The long history of elevated interferon (IFN)-α in association with disease activity in patients who have systemic lupus erythematosus (SLE) has assumed high significance in the past decade, with accumulating data strongly supporting broad activation of the type I IFN pathway in cells of patients who have lupus, and association of IFN pathway activation with significant clinical manifestations of SLE and increased disease activity based on validated measures. In addition, a convincing association of IFN pathway activation with the presence of autoantibodies specific for RNA-binding proteins has contributed to delineation of an important role for Toll-like receptor activation by RNA-containing immune complexes in amplifying innate immune system activation and IFN pathway activation. Although the primary triggers of SLE and the IFN pathway remain undefined, rapid progress in lupus genetics is helping define lupus-associated genetic variants with a functional relationship to IFN production or response in patients. Together, the explosion of data and understanding related to the IFN pathway in SLE have readied the lupus community for translation of those insights to improved patient care. Patience will be needed to allow collection of clinical data and biologic specimens across multiple clinical centers required to support testing of IFN activity, IFN-inducible gene expression and chemokine gene products as candidate biomarkers. Meanwhile, promising clinical trials are moving forward to test the safety and efficacy of monoclonal antibody inhibitors of IFN-α. Other therapeutic approaches to target the IFN pathway may follow close behind.
A role for type I interferon (IFN), predominantly IFN-α, in the pathogenesis of systemic lupus erythematosus (SLE) was first suggested based on the observation that serum from patients who had active SLE disease had augmented capacity to inhibit the death of virus-infected cells. Those data, published in the late 1970s, found an association between IFN activity and the standard serologic indicators of disease activity: anti-DNA antibody titer and low complement levels. Recent re-examination of the IFN system using newer technology has heightened attention to this important immune system mediator.
Along with clinical observations showing occasional induction of lupus-like autoantibodies and clinical disease in patients treated with recombinant IFN-α for hepatitis C infection or malignancy, microarray gene expression analysis showing broad activation of IFN-inducible genes in blood cells of patients who have lupus has suggested that IFN-α may be a central player in systemic autoimmune disease. As with any association between an immune system product and a clinical syndrome, it is important to address whether IFN-α plays a pathogenic role in the disease, is an innocent bystander, or is possibly playing a protective role. Increasing data relating IFN pathway activation to studies of genetic associations, clinical characterization of patients, murine models, and results from therapeutic interventions support an important, and possibly central, role for this cytokine family in SLE. Together, the emerging data support the validity of IFN-α as a therapeutic target in patients who have SLE.
Type I interferon in immune responses
The family of type I IFNs comprises the protein products of multiple related genes encoded on the short arm of human chromosome 9. IFN-β may be the prototype member of the family, but IFN-α has the largest number of isoforms. Over time, 13 IFN-α genes have evolved, presumably selected to effectively combat infection by viruses. The type I IFNs are rapidly induced by infection with DNA and RNA viruses, either through intracellular nucleic acid receptors or after engagement of a Toll-like receptor (TLR) that recognizes nucleic acids, such as TLR3 for double-stranded RNA, TLR7 or -8 for single-stranded RNA, or TLR9 for demethylated CpG-rich DNA. IFN-α can be synthesized by many cells, but plasmacytoid dendritic cells (pDCs) represent the cell type most capable of high-level IFN-α production.
The broad expression of type I IFN receptor (IFNAR) on many cell types contributes to the diverse cellular responses induced by this cytokine family. Although T cells may be the conductors of the adaptive immune response orchestra, IFN-α is an innate immune system product that orchestrates the immune system’s initial response to viral infection before T cell activation. Immune system functions implemented by IFN-α include differentiation of monocytes into dendritic-like cells capable of effective antigen presentation, induction of natural killer and natural killer T cells, promotion of IFN-γ production, support for B-cell differentiation into class-switched antibody-producing cells, and in some cases induction of apoptosis, resulting in release of cell debris, including potentially stimulatory self-antigens.
The presenting clinical features of SLE can sometimes resemble those of some virus infections, and many of the immunologic alterations that are characteristic of SLE are similar to the immunologic effects of virus-induced IFN.
Insights from gene expression studies in patients who have systemic lupus erythematosus
The advances in technology that permitted assessment of a broad spectrum of gene products in a population of cells emerged in the late 1990s and were used to show associations between the characteristic pattern of mRNA transcripts in monoclonal populations of cancer cells and disease prognosis. Initially, experts believed that this experimental approach would not yield clinically relevant insights from microarray gene expression studies in complex diseases, including the systemic autoimmune diseases, because of the variability in representation of diverse cell types among individuals.
In fact, microarray studies of peripheral blood mononuclear cells (PBMC) from patients who had lupus were highly informative. Several laboratories analyzed large data sets and detected a pattern of gene expression rich in transcripts induced by IFNs. Among the overexpressed mRNAs were some that are well-known as targets of type I IFN, such as MX1, the OAS family, and IFIT1. However the broad pattern of increased expression of IFN-regulated genes included many induced by both type I IFNs and type II IFN (IFN-γ).
To determine the relative roles of type I and type II IFN in the IFN signature, additional studies using more quantitative real-time polymerase chain reaction (PCR) focused more narrowly on genes preferentially regulated by either type I or type II IFN. Those data clearly showed the predominant picture of increased levels of type I IFN–induced genes in lupus PBMCs. Moreover, the level of expression of those gene products across a population of patients who had lupus showed a high level of statistically significant correlation between each type I IFN-induced transcript and the others. This pattern strongly suggested that type I IFN present in vivo in many patients who had lupus was driving a broad gene expression program, similar to what has been seen in patients treated with either recombinant IFN-α or IFN-β for hepatitis C or multiple sclerosis.
Some patients who had lupus also showed increased expression of genes preferentially regulated by IFN-γ, such as CXCL9 (monokine induced by gamma interferon [MIG]), but they were less frequent than those who showed activation of the type I IFN–induced genes.
The type I IFN family includes not only multiple IFN-α isoforms but also products of related genes, including IFN-β. To determine which of these type I IFNs was most responsible for expression of the IFN-inducible genes, a functional assay of type I IFN activity in plasma or serum was developed and preferential inhibition of that activity in SLE plasmas by neutralizing antibodies to IFN-α was observed. In contrast, only modest inhibition of type I IFN activity was seen when antibodies to IFN-β or IFN- were included in the cultures. The data led the authors to suggest that IFN-α represents the major type I IFN active in vivo in patients who have SLE, but that other isoforms probably contribute a small fraction of the type I IFN activity that alters immune system function in patients who have lupus.
The proportion of patients who have lupus who show the IFN signature varies among reports. In some studies of unselected adult patients, fewer than 50% show this gene expression pattern, whereas a study of pediatric patients who had lupus, most of whom were recently diagnosed and many who had not yet been treated aggressively, saw the IFN signature in nearly all patients.
An association of IFN pathway activation with several clinical features of lupus, particularly a history of renal disease and anemia, has been shown in several cohorts, and a relative underrepresentation of IFN pathway activation has been seen in patients who have antiphospholipid antibodies. Given the acknowledged diversity of disease manifestations in patients who have lupus, along with the fluctuating course of disease, it is not surprising that differences in prevalence of the IFN signature are seen in cross-sectional studies.
The demonstration of near-universal activation of the IFN pathway in pediatric patients who have lupus, with fewer adult patients showing this pattern, raises a question of whether the production or response to IFN-α is a function of age. In that regard, a study characterizing plasma type I IFN activity in patients who have SLE and healthy first-degree relatives based on age of the subjects showed similar patterns in women and men, but distinct levels of activity based on age. The age at which plasma IFN activity was greatest corresponded to the peak reproductive years, with women between ages 12 and 22 years showing higher levels than those younger than 12 or older than 22. Women who had lupus and their first-degree relatives showed the lowest levels of type I IFN activity after 50 years of age. Men showed a similar pattern, but with a peak age range several years older (16–29 years). IFN levels were not significantly different between women and men in either the patients or relatives.
Together, these data suggest that age of the pediatric lupus cohort likely contributed to the higher prevalence of IFN signature. The molecular basis of this interesting age-related pattern of IFN pathway activation is not known.
Insights from gene expression studies in patients who have systemic lupus erythematosus
The advances in technology that permitted assessment of a broad spectrum of gene products in a population of cells emerged in the late 1990s and were used to show associations between the characteristic pattern of mRNA transcripts in monoclonal populations of cancer cells and disease prognosis. Initially, experts believed that this experimental approach would not yield clinically relevant insights from microarray gene expression studies in complex diseases, including the systemic autoimmune diseases, because of the variability in representation of diverse cell types among individuals.
In fact, microarray studies of peripheral blood mononuclear cells (PBMC) from patients who had lupus were highly informative. Several laboratories analyzed large data sets and detected a pattern of gene expression rich in transcripts induced by IFNs. Among the overexpressed mRNAs were some that are well-known as targets of type I IFN, such as MX1, the OAS family, and IFIT1. However the broad pattern of increased expression of IFN-regulated genes included many induced by both type I IFNs and type II IFN (IFN-γ).
To determine the relative roles of type I and type II IFN in the IFN signature, additional studies using more quantitative real-time polymerase chain reaction (PCR) focused more narrowly on genes preferentially regulated by either type I or type II IFN. Those data clearly showed the predominant picture of increased levels of type I IFN–induced genes in lupus PBMCs. Moreover, the level of expression of those gene products across a population of patients who had lupus showed a high level of statistically significant correlation between each type I IFN-induced transcript and the others. This pattern strongly suggested that type I IFN present in vivo in many patients who had lupus was driving a broad gene expression program, similar to what has been seen in patients treated with either recombinant IFN-α or IFN-β for hepatitis C or multiple sclerosis.
Some patients who had lupus also showed increased expression of genes preferentially regulated by IFN-γ, such as CXCL9 (monokine induced by gamma interferon [MIG]), but they were less frequent than those who showed activation of the type I IFN–induced genes.
The type I IFN family includes not only multiple IFN-α isoforms but also products of related genes, including IFN-β. To determine which of these type I IFNs was most responsible for expression of the IFN-inducible genes, a functional assay of type I IFN activity in plasma or serum was developed and preferential inhibition of that activity in SLE plasmas by neutralizing antibodies to IFN-α was observed. In contrast, only modest inhibition of type I IFN activity was seen when antibodies to IFN-β or IFN- were included in the cultures. The data led the authors to suggest that IFN-α represents the major type I IFN active in vivo in patients who have SLE, but that other isoforms probably contribute a small fraction of the type I IFN activity that alters immune system function in patients who have lupus.
The proportion of patients who have lupus who show the IFN signature varies among reports. In some studies of unselected adult patients, fewer than 50% show this gene expression pattern, whereas a study of pediatric patients who had lupus, most of whom were recently diagnosed and many who had not yet been treated aggressively, saw the IFN signature in nearly all patients.
An association of IFN pathway activation with several clinical features of lupus, particularly a history of renal disease and anemia, has been shown in several cohorts, and a relative underrepresentation of IFN pathway activation has been seen in patients who have antiphospholipid antibodies. Given the acknowledged diversity of disease manifestations in patients who have lupus, along with the fluctuating course of disease, it is not surprising that differences in prevalence of the IFN signature are seen in cross-sectional studies.
The demonstration of near-universal activation of the IFN pathway in pediatric patients who have lupus, with fewer adult patients showing this pattern, raises a question of whether the production or response to IFN-α is a function of age. In that regard, a study characterizing plasma type I IFN activity in patients who have SLE and healthy first-degree relatives based on age of the subjects showed similar patterns in women and men, but distinct levels of activity based on age. The age at which plasma IFN activity was greatest corresponded to the peak reproductive years, with women between ages 12 and 22 years showing higher levels than those younger than 12 or older than 22. Women who had lupus and their first-degree relatives showed the lowest levels of type I IFN activity after 50 years of age. Men showed a similar pattern, but with a peak age range several years older (16–29 years). IFN levels were not significantly different between women and men in either the patients or relatives.
Together, these data suggest that age of the pediatric lupus cohort likely contributed to the higher prevalence of IFN signature. The molecular basis of this interesting age-related pattern of IFN pathway activation is not known.
Association of interferon-α with disease activity in systemic lupus erythematosus
The first studies of IFN-α in SLE from the 1970s indicated that circulating levels of the cytokine were associated with serologic activity of the disease. As microarray data from studies of PBMC in carefully characterized patients emerged from several laboratories in the early 2000s, measurements using standard tools such as systemic lupus erythematosus disease activity index and a quantitative real-time PCR measurement of IFN-inducible gene expression clearly showed a relationship to clinical disease activity.
What has been less certain is the degree of fluctuation of IFN pathway activation over time in individual patients. The question remains whether the absence of IFN-inducible gene expression in PBMC or increased plasma type I IFN activity defines a distinct subset of lupus, reflects low disease activity, or indicates chronic disease that was earlier characterized by IFN pathway activation but is now burned-out, Arriving at an answer to this question has been difficult because of the inherent challenges of longitudinal clinical research studies and the technical difficulties and expense of quantifying IFN pathway activation. Validated disease activity measures are rarely applied to patients followed up in routine clinical care. Although several clinical investigator teams have established large cohorts of well-characterized patients and have followed up those patients over several years, appropriate biologic samples are rarely available and the clinical investigators who collect that data are rarely the same investigators who perform the laboratory analyses of IFN-inducible gene expression or plasma type I IFN activity.
Most commercially available enzyme-linked immunosorbent assay (ELISA) platforms for measuring IFN-α have not been useful for gaining insights into patterns of disease activity, most likely because of the limited range of IFN isoforms measured or possibly the presence of inhibitors or other plasma or serum components that obfuscate accurate measurement of the functionally active IFN. The current available data do not definitively show whether and how IFN pathway activation relates to changes in immune function or disease activity, but the authors’ group documents fluctuations in IFN-inducible gene expression in PBMC over time, in some cases, with close parallel to fluctuations in disease activity scores or response to therapy. Additional longitudinal data will help determine whether IFN pathway activation can sometimes precede flares in disease activity and whether a causal link between those events can be established. Characterizing the changes in gene expression, serologic activity, or immune function that bridge a discrete increase in IFN pathway activation and a flare in disease activity could provide invaluable novel insights into lupus pathogenesis.
Advances in research into mechanisms of IFN pathway activation in systemic lupus erythematosus
Genetic Contributions to Type I Interferon Production or Response
The major histocompatibility complex (MHC) provides the most significant contribution to the genetic variations that result in increased risk for developing SLE. Alleles of MHC-encoded class II molecules are likely involved in the capacity to generate autoantigen-specific immune responses and production of autoantibodies. Recently developed candidate gene studies have identified several additional lupus-associated gene variants, and large-scale genome-wide association studies (GWAS) and their follow-up investigations, particularly two seminal collaborative studies published in 2008, have shown additional lupus-associated variants that identify up to 25 more genes, intronic regions, or gene loci.
The single nucleotide polymorphisms (SNPs) identified in the published GWAS datasets represent common variants that are widespread in the population and confer a very modest increased risk for SLE. Additionally, several rare genetic variants associated with much greater risk for lupus-like disease have been found. Together, the increasing data on common variants conferring low risk and rare variants conferring higher risk for lupus indicate the most important molecular pathways involved in lupus pathogenesis.
When the list of lupus-associated gene variants is considered in the context of their known biologic function, the aggregate data collected strongly support the essential role of the immune system in disease pathogenesis. Characterizing the precise functions that are altered by the nucleic acid variations enriched among patients who have lupus is the next major challenge, and will require the tools of genetics, molecular biology, and cell biology to provide a new understanding of lupus disease mechanisms and identify new therapeutic targets.
However, the data generated so far can be synthesized to propose several essential components of the disease, all of which reflect genetic factors that might augment likelihood of disease development. These include increased generation or impaired clearance of self-antigens, particularly nucleic acids, and capacity to activate autoantigen-specific immune response, including T- and B-cell activation and differentiation to plasma cells. Most relevant to this review are a growing number of lupus-associated genetic variants, both common and rare, that impact type I IFN production or response.
A role for genetic variation in the increased production of type I IFN seen in many patients who have SLE was first supported by a family study in which plasma type I IFN activity was quantified in patients, their first-degree relatives, and unrelated individuals. A significant increase in IFN level was documented in healthy first-degree relatives of patients compared with unrelated subjects. Moreover, high IFN levels tended to cluster in families. The conclusion that increased plasma type I IFN was a heritable trait led to subsequent efforts to relate lupus-associated genetic variants to activation of the IFN pathway and the publication of a series of papers by Timothy Niewold identifying contributions of specific gene variants to activation of that pathway.
Abundant data have supported a complex set of SNPs in the interferon regulatory factor 5 (IRF5) gene that are associated with SLE. Dissection of that association points to a role for particular autoantibody specificities, such as anti-Ro and anti-DNA, in the association with the IRF5 risk haplotype. Moreover, that risk haplotype shows increased association with SLE and increased type I IFN activity in plasma of patients who have lupus who express those autoantibodies. Together those data link autoantibodies that target nucleic acids or nucleic acid-binding proteins, IRF5, and type I IFN production. With the knowledge that IRF5 is a signaling molecule downstream from several of the intracellular TLRs, the data support the concept that TLRs activated by DNA and RNA signal through IRF5 to induce type I IFN.
Additional lupus-associated gene variants that might modulate the TLR pathway and IFN production include IRF7 and TNFAIP3, encoding A20, an inhibitor of the TLR pathway. An association between the lupus risk variants of PTPN22, a lymphocyte phosphatase, and secreted phosphoprotein 1 (SSP1; osteopontin) and plasma type I IFN activity have also been reported, although the exact mechanisms with which those variants impact IFN production have not been elucidated. A relationship between polymorphisms in FCGRIIA, one of the first lupus-associated genes to be identified, and IFN production has not been investigated, although that might be a productive research direction considering the role that Fc receptor plays in internalizing immune complexes that induce IFN through TLRs.
The response to type I IFN depends on sequential interaction of the cytokine with the two chains of IFNAR, the type I IFN receptor, and activation of a series of kinases, including members of the signal transducer and activator of transcription (STAT) and Janus kinase (Jak) families. Although STAT1 has been most often implicated in signaling downstream of IFNAR, the STAT4 gene has been associated with SLE in several GWAS. A study of patients who had SLE with the risk allele of STAT4 showed normal or even decreased levels of plasma type I IFN activity but a significant association with increased IFN-inducible gene expression in the PBMC. That is, for a given amount of type I IFN, patients who had the STAT4 risk allele seemed to have augmented transcription of genes regulated by type I IFN.
In some individuals the genetic makeup may favor autoantibody production over IFN pathway activation. A study of serum from mothers of babies who had the neonatal lupus syndrome, in whom anti-Ro antibodies were universally present in high titer, showed that the antibodies were accompanied by high IFN activity only in those who had clinical features of SLE or Sjögren’s syndrome. These clinical data further support the concept that development of clinical lupus has several prerequisites. Some individuals have a genetic load for activation of the IFN pathway and others have increased capacity to form autoantibodies. The individuals who engage both arms of the immune system—the production of IFN through the innate immune response and the production of autoantibodies through the adaptive immune response—are most likely to develop clinical disease.
The third important prerequisite, generation or impaired clearance of cell-derived self-antigens, may complete the requirements for disease pathogenesis. In that regard, although the lupus-associated genetic variants represent common variants that result in a modest increased risk for SLE, several recently described rare variants are associated with a greater risk for disease and seem to generate increased self-antigen. One of these genes, TREX1, encodes a DNAse, and another encodes an RNase. Normal function of those gene products is required to dispose of endogenous nucleic acids that might otherwise stimulate an innate immune response. That concept is supported by studies in mice deficient in TREX1 that show increased production of type I IFN.
The accumulated data implicate gene variants involved in the intracellular nucleic acid–response TLR pathways, and at least one gene that might contribute to signaling through the type I IFN receptor pathway, in susceptibility to SLE. A role for genetic variation in components of the TLR-independent cellular pathways in induction of type I IFN production is under investigation. In addition, rare variants are being identified that could result in generation of stimuli for immune responses that result in increased type I IFN. Additional research is required to determine whether analysis of lupus risk allelic variants or type I IFN activity, or the IFN signature will be practically useful for predicting increased susceptibility to lupus in individuals otherwise at risk (eg, sisters of patients who have lupus).
Molecular Pathways Mediating Production of Interferon-α
Numerous laboratories have supported an important capacity of nucleic acid–containing immune complexes to activate immune responses, particularly the innate immune response but also B-cell proliferation, through TLR pathways. Investigators in the Ronnblom laboratory were the first to study the circulating factors in lupus serum or plasma that induced IFN-α production in vitro. They showed that apoptotic or necrotic cell debris, when associated with SLE serum, could induce IFN. That response was inhibited by blockade of Fc receptors and chloroquine. At least one proposed mechanism of chloroquine’s effects in vitro is its modification of the acidification of intracellular vesicles, which is likely to impact signaling downstream of TLRs engaged by nucleic acid ligands. A similar mechanism is likely to be operative in patients treated with hydroxychloroquine.
Studies from other groups have refined the mechanisms that account for immune complex stimulation of IFN production and have implicated TLR7, which is responsive to single-stranded RNA, and TLR9, which is responsive to hypomethylated CpG-rich DNA, in the induction of IFN by nucleic acid–containing immune complexes. Although this mechanism is difficult to document in vivo in patients who have lupus, a striking association of IFN pathway activation with the presence of RNA-binding protein-specific autoantibodies (anti-RBP), such as anti-Ro, La, Sm, or ribonucleoprotein, has been observed. The data from studies of the IRF5 lupus risk haplotype also support a functional link between anti-RBP and induction of IFN through the TLR7 pathway.
Of course the induction of IFN by immune complexes cannot fully account for the increased production of type I IFN seen in SLE, because increased plasma IFN is observed in relatives who have lupus who do not express lupus autoantibodies, and some patients who have lupus with no measurable anti-RBP or anti-DNA antibodies do have an IFN signature. However, the documented capacity of chloroquine to inhibit immune complex–mediated IFN production in vitro and the convincing data showing reduced frequency and severity of lupus flares in patients maintained on hydroxychloroquine support an important contribution of TLR signaling to clinical disease activity. The author’s current view is that induction of IFN-α by nucleic acid–containing immune complexes represents an important mechanism of augmenting type I IFN production that is influenced by genetic factors. However, it does not fully account for type I IFN produced early in the course of preclinical disease, before the development of autoantibodies. Additional studies addressing a role for environmental triggers, including virus infection, that act on a susceptible genetic substrate should provide more detailed understanding of lupus pathogenesis.
Interferon-α in Murine Lupus Models
In contrast to many aspects of lupus pathogenesis in which murine models have led the way in defining important disease mechanisms, the most significant milestones in elucidating the role of type I IFN in this disease pathway have derived from studies of patients. In fact, data from the standard murine lupus models pointed to a predominant role of IFN-γ rather than IFN-α in lupus pathology, based on knock-out and transgenic studies and documentation of increased IFN-γ levels in kidneys of mice with glomerulonephritis. A role for IFN-γ in many patients who have lupus is also supported, but the data from the human system emphasize type I rather than type II IFN as the major player.
With attention focused on the type I IFN pathway in human lupus, new data from murine experiments are now supporting an important contribution of IFN-α to mouse lupus. This role has been shown through administering IFN-α to mice in the form of an adenoviral construct that leads to sustained increased levels of that cytokine. The result is accelerated development of autoimmunity, renal disease, and death. A proposed effect of adenoviral IFN or type I IFN induced by poly (I-C), a surrogate double-stranded RNA ligand for TLR3, on recruitment of activated monocytes to kidneys may also suggest additional mechanisms of tissue fibrosis and damage that could be modified by targeting IFN-α.
A murine model with very high fidelity to human lupus, or at least the aspects of human lupus associated with IFN-α, has been studied in detail by the laboratory of Westley Reeves. Administering 2,6,10,14-tetramethylpentadecane (pristane) to healthy mice activates immature monocytes that produce type I IFN that is dependent on signaling through TLR7. Typical lupus autoantibodies are produced and the mice develop lupus nephritis. Although the precise mechanisms accounting for activation of the monocyte targets by pristane are not yet described, this model is arguably the ideal experimental system for gaining new understanding of a potential role for environmental triggers in initiating IFN pathway activation and disease in a manner highly similar to that observed in human patients.
Related posts:
Cutaneous Lupus and the Cutaneous Lupus Erythematosus Disease Area and Severity Index Instrument
Pediatric Lupus—Are There Differences in Presentation, Genetics, Response to Therapy, and Damage Accrual Compared with Adult Lupus?
The Metabolic Syndrome in Systemic Lupus Erythematosus
Cell-Bound Complement Biomarkers for Systemic Lupus Erythematosus: From Benchtop to Bedside
Glutamate Receptor Biology and its Clinical Significance in Neuropsychiatric Systemic Lupus Erythematosus
Endothelial Function and its Implications for Cardiovascular and Renal Disease in Systemic Lupus Erythematosus
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
