Subcellular mapping of living cells via synchrotron microFTIR and ZnS hemispheres
FTIR imaging is a label-free, non-destructive method valuably exploited in the study of the biological process in living cells. However, the long wavelength/low spatial resolution and the strong absorbance of water are still key constrains in the application of IR microscopy ex vivo. In this work, a new retrofit approach based on the use of ZnS hemispheres is introduced to significantly improve the spatial resolution on live cell FTIR imaging. By means of two high refractive index domes sandwiching the sample, a lateral resolution close to 2.2 μm at 6 μm wavelength has been achieved, i.e. below the theoretical diffraction limit in air and more than twice the improvement (to ~λ/2.7) from our previous attempt using CaF2 lenses. The ZnS domes also allowed an extended spectral range to 950 cm−1, in contrast to the cut-off at 1050 cm−1 using CaF2. In combination with synchrotron radiation source, microFTIR provides an improved signal-to-noise ratio through the circa 12 μm thin layer of medium, thus allowing detailed distribution of lipids, protein and nucleic acid in the surround of the nucleus of single living cells. Endoplasmic reticula were clearly shown based on the lipid ν(CH) and ν(C=O) bands, while the DNA was imaged based on the ν(PO2−) band highlighting the nucleus region. This work has also included a demonstration of drug (doxorubicin) in cell measurement to highlight the potential of this approach.
KeywordsFT-IR imaging Fourier transform infrared High definition Cell systems/single cell analysis Immersion objective Anti-cancer drugs
MicroFTIR is a chemically specific, non-destructive analytical method which does not require any external labelling for molecule tracking. IR spectroscopy advantageously uses low energy photons thereby avoiding unwanted fluorescence, photo-bleaching or light-induced sample damage. Measured IR absorbance follows the Beer-Lambert’s law and therefore FTIR can be used as a quantitative technique. Significantly, it can be used to analyse living cells in situ without interfering on natural biological process. Monitoring living cells in their culture medium by FTIR imaging has been recognised as a powerful method for living cell studies [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]. Information obtained from living cells is more physiologically significant than from fixed cells. Light scattering from live cell measurement is inherently far less severe compared to dried cell measurements because of the smaller difference in refractive index between the cell and the surrounding medium .
When measuring live cells by FTIR, three major challenges to be considered are the strong IR absorbance of water in the cell culture medium and within the cell itself, the biocompatibility of the IR substrate and the limited spatial resolution relative to visible microscopy and in terms of subcellular structures. To experimentally overcome the strong absorbance of water, a number of approaches have been derived including using the attenuated total reflection (ATR) method, which probes the attached living cell without significant contribution from the water in the culture medium [3, 4, 6, 7], the use of a small path length liquid cells or microfluidic transmission cell with spacer matching to the thickness of the cell [12, 13, 14, 15]. Synchrotron-based FTIR microscopy is also often used to achieve diffraction limited resolution and to enhance the signal to noise ratio [5, 12, 16, 17, 18, 19, 20, 21]. In ATR mode, the live cell layer attached to the ATR element absorbs the evanescent wave generated during the internal reflection, which probes up to a few micrometres into the cell depending on the refractive index, n, of the element and the sample, and the angle of incidence . Germanium was found to be a suitable substrate for cell attachment, and its high refractive index significantly increases the numerical aperture of the optical system (up to fourfold) which allows the imaging of subcellular features [3, 22]. However, the high refractive index of Ge also results in a relatively small depth of penetration and path length therefore the spectra may neither be presentative of the cell bulk content nor have high absorbance signal. Furthermore, we recently found that bare Ge element can be damaged by the attached living cells and a coating is needed to reduce the damage . When higher SNR and a full bulk measurement of cells is desirable, transmission mode measurement should be used. In this mode, living cells are seeded on an infrared transparent window and confined into a liquid cell by two windows. Infrared spectra of individual single living cells can be measured using an infrared microscope by either mapping with an aperture size matching or smaller than the size of the cell or with an imaging focal plane array detector [20, 21, 24]. However, microscopic imaging measurement through a standard IR liquid transmission cell with standard IR windows (usually 4–6 mm thick) has shown to be heavily affected by chromatic aberration due to refraction of IR light at the air-window interface [25, 26]. By placing a CaF2 lens on the transmission cell window with a correct centre thickness to create a “hemispherical lens” on the sample, the chromatic aberration can be eliminated with the added advantage of increasing the magnification and spatial resolution of the resultant image . It has been shown that this approach allows imaging of single cells in wax printed microfluidic channels in a standard transmission IR cell . The improvement on spatial resolution via the “hemisphere” was found to be determined by the refractive index of the lens, i.e. ~ 1.4 times when CaF2 was used. It is expected that a higher magnification and spatial resolution can be obtained when a hemisphere with higher refractive index is used. ZnSe is a popular IR transparent substrate that has a refractive index of 2.4 but was found to be not compatible to a large number of cell lines . ZnS, in contrary to a recent article , was found to be highly compatible to a large number of cell lines [6, 7, 28] and has a high refractive index of 2.25. Using ZnS hemispheres in FTIR microscopy, therefore, should increase the spatial resolution by 2.25-fold, allowing to achieve subcellular FTIR image of live cells in transmission mode, which is not demonstrated before. While the method improves the overall quality of the image and spectrum, the SNR could be limited due to the strong absorbance of water in the mid-IR region and the small aperture required for high spatial resolution measurement. The higher refractive index of ZnS (compares to CaF2 and when used without anti-reflection coating) also increases reflection loss potentially resulting in a lower throughput of light. The main aim of this study is to demonstrate the combined ZnS hemispherical lens approach with the synchrotron radiation that can maximise the signal and achieve both spectral quality and spatial resolving power for live subcellular studies. As a first step, we measured the 1951 United State Airforce (USAF) resolution target to demonstrate the achieved lateral spatial resolution with two different aperture and step size. Then, we image the distribution of different chemical components in full thickness living cells to identify the subcellular organelles in the surrounding of the nucleus using synchrotron FTIR microspectroscopy as an example of the potential applications. This will pave the way to study in situ drug interactions with live cells for the monitoring of both drug and biochemical changes in cells during drug treatment at subcellular level.
Synchrotron FTIR microscopy
The experiment was carried out at the B22 beamline (MIRIAM) of the Diamond Light Source synchrotron facility. The system is comprised of a Vertex 80v FTIR spectrometer and a Hyperion 3000 microscope system (Bruker optics) with a 36× reverse Cassegrain reflective objective (NA = 0.5 in air) and a matching condenser. A high sensitivity mid-band mercury cadmium telluride (MCT) single element detector was used for the mapping experiment. This is 50 μm pitch size and has a cut-off at circa 650 cm−1. A multilayer Ge filter was used to reduce the spectral range below 4000 cm−1 of the incoming beam, and the detector non-linearity was software corrected during acquisition.
Resolution target measurement
An 8-mm diameter ZnS hemisphere was placed directly on the 1951 USAF resolution target (Newport Corporation) with the centre of the hemisphere aligned to the group 7 element 6 feature which contains 228.1 line pair per mm, i.e. line width of 2.19 μm. The signal was optimised in reflection mode on a reflective area ~ 15 μm away from the line feature where a background measurement was also measured. Water was added to fill the gap between the hemisphere and the resolution target to improve the visibility of the feature. Images were obtained by scanning using a 6 μm × 6 μm aperture size in air (i.e. 2.7 μm × 2.7 μm through the ZnS hemisphere) with a step size of 2 μm (i.e. 0.89 μm through the ZnS hemisphere) and a set aperture size of 3 μm × 3 μm in air (i.e. 1.3 μm × 1.3 μm through the ZnS hemisphere) with a set step size of 1 μm (i.e. 0.44 μm through the ZnS hemisphere). Measurements were obtained by averaging 16 scans at 4 cm−1 spectral resolution giving a scanning time of 2–3 s per spectrum.
Live cell measurement
A549 cells (86012804 Sigma) were used in this experiment. Cells were grown in the Diamond B22 cell culture lab nearby the beamline. They were DMEM supplemented with 10% foetal bovine serum (FBS), 2 mM L-glutamine and 100 unit/mL penicillin streptomycin in a T25 tissue culture flask and incubated in a 5% CO2 incubator. Cells are harvested when reached ~ 80% confluency. The harvested cells were suspended in fresh DMEM medium at 100,000 cells per mL and 100 μL of the cell suspension was seeded directly on the flat side of the ZnS hemisphere. The ZnS hemisphere carrying the cell suspension was incubated in the CO2 incubator for 12–24 h to allow the cells to attach to the ZnS surface. Once the cells are attached (observed through an inverted optical microscope), the DMEM medium was removed and CO2-independent L15 medium, supplemented with 10% FBS, 2 mM L-glutamine and 100 unit/mL of penicillin streptomycin, was added to the attached cell. The ZnS hemisphere with the attached cells was carefully loaded into a prototype transmission device where the cells are then sandwiched between the two ZnS hemispheres with a 12-μm spacer. The prototype can hold ~ 200 μL of medium to maintain the viability of cells during measurement. Once the hemispheres were aligned in the microscope, a background was measured in an area with just medium but no cell attached. (Note that the medium contains mostly water with 0.3 wt% amino acids, 0.4 wt% protein, 0.02 wt% phosphate and 0.1 wt% carbohydrates, i.e. the compounds that may interfere with the spectrum are at a much lower concentration than in the living cell). A typical spectrum measured from the background region is shown in Electronic Supplementary Material (ESM) Fig. S1, which demonstrate the lack of interference from any organic compounds. FTIR images of the cell between the two ZnS hemispheres were then captured by mapping: i.e. sample scanning via 23 μm × 23 μm aperture size in air (i.e. 10 μm × 10 μm through the ZnS hemisphere) with a step size of 7 μm (i.e. 3.1 μm through the hemisphere) and a 6 μm × 6 μm aperture size in air (i.e. 2.7 μm × 2.7 μm through the hemisphere) with a step size of 2 μm (i.e. 0.89 μm through the hemisphere). Measurements were obtained by averaging 50 scans at 4 cm−1 spectral resolution giving a scanning time of 10 s per spectrum. Cell viability in the prototype transmission device for more than 12 h was confirmed using trypan blue assay (ESM Fig. S2).
Results and discussion
Resolution target imaging
Importantly, the results have demonstrated a significant improvement over previous non-ATR-based high resolution FTIR imaging studies that also published the contrast profiles of the USAF target including using multi-beam synchrotron imaging , CaF2 hemispheres  or the high magnification approach in transmission mode [31, 32]. A similar improvement in spatial resolution was previous observed in ATR mode  and more recently when a Si immersion lens was used in back scattering mode (non-USAF target measurement) . However, the latter work was not demonstrated with live cells but with polystyrene beads. While ATR mode has the potential to achieve higher spatial resolution, especially with a Ge lens , it is important to note that ATR mode measures the surface layer of the attached cell and the layer is thin with high refractive index elements (e.g. Ge with an angle of incidence of 30° for NA of 0.5 or 37° for NA of 0.6 produces a depth of penetration of ~ 1.1 and 0.8 μm, respectively, in the cell layer even at the longer wavelength of 1000 cm−1). This is in contrast to the full thickness of the sample (e.g. living cell) is measured in transmission mode. Also, the objective is half illuminated in ATR mode in a similar way as the transflection mode measurement demonstrated in Fig. 2. This is in contrast to the fully illuminated objective for the measurements made in transmission mode, as in the case of live cell measurement presented later and illustrated in Fig. 1. The high lateral spatial resolution is therefore maintained, unlike in ATR or transflection mode, along both vertical and horizontal axes of the image.
The improvement in the spatial resolution is due to the increase in NA by introducing the ZnS hemisphere in the path of the IR beam just above the sample. Interestingly, the NA of the objective is 0.5, which is increased to 1.125 through the ZnS hemisphere, and the expected spatial resolution is calculated to be 3.3 μm based on the Rayleigh criterion. However, the results have clearly demonstrated that the spatial resolution exceeded the estimation using the Rayleigh criterion by at least 30%, which is in agreement with previous works by others using reverse Cassegrain objective with an FPA detector . While that previous work was based on focal plane array detector, this work has also demonstrated a similar better-than-Rayleigh criterion spatial resolution can be observed by the mapping approach.
Live cell imaging
The results have shown that the living cell can be easily located by plotting the amide II absorbance (1590–1490 cm−1) across the mapped region. In this experiment, the imaging area is focused around the nucleus region of the cell where the thickness of the cell is greatest and many large intracellular organelles are located. The image shows the amide II, a key protein band, was used here rather than amide I because the absorbance of water with this transmission cell masked the amide I spectral region. During this experiment, only a slight pressure was applied between the two hemispheres to avoid physical damage to the living cells. There are rooms to refine the experimental procedures to incrementally increase the pressure to reduce the path length through the medium above cell for a better SNR. This will be refined in the future development of the prototype transmission device. The use of a thinner spacer will also reduce the contribution from the medium above the attached cell further improve the SNR of the measurement. Previous works have shown that cell remains viable for several hours when sandwiched between windows with a gap of a few μm [20, 36] and > 60 h when subjected in a flow device . Nevertheless, the current procedure enables the demonstration of the principle of subcellular imaging of living cells with the improved spatial resolution.
It is interesting to note that the distribution of the overall lipid represented by the CH2 and CH3 stretching mode vibration in the 2970–2840 cm−1 region has shown a smaller, more localised distribution than then the amide II-based protein map suggesting a lipid-rich organelle is present near the middle of the cell. The representative spectra (ratioed to the background spectrum measured from a near-by area that is not occupied by cells) extracted from the lipid-rich and lipid-poor region of the cell are shown in Fig. 3. While the amide II band shows a similar absorbance, the lipid bands from the lipid-rich region show a stronger absorbance than from the lipid-poor region, which demonstrates the images shown reflect the differences in the spectra. Compared to the earlier study where CaF2 hemispheres were used for live cell imaging, which has a spectral range cut-off at around 1100 cm−1, Fig. 3 shows that the spectral range can reach down to 950 cm−1 when using the ZnS hemispheres, which is critical for the study of DNAs (1087 cm−1 and ~ 970 cm−1) and carbohydrates such as glycogen (~ 1150–1000 cm−1).
Drug in live cells
Synchrotron FTIR combined with the hemispherical lens approach is a powerful method for live cell FTIR imaging with subcellular spatial resolution. The high brightness and the ability to focus the IR light to a diffraction limited spot with synchrotron has enabled the use of small apertures without significant loss in energy throughput. When combined with the use of ZnS hemispheres, which increase the NA of the objective and further reduce the diffraction limited spot-size by 2.25×, features as small as 2.19 μm were laterally resolved at ~ 6 μm wavelength, i.e. a lateral spatial resolution of at least λ/2.7. Subcellular features, thereby, can be imaged with the improved spatial resolution. Different images of lipids, based on the alkane chain ν(CH) band (2970–2840 cm−1), carbonyl band (1768–1708 cm−1) and DNA band (1099–1082 cm−1), within a living A549 cell surrounding the nucleus are were obtained. The ability to monitor the spatial distribution of various components inside the single cell in this experiment clearly demonstrated the advantages of high spatial resolution image of live cells. A preliminary test of drug treated in an anti-cancer agent doxorubicin is also shown, demonstrating the potential to study the dynamic changes in cells as a result of drug treatment. The study of changes in the same cells upon drug treatment is not possible with fixed cell measurements.
We thank Diamond Light Source for access to beamline SM14912 that contributed to the results presented here. We thank the Prince Sattam Bin Abdulaziz University for Ali Altharawi’s PhD scholarship. We also thank the University of Lisbon for sponsoring the travel of Dr. Fale.
Compliance with ethical standards
Conflict of interest
There is no potential conflict of interest from this work.
- 4.Alam MK, Timlin JA, Martin LE, Williams D, Lyons CR, Garrison K, et al. Spectroscopic evaluation of living murine macrophage cells before and after activation using attenuated total reflectance infrared spectroscopy. Vib Spectrosc. 2004;34(1):3–11. https://doi.org/10.1016/j.vibspec.2003.07.002.CrossRefGoogle Scholar
- 7.Fale PLV, Altharawi A, Chan KLA. In situ Fourier transform infrared analysis of live cells’ response to doxorubicin. BBA-Mol Cell Res. 2015;1853(10 Part A):2640–8.Google Scholar
- 8.Chan KLA, Fale PLV. Label-free optical imaging of live cells. In: Meglinski I, editor. Biophotonics for medical applications. Woodhead Publishing; 2015. p. 215–41.Google Scholar
- 9.Vongsvivut J, Heraud P, Gupta A, Thyagarajan T, Puri M, McNaughton D, et al. Synchrotron-FTIR microspectroscopy enables the distinction of lipid accumulation in thraustochytrid strains through analysis of individual live cells. Protist. 2015;166(1):106–21. https://doi.org/10.1016/j.protis.2014.12.002.CrossRefPubMedGoogle Scholar
- 10.Munro KL, Bambery KR, Carter EA, Puskar L, Tobin MJ, Wood BR, et al. Synchrotron radiation infrared microspectroscopy of arsenic-induced changes to intracellular biomolecules in live leukemia cells. Vib Spectrosc. 2010;53(1):39–44. https://doi.org/10.1016/j.vibspec.2010.02.004.CrossRefGoogle Scholar
- 17.Birarda G, Grenci G, Businaro L, Marmiroli B, Pacor S, Vaccari L. Fabrication of a microfluidic platform for investigating dynamic biochemical processes in living samples by FTIR microspectroscopy. Microelectron Eng. 2010;87(5–8):806–9. https://doi.org/10.1016/j.mee.2009.11.081.CrossRefGoogle Scholar
- 23.Fale PLV, Chan KLA. Preventing damage of germanium optical material in attenuated total reflection-Fourier transform infrared (ATR-FTIR) studies of living cells. Vib Spectrosc 2016;in press.Google 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.