Standardization of the Papanicolaou stain. I. A comparison of five nuclear stains.

The staining characteristics of five nuclear stains used in a Papanicolaou staining procedure were investigated. Alcohol-fixed cervical smears were stained with a modified Papanicolaou procedure using hematoxylin, alcoholic thionin bromide, alcoholic Victoria blue B, gallocyanin or the thionin Feulgen reagent (thionin-SO2) as the nuclear stain. The same anionic counterstain was used for all slides, and the optical densities of cell nuclei and cytoplasm were measured with the IBAS 2000 image analyzer. Alcoholic thionin gave the most intense nuclear stain, with a very high reproducibility of the staining pattern. Hematoxylin showed the highest coefficient of variation of the staining intensity. Both hematoxylin and gallocyanin gave some nonspecific cytoplasmic staining. Thionin-SO2 allowed a quantitative assessment of DNA, but gave a low staining intensity. Staining with the metal complex dyes interfered with subsequent staining with the pararosaniline Feulgen reagent. Alcoholic thioinin is thus recommended as a nuclear stain for cervical cytology in the Papanicolaou procedure, both for image analysis and for visual microscopy.     

Evaluation of applying feulgen stain for DNA analysis on destained hematoxylin-eosin-stained cytologic smears

OBJECTIVE: To evaluate the feasibility and reliability of DNA analysis performed on the original hematoxylin-eosin (HE)-stained cytologic specimens by destaining the slides and restaining with the Feulgen method. STUDY DESIGN: Image cytometric analysis of DNA ploidy status was performed in a comparative study on 20 cytologic preparations from 10 nasopharyngeal carcinomas. Ten smears were stained directly by the Feulgen method, and the others were Feulgen stained after HE destaining. RESULTS: There was 90% overall concordance in DNA determination and a good correlation (r = .97, P < .001) between the DNA indices determined by the 2 methods. Discordance was probably due to tumor heterogeneity. CONCLUSION: This study demonstrated that image cytometric DNA analysis on previously routinely HE-stained cytologic preparations is feasible and reliable. This method permits retrospective studies on archival cytologic material from patients with long term follow-up data.     

Quantitative cytochemical determination of desoxyribonucleic acid with the Feulgen nucleal reaction

The possibility of using the Feulgen nucleal reaction for a quantitative cytochemical estimation of desoxyribonucleic acid (DNA) was investigated. The intensity of the reaction in nuclei was determined by absorption measurements with the microscope. The accuracy of such measurements was tested by comparison with measurements on the same material with a Beckman spectrophotometer. The values obtained with the microscope agreed within a few per cent with those obtained with the Beckman spectrophotometer. Furthermore, the errors introduced by uneven distribution of absorbing material, by variations in the numerical aperture of the system, and by variation in the area used on the phototube were investigated empirically. The following variables were studied with regard to their effect on the intensity of the Feulgen reaction: type of fixation, time of hydrolysis after acetic acid-alcohol and formalin fixation, time of staining in leucobasic fuchsin, method of preparation of leucobasic fuchsin. The intensity of the Feulgen reaction in liver and erythrocyte nuclei of various vertebrates, fixed in acetic acid-alcohol, was then compared with the DNA content of these nuclei as determined by chemical analysis on a known number of nuclei. The intensity of the reaction was found to be proportional to the DNA content of the nuclei, if nuclei of similar structure and DNA concentration were compared. In nuclei of different structure and DNA concentration (i.e. liver and erythrocyte nuclei), fixed in acetic acid-alcohol, the intensity of the Feulgen reaction was, however, not proportional to the DNA content. This difficulty was overcome by isolating nuclei in sucrose and by fixing them in formalin. Uniform distribution of DNA and therefore uniform coloring after the Feulgen reaction were thus obtained. In such nuclei with uniform distribution of absorbing material the Feulgen reaction was found to be proportional to the DNA content of nuclei, even if they differed greatly in their DNA concentration. The Feulgen nucleal reaction is not quantitative in an absolute sense. For absolute determinations nuclei of known DNA content must be treated together with the unknown material to serve as standard. From these data it therefore appears possible to determine cytochemically relative amounts of DNA in cellular structures by measuring their absorption after treatment with the Feulgen nucleal reaction.     

Nonspecific light loss and intrinsic DNA variation problems associated with feulgen DNA cytophotometry.

Nonspecific light loss by the cell-wall-plus-cytoplasm (CWC) can cause a 50% increase in Feulgen absorption units in peanut root-tip nuclei as determined by scanning at 450 nm, whereas this phenomenon is not evident with chicken erythrocytes. A two wavelength scanning method of subtracting nonspecific 450 nm absorption from 550 nm Feulgen absorption values eliminated the nonspecific light loss in CWC, However, the two wavelength scanning method is time consuming and somewhat impractical with a regular scanning microdensitometer such as Vickers M85. Elimination of the problem of nonspecific light loss is suggested by careful determination of background setting with the spot position close to the nucleus in CWC. The accuracy of the CWC background setting method was further tested by comparison with subtraction method. The use of plant nucleis as an internal standard in plant DNA measurements was also evaluated. Significant variation among the replicate slides due to the variation in pine nuclear DNA amounts was observed and plant nuclei generally are not reliable internal standards. Mature chicken erythrocytes are recommended as an internal standard because the cell type and metabolic state is known.     

Scanning Feulgen-deoxyribonucleic acid cytophotometry of Papanicolaou destained preparations.

Feulgen deoxyribonucleic acid cytophotometry of Papanicolaou destained specimens revealed a differential loss in Feulgen reactivity among human buccal and cervical smears, cultured embryonic lung fibroblasts and invasive cervical carcinoma cells. Loss in Feulgen reactivity in Papanicolaou destained fibroblasts and polyploid nuclei of malignant lesions was observed to result in underestimates of relative Feulgen deoxyribonucleic acid and nuclear area values using scanning integrating microdensitometry. Thus, Papanicolaou stained preparations may not be suitable for deoxyribonucleic acid quantification of high ploidy lesions since distributional absorption error is unpredictably influenced by such factors as ploidy level, nuclear size, chromatin dispersion and differential aldehyde loss during destaining. Feulgen deoxyribonucleic acid cytophotometry of Papanicolaou stained preparations can be useful for differentiating benign from malignant lesions if extent of aneuploidy (as reflected in abnormal deoxyribonucleic acid frequency distribution profile) is used as a diagnostic indicator.     

Comparison of three different methods for automated classification of cervical cells.

Over 4600 exfoliated squamous cervical cells taken from appropriate Papanicolaou samples were classified as normal, mildly dysplastic, moderately dysplastic and severely dysplastic by an experienced cytopathologist. The slides were de-stained and subsequently re-stained with Feulgen Thionin-SO2 stain. Images of the nuclei were then captured, recorded and processed employing an image cytometry device. Automated classification of the cells was carried out using three different methods--discriminant function analysis, a decision tree classifier and a neutral network classifier. The discriminant function analysis method, which combined all dysplastic cells into an abnormal group, achieved a combined error rate of less than 0.4% for moderate and severe dysplastic cells, and less than 40% for mildly dysplastic cells. All three methods yielded comparable results, which approached those of human performance.     

Quantitative aspects of the cytochemical Feulgen-DNA procedure studied on model systems and cell nuclei

Quantitative aspects of DNA losses during fixation and pararosaniline(SO2)-Feulgen staining of microscopic preparations were studied. The preparation of a new cytochemical model, consisting of DNA-protein layers (with thicknesses between 0.1 and 5.0 micrometer) on microscopic glass slides is described and potentialities and limitations of this model are discussed. Polyacrylamide films into which high molecular weight calf thymus DNA or chicken erythrocyte nuclei had been constrained served as another model. As biological objects chicken erythrocyte nuclei and rat liver nuclei either in suspension or on microscopical glass slides were used. The experimental results indicate a loss of about 5% of the DNA due to the fixation procedure applied. Hydrolysis in 5 N HCl at room temperature, staining with the pararosaniline-Schiff medium and rinsing with sulfurous acid induced losses of DNA too, varying in amount depending on the type of preparation used. About 10% of the original DNA content is lost in total from chicken erythrocyte nuclei and rat liver nuclei dried on microscopical glass slides, from chicken erythrocyte nuclei constrained in polyacrylamide films, and from DNA-protein layers on microscopic glass slides. For nuclei fixed and stained in suspension the total losses amount to about 40%. The differences in losses between various types of preparations are discussed. Biochemically, the content of DNA originally present per chicken erythrocyte nucleus was determined to be 2.52 pg, a value, which is in good accordance with reliable biochemical data published already. It is shown that calibration of cytochemical staining intensities into biochemical units or absolute amounts of material by use of a model system, is only reliable when it is known or to be expected that both the loss of material due to fixation and staining, and the stoichiometric relation between material present and dye molecules is identical. The same holds for the application of internal biological reference systems.     

A quantitative evaluation of cytophotometric DNA analysis in retrospective studies using archival tumor specimens.

Methods developed for the cytophotometric analysis of archival tumor specimens used in retrospective studies were evaluated quantitatively. May-Grünwald-Giemsa-stained cytologic slides up to 20 years old could be restained by the Feulgen reaction with excellent results if they were destained in methanol and refixed in formaldehyde prior to Feulgen staining. Storage time had only a minor influence on Feulgen stainability. However, a considerable variation in the intensity of the Feulgen stain was observed between different slides stained simultaneously; this variation was not related to storage time. As a consequence of this variation, the use of internal staining controls, such as granulocytes, is an absolute necessity in the quantitative comparison of different slides. By expressing DNA data from the tumor cells in relative values (c values) related to the internal staining control (with a defined mean value of 2c), Feulgen ploidy level determinations could be made as accurately from measurements on old, destained slides as on cells obtained from fresh tumor material. The ploidy level could also be accurately determined in most cases of prostatic carcinoma from measurements on histopathologic sections.     

Gallocyanin chromalum as a nuclear stain in cytology. I. A cytophotometric comparison of the Husain-Watts Gallocyanin chromalum staining protocol with the Feulgen procedure

Summary In the present study, the staining characteristics of the Gallocyanin chromalum technique devised by Husain and Watts are compared with the Feulgen reaction. Liver imprints, blood smears, and cervical smears were fixed in ethanol and stained with either the Husain and Watts Gallocyanin chromalum reagent or the Feulgen-Schiff reagent. The slides were then post-treated with 70% ethanol-HCl pH 1.0, or with phosphotungstic acid for 0.5–30 min. The integrated optical density of cell nuclei was measured with a VIDAS image analyzer. In the material stained with the Husain and Watts procedure, some Gallocyanin chromalum was removed from the nuclei in the early phase (5 min) of all the post-treatment steps, followed by a plateau phase where the integrated optical density remained constant for 30 min. In this phase, the nuclear absorbance was highly reproducible and of the same size regardless of the post-treatment. Both the Husain and Watts procedure and the Feulgen-reaction gave quantitative staining of DNA. The Gallocyanin chromalum stain after Husain and Watts is a quick staining procedure for quantitative evaluation of DNA in cytological material. Proper rinsing of the slides is necessary for a good reproducibility of results.     

Automated Feulgen staining with a temperature-controlled staining machine

A Shandon Varistain 24-3 staining machine was modified in order to run automated DNA Feulgen staining. Initial studies showed a strict dependence of the staining intensity (integrated optical density [IOD]) on the temperature of the DNA hydrolysis in 4 N HCl: a difference of 0.5 degrees C around the optimum hydrolysis temperature of 27.5 degrees C resulted in IOD differences of up to 7.8% in epithelial cells and up to 12.0% in lymphocytes. A temperature-controlled stainless steel cuvette, covered with a 4 N HCl-resistant material, was developed and integrated into the machine. Temperature measurements were performed at different positions in the cuvette and on glass slides with copper-constantan electrodes fixed on them; no temperature gradient could be detected within the cuvette. The adjusted temperature of 27.5 degrees C remained constant over 24 hours. The coefficient of variation (CV) of the staining intensity in lymphocytes between different areas on the same slide and between different slides of the same staining cycle was less than 0.6%. The CV between different staining cycles was 5.9%. This system for automated Feulgen staining thus gives reproducible and reliable results and may be introduced into routine diagnostic procedures.     

Quantitative aspects of the cytochemical Feulgen-DNA procedure studied on model systems and cell nuclei

Summary Quantitative aspects of DNA losses during fixation and pararosaniline(SO₂)-Feulgen staining of microscopic preparations were studied. The preparation of a new cytochemical model, consisting of DNA-protein layers (with thicknesses between 0.1 and 5.0 m) on microscopic glass slides is described and potentialities and limitations of this model are discussed.Polyacrylamide films into which high molecular weight calf thymus DNA or chicken erythrocyte nuclei had been constrained served as another model. As biological objects chicken erythrocyte nuclei and rat liver nuclei either in suspension or on microscopical glass slides were used.The experimental results indicate a loss of about 5% of the DNA due to the fixation procedure applied. Hydrolysis in 5 N HCl at room temperature, staining with the pararosaniline-Schiff medium and rinsing with sulfurous acid induced losses of DNA too, varying in amount depending on the type of preparation used. About 10% of the original DNA content is lost in total from chicken erythrocyte nuclei and rat liver nuclei dried on microscopical glass slides, from chicken erythrocyte nuclei constrained in polyacrylamide films, and from DNA-protein layers on microscopic glass slides. For nuclei fixed and stained in suspension the total losses amount to about 40%. The differences in losses between various types of preparations are discussed. Biochemically, the content of DNA originally present per chicken eythrocyte nucleus was determined to be 2.52 pg, a value, which is in good accordance with reliable biochemical data published already. It is shown that calibration of cytochemical staining intensities into biochemical units or absolute amounts of material by use of a model system, is only reliable when it is known or to be expected that both the loss of material due to fixation and staining, and the stoichiometric relation between material present and dye molecules is identical. The same holds for the application of internal biological reference systems.Supported by grant no. 28-394 of the Praeventiefonds, The HagueIn receipt of a grant from Het Ministerie van Onderwijs en Wetenschappen, Afdeling Buitenlandse Betrekkingen, The Netherlands     

Sensitivity of HPV tests on stained vs. unstained cervical smears

OBJECTIVE: To examine the effect of Pspanicolaou staining of cervical smears on the sensitivity of molecular biologic HPV tests. STUDY DESIGN: Two sensitive HPV tests were used, HPV DNA sequence analysis after polymerase chain reaction (PCR) amplification and the Hybrid Capture II method (HC II) (Digene Diagnostics Inc., Silver Spring, Maryland, USA). Papanicolaou-stained and unstained smears taken simultaneously were examined from 265 women readmitted for examination due to an atypical squamous cells of undetermined significance diagnosis. RESULTS: After an HPV test with the PCR method on unstained slides, 66% of the women were HPV positive, whereas the same women were HPVpositive in 54% when Papanicolaou-stained slides were analyzed. However, this difference was not statistically significant (p > 0.1). With the HC II method, 55% of unstained smears were HPV positive whereas 29% were HPV positive, when Papanicolaou-stained slides were examined. This difference was significant (p < 0.001). The same strong differences in sensitivity were observed when both the PCR and HC II methods were studied on the same Papanicolaou stained glass slides, whereas on unstained slides no significant difference was found. CONCLUSION: The results demonstrate that Papanicolaou staining of a cervical smear significantly decreases the sensitivity of an HPV test performed with the HC II method, whereas the PCR method is less affected. With the Papanicolaou method, the hematoxylin bath is followed by HCl treatment, and strong acid treatment destroys DNA.     

In vivo relationship between morphonuclear parameters and hormone sensitivity of the murine MXT mammary tumor

The initially hormonosensitive (HS) MXT mouse mammary tumor spontaneously evolved to hormonoin-dependence (HI). Using a SAMBA 200 cell image processor, we compared the DNA content and the chromatin structure of HS and HI tumor cells squashed onto histologic slides; the nuclei were colored by the Feulgen reaction. We compared HI and HS nuclei by a discriminant analysis using the 15 parameters obtained on each nucleus. We show that the percentage of well-classified granulocytes (2n DNA content control) versus HS or HI nuclei exceeded 99. On the other hand, this percentage did not reach 70 when we compared HS and HI. The cell cycle analysis revealed that the percentage of cells in S and G2 + M phases were significantly higher in HI tumors than in the HS. Hence, HI and HS MXT tumor nuclei seem to be morphologically identical, but are significantly different if we refer to cell proliferation rates.     

Image analysis and DNA content of urothelial cells infected with human polyomavirus

Human polyomavirus (HPV)-infected cells in the urinary sediment are characterized by large homogeneous basophilic nuclear inclusions, which may mimic the nuclear changes in urothelial cancer. The virus is composed of double-stranded DNA and produces intense green fluorescence of nuclei stained with acridine orange. DNA measurements of Feulgen-stained smears of urinary sediment disclosed that HPV-infected cells have aneuploid DNA values and could not be differentiated from cancer cells on the basis of DNA content alone. On the other hand, computer discriminant analysis performed on high-resolution images of HPV-infected and malignant urothelial cells stained by both the Papanicolaou and Feulgen methods showed that excellent discrimination between the two groups of cells could be achieved with either stain. The misclassification rates ranged from 3% to 9%. This differentiation was almost entirely based upon computer features pertaining to the texture of the nuclear chromatin. This study documented still further the diagnostic value of high-resolution image analysis of cells in the human urinary sediment.     

Studies on isolated cell components III. A cytological study of the effects of heparin on isolated nuclei

The gel released from isolated rat liver nuclei in response to heparin treatment has been found to stain with methylene blue, azure A, and methyl green when the dyes were added to the salt-sucrose nuclear isolation medium.Azure A and methylene blue caused rapid nuclear shrinkage to as little as $\text{1}\text{4}$ the original nuclear volume. Subsequent treatment with heparin caused the nuclei to fade rapidly and swell to approximately $\text{5}\text{4}$ of the original volume. With methylene blue stained nuclei heparin caused the extrusion of deeply stained, slightly birefringent rods through apertures on the nuclear surface. Methyl green also caused nuclear shrinkage, but to a lesser degree.Studies with the Feulgen reaction demonstrated structural damage in isolated rat liver nuclei as a result of heparin action. The viscous material released by heparin was shown to be Feulgen positive by resort to hydrolysis without prior fixation, since after customary fixatives the presence of a Feulgen positive reaction outside the nucleus could not be clearly demonstrated. The possibility is suggested that the Feulgen reaction following the customary fixatives depends in part on the manner in which the DNA is bound.The nuclei of leucocytes with visually intact cell membranes included in the nuclear preparations failed to show structural damage due to heparin and it is suggested that either the cell membrane provides some protection against heparin action or that damaged cells are more susceptible to this action.Observations made provide additional basis for the conclusion that heparin replaces DNA in the nucleo-histone of the nucleus, resulting in the structural damage observed, and releasing DNA in the form of a soluble viscous protein containing complex.     

Cytophotometric Evidence of Non-S-Phase Extra-Dna In Human Neuronal Nuclei

After Feulgen staining with acriflavine-Schiff, the DNA content of glial and neuronal nuclei from various sites of the human CNS (pre- and post-central gyrus, cerebellar cortex and spinal cord) were determined by fluorescence cytophotometry. the specimens were obtained from twelve adult human autopsy cases. Glial cell nuclei always revealed a biomodal DNA distribution pattern with a large 2c and a smaller 4c peak. the 4c peak was most prominent in the cerebellum. A few 8c glial nuclei were found. Neuronal cell nuclei disclosed unimodal DNA histograms with hyperdiploid means in the range 2.2–2.5c (1.8–2.9c for the individual populations). Tetraploid 4c DNA values were not observed, neither in Purkinje cells, nor in pyramidal cells. In eleven out of a total of forty-four slides the higher DNA means of neuronal nuclei were found to be statistically significant (P < 0.05) when compared with a population of 2c hepatocytes on the same slide. The results indicate the existence of some ‘extra DNA’ in human neuronal cell nuclei, the biological significance of which has still to be elucidated. It is however, suggested that it may play an important role in the functional activity of the CNS.     

INTRAMITOCHONDRIAL FIBERS WITH DNA CHARACTERISTICS : II. Enzymatic and Other Hydrolytic Treatments

The effects of proteolytic enzymes, ribonuclease, and deoxyribonuclease upon a fibrous component of chick embryo mitochondria, which was previously shown to have many fixation and staining properties characteristic of the bacterial nucleoplasm, are reported. Pepsin digestion of formaldehyde-fixed tissues removed the membranes and matrices of mitochondria, but a pepsin-resistant fibrous material remained which was heavily stained by uranyl and lead ions. Experiments on a DNA "model system" showed that DNA treated with osmium tetroxide can be depolymerized by deoxyribonuclease. Zinc ions strongly inhibited the depolymerization of DNA. Digestion of osmium tetroxide-fixed tissues (fixed only briefly) with deoxyribonuclease for 1 hour greatly reduced the Feulgen staining of the nuclei, and after 4 hours the Feulgen reaction was completely abolished. The reduction and the disappearance of the Feulgen reaction in nuclei was paralleled by partial to complete digestion of the mitochondrial fibers in the regions studied (after 1 and 4 hours, respectively), without any other obvious changes in cellular structures. When deoxyribonuclease was inhibited by the addition of zinc ions, the nuclear Feulgen reaction was not diminished, nor were the mitochondrial fibers removed. Buffer control incubations for deoxyribonuclease and ribonuclease did not alter the structure or staining properties of the mitochondrial fibers, nor did incubation with ribonuclease. The latter reaction digested the cytoplasmic and nucleolar ribosomes after a 4-hour incubation period, in parallel with the abolishment of toluidine blue staining. The results contribute further evidence that these mitochondria contain deoxyribonucleic acid.     

Quantitative DNS-Bestimmungen von Ameisenhirnzellen

Brains of workers and males of the ant species Lasius fuliginosus were Feulgen stained. The amount of Feulgen dye of diploid nuclei of the “Perikaryenschicht” were measured microspectrophotometrically. When the slides were kept in the dark at 4° C the colour intensity did not decrease during half a year. Hydrolysis in Schiffs reagent with 0,09 N HCl did not influence the colour significantly. When the same method was used the results were more or less reproducible; 6 of 8 series gave identical values. — Feulgen dye and nuclear volume of diploid cells of ants from a natural habitat decrease from spring to autumn by ca 30 % and 10% respectively. — Ants which were kept in the laboratory for nearly two years did not show such a regular change in the Feulgen dye in their brain cells. These nuclei never reached the colour intensity of those ants from a natural habitat. — No difference in the Feulgen intensity has been found between ants kept in 4° C and at room temperature, respectively. The staining maximum was reached after 1 h by animals from the natural habitat, after 2 h by those from the laboratory. As expected the amount of Feulgen dye of haploid male cells was half of that of diploid worker cells, but the nuclear volume of haploid male cells was larger than half of that of worker cells (Zusammenfassung see p. 93).     

Senescence vs. changes in nuclear DNA, protein and dry mass contents in leaf mesophyll of two species differing in occurrence of endomitotic polyploidy

DNA and Naphtol yellow S-staining (F-NYS) protein contents were measured cytophotometrically using the Feulgen method in the nuclei of the mesophyll from the basal and apical zone of young and old leaves in two perennial monocotyledonous species: Rhoeo discolor and Clivia miniata, differing in presence or absence of DNA endoreplication. Dry mass content was determined interferometrically using an uniform field with large image shearing method. It has been shown that nuclei with 2C DNA and below 2C DNA content dominate in old leaves. The decrease in dry mass content of nuclei correlated with the decrease in NYS protein content. Parallelly a significant increase in NYS protein and DNA contents observed in chromocenters Rhoeo discolor was proportional to the increase in their dry mass. The decrease in nuclear DNA content in mesophyll of old leaves in endoreplicating species was the same as in non-edoreplicating one, however the senescence was more intensive in endoreplicating species.     

The staining behavior of oocyte cytoplasm with the alcoholic Feulgen variant and with methyl green

Summary A variant of the Feulgen reaction which has been proposed as a method for demonstrating cytoplasmic DNA in oocytes has been tested on ovarian material from a variety of species. While Schiff positive staining was developed, this was not removable by pretreatment with DNase and could be reproduced by using oxidants used in the pseudoplasmal reaction. This method was not considered useful for demonstrating cytoplasmic DNA.When chloroform extracted solutions of methyl green were used to stain ovaries, cytoplasmic staining identical in pattern to that obtained with other basic dyes was observed. The cytoplasmic staining was prevented by pretreatment of sections with RNase, but was not affected by DNase pretreatment. In somatic cells with high concentrations of cytoplasmic RNA, only nuclear staining was observed. This nuclear staining was labile to DNase but not to RNase.This work was supported by U.S. Public Health Service grants GM-10003-03 and K-3-6176-03.Contribution number 376 of the Bermuda Biological Station.