Micro fluorescence in situ hybridization (μFISH) for spatially multiplexed analysis of a cell monolayer
- 3.1k Downloads
We here present a micrometer-scale implementation of fluorescence in situ hybridization that we term μFISH. This μFISH implementation makes use of a non-contact scanning probe technology, namely, a microfluidic probe (MFP) that hydrodynamically shapes nanoliter volumes of liquid on a surface with micrometer resolution. By confining FISH probes at the tip of this microfabricated scanning probe, we locally exposed approximately 1000 selected MCF-7 cells of a monolayer to perform incubation of probes — the rate-limiting step in conventional FISH. This method is compatible with the standard workflow of conventional FISH, allows re-budgeting of the sample for various tests, and results in a ~ 15-fold reduction in probe consumption. The continuous flow of probes and shaping liquid on these selected cells resulted in a 120-fold reduction of the hybridization time compared with the standard protocol (3 min vs. 6 h) and efficient rinsing, thereby shortening the total FISH assay time for centromeric probes. We further demonstrated spatially multiplexed μFISH, enabling the use of spectrally equivalent probes for detailed and real-time analysis of a cell monolayer, which paves the way towards rapid and automated multiplexed FISH on standard cytological supports.
KeywordsFluorescence in situ hybridization Microfluidic probe Spatial multiplexing Microfluidics
Multiplexed FISH is often used to locate multiple targets in a cell nucleus simultaneously using probes of different colors, in order to study sparse cytological substrates (Lichter 1997). However, the number of spectrally distinct fluorophores for multiplexed FISH is limited. To visualize additional targets simultaneously, fluorescent probes of different fluorescence intensities were used to distinguish the complete set of human chromosomes in situ (Tanke et al. 1999). Quantum dots (QD) allow the resolution of even more colors as they have narrow emission bands with minimal spectral overlap (Pathak et al. 2001). However, QD are bulky (15–20 nm), and QD-DNA constructs have a lower mobility than free DNA (Ioannou et al. 2009) and are therefore seldom used for direct detection of targets in FISH (Zhang et al. 2013).
Despite the merits of FISH-based approaches, their widespread use for diagnostics is limited for several reasons. Conventional implementations of FISH comprise multiple steps, such as sample-specific enzymatic treatment, fixation, denaturation, manual pipetting of probes, incubation, post-hybridization washes to remove unbound probes, and endpoint visualization, Fig. 1. Those steps are mostly performed manually, are time consuming, and require experienced personnel. In addition, FISH remains an expensive test owing to the high price of probes. Further, expensive imaging platforms are necessary to visualize spectrally distinct probes for multiplexed FISH. To make FISH assays more pervasive in diagnostics, there is therefore a need to reduce the cost per test, introduce automation, and simplify the implementation of multiplexed analysis.
A few microfluidic platforms were developed for instance, to automate FISH analysis (Sieben et al. 2008; Tai et al. 2013) and to increase throughput (Sieben et al. 2007). Within such platforms, interphase FISH has been integrated with a range of analytical techniques, such as flow cytometry (Liu et al. 2011), chemistrode (Liu et al. 2009), Förster resonance energy transfer (FRET) (Packard et al. 2011) and immunostaining (Zhang et al. 2006). Microfluidics-based devices have also been used to analyze specific chromosomal translocations within spread metaphases on slides (Vedarethinam et al. 2010; Shah et al. 2011). In the context of diagnostics, devices primarily comprising closed microfluidic channels or reservoirs have been developed to miniaturize FISH assays for prenatal diagnostics (Ho et al. 2012) and the analysis of circulating tumor cells (Lim et al. 2012; Mottet et al. 2014; Perez-Toralla et al. 2015; Gogoi et al. 2016) or cancer malignancies (Zanardi et al. 2010; Kurz et al. 2011; Mughal et al. 2014). All these implementations need cell immobilization within microfluidic systems or surface treatment to ensure cell adhesion. Other microfluidic implementations are suitable for cell suspensions (Zhang et al. 2006; Liu et al. 2009; Liu et al. 2011; Packard et al. 2011), but unsuitable for the analysis of cytological samples forming monolayers, such as adherent cells, fixed cytological samples and tissue sections. Working with such adherent biological substrates is vital as the retrieval of immobilized cells from their substrate prior to flowing them in closed channels may introduce nuclear damage, thereby impeding hybridization and subsequently the result itself. Towards using microfluidics for tissue section analysis, Histo Flex (Søe et al. 2011) implemented localized and multiplexed RNA FISH by placing an elastomeric lid patterned with microfluidic channels on tissue sections and then flowing reagents to perform local FISH on certain cells. However, this approach limits the flexibility in adapting to morphological variations between samples.
In this paper, we demonstrate an interactive and versatile implementation of μFISH on cell monolayers at the micrometer-scale using a microfluidic probe (MFP). The MFP is a non-contact, scanning probe technology that localizes nanoliter volumes of liquids on substrates using hydrodynamic flow confinement (HFC) (Kaigala et al. 2011). With this technology, we can precisely localize (bio)chemicals on standard biological substrates in the “open space”, i.e., without the need for inserting the biological entities into closed channels. Prior demonstrations with MFPs (Queval et al. 2010; Ainla et al. 2012; Sarkar et al. 2014) are micro-immunohistochemistry on tissue sections (Lovchik et al. 2012), cell inactivation (Kaigala et al. 2011), local lysis of cells for expression analysis and cell patterning (Kashyap et al. 2016, Rapid subtractive patterning of live cell layers with a microfluidic probe, unpublished), and high-quality protein patterning (Autebert et al. 2016). In this work, we specifically use a vertically-oriented probe head and hierarchical HFC (hHFC) to expose a cell monolayer to multiple processing liquids simultaneously (Autebert et al. 2014). The localization and scanning capabilities of this technology allow incubation of FISH probes on selected cells on a monolayer. We also perform spatially multiplexed μFISH, which cannot be performed using conventional FISH protocols as the latter entail covering the entire cell monolayer with the probe mix, Fig. 1. In this spatially multiplexed μFISH, we incubate spectrally equivalent probes on distinct areas of a cell monolayer, and visualize the FISH signals with a simplified detection system.
2 Materials and methods
2.1 MFP head and platform
The MFP platform comprises linear stages to control the position of the head relative to the substrate, syringe pumps, an environmental chamber, and a microscope, Fig. 2c. The head was mounted to the z-stage of the platform via a mounting plate and linked to the syringes (Hamilton, 1705 TLLX) with a connector (Dolomite, 8-ports circular connector), tubing and adapters (IDEX, Tygon). The stages are computer-controlled and motorized (Lang GmbH, Hüttenberg, Germany), with a sub-micrometer accuracy. The apex-to-surface distance as well as the x- and y-coordinates were monitored using the encoded position of the stage. Flow rates were set with Nemesys pumps (Cetoni GmbH, Korbussen, Germany) and Qmix software (Cetoni GmbH, Korbussen, Germany). 1000 μL and 250 μL syringes were filled with 1× saline sodium citrate (SSC) (from 20× SSC stock, 3.0 M NaCl, 0.3 M sodium citrate, pH 7.0, Thermo Fisher Scientific, Cat. No. 15557-044). An environmental chamber (Life Imaging Services GmbH, “The Cube and the Box”) enclosed the platform and the microscope, Supplemental Fig. S1.
2.2 Cell handling and preparation
MCF-7 cells (American Type Culture Collection (ATCC), HTB-22, breast adenocarcinoma) were cultured as recommended by ATCC. Cells were seeded in chamber slides (Thermo Fisher Scientific, Cat. No. 10717931) at a density of 105 –106 cells cm−2. After 2–3 days, the monolayer was washed with phosphate-buffered saline (PBS, pH 7.4, Thermo Fisher Scientific, Cat. No. 10010-023) and heat immobilized at 82 °C for 2 min. Next, the cells were rinsed with PBS and digested using pepsin at 37 °C for 8–10 min (Leica Biosystems, LK-101 A). Subsequently, they were rinsed with 2× SSC and fixed using Carnoy’s fixative (ethanol: acetic acid 3:1 (v/v). Sigma Aldrich, Cat. No. 02860 and 537010) at 4 °C for 40 min. Finally, the cells were dried for 5 min at room temperature (RT), washed with 2× SSC for 1 min twice, and then renatured in 2× SSC at 37 °C for 20 min.
2.3 Conventional FISH protocol
The compartments of the chamber slides were removed, and 10 μL FISH hybridization mix was deposited onto the cells (2 μL FISH probes in 8 μL FISH hybridization buffer). Buffer (KBI-FHB) and centromeric probes with a Platinum Bright550 dye (KBI-20017R and KBI-20007R) were purchased from Leica Biosystems. These cells with the probes were coverslipped (Menzel, Braunschweig, Germany). Probes and chromosomes were denatured at 75 °C for 5–10 min and then incubated at 37 °C in a dark, humidified chamber. The coverslip was removed, and non-specifically bound probes on the chamber slide were removed by washing twice with 0.1 % IGEPAL CA-630 in 2× SSC (v/v) for 1 min at RT and with 0.3 % IGEPAL CA-630 in 0.4× SSC (v/v) at 72 °C for 2 min (Sigma Aldrich, Cat. No. I8896). Subsequently, an additional wash was performed with 1× SSC at RT for 1 min. The cells were then mounted with mounting medium containing DAPI (Thermo Fisher Scientific, Cat. No. S36938) for inspection.
2.4 μFISH protocol
The compartments of the chamber slides were removed, the cells were immersed in 10 μL FISH buffer, sealed with a coverslip, and the chromosomes were denatured at 75 °C for 5–10 min in FISH hybridization buffer. 5 μL FISH hybridization mix (Conventional FISH protocol section) was denatured separately at 75 °C for 5 min in a PCR tube (VWR, Cat. No. 20170-012). Hoechst dye was added to the probes at a concentration of 0.2 μg mL−1 (Thermo Fisher Scientific, Cat. No. H3570). These probes were pipetted onto a sheet of parafilm and aspirated into the inner aperture (I2) of the head. Concurrently, the coverslip was removed from the chamber slide, and the cells were immersed in 1× SSC. This chamber slide was transferred to the sample holder of the MFP platform, and the head was positioned ~20 μm above the monolayer. The flow confinement was established using the flow rules defined in “MFP head and platform” section. The probes described above were injected from I2, and 1× SSC was injected from I1. Aspiration of probes and 1× SSC was performed from A1 and A2, respectively, Fig. 2b. After 10 min interaction of the hHFC with the cells (equivalent to 10 min incubation), Q i2 and Q a1 were stopped, and the cells were washed with 1× SSC flowing between the outer apertures for 2 min (Q i1 and Q a2). The head was positioned away from the slide before imaging. For multiplexing, the head was positioned away from substrate after the first probe hybridization wash, and 20 μL of 1× SSC was purged from A1 and I2, and 50 μL from A2 and I1, to avoid cross-contamination of the different probes used. Subsequently, the probe-loading procedure explained above was repeated.
2.5 Image acquisition and processing
Both the endpoint observation for conventional FISH and the real-time observation for μFISH were performed using an inverted microscope at 10×, 40× and 60× magnification (Nikon Eclipse Ti-E with objectives CFI Plan Fluor DLL 10×, ELWD 40× and ELWD 60×, respectively). An LED lamp (Sola, Lumencor) was used for illumination. Image acquisition was performed using an ORCA-flash 4.0 camera. For imaging, the NIS Elements Basic Research software (Nikon Instruments Europe, V.4.0) was used. Brightness and contrast adjustments of raw images as well as merging were done using the open-source FIJI (ImageJ) software (http://fiji.sc/Fiji), Supplemental Fig. S4.
3 Results and discussion
A specific challenge to implement μFISH with the MFP is the viscosity of the probe mix. The FISH probe mix is viscous, largely owing to the presence of dextran sulfate, which serves to increase the effective probe concentration and consequently improves the hybridization efficiency (Lederman et al. 1981). Localizing the viscous probe mix with the MFP has two important implications. First, the localization of a liquid at the apex of the head is due to the non-symmetric flow rates between the injection and aspiration apertures (Q a, tot ≥ 3 × Q i, tot) resulting in a continuous flow of the viscous processing liquid on the cell monolayer. This interaction of the localized viscous processing liquid exerts increased shear (shear increases linearly as a function of viscosity), which leads to delamination of the cell monolayer. To prevent this delamination of cells, we operated the MFP at 0.2 μL min−1, the lowest flow rate our fluidic system permitted. Second, at these low flow rates, we increased the injection and aspiration flow rates of the shaping liquid (Q i1, Q a2) to ensure that the viscous processing liquid remains confined. By operating the MFP under these conditions, we were able to confine the viscous processing liquid and to prevent cell delamination, even when incubation was performed for an extended duration of 2 h. We noted that diluting the probe mix with 1× SSC reduces the viscosity, but results in an increased incubation time.
A useful attribute of the MFP-based μFISH implementation is the ability to observe the FISH signal in real time. We observed centromere-specific FISH signals within 3 min of initiating the probe incubation step (Supplemental Fig. S3) with the MFP, which is approximately a 120-fold reduction in the incubation time recommended by the probe supplier (>6h). This 3 min incubation translates to a probe consumption of 0.6 μL with Q i2 of 0.2 μL min−1. Although surface coverage in μFISH is different from conventional FISH, up to 16 slides could be processed using our device, with 3 min incubation each and a total volume of 10 μL FISH probes, which is the volume used in conventional FISH for a single slide. Moreover, we compared the FISH signals under three different conditions (Fig. 3b): (a) conventional FISH with 8 h incubation, (b) conventional FISH with 10 min incubation, and (c) μFISH with 10 min incubation. We observed that the μFISH results with 10 min MFP-based incubation were better (higher intensity of spots and lower background) than those for 10 min of conventional incubation (Fig. 3d) and comparable to those of 8 h of conventional incubation (Fig. 3c). We hypothesize that the reduced incubation time is largely due to the continuous replenishment of the FISH probe mix on the cell monolayer. In contrast, in conventional FISH the probes diffuse overnight on top of the cells, resulting in a depletion layer, further lowering the concentration of probes accessing the cells.
Independently, using our MFP-based FISH implementation, we also observed a significant reduction in the cell-rinsing time compared with conventional rinsing. We substituted a single rinsing step for conventional rinsing methods comprising three 6-min steps and including detergents. At the end of probe incubation, we stopped the flow of FISH probes in the inner two apertures, whereas the flow of shaping liquid (SSC) continued. This continuous flow of SSC removes unbound probes within 1 min of rinsing, resulting in a very low background. Conventional rinsing uses detergents that can alter the structural integrity of cell membranes. Efficient rinsing with the MFP obviates the use of such detergents, thereby leaving the cytoskeleton minimally affected, which might be critical in certain applications, and eliminates the need for further manual manipulations of the cells prior to observation.
4 Concluding remarks
The μFISH implementation demonstrated here is a versatile technique to perform rapid, spatially multiplexed analysis of nucleic acids in adherent cells. This technique is characterized by significantly shorter incubation time, has better probe budgeting and utilization, is compatible with other steps in the standard workflow of conventional FISH (Supplemental Fig. S2), and most importantly, facilitates performing simultaneous tests on a given budget of sample. Probe consumption can be further reduced by implementing liquid recirculation, i.e., by circulating the probes back and forth between the inner apertures of the head as described in (Autebert et al. 2016). However, we note that recirculation of viscous probes within the MFP is yet to be explored and may impose special considerations on the fluid-handling system.
Specifically, MFP-based μFISH with CEP7 probes on MCF-7 cells resulted in ~120-fold reduction of hybridization time compared with a conventional FISH implementation, likely due to convective-enhanced transport of probes and more efficient rinsing than in conventional FISH implementations. Approaches to improve these characteristics include optimizing flow rates, pinching the flow confinement, including agents to influence molecular crowding, and optimizing the temperature, for example. Further, the ability to perform real-time observation makes it feasible to periodically measure the evolution of the FISH signal and estimate optimal signal-to-background values.
In our experiments, we used an MFP head design that results in an HFC footprint of ~0.096 mm2, which translates into a coverage of ~1000 cells on the surface. We chose to interact with 1000 cells as this is statistically representative of a multiplexed cytological smear. This HFC footprint area can be scaled up to the millimeter-scale with some modifications of the size and design of apertures (unpublished work). More interesting, performing μFISH with the MFP on fewer cells in order to get a higher resolution spatial profile across as sample is likely to be useful in the context of selected research questions. In terms of microfabrication of the heads, apertures sizes of 2 × 2 μm2 have been demonstrated (Kaigala et al. 2011) that could result in footprints of 10 μm2; however, a challenge observed in preliminary experiments is that channels of small dimensions are prone to clogging with cell debris during operation. This can be overcome by liquid shaping-based solutions (Autebert et al. 2014) combined with a reasonable reduction in aperture size. Nonetheless, with the current HFC footprint dimensions, it is theoretically possible to perform several hundred individual tests on a cytological sample of an area of 1 cm2 using spectrally equivalent probes, which in practice could still be done on a timescale of hours owing to the short incubation times and better rinsing with the MFP-based μFISH. In our demonstration, we successfully exposed distinct footprints sequentially to distinct FISH probes, aided by manual switching of the probes. In addition, multiple regions on the monolayer could be exposed to different processing liquids in parallel using a head with multiple pairs of apertures for multiple flow confinements. We believe that such a multiplexed analysis will significantly improve the range of cytological tests feasible not only on cell monolayers, but also on tissue sections, resulting in improved and more accurate tissue profiles for diagnosis. Finally, μFISH, which can be used with any type of FISH probe, allows assessing a wide variety of targets within one sample. We are confident that in addition to DNA-level in situ analysis of biological samples, it can, for example, be used in the future for combined quantification of RNA, and therefore provide a comprehensive profile for detailed cytological analysis.
Web of Science, query “fluorescence in situ hybridization” (from 18.02.2016).
This work was supported in part by the European Research Council (ERC) Starting Grant, under the 7th Framework Program (Project No. 311122, BioProbe). We thank Aditya Kashyap, Julien Cors, David Taylor and Radostina Eneva for technical assistance and discussions during the development of the protocol and the platform. Prof. Andrew de Mello (ETH Zurich), and Dr. Bruno Michel and Dr. Walter Riess are acknowledged for their continuous support.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
- D. Evanko (2007) Fully cooked FISH. In: Nat. Milestones. http://www.nature.com/milestones/miledna/pdf/miledna03.pdf
- P. Gogoi, S. Sepehri, Y. Zhou, M. A. Gorin, C. Paolillo, E. Capoluongo, K. Gleason, A. Payne, B. Boniface, M. Cristofanilli, T. M. Morgan, P. Fortina, K. J. Pienta, K. Handique, Y. Wang, Development of an automated and sensitive microfluidic device for capturing and characterizing circulating tumor cells (CTCs) from clinical blood samples. PLoS One 11, e0147400 (2016). doi: 10.1371/journal.pone.0147400 CrossRefGoogle Scholar
- L. S. Lim, M. Hu, M. C. Huang, W. C. Cheong, A. T. L. Gan, X. L. Looi, S. M. Leong, E. S.-C. Koay, M.-H. Li, Microsieve lab-chip device for rapid enumeration and fluorescence in situ hybridization of circulating tumor cells. Lab Chip 12, 4388–4396 (2012). doi: 10.1039/c2lc20750h CrossRefGoogle Scholar
- W. Liu, H. J. Kim, E. M. Lucchetta, W. Du, R. F. Ismagilov, Isolation, incubation, and parallel functional testing and identification by FISH of rare microbial single-copy cells from multi-species mixtures using the combination of chemistrode and stochastic confinement. Lab Chip 9, 2153–2162 (2009). doi: 10.1039/b904958d CrossRefGoogle Scholar
- G. Mottet, K. Perez-Toralla, E. Tulukcuoglu, F.-C. Bidard, J.-Y. Pierga, I. Draskovic, A. Londono-Vallejo, S. Descroix, L. Malaquin, J. Louis Viovy, A three dimensional thermoplastic microfluidic chip for robust cell capture and high resolution imaging. Biomicrofluidics 8, 1–14 (2014). doi: 10.1063/1.4871035 CrossRefGoogle Scholar
- M. M. Packard, Shusteff M., Alocilja E. (2011) Novel, rapid DNA-based on-chip bacterial identification system combining dielectrophoresis and amplification-free fluorescent resonance energy transfer assisted in-situ hybridization (FRET-ISH). In: Mohseni H, Agahi MH, Razeghi M (eds) Proceedings of SPIE, Bellingham. p 80990 JGoogle Scholar
- K. Perez-Toralla, G. Mottet, E. T. Guneri, J. Champ, F.-C. Bidard, J.-Y. Pierga, J. Klijanienko, I. Draskovic, L. Malaquin, J.-L. Viovy, S. Descroix, FISH in chips: turning microfluidic fluorescence in situ hybridization into a quantitative and clinically reliable molecular diagnosis tool. Lab Chip 15, 811–822 (2015). doi: 10.1039/C4LC01059K CrossRefGoogle Scholar
- H. J. Tanke, J. Wiegant, R. P. van Gijlswijk, V. Bezrookove, H. Pattenier, R. J. Heetebrij, E. G. Talman, A. K. Raap, J. Vrolijk, New strategy for multi-colour fluorescence in situ hybridisation: COBRA: COmbined binary RAtio labelling. Eur. J. Hum. Genet. 7, 2–11 (1999). doi: 10.1038/sj.ejhg.5200265 CrossRefGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.