Noninvasive Detection of iC3b/C3d Deposits in the Kidney Using a Novel Bioluminescent Imaging Probe

Significance Statement

Histologic quantification of complement C3 deposits in kidney biopsies provides prognostic information in patients with glomerulonephritis. Unfortunately, kidney biopsies are invasive procedures that cannot be performed regularly and only provide a snapshot of a small portion of one kidney at the time of sampling. We have developed a method to noninvasively detect specific C3 fragment deposition throughout both kidneys, using a monoclonal antibody targeting tissue-bound iC3b/C3d linked to a bioluminescent resonance energy transfer construct that emits near-infrared light. In a mouse model of glomerulonephritis, the probe detected iC3b/C3d in kidneys of live mice by bioluminescent imaging. This demonstrates that noninvasive imaging with an anti-iC3b/C3d probe can be used to monitor inflammation in the kidneys.

Uncontrolled complement activation mediates glomerular injury in many diseases. During activation, C3 fragments are covalently fixed to nearby tissues, providing a biomarker of inflammation. Because the abundance of these C3 fragments reflects ongoing immune activity, kidney tissue is routinely immunostained for C3c fragments as a marker of autoimmune and inflammatory processes. Molecular imaging probes for C3 fragments offer the promise of noninvasively detecting these fragments in live animals. However, a challenge in detecting tissue C3 fragments in vivo (e.g., C3b, iC3b, or C3d) is that intact C3 is an abundant plasma protein, and so, the probes must distinguish the tissue-bound fragments from C3 in plasma. To overcome this obstacle, we developed a panel of anti-C3d monoclonal antibodies that targets tissue-bound iC3b/C3d (referred to as C3d).¹

We previously labeled one of the monoclonal antibodies (mAb 3d29) with a fluorescent tag to detect deposits in the eyes of live mice,¹ and we have also labeled it with iodine-124 as a positron emission tomography probe.² Unfortunately, hemoglobin absorbs photons emitted from most fluorescent or bioluminescent probes, and so, they cannot be detected in deep organs. Radiolabeled probes, on the other hand, use dangerous and expensive isotopes. Furthermore, large proteins, such as antibodies, circulate for days or weeks, and the isotope requires a sufficiently long half-life to permit clearance of unbound probe from the blood pool (Supplemental Figure 1).³ One solution to these limitations is to use reporters that emit light of near-infrared (NIR) or orange-red wavelengths, which penetrate tissues to a depth of 5 mm to 3 cm.⁴ Although this range is not sufficient for imaging deep tissues in humans, it can be used to detect molecular targets in rodents. Furthermore, bioluminescence resonance energy transfer constructs that emit light at these wavelengths are also advantageous because they generate light through a chemical reaction, do not require photoexcitation from an external source, and have very little background signal. One such construct (“Antares”) contains a NanoLuc luciferase linked to two red-shifted orange fluorescent proteins that emit light at a wavelength of approximately 600 nm.⁵

To noninvasively detect C3 fragment deposits in the mouse kidney without radioactive isotopes, we inserted an AviTag into the Antares construct to label streptavidin-conjugated antibody (Figure 1A and Supplemental Figure 2). When coelenterazine substrate was added to the probe, it emitted light at 460 nm and approximately 590 nm peaks and was bright enough to be seen by the naked eye (Figure 1B and Supplemental Figure 2C). Using ELISA assays, we confirmed that Antares-labeled mAb 3d29 retained its binding specificity for C3d and generated light at the expected wavelengths (Figure 1C and Supplemental Figure 3). After comparing different doses and time points in vivo (Supplemental Figure 4), we tested whether the Antares-conjugated anti-C3d could be used to detect glomerular C3d deposits in a mouse model of C3 glomerulopathy (Cfh^−/− mice), which has abundant glomerular deposits of C3d (Supplemental Figure 5).⁶ Mice were injected with 100 μg of Antares-labeled mAb 3d29 and imaged after 1 week. The kidneys of Cfh^−/− mice emitted strong signal compared with those of C57BL/6 and C3^−/− control mice (Figure 1, D and E). The kidneys of Cfh^−/− mice injected with a control antibody or untargeted Antares emitted little signal (Supplemental Figure 6).

Figure 1.

A NIR-BRET–labeled probe detects C3d in live mice. (A) Diagram of Antares-labeled antibody. The Antares construct is linked to streptavidin-labeled antibody by means of an AviTag. Addition of sCTZ causes nLUC to emit light at a wavelength of 460 nm. This light excites the adjacent red-shifted OFP to emit light at 589 nm. (B) After addition of coelenterazine substrate, the Antares construct emits light with peaks at approximately 460 nm (blue arrow) and approximately 589 nm (red arrow), corresponding to the expected emission peaks for NanoLuc and the OFPs, respectively. Repeated exposure of the same sample to coelenterazine substrate led to attenuation of the amount of light emitted across the spectrum. (C) ELISA plates were coated with C3d or buffer alone; then, they were incubated with Antares-labeled mAb 3d29 (3d29-Ant), Antares alone, or PBS. sCTZ was added, and the emission at 600 nm was measured. Emission was strongest in the C3d-coated wells incubated with mAb-Antares (P < 0.001 for mAb-Antares versus Antares alone or PBS). (D) Cohorts of factor H-deficient (Cfh^−/−), C57BL/6, and C3 deficient (C3^−/−) mice were injected with 100 μg of Antares-labeled mAb 3d29. The mice were imaged after 1 week, and a representative mouse for each group is shown. Signal in the paws and ears of the C3^−/− mice is probably caused by unbound probe in the circulation of these mice. (E) Comparison of signal in the different strains of mice demonstrates that stronger signal is seen in the kidneys of Cfh^−/− mice (mean of 8.836×10⁶ photons; 95% CI, 3.703 to 13.97) compared with C57BL/6 mice (1.929×10⁶ photons; 95% CI, 0.204 to 0.8136) and C3^−/− mice (0.6987×10⁶ photons; 95% CI, 0.4247 to 0.9727). P < 0.01 for Cfh^−/− mice versus the other two strains. (F) A three-dimensional reconstruction with virtual organ library overlay shows that signal in Cfh^−/− mice injected with Antares-labeled mAb 3d29 localizes to the kidneys. BRET, bioluminescence resonance energy transfer; CI, confidence interval; nLUC, NanLuc; OFP, orange fluorescent protein; sCTZ, coelenterazine. Figure 1 can be viewed in color online at www.jasn.org.

Deposition of C3 fragments is not specific to any disease, but it may provide generalizable readout of glomerular inflammation and immunologic activity. By linking the Antares reporter to an antibody that targets tissue-bound C3d, we were able to detect complement activation in kidneys of live mice. The probe and imaging procedure were well tolerated, and preliminary toxicologic evaluation did not reveal any adverse effects (Supplemental Figure 7).

This method does have several limitations. It is likely that repeated injection of Antares will lead to the development of antiprobe antibodies. In addition, although factor H-deficient mice are a well-established model of C3G, kidney disease progression is slow in this model, limiting our ability to test whether C3d imaging reflects disease severity.⁶ Finally, although NIR light penetrates tissues better than light of shorter wavelengths, it still only penetrates <3 cm.⁷ Thus, Antares-labeled probes are useful for research but would be of limited use in humans.

In summary, we have developed a method of noninvasively detecting and quantifying C3d deposits throughout both kidneys of live mice. Because microscopy only samples a small region of the kidney, this molecular imaging method may more accurately reflect the overall degree of inflammation in the kidney and elsewhere in the body. Furthermore, insertion of the AviTag into the Antares construct makes it easy to label any streptavidin-conjugated protein, and this method can easily be adapted to track other antibodies or proteins in animal models.

Supplementary Material

jasn-34-1151-s001.pdf ^{(4.3MB, pdf)}

Disclosures

V.M. Holers is consultant for Q32 Bio, Inc., a company developing complement inhibitors. Also hold stock and may receive royalty income from Q32 Bio, Inc. V.M. Holers also reports Ownership Interest: Touch of Life Technologies, Inc (ToLTech) and Patents or Royalties: Alexion Pharmaceuticals, Inc. and Q32 Bio. L. Kulik reports Patents or Royalties: Q32 Bio. F. Poppelaars has consulted for Invizius and owns stock in Appelis, Chemocentryx, InflaRx N.V., and Omeros Corporation. J.M. Thurman is consultant for Q32 Bio, Inc., a company developing complement inhibitors. Also hold stock and may receive royalty income from Q32 Bio, Inc. J.M. Thurman also reports Honoraria: VMLY&R Health and Patents or Royalties: Alexion Pharmaceuticals and Q32 Bio, Inc. All remaining authors have nothing to disclose.

Funding

This work was supported by National Institutes of Health Grants R01DK076690 and R01DK113586 (J.M.T.), R01DK125823 (J.M.T. and V.M.H.), and Department of Defense Grant LR180050 (to J.M.T.). This project was also supported in part by the Animal Imaging Shared Resource of the University of Colorado Cancer Center (P30CA046934). F. Poppelaars was supported by a grant from the American Society of Nephrology Foundation for Kidney Research Ben J. Lipps Research Fellowship Program. This work was supported by Tekke Huizinga Fonds (awarded to Felix Poppelaars, http://tekkehuizingafonds.nl/).

Author Contributions

Conceptualization: Joshua M. Thurman.

Formal analysis: Brandon Renner, Joshua M. Thurman.

Investigation: V. Michael Holers, Liudmila Kulik, Jennifer Laskowski, Felix Poppelaars, Brandon Renner, Natalie J. Serkova.

Methodology: V. Michael Holers, Brandon Renner, Joshua M. Thurman.

Writing – original draft: Joshua M. Thurman.

Writing – review & editing: V. Michael Holers, Liudmila Kulik, Jennifer Laskowski, Felix Poppelaars, Brandon Renner, Natalie J. Serkova.

Data Sharing Statement

All data used in this study are available in this article.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/JSN/E406.

Supplemental Figure 1. Positron emission tomography of factor H-deficient mice (Cfh^−/− mice) after injection with ¹²⁴I-labeled mAb 3d29.

Supplemental Figure 2. Generation of AviTag-labeled Antares.

Supplemental Figure 3. Characterization of 3d29-Antares.

Supplemental Figure 4. Comparison of different doses of 3d29-Antares and different imaging time points.

Supplemental Figure 5. C3d deposition in kidneys of Cfh^−/−, wild-type, and C3^−/− mice.

Supplemental Figure 6. Comparison of 3d29-Antares signal with that of control probes.

Supplemental Figure 7. Injection with 3d29-Antares did not have detectable liver or kidney toxicity.