Growing Opportunities in Biomedical Sciences
- Fig. 1: Main elements of a bioAFM. Reprinted from Micros. Res. Tech. 80(1) 75-84 (2016) doi: 10.1002/jemt.22776.
- Fig. 2: Illustrative force-indentation curve showing recorded force as the cantilever is ramped towards (black) and away from (blue) the sample. This first part appears as a flat line in the force curve, because the tip is still too far away from the sample to experience any interaction force (x axis plotted as ‘distance’).The tip is considered to be “in contact” with the sample when repulsive forces are first observed (dashed vertical line). Cantilever movement proceeds until a preset maximum force is reached (x axis plotted as indentation δ), and then the direction of travel is reversed and the cantilever is moved away from the sample. The process described is typically performed continuously as a loop. From such force-indentation curves, mechanical parameters such as Young’s modulus (E), viscosity (η) or adhesion can be computed. Modified from Micros. Res. Tech. 80(1) 75-84 (2016) doi: 10.1002/jemt.22776.
Atomic Force Microscopes (AFM), once typically found in the basements of physics’ schools, are now beginning to populate life sciences research centers. Undeniably, the rise of mechanobiology research has found an ideal partner in AFM, to achieve not only high-resolution imaging of living samples, but also mechanical characterization of cells and tissues. Now that the study of cellular mechanical state is gaining relevance also in biomedical research, AFM is poised to make further contributions in the study of (patho)physiology and disease.
Deceitfully Simple in its Design, Surprisingly Versatile in its Capabilities
The basics of AFM operation may seem too simple to be true, but AFM is, at its core, a feat of engineering, electronics and nanofabrication. AFM is based on measuring the attractive and repulsive forces acting between a sharp tip and a sample’s surface. The tip (nm-scale radius) is attached to a flexible cantilever, which bends when forces are present (fig 1). Cantilevers are microscopic (µm scale) and are etched at the side of silicon nitride chips. The chip, which is macroscopic (mm scale), is firmly attached to a set of piezoelectric elements, which allow ultra‐precise positioning of the tip with respect to the sample. Bending of the cantilever is detected by a laser light, which reflects on the cantilever and is then detected by a 4-quadrant photodetector. The difference between the photovoltages output by the quadrants is used to measure precisely cantilever bending, and by knowing the cantilever’s stiffness, one can monitor in real-time the forces acting between tip and sample.
AFM was first proposed as an alternative to Scanning Tunneling Microscopy, to broaden the range of experimental conditions and measurable samples. This premise remains, and AFM now appeals to materials scientists and clinicians alike, being an enabling technique from high-resolution nanopatterning of materials to mechanical characterization of human biopsies. It would be naive to think a single system can perform optimally in such dissimilar tasks. Current AFMs (as well as the cantilevers used as probes) display a large degree of specialization.
This ranges from short-range ultrafast piezoelectrics that drive stiff cantilevers with very large resonance frequencies and nanometer-scale tips, to longer-range but slower piezoelectrics and soft cantilevers with microscale colloidal probes for the mechanical characterization of living cells. Respectively, two example outputs would be a video-rate sequence displaying the conformational changes of a protein , or a timeline recording the stiffening of a cell stimulated with a contractile agonist .
A less noted feature of AFM is its ‘lean’ design: only the tip of the cantilever needs to be in close proximity to the sample , while the piezoelectrics, laser and photodetector are housed away from the sample. Given the whole half-space under the sample is not occupied; AFMs fit seamlessly on the stage of optical inverted microscopes  and may even perform in situ measurements . Using the space to the sides of the sample has been also explored, with custom-made and commercial systems allowing side-view optical/fluorescence imaging of the probed sample .
Beyond Topographical Imaging and into Mechanical Characterization of Biological Samples
Being developed as a high-resolution imaging technique, AFMs initial output was the topography of the sample’s surface, with attention to minimizing the mechanical interaction between tip and sample. In stark contrast, pioneering work during the 90s proposed to use an AFM tip as a force-controlled microindentor that would allow measuring the mechanical properties of the probed sample [7-9]. Focusing on the study of living cells, the principles of operation evolved, giving priority to large contact areas (several µm2), large interaction forces (several nN), micron-scale ranges for sample indentation and cantilever deflection, and blunted tip geometries. These were achieved at the expense of spatial resolution and speed. To estimate mechanical properties, the cantilever is slowly ramped towards/away from the sample in the normal direction, while force-indentation curves are being recorded. The key elements of a force-indentation curve and how they reflect different mechanical phenomena are illustrated in figure 2. When AFM manufacturers incorporated this approach, force-volume mode was born. Its use, as well as studies using custom-built AFMs, lead to significant contributions in the then emerging field of mechanobiology. Unfortunately, AFM-based nanomechanics was considered to be low throughput, reliable only on thick areas around the cell nucleus and poised to produce relative estimates of cellular stiffness. These, together with the poor user-friendliness of early AFM systems and the cumbersome integration with optical microscopes, limited and delayed broad adoption of AFM nanomechanics for cell biology and biomedical research.
The Three Cornerstones of AFM-based Mechanical Characterization of Cells and Tissues
The current success of AFM nanomechanics reflects efforts by the bioAFM community to move beyond its traditional limitations. The last decade has seen simultaneous improvements in AFM systems, contact mechanics models, and automated calibration and data analysis pipelines.
The new force-feedback modes for AFM nanotopography and mechanics are based on recording force-displacement curves at relatively high oscillatory rates (~1kHz) and with high spatial resolution, to simultaneously map multiple properties of the probed sample. The approach can be said to combine the higher speed and resolution of tapping mode, with the data recording capabilities associated with force-volume mode. Acquiring and saving the whole force-displacement curve for each pixel provides finer control of the tip-sample interaction during the experiment, and allows a plethora of mechanical parameters to be computed afterwards via post-processing. Typically, new AFM systems incorporate heating and liquid perfusion capabilities, work on top of inverted fluorescence microscopes, and provide experiment planner pipelines that enable multiple cells to be probed, or the overlay of fluorescence images with mechanotopography maps. Efforts to integrate with advanced super-resolution methods are ongoing [10, 11].
Now, getting more data doesn’t necessarily mean getting better data. The advent of force-feedback mode was preceded by efforts to develop contact mechanics models better suited to the mechanical nature of living cells. Cells are viscoelastic, heterogeneous along their depth, and typically cultured over stiff substrates. A variety of models has been proposed to tackle each of these mechanical features. Among them, some are of particular interest, because they can be directly applied to standard force-indentation curves without need for protocol modifications. Their combined application allows computing several distinct mechanical parameters for each pixel or probed cell, paving the way for single-cell multiplex mechanical characterization. Briefly, models can be found to compute Young’s modulus correcting for the hard substrate effect , cell viscosity [13, 14], depth-dependent Young’s modulus  or cell adhesion . With small protocol modifications, models have been proposed to compute other sets of multiplex mechanical parameters, such as poroelasticity , complex shear modulus  or cortical tension and intracellular pressure .
Thirdly, the vast amount of force-indentation curves generated for each mapped cell and the fact that we can now readily probe tens of cells per petri dish requires pipelines for automated data analysis. Similarly, if we are to compare results between different experiments or even between different laboratories, a fair amount of standardization and proper calibrations are necessary. Recent studies have tackled calibration issues stemming from the inaccurate determination of the photodetector’s deflection sensitivity  and the cantilever tilt . Lastly, other studies have focused on crucial bottlenecks in data analysis. In particular, strategies for accurate determination of the contact point (the exact location where the tip gets in contact with the sample) have been outlined , and even an open source package is available that aims to combine full automation and model versatility .
Where Now for BioAFM in Biomedical Sciences?
Being established as a relevant theme in basic biological research, mechanobiology is now moving into the field of predictive biomedicine, with recent studies showing that cell and tissue mechanical properties can be used as biomarkers to predict cellular age or cancer prognosis [24, 25]. Further contributions in disease diagnostic, drug screening and personalized medicine may be made with the advent of machine learning algorithms and big data in general. If AFM is to remain an enabling tool, the bioAFM community faces renewed challenges to turn AFM into a more multiplex and high-throughput technique, to further its impact in life sciences and biomedicine. This appears an attainable endeavor, given the momentum gathered in the last years among the bioAFM community , together with the increasing interdisciplinarity shown by AFM users.
The author acknowledges funding from the European Commission (PCIG14-GA-2013-631011 CSKFingerprints), BBSRC (BB/P006108/1) and Dunhill Medical Trust (R454/1115).
1School of Engineering and Materials Science, Queen Mary University of London, UK
 T. Uchihashi, et al., High-speed atomic force microscopy reveals rotary catalysis of rotorless F(1)-ATPase, Science 333 (6043) 755-8 (2011) doi: 10.1126/science.1205510
 C.M. Cuerrier et al., Effect of thrombin and bradykinin on endothelial cell mechanical properties monitored through membrane deformation, J. Mol. Recognit. 22(5) 389-96 (2009) doi: 10.1002/jmr.953
 D. Guan, et al., Noncontact Viscoelastic Imaging of Living Cells Using a Long-Needle Atomic Force Microscope with Dual-Frequency Modulation. Physical Review Applied,. 8(4) 044010 (2017) doi: 10.1103/PhysRevApplied.8.044010
 N. Gavara and R.S. Chadwick, Relationship between cell stiffness and stress fiber amount, assessed by simultaneous atomic force microscopy and live-cell fluorescence imaging. Biomechanics and modeling in mechanobiology, 1-13 (2015) doi: 10.1007/s10237-015-0706-9
 D.E. Koser, et al., Mechanosensing is critical for axon growth in the developing brain. Nat. Neurosci. 19(12) 1592-1598 (2016) doi: 10.1038/nn.4394
 K. Beicker, et al., Vertical Light Sheet Enhanced Side-View Imaging for AFM Cell Mechanics Studies, Scientific Reports 8(1) 1504 (2018) doi: 10.1038/s41598-018-19791-3
 J.H. Hoh, et al., Atomic force microscopy and dissection of gap junctions, Science 253(5026) 1405-8 (1991)
 A.L. Weisenhorn, et al., Deformation and height anomaly of soft surfaces studied with an AFM, Nanotechnology 4(2) 106 (1993)
 M. Radmacher, et al., Measuring the viscoelastic properties of human platelets with the atomic force microscope, Biophysical Journal 70(1) 556-567. (1996) doi: 10.1016/S0006-3495(96)79602-9
 J.V. Chacko, F.C. Zanacchi, and A. Diaspro, Probing cytoskeletal structures by coupling optical superresolution and AFM techniques for a correlative approach, Cytoskeleton (Hoboken) 70(11) 729-40. (2013) doi: 10.1002/cm.21139
 N. Curry, et al., Correlative STED and Atomic Force Microscopy on Live Astrocytes Reveals Plasticity of Cytoskeletal Structure and Membrane Physical Properties during Polarized Migration, Frontiers in Cellular Neuroscience 11, 104 (2017) doi: 10.3389/fncel.2017.00104
 N. Gavara, and R.S. Chadwick, Determination of the elastic moduli of thin samples and adherent cells using conical atomic force microscope tips, Nat. Nano 7(11) 733-736. (2012) doi: 10.1038/nnano.2012.163
 L.M. Rebelo, et al., Comparison of the viscoelastic properties of cells from different kidney cancer phenotypes measured with atomic force microscopy, Nanotechnology 24(5) 055102 (2013) doi: 10.1088/0957-4484/24/5/055102.
 Y.M. Efremov, et al., Measuring nanoscale viscoelastic parameters of cells directly from AFM force-displacement curves. Scientific Reports 7(1) 1541 (2017)
 J.R. Staunton, et al., Correlating confocal microscopy and atomic force indentation reveals metastatic cancer cells stiffen during invasion into collagen I matrices, Scientific Reports 6, 19686 (2016) doi: 10.1038/s41598-017-01784-3
 F. Rico, et al., Cell dynamic adhesion and elastic properties probed with cylindrical atomic force microscopy cantilever tips, J. Mol. Recognit. 20(6) 459-66. (2007)
 E. Moeendarbary, et al., The cytoplasm of living cells behaves as a poroelastic material, Nature materials 12(3) 253-261 (2013) doi: 10.1038/nmat3517
 R. Sunyer, et al., The temperature dependence of cell mechanics measured by atomic force microscopy, Physical biology 6(2) 025009-025009. (2009) doi: 10.1088/1478-3975/6/2/025009
 A.X. Cartagena-Rivera, et al., Actomyosin Cortical Mechanical Properties in Nonadherent Cells Determined by Atomic Force Microscopy, Biophysical Journal 110(11) 2528-2539 (2016) doi: 10.1016/j.bpj.2016.04.034
 H. Schillers, et al., Standardized Nanomechanical Atomic Force Microscopy Procedure (SNAP) for Measuring Soft and Biological Samples, Scientific Reports 7(1) 5117 (2017) doi: 10.1038/s41598-017-05383-0
 R.S. Gates, Experimental confirmation of the atomic force microscope cantilever stiffness tilt correction, Review of Scientific Instruments 88(12) 123710 (2017) doi: 10.1063/1.4986201
 N. Gavara, Combined strategies for optimal detection of the contact point in AFM force-indentation curves obtained on thin samples and adherent cells, Scientific reports 6 (2016) doi: 10.1038/srep21267
 P. Hermanowicz, et al., AtomicJ: an open source software for analysis of force curves. Rev. Sci. Instrum. 85(6) 063703. (2014) doi: 10.1063/1.4881683
 J.M. Phillip, et al., Biophysical and biomolecular determination of cellular age in humans, Nature Biomedical Engineering 1, 0093 (2017) doi: 10.1038/s41551-017-0093
 J. Zemla, et al., Atomic force microscopy as a tool for assessing the cellular elasticity and adhesiveness to identify cancer cells and tissues, Semin. Cell Dev. Biol. 73, 115-124 (2018) doi: 10.1016/j.semcdb.2017.06.029
 Susana R. Sousa, Pierre Parot, Jean‐Luc Pellequer, Seventh International AFMBioMed Conference on AFM in Life Sciences and Medicine, April 11 to 15, 2016, Porto, Portugal, Journal of Molecular Recognition 30(12) e2681 (2017) doi: 10.1002/jmr.2681
 N. Gavara, A Beginner's guide to atomic force microscopy probing for cell mechanics, Micros. Res. Tech. 80(1) 75-84 (2016) doi: 10.1002/jemt.22776