Harmony and discord in endothelial calcium entry

This letter addresses the recent report from Abdullaev et al¹. In 2006, the seminal work of Feske et al² identified membrane proteins which were named Orais, after the Greek keepers of Heaven’s gate: Harmony (Orai1), Justice (Orai2) and Peace (Orai3). The proteins emerged through a study of severe combined immune deficiency (SCID), which is caused by a defect in Ca²⁺-entry of T-lymphocytes. A mutation in Orai1 gene underlies the defect.

There has been particular interest in Orais because of their relationship to the widely-occurring store-operated Ca²⁺-entry (SOCE) phenomena, which reflect Ca²⁺-permeable plasma membrane mechanisms that respond to the depletion status of endoplasmic reticular Ca²⁺ stores. In lymphocytes the store-operated mechanism is commonly observed as a Ca²⁺-selective and inwardly rectifying current, I-CRAC. The mechanism is defective in SCID and rescued by expression of wild-type Orai1². In predicted structure Orais belong to the tetraspanin protein family³ rather than resembling known ion channels but mutations in Orais modify ion-selectivity, suggesting Orais generate ion permeation⁴.

Endothelial I-CRAC-like phenomenon

Abdullaev et al’s search for I-CRAC in cultured endothelial cells¹ adds to an expanding view that the importance of Orai1 and I-CRAC is not restricted to the immune system. A striking feature of I-CRAC is its small amplitude; well below the size of most other whole-cell currents and close to the resolving power of whole-cell patch-clamp. Abdullaev et al¹ are to be commended for persisting when the current turned out to be at least 5 times smaller than the current of immune cells. Unfortunately, the current could only be convincingly shown in the absence of divalent cations, leaving open the possibility that Ca²⁺ shuts it down and, strictly, not enabling it to be described as I-CRAC (i.e. a Ca²⁺ current). Nevertheless, the current was lost when Orai1 was knocked-down by RNAi. An ionic current was also observed in the presence of a high concentration of extracellular Ca²⁺, but barely above background noise, making it difficult to define. Therefore, Abdullaev et al¹ observed an I-CRAC-like phenomenon linked to Orai1 but its existence in the presence of Ca²⁺ is unsure.

Although an I-CRAC-like phenomenon was observed using a common I-CRAC protocol¹, there is little direct evidence it depended on store-depletion. All recordings of a clear I-CRAC-like signal used a patch pipette containing a high concentration of the Ca²⁺ chelator, BAPTA. While BAPTA depletes intracellular stores, it also lowers the global cytosolic Ca²⁺ concentration. Therefore, observation of current with BAPTA in the pipette is not proof that the current arose because of store-depletion. Therefore, Abdullaev et al¹ also used another method for store-depletion (application of the SERCA inhibitor, thapsigargin). However, the observed current had unexpected characteristics: It developed abruptly and then declined (unlike I-CRAC of immune cells), the current-voltage relationship was not like that of I-CRAC, and gadolinium-sensitivity was not demonstrated because the current declined before gadolinium was applied. Ca²⁺- activated current could have been a confounding factor because the Ca²⁺-free BAPTA available for buffering may have been relatively little, especially given the dialysis time required for such large cells.

TRPC discord

Abdullaev et al¹ addressed the contribution of canonical transient receptor potential (TRPC) proteins because previous studies of endothelial SOCE all suggested TRPCs contribute the underlying ion permeation pathway^5-9. In studies of HUVECs (the same cell type as used by Abdullaev et al¹) store-depletion evoked non-selective cationic current that was abolished by dominant-negative TRPC3 construct. In other studies of HUVECs (and other endothelial cells), store-depletion evoked a similar non-selective cationic current that was enhanced by over-expressing TRPC1 and suppressed by antisense DNA or antibody targeting TRPC1. In other endothelial cells, a mutant form of TRPC4 inhibited Ca²⁺ current observed during store-depletion. Current resembling I-CRAC was observed in endothelial cells from wild-type but not TRPC4 knockout mice.

The TRPC hypothesis has attractions because TRPCs are bona fide ion pore-forming proteins with Ca²⁺ permeability. However, TRPCs do not have the Ca²⁺-selectivity and inward rectification of I-CRAC and are activated by a multitude of factors¹⁰, such that association with or dependence on store-depletion could be a relatively minor aspect of their biology. Nevertheless, independent investigators have published data suggesting TRPCs are involved in endothelial SOCE as well as associated I-CRAC-like or non-selective cationic currents. Resolutions to the apparent discord may come through answers to very specific questions, such as: (1) Do TRPCs activate when physiological concentrations of physiological agonists evoke store-depletion in physiological conditions? Probably yes. (2) Are TRPCs activated by only some experimental protocols designed to evoke store-depletion? Probably yes. (3) When the only event is store-depletion, do TRPCs respond? Although many investigators are keen to answer this question, it is currently impossible to be sure that store-depletion is the only event. (4) Do protocols isolating I-CRAC shut down other mechanisms? Almost certainly yes.

Abdullaev et al exclude TRPCs based on RNAi experiments aimed at knocking down expression of TRPC1 or TRPC4¹, two of the six human TRPCs. Demonstration of effective knock-down is important in such a situation. Unfortunately, the size of the protein suggested to be TRPC1 is 20 % larger than TRPC1’s predicted mass, over-expressed TRPC1 clone, or native TRPC1 validated in studies of TRPC1−/− mice. While glycosylation of TRPC1 is suggested, supporting data are not provided and the only potential N-linked glycosylation site is weak. While the band intensity was reduced by TRPC1 siRNA, other bands labeled by the antibody were also decreased, as was the amount of ß-actin. Although TRPC1 RNAi affected cell proliferation, the western blotting fails to confer confidence that the effect resulted from knock-down of TRPC1. The TRPC4 data, however, appear convincing.

What are we to conclude when previous studies provide data suggesting TRPCs contribute to SOCE of endothelial cells^5-9, where as Abdullaev et al’s data indicate no role¹? Abdullaev et al¹ suggest that most previous studies relied on antibodies that yielded non-specific effects, but the antibody data comprise only a small component of the endothelial cell evidence. Alternatively, could the details of the experimental protocol be crucial in determining whether an investigator observes a TRPC contribution? This would not be surprising but, if so, where does the difference lie and does it matter for understanding the biology?

Because I-CRAC is so small, particular conditions are needed to remove other, larger, unwanted currents. Might these conditions compromise the function of TRPCs? The Ca²⁺ concentration may be critical. Store-depletion can lead to a rise in cytoplasmic Ca²⁺, stimulating Ca²⁺-activated ion channels. However, this is not what is usually meant by store-operated channels, rather that the channels sense only the depletion status of the stores. To exclude changes in Ca²⁺ the experimenter must tightly buffer cytosolic Ca²⁺. “Ca²⁺ add-back” experiments using a Ca²⁺ indicator dye do not meet this condition. Commonly I-CRAC is studied with inclusion of a high concentration of BAPTA in the patch pipette, which helps to deplete the stores and buffer Ca²⁺, but also reduces cytosolic Ca²⁺ to sub-physiological levels. Less commonly, I-CRAC is evoked by a SERCA inhibitor during buffering of cytosolic Ca²⁺ at the physiological concentration. Almost always, two non-physiological conditions are employed in the extracellular medium: a high concentration of Ca²⁺ (10-20 mM) or a DVF (divalent cation free) solution. Such manoeuvres aid observation of I-CRAC but may divert experimenters from TRPCs, which have complex Ca²⁺ dependencies and regulation.

An additional factor to consider is IP₃. Notably, it was included in the patch pipette for the TRPC4−/−⁶ but not Abdullaev et al¹ studies. TRPC channels couple to IP₃ receptors and so involvement of TRPCs in SOCE may conceivably depend on a permissive concentration of IP₃. However, non-selective cationic store-operated current has been observed in endothelial cells without IP₃ in the patch pipette^5,8,9 and activation of TRPCs by other stimuli often does not depend on including IP₃ in the patch pipette. If permissive IP₃ is nevertheless required (or just important), we know that it would usually be present when agonists cause Ca²⁺-release. In contrast, physiological concentrations of agonists may evoke Ca²⁺-release without there being appreciable store-depletion. It is unclear how often cells experience store-depletion of the type evoked by thapsigargin or intracellular BAPTA.

Potential for harmony

If Orai1 does indeed confer a widespread I-CRAC, Abdullaev et al’s data show us that the current may often be so small in physiological conditions that it is undetectable.

Nevertheless, the Ca²⁺-entry may be sufficient to alter behaviour of other ion channels, especially if they are physically coupled to Orai. Such close relationships between ion channels are precedented, for example between voltage-gated Ca²⁺ channels and Ca²⁺-activated K⁺ channels, and different ligand-gated ion channels. Similarly, Orais and TRPCs may form complex molecular arrangements or webs. There is biochemical evidence for Orais interacting with TRPCs, either directly or via the auxiliary STIM1 protein and knock-down of Orai1 can abolish current carried by over-expressed TRPC1^11-13. Perhaps Orais and TRPCs are separable by experimental conditions but often partners in physiology. There is much work to do and Abdullaev et al’s study makes an important contribution to the campaign to solve these complex processes.