EP4049029A1

EP4049029A1 - Diagnostic and prognostic biomarkers of disease remission in rheumatoid arthritis

Info

Publication number: EP4049029A1
Application number: EP20796556.7A
Authority: EP
Inventors: Mariola KUROWSKA-STOLARSKA; Iain Mcinnes; Stefano ALIVERNINI; Elisa GREMESE
Original assignee: University of Glasgow
Current assignee: University of Glasgow
Priority date: 2019-10-21
Filing date: 2020-10-21
Publication date: 2022-08-31
Also published as: GB201915222D0; WO2021078799A1

Abstract

The present invention is based on the identification of a network of biomarkers which are characteristic of disease remission in patients with rheumatoid arthritis. The biomarkers provide a sensitive read-out of patient pathological status which allows treatment to be tailored, administered, evaluated and withdrawn in the appropriate clinical circumstances avoiding disease relapse, flare and unnecessary treatment with toxic and costly drugs. Accordingly, the present invention relates to novel diagnostic assays for disease remission in patients with rheumatoid arthritis. The invention also relates to methods, devices and kits for identifying patients in remission, evaluating therapeutic effectiveness in achieving remission and predicting the likelihood of relapse.

Description

Diagnostic and prognostic biomarkers of disease remission in rheumatoid arthritis

Field of the Invention

The present invention is based on the identification of a network of biomarkers which are characteristic of disease remission in patients with rheumatoid arthritis. Accordingly, the present invention relates to novel diagnostic assays for disease remission in patients with rheumatoid arthritis. The invention also relates to methods, devices and kits for identifying patients in remission, evaluating therapeutic effectiveness in achieving remission and predicting the likelihood of relapse.

Background to the Invention

Rheumatoid Arthritis (RA), the most common inflammatory arthropathy, is characterised by breach of self-tolerance to post-translationally modified self-proteins and chronic synovitis. Current therapies mainly target inflammatory cytokine and receptor pathways, or cells of adaptive immunity. While these deliver therapeutic benefit, partial or non- response in up to ~50% of patients remain a significant challenge. Furthermore, approximately half of those who respond will relapse within months of treatment- reduction or cessation (Nagy et al., 2015, Arthritis Res Ther 17, 181; Alivernini et al., 2016, Arthritis Res Ther 18, 39). Treatment-refractory rheumatoid arthritis (RA) is a major clinical challenge.

However, long-term drug-free remission does occur and provides proof-of-concept that articular immune-homeostasis can be reinstated in RA (Alivernini et al., 2017, Ann Rheum Dis, doi: 10.1136/annrheumdis-2016-210424). The cellular and molecular mechanisms mediating such remission are unknown yet could offer remarkable opportunity to enact cure, informing both RA and other immune-mediated disease therapeutics.

The healthy synovial membrane is a specialized, multifunctional structure consisting of a lining layer of synovial fibroblasts and macrophages, a supporting sublining layer of loose connective tissue with sublining fibroblasts, and a rich network of nerves and blood/lymphatic vasculature (Firestein et al., 2017, Kelley and Firestein's textbook of rheumatology. Tenth edition edn.). The origin of human synovial tissue resident macrophages (STMs) is unclear; however recent mouse studies suggest they originate from embryonic precursors, distinct from adult bone-marrow monocyte precursors, that populate the synovium during early embryogenesis and proliferate in situ (Misharin et al., 2014, Cell Rep 9, 591-604; Gomez Perdiguero et al., 2015, Nature 518, 547-551 ; Tu etal., 2019, Front Immunol 10, 1146). Their joint-specific functions in humans remain undefined but emerging murine evidence suggests that they help maintain joint immune-homeostasis (Misharin etal., 2014, Cell Rep 9, 591-604; Culemann et al., 2019 Nature, doi: 10.1038/s41586-019-1471 -1 ). Recent elegant work by Culemann et al. found that murine lining layer macrophages are long-lived, locally-renewing and form a protective epithelial-like barrier (Culemann etal., 2019 Nature, doi: 10.1038/s41586-019- 1471-1). In active RA, the synovial membrane becomes populated by many subsets of leukocytes including macrophages (Mandelin et al., 2018, Arthritis Rheumatol 70, 841- 854; Kurowska-Stolarska et al., 2017, RMD Open 3, e000527, doi: 10.1136/rmdopen- 2017-000527; Udalova etal., 2016, Nat Rev Rheumatol 12, 472-485, doi: 10.1038/nrrheum.2016.91). Human (Herenius etal., 2011, Ann Rheum Dis 70, 1160-1162) and mouse (Misharin et al., 2014, Cell Rep 9, 591-604; Weiss et al., Proc Natl Acad Sci U S A 112, 11001-11006) studies suggest that the latter may differentiate locally from blood-derived monocytes recruited by synovial chemokines (Yeo et al.,

2016, Ann Rheum Dis 75, 763-771), and comprise the main producers of pathogenic TNF (Alivernini etal., 2017, Ann Rheum Dis, doi: 10.1136/annrheumdis-2016-210424; Kurowska-Stolarska et al., 2017, RMD Open 3, e000527, doi: 10.1136/rmdopen-2017- 000527; Firestein etal., 2017, Immunity 46, 183-196; Kurowska-Stolarska et al., 2011, Proc Natl Acad Sci U S A 108, 11193-11198). Recently, single-cell transcriptome sequencing (scRNAseq) has revealed complex heterogeneity of both synovial fibroblasts (Croft et al., 2019, Nature 570, 246-251; Zhang et al., 2019, Nat Immunol 20, 928-942, doi: 10.1038/S41590-019-0378-1; Mizoguchi etal., 2018, Nat Commun 9, 789, doi: 10.1038/s41467-018-02892-y; Stephenson etal., 2018, Nat Commun 9, 791, doi: 10.1038/s41467-017-02659-x) and synovial tissue macrophages (Zhang et al.,

2019, Nat Immunol 20, 928-942, doi: 10.1038/s41590-019-0378-1 ; Kuo etal., 2019, Sci Transl Med 11, doi: 10.1126/scitranslmed.aau8587) from patients with active RA and osteoarthritis. A distinct biology associated with different synovial fibroblast clusters has since been experimentally validated and will have important implications for therapies aimed at distinct modulation of inflammation or tissue repair (Croft etal., 2019, Nature 570, 246-251). In contrast, identifying the phenotypic and functional heterogeneity of human synovial myeloid populations has been hampered by the relatively few cells sequenced, restricted functional validation ex vivo, and the lack of comparative phenotype profiles of macrophages in healthy and remission human synovium (Zhang et al. , 2019, Nat Immunol 20, 928-942, doi: 10.1038/s41590-019-0378-1 ; Kuo etal., 2019, Sci Transl Med 11, doi: 10.1126/scitranslmed.aau8587). Our earlier studies (Alivernini et al., 2016, Arthritis Res Ther 18, 39; Alivernini et al., 2017, Ann Rheum Dis, doi: 10.1136/annrheumdis-2016-210424; Gremese et al., 2018, Ann Rheum Dis 77, 1391-1393; Najm etal., 2018, Arthritis Res Ther 20, 265, doi: 10.1186/s13075-018- 1762-1) demonstrated that most synovial inflammation resolves in RA in sustained clinical and ultrasound remission (defined by attenuation of synovial hypertrophy, normalised blood flow and histology (Alivernini etal., 2016, Arthritis Res Ther 18, 39; Alivernini et al., 2017, Ann Rheum Dis, doi: 10.1136/annrheumdis-2016-210424)). However, clusters of synovial tissue macrophages persist, predominantly in the lining layer (Alivernini etal., 2017, Ann Rheum Dis, doi: 10.1136/annrheumdis-2016-210424). The phenotype and function of these cells are unknown.

Currently, the lack of reliably sensitive tests for remission in RA results in the unnecessary continuation of ineffective medication. As a consequence, many patients are exposed, unnecessarily, to drugs with an array of unpleasant side effects without any therapeutic benefit. Because there is no cheap, quick and reliable measure of sustained remission, even the administration of effective treatment in some patients often continues long after a sustained remission state has been achieved, resulting in unnecessary, toxic and costly treatment. Furthermore, many patients who achieve a state of remission are prone to relapse (or flare). The inability to reliably distinguish between, and predict, those patients who will experience flare when treatment is discontinued and those that will remain in remission long-term, results in premature treatment withdrawal in some patients and unnecessary continuation of treatment in others. Consequently, the treatment of rheumatoid arthritis is currently costly, imprecise and poorly tailored to the needs of the individual patient.

Summary of the Invention

The inventors explored the phenotypic and functional changes in synovial tissue macrophage (STM) subpopulations spanning health, inflammation and disease remission and surprisingly identified cellular and molecular mechanisms that actively maintain remission in rheumatoid arthritis (RA) mediated via tissue resident synovial tissue macrophages.

This discovery has formed the basis for the development of a sensitive, reliable and non-invasive test of the pathological status of patients having rheumatoid arthritis, which permits the rapid and accurate diagnosis of sustained remission, and reliable prediction of the likelihood of a patient in remission experiencing flare if treatment is removed. Importantly this allows treatment to be administered, evaluated and withdrawn in the appropriate clinical circumstances avoiding disease relapse, flare and unnecessary treatment with toxic and costly drugs. Further applications include the ability to select tailored treatment regimens and to rapidly and reliably evaluate candidate therapeutic agents for the treatment of rheumatoid arthritis.

Accordingly, the present invention provides a method for determining remission in a subject having rheumatoid arthritis, as well as methods of determining the likelihood of a subject having rheumatoid arthritis relapsing, experiencing flare, or remaining in sustained remission upon discontinuation of treatment, and methods for evaluating the response to a treatment regimen. The invention also provides devices and kits for use in the determination of remission in a subject having rheumatoid arthritis.

Accordingly, the present invention provides a method for determining remission in a subject having rheumatoid arthritis, the method comprising the steps of: a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; and c) comparing the level of each of the biomarkers determined in (b) with one or more reference values, wherein a difference in the level of MerTK and CD206 in the biological sample compared to the one or more reference values is indicative of disease remission.

In embodiments of the invention it will be appreciated that given the surprising power of the combination of two biomarkers of the present invention to differentiate between remission and active rheumatoid arthritis the method may at a minimum involve determining the level of MerTK and CD206 in the biological sample and comparing the level of each of MerTK and CD206 with one or more reference values.

Determining the levels of just MerTK and CD206 in the biological sample provides a simple and quick but surprisingly powerful indicator of sustained remission.

Typically, for example where a greater level of accuracy is required, the level of MerTK and CD206 may be investigated in combination with that of TREM2 and/or CD163 in a biological sample.

Accordingly, the method may further comprise determining the level of TREM2 in the biological sample; wherein a difference in the level of TREM2 in the biological sample compared to the one or more reference values is indicative of disease remission.

Additionally, or alternatively, the method may further comprise determining the level of CD163 in the biological sample; wherein a difference in the level of CD163 in the biological sample compared to the one or more reference values is indicative of disease remission.

The following combinations represent preferred combinations of biomarkers: MerTK, CD206 and CD163; MerTK, CD206 and TREM2; MerTK, CD206, CD163 and TREM2.

In a particularly preferred embodiment the invention provides a method for determining remission in a subject having rheumatoid arthritis, the method comprising the steps of: a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK, CD206, CD163 and TREM2 in the biological sample; and c) comparing the level of each of the biomarkers determined in (b) with one or more reference values, wherein a difference in the level of MerTK, CD206, CD163 and TREM2 in the biological sample compared to the one or more reference values is indicative of disease remission.

The combination of MerTK, CD206, CD163 and TREM2 performs surprisingly well, and can be used to provide very accurate indication of remission in patients with rheumatoid arthritis, from a biological sample comprising a synovial cell, or an extract or sub-cellular fraction thereof.

Accordingly, in embodiments of the invention the level of each of MerTK, CD206, TREM2 and CD163 is determined.

It will be appreciated that the abovementioned combinations of biomarkers represent various minimal marker sets, and additional biomarkers can also be included.

In embodiments, the invention involves assessing the levels of METK and CD206, and optionally CD163 and/or TREM2 in biological samples described herein. In particular, changes in levels of METK and CD206, and optionally CD163 and/or TREM2 may be assessed, and in preferred embodiments this change is differentially upwards for each of those markers in subjects which are in remission, compared for example to the levels of those biomarkers in control samples which are characteristic of active disease. For the avoidance of doubt, it should be noted that the present invention can be used for both initial diagnosis of remission and for ongoing monitoring of patient disease status e.g. response to treatment.

In embodiments remission is characterised by an elevated level of MerTK and CD206 compared to the one or more reference values, preferably compared to the levels of those biomarkers in one or more control samples which are characteristic of active disease. In certain embodiments remission is characterised by an elevated level of MerTK, CD206 and TREM2 compared to one or more reference values, preferably compared to the levels of MerTK, CD206 and TREM2 in one or more control samples which are characteristic of active disease. In certain embodiments remission is characterised by an elevated level of MerTK, CD206 and CD163 compared to one or more reference values, preferably compared to the levels of MerTK, CD206 and CD163 in one or more control samples which are characteristic of active disease. In preferred embodiments, remission is characterised by an elevated level of MerTK, CD206,

TREM2 and CD163 compared to the one or more reference values, preferably compared to the levels of MerTK, CD206, TREM2 and CD163 in one or more control samples which are characteristic of active disease.

Differential Expression Throughout, biomarkers in the biological sample(s) from the subject are said to be differentially expressed, where they are significantly up- or down- regulated compared to one or more reference values. Biomarkers in the biological sample(s) from the subject are said to be differentially expressed and indicative of remission, where they are significantly up- regulated. Biomarkers of the invention are said to be significantly up-, or down- regulated (i.e. increased or decreased), when after scaling of biomarker expression levels in relation to sample mean and sample variance, they exhibit an adjusted p value < 0.05 and a log fold change of > ± 1.5, preferably p<0.01 and log2- fold change > ±1.5 between samples. Remission may be identified from a biological sample by an increase in expression level, scaled in relation to sample mean and sample variance, relative to one or more reference values, for example relative to their expression level in an control sample characteristic of active rheumatoid arthritis. Depending on the circumstances and desired comparison, a suitable active disease control may be a treatment naive active disease control or a treatment resistant active disease control or both. Clearly, variation in the sensitivity of individual biomarkers, subject and samples mean that different levels of confidence are attached to each biomarker.

In the case of increased expression level (up-regulation relative to reference values), biomarkers exhibit an adjusted p value < 0.05 and a log2 fold change of 1.5 or more compared with the reference value. Preferably, biomarkers will exhibit an adjusted p value < 0.05 or < 0.01 and a log2 fold change of 1.6 or more, 1.7 or more, 1.8 or more,

1.9 or more, 2 or more, 2.1 or more, 2.2 or more, 2.3 or more, 2.4 or more, 2.5 or more, 2.6 or more, 2.7 or more, 2.8 or more, 2.9 or more, 3 or more, 3.1 or more, 3.2 or more, 3.3 or more, 3.4 or more, 3.5 or more, 3.6 or more, 3.7 or more, 3.8 or more, 3.9 or more, 4 or more, 4.5 or more, or 5 or more relative to one or more reference values, for example relative to their expression level in a negative i.e. active disease control.

Conversely, in the case of decreased expression level (down-regulation relative to reference values), biomarkers will preferably exhibit a log2 fold change of -1.6 or less compared with the reference value. Optionally, biomarkers will exhibit an adjusted p value < 0.05 or < 0.01 and a log2 fold change of -1.7 or less, -1.8 or less, -1.9 or less, -2 or less, -2.1 or less, -2.2 or less, -2.3 or less, -2.4 or less, -2.5 or less, -2.6 or less, -2.7 or less, -2.8 or less, -2.9 or less, -3 or less, -3.1 or less, -3.2 or less, -3.3 or less, -3.4 or less, -3.5 or less, -3.6 or less, -3.7 or less, -3.8 or less, -3.9 or less, or -4 or less, -4.5 or less, or 5 or less relative to one or more reference values, for example relative to their expression level in a positive i.e. remission control.

In addition to the degree or magnitude of biomarker expression, i.e. MerTK and CD206, and optionally TREM2 and/or CD163, as between active rheumatoid arthritis and rheumatoid arthritis in remission, the inventors have discovered that the relative proportion of synovial tissue macrophages expressing MerTK and CD206 (MerTK/CD206^pos) is also a sensitive indicator of pathological status in patients having rheumatoid arthritis. Specifically, they have identified that patients with rheumatoid arthritis in sustained remission have a greater number of MerTK/CD206^pos synovial tissue macrophages compared to those with active rheumatoid arthritis. Furthermore, patients with rheumatoid arthritis in sustained remission have a greater number of MerTK/CD206/CD163^pos synovial tissue macrophages compared to those with active rheumatoid arthritis. Furthermore, patients with rheumatoid arthritis in sustained remission have a greater number of MerTK/CD206/CD163/TREM2^pos synovial tissue macrophages compared to those with active rheumatoid arthritis.

Patients with active rheumatoid arthritis on the other hand, were found to have a greater number of synovial tissue macrophages which do not express MerTK or CD206 (MerTK/CD206^neg) compared to those with rheumatoid arthritis in sustained remission.

In particular, a greater number of MerTK/CD206/CD163^neg synovial tissue macrophages, and/or MerTK/CD206/CD163/TREM2^neg synovial tissue macrophages is characteristic of patients with active rheumatoid arthritis.

In embodiments, the numbers of synovial tissue macrophages expressing METK and CD206, and optionally CD163 and/or TREM2 in a biological sample described herein may be determined and compared to the numbers of synovial tissue macrophages in the same biological sample not expressing METK and CD206, and optionally CD163 and/or TREM2 wherein a greater number of synovial tissue macrophages expressing METK and CD206, and optionally CD163 and/or TREM2 is indicative of sustained remission.

Predicting Flare

As explained above, many patients with rheumatoid arthritis who achieve a state of remission are prone to relapse (or flare). Identifying those patients in remission which are prone to flare when treatment is withdrawn is a key aspect of disease management, a reliable test for which until now has been elusive.

Whilst an elevated level of MerTK and CD206 and optionally TREM2 and/or CD163 in biological samples described herein obtained from patients with rheumatoid arthritis is characteristic of sustained remission in those patients, the inventors have surprisingly discovered that the synovial tissue macrophage population profile of patients having rheumatoid arthritis independently provides a prognostic biomarker which is predictive of disease flare after treatment is withdrawn or changed. In particular, the inventors identified that patients who maintain remission when anti-arthritic treatment is discontinued have a distinct synovial tissue macrophage composition compared to those who subsequently flare.

Specifically, a low proportion of MerTK^pos synovial tissue macrophages in rheumatoid arthritis patients in remission is predictive of flare after treatment cessation, whereas a higher proportion of MerTK/CD206^pos synovial tissue macrophages and correspondingly lower MerTK/CD206^neg is predictive of sustained drug-free remission.

In certain embodiments, the invention provides a method for determining the likelihood of a subject having rheumatoid arthritis relapsing, experiencing flare, or remaining in sustained remission upon discontinuation of treatment; the method comprising the steps of: a) providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages; b) comparing the number of synovial tissue macrophages in the biological sample which express each of the biomarkers MerTK and CD206 to the number of synovial tissue macrophages in the biological sample which do not express MerTK and CD206; wherein a greater number of synovial tissue macrophages which do not express each of the biomarkers MerTK and CD206 is indicative of an increased likelihood of a subject relapsing or experiencing flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express each of the biomarkers MerTK and CD206 is indicative of sustained remission. Optionally, the method may further comprise comparing the number of synovial tissue macrophages in the biological sample which express TREM2 to the number of synovial tissue macrophages in the biological sample which do not express TREM2.

Alternatively or additionally, the method may further comprise comparing the number of synovial tissue macrophages in the biological sample which express CD163 to the number of synovial tissue macrophages in the biological sample which do not express CD163.

Suitably, a proportion of MerTK/CD206^pos or MerTK/CD206/TREM2^pos or MerTK/CD206/CD163^pos or MerTK/CD206/TREM2/CD163^pos synovial tissue macrophages in the biological sample of less than or equal to 48% is predictive of flare after treatment cessation. Optionally, a proportion of MerTK/CD206^pos or MerTK/CD206/TREM2^pos or MerTK/CD206/CD163^pos or

MerTK/CD206/TREM2/CD163^pos synovial tissue macrophages in the biological sample of less than or equal to 47.5%, less than or equal to 47%, less than or equal to 46%, less than or equal to 45%, less than or equal to 44%, less than or equal to 43%, less than or equal to 42%, less than or equal to 41%, less than or equal to 40%, less than or equal to 35%, less than or equal to 30%, less than or equal to 25%, less than or equal to 20%, less than or equal to 15%, less than or equal to 10%, less than or equal to 5%, less than or equal to 2.5%, less than or equal to 2%, less than or equal to 1%, less than or equal to 0.5%, less than or equal to 0.25% is predictive of flare after treatment cessation.

Furthermore, the inventors have identified the ratio of MerTK/CD206^pos to MerTK/CD206^neg synovial tissue macrophages as an independent factor predicting disease flare after treatment discontinuation in rheumatoid arthritis patients.

In certain embodiments the invention provides a method for determining the likelihood of a subject with rheumatoid arthritis and receiving treatment experiencing flare, or remaining in sustained remission upon discontinuation of treatment; the method comprising the steps of providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages and determining a ratio of either: synovial tissue macrophages in the biological sample which express MerTK and CD206 to synovial tissue macrophages in the biological sample which do not express MerTK or CD206; or synovial tissue macrophages in the biological sample which express MerTK, CD206 and TREM2 to synovial tissue macrophages which do not express any of MerTK, CD206 or TREM2; or synovial tissue macrophages in the biological sample which express MerTK, CD206 and CD 163 to synovial tissue macrophages which do not express any of MerTK, CD206 or CD163; or synovial tissue macrophages in the biological sample which express MerTK, CD206 and TREM2 and CD163 to synovial tissue macrophages which do not express any of MerTK, CD206, TREM2 or CD163; wherein a ratio of the biomarker expressing synovial tissue macrophages to the non- biomarker expressing synovial tissue macrophages of less than or equal to 2.5 is predictive of flare occurring upon discontinuation of treatment.

A ratio of synovial tissue macrophages in the biological sample expressing the relevant biomarker combinations to those in the biological sample not expressing any of the relevant biomarkers of less than or equal to 2.5 is predictive of flare occurring after treatment is discontinued. Optionally, depending on the particular biomarker combinations, the the Optionally, the ratio may be predictive of flare after treatment cessation if it is less than or equal to 2.4, less than or equal to 2.3, less than or equal to

2.2, less than or equal to 2.1 , less than or equal to 2.0, less than or equal to 1.9, less than or equal to 1.8, less than or equal to 1.7, less than or equal to 1.6, less than or equal to 1.5, less than or equal to 1.4, less than or equal to 1.3, less than or equal to

1.2, less than or equal to 1.1 , less than or equal to 1.0, less than or equal to 0.9, less than or equal to 0.8, less than or equal to 0.7, less than or equal to 0.6, less than or equal to 0.5.

Given the sensitivity and reliability of the biomarkers in determining remission, it will be appreciated that the levels of MerTK, CD206, and optionally TREM2 and/or CD163 in a biological sample and the relative proportions of synovial tissue macrophages in a biological sample expressing those biomarkers may conveniently be used as rapid, sensitive and reliable proxies for making clinical decisions, whether alone or in combination with other measures. For example, evaluating the effectiveness of a particular treatment regimen, determining whether a patient receiving treatment for rheumatoid arthritis is suitable for having treatment withdrawn or in screening for candidate therapeutic agents.

Accordingly, in certain embodiments the invention provides a method for evaluating the therapeutic efficacy of a candidate therapeutic agent for rheumatoid arthritis, the method comprising; comparing the level of the biomarkers MerTK and CD206 in biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the candidate therapeutic agent; wherein an increase in the level of MerTK and CD206 in the biological sample taken after the administration of the candidate therapeutic agent relative to the level of MerTK and CD206 in the biological sample taken before the administration of the candidate therapeutic agent is indicative of effective treatment.

Optionally, the method may further comprise comparing the level of TREM2 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the candidate therapeutic agent; wherein an increase in the level of TREM2 in the biological sample taken after the administration of the candidate therapeutic agent is increased relative to the level of TREM2 in the biological sample taken before the administration of the candidate therapeutic agent is indicative of effective treatment.

Optionally, the method may further comprise comparing the level of CD163 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the candidate therapeutic agent; wherein an increase in the level of CD163 in the biological sample taken after the administration of the candidate therapeutic agent is increased relative to the level of CD163 in the biological sample taken before the administration of the candidate therapeutic agent is indicative of effective treatment.

Preferably, the method comprises comparing the level of MerTK, CD206, TREM2 and CD163 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the candidate therapeutic agent; wherein an increase in the level of MerTK, CD206, TREM2 and CD163 in the biological sample taken after the administration of the candidate therapeutic agent is increased relative to the level of MerTK, CD206, TREM2 and CD163 in the biological sample taken before the administration of the candidate therapeutic agent is indicative of effective treatment

In certain embodiments, the invention also provides a method for evaluating response to treatment in a subject having rheumatoid arthritis, the method comprising; a) determining the level of each of the biomarkers MerTK and CD206 in a first biological sample comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from the subject at a first time point prior to administration of a treatment for rheumatoid arthritis; b) administering to the subject a treatment for rheumatoid arthritis; c) determining the level of each of the biomarkers MerTK and CD206 in a second biological sample comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from the subject at a subsequent time point following administration of the treatment for rheumatoid arthritis; and d) comparing the level of the biomarkers determined in (a) with the level of the corresponding biomarkers determined in (c), wherein an increase in the levels of MerTK and CD206 in (c) relative to (a) identifies the subject as having a positive response to treatment and a decrease or no change in the in the levels of MerTK and CD206 in (c) relative to (a) identifies the subject as having no response to treatment or a negative response to treatment.

Optionally, the method may further comprise determining the level of TREM2 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the treatment for rheumatoid arthritis; and comparing the level of TREM2 determined in (a) with the level of TREM2 determined in (c), wherein an increase in the levels of TREM2 in (c) relative to (a) identifies the subject as having a positive response to treatment and a decrease or no change in the in the levels of TREM2 in (c) relative to (a) identifies the subject as having no response to treatment or a negative response to treatment. Optionally, the method may further comprise determining the level of CD163 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the treatment for rheumatoid arthritis; and comparing the level of CD163 determined in (a) with the level of CD163 determined in (c), wherein an increase in the levels of CD163 in (c) relative to (a) identifies the subject as having a positive response to treatment and a decrease or no change in the in the levels of CD163 in (c) relative to (a) identifies the subject as having no response to treatment or a negative response to treatment.

Preferably, the method comprises determining the level of MerTK, CD206, TREM2 and CD163 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the treatment for rheumatoid arthritis; and comparing the level of MerTK, CD206, TREM2 and CD163 determined in (a) with the level of MerTK, CD206, TREM2 and CD163 determined in (c), wherein an increase in the levels of MerTK, CD206, TREM2 and CD163 in (c) relative to (a) identifies the subject as having a positive response to treatment and a decrease or no change in the in the levels of CD163 in (c) relative to (a) identifies the subject as having no response to treatment or a negative response to treatment.

In certain embodiments, the invention also provides a method for determining in a subject receiving treatment for rheumatoid arthritis, whether the subject is suitable for treatment withdrawal, the method comprising; a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; and c) comparing the level of each of the biomarkers determined in (b) with the level of each of the biomarkers MerTK and CD206 in a control sample, wherein an increase in the level of MerTK and CD206 in the biological sample compared to the control sample is indicative of suitability for treatment withdrawal.

Optionally, the method may further comprise determining the level of TREM2 in the biological sample and comparing the level of TREM2 in the biological sample with the level of TREM2 in a control sample, wherein an increase in the level of TREM2 in the biological sample compared to the control sample is indicative of suitability for treatment withdrawal.

Optionally, the method may further comprise determining the level of CD163 in the biological sample and comparing the level of CD163 in the biological sample with the level of CD163 in a control sample, wherein an increase in the level of CD163 in the biological sample compared to the control sample is indicative of suitability for treatment withdrawal.

In certain embodiments, the invention provides a method for determining in a subject receiving treatment for rheumatoid arthritis, whether the subject is suitable for treatment withdrawal, the method comprising; a) providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages; b) comparing the number of synovial tissue macrophages in the biological sample which express each of the biomarkers MerTK and CD206 to the number of synovial tissue macrophages in the biological sample which do not express MerTK and CD206; wherein a greater number of synovial tissue macrophages which express each of the biomarkers MerTK and CD206 indicates that the subject is suitable for treatment withdrawal.

Optionally, the method may further comprise comparing the number of synovial tissue macrophages in the biological sample which express TREM2 to the number of synovial tissue macrophages in the biological sample which do not express TREM2; wherein a greater number of synovial tissue macrophages which express TREM2 indicates that the subject is suitable for treatment withdrawal.

Optionally, the method may further comprise comparing the number of synovial tissue macrophages in the biological sample which express CD163 to the number of synovial tissue macrophages in the biological sample which do not express CD163; wherein a greater number of synovial tissue macrophages which express CD163 indicates that the subject is suitable for treatment withdrawal. In certain embodiments the invention provides a method for determining the likelihood of a subject receiving treatment for rheumatoid arthritis experiencing flare, or remaining in sustained remission upon discontinuation of treatment; the method comprising the steps of: a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; and c) comparing the level of each of the biomarkers determined in (b) with the level of each of the biomarkers MerTK and CD206 in a control sample; d) wherein an increase in the level of MerTK and CD206 in the biological sample compared to the control sample is indicative of long-term disease remission; and e) wherein a decrease in, or no change in, the level of MerTK and CD206 in the biological sample compared to the control sample is predictive of flare upon discontinuation of treatment.

In certain embodiments the invention provides a method for determining the likelihood of a subject receiving treatment for rheumatoid arthritis experiencing flare, or remaining in sustained remission upon discontinuation of treatment; the method comprising the steps of: a) providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages; b) comparing the number of synovial tissue macrophages in the biological sample which express each of the biomarkers MerTK and CD206 to the number of synovial tissue macrophages in the biological sample which do not express MerTK and CD206; wherein a greater number of synovial tissue macrophages which do not express each of the biomarkers MerTK and CD206 is predictive of flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express each of the biomarkers MerTK and CD206 is indicative of sustained remission. Optionally, the method may further comprise comparing the number of synovial tissue macrophages in the biological sample which express TREM2 to the number of synovial tissue macrophages in the biological sample which do not express TREM2; wherein a greater number of synovial tissue macrophages which do not express TREM2 is predictive of flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express TREM2 is indicative of sustained remission.

Optionally, the method may further comprise comparing the number of synovial tissue macrophages in the biological sample which express CD163 to the number of synovial tissue macrophages in the biological sample which do not express CD163; wherein a greater number of synovial tissue macrophages which do not express CD163 is predictive of flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express CD 163 is indicative of sustained remission.

Optionally, the method may comprise comparing the number of synovial tissue macrophages in the biological sample which express MerTK, CD206 and CD163 to the number of synovial tissue macrophages in the biological sample which do not express MerTK, CD206 and CD163; wherein a greater number of synovial tissue macrophages which do not express MerTK, CD206 and CD163 is predictive of flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express MerTK, CD206 and CD163 is indicative of sustained remission.

Optionally, the method may comprise comparing the number of synovial tissue macrophages in the biological sample which express MerTK, CD206 and TREM2 to the number of synovial tissue macrophages in the biological sample which do not express MerTK, CD206 and TREM2; wherein a greater number of synovial tissue macrophages which do not express MerTK, CD206 and TREM2 is predictive of flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express MerTK, CD206 and TREM2 is indicative of sustained remission.

Optionally, the method may comprise comparing the number of synovial tissue macrophages in the biological sample which express MerTK, CD206, TREM2 and CD163 to the number of synovial tissue macrophages in the biological sample which do not express MerTK, CD206, TREM2 and CD163; wherein a greater number of synovial tissue macrophages which do not express MerTK, CD206, TREM2 and CD163 is predictive of flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express MerTK, CD206, TREM2 and CD163 is indicative of sustained remission.

In certain embodiments the invention provides a method for determining the likelihood of a subject receiving treatment for rheumatoid arthritis experiencing flare, or remaining in sustained remission upon discontinuation of treatment; the method comprising the steps of: a) providing a biological sample obtained from the subject, the sample comprising a plurality of synovial tissue macrophages; and b) determining a ratio of synovial tissue macrophages which express MerTK and CD206 to those which do not express MerTK or CD206; wherein a ratio of MerTK and CD206 expressing synovial tissue macrophages to MerTK and CD206 negative synovial tissue macrophages of less than or equal to 2.5 is predictive of flare upon discontinuation of treatment.

Optionally, the method comprises determining a ratio of synovial tissue macrophages which express MerTK, CD206 and TREM2 to those which do not express MerTK,

CD206 or TREM2; wherein a ratio of MerTK, CD206 and TREM2 expressing synovial tissue macrophages to MerTK, CD206 and TREM2 negative synovial tissue macrophages of less than or equal to 2.5 is predictive of flare upon discontinuation of treatment.

Optionally, the method comprises determining a ratio of synovial tissue macrophages which express MerTK, CD206 and CD163 to those which do not express MerTK,

CD206 or CD163; wherein a ratio of MerTK, CD206 and CD163 expressing synovial tissue macrophages to MerTK, CD206 and CD163 negative synovial tissue macrophages of less than or equal to 2.5 is predictive of flare upon discontinuation of treatment.

Optionally, the method comprises determining a ratio of synovial tissue macrophages which express MerTK, CD206, TREM2 and CD163 to those which do not express MerTK, CD206, TREM2 or CD163; wherein a ratio of MerTK, CD206, TREM2 and CD163 expressing synovial tissue macrophages to MerTK, CD206, TREM2 and CD163 negative synovial tissue macrophages of less than or equal to 2.5 is predictive of flare upon discontinuation of treatment.

In certain embodiments the invention provides a method of treating a patient having rheumatoid arthritis, comprising the steps of; a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; c) comparing the level of each of the biomarkers determined in (b) with the level of each of the biomarkers MerTK and CD206 in a control sample; and d) administering a therapeutic agent where the level of MerTK and CD206 in the biological sample is elevated compared to the control sample; or e) providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages; f) comparing the number of synovial tissue macrophages in the biological sample of (e) which express each of the biomarkers MerTK and CD206 to the number of synovial tissue macrophages in the biological sample of (e) which do not express MerTK and CD206; and g) administering a therapeutic agent where a greater number of synovial tissue macrophages in the biological sample of (e) do not express MerTK and CD206.

In certain embodiments the invention provides a method of treating a patient having rheumatoid arthritis, wherein the patient is already receiving treatment for rheumatoid arthritis, comprising the steps of; a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; c) comparing the level of each of the biomarkers determined in (b) with the level of each of the biomarkers MerTK and CD206 in a control sample; and d) administering a different therapeutic agent where the level of MerTK and CD206 in the biological sample is elevated compared to the control sample; and e) optionally withdrawing the original treatment regime; or f) providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages; g) comparing the number of synovial tissue macrophages in the biological sample of (f) which express each of the biomarkers MerTK and CD206 to the number of synovial tissue macrophages in the biological sample of (f) which do not express MerTK and CD206; and h) administering a different therapeutic agent where either; i) a greater number of synovial tissue macrophages in the biological sample of (f) do not express MerTK and CD206; or ii) a ratio of MerTK and CD206 expressing synovial tissue macrophages to MerTK and CD206 negative synovial tissue macrophages in the biological sample of (f) is less than or equal to 2.5; and i) optionally withdrawing the original treatment regime.

In embodiments where the patient is receiving treatment with a therapeutic agent or where the method comprises the administration of a therapeutic agent, then suitable therapeutic agents may be selected from: non-biologic DMARDs such as Methotrexate, Sulfasalazine Hydroxychloroquine Leflunomide, Azathioprine, Penicillamine, Gold Injections, Ciclosporin; biological DMARDs including tumor necrosis factor (TNF) inhibitors such as etanercept, adalimumab, infliximab, certolizumab pegol, and golimumab; kinase inhibitors, including tofacitinib and baricitinib; or biological DMARDs with different targets, including anakinra, abatacept, rituximab, and tocilizumab; or any combination thereof.

Preferably the therapeutic agent comprises an anti-TNFa agent.

Reference Values

Throughout the term "reference value" may refer to a pre-determ ined reference value, for instance specifying a confidence interval or threshold value for the diagnosis of remission in a subject. Alternatively, the reference value may be derived from the expression level of a corresponding biomarker or biomarkers in a 'control' biological sample, for example a positive (remission), negative (active disease) or other (healthy) control. It will be understood that active disease controls may be treatment naive or treatment resistant. The control biological sample may suitably be a corresponding biological sample derived from the same subject at a different time point. The control biological sample may be a corresponding biological sample type derived from a different subject, for example, a subject with active disease, a healthy subject or a subject in remission. Additionally or alternatively the expression levels of the biomarkers in a biological sample may be compared to a reference value which may be, for example, a positive (remission), negative (e.g. active disease) or other (e.g. healthy) control, which may represent an average value for a particular population of subjects. Such a population of subjects, might for example, share particular characteristics with the subject of interest, such as age, sex, ethnicity, certain genetic characteristics and/or previous disease or treatment history. Furthermore, the reference value may be an 'internal' standard or range of internal standards, for example a known concentration of a protein, transcript, label or compound. Alternatively, the reference value may be an internal technical control for the calibration of expression values or to validate the quality of the sample or measurement techniques. This may involve a measurement of one or several transcripts within the sample which are known to be constitutively expressed or expressed at a known level (e.g. an invariant level). Accordingly, it would be routine for the skilled person to apply these known techniques alone or in combination in order to quantify the level of biomarker in a sample relative to standards or other transcripts or proteins or in order to validate the quality of the biological sample, the assay or statistical analysis. Subjects

Suitably the subject is a mammal. Preferably, the subject is a human. More preferably, the subject is an adult human. Preferably the subject is a human having been diagnosed as having rheumatoid arthritis. Typically, the subject is a human having been diagnosed as having rheumatoid arthritis for at least 6 months, at least 9 months, at least 12 months, at least 15 months, at least 18 months, at least 2 years, at least 5 years, at least 10 years, at least 15 years, at least 20 years.

Preferably, the subject is receiving treatment for rheumatoid arthritis, wherein the treatment is selected from methotrexate, sulfasalazine, hydroxychloroquine, leflunomide, azathioprine, penicillamine, gold Injections, ciclosporin, etanercept, adalimumab, infliximab, certolizumab pegol, golimumab, tofacitinib, baricitinib, anakinra, abatacept, rituximab, and tocilizumab, or any combination thereof. In preferred embodiments, the subject having rheumatoid arthritis is receiving treatment comprising a TNF-inhibitor and/or Methotrexate, preferably wherein the TNF-inhibitor is adalimumab or etanercept. In preferred embodiments, the subject having rheumatoid arthritis is receiving treatment consisting of TNF-inhibitor and/or Methotrexate, preferably wherein the TNF-inhibitor is adalimumab or etanercept.

Biological Samples

In preferred embodiments the methods of the invention are carried out in vitro, but it will be appreciated that the methods and assays of the invention are also capable of being carried out in vivo. The term “in vitro " is intended to encompass procedures performed with cells or extracts therefrom in culture whereas the term “in vivo " is intended to encompass procedures with/on intact multi-cellular organisms.

Each of the methods of the invention may involve obtaining a sample of biological material from the subject, or may be performed on a pre-obtained sample, e.g. one which has been obtained previously for other clinical purposes. Each of the methods and assays of the invention may include the step of processing the sample before analysis of biomarker levels is carried out. This may include for example, filtering and/or enriching the sample, for example in synovial tissue macrophages, or processing of the sample using fluorescence activated cell sorting (FACS) to obtain, for example particular subpopulations of synovial tissue macrophages e.g. those expressing CD64, CD1 1 b, MHC11 , and HLA-DR or for example to obtain DNA, cDNA, mRNA and/or protein.

In embodiments of the invention suitable biological samples will be those comprising a synovial cell, or an extract or sub-cellular fraction thereof. In embodiments of the invention several different types of biological sample could be used, e.g. a synovial tissue or fluid sample comprising at least one synovial cell. Alternatively, the biological sample may comprise an extract from a synovial cell, or sub-cellular fraction thereof. Typically, the biological sample comprises synovial myeloid cells. Preferably the biological sample comprises synovial tissue macrophages, or an extract or sub-cellular fraction thereof. Depending on the method used for measuring the levels of biomarker expression and/or the proportions of synovial tissue macrophages expressing those biomarkers, the biological sample may require different amounts of material. Often the biological sample will contain a plurality of synovial tissue macrophages, for example at least 1000, at least 2000, at least 3000, at least 4000, at least 5000, at least 6000, at least 7000, at least 8000, at least 9000, at least 10000, at least 15000, at least 20000 or at least 25000. Synovial tissue macrophages may be identified and/or distinguished from other cell types in any suitable manner, for example by expression of CD64,

CD1 1 b, MHCII and the absence of other cell-lineage markers.

Further Physiological Measurements The methods, devices and kits of the present invention may additionally make use of a range of biological samples and/or measurements taken from a subject to further determine or confirm the precise pathological status of the patient. In certain embodiments the methods of the invention may further involve investigating blood C- reactive protein levels. In certain embodiments the methods of the invention may further involve conducting physiological measurements selected from; tender joint count (TJC), swollen joint count (SJC) and/or blood C-reactive protein (CRP) levels. In certain embodiments the methods of the invention may further involve conducting a patient global assessment (PGA) and/or evaluator global assessment (EGA). Suitably the patient global assessment, may involve obtaining an answer on a 0-10 scale from a patient to the following question: considering all of the ways your arthritis has affected you, how do you feel your arthritis is today? (0-10, where 0 = very well and 10 = very poorly). In certain embodiments the methods of the invention may further involve establishing a simplified disease activity index (SDAI), where for example, the SDAI is the arithmetic sum of SJC +TJC+PGA+EGA+CRP, wherein the 28 joint count is used for joint assessment, the global evaluations are employed in cm (rather than mm), and CRP as mg/dl.

Remission

Remission is used herein to describe a diminution of the severity of rheumatoid arthritis compared with an active disease state. Commonly, this includes the attenuation of synovial hypertrophy, normalised blood flow and/or normalised histology compared with an active disease state. Remission in rheumatoid arthritis may be characterised by satisfaction of various criteria, for example sustained clinical remission (DAS28<2.6 for 3 sequential determinations each 3 months apart) or sustained ultrasound remission (Power Doppler negativity at US assessment for 3 sequential determinations each 3 months apart). In the determination of remission, clinical and laboratory evaluations may include the number of tender and swollen joints of 28 examined, Erythrocyte Sedimentation Rate (ESR), C-Reactive Protein (CRP) and Disease Activity Score (DAS28). Commonly, remission may be characterised by the following criteria: tender joint count <1 , swollen joint count <1 , C-reactive protein <1 mg/dl and patient global assessment (PGA) <1 ; and/or simplified disease activity index (SDAI) <3.3 in accordance with the Boolean criteria (see Bykerk etal., 2012, Rheumatology (Oxford)

51 Suppl 6, vi16-20, doi:10.1093/rheumatology/kes281). Testing is typically performed by sequential determinations each 3 months apart.

Sustained remission is used herein to describe the maintenance of a remission state in patients having rheumatoid arthritis (in accordance with the above) for a period of at least 9 months. Preferably, sustained remission in patients having rheumatoid arthritis refers to a state where each of the DAS or Boolean criteria are satisfied and wherein the state is maintained for a period of at least 9 months, at least 12 months, at least 15 months, at least 18 months, at least 21 months, at least 24 months, at least 36 months, or at least 48 months. More preferably the remission state in patients having rheumatoid arthritis (in accordance with the above) is maintained for a period of at least 1 year, at least 2 years, at least 3 years, at least 4 years, at least 5 years, at least 10 years, at least 15 years, at least 20 years, or at least 25 years. Remission may be attained with treatment or be drug-free. Preferably, the remission state is maintained for the above periods in the absence of treatment with disease-modifying anti-rheumatic drugs (DMARDs). Biomarkers

Suitably the biomarkers are selected from the group consisting of: the biomarker protein; and a nucleic acid molecule encoding the biomarker protein.

It is preferred that the biomarker is a nucleic acid molecule, and highly preferred that it is an mRNA molecule. There are numerous suitable techniques known in the art for the quantitative measurement of mRNA levels in a given biological sample. The levels of the biomarkers in the biological sample may be investigated for example by RNA sequencing e.g. RNA sequencing of synovial tissue macrophages.

In embodiments where the biomarker is an mRNA molecule the MERTK biomarker will preferably have a polynucleotide sequence of at least 90% sequence identity to SEQ ID NO: 5. Alternatively, in embodiments where the biomarker is an mRNA molecule the MERTK biomarker will preferably have a polynucleotide sequence of at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5% or 100% sequence identity to SEQ ID NO: 5.

In embodiments where the biomarker is an mRNA molecule the CD206 biomarker will preferably have a polynucleotide sequence of at least 90% sequence identity to SEQ ID NO: 6. Alternatively, in embodiments where the biomarker is an mRNA molecule the CD206 biomarker will preferably have a polynucleotide sequence of at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5% or 100% sequence identity to SEQ ID NO: 6.

In embodiments where the biomarker is an mRNA molecule the TREM2 biomarker will preferably have a polynucleotide sequence of at least 90% sequence identity to SEQ ID NO: 7. Alternatively, in embodiments where the biomarker is an mRNA molecule the TREM2 biomarker will preferably have a polynucleotide sequence of at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5% or 100% sequence identity to SEQ ID NO: 7.

In embodiments where the biomarker is an mRNA molecule the CD163 biomarker will preferably have a polynucleotide sequence of at least 90% sequence identity to SEQ ID NO: 8. Alternatively, in embodiments where the biomarker is an mRNA molecule the CD163 biomarker will preferably have a polynucleotide sequence of at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5% or 100% sequence identity to SEQ ID NO: 8.

In embodiments where the biomarker is a protein the MERTK biomarker will preferably have an amino acid sequence of at least 90% sequence identity to SEQ ID NO: 9. Alternatively, in embodiments where the biomarker is a protein the MERTK biomarker will preferably have an amino acid sequence of at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5% or 100% sequence identity to SEQ ID NO: 9.

In embodiments where the biomarker is a protein the CD206 biomarker will preferably have an amino acid sequence of at least 90% sequence identity to SEQ ID NO: 10. Alternatively, in embodiments where the biomarker is a protein the CD206 biomarker will preferably have an amino acid sequence of at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5% or 100% sequence identity to SEQ ID NO:

10.

In embodiments where the biomarker is a protein the TREM2 biomarker will preferably have an amino acid sequence of at least 90% sequence identity to SEQ ID NO: 11. Alternatively, in embodiments where the biomarker is a protein the TREM2 biomarker will preferably have an amino acid sequence of at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5% or 100% sequence identity to SEQ ID NO:

11.

In embodiments where the biomarker is a protein the CD163 biomarker will preferably have an amino acid sequence of at least 90% sequence identity to SEQ ID NO: 12. Alternatively, in embodiments where the biomarker is a protein the CD163 biomarker will preferably have an amino acid sequence of at least 95%, at least 96%, at least 97%, at least 98%, at least 99%, at least 99.5% or 100% sequence identity to SEQ ID NO:

12.

The levels of the biomarkers in the biological sample may be investigated for example by using specific binding partners. Suitably the binding partners are selected from the group consisting of: complementary nucleic acids; aptamers; antibodies or antibody fragments. Suitable classes of binding partners for any given biomarker will be apparent to the skilled person. Suitably the levels of the biomarkers in the biological sample are detected by direct assessment of binding between the target molecules and binding partners. Suitably the levels of the biomarkers in the biological sample are detected using a reporter moiety attached to a binding partner. Preferably the reporter moiety is selected from the group consisting of: fluorophores; chromogenic substrates; and chromogenic enzymes.

Binding Partners

In certain embodiments of the invention, expression levels of the biomarkers in a biological sample may be investigated using binding partners which bind or hybridize specifically to the biomarkers or a fragment thereof. In relation to the present invention the term 'binding partners' may include any ligands, which are capable of binding specifically to the relevant biomarker and/or nucleotide or peptide variants thereof with high affinity. Said ligands include, but are not limited to nucleic acids (DNA or RNA), proteins, peptides, antibodies, synthetic affinity probes, carbohydrates, lipids, artificial molecules or small organic molecules such as drugs. In certain embodiments the binding partners may be selected from the group comprising: complementary nucleic acids; aptamers; antibodies or antibody fragments. In the case of detecting mRNAs, nucleic acids represent highly suitable binding partners.

In the context of the present invention, a binding partner specific to a biomarker should be taken as requiring that the binding partner should be capable of binding to at least one such biomarker in a manner that can be distinguished from non-specific binding to molecules that are not biomarkers. A suitable distinction may, for example, be based on distinguishable differences in the magnitude of such binding.

In preferred embodiments of the methods, devices and/or kits of the invention, the biomarker is a nucleic acid, preferably an mRNA molecule, and the binding partner is selected from the group comprising; complementary nucleic acids or aptamers.

Suitably the binding partner is a nucleic acid molecule (typically DNA, but it can be RNA) having a sequence which is complementary to the sequence the relevant mRNA or cDNA against which it is targeted. Such a nucleic acid is often referred to as a 'probe' (or a reporter or an oligo) and the complementary sequence to which it binds is often referred to as the 'target'. Probe-target hybridization is usually detected and quantified by detection of fluorophore-, silver-, or chemiluminescence-labeled targets to determine relative abundance of nucleic acid sequences in the target.

Probes can be from 25 to 1000 nucleotides in length. However, lengths of 30 to 100 nucleotides are preferred, and probes of around 50 nucleotides in length are commonly used with success in complete transcriptome analysis.

While the determination of suitable probes can be difficult, e.g. in very complex arrays, there are many commercial sources of complete transcriptome arrays available, and it is routine to develop bespoke arrays to detect any given set of specific mRNAs using publically available sequence information. Commercial sources of microarrays for transciptome analysis include Ilium ina and Affymetrix. Nucleotide probe sequences may be designed to any sequence region of the biomarker transcripts (accession numbers listed in Table 1) or a variant thereof. The person skilled in the art will appreciate that equally effective probes can be designed to different regions of the transcript and that the effectiveness of the particular probes chosen will vary, amongst other things, according to the platform used to measure transcript abundance and the hybridization conditions employed. It will therefore be appreciated that probes targeting different regions of the transcript may be used in accordance with the present invention.

In other suitable embodiments of the invention, the biomarker may be a protein, and the binding partner is selected from the group comprising; antibodies, antibody fragments or aptamers.

Polynucleotides encoding any of the specific binding partners of biomarkers of the invention recited above may be isolated and/or purified nucleic acid molecules and may be RNA or DNA molecules.

Throughout, the term "polynucleotide" as used herein refers to a deoxyribonucleotide or ribonucleotide polymer in single- or double-stranded form, or sense or anti-sense, and encompasses analogues of naturally occurring nucleotides that hybridize to nucleic acids in a manner similar to naturally occurring nucleotides. Such polynucleotides may be derived from Homo sapiens, or may be synthetic or may be derived from any other organism. Commonly, polypeptide sequences and polynucleotides used as binding partners in the present invention may be isolated or purified. By "purified" is meant that they are substantially free from other cellular components or material, or culture medium. "Isolated" means that they may also be free of naturally occurring sequences which flank the native sequence, for example in the case of nucleic acid molecule, isolated may mean that it is free of 5' and 3' regulatory sequences. In a preferred embodiment the nucleic acid is mRNA. There are numerous suitable techniques known in the art for the quantitative measurement of mRNA transcript levels in a given biological sample. These techniques include but are not limited to; single cell RNA sequencing (scRNAseq), "Northern" RNA blotting, Real Time Polymerase Chain Reaction (RTPCR), Quantitative Polymerase Chain Reaction (qPCR), digital PCR (dPCR), multiplex PCR, Reverse Transcription Quantitative Polymerase Chain Reaction (RT-qPCR), branched DNA signal amplification or by high- throughput analysis such as hybridization microarray, Next Generation Sequencing (NGS) or by direct mRNA quantification, for example by "Nanopore" sequencing. Alternatively, "tag based" technologies may be used, which include but are not limited to Serial Analysis of Gene Expression (SAGE). The levels of biomarker mRNA transcript in a given biological sample may be determined by hybridization to specific complementary nucleotide probes on a hybridization microarray or "chip", by Bead Array Microarray technology or by RNA-Seq where sequence data is matched to a reference genome or reference sequences. In preferred embodiments, where the nucleic acid is mRNA, the levels of biomarker transcript(s) will be determined by scRNAseq, PCR, qPCR, dPCR or multiplex PCR. Preferably, mRNA transcript abundance will be determined by scRNAseq of synovial tissue macrophages. Nucleotide primer sequences may be designed to any sequence region of the biomarker transcripts (accession numbers listed in Table 1) or a variant thereof. The person skilled in the art will appreciate that equally effective primers can be designed to different regions of the transcript or cDNA of biomarkers listed in Table 1, and that the effectiveness of the particular primers chosen will vary, amongst other things, according to the platform used to measure transcript abundance, the biological sample and the hybridization conditions employed. It will therefore be appreciated that primers targeting different regions of the transcript may also be used in accordance with the present invention. However, the person skilled in the art will recognise that in designing appropriate primer sequences to detect biomarker expression, it is required that the primer sequences be capable of binding selectively and specifically to the cDNA sequences of biomarkers corresponding to the nucleotide accession numbers listed in Table 1 or fragments or variants thereof. Many different techniques known in the art are suitable for detecting binding of the target sequence and for high-throughput screening and analysis of protein interactions. According to the present invention, appropriate techniques may include (either independently or in combination), but are not limited to; co-immunoprecipitation, bimolecular fluorescence complementation (BiFC), dual expression recombinase based (DERB) single vector system, affinity electrophoresis, pull-down assays, label transfer, yeast two-hybrid screens, phage display, in vivo crosslinking, tandem affinity purification (TAP), ChIP assays, chemical cross- linking followed by high mass MALDI mass spectrometry, strep-protein interaction experiment (SPINE), quantitative immunoprecipitation combined with knock-down (QUICK), proximity ligation assay (PLA), bio-layer interferometry, dual polarisation interferometry (DPI), static light scattering (SLS), dynamic light scattering (DLS), surface plasmon resonance (SPR), fluorescence correlation spectroscopy, fluorescence resonance energy transfer (FRET), isothermal titration calorimetry (ITC), microscale thermophoresis (MST), chromatin immunoprecipitation assay, electrophoretic mobility shift assay, pull-down assay, microplate capture and detection assay, reporter assay, RNase protection assay, FISH/ISH co-localization, microarrays, microsphere arrays or silicon nanowire (SiNW)- based detection. Where biomarker protein levels are to be quantified, preferably the interactions between the binding partner and biomarker protein will be analysed using antibodies with a fluorescent reporter attached. In certain embodiments of the invention, the expression level of a particular biomarker may be detected by direct assessment of binding of the biomarker to its binding partner. Suitable examples of such methods in accordance with this embodiment of the invention may utilise techniques such as electro-impedance spectroscopy (EIS) to directly assess binding of binding partners (e.g. antibodies) to target biomarkers (e.g. biomarker proteins).

In certain embodiments of the present invention the binding partner may be an antibody, or antibody fragment, and the detection of the target molecules utilises an immunological method. In certain embodiments of the methods or devices, the immunological method may be an enzyme-linked immunosorbent assay (ELISA) or utilise a lateral flow device.

A method of the invention may further comprise quantification of the amount of the target molecules indicative of expression of the biomarkers that is present in the patient biological sample. Suitable methods of the invention, in which the amount of the target molecule present has been quantified, and the volume of the patient sample is known, may further comprise determination of the concentration of the target molecules present in the patient sample which may be used as the basis of a qualitative assessment of the patient's condition, which may, in turn, be used to suggest a suitable course of treatment for the patient.

Reporter moieties

In embodiments of the present invention the expression levels of the protein in a biological sample may be determined. In some instances, it may be possible to directly determine expression, e.g. as with GFP or by enzymatic action of the protein of interest (POI) to generate a detectable optical signal. However, in some instances it may be chosen to determine physical expression, e.g. by antibody probing, and rely on separate test to verify that physical expression is accompanied by the required function.

In certain embodiments of the invention, the expression levels of a particular biomarker will be detectable in a biological sample by a high-throughput screening method, for example, relying on detection of an optical signal, for instance using reporter moieties. For this purpose, it may be necessary for the specific binding partner to incorporate a tag, or be labelled with a removable tag, which permits detection of expression. Such a tag may be, for example, a fluorescence reporter molecule translationally-fused to the protein of interest (POI), e.g. Green Fluorescent Protein (GFP), Yellow Fluorescent Protein (YFP), Red Fluorescent Protein (RFP), Cyan Fluorescent Protein (CFP) or mCherry. Such a tag may provide a suitable marker for visualisation of biomarker expression since its expression can be simply and directly assayed by fluorescence measurement in vitro or on an array. Alternatively, it may be an enzyme which can be used to generate an optical signal. Tags used for detection of expression may also be antigen peptide tags. Similarly, reporter moieties may be selected from the group consisting of fluorophores; chromogenic substrates; and chromogenic enzymes. Other kinds of label may be used to mark a nucleic acid binding partner including organic dye molecules, radiolabels and spin labels which may be small molecules.

In certain embodiments, the levels of a biomarker or several biomarkers will be quantified by measuring the specific hybridization of a complementary nucleotide probe to the biomarker of interest under high-stringency or very high-stringency conditions. Preferably, probe-biomarker hybridization will be detected and quantified by detection of fluorophore-, silver-, or chemiluminescence-labelled probes to determine relative abundance of biomarker nucleic acid sequences in the sample. Alternatively, levels of biomarker mRNA transcript abundance can conveniently be determined directly by RNA sequencing or nanopore sequencing technologies.

The methods or devices of the invention may make use of molecules selected from the group consisting of: the biomarker protein; and nucleic acid encoding the biomarker protein.

Nucleotides and Hybridization Conditions Throughout, the term "polynucleotide" as used herein refers to a deoxyribonucleotide or ribonucleotide polymer in single- or double-stranded form, or sense or anti-sense, and encompasses analogues of naturally occurring nucleotides that hybridize to nucleic acids in a manner similar to naturally occurring nucleotides.

Nucleotide probe sequences may suitably be designed to any sequence region of the biomarker transcripts (accession numbers listed in Table 1) or a variant thereof. This is also the case with nucleotide primers used where detection of expression levels is determined by PCR-based technology. The person skilled in the art will appreciate that equally effective (and in some cases more beneficial) probes can be designed to different regions of the transcript, and that the effectiveness of the particular probes chosen will vary, amongst other things, according to the platform used to measure transcript abundance and the hybridization conditions employed. It will therefore be appreciated that probes targeting different regions of the transcript may also be used in accordance with the present invention.

It will be recognised that in designing appropriate probe sequences to detect biomarker expression, it is required that the probe sequences be capable of binding selectively and specifically to the transcripts or cDNA sequences of biomarkers corresponding to the nucleotide accession numbers listed in Table 1 or fragments or variants thereof.

The probe sequence will therefore be hybridizable to that nucleotide sequence, preferably under stringent conditions, more preferably very high stringency conditions. The term "stringent conditions" may be understood to describe a set of conditions for hybridization and washing and a variety of stringent hybridization conditions will be familiar to the skilled reader. Hybridization of a nucleic acid molecule occurs when two complementary nucleic acid molecules undergo an amount of hydrogen bonding to each other known as Watson-Crick base pairing. The stringency of hybridization can vary according to the environmental (i.e. chemical/physical/biological) conditions surrounding the nucleic acids, temperature, the nature of the hybridization method, and the composition and length of the nucleic acid molecules used. Calculations regarding hybridization conditions required for attaining particular degrees of stringency are discussed in Sambrook et al. (2001, Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY); and Tijssen (1993, Laboratory Techniques in Biochemistry and Molecular Biology — Hybridization with Nucleic Acid Probes Part I, Chapter 2, Elsevier, NY). The Tm is the temperature at which 50% of a given strand of a nucleic acid molecule is hybridized to its complementary strand. In any of the references herein to hybridization conditions, the following are exemplary and not limiting:

Very High Stringency (allows sequences that share at least 90% identity to hybridize)

• Hybridization: 5x SSC at 65°C for 16 hours

• Wash twice: 2x SSC at room temperature (RT) for 15 minutes each

• Wash twice: 0.5x SSC at 65°C for 20 minutes each

High Stringency (allows sequences that share at least 80% identity to hybridize)

• Hybridization: 5x-6x SSC at 65°C-70°C for 16-20 hours

• Wash twice: 2x SSC at RT for 5-20 minutes each

• Wash twice: 1x SSC at 55°C-70°C for 30 minutes each Low Stringency (allows sequences that share at least 50% identity to hybridize)

• Hybridization: 6x SSC at RT to 55°C for 16-20 hours

• Wash at least twice: 2x-3x SSC at RT to 55°C for 20-30 minutes each.

Devices and Kits

In a further embodiment, the invention provides a device for use in the determination of remission in a subject having been determined to have rheumatoid arthritis, the device comprising: i) a loading area for receipt of a biological sample; ii) binding partners specific for target molecules indicative of the level of biomarkers MerTK and CD206; and iii) detection means to detect the levels of said biomarker present in the sample. The device may optionally comprise binding partners specific for target molecules indicative of the level of biomarkers TREM2 and/or CD163.

The device is adapted to detect and quantify the levels of said biomarkers present in the biological sample.

The binding partners are preferably nucleic acid primers adapted to bind specifically to the cDNA transcripts of biomarkers, as discussed above. The detection means suitably comprises means to detect a signal from a reporter moiety, e.g. a reporter moiety as discussed above.

For example, the device comprises specific binding partners to the biomarkers being detected which enable the cDNA transcripts of the biomarkers to be amplified, e.g. by PCR. A variety of suitable PCR amplification-based technologies are well known in the art.

Optionally the device may be configured to determine the levels of MerTK, CD206, TREM2 and CD163 only.

In a further embodiment, the invention provides a kit of parts for determining whether an individual having rheumatoid arthritis is in sustained remission, wherein the kit comprises reagents for establishing the level of MerTK and CD206; wherein an elevated level of MerTK and CD206 compared to the one or more reference values is indicative of sustained remission.

Optionally, the kit may also comprise reagents for establishing the level of TREM2; wherein an elevated level of TREM2 compared to the one or more reference values is indicative of sustained remission.

Optionally, the kit may also comprise reagents for establishing the level of CD163; wherein an elevated level of CD163 compared to the one or more reference values is indicative of sustained remission.

Suitably, the kit may also comprise reagents for establishing the level of said biomarkers by RT-qPCR, microarray analysis, digital PCR, whole transcriptome shotgun sequencing, direct multiplexed gene expression analysis or whole transcriptome sequencing. A kit of parts for determining rheumatoid arthritis sustained remission, wherein the kit comprises: j) at least one binding partner that selectively binds to the MerTK biomarker, or a fragment thereof; k) at least one binding partner that selectively binds to the CD206 biomarker, or a fragment thereof;

L) a positive control for the detection of said biomarkers; m) at least one binding partner that selectively binds to a nucleic acid or protein which operates as an internal control; and n) optionally an internal standard.

Preferably, the kit further comprises at least one binding partner that selectively binds to the TREM2 biomarker, or a fragment thereof and a positive control for the detection of TREM2.

Preferably, the kit further comprises at least one binding partner that selectively binds to the CD163 biomarker, or a fragment thereof and a positive control for the detection of CD163.

Optionally the kit may be configured to determine the levels of MerTK, CD206, TREM2 and CD163 only.

Sequences Table 1. sequences and accession numbers used in the current study (see NCBI- GenBank Flat File Release 234.0), updated 15 October 2019.

The invention includes the combination of the aspects and preferred features described except where such a combination is clearly impermissible or expressly avoided.

The section headings used herein are for organizational purposes only and are not to be construed as limiting the subject matter described.

Aspects and embodiments of the present invention will now be illustrated, by way of example, with reference to the accompanying figures. Further aspects and embodiments will be apparent to those skilled in the art. All documents mentioned in this text are incorporated herein by reference.

Throughout this specification, including the claims which follow, unless the context requires otherwise, the word “comprise”, and variations such as “comprises” and “comprising”, will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integers or steps.

It must be noted that, as used in the specification and the appended claims, the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. Ranges may be expressed herein as from “about” one particular value, and/or to “about” another particular value. When such a range is expressed, another embodiment includes from the one particular value and/or to the other particular value. Similarly, when values are expressed as approximations, by the use of the antecedent “about,” it will be understood that the particular value forms another embodiment.

Where a nucleic acid sequence is disclosed herein, the reverse complement thereof is also expressly contemplated.

Brief Description of the Figures

Embodiments and experiments illustrating the principles of the invention are further described hereinafter with reference to specific examples and the accompanying figures in which:

Figure 1 shows that synovial tissue of RA patients in sustained clinical and ultrasound remission is enriched in MerTK/CD206^pos macrophages, with decline predictive of flare. (a) Representative expression of MerTK and CD206 on STMs distinguishes two main populations of STMs (b) Distribution of MerTK/CD206^pos and MerTK/CD206^neg STMs in healthy controls (n=10 subjects), RA naive-to-treatment (n=43 patients), RA resistant- to-treatment (n=30 patients), and patients in sustained clinical and ultrasound remission (n=36 patients) (c) Representative expression of CD163 shows that this receptor is expressed exclusively on CD206/MerTK^pos STMs (d-e) Distribution of CD163 expressing cells in healthy donors and RA patients as described in b. (f) Spearman correlation analyses between DAS28 value and the frequencies of MerTK/CD206^pos, CD163/CD206^pos and MerTK/CD206/CD163^pos STMs in RA patients as in b. (g) Comparison of the levels of MerTK expression, (h) distribution of CD163/CD206^pos, (i) MerTK/CD163/CD206^pos and (j) total CD163^pos STMs in RA patients in sustained DAS- based remission (n=24) compared to RA patients in Boolean remission (n=11). (k) Comparison of the distribution of STM phenotypes in RA patients who maintained in remission (n=11 ) compared to those who flared (n=11 ) after treatment cessation (n=22, total number of patients who consented to treatment modification). (I) Comparison of the distribution of MerTK/CD206and CD163/CD206 positive and negative cells in patients as in k. This shows that patients who maintained in remission have higher % of MerTK/CD206^pos compared to those who flared after treatment cessation (m) Odds Ratio (OR) and 95% Confidential Interval (95% Cls) are presented and show that the proportion of MerTK/CD206^pos STMs <47.5% or the MerTK/CD206^pos to MerTK/CD206^neg ratio <2.5 emerged as independent factors predicting disease flare after treatment discontinuation (n-p) Representative photo of immunohistochemistry of CD68 (Brown) and immunofluorescence staining of CD68 (green), MerTK (red) and nuclei (blue) in synovial tissue biopsies samples of RA patient in sustained remission (n) and with active disease (o-p). MerTK^pos STMs are mainly localized in the lining layer in remission (n). MerTK^posCD68⁺ cells are dispersed (o) or not present in tissues from RA patients with active disease (p). The dotted line indicates the reference edge of synovial lining/sublining areas (Magnification 20X). White arrows indicate CD68^pos/MerTK^pos cells and white dotted arrows indicate CD68^pos/MerTK^neg cells, respectively (Magnification 40X). Data are mean ± s.e.m.; p value are provided on the graphs or marked with * (< 0.05). The difference in (a-l) in individual STM populations between distinct joint conditions were evaluated using one-way ANOVA with Tukey correction for multiple comparison or two-tailed nonparametric unpaired Mann-Whitney test if 2 groups were compared. Figure 2 shows that single cell transcriptomics defines distinct STM subpopulations and phenotypes in different human joint immuno-conditions. (a) UMAP visualization of 9 clusters of synovial tissue macrophages identified as a result of analysis of scRNAseq STM data. Each cell is represented by an individual point and colored by cluster identity (b) Heatmap illustrating scaled expression of the top 20 differentially expressed genes per cluster (rows are genes, columns represent cells). Marker genes of interest are annotated, and the total number of genes differentially expressed by each cluster is also shown next to the cluster presented on the heatmap. All genes are expressed in at least 40% of cells in each cluster. Average log-fold change >0.25. (c) The expression of clusters’ marker genes is shown on log-normalized violin plots; median highlighted by black dot. (d) The top 50 PCs were embedded in the top 3 diffusion-map components to demonstrate the transcriptional relationship between the clusters (e) Hierarchical clustering of identified STMs visualized as a dendrogram (f) Expression of MerTK and CD163 in the 9 STM clusters (g) Proposed classification of human STMs based on scRNAseq and flow cytometry data (h) Split UMAP visualization demonstrating relative changes in the STM phenotypes between disease groups. The numbers were normalized to ~5,000 cells per group (i) Bar and dot plots illustrating the change in cluster distribution across conditions. Statistically significant differences between a given condition and at least one other condition are highlighted with ^* (two-way ANOVA, corrected for multiple comparisons with Tukey test). Each dot represents individual patient (j) Expression of S100A9 and SPP1 in the synovium of the validation PEAC cohort correlates with disease activity (k-m) Illustration of TREM2^pos and FOLR2^pos subsets on UMAP (k-l), and flow cytometry validation of TREM2^pos and FOLR2^pos STM clusters (m) in active RA and RA in remission.

Figure 3 shows that MerTK/CD206^pos and MerTK/CD206^neg STM populations have distinct pro- and anti-inflammatory phenotypes, respectively (a) Sorting strategy for MerTK positive and negative STMs (n=47 patients), and overview of functional studies. In vitro production by MerTK positive and negative STMs of (b-c) pro- and anti- inflammatory mediators, and (d) resolvin D1 . (e) Expression levels of MerTK on MerTK/CD206^pos STMs are reduced in active RA. (f) Visualisation of S100A12 mRNA expression on STM UMAP (g) Production of S100A12 by LPS-stimulated MerTK/CD206^neg and MerTK/CD206^pos STMs, FACS-sorted from biopsies of patients with active RA or RA in remission (h) Soluble mediators differentially expressed in the MerTK^neg -CD52^pos/S100A12^pos cluster between conditions are shown (p-value < 0.05, adjusted by Bonferroni Correction).

Figure 4 shows that TREM2^pos and FOLR2^pos clusters of MerTK^pos STMs from RA patients in remission have a unique transcriptomic signature (a) Heatmap illustrating scaled expression of the top 30 marker genes of each condition within the TREM2^|0W, TREM2^high and FOLR2/LYVE1^pos clusters. Rows are genes while columns show equal pseudo-bulk expression per condition within each cluster. All genes are expressed in at least 60% of cells in that condition with an average log fold-change >0.25 and p< 0.05 after correction for multiple comparisons (b) Table demonstrating the number of genes differentially expressed between conditions for each cluster (c) Violin plots illustrating the expression of top marker genes of TREM2^pos and FOLR2^pos clusters unique for healthy or remission STMs (d) Transcription factors identified in TREM2^pos and FOLR2^pos clusters in remission are inhibited in macrophages incubated with MerTK inhibitor. ^* p< 0.05 paired t-test. (e) Expression of KLF4 and NRA42 in different synovial tissue of PEAC cohort negatively correlates with DAS28.

Figure 5 shows that MerTK positive macrophages control inflammatory response of synovial fibroblasts (a) Schematic of a direct co-culture system of macrophages (MQ) with synovial fibroblasts (FLS). (b) mRNA expression levels of IL-6 and MMP-1 in FLS (FACS-sorted from co-culture with macrophages). MerTK inhibition of LPS and of LPS+Dex pre-treated macrophages enhanced their activation of FLS. (c-d) Levels of mediators in co-culture supernatants show that MerTK inhibition in macrophages enhanced the production of MMP1 , MMP3 and IL-6 by FLS. (e) Schematic of an indirect co-culture system of macrophages with synovial fibroblasts (FLS). (f) MerTK inhibition of LPS pretreated macrophages increased the concentrations of MMP1 , MMP3 and IL-6 in co-culture supernatants. Data are presented as dot-plot of individual experiments (n=5) with a bar indicating the mean. ^* and & represent p<0.05 in one-way Anova, followed by correction for multiple comparisons, or a paired t-test if two groups were compared (g) Heatmap demonstrating scaled log-expression of soluble mediators that were differentially expressed between FLS co-cultured with macrophages pre- stimulated with LPS and MerTK inhibitor, compared with FLS co-cultured with macrophages pre-treated with LPS only. Genes were adjusted for multiple comparison (p<0.01 and log2-fold change > +/-1.5). Figure 6 shows that lining layer FLS of RA patients in remission express reduced levels of MMPs and chemokines compared to active RA while sublining FLS are source of GAS6. (a) UMAP visualization of heterogeneity of synovial fibroblasts in active RA (n=4) and RA in remission (n=3). Each cell (n=13,949) is visualized by an individual point and colored by cluster identity (b) Stacked bar plots demonstrate the relative proportion of cells within each cluster per clinical group (c) Bar plots illustrate the change in cluster distribution between clinical groups. Each point represents an individual sample (d) Heatmap illustrates scaled, batch-normalized expression of the top 10 differentially expressed genes per cluster, showing that THY1 + sublining clusters express GAS6 particularly in cluster 5 (CXCL14^pos). Rows are genes and columns represent cells. All genes are expressed in at least 40% of cells in a given cluster. Average log-fold change > 0.25. (e) Violin plots showing the distribution of log-normalized expression of genes of interest. The median value is represented by a black dot. (f) Differentially expressed genes in the lining layer FLS of remission RA as compared to active RA. Scatter-plot illustrates pseudo-bulk expression of each differentially expressed gene and the top 10 genes for each condition are annotated. Heatmap shows the scaled, sample-specific pseudo-bulk expression of the top 20 differentially expressed genes, rows are genes and columns represent-sample pseudo-bulk expression. All genes are expressed in at least 60% of cells in that group. Average log-fold change > 0.25 with p< 0.05 after correction for multiple comparison, (g-h) Violin plots showing log-normalized expression of genes for MMPs, chemokines and anti-inflammatory mediators in lining layer FLS. The median value is represented by white dot. (i) mRNA expression of GAS6 in distinct sublining FLS clusters comparing active and remission RA. The THY1^high cluster shows greater expression of GAS6 in remission as compared to active disease. Median values are represented by white dots (j) GAS6 is produced by cultured FLS derived from biopsies of RA patients; treatment naive, treatment resistant, and RA in remission. Each dot represents a separate FLS line (n=5 per condition) ^*p < 0.05, two-way Anova with correction for multiple comparisons.

Figure 7 shows strategy for the analysis of STMs (a) A list of antibodies used for gating STMs (b) A gating strategy for STM analysis (c) Representative expression of MerTK and CD206 on STMs from healthy, RA patients with active disease (naive-to-treatment or resistant-to-treatment) and patients in disease remission (d) Illustration (Cytobank software) of CD163 expression on STM populations defined by MerTK/CD206 expression showing that CD163 is exclusively expressed on MerTK/CD206pos. CD163 expression is presented as arbitrary units.

Figure 8 shows the distribution of different STM populations in healthy and RA synovium and their correlations with clinical parameters. Healthy donors (n=10) are shown in green, treatment-naive RA are shown in orange (n=43), treatment-resistant RA are shown in red (n=30) and RA in sustained clinical and ultrasound remission are shown in blue (n=36). (a) Distribution of distinct STM populations defined by MerTK, CD163 and CD206 is presented (b) Spearman correlation analyses between Disease Activity Score over 28 joints (DAS28-CRP) or semiquantitative synovitis score (Krenn score) or Simplified Disease Activity Index (SDAI) or power doppler grade and the frequencies of RA MerTK/CD206pos, MerTK/CD206neg, CD163/CD206pos, CD163/CD206neg cells are presented. Data in (a) are mean ± s.e.m.; p-values are provided on the graphs or marked with *(<0.05). The difference in individual STM populations between different joint conditions were evaluated using one-way ANOVA with Tukey correction for multiple comparison or two-tailed nonparametric unpaired Mann-Whitney test if 2 groups were compared.

Figure 9 shows the distribution of MerTK/CD163 positive and negative STM populations in healthy and RA synovium and their correlations with clinical parameters. Healthy donors (n=10) are shown in green, treatment-naive RA are shown in orange (n=43), treatment-resistant RA are shown in red (n=30) and RA in sustained clinical and ultrasound remission are shown in blue (n=36). (a) Distribution of STM populations defined by MerTK, CD163 and CD206 is presented (b) Spearman correlation analyses between Disease Activity Score over 28 joints (DAS28-CRP) or semiquantitative synovitis score (Krenn score) or Simplified Disease Activity Index (SDAI) or power doppler grade and the frequencies of RA MerTK/CD163pos, MerTK/CD163neg, and MerTK/CD163/CD206pos are presented. Data in (a) are mean ± s.e.m.; p-values are provided on the graphs or marked with *(<0.05). The difference in individual STM populations between different joint conditions were evaluated using one-way ANOVA with Tukey correction for multiple comparison or two-tailed nonparametric unpaired Mann-Whitney test if 2 groups were compared.

Figure 10 shows the distinct distribution of STM populations defined by CD163, MerTK and CD206 between different types of remission (a-d) Comparison of STMs distribution between patients with disease remission defined by DAS28 (n=24 RA patients) and those who met Boolean remission criteria (n=11 ). The analyses include single marker positive or negative STMs in (a), MerTK/CD206, CD163/CD206 positive and negative STMs in (b), MerTK/CD163 positive and negative STMs in (c) and MerTK/CD163/CD206pos STMs in (d). Comparison of STMs distribution between RA patients in sustained clinical and US remission who flared (n=11 ) or maintained in disease remission (n=11) after treatment discontinuation. The analyses include CD163, MerTK and CD206 single marker positive and negative STMs in (e), MerTK/CD206 positive and negative STMs in (f), CD163/CD206 positive and negative STMs in (g), MerTK/CD163 positive and negative STMs in (h) and MerTK/CD163/CD206pos STMs in (i). (j) ROC curves for cut-off values of distribution data for CD206pos, MerTK/CD206pos, MerTK/CD206neg, CD163/CD206pos, CD163/CD206neg STM and MerTK/CD206pos to MerTK/CD206neg ratio in patients described in e-i. (k) Comparison of frequencies of occurrence of flare in RA patients who discontinued treatment based on STM distribution. Patients are stratified on the bases of the odds ratio cut-off value for different STM populations. Data in (a-i) are mean ± s.e.m.; p value are provided on the graphs. The difference in individual STM populations between different remission states were evaluated using two-tailed nonparametric unpaired Mann-Whitney test.

Figure 11 shows a trajectory analysis of synovial tissue macrophage populations (a) Hierarchical clustering of the average expression of each synovial macrophage cluster, visualized as a dendrogram, illustrates how the clusters are related (b-c) Single-cell trajectory analysis of synovial macrophage populations (Monocle 2.99) integrating cluster information (b) and pseudotime (progressively changing transcriptional state). A downsample of cells (n=10,000) were used in the Monocle analysis pipeline. Clusters identified by Seurat pipeline were used for differential expression analysis before performing dimensional reduction to generate a DDRTree. Cells were then ordered according to their position in the trajectory. The resulting trajectory plot is shown with cells coloured by (b) cluster identity or (c) pseudotime. This analysis shows branch points (decision points) of cellular differentiation into MerTKpos: FOLR2pos, TREM2pos and MerTKneg: HLApos, S100A12pos and SPP1 pos.(d) To find genes which change as a function of pseudotime, differential expression was performed. The heatmap shows the result of this analysis. Hierarchical clustering identifies 6 gene clusters with distinct expression patterns that clearly differentiate MerTKpos clusters from MerTKneg clusters in pseudotime function (e) Illustration of relative expression of the top 20 differentially expressed genes as a function of pseudotime. Cells are coloured by synovial tissue macrophage cluster identity. It must be noted the pseudotime function is used here to illustrate the developmental distance between different populations.

Figure 12 shows changes in the STMs clusters in health, synovitis and resolution of synovitis (a) Bar plot illustrating the change in cluster distribution (clusters are ordered here as in the heatmap on Figure 2b) across conditions. Statistical differences are highlighted and were calculated by two-way ANOVA with correction for multiple comparisons (Tukey test) (b) CLECIOA-positive STM cluster localization on the STM UMAP. (c) Expression of the top markers of HBEGF positive synovial myeloid cells described by Kuo et al. (2019) indicating that this population might be embedded in the MerTK-negative CLECIOA-positive synovial tissue resident dendritic cells (d) Expression of HBEGF is increased in CLEC10A positive cells in early undifferentiated arthritis.

Figure 13 shows a pathway analysis of differentially expressed genes between clusters reveals distinct functionality in effector pathways of synovial tissue macrophage subpopulations. (a-b) Heatmaps illustrating scaled pseudo-bulk expression of significantly enriched pathways within four MerTKpos clusters (a) and four MerTKneg clusters plus ICAMIpos cluster of MerTK positive STMs (b). Rows are genes and columns represent average expression for cells in each cluster by subject group. Differentially expressed genes between clusters were used to perform GO and IPA analysis. Upregulated genes from pathways of interest are annotated. All genes are significantly expressed in at least 60% of cells in that cluster. DE Genes identified by Seurat’s function were filtered afterwards to ensure that the p-value adjusted by Bonferroni Correction is significant (p-value < 0.05). Average log fold change >0.25.

Figure 14 shows that MerTK positive clusters in RA patients in disease remission show a unique gene expression pattern. Bar plots illustrating the number of intersecting genes differentially expressed between healthy and UPA, healthy and naive active RA, healthy and resistant RA, and healthy and RA in remission for each MerTKpos clusters of interest (TREM2low, TREM2high, and FOLR2/LYVE1pos). Red bar plots represent common upregulated genes and green bar plots represent common downregulated genes. Expression of genes identified as resolved (upregulated in active RA and downregulated in disease remission), super-inflamed (upregulated in active RA and in disease remission), restored (downregulated in active RA and restored to normal in disease remission), super-repressed (downregulated in active RA and in disease remission) are illustrated per cluster as a heatmap displaying the pseudo-bulk expression per group. Gene expression pattern was dissected using R package.

Figure 15 shows that MerTK positive macrophages control the inflammatory response of synovial fibroblasts (a) Representative histograms showing FACS-MerTK expression on monocyte-derived macrophages stimulated with LPS or LPS plus dexamethasone (LPS +Dex) as compared to isotype control (iso) (b) Expression of GAS6 mRNA in FLS FACS-sorted from direct co-cultures with macrophages as described in the brief description of Figure 5. This shows that GAS6 expression is reduced in FLS interacting with macrophages pre-treated with MerTK inhibitor (c) Levels of mediators in a direct co-culture systems showing that inhibition of MerTK in macrophages does not affect MMP2 or MMP9 production (d) Levels of mediators in a trans-well co-culture system showing that inhibition of MerTK in macrophages does not affect MMP2 and MMP9 production. Data are presented as dot plots of individual experiments with a bar indicating the mean. * and & p=<0.05 in One way ANOVA followed by correction for multiple comparison (*0 or a paired t-test (&) if two groups were compared.

Figure 16 shows a comparison of human and mouse single-cell transcriptional profiling of synovial macrophages (a) UMAP projections for human and mouse scRNAseq data analysed separately. Mouse data was clustered at a resolution of 0.3 and UMAP projection represents the top 12 PCs. Human data for this comparison included samples from healthy tissue, undifferentiated arthritis (UPA) and naive, active RA to align with disease conditions modelled in mouse data (b) Dendrogram representing the relationship between human macrophage phenotypes and mouse clusters identified by Culemann et al., 2019 Nature, doi: 10.1038/s41586-019-1471 -1. This plot was generated from the hierarchical clustering of the average expression of orthologous genes by each population (c) Original UMAP projection following the integration of synovial tissue macrophages from both human sample cohorts (discovery and validation) but before the removal of doublets and contaminant non macrophage lineage positive cells. The populations selected for analysis are demarcated. For the current analysis the proliferating cells are included (d) The UMAP projection of synovial tissue macrophages (with proliferating cells) following the merging and renaming of clusters. (e) Violin plots show expression of FOLR2 and CSF1 R by the cycling macrophage population characterised by STMN1 expression.

Figure 17 shows an investigation into potential barrier function in human synovial tissue macrophages: MerTK positive clusters are enriched in tight junction proteins. Fleatmap illustrating scaled pseudo-bulk expression of significantly enriched pathways by each patient group within each of identified synovial tissue macrophage clusters. Rows represent genes with a potential contribution to synovial lining layer barrier function (GO pathway- involved tight junction assembly and organization). Columns equal average expression for cells in each cluster by subject group. Genes identified in mouse synovial lining macrophages by Culemann etal., 2019, Nature, doi: 10.1038/s41586-019-1471-1 as tight junction proteins are in blue boxes. Among them TJP1 is expressed in human MerTK/TREM2 and in MerTK/FOLR2/LYVE1 positive STM subsets.

Figure 18 shows that distinct synovial tissue macrophage subsets regulate inflammation and provide a cellular and molecular mechanism for disease remission in rheumatoid arthritis. The HEALTHY synovial membrane mainly consists of two subsets of MerTKpos STMs: TREM2pos and FOLR2/LYVE1pos. Their transcriptomics suggest immunoregulatory functions, e.g. production of retinoic acid or expression of B7-like inhibitory molecules (VISG4). In ACTIVE RA (treatment-naive and treatment-resistant), synovial membrane is infiltrated by MerTKneg CD52pos STMs with two distinct phenotypes producing either pathogenic S100A alarmins (e.g. S100A12) and IL-ipb, or osteopontin. Both are also the main source of pathogenic TNF and IL-6, and potent inducers of pro-inflammatory (chemokines) and destructive (MMPs) properties of synovial stromal fibroblasts (FLS). RA in REMISSION (maintained after treatment cessation) is characterized by restoration of the TREM2pos and FOLR2/LYVE1pos MerTKpos subsets with transcriptome and regulatory properties that are different from those of healthy STMs; they poorly produce pro-inflammatory cytokines, which is further negatively regulated by GAS6 binding to its receptor MerTK. Instead these MerTKpos STMs produce resolvins (inflammation-resolving lipid mediators) and IL-10. Their transcriptome is characterized by MerTK-dependent transcription factors that are negative-regulators of inflammation. MerTKpos macrophages forestall pathogenic activation of FLS. RA patients in remission, with STMs composed of less than 47.5% MerTKpos, or ratio of MerTKpos to MerTKneg less than 2.5 have a higher likelihood of FLARE after treatment cessation. The MerTKneg STMs in these patients are of CD52/S100A12pos phenotype and release S100A12 alarmin upon stimulation, suggesting a role in the initiation of disease flare. In summary, sustained remission of arthritis appears to be an active process maintained by tissue-resident subsets of MerTKpos STM subpopulations (e.g. TREM2pos and FOLR2/LYVE1pos) that govern the functions of pro-inflammatory CD52/S100A12pos STMs and synovial fibroblasts to reinstate and maintain homeostasis. BM, bone marrow; SM, synovial membrane; FLS, fibroblast like synoviocytes; MMPs, metalloproteinases; SPP1, osteopontin; KLFs krueppel like factors; NR4As, nuclear receptor subfamily 4 group A; ATF3, c AMP- dependent transcription factor 3; MerTK, tyrosine-protein kinase Mer; TREM2, triggering receptor expressed on myeloid cells 2; LYVE1, lymphatic vessels endothelial hyaluronan receptor 1; FOLR2, folate receptor beta; GAS6, growth arrest-specific 6; S100A12, S100 calcium-binding protein A12; THY1, CD90.

Figure 19 shows the integration of scRNAseq data (a) Metadata for CD64/CD11 b positive synovial tissue macrophages sequenced in the Discovery Cohort at Oxford Genomics Centre (b) Metadata for all synovial cell types sequenced in the Validation Cohort at Glasgow Polynomics. (c) Myeloid cells sequenced in the Validation cohort are separated computationally (based on positive expression of CD14, MARCO and LYZ) and integrated with synovial macrophages sequenced in the Discovery Cohort.

Figure 20 shows quality control and sample filtering (a) Original UMAP projection following the integration of synovial tissue macrophages from both sample cohorts (as described in brief description of Figure 19), but before the removal of doublets, non- macrophage lineage positive cells and cycling cells. Populations retained for further analysis are highlighted (b) UMAP projection of synovial tissue macrophages following doublet removal and sample filtering - removal of SA139 (low seq. depth) and SA225 (remaining batch effect) (c) Final UMAP projection of synovial tissue macrophages following merging and renaming of clusters. Clusters were confirmed by ensuring their distinct transcriptional profile by differential expression and by gene ontology analysis (d) Original PCA analysis of pseudobulk expression of cells within each cluster by sample revealed that sample SA225 was responsible for PC1 therefore this sample was removed from further analysis (e) PCA analysis of sample-cluster pseudobulk after removal of SA225. Examples

In the following Examples, the inventors describe experimental results which illustrate the principles of the invention.

Example 1 : Svnovial tissue of RA patients in sustained remission is enriched in

MerTK/CD206^pos STMs, with decline predictive of flare

To determine the phenotypic spectrum of human synovial tissue macrophages (STMs) spanning health, immune-mediated synovitis and resolution, we characterised expression of candidate immune-receptors. These included CD163; previously found on healthy STMs (Kurowska-Stolarska et al., 2017, RMD Open 3, e000527, doi:10.1136/rmdopen-2017-000527; Singh et al., 2004, Ann Rheum Dis 63, 785-790), and MerTK and CD206; key markers of murine tissue-resident macrophages with immune-homeostatic function (Davies et al., 2013, Nat Immunol 14, 986-995; Gonzalez et a!., 2017, J Exp Med 214, 1281-1296; Hogg et al., 1985, Immunology 56, 673-681 ). Using these markers, we investigated the relative composition of STMs in dispersed synovial tissue biopsies from RA patients (45 treatment-naive active RA, 31 treatment- resistant active RA, 36 RA in sustained remission) and 10 healthy donors (Supplementary Table 1 Alivernini et al. 2020 Nature Medicine 26:1295-1306. doi: 10.1038/s41591 -020-0939-8). STMs were identified firstly by their expression of CD64, CD11 b, MHCII and the absence of other cell-lineage markers (Fig 7a-b). Two predominant STM populations were then identified by co-expression of MerTK and CD206: MerTK/CD206^pos and MerTK/CD206^neg (Fig.1a and extended data Fig.8a). The relative proportions of the main populations (Fig.lb) indicated that STMs from healthy donors were almost exclusively MerTK/CD206^pos. A predominance of MerTK/CD206^pos also distinguished RA patients in sustained remission from those with active RA. These STMs were fewer in RA patients with active disease, with commensurately increased MerTK/CD206^neg STMs (Fig.1a-b). CD163 was co- expressed as a subpopulation of MerTK/CD206^pos STMs (Fig.1c and Fig.7c-d) and this MerTK/CD163/CD206^pos STM population was increased in RA patients in sustained remission compared to active RA (Fig.1d-e and Fig.9a). The relative proportion of the MerTK/CD206^pos, CD163/CD206^pos and MerTK/CD163^pos STMs phenotypes that dominated in patients in disease remission, correlated negatively with clinical parameters of RA disease activity including DAS28 (CRP) (Fig.lf), SDAI, synovial hypertrophy and changes in US-measured synovial blood flow whereas MerTK/CD206^neg and CD163/CD206^neg STMs correlated positively (extended data Fig.8b and 9b).

We validated the relationship between STM subpopulations and tissue resolution using the highly stringent Boolean criteria for disease remission in RA (Bykerk et al., 2012, Rheumatology (Oxford) 51 Suppl 6, vi16-20, doi:10.1093/rheumatology/kes281 ). Among the 36 RA patients in clinical DAS28-defined remission, 11 also met the Boolean remission criteria at the time of biopsy. Their STMs were characterized by an increased density of MerTK expression on the MerTK/CD206^pos population (Fig.lg), and an increased proportion of CD163-expressing STMs as compared to equivalent populations in sustained DAS28-defined remission (Fig.1h-j and extended data Fig.10a-d). This indicates that ‘deeper remission’ is associated with changes in STM populations.

Next, we investigated the clinical significance of MerTK^pos and MerTK^neg STM populations associated with subsequent occurrence of flare after treatment withdrawal. All 36 RA patients that achieved sustained clinical and ultrasound remission received identical prior treatment (MTX plus TNF-inhibitor; Suppl. Tab.1). Among them, 22 RA patients consented to taper then discontinue treatment. Remission was maintained in 11 , whereas 11 patients flared. Those patients who maintained remission had, at the time of synovial tissue biopsy, a distinct STM composition compared to those who subsequently flared (Fig.1k). The most prominent difference was a higher proportion of MerTK/CD206^pos STMs and correspondingly lower MerTK/CD206^neg (Fig. 11 and extended data Fig.10e-i).

We hypothesised that the STM population profile may provide a prognostic biomarker predictive of disease flare after treatment change. To test this, logistic regression analysis was performed on the distribution data of MerTK/CD206^pos and MerTK/CD206^neg STM populations (Fig.1 Ol-k). The individual proportion of MerTK negative STMs was insufficient to predict flare. However, the proportion of MerTK/CD206^pos STMs <47.5% or the MerTK/CD206^pos to MerTK/CD206^neg ratio <2.5 emerged as independent factors predicting disease flare after treatment tapering and discontinuation in RA patients [Odds ratio: 13.5 (95%CI: 2.3-80.8) and 16.2 (95%CI: 2.61-100.45)], respectively) (Fig.lm). These data suggest that reinstatement of MerTK/CD206^pos STMs in synovium and their potential governance over any other STM phenotypes may be key for maintaining synovial tissue homeostasis in RA. Next, we explored the synovial tissue localization of MerTK^pos STMs via histology comparing RA in remission with active RA. IHC/IF demonstrated that MerTK^pos STMs localized mainly in the lining layer both in remission and in active RA (Fig.1n-p). In accordance with flow-cytometry, the majority of STMs (CD68^pos) in remission RA are MerTK^pos and form a tight layer (Fig.ln). In contrast, in active RA the CD68^pos cells had heterogeneous expression of MerTK with many lining layer areas lacking any MerTK expression (Fig.lp); if present, the MerTK^posCD68⁺ cells are dispersed (Fig.lo)

Together, these data suggest that MerTK/CD206^pos STMs variably characterize RA synovial tissue in distinct clinical phases and are indicative of persistent remission.

Figure 1 shows that synovial tissue of RA patients in sustained clinical and ultrasound remission is enriched in MerTK/CD206^pos macrophages, with decline predictive of flare. The content of panels (a)-(p) are described below (a) representative expression of MerTK and CD206 on STMs distinguishes two main populations of STMs (b) Distribution of MerTK/CD206^pos and MerTK/CD206^neg STMs in healthy controls (n=10 subjects), RA naive-to-treatment (n=43 patients), RA resistant-to-treatment (n=30 patients), and patients in sustained clinical and ultrasound remission (n=36 patients) (c) Representative expression of CD163 shows that this receptor is expressed exclusively on CD206/MerTK^pos STMs (d-e) Distribution of CD163 expressing cells in healthy donors and RA patients as described in b. (f) Spearman correlation analyses between DAS28 value and the frequencies of MerTK/CD206^pos, CD163/CD206^pos and MerTK/CD206/CD163^pos STMs in RA patients as in b. (g) Comparison of the levels of MerTK expression, (h) distribution of CD163/CD206^pos, (i) MerTK/CD163/CD206^pos and (j) total CD163^pos STMs in RA patients in sustained DAS-based remission (n=24) compared to RA patients in Boolean remission (n=11 ). (k) Comparison of the distribution of STM phenotypes in RA patients who maintained in remission (n=11 ) compared to those who flared (n=11 ) after treatment cessation (n=22, total number of patients who consented to treatment modification). (I) Comparison of the distribution of MerTK/CD206 and CD163/CD206 positive and negative cells in patients as in k. This shows that patients who maintained in remission have higher % of MerTK/CD206^pos compared to those who flared after treatment cessation (m) Odds Ratio (OR) and 95% Confidential Interval (95% Cls) are presented and show that the proportion of MerTK/CD206^pos STMs <47.5% or the MerTK/CD206^pos to MerTK/CD206^neg ratio <2.5 emerged as independent factors predicting disease flare after treatment discontinuation. (n-p) Representative photo of immunohistochemistry of CD68 (Brown) and immunofluorescence staining of CD68 (green), MerTK (red) and nuclei (blue) in synovial tissue biopsies samples of RA patient in sustained remission (n) and with active disease (o-p). MerTK^pos STMs are mainly localized in the lining layer in remission (n). MerTK^posCD68⁺ cells are dispersed (o) or not present in tissues from RA patients with active disease (p). The dotted line indicates the reference edge of synovial lining/sublining areas (Magnification 20X). White arrows indicate CD68^pos/MerTK^pos cells and white dotted arrows indicate CD68^pos/MerTK^neg cells, respectively (Magnification 40X). Data are mean ± s.e.m.; p value are provided on the graphs or marked with ^* (< 0.05). The difference in (a-l) in individual STM populations between distinct joint conditions were evaluated using one-way ANOVA with Tukey correction for multiple comparison or two-tailed nonparametric unpaired Mann-Whitney test if 2 groups were compared.

Example 2: scRNAseq of STMs defines heterogeneity within MerTK/CD206^pos and

MerTK/CD2O6^neg populations associated with synovial homeostasis, progression and resolution

To better understand the heterogeneity and changes in the patterns of human MerTK/CD206^pos and MerTK/CD206^neg STM populations during development and resolution of arthritis, we performed unbiased molecular profiling (scRNAseq), followed by validation with flow cytometry, of broad STMs compartments (CD11 b/CD64^pos) from treatment-naive RA (n=5), treatment-resistant RA (n=6) and RA in sustained remission (n=6). As comparators for RA, we included STMs from healthy (n=4) and pathological control (autoantibody-negative, undifferentiated peripheral arthritis (Alivernini et al., 2018, Front Med (Lausanne) 5, 186, doi:10.3389/fmed.2018.00186), n=4)

(Supplementary Table 2-3 Alivernini et al. 2020 Nature Medicine 26:1295-1306. doi: 10.1038/S41591 -020-0939-8). The transcriptome of 32,141 STMs (>5, 000/condition) revealed 9 STM clusters (phenotypes) (Fig.2a, Supplementary Tables 2-4 Alivernini et al. 2020 Nature Medicine 26:1295-1306. doi: 10.1038/s41591 -020-0939-8), each characterized by the expression of 63-432 unique genes (Fig.2b-c).

The developmental relationship between these 9 clusters revealed by diffusion map (Fig.2d), cell-trajectory analysis (Fig.11), hierarchical-clustering analyses (Fig.2e) and by gene expression of MerTK/CD163 (CD206 undetectable) (Fig.2f), classifies them into 4 subsets (TREM2^pos, FOLR2^pos’ HLA^pos and CD52^pos) comprising two main MerTK^pos and MerTK^neg STM populations as defined by flow cytometry (Fig 1 ). The TREM2^pos and FOLR2^pos are MerTK^pos STMs whereas the FILA^pos and CD52^pos comprise MerTK^neg STMs {the full STM taxonomy is proposed in Fig.2g). Analysis of 9 clusters provided insight into phenotypic changes within each of the 2 subpopulations of MerTK positive and negative STMs. The TREM2^pos subpopulation contains two phenotypes; TREM2^|0W and TREM2^high that are further distinguished by co-expression of TIMD4 and CD163. The FOLR2^pos subpopulation contains three distinct phenotypes categorized by top marker genes as ID2^pos, LYVE1^pos or ICAM1^pos. The MerTK negative FILA^pos subpopulation contains two clusters distinguished by either an interferon signature (ISG15^pos cluster) or an antigen presenting cell signature (CLEC10A^pos cluster). The latter resemble CD1c⁺ dendritic cells (DC) (Villani et al., 2017, Science 356, doi: 10.1126/science. aah4573) and likely represents synovial tissue resident DCs. The CD52^pos subpopulation of MerTK^neg STM is enriched in either alarmins (S100A12^pos cluster) or osteopontin (SPP1/CD9^pos cluster). The SPP1^pos and the ISG15^pos clusters (i.e MerTK negative) were previously noted in the synovium of active RA (Zhang et al., 2019, Nat Immunol 20, 928-942, doi: 10.1038/s41590-019- 0378-1 ) thus validating our analysis strategy.

To discover condition-specific STM profiles, and genes indicative of mechanisms of homeostasis, pathogenesis and resolution of arthritis, we compared differences in the relative proportions of the 9 clusters and their unique pathways between clinical states (Fig.2h-i, Fig.12). The healthy synovium contains predominantly MerTK^pos STMs (validating flow cytometry data from Fig.1 ) comprising TREM2^pos and FOLR2^pos subpopulations. Pathway analysis revealed that these subpopulations are enriched in complement and defensin pathways (contrasting with MerTK^neg STMs), suggesting efferocytosis and anti-microbial functions (Fig.13a). They also express genes of retinoic acid production ( ALDH1A1 , RBP4) which drive regulatory T-cell differentiation (Nolting et al., 2009, J Exp Med 206, 2131-2139) and show high expression of the B7-related co-inhibitory molecule VSIG4 which inhibits effector T-cells (Vogt et al., 2006, J Clin Invest 116, 2817-2826), suggesting a role in the local regulation of adaptive immunity. Healthy synovium had the highest proportion of the TREM2^high cluster compared to other states, consistent with recent murine studies suggesting that TREM2^pos cells form a protective epithelial-like lining barrier (Culemann et al., 2019 Nature, doi: 10.1038/S41586-019-1471 -1 ). In addition, human TREM2^high STMs have a distinct transcriptome indicative of phagocytosis e.g. high expression of scavenger receptors (e.g. TIM4, MARCO) and lipid (e.g. cholesterol) binding proteins (APOE, APOC1, FABP5), and components of the phagosome, together suggesting a role in clearing microbes, apoptotic cells and oxysterols (Fig.13a). High expression of MERTK and LILRB5, which inhibit TLR/cytokine (Rothlin et al., 2007, Cell 131, 1124-1136) and integrin/FcyR (van der Touw et al., 2017, Cancer Immunol Immunother 66, 1079-1087) driven activation respectively, suggests they may also restrain inflammation (Fig.13a). Interestingly, early undifferentiated arthritis (UPA) showed increased proportions of the MerTK^pos-TREM2^low STM cluster (Fig.2g-i) which is closely related to TREM2^high cluster in terms of transcriptomics and cell trajectory (Fig.2b,e, Fig.11) but have increased oxidative phosphorylation as evidence of a change in metabolism, and activation of the cytoskeleton (Fig.13a) and may represent an early activation phenotype of the protective TREM2^high STMs.

Treatment-naive and treatment-resistant active RA had increased proportions of the MerTK^neg -CD52/SPP1^pos cluster, and treatment-resistant RA additionally had an increased MertK^neg -CD52/S100A12^pos cluster (Fig.2i). Their transcriptomes indicate pro-inflammatory phenotypes e.g. increased expression of glycolytic enzymes ( LDHA , ALDOA, PKM, EN01 ; Fig.13b) indicating that their activation is fuelled by glycolysis. The top marker of the SPP1^pos cluster (osteopontin) has multiple pro-inflammatory and bone-resorbing properties (Kahles et al., 2014, Mol Metab 3, 384-393) and high levels of cytoskeletal proteins and integrins suggesting a migratory phenotype (Fig.13b). This confirms but significantly extends a prior description of SPP1^pos STMs in active RA (Zhang et al., 2019, Nat Immunol 20, 928-942, doi: 10.1038/s41590-019-0378-1). The S100A12^pos STM cluster is a novel finding of importance in active RA because of their abundance and their high expression of inflammation triggering alarmins S100A8/9/12. These are chemoattractant for neutrophils and monocytes and can bind RAGE/TLR4 on fibroblasts and monocytes to induce pro-inflammatory cytokines IL-6 and TNF (Wang et al., 2018, Front Immunol 9, 1298, doi:10.3389/fimmu.2018.01298). We validated the STM signatures of active disease with the Pathobiology of Early Arthritis Cohort (PEAC; synovial biopsy RNA from 90 RA patients (Humby et al., 2019, Ann Rheum Dis 78, 761- 772, doi: 10.1136/annrheumdis-2018-214539; Lewis et al., 2019, Cell Rep 28, 2455- 2470 e2455, doi:10.1016/j.celrep.2019.07.091). This independent analysis confirmed that the expression of the top two markers of the MerTK^neg clusters SPP1 and S100A9 correlates positively with disease activity (Fig.2j). In sustained remission RA, the MerTK^neg-CD52/SPP1^pos cluster was absent, yet the MerTK^neg-CD52/S100A12^pos cluster persists. This may uncover the identity of the persisting MerTK^neg cells in patients in remission who flare (Fig.11). Moreover, in contrast to active RA, and in accordance with our flow cytometry (Fig.1b), RA patients in remission have an increase in MerTK^pos STMs shown by scRNAseq as the FOLR2/LYVE1^pos cluster. Specific expression of LYVE1 suggests they are perivascular tissue macrophages (Lim et al., 2018, Immunity 49, 326-341 e327, doi:10.1016/j.immuni.2018.06.008). Their distinct transcriptome (e.g. BLVRB, HMOX1) suggests heme-degradation and iron homeostasis functions. Also, their transcriptome is selectively enriched in regulators of tissue collagen turnover (e.g. STAB1, TGFBI), (Rantakari et al., 2016, Proc Natl Acad Sci U S A 113, 9298-9303, doi:10.1073/pnas.1604780113; Nacu et al., 2008, J Immunol 180, 5036-5044) antiprotease enzymes (e.g. A2M), coagulation factors (F13A1) and regulators of VEGFR on endothelial cells ( SERPINF1 ) (Fig.13a), together suggesting a role of this cluster in the control of synovial tissue remodelling and homeostasis.

To validate the scRNAseq-based classification of remission compared with active RA STM phenotypes, we used the newly identified MerTK^pos subpopulation markers (TREM2 and FOLR2) in conjunction with MerTK on STMs from additional biopsies from RA patients with active disease and remission using flow-cytometry (Fig.2k-I). This confirmed that TREM2 and FOLR2 were exclusively expressed on MerTK^pos STM. In addition, this analysis revealed that both FOLR2^pos and TREM2^pos subpopulations contribute to an increase in MerTK^pos STMs observed in patients in sustained disease remission (Fig.2m).

Across all clinical states, the MerTK^neg-CLEC10a^pos, MerTK^pos-ID2^pos and MerTK^pos-ICAM1^pos clusters occur in similar proportions. Of interest, the CLEC10a^pos cluster is enriched in antigen presentation pathway genes and DC markers (Villani et al., 2017, Science 356, doi: 10.1126/science.aah4573) and in DC transcription factors (e.g. NR4A3) (Boulet et al., 2019, Proc Natl Acad Sci U S A 116, 15150-15159), strongly suggesting that this population represents synovial tissue resident dendritic cells (Fig.2i). This cluster has a recently described myeloid phenotype expressing HBEGF, EREG and PLAUR (Kuo et al., 2019, Sci Transl Med 11, doi: 10.1126/scitranslmed.aau8587) that potentially promotes synovial fibroblasts invasiveness in active RA. We confirmed that the expression of FIBEGF in this cluster is increased only in early inflammation compatible with a role in initiating synovitis (Fig. 12b). Together, these comprehensive data systematically map the transcriptomic heterogeneity in MerTK^neg and MerTK^pos STM populations spanning different clinical states, suggesting distinct inflammatory and regulatory functions that may actively contribute to RA synovitis or disease remission. Next, we investigated functional complexity ex vivo.

Figure 2 shows that single cell transcriptomics defines distinct STM subpopulations and phenotypes in different human joint immuno-conditions. Panel (a) shows UMAP visualization of 9 clusters of synovial tissue macrophages identified as a result of analysis of scRNAseq STM data. Each cell is represented by an individual point and colored by cluster identity. Panel (b) shows a heatmap illustrating scaled expression of the top 20 differentially expressed genes per cluster (rows are genes, columns represent cells). Marker genes of interest are annotated, and the total number of genes differentially expressed by each cluster is also shown next to the cluster presented on the heatmap. All genes are expressed in at least 40% of cells in each cluster. Average log-fold change >0.25. Panel (c) shows the expression of clusters’ marker genes is shown on log-normalized violin plots; median highlighted by black dot. Panel (d) shows the top 50 PCs were embedded in the top 3 diffusion-map components to demonstrate the transcriptional relationship between the clusters. Panel (e) shows hierarchical clustering of identified STMs visualized as a dendrogram. Panel (f) shows expression of MerTK and CD 163 in the 9 STM clusters. Panel (g) shows a proposed classification of human STMs based on scRNAseq and flow cytometry data. Panel (h) shows a split UMAP visualization demonstrating relative changes in the STM phenotypes between disease groups. The numbers were normalized to ~5,000 cells per group. Panel (i) shows bar and dot plots illustrating the change in cluster distribution across conditions. Statistically significant differences between a given condition and at least one other condition are highlighted with * (two-way ANOVA, corrected for multiple comparisons with Tukey test). Each dot represents individual patient. Panel (j) shows expression of S100A9 and SPP1 in the synovium of the validation PEAC cohort correlates with disease activity. Panels (k-m) show an illustration of TREM2^pos and FOLR2^pos subsets on UMAP (k-l), and flow cytometry validation of TREM2^pos and FOLR2^pos STM clusters (m) in active RA and RA in remission. Example 3: MerTK^neg STMs produce pro-inflammatory cytokines and alarmins while

MerTK^pos STMs produce inflammation-resolving mediators

To evaluate the functions of MerTK^pos and MerTK^neg STM populations, we tested their response to components of the inflammatory synovial microenvironment, and also compared their functions in patients in remission with those with active RA. We FACS- sorted MerTK/CD206^pos and MerTK/CD206^neg STMs from patients with active RA, and MerTK/CD206^pos STMs from RA patients in remission and compared their response to ex vivo stimulation with LPS (as a surrogate for danger-signal ligand binding TLR4) and/or GAS6 (endogenous ligand for suppressive MerTK).

Culture supernatants were assayed for multiple cytokines, chemokines and resolving mediators (Fig.3a). MerTK^neg and MerTK^pos STMs differed significantly in their responses to TLR4-stimulation (Fig.3b-d). MerTK^neg STMs produced significantly more IL-6, TNF, IL-1 β, CCL2 and CCL3 than MerTK^pos STM from either active RA or RA in remission. After stimulation, both populations produced comparable concentration of tissue remodelling cytokines (IL-13), and cytokines involved in the immune response against pathogens (e.g. IL-8, IL-12p70, IFNa2). In contrast to MerTK^pos STMs from active RA, the MerTK^pos STMs from RA in remission did not spontaneously produce any of the pro-inflammatory mediators tested, and the concentrations of cytokines produced per cell upon stimulation were lower than from STMs from active RA. Interestingly, following LPS stimulation, IL-10 was produced in similar concentrations by all STM populations regardless of clinical state. In contrast to pro-inflammatory cytokines, the inflammation resolving lipid mediator resolvin D1 was released only by MerTK^pos STMs, and the concentrations were strikingly higher in culture supernatants of STMs from disease remission (Fig.3d). This differential production of resolvins is consistent with the transcriptomic profile of the MerTK/TREM2^high STM cluster which is enriched in pathways involved in the production of lipid mediators (Fig.13a). This, together with the high ratio of IL-10 to TNF produced by MerTK^pos STMs, suggests that their functions include preventing excessive synovial inflammation to favour resolution.

As expected, there was no immunomodulatory effect of GAS6 on its receptor- negative MerTK^neg STM population. Interestingly, there was no influence of GAS6 on the LPS-induced cytokine production by MerTK^pos STMs from patients with active RA. Rather, GAS6 reduced further the low production of LPS-induced pro-inflammatory cytokines, especially IL-6, by MerTK^pos STMs from RA patients in remission. This difference in response to GAS6 could be attributable to the lower surface density of MerTK that we identified by FACS on MerTK^pos STMs from active RA (Fig.3e). Together, these data indicate that MerTK^neg STMs have an inflammatory phenotype, whereas MerTK^pos STMs, in particular those from RA patients in remission, have a resolving phenotype, and utilize a GAS6/MerTK negative-feedback regulatory loop to attenuate the response to inflammatory stimuli.

Our scRNAseq profiling also revealed the identity of MerTK^neg STMs that can persist in RA patients in remission. This MerTK^neg -CD52/S100A12^pos cluster expressing alarm ins (Fig.2i) may contribute to flares of arthritis. To test whether the MerTK^neg STMs release alarmins, and to compare the production of alarmins by MerTK^neg STMs from RA patients in remission with those with active RA, we investigated S100A12 in culture supernatants from LPS-stimulated MerTK^neg and MerTK^pos STMs FACS-sorted from biopsies of RA patients with active RA and RA in remission. As expected, MerTK^pos STMs from RA patients in remission produced negligible concentrations (2.1±1 4pg/ml) of S100A12. In contrast, MerTK^neg STMs produced high levels (155±43pg/ml) that are similar to those produced by MerTK^neg STMs from active RA (176±39pg/ml). Consistent with this, the transcriptomic analysis of the MerTK^neg -CD52/S100A12^pos STM cluster confirmed high expression of S100A12, 8 and 9 in remission RA equivalent to that of RA patients with treatment-naive active RA (Fig.3f-h). Thus, the MerTK^neg -CD52/S100A12^pos cluster, when present in patients in remission, has the potential to produce alarmins and initiate inflammation and flare, with the same potency as STMs from patients with active RA.

In summary, MerTK^neg and MerTK^pos STM populations have distinct pro- inflammatory and resolving properties, respectively. MerTK^neg STMs from RA patients in remission can produce pro-inflammatory alarmins upon stimulation and may contribute to flare of arthritis upon treatment modification if not counterbalanced by the governing functions of MerTK^pos STM.

Figure 3 shows that MerTK/CD206^pos and MerTK/CD206^neg STM populations have distinct pro- and anti-inflammatory phenotypes, respectively. Panel (a) shows the sorting strategy for MerTK positive and negative STMs (n=47 patients), and overview of functional studies. In vitro production by MerTK positive and negative STMs of (b-c) pro- and anti-inflammatory mediators, and (d) resolvin D1. Panel (e) shows expression levels of MerTK on MerTK/CD206^pos STMs are reduced in active RA. Panel (f) shows visualisation of S100A12 mRNA expression on STM UMAP. Panel (g) shows production of S100A12 by LPS-stimulated MerTK/CD206^neg and MerTK/CD206^pos STMs, FACS- sorted from biopsies of patients with active RA or RA in remission. Panel (h) shows soluble mediators differentially expressed in the MerTK^neg -CD52^pos/S100A12^pos cluster between conditions are shown (p-value < 0.05, adjusted by Bonferroni Correction). regulatory signature

To investigate the molecular signature underlying the resolving phenotype of the MerTK^pos clusters in remission, namely TREM2^|0W, TREM2^high and FOLR2^high/LYVE^pos we compared their transcriptomes between states. In particular we sought their transcriptomic trajectory from health, through joint inflammation to resolution. For example, these evolving programs included pathways induced during RA inflammation that resolve in disease remission (e.g. glycolysis: ALDOA, EN01), or pathways inhibited by inflammation that return to levels similar to those of healthy STMs (e.g. scavenger receptor MARCO, and the leukotriene and resolvin regulator ALOX5AP (Fig. 14). Flowever, some programs underwent transcriptomic changes during inflammation but do not return to normal levels in remission. These include sustained upregulation in the antigen presentation pathway and sustained repression of the regulatory signature typical of healthy STMs (e.g. retinoic acid pathway (Nolting et al., 2009, J Exp Med 206, 2131-2139) and a B7-relate co-inhibitory molecule VSIG4 (Vogt et al., 2006, J Clin Invest 116, 2817-2826)), which suggests long-term epigenetic imprinting triggered by inflammation that does not resolve (Fig.4a-b; Fig.14).

Interestingly, the MerTK^pos clusters: TREM2^pos, TREM2^high (and to a lesser extent FOLR2/LYVE1^pos) were found in remission to have a unique regulatory transcriptomic signature that is different from the regulatory transcriptomic signature of healthy STMs (Fig.4a and c). This signature is characterized by upregulation of transcription factors (KLF2, KLF4, NR4A1 , NR4A2 and ATF3) and upregulation of dual- specificity phosphatase 1 (DUSP1 ). These emerged in the top 30 upregulated genes in each of the MerTK^pos clusters in remission. Murine studies suggest that these transcription factors and DUSP1 are negative regulators of inflammation that can reinstate tissue homeostasis. Specifically, DUSP1 drives destabilization of pro- inflammatory mRNA transcripts (Smallie et al., 2015, J Immunol 195, 277-288) and lack of DUSP1 increases susceptibility to experimental arthritis (Vattakuzhi et al., 2012, Arthritis Rheum 64, 2201-2210, doi:10.1002/art.34403). KLF2 and KLF4 coordinate the expression of receptors that recognize and remove apoptotic cell (e.g. MARCO, TIM4) and inhibitors (e.g. SOCSs, A20) that limit the responses to intracellular TLR ligands (Roberts et al., 2017, Immunity 47, 913-927 e916, doi:10.1016/j.immuni.2017.10.006) while ATF3 inhibits type I interferon production induced by those ligands (Labzin et al., 2015, J Immunol 195, 4446-4455). NR4A1 and NR4A2 coordinate a metabolic switch from pathological glycolysis to homeostatic oxidative phosphorylation and trans-repress NFKB to limit the pro-inflammatory response to extracellular danger signals, respectively (Koenis et al., 2018, Cell Rep 24, 2127-2140 e2127, doi:10.1016/j.celrep.2018.07.065; Hanna et al., 2012, Circ Res 110, 416-427, doi:10.1161/CIRCRESAHA.111.253377; Mahajan et al., 2015, J Biol Chem 290, 18304-18314, doi: 10.1074/jbc.M115.638064). We confirmed that this remission-specific transcriptomic signature is linked to upstream activation of MerTK by demonstrating in vitro that their expression was reduced by a MerTK inhibitor (Fig 4d). A subsequent analysis of this signature in RA synovium using the independent PEAC cohort (Humby et al., 2019, Ann Rheum Dis 78, 761-772, doi: 10.1136/annrheumdis-2018-214539) confirmed that KLF4 and NR4A2 expression in RA synovium correlates negatively with disease activity (Fig.4e).

In summary, these data clearly indicate that MerTK^pos STM clusters in RA disease remission have regulatory functions characterised by a unique set of transcription factors.

Figure 4 shows TREM2^pos and FOLR2^pos clusters of MerTK^pos STMs from RA patients in remission have a unique transcriptomic signature. The content of panels (a)- (e) is described below (a) Heatmap illustrating scaled expression of the top 30 marker genes of each condition within the TREM2^low, TREM2^high and FOLR2/LYVE1^pos clusters. Rows are genes while columns show equal pseudo-bulk expression per condition within each cluster. All genes are expressed in at least 60% of cells in that condition with an average log fold-change >0.25 and p< 0.05 after correction for multiple comparisons (b) Table demonstrating the number of genes differentially expressed between conditions for each cluster (c) Violin plots illustrating the expression of top marker genes of TREM2^pos and FOLR2^pos clusters unique for healthy or remission STMs (d) Transcription factors identified in TREM2^pos and FOLR2^pos clusters in remission are inhibited in macrophages incubated with MerTK inhibitor. * p< 0.05 paired t-test. (e) Expression of KLF4 and NRA42 in different synovial tissue of PEAC cohort negatively correlates with DAS28. Example 5: Macrophage MerTK controls activation of synovial fibroblasts

We next tested whether MerTK-expressing macrophages can modulate the pro- inflammatory and destructive properties of the synovial stromal compartment. We generated direct-contact and trans-well macrophage-fibroblast co-cultures to mimic synovial myeloid-FLS (synovial fibroblast-like synoviocytes; FLS) interactions. We employed surrogate macrophage populations with different levels of membrane MerTK by culturing monocytes derived macrophages (/) with LPS (to mimic STMs with low MerTK) and (i7) with LPS plus (Dex)amethasone (to mimic STMs with high MerTK) (Fig.5a and Fig.15a). In addition, to inhibit MerTK function in these macrophages, some cells were also pre-incubated with a MerTK-selective inhibitor (Liu et al., 2013, Eur J Med Chem 65, 83-93). After direct co-culture, the FLS and macrophages were separated by FACS-sorting using cell specific markers and FLS specific expression of IL-6 and MMP-1 evaluated by qPCR (Fig.5b). Co-culturing FLS with macrophages pre- treated with LPS increased FLS expression of MMP1 and IL-6. Pre-treatment of these inflammatory macrophages with a MerTK inhibitor further increased FLS expression of MMP1 and IL-6. Co-culture of FLS with macrophages pre-treated with LPS plus Dex did not affect the low constitutive expression of IL-6 and MMP-1 in FLS. However, inhibition of MerTK in these macrophages increased IL-6 and MMP-1 expression in FLS showing that MerTK was at least partially responsible for their homeostatic phenotype (Fig.5b).

Commensurate with these observations, increased concentrations of MMP1 , MMP3 and IL-6 were detected in supernatants of co-cultures of FLS with macrophages that were pre-incubated with MerTK inhibitor as compared to the respective controls (Fig.5c-d). Other soluble mediators e.g. MMP2 and macrophage-derived MMP9 were not affected by MerTK inhibition (Fig.15c). Parallel experiments using the trans-well system (Fig.5e-f and Fig.15d) showed that direct contact between macrophages and FLS was not required for macrophage MerTK modulation of fibroblast production of MMP1 , MMP3 and IL-6. Taken together, these data suggest that MerTK-expressing macrophages limit the proinflammatory (cytokine) and tissue-destructive properties (MMP) of synovial fibroblasts.

To further explore the pathogenic phenotype(s) of FLS regulated via macrophage MerTK expression, we compared the transcriptome of FLS co-cultured with LPS pre- treated macrophages with FLS co-cultured with macrophages pre-treated with LPS and MerTK inhibitor. Eighty-two differentially expressed genes were identified enriched in two pathways. String pathway analysis highlighted the cytokine pathway (15 of 216 pathway genes, p=4.10¹⁵) and the chemokine pathway (9 of 48 pathway genes, p=1 .79 ¹²). In particular, the lack of MerTK-mediated inhibition of macrophages permits increased FLS-expression of numerous chemokines that recruit neutrophils (CXCL8, CXCL1 , CXCL2, CXCL5, GCF3), monocytes (CCL3, CXCL3), T-cells (CCL20, CXCL10), and increased expression of COX2 (PTGS2) generates PGE2 that mediates pain and blood-vessel permeability. In contrast, repair mediators (FGF14) and extracellular matrix (COL21a) were downregulated (Fig.5g-h). Together these data indicate that MerTK^pos macrophages may re-instate joint immune homeostasis by actively limiting the inflammatory and enhancing tissue remodelling functions of FLS.

Figure 5 shows that MerTK positive macrophages control inflammatory response of synovial fibroblasts. The content of panels (a)-(g) is described below (a) Schematic of a direct co-culture system of macrophages (MQ) with synovial fibroblasts (FLS). (b) mRNA expression levels of IL-6 and MMP-1 in FLS (FACS-sorted from co-culture with macrophages). MerTK inhibition of LPS and of LPS+Dex pre-treated macrophages enhanced their activation of FLS. (c-d) Levels of mediators in co-culture supernatants show that MerTK inhibition in macrophages enhanced the production of MMP1, MMP3 and IL-6 by FLS. (e) Schematic of an indirect co-culture system of macrophages with synovial fibroblasts (FLS). (f) MerTK inhibition of LPS pretreated macrophages increased the concentrations of MMP1, MMP3 and IL-6 in co-culture supernatants. Data are presented as dot-plot of individual experiments (n=5) with a bar indicating the mean. * and & represent p<0.05 in one-way Anova, followed by correction for multiple comparisons, or a paired t-test if two groups were compared (g) Heatmap demonstrating scaled log-expression of soluble mediators that were differentially expressed between FLS co-cultured with macrophages pre-stimulated with LPS and MerTK inhibitor, compared with FLS co-cultured with macrophages pre-treated with LPS only. Genes were adjusted for multiple comparison (p<0.01 and log2-fold change > +/- 1.5).

Example 6: Lining layer synovial fibroblasts in sustained disease remission show a decrease in mediators regulated by MerTK expressing macrophages

Finally, to investigate whether lining-layer FLS of patients in remission have a transcriptome indicative of local interaction with MerTK^pos STMs and to identify the potential local stomal source of MerTK ligand GAS6, we performed scRNA sequencing (lOxGenomics) of synovial fibroblasts from RA patients in sustained remission and from patients with active RA. Unsupervised clustering of 13,949 FLS confirmed the existing classification of FLS (Croft et al., 2019, Nature 570, 246-251; Zhang et al., 2019, Nat Immunol 20, 928-942, doi: 10.1038/s41590-019-0378-1 ; Mizoguchi et al., 2018, Nat Commun 9, 789, doi: 10.1038/s41467-018-02892 -y) that distinguished lining-layer FLS clusters expressing MMPs, and 4 sublining-layer clusters; HLA^high, THY1^high, THY1/CXCL14^pos and THY1/CD34^pos expressing collagens and immune-mediators (Fig.6a-e).

There were no differences in the relative proportion of these clusters between active RA and RA in sustained remission, however, crucially, their transcriptome differed. In particular, we observed decreased expression of mediators previously identified as regulated by macrophage MerTK (Fig.5). These include metalloproteinases (MMP1, MMP3) and chemokines ( CXCL1 , CXCL8). In contrast, mediators of tissue repair and resolution (e.g. LTBP4, IGFBP5/6 and AXL) were increased. Thus, lining layer FLS of RA patients in disease remission show a transcriptomic signature of cell with a ‘resolved’ phenotype potentially from interaction with MerTK^pos STMs.

We also sought the potential source of the MerTK ligand GAS6 in the synovium. Our single-cell transcriptome analysis of FLS revealed that sublining FLS clusters express GAS6 mRNA. In particular, a small THY1/CXCL14^pos cluster showed abundant expression of GAS6, although this did not differ between conditions (Fig.6e). However, an increase in GAS6 was observed in THY1^high cluster of patients in disease remission as compared to active RA, suggesting a potential increase in GAS6 levels in specific tissue niches of resolving synovium (Fig.6i). In vitro cultures of FLS lines derived from biopsies of patients with active RA and RA patient in sustained remission confirmed that FLS release significant amounts of GAS6. The production of GAS6 by these lines was further enhanced upon dexamethasone exposure (Fig.6j). Thus, GAS6 derived from sublining FLS likely contributes to the homeostatic regulatory functions of lining layer MerTK^pos STMs.

Figure 6 shows that lining layer FLS of RA patients in remission express reduced levels of MMPs and chemokines compared to active RA while sublining FLS are source of GAS6. The content of panels (a)-(j) is described below (a) UMAP visualization of heterogeneity of synovial fibroblasts in active RA (n=4) and RA in remission (n=3). Each cell (n=13,949) is visualized by an individual point and colored by cluster identity (b) Stacked bar plots demonstrate the relative proportion of cells within each cluster per clinical group (c) Bar plots illustrate the change in cluster distribution between clinical groups. Each point represents an individual sample (d) Heatmap illustrates scaled, batch-normalized expression of the top 10 differentially expressed genes per cluster, showing that THY1+ sublining clusters express GAS6 particularly in cluster 5 (CXCL14^pos). Rows are genes and columns represent cells. All genes are expressed in at least 40% of cells in a given cluster. Average log-fold change > 0.25. (e) Violin plots showing the distribution of log-normalized expression of genes of interest. The median value is represented by a black dot. (f) Differentially expressed genes in the lining layer FLS of remission RA as compared to active RA. Scatter-plot illustrates pseudo-bulk expression of each differentially expressed gene and the top 10 genes for each condition are annotated. Heatmap shows the scaled, sample-specific pseudo-bulk expression of the top 20 differentially expressed genes, rows are genes and columns represent-sample pseudo-bulk expression. All genes are expressed in at least 60% of cells in that group. Average log-fold change > 0.25 with p< 0.05 after correction for multiple comparison, (g-h) Violin plots showing log-normalized expression of genes for MMPs, chemokines and anti-inflammatory mediators in lining layer FLS. The median value is represented by white dot. (i) mRNA expression of GAS6 in distinct sublining FLS clusters comparing active and remission RA. The THY1^high cluster shows greater expression of GAS6 in remission as compared to active disease. Median values are represented by white dots (j) GAS6 is produced by cultured FLS derived from biopsies of RA patients; treatment naive, treatment resistant, and RA in remission. Each dot represents a separate FLS line (n=5 per condition) *p < 0.05, two-way Anova with correction for multiple comparisons.

Example 7: Discussion

This study provides the first comprehensive, comparative description of the functional biology of human synovial tissue macrophages in health, active RA synovitis and sustained RA disease remission. Our data define the presence, and functional consequences of discrete human macrophage subsets, and inform future biomarker and therapeutic target opportunities in RA and likely wider immune-mediated inflammatory diseases. Notably, we provide a novel cellular and molecular explanation suggesting that RA remission as an actively restrained state.

We combine integrated analysis of scRNAseq of >32,000 STMs from 25 synovial biopsies, FACS-phenotyping of STMs from 112 biopsies, and functional analysis of STMs from 47 biopsies. This spanned healthy homeostasis and early undifferentiated joint inflammation (UPA), treatment-naive early RA, treatment-resistant RA and RA in sustained remission. Multiparameter flow cytometry showed that STMs consisted of two main populations; positive and negative for MerTK/CD206. ScRNAseq confirmed that these populations are positioned on distinct cell trajectory branches, and uncovered deep phenotypic and functional heterogeneity in both, revealing RA stage-specific mechanisms of pathogenesis and of remission. The MerTK/CD206^pos STMs are dominant in healthy tissue and in RA in disease remission, whereas MerTK/CD206^neg STMs are enriched in active RA. Crucially, their relative proportion in remission is predictive of persistent remission or flare upon drug withdrawal which commensurate with their functional roles. Patients in disease remission whose STMs are composed of less than 47.5% MerTK/CD206^pos or alternatively their MerTK/CD206^pos to MerTK/CD206^neg ratio is less than 2.5 have higher likelihood of flare after treatment cessation. This can be explained by distinct functions of these two populations. The MerTK/CD206^neg STMs produce proinflammatory cytokines and alarm ins. In contrast, the MerTK/CD206^pos cells produce lipid mediators that resolve inflammation and their MerTK pathway restrains activation of the stromal compartment indicating that intercellular crosstalk between MerTK/CD206^pos and synovial fibroblasts during remission maintains joint immune-homeostasis.

We show that regulatory MerTK^pos STMs consist of distinct subsets including MerTK/TREM2^pos and MerTK/FOLR2/LYVE1^pos STMs, both increased in remission while MerTK/TREM2^pos is the dominant subpopulation of MerTK^pos in the healthy synovium. Their transcriptomes suggest distinct (but complementary) functions controlling the local immune-response and tissue homeostasis, respectively. Hierarchical clustering of orthologous human and mouse cluster-specific transcripts (Fig. 16-17) indicates that human TREM2^high STMs are homologs of mouse lining-layer Trem2/Cx3cr1 (Culemann et al., 2019 Nature, doi: 10.1038/s41586-019-1471-1) STMs, and human FOLR2/LYVE1^pos STMs may closely resemble interstitial Relmct^pos STMs. These murine counterparts differentiate from locally proliferating precursors and are key for maintaining immune homeostasis. Furthermore, in remission RA these clusters gain a unique phenotypic and transcriptomic signature that is different from similar cells in active RA and in healthy. We show that this is driven by MerTK activation, potentially by GAS6 produced locally by Thy1^pos synovial fibroblasts, and includes low production of pro-inflammatory cytokines, high production of resolvins and an increased expression of the set of transcription factors KLF2/4, NR4A1/2 and ATF3 that inhibit inflammation.

Of the four pro-inflammatory MertK^neg clusters, we identified that the CD52/S100A12^pos cluster can persist in RA patients in disease remission, and that these cells produce high levels of the S100A12 alarm in upon stimulation - this provides a potential mechanism whereby flare can be initiated. Thus, sustained remission of arthritis appears to be an active process maintained by tissue-resident subsets of MerTK^pos STM subpopulations (e.g. TREM2^pos and FOLR2/LYVE1^pos) governing pro- inflammatory CD52/S100A12^pos STMs and synovial fibroblasts to re-instate and maintain homeostasis. Of interest, recent mouse and human studies have suggested a protective function for TREM2⁺ resident-macrophages in adipose tissues that counteracts inflammation and metabolic deregulation. This highlights a broad regulatory role of TREM2^pos tissue resident macrophages (Jaitin et al., 2019, Cell 178, 686-698).

Of the three MerTK/FOLR2^pos subset, relatively small FOLR2/ID2^pos and FOLR2/ICAM1^pos clusters remained unchanged between different joint immune states. We speculate that FOLR2/ID2^pos may be the human equivalent of mouse M-CSF driven in situ precursors that give rise to mouse RELMa^{pos 8} which is the homolog of human FOLR2. This is supported by their high expression of M-CSF-R, and ID2 which is a key driver of self-renewing haemopoietic stem cells (Freeman et ai, 2015, Blood 126, 2646- 2649)( Fig.13a). We also identified a small population of cycling STMN1^pos STMs that also expressed M-CSF-R, that cluster with TREM2^|0W phenotype, suggesting that they might be precursors of the resolving MerTK/TREM2^pos STMs subset (Fig.16e). Synovial macrophages proliferating in situ was recently reported in patients with inflammatory osteoarthritis (Wood et al., 2019, JCI Insight 4, doi: 10.1172/jci. insight.125325) supporting the concept of self-renewing STMs in the human synovium, and the possibility for therapy directed differentiation of STM phenotypes to reinstate and maintain synovial homeostasis.

Of interest, FOLR2/ICAM1^pos STMs, which constitutes ~0.025% of STMs and constitutively express high levels of mRNA for pro-inflammatory cytokines (e.g.TNF and IL-1 β), chemokines and NFKB, are present in healthy synovium and their frequency did not change in inflammation and disease remission. Their MerTK expression and position on the ontogeny dendrogram suggest that they are a part of the MerTK^pos population. Little is known about these intriguing cells that may form the joint’s first line of defence against pathogens. In summary, we identified dynamic phenotypic changes in synovial tissue macrophage subsets spanning health, RA inflammation and disease resolution and describe for the first time in a human autoimmune disease, active mechanisms mediating sustained disease remission facilitated via tissue resident STMs (summary in Fig.18).

Examole 8: Material and Methods

Example 8a: Patients recruitment and management

One-hundred and twelve patients fulfilling the American College of Rheumatology 2010 revised criteria for RA (Aletaha et al., 2010, Ann Rheum Dis 69, 1580-1588) were enrolled and underwent ultrasound-guided synovial tissue biopsy of the knee at the Division of Rheumatology of Fondazione Policlinico Universitario A. Gemelli IRCCS - Universita Cattolica del Sacro Cuore (SYNGem cohort), Rome, Italy. RA patients were stratified into naive to treatment (n=45), inadequately responder to Methotrexate (Treatment resistant RA) (n=31) and patients in sustained (for minimum 9 months) clinical and US remission under combination of MTX+TNF-inhibitor (n=36). Ten healthy donors attending for arthroscopy for meniscal tear or cruciate ligament damage, with normal synovium by MRI and macroscopically were included as a control group, at the University of Glasgow. The study protocol was approved by the local Ethic Committee of the Universita Cattolica del Sacro Cuore (6334/15) and by the West of Scotland Research Ethical Committee (19/WS/0111). All subjects provided signed informed consent. Demographic, clinical and immunological features of the study RA and healthy cohorts are summarized in Supplementary Table 1- 3 of Alivernini et al. 2020 Nature Medicine 26:1295-1306. doi: 10.1038/s41591 -020-0939-8. All treatment-resistant RA were taking stable doses of MTX (mean dose: 15.3+3.3 mg/week). All RA in sustained clinical (DAS28<2.6 for 3 sequential determinations each 3 months apart) and ultrasound remission (Power Doppler negativity at US assessment for 3 sequential determinations each 3 months apart) were selected based on published protocols (Alivernini et al., 2016, Arthritis Res Ther 18, 39; Alivernini et al., 2017, Ann Rheum Dis, doi: 10.1136/annrheumdis-2016-210424). For each RA patient enrolled, clinical and laboratory evaluations included the number of tender and swollen joints of 28 examined, Erythrocyte Sedimentation Rate (ESR), C-Reactive Protein (CRP) and Disease Activity Score (DAS28). Peripheral blood samples were tested for IgA and IgM-RF (Orgentec Diagnostika, Bouty-UK) and ACPA (Menarini Diagnostics-ltaly) using commercial Enzyme-Linked Immunosorbent Assay (ELISA) and ChemiLuminescence Immunoassay (CLIA) methods respectively. After study enrolment, RA patients in sustained clinical and US remission (n=22) were first tapered on TNF-inhibitor (adalimumab 40 mg/4 weeks or etanercept 50 mg/2 weeks) for 3 months. After 3 months of TNF-inhibitor tapering, patients who were still in US remission (Power-Doppler negative) discontinued TNF-inhibitor and were followed every 3 months while maintained on stable doses of Methotrexate (15.2+2.9 mg/week), with follow-up after treatment modification of 13.2+7.6 months during which there were no treatment modifications². The relapse rate was recorded for each RA patient in sustained clinical and US remission after treatment modification (Alten etal., 2011, J Rheumatol 38, 1745-1750).

Example 8b: Patients selection for single-cell RNA sequencing

Seventeen patients fulfilling the American College of Rheumatology 2010 revised criteria for RA (Aletaha et al., 2010, Ann Rheum Dis 69, 1580-1588)(5 treatment-naive RA, 6 treatment-resistant RA and 6 RA patients in sustained clinical and US remission, respectively) and 4 patients with Undifferentiated Peripheral Arthritis (UPA) (Machado et al., 2011, Ann Rheum Dis 70, 15-24, doi: 10.1136/ard.2010.130625) with at least one active knee joint, seronegative for IgA/lgM-Rheumatoid Factor (RF) and Anti- Citrullinated Peptide Antibody (ACPA) and naive to any pharmacological treatment were enrolled in the study at the Division of Rheumatology of Fondazione Policlinico Universitario A. Gemelli IRCCS - Universita Cattolica del Sacro Cuore, Rome, Italy. For each RA and UPA patient enrolled, clinical and laboratory evaluations included the number of tender and swollen joints on 28, Erythrocyte Sedimentation Rate (ESR), C- Reactive Protein (CRP) and Disease Activity Score-28 (DAS). Peripheral blood samples were tested for IgA and IgM-Rheumatoid Factor (RF) (Orgentec Diagnostika, Bouty-UK) and ACPA (Menarini Diagnostics-ltaly) using commercial Enzyme-Linked Immunosorbent Assay (ELISA) and ChemiLuminescence Immunoassay (CLIA) methods respectively. Each enrolled patient underwent US-guided synovial tissue biopsy and synovial tissue samples were processed following the protocol described immediately below. Four healthy donor tissues were included as control. Demographic, clinical and immunological features of patients and healthy donors’ samples used in scRNAseq are summarized in Supplementary Table 2-3 of Alivernini et al. 2020 Nature Medicine 26:1295-1306. doi: 10.1038/s41591 -020-0939-8.

Example 8c: Synovial tissue biopsies

All RA and UPA arthritis patients enrolled underwent ultrasound-guided synovial tissue biopsy of the knee following the already published protocol (Alivernini et al., 2016, Nat Commun 7, 12970, doi: 10.1038/ncomms12970) at the Division of Rheumatology of Fondazione Policlinico Universitario A. Gemelli IRCCS - Universita Cattolica del Sacro Cuore (SYNGem cohort), Rome, Italy. All patients underwent ultrasound evaluation of the knee using an ultrasound machine with a multi- frequency linear transducer (MyLab Twice, Esaote). Using the ultrasound view, the best point of entrance for the needle was identified on the lateral margin of the suprapatellar recess. Each patient was provided with a face-mask and cap and the whole procedure was under sterile conditions. Skin disinfection was done with iodine solution (performed twice, starting from the point of needle entrance up to 25 cm proximally and distally). If joint effusion was present arthrocentesis of the knee joint was performed using the lateral suprapatellar access. The skin, subcutaneous tissue and joint capsule was anaesthetized with 10 ml 2% lidocaine. Next, a 14G needle (Precisa 1410-HS Hospital Service Spa, Italy) was inserted into the joint. Regions of synovial hypertrophy were identified under grey-scale guidance to ensure sampling of representative synovial tissue. All synovial tissue specimens obtained (at least eight pieces for histology and twelve pieces for singe-cell RNA-sequencing and functional experiments) were placed on a nonwoven wet gauze for collection. For Histology, tissue specimens were fixed in 10% neutral-buffered formalin and embedded in paraffin. Briefly, paraffin-embedded synovial tissue specimens were sectioned at 3-4 pm. Sections were stained for Haematoxylin and Eosin as follows: sections were deparaffinized in xylene and rehydrated in a series of graded ethanol then stained in haematoxylin and counterstained in Eosin/Phloxine. Finally, sections were dehydrated, cleared in xylene and mounted with Bio Mount (Bio- Optica). Slides were examined using a light microscope (Leica DM 2000). The severity of synovitis was graded according to the three synovial membrane features (synovial lining cell layer, stromal cell density and inflammatory infiltrate), each ranked on a scale from none (0), slight (1) and moderate (2) to strong (3). The values of the parameters were summed and interpreted as follows: 0-1, no synovitis; 2-4, low-grade synovitis; and 5-9, high-grade synovitis (Krenn et al., 2006, Histopathology 49, 358-364, doi: 10.1111 /j.1365-2559.2006.02508.x). For Immunohistochemistiy, sections were stained with lgG2a mouse anti-human monoclonal antibody for CD68 (clone 514H12; antibody at 6.7mg/ml) (Leica Biosystem, Newcastle, UK) or IgG rabbit anti-human monoclonal antibody for MerTK (clone Y323, Abeam ab205718, dilution 1/1000) by immunostainer BOND MAX III (Leica, Newcastle, UK). Single immunohistochemical staining for CD68 or MerTK was performed as follows: 3 Dm sections were prepared from formalin-fixed paraffin-embedded tissue blocks and dried in a 60°C oven for 30 min. Sections were placed in a Bond Max Automated Immunohistochemistry Vision Biosystem (Leica Microsystems GmbH, Wetzlar, Germany) according to the following protocol: firstly, tissues were deparaffinized and pre-treated with the Epitope Retrieval Solution 1 (CITRATE buffer) or Solution 2 (EDTA-buffer) at 98°C for 10min according to the manufacturer’s instructions. After washing, peroxidase blocking was carried out for 10m in using the Bond Polymer Refine Detection Kit DC9800 (Leica Microsystems GmbH). Tissues were again washed and incubated with the primary antibody for 30min then incubated with polymer for 10min, developed with DAB-Chromogen and finally counterstained with hematoxylin. Slides were examined using a light microscope (Leica DM 2000). For Immunofluorescence, formalin-fixed RA synovial tissues were microwaved in a citric acid-pH 9.2 and pre-incubated with phosphate-buffered saline 10% bovine serum albumin (BSA) for 30 min. Sections were then stained with a primary antibody against CD68 (clone L26 mouse anti-human monoclonal antibody, at 1.2mg/ml, Leica Biosystem, Newcastle, UK), and anti-MerTK (rabbit IgG polyclonal Cy3- coniugated antibody anti-human MerTK, clone 5770, BIOSS bs-0182R-Cy3, dilution 1/100) at 37°C for 1h. Sections were rinsed and incubated with secondary conjugated antibody (Fluorescein isothiocyanate (FITC) conjugated goat anti-mouse IgG H&L, #ab6785, (Abeam) (dilution 1/1000) at RT for 1h. Slides were mounted and scanned on a fluorescent microscope (Nikon).

Example 8d: Synovial tissue processing for synovial tissue macrophage phenotyping, subset FACS-sorting and scRNA-sequencing Fresh synovial tissues were diced to 1-2 mm³ fragments with sterile disposable no.22 scalpels and transferred into a sterile universal container containing 10ml sterile RPMI with Penicillin/Streptomycin 100/U/ml and L-Glutamine 2mM (RPMI medium) in 1/33 dilution of Liberase at 0.15mg/ml, 0.78 Wunsch units/ml [TM Research Grade (Thermolysin, Medium, Roche Diagnostics (000000005401127001, Sigma)]. Tissue pieces were incubated at 37^°C, 5% CO2 in a humidified atmosphere 30-45min rotating on a Miltenyi MACSmix tube-rotator and shaken vigorously by hand twice during this incubation. After incubation, the digested mixture was filtered using an Easy Strain 100pM cell-strainer into a 50ml falcon tube. Residual cell clumps retained on the filter were gently massaged using the rubber end of a 1 ml syringe plunger to optimise cell retrieval. Complete medium (RPMI above plus 10%FCS) was poured through the filter into the falcon tube up to 40ml then centrifuged 1800 RPM for 10min at 4°C and the supernatant was carefully removed. One ml of complete medium was added to gently resuspend the cell pellet using a wide opening 1ml pipette tip to minimize cell damage, then the resuspended cells were transferred to a sterile Eppendorf tube. Twenty ul aliquot was used to count the cells. Then cells were centrifuged at 1500rpm for 5m in at 4°C. The supernatant was removed, and cells were either aliquoted for STM phenotyping and/or STM FACS-sorting as described below, or for subsequent scRNA- sequencing (cells from 25 patients/healthy donors described above) cells were added to 1ml of ice-cold freezing mix [Bambanker (302-14681 ; Wako)], immediately frozen at - 80°C then stored in liquid nitrogen.

Example 8e: Phenotyping and FACS-sorting of STM subsets

Digested biopsies were centrifuged at 1800rpm for 10min, resuspended and washed with FACS buffer, and transferred to a FACS tubes (BD Biosciences) in a final volume of 3ml FACS buffer (PBS/2%FSC/2mMEDTA). An 80 pi aliquot was taken aside for setting up live-dead gates (unstained cells). To the rest of the cells, Fixable Viability Dye eFIuor™ 780 (eBioscience) was added at 1 :1000 in PBS and incubated for 20min at 4^°C. Cells were then washed with FACS buffer. Four tubes were labelled: a) unstained b) the live-dead marker only c) Fluorescence Minus One Control (FMO) tube, FMO minus FITC, where cells were stained with antibodies specific for STM but not FITC- antibodies against all other lineage-positive cells; d) cells stained with antibodies against STMs and FITC-antibodies against any other lineage-positive. Staining was performed in a final volume of 500mI with antibody dilution 1/100 for 30min on ice. All the antibodies are listed in Fig.7a. Cells were washed twice with FACS buffer and resuspended in finial volume of 500mI, filtered through and Easy Strain 100μM cell- strainer and analyzed or sorted with the use of FACS ARIA III sorter (BD Bioscience). Synovial tissue macrophages were gated based on their membrane expression of CD45, CD64, CD11b, and FILA-DR after all other cell lineages (FMO-FITC gating) and cell-doublets were excluded (Dump channel). FMO-FITC cells were used to set up a gate for the exclusion of lineage positive cells (dump channel). The expression of MerTK, CD163, CD206, TREM2, FOLR2 and TIM4 were evaluated on gated CD64^posCD11 b^posHLA-DR^pos STMs (Fig. 7). In addition, MerTK/CD206^pos and MerTK/CD206^neg STM population were FACS-sorted from 47 synovial biopsies. The cells were sorted into FACS tubes containing 2ml of complete RPMI1640. Post-sorting purity of macrophages was performed, and all data generated were analyzed using FlowJo software (Tree Star Inc, OR, USA).

Example 8f: Ex-vivo stimulation of sorted STMs

MerTK/CD206^pos and MerTK/CD206^neg STM were FACS-sorted into complete medium and plated in 96-well of a flat-bottom cell-culture plate, pre-coated with collagen (Sigma; bovine collagen at 1 :300 dilution). The precoating protocol was as follow: wells were incubated with collagen at 37^°C, 5% CO2 for 2h and then washed twice with PBS. STMs were seeded at 1000 cells/well and stimulated with LPS (10ng/ml, Sigma, L6529) or human recombinant Gas 6 (100 ng/ml, R&D Systems, 885-GSB-050), or both in combination or left unstimulated for 24h in total volume of The supernatants were then harvested and assayed using an ultra-sensitive 19-plex assay (Meso Scale Discovery, Maryland, USA), Resolvin D1 (Cayman Chemical, 500380) and S10012A (DY 1052-05 R&D Systems).

Example 8g: Co-culture of macrophages with synovial fibroblasts

Direct co-culture

CD14^pos cells were isolated from PBMC using CD14^pos micro-beads and AutoMACSPro (Miltenyi BioTec) according to the manufacturer’s protocol. These were differentiated to monocyte-derived macrophages in complete medium containing M-CSF. Briefly, cells were plated at a density of 1x10⁶ per well in a 6-well cell-culture plate in 3ml of RPMI 1640 compete medium containing M-CSF (PeproTech, UK) at 50ng/ml. On day 3, the medium was replaced with fresh medium containing M-CSF. On day 6, cells were pre- treated with LPS (1 ng/ml) or dexamethasone (1 μM) or both, in the presence of absence of MerTK inhibitor UNC1062 (Liu et al., 2013, Eur J Med Chem 65, 83-93)(Aobious; 250 μM). After 24h, macrophages were de-attached and labelled with CellTrace Far Red (5 μM, Life Technologies) according to the manufacturer’s protocol. These cells were added at 2x10³ per 1 well of 96 well plate containing 2x10³ primary fibroblast-like synoviocyte (FLS). The fibroblasts were obtained from US-guided synovial tissue biopsies (Supplementary Table 6 of Alivernini et al. 2020 Nature Medicine 26:1295- 1306. doi: 10.1038/s41591 -020-0939-8) and had been labelled with CellTracer Violet (5 pM, Life Technologies) 24h before the co-culture with macrophages. After 24 or 48h co- culture, culture supernatant was collected for assay of mediators, and macrophages and synovial fibroblasts were de-attached and stained with antibodies against the synovial fibroblasts’ marker podoplanin and the macrophages marker CD64 (both at 1/100 dilution, details in Supp Fig.1a. Fibroblasts and macrophages were FACS-sorted into RLT buffer (Qiagen) containing 1 % β-mercaptoethanol based on their specific CellTracer staining and cell type specific markers and stored at -80°C for RNA isolation.

Trans-well system

CD14^pos monocytes were plated in a 24-well plate in 3ml complete medium contained M-CSF (PeproTech) at the concentration of 50ng/ml. On day 3, some cells were pre- treated with LPS (1ng/ml) for 4h. For the last 2h, MerTK specific inhibitor, UNC1062 (Liu et al., 2013, Eur J Med Chem 65, 83-93) (Aobious) was added at the concentration of 100 or 250 μM. Cells were then washed with PBS and Transwell inserts (0.4pm pore size) containing 3x10⁵ RA synovial fibroblasts was added to the wells to generate a co- culture system to test the effect of soluble mediators without the direct cell contact. After 48h, supernatants were collected while the cells in the macrophage and fibroblast compartments were lysed separately in RLT buffer with 1% β-mercaptoethanol (Qiagen) and stored at -80°C till the RNA isolation step. The MMP Luminex panel (PPX-05/ PROCARTAPLEX MMP1 , MMP2, MMP3, MMP9 and MMP13 plex) and IL-6 Elisa (both from Life Technologies) was performed on supernatant from direct and trans-well co- cultures.

FLS were derived from biopsies of RA patients’ treatment-naive, treatment-resistant and in sustained disease remission (Supplementary Table 7 of Alivernini et al. 2020 Nature Medicine 26:1295-1306. doi: 10.1038/s41591 -020-0939-8). FLS were expanded in complete RPMI1640 medium supplemented with 2mM Glutamax, 1mM sodium pyruvate and 1% non-essential amino acid (Life Technologies). FLS at passage 2-3 were then seeded on 48-wells cell culture plates at a density of 30x10³ cells/well in the complete medium containing 1% FCS. Cells were stimulated with dexamethasone 1μM or TNF or IL-1β or IL-10 or TGFβ or LPS at the concentration of either 10 or 100 ng/ml for 24h and 48h. GAS6 was quantified in culture supernatants using the Human GAS6 DuoSet ELISA kit (R&D Systems, Catalog # DY885B).

Example 8i: qPCR for MMPs, IL-6, GAS6 and transcription factors

RNA from macrophages and synovial fibroblasts was isolated using RNEasy micro-kit (Qiagen), and cDNA was prepared using a High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). TaqMan mRNA primers/probe assays and TaQman Gene Expression master mixes (both from Life Technologies) were used with for semi- quantitative determination of the genes of interest. Data is presented as relative value (/) 2^-ΔCT where ACt=Cycle threshold for 18S (housekeeping) minus Ct for gene of interests or (//) fold change, where ACt for selected control condition =1 or 100%.

We used the following primers/ probe TaqMan assays:

Hs00231069_m1/ATF3,

Hs00374226_m1/NR4A, Hs01117527_g1/NR4A2,

Hs01031979_m1/MERTK,

Hs00360439_g1/KLF2,

Hs00358836_m 1 /KLF4,

Hs01090305_m1/Gas6,

Hs00174131_m1/ IL-6,

Hs00899658_m1/MMP1,

Hs00968305_m1/MMP3,

Hs00957562_m1/MMP9.

Example 8i: Single Cell Sequencing of STM and whole synovial tissues Our experiments were performed across two different sequencing centres. The first set of samples, which we refer to as our ‘Discovery Cohort’ was sequenced at the Oxford Genomics Centre, Oxford UK. Synovial tissue myeloid cells were sorted before sequencing, isolating cells with positive expression of CD11b and CD64 and negative expression of a range of other lineage specific cell markers (CD3, CD19, CD20, CD56, CD49, CD117 and CD15) as described in section “Phenotyping and FACS-sorting of STM subsets”

Around 2000-10000 synovial tissue macrophages per sample were sorted into qPCR 0.2ml tubes coated with FSC and containing 10 pi of PBS/0.02%BSA according to lOxGenomics protocol available online. With this data, we compared the transcriptomic profile of synovial myeloid cells from 5 subject groups - healthy, patients with undifferentiated peripheral arthritis (UPA), treatment-naive active RA, treatment- resistant active RA and RA in sustained remission. In addition, we sequenced a second set of samples at Glasgow Polyomics, University of Glasgow, Glasgow, UK; our “Validation Cohort”. The patient groups included Undifferentiated Peripheral Arthritis (UPA), treatment-naive active RA, treatment-resistant active RA and RA in sustained remission and the synovial tissue samples were analyzed for both STMs and FLS. We used the ‘validation’ scRNAseq transcriptomic profile of synovial myeloid cells sequenced by Glasgow Polyomics to validate the ‘discovery’ transcriptomic profile of FACS-sorted STMs scRNAseq profile measured at Oxford. Detailed information on both cohorts is provided in Supplementary Table 2-3 of Alivernini et al. 2020 Nature Medicine 26:1295-1306. doi: 10.1038/s41591 -020-0939-8 and Fig.19a-c. Data from both these cohorts was integrated using the following methods. Processing Raw Reads

All steps in primary data analysis, including read alignment and generation of count matrices, were performed using the Cell Ranger (2.1) pipeline. Raw base call files (BCL) generated by sequencing were previously demultiplexed into FASTQ files per sample. The “cellranger count” tool mapped the reads against the Human genome (hg19) and performed UMI counting.

QC & Filtering

The Seurat package (3.0.1) (Stuart et al., 2019, Cell 177, 1888-1902 e1821, doi: 10.1016/j. cell.2019.05.031) in R was used to create an object (CreateSeuratObject, min.cells=5). Cell filtering involved removal of cells with less than 500 expressed genes (subset, subset=nFeatures_RNA > 500). We also set thresholds for level of gene expression, including expression of mitochondrial genes (percent. mt). This allows for exclusion of doublets and dying cells (see Supplementary Table 4 of Alivernini et al. 2020 Nature Medicine 26:1295-1306. doi: 10.1038/s41591 -020-0939-8 for exact values). The data was normalized using Seurat’s NormalizeData function. For the analysis of synovial macrophages only, these cells were computationally isolated with the subset function from other cell types in Validation Cohort samples based on expression of CD 14, MARCO and LYZ. These STM specific markers were selected based on Discovery cohort data. The top 2000 variable genes were then identified for all samples, using the FindVariableFeatures function.

Integration

Sample integration was performed following the Seurat vignette, integrating all genes that are common between samples, using the functions: FindlntegrationAnchors, and IntegrateData (features. to. intergrate to find all common genes). These “integrated” batch-corrected values were then set as the default assay and the gene expression values are scaled before running principle component analysis.

Clustering and Dimensional Reduction UMAP based on PCA cell embeddings was generated from integrated counts batch- normalized by Seurat and the first 12 principle components (PCs) are visualized (RunUMAP). The same PCs were used in determination of the k-nearest neighbours for each cell during SNN graph construction before clustering at a chosen resolution of 0.5 (FindNeighbors, FindClusters). Rhe Destiny (2.14.0) R package

(httpsJ/academic. PUD. com/bioinformatics/article/32/8/1241/1744143) was used to plot a diffusion. A count matrix with the average expression of each cluster was generated before using Seurat’s PlotClusterTree function to generate a dendrogram.

Sample Filtering

In order to assess the quality of each sample, we determined the pseudo-bulk expression of each cluster per sample and performed PCA analysis on the result. Sample SA139 was removed due to low sequencing depth in the macrophages. SA225 was removed due to separation from all other samples in the PCA reduced dimensional space, (Supplementary Table 4 of Alivernini et al. 2020 Nature Medicine 26:1295- 1306. doi: 10.1038/s41591 -020-0939-8 and Fig.20d).

Differential Expression Analysis

In order to identify cluster markers and variable genes between conditions of RA, the Seurat function FindAIIMarkers was used with the “test. use” function MAST (Finak et al., 2015, Genome Biol 16, 278, doi: 10.1186/s 13059-015-0844-5). As recommended in the best practice of Seurat, for DE comparison the non-batch normalized counts were used. For identification of cluster markers, we specify that any markers identified must be expressed by at least 40% of cells in the cluster (‘min. pet’ parameter 0.4). A list of genes characterizing each of the clusters was compiled. For differential expression analysis between conditions, we increase this value to 0.6 to reduce the risk of sample bias. We use the default values for all other parameters. Genes are considered significantly DE if the adjusted p-value (< 0.05) by Bonferroni Correction and multiple test correction (multiple by number of tests). To visualise heatmaps the pheatmap package was adapted.

Pathway Analysis

To investigate the function of each of our identified synovial macrophage phenotypes, we performed pathway analysis using StringDB (https://strinq-db.org/) and I PA. Pathways associated with positive DE marker genes were investigated for each cluster. For each cluster, the Reactome pathways were exported and compared between all clusters in a custom R script. The script a gene ratio (number of observed genes in the pathway divided by total number of genes in the pathways as provided by String-db) as well as the associated FDR value. Only pathways with p-values less than 0.05 are listed (Fig. 13). Similar approaches were used to analyse FLS scRNAseq data from “Validation cohort”. Raw data is accessible at EMBL-EBI with the accession number E- MTAB-8322.

Example 8k: Trajectory Analysis

The Monocle 2.99 package (implemented in R) was used to construct a single-cell trajectory of our identified synovial tissue macrophage clusters. A downsampled dataset of the initial Seurat objected was created by selecting 10,000 random cells from the Discovery cohort. The single cell trajectory was constructed by performing differential expression analysis between the macrophage clusters identified from the Seurat analysis (differentialGeneTest, default parameters). Differentially expressed genes with a q-value <0.01 were used to order the cells along the trajectory. Dimensional reduction was performed (reduceDimension) using the DDRTree method. Pseudotime calculations then allowed for identification of genes which are differentially expressed along the trajectory (differentialGeneTest, pseudotime vector as input). To generate the plots, we used the monocle2 functions plot_cell_trajectory, plot_genes_in_pseudotime and plot_pseudotime_ heatmap (Fig.11).

Example 8I: Bulk RNA seq of synovial fibroblasts

High-quality total RNAs (RIN >8) were used to construct lllumina mRNA sequencing libraries. cDNA synthesis and amplification were performed by using SMART-seq v4 Ultra Low Input RNA Kit for Sequencing (cat. no. 634890, Takara) starting with 10 ng of total RNA, following the manufacturers protocol. 10 ng of amplified cDNAs were sheared prior to preparing the final libraries using the Bioruptor® Pico system (Diagenode, 24 cycles of 30 sec on and 30 sec off). Dual indexed lllumina sequencing libraries were prepared by SMARTer® ThruPLEX® DNA-seq 48D Kit (cat. no. R400406, Takara) following the kit protocol. The pooled libraries were sequenced at Edinburgh Genomics (Edinburgh, UK) on a NovaSeq 6000 system using a read length of 100 bases in paired-end mode. The reads were mapped with STAR (version 020201) with default parameter against the Human genome version GRCh38, release 91. The read count matrix was constructed with featureCounts (Version 1.6.4) using default parameters. All differential expression analysis was performed in R using the DESeq2 package. All genes with an adjusted p value < 0.05 and a log fold change of > ± 1.5 were considered significantly differentially expressed. Raw data is accessible at EMBL- EBI with the accession number E-MTAB-8316. Example 8m: Comparison of Human and Mouse scRNAseq Data Single cell transcriptional profiling on murine synovial tissue macrophages from the K/BxN serum transfer induced arthritis model (STIA) was performed in a recent publication (Culemann et al., 2019 Nature, doi: 10.1038/s41586-019-1471-1). We downloaded the data (GSE134691) and integrated this mouse data with our human samples from healthy tissue, undifferentiated arthritis (UPA), naive active RA and treatment resistant active RA. This was performed in a stepwise-manner - firstly by disease group, by species and finally integrating across species - using Seurat’s current integration methods (FindlntegrationAnchors, Integrate Data). The combined dataset was then scaled, before performing dimensional reduction and clustering using top 15PCs and a resolution of 0.3. Cluster marker genes were identified, and clusters were re-named accordingly. In addition, the datasets were sub-setted to create separate Seurat objects containing an assay of gene expression normalized across species from the final integration step. The datasets were then clustered and analysed separately. Orthologs (genes present in both datasets, n=7954) were also identified using the intersect function in R and the average expression of such genes was calculated for each dataset, using the gene expression values from cross-species normalization. The outputs for each dataset were merged and a distance matrix (dist function) was generated before performing hierarchical clustering (hclust function). A dendrogram was plotted from the result to demonstrate the relationship between synovial macrophage clusters from different species (Fig. 16-17).

Example 8n: Analysis of candidate genes in PEAC cohort Detailed methodology and analytical pipeline of synovial tissue bulk RNA-Seq from 90 individuals with early treatment-naive rheumatoid arthritis from the Pathobiology of Early Arthritis Cohort (PEAC) are described previously (Lewis et al., 2019, Cell Rep 28, 2455- 2470 e2455, doi:10.1016/j.celrep.2019.07.091). The study was approved by the UK Health Research Authority (REC 05/Q0703/198, National Research Ethics Service Committee London - Dulwich) and all patients gave written informed consent. Total RNA 1 pg/sample was extracted from whole synovial tissue retrieved from an inflamed peripheral joint using Trizol/Chloroform method. Bulk RNA-seq (50 million paired-end 75 bp reads/sample) was performed on lllumina HiSeq2500 platform. RNA-Seq data are uploaded to ArrayExpress (accession E-MTAB-6141). Data are expressed as regularised-log2 transformed reads. Example 80: Statistical evaluation of STM ohenotvoinq and culture experiments To define the best cut-off value for MerTK^pos/CD206^pos, MerTK^neg/CD206^neg, CD163^pos/CD206^pos and CD163^neg/CD206^neg ST-derived macrophages associated with disease flare after treatment modification in RA patients in sustained clinical and US remission (n=11 RA patients in sustained clinical and US remission who experienced disease flare and n= 11 RA patients in sustained clinical and US remission who did not experience disease flare after treatment modification) ROC analysis was performed for each parameter. Logistic regression model was performed to determine the influence of the dependent variable “Disease flare occurrence” by the independent variables “fulfilling the cut-off values for MerTK^pos/CD206^pos, MerTK^neg/CD206^neg, CD163^pos/CD206^pos and CD163^neg/CD206^neg synovial macrophage subpopulations” in RA patients in clinical and US-remission. The values are expressed as Odds Ratio (OR) and 95% Confidential Interval (95% Cls), respectively. The Hosmer-Lemeshow test was used to assess the fitting of the model. The difference in individual STM populations or cytokines between more than 2 joint conditions were evaluated using one-way ANOVA with Tukey’s correction for multiple comparison or Kruskal-Wallis test with Dunn’s correction for multiple comparison. Two-tailed nonparametric unpaired Mann-Whitney test was used if 2 groups were compared. Two-way ANOVA with Tukey’s correction for multiple comparison was used to evaluate (/) the differences between multiple cell clusters in multiple conditions and (//) multiple conditions and different time points.

Claims

1. A method for determining remission in a subject having rheumatoid arthritis, the method comprising the steps of: a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; and c) comparing the level of each of the biomarkers determined in (b) with one or more reference values, wherein a difference in the level of MerTK and CD206 in the biological sample compared to the one or more reference values is indicative of disease remission.

2. A method as claimed in claim 1 , wherein the method further comprises determining the level of TREM2 in the biological sample; wherein a difference in the level of TREM2 in the biological sample compared to the one or more reference values is indicative of disease remission.

3. A method as claimed in claim 1 or claim 2, wherein the method further comprises determining the level of CD163 in the biological sample; wherein a difference in the level of CD163 in the biological sample compared to the one or more reference values is indicative of disease remission.

4. A method as claimed in any preceding claim, wherein remission is characterised by an elevated level of MerTK and CD206 compared to the one or more reference values.

5. A method as claimed in claim 4, wherein remission is characterised by an elevated level of TREM2 compared to one or more reference values.

6. A method as claimed in claim 4 or claim 5, wherein remission is characterised by an elevated level of CD163 compared to one or more reference values.

7. A method as claimed in any preceding claim, wherein the biological sample comprises synovial cells.

8. A method as claimed in any preceding claim, wherein the biological sample comprises synovial tissue macrophages.

9. A method as claimed in any preceding claim, wherein the one or more reference values comprise the level of a corresponding biomarker from one or more control samples.

10. A method as claimed in any preceding claim, wherein the one or more reference values comprise the level of a corresponding biomarker from a control sample which is characteristic of active rheumatoid arthritis.

11 . A method as claimed in claim 10 or claim 11 , wherein the one or more control samples includes a biological sample comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from the subject at an earlier time point and which is characteristic of active rheumatoid arthritis; and wherein an increase in the level of each of the biomarkers compared to the one or more reference values is indicative of a change in the disease status of the subject from active to remission.

12. A method as claimed in any preceding claim, wherein the method comprises comparing the level of the biomarkers of a first biological sample taken at a first time point with the level of the corresponding biomarkers of a second biological sample taken at a second time point; wherein the first time point is before administration of a therapeutic agent and the second time point is after administration of the therapeutic agent.

13. A method as claimed in claim 12, wherein the therapeutic agent is an anti-arthritis agent.

14. A method as claimed in claim 12 or 13, wherein the therapeutic agent comprises a tumor necrosis factor inhibitor and/or methotrexate.

15. A method as claimed in claim 14, wherein the tumor necrosis factor inhibitor is adalimumab or etanercept.

16. A method as claimed in any preceding claim, wherein the biomarker comprises a biomarker protein or a nucleic acid molecule encoding the biomarker protein.

17. A method as claimed in any preceding claim, wherein the biomarker is a nucleic acid molecule encoding the biomarker protein.

18. A method as claimed in any preceding claim, wherein the biomarker is an mRNA molecule encoding the biomarker protein.

19. A method as claimed in any preceding claim, wherein the levels of the biomarkers in a biological sample are determined using specific binding partners selected from the group consisting of complementary nucleic acids; aptamers; antibodies or antibody fragments.

20. A method as claimed in any preceding claim, wherein the levels of the biomarkers in the biological sample are determined using real time quantitative PCR (RT-qPCR), digital PCR, microarray analysis, whole transcriptome shotgun sequencing, direct multiplexed gene expression analysis or whole transcriptome sequencing.

21. A method as claimed in any preceding claim, wherein the levels of the biomarkers in the biological sample are determined using a nucleic acid probe having a sequence which is complementary to the sequence the relevant mRNA or cDNA against which it is targeted.

22. A method as claimed in any preceding claim, wherein, the MERTK biomarker has a polynucleotide sequence of at least 90% sequence identity to SEQ ID NO: 5.

23. A method as claimed in any preceding claim, wherein, the CD206 biomarker has a polynucleotide sequence of at least 90% sequence identity to SEQ ID NO: 6.

24. A method as claimed in any preceding claim, wherein, the TREM2 biomarker has a polynucleotide sequence of at least 90% sequence identity to SEQ ID NO: 7.

25. A method as claimed in any preceding claim, wherein, the CD163 biomarker has a polynucleotide sequence of at least 90% sequence identity to SEQ ID NO: 8.

26. A method as claimed in any preceding claim, wherein remission is sustained remission.

27. A method as claimed in any preceding claim, wherein the method further comprises investigating physiological measurements selected from; tender joint count, swollen joint count, level of C-reactive protein in the blood, and patient global assessment.

28. A method as claimed in any preceding claim, wherein the subject is a human.

29. A method for evaluating the therapeutic efficacy of a candidate therapeutic agent for rheumatoid arthritis, the method comprising; comparing the level of the biomarkers MerTK and CD206 in biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the candidate therapeutic agent; wherein an increase in the level of MerTK and CD206 in the biological sample taken after the administration of the candidate therapeutic agent relative to the level of MerTK and CD206 in the biological sample taken before the administration of the candidate therapeutic agent is indicative of effective treatment.

30. A method as claimed in claim 29, wherein the method further comprises comparing the level of TREM2 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the candidate therapeutic agent; wherein an increase in the level of TREM2 in the biological sample taken after the administration of the candidate therapeutic agent is increased relative to the level of TREM2 in the biological sample taken before the administration of the candidate therapeutic agent is indicative of effective treatment.

31. A method as claimed in claim 29 or claim 30, wherein the method further comprises comparing the level of CD163 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the candidate therapeutic agent; wherein an increase in the level of CD163 in the biological sample taken after the administration of the candidate therapeutic agent is increased relative to the level of CD163 in the biological sample taken before the administration of the candidate therapeutic agent is indicative of effective treatment.

32. A method for evaluating response to treatment in a subject having rheumatoid arthritis, the method comprising; a) determining the level of each of the biomarkers MerTK and CD206 in a first biological sample comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from the subject at a first time point prior to administration of a treatment for rheumatoid arthritis; b) administering to the subject a treatment for rheumatoid arthritis; c) determining the level of each of the biomarkers MerTK and CD206 in a second biological sample comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from the subject at a subsequent time point following administration of the treatment for rheumatoid arthritis; and d) comparing the level of the biomarkers determined in (a) with the level of the corresponding biomarkers determined in (c), wherein an increase in the levels of MerTK and CD206 in (c) relative to (a) identifies the subject as having a positive response to treatment and a decrease or no change in the in the levels of MerTK and CD206 in (c) relative to (a) identifies the subject as having no response to treatment or a negative response to treatment.

33. A method as claimed in claim 32, wherein the method further comprises determining the level of TREM2 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the treatment for rheumatoid arthritis; and comparing the level of TREM2 determined in (a) with the level of TREM2 determined in (c), wherein an increase in the levels of TREM2 in (c) relative to (a) identifies the subject as having a positive response to treatment and a decrease or no change in the in the levels of TREM2 in (c) relative to (a) identifies the subject as having no response to treatment or a negative response to treatment.

34. A method as claimed in claim 32 or 33, wherein the method further comprises determining the level of CD163 in the biological samples comprising a synovial cell, or an extract or sub-cellular fraction thereof obtained from a subject having rheumatoid arthritis before and after administration of the treatment for rheumatoid arthritis; and comparing the level of CD163 determined in (a) with the level of CD163 determined in (c), wherein an increase in the levels of CD163 in (c) relative to (a) identifies the subject as having a positive response to treatment and a decrease or no change in the in the levels of CD163 in (c) relative to (a) identifies the subject as having no response to treatment or a negative response to treatment.

35. A method for determining in a subject receiving treatment for rheumatoid arthritis, whether the subject is suitable for treatment withdrawal, the method comprising; a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; and c) comparing the level of each of the biomarkers determined in (b) with the level of each of the biomarkers MerTK and CD206 in a control sample, wherein an increase in the level of MerTK and CD206 in the biological sample compared to the control sample is indicative of suitability for treatment withdrawal.

36. A method as claimed in claim 35, wherein the method further comprises determining the level of TREM2 in the biological sample and comparing the level of TREM2 in the biological sample with the level of TREM2 in a control sample, wherein an increase in the level of TREM2 in the biological sample compared to the control sample is indicative of suitability for treatment withdrawal.

37. A method as claimed in claim 35 or 36, wherein the method further comprises determining the level of CD163 in the biological sample and comparing the level of CD163 in the biological sample with the level of CD163 in a control sample, wherein an increase in the level of CD163 in the biological sample compared to the control sample is indicative of suitability for treatment withdrawal.

38. A method for determining in a subject receiving treatment for rheumatoid arthritis, whether the subject is suitable for treatment withdrawal, the method comprising; a) providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages; b) comparing the number of synovial tissue macrophages in the biological sample which express each of the biomarkers MerTK and CD206 to the number of synovial tissue macrophages in the biological sample which do not express MerTK and CD206; wherein a greater number of synovial tissue macrophages which express each of the biomarkers MerTK and CD206 indicates that the subject is suitable for treatment withdrawal.

39. A method as claimed in claim 38, wherein the method further comprises comparing the number of synovial tissue macrophages in the biological sample which express TREM2 to the number of synovial tissue macrophages in the biological sample which do not express TREM2; wherein a greater number of synovial tissue macrophages which express TREM2 indicates that the subject is suitable for treatment withdrawal.

40. A method as claimed in claim 38 or 39, wherein the method further comprises comparing the number of synovial tissue macrophages in the biological sample which express CD163 to the number of synovial tissue macrophages in the biological sample which do not express CD163; wherein a greater number of synovial tissue macrophages which express CD163 indicates that the subject is suitable for treatment withdrawal.

41.A method for determining the likelihood of a subject receiving treatment for rheumatoid arthritis experiencing flare, or remaining in sustained remission upon discontinuation of treatment; the method comprising the steps of: a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; and c) comparing the level of each of the biomarkers determined in (b) with the level of each of the biomarkers MerTK and CD206 in a control sample; d) wherein an increase in the level of MerTK and CD206 in the biological sample compared to the control sample is indicative of long-term disease remission; and e) wherein a decrease in, or no change in, the level of MerTK and CD206 in the biological sample compared to the control sample is predictive of flare upon discontinuation of treatment.

42. A method for determining the likelihood of a subject receiving treatment for rheumatoid arthritis experiencing flare, or remaining in sustained remission upon discontinuation of treatment; the method comprising the steps of: a) providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages; b) comparing the number of synovial tissue macrophages in the biological sample which express each of the biomarkers MerTK and CD206 to the number of synovial tissue macrophages in the biological sample which do not express MerTK and CD206; wherein a greater number of synovial tissue macrophages which do not express each of the biomarkers MerTK and CD206 is predictive of flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express each of the biomarkers MerTK and CD206 is indicative of sustained remission.

43. A method as claimed in claim 42, wherein the method further comprises comparing the number of synovial tissue macrophages in the biological sample which express TREM2 to the number of synovial tissue macrophages in the biological sample which do not express TREM2; wherein a greater number of synovial tissue macrophages which do not express TREM2 is predictive of flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express TREM2 is indicative of sustained remission.

44. A method as claimed in claim 42 or 43, wherein the method further comprises comparing the number of synovial tissue macrophages in the biological sample which express CD163 to the number of synovial tissue macrophages in the biological sample which do not express CD163; wherein a greater number of synovial tissue macrophages which do not express CD163 is predictive of flare upon discontinuation of treatment and wherein a greater number of synovial tissue macrophages which express CD163 is indicative of sustained remission.

45. A method for determining the likelihood of a subject receiving treatment for rheumatoid arthritis experiencing flare, or remaining in sustained remission upon discontinuation of treatment; the method comprising the steps of: a) providing a biological sample obtained from the subject, the sample comprising a plurality of synovial tissue macrophages; and b) determining a ratio of synovial tissue macrophages which express MerTK and CD206 to those which do not express MerTK or CD206; wherein a ratio of MerTK and CD206 expressing synovial tissue macrophages to MerTK and CD206 negative synovial tissue macrophages of less than or equal to 2.5 is predictive of flare upon discontinuation of treatment.

46. A method as claimed in claim 45, wherein the method comprises determining a ratio of synovial tissue macrophages which express MerTK, CD206 and TREM2 to those which do not express MerTK, CD206 or TREM2; wherein a ratio of MerTK, CD206 and TREM2 expressing synovial tissue macrophages to MerTK, CD206 and TREM2 negative synovial tissue macrophages of less than or equal to 2.5 is predictive of flare upon discontinuation of treatment.

47. A method as claimed in claim 45, wherein the method comprises determining a ratio of synovial tissue macrophages which express MerTK, CD206 and CD163 to those which do not express MerTK, CD206 or CD163; wherein a ratio of MerTK, CD206 and CD163 expressing synovial tissue macrophages to MerTK, CD206 and CD163 negative synovial tissue macrophages of less than or equal to 2.5 is predictive of flare upon discontinuation of treatment.

48. A method as claimed in claim 45, wherein the method comprises determining a ratio of synovial tissue macrophages which express MerTK, CD206, TREM2 and CD163 to those which do not express MerTK, CD206, TREM2 or CD163; wherein a ratio of MerTK, CD206, TREM2 and CD163 expressing synovial tissue macrophages to MerTK, CD206, TREM2 and CD163 negative synovial tissue macrophages of less than or equal to 2.5 is predictive of flare upon discontinuation of treatment.

49. A method of treating a patient having rheumatoid arthritis, comprising the steps of; a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; c) comparing the level of each of the biomarkers determined in (b) with the level of each of the biomarkers MerTK and CD206 in a control sample; and d) administering a therapeutic agent where the level of MerTK and CD206 in the biological sample is elevated compared to the control sample; or e) providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages; f) comparing the number of synovial tissue macrophages in the biological sample of (e) which express each of the biomarkers MerTK and CD206 to the number of synovial tissue macrophages in the biological sample of (e) which do not express

MerTK and CD206; and g) administering a therapeutic agent where a greater number of synovial tissue macrophages in the biological sample of (e) do not express MerTK and CD206.

50. A method of treating a patient having rheumatoid arthritis, wherein the patient is already receiving treatment for rheumatoid arthritis, comprising the steps of; a) providing a biological sample obtained from a subject comprising a synovial cell, or an extract or sub-cellular fraction thereof; b) determining the level of each of the biomarkers MerTK and CD206 in the biological sample; c) comparing the level of each of the biomarkers determined in (b) with the level of each of the biomarkers MerTK and CD206 in a control sample; and d) administering a different therapeutic agent where the level of MerTK and CD206 in the biological sample is elevated compared to the control sample; and e) optionally withdrawing the original treatment regime; or f) providing a biological sample obtained from a subject, the sample comprising a plurality of synovial tissue macrophages; g) comparing the number of synovial tissue macrophages in the biological sample of (f) which express each of the biomarkers MerTK and CD206 to the number of synovial tissue macrophages in the biological sample of (f) which do not express MerTK and CD206; and h) administering a different therapeutic agent where either; i) a greater number of synovial tissue macrophages in the biological sample of (f) do not express MerTK and CD206; or ii) a ratio of MerTK and CD206 expressing synovial tissue macrophages to MerTK and CD206 negative synovial tissue macrophages in the biological sample of (f) is less than or equal to 2.5; and i) optionally withdrawing the original treatment regime.

51. A method as claimed in claim 49 or 50, wherein the therapeutic agent comprises a DMARD.

52. A method as claimed in any of claims 49 to 51 , wherein the therapeutic agent comprises an anti-TNFa agent.

53. A method as claimed in claims 49 to 52, wherein the therapeutic agent is selected from Infliximab, Etanercept, Adalimumab, Certolizumab-pegol, Golimumab, Rituximab, Tocilizumab, Abatacept, Methotrexate, Sulfasalazine Hydroxychloroquine Leflunomide, Azathioprine, Penicillamine, Gold Injections, Ciclosporin or any combination thereof.

54. A device for use in the determination of remission in a subject having been determined to have rheumatoid arthritis, the device comprising: i) a loading area for receipt of a biological sample; ii) binding partners specific for target molecules indicative of the level of biomarkers MerTK and CD206; iii) optionally binding partners specific for target molecules indicative of the level of biomarkers TREM2 and/or CD163; and iv) detection means to detect the levels of said biomarker present in the sample.

55. A device as claimed in claim 54, wherein the binding partners comprise nucleic acid primers adapted to bind specifically to the cDNA transcripts of biomarkers.

56. A kit of parts for determining whether an individual having rheumatoid arthritis is in sustained remission, wherein the kit comprises reagents for establishing the level of MerTK and CD206; wherein an elevated level of MerTK and CD206 compared to the one or more reference values is indicative of sustained remission.

57. A kit as claimed in claim 56, wherein the kit comprises reagents for establishing the level of TREM2; wherein an elevated level of TREM2 compared to the one or more reference values is indicative of sustained remission.

58. A kit as claimed in claim 56 or 57, wherein the kit comprises reagents for establishing the level of CD163; wherein an elevated level of CD163 compared to the one or more reference values is indicative of sustained remission.

59. A kit as claimed in any of claims 56 to 58 claim, wherein the kit comprises reagents for establishing the level of said biomarkers by RT-qPCR, microarray analysis, digital PCR, whole transcriptome shotgun sequencing, direct multiplexed gene expression analysis or whole transcriptome sequencing.

60. A kit of parts for determining rheumatoid arthritis sustained remission, wherein the kit comprises: a) at least one binding partner that selectively binds to the MerTK biomarker, or a fragment thereof; b) at least one binding partner that selectively binds to the CD206 biomarker, or a fragment thereof; c) a positive control for the detection of said biomarkers; d) at least one binding partner that selectively binds to a nucleic acid or protein which operates as an internal control; and e) optionally an internal standard.

61 .A kit of parts as claimed in claim 60, wherein the kit further comprises at least one binding partner that selectively binds to the TREM2 biomarker, or a fragment thereof and a positive control for the detection of TREM2.

62. A kit of parts as claimed in claim 60 or 61 , wherein the kit further comprises at least one binding partner that selectively binds to the CD163 biomarker, or a fragment thereof and a positive control for the detection of CD163.

63. A kit or device as claimed in any of claims 54 to 62, wherein the kit or device is configured to determine the levels of MerTK, CD206, TREM2 and CD163 only.