Abstract
Purpose:
To develop a more efficient impression cytology (IC) method for the transfer of ocular surface cells onto glass microscope slides for cytochemical, immunocytochemical and immunofluorescence studies.
Methods:
Cells are lifted off the ocular surface with a mixed cellulose ester membrane and then firmly attached to a glass slide using a novel triblock copolymer comprised of collagen type I, polyethylenimine and poly-L-lysine (CPP), and crosslinking cells and glass slide by heating and cooling. The membrane is removed intact after softening it with a butanol/ethanol solution. Transfer of cells is complete in about 10–15 minutes and are ready for staining. The efficiency of our cell transfer method was compared to current methods based on poly-L-lysine and albumin paste.
Results:
Our method ensured almost complete transfer of cells. In contrast, the transfer of rabbit conjunctiva cells onto poly-L-lysine-covered slides was 37.5±6.3 % lower, and onto albumin-paste covered slides 62.5±5.6 % lower (mean ± SD); the transfer of rabbit goblet cells was even less efficient. The new method was also more efficient for transfer of cells from human oral mucosa obtained by IC. Transferred cells were successfully stained with H&E, chemiluminescence, and immunofluorescence agents. Using our method, we stained ocular surface cells for S100A4 and ATF4, both of which play a role in the pathophysiology of dry eye disease. We obtained similar results with oral mucosal cells, suggesting the generalizability of our approach. We propose an explanation for the strong adhesion of cells to the glass slide, which is based on their interactions with the triblock copolymer.
Conclusions:
We developed a novel approach for the efficient and rapid transfer of cells obtained by IC onto glass microscope slides using a novel copolymer. Compared to available methods, our improved approach makes IC robust and simple, and should increase its diagnostic yield and clinical applicability.
Keywords: impression cytology, IC, ocular surface, dry eye disease, CPP copolymer
INTRODUCTION
Impression cytology (IC) is a minimally invasive method for the retrieval of the outermost layer of ocular surface cells using various types of membranes. Introduced in 1977, IC evolved into the primary method of evaluating human conjunctival and corneal epithelial cells 1,2. IC transfers cells from the upper layers of the ocular surface (mainly interpalpebral conjunctiva) by applying on it a membrane either directly or, more recently, using a specialized instrument (Eyeprim, Opia Technologies, France). These cells are then evaluated according to the needs of the study. Thus, IC can be conceptualized as consisting of three distinct steps: cell acquisition from the ocular surface, followed by transfer to a platform, where they are subjected to evaluation by morphological or molecular methods.
What makes IC appealing are its simplicity and versatility. IC, an effective and safe technique, has contributed significantly to both clinical and research ophthalmology. The range of its applications encompasses conventional cytological evaluation of ocular surface cells for squamous metaplasia; examination of molecular and cellular mediators of inflammation; and assessment of the expression of various genes by determining the levels of their transcripts or of the proteins they encode 3–5.
IC is still evolving methodologically in order to overcome some persisting limitations. The need for topical anesthesia to minimize patient discomfort during sampling has been viewed as a limitation, as it may affect the molecular analysis of the sample 6. Another limitation is the variability of the number of harvested cells available for evaluation, which depends on the examiner’s skill and the properties of the materials used, especially those of the membrane. Given the clinical and research utility of IC, investigators have attempted to optimize each of the three steps of IC mentioned above.
The properties of the membrane play a major role in IC cell transfer. The ideal membrane should maximize cell retrieval, maintain the structural integrity of the cells, and allow their quantitative release to the final destination, usually a microscope glass slide or a test tube containing appropriate buffers for the extraction of proteins or nucleic acids. Several membranes have been used including those based on cellulose acetate, polycarbonate, nitrocellulose or polyethersulfone 2,7,8. The Eyeprim device uses polyethersulfone membranes; despite initial enthusiasm, it appears to perform similarly to the traditional IC method, at least for some applications 9.
The inadequate release of cells from membranes led investigators to study their morphology directly on the membrane10. However, visualization of cells on membranes is hindered by the lack of membrane transparency and/or by the high background when cells are stained, especially when nitrocellulose or cellulose acetate membranes are stained with immunocytochemical stains. To overcome this problem, Krenzer and Freddo harvested cells on pure nitrocellulose membranes, which were then sprayed with a fixative, transferred to poly-L-lysine-coated glass slides and dried 11. This procedure, however, proved cumbersome, as it requires chemical dissolution of the membrane matrix and a long washing step with various solutions or enzymatic digestion until the membrane is totally dissolved. Equally cumbersome and prolonged (includes a 3-day washing step) was a method in which cells were transferred onto a light microscopy slide coated with poly-L-lysine 12. Finally, a method using a double-sided adhesive tape on glass slides had its own limitations13.
To improve the transfer of cells from the membrane to the glass microscope slide, we coated the glass slide with a novel triblock copolymer of collagen, polyethyleneimine, and poly-L-lysine, which serves as a glue. This glue provides a pillow-like structure that greatly improves the adhesion of the cells to the glass slide. Here, we report our results, which demonstrate that our approach ensures the quantitative transfer of ocular surface cells onto microscope slides. These cells can then be evaluated with several morphological methods, including conventional staining, immunocytochemistry, and immunofluorescence.
MATERIALS AND METHODS
Reagents
All general solvents and reagents were of HPLC grade or of the highest grade commercially available. Nitrocellulose membranes: mixed cellulose nitrate and acetate esters, 0.22 μm pore size, (cat. no. GSTF14250, Millipore Corp. Bedford, MA). Glutaraldehyde 50% aqueous solution (16320, Electron Microscopy Sciences, Hatfield, PA). Bovine collagen type I 5mg/mL (A1064401, Thermo Fisher Scientific, CA). Reagents: Sodium metasulfite (161519), Ponceau S (141194), ethanol (T038181000), 1-butanol (B7906), poly-L-lysine 0.1 % (w/v) in H2O (P8920), polyethylenimine branched Mn~10,000 (408727), polyethylene glycol 400 (202398), periodic Acid-Schiff PAS kit (395B), hematoxylin (HHS16) and eosin (HT110316), Rhodamine B isothiocyanate 283924 and fluorescein isothiocyanate (46950) were from Sigma (St Louis, MO). VECTASHIELD® Antifade Mounting Medium with DAPI (H-1500, Vector Laboratories, Burlingame, CA). S100A8/9 antibody (48M7C7, Norvus, St Louis, MO). ATF4 (11815, Cell Signalling, Beverly, MA). NFAT5 antibody (GTX23446, GeneTex). Antibodies: CD11b/ ITGAM (ab133357, Abcam, Cambridge, MA). HRP-anti-rabbit IgG (ab7090), HRP-anti-mouse (ab47827), FITC-anti-rabbit (ab6717), FITC-anti-mouse (ab6785), Texas Red anti-mouse (ab6787), Texas Red anti-rabbit (ab6719) antibodies, as well as goat IgG (ab37373), rabbit IgG (ab37415) and mouse IgG (ab37355) Isotype controls were from Abcam (Cambridge, MA). Histostain Plus Bulk kit (85–8943, Invitrogen, Frederick, MD). In all instances, we used the antibody dilutions recommended by their respective manufacturers; they ranged between 1:200 and 1:2,000.
Study animals
Nine male Dutch-belted rabbits (DB), each 12-weeks weighing ~2 kg, were from Envigo (Somerset, NJ) and housed singly in a room with a strict 12-h on/off light cycle, temperature of 22 ± 2° C, and humidity of 45 ± 5%. During experiments, animals had free access to standard chow and water without nutritional enrichments that might affect tear production or stability. Rabbits were acclimated to this environment for ≥2 weeks prior to any study. All animal studies were in accord with the Statement for the Use of Animals of the Association for Research in Vision and Ophthalmology and approved by our relevant institutional committees.
Human sample collection
We evaluated conjunctival cells obtained by IC from nine normal male subjects, 45–65 years old. The study was approved by the Institutional Review Board (IRB) of the Stony Brook University Ethics Committee. Written consent was obtained from all human volunteers.
Conventional impression cytology
We employed the conventional IC methodology as follows. Sample acquisition: In humans, cells were obtained from the superior and inferior bulbar conjunctiva (1 o’clock and 7 o’clock position) after topical anesthesia with proparacaine hydrochloride (Sandos). In rabbits, cells were obtained from the superior bulbar conjunctiva after topical anesthesia with 1% lidocaine; sampling of the inferior conjunctiva was difficult. We removed the excess fluid using flat, round-tipped forceps. Then a 11mm x 8 mm MF Mixed Cellulose Ester membrane (MCE membrane type 0.22uM GSTF14250, Millipore Corp., Bedford, MA, USA) was applied to the bulbar conjunctiva with gentle pressure for 5 s, then an additional 20-μL drop of the proparacaine solution was added to the center of the membrane to soften it (for rabbits we used lidocaine). Once softened, the membrane was removed with a peeling motion. During conventional IC, at this stage, some practitioners transfer the cells onto a glass microscope slide by applying the membrane on the glass slide and then staining the cells. This approach often leads to unsatisfactory results. Consequently, others fix the cells on the membrane, which is then applied onto a glass microscope slide with double-sided adhesive tape. Following that, cells are stained in situ, e.g., with PAS to identify goblet cells. In rabbits, we followed the latter approach.
Proposed method for impression cytology
The proposed method is outlined in Fig. 1. Cell acquisition and staining are essentially the same as in the conventional method. The main difference between the two methods lies in the transfer of the cells to the glass microscope slide.
Figure 1. The improved IC method.
Ocular surface cells, obtained by applying a membrane onto the eye (a), are fixed by transferring the membrane into a tube with Fixing Solution (FS; b). After washing the membrane in Washing Solution (WS; c), it is immersed in the CPP copolymer (d) and applied to a microscope glass slide with a drop (20 μL) of the same copolymer (e). Then an assembly is created (e’) to firmly attach the cells on the microscope slide. Specifically, on top of the glass slide (1) with the cells still attached to the filter (2) through the copolymer (brown) we place tissue paper (3) and another glass slide (4). Firm pressure is then applied with a binder clip (5). After that, the glass slide (4) and the tissue paper above the membrane are removed and the glass slide (1) is heated and cooled as in Methods, whereupon the cells become firmly attached onto the microscope slide. Then the Membrane Softening Solution (MSS; blue color) is applied to fully cover the membrane for 10 seconds (f), and the membrane is removed intact with a pair of tweezers (g). The transfer of cells, essentially complete, is accomplished in about 10–15 minutes. The cells can be evaluated using various stains or by immunocytochemistry and immunofluorescence-based methods (h).
Working solutions
Fixing Solution (FS):
2% (v/v) glutaraldehyde in PBS. If specific protocols require a lower concentration of glutaraldehyde, e.g., 1%, then adding paraformaldehyde to 4% (w/v) provides adequate fixing of the cells.
Washing Solution (WS):
75% (v/v) ethanol in ddH2O.
CPP copolymer (Collagen/Polyethylenimine/Poly-L-lysine):
Bovine collagen type I 2.5 mg/mL, branched polyethylenimine 0.05% (w/v), Poly-L-lysine 0.05% (w/v), 33% (v/v) ethanol in ddH2O.
To avoid coalescing of its components, this mixture is prepared as follows. Solutions in ddH2O of bovine collagen type I 5 mg/mL, polyethylenimine 0.1 % (w/v), and poly-L-lysine 0.1 % (w/v), and absolute ethanol are added in a 2:1:1:2 ratio (v/v/v/v) in the sequence they are presented here and mixed by vortexing.
Membrane Softening Solution (MSS):
20% 1-butanol in 75% ethanol.
Depending on the properties of a given membrane, the composition of this solution may be modified to achieve the required softening for its removal. For example, the concentration of butanol can be increases to as high as 30% or lowered by adding ddH2O to as low as 10%. Only through trial and error the optimal composition of this solution can be determined. Too high a concertation of butanol may degrade the membrane such that its complete removal is not feasible; the opposite happens if its concentration is too low for a given membrane. Depending on the membrane composition, additional softening agents (10%) can be added including 4-hydroxybutyl nitrate, 4-chloro-1-butanol or γ-undecalactone.
The time of exposure of the membrane to the Membrane Softening Solution can be critical. The higher the butanol concentration the shorter the duration of its contact with the membrane ought to be to ensure its complete removal. Such exposure may vary between 2 and 30 seconds or even longer.
Immunofluorescence Blocking Solution:
5% normal bovine serum, 0.3% Triton X-100, supplemented with 250 mM L-glycine and 250 mM L-arginine in PBS. This solution blocks the autofluorescence of free cell aldehyde groups.
Freshly prepared working solutions are recommended, however, they are stable for at least 3 months when stored at 4°C in a hermetically sealed container.
In case it is desirable to determine on which side of a membrane the cells have been attached (for example, if there is doubt when membranes are not properly marked), the cells can be readily visualized by dipping the membrane for a few seconds into a solution of 0.05% Ponceau S in PBS. Ponceau S reversibly stains cell proteins and can be washed away in PBS; such washing does not interfere with subsequent steps of this method.
Cell transfer onto glass microscope slides
Each step of our method is outlined in Fig. 1
Cell acquisition and fixing:
Upon removal of the IC membrane from the eye, cells are fixed by immersing the membrane in Fixing Solution for 20 minutes. If needed, membranes can be kept in Storage Buffer for up to 2 weeks at 4C.
Washing the membrane:
The Fixing Solution is washed away by dipping the membrane in a tube with 5–10 mL Washing Solution with slow but constant agitation for about one minute.
Application of the copolymer:
After washing, the CPP copolymer is applied to the membrane by immersing it in a tube with an adequate amount of CPP solution followed by gentle shaking for a 10 sec. For optimal results, drying of the membrane between this and the next step should be avoided by not delaying the transfer, and 1 dop (20μL) of the same copolymer can be deposited directly in the middle of a fat-free glass slide. Following that the membrane is placed on the glass with the cells facing the slide. The membrane is then covered with 4 layers of soft tissue paper (preferably Kimwipes) and another glass slide is placed on top of the paper. Firm pressure in the middle of the glass is then applied for about 90 sec. (or longer if needed) with a binder clip. Fluid extruded during the application of pressure is removed with tissue paper.
After that, the covering glass slide and the tissue paper above the membrane are removed. The glass slide that has the membrane attached to it is dried in a heating block (56°C for 3 min) and cooled on ice for at least 3 min. These two steps allow the polymerization of the triblock components and the binding of cells, copolymer, and glass. If needed, the microscope slide with the membrane can be safely stored overnight at 4°C.
Membrane softening:
To soften the membrane glued on the glass slide, a 40-μL drop of Membrane Softening Solution is applied onto the membrane, evenly distributed with a small plastic rod or a pipette tip, and immediately dried with absorbing tissue paper to remove the excess solution. If the membrane is not dried, it will be difficult to remove it intact.
Membrane removal:
About 10 seconds after the application of the Membrane Softening Solution, the membrane is grasped at one of its corners with a pair of tweezers (preferably ~20 cm long) and peeled with a rolling movement towards its opposite side. If membrane removal is delayed, the membrane becomes too soft to be removed in toto, usually breaking into small pieces.
Following removal of the membrane, the cells on the glass slide are stained with conventional methods such as H&E or immunocytochemistry. If immunofluorescence staining is desired, the slide should be also washed with the Immunofluorescence blocking solution.
Two additional points merit mention. First, in addition to glutaraldehyde, cells can be fixed in 4% paraformaldehyde, 55% ethanol or 75% isopropanol (2-propanol) applied for 20 minutes. Methanol, ethanol (>60%), butanol, pentanol and higher alcohols, as well as long chain hydrophobic aldehydes like propionaldehyde are not appropriate fixatives, as they adversely affect the membranes, which in their presence are either excessively softened or disintegrate.
Second, IC membranes can be stored for up to 2 weeks at 4C in the following Storage Buffer: 0.2% aldehyde (glutaraldehyde or paraformaldehyde) with 2% isopropanol (2-propanol) or at least one of them in PBS. Before transferring cells to a glass slide, stored membranes should be processed from step b (fig.1) by incubating for in Fixing Solution (4% glutaraldehyde in PBS) that is required for binding by CPP on a glass.
Immunocytochemistry and immunofluorescence staining
The application of these two methods to cells was as previously described 14,15. Briefly, cells transferred onto glass slides were washed with PBST (PBS + 0.05% Tween 20) 3 times and permeabilized with 0.1% Triton X-100 in PBS for 15 min. Cells are blocked with normal serum or 5% bovine serum albumin for 1 h, and then incubated with an appropriate dilution of primary antibodies at 4 °C overnight. After three washes with PBST for 15 min, cells are incubated for 1 h with horseradish peroxidase-conjugated secondary antibody at room temperature. Cells are then developed with chromogen (DAB) for 5–10 min at room temperature after three washes are counterstained with hematoxylin. The slides are dehydrated and mounted with liquid mounting media. For immunofluorescence detection, cells are processed with the same procedure as above but incubated for 1 h with fluorophore-conjugated secondary antibodies (usually diluted 1:400) instead and covered with mounting medium with 4’−6-diamidino-2-phenylindole (DAPI). All stained slides are examined and photographed using a Nikon fluorescence microscope (Nikon Eclipse Ni, Tokyo, Japan) with a Wide Zoom Camera.
RESULTS AND DISCUSSION
Our novel method, outlined in fig. 1, was applied successfully to the IC study of rabbit and human eyes. We achieved consistently essentially complete transfer of the ocular surface cells from membranes onto microscope slides.
Fig. 2 (top row) illustrates the application of this method to the rabbit eye. As expected, the nitrocellulose membrane (NCM) was covered with ocular surface cells. With our transfer method, these cells were quantitatively deposited onto the glass slide. The extent of this transfer is demonstrated by the continuous carpet of cells on the glass slide visualized by PAS staining, and by the practically complete absence of cells on the NCM following the transfer step. The transfer of cells can be accomplished in 10–15 minutes.
Figure 2. Application of the IC method to rabbits.
Images of the three stages of our method are shown for each of the three types of glass slide described in the text. Cells were stained with PAS & haematoxylin on the membrane as they are lifted off the ocular surface (left column); after they are deposited onto the glass slide (middle column), and again on the membranes after deposition of cells to the respective glass slide (right column). Top row: Cells were processed using our method employing the CPP copolymer. A1: An abundance of goblet (big pink dots) and conjunctival (small blue dots) cells; the latter are less prominent. Inset: goblet cells stained with PAS (pink). A2: Image of IC conjunctiva cells deposited on a glass slide. PAS staining. Inset: higher magnification of the same. A3: Image of a membrane after the cells were deposited on a glass slide. The paucity of cells indicates their quantitative transfer to the slide. Middle row: Transfer of cells using conventional IC to glass slides covered with poly-L-lysine (PLL). Fewer cells are on the glass slide (B2) and more cells are left behind on the membrane (B3) compared to the corresponding images of the top row. Bottom row: Images of cells transferred to glass slides covered with albumin paste (ALB). Compared to the corresponding top row images, even though the same number of cells was uploaded to the membrane from the ocular mucosa (C1), there were fewer cells on the glass slide (C2), with many more remaining attached to the membrane (C3). Scale bars = 50 μm.
We compared our method to two conventional approaches, which use glass microscope slides covered with either poly-L-lysine or albumin paste. As can be seen in fig 2 (middle and lower rows), representative cytological pictures indicate that the transfer of cells from the NCM to these slides is suboptimal. Consistent with this observation is that a large number of cells remain on the corresponding NCMs after the transfer to the glass slides, which contrast with the absence of cells on the NCM when using the new method.
We quantified the efficiency of cell transfer by each of these three methods (fig. 3). Compared to our method, the transfer of conjunctiva cells onto poly-L-lysine-covered slides was 37.5±6.3 % lower (mean ± SD for this and all subsequent values), and onto albumin-paste covered slides 62.5±5.6 % lower. The transfer of goblet cells was even less efficient. Compared to our method, it was 57.1±4.3 % lower onto poly-L-lysine-covered slides, and 89.3 ±1.6 % lower onto albumin-paste covered slides (p<0.01 for both). Of note, no transfer of cells was possible onto fat-free microscope slides without coating them with agents such as poly-L-lysine.
Figure 3. Quantitative assessment of the transfer efficiency.
Quantification of the efficiency of cell transfer from the membrane to glass slides using CPP, PLL or ALB. Rabbit conjunctiva (left panel) and goblet cells (right panel) were counted in the 2 mm2 field of microscope view shown in fig. 2, middle column. Values: mean ± SD based on 3 separate fields of 6 membranes (n=18); each membrane was from a different rabbit. *, p<0.05. **, p<0.01.
The method was applied to human subjects as well. As indicated in fig. 4, the transfer of human cells was also complete. Similar to the rabbit cells, the transfer of conjunctiva cells to glass slides covered with poly-L-lysine or albumin paste was incomplete. This figure also shows an interesting variation of our approach. Cells can be stained on the membrane, immediately after they are lifted off the ocular surface and then deposited onto the glass slide, requiring no further staining (Suppl. fig. 1). Viewing cells on the membrane hinders their morphological evaluation because the optical properties of the membrane differ from those of glass. All cells shown in this figure were stained with H&E.
Figure 4. Application of the IC method to humans.
Hematoxylin and eosin-stained human conjunctival cells obtained by IC and deposited on class slides covered with different substrates. Left column: Images of conjunctival cells stained with haematoxylin & eosin on the membrane as they are lifted off the ocular surface Right column: Images of such cells after they are deposited onto the glass slide. Glass slides were covered with CPP copolymer (top row) or poly-L-lysine (PLL; middle row), or albumin paste (ALB; bottom row). The differences in transfer efficiency are evident Scale bar = 50 μm.
In addition to PAS and H&E staining, cells transferred by our method can be evaluated using a variety of methods. We have successfully stained conjunctival cells with additional stains, including crystal violet, Coomassie blue, and Ponceau S (Suppl. fig. 2) and in a preliminary manner with Masson trichrome, Prussian blue, toluidine blue O and Sudan black B. In addition, we stained these using immunocytochemistry and immunofluorescence to identify cellular proteins. For example, ocular cells were stained by immunocytochemistry for the expression of the S100A8/9 protein, which plays a role in the pathophysiology of dry eye disease 16,17, a common ocular disease 18 (fig. 5). An example of immunofluorescence staining of conjunctiva cells was the detection of the transcription factor ATF4, of importance to ocular diseases 17,18, shown in fig. 6. More cellular proteins have been detected in transferred cells by immunofluorescence, such as S100A8/9, p65, NFAT5, and ITGAM (Suppl. fig. 3–5). Furthermore, we also studied by confocal microscopy cells obtained with this method (data not shown). It is likely that the glutaraldehyde-fixed cells could be studied by electron-microscopy as well. Finally, fluorescent labeled small molecules that enters conjunctiva cells can be detected in cells obtained by IC and transferred to glass slides (Suppl. fig. 6).
Figure 5. Immunohistochemical detection of S100A8/9.
The expression of the protein S100A8/9 (brown) in human conjunctival cells obtained with our method was detected by immunohistochemistry, as in Methods. Nuclei were stained with hematoxylin. Insert: the primary antibody was a non-specific isotypic antibody (negative control). Scale bar = 50 μm.
Figure 6. Immunofluorescence detection of ATF4.
The expression of the protein ATF4 in human conjunctiva cells obtained with our method was detected by immunofluorescence, as in Methods. A. Cell nuclei (blue channel) were stained with DAPI; and ATF4 (red channel) was identified with a mAb reacted with a Texas Red-labelled secondary Ab. B. Same as in A but at a higher magnification, demonstrating the strong but varied expression of ATF4. Notable is the complete cell transfer, evidenced by the uninterrupted spread of cells. C. As in A above except for the use of a non-specific isotypic control primary antibody. Scale bar = 50 μm.
Of interest, goblet cells among those transferred with our method can be detected by using fluorescent stains such as fluorescein isothiocyanate or rhodamine isothiocyanate. Both of them fail to stain the mucins of the goblet cells and generate a “negative image” of the goblet cells (Suppl. fig. 7).
Our improved IC methodology can be applied to extra-ocular tissues. For example, we obtained cells from the oral cavity by IC using our transfer step and stained them with fluorescein isothiocyanate used as a conjugate with a novel salirasib derivative (fig. 7) or control conjugate of fluorescein isothiocyanate and ethanolamine (19,20). Staining methods mentioned above can be successfully applied to these cells, e.g., Suppl. fig. 2.
Figure 7. Application of the IC method to human oral cavity cells.
Human cells obtained from the oral cavity were transferred to glass microscope slides using our IC method. Upper row: Cells were stained with DAPI and a conjugate of fluorescein isothiocyanate with a novel drug that penetrates the cells. Lower row: Cells were stained with DAPI or a control conjugate of fluorescein isothiocyanate with ethanolamine showing barely detectable staining. Scale bar = 50 μm.
We compared the immunofluorescence staining of conjunctival cells directly on NCMs to that after their transfer onto class slides. As seen in fig. 8, staining cells directly on NCMs was accompanied by a high background that made the evaluation of cellular features difficult or impossible. Such background was not encountered when cells were deposited on glass slides.
Figure 8. The high fluorescence background of membranes used in IC.
MCE membranes with human conjunctiva cells obtained by IC were stained on the membrane by immunofluorescence for the detection of ATF4 using a specific mAb as in Methods. The secondary antibody carried a different fluorophore (left to right, DAPI, fluorescein, and Texas Red). Images were obtained using a fluorescence microscope with the membrane placed on the left half of each glass slide. The high fluorescence background markedly decreased the quality of cell imaging, rendering the approach impractical. Scale bar = 50 μm.
Fig. 9 indicates the chemical interactions that likely underlie critical steps of the method. The attachment of cells to the membrane is facilitated by an array of hydrophobic interactions and electrostatic attractions between the cell surface and the nitrocellulose membrane, which have been reported in detail21,22. Prominent among them are hydrophobic (hp) and van der Waals interactions of lipophilic molecules, e.g., extracellular lipoproteins. The electrostatic interactions include dipole-dipole (δ+δ- : δ+δ-) forces; attractions between the permanent dipoles of the nitrate groups (at neutral pH they are negatively charged) and the dissociated polar and polarizable groups of the cellular proteins; dipole-zwitterions; and dipole-ion attractions between the nitrate groups of the membrane and cations of the cell surface molecules, like arginine, lysine or histidine residues. Of note, the adsorption of proteins to membranes of different chemical composition has been the subject of significant research23,24.
Figure 9. Proposed chemical interactions underlying the improved IC method.
Left: Lifting off the cells by the membrane and their fixation. In this example, the membrane consists of MCE, containing cellulose acetate (CA) and cellulose nitrate (CN). Cells attach to it by hydrophobic interactions, and electrostatic attractions including dipole-dipole and dipole-ion forces. Bottom of this panel: Fixing of the cells with glutaraldehyde. The aldehyde group of glutaraldehyde binds free amine groups (shown as -NH3+) of proteins on the cell surface. A free and a bound glutaraldehyde molecule are shown. Right: The interaction of free aldehyde groups of glutaraldehyde (in already fixed cells) with amine groups of the CPP copolymer (collagen, polyethylenimine, and poly-L-lysine, as shown). The positively charged triblock copolymer is sandwiched between the cells and the glass slide, both negatively charged. Note: The Membrane Softening Solution (MSS; middle of left panel) can disrupt the attractive forces between cell and membrane, but not between copolymer and cell or glass.
Glutaraldehyde fixes the cells by binding to -NH2 groups of amino acid residues of proteins exposed on cell surface 25.
The strong adhesion of the cells to the glass slide is achieved through electrostatic attractions (and perhaps other forces) between the one side of the triblock copolymer and the negatively charged surface and between the other side of the copolymer and the negatively charged surface of the glass slide (the copolymer is sandwiched between them).
The nature of the interactions between cells, membrane, copolymer, and glass perhaps explains how our Membrane Softening Solution can help remove the membrane off the attached cells while not disrupting their attachment to the glass through the copolymer. We hypothesize that the electronegative, polar Softening Solution lowers the hydrophobic attractions and neutralizes bond dipole moments of cellulose nitrate, thus facilitating the disruption of dipole-dipole, dipole-zwitterion and dipole-ion attractions. The result is the dissociation of the membrane from the cells. But since the Membrane Dissociation Solution has no such effect on the interactions of the copolymer with the cells and the glass microscope slide, cells remain on the slide.
Our findings convincingly support the notion that our novel triblock copolymer greatly improves the critical step of cell transfer from the membrane to the glass slide where they can be evaluated. Indeed, with our method such transfer is essentially complete. This improved transfer step will likely minimize the number of membrane applications onto the ocular surface to obtain an adequate sample, making IC faster and less burdensome to both patient and examiner.
The improved method is simple to perform and robust, only requiring the quick optimization of the Membrane Softening Solution for each type of membrane and of the duration of its application. The wide applicability of this method is demonstrated by the examples shown here that represent ocular and extraocular cells stained with conventional as well as immunocytochemical and immunofluorescence staining.
In conclusion, we present a novel improvement of IC that makes it robust, simpler and with increased diagnostic yield compared to its current variations.
Supplementary Material
Supplementary Figure 1. Staining of cells directly on the impression cytology filter
Human conjunctiva cells were stained with H&E as in Methods directly on the IC membrane. Cells were fixed on the membrane within 5 minutes after they were harvested, and staining began 20 minutes later. Once staining was complete, you follow steps b to g in Figure 1 when cells are ready to be viewed to be viewed on the microscope. A: Top panel: the image is of a stained membrane placed on the glass slides. Bottom panel: the image is of the cells on the glass slide after the membrane has been removed. These H&E-stained cells were viewed on the microscope (B). Scale bar = 50 μm.
Supplementary Figure 2. Staining of IC cells with various stains
Cells obtained by IC from the ocular surface or the oral cavity were stained as in Methods with crystal violet (A, rabbit conjunctival cells; and B, human oral cells) or stained with hematoxylin and Ponceau S (C, human conjunctival cells) or with Coomassie blue (D, human conjunctival cells) Scale bar = 50 μm. Insert in panel D: higher magnification (white scale bar = 25 μm).
Supplementary Figure 3. Immunofluorescence staining of IC cells for S100A8/9 and NF-κB
Rabbit conjunctival cells obtained and processed per our method were stained for S100A8/9 and p65 expressed using the corresponding fluorescent mAbs. Nuclei were stained with DAPI. p65, a component of NF-κB, was detected using specific primary mAb and a FITC-ladled secondary antibody (green fluorescence), and S100A8/9, a protein involved in DED, was detected with specific primary mAb and a Texas red-ladled secondary antibody (red fluorescence). The merged images are shown in the last panel in which secreted S100A8/9 (red) are identified by the white arrow. Scale bar = 50 μm.
Supplementary Figure 4. Immunofluorescence staining of IC cells for NFAT5
Human conjunctiva cells were stained for the expression of the transcription factor NFAT5 with a specific primary Ab and a FITC-ladled secondary antibody (green fluorescence). Images of DAPI and NFAT5 are shown as well as the merged image. Scale bar = 50 μm.
Supplementary Figure 5. Immunofluorescence staining of IC cells for ITGAM
Human conjunctiva cells were stained for the expression of the integrin subunit α-M (ITGAM aka CD11b) with a specific primary Ab and a FITC-ladled secondary antibody (green fluorescence). Images of DAPI and ITGAM are shown as well as the merged image. Scale bar = 50 μm.
Supplementary Figure 6. Fluorescence detection of cell-penetrating drugs in conjunctiva cells
Human conjunctival cells obtained by IC were transferred to glass microscope slides. Cells were stained with DAPI and a conjugate of rhodamine B isothiocyanate with a novel drug that penetrates the cells. Scale bar = 50 μm.
Supplementary Figure 7. Visualization of ocular surface goblet cells by fluorescent staining
Goblet cells transferred from the ocular surface using our method can be recognized, in addition to PAS staining, by an alternative approach. When the transferred layer of ocular surface cells is stained with either fluorescein isothiocyanate (Fluorescein ITC) or Rhodamine B isothiocyanate (Rhodamine B ITC) the goblet cells remain unstained (black dots). The merged image accentuates the goblet cells. Scale bar = 50 μm.
ACKNOWLEDGEMENTS
Basil Rigas is the guarantor of this work and, as such, had access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. We thank Dr. Ernest Natke for his assistance with the manuscript. This work was supported by the National Institutes of Health [Grant R44EY031193].
Funding:
This work was support by the National Institutes of Health [Grant R44EY031193] and by funds provided by the Department of Ophthalmology, Renaissance School of Medicine, Stony Brook University, Stony Brook, NY, USA
Footnotes
DECLARATION OF INTEREST
The authors declare no competing interests except for BR who has an equity position in Medicon Pharmaceuticals, Inc. and Apis Therapeutics, LLC; and LH, who has served as an employee of Medicon Pharmaceuticals, Inc. and is currently an employee with an equity position in Apis Therapeutics, LLC.
REFERENCES
- 1.Egbert PR, Lauber S, Maurice DM. A simple conjunctival biopsy. Am J Ophthalmol 1977;84:798–801. [DOI] [PubMed] [Google Scholar]
- 2.Hagan S. Biomarkers of ocular surface disease using impression cytology. Biomark Med 2017; 11:1135–1147. [DOI] [PubMed] [Google Scholar]
- 3.Singh R, Joseph A, Umapathy T, Tint NL, Dua HS. Impression cytology of the ocular surface. Br J Ophthalmol 2005;89:1655–1659. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lopin E, Deveney T, Asbell PA. Impression cytology: recent advances and applications in dry eye disease. Ocul Surf 2009;7:93–110. [DOI] [PubMed] [Google Scholar]
- 5.Thia Z-Z, Tong L. Update on the role of impression cytology in ocular surface disease. Taiwan J Ophthalmol 2019;9:141–149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Hyun C, Filippich L, Hughes I. The inhibitory effect of pentobarbitone on reverse transcription-PCR. J Biochem Biophys Methods. 2005;62:63–68. [DOI] [PubMed] [Google Scholar]
- 7.Roy P, Cimbolini N, Feraille L, Elena P-P. Evaluation of membranes material for impression cytology. Invest Ophthalmol Vis Sci. ARVO. 2011;52:1937–1937. [Google Scholar]
- 8.Nelson JD, Havener VR, Cameron JD. Cellulose acetate impressions of the ocular surface. Dry eye states. Arch Ophthalmol. 1983;101:1869–1872. [DOI] [PubMed] [Google Scholar]
- 9.López-Miguel A, Gutiérrez-Gutiérrez S, García-Vázquez C, Enríquez-de-Salamanca A. RNA collection from human conjunctival epithelial cells obtained with a new device for impression cytology. Cornea. 2017;36:59–63. [DOI] [PubMed] [Google Scholar]
- 10.Calonge M, Diebold Y, Sáez V, Enríquez de Salamanca A, García-Vázquez C, Corrales RM, Herreras JM. Impression cytology of the ocular surface: a review. Exp Eye Res 2004;78:457–472. [DOI] [PubMed] [Google Scholar]
- 11.Krenzer KL, Freddo TF. Cytokeratin expression in normal human bulbar conjunctiva obtained by impression cytology. Invest Ophthalmol Vis Sci 1997;38:142–152. [PubMed] [Google Scholar]
- 12.Luzeau R, Carlier C, Ellrodt A, Amédée-Manesme O. Impression cytology with transfer: an easy method for detection of vitamin A deficiency. Int J Vitam Nutr Res 1988;58:166–170. [PubMed] [Google Scholar]
- 13.Haller-Schober E-M, Schwantzer G, Berghold A, Fischl M, Theisl A, Horwath-Winter J. Evaluating an impression cytology grading system (IC score) in patients with dry eye syndrome. Eye. 2006;20:927–933. [DOI] [PubMed] [Google Scholar]
- 14.Mattheolabakis G, Papayannis I, Yang J, Vaeth BM, Wang R, Bandovic J, Ouyang N, Rigas B, Mackenzie GG. Phospho-aspirin (MDC-22) prevents pancreatic carcinogenesis in mice. Cancer Prev Res (Phila). 2016;9:624–634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mattheolabakis G, Wang R, Rigas B, Mackenzie GG. Phospho-valproic acid inhibits pancreatic cancer growth in mice: enhanced efficacy by its formulation in poly-(L)-lactic acid-poly(ethylene glycol) nanoparticles. Int J Oncol 2017;51:1035–1044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tong L, Lan W, Lim RR, Chaurasia SS. S100A proteins as molecular targets in the ocular surface inflammatory diseases. Ocul Surf 2014;12:23–31. [DOI] [PubMed] [Google Scholar]
- 17.Wang S, Song R, Wang Z, Jing Z, Wang S, Ma J. S100A8/A9 in inflammation. Front Immunol 2018; 9:1298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Clayton JA. Dry eye. N Engl J Med 2018;378:2212–2223. [DOI] [PubMed] [Google Scholar]
- 19.Swoboda G, Hasselbach W. Reaction of fluorescein isothiocyanate with thiol and amino groups of sarcoplasmic atpase. Z Naturforsch C Biosci 1985:40: 863–875. [PubMed] [Google Scholar]
- 20.Heilkenbrinker A, Reinemann C, Stoltenburg R, Walter JG, Jochums A, Stahl F, Zimmermann S, Strehlitz B, Scheper T. Identification of the target binding site of ethanolamine-binding aptamers and its exploitation for ethanolamine detection. Anal Chem 2015:87: 677–685. [DOI] [PubMed] [Google Scholar]
- 21.Low SC, Shaimi R, Thandaithabany Y, Lim JK, Ahmad AL, Ismail A. Electrophoretic interactions between nitrocellulose membranes and proteins: biointerface analysis and protein adhesion properties. Colloids Surf B Biointerfaces. 2013;110:248–253. [DOI] [PubMed] [Google Scholar]
- 22.Přistoupil TI, Kramlová M, Štěrbíková J. On the mechanism of adsorption of proteins to nitrocellulose in membrane chromatography. J Chromatogr. 1969;42:367–375. [DOI] [PubMed] [Google Scholar]
- 23.Holstein CA, Chevalier A, Bennett S, Anderson CE, Keniston K, Olsen C, Li B, Bales B, Moore DR, Fu E, Baker D, Yager P. Immobilizing affinity proteins to nitrocellulose: a toolbox for paper-based assay developers. Anal Bioanal Chem 2016;408:1335–1346. [DOI] [PubMed] [Google Scholar]
- 24.Kurien BT, Scofield RH. Other notable methods of membrane protein detection: a brief review. Methods Mol Biol 2015;1314:357–370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hopwood D. Theoretical and practical aspects of glutaraldehyde fixation. Histochem J 1972;4:267–303. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure 1. Staining of cells directly on the impression cytology filter
Human conjunctiva cells were stained with H&E as in Methods directly on the IC membrane. Cells were fixed on the membrane within 5 minutes after they were harvested, and staining began 20 minutes later. Once staining was complete, you follow steps b to g in Figure 1 when cells are ready to be viewed to be viewed on the microscope. A: Top panel: the image is of a stained membrane placed on the glass slides. Bottom panel: the image is of the cells on the glass slide after the membrane has been removed. These H&E-stained cells were viewed on the microscope (B). Scale bar = 50 μm.
Supplementary Figure 2. Staining of IC cells with various stains
Cells obtained by IC from the ocular surface or the oral cavity were stained as in Methods with crystal violet (A, rabbit conjunctival cells; and B, human oral cells) or stained with hematoxylin and Ponceau S (C, human conjunctival cells) or with Coomassie blue (D, human conjunctival cells) Scale bar = 50 μm. Insert in panel D: higher magnification (white scale bar = 25 μm).
Supplementary Figure 3. Immunofluorescence staining of IC cells for S100A8/9 and NF-κB
Rabbit conjunctival cells obtained and processed per our method were stained for S100A8/9 and p65 expressed using the corresponding fluorescent mAbs. Nuclei were stained with DAPI. p65, a component of NF-κB, was detected using specific primary mAb and a FITC-ladled secondary antibody (green fluorescence), and S100A8/9, a protein involved in DED, was detected with specific primary mAb and a Texas red-ladled secondary antibody (red fluorescence). The merged images are shown in the last panel in which secreted S100A8/9 (red) are identified by the white arrow. Scale bar = 50 μm.
Supplementary Figure 4. Immunofluorescence staining of IC cells for NFAT5
Human conjunctiva cells were stained for the expression of the transcription factor NFAT5 with a specific primary Ab and a FITC-ladled secondary antibody (green fluorescence). Images of DAPI and NFAT5 are shown as well as the merged image. Scale bar = 50 μm.
Supplementary Figure 5. Immunofluorescence staining of IC cells for ITGAM
Human conjunctiva cells were stained for the expression of the integrin subunit α-M (ITGAM aka CD11b) with a specific primary Ab and a FITC-ladled secondary antibody (green fluorescence). Images of DAPI and ITGAM are shown as well as the merged image. Scale bar = 50 μm.
Supplementary Figure 6. Fluorescence detection of cell-penetrating drugs in conjunctiva cells
Human conjunctival cells obtained by IC were transferred to glass microscope slides. Cells were stained with DAPI and a conjugate of rhodamine B isothiocyanate with a novel drug that penetrates the cells. Scale bar = 50 μm.
Supplementary Figure 7. Visualization of ocular surface goblet cells by fluorescent staining
Goblet cells transferred from the ocular surface using our method can be recognized, in addition to PAS staining, by an alternative approach. When the transferred layer of ocular surface cells is stained with either fluorescein isothiocyanate (Fluorescein ITC) or Rhodamine B isothiocyanate (Rhodamine B ITC) the goblet cells remain unstained (black dots). The merged image accentuates the goblet cells. Scale bar = 50 μm.