Sjogrens syndrome—the non-apoptotic model of glandular hypofunction

The ‘classical’ model to explain glandular hypofunction in Sjögrens syndrome (SS) is tissue loss secondary to immune attack mediated by a combination of apoptosis and cytotoxic cell death (reviewed recently by Ramos-Casals and Font [1]). In this model, the process of glandular destruction is made self-sustaining by the continued production of novel ‘self’ antigens, secondary to apoptotic bleb formation or cell-death [2]. Failure of the organ is thought to follow directly from tissue loss. For over six decades, research into SS has been directed towards understanding the mechanisms of destruction of salivary and lacrimal acinar tissue. Over this time, the only role envisaged for salivary glands themselves in the pathological process, has been that of a ‘reactor’ for infiltrating lymphocytes. Three recent observations cast a serious doubt on the principles upon which the classical model for salivary gland hypofunction in SS is based. Firstly, apoptosis of the epithelial cells in the salivary glands has been shown to be a rare event [3]. Secondly, many patients with SS who have little or no glandular function (as evidenced by markedly diminished or absent saliva output) nevertheless retain large amounts of normal-appearing acinar tissue in their salivary glands [4]. Thirdly, this residual tissue is functional in vitro [5, 6], but with a reduced sensitivity to muscarinic stimulation [5]. The latter is a crucial observation that is reflected in vivo, as many SS patients can be stimulated to secrete using systemic sialogogues [7]. According to the classical model, SS patients with no salivary flow should have no functional salivary tissue. This is clearly not the case. A ‘non-apoptotic’ model for glandular hypofunction in SS that may accommodate these findings has been proposed [4, 8–10]. In this model, glandular atrophy follows chronic immune-mediated inhibition of acinar secretory function. The many and varied mechanisms of glandular destruction, identified in association with the ‘classical’ model apply equally well to the ‘non-apoptotic’ model. What is different is that the atrophy is a consequence of salivary gland hypofunction and not the cause of it. What is most significant from any clinical perspective is that glandular destruction is clearly an irreversible process, but immune-mediated glandular hypofunction may not be. Within a ‘non-apoptotic’, ‘non-classical’ model of glandular hypofunction in SS, there is a fresh hope for both patients and clinicians because identification of the mechanisms responsible for inhibition of the secretory process may lead to the development of simplified and more sensitive diagnostic tools and ultimately to disease-modifying treatments. One of the keys to unlocking the relationship between the immune response and the secretory process in salivary glands is a detailed understanding of the fluid secretory process itself. The standard ‘simple’ model for the control of fluid secretion in salivary acinar cells is a linear series of steps as follows (Fig. 1): This model is accurate, but so simplified that to use it as a guide to understand the pathology of SS is a recipe for confusion. For example, step (ii) of the simple model above is hardly an adequate summary of the dynamic interplay of Ca2+ release and re-uptake mechanisms that underlie fluid secretion, which have been elucidated over many years and detailed in hundreds of publications from different research groups. The only prediction that it is possible to make using the simple model of secretion is that preventing Ca2+ release would prevent secretion. One way to appreciate how misleading and restrictive is this approach, is to expand step (ii) of the simple model into something closer to reality. The ACh binds to type-3 muscarinic acetylcholine receptors (M3R) on the surface of the acinar cells [11, 12]. receptor activation stimulates production of the second messenger inositol 1,4,5-trisphosphate (IP3), which diffuses into the cytoplasm and binds to IP3 receptors (IP3R) on the endoplasmic reticulum causing the release of Ca2+ into the cytoplasm; increased cytoplasmic Ca2+ activity opens Ca2+-sensitive Cl− channels on the apical membrane of the cell (and the Ca2+-sensitive K+ channels on the basolateral membrane of the cell) and Cl−, which is maintained above electrochemical equilibrium by a Na+-dependent cotransport process, passes from the cell into the duct lumen; Na+ follows Cl− across the cell in order to maintain electrochemical neutrality and the resulting osmotic gradient drags the water into the duct lumen [13]. A simple model for secretion. See text for details. Prolonged cell-wide increases in [Ca2+]i are cytotoxic. Acinar cells survive the high elevated levels of [Ca2+]i necessary for secretion, because the Ca2+ signals are neither prolonged nor cell-wide. Two patterns of Ca2+ signal that are constrained in space and/or time, but capable of supporting fluid secretion have been identified in salivary acinar cells. The first is composed of a series of brief transient increases in [Ca2+]i that are restricted to the apical pole of the cell [14–16]. The second is one or more waves of the increased [Ca2+]i that spread across the cell from the apical pole. Whilst at first seeming quite different, these two patterns of Ca2+ signal depend on the same cellular mechanisms and the cell can easily shift from one pattern to the other. The key mechanism underlying both patterns of Ca2+ signalling in acinar cells is Ca2+-induced-Ca2+ release (CICR), the process whereby the presence of Ca2+ in the cytoplasm causes further release of Ca2+ from the endoplasmic reticulum (Fig. 2). Ca2+-induced Ca2+ release. (1) and (2) Both IP3 and Ca2+ increase the open state probability of the IP3R. (3) Increased [Ca2+]i in the vicinity of the IP3R following IP3-stimulated Ca2+ release causes the channel to remain open and thus allows further Ca2+ release. Activation of a single channel also has the effect of making adjacent channels more IP3-sensitive. This positive feedback process gives Ca2+ signals their characteristic rapid kinetics and provides a powerful mechanism for amplification and propagation of the Ca2+ signals. CICR in acinar cells is mediated by both the IP3R and also by the ryanodine receptor (RyR), which is an IP3-independent Ca2+ release channel with similar Ca2+ sensitivities. Together these channels gate Ca2+ release from the endoplasmic reticulum, regulated by the levels of IP3 and cyclic ADP ribose, the natural agonist of the RyR [17–19]. In the absence of the either agonist, there is little or no CICR. Low levels of IP3 or Cyclic adenosine diphosphate ribose (cADPr) permit brief localized CICR-mediated Ca2+ signals and higher levels allow these signals to self-propagate as a Ca2+ wave across the cell (Fig. 3). Signal propagation by CICR. In the presence of elevated levels of IP3 and/or cADPr, Ca2+ release at the apical pole (1) of the cell triggers a regenerative wave of Ca2+ release (4) that utilizes both IP3R and RyR to advance through the cell. Simply introducing CICR into the mechanism for Ca2+ signalling has many consequences for the model of secretion, including: positive feedback mechanisms require very tight regulatory mechanisms to ensure that they do not run out of control. CICR is regulated first by the IP3R and RyR that respond to very high Ca2+ levels by closing [20], second by the Ca2+ buffers and organelles in the cytoplasm that prevent passive spread of Ca2+ within the cell and third by the Ca2+ ATPases that remove Ca2+ from the cytoplasm. Fluid secretion is therefore, critically dependent on any process that modifies Ca2+ buffering or Ca2+ ATPase activity [21]. whilst ACh binding and increased levels of IP3 appear necessary for initiation of the Ca2+ signal, both IP3 and cADPr appear capable of regulating signal amplification and propagation. Cyclic ADPr levels are increased by cGMP-dependent activation of ribosyl cyclase and thus, secretion is also subject to regulation by mechanisms that alter cGMP, for example the cellular levels of nitric oxide (NO). Recent data also suggest a role for cAMP in the control of Ca2+ signalling via regulation of the affinity of the IP3R for IP3 [22]. all Ca2+ signals in acinar cells, whether they propagate across the cell or not, originate at the apical pole of the cell. Ca2+ ‘tunnels’ through the endoplasmic reticulum from the basal pole to be released at the apical pole [23]. Secretion is therefore also subject to regulation via those elements of the cytoskeleton, which maintain the endoplasmic reticulum as a continuous structure within the cell. One important aspect of this modest expansion of the model of secretion, that has many ramifications for pathological interruption of the secretory process, is recognition that the secretory process involves powerful amplifying mechanisms that are not directly connected to muscarinic ACh receptor activation. For example, reduced levels of cADPr, disruption of the cytoskeleton, increased cellular buffering or reduced ATPase activity would all disrupt CICR and produce acinar cells with an inability to secrete without affecting the ability of the cell to generate IP3 in response to ACh stimulation. In short, simply determining that acinar cells respond to muscarinic receptor activation with increased Phospholipase (PLC) activity [24] and/or increased [Ca2+]i [6] is no guarantee that the cells are functioning ‘normally’. What is important is that the cells generate a Ca2+ signal that is spatially and temporally appropriate to the stimulus [5, 25]. Points 1 (ACh binding to receptor) and 4 (water movement) of the ‘simple model of secretion’ (Fig. 1) also benefit from a closer scrutiny. Salivary fluid secretion is almost entirely regulated by ACh release from parasympathetic nerves, but without the very tight anatomical relationship between nerve ending and effector cell seen in, for example, a neuromuscular junction. ACh must therefore diffuse at least 100 nm in order to bind to the ACh receptors [26]. This ‘epilemmal’ (outside the parenchymal basement membrane) [26] relationship between nerves and cells will make the secretory process vulnerable to disruption between neurotransmitter release and neurotransmitter binding by, for example, acetylcholinesterase [27, 28] or antimuscarinic receptor antibodies [24, 29, 30]. The penultimate step of the secretory process is creation of a trans-acinus osmotic gradient to drag fluid into the ducts. Water movement across the acinus may be either paracellular (between the cells) or transcellular (across the cells). The most straightforward evidence for transcellular water transport is that flow rates are reduced by 65% in knockout mice where aquaporin-5 (AQP-5), the apical membrane water channel, has been eliminated [31]. The rate of fluid secretion will be dependent on both the magnitude of the osmotic gradient, which is controlled ultimately by the Ca2+ signal and also on the water permeability of the acinar cell membrane. This last may also be regulated by ACh binding to muscarinic acetylcholine receptors via ACh-dependent insertion of AQP-5 into the apical membrane [32, 33]. However, these findings are controversial (see subsequently). In summary, whilst muscarinic receptor activation and consequent production of IP3 remain at the heart of any model of stimulus secretion coupling in acinar cells, there are many other factors that could contribute to salivary gland hypofunction in SS including antimuscarinic receptor antibody binding [24], cADPr [34], cAMP [22], cGMP [35], ATPase activity [21], mitochondria [36] and anything that affects the cytoskeleton and cellular polarity [37–39]. With the ‘expanded’ model of stimulus-secretion coupling as a guide, current studies of salivary acinar cell pathology may be divided into three areas: (i) receptor activation, (ii) mobilization of Ca2+ and (iii) fluid movement. Alterations in each may play a role in the exocrine pathology of SS. Receptor hypersensitivity followed by cellular atrophy is a well-established consequence of glandular denervation [40] and there is evidence that M3R numbers increase in the labial salivary glands of patients with SS [41]. Hypersensitivity of the lacrimal and submandibular glands to agonist has also been observed in young MRL/lpr mice, an animal model of SS characterized by defects of the Fas receptor [42]. These findings are consistent with a reduction in the amount of ACh reaching the M3R for which possible causes include: loss of innervation. This most simple explanation of receptor hypersensitivity is not supported by data from either the MRL/lpr mouse model [43] or human SS labial glands [44] that show no reduction in the density of the parasympathetic or neuropeptide neuronal innervation associated with SS. reduction in ACh release. Data from in vitro experiments using submandibular and lacrimal gland lobules isolated from the MRL/lpr mouse have suggested that the cytokines IL-1α, IL-1β and TNF-α may have a role in impairing the release of ACh from the neurones [45]. However, these findings remain to be confirmed in mice and demonstrated in human salivary glands. increased breakdown of ACh in the epilemmal space. ACh will be subject to degradation by cholinesterases as part of normal regulatory processes from the point of release by the neurones. Increased levels of cholinesterases in the salivary glands would therefore accelerate ACh breakdown and reduce the amount reaching the acinar cells [28]. We have shown that patients with SS have significantly elevated levels of cholinesterase in saliva, which probably denotes increased levels of cholinesterases in the salivary glands themselves. Cholinesterases are present in both the serum and on the surface of lymphocytes [46]. Cholinesterase could therefore enter the gland within an inflammatory infiltrate or in association with lymphocytic infiltration. This speculative hypothesis also suggests a novel mechanism for the action of hydroxychloroquine, which is an effective cholinesterase inhibitor at therapeutic doses [27] used to treat SS patients. blockade of M3R by antibodies. Over the last few years, there have been an increasing number of reports detailing autonomic symptoms in patients with SS that could be related to blockade of muscarinic receptors. These symptoms include bladder irritability [29, 47], Aides tonic pupil [48], impaired microvascular responses to cholinergic stimulation [49] and a variable heart rate [50]. All of these symptoms, as well as those of xerostomia and xerophthalmia, could be the result of blockade of M3R by antibodies [24]. Of these four possible mechanisms for acinar cell hypofunction in SS involving a reduction in the amount of ACh reaching the cells, much interest has been focused on a possible role for antimuscarinic auto-antibodies (for detailed review see Dawson et al. [51]). There are three key criteria that must be satisfied in order to give a clear indication of a role for auto-antibodies in the pathology of any disease [52]. These criteria are: the antibody interferes with the function of the target tissue. SS IgG antibodies are known to be able to bind in vitro to muscarinic receptors present on acinar cell membrane fractions [24, 53]. More importantly, recent in vitro studies demonstrating functional impairment in vital tissue have shown that: SS IgG inhibited the ACh-evoked [Ca2+]i responses in mouse submandibular acinar cells [28] and in a salivary gland cell line [54] as well as perturbing muscarinic receptor-mediated smooth muscle contraction in both bladder and colon [10, 29, 55]. passive transfer of the antibody causes the disease in the recipient. Human SS IgG caused glandular hypofunction when infused into previously healthy animals [56] as did IgG from the non-obese diabetic (NOD) mouse [56] (an animal model of SS [57]). antibody is present at significant levels only in those individuals who have disease. This is the most difficult criterion to satisfy. Preliminary in vitro data, based on a small number of patients, indicate that SS IgG antibodies capable of causing glandular hypofunction appear with a high frequency in the IgG serum fraction of patients with both primary and secondary SS [24, 29]. Recent data utilizing a novel ELISA show antimuscarinic antibodies in 66/73 (90%) patients with primary SS [58, 59]. However, IgG antibodies with antimuscarinic activity also occur in scleroderma [55]. Antimuscarinic antibodies are difficult to detect using conventional immunological techniques [30, 60] and there is no simple way at present to screen a large number of patients, therefore the true population incidence and the specificity to SS remain unknown. In summary, there is evidence to support the hypothesis that a reduction in receptor activation, probably mediated by antimuscarinic antibodies (see [51] for review), is at least a contributing factor to the glandular hypofunction seen in SS. However, these data do not exclude supporting roles for increased cholinesterase activity within the salivary glands or cytokine-mediated inhibition of neuronal ACh release. that the cytokines and TNF-α capable of [Ca2+]i by with the activity of the Ca2+ ATPase in a human salivary cell line no similar has been observed in salivary gland acinar cells, where chronic to cytokines without effect on Ca2+ in single cells [5]. However, these findings do not exclude a role for cytokines in of the necessary for between cells of a single acinus There has been little of the pathological of disruption of the of second other in levels of cADPr have on the ability of acinar cells to respond to secretory in vitro There is a clear between regulation of cADPr production and SS through is by a number of cell including lymphocytes and salivary acinar cells and increased levels of have been demonstrated in both the saliva [44] and from patients with SS the cytokines and as well as SS IgG nitric oxide activity Increased levels cGMP production and cGMP cADPr production to elevated levels of and cADPr made acinar cells to ACh stimulation as by the model of secretion and consistent with of the of RyR to CICR. Hypersensitivity of acinar cells to stimulation is the of the seen in SS, but a chronic to elevated levels of cADPr, as would be the in SS, could lead to receptor and to Preliminary data indicate that the of acinar cells to ACh following chronic to is in to whether this effect is mediated by Water movement by simple across the of the cell membrane is much to support fluid secretion water via water channels known as that increase the water permeability of the by between and have been each a tissue and defects in the of these have been with including: and Salivary acinar cells on the apical cell membrane and on the basolateral membrane production in knockout mice is and of increased and reduced in by which suggests that has role in salivary secretion. with this hypothesis is the observation that the salivary acinar cells of knockout mice have significantly reduced water permeability could have a role in the pathology of SS water movement so as to the rate step of secretion. There are data that indicate that the of is reduced at the membrane of salivary acinar cells in patients with SS but evidence from a indicate that the of in SS is the role of in SS may have been made more difficult because on the acinar cell membrane been suggested to be by the action of ACh in combination with [32, 33]. However, these in vitro findings have been by data from a recent in animals that suggested is to the apical acinar cell membrane and not in response to the inability to increase water permeability as part of stimulus secretion coupling is a contributing mechanism for glandular hypofunction in SS. a mechanism is to be it could an for by localized because the salivary glands of with using a has been shown to salivary function 4 possible of between the immune and the secretory process that could lead to glandular (i) inhibition of neurotransmitter release by (ii) breakdown of ACh by increased levels of (iii) blockade of M3R by antimuscarinic of CICR by levels of cADPr as a result of Ca2+ and or of of between the immune and the secretory process that could lead to glandular (1) inhibition of neurotransmitter release by (2) breakdown of ACh by increased levels of (3) blockade of M3R by antimuscarinic (4) of CICR by levels of cADPr as a result of Ca2+ and or of A of has been to the detailed mechanisms of apoptosis in SS There can be no doubt that apoptosis in SS, but the of apoptosis in the secretory of SS is and there is a to the that apoptosis is the primary cause of secretory [4, 8–10]. much towards of the immune response in SS, the single most glandular hypofunction as a simple consequence of apoptosis and glandular atrophy is the observation that patients have glandular tissue and no natural salivary The ‘non-apoptotic’ model of glandular hypofunction is an to the at which the immune response can the secretory process, it possible from and and acinar cells. one single factor has been that could for the disruption to the secretory process seen in SS. In the evidence to suggests that the causes of salivary gland hypofunction are The current are: (i) to which of the identified possible mechanisms have a role in (ii) to mechanisms as they are identified and (iii) to an model that can for salivary gland A functioning ‘non-apoptotic’ model of glandular hypofunction would the first guide towards development of simple diagnostic and therapeutic for SS. The have no of The of of and of