Abstract
CD8+ T lymphocytes infiltrate the brain during congenital CMV infection and promote viral clearance. However, the mechanisms by which CD8+ T cells are recruited to the brain remain unclear. Using a mouse model of congenital CMV, we found a gut-homing chemokine receptor (CCR9) was preferentially expressed in CD8+ T cells localized in the brain postinfection. In the absence of CCR9 or CCL25 (CCR9’s ligand) expression, CD8+ T cells failed to migrate to key sites of infection in the brain and protect the host from severe forms of disease. Interestingly, we found that expression of CCR9 on CD8+ T cells was also responsible for spatial temporal positioning of T cells in the brain. Collectively, our data demonstrate that the CMV-infected brain uses a similar mechanism for CD8+ T cell homing as the small intestine.
Introduction
Cytomegalovirus is the most common congenital viral infection in the United States, occurring in up to 2% of all live births (1). The brain is a major target of congenital CMV, and clinical manifestations of the disease include vision impairment, sensorineural hearing loss, seizures, epilepsy, and psychomotor retardation (2–5). Importantly, more children acquire neurodevelopmental disabilities related to CMV than Down syndrome, fetal alcohol syndrome, or spina bifida (6), making congenital CMV the leading cause of birth defects in the United States. Although the urgent need for vaccines and therapeutics against congenital CMV is widely recognized, the immune mechanisms of protection in the brain remain undefined.
Because determining the factors that contribute to protective immune responses against congenital CMV is challenging in humans, infection of mice with murine CMV (MCMV) has been used to model human CMV (HCMV) (7–10). One obstacle in developing a mouse model of congenital HCMV infection is that MCMV does not cross the placenta in immunocompetent mice. However, the immune system and CNS in newborn mice are developmentally similar to a second-trimester human fetus (11). Therefore, a number of groups have used a model in which newborn mice are infected with MCMV (12, 13). Importantly, dissemination to the brain occurs only when mice are infected near birth, and the pathological abnormalities observed in the brain are indistinguishable from those seen in human infants (14).
Using the neonatal mouse model, Britt’s group (13) showed that CD8+ T cells are critical for the resolution of MCMV brain infections. Their conclusion was based on the following findings: (1) CD8+ T cells were the dominant immune cell type recruited to the brain, (2) MCMV continued to replicate in the brain until CD8+ T cells appeared, and (3) depletion of CD8+ T cells before infection resulted in 100% mortality. Importantly, Leruez-Ville’s group (15) detected a robust CD8+ T cell response in fetal HCMV-infected brains. Although these studies and others (13, 16) highlight a key role for CD8+ T cells in host protection, an important and unanswered question is how CD8+ T cells gain entry into the brain and traffic to sites of infection. In this study, we use a murine model of congenital CMV to identify key factors regulating the migration and localization of CD8+ T cells in the brain during infection.
Materials and Methods
Animals
C57BL/6NCR and NCI B6-Ly5.1/Cr mice were purchased from Charles River. B6.PL-Thy1(a)/Cyj (Thy1.1+), B6N.129-Ccr9tm1Lov/JmfJ (CCR9 knockout [CCR9−/−]), CD4cre, and Tcratm1Mom/J (TCRα−/−) mice were purchased from Jackson Laboratory. gBT-I TCR transgenic mice (specific for the HSV-1 glycoprotein gB498–505 peptide, SSIEFARL) (17) were provided by Dr. J. Nikolich-Zugich (University of Arizona, Tucson, AZ). All experiments in this study were conducted in accordance with the Institutional Animal Care and Use Committee at Cornell University.
MCMV infection
Recombinant MCMV expressing the MHC class I–restricted CTL epitope HSV gB498–505 (SSIEFARL), designated in the text as MCMV-gB, was provided by Dr. Cicin-Sain (Helmholtz Centre for Infection Research, Braunschweig, Germany). Newborn pups (<24 h postpartum) were infected i.p. with 100 PFUs.
Single adoptive transfer and cotransfer experiments
For single adoptive transfer experiments, splenocytes were magnetically enriched with CD8a microbeads (Miltenyi). Donor cells were transferred into congenic PepBoyJ recipients i.p. at day of birth. Recipient mice were infected with MCMV-gB 4 h after donor cells were injected. To inhibit circulating CCL25, we injected mice i.p. with a neutralizing anti-CCL25 blocking Ab or rat IgG2B Isotype Control (5 μg/mouse; R&D Systems) at days 1, 4, 7, 10, and 13 postinfection.
Flow cytometry
mAbs for mouse CD4 (GK1.5), CD8a (53-6.7) CCR9/CD199 (CW1.2), CCR6/CD196 (29-2L17), CXCR7 (8F11-M16), CXCR3/CD183 (CXCR3-173), CD103 (M290), Thy1.1/CD90.1 (OX-7), TCRβ (H57-597), Perforin (S16009A) CCR5/CD195 (HM-CCR5 [7A4]), CXCR4/CD184 (2B11), α4β7/LPAM-1 (DATK32), CD45.1 (A20), CD45.2 (104), TNF-α (MP6-XT22), IFN-γ (XMG1.2), and Granzyme B (GB11) were purchased from BD Biosciences, BioLegend, or ThermoFisher Scientific. Anti-mouse CD16/CD32 (93) and Fixable Viability Dye eFluor 780 were purchased from ThermoFisher Scientific. Kb:gB498-505 tetramer (tet) was obtained from the National Institutes of Health Tetramer Core Facility. For surface staining, cells were processed using IC fixation kit (Invitrogen). Intracellular staining was performed with Foxp3/Transcription Factor Staining Buffer (Invitrogen).
Quantitative real-time PCR
Total RNA from brains was isolated using a RNeasy Lipid Tissue mini kit (Qiagen) and cDNA generated using iScript RT Supermix for RT-qPCR (Bio-Rad) according to the manufacturer’s instructions. RNA transcripts of interest were then quantified from the cDNA using PerfeCta SYBR Green FastMix, Low ROX (Quanta Biosciences) according to the manufacturer’s instructions.
Immunohistochemistry
Brains were removed whole from euthanized, perfused mice and cut along the longitudinal fissure to separate the left and right hemispheres, placed into cassettes (Histo-tek), and processed for imaging using standard immunohistochemistry (IHC) techniques.
Results and Discussion
CCR9 is preferentially expressed on CD8+ T cells in the brains of MCMV-infected pups
In general, CD8+ T cells use chemokine receptors to migrate to peripheral sites of infection (18, 19). To identify the chemokine receptors associated with CD8+ T cell homing to the brain after MCMV infection, we used flow cytometry and measured the expression of various chemokine receptors on Ag-specific CD8+ T cells in the brain. Because MCMV elicits a wide variety of epitope-specific CD8+ T cells, we used a recombinant strain of MCMV that expresses the HSV-1 gB-8p peptide (denoted MCMV-gB) and results in an overwhelmingly dominant gB-specific CD8+ T cell response (20). Postinfection of newborn mice, we tracked gB-specific CD8+ T cells throughout the response using tet+ (Supplemental Fig. 1A). On day 17 (the peak of the CD8+ T cell response), we harvested the brains and spleens and compared the expression of different chemokine receptors on gB-specific CD8+ T cells that were previously shown to play a role in the recruitment of T cells to sites of infection (18, 19).
Unexpectedly, the only chemokine receptor that was preferentially expressed on gB-specific CD8+ T cells in the brain compared with the spleen was CCR9 (Fig. 1A, 1B). These results were surprising because CCR9 is believed to be primarily involved in lymphocyte homing to the gut epithelium and has never been associated with lymphocyte trafficking to the brain (21). To validate our findings, we used quantitative PCR and examined gene expression levels of CCR9 and its ligand (CCL25) in the brain at various times after MCMV infection. The expression levels of Ccr9 and Ccl25 increased after 7 d of infection and peaked on day 14, which is consistent with the kinetics of CD8+ T cell recruitment to the brain (Fig. 1C, 1D, Supplemental Fig. 1A). We also performed IHC and observed significantly more CCL25 staining in the brains of mice congenitally infected with MCMV than controls (Fig. 1E). These results suggest a possible role for the CCR9/CCL25 axis in the recruitment of CD8+ T cells to the brain after congenital MCMV infection.
Given the higher levels of CCR9 on CD8+ T cells in the brains of MCMV-infected mice, we wondered whether they might also express adhesion molecules that are typically found on gut-homing CD8+ T cells. Previous work has shown that CCR9+ CD8+ T cells follow a CCL25 chemokine gradient to the intestines, where they upregulate the integrins α4β7 and αE(CD103) β7 (22, 23). α4β7 and CD103 allow lymphocytes to bind to cell adhesion molecules on the endothelium and epithelium. Interestingly, we found elevated levels of the gut-homing markers α4β7 and CD103 on CD8+ T cells in the brain compared with the spleen (Supplemental Fig. 1B). Moreover, α4β7 and CD103 expression were preferentially found on the CD8+ T cells in the brain that also expressed CCR9 (Supplemental Fig. 1C). We also directly compared the phenotype of CD8+ T cells across different peripheral organs and found that CD8+ T cells in the brain more closely resemble those in the intestine than the spleen (Supplemental Fig. 1D). Thus, CD8+ T cells in the brains of MCMV-infected pups are phenotypically similar to CD8+ T cells homing to the gut.
CD8+ T cells use the CCR9/CCL25 axis to traffic into the brain
We next sought to determine the functional relevance of CCR9 on MCMV-specific CD8+ T cells during congenital CMV. To do this, we infected wild-type (WT) and CCR9−/− mice and compared the recruitment of CD8+ T cells to the brain after neonatal MCMV infection. We found a significantly higher number of gB-specific CD8+ T cells in the brains of WT mice at the peak of the response (day 17) (Supplemental Fig. 2A). In addition, the CD8+ T cells in the brains of WT mice expressed higher levels of CD103 and α4β7 compared with CCR9−/− animals (Supplemental Fig. 2B). These findings suggest CCR9 plays an important role in CD8+ T cell trafficking to the brain after MCMV infection. However, a caveat of this experiment is that CCR9 deficiency is not specifically limited to CD8+ T cells. Thus, we could not rule out the possibility that impaired trafficking of CD8+ T cells to the brains of CCR9−/− animals is due to a deficiency of CCR9 in other lymphocyte populations. To test whether CCR9 alters the migration of CD8+ T cells to the brain in a cell-intrinsic manner, we employed an adoptive cell transfer approach, whereby only the donor CD8+ T cells transferred into WT congenic recipient mice were deficient in CCR9 expression. This was done by crossing our CCR9−/− mice to mice that are transgenic for a rearranged TCR specific to the gB peptide (gBT-I mice). Equivalent numbers of donor CD8+ cells from neonatal WT or CCR9−/− gBT-I mice were injected into newborn congenic recipients, which were subsequently infected with MCMV-gB (Fig. 2A). At day 17 postinfection, we found significantly fewer donor CCR9−/− cells in the brains of recipient mice compared with WT controls (Fig. 2B). Importantly, we did not observe a significant difference in WT and CCR9−/− donor cells in the spleen (Supplemental Fig. 2C), suggesting that the reduction of CCR9−/− donor cells in the brains of infected mice is due to a trafficking defect.
We also compared the phenotype and function of WT and CCR9−/− donor cells in the brains of MCMV-infected mice. Consistent with our findings from the direct infection experiments, donor CCR9−/− cells expressed lower amounts of CD103 than WT donor cells (Supplemental Fig. 2D). To assess immune functionality, we restimulated WT and CCR9−/− donor cells from the brain with gB-peptide and compared their ability to produce cytokines and effector molecules. We did not observe any significant differences in the production of TNF-α, IFN-γ, Perforin, or Granzyme B (Supplemental Fig. 2E). Thus, although CCR9 facilitates homing of CD8+ T cells to the brain, it does not affect their ability to deploy effector functions at the site of infection.
To confirm that the specific ligand for CCR9 (CCL25) is required for CD8+ T cell trafficking to the brain, we performed another adoptive transfer experiment. This time, we adoptively transferred WT gBT-I donor cells into newborn congenic recipients that were either treated with a CCL25 blocking Ab (anti-CCL25) or IgG as a control every 3 d until day 13 postinfection (Fig. 2C). This blocking Ab has been previously validated for gut homing (24). In our study, we observed significantly more donor CD8+ T cells in the brains of mice given IgG control Ab than those receiving the CCL25 blocking Ab, suggesting that CCL25 is needed for CD8+ T cell trafficking to the brain during congenital MCMV infection (Fig. 2D). There was also significantly less CD103 and α4β7 expression on donor CD8+ T cells from recipients treated with anti-CCL25 compared with those treated with IgG (Supplemental Fig. 3A). Because the blocking Ab was systemically administered, it is possible that our treatment altered the ability of CD8+ T cells to expand in the periphery. However, this possibility seems unlikely given that we found similar numbers of donor cells in the spleens of recipients treated with IgG and anti-CCL25 (Supplemental Fig. 3B) and no difference in CCR9 expression (Supplemental Fig. 3C). Collectively, these findings suggest that the CCR9/CCL25 axis is a critical mechanism by which CD8+ T cells enter the brains of congenitally infected animals with MCMV.
CCR9 expression influences the microanatomical position of CD8+ T cells in MCMV-infected brains
An important question is whether CCR9 also influences the positioning of CD8+ T cells within the brain. Clinical imaging experiments have shown that congenital HCMV infection leads to brain malformations in particular regions, such as the thalamus and dentate gyrus of the hippocampus (25, 26) (Fig. 3A). To determine whether CCR9 influences the localization of CD8+ T cells in the brain, we infected WT and CCR9−/− mice at birth and stained brain sections on day 17 with a CD8 Ab. We found a significant influx of CD8+ T cells into the thalamus of both infected WT and CCR9−/− animals compared with uninfected controls (Fig. 3B, 3C). However, there was no difference in CD8 staining in the thalamus between infected WT and infected CCR9−/− mice (Fig. 3B, 3C). In contrast, we observed significantly more CD8 staining in the hippocampus of WT mice compared with CCR9−/− mice during infection (Fig. 3B, 3C). Importantly, we observed similar amounts of CCL25 staining in the hippocampus of WT and CCR9−/− animals (Supplemental Fig. 3D, 3E), suggesting that the reduction of CD8+ T cells in the hippocampus of CCR9−/− animals was due to a difference in CCR9 expression and not a difference in chemokine expression.
We next asked whether the altered distribution of CD8+ T cells within the brains of WT and CCR9−/− mice impacted the pattern of viral replication. To answer this question, we infected WT and CCR9−/− mice with MCMV that contained a GFP insert (MCMV-GFP), allowing us to visualize the virus through fluorescent microscopy. We examined the ratio of number of infected (GFP+) cells in the hippocampus to the thalamus. There was a significant increase in the ratio of GFP+ cells in the hippocampus compared with the thalamus (Fig. 3D, 3E). These data suggest that lack of CD8+ cells in the hippocampus may lead to an impaired ability to control viral replication in this compartment of the brain. Thus, not only does CCR9 expression affect trafficking of CD8+ T cells into the brain, but it also influences their location and potentially their ability to limit viral replication within certain regions of the brain.
CCR9+ CD8+ T cells limit disease severity in MCMV-infected hosts
A hallmark of congenital MCMV infection is smaller cerebellum size (14). To determine whether recruitment of CCR9+ CD8+ T cells is associated with good or poor clinical outcomes, we used our IHC samples and compared the cerebellum size in WT and CCR9−/− mice. We observed a pronounced reduction in cerebellum size in CCR9−/− mice postinfection (Fig. 4A). We also performed Nissl staining to identify neuron morphology and cellularity. Strikingly, we noticed malformations in the hippocampal region of CCR9−/− mice, which were particularly pronounced in the CA1 region (Fig. 4B). This hippocampal malformation in the CCR9−/− mice was characterized by a smaller band of neurons, and their morphology was atypical and had less defined nuclei and Nissl bodies than infected WT mice (Fig. 4B). To confirm that physiological differences between WT and CCR9−/− were due to a CD8+ T cell–specific phenomenon, we adoptively transferred WT or CCR9−/− donor CD8+ T cells into TCRα−/− mice and compared their ability to thrive during infection (Fig. 4C). Recipients that received CCR9−/− CD8+ T cells had a significant reduction in weight gain throughout the first 12 d of infection (Fig. 4D). On day 12, pup tails were measured as a surrogate for postnatal development. Mice that received CCR9−/− donor cells had significantly shorter tails than WT recipients (Fig. 4E). We also performed plaque assays and detected a higher viral load in the brains of CCR9−/− recipient mice on day 12 (Fig. 4F), which is the peak of viral replication in the brain during MCMV infection (13, 20). Collectively, these data suggest that CCR9+ CD8+ T cells curtail neurological and physiological sequalae during congenital MCMV infection.
In summary, we discovered that the CMV-infected brain uses similar mechanisms of CD8+ T cell homing as the small intestine. This conclusion is based on the following set of findings: (1) the gut-homing chemokine receptor CCR9 is selectively expressed in CD8+ T cells localized in the brain after MCMV infection; (2) the ligand for CCR9 (CCL25) is expressed in the brain at the peak of the CD8+ T cell response; (3) other gut-homing and tissue residency markers, α4β7 and CD103, are coexpressed on CCR9+ CD8+ T cells in the brains of neonates; and (4) inhibition of CCR9/CCL25 signaling impairs CD8+ T cell recruitment to the brain. Our findings are significant because the CCR9/CCL25 axis is believed to be primarily involved in lymphocyte homing to the gut epithelium and has never been associated with lymphocyte trafficking to the brain. Now that we have identified key molecules involved in the recruitment of CD8+ T cells to the brain, it may be possible to design more effective strategies to treat infants suffering from CMV brain infections.
In the future, it will be important to identify the factors leading to the upregulation of CCR9 on CD8+ T cells in the CMV-infected brain. In the mesenteric lymph nodes, retinoic acid produced by dendritic cells is largely responsible for the upregulation of CCR9 and α4β7 in lymphocytes, as well as homing to the gut (27). Because MCMV is a systemic infection, it is possible that CCR9 is induced in CD8+ T cells in the intestine before migrating to the brain. However, it is important to mention that retinoic acid is produced in the brain during early stages of development (28). Thus, it is also possible that CCR9 is upregulated in CD8+ T cells during priming in the cervical lymph nodes or in the brain after localization. It will also be important to determine whether other chemokine receptors play a role in CD8+ T cell homing to the brain during infection. Our experiments suggest that expression of CCR9 accounts for ∼50% of the CD8+ T cells that traffic into the brain during congenital CMV infection. How the other 50% of CD8+ T cells gain entry remains an open question. Knowledge acquired from these studies may provide us with a solid platform for enhancing T cell immunity in the brain during early stages of development.
Footnotes
This work was supported by the National Institute of Allergy and Infectious Diseases, National Institutes of Health (Grants R01AI105265 and R21AI110613).
The online version of this article contains supplemental material.
References
Disclosures
The authors have no financial conflicts of interest.