While it is well established that there are robust circadian rhythms of arginine vasopressin (AVP) in the cerebrospinal fluid (CSF), the route whereby the peptide reaches the CSF is not clear. A , AVP neurons constitute the largest fraction of the SCN neuronal population. Here, we show that processes of AVP-expressing SCN neurons cross the epithelium of the 3rd ventricular wall to reach the CSF (black arrows). Additionally, we report rostro-caudal differences in AVP neuron size and demonstrate that the localization of cells expressing the clock protein PER2 extend beyond the AVP population, thereby indicating that the size of this nucleus is somewhat larger than previously understood. B , Following lateral ventricle (LV) injection of cholera toxin β subunit (CTβ ; magenta) the retrograde tracer is seen in AVP neurons of the SCN, supporting the anatomical evidence that AVP neuronal processes directly contact the CSF.Arginine vasopressin (AVP) expressing neurons form the major population in the brain's circadian clock located in the hypothalamic suprachiasmatic nucleus (SCN). They participate in inter-neuronal coupling and provide an output signal for synchronizing daily rhythms. AVP is present at high concentrations in the cerebrospinal fluid (CSF) and fluctuates on a circadian timescale. While it is assumed that rhythms in CSF AVP are of SCN origin, a route of communication between these compartments has not been delineated. Using immunochemistry (ICC) and cell filling techniques, we determine the morphology and location of AVP neurons in mouse and delineate their axonal and dendritic processes. Cholera toxin β subunit (CTβ) tracer injected into the lateral ventricle tests whether AVP neurons communicate with CSF. Most importantly, the results indicate that AVP neurons lie in close proximity to the third ventricle, and their processes cross the ventricular wall into the CSF. We also report that contrary to widely held assumptions, AVP neurons do not fully delineate the SCN borders as PER2 expression extends beyond the AVP region. Also, AVP neurons form a rostral prong originating in the SCN medial-most and ventral-most aspect. AVP is lacking in the mid-dorsal shell but does occur at the base of the SCN just above the optic tract. Finally, neurons of the rostral SCN are smaller than those lying caudally. These findings extend our understanding of AVP signaling potential, demonstrate the heterogeneity of AVP neurons, and highlight limits in using this peptide to delineate the mouse SCN.

The suprachiasmatic nucleus (SCN) of the hypothalamus regulates daily rhythms in physiology and behavior. It is composed of a heterogeneous population of cells that together form the circuits underlying its master clock function. Numerous studies suggest the existence of two regions that have been termed core and shell. At a gross level, differences between these regions map to distinct functional differences, although the specific role(s) of various peptidergic cellular phenotypes remains unknown. In mouse, gastrin-releasing peptide (GRP) cells lie in the core, are directly retinorecipient, and lack detectable rhythmicity in clock gene expression, raising interest in their role in the SCN. Here, we provide evidence that calbindin-expressing cells of perinatal mouse SCN express GRP, identified by a green fluorescent protein (GFP+), but lack detectable calbindin later in development. To explore the intra-SCN network in which GRP neurons participate, individual GFP+ cells were filled with tracer and their morphological characteristics, processes, and connections, as well as those of their non-GFP-containing immediate neighbors, were compared. The results show that GFP+ neurons form a dense network of local circuits within the core, revealed by appositions on other GFP+ cells and by the presence of dye-coupled cells. Dendrites and axons of GFP+ cells make appositions on arginine vasopressin neurons, whereas non-GFP cells have a less extensive fiber network, largely confined to the region of GFP+ cells. The results point to specialized circuitry within the SCN, presumably supporting synchronization of neural activity and reciprocal communication between core and shell regions.

Immune checkpoint inhibitors targeting programmed cell death protein 1 (PD-1) have been highly successful in the treatment of cancer. While PD-1 expression has been widely investigated, its role in CD4+ effector T cells in the setting of health and cancer remains unclear, particularly in the setting of glioblastoma multiforme (GBM), the most aggressive and common form of brain cancer. We examined the functional and molecular features of PD-1+CD4+CD25-CD127+Foxp3-effector cells in healthy subjects and in patients with GBM. In healthy subjects, we found that PD-1+CD4+ effector cells are dysfunctional: they do not proliferate but can secrete large quantities of IFNγ. Strikingly, blocking antibodies against PD-1 did not rescue proliferation. RNA-sequencing revealed features of exhaustion in PD-1+ CD4 effectors. In the context of GBM, tumors were enriched in PD-1+ CD4+ effectors that were similarly dysfunctional and unable to proliferate. Furthermore, we found enrichment of PD-1+TIM-3+ CD4+ effectors in tumors, suggesting that co-blockade of PD-1 and TIM-3 in GBM may be therapeutically beneficial. RNA-sequencing of blood and tumors from GBM patients revealed distinct differences between CD4+ effectors from both compartments with enrichment in multiple gene sets from tumor infiltrating PD-1-CD4+ effectors cells. Enrichment of these gene sets in tumor suggests a more metabolically active cell state with signaling through other co-receptors. PD-1 expression on CD4 cells identifies a dysfunctional subset refractory to rescue with PD-1 blocking antibodies, suggesting that the influence of immune checkpoint inhibitors may involve recovery of function in the PD-1-CD4+ T cell compartment. Additionally, co-blockade of PD-1 and TIM-3 in GBM may be therapeutically beneficial.