Complement biosensors identify a classical pathway stimulus in complement-mediated thrombotic microangiopathy

Abstract

Complement-mediated thrombotic microangiopathy (CM-TMA) or hemolytic uremic syndrome, previously identified as atypical hemolytic uremic syndrome, is a TMA characterized by germ line variants or acquired antibodies to complement proteins and regulators. Building upon our prior experience with the modified Ham (mHam) assay for ex vivo diagnosis of complementopathies, we have developed an array of cell-based complement "biosensors" by selective removal of complement regulatory proteins (CD55 and CD59, CD46, or a combination thereof) in an autonomously bioluminescent HEK293 cell line. These biosensors can be used as a sensitive method for diagnosing CM-TMA and monitoring therapeutic complement blockade. Using specific complement pathway inhibitors, this model identifies immunoglobulin M (IgM)-driven classical pathway stimulus during both acute disease and in many patients during clinical remission. This provides a potential explanation for ∼50% of patients with CM-TMA who lack an alternative pathway "driving" variant and suggests at least a subset of CM-TMA is characterized by a breakdown of IgM immunologic tolerance.

Graphical abstract

Figure 1.

Complement biosensors identify a range of “normal” complement activation in healthy control sera. (A) Overview of creation of biosensors with different susceptibilities to complement and description of the assay. (B-E) Healthy control (HC) or heat-inactivated sera incubated with Livelight WT HEK293 cells (B), CD46KO cells (C), PIGAKO cells (D), double CD46, and PIGA KO (DKO) cells (E). Relative luminescence measured every 5 minutes for 2.5 hours. Data plotted as mean ± standard deviation (SD) for each triplicate. (F) Summary of relative luminescence at 1 hour for healthy controls (WT, CD46KO, and PIGAKO, n = 19; and for DKO, n = 18). Asterisk identifies outlier by ROUT analysis (Q = 1%). P values were calculated using 1-way analysis of variance (ANOVA) for Dunnett multiple-comparisons test. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001, ∗∗∗∗P < .0001. HI, heat inactivation; ns, not significant; RLU, relative luminescence units; ROUT, robust regression and outlier. Portions of figure were created with BioRender.com.

Figure 2.

Example CM-TMA tracings and summary of relative luminescence and blocking. (A,D) CM02 sera on CD46KO cells (A) or PIGAKO cells (D) in GVB⁺⁺ with or without heat inactivation (HI), eculizumab (ecu; 50 μg/mL), sutimlimab (sut; 30 μg/mL), or ACH-5548 (FDi). (B,E) Summary of relative 1-hour luminescence of healthy controls (HC), CM-TMA (acute or remission) with or without the addition of ecu, sut, ACH-4471/ACH-5548 (FDi; 1 μM), iptacopan (FBi; 1.5 μM), or “triple AP blockade” with FDi (1 μM), FBi (1.5 μM), and compstatin (18 μM) on CD46KO (B) or PIGAKO (E) (PIGAKO: HC, n = 18; acute TTP, n = 6; acute CM-TMA, n = 5; remission, n = 14; CM-TMA + ecu, n = 12; CM-TMA + sut, n = 11; CM-TMA + FDi, n = 11; CD46KO: HC, n = 18; acute TTP, n = 6; acute CM-TMA, n = 5; remission CM-TMA, n = 13; CM-TMA + ecu, n = 8; CM-TMA + sut, n = 8; CM-TMA + FDi, n = 7; CM-TMA FBi, n = 3; CM-TMA triple blockade, n = 3). Graphed as mean ± SD. Yellow fill indicates sample with pathogenic variant or variant of uncertain significance (VUS) in complement regulatory protein. Dashed line represents the 15th percentile of healthy control. (C,F) Relative 1-hour luminescence recovery after addition of inhibitor calculated at 1 hour as follows: (luminescence inhibitor-treated serum − luminescence untreated serum)/(luminescence heat-inactivated serum − luminescence untreated serum); (PIGAKO, n = 10 for each condition; CD46KO: CM-TMA + ecu, n = 12; CM-TMA + sut, n = 9; CM-TMA + FDi, n = 8; CM-TMA + FBi, n = 3; triple blockade, n = 3). (B,C,E-F) P values were calculated using 1-way ANOVA for Dunnett multiple-comparisons test. (G-H) AP only buffer inhibits complement activity. CM12 sera on PIGAKO in GVB⁺⁺ (G) or MgEGTA (AP only buffer with 13 mM Mg²⁺) (H) with or without inhibitors. (I) Example tracing of acute CM-TMA CM03 (FHAA) on CD46KO with or without inhibitors. (J) Example tracing of remission CM-TMA no mutation CM01 on CD46KO with addition of iptacopan (FBi; 1.5 μM). (K) Example tracing of CM13 on CD46KO with addition of “triple AP blockade” including with FDi (1 μM), FBi (1.5 μM), and compstatin (18 μM). (L) Example tracing of C3 glomerulopathy (C3G) with VUS in CD46 on CD46KO. Activity persists despite C3 being below limit of detection at time sample was collected (<15 mg/dL). Data plotted as mean ± SD for each triplicate. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. ns, not significant; RLU, relative luminescence units.

Figure 3.

AP assessment. (A-F) Assay in AP only buffer (13-mM MgEGTA in GVB⁰) on DKO cells (A-B,D-F) or CD46KO cells (C). (A) DKO cells were resuspended in MgEGTA with subsequent addition of healthy control serum with or without the addition of eculizumab, sutimlimab, FDi (ACH-5548), or a combination of sutimlimab and FDi. (B-C) Serum from CM03, patient with known FHAA titer of 57 104 AU in AP buffer on either DKO cells (B) or CD46KO cells (C) with or without the addition of FDi or sutimlimab. (D-G) Isolated AP inhibitor titrations. Pooled normal human serum (NHS) from complement technologies was incubated with increasing concentrations of eculizumab (D), FDi (ACH-5548) (E), FBi (iptacopan) (F), or C3 inhibitor (cp40) (G) in AP buffer on DKO cells. Example traces plotted as mean ± SD for each triplicate. RLU, relative luminescence units.

Figure 4.

C4d and C3c deposition and inhibition in HC and CM-TMA samples. (A-L) All experiments were done in the presence of C5 inhibitor, eculizumab (100 μg/mL). WT, PIGAKO, or CD46KO cells were treated with 20% serum in GVB⁺⁺ for 30 minutes and then stained singly for either C4d or C3c and deposition quantified by flow cytometry. The same healthy control (n = 4) and CM-TMA (n = 4 for WT; n = 5 for PIGAKO and CD46KO) were used across all cell types. Serum was treated, as indicated, with the sut (30 μg/mL) or ACH-5548 (1 μM; FDi). (A) Comparison of C4d deposition across cell types by percent positive cells determined by gating on heat-inactivated HC sample. (B) Relative median fluorescence intensity (MFI) of C3c deposition induced by CM-TMA serum vs healthy control serum. (C) Comparison of C3c deposition across cell types by percent positive cells determined by gating on heat-inactivated HC sample. (A-C) Intra–cell line differences between HC and CM-TMA P values were calculated with unpaired Mann-Whitney test. (D,G,J) C4d deposition on WT cells (D), PIGAKO (G), and PIGAKO (J) in HC or CM-TMA with or without addition of sut or FDi. (E,H,K) C3c MFI on WT cells (D), PIGAKO (G), PIGAKO (J) in HC or CM-TMA with or without addition of sutimlimab or FDi. (D-E,G-H,J-K) P values calculated using Friedman test for Dunn multiple comparisons. (F) Example histograms of CM07 on WT cells with or without HI, sut, or FDi. (I) Example histograms of CM12 on PIGAKO cells with or without HI, sut, or FDi. (L) Example histograms of CM06 on CD46KO cells with or without HI, sut, or FDi. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. ns, not significant.

Figure 5.

PIGAKO and DKO can be used to monitor therapeutic complement inhibition. (A) Example tracing of solid organ transplant-associated (SOT) TMA sample on therapeutic eculizumab ran on PIGAKO cells with or without addition of sutimlimab or eculizumab. (B) Example tracing of CM01 remission sample after being started on eculizumab on DKO cells. (C) Summary of relative luminescence at 1 hour for CM-TMA samples (n = 14) on treatment with C5 inhibition (acute and remission samples off treatment included for comparison). (D) Example tracing of CM21 on ravulizumab therapy with or without the addition of eculizumab, sutimlimab, or FDi on PIGAKO cells. (E) Example tracing of HC on DKO with increasing concentrations of eculizumab. (F) Example tracing of HC on PIGAKO with increasing concentrations of sutimlimab. (G) Example tracing of HC on DKO with increasing concentrations of C1 inhibitor. (H) Serum from acute CM-TMA sample (CM06) on eculizumab therapy was diluted with pooled normal human serum on PIGAKO. For panels A-B,D-H, example tracings plotted as mean ± SD for each triplicate. For panel C, P values were calculated using 1-way ANOVA for Dunnett multiple-comparisons test. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. ns, not significant; RLU, relative luminescence units.

Figure 6.

Serum treatment with DTT or IdeS and protein G spin columns identify IgM as the predominant immunoglobulin species leading to complement activation. (A) Nonreducing, no heat SDS-PAGE gel of IgG and IgM with increasing concentrations of DTT. IgG (3 μg), IgM (3 μg), or both were incubated with DTT (0 mM, 0.5 mM, 1 mM, and 3 mM) in phosphate-buffered saline (PBS) (pH 7.4) in a final reaction volume of 10 μL at room temperature for 30 minutes. Some experimental details removed throughout legend. IgM′ represents a partially reduced form of IgM. (B) Sera from patient with hemolysis, elevated liver enzymes, low platelets (HELLP) with or without the addition of additional C5 inhibitor (eculizumab), sutimlimab, DTT, or FDi on PIGAKO cells. (C) CM12 sera was processed over a protein G spin column to purify IgG (protein G eluate [GE]) or isolate flow-through (protein G flow-through [GF]). GE and GF underwent centrifugal concentration (as required), desalting, and buffer exchange into PBS. Final preparation was added to healthy control sera in the bioluminescent mHam on CD46KO cells with or without IdeS treatment. (D) CM04 sera prepared as per panel C. After preparation, GF samples were treated with IdeS, DTT (3-mM pretreatment; 0.6-mM final), or sutimlimab (30 μg/mL). Final preparation per well was added to HC sera in the bioluminescent mHam on CD46KO cells. (E) Example tracing of CM04 in bioluminescent mHam on CD46KO cells with or without HI, temperature control (untreated serum heated to 37°C for 30 minutes), eculizumab, sutimlimab, IdeS pretreatment, or DTT before treatment. (F) Summary of relative luminescence at 1 hour for CM-TMA samples treated with DTT (n = 5) or IdeS (n = 5) compared with acute, remission, or eculizumab spiked samples. (G) Healthy control sera spiked with polyclonal IgM (extra 5, 10, or 20 μg per well) and run in bioluminescent mHam on CD46KO cells. (H) Healthy control sera spiked with myeloma IgM (10 μg per well), polyclonal IgM (10 μg per well), myeloma IgG1 (10 μg per well), polyclonal IgG (10 μg or 60 μg per well) and run in bioluminescent mHam on PIGAKO cells. (I) Healthy control sera spiked with polyclonal IgM (5 or 10 μg per well), polyclonal IgG (20 or 60 μg per well), myeloma IgG1 (20 μg per well), or myeloma IgM (10 μg per well) and run in bioluminescent mHam on DKO cells. (B-E,G-I) Example traces plotted as mean ± SD for each triplicate. (F) P values were calculated using 1-way ANOVA for Dunnett multiple-comparisons test. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. HCh, heavy chain; LC, light chain; ns, not significant; RLU, relative luminescence units.

Figure 7.

CM-TMA sera has increased IgM binding to HEK293 cell surfaces and can presensitize cells even after HI of serum. Heat-inactivated sera was used for all cytometry and presensitization experiments to eliminate preloading of cells with C3 or C4 fragments. (A) WT cells were incubated for 45 minutes at 37°C with either heat-inactivated HC or CM-TMA sera (20% in GVB⁺⁺), washed, and evaluated for deposition of IgG (HC, n = 9; CM-TMA n = 17) or IgM (HC, n = 7; CM-TMA, n = 15) by flow cytometry. Acute CM-TMA samples (n = 6) indicated by red dots. Relative MFI calculated as ratio of sample MFI compared with the average of 4 healthy controls. P values were calculated using unpaired, 2-tailed t test, with Welch correction. (B-C) Example histograms of IgG (left) or IgM (right) staining from CM21 (B) and CM04 (C) on WT cells with (red) or without (blue) pretreatment of serum with DTT (3 mM pretreatment). HC stained IgG or IgM stained cells included as baseline reference population (gray). (D) Bioluminescent mHam tracing of HC1 on DKO. (H) Bioluminescent mHam tracing of HC2 on PIGAKO. (E-G,I-K) Specially selected low activity healthy control sera mix was used to facilitate complement activity in the bioluminescent mHam after either DKO (E-G) or PIGAKO (I-K) cells were presensitized (20% heat-inactivated sera treatment for 45 minutes in Dulbecco’s modified Eagle medium), washed, and resuspended in GVB⁺⁺ then run in the bioluminescent mHam. Baseline activity of the healthy control mix sera shown in each tracing as dark blue (HC sera mix) and dark red (HI HC sera mix). Pretreatment sera included HC1 (E), CM21 (F), CM04 (G), HC2 (I), CM07 (J), and CM06 (K). (D-K) Tracings plotted as mean ± SD for each triplicate. ns, not significant; RLU, relative luminescence units.

Figure 7.

CM-TMA sera has increased IgM binding to HEK293 cell surfaces and can presensitize cells even after HI of serum. Heat-inactivated sera was used for all cytometry and presensitization experiments to eliminate preloading of cells with C3 or C4 fragments. (A) WT cells were incubated for 45 minutes at 37°C with either heat-inactivated HC or CM-TMA sera (20% in GVB⁺⁺), washed, and evaluated for deposition of IgG (HC, n = 9; CM-TMA n = 17) or IgM (HC, n = 7; CM-TMA, n = 15) by flow cytometry. Acute CM-TMA samples (n = 6) indicated by red dots. Relative MFI calculated as ratio of sample MFI compared with the average of 4 healthy controls. P values were calculated using unpaired, 2-tailed t test, with Welch correction. (B-C) Example histograms of IgG (left) or IgM (right) staining from CM21 (B) and CM04 (C) on WT cells with (red) or without (blue) pretreatment of serum with DTT (3 mM pretreatment). HC stained IgG or IgM stained cells included as baseline reference population (gray). (D) Bioluminescent mHam tracing of HC1 on DKO. (H) Bioluminescent mHam tracing of HC2 on PIGAKO. (E-G,I-K) Specially selected low activity healthy control sera mix was used to facilitate complement activity in the bioluminescent mHam after either DKO (E-G) or PIGAKO (I-K) cells were presensitized (20% heat-inactivated sera treatment for 45 minutes in Dulbecco’s modified Eagle medium), washed, and resuspended in GVB⁺⁺ then run in the bioluminescent mHam. Baseline activity of the healthy control mix sera shown in each tracing as dark blue (HC sera mix) and dark red (HI HC sera mix). Pretreatment sera included HC1 (E), CM21 (F), CM04 (G), HC2 (I), CM07 (J), and CM06 (K). (D-K) Tracings plotted as mean ± SD for each triplicate. ns, not significant; RLU, relative luminescence units.