`of5 ml 0.6 % agarose in phosphate-buffered saline containing 1/80
`(v/v) SRBC. Centre well contains 2.5 ul 20 times concentrated
`culture medium of clone Sp-1/7-2 and 2.5 ul mouse serum. a,
`Sheep specific anti—mouse macroglobulin (MOPC 104E, Dr
`Feinstein); b, sheep anti-MOPC 21 (P3) IgGl absorbed with Adj
`PC-5; c, sheep anti-Adj PC-5 (IgG2a) absorbed with MOPC 21.
`After overnight incubation at room temperature the plate was
`developed with guinea pig serum diluted l:l0 in Dulbecco’s
`medium without serum.
`
`clones (3 "/,,) were positive by the direct plaque assay. The
`cloning efficiency in the experiment was 50 3/{,. In another experi-
`ment, however,
`the proportion of positive clones was con-
`siderably lower (about 0.2 ‘X,). In a third experiment the hybrid
`population was studied by limiting dilution analysis. From 157
`independent hybrids, as many as 15 had anti-SRBC activity.
`The proportion of positive over negative clones is remark-
`ably high.
`It
`is possible that spleen cells which have been
`triggered during immunisation are particularly successful
`in
`giving rise to viable hybrids. It remains to be seen whether
`similar results can be obtained using other antigenes.
`The cells used in this study are all of BALB/c origin and the
`hybrid clones can be injected into BALB/c mice to produce
`solid tumours and serum having anti-SRBC activity.
`It
`is
`possible to hybridise antibody-producing cells from different
`origins”. Such cells can be grown in vitro in massive cultures
`to provide specific antibody. Such cultures could be valuable
`for medical and industrial use.
`
`G. K6HLER
`C. MILSTEIN
`
`MRC Laboratory of Molecular Biology,
`Hills Road, Cambridge CB2 2QH, UK
`
`Received May 14; accepted June 26, I975.
`I Potter. M., Physiol. Rev., 52, 631-719 (1972).
`1 Horibata, K., and Harris. A. W., Exp] (‘(-11 Res.. 60. 61-70 (1970).
`3 Milstein. C.. and Munro, A. J.. in Defence and Recognition (edit. by Porter, R. R.),
`I99-228 (N/ITP Int. Rev. Sci., Butterworth, London, 1973).
`‘ Cotton. R. G. H., and Milstein. C., Nature. 244, 42-43 (1973).
`5 Schwaher, J., and Cohen, E. P., PI‘0('.I1(lN'l. A1-1:11.51-i. U.S.A., 7| . 2203-2207 ( I974).
`6 Littlefield, J. W., .S‘r'icnc-1». 145, 709 (1964).
`7 Svasti, J.. and Milstein, C.. Bioc/mm. J.. 128. 427-444 (1972).
`3 Milstein. C., Adetugbo, K., Cowan, N. .I., and Secher. D. 5.. Progress in Immuno-
`logy, ll,
`1 (edit. by Brent. L.. and Holborow, J.),
`I57-l68 (North-Holland.
`Amsterdam, I974).
`" Harris, H., and Watkins. J. F.. Nature, 205, 640-646 (1965).
`10 Awdeh, A. L., Williamson, A. R.. and Askonas. B. A., Nature, 219. 66-67 (1968).
`11 Milstein, C., Brownlee. G. G., Cartwright, E. M.. Janis. J. M.. and Proudfoot,
`N. 1., Nature. 252. 354-359 (1974).
`'2 Frangione, B.. and Milstein, C., Nature. 244. 597-599 (1969).
`1-‘ Jerne, N. K.. and Nordin. A. A.. S('i(’nz'z'. I40, 405 (1963).
`(1
`).
`N Cott;)_;13. R. G. H., Secher, D. S., and Milstein. (‘.. Eur. J. ImmIm., 3. 135-140
`
`Fig. 3 Autoradiograph of labelled components secreted by anti-
`SRBC specific hybrid lines. Fractionation before (8) and after (A)
`reduction was by IEF. pH gradient was 5.0 (bottom) to 9.0 (top)
`in the presence of6 M urea. Other conditions as in Fig. l.Super—
`natants from: a, hybrid clone Sp-1/7-2; /J, hybrid clone Sp-2/3-3;
`c, myeloma line P3-X67Ag8.
`
`©1975 Nature Publishing Group