703

We have shown that 90% of responding patients have We thank Dr Joe Tomita and Dr Louise Przywara from Diagnostic
ER-positive cells in aspirates; by contrast only 39% of Division, Abbott Laboratories, Chicago, USA, for interest and support; and
Dr Elwood Jensen for encouragement and advice.
non-responders had ER-positive cells. Thus, our method is
as predictive as the standard biochemical assay.! Correspondence should be addressed to R. C. C., Clinical Oncology
A particularly important advantage of the technique is Group, Room 122, Level 2, Jenner Wing G1, St George’s Hospital Medical
School, Cranmer Terrace, Tooting, London SW17 ORE.
that we can detect the receptor in samples from patients who
have recently discontinued tamoxifen therapy: this has not
been possible before since with the standard assay false REFERENCES
negatives are often obtained because tamoxifen binds to the 1. McGuire WL, De La Garza M. Improved sensitivity in the measurement of estrogen
available receptor sites, thus blocking the binding of tritiated receptor in human breast cancer. J Clin Endocrinol Metab 1973, 37: 986-89.
2. King WJ, Greene GL. Monoclonal antibodies localize estrogen receptor in the nuclei
oestradiol.’7
of target cells Nature 1984; 307: 745-47.
Some samples contained too few cells and thus we could 3. McClelland RA, Berger U, Miller LS, Powles TJ, Coombes RC.
not adequately assess receptor status. The repeated Immunocytochemical assay for estrogen receptor in patients with breast cancer:
relationship to a biochemical assay and to outcome of therapy J Clin Oncol 1986; 4:
procedure, since it does not cause much pain, is acceptable to 1171-76
most patients, and we now recommend that two aspirates 4. McClelland RA, Berger U, Wilson P, et al. Presurgical determination of estrogen
receptor status using immunocytochemically stained fine needle aspirate smears in
are done if immediate cytological confirmation is not
patients with breast cancer Cancer Res (in press).
available. The number of unassessable samples would then 5. Hayward JL, Rubens RD, Carbone PP, Heuson J-C, Kumaoka S, Segaloff A.
Assessment of response to therapy in advanced breast cancer. A project of the
be reduced.
programme on clinical oncology of the International Union Against Cancer,
Thus, the simple technique of fine-needle aspiration for Geneva, Switzerland. Br J Cancer 1977; 35: 292-98.
6. Stemberger LA, Hardy PH, Cuculis JJ, Meyer HG. The unlabeled antibody enzyme
immunocytochemical detection of ER provides a useful test method of immunohistochemistry. Preparation and properties of soluble antigen-
for the prediction of breast cancer response to endocrine antibody complex (horseradish peroxidase-antihorseradish peroxidase) and its use
in identification of spirochetes. J
Histochem Cytochem 1970; 18: 315-33.
therapy. Immunocytochemical tests for other antigens, such 7 Taylor RE, Powles TJ, Humphreys J, et al. Effects of endocrine therapy on
as the progesterone receptor and growth factors, should be
steroid-receptor content of breast cancer. Br JCancer 1982; 45: 80-85.
possible with this technique. Furthermore, the receptor 8. Mansi JL, Berger U, Easton D, et al Bone marrow micrometastases in patients with
primary breast cancer. an early predictor of bone metastases. Br Med J (in press).
content in the normal breast and other metastatic sites such 9. Berger U, Mansi JL, Wilson P, Coombes RC. Detection of estrogen receptor in bone
as liver and bone marrow could also be assessed.8,9 marrow from patients with metastatic breast cancer. J Clin Oncol (in press).

HUMAN PAPILLOMAVIRUS INFECTIONS IN in-situ hybridisation underestimates the total rate of HPV
WOMEN WITH AND WITHOUT ABNORMAL infections by a factor of 2 to 3.
CERVICAL CYTOLOGY
Introduction
E.-M. DE VILLIERS1 D. WAGNER2 EXPERIMENTAL data suggest that specific types of HPV
A. SCHNEIDER3 H. WESCH4 infection influence the development of human genital
H. MIKLAW5 J. WAHRENDORF6 cancer. Types 16 and 18, and also types 31, 33, and 35, have
U. PAPENDICK7 H. ZUR HAUSEN1 been identified in biopsy specimens from cervical, penile,
and vulvar cancer. The viral DNA becomes integrated into
Refereuzzentrum für humanpathogene Papillomviren, Deutsches
Krebsforschungszentrum, Heidelberg;1 Evangelisches the host celF and expresses specific functions3,4 such as the
Diakoniekrankenhaus, Freiburg,2 Labor für Gynäkologische synthesis of viral proteins.There is increasing evidence that
Zytologie und Histologie, Universitäts-Frauenklinik, Ulm;3 Institut disturbances in the complex interaction between specific
fur Nuklearmedizin, Deutsches Krebsforschungszentrum, cellular genes and persisting viral DNA influence cervical
Heidelberg;4 Salem Krankenhaus, Heidelberg;5 Institut für carcinogenesis.6 Existing data suggest that HPV type affects
Epidemiologie und Biometrie, Deutsches Krebsforschungszentrum, risk of maligant change in HPV infection, low risk being
Heidelberg;6 and Dr K. Thomae Pharmaceuticals, Biberach7 associated with types 6 and 11 and high risk with types 16
and 18.’
Summary 9295 smears, obtained from women
There is little information on the distribution of HPV
attending three gynaecological hospitals infection in women without cytologically or colposcopically
for routine screening, were examined for human
visible dysplastic lesions or carcinomas. A high prevalence of
papillomavirus (HPV) types 6 and 11 and HPV 16 and 18 HPV 16 infections has been reported in healthy women,
infections by filter in-situ hybridisation. The data were
especially in those aged > 40 years.’ We have examined
compared with cytological findings. In women with normal smears obtained from three different regions in southern
cytological smears HPV infection was identified in about Germany for the presence of HPV infection by filter in-situ
10% of women aged between 15 and 50 years and in less
than 5% of those aged over 50. In women with abnormal hybridisations,8 and report here the distribution of HPV
infection in various age-groups, the correlation between
smears (cervical intraepithelial neoplasia [CIN] I, II, and
HPV infection and cervical neoplasia in relation to the age of
III and invasive cancer) HPV infection was detected in
the affected women, and the sensitivity of the test system
35-40%; this rate seemed to be age-independent. The peak used.
incidence of CIN appeared several years after that of HPV
infection. In women aged > 30 years it also declined earlier Subjects and Methods
than did HPV positivity. The age-group distribution of
The subjects were 9295 patients attending the outpatient clinics
women with CIN I, II, and III differed significantly from
of the Evangelische Krankenhaus, Freiburg, the Frauenklinik of
that of patients with invasive cancer. Only about a third of the University ofUlm, and the Salem-Krankenhaus in Heidelberg,
HPV-positive patients remained virus-positive, probably mostly for a routine gynaecological examination.
because of fluctuations in virus production and the A cotton swab was used to obtain a combined endocervical and
insensitivity of the test system used. It is possible that filter ectocervical sample. When a regular cytological smear had been
704

TABLE CYTOLOGICAL FINDINGS IN RELATION TO AGE

prepared the cotton swab was put into 5 ml of phosphate buffered finding, and as positive for an HPV infection if the
saline, frozen, and then transported to the HPV-laboratory at the hybridisation results were positive on one occasion. Follow-
Deutsches Krebsforschungszentrum, Heidelberg.
up examinations 3 to 6 months later of patients initially
After being thawed, the suspension was mixed on a ’Vortex’
machine and filtered - through a nitrocellulose filter. After HPV-positive revealed only approximately 30% positivity
denaturation, neutralisation, and baking of the cell residue, the filter (data not shown). Hybridisation results were given as
was cut in half. One half was hybridised with a combination of positive, negative, and intermediate. It was not possible to
radiolabelled HPV 6 and 11 (HPV 6/11) DNA and the other with determine whether intermediate hybridisation results were
radiolabelled HPV 16 and 18 (HPV 16/18) DNA.8 Only conditions due to very low copy numbers of papilloma viral DNA or to
of high stringency (T m - 18OC) were used for hybridisation and viral non-specific binding of radioactivity to the filter because of
DNA without vector DNA was used for labelling. After being the presence of material such as mucus or blood. The
washed (under stringent conditions, ie, TJ the filters were
exposed to X-ray film for 5 days and the results were without prior
knowledge of the cytological findings.

Results
94% (8755) patients had normal smears. 2% (196) were
classified as having koilocytotic HPV infection, 2% (162)
showed signs of mild or moderate dysplasia (cervical
intraepithelial neoplasia [CIN] I or II), 1% (120) had
carcinoma in situ (CIN III), and in 1 % (62) had an invasive
carcinoma (table I). When filter halves were hybridised with
HPV 6/11 or HPV 16/18 DNA some filters were positive for
one or other of the groups of viruses and the rest were

positive for both groups. The latter require cautious

interpretation since in large quantity virus from one group
may cross-react with members of the other group even
under stringent conditions of hybridisation. The figures we
give below do not discriminate between HPV 6/11 and HPV
16/18 infections, although HPV 16/18 infections were
detected more often than those with HPV 6/11.
25% of the patients were examined more than once (table
II) for reasons varying from routine diagnostic tests to
re-examination of exisiting lesions. These patients were
classified according to the most abnormal cytological

TABLE II-CYTOLOGICAL DIAGNOSIS AMONG PATIENTS WITH

MORE THAN ONE SMEAR

Distribution ofHPV positivity by age-group among women (A) with

normal smear cytology; (B) with CIN; (C) with invasive
carcinoma (percentage of all the smears examined); and (D) with
CIN or invasive carcinoma (percentage of all smears with CIN or
Findings given as %. invasive carcinoma cytology).
 705

TABLE III-HPV DNA POSITIVITY BY IN-SITU HYBRIDISATION IN

contained HPV 16 or HPV 18 DNA or closely related virus
RELATION TO AGE AND CYTOLOGICAL DIAGNOSIS
types.1 We detected HPV DNA in only 40% of smears from
invasive cancer patients. An obvious explanation for this
discrepancy is that the method we used8 was less sensitive
than Southern blot hybridisation. However, the higher
degree of positivity that we found in patients with abnormal
cytology than in those with normal smears (even though
patients with abnormal cytology were examined more often
than those without detectable changes) is in keeping with
results obtained by Southern blot analysis.1
Approximately 10% of patients with normal smears were
positive for HPV DNA. We do not know whether this
results from virus production in individual cervical cells or
from latent infection with HPV DNA in a large number of
cells. In view of the apparently low sensitivity of the filter
hybridisation assay, it is likely that our results ’are an
underestimate of the real rate of infection, a conclusion
further supported by a report that 28% of pregnant women
have positive smears.16 In pregnant women hormonal
Numbers in parentheses refer to percentages.
stimulation was thought to encourage virus production.
In addition, only about 30% of women who were positive at
intermediate category was considered to be negative in the the first examination remained positive in subsequent
examinations in our study. Similarly, some patients who
analysis that follows. were negative at the first test became positive later. The most
Patients initially negative by cytology but positive for
HPV DNA were often diagnosed as having koilocytotic likely explanation for these conversions is fluctuations in
HPV infection at subsequent cytological examinations. virus production because the changes have also been noted
Since the hybridisation results from previous tests had been with highly sensitive detection procedures (Gissniann L,
communicated to the cytologists, we excluded the category Schneider A, unpublished). The factors discussed here
koilocytotic HPV infection from further analyses in case the suggest that our results probably underestimate the
real rate of HPV infection by a factor of
findings were biased. 2 to 3. However, careful validation studies with different
Out of 8755 patients with negative cytological smears, 9%
were positive for HPV DNA. There seemed to be an
methods of HPV analysis are needed to substantiate this
age-dependent pattern of HPV-positivity (figure and table possibility. The high prevalence of HPV again indicates that
a papillomavirus infection is not the sole factor responsible
III). The highest percentage of HPV-positive patients
for the transformation of normal cervical tissue to a
(10-13%) occurred in age-groups between 15 and 50 years.
Thereafter this percentage declined, to 2-5% in age-groups malignant tumour. Other factors, perhaps those that interact
above 55 years. This drop may reflect a decrease in virus with control mechanisms within the infected cell,6 may be
shedding with age, which may be related to hormonal required.
We thank Miss Oanh Nguyen for technical assistance and Dr Volker
changes in post-menopausal women, or to a low exposure to Schneider for reading the manuscript. This research was supported in part by
reinfections. the Bundesministerium fur Jugend, Familie, Frauen und Gesundheit, Bonn.
The peak prevalence of HPV positivity preceded the peak
Correspondence should be addressed to E.-M. deV.
prevalence of CIN smears and HPV positivity among CIN REFERENCES
patients was commonest among those aged 20-30 years 1. zur HausenH, Schneider A. The role of papillomaviruses in human anogenital cancer.
(figure and table III). Among those aged above 30 years the In: Salzmann NP, Howley PM, eds. The Papovavindae, Vol 2. The
rate of CIN-positivity declined more rapidly than did papillomaviruses. New York: Plenum, 1987: 245-63.
2. Durst M, Schwarz E, Gissmann L, Rowekamp W Integration and persistence of
HPV-positivity. Most invasive cancer smears came from human papillomavirus DNA in genital tumours. Banbury Rep 1986; 21: 273-80.
patients aged 55 to 80 (figure and table ill). 3. Schwarz E, Schneider-Gadicke A, Roggenbuck B, Mayer W, Gissmann L, zur
When CIN and invasive carcinoma patients were Hausen H. Expression of human papillomavirus DNA m cervical carcinoma cell
lines. Banbury Rep 1986; 21: 281-90.
combined 35-40% were HPV positive; this rate showed no 4. Schwarz E, Schneider-Gadicke A. Expression of human papillomavirus type 18 DNA
consistent pattern of age-dependence (figure). This high in cervical carcinoma cell lines. Cancer cells, DNA tumor viruses. Cold Spring
Harbor. Cold Spring Harbor Laboratory, 1986; 4: 581-87.
prevalence of HPV positivity compared with that among 5. Seedorf K, Olterdorf T, Krammer G, Rowekamp W. Identification of early proteins
patients without cytological abnormalities seems to argue for of the human papillomavirus type 16 (HPV 16) and type 18 (HPV 18) in cervical
carcinoma cells. EMBO J 1987; 6: 139-44.
a specific association between HPV infection and 6. zur Hausen H. Intracellular surveillance of persisting viral infections. Lancet 1986; ii:
proliferative changes and supports experimental data on 489-91.
human cell transformation and immortalisation by HPV 7. Meanwell CA, Cox MF, Blackledge G, Maitland NJ. HPV 16 DNA in normal and
malignant cervical epithelium: implications for the etiology and behaviour of
infections (refs 9 and 10 and Diirst M, Dzarlieva- cervical neoplasia. Lancet 1987; i: 703-07.
Petrusevska RT, Boukamp P, Fusenig NE, Gissmann L, 8. Wagner D, Ikenberg H, Bohm N, Gissmann L. Type-specific identification of human
papillomavirus in cervical smears by DNA m situ hybridization. Obstet Gynecol
unpublished). 1984; 6: 767-72.
9. Kreider JW, Howett MK, Wolfe SA, et al. Morphological transformation in vivo of
Discussion human uterine cervix with papillomavirus from condylomata acuminata. Nature
1985; 317: 639-41.
Previous reportsll-ls on the prevalence of HPV DNA in 10. Pirisi L, Yasumoto S, Feller M, Doniger J, DiPaolo JA. Transformation of human
fibroblasts and keratinocytes with human papillomavirus type 16 DNA. J Virol
normal cervical tissue were based mostly on small numbers
1987; 61: 1061-66.
of patients. More than 9000 patients were included in this 11. Burk RD, Kadish HS, Calderin S, Romney SL. Human papillomavirus infection of
the cervix detected by cervicovaginal lavage and molecular hybridization:
study. Correlation with biopsy results and Papanicolaou smear. Am J Obstet Gynecol
Hybridisation data based on Southern blot analysis have 1986; 154: 982-89.
shown that up to 80% of cervical carcinoma biopsies
706

DNA POLYMORPHISM OF HUMAN enzyme deficiency will have symptoms. So far, detection of
PORPHOBILINOGEN DEAMINASE GENE IN gene carriers and avoidance of precipitants is the most
ACUTE INTERMITTENT PORPHYRIA effective way to manage AIP.
At present, determination of erythrocyte PBG-deaminase
D. H. LLEWELLYN G. H. ELDER
N. A. KALSHEKER O. W. M. MARSH activity is the best method to detect gene carriers,5,6 but this
P. R. HARRISON B. GRANDCHAMP approach has limitations imposed by overlap of enzyme
activities between patients and controls,’ dependence of
C. PICAT Y. NORDMANN
P. H. ROMEO enzyme activity on red cell age,8 and the apparent restriction
M. GOOSSENS
of some AIP mutations to the liver.9 These difficulties are
Department of Medical Biochemistry, University of Wales College likely to be resolved by detection of mutations at the DNA
of Medicine, Cardiff; Beatson Institute for Cancer Research, level.
Bearsden, Glasgow; Molecular Genetics Laboratory, Faculty of Two isoenzymes of PBG-deaminase, which differ in their
Medicine Xavier Bichat, Paris, France; Biochemistry Laboratory, N-terminal aminoacid sequences, have been identified in
Hôpital Louis Mourier, Colombes, France; and INSERM U-91, mammals; one is present in erythroid and the other in
Hôpital Henri Mondor, Créteil, France non-erythroid tissues. A single gene codes for both
two-allele isoenzymes. In man, it is situated on the long arm of
Summary A common MspI restriction chromosome 1111 and seems to be transcribed from two
fragment length polymorphism of the different promoters to produce erythroid and non-erythroid
human erythroid porphobilinogen (PBG)-deaminase gene
mRNAs that differ at their 5’ ends.l° The alterations in
was investigated in 33 unrelated patients with acute
DNA sequence that underlie AIP have not been identified,
intermittent porphyria (AIP) and 20 controls. The
but immunochemical studies of the erythrocyte enzyme
polymorphism was tightly linked (lod score 3.14; no have revealed crossreacting immunological material
recombinants) to the locus for AIP as identified by (CRIM)-negative and CRIM-positive mutations,9,12 the
measurement of erythrocyte PBG-deaminase activity. The
former being present in about 85% of affected families.9
frequency of the polymorphism in the AIP patients did not We have investigated an Msp restriction fragment length
differ significantly from that in the controls. No common
polymorphism (RFLP)13 of the human PBG-deaminase
polymorphisms for eight other restriction endonucleases gene in AIP.
were found in either group. In 30 of the AIP patients no

crossreacting immunological material (CRIM) was .

Subjects and Methods
produced by the mutant PBG-deaminase allele. The MspI Subjects
polymorphism enabled each PBG-deaminase allele to be
distinguished in subjects heterozygous for the 33 unrelated patients with AIP were investigated. All had
decreased erythrocyte PBG-deaminase activities (mean 22 nmol h
polymorphism; thus a major gene deletion was excluded as
the cause of the CRIM-negative mutation in all of the 18 ml-1, range 14-28); 32 presented with acute porphyria and raised
families that contained an affected CRIM-negative urinary PBG excretion and 1 was symptomless, but his father died
from AIP. The families of 18 of these patients were investigated. All
individual heterozygous for the polymorphism. In suitable controls were unrelated and selected at - random from
families, the MspI polymorphism provides a more certain haematologically normal patients. Analysis of DNA and other
way of identifying carriers of the AIP gene than current measurements were done in Cardiff (21 patients, 11 families) or in
enzymatic methods and major gene deletions are unlikely to Paris (12 patients, 7 families).
be present in more than a small proportion of the
commonest type of AIP, the CRIM-negative form. DNA-hybridisation Analysis
High molecular weight DNA was prepared from white blood
Introduction cells14 and digested with restriction endonucleases (BRL). Upon
completion of digestion, the DNA was precipitated with ethanol
ATTACKS of intermittent porphyria (AIP) severe
acute
and resuspended in loading buffer, and electrophoresis was done
enough require hospital admission occur in about 1 in
to
through an agarose gel. The DNA was then transferred overnight
75 000 of the population, have a mortality rate of around onto ’Zeta’-probe membranes (Bio-rad) in alkaline conditions (0-4
5%, and may lead to long-term morbidity.i-4 Most attacks mol/1 NaOH). A 7 kb EcoRI fragment containing the section of the
are precipitated by drugs, alcohol, calorie restriction, or gene encoding the erythroid enzyme was used as the hybridisation
endocrine factors. The condition results from a partial probe (fig 1). Probes were labelled to a high specific activity (greater
than 1 x 109 cpm/jig) with the random priming method.15 Filters
deficiency of one of the enzymes of haem biosynthesis,
were hybridised with the probe at 68°C for 16 h in a high-stringency
porphobilinogen (PBG)-deaminase (EC 4.3.1.8), which is buffer containing 10% (w/v) dextran sulphate. After hybridisation,
inherited as an autosomal dominant trait. In the absence of
the filters were removed and washed extensively at high stringency
precipitants probably less than 20% of those who inherit the before being autoradiographed at - 70°C for 24-72 h with
intensifying screens. Autoradiographs were scanned with an LKB
’Ultrascan XL’ laser densitometer."
12 MacNab JC, Walkinshaw SA, Cordmer JW, Clements JB. Human papillomavirus in
clinically and histologically normal tissue of patients with genital cancer. N Engl J Mapping of the Polymorphic MspI Site
Med 1986; 351: 1052-58
The position of the polymorphic Mspl site was defined by both
13 Schneider A, Schuhmann R, de Villiers E.-M, Knauf W, Gissmann L Klinische
Bedeutung von humanen Papilloma-Virus- (HPV)-Infektionen im unteren
restriction mapping experiments and Southern blotting analysis
Genitaltrakt Geburtsch Frauenheill 1986, 6: 261-66. with fragments of the 7 kb EcoRl fragment (fig 1) as probes.
14 Toon PG, Arrand JR, Wilson JW, Sharp DS. Human papillomavirus infection of the
uterine cervix of women without cytological signs of neoplasia. Br Med J 1986; 293: PBG-deaminase Measurements
1261-64
15 Wickenden C, Steele A, Malcolm AD, Coleman DV Screening for wart virus PBG-deaminase activity was measured by a modification of the
infection in normal and abnormal cervices by DNA hybridisation of cervical
method of Magnusson et al.16 Between-batch coefficient of variation
scrapes. Lancet 1985; 1: 65-67.
16. Schneider A, Hotz M, Gissmann L. Increased prevalence of human papillomaviruses was 78%. The concentration of immunoreactive PBG-deaminase
in the lower genital tract of pregnant women Int J Cancer (in press) was measured by a radiochemical electroimmunoassay or an