1187

The recent spread of slim disease (AIDS) not only from Tanzania(as suggested by Serwadda et al) but also from Rwanda to southernUanda and Kampala must be considered.

Department of Pathology, Gesellschaft für Strahlen undUmweltforschung mbH München,

D-8042 Neuherberg, West Germany

Institute of Clinical Chemistry,Katharinenhospital, Stuttgart

Rubaga Hospital,Kampala,Uganda

K.-H. MARQUART

H. A. G. MÜLLER

J. SAILERR. MOSER

1 Göbel E. Kryptosporidiose bei Mensch und Tier: Nachweis und Intestinalentwicklungdes Parasiten Mitt Österr Ges Tropenmed Parasitol 1983; 5: 49-53.

2 Francioli P, Clément F, Glauser MP Beta-2 microglobulin: A sensitive non-specificmarker of acquired immune deficiency syndrome. Eur J Clin Microbiol 1984, 3:68-69.

3 Fuchs D, Hausen A, Reibnegger G, et al. Urinary neopterin in the diagnosis of acquiredimmune deficiency syndrome. Eur J Clin Microbiol 1984; 3: 70-71.

4 Piot P, Quinn TC, Taelman H, et al. Acquired immunodeficiency syndrome in aheterosexual population in Zaire. Lancet 1984, ii: 65-69

5. Van de Perre P, Rouvroy D, Lepage P, et al. Acquired immunodeficiency syndrome inRwanda Lancet 1984; ii: 62-65.

LIVER DISEASE IN HAEMOPHILIA

SIR,-Professor Mannucci and Dr Colombo (Oct 5, p 774)contrast our report’ of a high incidence of severe and sometimesprogressive liver disease in a group of 34 haemophiliacs with theirstudy of 10 patients.2 They speculate that age (mean 32 years in ourseries, 12 years in theirs), and possibly also delta infection, mightaccount for some of the differences between our findings. ,Age may be a factor. The relation between age and histological

diagnosis in our patients was examined by analysis of variance, andthere was a significant difference between the mean ages of patientswith chronic persistent hepatitis (CPH), chronic active hepatitis(CAH), and cirrhosis (25, 35, and 48 years respectively; p<0’01).Older patients may, in some cases, merely have had longer for thenatural history of their liver disease to unfold. On the other hand,since most of our 34 patients of all ages probably contractedhepatitis in 1973-76, a variation in host-response to viral infectionwith age may also be important.A crucial difference in patient selection also underlies the

differences between our results since all our patients had

persistently abnormal transaminases at the time of biopsy. Incontrast, 8 of the 10 patients described by Mannucci et al hadnormal or intermittently abnormal transaminases at the time of theirfirst follow-up biopsy. We believe that haemophiliacs with

persistently abnormal liver biochemistry may be affected morefrequently by progressive and severe liver disease than the rest.When discussing our results, we therefore related the number withsevere liver disease, not to the 34 on whom biopsies were done, butto the 79 patients from whom they had been selected. We were thusable to derive a frequency for cirrhosis of at least 11’5%, forexample. Physical signs and liver tests on the remaining patients notinvestigated by biopsy suggest that cirrhosis awaits histologicalconfirmation in a further 5 or 6 patients. The frequency of cirrhosismay thus be more than 15%, a figure in close accord with that from aretrospective study of 155 unselected liver biopsy and necropsyspecimens collected worldwide by Aledort et a1. Even this studymay underestimate the incidence of cirrhosis in haemophilia sincemany of the liver biopsies reviewed dated from the mid-1970s oreven earlier. The incidence of acute hepatitis in haemophiliaincreased rapidly in 1973-75, with the introduction ofconcentrates, and the time scale for the development of liver diseasedated from then for most patients. Cirrhosis may take many years todevelop and so many patients with biopsy evidence of CPH or CAHm the 1970s will now have progressed to cirrhosis.’ Non-haemophiliacs followed up after post-transfusional non-A, non-Bhepatitis (NANB) are similarly affected, cirrhosis arising in15-25o of individuals. 4-6In contrast with the experience in the Italian population, neither

delta infection nor hepatitis B is a major problem among ourpatients, since only 2 are HBsAg-positive. The delta antigen hasbeen detected in 6 of our 34 liver biopsy specimens and this has notbeen associated with progressive disease. In our population, NANB

hepatitis is the predominant cause ofhaemophilic liver disease, andthis gives rise to serious and sometimes progressive liver disease in asubstantial minority of patients.Mannucci and Colombo cite several reports, all dating from 1976,

stating that liver disease was an insignificant cause of death inhaemophiliacs. A decade later this statement warrants re-

examination.

We thank Prof M. Rizzetto for testing our specimens for the delta antigen.

University Departments of Haematology,Medicine, and Pathology,

Royal Hallamshire Hospital,Sheffield S10 2JF

C. R. M. HAYF. E. PRESTOND. R. TRIGER

J. C. E. UNDERWOOD1. Hay CRM, Preston FE, Triger DR, Underwood JCE. Progressive liver disease in

haemophilia: an understated problem? Lancet 1985; i: 1495-98.2. Mannucci PM, Colombo M, Rizzetto M. Non-progressive course of non A non B

chronic hepatitis in multitransfused hemophiliacs. Blood 1982; 60: 655-58.3. Aledort LM, Levine PH, Hilgartner M, et al. A study of liver biopsies and liver disease

amongst hemophiliacs. Blood 1985; 66: 367-72.4. Realdi G, Alberti A, Rugge M, et al. Long term follow-up of acute and chronic non A

non B post transfusional hepatitis: Evidence of progression to cirrhosis. Gut 1982;23: 270-75.

5. Berman M, Alter HJ, Ishak KG, Purcell RH, Jones EA. The chronic sequelae of non Anon B hepatitis. Ann Intern Med 1979; 91: 1-6.

6. Koretz RL, Stone O, Gitnick GL. The long-term course of non A non B posttransfusional hepatitis. Gastroenterology 1980; 79: 893-98.

COMMON ACTIVATED HELPER-T-CELL ORIGIN FORLYMPHOMATOID PAPULOSIS, MYCOSIS

FUNGOIDES, AND SOME TYPES OF HODGKIN’SDISEASE

SiR.&mdash;Dr Kadin should be congratulated for his attractive

hypothesis linking the above diseases (Oct 9, p 864); indeed, otherresearch groups are developing broadly similar concepts using thesame monoclonal antibodies. Kadin’s hypothesis can already besubstantially expanded using published data, which in part supporthis proposals.MacKiel has suggested that the initial event in mycosis fungoides

(MF) is retrovirus infection of epidermal Langerhans’ cells. C-typevirus-like particles have been identified in the skin and lymph nodesof such patients,2,3 and HTLV has been isolated from a case of theclosely related Sezary syndrome,4 while the HTLV p 19 antigen hasbeen identified in skin and lymph nodes of a case of MF.

5

Although HTLV I and II contain oncogenes and are associatedwith T cell proliferation, HTLV-III contains no oncogenes and ispurely T cell cytotoxic. It is, therefore, reasonable to modifyKadin’s hypothesis by suggesting that in some instances retrovirus-induced T cell immune paresis could permit secondary DNA virusinfection (such as Epstein-Barr and cytomegalovirus) with resultingcellular proliferation and malignancy. Intensive research into thepossible presence of new HTLV subtypes in this group of diseases isurgently required.The similarities between Hodgkin’s disease (HD) and MF are

well illustrated by a histological classification of MF proposed byGoos and Christopher. MF is regarded as showing four sub-types with poikilodermatous, mixed cellularity, lymphocytepredominance, and lymphocyte depleted features. The latter

aggressive subtype displays numerous large atypical cells withbinucleate forms resembling Reed-Sternberg (RS) cells andmorphological distinction from HD is difficult. The differentialdiagnosis is further complicated by the expression of the HD-associated antigen Ki-1 in MF.8 The monoclonal antibody Leu-Ml 1(Becton Dickinson) also reacts with RS cells9 and we havedemonstrated positive staining in some of the RS-type cells of MF(unpublished). The expression of such antigens in advanced formsof MF supports the view that the disease dedifferentiates duringevolution from its usual banal origin. However, these findingsimply that the diagnostic accuracy of saying one disease changesinto another is limited and fraught with problems.Both MF and lymphomatoid papulosis are characterised by

close association between lymphoid and antigen presenting(Langerhans’) cells, but Kadin does not incorporate this area intohis hypothesis. Although classically MF has been regarded as aprimary cutaneous disorder, there is now reasonable evidence that it

1188

is a pathological variant of the cycle involving lymphocytes andLangerhans’ cells migrating between the skin and local lymphnodes.3,IO The vital pivotal role of the lymph node in MF has beenfurther emphasised by the demonstration ofclonal rearrangementsof beta T-cell receptor gene DNA within MF nodes at the earlydermatopathic stage. 1 An altered lymphocyte/Langerhans’ cellinteraction has been suggested in lymphomatoid papulosis,I2 andKadin’s observation of dermatopathic change in this disease may beof pathogenic importance. It may also be significant that when cellsexpressing Ki-1 and Leu-Ml antigens are plentiful (HD, advancedMF, lymphomatoid papulosis type A), antigen-presenting cellsare sparse.3,I2,13 This suggests that normal immune controlmechanisms may have been lost, resulting in cellulardedifferentiation and proliferation. A similar pattern is seen inAIDS when apparent elimination of nodal antigen-presenting cellsby HTLV-III leads to progressive immune deficiency andexacerbation of the disease. Virus-induced alteration of the numberor function of antigen-presenting cells could explain the localisationof HD to lymphoid tissue and the extreme rarity of the primarycutaneous disease.We agree with Kadin that the three diseases have a common

activated helper T-cell origin, but we suggested that alteration in thedermatonodal cycle with abnormal lymphocyte-antigen presentingcell interactions induced by viral infection has a major pathogenicrole.

Sheffield

Department of Pathology,Sheffield University Medical School,Sheffield S10 2RX

Department of Pathology,Royal Hallamshire Hospital,Sheffield

DAVID SLATER

NICHOLAS ROONEY

PATRICIA MILLS

1. MacKie RM. Initial event in mycosis fungoides of the skin is viral infection of

epidermal Langerhans’ cells. Lancet 1981; ii. 283.2. Van der Loo EM, van Muijen GNP, Van Vloten WA, et al. C-type virus particles

specifically localised in Langerhans’ cells and related cells of skin and lymph nodesof patients with mycosis fungoides and S&eacute;zary’s syndrome. Vichows Arch B CellPathol 1979; 31: 193.

3 Slater DN, Rooney N, Bleehen S, Hamed A. The lymph node in mycosis fungoides: alight and electron microscopy and immunohistological study supporting the

Langerhans’ cell-retrovirus hypothesis Histopathology 1985; 9: 587.4. Grillot-Courvalin C, Brouet JC, Chermann JC Functional T4 helper cells in patients

with HTLV related S&eacute;zary syndrome. Lancet 1983, ii. 619-20.5. Turbitt ML, MacKie RM P19 antigen in skin and lymph nodes of patients with

advanced mycosis fungoides. Lancet 1985; ii: 9456 Goos M, Christopher E Clinical, morphological and functional aspects of cutaneous T

cell lymphoma. In: Goos M, Christopher E, eds. Lymphoproliferative diseases ofthe skin Berlin Springer, 1982

7 Harrington CI, Slater DN. Mycosis fungoides with blast cell transformation ArchDermatol 1978, 114: 611

8 Stein H, Mason DY, Gerdes J, et al. The expression of the Hodgkin’s disease associatedantigen Ki-1 in reactive and neoplastic lymphoid tissue. Blood (in press)

9. Hsu S, Jaffe E. Leu M1 and peanut agglutinin stain the neoplastic cells of Hodgkin’sdisease Am J Clin Pathol 1984, 82: 29.

10. Slater DN Lymphoproliferative conditions of skin. In. Anthony PP, MacSweenRNW, eds Recent advances in histopathology Vol XII Edinburgh: ChurchillLivingstone, 1984 83

11. Weiss LM, Hu E, Wood GS, Moulds C, Cleary ML, Warnke R, Sklar J Clonal

rearrangements of T-cell receptor genes in mycosis fungoides and dermatopathiclymphadenopathy. N Engl J Med 1985, 313: 539.

12. Willemze R, Scheffer E, Ruiter DJ, Van Vloten WA, Meijer CJLM. Immunological,cytochemical and ultrastructural studies in lymphomatoid papulosis. Br J Dermatol1983; 108: 381

13 Rooney N, Underwood JCE, Slater DN. Langerhans’ cell in lymphoid andnonlymphoid tissue. J Pathol 1985; 145: 94A.

SIR,-We have seen the clinical development, accompanied bydiagnostic histological changes, of MF in a patient with establishedlymphomatoid papulosis of 10 years’ duration. In cryostat sectionsfrom the MF lesions cells bore positive markers for the activated Thelper cell (CD2 [DAKO TII E rosette receptor] CD4 [Leu 3ax 3bhelper inducer phenotype]). They were negative for the Leu-MImarker which defines cells of myeloblastic origins and reacts withthe Reed-Sternberg (RS) cells of HD.1 Kadin and colleagues,2besides reporting the positive staining of the characteristic

lymphomatoid papulosis cell with the Ki-1 monoclonal antibody(reacting with the RS cell), have indicated that a high percentage ofsuch cells bear the Leu-MI marker. Unlike Ki-1, Leu-MI is active inparaffin embedding tissues, 3and this enabled us to examine a series

of previous biopsy specimens from this patient taken before thedevelopment of MF. In this retrospective study we ensuredexclusion of false positives to myelomonocytic cells by histological,histocytochemical, and double antibody controls and havedemonstrated consistent Leu-MI positivity at the time of

unequivocal lymphomatoid papulosis with striking loss of activityon transformation to MF.We subsequently examined a further four cases of lymphomatoid

papulosis, all of whom have been Leu-MI positive from the earlieststages of the disease. In a further case Leu-MI positivity was lostwith increasingly aggressive disease. Our experience supportsKadin’s hypothesis about the possible development of MF from aclone of lymphomatoid papulosis T helper cells.Given the known clinical associations of the disorders the

remainder of Kadin’s hypothesis is attractive. However, the

concept that the Ki-1 (and presumably analogously the Leu-MI) RScells represent transformed "cerebrifbrm helper cells requirescautious evaluation. The hypothesis is clearly not intended to beuniversally applicable and consideration must be given to theprobability that membrane markers representing gene productsmay not be restricted to lineage. Indeed malignant cells may, as aconsequence of aberrant gene expression, bear multiple lineagephenotypic markers. We doubt if, even with morphology, suchsingle gene markers, particularly when expressed in "malignant"disease, are adequate to define cellular lineage.

R. D. A. is in receipt of MRC training fellowship.

Departments of Pathologyand Dermatology,

University of Aberdeen,Aberdeen AB9 2ZD

R. D. ALDRIDGEH. SEWELLR. A. MAIN

1 Hsu S-M, Yang K, Jaffe ES. Phenotypic expression of Hodgkin’s and Reed-Sternbergcells in Hodgkin’s disease Am J Pathol 1985, 118: 209-17.

2 Kadin M, Nasu K, Sako D, Said J, Vonderheid E. A cutaneous proliferation of activatedhelper T cells expressing Hodgkin’s disease associated antigens. Am J Pathol 1985,119: 315-25

3 Pinkus GS, Thomas P, Said J Leu-MI a marker for Reed-Sternberg cells in Hodgkin’sdisease Am J Pathol 1985; 119: 244-52.

4. Stass S, Mirro J. Unexpected heterogeneity in acute leukaemia mixed lineage andlineage switch. Hum Pathol 1985, 16: 864-66.

INACTIVATION OF HTLV-III/LAV DURING PLASMAFRACTIONATION

SIR,-The HTLV-III/LAV is thought to survive the processing offactor VIII from contaminated plasma. However, the viability ofthevirus in cold ethanol fractionation and the potential infectivity ofother plasma products is unknown. Spire et all found that reversetranscriptase activity of LAV is lost rapidly when the virus in bufferis exposed to 19% ethanol. We report here our results on thestability of HTLV-III infectivity to ethanol.We began by measuring the residual infectivity (median

infectious dose, IDso 2) of HTLV-III seeded into RPMI-1640medium or plasma (pH 7 4&plusmn;0’ 2) and exposed to ethanol for 10 min.IDso reductions with 20% ethanol varied. The ID50 in medium at23&deg;C was reduced 103-fold. In one single-donor plasma sample at23&deg;C, infectivity was reduced 102-fold but in three donor plasmasamples at - 5&deg;C infectivity was reduced by 102’7 and 100’7 in twoand not reduced in the third. These results suggest that HTLV-III ismore stable to 20% ethanol in plasma than in medium and that somecharacteristic(s) of the plasma, possibly lipid content or proteincomposition, can stabilise the virus to ethanol. Lower temperaturesmay also stabilise the virus. 50% ethanol has consistently reducedvirus titre to levels below the detection limit of our assay when

exposure was at 23&deg;C in either culture media or plasma (untreatedID50, 10-10; treated ID5o, <101). However, at -5&deg;C in

plasma, titre reductions were only 10’ and 10’’BOur next experiment was designed to simulate as closely as pos-

sible on a small scale the precipitation of fractions I+11+111 by 20%ethanol at pH 6 - 9 and - 5&deg;C, as in a common manufacturing proce-dure.3 A sample of cryoprecipitate-poor plasma from a pool of aboutfive thousand donations previously found not to contain antibodiesto HTLV-III was seeded with the virus. Samples were drawn perio-dically after addition of ethanol to a final concentration of 20%, thenimmediately diluted 1:10 in RPMI-1640 medium and frozen to stop