Figure E1. Extended gating strategy for T and B cell subpopulations. (A) Whole blood was processed and analyzed via mass cytometry as in Figure 1. CD4, CD8, CD45RA, CD45RO, CD27, and CD57 were used to subset T cells. (B) CD19, HLADR, and CD27 were used to subset B cells. Representative data from one donor are shown.

Figure E2. Signaling network heatmap of TLR signaling in myeloid and lymphoid cells. Following whole blood stimulation with the indicated TLR ligands, cells were processed as described in Figure 1. Single-cell signaling differences between unstimulated and stimulated conditions were derived and organized as described in Figure 2A-2D. Responses from an additional donor, distinct from the donor in Figure 2E, are shown.

Figure E3. Variance in signaling responses to TLR stimuli in (A) myeloid and (B) lymphoid cell subsets. Following whole blood stimulation with the indicated TLR ligands, cells were processed as described in Figure 1. (A) Myeloid and (B) lymphoid cell populations were defined as in Figure 1. Median count values from unstimulated and stimulated cell populations were transformed on an arcsinh scale and subtracted to create a response index (arcsinh difference). The absolute value of arcsinh difference was used to place phosphorylation and degradation responses on the same scale. Average responses from six donors + SD are shown.

Figure E4. Signaling network heatmap of TLR signaling patterns in lymphoid cells. Following whole blood stimulation with the indicated TLR ligands, cells were processed as described in Figure 1. Single-cell signaling differences between unstimulated and stimulated states were derived and visualized as described in Figure 2A-2D, from data from same donor as in Figure 2E. Signaling responses to TLR ligands were observed only in B cells, T cells, and NK cells.

Figure E5. Single-cell signaling analysis of NK cell responses to lipopeptides. (A) Whole blood was stimulated with PAM2, PAM3, or control saline buffer for 30mins prior to flow cytometric analysis. A subpopulation of CD56hi NK cells induces IκBα degradation as well as CREB and p38 phosphorylation in response to lipopeptide stimulation. Representative data from one donor are shown. (B) Average NK cell responses from ten donors + SD are shown.

Figure E6. Single-cell signaling analysis of purified NK cell responses to lipopeptides. Whole blood and purified NK cells from the same healthy donor were stimulated with PAM2, PAM3, or control saline buffer for 30mins prior to flow cytometric analysis. (A) NK cells were isolated via RosetteSepTM, followed by magnetic bead enrichment for CD56 positive selection (supplementary methods), yielding a 96.75% pure population. The NK cell population constitutes 0.75% of peripheral whole blood lymphocytes in the same healthy volunteer donor. (B) CD56hi NK cell population, as gated from whole blood or purified NK cells, induce IκBα degradation but no pErk phosphorylation in response to lipopeptide stimulation (primarily PAM2). Representative data from one donor is shown.

Figure E7. Single-cell signaling analysis of T cell responses to lipopeptides. (A) Whole blood was stimulated with PAM2, PAM3, or control saline buffer for 30mins prior to flow cytometric analysis. A subpopulation of CD4 T cells degrades IκBα in response to lipopeptide stimulation. CD8 T cell responses also occurred but at lower levels. Representative data from one donor are shown. (B) Average CD4 (blue) and CD8 (red) T cell responses from ten donors + SD are shown. (C) T cells were enriched to 96% purity (supplementary methods) and stimulated as in part A. Average responses from six donors + SD are shown.

Figure E8. Illustration of intracellular cytokine staining analysis of CD14hi monocytes with radar plots. (A) Whole blood was stimulated with LPS for 6hrs in the presence of secretion blockers and analyzed via mass cytometry. CD14hi monocyte population was defined as in Figure 1. 95th percentile bisector gates were drawn on saline buffer control samples and used to quantify cytokine responses (positivity) to LPS stimulation. (B) Radar plots with linear scaling from 0 to 100% cytokine positivity (20% increments per radial) are used to represent nine cytokine responses. Representative data from one donor are shown.

Figure E9. TLR-induced cytokine responses in granulocyte and lymphoid cell subsets. Whole blood was stimulated with the indicated TLR ligands and processed as in Figure 4. Granulocytes, basophils, and lymphoid cell subsets were identified as indicated in Figure 1. TLR-induced cytokine signatures are represented as in Figure 4. Average values based on responses from nine healthy donors are shown.

Figure E10. Comparison of TLR-induced cytokine responses between CD14hi monocytes and CD1c+ DCs. Whole blood was stimulated with the indicated TLR ligands and processed as in Figure 4. CD14hi monocytes (blue) and CD1c+ DCs (red) exhibited overlapping stimulus-specific cytokine responses, with the exception IL-12 produced in higher frequency by DCs. Representative data from one donor are shown.

Figure E11. Intra- and inter-individual variance in TLR-induced CD14hi monocyte cytokine responses. (A) Whole blood from a single healthy donor was drawn once a month over a 4-month period. Blood samples were stimulated with the indicated TLR ligands, and processed as in Figure 4. Intracellular cytokine production was quantified as described in Figure E8. Average cytokine production responses from 4 samplings + SD are shown. (B) Average monocyte cytokine production responses (as described in part A) from nine distinct donors + SD are shown.

Figure E12. CD14hi monocytes from SLE patients produce MCP1 in the absence of exogenous stimulation. Whole blood was isolated from newly diagnosed untreated SLE patients and healthy gender-matched control subjects (distinct from healthy controls in Figure 6). Samples were either RBC lysed/WBC fixed immediately upon blood draw (grey), or incubated for 6hrs in the presence of protein secretion inhibitors (blue). Histograms for MCP1 for four SLE patients and their gender-matched controls are shown.

Figure E13. Variance in CD14hi monocytes cytokine production in SLE patients and healthy controls. Whole blood was isolated from newly diagnosed untreated SLE patients and healthy control subjects (distinct from healthy controls in Figure 6). Intracellular cytokine production at time 6hrs (no exogenous stimulation) was quantified as described in Figure E8, using cells from time zero for the 95th percentile threshold. Average cytokine production + SD from eight distinct SLE patients and eight gender-matched healthy donors are shown. Asterisks denote cytokines that were statistically significantly different between SLE patients and healthy control groups (p-value= 0.001 for TNFα, p-value= 0.002 for Mip1β, p-value= 7.6E-10 for MCP1, see methods for statistical calculation).

Figure E14. CD14hi monocytes from SLE patients demonstrate diverse combinatorial cytokine profiles. CD14hi monocyte combinatorial cytokine polyfunctionality was assessed in eight SLE patients and eight gender-matched healthy controls. Intracellular cytokine production at time 6hrs (no exogenous stimulation) was quantified as described in Figure E8, using cells from time zero for the 95th percentile threshold. 47 out of 512 different cytokine combinations were expressed by at least 1% of CD14hi monocytes from SLE patients and healthy controls. Rows and columns in this combinatorial heatmap follow the same configuration as in Figure 5. Hierarchical clustering of cytokines and donors relate cytokine co-expression (left panel) and donors with similar cytokine combination profiles (right panel), respectively.

Table E1. Summary of demographic data for healthy volunteer donors. These healthy volunteer donors were either part of the initial “reference framework” dataset, or gender-matched healthy controls for each of SLE patients. “As needed” medications were non-prescription medications used sporadically by study participants. For inclusion/exclusion criteria, see Supplementary Methods, Study Participants.

Table E2. Summary of demographic, laboratory, and clinical data for eight newly diagnosed untreated SLE patients. This data is classified into the eleven categories that are part of the ACR diagnostic criteria. For inclusion/exclusion criteria, see Supplementary Methods, Study Participants. Abbreviations: ESR, erythrocyte sedimentation rate; C3/C4, complement components 3 & 4; ANA, anti nuclear antibody titer; dsDNA, anti double-stranded DNA antibody titer; APL, antiphospholipid antibodies.

Table E3. Summary of antibodies used for mass and fluorescence-based cytometry analysis. Antibody information includes clone and manufacturer company.

Table E4. Summary of human TLRs and their ligands. TLRs’ ligand pattern recognition molecular associated patterns (PAMPs), examples of commercially available ligands, and their cellular distributions based on mRNA studies as described in Hornung et al., 2002.

1