| Experiences with BCG Vaccination in the
UK and Malawi Dr
Hazel DOCKRELL, Dr Rosemary Weir, Dr Gillian Black, Dr Anne-Ben-Smith, Dr Patricia
Gorak-Stolinska, Sian Floyd, Dr Mia Crampin, Dr Paul Fine.
Immunology and Infectious Disease Epidemiology
Units, London School of Hygiene & Tropical Medicine, UK
Introduction
I will be discussing a number of studies carried out by the London School of Hygiene
& Tropical Medicine in the UK and Malawi. A number of groups are developing
new TB vaccines that are about to enter clinical trials. Immunologists need to
identify the best correlates of protection to evaluate these new TB vaccines.
To this end, we need to exploit the observation that BCG vaccination is more effective
in certain populations than in others. Paul Fine has been working in northern
Malawi for about 25 years. We have thus compared the impact of BCG vaccination
on young adults who have not previously been BCG vaccinated in Malawi, where vaccination
is not inducing protection against pulmonary tuberculosis, with the situation
in the UK. In the UK, schoolchildren are still being BCG vaccinated at around
age 13, and epidemiological studies have shown that this does induce protection
against pulmonary tuberculosis.
Choice of Correlates
The immune
response to tuberculosis is very complex. We know that T cells and cell mediated
immunity is important. However, we still do not know what is the best correlate
of protection. We know that it will not be simple or straightforward. We know
that it involves T cells, and that certain cytokines are important. We have thus
been testing the use of simple whole blood assays to measure cytokine responses.
In order to be of use in the field, we need to use a simple assay, such as a diluted
whole blood assay. This assay involves taking peripheral blood, which is heparinised
and diluted in a tissue culture medium. Following antigen stimulation, cytokines
can be measured in the culture supernatants by ELISA. This assay has been proved
to work well under field conditions. When carrying out assays in different
sites, for example in comparing the UK with Malawi, and over long periods of time,
it is important to have assays that can be well controlled. We can thus use positive
control supernatants to control variations in the assays over time and in different
sites.
Details
of the Studies
We recruited
adolescents and young adults with no history of previous BCG vaccination. Before
inclusion in the study, they were skin tested with tuberculin. In Malawi, they
were HIV tested. Individuals were excluded if they had a skin test induration
of over 10 mm, were HIV positive or were pregnant in Malawi, or if they had a
Heaf test response of over 1 in the UK. The individuals were then randomised:
two-thirds were BCG vaccinated, and one-third were left without vaccination for
one year in the UK or given a placebo in Malawi. They were then re-tested one
year later. In Malawi, we recruited 633 individuals, aged 10-28 years, of
whom 562 were eligible. We had over 80% follow up one year later. In the UK, we
recruited 435 individuals, aged 12-15 years and had over 80% follow up one year
later.
IFNg
Responses
We then compared the
IFNg response in the two populations. In the UK, before vaccination, most IFNg
responses to PPD were negative. One year later, the control group did not change
but the vaccine group showed a massive change from 80% non-responders to over
80% responders. In Malawi, about 60% of the group gave positive IFNg responses
to PPD without vaccination. One year later, there was a very slight increase in
both the placebo and vaccine groups.
Overall, the responses
were more variable in Malawi than in the UK. The number of individuals vaccinated
with placebo with no change in response was quite small in Malawi. In the UK,
many more of the control group stayed the same. The IFNg responses to
PPD before and after vaccination in Malawi were not very different to the responses
in the UK after vaccination. Therefore, if we use IFNg as a correlate of protection,
only the increase from before to after vaccination in the UK is a correlate, because
the responses in Malawi are very similar to the post-vaccination responses in
the UK. We believe that the response in Malawi is due to exposure to environmental
mycobacteria. However, there are many different views as to whether that type
of exposure is protective. In order to control for vaccination
status, we measured the BCG scar size in the vaccinees. In terms of BCG scars,
the proportion of vaccinated individuals that developed scars was quite similar
in both populations (98.5% in Malawi and 96.7% in the UK).
Assay Timing
Another major
problem associated with human studies is that people cannot be tested on a weekly
basis, as is the case for animal experiments. Instead, we have to choose one or
two time points for a vaccine study. These immune responses are very time dependent.
For example, the increase in responses is much greater after three months than
after one year. The timing of the assay is therefore very critical in these studies,
and will have to be carefully determined.
Mycobacterial
Sensitivity
The fact that the responses to PPD in
Malawi were influenced by exposure to environmental mycobacteria indicates that
antigens in the BCG vaccine could cross-react with the environmental mycobacteria.
We therefore carried out studies in infants, who would not already be sensitised
to these mycobacteria. In Malawi, infants are BCG vaccinated at birth. Three months
later, the responses to mycobacterial antigens were strong in most, but not all
infants. In the UK, infants who are vaccinated are three or four weeks old - they
are not vaccinated at birth. Overall, we found that these immunologically naïve
infants respond well to BCG vaccination.
Measurement of
additional cytokines Are there other components of the immune
response that should be considered in addition to IFNg? There are many factors
that might influence immune responses which might affect our comparison of UK
school children with Malawian adolescents. For example, many Malawian adolescents
are infected with helminths. They also have more malaria and diahorreal disease.
The more species of helminths carried, the more the IFNg response to PPD drops,
and the more the IL-5 response tends to increase. We therefore tried to measure
other cytokines such as IL-4 or IL-5 responses. More people in Malawi than in
the UK make a IL-5 response but there is no increase with vaccination in either
population. There are also differences in innate immunity. For example, the UK
individuals made less IL-10. Comparison of whole blood
assays with skin testing Are these tests any better than doing
a skin test? Is there any advantage in doing an in vitro test? There is a very
good association between Mantoux skin test induration and the amount of IFNg made
in response to PPD. People with high skin test induration also tend to respond
more strongly to a tuberculosis specific antigen such as ESAT-6. IFNg
assays are robust and perform very well in field settings. They tend to be more
sensitive than skin test responses. However some people are discordant, and will
make one response but not the other. These assays can also be used to detect responses
to individual recombinant antigens. Conclusions
Comparing the UK and Malawi populations has shown that it is the increase in IFNg
that correlates with BCG protection rather than the absolute amount. Measuring
absolute amounts of IFNgis therefore insufficient. In Malawi, where many
of the population are pre-sensitised to mycobacterial antigens, BCG vaccination
only provides a temporary boosting of the response. Prior sensitivity to mycobacterial
antigens has implications for both vaccine design and timing of testing. We have
to bear in mind that the status of the immune system is quite different in low
income settings. For example, an accumulating body of evidence shows that people
in the developing world have more memory T cells than people of the same age in
Europe
We need to combine these types of assays with more direct
measures of mycobacterial killing. We also need to measure other cytokines and
not just IFNg. A new generation of commercial whole blood assays is now available.
The blood is taken neat and incubated overnight. More blood is required for overnight
assays. However, the six-day assays require a tissue culture hood and a CO2 incubator.
More work is needed to compare the ex vivo overnight IFNg assays with assays with
longer incubation periods. In the future, we may also be able to move to simpler
methods of cytokine detection, using multiplex assays or dipsticks. Overall,
diluted whole blood assays have the potential to be a tool for vaccine trials.
Testing large numbers of people requires a simple assay. For example, assays using
flow cytometry require a sophisticated laboratory and cannot be used on a large
scale in the field. Our work has shown that we cannot only
carry out vaccine studies in non-exposed Western populations and that direct comparisons
of populations who are, or are not, protected by BCG vaccination can provide an
informative model in which to develop correlates of protection suitable for use
in trials on new TB vaccines. From the Floor
Have you measured the assays in Asia or other European populations?
Dr Hazel DOCKRELL
Rosemary Weir began using the assays in studies of leprosy in Nepal. They have
also been used in Pakistan, the Philippines, China and Uganda. We have not yet
been able to carry out studies similar to the Malawi and UK studies I have described
in any other countries, as these are large studies that take many years to perform.
|