CD4 T cells play a major role in autoimmune diseases such as systemic lupus erythematosus (SLE) by accumulating as autoreactive disease-promoting memory cells and amplifying inflammation as well as providing B cell help (Suárez-Fueyo et al., 2016). It is known that dysregulated cytokine production, metabolic alterations and aberrant cell signalling cause abnormalities in T cell differentiation and hyperactivation of B cells and thus contribute to lupus disease (Moulton et al., 2017). For SLE, there are several studies in mouse models and humans, showing that the progressive disease is associated with lower expression of TNF-α, IFN-γ, and IL-2 (Humrich et al., 2010) as well as higher expression of PD-1 (Jiao et al., 2014), IL-21 (Wang et al., 2014), and IL-10 (Facciotti et al., 2016). Despite the detection of these individual parameters, a comprehensive combinatorial characterization of T-cell subsets in SLE is still lacking.

Recent articles have revealed enormous and unexpected heterogeneity and complexity in T cell subpopulations with partially overlapping properties and functions of the cell subsets, such as follicular T helper (Tfh) and Tfh-like cell subsets (Trüb et al., 2017; Wong et al., 2015; Wong et al., 2016). Of particular interest are the recently observed human T peripheral helper cells (Tph), which expand in autoimmune diseases (Bocharnikov et al., 2019; Christophersen et al., 2019; Ekman et al., 2019; Rao et al., 2017). These and other studies show on the one hand the enormous potential to gain new biological insights from multidimensional cytometry data. On the other hand, they reveal the tremendous demand for new analysis and visualization approaches, especially for the identification of small but different characteristics in apparently homogeneous subpopulations, which is emphasized by numerous review articles (Kvistborg et al., 2015; Mair et al., 2016; Newell and Cheng, 2016; Saeys et al., 2016). Most of the current analysis strategies are based on clustering and the respective frequencies of the clusters (Levine et al., 2015; Qiu et al., 2011; Spitzer et al., 2015; Wong et al., 2016). The main problem with these methods is that they tend to find mainly cell-populationclusters which are significantly high in phenotypical contrast. If the cell subpopulations are very similar or even partially overlapping, as in the present investigation, then a cluster approach is usually not very meaningful (Newell and Cheng, 2016; O'Neill et al., 2013; Saeys et al., 2016).

To circumvent this problem, we developed the bin-based ‘pattern recognition of immune cells (PRI)’ strategy. PRI is a reproducible analysis and visualization approach to examine the combinatorial expression patterns of proteins in a mixture of similar cell subpopulations. Here, we have combined PRI with comprehensive cytometric measurements of Th cells from NZBxW F1 mice, a common lupus model. The results of our approach provide evidence for a super-functional Th cell subpopulation mainly characterized by IL-21⁺ IFN-γ^hi PD-1^low, TNF-α^hi IL-2^hi, and functional properties of Tfh-like cells.

For many autoimmune diseases, the underlying cellular mechanisms are still incompletely understood. To develop novel targeted therapies, it is important to identify potentially pathogenic immune cell subsets. To this end, appropriate methods for the analysis and visualization of high-dimensional cytometric data are crucial to document global changes in the immune cell pattern during the course of the disease. Our study has generated three important findings: it showed the added value of our PRI-visualization, it identified a super-functional Tfh-like cell subpopulation and it indicated its potential functional effect on SLE.

First, PRI enabled the identification and characterization of the super-functional IL-21⁺ IFN-γ^hi cell subset by its main functionalities: (i) simple visualization of the combinatorial properties of three markers and heat map representation of different statistical attributes of the third marker by binning conventional two-parametric dot plots (e.g. Figure 1E, Figure 6E); (ii) semi-continuous, combinatorial display that allows straightforward perception of correlating patterns (e.g. Figure 3D–E); (iii) reproducible pseudo-multi-parametric display of different markers so that subpopulations can be intuitively observed and compared, even between samples of different staining panels, which is not applicable with cluster algorithms (e.g. Figure 4A and B; Qiu et al., 2011; Shen et al., 2018) and (iv) exploitation of auxiliary information in percentages is enabled by the combinatorial approach, which would otherwise only be accessible with further gating steps. It should also be noted, however, that with PRI many cells and a broad bin range together with the right marker combination are necessary to obtain meaningful information from the plots.

PRI paves the way for reproducible and automatable cytometric analyses, which are useful not only for basic research but also for clinical studies to identify disease mechanisms and progressions, to discover biomarkers, to analyze therapy responder/non-responder, and to monitor stable quality and effectiveness of manipulated cells. In addition, PRI provides a statistically robust pattern map for visual inspections and serves as input for prospective pattern recognition in machine learning approaches with feature engineering steps.

Second, the super-functional IL-21⁺IFN-γ^hi cell subset identified here is a novel Tfh-like subpopulation, which differs from the subpopulations described so far by a low expression of PD-1. In concordance with Tfh, Tph (Rao et al., 2017) and other Tfh-like cell subsets (Bubier et al., 2009; Hutloff, 2018; Vu Van et al., 2016; Weinstein et al., 2016) splenic super-functional T cells are IL-21⁺, CD40L^hi, and ICOS⁺ (Figure 4—figure supplement 1). Their affiliation to Tfh or extrafollicular Tfh cells could be ruled out, as the majority of super-functional T cells do neither express Bcl6 or CXCR5, nor CXCR4 or PSGL-1^low (Figure 4; Odegard et al., 2008). They rather seem to exhibit polyfunctional characteristics such as a strong co-expression of Th1 and Tfh-like effector proteins with high activation potential.

Poly-functional T cells were first described as being protective in viral infections (Kannanganat et al., 2007; Seder et al., 2008). However, polyfunctional T cell subpopulations were identified as pathogenic in autoimmune diseases (Basdeo et al., 2015). Most of the previous studies on poly-functional T cells focused on IFN-γ^hi, TNF-α^hi, IL-2^hi cells in the context of infections and immunizations (Mateus et al., 2019; Sauzullo et al., 2014; Westerhof et al., 2019). Compared to the triple positive poly-functional T cells, super-functional T cells here were even superior regarding their additional expression of IL-21, ICOS, and CD40L^hi and their demonstrated B cell helper activity (Figure 7). In vitro, PD-1^low CXCR3⁺ CXCR5^- Tsh cells and PD-1^hi cells, containing the CXCR5⁺ Tfh cells, displayed dichotomic functions regarding the induction of GC-like and plasma cell phenotypes, respectively. In keeping with the chemokine profile of these two CD4 T cell subsets, this result suggests that in vivo PD-1^low CXCR3⁺ CXCR5^- Tsh cells could provide help in the inter-follicular space juxtaposed to the GC and enriched in the CXCR3 ligand CXCL9 (Matloubian and Cyster, 2012). There, B cells become blasts, proliferate and differentiate into GC B cells that migrate to CXCR5 ligand-rich follicles, where further support by the Tfh promotes affinity maturation and plasma cells differentiation (Haynes et al., 2007; Vinuesa et al., 2016). Alternatively, Tsh could participate to expand activated B cells in an extrafollicular GC-like reaction, as recently described during Salmonella infections (Di Niro et al., 2015).

Considering the observed extrafollicular co-localization of T cells and B cells in the spleen of NZBxW mice (data not presented) and the peripheral localization of super-functional T cells (Figure 5D), we speculate that they can even provide B cell help directly in inflamed tissues, thereby exacerbating tissue destruction. Due to their low expression of inhibitory receptors (especially PD-1, CTLA4), their activation is potentially facilitated compared to Tfh cells, as these express significantly higher amounts of PD-1, CTLA4, TIGIT and BTLA (Figure 4; Butte et al., 2007; Walunas et al., 1994).

Third, the super-functional IL-21⁺IFN-γ^hi cell subset could promote SLE progression through B cell help dependent and independent effects. Similar to Tfh-like cells, which were found in SLE and other autoimmune diseases (Horiuchi and Ueno, 2018; Hutloff, 2018; Rao, 2018; Vu Van et al., 2016), the super-functional subset delivers effective B-cell help for IgG production in an IL-21- and CD40L-dependent manner. Tissue-resident Tfh-like cells are proposed to be 'the most pathogenic T cell subset, as they select autoreactive B cells in the uncontrolled environment of lymphocytic tissue infiltrates and drive the local differentiation of plasmablasts producing pathogenic antibodies directly in the affected tissues' (Hutloff, 2018).

Due to their presence also in non-lymphoid organs (Figure 5) and their IL-21 and IFN-γ co-expression, the super-functional cell subset might be superior in activation-induced processes that drive autoimmunity by expanding pathogenic T cell subsets in non-lymphoid tissue and promoting inflammation. This hypothesis is supported by several studies focusing either on IL-21 or IFN- γ effects. 

IL-21 uniquely contributes to SLE and other autoimmune diseases, because pharmacological and genetic abrogation of IL-21 signaling in mice stopped inflammation in non-lymphoid organs and protected them from autoimmune diseases, respectively (Choi et al., 2017; Liu and King, 2013). In line with the reported IL-21–mediated proliferation (Zeng et al., 2007) and its autocrine and paracrine action (Nurieva et al., 2007) is our observation that the IL-21⁺ IFN-γ+ double-producing population had the highest proliferative capacity in vivo compared to single-producing (IL-21⁺ IFN-γ^–, IL-21^– IFN-γ+) and double-negative (IL-21^– IFN-γ^–) populations measured by Ki67 expression (data not shown).

IFN- γ is also considered to be a key molecule in the pathogenesis of SLE (Peng et al., 2002; Pollard et al., 2013; Tsokos et al., 2016). Consequently, excessive IFN-γ production was required to sustain lupus-associated Tfh cell accumulation in mice (Lee et al., 2012). Contrary to this and other reports on lupus-prone mice (Enghard et al., 2006; Humrich et al., 2010), we observed no increase but a slight decrease in IFN-γ frequencies with disease progression (Figure 2), possibly due to different subpopulations of Tmem and Th cells being studied.

In addition to IL-21 and IFN-γ^hi co-expression, the super-functional cell subset is characterized by high IL-2 and TNF-α but no IL-10 expression. IL-2 is discussed to be rather protective than pathogenic in SLE because the acquired IL-2 deficiency is a crucial general event in the pathogenesis of lupus and mainly leads to a disorder of the homeostasis of regulatory T cells. Correspondingly, a low dose IL-2 therapy selectively corrected Treg cell defects and greatly expanded the Treg cell subpopulation (Humrich et al., 2015; von Spee-Mayer et al., 2016). In contrast, anti-TNF-α therapy with infliximab has produced inconsistent results so far and is not routinely used to treat SLE (Aringer and Smolen, 2012). IL-10 has pleiotropic effects, which can be anti-inflammatory and pro-inflammatory. In SLE patients, however, the higher IL-10 level associated with the disease is considered pathogenic and its blockage improves the disease (Llorente et al., 2000). Recently, an IL-10/IFN-γ co-producing pathogenic Th cell sub-population (CXCR5^- CXCR3⁺ PD-1^hi) was identified in SLE patients (Caielli et al., 2019). These cells differ from our super-functional cells primarily because they do not co-express IL-21, rather they provide B cell help independently of IL-21, and are PD-1 high.

Transient co-expression of IL-21 and IFN-γ has already been described for subsets of early Th1-Tfh and germinal centre Tfh cells (Fang et al., 2018; Nakayamada et al., 2011; Weinstein et al., 2016). However, it is still unclear to what extent double producing cells functionally differ from IFN-γ and IL-21 single producers (Song and Craft, 2019). It is worth mentioning in this context that the majority of IL-21⁺ Tmem cells in this study were IFN-γ+ (Figure 3F; young: 56 ± 3% versus old diseased: 71 ± 4%) and non-Tfh cells (Figure 6). They were found in both lymphoid and peripheral organs of NZBxW mice (Figure 5).

Interestingly, IL-21 was reported to inhibit TCR-induced Th1 differentiation (Kastirr et al., 2014; Suto et al., 2006; Wurster et al., 2002) and, vice versa, forced expression of the Th1 master transcription factor T-bet inhibited IL-21 expression (Kastirr et al., 2014). Therefore, the special co-expression of IL-21 and IFN-γ could be triggered by chronic inflammatory conditions, leading to a dysregulated cytokine landscape (Humrich et al., 2010), specific expansion of double positive cells (Kastirr et al., 2014), and plasticity of Th cell phenotype (Carpio et al., 2015).

Several reports underline a high diversity and plasticity of Tfh and Tfh-like cells (Lüthje et al., 2012; Song and Craft, 2019; Wong et al., 2015). Song and Craft even proposed in between the two archetypes are differentiated cells that have varied degrees of Th and Tfh cell characteristics which may not exist in distinct subsets" (Song and Craft, 2019). They proposed to search for alternatives to further dividing Th, Tfh and Tfh-like cell subsets. Our PRI approach is an attempt in this direction to analyze and visualize a continuum between cell subsets without further dividing subpopulations by gating (Figure 4, Figure 3—figure supplement 1D).

In this study, we defined a novel subset of Tfh-like cells – super-functional IL-21⁺ IFN-γ^hi PD-1^low Tmem cells – that exhibit superior production of cytokines in both quantity and number and provide B cell help. This suggests that efforts to decrease their frequency or manipulate their main properties such as IL-21/IFN-γ co-expression may represent an important therapeutic strategy.

Flow cytometry data have been deposited in FlowRepository under the accession code FR-FCM-Z2C8. All data generated or analysed during this study are included in the manuscript and supporting files. Source data files have been provided for Figure 1D (Figure 1-source data 1); Figure 2A,B,D,E (Figure 2-source data 1-4) and Figure 2-figure supplement 1B (Figure 2-figure supplement 1-source data 1); Figure 2-figure supplement 2B,E (Figure 2-figure supplement 2-source data 1-2); Figure 2-figure supplement 3D (Figure 2-figure supplement 3-source data 1); Figure 3C (Figure 3-source data 1); Figure 3-figure supplement 1A-C (Figure 3-figure supplement 1-source data 1-2); Figure 5A-D (Figure 5-source data 1-4); Figure 6B,C (Figure 6-source data 1-2) and Figure 6-figure supplement 1A (Figure 6-figure supplement 1-source data 1); and Figure 7B,C (Figure 7-source data 1-2).

1. 1. Gryzik
   2. Hoang

   (2019) Flow Repository

   ID FR-FCM-Z2C8. PRI Tfh-like.

   http://flowrepository.org/id/FR-FCM-Z2C8

1. 1. Aringer M
   2. Smolen JS

   (2012) Therapeutic blockade of TNF in patients with SLE-promising or crazy?

   Autoimmunity Reviews 11:321–325.

   https://doi.org/10.1016/j.autrev.2011.05.001

   • PubMed
   • Google Scholar
2. 1. Basdeo SA
   2. Moran B
   3. Cluxton D
   4. Canavan M
   5. McCormick J
   6. Connolly M
   7. Orr C
   8. Mills KH
   9. Veale DJ
   10. Fearon U
   11. Fletcher JM

   (2015) Polyfunctional, pathogenic CD161+ Th17 lineage cells are resistant to regulatory T Cell-Mediated suppression in the context of autoimmunity

   The Journal of Immunology 195:528–540.

   https://doi.org/10.4049/jimmunol.1402990

   • PubMed
   • Google Scholar
3. 1. Bendfeldt H
   2. Benary M
   3. Scheel T
   4. Frischbutter S
   5. Abajyan A
   6. Radbruch A
   7. Herzel H
   8. Baumgrass R

   (2012) Stable IL-2 decision making by endogenous c-Fos amounts in peripheral memory T-helper cells

   Journal of Biological Chemistry 287:18386–18397.

   https://doi.org/10.1074/jbc.M112.358853

   • PubMed
   • Google Scholar
4. 1. Bocharnikov AV
   2. Keegan J
   3. Wacleche VS
   4. Cao Y
   5. Fonseka CY
   6. Wang G
   7. Muise ES
   8. Zhang KX
   9. Arazi A
   10. Keras G
   11. Li ZJ
   12. Qu Y
   13. Gurish MF
   14. Petri M
   15. Buyon JP
   16. Putterman C
   17. Wofsy D
   18. James JA
   19. Guthridge JM
   20. Diamond B
   21. Anolik JH
   22. Mackey MF
   23. Alves SE
   24. Nigrovic PA
   25. Costenbader KH
   26. Brenner MB
   27. Lederer JA
   28. Rao DA

   (2019) PD-1hiCXCR5– T peripheral helper cells promote B cell responses in lupus via MAF and IL-21

   JCI Insight 4:e130062.

   https://doi.org/10.1172/jci.insight.130062

   • Google Scholar
5. 1. Bubier JA
   2. Sproule TJ
   3. Foreman O
   4. Spolski R
   5. Shaffer DJ
   6. Morse HC
   7. Leonard WJ
   8. Roopenian DC

   (2009) A critical role for IL-21 receptor signaling in the pathogenesis of systemic lupus erythematosus in BXSB-Yaa mice

   PNAS 106:1518–1523.

   https://doi.org/10.1073/pnas.0807309106

   • PubMed
   • Google Scholar
6. 1. Butte MJ
   2. Keir ME
   3. Phamduy TB
   4. Sharpe AH
   5. Freeman GJ

   (2007) Programmed death-1 ligand 1 interacts specifically with the B7-1 costimulatory molecule to inhibit T cell responses

   Immunity 27:111–122.

   https://doi.org/10.1016/j.immuni.2007.05.016

   • PubMed
   • Google Scholar
7. 1. Caielli S
   2. Veiga DT
   3. Balasubramanian P
   4. Athale S
   5. Domic B
   6. Murat E
   7. Banchereau R
   8. Xu Z
   9. Chandra M
   10. Chung CH
   11. Walters L
   12. Baisch J
   13. Wright T
   14. Punaro M
   15. Nassi L
   16. Stewart K
   17. Fuller J
   18. Ucar D
   19. Ueno H
   20. Zhou J
   21. Banchereau J
   22. Pascual V

   (2019) A CD4⁺ T cell population expanded in lupus blood provides B cell help through interleukin-10 and succinate

   Nature Medicine 25:75–81.

   https://doi.org/10.1038/s41591-018-0254-9

   • PubMed
   • Google Scholar
8. 1. Carpio VH
   2. Opata MM
   3. Montañez ME
   4. Banerjee PP
   5. Dent AL
   6. Stephens R

   (2015) IFN-γ and IL-21 double producing T cells are Bcl6-Independent and survive into the memory phase in plasmodium chabaudi infection

   PLOS ONE 10:e0144654.

   https://doi.org/10.1371/journal.pone.0144654

   • PubMed
   • Google Scholar
9. 1. Cheng Q
   2. Mumtaz IM
   3. Khodadadi L
   4. Radbruch A
   5. Hoyer BF
   6. Hiepe F

   (2013) Autoantibodies from long-lived 'memory' plasma cells of NZB/W mice drive immune complex nephritis

   Annals of the Rheumatic Diseases 72:2011–2017.

   https://doi.org/10.1136/annrheumdis-2013-203455

   • PubMed
   • Google Scholar
10. 1. Choi JY
    2. Ho JH
    3. Pasoto SG
    4. Bunin V
    5. Kim ST
    6. Carrasco S
    7. Borba EF
    8. Gonçalves CR
    9. Costa PR
    10. Kallas EG
    11. Bonfa E
    12. Craft J

    (2015) Circulating follicular helper-like T cells in systemic lupus erythematosus: association with disease activity

    Arthritis & Rheumatology 67:988–999.

    https://doi.org/10.1002/art.39020

    • PubMed
    • Google Scholar
11. 1. Choi JY
    2. Seth A
    3. Kashgarian M
    4. Terrillon S
    5. Fung E
    6. Huang L
    7. Wang LC
    8. Craft J

    (2017) Disruption of pathogenic cellular networks by IL-21 blockade leads to disease amelioration in murine lupus

    The Journal of Immunology 198:2578–2588.

    https://doi.org/10.4049/jimmunol.1601687

    • PubMed
    • Google Scholar
12. 1. Christophersen A
    2. Lund EG
    3. Snir O
    4. Solà E
    5. Kanduri C
    6. Dahal-Koirala S
    7. Zühlke S
    8. Molberg Ø
    9. Utz PJ
    10. Rohani-Pichavant M
    11. Simard JF
    12. Dekker CL
    13. Lundin KEA
    14. Sollid LM
    15. Davis MM

    (2019) Distinct phenotype of CD4⁺ T cells driving celiac disease identified in multiple autoimmune conditions

    Nature Medicine 25:734–737.

    https://doi.org/10.1038/s41591-019-0403-9

    • PubMed
    • Google Scholar
13. 1. Cossarizza A
    2. Chang HD
    3. Radbruch A
    4. Akdis M
    5. Andrä I
    6. Annunziato F
    7. Bacher P
    8. Barnaba V
    9. Battistini L
    10. Bauer WM
    11. Baumgart S
    12. Becher B
    13. Beisker W
    14. Berek C
    15. Blanco A
    16. Borsellino G
    17. Boulais PE
    18. Brinkman RR
    19. Büscher M
    20. Busch DH
    21. Bushnell TP
    22. Cao X
    23. Cavani A
    24. Chattopadhyay PK
    25. Cheng Q
    26. Chow S
    27. Clerici M
    28. Cooke A
    29. Cosma A
    30. Cosmi L
    31. Cumano A
    32. Dang VD
    33. Davies D
    34. De Biasi S
    35. Del Zotto G
    36. Della Bella S
    37. Dellabona P
    38. Deniz G
    39. Dessing M
    40. Diefenbach A
    41. Di Santo J
    42. Dieli F
    43. Dolf A
    44. Donnenberg VS
    45. Dörner T
    46. Ehrhardt GRA
    47. Endl E
    48. Engel P
    49. Engelhardt B
    50. Esser C
    51. Everts B
    52. Dreher A
    53. Falk CS
    54. Fehniger TA
    55. Filby A
    56. Fillatreau S
    57. Follo M
    58. Förster I
    59. Foster J
    60. Foulds GA
    61. Frenette PS
    62. Galbraith D
    63. Garbi N
    64. García-Godoy MD
    65. Geginat J
    66. Ghoreschi K
    67. Gibellini L
    68. Goettlinger C
    69. Goodyear CS
    70. Gori A
    71. Grogan J
    72. Gross M
    73. Grützkau A
    74. Grummitt D
    75. Hahn J
    76. Hammer Q
    77. Hauser AE
    78. Haviland DL
    79. Hedley D
    80. Herrera G
    81. Herrmann M
    82. Hiepe F
    83. Holland T
    84. Hombrink P
    85. Houston JP
    86. Hoyer BF
    87. Huang B
    88. Hunter CA
    89. Iannone A
    90. Jäck HM
    91. Jávega B
    92. Jonjic S
    93. Juelke K
    94. Jung S
    95. Kaiser T
    96. Kalina T
    97. Keller B
    98. Khan S
    99. Kienhöfer D
    100. Kroneis T
    101. Kunkel D
    102. Kurts C
    103. Kvistborg P
    104. Lannigan J
    105. Lantz O
    106. Larbi A
    107. LeibundGut-Landmann S
    108. Leipold MD
    109. Levings MK
    110. Litwin V
    111. Liu Y
    112. Lohoff M
    113. Lombardi G
    114. Lopez L
    115. Lovett-Racke A
    116. Lubberts E
    117. Ludewig B
    118. Lugli E
    119. Maecker HT
    120. Martrus G
    121. Matarese G
    122. Maueröder C
    123. McGrath M
    124. McInnes I
    125. Mei HE
    126. Melchers F
    127. Melzer S
    128. Mielenz D
    129. Mills K
    130. Mirrer D
    131. Mjösberg J
    132. Moore J
    133. Moran B
    134. Moretta A
    135. Moretta L
    136. Mosmann TR
    137. Müller S
    138. Müller W
    139. Münz C
    140. Multhoff G
    141. Munoz LE
    142. Murphy KM
    143. Nakayama T
    144. Nasi M
    145. Neudörfl C
    146. Nolan J
    147. Nourshargh S
    148. O'Connor JE
    149. Ouyang W
    150. Oxenius A
    151. Palankar R
    152. Panse I
    153. Peterson P
    154. Peth C
    155. Petriz J
    156. Philips D
    157. Pickl W
    158. Piconese S
    159. Pinti M
    160. Pockley AG
    161. Podolska MJ
    162. Pucillo C
    163. Quataert SA
    164. Radstake T
    165. Rajwa B
    166. Rebhahn JA
    167. Recktenwald D
    168. Remmerswaal EBM
    169. Rezvani K
    170. Rico LG
    171. Robinson JP
    172. Romagnani C
    173. Rubartelli A
    174. Ruckert B
    175. Ruland J
    176. Sakaguchi S
    177. Sala-de-Oyanguren F
    178. Samstag Y
    179. Sanderson S
    180. Sawitzki B
    181. Scheffold A
    182. Schiemann M
    183. Schildberg F
    184. Schimisky E
    185. Schmid SA
    186. Schmitt S
    187. Schober K
    188. Schüler T
    189. Schulz AR
    190. Schumacher T
    191. Scotta C
    192. Shankey TV
    193. Shemer A
    194. Simon AK
    195. Spidlen J
    196. Stall AM
    197. Stark R
    198. Stehle C
    199. Stein M
    200. Steinmetz T
    201. Stockinger H
    202. Takahama Y
    203. Tarnok A
    204. Tian Z
    205. Toldi G
    206. Tornack J
    207. Traggiai E
    208. Trotter J
    209. Ulrich H
    210. van der Braber M
    211. van Lier RAW
    212. Veldhoen M
    213. Vento-Asturias S
    214. Vieira P
    215. Voehringer D
    216. Volk HD
    217. von Volkmann K
    218. Waisman A
    219. Walker R
    220. Ward MD
    221. Warnatz K
    222. Warth S
    223. Watson JV
    224. Watzl C
    225. Wegener L
    226. Wiedemann A
    227. Wienands J
    228. Willimsky G
    229. Wing J
    230. Wurst P
    231. Yu L
    232. Yue A
    233. Zhang Q
    234. Zhao Y
    235. Ziegler S
    236. Zimmermann J

    (2017) Guidelines for the use of flow cytometry and cell sorting in immunological studies

    European Journal of Immunology 47:1584–1797.

    https://doi.org/10.1002/eji.201646632

    • PubMed
    • Google Scholar
14. 1. Di Niro R
    2. Lee SJ
    3. Vander Heiden JA
    4. Elsner RA
    5. Trivedi N
    6. Bannock JM
    7. Gupta NT
    8. Kleinstein SH
    9. Vigneault F
    10. Gilbert TJ
    11. Meffre E
    12. McSorley SJ
    13. Shlomchik MJ

    (2015) Salmonella infection drives promiscuous B cell activation followed by extrafollicular affinity maturation

    Immunity 43:120–131.

    https://doi.org/10.1016/j.immuni.2015.06.013

    • PubMed
    • Google Scholar
15. 1. Ekman I
    2. Ihantola EL
    3. Viisanen T
    4. Rao DA
    5. Näntö-Salonen K
    6. Knip M
    7. Veijola R
    8. Toppari J
    9. Ilonen J
    10. Kinnunen T

    (2019) Circulating CXCR5^-PD^-1^hi peripheral T helper cells are associated with progression to type 1 diabetes

    Diabetologia 62:1681–1688.

    https://doi.org/10.1007/s00125-019-4936-8

    • PubMed
    • Google Scholar
16. 1. Enghard P
    2. Langnickel D
    3. Riemekasten G

    (2006) T cell cytokine imbalance towards production of IFN-gamma and IL-10 in NZB/W F1 lupus-prone mice is associated with autoantibody levels and nephritis

    Scandinavian Journal of Rheumatology 35:209–216.

    https://doi.org/10.1080/03009740500417791

    • PubMed
    • Google Scholar
17. 1. Facciotti F
    2. Gagliani N
    3. Häringer B
    4. Alfen JS
    5. Penatti A
    6. Maglie S
    7. Paroni M
    8. Iseppon A
    9. Moro M
    10. Crosti MC
    11. Stölzel K
    12. Romagnani C
    13. Moroni G
    14. Ingegnoli F
    15. Torretta S
    16. Pignataro L
    17. Annoni A
    18. Russo F
    19. Pagani M
    20. Abrignani S
    21. Meroni P
    22. Flavell R
    23. Geginat J

    (2016) IL-10-producing forkhead box protein 3-negative regulatory T cells inhibit B-cell responses and are involved in systemic lupus erythematosus

    Journal of Allergy and Clinical Immunology 137:318–321.

    https://doi.org/10.1016/j.jaci.2015.06.044

    • PubMed
    • Google Scholar
18. 1. Fang D
    2. Cui K
    3. Mao K
    4. Hu G
    5. Li R
    6. Zheng M
    7. Riteau N
    8. Reiner SL
    9. Sher A
    10. Zhao K
    11. Zhu J

    (2018) Transient T-bet expression functionally specifies a distinct T follicular helper subset

    Journal of Experimental Medicine 215:2705–2714.

    https://doi.org/10.1084/jem.20180927

    • PubMed
    • Google Scholar
19. 1. Finak G
    2. Perez JM
    3. Weng A
    4. Gottardo R

    (2010) Optimizing transformations for automated, high throughput analysis of flow cytometry data

    BMC Bioinformatics 11:546.

    https://doi.org/10.1186/1471-2105-11-546

    • PubMed
    • Google Scholar
20. 1. Fuhrmann F
    2. Lischke T
    3. Gross F
    4. Scheel T
    5. Bauer L
    6. Kalim KW
    7. Radbruch A
    8. Herzel H
    9. Hutloff A
    10. Baumgrass R

    (2016) Adequate immune response ensured by binary IL-2 and graded CD25 expression in a murine transfer model

    eLife 5:e20616.

    https://doi.org/10.7554/eLife.20616

    • PubMed
    • Google Scholar
21. 1. Hayashi Y
    2. Utsuyama M
    3. Kurashima C
    4. Hirokawa K

    (1989)

    Spontaneous development of organ-specific autoimmune lesions in aged C57BL/6 mice

    Clinical and Experimental Immunology 78:120–126.

    • PubMed
    • Google Scholar
22. 1. Haynes NM
    2. Allen CD
    3. Lesley R
    4. Ansel KM
    5. Killeen N
    6. Cyster JG

    (2007) Role of CXCR5 and CCR7 in follicular th cell positioning and appearance of a programmed cell death gene-1high germinal center-associated subpopulation

    The Journal of Immunology 179:5099–5108.

    https://doi.org/10.4049/jimmunol.179.8.5099

    • PubMed
    • Google Scholar
23. 1. Horiuchi S
    2. Ueno H

    (2018) Potential pathways associated with exaggerated T follicular helper response in human autoimmune diseases

    Frontiers in Immunology 9:1630.

    https://doi.org/10.3389/fimmu.2018.01630

    • PubMed
    • Google Scholar
24. 1. Humrich JY
    2. Morbach H
    3. Undeutsch R
    4. Enghard P
    5. Rosenberger S
    6. Weigert O
    7. Kloke L
    8. Heimann J
    9. Gaber T
    10. Brandenburg S
    11. Scheffold A
    12. Huehn J
    13. Radbruch A
    14. Burmester GR
    15. Riemekasten G

    (2010) Homeostatic imbalance of regulatory and effector T cells due to IL-2 deprivation amplifies murine lupus

    PNAS 107:204–209.

    https://doi.org/10.1073/pnas.0903158107

    • PubMed
    • Google Scholar
25. 1. Humrich JY
    2. von Spee-Mayer C
    3. Siegert E
    4. Alexander T
    5. Hiepe F
    6. Radbruch A
    7. Burmester GR
    8. Riemekasten G

    (2015) Rapid induction of clinical remission by low-dose interleukin-2 in a patient with refractory SLE

    Annals of the Rheumatic Diseases 74:791–792.

    https://doi.org/10.1136/annrheumdis-2014-206506

    • PubMed
    • Google Scholar
26. 1. Hutloff A

    (2018) T follicular Helper-Like cells in inflamed Non-Lymphoid tissues

    Frontiers in Immunology 9:1707.

    https://doi.org/10.3389/fimmu.2018.01707

    • PubMed
    • Google Scholar
27. 1. Hwang S
    2. Cobb DA
    3. Bhadra R
    4. Youngblood B
    5. Khan IA

    (2016) Blimp-1-mediated CD4 T cell exhaustion causes CD8 T cell dysfunction during chronic toxoplasmosis

    The Journal of Experimental Medicine 213:1799–1818.

    https://doi.org/10.1084/jem.20151995

    • PubMed
    • Google Scholar
28. 1. Jiao Q
    2. Liu C
    3. Yang Z
    4. Ding Q
    5. Wang M
    6. Li M
    7. Zhu T
    8. Qian H
    9. Li W
    10. Tu N
    11. Fang F
    12. Ye L
    13. Zhao Z
    14. Qian Q

    (2014) Upregulated PD-1 expression is associated with the development of systemic lupus erythematosus, but not the PD-1.1 Allele of the PDCD1 Gene

    International Journal of Genomics 2014:1–6.

    https://doi.org/10.1155/2014/950903

    • PubMed
    • Google Scholar
29. 1. Kannanganat S
    2. Ibegbu C
    3. Chennareddi L
    4. Robinson HL
    5. Amara RR

    (2007) Multiple-cytokine-producing antiviral CD4 T cells are functionally superior to single-cytokine-producing cells

    Journal of Virology 81:8468–8476.

    https://doi.org/10.1128/JVI.00228-07

    • PubMed
    • Google Scholar
30. 1. Kasagi S
    2. Kawano S
    3. Okazaki T
    4. Honjo T
    5. Morinobu A
    6. Hatachi S
    7. Shimatani K
    8. Tanaka Y
    9. Minato N
    10. Kumagai S

    (2010) Anti-programmed cell death 1 antibody reduces CD4⁺PD-1⁺ T cells and relieves the lupus-like nephritis of NZB/W F1 mice

    Journal of Immunology 184:2337–2347.

    https://doi.org/10.4049/jimmunol.0901652

    • PubMed
    • Google Scholar
31. 1. Kastirr I
    2. Maglie S
    3. Paroni M
    4. Alfen JS
    5. Nizzoli G
    6. Sugliano E
    7. Crosti MC
    8. Moro M
    9. Steckel B
    10. Steinfelder S
    11. Stölzel K
    12. Romagnani C
    13. Botti F
    14. Caprioli F
    15. Pagani M
    16. Abrignani S
    17. Geginat J

    (2014) IL-21 is a central memory T cell-associated cytokine that inhibits the generation of pathogenic Th1/17 effector cells

    The Journal of Immunology 193:3322–3331.

    https://doi.org/10.4049/jimmunol.1400775

    • PubMed
    • Google Scholar
32. 1. Kvistborg P
    2. Gouttefangeas C
    3. Aghaeepour N
    4. Cazaly A
    5. Chattopadhyay PK
    6. Chan C
    7. Eckl J
    8. Finak G
    9. Hadrup SR
    10. Maecker HT
    11. Maurer D
    12. Mosmann T
    13. Qiu P
    14. Scheuermann RH
    15. Welters MJ
    16. Ferrari G
    17. Brinkman RR
    18. Britten CM

    (2015) Thinking outside the gate: single-cell assessments in multiple dimensions

    Immunity 42:591–592.

    https://doi.org/10.1016/j.immuni.2015.04.006

    • PubMed
    • Google Scholar
33. 1. Larsen M
    2. Sauce D
    3. Arnaud L
    4. Fastenackels S
    5. Appay V
    6. Gorochov G

    (2012) Evaluating cellular polyfunctionality with a novel polyfunctionality index

    PLOS ONE 7:e42403.

    https://doi.org/10.1371/journal.pone.0042403

    • PubMed
    • Google Scholar
34. 1. Lee SK
    2. Silva DG
    3. Martin JL
    4. Pratama A
    5. Hu X
    6. Chang PP
    7. Walters G
    8. Vinuesa CG

    (2012) Interferon-γ excess leads to pathogenic accumulation of follicular helper T cells and germinal centers

    Immunity 37:880–892.

    https://doi.org/10.1016/j.immuni.2012.10.010

    • PubMed
    • Google Scholar
35. 1. Levine JH
    2. Simonds EF
    3. Bendall SC
    4. Davis KL
    5. Amir el AD
    6. Tadmor MD
    7. Litvin O
    8. Fienberg HG
    9. Jager A
    10. Zunder ER
    11. Finck R
    12. Gedman AL
    13. Radtke I
    14. Downing JR
    15. Pe'er D
    16. Nolan GP

    (2015) Data-Driven phenotypic dissection of AML reveals Progenitor-like cells that correlate with prognosis

    Cell 162:184–197.

    https://doi.org/10.1016/j.cell.2015.05.047

    • PubMed
    • Google Scholar
36. 1. Liu SM
    2. King C

    (2013) IL-21-producing th cells in immunity and autoimmunity

    The Journal of Immunology 191:3501–3506.

    https://doi.org/10.4049/jimmunol.1301454

    • PubMed
    • Google Scholar
37. 1. Llorente L
    2. Richaud-Patin Y
    3. García-Padilla C
    4. Claret E
    5. Jakez-Ocampo J
    6. Cardiel MH
    7. Alcocer-Varela J
    8. Grangeot-Keros L
    9. Alarcón-Segovia D
    10. Wijdenes J
    11. Galanaud P
    12. Emilie D

    (2000) Clinical and biologic effects of anti-interleukin-10 monoclonal antibody administration in systemic lupus erythematosus

    Arthritis & Rheumatism 43:1790–1800.

    https://doi.org/10.1002/1529-0131(200008)43:8<1790::AID-ANR15>3.0.CO;2-2

    • PubMed
    • Google Scholar
38. 1. Lüthje K
    2. Kallies A
    3. Shimohakamada Y
    4. Belz GT
    5. Light A
    6. Tarlinton DM
    7. Nutt SL

    (2012) The development and fate of follicular helper T cells defined by an IL-21 reporter mouse

    Nature Immunology 13:491–498.

    https://doi.org/10.1038/ni.2261

    • PubMed
    • Google Scholar
39. 1. Mair F
    2. Hartmann FJ
    3. Mrdjen D
    4. Tosevski V
    5. Krieg C
    6. Becher B

    (2016) The end of gating? an introduction to automated analysis of high dimensional cytometry data

    European Journal of Immunology 46:34–43.

    https://doi.org/10.1002/eji.201545774

    • Google Scholar
40. 1. Mateus J
    2. Guerrero P
    3. Lasso P
    4. Cuervo C
    5. González JM
    6. Puerta CJ
    7. Cuéllar A

    (2019) An animal model of acute and chronic chagas disease with the reticulotropic Y strain of Trypanosoma cruzi That Depicts the Multifunctionality and Dysfunctionality of T Cells

    Frontiers in Immunology 10:918.

    https://doi.org/10.3389/fimmu.2019.00918

    • PubMed
    • Google Scholar
41. 1. Matloubian M
    2. Cyster JG

    (2012) Th1 cell induction in lymph nodes according to a red-blue chemokine map

    Immunity 37:954–956.

    https://doi.org/10.1016/j.immuni.2012.11.007

    • PubMed
    • Google Scholar
42. 1. McKinney EF
    2. Lee JC
    3. Jayne DR
    4. Lyons PA
    5. Smith KG

    (2015) T-cell exhaustion, co-stimulation and clinical outcome in autoimmunity and infection

    Nature 523:612–616.

    https://doi.org/10.1038/nature14468

    • PubMed
    • Google Scholar
43. 1. Moulton VR
    2. Suarez-Fueyo A
    3. Meidan E
    4. Li H
    5. Mizui M
    6. Tsokos GC

    (2017) Pathogenesis of human systemic lupus erythematosus: a cellular perspective

    Trends in Molecular Medicine 23:615–635.

    https://doi.org/10.1016/j.molmed.2017.05.006

    • PubMed
    • Google Scholar
44. 1. Nakayamada S
    2. Kanno Y
    3. Takahashi H
    4. Jankovic D
    5. Lu KT
    6. Johnson TA
    7. Sun HW
    8. Vahedi G
    9. Hakim O
    10. Handon R
    11. Schwartzberg PL
    12. Hager GL
    13. O'Shea JJ

    (2011) Early Th1 cell differentiation is marked by a tfh cell-like transition

    Immunity 35:919–931.

    https://doi.org/10.1016/j.immuni.2011.11.012

    • PubMed
    • Google Scholar
45. 1. Newell EW
    2. Cheng Y

    (2016) Mass cytometry: blessed with the curse of dimensionality

    Nature Immunology 17:890–895.

    https://doi.org/10.1038/ni.3485

    • PubMed
    • Google Scholar
46. 1. Nurieva R
    2. Yang XO
    3. Martinez G
    4. Zhang Y
    5. Panopoulos AD
    6. Ma L
    7. Schluns K
    8. Tian Q
    9. Watowich SS
    10. Jetten AM
    11. Dong C

    (2007) Essential autocrine regulation by IL-21 in the generation of inflammatory T cells

    Nature 448:480–483.

    https://doi.org/10.1038/nature05969

    • Google Scholar
47. 1. Nusser A
    2. Nuber N
    3. Wirz OF
    4. Rolink H
    5. Andersson J
    6. Rolink A

    (2014) The development of autoimmune features in aging mice is closely associated with alterations of the peripheral CD4⁺ T-cell compartment

    European Journal of Immunology 44:2893–2902.

    https://doi.org/10.1002/eji.201344408

    • PubMed
    • Google Scholar
48. 1. O'Neill K
    2. Aghaeepour N
    3. Spidlen J
    4. Brinkman R

    (2013) Flow cytometry bioinformatics

    PLOS Computational Biology 9:e1003365.

    https://doi.org/10.1371/journal.pcbi.1003365

    • PubMed
    • Google Scholar
49. 1. Odegard JM
    2. Marks BR
    3. DiPlacido LD
    4. Poholek AC
    5. Kono DH
    6. Dong C
    7. Flavell RA
    8. Craft J

    (2008) ICOS-dependent extrafollicular helper T cells elicit IgG production via IL-21 in systemic autoimmunity

    Journal of Experimental Medicine 205:2873–2886.

    https://doi.org/10.1084/jem.20080840

    • PubMed
    • Google Scholar
50. 1. Peng SL
    2. Szabo SJ
    3. Glimcher LH

    (2002) T-bet regulates IgG class switching and pathogenic autoantibody production

    PNAS 99:5545–5550.

    https://doi.org/10.1073/pnas.082114899

    • PubMed
    • Google Scholar
51. 1. Pollard KM
    2. Cauvi DM
    3. Toomey CB
    4. Morris KV
    5. Kono DH

    (2013)

    Interferon-gamma and systemic autoimmunity

    Discovery Medicine 16:123–131.

    • PubMed
    • Google Scholar
52. 1. Qiu P
    2. Simonds EF
    3. Bendall SC
    4. Gibbs KD
    5. Bruggner RV
    6. Linderman MD
    7. Sachs K
    8. Nolan GP
    9. Plevritis SK

    (2011) Extracting a cellular hierarchy from high-dimensional cytometry data with SPADE

    Nature Biotechnology 29:886–891.

    https://doi.org/10.1038/nbt.1991

    • PubMed
    • Google Scholar
53. 1. Rao DA
    2. Gurish MF
    3. Marshall JL
    4. Slowikowski K
    5. Fonseka CY
    6. Liu Y
    7. Donlin LT
    8. Henderson LA
    9. Wei K
    10. Mizoguchi F
    11. Teslovich NC
    12. Weinblatt ME
    13. Massarotti EM
    14. Coblyn JS
    15. Helfgott SM
    16. Lee YC
    17. Todd DJ
    18. Bykerk VP
    19. Goodman SM
    20. Pernis AB
    21. Ivashkiv LB
    22. Karlson EW
    23. Nigrovic PA
    24. Filer A
    25. Buckley CD
    26. Lederer JA
    27. Raychaudhuri S
    28. Brenner MB

    (2017) Pathologically expanded peripheral T helper cell subset drives B cells in rheumatoid arthritis

    Nature 542:110–114.

    https://doi.org/10.1038/nature20810

    • PubMed
    • Google Scholar
54. 1. Rao DA

    (2018) T cells that help B cells in chronically inflamed tissues

    Frontiers in Immunology 9:1924.

    https://doi.org/10.3389/fimmu.2018.01924

    • PubMed
    • Google Scholar
55. 1. Saeys Y
    2. Van Gassen S
    3. Lambrecht BN

    (2016) Computational flow cytometry: helping to make sense of high-dimensional immunology data

    Nature Reviews Immunology 16:449–462.

    https://doi.org/10.1038/nri.2016.56

    • PubMed
    • Google Scholar
56. 1. Sauzullo I
    2. Scrivo R
    3. Mengoni F
    4. Ermocida A
    5. Coppola M
    6. Valesini G
    7. Vullo V
    8. Mastroianni CM

    (2014) Multi-functional flow cytometry analysis of CD4⁺ T cells as an immune biomarker for latent tuberculosis status in patients treated with tumour necrosis factor (TNF) antagonists

    Clinical & Experimental Immunology 176:410–417.

    https://doi.org/10.1111/cei.12290

    • PubMed
    • Google Scholar
57. 1. Seder RA
    2. Darrah PA
    3. Roederer M

    (2008) T-cell quality in memory and protection: implications for vaccine design

    Nature Reviews Immunology 8:247–258.

    https://doi.org/10.1038/nri2274

    • PubMed
    • Google Scholar
58. 1. Shen Y
    2. Chaigne-Delalande B
    3. Lee RWJ
    4. Losert W

    (2018) CytoBinning: Immunological insights from multi-dimensional data

    PLOS ONE 13:e0205291.

    https://doi.org/10.1371/journal.pone.0205291

    • Google Scholar
59. 1. Snapper CM
    2. Paul WE

    (1987) Interferon-gamma and B cell stimulatory factor-1 reciprocally regulate ig isotype production

    Science 236:944–947.

    https://doi.org/10.1126/science.3107127

    • PubMed
    • Google Scholar
60. 1. Song W
    2. Craft J

    (2019) T follicular helper cell heterogeneity: time, space, and function

    Immunological Reviews 288:85–96.

    https://doi.org/10.1111/imr.12740

    • PubMed
    • Google Scholar
61. 1. Spitzer MH
    2. Gherardini PF
    3. Fragiadakis GK
    4. Bhattacharya N
    5. Yuan RT
    6. Hotson AN
    7. Finck R
    8. Carmi Y
    9. Zunder ER
    10. Fantl WJ
    11. Bendall SC
    12. Engleman EG
    13. Nolan GP

    (2015) IMMUNOLOGY an interactive reference framework for modeling a dynamic immune system

    Science 349:1259425.

    https://doi.org/10.1126/science.1259425

    • PubMed
    • Google Scholar
62. 1. Suárez-Fueyo A
    2. Bradley SJ
    3. Tsokos GC

    (2016) T cells in systemic lupus erythematosus

    Current Opinion in Immunology 43:32–38.

    https://doi.org/10.1016/j.coi.2016.09.001

    • PubMed
    • Google Scholar
63. 1. Suto A
    2. Wurster AL
    3. Reiner SL
    4. Grusby MJ

    (2006) IL-21 inhibits IFN-gamma production in developing Th1 cells through the repression of eomesodermin expression

    The Journal of Immunology 177:3721–3727.

    https://doi.org/10.4049/jimmunol.177.6.3721

    • PubMed
    • Google Scholar
64. 1. Swain SL

    (1994) Generation and in vivo persistence of polarized Th1 and Th2 memory cells

    Immunity 1:543–552.

    https://doi.org/10.1016/1074-7613(94)90044-2

    • PubMed
    • Google Scholar
65. 1. Tilstra JS
    2. Avery L
    3. Menk AV
    4. Gordon RA
    5. Smita S
    6. Kane LP
    7. Chikina M
    8. Delgoffe GM
    9. Shlomchik MJ

    (2018) Kidney-infiltrating T cells in murine lupus nephritis are metabolically and functionally exhausted

    Journal of Clinical Investigation 128:4884–4897.

    https://doi.org/10.1172/JCI120859

    • PubMed
    • Google Scholar
66. 1. Trüb M
    2. Barr TA
    3. Morrison VL
    4. Brown S
    5. Caserta S
    6. Rixon J
    7. Ivens A
    8. Gray D

    (2017) Heterogeneity of phenotype and function reflects the multistage development of T follicular helper cells

    Frontiers in Immunology 8:489.

    https://doi.org/10.3389/fimmu.2017.00489

    • PubMed
    • Google Scholar
67. 1. Tsokos GC
    2. Lo MS
    3. Costa Reis P
    4. Sullivan KE

    (2016) New insights into the immunopathogenesis of systemic lupus erythematosus

    Nature Reviews Rheumatology 12:716–730.

    https://doi.org/10.1038/nrrheum.2016.186

    • PubMed
    • Google Scholar
68. 1. Utzschneider DT
    2. Charmoy M
    3. Chennupati V
    4. Pousse L
    5. Ferreira DP
    6. Calderon-Copete S
    7. Danilo M
    8. Alfei F
    9. Hofmann M
    10. Wieland D
    11. Pradervand S
    12. Thimme R
    13. Zehn D
    14. Held W

    (2016) T cell factor 1-Expressing Memory-like CD8(+) T cells sustain the immune response to chronic viral infections

    Immunity 45:415–427.

    https://doi.org/10.1016/j.immuni.2016.07.021

    • PubMed
    • Google Scholar
69. 1. Vinuesa CG
    2. Linterman MA
    3. Yu D
    4. MacLennan IC

    (2016) Follicular helper T cells

    Annual Review of Immunology 34:335–368.

    https://doi.org/10.1146/annurev-immunol-041015-055605

    • PubMed
    • Google Scholar
70. 1. von Spee-Mayer C
    2. Siegert E
    3. Abdirama D
    4. Rose A
    5. Klaus A
    6. Alexander T
    7. Enghard P
    8. Sawitzki B
    9. Hiepe F
    10. Radbruch A
    11. Burmester GR
    12. Riemekasten G
    13. Humrich JY

    (2016) Low-dose interleukin-2 selectively corrects regulatory T cell defects in patients with systemic lupus erythematosus

    Annals of the Rheumatic Diseases 75:1407–1415.

    https://doi.org/10.1136/annrheumdis-2015-207776

    • PubMed
    • Google Scholar
71. 1. Vu Van D
    2. Beier KC
    3. Pietzke LJ
    4. Al Baz MS
    5. Feist RK
    6. Gurka S
    7. Hamelmann E
    8. Kroczek RA
    9. Hutloff A

    (2016) Local T/B cooperation in inflamed tissues is supported by T follicular helper-like cells

    Nature Communications 7:10875.

    https://doi.org/10.1038/ncomms10875

    • PubMed
    • Google Scholar
72. 1. Walunas TL
    2. Lenschow DJ
    3. Bakker CY
    4. Linsley PS
    5. Freeman GJ
    6. Green JM
    7. Thompson CB
    8. Bluestone JA

    (1994) CTLA-4 can function as a negative regulator of T cell activation

    Immunity 1:405–413.

    https://doi.org/10.1016/1074-7613(94)90071-X

    • PubMed
    • Google Scholar
73. 1. Wang L
    2. Zhao P
    3. Ma L
    4. Shan Y
    5. Jiang Z
    6. Wang J
    7. Jiang Y

    (2014) Increased interleukin 21 and follicular helper T-like cells and reduced interleukin 10+ B cells in patients with new-onset systemic lupus erythematosus

    The Journal of Rheumatology 41:1781–1792.

    https://doi.org/10.3899/jrheum.131025

    • PubMed
    • Google Scholar
74. 1. Weinstein JS
    2. Herman EI
    3. Lainez B
    4. Licona-Limón P
    5. Esplugues E
    6. Flavell R
    7. Craft J

    (2016) TFH cells progressively differentiate to regulate the germinal center response

    Nature Immunology 17:1197–1205.

    https://doi.org/10.1038/ni.3554

    • PubMed
    • Google Scholar
75. 1. Westerhof LM
    2. McGuire K
    3. MacLellan L
    4. Flynn A
    5. Gray JI
    6. Thomas M
    7. Goodyear CS
    8. MacLeod MK

    (2019) Multifunctional cytokine production reveals functional superiority of memory CD4 T cells

    European Journal of Immunology 49:2019–2029.

    https://doi.org/10.1002/eji.201848026

    • PubMed
    • Google Scholar
76. 1. Wong MT
    2. Chen J
    3. Narayanan S
    4. Lin W
    5. Anicete R
    6. Kiaang HT
    7. De Lafaille MA
    8. Poidinger M
    9. Newell EW

    (2015) Mapping the diversity of follicular helper T cells in human blood and tonsils using High-Dimensional mass cytometry analysis

    Cell Reports 11:1822–1833.

    https://doi.org/10.1016/j.celrep.2015.05.022

    • PubMed
    • Google Scholar
77. 1. Wong MT
    2. Ong DE
    3. Lim FS
    4. Teng KW
    5. McGovern N
    6. Narayanan S
    7. Ho WQ
    8. Cerny D
    9. Tan HK
    10. Anicete R
    11. Tan BK
    12. Lim TK
    13. Chan CY
    14. Cheow PC
    15. Lee SY
    16. Takano A
    17. Tan EH
    18. Tam JK
    19. Tan EY
    20. Chan JK
    21. Fink K
    22. Bertoletti A
    23. Ginhoux F
    24. Curotto de Lafaille MA
    25. Newell EW

    (2016) A High-Dimensional atlas of human T cell diversity reveals Tissue-Specific trafficking and cytokine signatures

    Immunity 45:442–456.

    https://doi.org/10.1016/j.immuni.2016.07.007

    • PubMed
    • Google Scholar
78. 1. Wurster AL
    2. Rodgers VL
    3. Satoskar AR
    4. Whitters MJ
    5. Young DA
    6. Collins M
    7. Grusby MJ

    (2002) Interleukin 21 is a T helper (Th) cell 2 cytokine that specifically inhibits the differentiation of naive th cells into interferon gamma-producing Th1 cells

    Journal of Experimental Medicine 196:969–977.

    https://doi.org/10.1084/jem.20020620

    • PubMed
    • Google Scholar
79. 1. Zeng R
    2. Spolski R
    3. Casas E
    4. Zhu W
    5. Levy DE
    6. Leonard WJ

    (2007) The molecular basis of IL-21-mediated proliferation

    Blood 109:4135–4142.

    https://doi.org/10.1182/blood-2006-10-054973

    • PubMed
    • Google Scholar

Acceptance summary:

Some T helper cell populations have been shown to expand in autoimmune diseases such as systemic lupus erythematosus and might thus contribute to the unfolding of the disease condition. To zoom on those T cells that have phenotype very similar to other T cell subpopulations, the authors combined a bin-based "pattern recognition of immune cells" strategy with comprehensive cytometric measurements of T helper cells from found in lupus-prone mice. On that basis, they defined a novel subset of super functional T helper cells that exhibit superior B cell help and may constitute therapeutic targets.

Decision letter after peer review:

Thank you for submitting your article "Identification of a super-functional Tfh-like subpopulation in murine lupus by pattern perception" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Satyajit Rath as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Joe Craft (Reviewer #1).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

The authors used CytoBinning, a recently described flow cytometry analysis approach, to explore the phenotype and cytokine expression profile of splenic memory CD4⁺ T cells from NZBxW mice at different stages of disease. The analysis approach combines automation of a traditional workflow and machine learning, linking high dimensional data back to two biomarkers which can be represented as 2D scatter plots and enables the identification of subtle shifts in immune phenotypes. Using this approach, the authors identify a subpopulation of IL21⁺ IFNγ⁺ PD1^low CD40L^high CXCR5^- Bcl6^- CD4⁺ T cells expanded in older, diseased mice that expresses high levels of TNFα and IL2. Considering that in co-culture experiments they observe that it provides B cell help though IL21 and CD40L, the authors define this population as a "super-functional Tfh-like cells". Overall, this interesting manuscript surmises that these newly identified cells are critical for the disease phenotype in lupus, including promotion of autoantibody production.

As a major aim of the manuscript is the presentation of a new method to analyse flow cytometry data, further information should be made available to allow a direct comparison of this new method with previous methods. Moreover, the "super-functional Tfh-like cells" should be more precisely compared to the T peripheral helper (Tph) cells, first described in rheumatoid arthritis by Rao and colleagues (Rao et al., 2017), and this year in SLE (Bocharnikov et al., 2019), celiac disease (Christophersen et al., 2019) and T1DM (Ekman et al., 2019).

Essential revisions:

1) Methodological issues.

1. 1) While the insights revealed using the combinatorial approach with data binning are of interest, sifting through the data is more challenging than it need be. Certain panels of figures are mislabeled (2H does not exist, 3C or D is incorrectly noted, and so forth), and the authors are asked to proof the text and figures. As well, Figure 2E and Figure 2—figure supplement 1B are nigh on impossible to parse, given overlap in colors (not even accounting for red-green discernment). In a like vein, certain figures are challenging to interpret, given the light shades of gray that appear to overlap. The concerns stand out all the more, given the nice visuals of the binning data.

1.2) Each PRI plot superimposes data from several samples (biological replicates). While the authors show that data are reproducible for individual mice, it is likely that for some markers or some biological experiments, great inter-individual variation can occur. The provided statistics fail to represent this issue. A better clarification on how PRI plots can be used to show the range of variability among a population should be provided.

1.3) The audience of this manuscript will be far more familiar with traditional methods of FACS analysis than PRI. As a consequence, in order to establish the robustness of PRI-based quantification and familiarity with the approach, Figure 3A should include also samples from young mice. It seems from the analysis of the flow cytometry plots of old mice (Figure 3A) that all IL-21⁺ cells co-express IL-2, TNF, and IFN while being CXCR5^-, PD-1int, and IL-10^-. However, PRI assessment of the same samples (Figure 3F) provides a visual representation suggesting that the vast majority of IFNγ⁺ cells are also producing IL-21, while in the dot plots (Figure 3A) only ~1/7 of IFNγ⁺ cells produce IL-21. Furthermore, looking at the PRI of young mice it is unclear whether the IL-21 producing cells maintain the same phenotype as the cells from old mice, like the authors claim (subsection “Bin plots revealed an IL-21⁺ super-functional T cell subpopulation”). For all of the above Figure 3A should allow independent assessment of young/old phenotype in addition to PRI plots of Figure 3F.

1.4) Additionally, the most biologically relevant statistical assessment is in respect to cell populations (defined to the best of our ability) rather than individual markers. If the authors claim the importance of a novel cell subset defined as IL-21⁺, IFNγ⁺… the relevant statistical analysis of Figure 3E should be based on the frequency of that cell subset in young vs. old mice rather than (or in addition to) the presented analysis based on individual markers.

1.5) The phenotype of IL-21-producing cells displayed in Figure 4 is not easy to interpret given the poor identification of the IL21⁺ cells, as they are contained in bins together with other cells. Lack of statistics is also a concern. As a key objective of the manuscript is to show how PRI compares with FACS analysis, the figure would improve with traditional dot plots of gated IL-21⁺ cells (or even how IL21⁺IFNγ⁺⁺ compared with IL-21^-IFN⁺⁺ for the different markers – to a large extent these two cell populations are in the same bin region). Statistics could then be provided and compared with PRI data.

1.6) It is also puzzling why the authors do not compare the phenotype of IL-21⁺ cells and IL-21^-IFN⁺ and IL21^-IFN^- instead of a complex matrix of comparisons based on IFNγ and PD-1 expression (Figures 5, 6 and Figure 6—figure supplement 1). A major goal of the manuscript is to define a IL-21⁺IFNγ⁺ cell population that can be easily gated. Data should be provided for the population of interest in addition to arbitrary quadrants.

1.7) Figure 2C. How was the determination of PD-1^-, PD-1^low and PD-1 positivity determined? The gating does not follow the cell density plots.

1.8) The authors note the data are reproducible. but how was reproducibility shown statistically?

2) Biological issues

2.1) The authors focused only on the "super-functional Tfh-like cells", even though in the sicker mice there is not only and expansion of IL21⁺ IFNγ⁺ PD1^low cells but also of IL21⁺ IFNγ⁺ PD1^high cells (Figure 3E) as well IL10⁺ IFNγ⁺ PD1^high cells. To prove that this IL21⁺ INFγ⁺ PD1^low CD40L^high CXCR5^- Bcl6^- T cell population helps B cells though IL21 and CD40L, they sort non Tfh-Th1 like cells (CD4⁺, CXCR5^-, CD44^hi, CXCR3⁺ cells) and check their ability to support B cell activation and Ig production. Two main concerns arise from these set of experiments. First, the sorting strategy should include PD1 in order to enrich the non Tfh-Th1 like mem T cells (CD4⁺, CXCR5^-, CD44^hi, CXCR3⁺) within the PD1^low population (CD4⁺, PD1^low, CXCR5^-, CD44^hi, CXCR3⁺ cells). Second, the B cell helper assay should be repeated and compared with other cell populations, including conventional Tfh1 cells (CD4⁺, PD1^hi, CXCR5⁺, CD44^hi, CXCR3⁺ cells) and include naïve versus total B cells. In addition, the authors should perform a more thorough characterization of the main three expanded cell subpopulations by RNAseq to better define the uniqueness of the novel "super- functional" cells.

2.2) The authors claim IFNγ⁺IL-21-producing cells represent a novel population distinct from previous defined populations with "super-functional" properties.

2.2.1) One cell population that has been shown important in human SLE and bears some similarities was not discussed: T peripheral helper (Tph) cells, first described in rheumatoid arthritis by Rao and colleagues (Rao et al., 2017), and this year in SLE (Bocharnikov et al., 2019), in celiac disease (Christophersen et al., 2019) and T1DM (Ekman et al., 2019). These cells are CXCR5^- Bcl6^- and characterized by IL-21 production, a critical difference to the population discussed in this manuscript is that Tph are PD-1^hi. However, if the two populations are different, they should both be present. Do the authors claim the only T cell population producing IL-21 is the newly discovered? Is there an explanation for poor identification of other IL-21-producing T cells among CXCR5^- cells, namely Tph?

2.2.2) Considering that the super-functional T cell population identified by the authors is similar to Tph cells, it would be of interest to know if the super-functional T cells are Blimp1⁺, which is likely the case, recognizing the potential limitation of flow staining for this transcriptional repressor in mouse.

2.2.3) While the T-B cell helper assay adds a functional component to the paper, how physiological this experiment really is? Do the authors think that their T cells drive "naïve" B cells to produce Ig in vivo? Do super-functional T cells come into contact with naïve B cells in secondary lymphoid organs or elsewhere? Wouldn't naïve B cells need an IgM as well as TLR and/or cytokine signals first, to move into proximity with activated T cells?

2.2.4) The functional assay cannot be easily interpreted with the available information. Since sorting was done with an indirect strategy (based on CXCR3⁺CXCR5^- T cells) it will be important to show what is the frequency of the IL-21⁺IFNγ⁺PD-1int cells among the sorted population. A positive control with bona fide CXCR5⁺PD1^hi Tfh cells should be provided to compare the potency of Tfh and the claimed "super-functional" population. Titrations may be provided to assess relative potency, if needed. It is hard to claim super-functional properties without a functional comparison with the current standard of B cell help (i.e. Tfh cells). FACS plots with B cells stained with anti-GL7 and anti-IgG1 could also make the results more compelling, given the low levels of IgG antibodies produced.

2.2.5) Frequently, BCR agonists (anti-IgM for instance) are provided in addition to T cell stimulation for in vitro assays (see for instance Sage, Nature Immunol 2016 17:1436), this may improve the assay read-out.

2.2.6) More clarity about disease definition would be helpful. "Chronic inflammation, high proteinuria", is not so specific, not is age of mice. Likewise, were autoantibodies and renal histology assessed?

Summary:

The authors used CytoBinning, a recently described flow cytometry analysis approach, to explore the phenotype and cytokine expression profile of splenic memory CD4⁺ T cells from NZBxW mice at different stages of disease. […] Overall, this interesting manuscript surmises that these newly identified cells are critical for the disease phenotype in lupus, including promotion of autoantibody production.

Binning is a way of discretizing data and can be done by adaptive binning (e.g. quantiles) or fixed-width binning (e.g. rounding). The mentioned CytoBinning approach, first described by Roederer et al., 2001 (Cytometry 2001, 45:37), use the percentile-based binning. In contrast, our approach uses fixed-width bins with a varying number of cells. This visualisation was chosen, because the cells tend to accumulate near the center of the data range but flatten out fast near the boundary of the range. A percentile-based binning would emphasize on the cell distribution. In contrast, we focus on the expression distribution and pattern mainly at the range boundaries. We do not only use the regular cell count, but we particularly make use of the mean fluorescence intensities and the frequency of positive cells of an additional third marker or several markers to provide valuable additional information. This is the combinatorics of 3 or more markers. Furthermore, a fixed-width binning allows for a good intuitive comparison of different samples.

As a major aim of the manuscript is the presentation of a new method to analyse flow cytometry data, further information should be made available to allow a direct comparison of this new method with previous methods.

We accepted the advice and integrated additional conventional analyses into several figures. In Figure 3A and B we included all respective contour plots for young mice. We confirmed the bin plot data of new Figure 3F and G by conventional gating analysis shown in Figure 3A and B, and the results of statistical analysis in Figure 3C. For comparison we show the statistical analysis from the bin plots in Figure 3—figure supplement 1B.

Moreover, the "super-functional Tfh-like cells" should be more precisely compared to the T peripheral helper (Tph) cells, first described in rheumatoid arthritis by Rao and colleagues (Rao et al., 2017), and this year in SLE (Bocharnikov et al., 2019), celiac disease (Christophersen et al., 2019) and T1DM (Ekman et al., 2019).

Thank you for the recommendation. Tfh and Tph cells are both characterized by very high PD-1 expression (which distinguishes them from the PD-1^low super-functional Tfh-like (Tsh) cells), high production of IL-21, and either high or absent expression of CXCR5. We included new data and figures for these three markers (Figure 3C lower row, Figure 3G, Figure 3—figure supplement 1B and Figure 5D). In addition, we inserted additional references into the main text of the manuscript.

Essential revisions:

1) Methodological issues.

1. 1) While the insights revealed using the combinatorial approach with data binning are of interest, sifting through the data is more challenging than it need be. Certain panels of figures are mislabeled (2H does not exist, 3C or D is incorrectly noted, and so forth), and the authors are asked to proof the text and figures.

Sincerely sorry for the mistake. We corrected it.

As well, Figure 2E and Figure 2—figure supplement 1B are nigh on impossible to parse, given overlap in colors (not even accounting for red-green discernment). In a like vein, certain figures are challenging to interpret, given the light shades of gray that appear to overlap. The concerns stand out all the more, given the nice visuals of the binning data.

We totally see the point. We improved the colours and the graphics for both pie charts. For simplicity, we moved the pie chart (former Figure 2E) to Figure 2—figure supplement 2E).

1.2) Each PRI plot superimposes data from several samples (biological replicates). While the authors show that data are reproducible for individual mice, it is likely that for some markers or some biological experiments, great inter-individual variation can occur. The provided statistics fail to represent this issue. A better clarification on how PRI plots can be used to show the range of variability among a population should be provided.

In the original version of the manuscript, each bin plot visualised the cytometric data of a single mouse. To minimize inter-individual deviations, data from different mice were combined after measurement and were visualised in one bin plot. Therefore, for the new Figure 3 pictures we used concatenated cells from 3 mice each (young and old diseased). To exemplary demonstrate the robustness of the method, we compared the pattern of concatenated cells from the new experiments with patterns of 4 single mice from older experiments (Figure 2—figure supplement 3B). Thus we inserted the new Figure 2—figure supplement 3C.

1.3) The audience of this manuscript will be far more familiar with traditional methods of FACS analysis than PRI. As a consequence, in order to establish the robustness of PRI-based quantification and familiarity with the approach, Figure 3A should include also samples from young mice.

We included conventional plots of young mice into Figure 3A and B.

It seems from the analysis of the flow cytometry plots of old mice (Figure 3A) that all IL-21⁺ cells co-express IL-2, TNF, and IFN while being CXCR5^-, PD-1int, and IL-10^-. However, PRI assessment of the same samples (Figure 3F) provides a visual representation suggesting that the vast majority of IFNγ⁺ cells are also producing IL-21, while in the dot plots (Figure 3A) only ~1/7 of IFNγ⁺ cells produce IL-21.

We admit that too many frequencies given in the old Figure 3F were confusing. Therefore, we depicted here (Author response image 1) and in the novel Figure 3F and G only the most important green numbers (Z⁺ producers of all CD44⁺ cells in the respective quadrants).

The old figures showed that 5.7% (Author response image 1 left picture) and 6.8% (Author response image 1 right picture) of CD44⁺ cells are IL21⁺IFN-γ⁺ (DP) cells and 5.4% are triple-producers (with PD-1⁺). Most of the IFN-γ producing cells (32%) were IL-21^- and PD-1⁺.

Author response image 1

Download asset Open asset

Furthermore, looking at the PRI of young mice it is unclear whether the IL-21 producing cells maintain the same phenotype as the cells from old mice, like the authors claim (subsection “Bin plots revealed an IL-21⁺ super-functional T cell subpopulation”). For all of the above Figure 3A should allow independent assessment of young/old phenotype in addition to PRI plots of Figure 3F.

We included plots of young mice into Figure 3A and G.

The frequencies of IL-21⁺ and IL21⁺IFN-γ⁺ among the Tmem cells (CD44⁺) are not significantly different between young and old mice but show a slight tendency to be a bit higher in old diseased mice as shown both with PRI-data (Figure 3—figure supplement 1B) and FlowJo-data (Figure 3C). However, the properties of these cells to co-produce PD-1 increased significantly (Figure 3—figure supplement 1C).

1.4) Additionally, the most biologically relevant statistical assessment is in respect to cell populations (defined to the best of our ability) rather than individual markers. If the authors claim the importance of a novel cell subset defined as IL-21⁺, IFNγ⁺… the relevant statistical analysis of Figure 3E should be based on the frequency of that cell subset in young vs. old mice rather than (or in addition to) the presented analysis based on individual markers.

In fact, this consideration is really logical. Therefore, we provided two new experiments (with 10 mice in total) and included all relevant markers within one staining panel. This allowed us to analyse in parallel the subsets Tfh, Tph, and Tsh. The new Figure 3F, G, and Figure 3—figure supplement 1B show the PRI data of concatenated measured cells (3 mice each for young and old mice). The conventional flow cytometric analysis is shown in Figure 3C.

1.5) The phenotype of IL-21-producing cells displayed in Figure 4 is not easy to interpret given the poor identification of the IL21⁺ cells, as they are contained in bins together with other cells.

We agree to this. However, we experienced that taking IFN-γ instead of IL-21 is better for visualisation of preferred expression areas of the most Z-parameters. The range of IFN-γ is much greater than the range from IL-21. Therefore, tendencies and differences in the populations are better visualised as can be seen e.g. for CD40L intensities (Figure 4—figure supplement 1A) which correlates with the IFN-γ intensities.

Lack of statistics is also a concern. As a key objective of the manuscript is to show how PRI compares with FACS analysis, the figure would improve with traditional dot plots of gated IL-21⁺ cells (or even how IL21⁺IFNγ⁺⁺ compared with IL-21^-IFN⁺⁺ for the different markers – to a large extent these two cell populations are in the same bin region). Statistics could then be provided and compared with PRI data.

The new Figure 3C and the new Figure 3—figure supplement 1B, offer the possibility of a direct comparison of the statistical data obtained by conventional gating and PRI analysis, respectively.

1.6) It is also puzzling why the authors do not compare the phenotype of IL-21⁺ cells and IL-21^-IFN⁺ and IL21^-IFN^- instead of a complex matrix of comparisons based on IFNγ and PD-1 expression (Figures 5, 6 and Figure 6—figure supplement 1). A major goal of the manuscript is to define a IL-21⁺IFNγ⁺ cell population that can be easily gated. Data should be provided for the population of interest in addition to arbitrary quadrants.

The results from subsequent gating are shown in the new Figure 3C.

1.7) Figure 2C. How was the determination of PD-1^-, PD-1^low and PD-1 positivity determined? The gating does not follow the cell density plots.

For PD-1^+/- gating we used the FMO controls. For determination of PD-1^high, we took the decline in co-expression of PD-1^high and IL-2 intensities in old diseased mice.

1.8) The authors note the data are reproducible. but how was reproducibility shown statistically?

Plotting the same data file produces the same pattern and numbers (100% identical) if all parameters, such as the defined threshold value of the parameter Z, the minimum number of cells per bin and the bin width are maintained. By applying different threshold values for parameter Z (IL-10) as shown in Figure 2—figure supplement 3D, the reviewer can get an impression about the changes of the pattern.

2) Biological issues

2.1) The authors focused only on the "super-functional Tfh-like cells", even though in the sicker mice there is not only and expansion of IL21⁺ IFNγ⁺ PD1^low cells but also of IL21⁺ IFNγ⁺ PD1^high cells (Figure 3E) as well IL10⁺ IFNγ⁺ PD1^high cells.

Yes, the reviewer is right (seen also in the new Figure 3C), there are also more IL21⁺ IFNγ⁺ PD1^high and IL10⁺ IFNγ⁺ PD1^high cells in old mice compared to young mice. However, the majority of IL-21⁺ cells are PD-1^low and IFNγ⁺ , which defines a previously unknown subset. Therefore, we concentrated on characterisation of this novel cell subset.

To prove that this IL21⁺ INFγ⁺ PD1^low CD40L^high CXCR5^- Bcl6^- T cell population helps B cells though IL21 and CD40L, they sort non Tfh-Th1 like cells (CD4⁺, CXCR5^-, CD44^hi, CXCR3⁺ cells) and check their ability to support B cell activation and Ig production. Two main concerns arise from these set of experiments. First, the sorting strategy should include PD1 in order to enrich the non Tfh-Th1 like mem T cells (CD4⁺, CXCR5^-, CD44^hi, CXCR3⁺) within the PD1^low population (CD4⁺, PD1^low, CXCR5^-, CD44^hi, CXCR3⁺ cells). Second, the B cell helper assay should be repeated and compared with other cell populations, including conventional Tfh1 cells (CD4⁺, PD1^hi, CXCR5⁺, CD44^hi, CXCR3⁺ cells) and include naïve versus total B cells.

To address this concern, we conducted a new set of T/B co-culture assays with cells sorted according to their PD-1 expression.

We performed new T/B co-culture assays to compare directly the B helper abilities of Tsh cells defined as CD4⁺, PD1^low CXCR5^- CD44⁺ CXCR3⁺ cells, to that of CD4⁺PD1^low CXCR5^- CD44⁺ CXCR3^- cells and PD-1^high cells, defined as CD4⁺PD1^high CXCR5^+/- CD44⁺ containing both Tfh and Tph cells. These experiments are reported in Figure 7C and Figure 7—figure supplement 2. Due to the rarity of Tfh and Tph populations and the limited availability of old enough NZBxW F1 mice, we had to resort to aged (15-24 months) C57Bl/6 mice known to have increased autoimmune features (see for example Nusser et al., 2014). We had found that 24-month old C57Bl/6 mice presented with increased numbers of CD44⁺ cells and Tsh with the same features as sick NZBxW F1 mice. The strategy to sort the compared populations, as well as their respective productions of CD40L, IL-21 and IFN-g are presented in Figure 7—figure supplement 2A and C.

The limiting numbers of PD-1^hi cells restricted our study to the use of total B cells defined as CD19⁺ B220⁺ cells (Figure 7—figure supplement 2B).

These experiments clearly show dichotomic roles for Tsh and PD-1^high cells. While Tsh cells promote higher B cells expansion and GC-like differentiation of the B cells (GL7⁺ Fas⁺), PD-1^high cells promote cell survival and sustains better plasma cell differentiation (Figure 7C and Figure 7—figure supplement 2D). These results showing the superiority of Tsh cells in providing help for B cell proliferation is now described and discussed in the manuscript.

In addition, the authors should perform a more thorough characterization of the main three expanded cell subpopulations by RNAseq to better define the uniqueness of the novel "super-functional" cells.

We agree that RNAseq data might reveal additional unique markers of this novel subset. Here, we rather concentrated on comprehensive protein pattern (in particular cytokine pattern) characterisation. We feel that this analysis already provided unique markers not only for definition of this subset but also unique functional properties.

2.2) The authors claim IFNγ⁺IL-21-producing cells represent a novel population distinct from previous defined populations with "super-functional" properties.

2.2.1) One cell population that has been shown important in human SLE and bears some similarities was not discussed: T peripheral helper (Tph) cells, first described in rheumatoid arthritis by Rao and colleagues (Rao et al., 2017), and this year in SLE (Bocharnikov et al., 2019), in celiac disease (Christophersen et al., 2019) and T1DM (Ekman et al., 2019). These cells are CXCR5^- Bcl6^- and characterized by IL-21 production, a critical difference to the population discussed in this manuscript is that Tph are PD-1^hi. However, if the two populations are different, they should both be present. Do the authors claim the only T cell population producing IL-21 is the newly discovered? Is there an explanation for poor identification of other IL-21-producing T cells among CXCR5^- cells, namely Tph?

The reviewer is right, both cell subsets Tph and Tsh are increased in aged NZBW mice and produce IL-21. To analyse Tph cells in parallel to Tfh and Tsh cells, we performed 2 new experiments which included staining for CXCR5 (seen in Figure 3). In addition we re-analysed data of Figure 5 to directly compare the frequencies of both IL-21-producing subsets. Tsh cells are more than twice the number as Tph cells in the NZBW model.

2.2.2) Considering that the super-functional T cell population identified by the authors is similar to Tph cells, it would be of interest to know if the super-functional T cells are Blimp1⁺, which is likely the case, recognizing the potential limitation of flow staining for this transcriptional repressor in mouse.

We tried Blimp1⁺ staining, but were not successful. An example is shown where we compared staining obtained for PD-1^low CXCR5^- IFN-g⁺ IL-21⁺ (Tsh) cells, PD-1^high CXCR5⁺ (Tfh) cells and PD-1^high CXCR5^- cells. Compared to FMO staining, the antibody gave a similar signal for the entire population, including supposedly negative Tfh-cells, whereas T-bet staining as control gave a differential signal.

Author response image 2

Download asset Open asset

2.2.3) While the T-B cell helper assay adds a functional component to the paper, how physiological this experiment really is? Do the authors think that their T cells drive "naïve" B cells to produce Ig in vivo? Do super-functional T cells come into contact with naïve B cells in secondary lymphoid organs or elsewhere? Wouldn't naïve B cells need an IgM as well as TLR and/or cytokine signals first, to move into proximity with activated T cells?

We agree with the reviewer. However, we would like to point out that the coculture experiments have now been repeated with total B cells and that it is quite impressive that Tsh cells can help naive B cells.

2.2.4) The functional assay cannot be easily interpreted with the available information. Since sorting was done with an indirect strategy (based on CXCR3⁺CXCR5^- T cells) it will be important to show what is the frequency of the IL-21⁺IFNγ⁺PD-1int cells among the sorted population. A positive control with bona fide CXCR5⁺PD1^hi Tfh cells should be provided to compare the potency of Tfh and the claimed "super-functional" population. Titrations may be provided to assess relative potency, if needed. It is hard to claim super-functional properties without a functional comparison with the current standard of B cell help (i.e. Tfh cells). FACS plots with B cells stained with anti-GL7 and anti-IgG1 could also make the results more compelling, given the low levels of IgG antibodies produced.

A new set of co-culture assays with a sorting strategy based on the expression of PD-1 and CXCR3 (Figure 7C and Figure 7—figure supplement 2) addresses the reviewer’s concern. This new test confirmed the high helper functions of PD-1^low CXCR3⁺. Moreover, we characterized the respective cytokines and CD40L productions of the sorted CD4 T cell subsets to allow a direct correlation between their profiles and helper capacity (Figure 7—figure supplement 2C). These experiments provide a solid proof that Tsh cells have superior helper ability to support B cell proliferation over Tfh, suggesting a role in the production of self-reactive B cells in autoimmune contexts.

We took the advice of the reviewer and analysed the co-cultured B cells at the endpoint of the We took the advice of the reviewer and analyzed the co-cultured B cells at the endpoint of the experiment (day 5) by flow cytometry, using GL7 and Fas as marker of activated GC-like cells, and CD138 as an alternative to IgG1 for definition of antibody-producing cells. We preferred CD138 over IgG1 because IgG1 is mainly produced in Th2 contexts and Tph display a more Th1 phenotype with IFN-g, which induces diverse Ig class switching. Measuring IgG1 may have led to underestimate the production of Ig-producing cells. This new readout highlighted differences in the quality of T cell help provided by the different subsets, but also provided an explanation for the low levels of IgG detected by ELISA, given the low percentage of CD138⁺ cells.

2.2.5) Frequently, BCR agonists (anti-IgM for instance) are provided in addition to T cell stimulation for in vitro assays (see for instance Sage, Nature Immunol 2016 17:1436), this may improve the assay read-out.

Thank you for the advice. However, because we worked with limiting numbers of T cells, we could not endeavour to set an experimental system in which T help and BCR agonists are mutually titrated to optimize the co-culture assay. We plan to explore Tsh helper functions in antigen-dependent contexts involving BCR engagement in the future.

2.2.6) More clarity about disease definition would be helpful. "Chronic inflammation, high proteinuria", is not so specific, not is age of mice. Likewise, were autoantibodies and renal histology assessed?

Disease definition concerning age, proteinuria, and auto-antibodies is depicted in Figure 2A and in the main text. Auto-antibodies increase with age and disease score. Renal histology was not assessed.

Author details

1. Stefanie Gryzik

   German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany

   Contribution

   Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft

   Contributed equally with

   Yen Hoang

   For correspondence

   s_gryzik@web.de

   Competing interests

   No competing interests declared

2. Yen Hoang

   1. German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany
   2. University of Potsdam, Potsdam, Germany

   Contribution

   Conceptualization, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft

   Contributed equally with

   Stefanie Gryzik

   For correspondence

   yen.hoang@drfz.de

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-8956-1709

3. Timo Lischke

   German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany

   Contribution

   Conceptualization, Investigation, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0413-4252

4. Elodie Mohr

   German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany

   Contribution

   Formal analysis, Investigation, Methodology, Writing - review and editing

   For correspondence

   elodie.mohr@drfz.de

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-3406-7302

5. Melanie Venzke

   German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany

   Contribution

   Resources

   Competing interests

   No competing interests declared

6. Isabelle Kadner

   1. German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany
   2. University of Potsdam, Potsdam, Germany

   Contribution

   Software

   Competing interests

   No competing interests declared

7. Josephine Poetzsch

   1. German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany
   2. University of Potsdam, Potsdam, Germany

   Contribution

   Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4238-3866

8. Detlef Groth

   University of Potsdam, Potsdam, Germany

   Contribution

   Methodology

   Competing interests

   No competing interests declared

9. Andreas Radbruch

   1. German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany
   2. Charité, Campus Mitte, Berlin, Germany

   Contribution

   Funding acquisition

   Competing interests

   No competing interests declared

10. Andreas Hutloff

    German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany

    Contribution

    Supervision, Methodology, Writing - review and editing

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-0572-8151

11. Ria Baumgrass

    1. German Rheumatism Research Center (DRFZ), A Leibniz Institute, Berlin, Germany
    2. University of Potsdam, Potsdam, Germany

    Contribution

    Conceptualization, Resources, Supervision, Funding acquisition, Investigation, Writing - original draft, Project administration

    For correspondence

    baumgrass@drfz.de

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-3289-1608

• 1,982

  Page views

• 261

  Downloads

• 5

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.