PHENOLIC GLYCOLIPID I OF MYCOBACTERIUM LEPRAE-ANTIBODY LEVELS IN HOUSEHOLD CONTACTS OF LEPROSY PATIENTS

Abstract r The levels of serum antibody to Mycobactetium leprae specific phenolic glycolipid antigen PG1 were determined in 483 leprosy patients and 1403 of their household contacts in a 5 year followup programme. Thirteen percent of the household contacts were seropositive and 3.4% had very high levels ot antibodies. Household rontact seropositivity was not related to the index patients antibody titre or the disease type. The determination of PC1 antibody is useful in the study of factors affecting leprosy transmission.


Introduction
Leprosy is a chronic granulomatous infection in man caused by Myco-,bacterium leprae.Clinical leprosy presents a wide spectrum ranging from I paucibacillary tuberculoid t o rnultibacillary lepromatous disease.'Most attempts to use serum antibody levels as a correlate of infection with Mycobacterium leprae have been disa pointin due to non-availability of antigens specific for leprosy bacillus.'*x 1+B3 A phenolic glycolipid (PG1) isolated from armadillo-derived M. leprae and characterized8t9 has been shown to be M. leprae specific.819f15The specific antigenic determinant on PG1 is the terminal 3,6-di-0-methyl glucose r e ~i d u e .~' ~' An ELISA procedure developed to measure human serum antibodies to this P G ' ~ permitted investigation of potential correlations between anti PG1-antibody and M. leprae infection.
We have previously reported that in patients with clinical leprosy, the levels and titres of antibody to PG1 are generally higher in multibacillar than paucibacillary patients and in treated than untreated patients.2*3p4v 1 Anti-PG1 antibodies were detected in 34% of household contacts with reference to controls from Seattle, U.S.A. and in a sub-population of these contacts, .theanti-PGI antibody was useful in predicting the development of clinical In the present communication, we report further information on the occurrence and significance of anti-PG1 antibody in household contacts of leprosy patients.

The study population
The study population consisted of 483 leprosy patients and 1403 of their household contacts in Sri Lanka.The subjects were selected randomly from 12 different geographic locations and admitted t o the study if they volunteered to participate in the follow up clinical examination and blood collection programme.All patients and contacts were examined by at least two clinicians with the assistance of paramedical workers of the Anti-Leprosy Campaign of Sri Lanka, blood collected at least at 3-6 month intervals and followed up over a period of 5 years commencing 1980.Sufficient clinical and histological data were available to classify 304 of the 483 patients according to the Ridley Jopling classification.15Information on household contacts was collected from the medical records and from the patients and their contacts who were seen at their home or at central clinics organised specially for this purpose.
Blood samples were collected by venupuncture, serum separated and stored at -20°C in the presence of 0.01% Sodium Azide and 0.01% Thiomenal.Control sera were collected from 118 residents of the same geographic areas who had no known exposure to leprosy patients.
The total number of household contacts with serum available was 1190.Two sera were available for 670 contacts and 3 or more sera were available for 333 contacts.

Serum antibody t o PG1
Serum antibody .toPG1 was determined by ELISA using deacylated.PG1 as described previously.1:4An absorbance.o f ' 0.22 at 490 nm wavelength (A390) in the ELISA represented more than mean + 2 Staridard Deviations for the 11 8 control sera from Sri Lanka and was considered.asseropositive.

Data analysis
All data were stored in a Computer Database using the software package dbase 11.Statistical analysis was performed by using the package SPSS adapted for Wang Professional Computer.

' Serum antibody and seropositivity rates in household contacts
In Table "1.are shown the overall seropositivity rates, mean antibody levels $or serum 1, serum 2 and seium 3 for contacts with 2 or more sera available and the correlation coefficients between follow up'sera.The overall seroposi-  tivity rate was 13% for total contacts (A490 > = 0.22 =Mean + 2 S.D. of Sri Lanka controls) and very high antibody levels (A490 > = 0.44) were seen in 3.4%.A total of 83 contacts showed decreasing antibody levels (change > 0.1 1) and 23 showed increasing antibody levels (change > 0.11).
In some household contacts, the serum antibody to PG1 showed considerable variation, but there was a significant correlation between the follow up sera.This correlation was of a much higher level in the case of patients than in contacts.

3.2
Correlations between antibody levels in household contacts with the antibody levels in the index patients.
In Table 2 are shown the relations observed between the index patients antibody levels and that of household contacts.The mean AqgO of sepin 1 and serum 2 from those contacts who had a minimum of 2 sera were used in this computation.There was approximately uniform distribution of both seronegative and seropositive household.contacts in households of high, moderate and low antibody index patients.

Correlation between contact seropositivity rates and index patient disease t y p .
As shown in Table 3, household contact seropositivity was not related to the index patients disease type.

Clustering of seropositive contacts t o households.
It was observed that the majority of seropositive contacts were restricted t o certain households.Out of 21 3 households of family size between-2 and 6 contacts, a l l the seropositive contacts (total 99) were in 68 (32%) households, while the remaining 145 households (total contacts 521) were totally seronegative.This segregation was clearer in small households ( c ~n t a c i size 2-3) than in large households.The seropositivity rate in the 'positive' households was 32% as compared to the overall seropositivity of 14% for this group of contacts.These results are summarised in Table 4.

Development of clinical leprosy in household contactl and relation t o index patients disease type and anti-PG1 -antibody levels.
During the observation period of approximately 5 years, 1 3 contacts developed clinical disease.The antibody levels of both the contact and the index patient and the disease type etc, are summarised in Table 5.There wasno correlation between the anti-PG1 antibody levels in contacts who developed clinical leprosy and the index patients disease type or the antibody levels.
between A 490 and index patients antibody levels.Data presented in the present communication show that the determination of anti-PGl antibody in household contacts provides information that could be useful in defining patterns and epidemiological aspects of leprosy transmission.The demonstration of very high levels of anti-PG1 antibody in .
3.4% of household contacts strongly implies continuous antigenic stimulation.Most importantly, if the contacts with very high levels of anti-PGI antibody are in fact harbouring M. leprae, they form a hazardous reservoir of infection.
The absence of correlations between antibody levels in household contacts and the index patients tyge of disease and antibody levels have been reported for other antigens.'* 3 1 1The .absence of such correlatiohs between contacts who developed clinical &ease and the index patient has not been reported previously.The lack of such correlations, taken together with clustering of seropositive household contacts to certain families and geographic distribution of contact seropositivity imply that environmental factors other than exposure to the patient in the family probably play a role in leprosy transmission.This could also mean that the actual source of infection is not the index patient.The development of predominantly TT/BT disease in TT/BT households was also observed.In one household, all three household contacts developed TT/BT disease over a period of approximately 10 years.

2 Total
and % seropositivity in the second serum (A490 > 0.21) Number of seropositivity in serum 1, decreased in Serum 2

Table 3 .+ 2 .
Antibody to PGl 1 of Mycobacterium leprae: correlation between household contacts seropositlvlty and disease type +I of the index patient.patients, both treated and untreated, treatment with DDS only.S.D.) of Sri Lankacontrols.

Table 1 .
Serum antibody t o PG1:Seropositivity rates and correlations between follow up sera in patients and contacts.

Table 2 .
Correlation between mean antibody levels t e P G 1 in patients and their household contacts. *a