First analytical confirmation of drug-induced crystal nephropathy in felines caused by GS-441524, the active metabolite of Remdesivir

Abstract

GS-441524 is an adenosine nucleoside antiviral demonstrating significant efficacy in the treatment of feline infectious peritonitis (FIP), an otherwise fatal illness, resulting from infection with feline coronavirus. However, following the emergence of COVID-19, veterinary development was halted, and Gilead pursued clinical development of a GS-441524 pro-drug, resulting in the approval of Remdesivir under an FDA emergency use authorization. Despite lack of regulatory approval, GS-441524 is available without a prescription through various unlicensed online distributors and is commonly purchased by pet owners for the treatment of FIP. Herein, we report data obtained from the analytical characterization of two feline renal calculi, demonstrating the propensity for GS-441524 to cause renal toxicity through drug-induced crystal nephropathy in vivo. As definitive diagnosis of drug-induced crystal nephropathy requires confirmation of the lithogenic material to accurately attribute a mechanism of toxicity, renal stone composition and crystalline matrix were characterized using ultra-performance liquid chromatography mass spectrometry (LCMS), ultra-performance liquid chromatography photodiode array detection (UPLC-PDA), nuclear magnetic resonance (NMR) spectroscopy, Fourier-transform infrared spectroscopy (FTIR), and X-ray powder diffraction (XRD). This work serves to provide the first analytical confirmation of GS-441524-induced crystal nephropathy in an effort to support toxicologic identification of adverse renal effects caused by administration of GS-441524 or any pro-drug thereof.

Graphical Abstract

1. Introduction

Drug-induced crystal nephropathy is a phenomenon that occurs due to either intratubular precipitation of a given drug or drug metabolite, or the crystallization of endogenous substances secondary to a drug-induced physiologic changes [1]. Crystal nephropathy resulting from intratubular precipitation is commonly associated with drugs characterized by poor water solubility and significant renal elimination, resulting in intratubular saturation and subsequent crystallization. [1–4]. The risk of crystallization may be increased in patients with compromised renal function or in those with preexisting renal abnormalities that provide a nidus for homogenous nucleation to occur [5]. GS-441524 is poorly water soluble [6, 7], and elimination of GS-441524 relies predominantly on renal clearance, as demonstrated across multiple species [6]. In addition, current studies indicate that patients with reduced renal function display significant accumulation of GS-441524 [8, 9], a factor which may increase the risk of adverse renal effects.

While acute kidney injury resulting from direct nephrolith-associated tubular obstruction may be of concern in some patients, the primary mechanism of injury associated with drug-induced crystal nephropathy involves inflammation and cytotoxic damage due to crystal deposition within the tubular epithelium and renal interstitium [4, 7, 17–19]. Given the propensity for renal damage to occur in the absence of profound crystallopathy, this mechanism of toxicity may be difficult to identify, and currently, both the incidence and clinical relevance of GS-441524-induced crystal nephropathy remains unknown.

Clinical diagnosis of drug-induced crystal nephropathy is often complex, as drugcontaining calculi, including the analytes described in this study, appear radiolucent on abdominal radiograph, although their presence may be detected by ultrasonography [5]. Alternatively, atypical urinary crystals may be observed on urine sediment analysis or identified through histopathologic findings. Definitive diagnosis of drug-induced crystal nephropathy is based on stone composition analysis, predominantly by means of mass spectrometry, Fourier-transform infrared spectroscopy, or X-ray diffraction [5]. This requires stone samples to be obtained from the patient, through collection of urinary crystals, renal biopsy, or through post-mortem analysis.

We recently described the first clinical case report of crystal nephropathy in felines following administration of GS-441524 for the treatment of FIP [10]. Therefore, the main objective of this study was to provide comprehensive analytical evidence supporting definitive diagnosis of GS-441524-induced crystal nephropathy for these corresponding clinical cases. In this study, renal stone samples were of sufficient quantity and purity to undergo characterization by LCMS, UPLC-PDA, NMR, FTIR, and XRD. While a variety of analytical methods have been described in the literature for identification and quantification of GS-441524 in plasma or serum samples [11], to our knowledge, this study not only represents the first analytical confirmation of GS-441524-associated nephroliths, but also provides characterization of these nephroliths across a diverse set of analytical techniques.

2. Material and methods

2.1. Regents and standards

UPLC grade analytical solvents (Solvent A: water: 0.1% formic acid and Solvent B: acetonitrile: 0.1% formic acid) were purchased from Sigma Aldrich (Steinheim, Germany) and VWR International (Darmstadt, Germany). Analytical standards were purchased from Cayman Chemical (Ann Arbor, MI, USA; Item No. 30469, Batch No. 0660232 and 0588813). Renal stone #1 was obtained from Charleston Veterinary Referral Center and renal stone #2 was obtained from The University of Georgia College of Veterinary Medicine. Case studies and corresponding clinical data are reported separately [10].

2.2. Sample preparation

All analytical samples were stored dry at −20°C prior to analysis and either analyzed as a dry powder or dissolved in appropriate solvent prior to experimentation. For UPLC-PDA and LCMS analysis, renal stone samples were ground to a fine powder using a glass homogenizer and then dissolved in HPLC-grade H₂O at a concentration of 50 μg/mL (UPLC-PDA) or 80 μg/mL (LCMS). Samples were solubilized by agitating the solution using a vortex mixer (Fisher Scientific, Waltham, MA, USA) and heating to 60°C for 5 minutes in a sonicated water bath (VWR, Radnor, PA, USA). Prior to analysis, samples were filtered through an Acrodisc 0.22 μm wwPTFE syringe filter (Cytiva, Marlborough, MA, USA). For NMR analysis, samples were dissolved in DMSO-d6 at a concentration of 10 mg/mL (analytical standard) or ~2 mg/mL for renal stone samples.

2.3. Instrumentation

UPLC-PDA analysis was recorded on a Waters Acquity H-Class Plus Ultra-Performance liquid chromatographic system, which encompassed a quaternary solvent manager, a sample manager, and a photo diode array (PDA) detector (Waters Corp, Milford, MA, USA). Separation was achieved on a C18 reverse-phase column (Acquity UPLC BEH C18 1.7 M, 2.1 × 100 mm) with a column temperature of 30°C and a sample injection volume of 10 μL. UV absorbance was monitored over a range of 210–450 nm with a sampling rate of 20 points/sec and recorded at 250 nm and using Waters Empower 3 Pro software (v 7.00.00.99). For this study, we adapted a similar chromatographic method to previous studies aimed at detection of GS-441524 in plasma and serum samples [12–14]. The mobile phase consisted of 0.1% formic acid in water (phase A) and 0.1% formic acid in acetonitrile (phase B). Chromatographic separation was achieved at a flow rate of 0.3 mL/min following a gradient elution profile as follows: 0–0.35 min, 100% A; 0.35–1.5 min, 100%−95% A; 1.5–4 min, 95%−70% A; 4–6 min, 70%−70% A; 6–7 min, 70%−0% A. Subsequent to each run, the column was regenerated with 100% phase A for 3 minutes. Between each injection, the autosampler needle and injection port were washed with acetonitrile (strong wash) and methanol-water (20:80, v/v) (weak wash). To prevent of carry-over, a blank sample consisting of a 10 μL injection of methanol was analyzed between each analytical sample injection. In order to verify the concentration of GS-441524 present in renal stone samples, a calibration curve was established using a concentration range of 10 μg/mL to 70 μg/mL which demonstrated linearity based on peak AUC at 254 nm (R² = 0.996) (data available in supplementary material).

LCMS analysis was recorded on Waters Acquity H-Class Plus Ultra-Performance liquid chromatographic system, which encompassed a quaternary solvent manager, a sample manager, and a QDa mass detector. LCMS parameters included ion-source temperature of 600°C, a sampling rate of 8 Hz, an ESI capillary voltage of 0.8 kV, and a cone voltage ramp of 15–50 (0.047 V/Da). Samples were injected at a volume of 2 μL, and mass spectra were monitored in both positive and negative ionization modes between 100–800 m/z. Separation was achieved using a C18 reverse-phase column (Acquity UPLC BEH C18 1.7 M, 2.1 × 100 mm) held at 35°C. Analysis was performed on Waters MassLynx mass spectrometry software (V 4.2).

NMR spectroscopy was performed on a Bruker Avance Ultrashield 600 MHz NMR spectrometer (Bruker Corp, Billerica, MA, USA) controlled by TopSpin 3.1 software. Tetramethylsilane (TMS) was used as an internal standard. Samples were prepared in DMSO-d6 and chemical shifts are reported as δ values with reference to TMS. NMR spectra were analyzed using TopSpin software (V 4.1.3).

XRD analysis was performed on a Malvern Panalytical Benchtop X-ray Powder Diffractometer Aeris Research Edition (Malvern Panalytical, Worcestershire, United Kingdom) and analyzed on HighScore Plus software (V 4.6a). FTIR analysis was performed on a Perkin Elmer Spectrum 2 Fourier-transform infrared spectrophotometer (PerkinElmer, Shelton, CT, USA) and analyzed on PerkinElmer Spectrum 2 software (V 10.5.4). Microscopy was performed on an Olympus SZ series stereoscope (Olympus Corp, Shinjuku, Tokyo, Japan) and images were captured with a SWIFT EP microscope adapter, Model num. EP5R (Microscope World, Carlsbad, CA, USA).

3. Results and discussion

3.1. Qualitative analysis

Sample #1 consisted of an irregularly shaped calculus and a few irregularly shaped calculus fragments with a total weight is 85.7 mg. Sample #2 consisted of two, minute, crenated, irregularly shaped calculi, weighing a total of 1.3 mg. Both renal stone samples were comprised of a distinct crystalline matrix with a complex lattice structure, indicative of crystal nephropathy resulting from homogenous nucleation of GS-441524 (Figure 1).

Figure 1.

Stereo microscopy images demonstrating compact crystallization pattern of feline renal stone #1, indicative of homogenous nucleation of GS-441524.

3.2. Ultra-performance liquid chromatography photodiode array detection

The chromatographic method described above resulted in a single distinct peak for all analytical samples, with a column retention time of 3.8 minutes and a maximum absorption wavelength of 238.6 nm (Figure 3). Of note, analysis of higher concentrations (> 80 μg/mL) resulted in minor deviations in retention time and a less defined UV spectra with a maximum absorption wavelength of 245 nm. Quantitative analysis of renal stone samples ranging 20 μg/mL to 60 μg/mL were consistent with values determined by the standard calibration curve indicative of a high degree of purity. This determination is consistent with all analytical findings in this study, which did not detect the presence of any additional compounds within the renal stone samples.

Figure 3.

UPLC-PDA comparative ultra-violet spectra of GS-441524 analytical standard, renal stone #1, and renal stone #2.

3.3. Ultra-performance liquid chromatography mass spectrometry

Both renal stone samples demonstrated consistent results with the GS-441524 analytical standard, with a base peak of m/z 292 [M+H]⁺, as well as the presence of a sodium adduct at m/z 314 [M+Na]⁺. GS-441524 was also detected in negative ionization mode with peaks at m/z 290 [M-H]⁻ and 291 [M-2H]⁻. ESI mass spectra in positive (ES+) and negative (ES−) ionization modes are summarized in Figure 3A. A variety of studies including both mechanistic [15] and methodological [12–14, 16, 17] investigations have been reported for detection of GS-441524. In this study, all analytical samples displayed fragmentation patterns consistent with previously reported literature. A corresponding proposed fragmentation scheme is depicted in Figure 3B, based on the mechanistic investigations by Dadinaboyina et al. [15]. The depicted fragmentation pattern is described as follows: (i) loss of HCN, corresponding to m/z 265 [M+H]⁺ and m/z 287 [M+Na]⁺ in ES+, with values of m/z 264 [M-H]⁻ and m/z 263 [M-2H]⁻ in ES-. (ii) Loss of C₃H₆O₃ corresponding to m/z 202 [M+H]⁺, (iii) Loss of H₂O, corresponding to m/z 245 [M-H] ⁻. The fragments displayed in Figure 3 represent 97±0.15% of the total ion current (TIC) in ES+ for all three analytical samples (monitored 100–800 m/z, analyzed 150–350 m/z, centered across major peak). Raw spectra available in supplementary material (S14–S15).

3.4. NMR spectroscopy

NMR peak assignments were determined based on chemical shifts and splitting patterns and confirmed using Heteronuclear Multiple Quantum Coherence (HMQC), ¹H-¹H Correlated Spectroscopy (COSY), and ¹³C Attached Proton Test (APT) NMR analysis (Figure 4). Exchangeable protons were confirmed through the addition of 20 μL D₂O to samples containing DMSO-d6. Both ¹H and ¹³C NMR spectra for GS-441524 have been previously reported and the results from this study were largely consistent with published data [18–20]. Overall, chemical shifts and splitting patterns were highly consistent across all analytes. As renal stone samples were analyzed at a low concentration without any attempt at further purification, slight broadening of peaks and minor variations in chemical shifts (less than 0.03 ppm) were observed between samples on ¹H NMR. Comparative ¹H NMR chemical shifts and carbon assignments are reported in Table 1 and are detailed as follows: (i) a singlet at δ 7.91 ppm corresponding to C₅, (ii) a broad NH₂ peak at δ 7.88 ppm, overlapping with the C₅ peak, (iii) a doublet at δ 6.90 ppm corresponding to C₁, (iv) a doublet at δ 6.87 ppm corresponding to C₉, (v) a doublet at δ 6.07 ppm corresponding to OH₁₈, (vi) a doublet at δ 5.18 ppm corresponding to OH₁₇, (vii) a doublet of doublets at δ 4.90 ppm corresponding to OH₁₆, (viii) a doublet of doublets at δ 4.65 ppm corresponding to C₁₁, (ix) a doublet of doublet of doublets at δ 4.06 ppm corresponding to C₁₃, (x) a doublet of doublet of doublets at δ 3.96 ppm corresponding to C₁₂, (xi) The C₁₅ protons are diastereotopic relative to the C₁₃ and OH₁₆ protons, leading to two separate doublet of doublet of doublets at δ 3.64 ppm and δ 3.51 ppm, reflected by both geminal coupling to each other and vicinal coupling to C₁₃ and OH₁₆ protons. The ¹³C chemical shifts were consistent across all analytes. Due to limited quantity of renal stone #2, not all peaks resolved on ¹³C NMR, however, detectable peaks for all analytes are reported in Table 2.

Figure 4.

Heteronuclear Multiple Quantum Coherence (HMQC) NMR spectrum of GS-441524 with corresponding carbon assignments.

Table 1.

Comparative ¹H NMR chemical shifts and carbon assignment for analytical samples in DMSO-d6 (600 MHz).

	Analytical Standard	Renal Stone #1	Renal Stone #2
	ppm	ppm	ppm		J (Hz)
C 5	7.91	7.92	7.91	s, 1H
NH 2	7.88	7.87	7.87	br, 2H
C 1	6.90	6.90	6.90	d, 1H	4.5
C 9	6.87	6.88	6.87	d, 1H	4.5
OH 18	6.07	6.10	6.08	d, 1H	6.2
OH 17	5.18	5.19	5.19	d, 1H	5.2
OH 16	4.90	4.92	4.90	dd, 1H	6.1, 5.1
C 11	4.65	4.64	4.64	dd, 1H	6.2 , 5.5
C 13	4.06	4.06	4.06	ddd, 1H	5.2, 4.8, 3.5
C 12	3.96	3.95	3.96	ddd, 1H	5.5, 5.2, 5.2
C l5α	3.64	3.64	3.64	ddd, 1H	12.1, 5.1, 3.5
C 15β	3.51	3.51	3.51	ddd, 1H	12.1, 6.1, 4.8

s = singlet, br = broad peak, d = doublet, dd = doublet of doublets, ddd = doublet of doublet of doublets. Reported J values correspond to ¹H NMR spectra for GS-441524 analytical standard.

Table 2.

Comparative ¹³C APT NMR chemical shifts and carbon assignment for analytical samples in DMSO-d6 (600 MHz).

	Analytical standard	Renal stone #1	Renal stone #2
C3	156.6	156.6	156.6
C5	147.8	147.8	-
C8	123.8	123.8	123.8
CN	117.3	117.3	117.3
C2	116.5	116.5	-
C9	110.7	110.7	110.7
C1	100.7	100.7	100.7
C13	85.4	85.4	85.4
C10	78.6	78.6	-
C11	74.2	74.2	74.2
C12	70.0	70.0	70.0
C15	60.9	60.9	60.9

3.5. Fourier-transform infrared spectroscopy

Renal stone samples were analyzed by FTIR, and the resulting spectra were compared to GS-441524 analytical standard (Figure 5B.). Both analytical samples produced IR absorbance patterns consistent with the analytical standard, including of a series of characteristic strong absorbance bands between 3200–3500 cm⁻¹, indicative of N-H/ O-H stretching vibrations, as would be expected given the presence of the primary amine moiety and ribose hydroxyl groups present in GS-441524. In addition, all samples demonstrated weak CN stretching at 2244 cm⁻¹, indicative of the presence of a nitrile moiety. Major distinguishable bands > 20000 cm⁻¹ for all analytes are displayed in Table 3. Spectra and complete peak list for all analytes are reported in supplementary material (S16–S19).

Figure 5.

A) XRD diffractogram for renal stone #2 (top) and GS-441524 analytical standard (bottom) between 0–50°2θ. Complete XRD peak list available in supplementary material. B) FTIR spectra for renal stone #1 (top) and GS-441524 analytical standard (bottom). FTIR spectra for all samples and complete peak list available in supplementary material.

Table 3.

Fourier transform infrared spectroscopy major distinguishable bands > 20000 cm⁻¹

	Analytical standard	Renal stone #1	Renal stone #2
3200–3500 cm−1 NH2, OH	3476.63	3476.97	3476.78
	3430.52	3430.50	3430.70
	3340.21	3340.33	3340.56
	3239.63	3239.92	3239.69
3000–3200 cm−1 C=C-H	3189.02	3190.32	3190.87
	3092.48	3092.63	3092.65
2700–3000 cm−1 C-H	2949.91	2949.74	2949.87
	2912.24	2913.13	2913.18
	2854.39	2854.22	2854.68
	2770.32	2770.28	2770.09
	2725.73	2726.84	2725.83
2244 cm−1 CN	2244.24	2244.34	2244.31

3.6. X-ray diffraction

Overlaid X-ray diffractograms for GS-441524 analytical standard and renal stone #2 are depicted in Figure 5A. While characterization of crystal structure was not within the scope of this study, these data provide valuable evidence supporting the assertion that the feline uroliths analyzed in this study were formed by means of homogenous nucleation of GS-441524, rather than the formation of a unique co-crystal or salt adduct thereof. XRD analysis resulted in the identification of 30 distinct shared peaks of intensity greater than 1%, between 0–54°2θ (S21).

4. Conclusion

Consistent results were observed across all analytical techniques performed in this study, demonstrating that the identified renal stones were solely comprised of GS-441524. The unique purity of biologic samples formed by crystallization allowed for comprehensive analytical characterization, without requiring additional steps of purification or extraction. This work serves to provide both definitive confirmation of the chemical makeup of the tested analytes, along with corresponding spectral data, in an effort to support the identification of biologic samples consisting of GS-441524. As the results of this study confirm the propensity for GS-441524 to cause adverse renal effects through homogenous nucleation in vivo, further clinical studies are warranted, as GS-441524-induced crystal nephropathy may represent an underdiagnosed and underreported condition.

Supplementary Material

1

NMR Analysis

S1. 1H NMR, 600 MHz, D6-DMSO, Analytical Standard

S2. 1H NMR, 600 MHz, D6-DMSO, Renal stone #1

S3. 1H NMR, 600 MHz, D6-DMSO, Renal stone #2

S4. COSY 1H NMR, 600 MHz, D6-DMSO, Analytical Standard

S5. NOESY 1H NMR, 600 MHz, D6-DMSO, Analytical Standard

S6. 1H NMR, 600 MHz, D6-DMSO + D2O, Analytical Standard

S7. 13C APT NMR, 600 MHz, D6-DMSO, Analytical Standard

S8. 13C APT NMR, 600 MHz, D6-DMSO, Renal stone #1

S9. 13C APT NMR, 600 MHz, D6-DMSO, Renal stone #2

UPLC-PDA Analysis

S10. UPLC-PDA spectra for GS-441524 analytical standard

S11. UPLC-PDA spectra for renal stone #1

S12. UPLC-PDA spectra for renal stone #2

S13. UPLC-PDA Calibration Curve

LCMS Analysis

S14. Mass spectra for all samples (ES+)

S15. Mass spectra for all samples (ES−)

FTIR Analysis

S16. FTIR spectra for GS-441524 analytical standard

S17. FTIR spectra for renal stone #1

S18. FTIR spectra for renal stone #2

S19. FTIR peak list for all samples

XRD Analysis

S20. XRD diffractogram

S21. XRD peak list

Figure 2.

A) Mass spectra of GS-441524 analytical standard, renal stone #1, and renal stone #2 in both positive and negative ionization modes. Major peaks ≥ 1% total ion current (TIC) between 200–350 m/z are shown. B) Proposed fragmentation pattern for GS-441524 based on mechanistic investigations of remdesivir by Dadinaboyina et al. [19].

Highlights.

This work describes the first analytical confirmation of GS-441524 toxicity in vivo.

Samples were obtained from felines with suspected GS-441524 crystal nephropathy.

Renal stones were characterized by UPLC-PDA, LCMS, NMR, XRD, and FTIR.

Consistent spectral results were observed across all analytical techniques.

Data Availability

Complete spectral analysis provided in supplemental material. Additional data available upon request.