Distinctive effects of T cell subsets in neuronal injury induced by co-cultured splenocytes in vitro and by in vivo stroke in mice

Abstract

Background and purpose

T cells and their subsets modulate ischemic brain injury. We studied the effects of the absence of T cell subsets on brain infarction after in vivo stroke and then used an in vitro co-culture system of splenocytes and neurons to further identify the roles of T cell subsets in neuronal death.

Methods

Stroke was induced by MCA suture occlusion in mice and infarct sizes were measured 2 days post-stroke.

Splenocytes were co-cultured with neurons, and neuronal survival was measured 3 days later.

Results

A deficiency of both T and B cells (SCID) and the paucity of CD4 or CD8 T cells equally resulted in smaller infarct sizes as measured 2 days post-stroke. Although a functional deficiency of regulatory T cells had no effect, impaired Th1 immunity reduced infarction and impaired Th2 immunity aggravated brain injury, which may be due to an inhibited and enhanced inflammatory response in mice deficient in Th1 and Th2 immunity, respectively. In the in vitro co-culture system, WT splenocytes resulted in dose-dependent neuronal death. The neurotoxicity of splenocytes from the above immunodeficient mice was consistent with their effects on stroke in vivo , except for the mice with the paucity of CD4 or CD8 T cells, which did not alter the ratio of neuronal death.

Conclusion

T cell subsets play critical roles in brain injury induced by stroke. The detrimental versus beneficial effects of Th1 cells and Th2 cells both in vivo and in vitro reveal differential therapeutic target strategies for stroke treatment.

Introduction

Inflammation-mediated brain injury is aggravated by both the innate and adaptive immune systems.¹ The innate immune response mainly involves macrophages and neutrophils, while the adaptive immune response is regulated by T cells and B cells. T cells consist of CD4 and CD8 (cytotoxic T or Tc) T cells. Based on the cytokine profiles and functionality, CD4 T cells are divided into type 1 helper T cells (Th1), Th2 cells and regulatory T (Treg) cells. ² While Th1 cells are pro-inflammatory, both Th2 cells and Treg cells are anti-inflammatory. ³ The adaptive immune system and innate immune cells may crosstalk to regulate brain injury.

The detrimental effects of macrophages and neutrophils in ischemic brain injury have been extensively studied.^4–11 Macrophages, which are derived from brain resident microglia and/or circulating monocytes, are activated and recruited a few hours after stroke onset. Macrophages significantly exacerbate brain infarction. Neutrophils are also recruited into the ischemic brain and contribute to brain injury. Nevertheless, only a few groups have studied the roles of T cells and their subsets, as well as functional subsets of CD4 T cells in ischemic brain injury. Hurn et al. showed T cell deficiency robustly reduced infarct size post-stroke.¹² Kleinschnitz et al. and Yalmaz et al. showed that the absence of CD4 T cells or CD8 T cells resulted in a similar reduction in infarct size as measured 1 day post-stroke.^{13, 14} In addition, Ansel et al. showed that Treg cells were responsible for late phase of infarction as measured at 1 week but not for acute infarction measured 1 to 3 days post-stroke.¹⁵ Most recently, IL4 deficiency in mice, which results in Th2 impairment,¹⁶ led to a larger infarction measured at 2 days post-stroke.¹⁷ How Th1 deficiency affects brain injury has not been reported, and no studies have compared the deficiency of T cell subsets or functionally distinct CD4 T cell subsets on stroke-induced brain injury in a single experimental setting. We thus investigated these subsets and directly compared their effects on stroke.

Little is known about the protective or detrimental mechanisms of these distinctive T cell and CD4 T cell subsets in brain injury induced by stroke. T cells are recruited into the ischemic brain as early as 24h after stroke.^{12, 14} However, whether physical contact between T cells and neurons causes neuronal death is unknown. To further identify T cell subsets critical to brain injury we co-cultured splenocytes and neurons in a co-culture system, allowing a direct contact between lymphocytes and neurons. We compared the effects on neuronal death of the splenocytes from mice with the deficiency in total lymphocytes, CD8 T cells, CD4 T cells, or functional distinct CD4 T cells subsets (Th1, Th2 or Treg cells). Moreover, we examined acute brain infarct sizes post-stroke in these mice and compared differences between in vitro and in vivo results. Last, we investigated whether Th1 or Th2 deficiency affects brain injury by regulating the activity of innate immune cells such as macrophages and neutrophils in an in vivo experimental setting.

Materials and Methods

Animals

Mice from Jackson Laboratory (Bar Harbor, ME) were bred at the Stanford animal facility, except for the C57BL/6J WT mice. Immune-deficient animals used included SCID mice with a B6 genetic background, CD8 T cell-deficient mice (B6.129S2-Tap1^tm1Arp/J), CD4 T cell-impaired mice (B6.129S2-H2^dlAb1-Ea/J), Th1 impaired mice (B6.129S2-Mapk9^tm1Flv/J), Th2-impaired mice (C57BL/6-Il4^tm1Nnt/J), and Treg-impaired mice (B6.129X1-Ebi3^tm1Rsb/J). In addition, we used an EGFP mouse (C57BL/6-Tg(CAG-EGFP)131Osb/LeySopJ for tracking whether splenocytes were washed away in co-culture with neurons. All protocols were approved by the Stanford Institutional Animal Care and Use Committee and conducted according to the NIH Guide for Care and Use of Laboratory Animals.

Stroke model, neurological deficit scores and CBF measurement

Male mice, 10 to 12 weeks old (25 – 30g), were anesthetized with 2% isoflurane in 70% air and balanced O₂ using a face mask. Middle cerebral artery occlusion (MCAO) was induced by the insertion of a silicone-coated 6-0 monofilament (Doccol Corp, Redlands, CA) into the MCA internal carotid artery for 1 hr followed by reperfusion, as described¹⁷. Rectal temperature was maintained at 37±0.5 °C with a heating pad (Harvard Apparatus, Holliston, MA). Heart rate, oxygen saturation, and respiratory rate were monitored continuously (STARR Life Sciences Corp, Allison Park, PA). Animals with no observable deficits immediately after ischemia, those that died within 48 hrs, and those with subarachnoid hemorrhage at the time of death were excluded from analysis. Brains were sectioned coronally after 48 hrs into 4 slices and stained in 2% 2,3,5-triphenyletrazolium chloride (TTC). Infarct size was analyzed, normalized blindly to the non-ischemic hemisphere and expressed as a percentage and corrected for edema using the NIH Image program (Image J 1.37v). Neurological deficit scores were evaluated at 48 hrs according to a neurological grading score,¹⁸ from 0 ( no observable neurological deficit) to 4 (unable to walk spontaneously and a depressed level of consciousness). The evaluator was blinded to experimental treatments. In separate animals, regional cerebral blood flow (rCBF) was monitored through a microtip fiber optic probe (diameter 0.5 mm) connected through a Master Probe to a laser Doppler computerized main unit (PerFlux 5000, Perimed AB, Stockholm, Sweden). A small incision was made at the coordinate 1 mm caudal to the bregma and 3.3 mm lateral to the midline in the ischemic hemisphere to expose the skull, and the laser Doppler probe was attached to the exposed skull. CBF was measured 10 min before ischemia onset, during (30 min after stroke onset) ischemia and 5 min after reperfusion. All data were normalized to the values of CBF measured before stroke in wild type animals and expressed as relative ratios.

Co-culture of pure cortical neurons or mixed neurons and astrocytes with splenocytes

A primary cortical neuronal culture (purity of neurons were >99%) was conducted as described previously.¹⁹ Cortex from fetal Swiss Webster mice (Charles Rivers, Wilmington, MA) at 15 days gestation was dissected, collected and digested in 0.25% trypsin with EDTA (Invitrogen, Carlsbad, USA). Cells (4.5×10⁵/ml) were planted at 200µl/well in 96-well plates and incubated at 37°C, 5%CO₂. After 26–30 hrs, 60% of the plating medium was replaced by glia conditioned medium with B-27 serum-free supplement. Cytosine arabinoside (3 µM) was added to inhibit glial cell proliferation. Cells were continuously incubated without further medium changes until used for co-culture with splenocytes at 9 days in vitro.

To generate mixture of neurons and astrocytes primary culture, astrocyte cultures were first prepared as previously described²⁰. Briefly, cortices freed of meninges were dissected from newborn (days 1–3) Swiss-Webster mice, minced and treated with 0.25% trypsin with EDTA (Invitrogen, Carlsbad, CA, U.S.A) for 40 min at 37°C in a water bath. The cells were resuspended in DMEM containing high glucose, L-glutamine and supplemented with 10% equine serum, 10% fetal bovine serum, and 10 ng/ml epidermal growth factor (Sigma-Aldrich, St. Louis, MO, USA) and pipetted to a single-cell suspension. The cell suspension (200µl/well) was plated in Falcon Primaria 96-well plates (Becton-Dickinson, Franklin Lakes, NJ, USA) at a density of 2 hemispheres/10 ml. On day 10–11, when the astrocytes were 100% confluence, the medium was changed to high glucose DMEM with 10% equine serum. On day 14, after completely removing the astrocyte growth medium, cortical neurons were dissected and planted on the astrocyte monolayer by 0.9×10⁵/well in DMEM supplemented with 5% equine serum and 5% fetal bovine serum. Cells were then 1/2 fed by the above medium twice per week. On day 9 of the neuronal-astrocyte culture, cells were cocultured with splenocytes.

To prepare the splenocyte cultures, mouse spleens were chopped into small pieces, homogenized, filtered through a 70µm strainer twice and recovered to a 50 ml solution. The solution was centrifuged at 1200rpm for 5min at room temperature (RT). After removal of the supernatant, 5ml ACK Lysing buffer (GIBCO, Invitrogen, Carlsbad, CA, U.S.A) was added to the cell sediment to lyses erythrocytes. After incubation for 5min at room RT, RPMI1640 was added to stop the lyses reaction, and centrifuged at 1200rpm for 5min at RT. The cell pellet was re-suspended in 5ml Dulbecco's modified Eagle's medium (DMEM, Invitrogen, Carlsbad, CA, U.S.A) and counted.

Splenocytes were added into pure neuronal cultures or mixed neuron cultures on day 9 after they were prepared. The co-culture system was incubated for another 72 hrs. Splenocytes were removed completely by repeated wash and 100µl DMEM+5%ES+1%PS was added to each well. Next 10µl CCK-8 solution was added to each well and incubated at 37°, 5%CO2 for 1.5 hrs. The absorbance at 450nm was measured by microplate reader with reference wavelength at 650nm.

Immunofluorescent staining

Ischemic or sham-operated mice 48 hrs after stroke onset were euthanized with an overdose of isoflurane and perfused with icy PBS, then embedded in OCT tissue frozen medium. The brain was cut into 5µm sections and fixed with cold acetone. Immunofluorescent staining was carried out under moderate shaking. All washes and incubations were done in 0.1M PBS (pH 7.4). Sections were incubated for 1 hr with blocking solution (0.1M PBS, 5% equine serum). After washes, sections were incubated overnight at 4°C with rat anti-mouse primary antibody for CD68 (diluted 1:200; MCA1957GA, AbD Serotec, Kidlington, Oxford, UK), a marker for reactive macrophages/microglia, or a rabbit anti-human myeloperoxidase antibody (MPO, diluted 1:50, #A0398, Dako North America, Inc, Carpinteria, CA, USA). Sections were then rinsed and incubated for 2 hrs at RT with an Alexa 488-conjugated goat anti-rabbit (for MPO) or Alexa 488-conjugated goat anti-rat (for CD68 positive macrophages/microglia, diluted 1:200, Invitrogen, Carlsbad, CA, U.S.A) secondary antibodies. The sections were then washed and covered using Vectashield mounting medium with 4', 6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA). A negative control without primary antibodies was performed in parallel.

The expression of CD68 or MPO was investigated using the optical fractionator method on epifluorescent photomicrographs (Zeiss axiovert inverted scanning microscope, Zeiss, Germany). For each animal, 3 sections were choose for an average value per mouse and the number of immunoreactive cells (for both CD68 and MPO) in the predefined infarct area was counted using Image J software (Image J 1.37v, Wayne Rasband, available through National Institutes of Health). All counts were performed on coded sections to blind the investigator to the treatment group.

Statistics

All results were presented as mean ± S.E.M. Statistical analyses were performed by ANOVA followed by the Student-Newman-Keuls post hoc test. Tests were considered significant at P-values < 0.05.

Results

Distinctive effects of the deficiency of T cell subsets or CD4 T cell subsets in ischemic stroke in vivo

We measured changes in CBF during stroke and after reperfusion. CBF was reduced to about 27% at the measured cortex and recovered to about 87% after reperfusion compared with that before stroke. These values were not significantly different in mice with the deficiency in either T cell subset or functional CD4 T cell subset (Table-1).

Table-1.

CBF was measured before stroke onset, during stroke, and after reperfusion.

Genotype	Before MCAO	During MCAO	After MCAO
WT	1.000±0.038	0.2706±0.0124	0.8580±0.03281
SCID	1.031±0.036	0.2632±0.0078	0.8830±0.03304
Tap-1 KO(CD8-)	1.019±0.047	0.2642±0.0162	0.8617±0.04082
MHCIIKO(CD4-)	1.084±0.039	0.2660±0.0116	0.941±0.0417
JNK2 KO(Th1-)	1.082±0.0349	0.2642±0.0154	0.8476±0.03626
IL-4 KO(Th2-)	1.016±0.051	0.2503±0.016	0.8645±9.04335
Ebi3 KO(Treg-)	1.040±0.035	0.2531±0.0155	0.8941±0.03568

Infarct sizes were measured 2 days after stroke. A deficiency of both T cells and B cells, or CD4 , or CD8 T cells, resulted in similarly reduced infarct sizes (Fig.1). For CD4 T cell subsets, Th2 deficiency aggravated infarct size and Th1 deficiency inhibited infarct size. The functional deficiency of Treg, however, did not affect infarct size (Fig.1). Changes in neurological scores following stroke in these animals were consistent with infarct sizes.

Fig. 1.

The effects of immune deficiency on infarct size and neurological scores post-stroke. Data were divided into part I and part II to focus on comparing T cell subsets and CD4 T cell subsets, respectively. Part I includes the deficiency or paucity of T and B cells, CD4 or CD8 cells. Part II includes the paucity of CD4 T cells or their functional deficiency (Th1, Th2 or Treg). The deficiency for MHCII and Tap-1 that markedly reduce the number of CD4 T cells and CD8 T cells, respectively are indicated on the figure. A. Infarct sizes. Infarct sizes were measured 2 days post stroke. The top panels are representative ischemic brains with TTC staining. The bottom bar graphs represent average infarct sizes. Although Th1 deficiency resulted in smaller infarct volumes than the paucity of CD4 T cells, no significant difference was observed. B. Neurological scores. Neurological scores were measured 2 days after stroke. Bar graphs corresponding to infarct sizes compare the effects of T cell and CD4 T cell subset deficiencies on neurological scores. * WT, wild type. *, **, *** vs WT, P<0.05, 0.01, 0.001, respectively.

Th1 deficiency inhibited while Th2 deficiency promoted the inflammatory response

We examined the effects of Th1 or Th2 deficiency on macrophage and neutrophil activity 2 days after stroke. Immunostaining showed robust protein expression of CD68, the macrophage activity marker, and MPO, the neutrophil activity marker, in the ischemic brains of WT mice (Fig. 2A). However, Th1 deficiency inhibited while Th2 deficiency promoted their expression (Fig.2B).

Fig. 2.

Th1 deficiency inhibited while Th2 deficiency promoted inflammation after stroke. A. The top picture shows a representative coronal brain section with cresyl/violet staining on which the square represents the area where pictures of immunostaining were taken and cells were counted. The bottom two panels are representative immunostaining of the macrophage activity marker, CD68, and neutrophil activity marker, MPO, at 48 hrs after stroke in WT-, Th1- and Th2-deficient mice. B. Statistical results of CD68- and MPO-positive cell numbers. The immune positive cells were counted and the numbers for WT-, Th1- and Th2-deficient mice are portrayed in the bar graphs. *, ** vs WT, P<0.05, 0.01, respectively.

The neurotoxicity of splenocytes in vitro

We used the CCK-8 kit to measure neuronal survival after washing away co-cultured splenocytes in vitro. To confirm that splenocytes were completely removed before measuring neuronal survival, we added EGFP-positive splenocytes into cultured pure neurons to prove that almost all EGFP-positive splenocytes were removed from the co-culture after washing (Fig.3A). This ensures that measured cell survival was exclusively derived from the neurons.

Fig. 3.

Establishment of in vitro co-culturing system of splenocytes with pure neurons. A. Splenocytes dose-dependently caused neuronal death in pure neuronal culture. Splenocytes were added to pure neuronal culture at various ratios (1:0.4 to 1:10) of neurons to splenocytes. Higher density of splenocytes resulted in more neuronal death. Neuronal survival was measured by the CCK-8 kit 3 days after co-culturing, and transformed into a ratio of neuronal death. B. Splenocytes did not cause neuron death in mixed culture of neurons and astrocytes with the ratio of 1:5 (mixed neurons and astrocytes:splenocytes). Note that survival rate was not transferred into a ratio of neuronal death in this part of the figure. C. The microscopy study suggests that splenocytes were washed away from the co-culture system before neuronal survival was measured. To confirm that splenocytes were removed before measuring neuronal survival, EGFP positive splenocytes were added to the pure neuronal culture. Pictures were taken 3 days later before or after washing. The phase contrast image shows that neurons (arrows) and lymphocytes (arrow heads) co-existed in the culture. The EGFP fluorescent image further shows the presence of EGFP-positive lymphocytes in the culture before washing. However, after washing with media, almost all EGFP splenocytes were washed away, and only MAP-2 positive neurons remained. The picture taken from the culture with vehicle treatment without lymphocytes shows higher densities of MAP-2 positive neurons than the culture treated with lymphocytes. Scale bar, 20µm.

We determined the optimal ratio of splenocytes to neurons being able to induce neuronal death. Addition of splenocytes to the culture of pure neurons resulted in a dose-dependent neuronal death (Fig. 3B). For the remaining experiments, we selected the ratio of 1:5 (neurons:splenocytes), which resulted in approximately 40% neuronal death. We then added splenocytes to a mixed culture of neurons and astrocytes to test for neuron/astrocyte death using the same ratio of 1:5. In this case, splenocytes did not cause cell death. (Fig.3B).

We further investigated the effects of splenocytes from several immunodeficient animals on cultured neurons. The results show that splenocytes from SCID mice but not from CD4- or CD8-deficient mice, resulted in less neuronal death compared with splenocytes from WT mice (Fig.4). Th1 deficient splenocytes caused less neuronal death while Th2 deficient splenocytes significantly aggravated neuronal death. Splenocytes from functional Treg deficient mice had no effect (Fig.4).

Fig. 4.

The effects of splenocytes on neuronal death in the co-culture system. Similarly to Fig.1, the data were divided into Part I and Part II. Part I. Comparison of the paucity of T cell subsets on neuronal death. Splenocytes with total lymphocyte deficiency (T and B cells in SCID mice) and the paucity of CD4 T cells or CD8 T cells in co-culture resulted in less neuronal death. Part II: The effects of the deficiency of individual functional CD4 T cell subsets on neuronal death. The deficient genes and corresponding phenotypes are also labeled. The experimental numbers on the bar graphs represent the total well numbers for each cell type, which were repeated 3 times on different days. WT, wild type. **, *** vs WT, P<0.01, 0.001, respectively.

Discussion

To the best of our knowledge our study is the first to systemically examine the effects of splenocytes and several immunodeficiency on neuronal death in an in vitro co-culture system, which offers a valuable method to study the effects of lymphocytes on neuronal death. We found that WT splenocytes killed neurons in the pure neuronal culture in a dose-dependent manner. WT splenocytes at a ratio of 1:5 (neuron:lymphocytes) resulted in about 40% neuronal death. Nevertheless, WT splenocytes did not cause cell death when co-cultured with the mixture of astrocytes and neurons, suggesting that astrocytes may play a protective role against lymphocyte-induced neuron death, which is consistent with many previous studies on the neuroprotective effects of astrocytes. Using an in vitro co-culture system with pure neurons, we found that splenocytes from SCID mice, but not from the mice with the paucity of CD4 or CD8 T cells , resulted in less neuronal death compared with WT splenocytes. We also found that neuronal death was inhibited by Th1 deficiency, enhanced by Th2 deficiency, and unaltered by functional deficiency of Treg cells.

The in vitro results are largely consistent with the in vivo pathological outcomes, except for that the paucity of CD4 T cells and CD8 T cells in mice resulted in equal reductions in infarct sizes in vivo. The underlying mechanisms for this discrepancy between in vitro and in vivo experiments are unknown. Major factors contributing to the in vitro setting may include different physical interactions between neurons and T cells, various cytokine releases, the lack of ischemia, or the absence of in situ reactions of microglia and other cell types. However, our in vitro data may suggest that the direct physical contact of CD4 T cells or CD8 T cells with neurons is not critical for their detrimental effects. CD4 T cells and CD8 T cells may indirectly affect ischemic brain injury by altering the functions of macrophages, the final direct effectors on neuronal death in vivo. In addition, we found that functional deficiency of Treg cells did not alter neuronal death and acute infarction both in vitro and in vivo; these results are consistent with previous reports.¹⁵

We demonstrated the first evidence that Th1 cell deficiency resulted in a robust reduction in infarct size and neuronal death in vivo and in vitro. A previous study showed that the IL-4 deficiency in mice resulted in a larger infarction, suggesting that the Th2 response is neuroprotective.¹⁷ Both Th1 and Th2 are subsets of CD4 T cells. Th1 is pro-inflammatory by polarizing macrophages into classical activated macrophages (M1) while Th2 is anti-inflammatory by polarizing alternative activated macrophages (M2).²¹ In our study, it appears that Th1 deficiency resulted in even smaller infarct sizes than CD4 T cell deficiency, though significance is not reached. This is understandable because in Th1-deficient mice the pro-inflammatory effect of Th1 was abolished while the anti-inflammatory effects of Th2 remained. However, both pro- and anti-inflammatory factors are depleted in animals with CD4 T cells deficiency. Indeed, we found that Th1 deficiency reduced while Th2 deficiency promoted the inflammatory response, as evidenced by the differential protein expression of CD68 and MPO in the ischemic region after stroke in Th1- and Th2-deficient animals.

In conclusion, both in vitro and in vivo studies suggest distinctive effects of T cell subsets as well as CD4 T cell subsets in ischemic brain injury. The distinctive effects of the Th1 versus Th2 response in inflammation indicate selective therapeutic targets for stroke treatment.