Molecular Immunology, Vol. 27, No. 4, pp. 327- 333, 1990
`Printed in Great Britain.
`
`0161-5890/90 $3.00 + 0.00
`© 1990 Pergamon Press pie
`
`THE EFFECT OF TEMPERATURE ON THE BINDING
`KINETICS AND EQUILIBRIUM CONSTANTS OF
`MONOCLONAL ANTIBODIES TO CELL SURF ACE
`ANTIGENS
`
`RICKY w. JOHNSTONE, SARAH M. ANDREW, MARK P. HOGARTH, GEOFFREY A. PIETERSZ and
`IAN F. C. McKENZIE*
`Research Centre for Cancer and Transplantation, Department of Pathology, University of Melbourne,
`Parkville, Victoria 3052, Australia
`
`(First received 21 April 1989; accepted 4 October 1989)
`
`Abstract-The effect of temperature on the kinetic association and dissociation binding parameters, and
`equilibrium constants of four monoclonal antibodies to the murine Ly-2.1 and Ly-3.1 antigens has been
`studied using flow cytometry. All four monoclonal antibodies were conjugated to FITC and their
`association to, and dissociation from, the surface of murine thymoma cells was observed at 15 sec intervals,
`at temperatures between I and 37°C. The initial association rate constant and the dissociation rate
`constant for each antibody at each temperature were calculated from graphs of the first-order reactions
`and it was demonstrated that an increase in temperature caused an increase in both association rate and
`dissociation rate of the antibodies. Generally the increase in association rate with temperature was less
`than the increase in dissociation rate. Differences between antibodies to the same antigen (Ly-2. l) suggest
`that changes in membrane fluidity were not solely responsible for the changes in association rate. However,
`the equilibrium constants (Koq) did not always show a simple relationship of increasing temperature
`causing decreasing values for Koq. For one antibody the highest value for Koq was seen at l 7°C rather_ tha_n
`at 37°C and differences in Koq between individual antibodies were greater at J°C than at _37°<;. Kmet1c
`rate constants are usually measured at 4°C or room temperature, therefore for antibodies under
`consideration for in vivo use, measurements at 37°C are more appropriate.
`
`INTRODUCTION
`
`The bond between antibody and antigen is dependent
`on non-covalent forces, and it is a dynamic and
`complex interaction. The equilibrium constant, or
`affinity, of an antibody can be used to describe the
`antigen-antibody
`interaction while
`the
`two are
`bound, the Law of Mass Action relating the equi(cid:173)
`librium constant to the association and dissociation
`of antibody and antigen (Steward and Steensgaard,
`1985). Thus the equilibrium constant is dependent on
`the rate of antigen-antibody association and also the
`rate of their dissociation. The rate of these reactions
`should vary with temperature; the thermodynamic
`nature of the bonds involved in antigen-antibody
`interactions predicts such temperature dependency.
`As the bond between antibody and antigen is made
`up of various non-covalent forces, and is dependent
`on physical conformations and orientation within
`the binding site, it cannot be assumed that the
`equilibrium constant for every monoclonal antibody
`follows a predictable pattern of increasing tempera(cid:173)
`ture causing decreasing affinity values. The effect
`of temperature changes on the kinetic parameters
`(association and dissociation rate constants), and
`
`• Author to whom correspondence should be addressed.
`
`affinity of antibodies has not been the subject of
`detailed study, but in general it appears that the
`dissociation rate shows greater temperature depen(cid:173)
`dency than the association rate, and that differences
`in affinities arise mainly through variations in dissoci(cid:173)
`ation rates (Mason and Williams, 1980, 1986; Froese,
`1968). The temperature dependency of the affinity of
`antibodies for haptens or antigens in solid phase may
`be solely confined to effects on the bonds at the
`binding site. In the case of antibody binding at the
`cell surface, temperature effects on the membrane of
`the cell may also affect the behaviour of bound
`antibody (Linden et al., 1973). Accurate measure(cid:173)
`ment of association and dissociation rates for anti(cid:173)
`bodies binding to viable cells can be carried out using
`flow cytometry (Roe et aL, 1985). This study shows
`the effect of change in temperature on kinetics of
`antibody binding at the cell surface. Four mono(cid:173)
`clonal antibodies to the same (Ly-2.1) or related
`(Ly-3.1) antigens were used, and their association
`to and dissociation from cell surface antigens at a
`range of temperatures between I and 37°C were
`examined using direct immunofluorescence on a
`flow cytometer. The results show that variations in
`both association and dissociation rate constants can
`occur with changes in temperature and that there
`is not necessarily an inverse relationship between
`temperature and equilibrium constant.
`
`327
`
`

`

`328
`
`RICKY W. JOHNSTONE et al.
`
`MATERIALS AND METHODS
`
`Monoclonal antibodies and target cells
`Monoclonal antibodies to the murine Ly-2. l and
`Ly-3.1 antigens were used: 49-14.2 (14.2), lgGI;
`49-11.1 (11.1), IgG2a; and 49-31.1 (31.l), lgG3 bind
`to Ly-2.1 (Hogarth et al., 1982a ); 5034-29.5 (29.5),
`IgG! binds to Ly-3.1 (Sutton, 1984). The antibodies
`were affinity purified from mouse ascites fluid by
`Protein A Sepharose. The murine thymoma ITT( 1)
`75NS.E3 (E3) (Hogarth et al., 1982b; Smyth et al.,
`1986) which expresses both Ly-2.1 and Ly-3.1 was
`used as the target cell line. The cells were maintained
`in Dulbecco's Modified Eagle's Medium (DME) with
`10% newborn calf serum (Flow Laboratories,
`NSW, Australia) containing 50 U/ml penicillin and
`50 µg/ml steptomycin.
`
`Fluorescein labelling of antibodies and activity of the
`labelled preparatiofl
`Purified antibodies were conjugated to fluorescein
`isothiocyanate (FITC) [Sigma, Dorset, U.K. (Pimm et
`al., 1982)] such that fluorescein to protein ratios of be
`tween I : I and 5: I were obtained, determined spectro(cid:173)
`photometrically (Forni, 1979). FITC antibodies were
`stored in phosphate buffered saline (PBS pH 7.4)
`with 1.0% sodium azide at 4°C. Each preparation
`was shown to have activity comparable with the
`unlabelled preparation using competition binding
`assays (Robins et al., 1986) on a flow cytometer
`(FACScan, Becton-Dickinson, Mountain View,
`CA, U.S.A.). The immunoreactive fractions and
`equilibrium binding characteristics of the particular
`antibody preparations used had been determined
`previously (Andrew et al., 1990). The immuno(cid:173)
`reactive fractions were: 14.2: 52.65%; 11.1: 49.30%;
`31.l: 78.41 %; and 29.5: 54.28%. These figures were
`taken into account in calculations for absolute
`amounts of antibody bound per cell, corrections
`being made based on the assumption that only the
`immunoreactive fraction of the antibody had binding
`activity. Equilibrium binding experiments performed
`on the F ACScan were used to determine the amount
`of antibody required to saturate a given number of
`cells, in each case 5 µg of purified antibody saturated
`4 x 105 cells.
`
`Association experiments
`Association of FITC-antibody to E3 cells with
`time was observed over a range of temperatures using
`flow cytometry. Cells were prepared by washing three
`times in PBS with 2% newborn calf serum, viability
`was monitored with Trypan blue; the final concen(cid:173)
`tration of the cells was 4 x 105 cells per 0.5 ml. An
`initial fluorescence reading (time, t = 0) of 2000 cells
`was taken, then 5 µg of FITC-antibody in 200 µI
`PBS was added to the cells, and fluorescence readings
`of a mean of 2000 cells were obtained at 15 sec
`intervals on the FACScan. The temperature of the
`reaction mixture was controlled by immersion of the
`
`F ACS tube m a beaker contammg water at the
`appropriate
`temperature which was monitored
`continuously with a thermometer (the variation in
`temperature was within 1 °C of the experimental
`temperature). Experiments were carried out at I , 10,
`17, 25 and 37°C.
`
`Dissociation experiments
`The dissociation of FITC-antibody with time was
`measured using flow cytometry on E3 cells. Cells were
`prepared as above, and incubated with 5 µg of FITC(cid:173)
`labelled preparations of each antibody on ice for 2 hr.
`An initial fluorescence reading was taken at time,
`t = 0, then a 40-fold excess of unlabelled intact
`antibody, taken from the same batch as the FITC(cid:173)
`labelled preparation, was added to the tube. This
`large excess of unlabelled antibody ensured that there
`was little chance of reassociation of FITC-labelled
`antibody during the course of the experiment. Dis(cid:173)
`sociation of FITC-labelled antibody from E3 cells
`was observed by taking fluorescence readings from
`2000 cells at 15 sec intervals; again, the experiments
`were carried out a I, 10, 17, 25 and 37°C.
`For all flow cytometric observations, the FACScan
`was calibrated with both Calibrite beads (Becton(cid:173)
`Dickson; Bioclone Australia Pty, Marrickville, NSW,
`Australia) and Quantitative Fluorescein Microbead
`Standards (Flow Cytometry Standards Corp., Re(cid:173)
`search Triangle Park, NC U.S.A.). The machine
`settings on the F ACScan were standardized and the
`same settings were used for each experiment. The
`fluorescence units were calibrated, one fluorescence
`unit represented 2160 molecules of FITC.
`
`Calculation of constants
`The association rate constant was obtained from a
`graph of antibody molecules bound per cell versus
`time, which were calculated using the mean fluor(cid:173)
`escence bound per cell, the F: P ratio and the cali(cid:173)
`bration of the FACScan. The association constant,
`k 1 = gradient/2CQN,
`
`where: C = molar concentration of cells [calculated
`from concentration of cells in mol/1 = number of cells
`per litre/ Avogadro's number (Trucco and de Petris,
`1981 )], Q = input concentration of antibody
`in
`molecules per cell, and N = number of antigenic sites
`per cell (Roe et al. , 1985). Ly-2. l and Ly-3.1 have
`been shown to have 1.9 x 105 and 3.8 x !05 antigenic
`sites per E3 cell respectively (Andrew et al., 1990).
`The dissociation rate constant, k2 , is the slope of a
`graph of log U / U0 vs time, where U = mean fluores(cid:173)
`cence at time t, and U0 = initial mean fluorescence
`(Roe et al., 1985).
`The equilibrium constant, Keq, was calculated from
`
`Keq = association constant,
`
`k 1 /dissociation constant, k2 •
`
`

`

`Temperature effects on antibody binding kinetics
`
`329
`
`All graphs were drawn using the simple curve of best
`fit obtained from Cricket Graph Software on an
`Apple MacIntosh personal computer. The slopes of
`the curves were obtained from regression analysis
`using the same software package.
`
`RESULTS
`
`Variation in antibody association with temperature
`Results of association experiments carried out at
`various temperatures with each monoclonal antibody
`and E3 cells are shown in Fig. I. The increase in
`fluorescence with
`time indicates an increase in
`bound antibody per E3 cell with only the initial,
`
`monovalent, association rate being measured. The
`association curves seen in Fig. I were found to be
`reproducible with very similar straight line curves
`obtained upon repetition of the experiment. For
`all monoclonal antibodies studied, increasing the
`in
`the rate of
`temperature caused an increase
`association of antibody to the cells.
`Association rate constants for each antibody at
`each temperature (Table 1) show considerable vari(cid:173)
`ation with temperature. The temperature-related
`increases in association rates for individual mono(cid:173)
`clonals do not show a specific, constant pattern and
`there are differences in k 1 values between the antibody
`groups. The effect of temperature was considerably
`
`A
`
`B
`
`...I
`...I 120000
`w
`0 c z 100000
`::,
`0 m
`1/) 80000
`w
`...I
`::>
`0
`w
`...I
`0
`:IE
`► 0
`0 m 20000
`~
`<
`
`60000
`
`40000
`
`0
`
`...I
`...I w 120000
`~ 100000
`z ::,
`0 m 80000
`1/)
`w
`...I
`::, 60000
`0
`~
`0 ::s 40000
`► 0
`20000
`0 m
`§
`<
`
`0
`
`0
`
`100 200 300 400 500
`TIME(SECS)
`
`D
`
`::l 30000
`
`i 25000
`
`0
`Ill 20000
`1/)
`~ ::> 15000
`~ 10000
`:a
`8 m
`
`0
`
`§
`<
`
`5000
`
`500 1 000 1500 2000 2500 3000
`
`TIME(SECS)
`
`0
`
`100 200 300 400 500
`
`TIME(SECS)
`
`O
`
`200 400 600 800 1000
`TIME(SECS)
`
`Fig. I. Variation in antibody association rate with temperature. Molecules of antibody bound per cell
`(calculated from F: P ratios and flow cytometer calibration) are plotted against time (sec) for each
`temperature, fluorescence readings of 2000 cells were taken at 15 sec intervals. A, antibody 11 . 1; B,
`antibody 14.2; C, antibody 31.1; D, antibody 29.5. The temperatures are: 1°C (■), I0°C (.A.), l 7°C ( x ).
`25°C (l:,,) and 37°C (□). (Note: different scales for both axes in A and D.)
`
`

`

`330
`
`RICKY W. JOHNSTONE et al.
`
`Table I. Calculated values for association rate constants (k 1 ) at
`various temperatures
`
`Monoclonal antibody
`(subclass)
`11.1 (lgG2a)
`142 (IgGI)
`31.1 (lgG3J
`29.5 (lgG I)
`
`Values of k, ( x 104)1/mol/sec
`17' C
`25"C
`IOC
`2.125
`1.781
`1.222
`2.191
`0.809
`0.196
`0.036
`0.095
`
`1.866
`1.472
`0.312
`0.057
`
`IC
`1.407
`0.944
`0.104
`0.018
`
`JTC
`3.24]
`2.320
`1.881
`0.098
`
`greater in the case of 3 I.I (Fig. IC) and 29. 5 (Fig. ID)
`IA, 18).
`than on the other two antibodies (Fig.
`The association rate constant at 37cc
`for 31.1
`(1.88 1/mol/sec) was
`18
`times
`that
`at
`I "C
`I I.I, the
`(0. I 04 1/mol/sec), whereas for 14.2 and
`increase in rate constant was 2-fold between I and
`
`37°C (Table I). These latter two antibodies showed
`very little increase in association rate constant be(cid:173)
`tween 25 and 37°C (Table I). From these results it is
`clear that different antibodies are affected differently
`by increases in temperature with respect to their
`association characteristics, even when they are associ(cid:173)
`ating to the same antigen. These differences do not
`appear to be subclass related, as the two IgG l
`antibodies behaved differently (i.e. 14.2: k 1 = 2.320,
`29.5: k 1 = 0.098, at 37°C).
`Considerable variation in k 1 was observed between
`the antibodies, 29.5 having the slowest rate of
`association over all and 11.1
`the fastest. Thus,
`there exists a direct relationship between antibody
`
`A
`
`0
`
`0
`::>
`3
`£
`-1
`
`B
`
`0
`~
`2.
`.E
`
`0
`
`-1
`
`0
`
`1000 2000 3000
`TIME(SECS)
`
`4000
`
`1000 2000 3000
`TIME(SECS)
`
`4000
`
`C
`
`0 I £
`
`-1
`
`D
`
`0
`
`6' I -1
`
`0
`
`1 000
`500
`TIME(SECS)
`
`1500
`
`2000
`
`2000
`1000
`TIME(SECS)
`
`3000
`
`Fig. 2. Variation in antibody dissociation rate with temperature. The natural log of fluorescence values
`(mean of 2000 cells) divided by initial fluorescence (lnU / V0 ) is plotted against time; readings being taken
`at 15 sec intervals. A, antibody 11.1; B, antibody 14.2; C, antibody 31.1; D, antibody 29.5. The
`temperatures are: 1 °C (■), 10°c (A), !7°C ( X ), 25°C (6) and 37°C (□). (Note: dissociation of antibody
`29.5 at 1°c was too slow to be accurately measured.)
`
`

`

`Temperature effects on antibody binding kinetics
`
`331
`
`l ' C
`
`Table 2. Calculated values for dissociation rate constants (k2 ) at 1 1 2 at various temperatures
`Values of k2 ( x JQ- ')/sec, and 11;2 min
`25°C
`IO°C
`l7°C
`k,
`k,
`k,
`Monoclonal antibody
`k,
`1112
`1,:2
`I 1/2
`1112
`I I.I (IgG2a)
`14
`8.3
`32
`12.0
`3.64
`112
`1.03
`64
`14.2 (lgGI)
`1.72
`5.9
`38
`20
`12.0
`3.08
`31.J (IgG3)
`8.1
`18
`14
`40
`2.87
`31
`6.3
`a
`a
`29.5 (lgGI)
`46
`2.57
`178
`3.68
`0.65
`"Dissociation of 29.5 at this temperature was too slow to be reasonably assessed.
`
`37'C
`11:2
`2.0
`2.2
`2.0
`5.0
`
`k,
`57
`49
`58
`23.0
`
`IQ
`10
`3.8
`32
`
`association rate constants and temperature for all of
`the monoclonal antibodies studied. The amount of
`rate increase with temperature appears to be an
`inherent property of the individual antibodies and is
`not immunoglobulin subclass related.
`
`Variation in antibody dissociation with temperature
`Curves showing variation in dissociation of each
`antibody from E3 cells with temperature are shown
`in Fig. 2. As the fluoresceinated antibodies dissociate
`from the cell, unlabelled antibodies take their place,
`thus decreasing the mean linear fluorescence per
`cell. With time, the fluorescence per cell decreases,
`indicating a decrease in bound fluoresceinated anti(cid:173)
`body. An increase in temperature caused an increase
`in the rate of dissociation of antibody from cells. The
`effect of temperature on dissociation was reflected in
`the dissociation constants, and t 112 (time taken for
`50% of the antibody to dissociate from the cell
`surface) shown in Table 2. The value for t 112 was
`calculated from k 2 = 0.693/t1,2 (Mason and Williams,
`I 986). For all four antibodies (Fig. 2), there was a
`marked increase in dissociation with an increase in
`temperature. At l °C (Fig. 2), there was very little
`11.1, 14.2 and 31 .1 antibody dissociation from the cell
`surface whilst the rate of 29.5 dissociation at I °C was
`too slow to allow accurate measurements. For all
`antibodies, t112 was short at 37°C, the anti-Ly-2.1
`antibodies (I I.I, 14.2 and 31.1) having t1:2 of approxi(cid:173)
`mately 2 min. The anti-Ly-3.1 antibody (29.5)
`showed very slow dissociation kinetics compared to
`the other three antibodies with a t1•2 of 5.0 min at
`37°C. Again, a different effect of temperature between
`the different antibodies was observed; antibody 1 I.I
`was most affected with the dissociation constant at
`37°C being over 50 times that at I 0 C. Antibody 31. l
`was the least affected by temperature with respect to
`) = 20 x k 2 (37°C).
`dissociation rate constant, k 2 (I 0
`Comparing variations of dissociation rate constants
`between the various antibodies, the antibody with the
`slowest association rate, namely 29.5, also had the
`slowest dissociation rate. Thus, there is a direct
`relationship between temperature and the rate of
`
`dissociation for the four antibodies studied. The
`increase in dissociation with
`temperature varies
`between the antibodies and there appears not to be a
`constant pattern of dissociation rate increase.
`
`Variation in equilibrium constant with temperature
`The equilibrium constants for each antibody at
`the various temperatures are shown in Table 3.
`For antibodies 14.2, 11 .1 and 29.5 the equilibrium
`constants decreased with increasing temperature, the
`greatest effect being seen with antibody 11.1 where
`K.q (1 °C) = 24 x K.q (3 7"C). In the case of antibody
`31.1 very little change in K,9 with temperature was
`observed, the highest value for Keq occuring at I T C;
`K.q (37°C) was slightly lower than K.9 (l "'C). The
`reason for this anomaly lies in the calculated equi(cid:173)
`librium constant being taken from the association/
`dissociation constant; the effect of temperature on the
`association rate of 31.1 was considerably greater than
`for the other antibodies.
`Generally, however, antibody-antigen binding re(cid:173)
`actions follow basic thermodynamic principles, an
`increase in temperature causing an increase in both
`association and dissociation with a resultant decrease
`in antibody affinity. Not all antibodies showed an
`inverse relationship between temperature and equi(cid:173)
`librium constant, as with antibody 31 .1 due to the
`nature of the derivation of K.9 .
`
`DISCUSSION
`
`The major aim of the study was to examine the
`effect of temperature on antibody association and
`dissociation rates and subsequently on the affinity
`(equilibrium constant) of antibodies to cell surface
`antigens. While such studies are conventionally
`performed at 4°C or room temperature, the increased
`use of monoclonal antibodies at 37°C in in vivo
`therapy and imaging makes it appropria te to conduct
`in vitro studies at higher temperatures. This report
`shows that temperature has a considerable effect on
`antibody kinetics which are different for each anti(cid:173)
`body and are not related to immunoglobulin isotype.
`
`Table 3. Values for equilibrium constant (K,9) at various temperatures
`Monoclonal antibody
`Values for K"' 1/mol
`ITC
`(subclass)
`2.25 x 107
`11.1 (IgG2a)
`2.49 X JQ7
`14.2 (lgGI)
`3.85 x 106
`31.I (lgG3)
`2.22 X 106
`29.5 (lgGI)
`"Not calculated.
`
`1°c
`1.37 >< 108
`5.49 >< 107
`3.62 >< IO'
`a
`
`10°C
`4.90 x 107
`3.97 X 107
`3.11 x IO'
`5.54 x 106
`
`25 'C
`1.77 x 107
`1.83 X JQ7
`2.60 x IO'
`2.55 x 106
`
`3TC
`5.96 x IO'
`4 .73 X 106
`3.24 x 106
`4.25 x 105
`
`

`

`332
`
`RICKY w. JOHNSTONE et al.
`
`The study clearly shows that there are considerable
`temperature-induced variations in the kinetic binding
`parameters of antibodies to cell surface antigens.
`These variations consisted of an increase in tempera(cid:173)
`ture causing an increase in the rate of association and
`dissociation for all four antibodies studied. In all
`cases the initial rates of association and dissociation
`were studied, so that the reactions followed first(cid:173)
`order kinetics. This not only simplified the analysis
`of the results but also meant that the true affinity of
`the (possibly bivalent) binding of intact antibody
`to cell surface antigens was not obtained. Thus
`the calculated equilibrium constants represent an
`"apparent affinity" and are likely to be biased
`towards monovalent cell surface interactions with a
`single binding site of the antibody. However, within
`this system, the results are comparable particularly as
`three of the antibodies react with the same cell surface
`antigen (Ly-2.1) and Ly-3. l is physically associated
`with Ly-2.1 on the cell surface (Sutton, 1984).
`This study shows that an increase in temperature
`caused an increase in monoclonal antibody associ(cid:173)
`ation to, and dissociation from cell surface antigens.
`In the case of antibody 31.1 , an 18-fold increase in
`association rate was observed when the temperature
`was increased from I to 37°C, however, antibody 11.1
`showed only a 2-fold increase for the same tem(cid:173)
`perature change. This clearly shows that although
`antibody-antigen reactions do follow basic thermo(cid:173)
`dynamic principles, the variation in association rate
`with temperature is an inherent characteristic of the
`particular monoclonal antibody. As with association
`rate variations, differences in the rate of dissociation
`with temperature was observed between the four
`antibodies. Monoclonal antibody 29.5 had a dissoci(cid:173)
`ation rate that was so slow at I ''C, accurate and
`reproducible measurements could not be made. How(cid:173)
`ever, all antibodies showed an increase in rate of
`dissociation as the temperature was increased.
`Using W3/25 antibody and studying dissociation
`from thymocyte membranes, Mason and Williams
`(1980) observed biphasic dissociation kinetics at
`higher temperatures (18 and 26°C) in experiments
`conducted over a longer time period (240 min) than
`the higher temperature experiments presented here.
`The antibody studies described herein were of lower
`affinity than those used in the previous study and had
`considerably faster dissociation kinetics. It may be
`that the low affinity antibodies follow a monophasic
`dissociation pattern, perhaps even showing Fab-like
`binding and hence biphasic binding reactions such as
`those observed by Mason and Williams (1980) did
`not occur. Further experiments including high affinity
`antibodies and Fab fragments may demonstrate this.
`In addition, these authors ( 1980, 1986) showed values
`for dissociation constants using W3/25 and 0X7 (Fab
`fragments) with 10-20-fold increase in dissociation
`seen with temperature rises of 4--26cC and 4--18"C
`respectively. In the current study, three of the four
`antibodies showed similar increases in dissociation
`
`rates (5-10-fold) up to 25°C. A large increase in rate
`of dissociation then occured between 25 and 37°C,
`there being a 4--6-fold increase in dissociation rate
`between these temperatures. The reason for this is
`unclear but it appears that this was not connected
`with a sudden increase in membrane fluidity or
`changes in state of the membrane phospholipids
`which have been reported to occur at characteristic
`temperatures (Linden et al., 1973), as one anti-Ly-2.1
`antibody (31.1) sho·.ved only a further 2-fold increase
`in dissociation rate between these two temperatures.
`It is possible that changes in individual epitopes may
`be differentially related to temperature but this is
`unlikely.
`As well, increases in association rate with increas(cid:173)
`ing temperature have also been reviewed by Mason
`and Williams (1986) with antibodies 0X7 and W3/25
`at two different temperatures. Increases in tempera(cid:173)
`ture from 4 to l 8°C lead to 2-3-fold increases in the
`values for association rate constants. In the current
`study, 2.5- and 2.3-fold increases in association
`constant (k 1) were obtained for antibodies 14.2 and
`11.1 respectively but a 5.4-fold increase was observed
`for 29.5, and in the case of 31.1, the association rate
`at 3TC was 18 times that at 1°C. The temperatures
`between which the greatest change took place varied
`also, the greatest change being seen between 17 and
`25°C for 14.2 and 31.1, but at lower temperatures for
`29.5 and between 25 and 3TC for 11 .1. It is unlikely
`therefore that the change in association rate with
`temperature can be accounted for by a fall in the
`viscosity of the surrounding medium or by simple
`alterations in the fluidity of the membrane which
`would be a characteristic of the cell.
`The equilibrium constant is dependent on both
`association and dissociation rate constants and it has
`been suggested that increasing temperature may be
`expected to produce a decrease in the equilibrium
`constant (or affinity) because the small increase in
`association rate is more than compensated by a large
`increase in dissociation rate. Mason and Williams
`(1986) state that there is no theoretical reason to
`suppose this would be universally the case. We have
`found a graded decrease in the value for the equi(cid:173)
`librium constant in three of the four antibodies
`studied. In the case of antibody 31.1 although a small
`decrease in K,q was observed between I and 37°C, the
`value for Keq was highest at I T C. It is interesting to
`note that differences in K.q between the three anti(cid:173)
`Ly-2.1 antibodies at I "C were greater than the differ(cid:173)
`ences between them at 3TC. Antibody I I.I had a Keq
`38 times the Keq of 31.1 at I "C , but only twice the
`value at 37°C. It is possible that when considering
`antibodies to the same antigen, the higher affinity
`antibodies are more affected by temperature changes
`than the lower affinity antibodies, although this could
`be merely a reflection on the effect of temperature on
`the association rate of the antibodies concerned.
`The study highlights the complex phenomena oc(cid:173)
`curring within the antibody- antigen binding site. In
`
`

`

`Temperature effects on antibody binding kinetics
`
`333
`
`most cases it would appear that antibody affinity does
`decrease with increasing temperature, but as usual
`with monoclonal antibodies, generalized statements
`cannot be made. To discount subclass associated
`differences, we included two antibodies of the same
`subclass but to different antigens; these two anti(cid:173)
`bodies showed unrelated changes in kinetic binding
`parameters with temperature. For each of the four
`antibodies studied both association and dissociation
`rates increased with temperature, increases in associ(cid:173)
`ation rates were less in each case than the increase
`in dissociation rate. Some effect of temperature was
`seen on antibody affinity but this was not always
`consistent.
`The study shows that although in most cases
`antibody-antigen reactions follow basic thermo(cid:173)
`dynamic principles resulting in increased rates of
`association and dissociation with temperature, not
`all antibodies are affected by temperature changes
`to the same degree. This could be an important fact
`when deciding which monoclonal antibody is best
`for clinical therapy and imaging of neoplasms. The
`results indicate that in vitro binding experiments
`at 37"C need to be carried out if an antibody is to
`be used clinically on the basis of its affinity to a
`particular antigen.
`
`REFERENCES
`
`Andrew S. M., Johnstone R. W., Pietersz G. A. and
`McKenzie I. F. C. ( 1990) A simple, rapid method for the
`measurement of the kinetic binding parameters, and
`apparent affinity of monoclonal antibodies for cell surface
`antigens using flow cytometry. J.
`Immunol. Meth.
`(in press).
`Forni L. (1979) Reagents for immunofluorescence and their
`use for studying lymphoid populations. In Immunological
`Methods (Edited by I. Lefkovits and B. Perris), p. 151.
`Academic Press, New York.
`Froese A. ( 1968) Kinetic and equilibrium studies on 2,4-
`dinitrophenyl hapten- a ntibody systems. Jmmunochem(cid:173)
`isrry 5, 253-264.
`
`Hogarth P. M ., Edwards J., McKenzie I. F . C., Goding
`J. W. and Liew F. Y. (1982a) Monoclonal antibodies to
`the murine Ly-2.1 cell surface antigen. Immunology 46,
`135-144.
`Hogarth P. M., Henning M. M. and McKenzie I. F . C.
`( 1982b) Alloantigenic phenotype of radiation induced
`thymomas in mice. J. Natl Cancer Inst. 69, 619--{i25.
`Linden C. D., Wright K. C., McConnell H . H. and Fox
`C. F. (1973) Lateral phase separations in membrane lipids
`and the mechanism of sugar transport in Escherichia coli.
`Proc. Natl Acad. Sci. U.S.A . 70, 2271-2275.
`Mason D. W. and Williams A. F . (1980) The kinetics of
`antibody binding to membrane antigens in solution and
`at the cell surface. Biochem. J. 187, 1-20.
`Mason D. W. and Williams A. F. (1986) Kinetics of
`antibody reactions and the analysis of cell surface
`antigens. In Handbook of Experimental Immunology,
`Vol. I, Immunochemistry, 4th Edn (Edited by D. M.
`Weir), p. 38.1. Blackwell, Oxford.
`Pimm M. V., Embleton M. J., Perkins A. C., Price M . R.,
`Robins R. A., Robinson G. R. and Baldwin R. W. (1982)
`In vivo localization of anti-osteogenic sarcoma 79IT
`monoclonal antibody in osteogenic sarcoma xenografts.
`Int. J. Cancer 30, 75- 85.
`Robins R. A., Laxton R. R., Garnett M., Price M. R.,
`Baldwin R. W. (1986) Measurement of tumour reactive
`antibody and antibody conjugate by competition, quanti(cid:173)
`tated by flow cytofluorimetry. J . lmmunol. Meth. 90,
`165- 172.
`Roe R., Robins R. A., Laxton R. R. and Baldwin R. W.
`( 1985) Kinetics of divalent monoclonal antibody binding
`to tumour cell surface antigens using flow cytometry
`standardization and mathematical analysis. Molec. Im(cid:173)
`munol. 22, 11- 21.
`Smyth M. J., Pietersz G . A., Classon 8. J. and McKenzie
`I. F. C. (1986) Specific targeting of chlorambucil to
`tumours using monoclonal antibodies. J . Natl Cancer
`Inst. 76, 503-510.
`Steward M. W. and Steensgaard J. (1985) In Antibody
`Affinity: Thermodynamic Aspects and Biological Signifi(cid:173)
`cance. CRC Press, Boca Raton.
`Sutton V. R. (1984) Cell surface antigens. Ph.D. Thesis,
`University of Melbourne.
`Trucco M. and de Petris S. ( 1981) Determination of
`equilibrium binding parameters of monoclonal antibodies
`specific for cell surface antigens. In Immunological
`Methods II (Edited by I. Lefkovits and B. Perris), pp.
`1- 20. Academic Press, New York.
`
`