1543
this small study, classification of HIV disease was not considered.3 The mechanism whereby didanosine causes reduced resting salivary flow is unknown and we proposed one possible explanation. Valentine’s suggestion is also possible. In Valentine’s study of patients on didanosine with increased serum amylase,4 there was no consistent evidence for an association between increased salivary amylase and subjective xerostomia. In the rheumatological features described by Kame,s and cited by Valentine et al, xerostomia alone was not described directly as a feature of HIV infection but as part of a sicca syndrome. Another study mentioned by Valentine was done to assess whether stage of HIV infection or didanosine was related to serum amylase concentration and not to subjective or objective measurements of xerostomia.6 We emphasise that our study was designed as a preliminary investigation to obtain objective measurements of salivary flow in patients taking didanosine and reporting subjective xerostomia. We can conclude that didanosine does affect salivary flow rates by objective as well as subjective assessment, although the exact mechanism of action is not yet understood. Again, we emphasise that didanosine recipients who have xerostomia should be identified: they should be evaluated for salivary disease, and symptomatic treatment should be given as appropriate. Oral AIDS Center, Department of Stomatology, University of California San Francisco, San Francisco, California 94143, USA
CAROLINE L. DODD DEBORAH GREENSPAN JANICE L. WESTENHOUSE MITCHELL H. KATZ
1. Schiødt
M, Greenspan D, Daniels TE, et al. Parotid gland enlargement and xerostomia associated with labial sialoadenitis in HIV-infected patients. J Autoimmun 1989; 2: 415-25. 2. Schiødt M. HIV-associated salivary gland disease: a review. Oral Surg Oral Med Oral Pathol 1992; 73: 164-67. 3. Centers for Disease Control: Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. MMWR 1987; 36 (suppl 1S): 3S-15S. 4. Valentine C, Sherwood R, Deenmamode J. Didanosine and amylase monitoring. Lancet 1992; 339: 999. 5. Kame BR. Rheumatologic manifestations of infection with human immunodeficiency virus (HIV). Ann Intern Med 1989; 111: 158-67. 6. Sweeney J, Valentine C, Gompels M, et al. Nucleoside analogues and patterns of salivary, pancreatic and total amylase in HIV-positive individuals. Presented at VIII International Conference on AIDS, Amsterdam, July 1992 (abstr MoB
0078).
In-vitro
is unreliable to study the activity of zidovudine against mycoplasmas in vitro and such observations cannot be extrapolated to the natural biochemical setting. The ideal is a culture medium as closely related to blood as possible. This drug can no longer be considered as the cause of difficulties encountered during isolation of myocplasmas from blood. Fortunately, to detect myocplasmas in HIV-infected individuals and to evaluate the potential role of mycoplasma in the development of AID S,3new methods, such as the PCR, are now available.5 GÉRARD PAPIEROK ALAIN LAFEUILLADE SANDRINE DOMBRECHT GÉRARD PAUTRAT PIERRE PELLEGRINO CLAUDE ESCARGUEL ROBERT QUILICHINI
International Mycoplasma Laboratory, 83870 Signes, France; Department of Internal Medicine,
Hôpital Chalucet, Toulon, and Laboratory of Bacteriology, Hôpital Général, Hyères
1. Lafeuillade A, Papierok G, Pautrat G. In-vitro activity of zidovudine against mycoplasma. Lancet 1992; 339: 131. 2. Taylor-Robinson D, Furr PM. In-vitro activity of zidovudine against mycoplasma. Lancet 1992; 339: 686. 3. Montagnier L, Berneman D, Guetar, et al. Inhibition of HIV prototype strains infectivity by antibodies directed against a peptidic sequence of myocplasma. CR Acad Sci Paris III 1990; 311: 425-30. 4. Lo SC, Tsai S, Benish JR, et al. Enhancement of HIV-1 cytocidal effects in CD4 + lymphocytes by the AIDS-associated mycoplasma. Science 1991; 251: 1074-76. 5. Wang R, Hu W, Dawson MS, et al. Selective detection of Mycoplasma fermentans using the polymerase chain reaction (PCR). 91st general meeting of American Society for Microbiology (Dallas, May 5, 1991); abstr G5.
Is cervical
cancer
monoclonal?
SIR,--Genetic, virological, and molecular biological studies of cancer are usually interpreted as showing that carcinogenesis proceeds from a single transformed cell or from an infected complex of cells. This interpretation does not explain simple light-microscopic studies of the morphogenesis of cancer, especially cervical cancer. Reports that cancers of the gastrointestinal tract and skin develop in circumscribed fields sharply demarcated from healthy epithelium date back to 1900. The same applies to the cervix. the induction of
With A. G. Ostor I have reviewed
more
than 4000 conisation
specimens containing cervical intraepithelial neoplasia (CIN) and processed as serial sections of the entire cone.2,3The goal was to
activity of zidovudine against mycoplasma
SIR,-Earlier this year we reported that growth of mycoplasma is inhibited by zidovudine in vitro. Taylor-Robinson and Furr2 were unable to confirm this so we contacted them and found that we were using exactly the same strains but that the growth media differed. The media used for the culture of mycoplasmas are complex and we considered the possibility of synergy between one of the components of the medium we used and zidovudine. We have tested the activity of zidovudine against two collection strains of mycoplasma (Mycoplasma farmentans strain PG18 and M pyrum strain ATCC 25960) in various broth and agar media. The modified SP4 medium used in our previous experiments was the control. This medium revealed an inhibitory effect by zidovudine. Replacement of the foal serum with fetal calf serum and removal of the ampicillin and co-trimoxazole did not alter the inhibitory activity of zidovudine against mycoplasmas. However, by altering the concentration of CMRL (CMRL 1066 from Gibco), a nutritive compound containing aminoacids, vitamins, and nucleic acids, radically different results were obtained. With a low concentration (05%) of CMRL, we obtained an inhibitory activity of zidovudine; with higher concentrations (5%) mycoplasma growth was promoted. For each experiment, three concentrations of zidovudine (10, 1, and 01ug/ml) were compared with a control tube without zidovudine. The zidovudine enhancement and inhibition effects were demonstrated for all three concentrations. We are now investigating the possibility of competition between zidovudine and nucleic acids. The findings reported here could explain the discordance between the data of Taylor-Robinson and Furr and ours. Clearly it
reconstruct
the localisation of the lesions and their relation
to
the
healthy epithelium. The observations were striking. Reserve cells appear as circumscribed fields in the columnar epithelium. The multiplication of reserve cells usually leads to benign squamous epithelium and only occasionally to CIN. The resulting epithelia occupy the entire field of the reserve cells and are sharply demarcated from the adjacent epithelia. If CIN arises
simultaneously or subsequently in neighbouring fields, there are always sharp borders between morphologically distinct types of CIN. This was confirmed in a study of 152 conisation specimens, in 97% of which CIN was located in the field of cervical glands or in the area of original squamous epithelium. CIN never exceeded the original border between columnar and squamous epithelium that is marked by the so-called last cervical gland.2,3 In 16% of these specimens, CIN was found to have developed within original squamous epithelium by atypical proliferation of basal cells. By contrast with reserve cells, which appear beneath the pre-existing columnar cells, atypical basal cell hyperplasia does not show a sharp demarcation from the onset. However, the endpoint of its development is also clearly confined CIN. Theories on the monoclonal origin of cancer, derived from experimental findings,4,s do not explain these morphological facts. The patterns described above exclude the possibility that a cancer, if it develops via an intraepithelial stage, stems from a single transformed cell or a small complex of cells. If it did, stages of intraepithelial spread would have been seen in cancerous epithelium. However, carcinogenesis always affects all cells of proliferating ability in one or more epithelial segments. If carcinogenesis were due to the loss of a protective factor, such as the p53 proteinthere would have to be a mechanism to induce this loss simultaneously in thousand of cells. And infection would have to
this small study, classification of HIV disease was not considered.3 The mechanism whereby didanosine causes reduced resting salivary flow is unknown and we proposed one possible explanation. Valentine’s suggestion is also possible. In Valentine’s study of patients on didanosine with increased serum amylase,4 there was no consistent evidence for an association between increased salivary amylase and subjective xerostomia. In the rheumatological features described by Kame,s and cited by Valentine et al, xerostomia alone was not described directly as a feature of HIV infection but as part of a sicca syndrome. Another study mentioned by Valentine was done to assess whether stage of HIV infection or didanosine was related to serum amylase concentration and not to subjective or objective measurements of xerostomia.6 We emphasise that our study was designed as a preliminary investigation to obtain objective measurements of salivary flow in patients taking didanosine and reporting subjective xerostomia. We can conclude that didanosine does affect salivary flow rates by objective as well as subjective assessment, although the exact mechanism of action is not yet understood. Again, we emphasise that didanosine recipients who have xerostomia should be identified: they should be evaluated for salivary disease, and symptomatic treatment should be given as appropriate. Oral AIDS Center, Department of Stomatology, University of California San Francisco, San Francisco, California 94143, USA
CAROLINE L. DODD DEBORAH GREENSPAN JANICE L. WESTENHOUSE MITCHELL H. KATZ
1. Schiødt
M, Greenspan D, Daniels TE, et al. Parotid gland enlargement and xerostomia associated with labial sialoadenitis in HIV-infected patients. J Autoimmun 1989; 2: 415-25. 2. Schiødt M. HIV-associated salivary gland disease: a review. Oral Surg Oral Med Oral Pathol 1992; 73: 164-67. 3. Centers for Disease Control: Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. MMWR 1987; 36 (suppl 1S): 3S-15S. 4. Valentine C, Sherwood R, Deenmamode J. Didanosine and amylase monitoring. Lancet 1992; 339: 999. 5. Kame BR. Rheumatologic manifestations of infection with human immunodeficiency virus (HIV). Ann Intern Med 1989; 111: 158-67. 6. Sweeney J, Valentine C, Gompels M, et al. Nucleoside analogues and patterns of salivary, pancreatic and total amylase in HIV-positive individuals. Presented at VIII International Conference on AIDS, Amsterdam, July 1992 (abstr MoB
0078).
In-vitro
is unreliable to study the activity of zidovudine against mycoplasmas in vitro and such observations cannot be extrapolated to the natural biochemical setting. The ideal is a culture medium as closely related to blood as possible. This drug can no longer be considered as the cause of difficulties encountered during isolation of myocplasmas from blood. Fortunately, to detect myocplasmas in HIV-infected individuals and to evaluate the potential role of mycoplasma in the development of AID S,3new methods, such as the PCR, are now available.5 GÉRARD PAPIEROK ALAIN LAFEUILLADE SANDRINE DOMBRECHT GÉRARD PAUTRAT PIERRE PELLEGRINO CLAUDE ESCARGUEL ROBERT QUILICHINI
International Mycoplasma Laboratory, 83870 Signes, France; Department of Internal Medicine,
Hôpital Chalucet, Toulon, and Laboratory of Bacteriology, Hôpital Général, Hyères
1. Lafeuillade A, Papierok G, Pautrat G. In-vitro activity of zidovudine against mycoplasma. Lancet 1992; 339: 131. 2. Taylor-Robinson D, Furr PM. In-vitro activity of zidovudine against mycoplasma. Lancet 1992; 339: 686. 3. Montagnier L, Berneman D, Guetar, et al. Inhibition of HIV prototype strains infectivity by antibodies directed against a peptidic sequence of myocplasma. CR Acad Sci Paris III 1990; 311: 425-30. 4. Lo SC, Tsai S, Benish JR, et al. Enhancement of HIV-1 cytocidal effects in CD4 + lymphocytes by the AIDS-associated mycoplasma. Science 1991; 251: 1074-76. 5. Wang R, Hu W, Dawson MS, et al. Selective detection of Mycoplasma fermentans using the polymerase chain reaction (PCR). 91st general meeting of American Society for Microbiology (Dallas, May 5, 1991); abstr G5.
Is cervical
cancer
monoclonal?
SIR,--Genetic, virological, and molecular biological studies of cancer are usually interpreted as showing that carcinogenesis proceeds from a single transformed cell or from an infected complex of cells. This interpretation does not explain simple light-microscopic studies of the morphogenesis of cancer, especially cervical cancer. Reports that cancers of the gastrointestinal tract and skin develop in circumscribed fields sharply demarcated from healthy epithelium date back to 1900. The same applies to the cervix. the induction of
With A. G. Ostor I have reviewed
more
than 4000 conisation
specimens containing cervical intraepithelial neoplasia (CIN) and processed as serial sections of the entire cone.2,3The goal was to
activity of zidovudine against mycoplasma
SIR,-Earlier this year we reported that growth of mycoplasma is inhibited by zidovudine in vitro. Taylor-Robinson and Furr2 were unable to confirm this so we contacted them and found that we were using exactly the same strains but that the growth media differed. The media used for the culture of mycoplasmas are complex and we considered the possibility of synergy between one of the components of the medium we used and zidovudine. We have tested the activity of zidovudine against two collection strains of mycoplasma (Mycoplasma farmentans strain PG18 and M pyrum strain ATCC 25960) in various broth and agar media. The modified SP4 medium used in our previous experiments was the control. This medium revealed an inhibitory effect by zidovudine. Replacement of the foal serum with fetal calf serum and removal of the ampicillin and co-trimoxazole did not alter the inhibitory activity of zidovudine against mycoplasmas. However, by altering the concentration of CMRL (CMRL 1066 from Gibco), a nutritive compound containing aminoacids, vitamins, and nucleic acids, radically different results were obtained. With a low concentration (05%) of CMRL, we obtained an inhibitory activity of zidovudine; with higher concentrations (5%) mycoplasma growth was promoted. For each experiment, three concentrations of zidovudine (10, 1, and 01ug/ml) were compared with a control tube without zidovudine. The zidovudine enhancement and inhibition effects were demonstrated for all three concentrations. We are now investigating the possibility of competition between zidovudine and nucleic acids. The findings reported here could explain the discordance between the data of Taylor-Robinson and Furr and ours. Clearly it
reconstruct
the localisation of the lesions and their relation
to
the
healthy epithelium. The observations were striking. Reserve cells appear as circumscribed fields in the columnar epithelium. The multiplication of reserve cells usually leads to benign squamous epithelium and only occasionally to CIN. The resulting epithelia occupy the entire field of the reserve cells and are sharply demarcated from the adjacent epithelia. If CIN arises
simultaneously or subsequently in neighbouring fields, there are always sharp borders between morphologically distinct types of CIN. This was confirmed in a study of 152 conisation specimens, in 97% of which CIN was located in the field of cervical glands or in the area of original squamous epithelium. CIN never exceeded the original border between columnar and squamous epithelium that is marked by the so-called last cervical gland.2,3 In 16% of these specimens, CIN was found to have developed within original squamous epithelium by atypical proliferation of basal cells. By contrast with reserve cells, which appear beneath the pre-existing columnar cells, atypical basal cell hyperplasia does not show a sharp demarcation from the onset. However, the endpoint of its development is also clearly confined CIN. Theories on the monoclonal origin of cancer, derived from experimental findings,4,s do not explain these morphological facts. The patterns described above exclude the possibility that a cancer, if it develops via an intraepithelial stage, stems from a single transformed cell or a small complex of cells. If it did, stages of intraepithelial spread would have been seen in cancerous epithelium. However, carcinogenesis always affects all cells of proliferating ability in one or more epithelial segments. If carcinogenesis were due to the loss of a protective factor, such as the p53 proteinthere would have to be a mechanism to induce this loss simultaneously in thousand of cells. And infection would have to
