APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 2008, p. 25592564 0099-2240/08/$08.000 doi:10.1128/AEM.02839-07 Copyright 2008, American Society for Microbiology. All Rights Reserved.

Vol. 74, No. 9

MINIREVIEW
Specic Molecular Recognition and Nonspecic Contributions to Bacterial Interaction Forces
Henk J. Busscher,1 Willem Norde,1,2 and Henny C. van der Mei1*
University of Groningen, University Medical Center Groningen, Department of Biomedical Engineering, Antonius Deusinglaan 1, 9713 AV Groningen, The Netherlands,1 and Wageningen University, Laboratory of Physical Chemistry and Colloid Science, P.O. Box 8038, 6700 EK Wageningen, The Netherlands2 Bacteria adhere to surfaces by virtue of their interaction forces with a substratum surface. A few decades ago, a paper on bacterial adhesion to surfaces typically would either commence with the statement (15) bacterial adhesion to surfaces is mediated by highly specic, stereo-chemical interactions between complementary components on the interacting surfaces or (16, 24) bacterial adhesion is mediated by a complicated interplay between attractive Lifshitz-Van der Waals forces and repulsive or attractive electrostatic and acid-base forces, originating from the interacting surfaces. Generally, the specic approach was favored by microbiologists and biochemists, while physico-chemists usually took a nonspecic approach. The two approaches were reconciled with each other (7, 8) by the realization that both interactions originate from the same, fundamental physico-chemical forces (Lifshitz-Van der Waals, electrostatic, and acid-base interaction) (37). Nonspecic, Lifshitz-Van der Waals interactions operate over longer distances (several tens of nanometers) and originate from all atoms in the interacting entities. The summation of the relatively weak pairwise interactions between all atoms in an adhering bacterium and a substratum yields the nal interaction force, similar to the origin of the gravitational force of the earth. Specic interactions, making up for molecular recognition between ligand and receptor molecules, operate over spatially well-conned stereochemical regions, established for instance by interactions between acid, electron-accepting and basic, electron-donating groups or oppositely charged domains, at close approach (up to several nanometers). Characterization of the bacterial cell and substratum surfaces in terms of their zeta potentials and surface free energies (from measured contact angles with liquids) offers the possibility to calculate the electrostatic and Lifshitz-Van der Waals contributions to the interaction force between two entities in an approach called the DLVO (Derjaguin, Landau, Verwey, and Overbeek) theory (5, 16). In the so-called extended DLVO theory (38), acid-base interaction forces are accounted for in addition to Lifshitz-Van der Waals and electrostatic forces. Application of physico-chemical theories toward explaining bacterial adhesion to surfaces has not always been
* Corresponding author. Mailing address: Department of Biomedical Engineering, University Medical Center Groningen, Antonius Deusinglaan 1, 9713 AV Groningen, The Netherlands. Phone: 31-503633140. Fax: 31-50-3633159. E-mail: h.c.van.der.mei@med.umcg.nl. Published ahead of print on 14 March 2008. 2559
successful, not even adhesion to inert (nonbiological) surfaces. After evaluating over 250 references, Bos et al. (5) concluded that the only general conclusion to be drawn was that negatively charged bacteria adhere more rapidly to a positively charged than to a negatively charged substratum surface. The aws of a physico-chemical approach based on overall surface characteristics become especially clear when considering bacterial adhesion to protein-coated surfaces, as illustrated by the example in Table 1. First of all, it should be noted from Table 1 that the presence or absence of antigen I/II on the streptococcal cell surface has only minor effect, if any, on the cell surface hydrophobicity by water contact angles or on the bacterial zeta potential. In a nonspecic approach, one would expect similar adhesion of the two strains to a given surface, which is indeed the case for bare glass. Based on nonspecic interactions, however, it cannot be explained why the strain with antigen I/II adheres in almost four-fold-higher numbers than the strain without antigen I/II after the glass is coated with a salivary conditioning lm. Clearly, neither the water contact angles nor zeta potentials are able to probe the presence of localized, microscopic attractive domains that constitute the molecular recognition groups on the interacting cell surfaces. With the introduction of the atomic force microscope (AFM), it has become possible to probe the physico-chemical properties of the bacterial cell surface at a microscopic level, including interaction forces between surfaces (9, 12). In this respect, it is important to realize that different interpretations can be given to the word specic. In microbial adhesion, the word specic is generally associated with molecular recognition phenomena, but sometimes it is also used to designate short-range stereochemical interactions, such as acid-base bonding (1), that are in our view not considered to be specic in the sense of molecular recognition (notwithstanding that acid-base interactions can contribute to molecular recognition). Along these lines, binding forces between bronectin-coated AFM tips and tissue-invasive and noninvasive Staphylococcus aureus strains have been compared, rather than establishing contributions of specic molecular recognition and nonspecic forces to the interaction (41). It is the aim of this minireview to provide a comparison of the specic and nonspecic contributions to the forces that mediate bacterial adhesion to inert and protein-coated surfaces, based on AFM data. To this end, we will rst briey describe how the AFM can be used to measure real-life bac-

APPL. ENVIRON. MICROBIOL.
TABLE 1. Two oral streptococci that solely differ in the presence of surface antigen I/II have similar overall surface characteristicsa
Zeta potential (mV) Adhesion (106 cm2) to: Glass Salivary conditioning lm

S. mutans strain

Water contact angle ()
LT11 (with antigen I/II) IB03987 (without antigen I/II)

12.1 11.8

9.6 2.5
a Surface characteristics include water contact angles, zeta potentials, and the ability in line to adhere equally well to an inert glass surface but not to an adsorbed lm of salivary proteins. Data were taken from studies by Xu et al. (40) and Petersen et al. (25).
terial interaction forces and single-bond molecular recognition forces. Data sets referring to bacterial interaction forces and single-bond molecular forces have appeared as separate classes of data in the literature and have never been combined. After a review of both data sets, they are combined in the section Synthesis and Conclusion in order to compare specic- and nonspecic contributions to the forces mediating bacterial adhesion. BACTERIAL ADHESION FORCES PROBED BY AFM The AFM provides an excellent tool to measure interaction forces between surfaces. Usually, approach or retraction forces are measured between the tip (27), a colloid probe (28), or a bacterium attached to the AFMs cantilever and a second surface, which can be either an inert protein-coated or immobilized bacterial cell surface (18, 35). When retracting the probe from a surface, the probe will stay in contact with the surface until the elastic restoring cantilever force of the bent cantilever overcomes the interaction forces between the surfaces under study, yielding an estimate of the attractive force at a certain distance (9). The AFM can be used in two entirely different ways. (i) By attaching bacteria to the cantilever, interaction forces can be determined between a bacterium and a substratum surface or another immobilized bacterium. These measurements are subject to considerable uncertainty regarding the actual contact area between the interacting surfaces, but this disadvantage

FIG. 2. Example of single-bond interaction forces during retraction. Disruption of each bond yields a small adhesion force during retraction. (Adapted in part from reference 34. Copyright 2003 American Chemical Society).
has to be weighed against the fact that such measurements mimic real-life conditions (see Fig. 1 for an example). Moreover, probing of interaction forces in the absence of a sharp tip may avoid damage to the inner cell surface, as has been described to happen for naturally immobilized Staphylococcus epidermidis cells on glass (21, 22). (ii) By functionalizing the AFM tip and the surface to be probed with relevant biomolecules (13, 17), single-bond molecular recognition forces can be measured by careful analysis of individual detachment events during retraction (see Fig. 2 for an example). The measurement of molecular recognition forces between ligands and receptors is highly precise and relevant to understand the specic contributions to real-life interactions between biological surfaces. REAL-LIFE BACTERIAL INTERACTION FORCES In Table 2 we have compiled some real-life bacterial interaction forces in the absence and presence of specic force contributions, as derived from AFM measurements. Nonspecic contributions to the interaction forces are always present, but specic contributions depend on the absence or presence of specic recognition, mediated by target and receptor molecules at the interacting surfaces.
FIG. 1. Example of interaction forces between a protein-coated tip and a bacterial cell surface during retraction. Interaction forces of multiple bonds between adsorbed proteins and bacterial cell surfaces are accumulated into one adhesion force peak.
VOL. 74, 2008 TABLE 2. Bacterial interaction forces in the absence and presence of specic recognitiona
Bacterial interaction (nN): Bacterial adhesion parameter With specic contribution Without specic contribution

Reference

Streptococci to salivary lms pH 5.8 pH 6.8
40 Median, 0.0 Range, 1.2 Median, 0.4 Range, 2.9 Median, 0.0 Range, 0.1 Median, 0.1 Range, 0.Median, 0.0 Range, 5.0 Median, 0.1 Range, 4.9 Mean, 3.0 to 4.0 Median, 0.0 Range, 1.5 Median, 0.1 Range, 2.1 Mean, 1.0 26
Streptococci to laminin lms pH 5.8 pH 6.8
Coaggregation between actinomyces and streptococci Aggregation between enterococci Summary pH dependence Force value

Mean, 2.3 to 2.6

Mean, 1.2 to 1.5

Increases with pH 3 to 5

Increases with pH 0 to 2
a Note that some of the older data were analyzed according to parametric statistics (force contributions represented by mean values), whereas more recent data analyses make use of nonparametric statistics (force contributions represented by median and range values of the distribution).
The data for streptococcal adhesion to salivary and laminin lms involve Streptococcus mutans LT11 and S. mutans IB03987, strains with and without the specic recognition surface protein antigen I/II, respectively. Antigen I/II is involved in the adhesion of streptococci to salivary lms and extracellular matrix proteins, like laminin. Streptococci with antigen I/II adhere stronger to salivary pellicles and laminin lms than streptococci without antigen I/II, while furthermore the adhesive forces increase with increasing pH. Nonparametric statistics have demonstrated these differences in median force val-

FIG. 3. Nonparametric distribution of adhesion forces (Fadh) between a protein-coated AFM tip and a bacterial cell surface, showing a trimodal distribution with a large range value.
ues to be signicant (6, 40), although the largest effects are seen on the range values of the distributions. The nonparametric, wide distributions generally observed in interaction force measurements by AFM (see Fig. 3 for an example) suggest that the surface characteristics are not homogeneously distributed over all bacterial cells probed during AFM. Indeed, culture heterogeneities are quite common and with the introduction of instrumentation that can measure properties of individual bacteria, like AFM, this is becoming more and more obvious. Furthermore, so-called zeta sizing has demonstrated that subpopulations with different cell surface charges exist within axenic cultures (3, 36). Different subpopulations within one culture can also differ in agellation (29), natural competence (11), or autouorescence (19). It is likely that the largest force values, as indicated by the range values in the nonparametric AFM interaction force distributions, represent a subpopulation that must be considered most relevant for adhesion: if only 1% of a culture would be represented by the range value, in a suspension of 106 bacteria per ml, this would represent 104 bacteria per ml with a strong afnity for a substratum! Alternatively, a nonparametric, wide interaction force distribution may reect a heterogeneity over the surface of one individual bacterium (6, 13, 26, 40), such as for instance described for Pseudomonas putida cells, where AFM has demonstrated a range of adhesion afnities and polymer lengths on a single bacterium (10). However, also if the wide nonparametric distributions would be indicative of heterogeneity over a single cell surface, this would still point to the importance of the range value since the highest force values measured on a single cell surface are most relevant in
TABLE 3. Overview of single-bond molecular recognition forces
Molecular pair Interaction force (nN/single bond) Reference
Avidin-biotin Avidin-iminobiotin Streptavidin-biotin Avidin-desthiobiotin Streptavidin-iminobiotin Protein D-galactose Protein D-mannose Vascular smooth muscle cell receptor-bronectin Arg-Gly-Asp ligandsintegrins on osteoblast Saccharomyces carlsbergensiscarbohydrate S. carlsbergensismannosespecic lectin Mycobacterium bovis-heparin Bacillus subtilis spore-antibody CotA Fv fragment of antilysozymelysozyme Interaction force on avg
0.160 0.085 0.257 0.094 0.135 0.038 0.054 0.039 0.0320.097 0.121 0.117 0.053 0.055 0.050 0.095
gens. The protein has various carbohydrate recognition domains which can take part in the specic interactions with carbohydrates (32). The interaction between bronectin and 51-integrin is important for the focal adhesion of vascular smooth muscle cells. The unbinding force of 0.039 nN was measured for a single bond between bronectin and 51integrin (30). The Arg-Gly-Asp (RGD) sequence within a protein has considerable inuence upon the nal binding force with integrins; forces between 0.032 and 0.097 nN have been observed (20). Aggregation of yeast like Saccharomyces carlsbergensis is important in their occulation, which controls fermentation in brewing and wine making. The lectin-carbohydrate single-bond interaction forces originating from S. carlbergensis amounts of 0.117 to 0.121 nN (33). A 0.055-nN unbinding force was measured for protein A from the outer surface of a bacillus spore (31). Antigen binding of individual Fv fragments of antilysozyme antibodies (Fv) to lysozyme was accompanied by a single-bond interaction force of 0.050 nN (4). Overall inspection of Table 3 shows that for specically interacting molecules, single-bond interaction forces reach, at the most, several tenths of an nN. SYNTHESIS AND CONCLUSIONS The aim of this review is to provide a comparison of the specic and nonspecic contributions to the forces that mediate bacterial adhesion to inert and protein-coated surfaces. Although, of course, exact values depend on the strain-substratum combination, it can be concluded that in the absence of specic contributions, bacterial interaction forces operate in the regimen up to 2 nN. In the presence of specic contributions, forces about 2 to 3 times stronger are observed, which implies that the specic contribution to an interaction force amounts between 1 and 5 nN. At the single-bond level, the molecular recognition forces that make up for a specic contribution to the interaction forces in bacterial adhesion also differ depending on the ligand-receptor system evaluated, but here too a general conclusion can be drawn, namely that these forces operate in the subnanometer regime and on average amount to 0.095 nN, albeit with variation dependent on the type of ligand-receptor pair involved. By comparison of the specic contribution to bacterial interaction forces (Table 2) with the single-bond recognition forces (Table 3), it can be calculated the specic contribution must involve 10 to 50 specic ligand-receptor bonds (that is the number of ligand-receptor pairs interacting with 0.095 nN per pair, that make up for a total specic contribution of 1 to 5 nN). The pH dependence of the specic contribution to the interaction forces measured points to the electrostatic nature of these interactions, which most likely involves ion pairing. Exact spatial stereochemistry between recognition molecules then allows for the nal specicity in bacterial selection for a given substratum surface. The nal question to be addressed is whether 10 to 50 ligand-receptor bonds involved in adhesion of one bacterium to a substratum surface is a realistic estimate. We previously addressed this question for streptococcal adhesion to laminincoated substrata (6), but this review allows more general evaluation. Based on a contact area between a micrometer-sized

terms of mediating adhesion. AFM has also been applied to measure the forces between (co-)aggregating bacteria (26) and shows that in case streptococci and actinomyces have specic recognition molecules for each other, they coaggregate more strongly (3 to 4 nN) than when the specic recognition molecules are lacking(1 nN). Analogously, enterococci may or may not have so-called aggregation proteins at their surface. Pairs with aggregation proteins aggregate more strongly (2.3 to 2.6 nN) than strains without the aggregation protein (1.2 to 1.5 nN (39). Moreover, when the specic recognition molecules on strains with aggregation proteins were blocked using antibodies, the interaction force decreased to values observed for strains lacking the aggregation protein. Overall, inspection of Table 2 shows that bacterial interaction forces in the absence of specic recognition amount on average 0 to 2 nN. When specic contributions exist in addition to the always present nonspecic contributions, the interaction forces reach average values of 3 to 5 nN. Referring to Table 2, it is concluded that a factor of a 2 to 3 difference in interaction force can have a tremendous impact on bacterial adhesion to surfaces. SINGLE-BOND MOLECULAR RECOGNITION FORCES In Table 3, we have compiled single-bond molecular recognition forces, as derived from AFM using functionalized tips. The interaction between avidin and streptavidin with biotin analogs is one of the most well-known specic recognition phenomena, and although there are differences between single-bond interaction forces among different biotin analogs (Table 3), the order of magnitude is in the subnanometer region for all analogs studied (14, 23), but well above the generally accepted lower limit for reliable single-bond force measurements of 0.005 nN. Protein D is found in the alveolar uid, where it takes part in the immune defense of the lungs against invading patho-

VOL. 74, 2008

bacterium and the AFM tip of m2, the distribution of 10 to 50 ligand-receptor pairs over this contact area would yield the conclusion that to sites would be present over an entire bacterial cell surface. Assuming a projected area of 100 nm2 per binding site, as valid for a molecule like immunoglobulin G (IgG) (2), this implies that a bacterial cell surface is covered fully by specic binding sites, and in fact requires that the specic binding sites are arranged along structural surface features in order to allow a t. Very often this is indeed the case (6). The above conclusion of full coverage depends strongly on the size of the specic binding site assumed, and a projected area of the binding site 2 to 3 times smaller would yield the conclusion of partial surface coverage of the bacterial cell surface by the specic binding sites. This may, in certain cases, be more realistic, especially because full coverage of a bacterial cell surface by a class of specic recognition molecules should in principle be reected strongly in the overall properties of the bacterial cell surface. As in general (see also Table 1), the absence or presence of specic recognition molecules is hardly expressed in overall physico-chemical cell surface properties as hydrophobicity and charge, we consider it more likely that the estimated to sites per bacterial cell surface only yield partial surface coverage: i.e., the projected surface area of a ligand-receptor pair should be considerably smaller than 100 nm2. To conclude, this review provides a further elaboration of our understanding of bacterial adhesion mechanisms and points to the need to evaluate adhesion mechanisms on a microscopic or even nanoscopic level, in addition to evaluations based on macroscopic characteristics such as surface hydrophobicity and charge.

ACKNOWLEDGMENT This review is dedicated to Paul G. Rouxhet, on the occasion of his retirement, 28 September 2007.
REFERENCES 1. Abu-Lail, N. I., and T. A. Camesano. 2006. Specic and nonspecic interaction forces between Escherichia coli and silicon nitride, determined by Poisson statistical analysis. Langmuir 22:72967301. 2. Bagchi, P., and S. M. Birnbaum. 1981. Effect of pH on the adsorption of immunoglobulin G on anionic poly(vinyl) model latex particles. J. Colloid Interface Sci. 83:460478. 3. Baygents, J. C., J. R. Glynn, O. Albinger, B. K. Biesemeyer, K. L. Ogden, and R. G. Arnold. 1998. Variation of surface charge density in monoclonal bacterial populations: implications for transport through porous media. Environ. Sci. Technol. 32:15961603. 4. Berquand, A., N. Xia, D. G. Castner, B. H. Clare, N. L. Abbott, V. Dupres, Y. Adriaensen, and Y. F. Dufrene. 2005. Antigen binding forces of single antilysozyme Fv fragments explored by atomic force microscopy. Langmuir 21:55175523. 5. Bos, R., H. C. van der Mei, and H. J. Busscher. 1999. Physico-chemistry of initial microbial adhesive interactionsits mechanisms and methods for study. FEMS Microbiol. Rev. 23:179229. 6. Busscher, H. J., B. van de Belt-Gritter, R. J. B. Dijkstra, W. Norde, F. C. Petersen, A. A. Scheie, and H. C. van der Mei. 2007. Intermolecular forces and enthalpies in the adhesion of Streptococcus mutans and an antigen I/II-decient mutant to laminin lms. J. Bacteriol. 189:29882995. 7. Busscher, H. J., M. M. Cowan, and H. C. van der Mei. 1992. On the relative importance of specic and non-specic approaches to oral microbial adhesion. FEMS Microbiol. Rev. 8:199209. 8. Busscher, H. J., and A. H. Weerkamp. 1987. Specic and non-specic interactions in bacterial adhesion to solid substrata. FEMS Microbiol. Rev. 46: 165173. 9. Butt, H. J., R. Berger, E. Bonaccurso, Y. Chen, and J. Wang. 2007. Impact of atomic force microscopy on interface and colloid science. Adv. Colloid Interface Sci. 133:91104.
10. Camesano, T. A., and N. I. Abu-Lail. 2002. Heterogeneity in bacterial surface polysaccharides, probed on a single-molecule basis. Biomacromolecules 3:661667. 11. Dubnau, D. 1991. Genetic competence in Bacillus subtilis. Microbiol. Rev. 55:395424. 12. Dufrene, Y. F. 2003. Recent progress in the application of atomic force microscopy imaging and force spectroscopy to microbiology. Curr. Opin. Microbiol. 6:317323. 13. Dupres, V., C. Verbelen, and Y. F. Dufrene. 2007. Probing molecular recog nition sites on biosurfaces using AFM. Biomaterials 28:23932402. 14. Florin, E. L., V. T. Moy, and H. E. Gaub. 1994. Adhesion forces between individual ligand-receptor pairs. Science 264:415417. 15. Gibbons, R. J., I. Etherden, and E. C. Moreno. 1985. Contribution of stereochemical interactions in the adhesion of Streptococcus sanguis C5 to experimental pellicles. J. Dent. Res. 64:96101. 16. Hermansson, M. 1999. The DLVO theory in microbial adhesion. Colloids Surf. B Biointerfaces 14:105119. 17. Hinterdorfer, P., and Y. F. Dufrene. 2006. Detection and localization of single molecular recognition events using atomic force microscopy. Nat. Methods 3:347355. 18. Kasas, S., and A. Ikai. 1995. A method for anchoring round shaped cells for atomic force microscope imaging. Biophys. J. 68:16781680. 19. Kell, D. B., H. M. Ryder, A. S. Kaprelyants, and H. V. Westerhoff. 1991. Quantifying heterogeneity: ow cytometry of bacterial cultures. Antonie van Leeuwenhoek 60:145158. 20. Lehenkari, P. P., and M. A. Horton. 1999. Single integrin molecule adhesion forces in intact cells measured by atomic force microscopy. Biochem. Biophys. Res. Commun. 259:645650. 21. Mendez-Vilas, A., A. M. Gallardo-Moreno, and M. L. Gonzalez-Mart n. 2006. Nano-mechanical exploration of the surface and sub-surface of hydrated cells of Staphylococcus epidermidids. Antonie van Leeuwenhoek 89: 373386. 22. Mendez-Vilas, A., A. M. Gallardo-Moreno, R. Calzado-Montero, and M. L. Gonzalez-Martin. 2007. AFM probing in aqueous environment of Staphylococcus epidermidis cells naturally immobilized on glass: physico-chemistry behind successful immobilization. Colloids Surf. B Biointerfaces. doi: 10.1016/j.colsurfb.2007.11.011. 23. Moy, V. T., E. L. Florin, and H. E. Gaub. 1994. Interaction forces and energies between ligands and receptors. Science 266:257259. 24. Mozes, N., F. Marchal, M. P. Hermesse, J. L. van Haecht, L. Reulieux, A. J. Leonard, and P. G. Rouxhet. 1987. Immobilization of microorganisms by a support, interplay of electrostatic and non-electrostatic interactions. Biotechnol. Bioeng. 30:439450. 25. Petersen, F. C., S. Assev, H. C. van der Mei, H. J. Busscher, and A. A. Scheie. 2002. Functional variation at the antigen I/II surface protein in Streptococcus mutans and Streptococcus intermedius. Infect. Immun. 70:249256. 26. Postollec, F., W. Norde, J. de Vries, H. J. Busscher, and H. C. van der Mei. 2006. Interactive forces between co-aggregating and non-co-aggregating oral bacterial pairs. J. Dent. Res. 85:231234. 27. Razatos, A., Y. L. Ong, M. M. Sharma, and G. Georgiou. 1998. Molecular determinants of bacterial adhesion monitored by atomic force microscopy. Proc. Natl. Acad. Sci. USA 95:1105911064. 28. Salerno, M. B., X. Li, and B. E. Logan. 2007. Adhesion characteristics of two Burkholderia cepacia strains examined using colloid probe microscopy and gradient force analysis. Colloids Surf. B Biointerfaces 59:4651. 29. Streger, S. H., S. Vainberg, H. Dong, and P. B. Hatzinger. 2002. Enhancing transport of Hydrogenophaga ava ENV735 for bioaugmentation of aquifers contaminated with methyl tert-butyl ether. Appl. Environ. Microbiol. 68: 55715579. 30. Sun, Z., L. A. Martinez-Lemus, A. Trache, J. P. Trzeciakowski, G. E. Davis, U. Pohl, and G. A. Meininger. 2005. Mechanical properties of the interaction between bronectin and 51-integrin on vascular smooth muscle cells studied using atomic force microscopy. Am. J. Physiol. Heart Circ. Physiol. 289:H2526H2535. 31. Tang, J., D. Krajcikova, R. Zhu, A. Ebner, S. Cutting, H. J. Gruber, I. Barak, and P. Hinterdorfer. 2007. Atomic force microscopy imaging and single molecule recognition force spectroscopy of coat proteins on the surface of Bacillus subtilis spore. J. Mol. Recognit. 20:483489. 32. Thormann, E., A. C. Simonsen, L. K. Nielsen, and O. G. Mouritsen. 2007. Ligand-receptor interactions and membrane structure investigated by AFM and time-resolved uorescence microscopy. J. Mol. Recognit. doi:10.1002/ jmr.850. 33. Touhami, A., B. Hoffmann, A. Vasella, F. A. Denis, and Y. F. Dufrene. 2003. Aggregation of yeast cells: direct measurement of discrete lectin-carbohydrate interactions. Microbiology 149:28732878. 34. Touhami, A., B. Hoffmann, A. Vasella, F. A. Denis, and Y. F. Dufrene. 2003. Probing specic lectin-carbohydrate interactions using atomic force microscopy imaging and force measurement. Langmuir 19:17451751. 35. Vadillo-Rodriguez, V., H. J. Busscher, W. Norde, J. de Vries, R. J. B. Dijkstra, I. Stokroos, and H. C. van der Mei. 2004. Comparison of atomic force microscopy interaction forces between bacteria and silicon nitride

39. Waar, K., H. C. van der Mei, H. J. M. Harmsen, J. de Vries, J. Atema-Smit, J. E. Degener, and H. J. Busscher. 2005. Atomic force microscopy study on specicity and non-specicity of interaction forces between Enterococcus faecalis cells with and without aggregation substance. Microbiology 151: 24592464. 40. Xu, C. P., B. van de Belt-Gritter, R. J. B. Dijkstra, W. Norde, H. C. van der Mei, and H. J. Busscher. 2007. Interaction forces between salivary proteins and Streptococcus mutans with and without antigen I/II. Langmuir 23:94239428. 41. Yongsunthon, R., V. G. J. Fowler, B. H. Lower, F. P. Vellano, E. Alexander, L. B. Reller, G. R. Corey, and S. K. Lower. 2007. Correlation between fundamental binding forces and clinical prognosis of Staphylococcus aureus infections of medical implants. Langmuir 23:22892292.
substrata for three commonly used immobilization methods. Appl. Environ. Microbiol. 70:54415446. 36. van der Mei, H. C., and H. J. Busscher. 2001. Electrophoretic mobility distributions of single-strain microbial populations. Appl. Environ. Microbiol. 67:491494. 37. Van Oss, C. J. 1995. Hydrophobicity of biosurfacesorigin, quantitativedetermination and interaction energies. Colloids Surf. B Biointerfaces 5:91 110. 38. Van Oss, C. J., R. J. Good, and M. K. Chaudhury. 1986. The role of Van der Waals forces and hydrogen bonds in hydrophobic interactions between biopolymers and low energy surfaces. J. Colloid Interface Sci. 111:378390.

Antagonistic Interplay between Antimitotic and G1-S Arresting Agents Observed in Experimental Combination Therapy
Korey R. Johnson, Kristy K. Young and Weimin Fan Clin Cancer Res 1999;5:2559-2565. Published online September 1, 1999.

Updated Version

Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/5/9/2559
Cited Articles Citing Articles
This article cites 37 articles, 16 of which you can access for free at: http://clincancerres.aacrjournals.org/content/5/9/2559.full.html#ref-list-1 This article has been cited by 6 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/5/9/2559.full.html#related-urls
E-mail alerts Reprints and Subscriptions Permissions
Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at pubs@aacr.org. To request permission to re-use all or part of this article, contact the AACR Publications Department at permissions@aacr.org.
Downloaded from clincancerres.aacrjournals.org on June 9, 2011 Copyright 1999 American Association for Cancer Research
Vol. 5, 2559 2565, September 1999
Clinical Cancer Research 2559
Antagonistic Interplay between Antimitotic and G1-S Arresting Agents Observed in Experimental Combination Therapy1
Korey R. Johnson, Kristy K. Young, and Weimin Fan2
Department of Pathology and Laboratory Medicine, Medical University of South Carolina, Charleston, South Carolina 29425

ABSTRACT

Paclitaxel is a naturally occurring antimitotic agent that has been shown to stabilize microtubules, induce mitotic arrest, and ultimately induce apoptotic cell death. The favorable clinical activity of paclitaxel has prompted considerable interest in combining paclitaxel with numerous other antineoplastic agents. Our previous studies have suggested 5-fluorouracil (5-FU), an antineoplastic agent that usually arrests tumor cells at the G1-S phase of the cell cycle, in combination with paclitaxel significantly represses paclitaxel-induced mitotic arrest and apoptosis. In the present study, we have extended this investigation to include several other antimitotic agents (vinblastine, colchicine, and nocodazole) in various combination schedules with the G1-S arresting agents 5-FU and hydroxyurea (HU). We found 5-FU, as well as HU, could significantly interfere with the overall cytotoxicity as compared with treatment with antimitotic agents alone. It appeared that 5-FU or HU severely limited the antimitotic agents cytotoxic effects on both mitotic arrest and apoptosis. No combination of a G1-S arresting agent with an antimitotic agent in any schedule produced an antitumor effect greater than that of the antimitotic agent alone. In addition, biochemical examination revealed that 5-FU and HU blocked the antimitotic agent-induced increase of p21WAF1/CIP1 protein levels, as well as prevented the hyperphosphorylation of the bcl-2 and c-raf-1 proteins. These findings suggest that careful considerations may be necessary when combining antineoplastic agents that exert their cytotoxic action at different phases of the cell cycle.

INTRODUCTION

Clinical protocols frequently combine chemotherapeutic agents that exhibit their cytotoxic action at different phases of the cell cycle (1 4). An optimal combination chemotherapy
Received 4/12/99; revised 6/21/99; accepted 6/21/99. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by NIH Grant CA71851 (to W. F.). 2 To whom requests for reprints should be addressed, at Department of Pathology and Laboratory Medicine, Medical University of South Carolina, 171 Ashley Avenue, Charleston, SC 29425; Phone: (843) 7925108; Fax: (843) 792-7762.

protocol results in increased therapeutic efficacy, decreased host toxicity, and minimal or delayed drug resistance (1, 5, 6). However, when antineoplastic agents with similar or different modes of action are combined, the outcome may be synergistic, additive, or antagonistic. Synergism implies that two drugs may produce greater therapeutic efficacy than an expected additive effect, whereas antagonism implies that the actual therapeutic activity produced by two drugs may be smaller than their expected additive effect (6, 7). Chemotherapeutic regimes frequently use the novel antimitotic agent paclitaxel. Paclitaxel possesses a unique mechanism of action in that it can bind to microtubulin, stabilize the tubulin polymer, and aid in further polymerization by shifting the dynamic equilibrium toward microtubule assembly (8 12). The microtubules bound by paclitaxel are unusually stable and thereby abrogate the dynamic reorganization process of the microtubule network required to form a functional spindle apparatus for the completion of mitosis and cell proliferation, ultimately leading to programmed cell death (8, 9, 1117). Paclitaxel has proven especially important clinically because of its activity in the treatment of metastatic breast carcinoma and drug refractory ovarian carcinoma (18 20). The promising clinical activity of paclitaxel has prompted interest in combining this agent with several other very effective antineoplastic agents. One such agent with which paclitaxel is presently being combined in clinical trials is the antimetabolite 5-FU.3 Specifically, the trials are combining these two agents in various schedules for the treatment of metastatic breast carcinoma, head and neck carcinoma, and gastrointestinal tract carcinomas (2125). Recently, our laboratory reported that the pretreatment or simultaneous addition of 5-FU (3) with paclitaxel in solid tumor cell lines significantly repressed the overall cytotoxicity as compared with paclitaxel treatment alone (26). It appeared 5-FU interfered with the cell killing activity of paclitaxel by preventing the tumor cells from entering the G2-M phase of the cell cycle. This finding raises concern as to the effectiveness of combining antimitotic agents with G1-S arresting agents. To determine whether this phenomenon applies to the combination of 5-FU with other antimitotic agents, we extended our investigation to include vinblastine, colchicine, and nocodazole. In addition, we combined all four of these antimitotic agents in various regimes with HU, another potent G1-S phase arresting agent that inhibits ribonucleotide reductase, resulting in depletion of dUMP necessary for DNA synthesis. Our results indicate that the combination of 5-FU or HU with any one of these antimitotic agents produces less antitumor activity than that seen with the treatment of an antimitotic agent alone. These findings suggest that 5-FU and HU may interfere with the

The abbreviations used are: 5-FU, 5-fluorouracil; HU, hydroxyurea; MTT, 3-(4,5-dimethlthiazol-2-yl)-2,5-diphenyltetrazolium bromide.
2560 Antagonistic Interplay of Antimitotic and G1-S Arresting Agents
cytotoxicity of the antimitotic agents simply by preventing the majority of cells from entering the G2-M phase of the cell cycle.

MATERIALS AND METHODS

Drugs and Cell Culture. 5-FU, HU, paclitaxel, nocodazole, vinblastine, and colchicine were purchased from Sigma Chemical Co. (St. Louis, MO) and dissolved in 100% DMSO to make stock concentrations of 10 mM, 2 M, 1 mM, 1 mg/ml, and 1 mM, respectively. Drugs were then diluted in culture media to obtain the desired concentrations. In all experiments, the various drugs were used at the following concentrations: 10 M 5-FU, 2 mM HU, 100 nM paclitaxel, 0.2 g/ml nocodazole, 100 nM vinblastine, and 100 nM colchicine. The human breast cancer Bcap37 (27) and the human epidermoid carcinoma KB cell lines (American Type Culture Collection, Rockville, MD) were propagated in RPMI 1640 supplemented with 10% fetal bovine serum (Sigma) and 1% penicillin/streptomycin. As described earlier (27, 28), the drugs were added when cells reached 60 70% confluency. Detection of Internucleosomal DNA Fragmentation. After incubation with the various drug regimes, approximately cells were harvested and suspended in lysis buffer containing 20 mM Tris, 5 mM EDTA, 0.5% Triton X-100, and 0.5 mg/ml proteinase K for 1 h on ice. The remaining steps for DNA fragmentation were performed as described previously (29). DNA samples were then analyzed by electrophoresis on a 1.2% agarose slab gel containing 0.2 g/ml ethidium bromide. Flow Cytometric Analysis. Cell sample preparation and propidium iodide staining were performed according to the methods described by Nicoletti et al. (30). At approximately 60 70% confluency, cells were treated with the various drug regimes for 24 and 48 h. Cells were collected and processed as described previously (26). Cell cycle distribution was determined using a Coulter Epics V instrument (Coulter Corp.), with an argon laser set to excite at 488 nm. Cytospin Preparation. Cells were cultured to 60 70% confluency, at which time they were treated with the various drug regimes for 24 and 48 h. Cells were harvested by trypsinization and washed twice with PBS. Cell numbers were determined using a hemacytometer, and 50,000 100,000 cells from each group were used for cytospin preparations. Slides were air dried and fixed in 100% acetone for 5 min prior to Giemsa-Wright staining and then were examined using brightfield microscopy (29). MTT Assay. Bcap37 and KB cells were harvested with trypsin and resuspended to a final concentration of cells/ml in fresh medium containing 10% FBS and 1% penicillin/streptomycin. Aliquots of 100 l from cell suspension were evenly distributed into 96-well tissue culture plates with lids (Falcon, Oxnard, CA). Designated columns were treated for 24, 48, and 72 h. One column from each plate contained medium alone, and another column contained cells in drug-free medium. Cell viability was assessed in accordance to the protocol described by Carmichael et al. (31). The absorbancies of individual wells were determined at 560 nm by a microplate reader (Molecular Devices,CA). Western Blots. Cells were treated with the various drug regimes for 48 h and promptly harvested by trypsinization.

Protein extraction and immunoblot procedures were performed as described previously (32). Briefly, the membranes were blocked with BLOTTO for 2 h at room temperature and subsequently probed with 0.5 g/ml of p53 murine monoclonal antibody (DO-1; Santa Cruz Biotechnology), 0.5 g/ml of p21 murine monoclonal antibody (WAF-1 AB-1; Calbiochem), 1.0 g/ml of bcl-2 murine monoclonal antibody (clone # 100; Santa Cruz), or with 0.5 g/ml of c-raf-1 murine monoclonal antibody (Transduction Laboratories), diluted in 3% BSA-PBST (3% BSA-PBS-0.05% Tween 20). After three 15-min incubations with BLOTTO, 0.1 g/ml of affinity-purified goat anti-mouse IgG conjugated to horseradish peroxidase (Jackson ImmunoResearch) was incubated with the membranes. Immunoreactive bands were then visualized using a chemiluminescence substrate for horseradish peroxidase (Amersham) and exposure to Kodak X-OMAT AR film.

RESULTS

5-FU and HU Interfere with the Capacity of Antimitotic Agents to Induce Apoptotic Cell Death. The hallmark feature indicative of the late stages of programmed cell death is the internucleosomal fragmentation of genomic DNA, which generates a characteristic ladder pattern of 200-bp intervals when analyzed using agarose gel electrophoresis. Therefore, to investigate the influence of 5-FU and HU on the capacity of the antimitotic agents to induce apoptotic cell death, we simultaneously applied 5-FU or HU with antimitotic agents and compared this to the treatment of the singular agents in the Bcap37 and KB cell lines. Fig. 1 demonstrates the ability of the antimitotic agents paclitaxel, nocodazole, vinblastine, and colchicine to induce DNA fragmentation in the Bcap37 cell line after 72 h of treatment, indicated by the characteristic ladder patterns. Treatment with either 5-FU or HU alone for 72 h was unable to induce DNA fragmentation (Fig. 1, Lanes 5, 7, 9, and 11). However, the simultaneous treatment of 5-FU or HU in combination with paclitaxel, nocodazole, vinblastine, or colchicine for 72 h dramatically reduced DNA fragmentation (Fig. 1, Lanes 5, 7, 9, and 11, respectively). Similar results were seen with the KB cell line (data not shown). 5-FU and HU Hinder the Ability of Antimitotic Agents to Induce G2-M Arrest. Morphological examination of Bcap37 and KB cell line cultures treated with paclitaxel, nocodazole, vinblastine, and colchicine revealed a significant number of apparent mitotic figures as quickly as 12 h and persisting throughout the 72-h observation. However, the simultaneous addition of 5-FU in combination with the various antimitotic agents significantly decreased the number of apparently mitotically arrested cells (data not shown). Thus, to determine whether 5-FU did significantly interfere with the capacity of the antimitotic agents to arrest cells mitotically, we prepared cytospin slides of both cell lines treated with the antimitotic agents alone or in simultaneous combination with 5-FU. We counted those cells that appeared to contain condensed chromosomes, i.e., mitotic figures, and summarized these results in Table 1. Coincidentally, we found that 5-FU dramatically inhibited the ability of all four antimitotic agents to induce mitotic arrest, as seen in the phase contrast morphological observation. To clarify the exact cell cycle distribution of the Bcap37 cells treated with

Clinical Cancer Research 2561
Fig. 1 Effects of the ability of the antimitotic or G1-S arresting agents to induce internucleosomal DNA fragmentation in Bcap37 cells. Bcap37 cells were cultured in 100 mm tissue culture dishes to 60 70% confluency. Cells were then treated with either 10 M 5-FU (A) or 2 mM HU (B) and were then treated with 100 nM Taxol (TX), 0.2 g/ml nocodazole (NOC), 100 nM vinblastine (VINB), or 100 nM colchicine (COLCH), alone or simultaneously with 5-FU (A) or HU (B). Total DNA was collected after 72 h exposure to the various drug regimes and electrophoresed on a 1.2% agarose slab gel containing 0.2% ethidium bromide.
Table 1 Effect of 5-FU and antimitotic agents on mitotic arresta % of cells at G2M phaseb Bcap 37 Drug exposure Control 5-FU Paclitaxel 5-FU Paclitaxel Nocodazole 5-FU Nocodazole Vinblastine 5-FU Vinblastine Colchicine 5-FU Colchicine 24 h 48 h 24 h KB 48 h 14 8
a This table is based on three separate experiments and presented as mean SE. b Cytospin slides were stained with Giemsa. Cells (300) were counted from each slide, and only those cells with typical morphological features of condensed chromosomes were counted as mitotically arrested cells.
the various drug regimes, we performed flow cytometric analysis. Fig. 2 clearly illustrates the capacity of all four antimitotic agents to arrest the majority of cells in the G2-M phase of the cell cycle. However, pretreatment with 5-FU for 6 h, followed by the addition of antimitotic agents, or even simultaneous addition of 5-FU in combination with the antimitotic agents, dramatically reduced the number of cells in the G2-M phase of the cell cycle. Interestingly, the pretreatment with the antimitotic agents for 6 h, followed by the addition of 5-FU, partially blocked the number of cells in the G2-M phase of the cell cycle. Similar results were found when HU was substituted for 5-FU in each protocol (data not shown). The Cytostatic Effect of 5-FU and HU Interferes with the Cytotoxicity of Paclitaxel in a Schedule-dependent Manner. Previous studies in our laboratory demonstrated that the IC50 for 5-FU in the Bcap37 and the KB cell lines was 10 M. We tested 10 M 5-FU in combination with IC90s of a variety of antimitotic agents, such as paclitaxel, nocodazole, vinblastine,
and colchicine, in various schedules of administration. Each antimitotic agent was incubated simultaneously with 5-FU, and in separate experiments, 5-FU was added 6 h before or after each antimitotic agent. Each antimitotic agent, as well as 5-FU, was applied alone. The drug regimes were carried out for 24, 48, and 72 h. At the end of each time course treatment, the MTT assay was performed as described in Materials and Methods to measure cell viability. We found that the pretreatment or simultaneous exposure of the tumor cells with 5-FU could significantly interfere with the cytotoxic effects of all four antimitotic agents (Fig. 3). However, pretreatment of the tumor cells with antimitotic agents could clearly attenuate the inhibitory effect of 5-FU. It should be noted that no combination of 5-FU with an antimitotic agent produced a cell killing efficiency as great as the treatment with the antimitotic agent alone. Similar MTT studies were performed in which HU was combined with all four antimitotic agents in the previously described drug regimes, and data were concurrent with those discussed above (data not shown). Clonogenic survival assays were performed in conjunction with the MTT assays, and these results also correlate with the above data (data not shown). 5-FU and HU Prevent Antimitotic Agent-induced Hyperphosphorylation of bcl-2 and c-raf-1, in Addition to Blocking the Increase in p21WAF1/CIP1 Protein Levels. Schandl et al. (33) recently demonstrated the ability of the antimitotic agents paclitaxel, nocodazole, vinblastine, and colchicine to induce the hyperphosphorylation of bcl-2. It has been proposed that the hyperphosphorylation of bcl-2 inactivates this protein and coincides with apoptotic cell death (34, 35). Therefore, we investigated the effect of 5-FU or HU on the ability of antimitotic agents to induce hyperphosphorylation of bcl-2 in the Bcap37 and KB cell lines. Fig. 4 (Lanes 3, 5, 7, and 9) clearly demonstrates the capacity of all four antimitotic agents to induce bcl-2 hyperphosphorylation. However, the simultaneous addition of 5-FU or HU with the various antimitotic agents attenuated their ability to hyperphosphorylate bcl-2 (Fig. 4, Lanes 4, 6, 8, and 10). Blagosklonny et al. (36, 37) recently published a study implicating the active phosphorylated form of c-raf-1, repre-

2562 Antagonistic Interplay of Antimitotic and G1-S Arresting Agents
Fig. 2 Flow cytometric analysis depicting the distribution of cells in various phases of the cell cycle. Bcap37 cells were treated for 48 h with 10 M 5-FU, 100 nM paclitaxel, or 100 nM vinblastine (VINB), alone or in the combinations indicated. Where applicable, pretreatments were performed for 6 h. Samples were analyzed by flow cytometric analysis as described in Materials and Methods. The distribution of cells in G0-G1, S, and G2-M phases of the cell cycle and apoptotic cells (Ap) are indicated above each corresponding peak. TX, Taxol; NOC, nocodazole; COLCH, colchicine.
sented by decreased electrophoretic mobility on SDS-PAGE, following paclitaxel treatment, to be associated with the hyperphosphorylation of bcl-2 and subsequent induction of apoptotic cell death. Our studies correlate with these findings in that paclitaxel (in addition to nocodazole, vinblastine, and colchicine) induced a slight decrease in electrophoretic mobility of c-raf-1 [previously demonstrated by Blagosklonny et al. (39) to be an active phosphorylated form of c-raf-1] in cohorts with bcl-2 hyperphosphorylation (Fig. 4, Lanes 3, 5, 7, and 9). As was seen with bcl-2, simultaneous addition of 5-FU or HU with the antimitotic agents blocked the hyperphosphorylation of craf-1 (Fig. 4, Lanes 4, 6, 8, and 10). In addition, several experiments have indicated that the antimitotic agents paclitaxel, vinblastine, and nocodazole could induce the expression of the cyclin-dependent kinase inhibitor p21WAF1/CIP1 (38, 39). We sought to investigate the possible changes in p21WAF1/CIP1 protein levels in the Bcap37 and KB cell lines, following the treatments with the various drug regimes. Fig. 4 (Lanes 3, 5, 7, and 9) clearly demonstrates the increase in protein levels of p21WAF1/CIP1 after treatment with all four antimitotic agents. Simultaneous treatment with either 5-FU or HU in combination with any one of the antimitotic agents appeared to abrogate this increase in p21WAF1/CIP1 protein levels.

DISCUSSION

In this study, we examined the effects of combining chemotherapeutic agents that exhibit their cytotoxic action at different phases of the cell cycle. In particular, we looked into the combination of an antimitotic agent (paclitaxel, nocodazole,
vinblastine, or colchicine) with an antimetabolite (5-FU or HU) to determine possible increases in overall cytotoxic effects, as compared with single-agent treatment alone. These combinations are particularly interesting in that paclitaxel is presently being combined with 5-FU and HU in clinical trials (2123). Antimitotic agents largely target the microtubule network in individual cells. These agents disrupt the dynamic reorganization of this network, resulting in aberrant mitotic aster formation. Thus, affected cells are unable to transverse successfully from metaphase to anaphase. Ultimately, the prolonged mitotic arrest in most of these cells leads to programmed cell death, or apoptosis. For example, paclitaxel has proven to be especially effective in its cell killing proficiency, apparently through this disruption of the microtubule network (8, 40). Morphological examination revealed that the majority of solid tumor cells treated with clinically relevant concentrations of paclitaxel sustained prolonged mitotic arrest, followed by eventual induction of apoptosis (27, 29, 41 42). In fact, recent evidence from our laboratory indicated that a sustained mitotic arrest is required for paclitaxel-induced apoptosis. When cells were arrested in the G1-S phase through treatment with 5-FU followed by paclitaxel, we noted a marked decrease in programmed cell death as compared with cells treated with paclitaxel alone. Subsequently, we found no apparent beneficial cytotoxic effects through any in vitro schedule combination of 5-FU and paclitaxel (26). The present study demonstrates that the combination of 5-FU with other antimitotic agents, such as vinblastine, colchicine, and nocodazole, once again produces no beneficial cytotoxic effect. Similar experiments were performed on the Bcap37 and KB cell lines, in which all four antimitotic agents were combined in a
Clinical Cancer Research 2563
Fig. 3 Cytotoxic effects exhibited by antimitotic and G1 arresting agents in the Bcap37 and KB cell lines. Approximately Bcap37 (A) and KB (BD) cells were cultured in 96-well microculture plates. Cells were incubated with 10 M 5-FU, 100 nM paclitaxel (A and B), 100 nM vinblastine (C), or 100 nM colchicine (D). Also, incubations were performed in which 5-FU was added simultaneously with paclitaxel (A and B), vinblastine (C), or colchicine (D). Furthermore, cells were pretreated with 5-FU for 6 h before the addition of paclitaxel (A and B), vinblastine (C), or colchicine (D) for an additional 24, 48, or 72 h. Lastly, cells were pretreated with paclitaxel (A and B), vinblastine (C), or colchicine (D) for 6 h before the addition of 5-FU for an additional 24, 48, or 72 h. MTT assays were performed as described in Materials and Methods. Bars, SD.

variety of schedules with HU, another potent G1-S-phase arresting agent. As was the case with 5-FU in combination with paclitaxel, HU inhibits the capacity of all four antimitotic agents to induce apoptotic cell death (Fig. 1B). Flow cytometric analysis and cytospin preparations demonstrate that both 5-FU and HU could arrest the majority of cells in the G1-S phase of the cell cycle, thereby preventing entry into the G2-M phase, unless the antimitotic agent was administered prior to the G1-S arresting agent. These findings correlate with the ability of these G1-S arresting agents to interfere with the capacity of the antimitotic agents to induce apoptosis (Fig. 1). Hence, it appears that the mechanism by which 5-FU and HU perturbate the cytotoxic effects of the antimitotic agents is merely by preventing the majority of the tumor cells from accumulating in the G2-M phase of the cell cycle. These data correlate with several previous studies that suggest that the greatest cytotoxic effects of these antimitotic agents is seen when the majority of cells are arrested in mitosis (14, 15, 43). An important consideration when combining chemotherapeutic agents that possess different mechanisms of action is the schedule in which the drugs are administered. An optimal combination may produce a synergistic effect in which the therapeutic efficacy is greater than the expected additive effect (6,
44). For example, a group of investigators recently discovered that the administration of paclitaxel prior to cisplatin produced a synergistic or additive effect, whereas the reverse sequence resulted in antagonism (45). Furthermore, Kano et al. (46) reported that the pretreatment or simultaneous addition of 5-FU with paclitaxel failed to produce a synergistic effect. However, Kano et al. (46) discovered an additive effect when paclitaxel was added prior to 5-fluorouracil. Our results from the cytotoxicity assay that examined the combination of the G1-S arresting agents 5-FU and HU with the antimitotic agents demonstrated the sequential importance of in vitro combination drug regimes in the Bcap 37 and KB cell lines. These data suggest that subadditive effects are generated when 5-FU or HU is added prior to or in simultaneous combination with an antimitotic agent. In fact, only when 5-FU or HU was added 6 h after antimitotic agent treatment did we see a cytotoxic effect nearly equivalent to the treatment of the antimitotic agents alone. Thus, the data from this study suggest that the treatment of antimitotic agents alone produced the greatest overall cytotoxic effect in the Bcap37 and KB cell lines. In light of the fact that the phosphorylation of the bcl-2 oncoprotein has recently been implicated in G2-M arrest (33) and the induction of apoptosis (34, 35), we performed Western

2564 Antagonistic Interplay of Antimitotic and G1-S Arresting Agents
In summary, this study has investigated the possible influence of two G1-S arresting agents, 5-FU and HU, on the cytotoxic effects of various antimitotic agents on human solid tumor cells in vitro. Our results demonstrate that both 5-FU and HU could interfere with the cytotoxic effects of antimitotic agents on mitotic arrest and apoptosis. Meanwhile, we found that the G1-S arresting agents specifically perturbated the capacity of the antimitotic agents to induce bcl-2 phosphorylation, c-raf-1 activation, and increase in p21WAF1/CIP1 protein levels. These data suggest that the G1-S arresting agents may prevent the majority of cells from progressing to the G2-M phase of the cell cycle, where antimitotic agents have been shown to exert their greatest cytotoxic effect (40). In light of these findings, careful consideration or experimental evaluation is necessary when combining antineoplastic agents that exert their cytotoxic action at different phases of the cell cycle.

REFERENCES

1. Devita, V. T. Principles of chemotherapy. In: V. T. Devita, S. Hellman, and S. A. Rosenberg (eds.), Cancer: Principles and Practice of Oncology. Ed. 4, pp. 276 292. Philadelphia: J. B. Lippincott Co., 1993. 2. Sartorelli, A. C. Some approaches to the therapeutic exploitation of metabolic sites of vulnerable neoplastic cells. Cancer Res., 29: 1019 1032, 1969. 3. Frei, E., III. Clinical studies of combination chemotherapy for cancer. In: T-C. Chou and D. C. Rideout (eds.), Synergism and Antagonism in Chemotherapy, Vol. 1, pp. 103108. San Diego: Academic Press, 1991. 4. Fan, W., Johnson, K. R., and Miller, M. C., III. In vitro evaluation of combination chemotherapy against human tumor cells (Review). Oncol. Rep., 5: 10351042, 1998. 5. Yarbro, J. W. The scientific basis of cancer chemotherapy. In: M. C. Perry (ed.), The Chemotherapy Source Book, Ed. 2, pp. 118. Baltimore: Williams and Wilkins, 1996. 6. Rideout, D. C., and Chou, T-C. Synergism, antagonism, and potentiation in chemotherapy: an overview. In: T-C. Chou and D. C. Rideout (eds.), Synergism and Antagonism in Chemotherapy, Vol. 1, pp. 3 60. San Diego: Academic Press, 1991. 7. Bertino, J. R., and Chou, T-C. Chemotherapy: synergism and antagonism. In: J. R. Bertino (ed.), Encyclopedia of Cancer, Vol. 1, 368 379. San Diego: Academic Press, 1997. 8. Schiff, P. B., Fant, J., and Horwitz, S. B. Promotion of microtubule assembly in vitro by taxol. Nature (Lond.), 22: 665 667, 1979. 9. Schiff, P. B., and Horwitz, S. B. Taxol stabilizes microtubules in mouse fibroblast cells. Proc. Natl. Acad. Sci. USA, 77: 15611565, 1980. 10. Schiff, P. B., and Horwitz, S. B. Taxol assembles tubulin in the absence of exogenous guanosine 5-triphosphate or microtubule-associated proteins. Biochemistry, 20: 32473252, 1981. 11. Kumar, N. Taxol induced polymerization of purified tubulin: mechanism of action. J. Biol. Chem., 256: 1043510441, 1981. 12. Rowinsky, E. K., Donehower, R. C., Jones, R. J., and Tucker, R. W. Microtubule changes and cytotoxicity in leukemia cell lines treated with taxol. Cancer Res., 48: 4093 4100, 1988. 13. Dustin, P. Microtubules. Sci. Am., 243: 66 76, 1980. 14. Woods, C. M., Zhu, J., McQueney, P. A., Bollag, D., and Lazarides, E. Taxol-induced mitotic block triggers rapid onset of a p53-independent pathway. Mol. Med., 1: 506 526, 1995. 15. Horwitz, S. B. Mechanism of action of Taxol. Trends Pharmacol. Sci., 13: 134 136, 1992. 16. Kung, A. L., Zetterberg, A., Sherwood, A. W., and Schimke, R. T. Cytotoxic effects of cell cycle specific agents: a result of cell cycle perturbation. Cancer Res., 50: 73077314, 1990.

Fig. 4 Western blot analysis for the bcl-2, c-raf-1, and p21WAF1/CIP1 proteins. Total whole-cell protein extract was collected from Bcap37 cells treated for 48 h with either 10 M 5-FU (A) or 2 mM HU, 100 nM paclitaxel, 0.2 g/ml nocodazole (NOC), 100 nM vinblastine (VINB), or 100 nM colchicine (COLCH; B), alone or in combination with 5-FU or HU. Equal amounts (100 g/lane) of cellular protein were fractionated on a 12% SDS-polyacrylamide gel and transferred to nitrocellulose membranes. The membranes were immunoblotted with a monoclonal antibody to p21WAF1 (WAF1 Ab-1; Calbiochem), bcl-2 (clone #100; Santa Cruz), or c-raf-1 (Transduction Laboratories) as described in Materials and Methods.
blot analysis to determine the phosphorylation state of bcl-2 in each treatment regime. Fig. 4 clearly indicates the ability of the antimitotic agents to hyperphosphorylate bcl-2. However, the simultaneous addition of the G1-S arresting agents with the antimitotic agents blocked the hyperphosphorylation of bcl-2. Coincidentally, c-raf-1 activation through phosphorylation has been demonstrated after antimitotic agent treatment and has been linked to bcl-2 phosphorylation and subsequent induction of apoptosis (34). Our findings demonstrate the capacity of the G1-S arresting agents to interfere with the c-raf-1 phosphorylation induced by the antimitotic agents (Fig. 4). The treatments with the G1-S arresting agents also interfered with the ability of the antimitotic agents to increase the protein levels of the cyclin-dependent kinase inhibitor p21WAF1/CIP1 (Fig. 4). All of these findings correlate with the interference of the G1-S arresting agents with the capacity of the antimitotic agents to induce apoptotic cell death.
Clinical Cancer Research 2565
17. Donaldson, K. L., Goolsby, G., Kiener, P. A., and Whal, A. F. Activation of p34cdc coincident with Taxol-induced apoptosis. Cell Growth Differ., 5: 10411050, 1994. 18. Rowinsky, E. K., Cazenvae, L. A., and Donehower, R. C. Taxol: a novel investigational antimicrotubule agent. J. Natl. Inst. Cancer, 82: 12471259, 1990. 19. Wiernik, P. H., Schwartz, E. L., Strauman, J. J., Dutcher, J. P., Lipton, R. B., and Paietta, E. Phase I clinical and pharmacokinetics study of Taxol. Cancer Res., 47: 2486 2493, 1987. 20. Holmes, F. A., Walter, R. S., Theriault, R. L., Forman, A. D., Newton, L. K., Raber, M. N., Buzdar, A. U., Frye, D. K., and Hortobagyi, G. N. Phase II trial of Taxol, an active drug in the treatment of metastatic breast cancer. J. Natl. Cancer Inst., 83: 17971805, 1991. 21. Klassen, U., Harstrick, A., Wilke, H., and Seeber, S. Preclinical and clinical study results of the combination of paclitaxel and 5-fluorouracil/ folinic acid in the treatment of metastatic breast cancer. Semin. Oncol., 23 (1 Suppl. 1): 44 47, 1996. 22. Paul, D. M., Garrett, A. M., Meshad, M., DeVore, R. D., Porter, L. L., and Johnson, D. H. Paclitaxel and 5-fluorouracil in metastatic breast cancer: The US experience. Semin. Oncol., 23(1 Suppl. 1): 48 52, 1996. 23. Johnson, D. H., Paul, D., and Hande, K. R. Paclitaxel, 5-fluorouracil, and folinic acid in metastatic breast cancer: BRE-26, a Phase II trial. Semin. Oncol., 24 (1 Suppl. 3): s3-22s3-25, 1997. 24. Vokes, E. E., Stupp, R., Haraf, D., Moran, W., Malone, D., Wenig, B., Sweeney, P., and Weichselbaum, R. R. Hydroxyurea with continuous infusion paclitaxel, 5-fluorouracil, and concomitant radiotherapy for poor-prognosis head and neck cancer. Semin. Oncol., 22 (3 Suppl. 6): 4752, 1995. 25. Ajani, J. A., Ilson, D. H., and Kelsen, D. P. The activity of paclitaxel in gastrointestinal tumors. Semin. Oncol., 22 (5 Suppl. 12): 46 53, 1995. 26. Johnson, K. R., Wang, L., Miller, M. C., III, Willingham, M. C., and Fan, W. 5-Fluorouracil interferes with paclitaxel cytotoxicity against human solid tumor cells. Clin. Cancer Res., 3: 1739 1745, 1997. 27. Cheng, L., Zheng, S., Raghunathan, K., Priest, D. G., Willingham, M. C., Norris, J. S., and Fan, W. Characterization of Taxol-induced apoptosis and altered gene expression in breast cancer cells. Cell. Pharmacol., 2: 249 257, 1995. 28. Fan, W., Cheng, L., Norris, J. S., and Willingham, M. C. Glucocorticoids selectively inhibit Taxol-induced apoptosis in a human breast cancer cell line. Cell. Pharmacol., 3: 435 440, 1996. 29. Fan, W., Everett, E. T., Tang, C., Cooper, T., Fang, Q., Bhalla, K., and Norris, J. S. Glucocorticoid-mediated inhibition of Taxol-induced apoptosis leiomyosarcoma cells. Cell. Pharmacol., 1: 205212, 1994. 30. Nicoletti, I., Migliorati, G., Pagliacci, M. C., Grignani, F., and Riccardi, C. A rapid and simple method for measuring thymocyte apoptosis by propidium iodide staining and flow cytometry. J. Immunol. Methods, 139: 271279, 1991. 31. Carmichael, J., Degraff, W. G., Gadzar, A. F., Minna, J. D., and Mitchell, J. Evaluation of a tetrazolium-based semiautomated colorimet-

ric assay: assessment of chemosensitivity testing. Cancer Res., 47: 935942, 1987. 32. Fan, W., Schandl, C. A., Cheng, L., Norris, J. S., and Willingham, M. C. Glucocorticoids modulate paclitaxel cytotoxicity in human solid tumor cells. Cell. Pharmacol., 3: 343348, 1996. 33. Schandl, C. A., Priest, D. G., and Willingham, M. C. Phosphorylation of bcl-2 is a cell cycle-regulated event. Cell. Pharmacol., 3: 367372, 1996. 34. Haldar, S., Jena, N., and Croce, C. M. Inactivation of bcl-2 by phosphorylation. Proc. Natl. Acad. Sci. USA, 92: 4507 4511, 1995. 35. Haldar, S., Chintapalli, J., and Croce, C. M. Taxol-induced bcl-2 phosphorylation and death of prostate cancer cells. Cancer Res., 56: 12531255, 1996. 36. Blagosklonny, M., Schulte, T., Nguyen, P., Trepel, J., and Neckers, L. M. Taxol-induced apoptosis and phosphorylation of bcl-2 protein involves c-raf-1 and represents a novel c-raf-1 signal transduction pathway. Cancer Res., 56: 18511854, 1996. 37. Blagosklonny, M. V., Giannakakou, P., El-Deiry, W. S., Kingston, D. G. I., Higgs, P. I., Neckers, L. M., and Fojo, T. Raf-1/bcl-2 phosphorylation: a step from microtubule damage to cell death. Cancer Res., 57: 130 135, 1997. 38. Tishler, R. B., Lamppu, D. M., Park, S., and Prince, B. D. Microtubule-active drugs Taxol, vinblastine, and nocodazole increase the levels of transcriptionally active p53. Cancer Res., 55: 6021 6025, 1995. 39. Blagosklonny, M. V., Schulte, T. W., Nguyen, P., Mimnaugh, E. G., Trepel, J., and Neckers, L. M. Taxol induction of p21WAF1 and p53 requires c-raf-1. Cancer Res., 55: 4623 4626, 1995. 40. Manfredi, J. J., and Horwitz, S. B. Paclitaxel: an anitimitotic agent with a new mechanism of action. Pharmacol. Ther., 25: 83125, 1984. 41. Liu, Y., Bhalla, K., Hill, C., and Priest, D. G. Evidence for involvement of tyrosine phosphorylation in taxol-induced apoptosis in human ovarian tumor cell line. Biochem. Pharmacol., 48: 12651272, 1994. 42. Willingham, M. C., and Bhalla, K. Trasnsient mitotic phase localization of bcl-2 oncoprotein in human carcinoma cells and its possible role in prevention of apoptosis. J. Histochem. Cytochem., 42: 441 450, 1994. 43. Milas, L., Hunter, N. R., Kurdoglu, B., Mason, K. A., Meyn, R. E., Stephens, L. C., and Peters, L. J. Kinetics of mitotic arrest and apoptosis in murine mammary and ovarian tumors treated with taxol. Cancer Chemother. Pharmacol., 35: 297303, 1995. 44. Kano, Y., Akutsu, M., Tsunoda, S., Suzuki, K., and Yazawa, Y. In vitro schedule dependent interaction between paclitaxel and cisplatin in human carcinoma cell lines. Cell. Chemother. Pharmacol., 37: 525530, 1996. 45. Kano, Y., Akutsu, M., Tsunoda, S., Ando, J., Matsui, J., Suzuki, K., Ikeda, T., Inoue, Y., and Adachi, K. Schedule-dependent interaction between paclitaxel and 5-fluorouracil in human carcinoma cell lines in vitro. Br. J. Cancer, 74: 704 710, 1996.