| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Unité Postulante "Mycologie Moléculaire," Institut Pasteur, Paris, France;
* Unité Postulante "Macrophages et Développement de l’Immunité," Institut Pasteur, Paris, France,
Unité "Rétrovirus & Transfert génétique," Institut Pasteur, Paris, France;
Service d’Anatomie et de Cytologie Pathologiques, Hôpital Lariboisiere, Paris, France;
Unité "Epidémiologie & Physiopathologie des Virus Oncogènes," Institut Pasteur, Paris, France;
|| CEA, Service de Neurologie, CRSSA, EPHE, Fontenay-aux-Roses, France;
** Groupe "Virologie Moléculaire & Vectorologie," Institut Pasteur, Paris, France; and
Unité "Récepteurs & Cognition," Institut Pasteur, Paris, France
1 Correspondence: Institut Pasteur, 25 rue du Dr. Roux, 75015 Paris, France. E-mail:mordelet{at}pasteur.fr; or granon{at}pasteur.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and astrogliosis. No apoptotic event was detected. We further demonstrated that Nef neurotoxicity is associated with cognitive deficits.—Mordelet, E., Kissa, K., Cressant, A., Gray, F., Ozden, S., Vidal, C., Charneau, P., Granon, S. Histopathological and cognitive defects induced by Nef in the brain.
Key Words: neuropathology • behavior • bone marrow-derived macrophages
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4
5
6)
. The neuropathological damages commonly known as AIDS cognitive disorders (ADC) are characterized by progressive deterioration of mental and motor functions. When observed, brain lesions mainly affect the white matter, the basal ganglia, the cerebral cortex, and the hippocampus (1
, 7
8
9)
. However, a non-negligible percentage of patients show no lesion but exhibit minor cognitive deficits that may constitute a pathology different from ADC (10)
. At a cellular level, the main reservoir for HIV replication are monocyte-derived macrophages/microglial cells (11
12
13)
. A low level of infected astrocytes, neurons, and brain capillary endothelial cells has been reported in brain specimens with ADC (14
15
16
17
18)
.
Several lines of recent in vitro and in vivo evidence suggest that different factors are involved in HIV-1 neuropathogenesis. Among them, viral products (Tat, gp120, gp160, Nef), cellular products (acid quinolinic, NO, etc.), cytokine induction (TNF-, IL-8, INF-
), and excitotoxic injuries may lead to neurodegeneration or neuronal loss (19
20
21
22
23)
.
Nef protein is precociously and strongly expressed during HIV-1 infection (24
, 25)
. Immunohistochemical staining has revealed an overexpression of Nef compared with other viral proteins in HIV-infected astrocytes and brain biopsies (26
, 27)
. In SIV-infected monkeys, it has been shown that Nef expression is essential to maintain a high replication level of the virus and promote the development of AIDS (28)
. In a transgenic mouse model, Skowronski and collaborators (29)
showed that the expression of Nef in lymphocytes TCD4+ induced modifications in the signaling pathway of those cells leading to their progressive death. Moreover, among the transgenic mice expressing the complete coding sequences of HIV-1 in T CD4+ cells, only the Nef transgenic mice developed a severe AIDS-like disease, detected in all organs, including the brain (30)
. Altogether, these data suggest a potentially important and specific role for Nef in cellular dysfunctions and its contribution to the development of the neuropathology associated with AIDS (31
32
33
34)
.
A highly significant increase in the number of macrophages in patients who suffered from ADC has been reported in postmortem studies, pointing to a possible role of macrophage infiltrates for the exacerbation of neuronal dysfunction (35)
. Several clinical and experimental studies have focused on the production of chemotactic factors during HIV infection (36
37
38
39)
. Recently, Koedel and collaborators (40)
have proposed that Nef is essential for AIDS neuropathology as a mediator for the recruitment of leukocytes. Therefore, the presence of leukocytes within the brain may serve as a vehicle for the virus and may perpetrate the disease through their production of neurotoxins. Overall, this corpus of results led us to investigate the role of the Nef protein when expressed by some preferential HIV target cells (macrophages). In parallel, we performed direct injection of vector particles known to target mainly astrocytes in order to see whether the expression of Nef in the astrocytes would be different from its expression in macrophages.
The main aim of our research was to determine in vivo whether Nef, expressed by macrophages when transplanted onto the rat brain could induce observable early toxicity. To address this, we had efficiently transduced ex vivo rat primary macrophages with a lentiviral vector (41)
and showed that, in vivo, transplants of transduced macrophages into the rat brain maintain long-term expression of a control protein (green fluorescent protein, or GFP) for up to 90 days without any significant sign of astrogliosis.
In the present study, we showed a sustained, stable, and noncytotoxic expression of Nef in bone marrow-derived macrophages (BMDM) in culture. This first step allowed us to graft primary macrophages expressing Nef onto the rat hippocampus and study their effects on behavior and histopathology. Our present results corroborate the presence of Nef in the hippocampus with its histopathological effects including monocyte/macrophage recruitment, expression of the proinflammatory cytokine TNF-, and astrogliosis. We found no correlation between the presence of Nef in the brain and certain apoptotic events. We show that Nef neurotoxicity is associated with cognitive deficits specific to hippocampal dysfunction.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 42)
. Briefly, cells were flushed out with a 25 gauge needle into ice-cold phosphate-buffered saline (PBS) without Mg2+, Ca2+. The marrow plugs were centrifuged (1500 g, 10 min, 4°C). After dissociation and elimination of the red cells, bone marrow cells were seeded onto 6-well dishes (4x106 cells/mL, Costar, Cambridge, MA, USA) in RPMI 1640 (Seromed, Berlin, Germany) supplemented with 2 mM L-glutamine, 2 0 mM NaHCO3, 5U/mL penicillin, 50 µg/mL streptomycin, and 10% heat-inactivated fetal calf serum. The conditioned medium from L929 (5%) as a source of colony-stimulating factor 1 (CSF-1) was added in RPMI supplemented (43)
. At day 3, adherent and differentiated BMDM were then used for transduction. For immunofluorescence studies, glass coverslips were placed into the wells of cultures plates. For in vivo transplantation, the cells were detached with trypsin-EDTA 1x, washed and resuspended in serum and CSF-1 free medium at 2.5 x 107 cells/mL.
Lentiviral vector
A three-plasmid expression system was used to generate vector particles by transient transfection of 293T cells using the calcium phosphate coprecipitation technique as described (44
, 45)
. We used the vector plasmid encoding an hCMV-driven expression cassette of the GFP or Nef cDNA (TRIP vector). The encapsidation plasmid (p8.2) provides all vector proteins and the VSV-G envelope expression plasmid (phCMV-G) permits the production of vector particles. The system was improved by the deletion of the U3 region of the 3' long terminal repeat (LTR) of the DNA used to produce the TRIP vector (46)
. The vector stocks were treated with DnaseI as described (45)
. Vector titration was assayed for p24 Gag antigen by ELISA assay (NEN Life Science Products, Boston, MA, USA). In vitro transduction experiments were done in 6-well plates. BMDM were transduced using 150 ng of p24 TRIP-GFP or 300 ng of p24 TRIP-Nef vector particles, respectively/106 cells.
Intracerebral injections of viral particles were performed using 1.5 µL of TRIP-GFP (100 ng/µL p24) and 4.3 µL of TRIP-Nef vector particles (70 ng/µL p24).
Quantitation of gene transfer efficiency
Transduction efficiency was evaluated using flow cytometry. The mononuclear phagocyte population (BMDM) was selected by granularity and cell size as described (41)
. After suitable gating, 10,000 events were collected for each experiment in FL1-height (488 wavelength) (CellQuest software; Becton Dickinson, Pont de Claix, France). M1 represents the mean of the signal corresponding to the negative-untransduced macrophage population and M2 the percentage of Nef- or GFP-transduced macrophages.
Immunoprecipitation and immunoblot analysis
Primary macrophages were transduced ex vivo for 48 h. Cell extracts were precipitated in lysis buffer, loaded, denaturated, and frozen. Supernatants were collected, concentrated (Centriplus, Amicon, Millipore, Yvelines, France), and treated as described above. Protein concentration was determined using BCA reagent kit (Pierce, Rockford, IL, USA). For each sample, proteins (50 µg) were fractionated on 16% SDS-PAGE gradient gels under nonreducing conditions and transferred to a cellulose membrane. Nef was labeled using Nef antibody (Hybridolab, Pasteur), followed by a secondary anti-mouse horseradish peroxidase. Proteins were detected with the chemiluminescence detection Kit (Amersham, Biosciences UK Limited, Little Chalfont, England, UK).
Intracerebral macrophage transplantation
Autologous engraftments were performed using BMDM transplanted into the brain of Long Evans rats (n=24). Male adult rats (250–300 g) were anesthetized with sodium pentobarbital (50 mg/kg; Sanofi, France) and positioned onto a stereotaxic frame as described (41)
. Holes were drilled into the appropriate locations and a 10 µL Hamilton syringe was used to inject the cellular suspension or the vector particles into the hippocampus (AP–2.6 and –4.2, L±1.4 and ±3, DV–3.3 and –3.5) at a slow time course (0.2 µL/min). At cessation of the injection, the needle was left in place for 15 min before withdrawal. Four different groups of animals were grafted bilaterally: 1) six rats with Nef-transduced macrophages and, as a control, six rats with GFP-transduced BMDM, 2) six rats injected with TRIP-Nef vector particles; as a control, six rats injected with TRIP-GFP vector particles. One animal injected with TRIP-Nef vector particles died during surgery. Animals were killed 2 months after cell transplantation or vector particle injection. Rats were perfused intracardially with 4% paraformaldehyde (PFA). Brains were removed, dipped in 20% sucrose solution, frozen in Tissue-tek OCT (Miles, Elkhart, IN, USA), and stored at –80°C.
Immunocytochemical procedures
In vitro, BMDM cells were washed, fixed with 4% PFA solution for 20 min at room temperature and rinsed with PBS. Cells were then treated with a blocking solution (10% goat serum and 0.2% Triton X-100 in PBS) and incubated with Nef antibody (Hybridolab, Pasteur). Next, the secondary antibody FITC-conjugated goat anti-mouse IgG (Beckman Coulter, Roissy, France) diluted 1:500 in PBS was added to the cells. The glass coverslips were mounted in fluosaved reagent (Calbiochem, Meudon, France) and stored at –20°C. The GFP-autofluorescent and FITC labelings were analyzed by optical microscope (Zeiss, le Pecq, France).
For in situ immunohistochemistry, coronal OCT-coated slices (cryostat, 25 µm) were rehydrated in PBS and incubated in 0.3% hydrogen peroxide in methanol for 20 min at room temperature in a moist chamber to inhibit endogenous peroxidase. Slices were then treated with a blocking solution containing 10% goat serum and 0.2% Triton X-100 in PBS. To characterize the macrophages, cells were incubated with mouse IgG anti-rat Mac-1 (Serotec, Oxford, UK, diluted 1:100 in blocking solution) (47
, 48)
. Antibodies specific for the astrocytic marker GFAP (Chemicon, Paris, France; diluted 1:400 in blocking solution) and TNF- (Chemicon; diluted 1:40) were used to detect inflammatory process (49
, 50)
. Next, a secondary biotinylated antibody (Amersham, France) was amplified by a streptavidin complex labeled with alkaline phosphatase (Dako, Cambridgeshire, UK). Slices were counterstained in a Mayer’s hematoxylin solution (Sigma, St. Louis, MO, USA) and analyzed by optical microscopy.
TUNEL assay
DNA strand breaks of apoptotic cells were identified in situ by the ApopTag kit (Intergen Compagny, Oxford, UK). Brain cryosections were treated with hydrogen peroxide (0.3%) in PBS for 5 min at room temperature to quench endogenous peroxidase. Slices were rehydrated with the equilibration buffer and incubated with the working strength TdT-based enzymatic solution laced with nucleotides for 1 h in a humidified chamber. The reaction was stopped using the working strength stop/wash buffer. Slices were counterstained with propidium iodide (1 mg/mL) and mounted under a glass coverslip. Pictures were obtained on optical microscope.
Behavioral methods
Apparatus
Open field
A white circular open field (110 cm in diameter and 35 cm height) was illuminated by two indirect 60W bulbs suspended on both side walls. Four different objects were placed into the open field and salient visual and auditory cues were placed in the environment to allow spatial orientation. A camera fixed to the ceiling above the apparatus was connected to a videotrack system (View-point, Lyon, France), allowing the experimenter to record and observe behavior out of the sight of the animals.
Cross maze
As described in ref 51
, the apparatus was made of four arms designing a cross (total size 1 mx78 cm) connecting a central platform (Fig. 1
A). cup containing sucrose pellets was presented at the end of the north arm of the maze. The experimental room contained a few extra-maze cues and was illuminated by two indirect 60W bulbs.
|
Experimental design and testing procedure
Open-field
The rats were daily manipulated for 10 min for 2 wk before surgery and 1 more wk during postoperative recovery. They were subjected to the open field test 28 days after grafting. The general procedure has been described in detail elsewhere (52)
. Briefly, rats were subjected to five successive 6 min sessions separated by a 3 min delay during which they were returned to their home cage. During session 1, rats explored an empty environment; for sessions 2–4, their environment contained four objects (A–D). During session 5, the spatial arrangement of the objects was altered to test the reaction-to-spatial change. The location of objects B and D was modified so that D replaced B and B was displaced at the periphery of the arena. To neutralize the possible effects of olfactory cues, the experimenter used plastic gloves when moving and manipulating the objects. The floor was cleaned with water between each session.
Cross maze
This experimental procedure (51)
started 38 days after surgery. Animals were food deprived until their weight was adjusted at 85% of normal body weight. All animals had access to water ad libitum. Place learning started once the rats had become acclimated to the maze (i.e., rats consumed food pellets in the maze for a maximum of 2 min). Two rats from macrophage group (one from GFP and one from Nef condition) were discarded because they lacked food motivation.
Each rat received three trials a day. Between each trial rats were returned to a waiting cage by the time another rat ran into the maze. The starting arm was either at the west or east arm and randomly chosen, so that on 2 consecutive days every rat started half of the time by the east arm and half of the time by the west arm. The south arm was neither baited nor a starting point but contained a food cup with unreachable food in order to avoid odor guidance. The time to reach the food goal was recorded. Once the rat began to eat, the timer was stopped and the animal was allowed to consume the food for 10 to 15 s. If a rat failed to reach the goal in 2 min, it was returned to a waiting cage.
After 7 days of learning, all rats reached the criterion (i.e., making no visit to an unbaited arm for 2 consecutive days; 6 trials). After a rat reached the criterion, it was submitted to an orientation test consisting of reaching the goal from modified starting points. In the modified configuration, the position of the goal arm and distal environmental cues remained unchanged (Fig. 1B
).
Data collection and analysis
Open-field
The video track software collected automatically locomotor activity in terms of distance and exploratory activity in terms of time of entry in areas containing objects.
Locomotor activity was compared between groups using a t test during the first session before introduction of objects.
Times spent in areas containing each object were averaged across individuals per group. A repeated measure ANOVA for paired comparison (Statview program) was carried out. The main factors were "G" as the between-subject measure and "session" as the within-subject measure. To ascertain whether rats showed habituation, we compared sessions 2, 3, and 4.
Reaction-to-spatial-change was compared between groups using a t test performed on the difference between the time spent investigating the exchanged objects and the nonexchanged ones: session 5 minus session 4.
Cross maze
In this task, the variable was the time to reach the goal. We carried out repeated measures of variance to ascertain the effects of the main factors and t tests for post hoc analysis.
During learning, session was the within-subject factor and group the between-subject factor.
In the orientation test, group was the between-subject factor and "condition" was the within-subject factor with two levels, standard configuration vs. test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A). We observed a signal showing that >70% of macrophages expressed Nef (M2) compared with untransduced BMDM (M1) (Fig. 2B
). This highly efficient transduction was obtained using a concentration of TRIP-Nef vector particles corresponding to 300 ng of p24 (see Materials and Methods). A control of transduction was assessed using GFP lentiviral particles at a concentration of 150 ng of p24. As shown in Fig. 2C
, >75% of BMDM was efficiently transduced with TRIP-GFP vector (M2) compared with untransduced cells (M1). The sustained Nef expression remains stable for at least 1 month in vitro without any cytotoxic effect noted in the BMDM culture (data not shown).
|
Characterization of transduced cells
Primary culture of BMDM (Fig. 3
A) was characterized by immunohistochemical procedure using Nef antibody or autofluorescence detection for GFP. As shown in Fig. 3B
, after 15 days of transduction the greatest part of BMDM expressed Nef. The same result was obtained using TRIP-GFP vector as a control of transduction (Fig. 3C
). All BMDM came from the mononuclear phagocyte lineage as shown using ED1 and Mac-1 antibodies (41)
.
|
Western blot analysis was performed under nondenaturating conditions on cellular extracts and supernatants of BMDM transduced with the lentiviral TRIP vector expressing Nef or GFP. In the cellular extracts of Nef-transduced BMDM (Fig. 3D
, line b), Nef can be detected using a monoclonal antibody as a 27 kDa protein. The same antibody was unable to reveal Nef expression in the concentrated supernatant of Nef-transduced BMDM (Fig. 3D
, line c). No signal was observed in the cellular extract of GFP-transduced BMDM (Fig. 3D
, line a).
Viability of transduced macrophages transplanted into the rat hippocampus
The high efficiency of Nef primary macrophages transduction allowed us to investigate the effect of Nef in the rat brain, so we performed engraftments of Nef-transduced BMDM into the hippocampus area (Fig. 4
A).
|
We observed that Nef was detected around the injection site of transplanted Nef-BMDM 2 months after the graft (Fig. 4B
). Similar results were obtained with GFP-transduced BMDM (Fig. 4C
). Analysis of direct vector particles injection showed that Nef or GFP particles surrounded the needle track in the hippocampus (Fig. 4D, E
, respectively).
Chemotactic and neurotoxic effects of transplanted Nef-BMDM in the rat hippocampus
Two months after the graft, transduced Nef and GFP BMDM were still alive at the injection site and could be detected using Mac-1 antibody (Fig. 5
A). The presence of Mac-1-positive cells around the injection site of Nef-transduced-BMDM (Fig. 5A
, bottom) suggests the recruitment of peripheral or resident monocytes/ macrophages as a consequence of Nef expression. Similar results were observed around the injection sites of Nef viral particles (data not shown). By contrast, in the brain of rats grafted with GFP-transduced BMDM, only grafted macrophages were stained with Mac-1 antibody (Fig. 5A
, top).
|
We examined each transplanted and injected brain slice for astrocytes reactivity to assess local astrogliosis using GFAP astrocytic marker. Two months after the graft, enhanced GFAP staining was detected around the Nef-transduced BMDM (Fig. 5B
, bottom) whereas no significant sign of astrogliosis was seen around the graft of GFP positive BMDM (Fig. 5B
, top).
As brain damage can be induced by direct or indirect mechanisms involving proinflammatory cytokines, we examined the presence of TNF- expression in the rat brain. We observed about the Nef-transduced BMDM a significant signal of TNF- 2 months after the engraftment (Fig. 5C
, bottom). Moreover, TNF- expression was detected around the astrogliosis only in the case of Nef-BMDM graft. No significant astrogliosis or TNF- expression was detected around the GFP-BMDM graft (Fig. 5C
, top).
The next step was to analyze whether Nef could induce an apoptotic effect in the rat brain as a consequence of the inflammatory process. Figure 5D
illustrates that no apoptosis corpus was observed 2 months after Nef-BMDM or GFP-BMDM transplantation into the rat hippocampus.
Behavioral results
Exploratory activity
Statistical ANOVA for exploratory activity is summarized in Table 1
. It shows no statistical difference between groups (macrophages GFP vs. macrophages Nef and particles GFP vs. particles Nef) for locomotor activity and habituation, although there is a tendency for animals receiving macrophages expressing Nef to exhibit a higher level of exploratory activity toward the objects (illustrated in Fig. 6
). We cannot exclude this lack of statistical difference between groups as an expression of a high variability between Nef individuals. There is no significant interaction group X session, indicating a similar habituation in both groups.
|
|
However, the reaction-to-spatial-change is significantly different for the GFP and Nef animals of the particle group (time of object exploration for S5-S4, t=0.04), and nearly reach significance for the GFP and Nef animals of the macrophage group (time of object exploration for S5-S4, t=0.054).
Place learning
Statistical analyses of variance for place learning and orientation test are summarized in Table 2
.
|
Statistical analyses conducted on "time to reach the goal" showed a main effect of session, no group effect and no significant group X session interaction, indicating that all animals learned the task in 7 days or less and did not differ in the time needed to reach the goal as illustrated in Fig. 7
A, B.
|
Orientation challenge
We compared the time to reach the goal for GFP and Nef animals (group effect) before modification of the spatial arrangement of the maze arms (performance at the 7th day of training) and during the orientation test for which configuration of the apparatus is modified (condition effect). Statistical results of the repeated measures of variance for the main effects are given in Table 2
. As illustrated in Fig. 7C, D
, these results suggest that the orientation test did not affect the performance of both groups in the same way.
In the macrophage group, post hoc analysis revealed a significant condition effect for group 2 (Nef) (t=2.8, ddl=4, P=0.048) but not for group 1 (GFP) (t<1, NS). The time to find the goal for the Nef group was longer during the orientation test than during the learning condition, but remained stable for the GFP animals whatever the configuration of the maze.
In the particle groups, post hoc analysis showed a condition effect for group 2 (Nef) (t=3.24, ddl=4, P=0.03) but not for group 1 (GFP) (t<1, NS).
Overall, these results show that only Nef animals (macrophage and particle) spent significantly more time reaching the goal in the modified configuration than in the standard configuration, suggesting that the ability to adapt their behavior to sudden configural changes is altered in Nef animals compared with GFP animals.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our in vitro results demonstrate that HIV-1 lentiviral vector allowed high and stable expression of Nef in transduced macrophages and correct maturation and expression of Nef in the BMDM. Moreover, neither cytotoxic effects nor secretion of Nef protein from the cell was observed in the culture over 1 month after transduction. This suggests that our in vivo observations are unlikely to result from the effect of the secreted protein but rather from the expression of Nef by the macrophages.
In vivo, we grafted BMDM expressing Nef into the hippocampus area to 1) detect HIV-1-gene sequences (in particular, gag and nef) into the hippocampus (53)
, and behavioral effects of hippocampal damage in the rat are described in the literature (see ref 54
for review) and because 2) earlier studies have described the central role of Nef in toxic mechanisms, notably by modulating the intracellular machinery of the host cell (55
56
57)
. It has been shown in vivo that HIV neuropathology may be correlated with the presence of Nef expressed in astrocytes (26
, 27
, 53)
. 3) We developed an efficient transduction method for targeting original HIV cells (macrophages) using a TRIP vector (41)
.
To mimic the human HIV neuropathology in an animal model, we grafted onto the rat hippocampus macrophages ex vivo transduced with TRIP-Nef vector.
One month after grafting, we began the investigation of behavioral functions in transplanted animals by first conducting open field exploratory behavior. Our results showed that reaction to spatial change is affected in animals receiving Nef. Indeed, we did not observe increased locomotor activity or impaired habituation of exploratory activity as would be expected if the hippocampus was lesioned (58
, 59)
. However, the reaction to spatial change is impaired in Nef animals, suggesting that spatial memory is affected in these rats. The ability to memorize a spatial configuration is known to rely on hippocampal function in animals (60
61
62)
and humans (63
, 64)
. Although supporting spatial representation is a main characteristic of the hippocampus (65)
, two primary reasons, not mutually exclusive, can explain this incomplete effect.
First, the construction of spatial representation is one of the major functions in rodents. Thus, it is supported by the functioning of several interconnected brain areas including the prefrontal and parietal cortices, the septum, CA1 and CA3 hippocampus, and the amygdala (54
,61
). Such a redundant function is unlikely to be affected by a minor lesion of one of the targeted area, as may have been the case only 1 month after grafting. Our histological results indicated only focal inflammation of the hippocampus around the graft (Fig. 5A
) whereas earlier work showed that complete lesion or inactivation of the hippocampus is necessary to induce robust impairment in place representation (58)
. Moreover, the Nef transgenic mouse model described by Hanna and colleagues (30)
supports the view that a certain level of expression of the transgene is necessary to achieve sufficient protein synthesis to induce an AIDS-like pathogenicity (30)
.
Second, the open field test might have been carried out too early for the hippocampus to be fully affected by Nef expression. This explanation is supported by the observation that the performance is quite variable among rats receiving Nef. This variability may be due to individual factors in Nef action within the hippocampus over time. This hypothesis is supported by the drastic behavioral impairments observed in the spatial learning task.
In the spatial learning experiment, animals receiving particles or macrophages expressing Nef showed normal spatial learning of the food goal location. Animals receiving the control protein GFP and those receiving Nef demonstrated significant decreases in the time needed to reach the food, suggesting that all animals learn to locate the goal progressively. However, several studies have shown that rats can locate spatial goals by computing intra- and extra-maze cues and that the hippocampus participates crucially in the integration of distal configural cues (51
, 62
, 66)
.
It is known that spatial tasks favor in rodents the development of strategies that affect learning processes (67
68
69
70)
. Animals spontaneously display place strategies that were reinforced in our protocol (51)
. During the orientation test, the angular relationships between the arms of the maze were altered such that the arm containing the food goal was the only place having stable spatial relationships with the external environment. Our results for this test showed that GFP rats still performed as efficiently as in the standard configuration, suggesting they were able to use a memorized configuration of the goal based on distal landmarks. On the contrary, Nef animals (for either the macrophage or particle groups) showed a marked impairment, as for these groups the time needed to reach the goal significantly increased after alteration of angular relationships between the arms of the maze. We previously showed that in this configuration, normal animals solve the task using primarily extra-maze cues, leaving intra-maze cues irrelevant (51)
. The ability to adapt the behavior to sudden configural changes based on spatial strategy implies using and updating the spatial relationships between the components of the environment. These cognitive abilities, known to rely on the integrity of the hippocampal (58
, 59
, 61)
, are specifically altered in Nef animals. We can therefore infer that animals receiving Nef exhibit cognitive defects resulting from hippocampal dysfunction.
At the completion of behavioral experiments, animals were killed and the brains were removed for histopathological analyses. Our results showed that 2 months after transplantation, the presence of Nef in the rat hippocampus induced a significant astrogliosis and a recruitment of monocytes/macrophages around the injection site of the graft. These histopathological effects are unlikely to be due to a general effect of a foreign protein expressed in the brain, as GFP-BMDM did not produce a similar astrogliosis. Our results are in accordance with those of Koedel and colleagues (40)
, who demonstrated that injection of soluble HIV-1 Nef in the brain induced the recruitment of leukocytes toward the inflamed region. However, as they injected a soluble form of the protein, they conclude that Nef represents the viral factor involved in HIV neuropathology. Our present data further show that not only is the protein itself able to induce a neuropathology but that, as in infected patients, the protein expressed by the HIV target cells (i.e., macrophages) induced an early noticeable neurotoxicity. Indeed, it has never been shown that Nef is secreted at a sufficient rate to be able to induce an inflammatory process. Therefore, we hypothesize that Nef modifies the cellular machinery of the HIV target cells, which secondarily spreads inflammatory processes toward the surrounding cells of the target area. In accordance with this hypothesis, we show a significant expression of the proinflammatory cytokine TNF- specifically around the injection site of the Nef grafts, an inflammatory process not observed in GFP-transduced cells. The TNF- secretion seems to be corroborated with the localization of astrogliosis, but could be the consequence of the expression of Nef by the macrophages. In fact, Nef neurotoxicity has been linked to the production of quinolinic acid by macrophages. The quinolinic acid, as the proinflammatory cytokine TNF-, has been shown to be produced by activated macrophages in several inflammatory brain diseases, including AIDS dementia complex (37
, 71
72
73)
.
In the literature, HIV-1 neurotoxic proteins (i.e., gp120 and tat) have been shown to induce apoptotic events (74
75
76)
. It has been further suggested that apoptosis is mediated through indirect mechanisms involving modifications of the signaling pathway of infected cells (77
78
79
80
81)
. However, the results concerning Nef are controversial. Rasola and collaborators (82)
concluded that Nef enhances programmed cell death, whereas other studies have described the role of Nef in promoting resting cell infection (83)
or argued that Nef is associated with anti-apoptotic signals (84
, 85)
. In our study, we did not find any apoptotic events associated with astrogliosis or TNF- expression.
To conclude, our results showed that Nef, through its expression by HIV target cells, is one of the viral factors that influence the early development of a neuropathology that can produce cognitive impairments. Moreover, in accordance with the results obtained in other laboratories (40
, 86)
, our data suggest that Nef may mediate chemotaxic effects and a reactive response by macrophages and astrocytes. Our study has investigated, for the first time, both behavioral and histopathological effects of a viral HIV-1 protein expressed by the natural HIV-1 target cells grafted onto a specific brain area known to be affected in HIV-demented patients. The rat model of HIV-associated neuropathology should be useful for the study of fundamental mechanisms of neurotoxicity exerted by a specific protein expressed by macrophages within the brain. Furthermore, it emphasizes the interest of correlating histopathological and behavioral data in evolutive neuropathologies affecting cognitive functions.
|
ACKNOWLEDGMENTS |
|---|
Received for publication June 24, 2004. Accepted for publication July 26, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
