Brain inflammation is a common feature of HIV-infected patients without HIV encephalitis or productive brain infection

Abstract

HIV-associated neurocognitive disorders (HAND) describes different levels of neurocognitive impairment, which are a common complication of HIV infection. The most severe of these, HIV-associated dementia (HIV-D), has decreased in incidence since the introduction of combination antiretroviral therapy (cART), while an increase in the less severe, minor neurocognitive disorder (MND), is now seen. The neuropathogenesis of HAND is not completely understood, however macrophages (MΦ)s/microglia are believed to play a prominent role in the development of the more severe HIV-D. Here, we report evidence of neuroinflammation in autopsy tissues from patients with HIV infection and varying degrees of neurocognitive impairment but without HIV encephalitis (HIVE). MΦ/microglial and astrocyte activation is less intense but similar to that seen in HIVE, one of the neuropathologies underlying HIV-D. MΦs and microglia appear to be activated, as determined by CD163, CD16, and HLA-DR expression, many having a rounded or ramified morphology with thickened processes, classically associated with activation. Astrocytes also show considerable morphological alterations consistent with an activated state and have increased expression of GFAP and vimentin, as compared to seronegative controls. Interestingly, in some areas, astrocyte activation appears to be limited to perivascular locations, suggesting events at the blood-brain barrier may influence astrocyte activity. In contrast to HIVE, productive HIV infection was not detectable by tyramide signal-amplified immunohistochemistry or in situ hybridization in the CNS of HIV infected persons without encephalitis. These findings suggest significant CNS inflammation, even in the absence of detectable virus production, is a common mechanism between the lesser and more severe HIV-associated neurodegenerative disease processes and supports the notion that MND and HIV-D are a continuum of the same disease.

Figure 1. CD68⁺ and CD16⁺ microglia frequency in the CNS of HIV infected subjects with and without encephalitis

The number of CD68⁺ or CD16⁺ microglia in FWM and BG was determined for each case by averaging the number of positive cells observed over 10 random microscopic fields at 200X magnification by two independent observers for a total of 20 microscopic fields. The mean of the two averages was accepted as the average number of CD68⁺ or CD16⁺ microglia per field for each case. Assuming equal variances, as verified by Levene's test, the relationship of the means for the two groups, HIV⁺/no E and HIVE, was determined by the Independent Samples t-test (GNU PSPP). The average frequency of CD68⁺ microglia in FWM and BG of patients with HIV infection was similar, regardless of the presence or absence of encephalitis. The difference in the average frequency of CD16⁺ microglia, however, does approach significance between the two groups (p=0.06), suggesting that while the number of microglia may be similar in HIV infected persons with and without encephalitis, the degree of MΦ/microglial activation is greater in those with HIVE. It is important to note, that the mean CD68⁺ or CD16⁺ cell frequency reported here for HIVE subjects may be less than actual, as cells that accumulated perivascularly (cuffs) and within nodular lesions, where it was difficult to determine a single cell border, were counted as one.

Figure 2. CD16⁺ MΦs/microglia are observed in the CNS of HIV⁺ patients with and without encephalitis

As reported previously, significant accumulation of CD16⁺ MΦs and microglia are observed in the CNS of patients with HIVE. These cells are found in the parenchyma (Panel C) and comprising perivascular cuffs (Panel F) and nodular lesions (Panel G). HIV⁺/no E also demonstrated considerable CD16 expression in the CNS (Panels B and E), but was less than that seen in HIVE (see Table 2). The degree of positivity varied in this patient population, where some patients demonstrated a low level, while others showed greater CD16 expression. Indeed, some patients in the HIV⁺/no E grouping demonstrated both high and low frequencies. This is in contrast to HIVE CNS, which showed a high frequency of CD16 expression throughout the brain section studied. CD16⁺ microglia in HIVE appear to be in a higher activation state, suggested by a rounded morphology with retracted branches and greater cytoplasm than normally seen in healthy CNS, than in HIV⁺/no E, which show a more ramified morphology (compare Panels B and C). CD16 was not observed in seronegative CNS (Panels A and D). All panels shown at 40X magnification under oil.

Figure 3. CD163⁺ MΦs/microglia accumulate in the CNS of HIV⁺ patients with and without encephalitis

As reported previously, significant accumulation of CD163⁺ MΦs and microglia are observed in the CNS of patients with HIVE within the parenchyma (Panel C), perivascular cuffs (Panel F) and nodular lesions (Panel G). To a lesser degree, CD163 positivity was also observed in the brain parenchyma of patients with HIV⁺/no E (Panel B). CD163⁺ microglia in HIV⁺/no E appear rounded with few branches. Rounded CD163⁺ cells are also observed in HIVE, in addition to numerous activated microglia with thickened, retracted processes (compare Panels B and C). In normal brain, CD163⁺ expression is seen on perivascular MΦs (Panel D), but is not normally expressed by resident microglia (Panel A). Comparatively, accumulation of CD163⁺ perivascular MΦs is seen in HIV⁺/no E (Panel E) but not to the same degree of that observed in HIVE (Panel F). All panels shown at 40X magnification under oil.

Figure 4. HLA-DR reveals pathological alterations in HIV⁺/no E

HLA-DR expression in HIVE was seen by MΦs/microglia that accumulated perivascularly (Panel F) and within nodular lesions (Panel G). In the parenchyma, HLA-DR was expressed by activated microglia, with thickened and retracted processes, most often observed in areas of pathology (Panel C). In the CNS of subjects with HIV⁺/no E, HLA-DR was, for the most part, not expressed by cells in the parenchyma (not shown). Interestingly, when parenchymal HLA-DR was observed, it was seen on aggregates of cells, or “soft nodules” (Panels A and B). Rare HLA-DR expression was seen on perivascular MΦs in HIV⁺/no E (Panel D), with the exception of five subjects, who demonstrated some degree of HLA-DR⁺ MΦ accumulation perivascularly, that appeared to develop into cuffs in one subject (Panel E). A nodule also appears to have formed in the region of the cuff shown in Panel E (arrow). All panels shown at 40X magnification under oil.

Figure 5. Productive HIV infection is not detected in the CNS of patients with HIV⁺/no E

In HIVE CNS, productive HIV infection was detected, as indicated by HIVp24 positivity (Panels B, D and E). This was not observed in any of the HIV⁺/no E cases studied (Panels A and C). All panels shown at 40X magnification under oil.

Figure 6. GFAP expression and astrocyte morphology in HIV infected patients with and without encephalitis

GFAP expression by astrocytes was observed in all groupings, but with apparent differences in the frequency of GFAP⁺ cells, expression intensity and astrocyte morphology. Increasing frequency of GFAP⁺ astrocytes was observed in the white matter (WM) of HIV⁺/no E and HIVE subjects, as compared to HIV⁻ subjects (compare Panel A with Panels B and C; compare Panel D with Panels E and F). The increased frequency of GFAP⁺ astrocytes corresponds to decreased GFAP expression and greater astrocyte hypertrophy, with the most severe seen in HIVE (Panels C and F). In grey matter (GM), GFAP expression by astrocytes in areas away from blood vessels was only observed in HIVE (compare Panels G and H with Panel I). These astrocytes demonstrate higher (brighter) expression of GFAP, than those found in WM, and have largely retained their star-life morphology (compare Panel I with Panel C).

Figure 7. GFAP positivity distinguishes perivascular astrocytes in cortical GM of patients with HIV⁺/no E

Perivascular astrocytes were often found to express higher levels of GFAP than those in the brain parenchyma in HIVE and HIV⁺/no E, suggesting events at the BBB influence astrocyte activation. This is most obvious in cortical GM of patients with HIV⁺/no E, which often displayed regions of perivascular astrocyte positivity that did not extend significantly into the parenchyma (Panels C and D). This was also observed, but to a considerably lesser degree in frequency and intensity, in seronegative CNS (Panels A and B).

Figure 8. Vimentin positivity is primarily limited to hypertrophic astrocytes in HIVE and HIV⁺/no E

Vimentin positivity was more frequent and intense in HIVE (Panels B and D) than HIV⁺/no E (Panels A and C), but limited to hypertrophic astrocytes in both groupings. Unlike GFAP, vimentin did not highlight astrocytes located around blood vessels in HIV⁺/no E (see Figure 7). HIV⁺/no E shows what appears to be endothelial staining, which is also observed in HIVE but with the additional astroglial positivity (Panels C and D). Rare hypertrophic vimentin⁺ astrocytes were observed in only one of the seronegative cases studied (see Table 3).