A fluorogenic near-infrared imaging agent for quantifying plasma and local tissue renin activity in vivo and ex vivo

Abstract

The renin-angiotensin system (RAS) is well studied for its regulation of blood pressure and fluid homeostasis, as well as for increased activity associated with a variety of diseases and conditions, including cardiovascular disease, diabetes, and kidney disease. The enzyme renin cleaves angiotensinogen to form angiotensin I (ANG I), which is further cleaved by angiotensin-converting enzyme to produce ANG II. Although ANG II is the main effector molecule of the RAS, renin is the rate-limiting enzyme, thus playing a pivotal role in regulating RAS activity in hypertension and organ injury processes. Our objective was to develop a near-infrared fluorescent (NIRF) renin-imaging agent for noninvasive in vivo detection of renin activity as a measure of tissue RAS and in vitro plasma renin activity. We synthesized a renin-activatable agent, ReninSense 680 FAST (ReninSense), using a NIRF-quenched substrate derived from angiotensinogen that is cleaved specifically by purified mouse and rat renin enzymes to generate a fluorescent signal. This agent was assessed in vitro, in vivo, and ex vivo to detect and quantify increases in plasma and kidney renin activity in sodium-sensitive inbred C57BL/6 mice maintained on a low dietary sodium and diuretic regimen. Noninvasive in vivo fluorescence molecular tomographic imaging of the ReninSense signal in the kidney detected increased renin activity in the kidneys of hyperreninemic C57BL/6 mice. The agent also effectively detected renin activity in ex vivo kidneys, kidney tissue sections, and plasma samples. This approach could provide a new tool for assessing disorders linked to altered tissue and plasma renin activity and to monitor the efficacy of therapeutic treatments.

the renin-angiotensin system (RAS) plays an important role in the regulation of blood volume, electrolyte homeostasis, and systemic vascular resistance, which together affect cardiac output and arterial pressure (11). Renin, a highly specific aspartyl protease, is primarily stored and released from juxtaglomerular (JG) cells within the afferent arterioles of the kidney glomeruli and plays an important regulatory role in RAS function. JG cells can sense a reduction in afferent arteriole pressure in the kidney leading to direct release of renin, and specialized adjacent cells (the macula densa) can also trigger JG renin release upon sensing of decreased plasma sodium chloride in renal tubules or changes in kidney sympathetic nerve signaling. Release of renin into the blood then mediates the first and rate-limiting step of the RAS cascade by cleaving angiotensinogen to generate angiotensin I (ANG I). ANG I is further cleaved by angiotensin-converting enzyme (ACE) to generate bioactive angiotensin II (ANG II), which binds to angiotensin receptor-1 (AT1R), causing downstream effects such as vasoconstriction, salt retention, aldosterone release, and hypothalamic stimulation of the thirst reflex. This places renin at the nexus of the RAS system as the essential regulator of ANG II levels that drive increases in blood pressure (10) and makes this enzyme an optimal biomarker for assessing changes in the RAS.

Extensive research has focused on the inhibition of various components of the RAS for treatment of hypertension and cardiovascular diseases. ACE and AT1R inhibitors (5, 13), AT1R blockers, and more recently renin inhibitors (31) have been used for the treatment of hypertension and attenuation of end-organ damage. In the development of these treatments, plasma renin activity (PRA) has been generally considered the most practical method for determining the activity of the RAS, using a variety of techniques, including immunoassays (3), fluorescence (FRET) substrates (24, 27), mass spectrometry (8), and magnetic sensors (34). In preclinical studies, these technologies have been limited to plasma assessment. However, recent discoveries of an additional local tissue RAS (1, 25, 30), in addition to the classic circulating RAS cascade, may make reliance only on plasma measurements incomplete.

Ideally, noninvasive in situ imaging of local tissue RAS activity, in addition to measurement of plasma renin, would help in the determination of therapeutic efficacy, but only limited efforts have been reported for in vivo imaging (16). In particular, tomographic fluorescence imaging of near-infrared-labeled imaging agents has the potential to be used in drug discovery RAS research by virtue of improved deep tissue penetration, quantitative readout (picomoles rather than relative light intensity), and the pairing with protease-activated NIR imaging agents. This approach has been shown to be effective for the assessment and monitoring of protease activity associated with various diseases, including cancer (20), atherosclerosis (4), lung inflammation, and arthritis (26, 33). In a similar fashion, noninvasive imaging and quantification of renin activity in vivo and monitoring of response to therapeutics could have a significant impact on the development of cardiovascular disease-targeted drugs in both preclinical and clinical settings.

Herein, we report the development, characterization, and functional evaluation of an activatable NIR fluorescence (NIRF) imaging agent that is cleaved/activated by rodent renin. This imaging agent has utility in plasma renin assessment, ex vivo tissue assessment, and noninvasive kidney imaging in live animals. Used in vitro, it can be added to plasma samples to effectively quantify PRA, consistent with an independent renin immunoassay, and to measure renin upregulation in kidney tissue sections as assessed by fluorescence microscopy. In vivo, the renin-imaging agent can be injected intravenously in living mice to image and quantify kidney renin activity in a low-salt diet mouse model of hyperreninemia (32), with easy ex vivo validation of fluorescence levels in tissue and plasma samples.

MATERIALS AND METHODS

Experimental animals.

Specific pathogen-free female C57BL/6 and SKH-1 mice (6–8 wk of age) were obtained from The Jackson Laboratories (Bar Harbor, ME) and housed in a controlled environment (72°F; 12:12-h light-dark cycle) under specific-pathogen free conditions with water and food provided ad libitum. All experiments were performed in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The study protocol was approved by the internal Institutional Animal Care and Use Committee based on established guidelines for animal care and use. No invasive or surgical procedures were used in these studies, but all imaging activities were performed under appropriate anesthesia to minimize animal distress.

Fluorogenic renin-imaging agent ReninSense.

The fluorogenic agent, ReninSense 680 FAST (ReninSense), consists of a modified rodent angiotensinogen-derived peptide sequence to which NIR fluorochromes (VivoTag-S680, PerkinElmer, Boston, MA) were linked to both the C and N termini (Fig. 1A). The agent displays an absorbance maximum at 675 nm and an emission maximum of 693 nm (Fig. 1B). The closely associated fluorochromes efficiently quench one another while the substrate is intact but release strong NIRF upon cleavage of the peptide by the targeted enzyme; when exposed to rodent renin, the agent shows more than a 26-fold increase in fluorescence intensity over the background (Fig. 1C). To prolong the plasma half-life of the agent for in vivo imaging applications, the substrate was further conjugated to a pharmacokinetic modifier composed of a polymer carrier at a ratio of 1 substrate per polymer molecule. The final molecular mass of ReninSense is ∼43,000 Da. UV-visible light absorbance and fluorescence emission spectra of the native and renin-activated agent were recorded on Cary 50 and Cary Eclipse spectrophotometers, respectively, in 1× PBS using a 665-nm light for fluorescence excitation.

Fig. 1.

Chemical design and properties of ReninSense. A; schematic of ReninSense 680 activation by renin. The agent consisted of an angiotensinogen-derived peptide sequence flanked by 2 near-infrared fluorochromes that were self-quenched in the native state and released strong fluorescence upon proteolytic cleavage of the agent. The fluorogenic peptide substrate was further conjugated to a pharmacokinetic modifier (labeled PKM) which gave the agent a total molecular mass of ∼43,000 Da. B: the agent is 94–96% self-quenched in the native state, and upon activation the peak absorbance increases in wavelength by ∼45 nm. C: fluorescence emission with a maximum at ∼690 nm (excitation at 665 nm) was increased >26-fold upon proteolytic activation.

In vitro profiling of protease cleavage of ReninSense.

Activation of ReninSense (1 μM) was determined with a panel of relevant enzymes (0.1 μM final concentration): active recombinant human (Proteos, Kalamazoo, MI), rat (Merck Laboratories, West Point, PA), and mouse renin (Innovative Research, Novi, MI), human neutrophil elastase (Innovative Research), human neutrophil cathepsin G (Enzo Life Sciences, Plymouth Meeting, PA), human liver cathepsin D (Enzo Life Sciences,), human plasma plasmin (Calbiochem, San Diego, CA), trypsin (Sigma, St. Louis, MO), and recombinant human cathepsin B (R&D Systems, Minneapolis, MN). The assay condition for renin was 50 mM MOPS, 100 mM NaCl, pH 7.4, with freshly added 0.002% Tween 20, and incubation at 37°C for up to 24 h. The assay buffer for elastase was 100 mM Tris (pH 7.5); for cathepsin G, 100 mM Tris (pH 7.5), 1.6 M NaCl; for cathepsin D, 100 mM formic acid, pH 3.3; and for plasmin and trypsin, 1× PBS, pH 7.4. For the activation of mouse prorenin, cathepsin B was first activated in 25 mM MES, pH 5, 5 mM DTT for 15 min at room temperature, then the activated cathepsin B was added to mouse prorenin in 25 mM MES (pH 5), 0.5 mM DTT and incubated overnight at room temperature (RT). ReninSense was activated by the active mouse renin in the buffer described above. Reactions were carried out in a 250-μl volume and set up in 96-well plates with black sides and bottom. Reactions were monitored at various time points at excitation/emission wavelengths of 663/690 nm with a cutoff at 665 nm using a Gemini Fluorescence Plate Reader (Molecular Devices, San Leandro, CA). The released fluorescence is shown after subtracting the agent-only background.

Mouse plasma stability and pharmacokinetic studies.

For in vitro plasma stability studies, ReninSense (1 μM) was incubated in normal mouse plasma (purchased from Innovative Research) diluted 1:4 in 1× PBS, pH 7.40 with 1 mM EDTA, at 37°C for 24 h. The extent of cleavage was assessed by the appearance of a distinct fragment peak upon HPLC analysis. For in vivo sample preparation, 2 nmol of ReninSense was injected into normal C57BL/6 mice and plasma samples were collected at 2 min, 30 min, and at 1, 3, 6, 24, 48, and 72 h to characterize plasma pharmacokinetics. Plasma from a normal mouse without ReninSense was used for background subtraction. These samples were stored frozen at −80°C until analytic testing was to be performed, at which time samples were allowed to thaw at room temperature for at least half an hour before being processed for HPLC analyses.

For HPLC analysis, aliquots (50 μl) of each plasma sample were placed in Eppendorf tubes. Cold methanol (150 μl) was added, and the tubes were vortexed followed by centrifugation at 12,000 rpm and 4°C for 10 min to precipitate the plasma proteins. An aliquot (110 μl) of each supernatant was transferred to vials for analysis on a Jupiter C18, 300-Å, 5-μm, 50 × 4.6-mm HPLC column (Phenomenex, Torrance, CA) with a Waters model 2695 HPLC that utilized a photodiode array (PDA) and fluorescence detection (scan from 225 to 800 nm). The aqueous mobile phase contained 25 mM triethylamine acetate, pH 7. Samples were eluted with acetonitrile at a flow rate of 2 ml/min. A gradient of 10–85% organic provided sufficient resolution. An additional mobile phase containing 0.1% formic acid in methanol was used in a wash step to prevent the buildup of proteins on the column. The wavelength corresponding to the absorbance maximum of the fluorophore (675 nm) was extracted from the PDA trace. For quantification, the fluorescence spectrophotometer monitoring (excitation 675 nm; emission 693 nm) was calibrated using ReninSense standards prepared at 0, 78, 156, 312, 625, 1,250, and 2,500 nM in mouse plasma, yielding standard curves with correlation coefficients >0.99. As expected from the properties of the polymer carrier added as a pharmacokinetic modifier, ReninSense shows a plasma half-life of ∼7 h (see Fig. 6A).

Fig. 6.

Pharmacokinetics and tissue biodistribution of ReninSense. C57BL/6 mice (n = 3/group) were injected with 2 nmol ReninSense, and blood was collected at 7 time points spanning 5 min to 72 h. A: total plasma ReninSense levels in mice (i.e., both activated and unactivated forms) were quantified by HPLC as described in materials and methods. B: in separate cohorts of mice (n = 3 mice/time point), tissues were excised from mice at different times post-ReninSense injection to determine the differential distribution of activated ReninSense in kidneys, liver, heart, lungs, muscle, and blood over time as assessed by epifluorescense intensity ex vivo. Note that this assessment is of only activated ReninSense and would not detect unactivated ReninSense.

In vitro fluorescence plate assays for PRA.

C57BL/6 mice (Jackson Laboratories) were fed a low-fluorescence and sodium-deficient diet (0.02% Na, LS diet, catalog no. 17095; Harlan Laboratories, Indianapolis IN), and water was replaced with a diuretic solution (amiloride, 0.1 mg/5 ml or 5 mg·kg⁻¹·day⁻¹; Sigma-Aldrich) for 2 days as described by Wagner et al. (32). Control mice were fed normal chow and water ad libitum. In some studies, a cohort of mice fed the sodium-deficient/amiloride (LS) diet were also administered a 3 nM mouse renin inhibitor, L-810 (Merck Canada). For in vivo studies, this agent was given in three doses of 30 mg/kg po at 8, 22, and 30 h following the start of the diet, and at 48 h the mice were euthanized by CO₂ inhalation. Blood samples from individual animals were harvested by cardiac puncture and then transferred to coagulation vials. The blood samples were centrifuged at 10,000 rpm for 5 min to collect plasma. Plasma was diluted 10× with 1× PBS, pH 7.4, and preincubated with or without the renin inhibitor L810 (300 nM) at RT for 30 min. Samples were assessed for PRA using an immunoassay to measure generation of ANG I (3). For assessment of PRA by fluorescence, ReninSense was the added at 1 μM final concentration and incubated at 37°C for 1, 3, 5, and 24 h before the fluorescence in each plate was read in a microplate reader using appropriate absorption and emission filters (excitation 675 nm and emission 693 nm).

Ex vivo activation of ReninSense in mouse kidney sections.

C57BL/6 mice on the LS diet for 2 days and control mice were euthanized by CO₂ inhalation, and kidneys were removed. In some studies, separate cohorts of mice on the LS diet received two doses of L-810 (30 mg/kg) po 8 and 24 h after the start of the diet. Individual kidneys were snap frozen in OCT and sectioned using a microtome. For consistency, only the right kidneys were used. As renin is released by the JG cells within the renal cortex, 10-μm kidney sections were taken from the cortical region of the kidney. Sections were incubated with ReninSense (1 μM) in the absence or presence of L-810 (300 nM) at 37°C with a coverslip in a humidified incubator for 3 h. Sections were observed under fluorescence microscopy using appropriate filters (Zeiss Axioskop 2 MOT Plus).

In vivo imaging of renin upregulation.

C57BL/6 mice, maintained on either the LS diet/diuretic or control diet based on a published model of hyperreninemia (32), were injected intravenously with ReninSense (2 nmol/mouse in 100 μl PBS) 24 h after diet initiation. In some studies, a cohort of mice fed the LS diet were also administered three doses of L-810 (30 mg/kg orally) 8, 22, and 30 h after the start of the diet. The cohorts remained on their respective diets until the end of the imaging study. Mice were then anesthetized using an intraperitoneal injection of ketamine (100 mg/kg) and xylazine (20 mg/kg) and depilated to minimize fur interference with the fluorescent signal. Nair lotion (Church & Dwight, Princeton, NJ) was applied thickly on skin over the torso (front, back, and sides) of each mouse, rinsed off with warm water, and reapplied until all fur had been removed. LS diet and control mice were then imaged using the FMT 2500 fluorescence tomography in vivo imaging system (PerkinElmer). For imaging, each anesthetized mouse was positioned in a prone position in the imaging cassette, adjusted to the proper depth to gently restrain the mouse, and the cassette was then inserted into the heated docking system (regulated at ∼37°C) in the FMT imaging chamber. A NIR laser diode transilluminated (i.e., passed light through the body of the animal to be collected on the opposite side) the mouse torso, with signal detection occurring via a thermoelectrically cooled CCD camera placed on the opposite side of the imaged animal. Appropriate optical filters allowed collection of both fluorescence and excitation data sets, and the multiple source-detector fluorescence projections were normalized to the paired collection of laser excitation data. The entire image acquisition sequence took ∼5–6 min/mouse. Epifluorescence imaging on either the FMT 2500 (epifluorescence mode) or the Kodak 2D epifluorescence imager was routinely performed before each tomographic imaging session using built-in LED front illuminators and collection of single camera images. In some studies, mice also received three doses of L-810 (30 mg/kg po) at 8, 22, and 30 h after the start of the diet (n = 7), a dose regimen characterized with in vivo efficacy in a mouse model of hypertension (Merck Canada; personal communication). Imaging times were effective between 6 and 24 h, consistent with in vitro activation kinetics, with better signal to background observed at 24 h.

FMT 2500 image analysis.

The collected fluorescence data were reconstructed by FMT 2500 system software (TrueQuant v2.0, PerkinElmer) for the quantification of the three-dimensional (3D) fluorescence signal within the kidneys. The system is calibrated with appropriate fluorophores to allow determination (in units of nM and total pmol) of the amount of fluorescence within each and every three-dimensional voxel comprising the volume of the scanned region of the imaged subject. 3D regions of interest (ROI) were drawn to encompass the fluorescent signal defining each kidney. Extra care was taken to consistently position the 3D ROIs consistently from mouse to mouse using anatomic landmarks from the 2D white light image of each mouse's body. A 300-nM threshold was applied identically to all animals, equivalent to 45% of the mean kidney fluorescence of the positive control LS diet mice. This decreased the quantification of the lowest-intensity, background fluorescence by ∼40% in the kidney but removed only 5–6% of the renin-induced fluorescence in positive control mice. The total amount of fluorescence in each kidney (in pmol) was automatically calculated relative to internal standards generated with known concentrations of appropriate NIR dyes. For visualization and analysis purposes, TrueQuant v.2.0 software provided 3D images.

Ex vivo imaging of excised kidneys.

Upon completion of the noninvasive in vivo imaging, mice were euthanized, kidneys were collected, and tissue epifluorescence was measured with the appropriate optical filters using either the FMT or Kodak Imaging Station 2D imaging capabilities. 2D ROIs were drawn around the kidneys, and a threshold was applied identically to all animals to remove low-grade background signals from images. ROIs were drawn to measure mean fluorescence intensity in the kidneys and control tissue (subtracted as background). Other organs and tissues were also collected from normal C57BL/6 mice injected with ReninSense and imaged to assess the general biodistribution of the agent at 24 h.

Plasma fluorescence of activated ReninSense after intravenous injection.

Upon completion of noninvasive in vivo imaging, blood samples from individual animals were collected and then transferred to vials containing EDTA to prevent coagulation. The blood samples were centrifuged at 10,000 rpm for 5 min to collect the resulting plasma. Plasma fluorescence was measured using a microplate fluorescence plate reader (50 μl plasma/mouse in a 96-well low-volume black plate). Plasma collected from normal mouse controls was used for background subtraction.

Statistical analysis.

Data are presented as means ± SE. Significance analysis of in vivo kidney fluorescence was conducted using a two-tailed unpaired Student's t-test when two groups were analyzed or a one-tailed analysis of variance followed by a Scheffé multiple comparison test, with P values <0.05 considered significant.

RESULTS

In vitro assessment of ReninSense activation by renin and other enzymes.

The activation of ReninSense was tested against active mouse, rat, and human renin as well as a panel of relevant enzymes (Fig. 2A). ReninSense was quickly cleaved by mouse renin, reaching a plateau of ∼4,500 released fluorescence units (FLU) at 3 h. Activation was slower in the presence of rat renin, reaching only ∼400 FLU at 3 h and 2,000 FLU at 24 h, at which time the plateau had yet not been reached. ReninSense was not cleaved appreciably by human renin. Other enzymes known to cleave angiotensinogen to ANG I (e.g., neutrophil elastase, cathepsin G, and cathepsin D) cleaved ReninSense with a efficiency similar to that of mouse renin, whereas the unrelated enzymes, plasmin, trypsin, neutrophil elastase, and ACE, did not cleave the agent (Fig. 2B). It is not currently defined whether ACE is able to further cleave the activated substrate following renin cleavage, although this would be the expected fate for an angiotensinogen peptide.

Fig. 2.

Enzymatic activation of ReninSense by renin. A: ReninSense was tested with active renin from a mouse, rat, and human. ReninSense 680 FAST was rapidly activated by mouse renin (top line, ●), relatively slowly by rat renin (middle line, ○), and not cleaved by human renin (bottom line, ▲). B: activation of ReninSense by renin, cathepsin D, cathepsin G, angiotensin-converting enzyme (ACE), trypsin, plasmin, and neutrophil elastase (NE) showing predicted activation by renin, cathepsin D, and cathepsin G as expected for an angiotensinogen-based renin substrate.

Stability in normal control plasma.

In the presence of plasma proteins, fluorescence-quenched substrate-based agents can potentially undergo conformational changes or can be cleaved by plasma proteases to partially or completely dequench the fluorescence. To assess stability in normal mouse plasma, 1 μM ReninSense was incubated with either PBS, 25% normal mouse plasma (i.e., plasma diluted 4-fold in PBS, pH 7.4, 1 mM EDTA), or recombinant mouse renin (0.1 μM) at 37°C for 24 h (Fig. 3). Cleavage products were then analyzed using HPLC and UV detection at 675 nm. ReninSense showed very good stability in both PBS and normal mouse plasma (Fig. 3, A and B) with no detectable cleavage at 24 h nor plasma dequenching of the fluorophores. [Note that in subsequent studies, 100% normal mouse plasma shows low-level, yet measureable, activation of ReninSense.] As expected, the agent was completely cleaved upon incubation with 0.1 μM rmRenin for 24 h under the same conditions (Fig. 3C), suggesting that ReninSense would not be subject to nonmechanistic activation or dequenching while in circulation in vivo but could be selectively activated in tissues secreting active renin or in plasma with high active renin levels.

Fig. 3.

Mouse plasma stability of ReninSense. ReninSense was incubated in PBS (A), normal mouse plasma alone (B), or with recombinant mouse renin (rmRenin; C) at 37°C for 24 h. Cleavage products were analyzed by HPLC as described in materials and methods.

Assessment of renin upregulation in plasma and tissue samples from mice.

To assess the ability of ReninSense to detect both plasma and tissue renin upregulation, plasma and kidney tissues were collected from sodium-sensitive inbred C57BL/6 mice known to increase local kidney renin activation and release into circulation when maintained on a LS diet and diuretic regimen (32). Control mice were maintained on normal chow for 24 h with no diuretic treatment. As an independent validation of renin upregulation in this mouse model, PRA in control and experimental mice was assessed by a competition-based enzyme immunoassay that quantified ANG I generated from angiotensinogen cleavage by plasma renin (3). The immunoassay verified the expected PRA increase for this established mouse model and established an objective data set for comparison with fluorescence data sets. As ReninSense mimics angiotensinogen as a substrate for renin, the direct incubation of ReninSense with plasma samples provided a simple and fast assay, requiring no secondary reagents, to detect PRA. The fluorescent signal from ReninSense incubated with plasma from control mice only slightly increased from 0 to 1 h (Fig. 4, ▲), consistent with the low but measureable levels of renin activity known to be in mouse plasma. In contrast, incubation of ReninSense with plasma from mice on the LS diet (Fig. 4B, ●) caused a more than threefold increase in the fluorescent signal by 24 h. The observed renin-induced cleavage was completely inhibited by L-810 (Fig. 4B, ○). The rate of activation of ReninSense 680 FAST was also assessed (Fig. 4C) between 2 and 24 h as a better means of comparison with the kinetic enzyme immunoassay, and this analysis revealed a similar change in PRA rates between the two assays; there was a lower rate of activation in normal plasma and a ∼20-fold increase in activation rate when plasma from low-salt diet mice was used.

Fig. 4.

In vitro activation of ReninSense by plasma from mice on a low-salt diet (LSD). A: mouse plasma under a normal diet or LSD was harvested and assessed for plasma renin activity (PRA) using a competition-based enzyme immunoassay as previously described (3). Values are means ± SE. B: ReninSense was activated by plasma from normal diet (triangles) or LSD (circles) mice with (filled symbols) or without (open symbols) a specific renin inhibitor, L-810. Note that normal diet, normal diet+L-80, and LSD+L-810 data overlap. C: to better compare with enzyme immunoassay data, the rate of fluorescence increase in ReninSense (RS680) was assessed between 2 and 24 h as an indication of agent activation rate in plasma. The microplate plasma fluorescence assay showed <5% coefficient of variation among replicate wells.

Renin activity was also readily detected by incubating ReninSense with 10-μm frozen kidney sections taken from mice on the LS or normal diet for 24 h. Tissue sections were incubated at 37°C with ReninSense to achieve an optimal fluorescent signal at 3 h. As expected, a visible level of activation of ReninSense was seen by NIRF microscopy in kidney sections from mice on the normal diet (Fig. 5A); however, the fluorescent signal was notably increased in kidneys from mice on the LS diet (Fig. 5B). The activation of ReninSense was completely blocked by coincubation with L-810 in kidneys from both groups of mice (Fig. 5, C and D), although some low-grade fluorescence signal did become visible with much longer exposure times (data not shown). By demonstrating an increased activation signal in kidney tissues from LS diet-treated mice, and its complete inhibition by a selective renin inhibitor, we have provided unambiguous evidence that ReninSense is a sensitive and specific tool for measuring renin activity in nonhomogenized kidney tissues. The widespread signal throughout the tissue, rather than specific JG accumulation, is consistent with renin release from cut sections into the media overlaying the tissue. Naturally, as the activation of ReninSense requires an active renin enzyme, this assay approach will not work for fixed tissues.

Fig. 5.

Activation of ReninSense by mouse kidney slices. Kidney renin activity was assessed in situ by incubating frozen kidney sections (10 μm thick) from normal diet (A and C) or LSD (B and D) mice with 1 μM ReninSense at 37°C for 3 h in the presence (C and D) or absence (A and B) of 300 nM selective renin inhibitor L-810. Fluorescent microscopy images of kidney cortical regions were captured with an acquisition time of 2 s. Final magnification was ×100.

Pharmacokinetics and activated ReninSense biodistribution.

Normal C57BL/6 mice were injected intravenously with 2 nmol of agent (n = 2–3 mice/time point), and plasma samples were collected at different times thereafter for analysis by RP-HPLC using absorbance detection at 675 nm. In addition, to determine the tissue biodistribution of activated ReninSense, mice on the LS diet and normal diet were injected with ReninSense and euthanized 24 h later, at which time kidneys, hearts, lungs, livers, muscle samples, and blood samples were harvested and assessed by 2D planar imaging. Based on a standard curve, the plasma half-life of ReninSense was ∼7 h (Fig. 6A), with the plasma signal dropping to near baseline at 72 h. Biodistribution images revealed generally higher signals in kidneys, lungs, heart, liver, and blood, as expected (Fig. 6B), particularly at earlier time points, but mechanistic binding was not studied in any tissues other than the kidneys. Salivary glands were not assessed, as C57BL/6 mice do not show upregulation of renin activity in this tissue unlike in other mouse strains. Kidney signals achieved a plateau between 6 and 48 h, then slowly decreased through 72 h, with the optimal imaging window defined by optimal imaging quality at 24 h due to significant clearance from circulation.

Assessment of selective ReninSense accumulation in the kidney in vivo.

Although ReninSense showed some apparent differential accumulation within the kidney, consistent with activation within the kidney, the possibility remained that this was due to extrarenal activation and subsequent circulation and clearance via the kidney. To attempt to address this possibility, we compared ReninSense to a commercially available protease-activatable control agent of nearly identical size and structure (differing only in substrate sequence and wavelength of fluorophore), with similar plasma pharmacokinetics (6- to 7-h half-life). This commercially available control agent, MMPSense 750 FAST (PerkinElmer), is cleaved by a family of matrix metalloproteases (MMP-2, -7, -9, -13, and others) and not by renin. Imaging in normal, hairless SKH-1 mice shows a similar bladder signal when ReninSense is compared with MMPSense; however, kidney signal accumulation is evident only with ReninSense (Fig. 7, A and B) and highly statistically significant compared with MMPSense. Thus, despite similar levels of excretion via the bladder, only ReninSense appears to accumulate in the kidneys. Mouse images of these two agents (focusing only on the signal within the kidneys and bladder) support the clear differences between the agents (Fig. 7, C and D), and the ReninSense:MMPSense ratios in kidney and bladder are 8- to 16- and ∼1.4-fold, respectively (Fig. 7E) between 6 and 24 h.

Fig. 7.

Differential kidney/bladder biodistribution of ReninSense and a control activatable agent. Normal, hairless SKH-1 female mice were maintained on a normal diet before intravenous injection with ReninSense and the control agent MMPSense 750 FAST. Mice were imaged using the FMT 2500 at 6 and 24 h, and fluorescence levels in the bladder and kidney were measured for ReninSense (A) and MMPSense (B). 3D fluorescence molecular tomographic (FMT) images of mice show the fluorescent signal within the kidneys and bladder regions (C and D) with signals in other tissues excluded for clarity. E: normalization of ReninSense imaging data to MMPSense data was performed to highlight the differential kidney accumulation that occurs with ReninSense. Values are means ± SE. **Statistical significance (P < 0.001) in the ReninSense signal compared with MMPSense.

Imaging renin upregulation in vivo.

The ability of ReninSense to be cleaved, and thus provide a readout for renin activity in vivo, was investigated in mice placed for 48 h on either a normal diet, LS diet, or LS diet with L-810 (30 mg/kg po at 8, 22, and 30 h). ReninSense was injected intravenously at 24 h, and mice were imaged tomographically by FMT 2500 at 48 h (i.e., 24 h after ReninSense injection). As shown in Fig. 8A, the expected basal kidney fluorescence was detected in mice fed normal chow. The pattern and magnitude of fluorescence within the kidneys of mice on the LS diet, however, were increased considerably relative to normal control mice, and this increase was blocked by L-810 nearly to the levels measured in control mice. Axial tomographic slices through the kidneys in all three groups further revealed the extent of the differences in the amount of ReninSense activation under different experimental conditions (Fig. 8A, insets at right). Immediately after FMT imaging, mice were euthanized, and both plasma and kidneys were collected for analysis. Tissue epifluorescence imaging of excised kidneys (Fig. 8B) confirmed the different levels of fluorescence seen noninvasively by FMT in the three study groups, although less of a difference among groups was seen quantitatively due to the inherent limitations of 2D imaging (i.e., little or no detection of a deep tissue signal, increased background signal). For further confirmation of imaging results, representative kidneys from each group were snap-frozen in OCT and sectioned for fluorescence microscopy assessment. Representative NIRF microscopic images of 10-μm sections show the basal levels of activated ReninSense fluorescence in kidneys of mice on the normal diet, the increase induced by the LS diet, and the inhibitory effect of L-810 (Fig. 8C). The widespread interstitial distribution of active ReninSense throughout the kidney, rather than specific localization to the JG apparatus, is consistent with the 43-kDa molecular mass of the agent that minimizes both cellular uptake and clearance via the kidney. In the noninvasive FMT imaging data sets, the amount of activated agent per kidney (in pmol) was quantified in 3D regions of interest drawn around the sites of fluorescence defining each kidney (Fig. 8D). The LS diet caused a significant increase in the ReninSense signal (90 ± 6 vs. 25 ± 5 pmol, P < 0.001), which was inhibited by L-810 (40 ± 8 pmol compared with mice on the LS diet only, P < 0.001). The fluorescence intensities of the kidney tissues (Fig. 8E) showed a significant increase induced by the LS diet (relative fluorescence units in controls, P < 0.001) and the inhibition by L-810 (P < 0.001 compared with mice on the LS diet only), suggesting the noninvasive FMT quantification of kidney renin activity-related fluorescence in living mice. Although we had hoped to see a somewhat restricted distribution of a signal to the juxtaglomerular cells, we were not surprised that the drug-like nature of the ReninSense agent, and its ability to distribute generally into tissue, would provide only a widespread signal through the kidney at 24 h. Plasma from all three groups of mice was also collected, and fluorescence was determined in a fluorescence microplate reader (Fig. 8F). The LS diet caused a significant increase in the plasma ReninSense signal (P < 0.001 compared with controls), and the observed increase was significantly blocked by L-810 (P < 0.001 compared with mice on the LS diet only). These data collectively show the utility of ReninSense as a tool for assessing kidney renin activity in vivo as well as ex vivo in kidney and plasma samples.

Fig. 8.

In vivo imaging of renin activity in kidneys. Mice were fed a normal diet, LSD, or LSD plus renin inhibitor L-810 for 24 h before ReninSense was injected intravenously. FMT imaging was performed using the FMT 2500 to quantify the kidney signals 24 h postinjection (A), and quantitative data are presented in D. Insets: axial tomographic slices of the fluorescent data throughout the kidney regions to better highlight the magnitude of deep kidney signals. All mice were then euthanized, and kidneys as well as plasma were harvested. Epifluorescence images were acquired for all kidneys using a Kodak Image Station 2D planar imager (B), and quantitative data are presented in E. Representative kidney slides (C) were examined under a fluorescent microscope. Plasma fluorescent signals are presented in F. Data represent 1 of 3 experiments, and all images show only the signal within the 3D kidney region for clarity. Values are mean ± SE. **Statistical significance (P < 0.001).

DISCUSSION

The traditional concept of the RAS is that of a regulatory hormonal system within the circulation that is essential for the maintenance of blood pressure. Renin is released by the kidney to cleave circulating angiotensinogen, producing ANG I. ANG I, after passage into the pulmonary vasculature, is converted to ANG II, which circulates via the blood to peripheral tissues to affect a variety of biological systems and ultimately increase blood pressure. In the past decade, however, the focus of interest on the RAS in hypertension and organ injury has changed to include the importance of the local RAS in specific tissues (1, 7). Functional local RAS can be found in multiple tissues, including the brain, heart, adrenal glands, adipose tissue, gonads, pancreas, kidney, and vasculature (1, 9, 21), and several indirect lines of evidence suggest that the activities of the circulating RAS and the tissue RAS are both important, yet sometimes dissociated. Several researchers using experimental models of diabetes have shown normal or decreased PRA that does not reflect the measurable kidney renin increases (2, 6, 15). Similar dissociations have been seen in hypertensive patients (22) as well as in patients with type 1 diabetes (12).

There is further complexity in RAS biology, with studies over the past decade showing that prorenin, generally considered the inactive precursor form of renin, may deposit in a number of tissues and exhibit enzymatic activity; circulating prorenin has been shown to bind intrinsic prorenin-binding receptors in tissues, leading to intracellular signaling cascades (23) as well as receptor-mediated nonproteolytic activation (14) that allows prorenin to cleave angiotensinogen within the tissue. This type of activity of prorenin, outside of the traditional concepts of the RAS, may further dissociate the alignment of plasma and tissue renin responses in cardiovascular and kidney diseases (18). In such cases, this local cleavage of angiotensinogen in extrarenal tissue would likely not be a major contributor to plasma renin activity, so standard plasma assays alone would not be as useful as understanding and measuring both plasma and tissue levels of renin activity.

To provide an imaging tool to assess the complexity of RAS biology and tissue localization in vitro and in vivo, we developed a NIR renin-activatable agent. ReninSense was designed with a specific angiotensinogen peptide sequence that was modified and flanked by NIR fluorochromes to provide a construct that was optically silent in its intact form but became fluorescent upon cleavage by mouse renin. Modifications with a pharmacokinetic modifier served to extend the plasma and tissue half-lives, allowing maximal tissue penetration and sufficient time for activation within the sites of renin activity. In vitro tissue and plasma studies showed clear low levels of ReninSense activation in normal samples, with dramatic increases in fluorescence measured in samples from hyperreninemic mice. In kidney tissue slices, the ReninSense fluorescence was nearly ablated in both groups by a selective inhibitor of mouse renin, suggesting that the background ReninSense activation in control tissues was also due to renin activity. In contrast, there were only very low levels of background fluorescence seen in the plasma samples of control mice, in agreement with prior plasma assessments showing low levels of ReninSense cleavage.

In vivo imaging studies, using intravenous injection of ReninSense into control and hyperreninemic mice, also yielded similar patterns of kidney fluorescence increases, and in vivo L-810 treatment was able to inhibit the fluorescent signal in hyperreninemic mice to the level of the control animals. It was clear, however, that the normal levels of kidney ReninSense signal were as high as 20–25% of the levels measured in hyperreninemic mouse kidneys. As ReninSense is expected to behave similarly to native angiotensinogen (i.e., cleaving into fragments that circulate through the body for further cleavage), we considered the possibility that the basal kidney signal could be due to kidney clearance of ReninSense fragments generated by plasma renin. Indeed, we saw widespread signals throughout the kidneys rather than a signal specifically localized to the JG cells. Based on biodistribution research comparing ReninSense to a closely matched MMP-activatable agent, it is clear that only the renin activatable agent showed preferential accumulation within the kidney despite nearly identical agent composition and closely matched excretion via the bladder. This suggests that FMT kidney imaging of ReninSense predominantly quantifies tissue-localized renin activity, although we cannot exclude some small contribution by a circulating activated agent. The widespread distribution of the ReninSense signal within the kidneys appears to reflect the physicochemical properties of the agent, which (due to the 43-kDa molecular mass) activates extracellularly, is poorly taken up by cells, minimally cleared via the kidneys, and has a sufficient tissue half-life to allow local diffusion through kidney tissue over the 18 h in which it is maximally activated. Further studies using models and imaging analysis techniques designed to dissociate plasma and kidney renin activity will address this question more completely.

The optimal imaging time point for ReninSense is between 6 and 24 h, although the optimal signal to background is at 24 h. The imaging time in combination with whole-body washout requiring 5–6 days limits the utility of this imaging agent for repeated, short-term assessment of dynamic changes in renin activity. However, the agent is ideal for rapid noninvasive imaging in animal models involving disease-related changes in renin activity. The obvious advantage of this approach over conventional assessments of plasma renin changes is that the renin activity is assessed in vivo, in situ, in the tissue producing the renin activity. In instances where multiple tissue sources may contribute to circulating plasma renin levels, such an in situ assessment may be critical for disease/therapy assessment. The current studies have focused on kidney renin, but future studies are addressing the possibilities of assessing renin activity associated with other tissues and/or other disease states.

ReninSense also shows significant utility as an in vitro and ex vivo tool for assessing renin activity in plasma and tissue samples, as can be seen in the data from Figs. 2, 3, 4, 5, and 8. Many of the existing assays, whether for direct renin levels or for renin activity, are exquisitely specific due to the use of capturing monoclonals, but they require multiple incubation steps over a time period from hours to overnight. In contrast, ReninSense can be incubated with plasma samples and, with no additional sample manipulation, can be measured in a microwell plate as an increase in fluorescence at either a single end point or as a kinetic assay (see Fig. 4) compared with a standard curve with defined amounts of active renin. Measurements of plasma renin activity can also be made by sampling the blood of mice receiving an intravenous injection of ReninSense, and relative differences in ex vivo tissue renin can be assessed in frozen tissue sections with intact renin enzymatic activity. Figure 8 shows this full spectrum of use for ReninSense noninvasively in situ (kidney quantification), in tissue imaging ex vivo (whole tissue epifluorescence imaging; fluorescence microscopy), and looking at in vivo activation in plasma by fluorescence microplate reader ex vivo.

In conclusion, we have demonstrated the ability of ReninSense and FMT imaging to noninvasively visualize and quantify renin upregulation in the kidney in a robust and consistent manner. The consistency of the quantitative tomography, as well as its excellent correlation with gross and microscopic tissue fluorescence readouts, provides a powerful tool for quantifying alterations in tissue renin activity associated with hyperreninemia. Indeed, success of this noninvasive imaging approach relied heavily on the ability to see sufficiently deeply into the mouse to accurately detect a kidney signal; 2D epifluorescence imaging of the mice was unable to achieve reliable detection and quantitation of a kidney signal. This agent could also be used to assess plasma renin activity by harvesting plasma samples from mice receiving agent injection. No additional reagents or sample processing are required for the assessment of plasma fluorescence. Our findings suggest that ReninSense will provide a useful research tool for understanding changes in renin activity that occur in hypertension as well as in other disease areas, including polycystic kidney disease (19), cardiovascular hypertrophy (17), diabetic nephropathy (29), and central nervous system autoimmunity (28), moving the field of RAS biomarkers from plasma proteomics to in-life in situ tissue measurements of biological RAS changes.

DISCLOSURES

All authors are employed by PerkinElmer/VisEn Medical or Merck, and this is clearly stated in the manuscript. Funding of these studies was shared in a collaboration between PerkinElmer and Merck. The present research documents the utility of a novel PerkinElmer renin-imaging agent and imaging technology in assessing hyperreninemia, but the authors receive no financial gain as a result of publication or pending patents. Merck authors have no financial stake in PerkinElmer.

AUTHOR CONTRIBUTIONS

Author contributions: J.Z., K.O.V., J.M., J.D., and B. Bao. performed experiments; J.Z., K.O.V., and J.D.P. analyzed data; J.Z., D.V.P., K.M., and J.D.P. interpreted results of experiments; J.Z., D.V.P., and J.D.P. drafted manuscript; J.Z., D.V.P., B. Bednar, M.R., and J.D.P. edited and revised manuscript; J.Z., D.V.P., K.O.V., J.M., J.D., B. Bao, M.D.P., D.X., D.M., M.K., B. Bednar, C.S., D.Z.G., K.M., W.Y., M.R., and J.D.P. approved final version of manuscript; D.V.P., M.D.P., D.X., D.M., M.K., B. Bednar, C.S., D.Z.G., K.M., W.Y., M.R., and J.D.P. provided conception and design of research; J.D.P. prepared figures.