| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Division of Biology and Medicine, Brown University, Providence, RI 02912;
Institute of Neurology, University of London, London, United Kingdom; and
Department of Pathology, Dartmouth Medical School, Lebanon, NH 03756
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In contrast, previous work of Cserr and Knopf (2, 3) suggests that intrathecal Ab synthesis within the brain may not just be a sequela associated with CNS diseases involving a compromised BBB; it can be elicited in a healthy animal with normal BBB permeability. A model was developed to study regulation of immune responses in the brain of normal rats (4). In this model, a small volume of Ag in saline is slowly microinfused through an indwelling cannula localized to a specific region of the brain by stereotactic placement. Ag infusion is conducted 1 wk following cannula implantation to allow the BBB to regain normal permeability (5). Specific Ab titers can be detected in serum within 1 wk postinfusion and persist for several months. In addition, specific Ab can be detected in CSF at 3 wk postinfusion, at levels in excess of values predicted by passive influx from plasma, via a normal leakage pathway determined by molecular size (6). In these rats, normal barrier permeability was confirmed by measuring relative concentrations of endogenous albumin levels in CSF and serum. Moreover, excess CSF Abs were never detected in rats peripherally immunized with Ag plus adjuvant, even when serum Ab titers generated by this hyperimmunization protocol were up to two orders of magnitude higher (6). Specialized Ig transport into the CNS has also been excluded (7). We have therefore concluded that intrathecal synthesis is required to account for the excess of specific Ab in the CNS following Ag stimulation to the brain.
Evidence for a humoral immune response within an Ag-stimulated but apparently healthy brain is in contrast to the immune privilege status of this organ with respect to allogeneic tumors and tissue transplants and their prolonged survival in normal brain (2). Historically, the survival of these tissues was attributed to an apparent lack of an effective immune response within the normal brain, which was a consequence of two unique anatomical features: the absence of classic lymphatic drainage within the nervous tissue and the presence of the BBB. Together, these features could block afferent and efferent arms of an immune response to Ag introduced into the brain. It is now clear that interactions do occur between the normal CNS and immune system regardless of these features (2, 3). Immune responses within the brain are regulated and are not a passive consequence of anatomical isolation, although a regulatory role for the barriers is probable (8, 9, 10, 11). Furthermore, a modern model for CNS immunity cannot assume that humoral and cell-mediated arms of the immune response are regulated in the same manner. In our normal rat model, CSF Abs were elicited by introducing T-dependent Ags into the brain, indicating that interaction between Ag-specific B and T cells, together with Ag (native and processed, respectively), must have occurred at some point. It is known that activated T cells can enter the normal CNS regardless of Ag specificity or MHC compatibility with the host; retention of activated T cells within the brain required the MHC-restricted recognition of specific Ag (8, 9). However, little has been done to assess the competency of activated B cells to cross into the normal brain.
In the present study, three series of experiments were performed to further characterize the intrathecal Ab response, using our normal rat brain model for placing Ag behind an intact BBB (12). First, the kinetics of intrathecal Ab synthesis was established following a single infusion of Ag into CSF through an indwelling cannula, the previously published model. CSF and serum Abs were also characterized by isoelectric focusing (IEF) and immunoblotting to immobilized Ag. In the latter two series of experiments, the protocol was modified to increase the numbers of circulating Ag-specific lymphocytes, by i.m. preimmunization with Ag plus CFA. Subsequently, Ag or saline (control) was infused through the indwelling cannula. In addition to determining the kinetics of the intrathecal response and characterizing specific Abs by immunoblotting, the cannula was placed into the brain parenchyma (caudate nucleus) to facilitate immunohistochemical analysis at the site of Ag deposition. The presence of Ag-specific B cells and CD4+ T cells and the up-regulation of MHC class II expression at the site of Ag infusion were assessed. The results of this study support the hypothesis that B cells have the ability to seek Ag, even behind the BBB, and to differentiate into plasma cells in the tissue in which the Ag is encountered.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All experiments were performed using male, viral Ag-free rats (either Sprague Dawley from Charles River Breeding Laboratories, Wilmington, MA, or Lewis, from National Cancer Institute, Bethesda, MD), weighing 150 to 400 g at time of cannula implantation. Anesthesia for surgical implantation was achieved using a mixture of ketamine HCl (60 mg/kg; Aveco, Fort Dodge, IA) and xylazine (35 mg/kg; Rugby Laboratories, Rockville Center, NY) injected i.m. into the hind flank. At sacrifice (except for immunohistochemistry experiments described below), the animals were given an overdose of sodium pentobarbital (100 mg/kg; Abbott Laboratory, No. Chicago, IL) by intracardial injection.
Cannula implantation
Lateral ventricle or caudate nucleus implantations have been described in previous publications (4, 5). In earlier experiments, PE-10 polyethylene tubing (Clay Adams, Parsippany, NJ) was implanted into the lateral ventricle; for latter experiments, a stainless steel cannula (Plastics One, Roanoke, VA) was implanted into a caudate nucleus. Infusion of Ag or control fluid through the cannula (d. 0) was performed 7 days postimplantation.
Immunizations
Protocol I: cannula was implanted in lateral ventricle (d. -7); human serum albumin (HSA; 90 µg in 9 µl saline) infused on day 0; Sprague Dawley rats (n = 16) sacrificed on day 5, 9, 10, 11, 14, 15, and 21 postinfusion. Blood and CSF were taken on day of sacrifice; serum and CSF were analyzed for anti-HSA and for rat serum albumin (RSA) concentrations, as described below.
Protocol II: cannula was implanted in lateral ventricle (d. -7); HSA (1 mg/ml saline) emulsified with an equal volume of CFA (Organon Teknika, West Chester, PA) injected i.m. (1 mg HSA., day -7, after completion of implantation surgery); HSA (90 µg in 9 µl saline) or saline (9 µl) infused on day 0; Sprague Dawley rats sacrificed on either day 5 (n = 5) or day 14 (n = 6) postinfusion. Serum and CSF were analyzed for anti-HSA and RSA and for oligoclonal banding by IEF and immunoblotting, as described below.
Protocol IIIA: cannula was implanted in caudate nucleus (d. -7); OVA in CFA injected i.m. (1 mg OVA, day -7, after completion of implantation surgery); OVA (45 µg in 0.75 µl saline, n = 4) or saline (0.75 µl, n = 2) infused on day 0; Lewis rats sacrificed on day 14. Serum and CSF were analyzed for anti-OVA and RSA. Brains were fixed and removed for histology, as described below.
Protocol IIIB: cannula was implanted in caudate nucleus (d. -40); OVA in CFA injected i.m. (1 mg OVA, day -40, after completion of implantation surgery, and again on day -7); OVA (45 µg in 0.75 µl saline, n = 4) or saline (0.75 µl, n = 2) infused on day 0; Lewis rats sacrificed on day 9. Serum and CSF were analyzed for anti-OVA and RSA. Brains were fixed and removed for histology, as described below.
Assays
Blood was removed either from the tail vein or, in more recent experiments, by cardiac puncture. After clotting, serum was collected by centrifugation and frozen at -70°C until assay. CSF was withdrawn from the cisterna magna as previously described (6). An ELISA was used to measure anti-HSA or anti-OVA binding to immobilized Ag and for RSA detection using immobilized anti-RSA (6). The Ab quotient (QAb) was calculated as the ratio of CSF Ab titer x 1000 to serum Ab titer. The RSA quotient (QRSA) was calculated using CSF and serum concentrations of this protein x 1000. The Ab index (IAb), calculated as (QAb)/(QRSA), normalizes for the typical barrier leakage of protein from blood to CSF (13). The QRSA is a measure of passive "leakage" or permeability of brain barriers to RSA, a plasma protein synthesized in the liver and not in the brain. The QAb is a measure of barrier leakage to plasma Abs plus excess Ab from other sources (e.g., intrathecal synthesis, receptor-mediated transport). In animals not immunized in the CNS, QAb is less than QRSA, since normal barrier permeability is inversely proportional to protein size, and IAb is <1.0 (range 0.3–0.8 in our laboratory). If the IAb in CNS-immunized animals exceeds this range, the excess Ab is attributed to intrathecal synthesis, since we are unable to demonstrate receptor-mediated transport in this system (7). This procedure for determining intrathecal Ab synthesis is standard in the neuroimmunology literature.
Immunoblotting to detect specific Ab or total Ig was conducted on serum
and CSF samples following IEF on agarose slab gels as previously
described (14). The sample volumes loaded on gels are described in the
figure legends. After electrophoresis for 1000 V-h, one-half of the gel
was blotted (30 min, room temperature) with a 0.45-µm nitrocellulose
membrane (Sartorius, Epsom, U.K.) that had previously been coated with
Ag (0.5% HSA in saline, 0.05% sodium azide), followed by blocking in
1% skim milk. The duplicate half of the gel was blotted to a
nitrocellulose membrane. The membranes were then blocked in 1% skim
milk in saline, rinsed in saline, and immersed in 50 ml of 0.1%
milk/saline containing 50 µl peroxidase-conjugated anti-rat
-globulins (Dako, High Wycombe, U.K.) and incubated at room
temperature overnight with rocking. The membranes were thoroughly
washed in tap water and saline, then developed using ethylamino
carbazole and hydrogen peroxide.
Immunohistochemistry
Rats under deep anesthesia were perfused via the aorta with 50 ml of Dulbecco’s modified saline followed by 200 ml of 4.0% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4. Following removal, the brain was sectioned in a horizontal plane and washed extensively in 0.2 M phosphate buffer, pH 7.4. Brain sections were equilibrated overnight with 30% sucrose for cryoprotection, then frozen in OCT mounting medium (Miles Laboratories, Elkhart, IN).
Cryostat sections (5 µm thick) were mounted on precleaned glass
slides, and immunohistochemical staining was performed, as extensively
detailed previously (15). Primary murine mAbs used for tissue staining
(Serotec, Washington, DC) were: OX-6 against rat MHC class II
molecules, R7.3 specific for rat 
ß TCRs, W3/25 anti-rat CD4,
OX-8 anti-rat CD8, OX-42 against rat CD11b/c, and ED-2 identifying
a specific subset of macrophages. In addition, OVA was conjugated to
FITC for direct use as a fluorescent ligand specific for Abs and
Ab-bearing cells reactive with OVA. Biotinylated anti-rat IgG
(Vector Laboratories, Burlingame, CA) was used to identify rat plasma
and B cells, and rat-absorbed biotinylated anti-murine IgG (Vector
Laboratories) was used to localize the murine primary Abs. To localize
the biotinylated Abs, either rhodamine-conjugated avidin (Sigma
Chemical, St. Louis, MO) or peroxidase-conjugated avidin (Vector
Laboratories) was employed.
A series of double-labeling studies was performed in which OVA was biotinylated and Abs against rat CD11b/c (OX-42) or rat MHC class II (OX-6) were coupled to FITC. A 1:1 mixture of biotinylated-OVA plus either FITC-OX-42 or FITC-OX-6 was layered over the tissue and incubated overnight at 4°C. Following rinsing of the slides, the tissue was exposed to streptavidin-conjugated Texas Red (Sigma) for 2 h at room temperature, rinsed, and coverslips mounted for viewing by confocal microscopy.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In this first set of experiments, the appearance of Ag-specific Ab in CSF was characterized over a 3-wk period following a single infusion of HSA into the CSF (cannula in lateral ventricle) of Sprague Dawley rats (protocol I). Anti-HSA titers and levels of RSA in both CSF and serum were measured from individual rats at the different times postinfusion and used to calculate the IAb, which is a measure of the Ab titer in CSF normalized for protein leakage from blood to CSF (see Materials and Methods). For comparison, IAb data from two other conditions were analyzed: for rats immunized i.m. with HSA plus CFA but with no Ag infusion into CSF; and for nonimmunized (naive) rats in which total rat Ig levels measured in CSF and serum were used to calculate the index.
Figure 1
A shows IAb values at
several times, beginning on day 5, when serum anti-HSA titers are
barely detectable, through day 21, when serum anti-HSA levels are
approaching maximal values (Table I) (4).
Anti-HSA is not detectable in CSF on day 5. By day 10, anti-HSA is
detectable in only 1 of 8 rats, and the IAb value for that case is not
greater than would be expected from passive leakage of serum Ab (Fig. 1A, NAIVE or IM values). Anti-HSA Abs in CSF are increased
on day 14, and IAb values are greater than expected for passive leakage
in two of five rats. On day 21, six of six rats had detectable
anti-HSA Abs in CSF, and four of these had elevated IAb values,
ranging from 1.1 to 16.8. The indices for the other rats were in the
range of normal passive leakage (0.3–0.6). Having excluded damage to
the BBB (QRSA is in the normal range, Table I) and excluding a
mechanism of active Ig transport (IAb values of peripherally immunized
rats is 0.3–0.6, Fig. 1A), these data indicate that
intrathecal Ab synthesis contributes to the elevated anti-HSA Ab
levels in CSF for rats receiving a single CSF infusion of the Ag. The
intrathecal synthesis becomes detectable around 2 wk postinfusion.
|
|
). IgG anti-HSA was detected in CSF
with oligoclonal banding patterns in all cases. Diluting the serum to
achieve anti-HSA titers in the same range as those in CSF, only two
samples had significant IgG anti-HSA staining. This lower staining
activity is attributed to the greater heterogeneity of anti-HSA in
serum relative to CSF. The staining for total IgG in these samples
(data not shown) showed sufficient intensity to be confident of
visualizing Ag-specific Ab. The CSF findings are definitive examples of
oligoclonal banding (16), except that in this model the rats are
normal, i.e., without overt CNS pathology. These results further
strengthen the conclusion that primary administration of Ag into CSF
elicits a specific humoral immune response within the CNS.
|
In previous work, we have shown that a single CSF infusion of HSA
elicits Ag-specific Ab production in cervical lymph nodes and spleen,
with the nodes being essential for induction of the response (4). This
peripheral response may be the source of B cells generating intrathecal
Ab synthesis in the experiment described above. To increase the number
of HSA-specific lymphocytes available for trafficking to the brain and
thus strengthening the intrathecal Ab response to CNS Ag, the
experimental protocol was modified to expand the pool of HSA-specific
lymphocytes circulating in blood. Rats were preimmunized with an i.m.
injection of HSA in CFA 1 wk before CSF infusion of either HSA or
saline (control). Recirculating lymphoblasts are expected to be at a
high concentration at the time of CSF infusion (17). CSF and serum
samples were collected on day 5 and 14 postinfusion and IAb values
determined (Fig. 1
B). In addition, CSF and serum
samples were subjected to the IEF/immunoblotting procedure (Fig. 3; Table II).
|
|
B), the
IAb values are greater in the preimmunized rats that received HSA by
CSF infusion, compared with the saline infusion controls. On day 5, the
IAb values are 2 to 4 times higher (HSA vs saline) and on day 14 are 2
to 20 times higher. These results are to be compared with the previous
experiment (Fig. 1A) in which there was no detectable
CSF Ab at day 5 and no significant IAb values above the normal range
for passive leakage until day 14. The day 14 data in Figure 1B resemble the day 21 data in Figure 1A. Thus,
peripheral preimmunization has significantly accelerated the onset of
intrathecal Ab synthesis following the infusion of HSA into the CSF. It
appears that the entry and/or retention of Ag-specific lymphocytes is
enhanced within the CNS in this preimmunization model and is dependent
upon infusion of Ag into CSF.
Confirmation of this result is obtained by analyzing CSF and serum
using IEF/immunoblotting to immobilized Ag, HSA (Fig. 3; Table II).
Comparing anti-HSA intensities of the undiluted CSF and diluted
serum samples by densitometry, rats receiving an HSA infusion into CSF
have higher Ab ratios (CSF/serum) relative to the saline-infused
controls (Table II). These results confirm the data obtained using
ELISA (Fig. 1B). Two of the three rats at day 5 and
all three rats at day 14 post-CSF infusion of HSA have anti-HSA
quotients that are 2- to more than 10-fold greater than the saline
controls. The oligoclonal bands in CSF were less pronounced using this
protocol, but an example is shown in Figure 3. Due to peripheral
preimmunization with Ag in CFA, serum Ab titers are much higher and
contribute to CSF Ab titers due to the normal passive leakage of plasma
proteins into CSF. The QRSAs (Table II) from these samples, used to
calculate IAb values in Figure 1B, establish that the brain
barrier membranes are not compromised by the peripheral immunization
with CFA present. These results confirm an earlier study using two
different Ags for immunization simultaneously, OVA in CSF and HSA
+ CFA peripherally. IAb for anti-OVA was above IAb for anti-HSA
in all cases, and the range of values for albumin quotients was within
normal limits (6).
Thus, onset of intrathecal Ab synthesis can be accelerated in rats preimmunized with HSA + CFA in the periphery. Stimulation is dependent upon infusion of the preimmunizing Ag into the CSF through an indwelling cannula. Since B cells are the source of Ab-secreting plasma cells, such cells would be expected to increase in numbers within the brain of rats due to enhanced cell traffic to the brain or higher retention of Ag-reactive cells, or both.
Immunohistochemical analysis at the site of Ag deposition in rats receiving a systemic immunization of OVA plus adjuvant, followed by OVA infusion into caudate nucleus (protocols IIIA and IIIB)
To facilitate immunohistochemical analysis of the Ag-stimulated CNS, the infusion site for Ag was changed to the caudate nucleus, thereby localizing Ag to a specified region within the brain parenchyma. Rats were preimmunized at a peripheral site by i.m. injection with OVA plus CFA and subsequently received an infusion of OVA in saline via the indwelling caudate cannula. Two control groups were also analyzed: one to determine the effects of an indwelling cannula with no peripheral or central exposure to Ag (negative control); the other to determine the effects of peripheral preimmunization only (saline-infused into cannula). In the latter control, no direct Ag stimulation was administered to the brain.
In the brains of negative control rats, scattered within the tissue
lining the cannula track and the zone around the cannula tip,
macrophages and rare CD4+ and CD8+
lymphocytes were found. Also in the vicinity of the cannula tip, there
were reactive microglial cells expressing MHC class II and CD4
molecules, as would be expected around a cerebral stab wound in a rat
after 2 wk (18). Immunohistochemical staining for the presence of rat
IgG revealed only a modest blush that was closely confined to the
region of the cannula (Fig. 4,
top). There was no evidence of anti-OVA IgG, nor
were there cells of macrophage or lymphocytic morphology that bound the
FITC-OVA conjugate. Importantly, the presence of a caudate cannula
resulted in minimal reaction on the part of the brain parenchyma, and
there was no evidence of an anti-OVA response of any type (although
macrophages phagocytosing cellular debris were present in the lumen of
the cannula track). Since the negative controls were not immunized with
OVA, these latter observations confirm the specificity of the FITC-OVA
conjugate.
|
, middle). Staining of the area with FITC-OVA revealed
that positivity with this probe was coextensive with the staining for
rat IgG, an expected result because normal leakage of the high titer
anti-OVA from serum had occurred. Moreover, two types of cellular
staining could be identified. Some cells showed a hazy, diffuse
positivity of the cell membrane and granular positivity of the
cytoplasm. By their location, number, morphology, and similarity to
OX-42 staining (CD11b+) cells in other sections, it was
possible to conclude that they represented perilesional macrophages
that had become coated with anti-OVA Abs. In addition to these
macrophages, there was an occasional round cell that had a rim of
bright positivity and diffuse, less intense cytoplasmic staining with
FITC-OVA. These cells were found around the cannula track and in the
parenchyma, albeit they were few. It is presumed that they represent a
lymphocyte population that is either B cell-specific for OVA or some
other form of lymphocyte that had bound anti-OVA IgG to its
membrane.
Double-labeling studies were then performed as a way of distinguishing
OX-42 (anti-CD11b, CR3)-positive cells bearing FcR from B
lymphocyte cell lineages. For this analysis, biotinylated OVA and
FITC-OX-42 conjugates were placed on the same tissue section, with
bound biotinylated-OVA detected by streptavidin-Texas Red. New sections
were prepared for this study using confocal microscopy. Cells staining
with single conjugate alone were not detected, confirming their rarity
in preimmunized rats receiving saline through the cannula (Fig. 5). The only stained cells detected were
brownish-orange to yellowish-orange in color, indicating colocalization
of binding both the red and green conjugates, perhaps at differing
ratios. Morphologically, these cells displayed the bound conjugates as
discrete aggregates, as if they had collected within vesicles. We
identify these cells as the perilesional macrophages described
above.
|
, bottom). Staining of adjacent sections with FITC-OVA
also demonstrated a wide zone of positivity with this probe that
corresponded to the area of IgG positivity. Inspection of the cellular
components in the region lining the cannula track demonstrated the
FITC-OVA-positive macrophages and B cells/lymphocytes noted above; yet
they were both present in much higher numbers. Moreover, in the
parenchyma around the cannula site, there was now a broad zone of
microglial positivity for MHC class II and CD4, as well as enhanced
prominence of astrocytes. In the parenchyma, lymphocytes and
macrophages (or activated microglial cells) were present in notably
greater numbers than in other groups of rats. In addition, it was
possible to detect a type of cell with the FITC-OVA that had not been
noted in prior groups of rats. These cells were encountered
infrequently, but could occasionally be found singly or in small groups
and occasionally in clusters of three or four around small blood
vessels. They had a small round nucleus and abundant cytoplasm that
stained brightly and homogeneously with FITC-OVA; these were judged to
be OVA-specific plasma cells.
To confirm the presence of plasma cells in these twice-immunized rats,
the double-staining technique described above was applied to new
sections of the tissue samples (Fig. 6,
upper and lower panels). In these
sections, rats infused through the cannula with OVA 9 days previously
(protocol IIIB) were used, since this was within the time period when
intrathecal synthesis was not detectable in protocol I. Quite
convincingly, individual cells staining one of three different colors
were detected, which easily distinguishes them from the saline-infused,
peripherally immunized controls described above. Cells staining bright
green, with dendritic-like extension
(CD11b+/anti-OVA-) were frequent and
probably included the activated microglial population detected above
(MHC class II+/CD4+). They were more extensive
when Ag was present (compare with Fig. 5). Cells staining orange
(CD11b+/anti-OVA+) with the vesicular
morphology are likely to be the macrophages prominent in the
saline-infused controls (Fig. 5). Cells staining red
(CD11b-/anti-OVA+) with either a central
nucleus or brighter red with an eccentric nucleus were also frequent.
The former are presumed to be B cells displaying surface Ig with
anti-OVA specificity. The latter are presumed to be plasma cells
that have been sliced through during sectioning, allowing access of the
biotinylated OVA to the cytoplasmic Ig destined for secretion. Thus,
the presence of plasma cells and B cells in the brain parenchyma is a
distinguishing and specific feature of the peripherally immunized rats
that have received an OVA challenge into the brain parenchyma.
Peripheral preimmunization accelerates the appearance of these B cells,
which require Ag to be present in the brain parenchyma to accumulate
and possibly to differentiate. Since circulating plasma cells are
rarely found, we presume that plasma cells found in the brain
parenchyma were derived from the B cells present in a
cytokine-dependent reaction.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and 6).
Morphologically, some of these Ag-binding cells satisfy the criteria
that define B cells and/or activated B cells, while others qualify as
Ab-secreting plasma cells. Corroboration of intrathecal Ab synthesis is
also provided by the elevated IAb data (Fig. 1B) and
immunoblotting studies (Figs. 2 and 3) in which rat serum albumin
quotients (QRSA) show a normal BBB permeability (Table II).
Furthermore, we have found no evidence to support specialized transport
of Abs into normal brain (7). Together, these findings strongly support
our hypothesis that appearance of excess Abs in CSF, 2 to 3 wk after a
single Ag administration into the CNS, is the result of intrathecal Ab
synthesis by T-dependent Ag-specific effector cells of B lymphocyte
lineage. Unlike B lymphocytes, plasma cells rarely circulate in the blood; therefore, Ag-specific plasma cells found in the brain are most likely derived from activated B cells that have trafficked to the brain after being generated in peripheral lymphoid organs (4). Retention within the CNS and activation (or reactivation) of these B cells appears to be Ag-dependent, as saline-infused preimmunized rats do not accumulate B cells with this Ag specificity. The range of B cell types detected in the brain indicates that T-dependent Ag-activated B lymphocytes trafficking into the brain enter a microenvironment that can constitutively support B cell differentiation into Ab-secreting plasma cells, the end stage effector cells of humoral immunity. Some form of "T cell help" may be required for this ultimate step; this has not been ascertained. Nonetheless, it appears that the B cell to plasma cell transition in the CNS environment requires the presence of Ag.
Finding a supportive environment within the CNS for a humoral immune response is in contrast to the fate of CTL precursors (preCTL) specific for a tumor implanted in the brain (19). As described above for B cells (which one may call "plasma cell precursors"), preCTL generated in the periphery infiltrate the brain parenchyma. But unlike the progression of plasma cell precursors to plasma cells in the intrathecal humoral response, preCTL are inhibited from differentiating into active CTL, effectors of this cell-mediated immune response. This inhibition of cell-mediated effectors accounts for the immune privileged status attributed to the CNS by investigators earlier in this century (see Ref. 3 and references cited therein).
To summarize, we conclude that intrathecal Ab synthesis can occur in the presence of an intact BBB and that it is a regulated response, dependent on two concurrent conditions. One condition is the induction of a population of Ag-activated B lymphocytes in the periphery. The other is the presence or retention of Ag in the brain in a form recognizable by the immune system. While others have reported on Ab production in the brain, e.g., EAE (20, 21) and multiple sclerosis (16), it is associated with CNS diseases displaying profound CNS inflammation, a complex mixture of T cell-mediated inflammation and ongoing target tissue damage.
The importance of a peripheral source of Ag-activated lymphocytes for
CNS humoral immunity was demonstrated by experiments characterizing the
kinetics of IAb elevation in rats receiving a protein CSF infusion with
or without peripheral preimmunization (Fig. 1, B and
A, respectively). For rats receiving only a CSF infusion
(protocol I), intrathecal synthesis did not precede the appearance of
Abs in blood (Table I) and was only detectable beginning 2 wk post-CSF
infusion of protein (Fig. 1A). These results suggest
that, following Ag infusion, a period of about 14 days is required for
a series of events to occur: efflux of Ag from brain to draining
lymphoid organs, Ag-specific lymphocyte recruitment and expansion
within these organs, trafficking of activated immune cells via blood to
the brain, and their subsequent retention plus B cell differentiation
to plasma cells. In contrast, peripheral preimmunization greatly
reduced the time interval between Ag infusion into the brain and
appearance of excess Abs in the CSF; in these rats, intrathecal
synthesis was already detectable by 5 days post-CSF infusion (Figs. 1B and 3; Table II). Since Ag efflux from the CNS has been
shown to occur relatively rapidly (t1/2 =
12 h (22)), it seems most probable that the most time-consuming
events for intrathecal synthesis to become established in the brain are
related to peripheral induction of the humoral immune response, i.e.,
lymphocyte recruitment, expansion, and differentiation, which must
occur first in draining lymphoid organs and again upon encountering the
same Ag in the CNS. By preimmunizing rats, a critical number of
Ag-specific B and T lymphocytes have already been generated. These
cells are available to traffic shortly following CSF infusion of Ag and
may even be part of the existing recirculating lymphocytes, competent
to cross cerebral endothelial or epithelial boundaries, at the time of
Ag infusion into the brain.
In previous studies, we have identified two peripheral sources of Ab-secreting cells, the cervical lymph nodes and the spleen, elicited by Ag infusion into brain parenchyma or CSF (4), similar to other results (23, 24). Kida et al. (25) have also shown that, in addition to cervical lymph nodes, lumbar lymph nodes draining the spinal cord collect ventricular CSF-infused India ink particles, but these authors did not assess Ab responses. Thus, remaining to be defined are the relative roles that lymph nodes and spleen play in induction plus maturation of the peripheral Ab response and in cell trafficking. Before infusion of T-dependent Ag into the brain, removal of cervical lymph nodes blunts the serum Ab response to brain-infused Ag, yet removal of the spleen does not (4, 26). Ag infused into brain reaches cervical nodes and spleen via drainage along the olfactory nerve into the cervical lymph and through arachnoid villi into the blood, respectively. However, concentrations of draining protein Ag in cervical lymph are much greater than in the blood (27); potentially so are the cell populations in nodes vs spleen with which Ag interacts (evidenced by the peripheral Ab response to a T-independent Ag) (26). It is possible that cervical nodes play a more prominent role early in the induction of Ab responses to T-dependent Ags. Following induction, Ag-activated B cells and/or Th cells from the cervical nodes could then traffic to the spleen, reencounter draining Ag, and elicit further changes leading to secretion of Abs by cells in the spleen.
Furthermore, it is not known at what developmental stage the peripheral B cells are able to traffic from secondary lymphoid organs to the brain. Based upon similarities that may exist with T cells (8), it is tempting to speculate that stimulation of peripheral B cells by Ag plus CFA and/or reintroduction of soluble Ag into the brain elicits B lymphoblast transformation, and these lymphoblasts (both B and T) are competent to circulate, randomly seeking Ag in any tissue. The presence of Ag in the CNS leads to retention of these blast cells at the site of deposition, and subsequently, to formation of plasma cells. Others have detected B lymphoblasts in the efferent lymphatic circulation, and they may be intermediates in the trafficking of B cells from germinal centers of secondary lymphoid organs to other sites in the body where specific Ag is present (17).
Our immunohistochemistry data clearly demonstrate that the presence of plasma cells is dependent upon Ag introduction into the brain. Remarkably, there is no indication of widespread disruption of the BBB at the time when B cells and plasma cells are detected in the brain. But, examination of brain tissue within the vicinity of Ag introduction does reveal a number of changes from normal brains, besides the presence of plasma cells and B cells, viz an enhanced state of reactivity of resident parenchymal cell elements. Some changes are a consequence of cannula implantation, producing a delimited parenchymal wound eliciting a restricted microglial and astrocytic response as part of the healing process. Additionally, locally devitalized tissue would require removal; thus, a population of macrophages would be expected. These wound-healing changes are also observed in cannula only and saline-infused controls and are well advanced toward resolution at the times of observation. The other changes we observe, only in Ag-stimulated brains, demonstrate the presence of an immune reactive region, but in no way did it resemble in severity or extent the inflammatory responses observed in EAE, a well-characterized, T cell-mediated disease. The reactions we observe are far milder and more diffuse. The presence of high levels of circulating anti-OVA Abs and activated lymphocytes to the immunizing Ag made no detectable difference in this cellular response.
Comparing the pericannicular areas of saline-infused to Ag-infused
brains, a dramatic increase in the area and staining density of IgG
(Fig. 4), including anti-OVA Abs detected with FITC-OVA, was
observed. The degree of coextensive labeling with anti-IgG and
FITC-OVA was more pronounced in OVA-infused preimmunized rats.
Spreading of immunospecific activity along the cannula track is
attributed to diffusion of infused Ag along the path of least
resistance. One possible explanation is that by the time the tissue was
examined, local production of Ag-specific Ab (by plasma cells that had
differentiated behind the intact BBB) was sufficient to give the area a
diffuse IgG positivity. Alternatively, although there is no indication
of widespread disruption of the BBB, one cannot totally exclude the
possibility of a focal site of increased barrier permeability.
Colocalization of Ag-specific B and T cells, plus release of cytokines,
could enhance cellular infiltration at the site and local seepage of
plasma Ig.
A comparison of paired CSF and serum samples for anti-HSA by IEF
and immunoblotting demonstrates an important difference (Figs. 2 and 3). Oligoclonal Ab banding was detectable in the CSF of rats receiving
HSA by microinfusion with or without preimmunization. Oligoclonality is
indicative of the expansion of certain B cell clones within the CNS and
has been observed in a number of diseases that involve the CNS, e.g.,
multiple sclerosis and neurosyphilis (28, 29). However, estimating that
intrathecal synthesis accounts for at least 50% and possibly up to
80% of the anti-OVA in the CSF of rats receiving HSA by
microinfusion (Table II, day 5), expansion of B cell clones is more
widespread, i.e., polyclonal. Preferential expansion of a few B cell
clones, to account for the bands, may be due to several factors,
including proximity to a source of stimulation (Ag, cytokines) and
affinity for or dosage of infused Ag. It is unknown whether the process
leading to intrathecal B cell expansion and differentiation to plasma
cells must be initiated by activated Ag-specific T lymphocytes, since
HSA and OVA are T-dependent Ags. T cells that have detected their
cognate Ag during passage through the brain are retained for a finite
period (8, 9) and may attract other elements of the immune response.
Since CD4+ and CD8+ lymphocytes were readily
detectable near the cannula, this is a possibility. Ag-specific B cells
finding their way to the site are probably already activated (resting
lymphocytes do not cross normal endothelia (30)) and competent to
differentiate into plasma cells upon encountering their cognate Ag in
the tissue. Although T cells are the most likely candidates for
stimulating plasma cell differentiation, the possibility that B cells
obtain sufficient "help" from other cellular elements resident in
the parenchyma cannot be excluded.
Issues of B cell traffic to the brain are more significant in light of two recent reports. B cell-deficient mice exposed to encephalitogenic peptide develop EAE but are unable to down-regulate the disease (31). A role for B cells in neuroinvasion is also evident using a mouse model of the prion disease scrapie, as B cell-deficient mice fail to develop the brain pathology when prion protein is introduced by the i.p., route but are susceptible when exposed by the intracerebral route (32).
In summary, the results of this study provide evidence to strongly support the conclusion that infusion of a T-dependent protein Ag into a brain with normal barrier permeability can lead to an intrathecal humoral immune response. The sources of lymphocytes mediating this response are secondary lymphoid organs stimulated by the Ag. The primary antigenic stimulus of the lymphoid organs may be derived from the site of CNS immunization (4, 31) or introduced by peripheral injection (this study). An important issue appears to be induction of a pool of migrating, activated lymphocytes capable of entering the normal CNS. Retention of migrating B lymphocytes within the CNS and their subsequent interactions leading to Ab secretion, for example, is clearly dependent upon the presence of Ag within the CNS. This model of inducing an Ab-specific immune response provides a powerful tool for dissecting the critical cellular, cytokine, and costimulatory signals that govern B cell/Ab responses within the brain. In conclusion, we propose that activated B cells, like T cells, are competent to locate their specific Ags at any site in the body—even within the central nervous system.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Paul M. Knopf, Department of Molecular Microbiology and Immunology, Division of Biology and Medicine, Box G-B413, Brown University, Providence, RI 02912. E-mail address:
3 Abbreviations used in this paper: CNS, central nervous system; BBB, blood-brain barrier; CSF, cerebrospinal fluid; IEF, isoelectric focusing; EAE, experimental autoimmune encephalomyelitis; HSA, human serum albumin; IAb, Ab index; RSA, rat serum albumin; QRSA, RSA quotient; QAb, Ab quotient; preCTL, CTL precursor.
Received for publication January 16, 1998. Accepted for publication March 19, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
4ß1 integrin. Nature 356:63.[Medline]
4 integrin by CD4 T cells is required for their entry into brain parenchyma. J. Exp. Med. 177:57.This article has been cited by other articles:
|
|
 |
|
  M. C. Dalakas Invited Article: Inhibition of B cell functions: Implications for neurology Neurology, June 3, 2008; 70(23): 2252 - 2260. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Serafini, B. Rosicarelli, D. Franciotta, R. Magliozzi, R. Reynolds, P. Cinque, L. Andreoni, P. Trivedi, M. Salvetti, A. Faggioni, et al. Dysregulated Epstein-Barr virus infection in the multiple sclerosis brain J. Exp. Med., November 26, 2007; 204(12): 2899 - 2912. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. P. Owens, K. M. Winges, A. M. Ritchie, S. Edwards, M. P. Burgoon, L. Lehnhoff, K. Nielsen, J. Corboy, D. H. Gilden, and J. L. Bennett VH4 Gene Segments Dominate the Intrathecal Humoral Immune Response in Multiple Sclerosis J. Immunol., November 1, 2007; 179(9): 6343 - 6351. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. A. Asuni, A. Boutajangout, D. Quartermain, and E. M. Sigurdsson Immunotherapy Targeting Pathological Tau Conformers in a Tangle Mouse Model Reduces Brain Pathology with Associated Functional Improvements J. Neurosci., August 22, 2007; 27(34): 9115 - 9129. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Cepok, B. Rosche, V. Grummel, F. Vogel, D. Zhou, J. Sayn, N. Sommer, H.-P. Hartung, and B. Hemmer Short-lived plasma blasts are the main B cell effector subset during the course of multiple sclerosis Brain, July 1, 2005; 128(7): 1667 - 1676. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Krumbholz, D. Theil, T. Derfuss, A. Rosenwald, F. Schrader, C.-M. Monoranu, S. L. Kalled, D. M. Hess, B. Serafini, F. Aloisi, et al. BAFF is produced by astrocytes and up-regulated in multiple sclerosis lesions and primary central nervous system lymphoma J. Exp. Med., January 18, 2005; 201(2): 195 - 200. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Ritchie, D. H. Gilden, R. A. Williamson, M. P. Burgoon, X. Yu, K. Helm, J. R. Corboy, and G. P. Owens Comparative Analysis of the CD19+ and CD138+ Cell Antibody Repertoires in the Cerebrospinal Fluid of Patients with Multiple Sclerosis J. Immunol., July 1, 2004; 173(1): 649 - 656. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Friese, M. Platten, S. Z. Lutz, U. Naumann, S. Aulwurm, F. Bischof, H.-J. Buhring, J. Dichgans, H.-G. Rammensee, A. Steinle, et al. MICA/NKG2D-Mediated Immunogene Therapy of Experimental Gliomas Cancer Res., December 15, 2003; 63(24): 8996 - 9006. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. P. Owens, A. M. Ritchie, M. P. Burgoon, R. A. Williamson, J. R. Corboy, and D. H. Gilden Single-Cell Repertoire Analysis Demonstrates that Clonal Expansion Is a Prominent Feature of the B Cell Response in Multiple Sclerosis Cerebrospinal Fluid J. Immunol., September 1, 2003; 171(5): 2725 - 2733. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. C. Anthony, D. H |
) calculated (see Materials
and Methods). Peripherally immunized control IAb values
were calculated for rats immunized for 3 wk by i.m. injection of HSA in
CFA, but not in the CSF (IM) (
). Other control rats were not
immunized (NAIVE), but their serum and CSF were analyzed for total rat
Ig and RSA, and their respective quotients (CSF:serum) were used to
calculate the Ig index (
). B, Sprague Dawley rats were
immunized by i.m. HSA in CFA and 7 days later received either HSA in
saline (
