SCANNING VOL. 00, 1–7 (2012)C© Wiley Periodicals, Inc.

Imaging and Measuring the Molecular Force of LymphomaPathological Cells Using Atomic Force Microscopy

MI LI1,2, XIUBIN XIAO3, LIANQING LIU1, NING XI1,4, YUECHAO WANG1, ZAILI DONG1, AND WEIJING ZHANG3

1State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang,China2Graduate University of Chinese Academy of Sciences, Beijing, China3Department of Lymphoma, Affiliated Hospital of Military Medical Academy of Sciences, Beijing, China4Department of Electrical and Computer Engineering, Michigan State University, East Lansing, Michigan

Summary: Atomic force microscopy (AFM) pro-vides a new technology to visualize the cellular to-pography and quantify the molecular interactions atnanometer spatial resolution. In this work, AFM wasused to image the cellular topography and measurethe molecular force of pathological cells from B-celllymphoma patients. After the fluorescence staining,cancer cells were recognized by their special mor-phological features and then the detailed topogra-phy was visualized by AFM imaging. The AFM im-ages showed that cancer cells were much rougherthan healthy cells. CD20 is a surface marker of Bcells and rituximab is a monoclonal antibody againstCD20. To measure the CD20-rituximab interactionforces, the polyethylene glycol (PEG) linker was usedto link rituximab onto the AFM tip and the verifi-cation experiments of the functionalized probe in-dicated that rituximab molecules were successfullylinked onto the AFM tip. The CD20-rituximab in-teraction forces were measured on about 20 patho-logical cells and the force measurement results indi-cated the CD20-rituximab binding forces were mainlyin the range of 110–120 pN and 130–140 pN. Theseresults can improve our understanding of the topog-raphy and molecular force of lymphoma pathologicalcells. SCANNING 00: 1–7, 2012. C© 2012 Wiley Peri-odicals, Inc.

Contract grant sponsor: National Natural Science Foundationof China; Contract grant number: 60904095, 61175103; Contractgrant sponsor: CAS FEA International Partnership Program forCreative Research Teams.

Address for reprints: Lianqing Liu, State Key Laboratory ofRobotics, Shenyang Institute of Automation Chinese Academy of Sci-ences, Shenyang 110016, ChinaE-mail: lqliu@sia.cn, xin@egr.msu.edu, zhangwj3072@163.com

Received 9 April 2012; Accepted with revision 4 May 2012

DOI 10.1002/sca.21033Published online in Wiley Online Library (wileyonlinelibrary.com)

Key words: CD20, rituximab, atomic forcemicroscopy, single-molecule force spectroscopy

Introduction

With the advancement of research methods in bio-chemistry, the intracellular chemical processes havebeen gradually clarified. However, the traditionalbulk biochemical experiments in test tubes may notreflect the complete processes in vivo owing to theirinherent ensemble averaging (Cecconi et al., 2005).The test tube measurements provide averaged infor-mation obtained on large ensembles of moleculesfrom many cells (Dufrene, 2009). Undoubtedly, theresults deduced from the ensemble measurements arecorrect for depicting the behavior of the ensemblemolecules, but on the other hand, the ensemble mea-surements hide the rare events and the variable prop-erties of individual molecules (Dupres et al., 2009).As opposed to test tube experiments, single-moleculetechniques allow one to look beyond ensemble av-erages (Zhuang et al., 2000) and thereby can revealthe events and properties that would otherwise beinaccessible, providing new insights for understand-ing the cellular physiological activities at individualcell/molecule levels. Consequently, utilizing single-molecule techniques to investigate the behavior of in-dividual cells/molecules has become a new researchhotspot in life sciences (Kodera et al., 2010; Katanand Dekker, 2011).

One of the most common single-molecule tech-niques is atomic force microscopy (AFM) (Binniget al., ’86), which is widely used in life sciences sinceits invention. The main advantage of AFM is thatit can directly image the topography of biologicalsamples at nanometer resolution under physiologicalconditions, without the need of labeling, staining, orfixation. Besides imaging, AFM is also a force probe

2 SCANNING VOL. 00, 0 (2012)

tool with picoNewton force sensitivity. This makes itsuitable to measure the binding forces of individualreceptor–ligand pairs and this technology is calledsingle-molecule force spectroscopy (SMFS) (Mulleret al., 2009). In SMFS, ligands (or receptors) are at-tached to the AFM cantilever tip and the cognate re-ceptors (or ligands) are tethered to a support. SMFScan also be applied to directly measure the specificbinding forces of receptors on cell surface (Puntheera-nurak et al., 2006; Lee et al., 2007; Shi et al., 2009). Upto now nearly all SMFS experiments were performedon isolated biomolecules (Dufrene et al., 2011) or oncells cultured in vitro (Puntheeranurak et al., 2006;Lee et al., 2007; Shi et al., 2009). Biomolecules in vivoare highly controlled by their environments (Dufreneet al., 2011) and hence the isolated proteins may altertheir properties. Cells grown on flat two-dimensionalcell culture can differ considerably in their morphol-ogy and character from those growing in tissues andorgans (Yamada and Cukierman, 2007). Thus, thecells cultured in vitro may not reproduce all the facetsof human responses to diseases. Consequently, inves-tigating the behaviors of cells from patients will have asignificant impact. In this work, we used AFM to im-age the cellular topography and measure the molecu-lar force directly on pathological cells from lymphomapatients.

Material and Methods

Sample Preparation

The pathological cells were prepared from the bonemarrow of B-cell lymphoma patients whose bonemarrow was intruded by lymphoma cells. Poly-L-lysine was used to immobilize cells onto the glassslide. Drop poly-L-lysine solution onto the fresh glassslides and store at room temperature for natural dry-ing. A paracentesis needle was used to acquire thebone marrow from the lymphoma patients. Then, thebone marrow was dropped onto the poly-L-lysine-coated glass slides and another glass slide was used todisperse the bone marrow and 4% formaldehyde wasused for fixation.

Probe Functionalization

The heterobifunctional linker molecule NHS-PEG2000-MAL (JenKem Company, China) was usedto link rituximab onto the AFM tip. The probe func-tionalization procedure followed an established pro-cedure (Stroh et al., 2004). The NHS end of the PEGlinker reacts with amines on the silicon tip, yielding astable amide bond, while the MAL group reacts witha protein thiol resulting in a disulfide linkage betweenPEG and protein (Ebner et al., 2007).

Scanning Electron microscopy (SEM) and FluorescenceMicroscopy

SEM and fluorescence microscopy were applied toverify that rituximab had been linked onto AFM tips.For SEM, the functionalized probe was attached ontoa copper puck. Then place the puck onto the stage ofthe SEM for being scanned. For control experiments,the normal probe was scanned by SEM as same as thefunctionalized probe. For fluorescence validation, theRhodamine B isothiocyanate (RBITC) labeled goatanti-human IgG was used. First, the functionalizedprobe was placed into a petri dish in PBS and thenthe IgG solution was added and incubated for 1 h.After the reaction, the probe was rinsed with PBS toremove the unbound IgG and then was placed ontothe stage of the fluorescence microscope. For controlexperiments, the normal probe was treated as same asthe functionalized probe.

For cancer cell recognition, rituximab and RBITC-labeled goat anti-human IgG were used. (1) Put theslide pathological cell sample into a petri dish with 10mL PBS. Add 1mL rituximab into the petri dish andincubate in 4◦C for 4 h. (2) Wash the sample threetimes with PBS. (3) Drop the solution of IgG ontothe sample and incubate in 4◦C for 1 h. (4) Wash thesample three times with PBS. (5) Put the sample ontothe stage of the fluorescence microscope and use thegreen excitation light.

AFM Imaging and Force Measurements

AFM imaging and measurements were performedusing a Bioscope Catalyst AFM (Veeco Company,Santa Barbara, CA) with silicon nitride probes. Thenominal spring constant of the cantilever used was0.06 N/m. The spring constant was calibrated by aThermal Tune Adapter (Veeco Company, Santa Bar-bara, CA). The normal probes were used to image thetopography of cancer cells from lymphoma patientsat room temperature in air under the guidance of flu-orescence. The imaging mode was contact mode. TheSMFS experiments were performed on pathologicalcells at room temperature in PBS. The functionalizedprobes were moved onto the patient cells guided bythe optical system and force curves were obtained onthe cell surface. About 1,000 force curves were ob-tained for each pathological cell.

Results

Cancer Cell Recognition and AFM Imaging

Figure 1 shows the fluorescence staining images ofcells from lymphoma patients. The optical images ofthe pathologic bone marrow sample from lymphoma

M. Li et al.: Imaging and measuring molecular force of lymphoma pathological cells using AFM 3

Fig 1. Fluorescence staining of cell sample from lymphoma patients. (A), (D) were optical images of the sample and (B), (E) werethe corresponding fluorescence images, respectively. The insets in (A), (D) were the higher resolution optical images. (C), (F) werethe higher resolution fluorescence images of the cancer cells indicated by the circle in (B), (E), respectively.

patients are provided in Figure 1A and D. Thecorresponding fluorescence images are provided inFigure 1B and E. We can see that most of the cellsof the pathologic sample did not exhibit red fluo-rescence. We know that CD20 is only expressed onB lymphocytes (B cells). While in the bone marrow,cells (such as red blood cells) that do not expressCD20 are in the majority. Hence, we saw only a fewcells were stained with red fluorescence. For thesecells that exhibited fluorescence, a proportion ofthem had prominent features, such as round nucleuswith central nucleolus (Bouabdallah et al., 2003)(Fig. 1C), or irregular nucleus with splitting nucleolusand these indicated that these cells were cancer cells(Fig. 1F). We can see that cancer cells were theminority of the pathologic cell sample and the cancercells had special morphological features on nucleus.

Guided by the fluorescence, the detailed topogra-phy of cancer cells was imaged by AFM, as shown in

Figure 2. From the fluorescence image (Fig. 2A, E),we can see a cancer cell (denoted by the circles). Then,the AFM probe was moved to the cancer cells. First,a large scan (100 μm) was performed, and the topog-raphy image (Fig. 2B, F) was obtained. The cancercell can easily be recognized in the AFM images (de-noted by the square in Fig. 2B, F). By zooming intoa small area (30 μm), the higher topography imagesof the cancer cell were observed (Fig. 2C, D, G andH). From the AFM images of the cancer cells, wecan see that comparing with the healthy cells on thesample, cancer cells exhibit different morphology. Thehealthy cells have smooth and plump surfaces, whilethe cancer cells have corrugated and hollow surfaces.By applying AFM processing software, the roughnessof the cell surface was computed. Figure 3A was theroughness curve of a local area (4 μm) on the can-cer cell surface (denoted by the square). Figure 3Bwas the roughness curve on the healthy cell surface

Fig 2. Under the guidance of fluorescence, the topography of cancer cells was imaged by AFM. (A), (E) were fluorescence imagesand (B–D), (F–H) were the corresponding AFM images, respectively. (B), (F) were the topography images with a big scan size. (C),(D) were the topography images with a small scan size (indicated by the squares in (B), (F)) and (D), (H) were the correspondingAFM deflection images.

4 SCANNING VOL. 00, 0 (2012)

Fig 3. Measuring the roughness of the cancer cells and healthy cells. (A), (B) were the roughness curves of the local areas (indicatedby the squares) of the cancer cell and healthy cell, respectively. (C) Contrast of the roughness of the cancer cells and healthy cells.Eight cancer cells and eight healthy cells were used to compute the roughness.

(denoted by the square). The roughness of the can-cer cell was 78.34 nm, while the roughness of thehealthy cell was 21.31 nm. To quantitatively measurethe cellular roughness, eight cancer cells and eighthealthy cells were chosen to compute the roughnessand the contrast of roughness of cancer cells andhealthy cells was shown in Figure 3C. We can see thatcancer cells were about four times rougher thanhealthy cells.

Probe Functionalization

To measure the CD20-rituximab binding forces,PEG molecules were used to link rituximabs ontothe AFM tip. After the functionalization procedure,we should check whether rituximab molecules hadbeen linked onto the tip. Figure 4 shows the SEMand fluorescence microscopy results of verificationexperiments. The SEM images of a normal tip andthe SEM images of a functionalized tip are providedin Figure 4A and B and Figure 4C and D, respec-tively. We can see that there were many particles on

the surface of the functionalized tip (denoted by thearrows in Fig. 4D) while the surface of the normaltip was smooth. The optical images were shown inFigure 4E and G and the corresponding fluorescenceimages were shown in Figure 4F and H, respectively.We can see that the fluorescence of functionalizedtip was bright and the fluorescence around the tip(denoted by white circle in Fig. 4H) was obviouslybrighter than other areas, while the fluorescence ofthe normal tip was dim. The tip shape was pyramidaland the tip height was in the range of 2.5–8.0 μm. Af-ter the functionalization procedure, rituximabs boundonto the tip surface through PEG linker. Rituximabsshould also bind onto the cantilever of the probe. Butthe density of rituximabs on the tip was much largerthan in the cantilever. Then, after the fluorescencestaining, the density of fluorescein-labeled secondaryantibody on the tip was much more than on the can-tilever and this resulted in that the fluorescence of thetip was markedly brighter than that of the cantilever.Hence, through the functionalization procedure, rit-uximab molecules were successfully linked onto theAFM tip.

Fig 4. Using SEM and fluorescence microscopy to verify whether rituximabs had been linked onto the AFM tip. (A), (B) werethe SEM images of the normal probe and (C), (D) were the SEM images of the functionalized probe. (B), (D) were the higherresolution SEM images (indicated by the squares in (A), (C), respectively). (E), (G) were the optical images of the normal probeand the functionalized probe respectively and (F), (H) were the corresponding fluorescence images.

M. Li et al.: Imaging and measuring molecular force of lymphoma pathological cells using AFM 5

Fig 5. Measuring the CD20-rituximab binding forces on pathological cells. (A) A typical force curve with CD20-rituximab bindingoccurred. (inset) Principle of the force measurement. (B) A typical force curve with no binding occurred. (C) Histogram of bindingforces of one cell. (D) The distribution of the binding forces of 22 different cells.

Molecular Force Measurements on Pathological Cells

The CD20-rituximab binding forces were mea-sured quantitatively on pathologic cells by using thefunctionalized tips. The method used here was usingthe functionalized tips to obtain force curves on cellsfrom the sample. If the force curves had prominentabrupt peaks, then the cell was chosen for measur-ing the CD20-rituximab binding forces. Figure 5Awas a typical force curve where the abrupt peak indi-cated the CD20-rituximab specific binding. The insetin Figure 5A was the principle of the force measure-ment. Rituximabs were linked onto the AFM tip andthen the tip was approached and contacted with theCD20 molecules on the cell surface, obtaining forcecurves during the approach-retract process. CD20 isa four-transmembrane B cell-specific molecule thatfunctions as a membrane-embedded Ca2+ channel(Lebien and Tedder, 2008). The binding of rituximabto CD20s on the B-cell surface can result in the deple-tion of the B cell by three mechanisms, including Fc-FcrR interactions, complement-dependent cytotoxi-city (CDC), and the programmed cell death (PCD)(Alduaij and Illidge, 2011). For each cell, ∼1,000force curves were obtained at the same loading rate1.99 μm/s. For these 1,000 force curves, only a smallproportion of them had typical molecular recognitionabrupt peaks and these curves were selected to com-pute the binding force. Figure 5B was a typical forcecurve when no CD20-rituximab binding occurred. Wecan see that there was also a peak in Figure 5B andthis is because of the nonspecific adhesion interac-tions between tip and cell surface. The magnitude ofthe peak in Figure 5B was very small and the shapeof the peak indicated they were nonspecific adhe-sion force. For contrast, we can see that the peak inFigure 5A had a step-like shape (indicated by the ar-row) and this shape was caused by the stretching ofPEG linker and this shape was the typical fashionof molecular binding (Hinderdorfer et al., ’96). Fig-ure 5C was the histogram of the binding forces of onecell and the Gaussian fit indicated that the bindingforce was 132 ± 51 pN. The binding forces of the 22different cells were shown in Figure 5D. We can seethat the binding forces were mainly in the range of

110–120 pN (nine cells) and 130–140 pN (seven cells).The binding forces of four cells were about 100 pNand the binding forces of two cells were about 145 pN.

Discussion

A century ago, Paul Ehrlich envisioned antibod-ies as “magic bullet” that would specifically trace andkill tumor cells (Schrama et al., 2006). His dream isnow not only a reality but a major aspect of clinicalmedicine (Schwartz, 2004): more than 25 antibod-ies are approved for human therapy and more than240 antibodies are currently in clinical developmentworldwide (Chan and Carter, 2010). The prosperityof antibody engineering was largely because of theunprecedented success of rituximab in the treatmentof B-cell lymphomas since its approval in 1997. Rit-uximab is a powerful drug against B-cell lymphoma,but many patients with CD20-positive malignanciesstill fail to respond to the rituximab-containing im-munochemotherapy (Alduaij and Illidge, 2011). Theunderlying mechanisms for the rituximab resistanceremain unknown (Taylor and Lindorfer, 2010) andthis problem has become a big challenge for the phar-maceutical industry to develop the next generation ofantibody-targeted drugs. The advent of AFM, espe-cially the rapid development of high-speed AFM inrecent years (Kodera et al., 2010), opens an oppor-tunity to understand the rituximab resistance at indi-vidual cell/molecule levels under physiological condi-tions and may provide novel insights into the cellularprocesses. Since the current researches of AFM weremainly on cells cultured in vitro, using AFM to inves-tigate the behavior of pathological cells may have asignificant impact.

Since its invention, AFM has been used for cellimaging (Butt et al., ’90). AFM provides higherspatial resolution than optical microscopy and caneven image single molecule on the membrane (Casusoet al., 2011). With AFM imaging, the topographychanges of single living cells can be observed (Rotschand Radmacher, 2000). Currently, most of the AFMstudies were performed on cells cultured in vitro.Investigating the topography of pathological cells

6 SCANNING VOL. 00, 0 (2012)

from patients has, to our knowledge, not been largelyreported before. Researchers have imaged the patho-logical erythrocytes from a patient with hereditaryspherocytosis and observed the changes of the cellshape (Dulinksa et al., 2006). Also, the mechanicalproperties of metastatic cancer cells and benign cellsfrom cancer patients have been investigated by AFM(Cross et al., 2007). Here, under the guidance offluorescence, the detailed topography of lymphomacancer cells were imaged by AFM and the AFM im-ages showed that cancer cells were corrugated whilethe healthy cells were smooth, as shown in Figure 2.Besides, the roughness of the cancer cells and healthycells was quantitatively measured. We chose eightcancer cells and eight healthy cells to compare theirroughness. The roughness measurement results indi-cated that cancer cells were about four times rougherthan healthy cells (Fig. 3). We know that the geneticalterations drive the progressive transformation ofnormal human cells into malignant cells (Hanahanand Weinberg, 2000). These genetic alterations maycause the cellular changes, such as the elasticityand the roughness. Researchers have indicated thatcancer cells were softer than normal cells (Crosset al., 2007). Here, we can see that cancer cells wererougher than normal cells.

With SMFS, many types of molecular bindingforces had been measured before. The binding force ofsingle biotin–avidin was about 160 pN (Florin et al.,’94), and the binding force of single antibody–antigen(human serum albumin and antibody) was about 244pN (Hinterdorfer et al., ’96). These two measurementswere performed on isolated proteins. The experimentson living CHO cell surface indicated that the bind-ing force of single antibody–antigen (SGLT1 and an-tibody) was about 100 pN (Puntheeranurak et al.,2006). Usually the binding force of single receptor–ligand interactions are about 20–200 pN (Mulleret al., 2009). Here, we measured the binding forcesof CD20-rituximab on 22 pathological cells. The re-sults indicated that the binding forces were mainlyin the range of 110–120 pN, 130–140 pN. Hence,the forces measured here were in the normal rangeof single receptor–ligand binding force. We knowthat for the lymphoma patients whose bone marrowwas intruded by lymphoma cells, there are normalB cells and malignant B cells in the bone marrow.These 22 cells measured here may include both nor-mal B cells and malignant B cells. Malignant B cellswere quite different from normal B cells (Klein andDalla-Favera, 2008; Kotani et al., 2010) and this dif-ference may cause different properties of the malig-nant cells and normal cells, such as antibody–antigenaffinity. Researchers have indicated that the distribu-tion of CD20 and cellular surface ultrastructre of nor-mal B cells were quite different than that of B-chroniclymphocytic leukemia cells (Wang et al., 2011). Be-

sides, B cells have many different types, includingpro-B cells, pre-B cells, mature B cells, memory Bcells, and plasma B cells (Lebien and Tedder, 2008).These different B cells may also have different sur-face structures that may induce different antibody–antigen affinities. Compared to the CD20-rituximabbinding force (about 80 pN) on lymphoma Raji cellscultured in vitro (Li et al., 2010), we can see that thebinding forces on pathological cells were much largerthan that on Raji cells. This may be because of thedifferent environments of pathological cells and Rajicells. Pathological cells were in vivo, while Raji cellswere cultured in vitro. This difference may cause thedifferent CD20-rituximab binding forces.

In summary, the rapid developments in AFMsingle-molecule techniques have allowed researchersto characterize the behavior and measure the bindingforces of single molecules. Here, we used AFM to in-vestigate the cellular morphology and molecular forcedirectly on lymphoma pathological cells. The AFMimages showed the special topography (corrugated,hollow) of cancer cells compared to healthy cells(smooth, plump) and the roughness measurementsindicated that cancer cells were about four timesrougher than healthy cells. The molecular force mea-surements indicated that the CD20-rituximab bindingforces of pathological cells were mainly in the rangeof 110–120 pN, 130–140 pN, significantly larger thanthat on Raji cells. These results will facilitate furtherresearches of rituximab’s different efficacies.

References

Alduaij W, Illidge TM. 2011. The future of anti-CD20monoclonal antibodies: are we making progress. Blood117:2993–3001.

Binnig G, Quate CF, Gerber C. 1986. Atomic force micro-scope. Phys Rev Lett 56:930–933.

Bouabdallah R, Mounier N, Guettier C, Molina T, Ribrag V,Thieblemont C, Sonet A, Delmer A, Belhadj K, GaulardP, Gisselbrecht C, Xerri L. 2003. T-cell/histiocyte-richlarge B-cell lymphomas and classical diffuse large B-celllymphomas have similar outcome after chemotherapy: amatched-control analysis. J Clin Oncol 21:1271–1277.

Butt HJ, Wolff EK, Gould SAC, Northern BD, Peterson CM,Hansma PK. 1990. Imaging cells with the atomic forcemicroscope. J Struct Biol 105:54–61.

Casuso I, Rico F, Scheuring S. 2011. Biological AFM: wherewe come from—where we are—where we may go. J MolRecognit 24:406–413.

Cecconi C, Shank EA, Bustamante C, Marqusee S. 2005. Di-rect observation of the three-state folding of a single pro-tein molecule. Science 309:2057–2060.

Chan AC, Carter PJ. 2010. Therapeutic antibodies for au-toimmunity and inflammation. Nat Rev Immunol 10:301–316.

Cross SE, Jin YS, Rao J, Gimzewski JK. 2007. Nanomechani-cal analysis of cells from cancer patients. Nat Nanotechnol2:780–783.

Dufrene YF. 2009. Atomic force microscopy: a powerfulmolecular toolkit in nanoproteomics. Proteomics 9:5400–5405.

M. Li et al.: Imaging and measuring molecular force of lymphoma pathological cells using AFM 7

Dufrene YF, Evans E, Engel A, Helenius J, Gaub HE, MullerDJ. 2011. Five challenges to bringing single-molecule forcespectroscopy into living cells. Nat Methods 8:123–127.

Dulinksa I, Targosz M, Strojny W, Lekka M, Czuba P, Bal-wierz W, Szymonski M. 2006. Stiffness of normal andpathological erythrocytes studied by means of atomic forcemicroscopy. J Biochem Biophys Methods 66:1–11.

Dupres V, Alsteens D, Andre G, Verbelen C, Dufrene YF.2009. Fishing single molecules on live cells. Nano Today4:262–268.

Ebner A, Wildling L, Kamruzzahan ASM, Rankl C, WrussJ, Hahn CD, Holzl M, Zhu R, Kienberger F, Blaas D,Hinterdorfer P, Gruber HJ. 2007. A new, simple methodfor linking of antibodies to atomic force microscopy tips.Bioconjug Chem 18:1176–1184.

Florin EL, Moy VT, Gaub HE. 1994. Adhesion forces betweenindividual ligand-receptor pairs. Science 264:415–417.

Hanahan D, Weinberg RA. 2000. The hallmarks of cancer.Cell 100:57–70.

Hinterdorfer P, Baumgartner W, Gruber HJ, Schilcher K,Schindler H. 1996. Detection and localization of individ-ual antibody-antigen recognition events by atomic forcemicroscopy. Proc Natl Acad Sci USA 93:3477–3481.

Katan AJ, Dekker C. 2011. High-speed AFM reveals the dy-namics of single biomolecules at the nanometer scale. Cell147:979–982.

Klein U, Dalla-Favera R. 2008. Germinal centres: role in B-cellphysiology and malignancy. Nat Rev Immunol 8:22–33.

Kodera N, Yamamoto D, Ishikawa R, Ando T. 2010. Videoimaging of walking myosin V by high-speed atomic forcemicroscopy. Nature 468:72–76.

Kotani A, Harnprasopwat R, Toyoshima T, Kawamata T, TojoA. 2010. miRNA in normal and malignant B cells. Int JHematol 92:255–261.

Lebien TW, Tedder TF. 2008. B lymphocyte: how they developand function. Blood 112:1570–1579.

Lee S, Mandic J, Vliet KJV. 2007. Chemomechanical map-ping of ligand-receptor binding kinetics on cells. Proc NatlAcad Sci USA 104:9609–9614.

Li M, Liu L, Xi N, Wang Y, Dong Z, Li G, Xiao X, Zhang W.2010. Detecting CD20-Rituximab specific interactions onlymphoma cells using atomic force microscopy. Sci ChinaLife Sci 53:1189–1195.

Muller DJ, Helenius J, Alsteens D, Dufrene YF. 2009. Forceprobing surfaces of living cells to molecular resolution.Nat Chem Biol 5:383–390.

Puntheeranurak T, Wildling L, Gruber HJ, Kinne RKH, Hin-terdorfer P. 2006. Ligands on the string: single-moleculeAFM studies on the interaction of antibodies and sub-strates with the Na+− glucose co-transporter SGLT1 inliving cells. J Cell Sci 119:2960–2967.

Rotsch C, Radmacher M. 2000. Drug-induced changesof cytoskeletal structure and mechanics in fibroblast:an atomic force microscopy study. Biophys J 78:520–535.

Schrama D, Reisfeld RA, Becker JC. 2006. Antibody targeteddrugs as cancer therapeutics. Nat Rev Drug Discov 5:147–159.

Schwartz RS. 2004. Paul Ehrlich’s magic bullets. N Engl J Med350:1079–1080.

Shi X, Xu L, Yu J, Fang X. 2009. Study of inhibition ef-fect of Herceptin on interaction between Heregulin andErbB receptors HER3/HER2 by single-molecule forcespectroscopy. Exp Cell Res 315:2847–2855.

Stroh C, Wang H, Bash R, Ashcroft B, Nelson J, GruberH, Lohr D, Lindsay SM, Hinterdorfer P. 2004. Single-molecule recognition imaging microscopy. Proc Natl AcadSci USA 101:12503–12507.

Taylor RP, Lindorfer MA. 2010. Antigenic modulation andrituximab resistance. Semin Hamatol 47:124–132.

Wang Q, Lu Y, Li S, Wang M, Cai J. 2011. Distribution andforce spectroscopy of CD20 antigen-antibody binding onthe B cell surface. Chin J Biotechol 27:131–136.

Yamada KM, Cukierman E. 2007. Modeling tissue morpho-genesis and cancer in 3D. Cell 130:601–610.

Zhuang X, Bartley LE, Babcock HP, Russell R, Ha T, Her-schlag D, Chu S. A single-molecule study of RNA catalysisand folding. 2000. Science 288:2048–2051.