Affinity Enhancement Bivalent Morpholinos for Pretargeting: Surface

We recently reported that one bivalent MORF (phosphorodiamidate morpholino, a DNA analogue oligomer) exhibited affinity enhancement (ratio of bivalent...
0 downloads 0 Views 142KB Size
Bioconjugate Chem. 2005, 16, 1098−1104

1098

Affinity Enhancement Bivalent Morpholinos for Pretargeting: Surface Plasmon Resonance Studies of Molecular Dimensions Jiang He,* Xinrong Liu, Surong Zhang, Guozheng Liu, and Donald J. Hnatowich Division of Nuclear Medicine, Department of Radiology, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, Massachusetts 01655. Received March 4, 2005; Revised Manuscript Received July 18, 2005

Bivalent effectors have been reported to provide superior pretargeting by affinity enhancement. We recently reported that one bivalent MORF (phosphorodiamidate morpholino, a DNA analogue oligomer) exhibited affinity enhancement (ratio of bivalent to monovalent equilibrium constants for binding) to immobilized complementary DNA (cDNA) by surface plasmon resonance (SPR). Because bivalent effectors using oligomers are easily synthesized with molecular spacing between binding sites, an important determinant of binding, adjustable simply by selecting linkers of different dimensions and/ or lengthening or shortening the oligomer chain length, they may have advantages over existing bivalent effectors. We synthesized four bivalent MORFs with different dimensions between binding sites and measured binding affinities and affinity enhancement by SPR. An 18 mer (MORF18) was made bivalent by dimerizing both with disuccinimidyl suberate (DSS) and disuccinimidyl glutarate (DSG) linkers. By again using DSS but adding seven nonbinding adenine bases and by eliminating six binding bases, a total of four bivalent effectors, DSS-MORF12, DSG-MORF18, DSS-MORF18, and DSS-MORF25, were prepared with two different hybridization affinities (i.e. MORF12 and MORF18/ 25) and three different spacings (i.e. 20, 25, and 100 Å) between binding sites. The biotinylated cDNA was immobilized on a sensor chip at 500 and 100 RU coating densities providing an average cDNA separation of 25 and 80 Å. As expected, bimolecular binding dominated monomolecular binding in all cases, especially at lower MORF effector concentrations and at higher coating densities. All bivalent MORFs showed equilibrium constants superior to their monovalent form and therefore affinity enhancement. DSS-MORF25 showed the highest equilibrium constant for bimolecular binding presumably because of its larger separation between binding sites. Nevertheless, DSS-MORF12 showed the largest affinity enhancement of almost 3 orders of magnitude presumably because the shorter chain lowered the equilibrium constant for monomolecular binding. We have shown that bivalent effectors may be easily synthesized using MORF. The results provide further evidence that the use of bivalent effectors can greatly improve MORF pretargeting and, finally, that bivalent MORFs with reduced equilibrium constants may actually exhibit higher affinity enhancement.

INTRODUCTION

Pretargeting has been intensively studied for tumor imaging and therapy (1-4). Two pretargeting methods have entered clinical trials using either bispecific antibodies with specificity for both tumor antigens and a radiolabeled effector or several avidin/streptavidinbiotin recognition systems (5-8). The success claimed for the former approach has been attributed in large part to affinity enhancement resulting from the bridging of two antibodies on the tumor surface by the radiolabeled bivalent effector (i.e. “affinity enhancement system” or AES). The bivalent effectors used successfully include AcLys(In-DTPA)-Tyr-Lys(In-DTPA)-Lys(TscG-Cys-)-NH2 (7, 9), Janus bivalent hapten (10), pGlu-Leu-Tyr-Glu-AsnLys(DTPA)-Pro-Arg-Arg-Pro-Tyr-Ile-Leu and DTPA-GlyGlu-Leu-Tyr-Glu-Asn-Lys(Ac)-Pro-Arg-Arg-Pro-Tyr-IleLeu (DTPA-Gly-NT) (11), and Ac-Phe-Lys(-DTPA)-TyrLys(DTPA)-NH2 (12) and have been variously radiolabeled with 99mTc, 188Re, 125I, 131I, 111In, and most recently 68Ga for radioimmunodetection and radiotherapy in tumor animal models (7, 13-16). One observation from these investigations has been the importance of molecular spacing between binding sites. It may be estimated that about * To whom correspondence should be addressed. Phone: 508856-2187. Fax: 508-856-4572. E-mail: [email protected].

25 Å between the binding sites has been successfully used previously (12). Furthermore, a recent publication has shown that a bivalent receptor binding agent showed a 200-fold increase in affinity as the spacer between binding sites varied from 4 to 20 carbon atoms (i.e. 5-25 Å) (17). This laboratory has previously reported encouraging results in tumored mice with pretargeting but using DNA and its analogues (collectively: oligomers) in which an antitumor antibody conjugated with an oligomer, currently a phosphorodiamidate morpholino (MORF), serves as the first injection while the radiolabel is administered later on the complementary oligomer, currently a complementary MORF (i.e. cMORF), as effector (18-21). One advantage of using oligomers for pretargeting is the ease with which bivalent effectors for affinity enhancement pretargeting may be constructed and the ease with which the molecular dimension separating the binding sites may be shortened or elongated. In a previous paper (22), we reported initial observations of affinity enhancement. We now describe further evaluations by surface plasmon resonance (SPR) of the influence of spacing, effector concentration, coating density, and oligomer chain length on affinity enhancement. Bivalent MORFs were synthesized with molecular dimension varied either by attaching linkers of different dimension or by elongating or shortening the oligomer chain.

10.1021/bc050061s CCC: $30.25 © 2005 American Chemical Society Published on Web 09/01/2005

Affinity Enhancement Bivalent Morpholino for Pretargeting

Figure 1. Molecular structures of DSS and DSG.

Specifically, an 18 mer amino-derivitized MORF (i.e. MORF18) was made bivalent by dimerizing both with disuccinimidyl suberate (DSS) and disuccinimidyl glutarate (DSG) linkers. Again using DSS as linker, the bivalent MORF was further elongated by adding seven bases (all adenines and therefore not participating in hybridization). Thus DSG-MORF18, DSS-MORF18, and DSS-MORF25 were prepared with 20, 25, and 100 Å between binding sites. Investigations of these three bivalent MORFs suggested that further improvements might result from reducing the number of hybridizing bases by subtracting six bases. Therefore, by again using DSS as linker, DSS-MORF12 was synthesized with 25 Å between binding sites but with a lower binding affinity by virtue of fewer bases. SPR was used to measure the affinity (i.e. equilibrium constant) for bivalent and monovalent binding of each of the four MORFs and affinity enhancement defined as the ratio of equilibrium constants of the bivalent over the corresponding monovalent MORF. To measure these properties, a biotinylated cDNA was immobilized on a sensor chip at two densities and the hybridization properties of bivalent and monovalent MORFs were measured. MATERIALS AND METHODS

All MORFs and cMORFs [collectively: (c)MORF] were purchased prepurified with amine modification on the 3′ equivalent end (Gene-Tools, Corvallis, OR) and were used as received. The MORF linker of this investigation was C(O)OCH2C6H4CH2NH2 except in the case of MORF25 in which only the C(O)CH2CH2NH2 linker was supplied by the manufacturer. Because of the longer dimensions of the bivalent MORF25s, the 7-8 Å difference in linker dimensions is immaterial. Biotinylated (c)DNA having the same sequences as the (c)MORF were purchased from Qiagen Inc. (Valencia, CA) and were also used as received. Biotinylated (c)DNAs were obtained with a flexible tetra(ethylene glycol) (TEG) linker of about 26 Å in length. Disuccinimidyl suberate (DSS) and disuccinimidyl glutarate (DSG) were purchased from Pierce (Rockford, IL) while anhydrous dimethylformamide (DMF), N-methyl-2-pyrrodinone (NMP), and diisopropylethylamine (DIEA) were from Sigma-Aldrich (St Louis, MO). All other chemicals were reagent grade and were used without purification. The DSS-MORF18 used in this investigation is identical to the bivalent MORF used previously by this laboratory (22). Synthesis of Bivalent MORF. The structures of DSS and DSG linkers are shown in Figure 1. When the bivalent MORF is prepared by the attachment of one MORF to each end of the linker, the DSG linker adds about 20 Å to the overall length of the bivalent MORF while the DSS linker adds about 25 Å. The preparation of bivalent MORFs was as previously described (22). Typically, about 1.5 mg (0.25 µmol) of amine-derivitized

Bioconjugate Chem., Vol. 16, No. 5, 2005 1099

MORF was dissolved in 1.0 mL of 1-methyl-2-pyrrolidinone (NMP). A 0.125 µmol amount of DSS or DSG in NMP solution at 2 mg/mL was added, followed by 0.66 µmol of N,N-diisopropylethylamine (DIEA). This mixture was incubated at room temperature for 2 d. The bivalent MORF was purified by Q Sepharose ion exchange (IE) HPLC. The collected fractions of bivalent MORFs were passed through a PD-10 column in 50 mM phosphate buffer. The molecular weight of the product was confirmed by MALDI-TOF before storage at freezer temperatures for future use. The spacings between the two binding regions are about 20, 25, and 100 Å for DSGMORF18, DSS-MORF18, and DSS-MOR25, respectively. The average distance between binding sites of the immobilized cDNA on the sensor chip may be estimated using the manufacturer’s value of 10000 pg streptavidin/ mm2 on the SA chip (i.e. 1.66 × 10-13 mol streptavidin/ mm2). Since the sensorgram response is directly related to the concentration bound, it may be estimated that saturating the streptavidin coating with cDNA would provide a response of about 4000 RU (response units). From this a density of about 99 × 10-12 mm2/DNA and 19.8 × 10-12 mm2/DNA can be estimated for a coating density of 100 RU and 500 RU. Therefore, the distance between the attachment of two neighboring cDNAs would be 80 Å and 25 Å for 100 RU and 500 RU, respectively. However, since the nonbinding regions of each immobilized cDNA is estimated to be about 26 Å and the immobilized cDNAs are capable of free rotation about their tether, with the exception of the shortest bivalent (DSG-MORF18) at the lowest cDNA density, all bivalent MORFs of this investigations were capable of bridging cDNAs at both densities. Chromatography. Ion exchange chromatography was used to purify the bivalent from unreacted monovalent MORFs as previously described (22), taking advantage of the fact that tyrosine bases will be negatively charged at pH 12. Ion exchange (IE) HPLC was performed using a HiTrap Q HP column (Amersham Biosciences, Uppsala, Sweden) with 0.01 N NaOH (solvent A) and 0.01 N NaOH in 2 N NaCl (solvent B) at a flow rate of 1.0 mL/min, going from 100% A to 70% A over 30 min. Before each analysis, the column was stabilized with 0.01 N NaOH and after each analysis the column was cleaned with solvent B. For small scale preparative purification, samples were loaded on the column by several injections in running solvent A but without exceeding the column’s maximum binding capability. The bivalent MORF was collected in fractions during the gradient elution. Surface Plasmon Resonance Measurements. SPR was performed on a BIAcore 2000 (BIAcore, Piscataway, NJ) instrument operating at room temperature as previously described (22). As before, (c)DNAs rather than (c)MORFs were immobilized because of earlier difficulties in immobilizing biotinylated (c)MORFs. Biotinylated (c)DNA at 20 nM was added in 5-10 µL aliquots to a new streptavidin-dextran-coated sensor chip (SA) at a flow rate of 20 µL/min only until a response of about 100 ((10) or 500 ((30) RU was reached. The absence of mass transfer effects was confirmed by running separately one concentration of free MORF at three different flow rates (10, 30, and 75 µL/min) and demonstrating identical response and curve shape for all three sensorgrams. When the identical measurement was performed on a sensor chip to which cDNA was added to a response of 750 RU, evidence of mass transfer effects was observed and accordingly this chip was not used in subsequent studies. Solutions of free monovalent and bivalent MORFs

1100 Bioconjugate Chem., Vol. 16, No. 5, 2005

He et al.

Figure 2. Normalized sensorgrams showing similar association kinetics but strikingly different dissociation kinetics for DSSMORF25 (trace 1), DSS-MORF18 (trace 2), DSG-MORF18 (trace 3), DSS-MORF12 (trace 4), MORF18 (trace 5), MORF25 (trace 6), and MORF12 (trace 7). Obtained at 10 nM concentration and at 100 RU coating density. Table 1. Properties of Bivalent MORFs (base sequences of native MORFs, MW calculated, MW found) bivalent MORF DSG-MORF18 DSS-MORF18 DSS-MORF25 DSS-MORF12

MORF sequence TCTTCTACTTCACAACTA TCTTCTACTTCACAACTA TCTTCTACTTCACAACTAAAAAAAA TCTTCTACTTCA

MW expected (found ( SD), Da 12424 (12397( 27) 12466 (12474( 8) 17034 (16998( 36) 8508 (8498( 10)

were prepared at six concentrations of MORF from 0.0 to 40 nM in the same running buffer (10 mM HEPES, 150 mM NaCl, 3.4 mM Na2EDTA, 0.005% P20, pH 7.4) and injected separately onto the active (cDNA) or control (DNA) surface at a flow rate of 30 µL/min to minimize the mass transfer effect. The chip surface was regenerated by injection of 100 mM HCl. To correct for nonspecific binding and refractive index changes, the control responses were subtracted from those obtained from the active surface. A minor baseline drift resulting from a slow dissociation of the complex on the active and control surfaces was eliminated by also subtracting sensorgrams obtained following the injection of running buffer (23). In the first series of experiments, samples of monovalent MORFs (MORF18 and MORF25) and bivalent MORFs (DSS-MORF18, DSG-MORF18, DSS-MORF25) were run at six concentrations on both the 100 and 500 RU chips and with a dissociation time of only 25 min. In the second series of experiments, all six concentrations were again run but with dissociation time of 1 h to obtain more accurate values for the dissociation rate constants. In the third series of experiments, the monovalent MORF12 and bivalent DSS-MORF12 were run at the six concentrations on both 100RU and 500RU chips for 25 min. In every case, the measurement was repeated four times and, as mentioned above, each sample was run on both the cDNA and DNA chips at the same surface density so that results with the latter could be used to correct for bulk refractive index changes (23). Sensorgram Kinetic Analysis. Sensorgram curves were evaluated using numerical integration algorithms (BIAevaluation 3.0, BIAcore, Piscataway, NJ). The algorithm may be programmed for simple monomolecular 1:1 Langmuir interactions of a monovalent effector (A) or for

a bivalent effector (AA) binding monomolecularly (i.e. A+LfAL where A is the effector, in this case MORF, and L is immobilized ligand, in this case cDNA) or binding bimolecularly (i.e. AA+LfAAL and AAL+ LfLAAL). In addition, the algorithm may be programmed to analyze globally (i.e. all effector concentrations considered together to provide one value for each of the two rate constants, association and dissociation) or locally (i.e. each effector concentration considered separately to provide one value for each of the two rate constants for each effector concentration). An added complication in the case of bivalent effector binding is the possibility that the binding may be both monomolecular and bimolecular and thus cannot be accurately modeled. To emphasis this fact, the rate constants in this report for bivalent MORFs are referred to as “apparent”. While different model and fitting methods were used by us previously, in this study, the simple monomolecular 1:1 Langmuir interaction model was used to compare the four bivalent MORFs with their corresponding monovalent form. It should be noted that in each comparison, the concentrations were adjusted to the same molarity of effector and therefore the MORF molarity of the bivalent effector was twice that of the monovalent effector. Finally, the program provides for each analysis a Chi2 (χ2) value that is a statistical measure of how closely the model fits the experimental data. In general, χ2 values lower than about 10 signify a good fit. RESULTS

Synthesis of Bivalent MORF. The amine derivatized MORFs were dimerized with DSS or DSG for 2 days in the presence of NMP and DIEA followed by purification by ion exchange chromatography. Changing from alkaline solution to phosphate buffer or water was performed on a PD-10 column. The correct structure of all four bivalent MORFs was confirmed by MALDI-TOF MS as shown in Table 1. Surface Plasmon Resonance Measurements. The SPR measurements were performed at both cDNA surface densities of 100 and 500 RUs and at six concentrations of the monovalent and bivalent MORFs in the concentration range from 0 to 40 nM. Figure 2 presents

Bioconjugate Chem., Vol. 16, No. 5, 2005 1101

Affinity Enhancement Bivalent Morpholino for Pretargeting

Table 2. Kinetics Rate Constants and Equilibrium Constants for Monovalent MORF12, MORF18, and MORF25 and the Apparent Kinetic Rate Constants and Equilibrium Constants for the Corresponding Bivalent MORFs Obtained at Both CDNA Surface Densitiesa density

cDNA effectors

ka (1/Ms) × 105

kd (1/s) × 10-5

Kequ (1/M) × 1010

100RU

DSG-MORF18 DSS-MORF18 DSS-MORF25 DSS-MORF12 MORF12 MORF18 MORF25 DSG-MORF18 DSS-MORF18 DSS-MORF25 DSS-MORF12 MORF12 MORF18 MORF25

1.15 (0.2) 8.10 (2.1) 6.25 (3.6) 3.31 (1.32) 0.206 (0.02) 2.62 (0.19) 1.89 (0.13) 2.78 (0.41) 13.7 (3.56) 5.54 (1.74) 12.86 (4.6) 0.38 (0.04) 4.80 (0.23) 3.31 (0.21)

6.25 (0.29) 4.85 (0.57) 2.97 (0.31) 13.5 (0.69) 419 (38.7) 24 (3.4) 20 (2.6) 1.18 (0.38) 3.55 (0.47) 0.98 (0.16) 26.2 (0.43) 826 (77) 48 (5.1) 31 (4.2)

0.18 (0.03) 1.67 (0.29) 2.10 (0.13) 0.246 (0.16) 0.00049 0.11 (0.02) 0.095 (0.02) 2.36 (0.35) 3.87 (0.61) 5.68 (0.72) 0.477 (0.13) 0.00046 0.10 (0.02) 0.11 (0.01)

500RU

χ2

affinity enhancement

6.7 5.1 3.6 7.9 0.93 6.1 7.8 14 25 9.5 1.7 1.2 2.01 4.9

1.6 15 22 502

24 39 52 1036

a Analysis by global fitting using the 1:1 Langmuir interaction model. k : association rate constant; k : dissociation rate constant; a d Kequ: equilibrium constant; X2: index for goodness of fit. Standard deviations in parentheses.

normalized sensograms of the four bivalent and the three monovalent MORFs at 10 nM concentration and at 100 RU cDNA coating density. The sensorgram of all seven MORFs have been overlayed to illustrate the striking differences in dissociation rate. Among the bivalent MORFs, the order of increased stability to dissociation is DSSMORF25 > DSSMORF18 > DSGMORF18 > DSSMORF12. Among the monovalent MORFs, this order is MORF25 ∼ MORF18 . MORF12 (there should be no significant difference between MORF25 and MORF18 since they have the same binding region in this study). Furthermore the bivalent MORF12 exhibited stability comparable to that of the larger monovalent MORF25. Finally, except for this similarity, all bivalent MORFs showed increased stability over all monovalent MORFs. The kinetic rate constants and equilibrium constants for monovalent MORF12, MORF18, and MORF25 and the apparent kinetic rate constants and equilibrium constants for the corresponding bivalent DSG-MORF18, DSS-MORF18, DSS-MORF25, and DSS-MORF12 obtained at both cDNA surface densities are presented in Table 2. Analysis was by global fitting using the simpler 1:1 Langmuir interaction algorithm. Included in the table are the ratios of affinity constants for bivalent over corresponding monovalent binding at both cDNA densities. As shown, the equilibrium constants for the three monovalent MORFs binding are little changed between cDNA densities. Figure 3 presented sensorgrams obtained at 5 nM for the bivalent and monovalent MORF12s only. At both surface densities, the bivalent MORF12 shows a profoundly higher association and lower dissociation phase and therefore a significantly improved equilibrium constant due to bimolecular binding. Considering first only DSG-MORF18, DSS-MORF18, and DSS-MORF25, although these three bivalent MORFs possess the same binding region, they show statistically significantly (p < 0.04) different affinity enhancement presumably as a result of the different molecular dimension. As shown in the table, at low cDNA coating density of 100 RU, both DSS-linked MORFs show affinity enhancements of about 20-fold compared to about 2 for the shorter DSG-MORF18. This result could be predicted on the basis of molecular dimensions. At the lowest 100 RU coating density, the separation between attached cDNAs is estimated to be about 80 Å. The maximum radius of each cDNA when fully extended is about 26 Å, leaving a gap of about 28 Å. Because of its longer 25 Å length, bivalent MORFs attached with the DSS linker are capable of bridging this gap. By contrast, bivalent

MORFs attached via the 20 Å DSG linker are too short and therefore are capable only of monovalent binding with the result that affinity enhancement is prevented. However, at high coating density, all three sets of bivalent MORFs showed superior Kequ with the highest improvement of 40-50-fold for DSS-MORF18 and DSSMORF25. These results are reasonable since high coating densities should favor bivalent binding of the bivalent MORFs. Only in the case of low coating density and the bivalent MORF with the shortest linker, DSG-MORF18, was the enhancement minimal over the monovalent MORF. Furthermore, the bivalent MORF12 showed much higher association rate constants along with lower dissociation rate constant both at 100 RU or 500 RU cDNA coating density compared with monovalent MORF12 to provide equilibrium constants 500 to1000-fold higher (i.e. 500 to1000 higher affinity enhancement) at both cDNA densities. The striking affinity enhancement due to bimolecular binding in the case of bivalent MORF is clearly seen in Figure 3 by comparison of sensorgrams with the monovalent MORF12 at both coating densities. Sensorgrams of DSS-MORF12 show greater response in association phase at each concentration (data not presented). Furthermore, the dissociation of monovalent MORF12 is essentially completed by about 1500 s while at this time DSS-MORF12 is still about 55% and 87% intact at 100 RU and 500 RU, respectively. DISCUSSION

Earlier this laboratory reported that binding of one bivalent MORF to immobilized cDNA was bimolecular under the conditions of that study and, presumably as a result, exhibited a faster association rate and a slower dissociation rate and therefore a higher affinity constant and affinity enhancement. Since the probability of monomolecular binding (AAL) compared to bimolecular binding (LAAL) will depend on the molecular dimensions of the bivalent effector (in this case MORF) and the concentration (spacing) and flexibility of the ligand molecules (in this case cDNA) on the immobilized surface (24), in this present investigation, SPR was used to compare the behavior of four bivalent MORFs differing in the spacing between binding sites with their corresponding monovalent form during hybridization to immobilized cDNA at two densities In general, low cDNA densities and higher MORF effector concentrations will favor monomolecular binding,

1102 Bioconjugate Chem., Vol. 16, No. 5, 2005

He et al.

Figure 3. Sensorgrams obtained at 5 nM for the bivalent and monovalent MORF12 at 100RU (panel A) and 500RU (panel B) surface densities showing the significantly improved association and dissociation constant of the bivalent MORF12 due to bimolecular binding. The MORF concentrations of the bivalent MORF are twice the effector concentrations.

in the first instance because of the greater spacing between cDNAs (thus making more difficult bridging by the effector) and in the latter instance because the cDNAs are more likely to be saturated with MORF bound monomolecularly (since bimolecular binding is a two-step process initially requiring monomolecular binding). Therefore, six concentrations not exceeding 40 nM (to avoid monomolecular binding saturation) of all effectors and two coating densities of 100 RU and 500 RU were used in this investigation. As shown in Table 2, presumably because of its shorter spacing between binding sites, the bivalent DSG-MORF18 showed affinity enhancement at the 500 RU cDNA density but not at the lower 100 RU density. By contrast, the bivalent MORFs DSS-MORF18 and DSS-MORF25 with their longer spacing between binding sites showed comparably higher affinity enhancement compared to their corresponding monovalent MORF18 and MORF25 respectively at both densities. It is important to point out that in the sensorgram analysis, quantitative evaluation was made difficult by

the high cDNA densities required to achieve bimolecular binding. These high densities can result in mass transport effects and, furthermore, can complicate the analysis by permitting rebinding during dissociation. In this investigation it was possible to prevent mass transport complications by adjusting the flow rate to be in a range where the response was shown to be flow-rate-independent. However, it was not possible to eliminate rebinding effects especially in case of effectors with very high intrinsic affinities. Thus rebinding contributed partially to the resulting kinetic analysis. This was the case for monovalent MORF18, monovalent MORF25, and all bivalents due to their high affinity. Rebinding effects may be suspected when the dissociation rates are slower at lower effector concentrations compared to higher concentrations. Normally all concentrations should give the same affinity constants. At very low concentration as used in this investigation, the cDNA is not saturated at the high coating density and thus rebinding is possible. Under these conditions, the dissociation rates are slower

Affinity Enhancement Bivalent Morpholino for Pretargeting

at low effector concentrations compared to higher concentrations. That the dissociation rate of monovalent 12 mer MORF at all six concentrations was identical even at the highest 500 RU density (data not presented) may imply little rebinding in the case of this effector. This result is presumably due to an intrinsic low affinity resulting from the shorter chain length. Therefore the “rebinding” (lower dissociation rate at lower concentration) for the bivalent 12 mer MORF (DSS-MORF12) must be due to the enhanced affinity from bimolecular binding. However, while rebinding will complicate accurate analysis of SPR results, it may favorably contribute to a longer retention of the effector in tumor. The importance of selecting the proper molecular dimensions in oligomer effectors intended for affinity enhancement pretargeting rests in the promise of greatly improved target/nontarget ratios. A higher equilibrium constant for bimolecular binding over monomolecular binding (i.e. affinity enhancement) would favor dissociation and minimal binding in blood and other nontarget tissues because of the lower cMORF densities in these sites at the time of effector administration. One result of this investigation is the observation that higher affinity of binding (i.e. higher equilibrium constants) of effector does not necessarily translate into higher affinity enhancement. Since affinity enhancement is the ratio of equilibrium constants for bimolecular over monomolecular binding, a high equilibrium constant for monomolecular binding may reduce the ratio to insignificance. This is apparently the case for DSS-MORF18 and DSSMORF25, both of which show higher equilibrium constants than DSS-MORF12 yet both show lower affinity enhancements. CONCLUSION

As shown herein, the distance between binding sites in MORF effectors may be adjusted simply by changing the linker length and/or by elongating or shortening the base chain. These results also show that bimolecular binding was occurring in the case of the bivalent MORF by the higher equilibrium constants for the bivalent MORF compared to their corresponding monovalent form. Furthermore, these results also provide further evidence that bivalency may be superior to monovalency in MORF pretargeting applications in the affinity enhancement observed in all cases of bivalent MORF and cDNA densities. However, since affinity enhancement is the ratio of equilibrium constants for bimolecular over monomolecular binding, that the bivalent MORF derivatized from shorter chain effector (i.e., DSS-MORF12) showed the greatest affinity enhancement suggests that higher equilibrium constant and higher affinity enhancement may not be synonymous. ACKNOWLEDGMENT

Financial support for this investigation was provided, in part, by the National Institutes of Health (CA 94994). LITERATURE CITED (1) Goodwin, D. A., Meares, C. F., McTigue, M., Mccall, M. J., and David, G. S. (1986) Rapid localization of haptens in sites containing previously administered antibody for immunoscintigraphy with short half-life tracers [abstract]. J. Nucl. Med. 27, 959. (2) Chang, C., Sharkey, R. M., Rossi, E. A., Karacay, H., McBride, W., Hansen, H. J., Chatal, J., Barbet, J., and Goldenberg, D. M. (2002) Molecular advances in pretargeting radioimmunotherapy with bispecific antibodies. Mol. Cancer Ther. 1, 553-563.

Bioconjugate Chem., Vol. 16, No. 5, 2005 1103 (3) Hnatowich, D. J., Virzi, F., and Rusckowski, M. (1987) Investigations of avidin and biotin for imaging applications. J. Nucl. Med. 28, 1294-302. (4) Kuijpers, W. H., Bos, E. S., Kaspersen, F. M., Veeneman, G. H., and van Boeckel, C. A. (1993) Specific recognition of antibody-oligonucleotide conjugates by radiolabeled antisense nucleotides: a novel approach for two-step radioimmunotherapy of cancer. Bioconjugate Chem. 4, 94-102. (5) Hamblett, K. J., Kegley, B. B., Hamlin, D. K., Chyan, M. K., Hyre, D. E., Press, O. W., Wilbur, D. S., and Stayton, P. S. (2002) A streptavidin-biotin binding system that minimizes blocking by endogenous biotin. Bioconjugate Chem. 13, 588-598. (6) Wilbur, D., Pathare, P. M., Hamlin, D. K., and Weerawarna, S. A. (1997) Biotin Reagents for Antibody Pretargeting. 2. Synthesis and in Vitro Evaluation of Biotin Dimers and Trimers for Cross-Linking of Streptavidin. Bioconjugate Chem. 8, 819-832. (7) Karacay, H., McBride, W. J., Griffiths, G. L., Sharkey, R. M., Barbet, J., Hansen, H. J., and Goldenberg D. M. (2000) Experimental pretargeting studies of cancer with a humanized anti-CEA x murine anti-[In-DTPA] bispecific antibody construct and a 99mTc-/188Re-labeled peptide. Bioconjugate Chem. 11, 842-854. (8) Gruaz-Guyon, A., Janevik-Ivanovska, E., Raguin, O., De Labriolle-Vaylet, C., and Barbet, J. (2001) Radiolabeled bivalent haptens for tumor immunodetection and radioimmunotherapy. Q. J. Nucl. Med. 45, 201-6. (9) Cardillo, T. M., Karacay, H., Goldenberg, D. M., Yeldell, D., Chang, C. H., Modrak, D. E., Sharkey, R. M., and Gold D. V. (2004) Improved targeting of pancreatic cancer: experimental studies of a new bispecific antibody, pretargeting enhancement system for immunoscintigraphy. Clin. Cancer Res. 10, 3552-3561. (10) Goodwin, D. A., Meares, C. F., McTigue, M., Chaovapong, W., Diamanti, C. I., Ransone, C. H., and McCall M. J. (1992) Pretargeted immunoscintigraphy: Effect of hapten valency on murine tumor uptake. J. Nucl. Med. 33, 2006-2013. (11) Hillairet De Boisferon, M., Raguin, O., Thiercelin, C., Dussaillant, M., Rostene, W., Barbet, J., Pelegrin, A., and Gruaz-Guyon, A. (2002) Improved tumor selectivity of radiolabeled peptides by receptor and antigen dual targeting in the neurotensin receptor model. Bioconjugate Chem. 13, 654662. (12) Boerman, O. C., van Eerd, J., Oyen, W. J. G., and Corstens, F. H. M. (2001) A 3-Step Pretargeting Strategy to Image Infection. J. Nucl. Med. 42, 1405-1411. (13) Barbet, J., Peltier, P., Bardet, S., Vuillez, J. P., Bachelot, I., Denet, S., Olivier, P., Leccia, F., Corcuff, B., Huglo, D., Proye, C., Rouvier, E., Meyer, P., and Chatal, J. F. (1998) Radioimmunodetection of medullary thyroid carcinoma using indium-111 bivalent hapten and anti-CEA x anti-DTPAindium bispecific antibody. J. Nucl. Med. 39, 1172-1178. (14) Gautherot, E., Rouvier, E., Daniel, L., Loucif, E., Bouhou, J., Manetti, C., Martin, M., Le Doussal, J. M., and Barbet, J. (2000) Pretargeted radioimmunotherapy of human colorectal xenografts with bispecific antibody and 131I-labeled bivalent hapten. J. Nucl. Med. 41, 480-487. (15) Griffiths, G. L., Chang, C. H., McBride, W. J., Rossi, E. A., Sheerin, A., Tejada, G. R., Karacay, H., Sharkey, R. M., Horak, I. D., Hansen, H. J., Goldenberg, D. M. (2004) Reagents and methods for PET using bispecific antibody pretargeting and 68Ga-radiolabeled bivalent hapten-peptidechelate conjugates. J. Nucl. Med. 45, 30-39. (16) Hosono, M., Hosono, M. N., Kraeber-Bodere, F., Devys, A., Thedrez, P., Fiche, M., Gautherot, E., Barbet, J., and Chatal, J. F. (1998) Biodistribution and dosimolecular study in medullary thyroid cancer xenograft using bispecific antibody and iodine-125-labeled bivalent hapten. J. Nucl. Med. 39, 1608-1613. (17) Sridhar, J., Wei, Z., Nowak, I., Lewin, N. E., Ayers, J. A., Pearce, L. V., Blumberg, P. M., and Kozilowski, A. P. (2003) New bivalent PKC ligands linked by a carbon spacer: Enhancement in binding affinity. J. Med. Chem. 46, 4196-4204. (18) Mang’era, K. O., Liu, G., Wang, Y., Zhang, Y., Liu, N., Gupta, S., Rusckowski, M., and Hnatowich, D. J. (2001)

1104 Bioconjugate Chem., Vol. 16, No. 5, 2005 Initial investigations of 99mTc-labeled morpholinos for radiopharmaceutical applications. Eur. J. Nucl. Med. 28, 16821689. (19) Liu, G., Mang’era, K., Liu, N., Gupta, S., Rusckowski, M., and Hnatowich, D. J. (2002) Tumor pretargeting in mice using 99mTc-labeled morpholino, a DNA analog. J. Nucl. Med. 43, 384-391. (20) Liu, G., Liu, C., Zhang, S., He, J., Liu, N., Gupta, S., Rusckowski, M., and Hnatowich, D. J. (2003) Investigations of technetium-99m morpholino pretargeting in mice. Nucl. Med. Commun. 24, 697-705. (21) He, J., Liu, G., Gupta, S., Zhang, Y., Rusckowski, M., and Hnatowich, D. J. (2004) Amplification targeting: a modified pretargeting approach with potential for signal amplificationproof of a concept. J. Nucl. Med. 45, 1087-1095.

He et al. (22) He, J., Liu, G., Vanderheyden, J.-L., Dou, S., Rusckowski, M., and Hnatowich, D. J. (2005) Affinity Enhancement Bivalent Morpholino for Pretargeting: Initial Evidence by Surface Plasmon Resonance. Bioconjugate Chem. 16, 338345. (23) He, J., Liu, G., Zhang, S., Vanderheyden, J.-L., Liu, N., Liu, C., Zhang, Y., Gupta, S., Rusckowski, M., and Hnatowich, D. J. (2003) A comparison of in vitro and in vivo stability in mice of two morpholino duplexes differing in chain length. Bioconjugate Chem. 14, 1018-1023. (24) Muller, K. M., Arndt, K. M., and Pluckthun, A. (1998) Model and simulation of multivalent binding to fixed ligands. Anal. Biochem. 261, 149-158.

BC050061S