Study of the neuropathology of HIV infection in an African Paediatric Cohort

BACKGROUND: In 1991-92 a large collaborative post-mortem study of adults and children was undertaken in the West African city of Abidjan, Cote d'lvoire, in order to document the impact of human immunodeficiency virus (HIV) infection on the already high mortality from infectious disease. As part of this original study, brain tissue preserved in paraffin wax blocks was retained from 78 HIV positive children (76 HIV-1 and 2 HIV-2) and 77 age-matched HIV-negative children. The baseline neuropathological findings were published and the material was stored as an archive at the Western General Hospital, Edinburgh. Further study of this unique collection of cases was clearly warranted.HYPOTHESES AND AIMS OF THE PRESENT STUDY: In the Initial Study, an attempt was made to quantify white matter cell numbers using a small sample of the whole cohort. The hypothesis was that it would be possible to obtain reproducible numbers for total cell counts in the white matter of HIV negative children and that the numbers were likely to be significantly higher in HIV positive children. The Main Study focussed on the extent of inflammatory change in HIV negative and positive cases, with the prediction that HIV positive children would show a higher level of inflammatory infiltrate than HIV negative children, despite the high level of background brain pathology in the latter group, and that this would be associated with evidence of greater white matter damage in the HIV positive cases. Lastly in the Apolipoprotein E (APOE) Study, it was predicted that the inflammatory response, particularly the degree of microglial activation would be influenced by the APOE genotype, being most evident in children possessing one or more APOE e4 alleles.MATERIALS AND METHODS: For the Initial Study, eight age-matched HIV negative and positive cases were selected from the African cohort, together with an additional HIV negative case at either end of the age spectrum. Four regions of the brain were selected from each of these 10 cases, including cerebral convexity, hippocampus, basal ganglia and the cerebellum. Sections were stained routinely with haematoxylin and eosin, Luxol fast blue and by immunohistochemistry. Total cell counts were performed in an area of one mm2, and subsequently for individual cell types identified using a combination of simple morphology and immunostaining for glial fibrillary acidic protein (GFAP). Although there is no immunostain that reliably identifies all of the members of a particular glial population, whether they be astrocytes, oligodendrocytes or microglia, separation of these cell types was attempted on the basis of GFAP positivity or on simple nuclear morphology. The results were tabulated in Excel.In the Main Study 40 further cases from the African cohort were examined. These included 20 HIV negative and 20 positive, chosen to represent three age groups, lower, middle and upper, up to the age of six years. The basal ganglia and the hippocampus were selected for examination in these cases. In about half of these cases, significant pathology other than HIV-associated changes was present and the spectrum of CNS disease included malaria, meningitis and non-HIV encephalitis. Sections were stained routinely with haematoxylin and eosin, Luxol fast blue and by immunohistochemistry for inflammatory cell markers (CD 14, CD 16 and CD68) or lymphocyte markers (CD8 and CD20). Sections were examined by routine microscopy, and by simple quantitation or image analysis.In the APOE study genotyping was undertaken first on the forty cases used in the main study. Twenty further cases were added subsequently, 10 HIV positive and 10 negative, and within the same age range as those in the main study. DNA was extracted from paraffin sections according to a standard protocol and amplified using polymerase chain reaction (PCR). The degree of inflammation, as detected by CD 16 and CD68 image analysis in these cases, was correlated with the different APOE genotypes. All cases used in this study were anonymised. Ethical permission was sought and obtained for this study (LREC/2003/6/6).RESULTS: The results of the Initial Study showed that the total cell numbers varied from case to case in the HIV negative group, and from one brain region to another, but that these differences were not statistically significant. The total cell counts also varied between individual HIV positive cases and these differences were not statistically significant either. However, comparison ofthe means for HIV negative cases with those of the HIV positive cases revealed significant increases in the latter group. Despite the obvious limitations of the morphological approach to individual cell typing, subset counts from the cerebral convexity, basal ganglia, hippocampus and cerebellum, putatively of astrocytes, oligodendrocytes, microglia and endothelial cells, showed significantly higher levels in the HIV positive cases.In the Main Study, changes in the white matter assessed by routine and Luxol fast blue staining, and by immunohistochemistry for myelin basic protein (MBP), P-amyloid precursor protein (BAPP) and GFAP, and graded according to a simple system, showed only limited differences between HIV positive and negative groups. A general increase in activation of macrophage/microglial cells and of lymphocyte numbers was found in both basal ganglia and hippocampus in the HIV positive group. This increase in the HIV positive group was detected despite a reasonable balance of background brain pathology between the HIV positive and negative subsets. In the case of CD68 immunostaining, the increase was statistically significant for the grey (p=0.002) and white (p=0.009) matter of the basal ganglia. For counts of CD8 positive perivascular cells, the increase was also significant in grey (p=0.001) and white matter (p=0.000) of the basal ganglia, and in grey (p=0.002) and white (p=0.000) matter of the hippocampus.In the APOE study, DNA extraction was successful in 47 cases in total. The distribution of APOE s3 and s4 alleles in this group proved in accordance with the known West African distributions where the s4 allele approaches a frequency of up to 29%. In this study, though the case numbers are small, the frequency of the e4 allele was 41% in HIV negative children and 23% in HIV positive children (p=0.01). These rates were significantly higher than in most Caucasian populations (p=0.001). Satisfactory immunostaining for inflammatory markers was achieved in 44 of the genotyped cases. Correlations with the degree of neuroinflammation in grey and white matter of the basal ganglia were confined to APOE s3/e3 and APOE e3/e4 cases because of the small numbers of cases with rarer APOE allelic patterns. For HIV negative children, there was little difference for CD 16 or CD68 staining between APOE e3/e3 and APOE e3/e4 cases, either in white or grey matter. For HIV positive children, subjects with e3/e4 genotypes did show a trend for higher levels of CD68 activation than those with 3/3 genotypes, particularly in the basal ganglia white matter, but failed to reach significance (p=0.07 for white matter and p=0.1 for grey matter).CONCLUSIONS: 1. Although the results of the initial study are limited in their possible application to any studies other than in this cohort, the increases in total and individual cell counts in the HIV positive group are intriguing. The limitations of a study based on morphological assessment are recognised. 2. In the main study the failure to demonstrate significant white matter change in either HIV negative or positive groups may mean that there is none, but wider sampling in each case and examination of a larger number of cases would increase confidence in this result. It is noted that incomplete myelination in these young subjects is a possible confounding factor in this study. Inflammatory changes, both innate and acquired, proved to be significantly greater in both grey and white matter of HIV positive compared with negative cases, there being a relative balance of co-existing disease in the two groups. This study shows, as predicted, that HIV infection is accompanied by significant microglial activation in these West African children, and that this occurs even in the absence of productive viral infection in the brain. This finding is in keeping with what is known in Western paediatric populations infected with HIV. 3. APOE genotyping revealed a distinctive allele distribution especially in respect of the s4 allele, which was found to be twice as common in this African cohort as in the Scottish population. Although there was trend for greater microglial activation in APOE c4 HIV positive children, the differences between these and APOE e3 HIV positive children did not reach significance