Changes in nucleic acid and protein content in nuclei of human cervical cells

Mechanisms of Ageing and Development, 27 (1984) 135-142 Elsevier Scientific Publishers Ireland Ltd.

135

CHANGES IN NUCLEIC ACID AND PROTEIN CONTENT IN NUCLEI OF HUMAN CERVICAL CELLS

PRADEEP K. BHATTACHARYAa and ARISTOTEL J. PAPPELIS b aDepartment of Biology, Indiana University Northwest, Gary, IN 46408 (U.S.A.), and bDepartment of Botany, Southern Illinois University, Carbondale, IL 62901 (U.S.A.) (Received September 6th, 1983) SUMMARY Nuclei in five classes of cervical cells observed in Pap smears were studied using quantitative epifluorescence microscopy. The five classes of cells were: parabasal (Pb) cells; intermediate cells with round (I-R), oval (I-O), and rod-pyknotic (I-RP) nuclei; and, pyknotic (P) cells. Six nuclear traits were measured: total nucleic acid, DNA, RNA, total protein, histone, and non-histone protein. The six nuclear indices increased as Pb cells became I-R cells (cell enlargement and maturation), and then decreased as I-R cells degenerated through the following senescence sequence: I-O ~ I-RP -+ P. We infer that these changes continue and result in anucleate, superficial cells. Pb cells are probably in early stages of DNA synthesis (S-phase of the cell cycle) since the mean for DNA increased as they became I-R cells. The following types of cells comprised the Pap smears studied: Pb, 7%; I-R, 19%; I-O, 55%; I-RP, 8%; P, 9%; superficial cells with nuclei devoid of nucleic acids, 1%; and, anucleate cells, 1%. We conclude that cervical exfoliative cytology provides a model system for the study of human cell development, maturation, senescence, and death in addition to its use in detecting early through late stages of cervical cancer. The high correlation between the nuclear indices studied suggests that several quantitative nuclear parameters other than DNA may be useful for cancer detection.

Key words." Total nucleic acids; DNA; RNA; Histone protein; Non-histone protein; Cervical cells

INTRODUCTION Gynecologic exfoliative cytology has been used for more than 30 years for cancer detection in asymptomatic patients [1,2]. The rationale for cancer detection lies in the morphology and physiology of the epithelial cells and their location in the various layers of the tissue. Stratified squamous epithelium that covers part of the ectocervix and vagina consists, more or less, of three layers: the deepest (germinal) layer which adheres 0047-6374/84/$03.00 Printed and Published in Ireland

© 1984 Elsevier Scientific Publishers Ireland Ltd.

136 to the basement nlembrane and regenerates the epitt~elium; the middle and exfoliate along with superficial cells. They differ from normal cells in theii cellula~ nuclear, and nucleolar morphologies which enables cytologists to predict maligna~t diseases. Normal cells observed in Pap (cervical l smears are like bucc,il ,:ells studied b\ Bhattacharya e t al. [31 as a model system lk)r human cell developlnenl, tnuturatiul~ senescence and death. The DNA content of normal (intermediate) cells observed m Pap smears have heel~ used as a nuclear index for cancer detection [ 4 - 8 ] . Malignant tumors have an aneuploid DNA distribution pattern and may arise from DNA stem lines ranging from hypoploid to hypertetraploid levels [4]. Epithelial dysplasia of the cervix shows a regulal I)NA distribution with diploid and tetraploid values. Thus, not only can cancer cells be distinguished front normal cells, carcinoma in situ can be distinguished from mvasive cat cimmu~ on the basis of DNA measurements. Auer e t al. [7] recently reported thai both DNA and nuclear protein content can be determined simultaneously cylophotometricall_x This type of data can be used to distinguish cancer cells from nornlal cells [ 8 ] The method may have application to the study of cellular senescence. Dutu e t al. [9] reviewed attempts to develop rapid, specific information useful in thu diagnosis of carcinonras of the cervix {analysis o f vaginal fluid, selected biochemical and cytochemical methods). In their study, they assessed the value and limitations of c\ I~ enzymologic investigations in the diagnosis of cervical carcinomas tested being practically absent. Nuclei were consistently negative. Their data revealed continuous degeneration during maturation and senescence of normal cervical cell> The biology of cellular senescence remains unexplained in general terms I ltJ] Bhattacharya e t al. [3] compared eight nuclear indices (total nucleic acid, DNA. RNA. total protein, histone protein, non-histone protein, protein-bound lysme, and proteinbound arginine) in five cell types observed in buccal smears (palabasal: intermediate cell> with round, oval, or rod-pyknotic nuclei: and superficial cells with p vknotic nucleiL All nuclear indices increased as parabasal cells become intermediate cells with round nuclei and decreased as cells degenerated to form pyknotic cells. The data supported the sequence of cellular and nuclear changes used in the cell development index I 1 li

137 Since the best example of human exfoliative cytology involves cervical cells for cancer detection, we believe that information regarding changes during development, maturation, and senescence like that for buccal cells [3] would assist researchers interested in these cellular sequences and others interested in developing rapid, early cancer detection systems using more than one nuclear trait as the diagnostic criterion. The purposes of this study were: to determine the relative amounts of nuclear proteins and nucleic acids (six nuclear indices) in five cell types observed in normal cervical smears which represent the various stages of development, maturation, senescence and death of cells in normal cervical tissues, and to determine the relative frequency of normal cervical cells in the smears using the cell development index developed to classify exfoliating epithelial cells.

MATERIALS AND METHODS Seven smears were selected at random from 150 duplicate, normal smears (Papanicolaou screening, fixed with Spray-Cyte, air-dried; primary slides of all seven donors were found to be normal) and cells were rehydrated with 4% neutral formalin for 4 - 5 h. The quantitative estimation (in arbitrary units) of the six nuclear indices was obtained using cytofluorometric methods [ 12] and a Zeiss Universal microsocope for combined verticallight epifluorescence excitation [Zeiss Vertical Illuminator III FL, FL-Neofluar objective lenses, stabilized Osram HBO 200 W mercury lamp, Schott (Mainz) UG and BG excitation filters, variable range barrier filters (410 to 650 nm)] and transmitted-light phase-contrast illumination equipped for photometric measurements as follows: Zeiss Microscope Photometer 01 with RCA 1 P-28 photomultiplier stabilized by an AC stand, Zeiss Digital Photometer- Indicator, and electronic shutter control for 7 ms ultraviolet excitation [13]. Uranium glass plate (Zeiss GG 17) was used as fluorescent standard for calibration of the fluorescence values. After staining, the cervical smears were covered with an oil/ glycerine/staining solution. Cells were first viewed by transmitted tungsten light (phase contrast) to locate appropriate cells and position nuclei for cytofluorescence measurements. Background reading in the cytoplasm adjacent to the measured nucleus was subtracted from the reading obtained for the nucleus. The five cell types studied were selected using the criteria for buccal cells [3,11]: parabasal cells (Pb); intermediate cells with round (I-R), oval (I-O), and rod-pyknotic (I-RP) nuclei; and, pyknotic cells (P). The buccal cell criteria [ 11] were adjusted for cervical cell and nuclear sizes and morphologies [1]. For each of seven slides, nuclei of five cells per cell type were measured (a total of 35 nuclei measured for each cell type). Selected cells were free from others and flat (no curled or folded margins). Measurements for total nucleic acids (TNA) and DNA following ribonuclease digestion (RNase A, Sigma, 80 K units/rag at a concentration of 1 mg pr 2 ml of 30% ethanol for 1 h at 50°C) were determined sequentially on the same slide (not necessarily the same cell) using the berberine sulfate method [14,15]. The RNA values were obtained by

13~ subtracting the DNA measurements from the TNA measurements, kxcitation filters/~,i and UG5 used with barrier filters 50 and 53 resulted in a greenish-yellow fluorescence in tile nuclei. The snrears were destained when TNA was extracted by hydrolyzing th,, cells "in 5'2~ trichloroacetic acid for 3 h a l 60°C. Destained, hydrolyzed smears were restained with 0.1~/ Sulfaflavine 112,14] in 0.02 M buffer solution at pH 8.0 for ,s0 mm for histone (H) measurements. The cells were rinsed and stained for t<~tal protein ~1PI measurement in 0.1~7~: Sulfaflavine in 0.02 M buffer at pH 2.8 for 30 rain. The buffer I~T both staining solutions was Mcllvaine's citric acid-phosphate buffer at the stated p|t. Zeiss UGI excitation filter and barrier filter 41 resulted in a yellowish-green fluorescence in tire nucleus. Non-histone protein tnftPl values were determined by subtracting the II values from the TP values. At all times, appropriate enzyme-extracted controls icellsl wilh :ind without acid hydrolysis were used to check primary cell fluorescence and reactioH specificit,,. [he Fk-Neofluar 40x objective lense, fixed Optovar, and fixed diaphragm were used for all of the fluorescent measurements. Also, during measurements, dial settings on the phot,~meter indicator unit were kept constant for each fluorochrome. The fluorochrome berberine sulfate (Fluka, Buchs, Switzerland l was obtained from Tridom Chemicals, Itauppage, Long Island, NY, and the Brilliant Sulfaflavine (('hromaGesselschaft Schmid and Co.l was obtained from Roboz Surgical Instruments ( ' ~ . Washington, DC. One hundred cells were randonrly selected in each smear and classified using the cell development index (CDI) criteria I111. For CDI computation, values were assigned to the. cell types oil a 1 through 9 scale as follows: Pb, I I - R , 4; 1-O, q: I-RP, (~:l', 7:ghosi ((;I (nucleus visible using phase-contrast microscopy but contains no DNA when viewed with berberine sulfate stain and epifluorescence microscopy), 8" and. anucleate ~A), % The values assigned were multiplied by tire percentage for cell types and then added together to obtain the CDI value. The data obtained for each ce,vical smear to compute CDI also were used to compute the karyopyknotic index (KI) and maturation index IMIt The K1 was the ratio of P and A to all cells. The MI was computed by multiplying the percentages by tile following rating scores and adding all scores: Pb, 0.0 I-R, I-O, I-RP, 0.5 ; and P, (;, and A, 1.0. Statistical analysis of tile data was by two-way analysis ol variance and Duncan~ multiple range test at the 0.05 probability level.

RESULTS The six nuclear traits were easily measured in fiat cells found free from others, However, it was difficult to find the five required for each cell type in this condition. The highest values for Ihe six nuclear indices occurred in 1-R cells (Table I). These values were arbitrarily set at 100c~ for comparative purposes We considered Pb cells to be in various stages of the cell cycle (G1, S, or G2) and the mean for these would be inappropriate as a

139 TABLE I NUCLEIC ACID AND PROTEIN CONTENT IN CERVICAL CELLS (FIVE CELL TYPES) Nuclear indices: nuclear total nucleic acid (TNA), DNA, RNA, total nuclear protein (TP), histone (H), and non-histone protein (nHP). Figures represent arbitrary values of fluorescence [expressed as percentage of fluorescence units of I-R (as 100%)] from berberine sulfate for TNA and DNA (after RNase treatment), and Brilliant Sulfaflavine (pH 2.8) for TP and (pH 8.0) H, corrected for background. The values for RNA were obtained by substracting DNA from TNA values. The values for nHP were obtained by subtracting H from TP values. All values are means for five nuclei per cell type in cervical smears from each of seven donors (a total of 35 nuclei measured for each cell type). Standard Error (S.E.). Cell types: parabasal, Pb; intermediate cells with round (I-R), oval (I-O), and rod-pyknotic (I-RP) nuclei; and, pyknotic cells, P. The values represent arbitrary units of fluorescence. Within each of the indices, means in the same row with the same letter are not statistically different, P = 0.05. The means are compared as a percentage of I-R (% of I-R). Nuclear index

Cell type Pb

I-R

I-0

I-RP

P

TNA S.E. % of I-R

31.5 b -+4.1 78

40.2 a -+5.1 100

31.6 b +-4.6 79

22.0 c -+3.0 55

16.7 d -+1.7 42

DNA S.E. % of I-R

22.0 b -+2.2 81

27.3 a +-2.6 100

20.6 b -+2.0 75

15.4 c +-1.9 56

12.1 c ±1.1 44

RNA % of I-R

9.5 ab 74

12.9 a 100

11.0 a 85

6.6 bc 51

4.6 c 36

TP S.E. % of I-R

42.9 b +-4.0 78

54.7 a -+6.3 100

40.6 a +-3.4 74

29.2 c -+3.2 53

23.4 d +2.0 43

H S.E. % of I-R

12.2 b -+0.8 81

15.0 a -+0.9 100

11.3 b ±0.6 75

8.4 c -+0.6 56

7.3 d -+0.4 49

nHP % of I-R

30.7 b 79

39.7 a 100

29.3 b 76

20.8 c 5"4

16.1 c 42

s t a n d a r d for c o m p a r i s o n o f cell t y p e s w i t h i n t h e same cervical smear or for similar data f r o m all s m e a r s studied. All o f t h e n u c l e a r indices increased as P b cells d e v e l o p e d i n t o I-R cells. T h e m e a n s for all n u c l e a r indices d e c r e a s e d significantly t h e r e a f t e r as I-R cells c o m p l e t e d m a t u r a t i o n a n d e n t e r e d the s e n e s c e n c e phase s e q u e n c e (I-O t o I-RP to P) e x c e p t : D N A , for I-RP a n d P; R N A , for I-R a n d I-O a n d for I-RP and P; and, nHP, for I-RP a n d P. A l t h o u g h a l m o s t all d i f f e r e n c e s were significant at t h e 1% level, we listed d i f f e r e n c e s in T a b l e I at t h e 5% level. T h e overall t r e n d s for m e a n s o f all nuclear indices from Pb to P cells were as follows: Pb, 79% o f I-R cells (set a r b i t r a r i l y at 100%); I-R, 100%; I-O, 77%; I-RP, 54%; a n d P, 43%. T h e g r e a t e r p a r t o f t h e T N A f l u o r e s c e n c e ( a b o u t 69%) was a c c o u n t e d for b y D N A

14(1

fluorescence. Similarly. the greater part oi the TP fluorescence labour 7 ! for by the nHP fluorescence.

} was accounted

The ratios t\)r TNA/DNA over the five types were: Pb, 0.97; I-R, 1.00: 1-O, 1.04; I-RP. 0.98; and. P, 0.96. The ratios for T N A / R N A over the five cell types were: Pb, 1.t(~, I-R 1.00; 1-O, 1.08; 1-RP, 1.00; and P, 0.96. The ratios for RNA/DNA over the five ceil types were: Pb, 0.83; I-R, 1.00; 1-O, 0.96; I-RP, 0 . 9 7 ; a n d P, 0.93. The ratios li~r TP/DNA over the five cell types were: Pb, 0.99, I-R, 1.00; I-O, 1.01: I-RP, 05}7: and, P, (LOS. The ratios for H/DNA over the five cell types were: Pb, 1.01: l-R, 1.00: 1-O, IX)I: I-RP 1.01; and, P, 1.09. The ratios for nHP/DNA over the five cell types were: Pb, 0.98, I - R 1.00; I-O, 1.01 ; I-RP, 0.97: and, P. 0.93. The ratios for nHP/RNA over tire five cell types were: Pb, 1.18; I-R, 1.00; 1-O, t.0(~; I-RP, 1.00; and, P, 1.00. The correlation coefficients lot the six nuclear indices over the five ceil types ranged from 0.87 to 1.00. Of the 15 possible r values, one was 0.87 ( D N A RNA). l h e r values for TNA nttP and T N A - T P were 0.93 and 0.94, respectively. Five others ranged ir,~m 0.97 to 0.99 and seven were 1.00. The latter included: TNA -DNA: DNA with TP, It. c,~ nHP; TP with H or nHP;and, !f nHP. The cell type means ti)r tire seven cervical smears studied were as follows: Pb, 7< : I-R, 19+S; 1-O, 555:;: I-RP, 8<;: P, o!);; G, 1~/<; and A, 1'.:~. No binucleate cells were ~t~served. The mean for CDI was 489 (range 469 to 532). The mean for KI was 0.11 (range 0.06 to 0.18). The mean for Mt was 52 (range 49.5 to 57.5).

DISCUSSION

Our results with cervical cells (six nuclear macromolecule indices) are remarkably like those we obtained with buccal cells 13]. Three minor differences were observed: cervical cells in this study had less TNA and RNA in Pb cells (relative to I-O cells) than observed for those cells in buccal smears: and, during the degeneration o f cervical cells, RNA was found to decrease more rapidly as 1-O cells became I-RP cells than was found for buccal cells. We infer from these surprising shnilarities in trends that the mechanisms involved in cervical and buccal cell development, maturation, senescence, and death must be nearly identical. The autolysis o f the nuclear macromolecules continues as P cells degenerate to form ghost cells (nuclei visible using phase-contrast microscopy but containing little or no nucleic acids detectable using fluorochromes) and anucleate cells. In both cervical and buccal cells, nucleoli degenerate after Pb cells enlarge to form intermediate cells. We infer that groups of genes that are active in basal and parabasal cells cease activity (chromosome condensation) and other groups o f genes involved in nuclear and cytoplasmic degeneration become active (chromosomes decondensation). Another "alternative inference is that chromosome decondensation enables genes for degradative enzymes to become active and these result in the autolysis of all macromolecules: i.< the genes that are active during development and maturation are destroyed as DNA and other macromolecules are degraded. This may be possible since no quiescent state for cervical or buccal epithelial cells has ever been described.

141 We reported nuclear degeneration (using the eight macromolecule indices studied in buccal cells) occurred in onion bulb epidermal cells in two model systems [16] : sequential leaf base senescence; and, apical cell senescence. These plant cells show great shnilarity during senescence to those of the human buccal and cervical tissues. Dhillon and Miksche [17] suggested that sequential leaf senescence in tobacco involved an increase of 25% heterochromatin or loss of euchromatin in older nuclei which resulted in decreased gene expression. Bowen and Lockshin [10] proposed that physiological activities observed in maturing and senescing cells probably involve function specific condensation and decondensation patterns in chromatin with age, The methods o f Dhillon and Miksche [17] and Nagl [18] used to study chromocenters and changes in the number o f these in nuclei during aging may prove useful in future studies o f buccal and cervical cell development, maturation, senescence, and death. The value o f multiple nuclear macromolecule measurement using the indices we used on cervical cells in this study and in buccal cells [3] in the study o f cancer remains to be determined. The recent work by Caspersson et al. [8] and Auer et al. [19], measuring both DNA and total nuclear protein, demonstrated that these two indices provided a background for judging the growth activity o f individual cells. Further, the results with these two indices improved the classification of malignancy grades in individual breast carcinomas (which is of considerable clinical importance). It appears possible that cervical and oral cells can be used to study two major biological problems: normal cell developnrent and senescence ; and, carcinogenesis and malignancy development. REFERENCES 1 M.E. Boon and M.L. Tabbers-Boumeester, Gynecological Cytology, University Park Press. Baltimore, 1980, pp. 1 192. 2 L.G. Koss, Diagnostic Cytology and its Histopathologic Bases, Vol. 1, 3rd edn., J.B. Lippincott Co., Philadelphia, pp. 157-411. 3 P.K. Bhattacharya, V.M. Russo and A.J. Pappelis, Changes in the nucleic acid and protein content in nuclei of human buccal cells. Anal, Quant. Cytol., 5 (1983) 124-128. 4 W. Sandritter, DNA cytophotometry in cellular pathology. Acta Histoehem. Cytochem., 13 (1980) 35-39. 5 G.L. Wied, P.H. Barrels, H.E. Dytch and M. Bibbo, Rapid DNA evaluation in clinical diagnosis. Acta Cytol., 27 (1983) 33-37. 6 I. Nishiya, I. Yoshiaki and M. Sasaki, Nuclear DNA content and the number of Barr bodies in premalignant and malignant lesions of the uterine cervix. Acta Cytol., 25 (1981) 407-411. 7 G. Auer, J. Ono and T.O. Caspersson, Determination of the fraction of G o cells in cytologic samples by means of simultaneous DNA and nuclear protein analysis. Anal. Quant. C),tol., 5 (1983) 1 4. 8 T.O. Caspersson, G. Auer, A. Fallenius and J. Kudynowski, Cytochemical changes in the nucleus during tumour development. Histoch. J., 15 (1983) 337-362. 9 R. Dutu, M. Nedelea, G. Veluda and V. Burculet, Cytoenzymologic investigations on carcinomas of the cervix uteri. Acta Cytol., 24 (1980) 160-166. 10 I.D. Bowen and R.A. Lockshin, Cell Death in Biology and Pathology, Chapman and ttall, New York, 1981, pp. 1-7. 11 C.K. Pappelis, J. Slobin, J. Corvallis, H.D. Detwiler, A.J. Pappelis and G.A. Pappelis, Comparison of buccal and nasal epithelial cells using a new cell development index and quantitativc interference microscopy. Acta Cytol., 20 (1976) 372-374.

142

12 F. Ruch and U. Leeman, Cytofluorometry. In V. Neuhof led.), Molecular Biology, Biochemistr,~ and Biophysics, Micromethods in Molecular Biology, Vol. 14. Springer-Verlag, Berlin. 1~)7~ pp. 329-346. 13 F. Ruch and L. Trappe, A microscope ftuorometer with short-time excitation and electronic shutter control. Funct. Photogr., 16 (1981) 14 16. 14 F. Ruch, Principles and some applications of cytofluorometry. In G.L. Wied and G.I.. Bahr (eds.), Introduction to Quantitative Cytoehemistrv-H, Academic Press, New York, 1970. pp. 431-450. 15 S.K. Curtis and R.R. Cowden. Four fluorochromes for the demonstration and microfiuorometri~" estimation of RNA. Histochemistr3,, 72 (1981 ) 39-48. 16 P.K. Bhattacharya and A.J. PappeLis, Nucleic acid, protein, and protein bound lysine and argininc patterns in epidermal nuclei of the mature and senescing onion bulb. Mech. Ageing Dev.. 21 (1983) 27 36. 17 S.S. Dhillon and J.P. Miksche, DNA changes during sequential leaf sene~ence of tobacco (Nieotiana tabacum ). Physiol. Plant., 51 (1981) 291-298. 18 W. Nagl, Nuclear ultrastructure: condensed chromatin in plants is species specific (karyotypical~, but not tissue specific (functional). Protoplasma, 100 (1979) 53 -71. 19 G. Auer, J. Ono and T.O. Caspersson, Cytochemical identification of quiescent and growthactivated tumor cells. Anal. Quant. C~vtol.. 5 (1983) 5--8.