- Email: [email protected]
Nanomechanical probing of microbubbles using the atomic force microscope
Available online at www.sciencedirect.com
Ultrasonics 46 (2007) 349–354 www.elsevier.com/locate/ultras
Nanomechanical probing of microbubbles using the atomic force microscope V. Sboros
b
a,*
, E. Glynos b, S.D. Pye c, C.M. Moran a, M. Butler a, J.A. Ross d, W.N. McDicken a, V. Koutsos b
a Medical Physics, School of Clinical Sciences and Community Health, University of Edinburgh, Edinburgh, UK Institute for Materials and Processes, Centre for Materials Science and Engineering, School of Engineering and Electronics, University of Edinburgh, Edinburgh, UK c Medical Physics, Royal Infirmary of Edinburgh, Edinburgh, UK d Clinical and Surgical Sciences, University of Edinburgh, Edinburgh, UK
Received 27 April 2007; received in revised form 18 June 2007; accepted 20 June 2007 Available online 30 June 2007
Abstract Atomic force microscopy (AFM) is a versatile mechanical nanosensor that can be used to quantify the mechanical properties of microbubbles (MBs) and the adhesion mechanisms of targeted MBs. Mechanical properties were investigated using AFM tipless cantilevers to microcompress the MBs. The range of compressive stiffness for biSphere was found to be between 1 and 10 N m1 using a cantilever with a spring constant of 0.6 N m1. This stiffness was shown to decrease with the MB size in a non-linear fashion. It is also possible to calculate a theoretical Young’s modulus of the shell. The adhesion properties of targeted lipid based MBs that use avidin–biotin chemistry for the attachment of targeting ligands were also studied. The MBs were attached to poly-L-lysine treated tipless cantilevers with spring constants ranging from 0.03 to 0.1 N m1. This system interrogated individual cells with pulling cantilever distance of 15 lm, and scan rate at 0.2 Hz. The depth of contact was not larger than 0.4 lm. The targeted MBs provided a significantly larger adhesion to the cells compared to control ones. Average adhesion force was dependent on depth of contact. Analysis of the data demonstrated a single distribution of adhesion events with median at 89 pN, which is in agreement with the literature for such interactions. The nanointerrogation of MBs using AFM provides new insight into their mechanical properties, and should be of assistance to MB design and manufacture. 2007 Elsevier B.V. All rights reserved.
1. Introduction As has been shown in a number of areas of pathology MBs are strong candidates as ultrasonic contrast agents used to quantify perfusion non-invasively [1]. The development of site-targeted MBs expands the field beyond perfusion imaging and allows attachment to specific markers of disease [1]. The expansion of MB technology to include localised drug and gene delivery is potentially very valuable
*
Corresponding author. Tel.: +44 131 2426292; fax: +44 131 2426314. E-mail address: [email protected] (V. Sboros).
0041-624X/$ - see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ultras.2007.06.004
[2,3]. The complex interaction that MBs have with ultrasound [4] is however, often overlooked and has yet to be theoretically modelled successfully. New imaging modalities that exploit MB destruction [5–8], and reject background tissue signals [9–11] are incorporated in ultrasonic equipment, but are based on a simple model of MB properties. As perfusion imaging deals with the smallest of vessels and pathology often deals with subtle changes in vessel numbers, a perfusion imaging tool needs to be very sensitive thus requiring optimised transmitted beams for matching MBs and their appropriate signal processing. The success of this effort relies on a thorough understanding of the physical properties of MBs and their interaction with
350
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
the ultrasound field. Towards achieving this goal, fast optical microscopy has recorded an array of MB behaviours [12] and has proved to be of substantial value. However, the viscoelastic properties of the shell, which has typical thickness between 10 and 200 nm, have not been measured accurately but have been estimated and remain the most important unknown in MB behaviour. Fitting the shell parameters into MB theoretical models is a common practice in previous theoretical investigations. Other parameters responsible for MB motion have been studied systematically and validated experimentally. However, the fitting approach used for the shell is explorative, and perhaps the main reason for the limited success of these models. Knowledge of the mechanical properties of the shell will elucidate its contribution to the MB motion in the presence of ultrasound. It is known that MBs do not scatter in a uniform manner and at low acoustic pressures only a small fraction of the MB population scatter efficiently [13]. Knowledge of the shell property may also offer an explanation for this non-uniformity. It is important to note that optical microscopy is limited in such measurements, primarily due to spatial resolution restrictions. In the past, MB models either assumed a fixed behaviour for the shell properties [14] or used a best fit to experimental data [15]. Today a possible solution is provided by the atomic force microscope (AFM), a versatile mechanical micro/nano sensor [16,17]. We have previously recorded topographies of the biSphere (Point Biomedical Corp., San Carlos, CA, USA) MB (Fig. 1) and showed that the measurement of the elastic properties of MBs is feasible [18]. In this paper we present new data that provide insight on the elastic properties of biSphere. In addition to determining the properties of the bubble shell, targeted MBs require a nanointerrogation technique that will provide a detailed analysis of the behaviour of tar-
geting mechanisms in the vicinity of target cells. AFM has been shown to provide adhesion measurements at the level of single hydrogen bonds [19]. The ability of targeted MBs to adhere to cells using a specific molecular interaction is investigated here. 2. Materials and methods 2.1. MB mechanical properties The methodology is similar to our previous investigation [18]. Force–distance (f–d) curves were acquired on biSphere MBs using the MFP-1D AFM system (Asylum Research, Santa Barbara, CA, USA) with 1 Hz scan rates and 3 lm ramp size. The MBs were mechanically interrogated with aluminium coated tipless cantilevers (NSC-12, MikroMash, Madrid, Spain) with spring constants around 0.6 N/m. Spring constant calibration was performed by means of thermal noise spectra in air [20] available on the MFP-1D. The MB effective spring stiffness was assessed by measuring the deflection of the AFM cantilever during compression. This deflection depends on the elastic property of the MB and the cantilever. In order to calculate the deformation of the MB the deflection of the cantilever needs to be subtracted. s ¼ d F =k c
ð1Þ
where s is the MB deformation during compression after initial contact between the cantilever and the MB, d is the cantilever deflection, F the force applied and kc the cantilever spring constant. Note that F/kc corresponds to the deflection of the cantilever at a hard surface at the same force F. Thus force–distance (or deflection) (f–d) measurements are converted to force–deformation (f–s) measurements. The AFM is mounted on an inverted optical microscope (TE2000U, Nikon UK Limited, Surrey, UK). Images of MBs under the cantilever can be viewed in reflection mode. A digital camera (Hamamatsu Orca-ER C4742-80, Hamamatsu Photonics, Hamamatsu city, Japan) is attached to the microscope and the images are transferred and processed on a PC using image analysis software (IPLab v3.7, BD Bioscience Bioimaging, Rockville, USA). Using a 60· objective it is possible to size a MB with half micron accuracy. 2.2. Adhesion properties of targeted MBs
Fig. 1. BiSphere MB topography. Methodology is described in (1817). The MB has 1.63 lm diameter. The outline of the cross section of the MB topography data fits a sphere. Note the big defect area, which reaches a maximum peak to peak roughness 0.66 lm. It can be seen that a different layer emerges inside this area. It is known that biSphere is made out of two separate layers, and the outer layer is typically a few nanometers. The gap presented here is much larger and reveals an inner layer. Structural studies of MBs may reveal important aspects of MB construction and should be of use to their design and manufacturing.
The hepatic endothelial cell line SkHep-1 was cultured using standard techniques on Petri dishes [21]. Half of the Petri dish surface was utilised for this purpose. Flow cytometry analysis was carried out using a Coulter XL flow cytometer, as described previously [22], and demonstrated relatively low expression of CD31 antigen on the surface of these cells (CD31– mean fluorescence intensity (mfi) = 0.454 compared with isotype control mfi = 0.245).
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
The manufacture of the in-house MBs has been described previously [23]. They are phospholipid-based and were used as controls. A modified process was used to produce targeted MBs, a biotinylated phospholipid (Avanti, Alabaster, USA) was incorporated into the lipid shell and linked to biotinylated CD31 antibody (Abcam, Cambridge, UK) via a streptavidin bridge (Sigma, Poole, UK). Tipless cantilevers (CSC-12Al, MikroMash, Madrid, Spain) of typical spring constants ranging between 0.03 and 0.1 N/m were treated with 1/10 poly-L-lysine solution for 5 min, then rinsed and dried. The MB suspension was placed in the half of the Petri dish not occupied by the cultured cells. The cantilever approached the MBs manually until contact was reached, the MB was trapped between the cantilever and the bottom of the dish, and a slight deflection was registered. Twenty seconds were allowed to pass before the cantilever was retracted. The MB attachment was confirmed by subsequent microscopic inspection. The cultured cells were then placed under the cantilever/ MB probe. F–d curves were acquired using the MFP-1D AFM (Asylum Research, Santa Barbara, CA, USA) and were converted into f–s curves as for the mechanical properties experiment. The scan rate of 0.2 Hz was used throughout the adhesion experiments. 3. Results and discussion 3.1. MB mechanical properties A typical f–s curve is displayed in Fig. 2a. Upon the approach (0.30–0.10 lm distance) of the cantilever to the
160
351
biSphere MB, the deflection is zero as there are no forces which would impose a deflection. When the cantilever contacts the MB at a distance of 0.10 lm it deflects and this is recorded in Fig. 2a as positive force. When the force reaches a maximum (here 168 nN), the direction of the cantilever changes and the force reduces until eventually the cantilever detaches from the MB. The MB stiffness was calculated from the compression curve after the short non-linear part (10 nN) immediately after contact. Only MBs with reproducible f–s curves were included in this study. This was always the case for biSphere and there was no change in the stiffness measurement over 100 compressions. The calculations used the section of the curve below 200 nm MB deformation, and it was found that eight f–s curves are adequate to provide an assessment of the uncertainty of the measurement. The stiffness of the MB in Fig. 2a was 2.20 ± 0.06 N m1. The stiffness for different MB sizes are plotted in Fig. 2b, and display a monotonic stiffness decrease with size. Further analysis may employ the use of a mechanical model in order to assign an elastic modulus to MB shell material [24]: p F ¼ pffiffiffi Eh2 e1=2 þ 4pER0 he3 ð2Þ 2 2 where F is the force applied to the MB, E the MB shell elastic modulus, h the thickness of the shell, R0 the radius of the MB at rest, and e the relative deformation (=d/2R0 where d is the MB deformation). An approximate thickness is given from the manufacturer to be proportional to the radius and approximately 7.5 nm for 1 lm diameter [25]. The first term of the right side of the equation accounts for the bending of the finite thickness of the shell, which is important at small deformations (<5%), while the second term accounts for the elasticity of the whole shell and gains significance at larger deformation. The Young modulus E
140
100 80 60 40 20 0 -200.00
0.05
0.10
0.15
0.20
0.25
0.30
Deformation (micrometers)
Fig. 2a. Typical f–s curve for a 4.7 lm diameter biSphere MB. Approach (blue curve) and retract (red curve) show some hysteresis maybe due to the viscous behaviour of the MB. Upon contact there is a degree of noise most probably created due to the hydrodynamics of the water-immersed cantilever–MB system at the specific scan rate (1 Hz). The measurement of the MB stiffness was made within contact part of the approach curve and was found to be 2.2 N/m. The fitted curve (dashed line) to the contact part of the approach curve, is created by fitting a mechanical model to the results in order to extract a value for the Young modulus of the shell (see equation and discussion in Section 3). (For interpretation of the references in colour in this figure legend, the reader is referred to the web version of this article.)
Microbubble stiffness (N/m)
Force (nN)
120 8 7 6 5 4 3 2 1 0 2
3
4
5
6
Diameter (micrometers)
Fig. 2b. BiSphere MB stiffness vs. diameter as captured using a cantilever with 0.6 N/m stiffness. The average stiffness is calculated over eight curves for each microbubble. The standard deviation of the stiffness measurement was generally below 5%, while the diameter measurement was made with 60· objective and provided a standard deviation of 0.3 lm. A monotonic decrease of stiffness with size increase is apparent in this graph. Moreover, MBs with similar size often have significantly different stiffness.
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
3.2. Adhesion properties of targeted MBs Control and targeted MB interrogation of SkHep-1 cells are shown in Fig. 3a and b, respectively. The curves lack
30 25 20
Force (nN)
is proportional to the applied force and can be deduced with a simple fit. From the data of Fig. 2a, E = 1.35 GPa can increase to 10 GPa for smaller sized MBs. Our deformation is less than 5% and the second term of the above equation has a small contribution. The fit of the force curve (dashed line in Fig. 2a) presents another problem, however. The equation above is designed mainly for soft materials and relatively large deformations and as a result has been used and tried in the regime of dominance of the second term [24]. This means that the approximation made for small deformations, which is more relevant for the data presented here, may be different. Further the biSphere has a softer second outer shell layer [25] which may affect the appearance of the curve. Further work is required to address this complex problem. It is proposed that the present calculations of the elastic modulus may have some qualitative value and be assigned to the inner structural layer of the shell assuming that this is the hardest layer. Another important aspect of the data presented in Fig. 2b is the monotonic increase in MB stiffness with the decrease in diameter and consequently with the shell thickness. This indicates a size effect on the measured mechanical properties of the microbubble shell. The observed behaviour of E, as derived from Eq. (2), is similar to Fig. 2b and to what has been previously observed in other nanosystems [26]: the elastic modulus increased non-linearly and significantly beyond the value expected from macroscopic measurements of the bulk material when the characteristic dimensions of the interrogated object obtained very small values. Further analysis of MB shell thickness may be required, for example by electron microscopy, to establish the dependence of thickness and structure on MB size. Among the MBs presented in Fig. 2b, there are several of similar size but with significantly different stiffness, which is evidence of mechanical/structural dispersion in the shell manufacture. The force data presented here refer to MB mechanical properties in the static regime. MBs are expected to have different viscoelastic properties when exposed to ultrasonic waves. However, it is very important to connect structural characteristics and mechanical behaviour at the relatively simple and accessible AFM low frequency regime. The usefulness of the current data lies in making this connection possible and in this way feedback can be generated for the manufacturing processes. E.g. it is apparent from the dispersion in our data that there is a large variation in the shell characteristics (thickness, local curvature, nanostructure, defects) for microbubbles of the same diameter. MB models have assumed uniform shell characteristics in the absence of such data. The present experimental set-up offers a tool that may provide a useful link between manufacturing and acoustic [13,27] or optical [28] experiments.
15 10 5 0 1
0 -5
b
2
3
5
4
6
Deformation (micrometers)
2
1
Force (nN)
352
0 0
2
4
6
8
10
12
14
16
18
-1
-2
-3
Deformation (micrometers)
Fig. 3. (a) F–s curve of an in-house phospholipids MB with an SkHep-1 cell. Approach (blue curve) and retract (red curve) provide some hysterisis due to the viscosity of both MB and cell. Hardly any adhesion is observed when the MB/cantilever system retracts and detaches from the cell. (b) F–s curve of an in-house targeted MB with CD31 with an SkHep-1 cell. A significant adhesion is registered when the MB/cantilever system retracts. The cumulative adhesion is 2.06 nN while 16 detachment events are recorded that belong to the 1.52 nN of the force. (For interpretation of the references in colour in this figure legend, the reader is referred to the web version of this article.)
the low frequency noise observed in Fig. 2a. This may be attributed to the difference in scanning rate, geometry and mechanical properties of the structures involved, further investigation would be required to confirm this. In both control and targeted MBs the approach and retraction contact areas do not coincide (i.e. positive forces do not occur at the same deformations), which may be attributed to the cell or MB viscous properties. Overall 62 curves were acquired in five control MB–cell pairs; 15% of the f–s curves recorded some adhesion. For the targeted MBs, 95 curves were acquired in eight MB–cell pairs. The adhesion was apparent in 85% of the curves, and was significantly greater than the adhesion observed in the control experiments. Different MB–cell pairs provided often significantly different cumulative adhesion. This difference may be due to changes in the depth of contact, cell life/condition and to the variability in the manufacture of targeted MBs. The shape of Fig. 3b is typical and records 16 vertical jumps that are single detachment events and account for 1.52 nN of the adhesion developed between the MB and the cell during this contact. The cumulative adhesion was
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
2.06 nN, which means that 0.54 nN of adhesion could not be differentiated as single de-adhesion events, and most probably are below the spatial and force resolution limits for the AFM under the specific cantilever and settings. Between unbinding events horizontal plateaux were nearly always observed. The force jumps may be associated with the force required to detach the MB and the cell in a specific adhesion site. Each site may develop one or more bonds. Analysing all the available adhesion jumps we observed that the majority of the bonds (>80%) lie between 50 pN and 150 pN and the median value was 89 pN. A full statistical analysis requires more measurements. These will establish the behaviour of events and conclude as to whether we observe a single or multiple distributions of individual bonds. The recorded events may not be specific to CD31 interactions, as there are several proteins on the cell surface that are likely to bind with the CD31 antibody. However, work on similar interactions such as the LFA-1/ICAM-1 interaction shows dominant unbinding forces around 100 pN [29]. Further work should investigate the specificity of the interactions measured here, and the ability of the targeted MBs to target specific cells. A larger number of f–d curves like the one presented in Fig. 3b may provide enough data for analysing the distance between two unbinding events [30] and may provide useful information on the spatial distribution of targeting sites on the MB as well as the function of targeting. 3.3. Summary Little is known about the MB shell as (a) an elastic body that is a major contributor to MB motion in an ultrasonic field, or (b) as a structure on which targeting sites may be grafted. We have demonstrated that both these aspects can be characterised using atomic force microscopy to measure: (a) the elastic properties of the biSphere shell and (b) the adhesion of targeted in-house MBs with an endothelial cell line. The techniques described here provide unique information about the physical and biological properties of MBs, and have the potential for further development. Acknowledgements The authors acknowledge the generous offering of cell culturing facilities by Dr. Alistair Elfick. Thanks to Carrie Cunningham, James Black and Ian Ansell for all their work throughout this project. Also, special thanks to Bob Short (Point Biomedical Corporation, San Carlos, CA, USA) for the valuable information and guidance throughout this work. The work was funded by the Engineering and Physical Sciences Research Council, UK (Project Grant 21639/01). References [1] J.R. Lindner, Microbubbles in medical imaging: current applications and future directions, Nat. Rev. 3 (2004) 527–532.
353
[2] R.V. Shohet, S. Chen, Y.T. Zhou, Z. Wang, R.S. Meidell, R.H. Unger, P.A. Grayburn, Echocardiographic destruction of albumin microbubbles directs gene delivery to the myocardium, Circulation 101 (2000) 2554–2556. [3] E.C. Unger, E. Hersh, M. Vannan, T.O. Matsunaga, T. McCreery, Local drug and gene delivery through microbubbles, Prog. Cardiov. Dis. 44 (2001) 45–54. [4] V. Sboros, K.V. Ramnarine, C.M. Moran, S.D. Pye, W.N. McDicken, Understanding the limitations of ultrasonic backscatter measurements from microbubble populations, Phys. Med. Biol. 47 (2002) 4287–4299. [5] T.R. Porter, F. Xie, Transient myocardial contrast after initial exposure to diagnostic ultrasound pressures with minute doses of intravenously injected microbubbles, Circulation 92 (1995) 2391–2395. [6] T.R. Porter, F. Xie, D. Kricsfeld, R.W. Armbruster, Improved myocardial contrast with second harmonic transient ultrasound response imaging in humans using intravenous perfluorocarbonexposed sonicated dextrose albumin, J. Am. Coll. Cardiol. 27 (1996) 1497–1501. [7] V. Uhlendorf, C. Hoffmann, Nonlinear acoustical response of coated microbubbles in diagnostic ultrasound, IEEE Ultrasonics Symp. 3 (1994) 1559–1562. [8] A. Bauer, R. Schlief, M. Zomack, A. Urbank, H.P. Niendorf, in: N.C. Nanda, R. Schlief, B.B. Goldberg (Eds.), Advances in Echo Imaging using Contrast Enhancement, Kluwer Academic Publisher, Dodrecht, The Netherlands, 1997. [9] D.H. Simpson, C.T. Chin, P.N. Burns, Pulse inversion doppler: a new method for detecting nonlinear echoes from microbubble contrast agents, IEEE Trans. UFFC 46 (1999) 372–382. [10] V. Mor-Avi, E.G. Caiani, K.A. Collins, C.E. Korcarz, J.E. Bednarz, R.M. Lang, Combined assessment of myocardial perfusion and regional left ventricular function by analysis of contrast-enhanced power modulation images, Circulation 104 (2001) 352–357. [11] W. Wilkening, B. Brendel, H. Jiang, J. Lazenby, H. Ermert, Optimized receive filters and phase-coded pulse sequences for contrast agent and nonlinear imaging, IEEE Ultrasonics Symp. (2001) 1733–1737. [12] M. Postema, A. van Wamel, C.T. Lancee, N. de Jong, Ultrasoundinduced encapsulated microbubble phenomena, Ultrasound Med. Biol. 30 (2004) 827–840. [13] V. Sboros, C.M. Moran, S.D. Pye, W.N. McDicken, The behaviour of individual contrast agent microbubbles, Ultrasound Med. Biol. 29 (2003) 687–694. [14] N. De Jong, R. Cornet, C.T. Lancee, Higher harmonics of vibrating gas-filled microspheres. Part one: simulations, Ultrasonics 32 (1994) 447–453. [15] P.A. Dayton, K.E. Morgan, A.L. Klibanov, G.H. Brandenburger, K.W. Ferrara, Optical and acoustical observations of the effects of ultrasound on contrast agents, IEEE Trans. UFFC 46 (1999) 220–232. [16] K.D. Jandt, Atomic force microscopy of biomaterials surfaces and interfaces, Surf. Sci. 491 (2001) 303–332. [17] Z. Reich, R. Kapon, R. Nevo, Y. Pilpel, S. Zmora, Y. Scolnik, Scanning force microscopy in the applied biological sciences, Biotechnol. Adv. 19 (2001) 451–485. [18] V. Sboros, E. Glynos, S.D. Pye, C.M. Moran, M. Butler, J. Ross, R. Short, W.N. McDicken, V. Koutsos, Nano-interrogation of ultrasonic contrast agent microbubbles using atomic force microscopy, Ultrasound Med. Biol. 32 (2006) 579–585. [19] J.H. Hoh, J.P. Cleveland, C.B. Prater, J.-P. Revel, P.K. Hansma, Quantized adhesion forces detected with the atomic force microscope, Mater. Sci. Eng. C 4 (1997) 233–240. [20] H.J. Butt, M. Jaschke, Calculation of thermal noise in atomic force microscopy, Nanotechnology 6 (1995) 1–7. [21] S.C. Heffelfinger, H.H. Hawkins, J. Barrish, L. Taylor, G.J. Darlington, SK HEP-1: A human cell line of endothelial origin, In Vitro Cell. Dev. Biol. 28A (1992) 136–142.
354
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
[22] N.M. Masson, I.S. Currie, J.D. Terrace, O.J. Garden, R.W. Parks, J.A. Ross, Hepatic progenitor cells in human foetal liver express the oval cell marker Thy-1, Am. J. Physiol. (Gastroint. Liv. Physiol.) 291 (2006) G45–G54. [23] C.M. Moran, J.A. Ross, C. Cunningham, M. Butler, T. Anderson, D. Newby, K.A.A. Fox, W.N. McDicken, Manufacture and acoustical characterisation of a high-frequency contrast agent for targeting applications, Ultrasound Med. Biol. 32 (2006) 421–428. [24] V.V. Lulevich, D. Andrienko, O.I. Vinogradova, Elasticity of polyelectrolyte multilayer microcapsules, J. Chem. Phys. 120 (2004) 3822–3826. [25] T.B. Ottoboni, E.G. Tickner, R.E. Short, R.K. Yamamoto, Hollow microspheres with controlled fragility for medical use, US patent 6776761 August 17, 2004. [26] S. Cuenot, S. Demoustier-Champagne, B. Nysten, Phys. Rev. Lett. 85 (2000) 1690.
[27] V. Sboros, S.D. Pye, C.A. MacDonald, J. Gomatam, C.M. Moran, W.N. McDicken, Absolute measurement of ultrasonic backscatter from single microbubbles, Ultrasound Med. Biol. 31 (2005) 1063–1072. [28] N. De Jong, P.J.A. Frinking, A. Bouakaz, M. Goorden, T. Schourmans, X. Jingping, F. Mastik, Optical imaging of contrast agent microbubbles in an ultrasound field with a 100-MHz camera, Ultrasound Med. Biol. 26 (2000) 487–492. [29] X. Zhang, E. Wojcikiexicz, V.T. Moy, Force spectroscopy of the leukocyte function-associated antigen-1/intercellular adhesion molecule-1 interaction, Biophys. J. 83 (2002) 2270–2279. [30] P. Hinterdorfer, W. Baumgartner, H.J. Gruber, K. Schilcher, H. Schindler, Detection and localization of individual antibody–antigen recognition events by atomic force microscopy, Proc. Natl. Acad. Sci. USA 93 (1996) 3477–3481.
Ultrasonics 46 (2007) 349–354 www.elsevier.com/locate/ultras
Nanomechanical probing of microbubbles using the atomic force microscope V. Sboros
b
a,*
, E. Glynos b, S.D. Pye c, C.M. Moran a, M. Butler a, J.A. Ross d, W.N. McDicken a, V. Koutsos b
a Medical Physics, School of Clinical Sciences and Community Health, University of Edinburgh, Edinburgh, UK Institute for Materials and Processes, Centre for Materials Science and Engineering, School of Engineering and Electronics, University of Edinburgh, Edinburgh, UK c Medical Physics, Royal Infirmary of Edinburgh, Edinburgh, UK d Clinical and Surgical Sciences, University of Edinburgh, Edinburgh, UK
Received 27 April 2007; received in revised form 18 June 2007; accepted 20 June 2007 Available online 30 June 2007
Abstract Atomic force microscopy (AFM) is a versatile mechanical nanosensor that can be used to quantify the mechanical properties of microbubbles (MBs) and the adhesion mechanisms of targeted MBs. Mechanical properties were investigated using AFM tipless cantilevers to microcompress the MBs. The range of compressive stiffness for biSphere was found to be between 1 and 10 N m1 using a cantilever with a spring constant of 0.6 N m1. This stiffness was shown to decrease with the MB size in a non-linear fashion. It is also possible to calculate a theoretical Young’s modulus of the shell. The adhesion properties of targeted lipid based MBs that use avidin–biotin chemistry for the attachment of targeting ligands were also studied. The MBs were attached to poly-L-lysine treated tipless cantilevers with spring constants ranging from 0.03 to 0.1 N m1. This system interrogated individual cells with pulling cantilever distance of 15 lm, and scan rate at 0.2 Hz. The depth of contact was not larger than 0.4 lm. The targeted MBs provided a significantly larger adhesion to the cells compared to control ones. Average adhesion force was dependent on depth of contact. Analysis of the data demonstrated a single distribution of adhesion events with median at 89 pN, which is in agreement with the literature for such interactions. The nanointerrogation of MBs using AFM provides new insight into their mechanical properties, and should be of assistance to MB design and manufacture. 2007 Elsevier B.V. All rights reserved.
1. Introduction As has been shown in a number of areas of pathology MBs are strong candidates as ultrasonic contrast agents used to quantify perfusion non-invasively [1]. The development of site-targeted MBs expands the field beyond perfusion imaging and allows attachment to specific markers of disease [1]. The expansion of MB technology to include localised drug and gene delivery is potentially very valuable
*
Corresponding author. Tel.: +44 131 2426292; fax: +44 131 2426314. E-mail address: [email protected] (V. Sboros).
0041-624X/$ - see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ultras.2007.06.004
[2,3]. The complex interaction that MBs have with ultrasound [4] is however, often overlooked and has yet to be theoretically modelled successfully. New imaging modalities that exploit MB destruction [5–8], and reject background tissue signals [9–11] are incorporated in ultrasonic equipment, but are based on a simple model of MB properties. As perfusion imaging deals with the smallest of vessels and pathology often deals with subtle changes in vessel numbers, a perfusion imaging tool needs to be very sensitive thus requiring optimised transmitted beams for matching MBs and their appropriate signal processing. The success of this effort relies on a thorough understanding of the physical properties of MBs and their interaction with
350
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
the ultrasound field. Towards achieving this goal, fast optical microscopy has recorded an array of MB behaviours [12] and has proved to be of substantial value. However, the viscoelastic properties of the shell, which has typical thickness between 10 and 200 nm, have not been measured accurately but have been estimated and remain the most important unknown in MB behaviour. Fitting the shell parameters into MB theoretical models is a common practice in previous theoretical investigations. Other parameters responsible for MB motion have been studied systematically and validated experimentally. However, the fitting approach used for the shell is explorative, and perhaps the main reason for the limited success of these models. Knowledge of the mechanical properties of the shell will elucidate its contribution to the MB motion in the presence of ultrasound. It is known that MBs do not scatter in a uniform manner and at low acoustic pressures only a small fraction of the MB population scatter efficiently [13]. Knowledge of the shell property may also offer an explanation for this non-uniformity. It is important to note that optical microscopy is limited in such measurements, primarily due to spatial resolution restrictions. In the past, MB models either assumed a fixed behaviour for the shell properties [14] or used a best fit to experimental data [15]. Today a possible solution is provided by the atomic force microscope (AFM), a versatile mechanical micro/nano sensor [16,17]. We have previously recorded topographies of the biSphere (Point Biomedical Corp., San Carlos, CA, USA) MB (Fig. 1) and showed that the measurement of the elastic properties of MBs is feasible [18]. In this paper we present new data that provide insight on the elastic properties of biSphere. In addition to determining the properties of the bubble shell, targeted MBs require a nanointerrogation technique that will provide a detailed analysis of the behaviour of tar-
geting mechanisms in the vicinity of target cells. AFM has been shown to provide adhesion measurements at the level of single hydrogen bonds [19]. The ability of targeted MBs to adhere to cells using a specific molecular interaction is investigated here. 2. Materials and methods 2.1. MB mechanical properties The methodology is similar to our previous investigation [18]. Force–distance (f–d) curves were acquired on biSphere MBs using the MFP-1D AFM system (Asylum Research, Santa Barbara, CA, USA) with 1 Hz scan rates and 3 lm ramp size. The MBs were mechanically interrogated with aluminium coated tipless cantilevers (NSC-12, MikroMash, Madrid, Spain) with spring constants around 0.6 N/m. Spring constant calibration was performed by means of thermal noise spectra in air [20] available on the MFP-1D. The MB effective spring stiffness was assessed by measuring the deflection of the AFM cantilever during compression. This deflection depends on the elastic property of the MB and the cantilever. In order to calculate the deformation of the MB the deflection of the cantilever needs to be subtracted. s ¼ d F =k c
ð1Þ
where s is the MB deformation during compression after initial contact between the cantilever and the MB, d is the cantilever deflection, F the force applied and kc the cantilever spring constant. Note that F/kc corresponds to the deflection of the cantilever at a hard surface at the same force F. Thus force–distance (or deflection) (f–d) measurements are converted to force–deformation (f–s) measurements. The AFM is mounted on an inverted optical microscope (TE2000U, Nikon UK Limited, Surrey, UK). Images of MBs under the cantilever can be viewed in reflection mode. A digital camera (Hamamatsu Orca-ER C4742-80, Hamamatsu Photonics, Hamamatsu city, Japan) is attached to the microscope and the images are transferred and processed on a PC using image analysis software (IPLab v3.7, BD Bioscience Bioimaging, Rockville, USA). Using a 60· objective it is possible to size a MB with half micron accuracy. 2.2. Adhesion properties of targeted MBs
Fig. 1. BiSphere MB topography. Methodology is described in (1817). The MB has 1.63 lm diameter. The outline of the cross section of the MB topography data fits a sphere. Note the big defect area, which reaches a maximum peak to peak roughness 0.66 lm. It can be seen that a different layer emerges inside this area. It is known that biSphere is made out of two separate layers, and the outer layer is typically a few nanometers. The gap presented here is much larger and reveals an inner layer. Structural studies of MBs may reveal important aspects of MB construction and should be of use to their design and manufacturing.
The hepatic endothelial cell line SkHep-1 was cultured using standard techniques on Petri dishes [21]. Half of the Petri dish surface was utilised for this purpose. Flow cytometry analysis was carried out using a Coulter XL flow cytometer, as described previously [22], and demonstrated relatively low expression of CD31 antigen on the surface of these cells (CD31– mean fluorescence intensity (mfi) = 0.454 compared with isotype control mfi = 0.245).
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
The manufacture of the in-house MBs has been described previously [23]. They are phospholipid-based and were used as controls. A modified process was used to produce targeted MBs, a biotinylated phospholipid (Avanti, Alabaster, USA) was incorporated into the lipid shell and linked to biotinylated CD31 antibody (Abcam, Cambridge, UK) via a streptavidin bridge (Sigma, Poole, UK). Tipless cantilevers (CSC-12Al, MikroMash, Madrid, Spain) of typical spring constants ranging between 0.03 and 0.1 N/m were treated with 1/10 poly-L-lysine solution for 5 min, then rinsed and dried. The MB suspension was placed in the half of the Petri dish not occupied by the cultured cells. The cantilever approached the MBs manually until contact was reached, the MB was trapped between the cantilever and the bottom of the dish, and a slight deflection was registered. Twenty seconds were allowed to pass before the cantilever was retracted. The MB attachment was confirmed by subsequent microscopic inspection. The cultured cells were then placed under the cantilever/ MB probe. F–d curves were acquired using the MFP-1D AFM (Asylum Research, Santa Barbara, CA, USA) and were converted into f–s curves as for the mechanical properties experiment. The scan rate of 0.2 Hz was used throughout the adhesion experiments. 3. Results and discussion 3.1. MB mechanical properties A typical f–s curve is displayed in Fig. 2a. Upon the approach (0.30–0.10 lm distance) of the cantilever to the
160
351
biSphere MB, the deflection is zero as there are no forces which would impose a deflection. When the cantilever contacts the MB at a distance of 0.10 lm it deflects and this is recorded in Fig. 2a as positive force. When the force reaches a maximum (here 168 nN), the direction of the cantilever changes and the force reduces until eventually the cantilever detaches from the MB. The MB stiffness was calculated from the compression curve after the short non-linear part (10 nN) immediately after contact. Only MBs with reproducible f–s curves were included in this study. This was always the case for biSphere and there was no change in the stiffness measurement over 100 compressions. The calculations used the section of the curve below 200 nm MB deformation, and it was found that eight f–s curves are adequate to provide an assessment of the uncertainty of the measurement. The stiffness of the MB in Fig. 2a was 2.20 ± 0.06 N m1. The stiffness for different MB sizes are plotted in Fig. 2b, and display a monotonic stiffness decrease with size. Further analysis may employ the use of a mechanical model in order to assign an elastic modulus to MB shell material [24]: p F ¼ pffiffiffi Eh2 e1=2 þ 4pER0 he3 ð2Þ 2 2 where F is the force applied to the MB, E the MB shell elastic modulus, h the thickness of the shell, R0 the radius of the MB at rest, and e the relative deformation (=d/2R0 where d is the MB deformation). An approximate thickness is given from the manufacturer to be proportional to the radius and approximately 7.5 nm for 1 lm diameter [25]. The first term of the right side of the equation accounts for the bending of the finite thickness of the shell, which is important at small deformations (<5%), while the second term accounts for the elasticity of the whole shell and gains significance at larger deformation. The Young modulus E
140
100 80 60 40 20 0 -200.00
0.05
0.10
0.15
0.20
0.25
0.30
Deformation (micrometers)
Fig. 2a. Typical f–s curve for a 4.7 lm diameter biSphere MB. Approach (blue curve) and retract (red curve) show some hysteresis maybe due to the viscous behaviour of the MB. Upon contact there is a degree of noise most probably created due to the hydrodynamics of the water-immersed cantilever–MB system at the specific scan rate (1 Hz). The measurement of the MB stiffness was made within contact part of the approach curve and was found to be 2.2 N/m. The fitted curve (dashed line) to the contact part of the approach curve, is created by fitting a mechanical model to the results in order to extract a value for the Young modulus of the shell (see equation and discussion in Section 3). (For interpretation of the references in colour in this figure legend, the reader is referred to the web version of this article.)
Microbubble stiffness (N/m)
Force (nN)
120 8 7 6 5 4 3 2 1 0 2
3
4
5
6
Diameter (micrometers)
Fig. 2b. BiSphere MB stiffness vs. diameter as captured using a cantilever with 0.6 N/m stiffness. The average stiffness is calculated over eight curves for each microbubble. The standard deviation of the stiffness measurement was generally below 5%, while the diameter measurement was made with 60· objective and provided a standard deviation of 0.3 lm. A monotonic decrease of stiffness with size increase is apparent in this graph. Moreover, MBs with similar size often have significantly different stiffness.
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
3.2. Adhesion properties of targeted MBs Control and targeted MB interrogation of SkHep-1 cells are shown in Fig. 3a and b, respectively. The curves lack
30 25 20
Force (nN)
is proportional to the applied force and can be deduced with a simple fit. From the data of Fig. 2a, E = 1.35 GPa can increase to 10 GPa for smaller sized MBs. Our deformation is less than 5% and the second term of the above equation has a small contribution. The fit of the force curve (dashed line in Fig. 2a) presents another problem, however. The equation above is designed mainly for soft materials and relatively large deformations and as a result has been used and tried in the regime of dominance of the second term [24]. This means that the approximation made for small deformations, which is more relevant for the data presented here, may be different. Further the biSphere has a softer second outer shell layer [25] which may affect the appearance of the curve. Further work is required to address this complex problem. It is proposed that the present calculations of the elastic modulus may have some qualitative value and be assigned to the inner structural layer of the shell assuming that this is the hardest layer. Another important aspect of the data presented in Fig. 2b is the monotonic increase in MB stiffness with the decrease in diameter and consequently with the shell thickness. This indicates a size effect on the measured mechanical properties of the microbubble shell. The observed behaviour of E, as derived from Eq. (2), is similar to Fig. 2b and to what has been previously observed in other nanosystems [26]: the elastic modulus increased non-linearly and significantly beyond the value expected from macroscopic measurements of the bulk material when the characteristic dimensions of the interrogated object obtained very small values. Further analysis of MB shell thickness may be required, for example by electron microscopy, to establish the dependence of thickness and structure on MB size. Among the MBs presented in Fig. 2b, there are several of similar size but with significantly different stiffness, which is evidence of mechanical/structural dispersion in the shell manufacture. The force data presented here refer to MB mechanical properties in the static regime. MBs are expected to have different viscoelastic properties when exposed to ultrasonic waves. However, it is very important to connect structural characteristics and mechanical behaviour at the relatively simple and accessible AFM low frequency regime. The usefulness of the current data lies in making this connection possible and in this way feedback can be generated for the manufacturing processes. E.g. it is apparent from the dispersion in our data that there is a large variation in the shell characteristics (thickness, local curvature, nanostructure, defects) for microbubbles of the same diameter. MB models have assumed uniform shell characteristics in the absence of such data. The present experimental set-up offers a tool that may provide a useful link between manufacturing and acoustic [13,27] or optical [28] experiments.
15 10 5 0 1
0 -5
b
2
3
5
4
6
Deformation (micrometers)
2
1
Force (nN)
352
0 0
2
4
6
8
10
12
14
16
18
-1
-2
-3
Deformation (micrometers)
Fig. 3. (a) F–s curve of an in-house phospholipids MB with an SkHep-1 cell. Approach (blue curve) and retract (red curve) provide some hysterisis due to the viscosity of both MB and cell. Hardly any adhesion is observed when the MB/cantilever system retracts and detaches from the cell. (b) F–s curve of an in-house targeted MB with CD31 with an SkHep-1 cell. A significant adhesion is registered when the MB/cantilever system retracts. The cumulative adhesion is 2.06 nN while 16 detachment events are recorded that belong to the 1.52 nN of the force. (For interpretation of the references in colour in this figure legend, the reader is referred to the web version of this article.)
the low frequency noise observed in Fig. 2a. This may be attributed to the difference in scanning rate, geometry and mechanical properties of the structures involved, further investigation would be required to confirm this. In both control and targeted MBs the approach and retraction contact areas do not coincide (i.e. positive forces do not occur at the same deformations), which may be attributed to the cell or MB viscous properties. Overall 62 curves were acquired in five control MB–cell pairs; 15% of the f–s curves recorded some adhesion. For the targeted MBs, 95 curves were acquired in eight MB–cell pairs. The adhesion was apparent in 85% of the curves, and was significantly greater than the adhesion observed in the control experiments. Different MB–cell pairs provided often significantly different cumulative adhesion. This difference may be due to changes in the depth of contact, cell life/condition and to the variability in the manufacture of targeted MBs. The shape of Fig. 3b is typical and records 16 vertical jumps that are single detachment events and account for 1.52 nN of the adhesion developed between the MB and the cell during this contact. The cumulative adhesion was
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
2.06 nN, which means that 0.54 nN of adhesion could not be differentiated as single de-adhesion events, and most probably are below the spatial and force resolution limits for the AFM under the specific cantilever and settings. Between unbinding events horizontal plateaux were nearly always observed. The force jumps may be associated with the force required to detach the MB and the cell in a specific adhesion site. Each site may develop one or more bonds. Analysing all the available adhesion jumps we observed that the majority of the bonds (>80%) lie between 50 pN and 150 pN and the median value was 89 pN. A full statistical analysis requires more measurements. These will establish the behaviour of events and conclude as to whether we observe a single or multiple distributions of individual bonds. The recorded events may not be specific to CD31 interactions, as there are several proteins on the cell surface that are likely to bind with the CD31 antibody. However, work on similar interactions such as the LFA-1/ICAM-1 interaction shows dominant unbinding forces around 100 pN [29]. Further work should investigate the specificity of the interactions measured here, and the ability of the targeted MBs to target specific cells. A larger number of f–d curves like the one presented in Fig. 3b may provide enough data for analysing the distance between two unbinding events [30] and may provide useful information on the spatial distribution of targeting sites on the MB as well as the function of targeting. 3.3. Summary Little is known about the MB shell as (a) an elastic body that is a major contributor to MB motion in an ultrasonic field, or (b) as a structure on which targeting sites may be grafted. We have demonstrated that both these aspects can be characterised using atomic force microscopy to measure: (a) the elastic properties of the biSphere shell and (b) the adhesion of targeted in-house MBs with an endothelial cell line. The techniques described here provide unique information about the physical and biological properties of MBs, and have the potential for further development. Acknowledgements The authors acknowledge the generous offering of cell culturing facilities by Dr. Alistair Elfick. Thanks to Carrie Cunningham, James Black and Ian Ansell for all their work throughout this project. Also, special thanks to Bob Short (Point Biomedical Corporation, San Carlos, CA, USA) for the valuable information and guidance throughout this work. The work was funded by the Engineering and Physical Sciences Research Council, UK (Project Grant 21639/01). References [1] J.R. Lindner, Microbubbles in medical imaging: current applications and future directions, Nat. Rev. 3 (2004) 527–532.
353
[2] R.V. Shohet, S. Chen, Y.T. Zhou, Z. Wang, R.S. Meidell, R.H. Unger, P.A. Grayburn, Echocardiographic destruction of albumin microbubbles directs gene delivery to the myocardium, Circulation 101 (2000) 2554–2556. [3] E.C. Unger, E. Hersh, M. Vannan, T.O. Matsunaga, T. McCreery, Local drug and gene delivery through microbubbles, Prog. Cardiov. Dis. 44 (2001) 45–54. [4] V. Sboros, K.V. Ramnarine, C.M. Moran, S.D. Pye, W.N. McDicken, Understanding the limitations of ultrasonic backscatter measurements from microbubble populations, Phys. Med. Biol. 47 (2002) 4287–4299. [5] T.R. Porter, F. Xie, Transient myocardial contrast after initial exposure to diagnostic ultrasound pressures with minute doses of intravenously injected microbubbles, Circulation 92 (1995) 2391–2395. [6] T.R. Porter, F. Xie, D. Kricsfeld, R.W. Armbruster, Improved myocardial contrast with second harmonic transient ultrasound response imaging in humans using intravenous perfluorocarbonexposed sonicated dextrose albumin, J. Am. Coll. Cardiol. 27 (1996) 1497–1501. [7] V. Uhlendorf, C. Hoffmann, Nonlinear acoustical response of coated microbubbles in diagnostic ultrasound, IEEE Ultrasonics Symp. 3 (1994) 1559–1562. [8] A. Bauer, R. Schlief, M. Zomack, A. Urbank, H.P. Niendorf, in: N.C. Nanda, R. Schlief, B.B. Goldberg (Eds.), Advances in Echo Imaging using Contrast Enhancement, Kluwer Academic Publisher, Dodrecht, The Netherlands, 1997. [9] D.H. Simpson, C.T. Chin, P.N. Burns, Pulse inversion doppler: a new method for detecting nonlinear echoes from microbubble contrast agents, IEEE Trans. UFFC 46 (1999) 372–382. [10] V. Mor-Avi, E.G. Caiani, K.A. Collins, C.E. Korcarz, J.E. Bednarz, R.M. Lang, Combined assessment of myocardial perfusion and regional left ventricular function by analysis of contrast-enhanced power modulation images, Circulation 104 (2001) 352–357. [11] W. Wilkening, B. Brendel, H. Jiang, J. Lazenby, H. Ermert, Optimized receive filters and phase-coded pulse sequences for contrast agent and nonlinear imaging, IEEE Ultrasonics Symp. (2001) 1733–1737. [12] M. Postema, A. van Wamel, C.T. Lancee, N. de Jong, Ultrasoundinduced encapsulated microbubble phenomena, Ultrasound Med. Biol. 30 (2004) 827–840. [13] V. Sboros, C.M. Moran, S.D. Pye, W.N. McDicken, The behaviour of individual contrast agent microbubbles, Ultrasound Med. Biol. 29 (2003) 687–694. [14] N. De Jong, R. Cornet, C.T. Lancee, Higher harmonics of vibrating gas-filled microspheres. Part one: simulations, Ultrasonics 32 (1994) 447–453. [15] P.A. Dayton, K.E. Morgan, A.L. Klibanov, G.H. Brandenburger, K.W. Ferrara, Optical and acoustical observations of the effects of ultrasound on contrast agents, IEEE Trans. UFFC 46 (1999) 220–232. [16] K.D. Jandt, Atomic force microscopy of biomaterials surfaces and interfaces, Surf. Sci. 491 (2001) 303–332. [17] Z. Reich, R. Kapon, R. Nevo, Y. Pilpel, S. Zmora, Y. Scolnik, Scanning force microscopy in the applied biological sciences, Biotechnol. Adv. 19 (2001) 451–485. [18] V. Sboros, E. Glynos, S.D. Pye, C.M. Moran, M. Butler, J. Ross, R. Short, W.N. McDicken, V. Koutsos, Nano-interrogation of ultrasonic contrast agent microbubbles using atomic force microscopy, Ultrasound Med. Biol. 32 (2006) 579–585. [19] J.H. Hoh, J.P. Cleveland, C.B. Prater, J.-P. Revel, P.K. Hansma, Quantized adhesion forces detected with the atomic force microscope, Mater. Sci. Eng. C 4 (1997) 233–240. [20] H.J. Butt, M. Jaschke, Calculation of thermal noise in atomic force microscopy, Nanotechnology 6 (1995) 1–7. [21] S.C. Heffelfinger, H.H. Hawkins, J. Barrish, L. Taylor, G.J. Darlington, SK HEP-1: A human cell line of endothelial origin, In Vitro Cell. Dev. Biol. 28A (1992) 136–142.
354
V. Sboros et al. / Ultrasonics 46 (2007) 349–354
[22] N.M. Masson, I.S. Currie, J.D. Terrace, O.J. Garden, R.W. Parks, J.A. Ross, Hepatic progenitor cells in human foetal liver express the oval cell marker Thy-1, Am. J. Physiol. (Gastroint. Liv. Physiol.) 291 (2006) G45–G54. [23] C.M. Moran, J.A. Ross, C. Cunningham, M. Butler, T. Anderson, D. Newby, K.A.A. Fox, W.N. McDicken, Manufacture and acoustical characterisation of a high-frequency contrast agent for targeting applications, Ultrasound Med. Biol. 32 (2006) 421–428. [24] V.V. Lulevich, D. Andrienko, O.I. Vinogradova, Elasticity of polyelectrolyte multilayer microcapsules, J. Chem. Phys. 120 (2004) 3822–3826. [25] T.B. Ottoboni, E.G. Tickner, R.E. Short, R.K. Yamamoto, Hollow microspheres with controlled fragility for medical use, US patent 6776761 August 17, 2004. [26] S. Cuenot, S. Demoustier-Champagne, B. Nysten, Phys. Rev. Lett. 85 (2000) 1690.
[27] V. Sboros, S.D. Pye, C.A. MacDonald, J. Gomatam, C.M. Moran, W.N. McDicken, Absolute measurement of ultrasonic backscatter from single microbubbles, Ultrasound Med. Biol. 31 (2005) 1063–1072. [28] N. De Jong, P.J.A. Frinking, A. Bouakaz, M. Goorden, T. Schourmans, X. Jingping, F. Mastik, Optical imaging of contrast agent microbubbles in an ultrasound field with a 100-MHz camera, Ultrasound Med. Biol. 26 (2000) 487–492. [29] X. Zhang, E. Wojcikiexicz, V.T. Moy, Force spectroscopy of the leukocyte function-associated antigen-1/intercellular adhesion molecule-1 interaction, Biophys. J. 83 (2002) 2270–2279. [30] P. Hinterdorfer, W. Baumgartner, H.J. Gruber, K. Schilcher, H. Schindler, Detection and localization of individual antibody–antigen recognition events by atomic force microscopy, Proc. Natl. Acad. Sci. USA 93 (1996) 3477–3481.