Evidence For and Against Direct Kidney Infection by SARS-CoV-2 in Patients with COVID-19

Abstract

Despite evidence of multiorgan tropism of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in patients with coronavirus disease 2019 (COVID-19), direct viral kidney invasion has been difficult to demonstrate. The question of whether SARS-CoV2 can directly infect the kidney is relevant to the understanding of pathogenesis of AKI and collapsing glomerulopathy in patients with COVID-19. Methodologies to document SARS-CoV-2 infection that have been used include immunohistochemistry, immunofluorescence, RT-PCR, in situ hybridization, and electron microscopy. In our review of studies to date, we found that SARS-CoV-2 in the kidneys of patients with COVID-19 was detected in 18 of 94 (19%) by immunohistochemistry, 71 of 144 (49%) by RT-PCR, and 11 of 84 (13%) by in situ hybridization. In a smaller number of patients with COVID-19 examined by immunofluorescence, SARS-CoV-2 was detected in 10 of 13 (77%). In total, in kidneys from 102 of 235 patients (43%), the presence of SARS-CoV-2 was suggested by at least one of the methods used. Despite these positive findings, caution is needed because many other studies have been negative for SARS-CoV-2 and it should be noted that when detected, it was only in kidneys obtained at autopsy. There is a clear need for studies from kidney biopsies, including those performed at early stages of the COVID-19–associated kidney disease. Development of tests to detect kidney viral infection in urine samples would be more practical as a noninvasive way to evaluate SARS-CoV-2 infection during the evolution of COVID-19–associated kidney disease.

Introduction

Although acute lung injury is the most prominent clinical manifestation in patients with severe coronavirus disease 2019 (COVID-19), AKI is also frequently observed. The reported incidence of AKI in COVID-19 ranges between 22% and 57% in patients who are hospitalized, and it is associated with high mortality (1–13). Angiotensin-converting enzyme 2 (ACE2), a protein that acts as the chief receptor for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) cell entry (14–17), is highly expressed in the kidney (18–20). However, it has been difficult to establish if the virus directly infects the kidney parenchyma, as recently pointed out by Khan et al. (21). The evidence in favor of, or against, direct kidney invasion by SARS-CoV-2 will be juxtaposed on Koch’s postulates, later revised for viral infections by Thomas Rivers in 1937, and more recently by Fredericks and Relman (22). We highlight the more relevant findings in support of or against kidney infectivity in lieu of a detailed point-by-point account of Koch’s postulates to establish whether SARS-CoV-2 can directly infect the kidney. This answer will be important in defining the pathophysiology of kidney injury seen in patients with severe COVID-19.

Although AKI is commonly identified in patients with COVID-19, more unique kidney manifestations, such as collapsing glomerulopathy, have also been described (23–38). AKI in patients with COVID-19 may be caused by factors common to a majority of cases of AKI in patients who are critically ill in the intensive care unit, including hypotension, sepsis, and exposure to nephrotoxins (39–42). There are, however, additional features that suggest a more complex pathophysiology (1,43). AKI in patients with COVID-19 could be mediated by overactivation of the innate immune system, cytokine release, complement activation, angiotensin II (Ang II) overactivity, the development of a hypercoagulable state, hypovolemia secondary to over diuresis, and/or increased central venous pressure secondary to high positive end-expiratory pressure (1,3,43–45). Analysis of 2600 patients admitted to the hospital with COVID-19 showed that after adjustment for demographics, comorbidities, vital signs, medications, and laboratory results, COVID-19 remained highly associated with AKI (46). Viral invasion of the kidney, if it occurs, could be an additional contributing factor to AKI and collapsing glomerulopathy. The prognosis may be worse than regular AKI, but further data are needed to understand the evolution of AKI and possible transition to CKD in some patients with COVID-19. This review examines the evidence in favor of and against SARS-CoV-2 kidney infection in patients with COVID-19 reported to date. Although most autopsy studies show no convincing evidence for SARS-CoV-2 in the kidney, the evidence in other studies is strong, perhaps because a more comprehensive analysis involving immunofluorescence (IF), RT-PCR, and in situ hybridization was performed.

Biology of Severe Acute Respiratory Syndrome Coronavirus 2 and Localization and Function of Full-Length Angiotensin-Converting Enzyme 2, its Main Receptor

SARS-CoV-2 belongs to the family of Coronaviridae, which are classified in α-, β-, γ, and δ virus (47). The positive-strand RNA genome of coronavirus is surrounded by a helical capsid, the nucleocapsid protein, which is surrounded by a lipid bilayer envelope (47). This envelope consists of membrane protein, envelope protein, and spike (S) protein (47). The S protein of SARS-CoV-2, a β coronavirus, has two components: S1, which contains the receptor-binding domain, and S2, which contains the fusion peptide (14,48). The S protein mediates cell entry and therefore is critical for the virus host range, but is also involved in inducing the host immune response (47). ACE2 in its full-length form is the main receptor that SARS-CoV-2 uses to enter host cells (2,14–17). In addition, ACE2 acts as monocarboxypeptidase cleaving phenylalanine from the C-terminus of its substrates, for example, causing the formation of Ang- (1–7) from Ang II (49–51). Other substrates of ACE2 include apelin-13, apelin-36, and the proinflammatory peptide des-arg⁹ bradykinin (52–54).

It is widely believed that surface membrane ACE2 decreases in COVID-19 as a result of ACE2 virus–protein complex internalization, and the deficit of ACE2 renders the infected organs more vulnerable to Ang II and des-arg⁹ bradykinin (2,55–57). Accumulation of these peptides can foster organ injury and adverse outcomes (2), especially in organs that express ACE2. In a study of ACE2 mRNA expression in 72 human tissues using real-time PCR, high levels were found in the kidney, testis, and cardiovascular tissues (58). In human lungs, mRNA expression of ACE2 can be detected on type II pneumocytes (59), and at the protein level by immunohistochemistry (IHC) on type I and type II pneumocytes (60). The latter cell type is considered the chief site of pathogenic infection by SARS-CoV-2. In the kidneys, ACE2 is abundantly expressed in the proximal tubule apical membrane (18–20) (Figure 1A). As demonstrated by IF and immunogold labeled electron microscopy (EM), mouse glomerular parietal and visceral epithelial cells (podocytes) also express full-length ACE2, but in much smaller amounts than the proximal tubule (Figure 1B) (18,19,61). In agreement, ACE2 has been detected in human proximal convoluted tubules and parietal epithelial cells of the Bowman’s capsule by IF (62).

Figure 1.

Immunofluorescence and immunogold analysis of angiotensin-converting enzyme 2 (ACE2) in the kidney. (A) Immunofluorescence staining of ACE2 (red) in proximal tubules. (B) ACE2 immunogold labeling in glomeruli. ACE2 labeled with 15 nm of gold particles is distributed in podocyte foot processes and slit diaphragm (A, arrows). The glomerular basement membrane (GBM) does not have ACE2 immunogold particles (modified from ref. 19 with permission).

SARS-CoV-2 infection is also governed by specific proteases found in each cell type (14), chiefly TMPRSS2. Single-cell transcriptome analysis revealed that TMPRSS2 is highly expressed in the distal nephron, not in the proximal tubule (1). In kidney organoids of embryonic origin, ACE2 and TMPRSS2 colocalization can be seen by IF in proximal tubule–like structures (63). The areas of colocalization in these organoids are in the presumed apical border area, where ACE2 is abundantly expressed in the adult. Clearly, studies in human kidneys are needed to clarify the issue of ACE2 and TMPRSS2 localization within the kidney. It is possible that other proteases, similar to TMPRSS2, prime SARS-CoV-2 for internalization with ACE2 in the proximal tubule. Other proteases, such as furin and cathepsin L, which are necessary for SARS-CoV-2 processing (64), are more ubiquitously expressed, and are found in the proximal tubule (65,66).

How Could Severe Acute Respiratory Syndrome Coronavirus 2 Reach the Kidney?

Access of the virus to the kidney is obviously not as direct as it is in the lungs, where the route SARS-CoV-2 uses to reach these organs is inhalation. The mechanisms proposed here are theoretical considerations because our understanding of kidney infection by SARS-CoV-2 is still evolving. Viremia would be the expected route for SARS-CoV-2 to reach the kidney. Viremia, even transient, could lead to SARS-CoV-2 entry by binding to ACE2 in podocytes, providing an initial nidus for subsequent viral invasion of the kidney parenchyma. Most patients with COVID-19, however, do not have documented viremia detected by RNA levels (67,68). There is evidence that higher viral RNA loads in plasma are associated with increased disease severity and mortality in patients with COVID-19 (67,68). The possibility of viral exposure to podocytes needs to be considered. SARS-CoV-2 RNA and S protein have been documented in the glomerulus in autopsies of patients with COVID-19 (69,70). This could reflect viral RNA stuck in the glomerular filtration barrier, rather than viral infection. Evidence for SARS-CoV-2 RNA or protein in the glomerular cells nevertheless has been documented using both in situ hybridization and IF (69,70).

One next needs to consider how SARS-CoV-2 could reach the lumen of the proximal tubule to bind to the abundant ACE2 receptors in the apical membrane (1,21), instead of the basolateral membrane, which is in potential contact with the virus present in the blood stream. One possible circumferential route to the apical membrane of the proximal tubule is via the tubular fluid. After SARS-CoV-2 infects podocytes, access to the tubular fluid and subsequent binding to ACE2 in the apical membrane of the proximal tubule is a potential route. Alternatively, the virus could reach the apical membrane of the proximal tubule during cellular injury of the proximal tubule in patients with AKI. In this scenario, the virus could traverse the cell to interact with apical ACE2 protein. In primary human airway epithelia, ACE2 is expressed apically, and SARS-CoV-2 infection predominantly occurs on the apical surface, but infection can occur on the basolateral surface at low efficiency (71). This perhaps involves low level transcytosis of ACE2 to the basolateral membrane of the cell. Another possibility that lacks experimental support but is worthy of consideration is that when kidney damage occurs, there may be aberrant expression of ACE2 in the basolateral membrane. This is a possibility given the report of altered CD147 expression, another putative receptor for SARS-CoV-2, in patients with COVID-19. CD-147 is expressed basolateral on proximal and distal tubular epithelial cells (72) and, thus, could mediate SARS-CoV-2 cell entry.

How Can Severe Acute Respiratory Syndrome Coronavirus 2 Be Detected in the Kidney?

Different methods can be applied to detect the presence of SARS-CoV-2 in the kidney (Table 1). Strong evidence for the presence of viral RNA can be derived from in situ hybridization (73), RT-PCR (74), or viral growth in plaque assays (75,76). However, overinterpretation of these results is possible, especially with RT-PCR–based strategies on whole kidney samples because they will not distinguish parenchymal infection from the presence of virus within blood or urine. The nucleocapsid or S protein of SARS-CoV-2, moreover, can be detected by either IHC or IF. We found 14 reports using nine different antibodies to detect either spike protein or nucleocapsid protein of SARS-CoV-1 or -2 (Table 2). One important limitation of IHC is the potential of crossreactivity. Detection of SARS-CoV-2 proteins by IHC was less sensitive and specific than detection of SARS-CoV-2 RNA by RT-PCR or in situ hybridization in one report (77). Others found no difference between detection of SARS-CoV-2 by IHC and in situ hybridization (73). Other factors that can influence detection of proteins include tissue fixation, unmasking procedures, antibody dilution, and detection systems (78). Advanced techniques, such as protein mass spectrometry, which are more sensitive to detecting the presence of SARS-CoV-2 protein in nasopharynx epithelial swabs (79), could be used in studies to detect SARS-CoV-2 in kidney or the urine. EM is another common method of detection of viral-like particles, however, does not provide definitive evidence for SARS-CoV-2 virions (80).

Table 1.

Methods of viral detection to demonstrate presence of severe acute respiratory syndrome coronavirus 2

Method	What Is Being Detected?
Electron microscopy	Virus-like particles
Immunohistochemistry	Viral protein
Immunofluorescence with confocal microscopy	Viral protein
In situ hybridization	Viral RNA
RT-PCR	Viral RNA
Plaque assay	Live virus
Protein mass spectrometry	Viral protein

Table 2.

Findings of severe acute respiratory syndrome coronavirus 2 protein or RNA in kidney of patients with coronavirus disease 2019

Author	Patients Stated to Have Severe Acute Respiratory Syndrome Coronavirus 2 Kidney Presence	Immunohistochemistry: Viral Protein	Immunofluorescence: Viral Protein	RT-PCR: Viral RNA	In situ Hybridization: Viral RNA	Postmortem Interval
Findings with evidence (postmortem)
Braun et al. (83)	38/63	Not done	Not done	38/63	Not done	Median: 5 days
Puelles et al. (69)	16/27	Not done	Exact number not given; spike glycoprotein antibody (3A2) (Abcam, ab272420), SARS-CoV SΔ10 within S2 domain protein (Genetex, GTX632604)	16/26	Exact number not given; RNA scopea	Average: 2.8 days
Remmelink et al. (85)	10/17	Not done	Not done	10/17	Not done	72–96 hrs
Bouquegneau (70)	12/16	9/16; 2019-nCoV N-Protein (NP) rabbit polyclonal antibody (ABclonal #A20016)	Not done	1/16	6/16; RNA scopea	< 3 hrs
Su et al. (9)	8/10	Not done	3/6; anti–SARS-CoV nucleo-protein antibody (40143T62; Sino Biologic, China)	Not done	Not done	1–6 hrs
Bradley et al. (81)	5/6	2/4; monoclonal antibody to the SARS-CoV-2 spike protein (GeneTex; Irvine, CA, USA)	Not done	3/3	Not done	<140 hrs
Hanley et al. (84)	3/5	Not done	Not done	3/5	Not done	<10 days
Schurink et al. (82)	1/11	1/11; noncommercial monoclonal mouse antibody and polyclonal rabbit antibody for SARS-CoV-2 nucleocapsid protein (Sino Biologic & Nanommune, Irvine, CA, USA)	Not done	Not done	Not done	Median: 15 hrs
Ichimura et al. (unpublished observations)	1/1	Not done	1/1; SARS-CoV nucleocapsid (rabbit, PA1–41098; Invitrogen, Waltham, MA, USA)	Not done	Not done	Not given
Diao et al. (13)	6/6	6/6; anti–SARS-CoV-2 nucleocapsid protein antibody (Sino Biologic, Beijing, China or ab273434, Abcam) and anti-SARS spike glycoprotein antibody (ab273433, Abcam)	6/6; anti–SARS-CoV-2 nucleocapsid protein antibody (Sino Biologic, Beijing, China or ab273434, Abcam) and anti-SARS spike glycoprotein antibody (ab273433, Abcam)	Not done	3/3; RNA scopea	< 24hrs
Findings with low evidence (biopsies)
Kudose et al. (97)	2/16	0/16; mouse monoclonal IgG1 antibody against S2 subunit from SARS-CoV-2 spike protein from clone 1A9 (GeneTex, Irvine, CA, USA) and rabbit monoclonal antibody against nucleocapsid protein clone 001 (Sino Biologic, Beijing, China) (40143-R001)	Not done	Not done	0/16 when performed automatically; 2/16 when performed manually; RNA scopea	—
Findings with no evidence (biopsies)
Sharma et al. (98)	0/10	0/10; antibody to SARS-CoV-2 nucleocapsid protein (Clone 1C7; Bioss, Woburn, MA)	Not done	Not done	Not done	—
Akilesh et al. (25)	0/8	0/4; antibody to SARS-CoV nucleocapsid protein (40143-T62, Sinobiological, Wayne, PA, USA)	Not done	Not done	0/4; RNA scopea	—
Huang et al. (99)	0/1	0/1; antibody to spike protein of SARS-CoV-2 (40150-R007), Sino Biologic, Beijing, China)	Not done	0/1	Not done	—
Wu et al. (24)	0/6	Not done	Not done	Not done	0/6; RNA scopea	—
Sharma et al. (32)	0/2	Not done	Not done	Not done	0/2; RNA scopea	—
Nasr et al. (33)	0/1	Not done	Not done	Not done	0/1; RNA scopea	—
Larsen et al. (24)	0/1	Not done	Not done	Not done	0/1; RNA scopea	—
Peleg et al. (28)	0/1	Not done	Not done	Not done	0/1; not given	—
Couturier et al. (34)	0/1	Not done	Not done	0/1	Not done	—
Kissling et al. (36)	0/1	Not done	Not done	0/1	Not done	—
Lazareth et al. (27)	0/1	Not done	Not done	0/1; Crystal Digital PCR-TM (Stilla Technologies, Villejuif, France)	Not done	—
Findings with no evidence (postmortem)
Golmai et al. (100)	0/12	0/12; primary mouse antibody for SARS-CoV-2 nucleocapsid protein (Clone 1C7, Bioss Woburn, MA, USA)	Not done	Not done	0/4; RNA scopea	Not given
Massoth et al. (77)	0/7	0/3; SARS nucleocapsid antibody (NB100–56576; Novus Biologicals)	Not done	0/5	0/7; RNA scopea	Not given
Brook et al. (101)	0/3	0/3; SARS Rabbit Polyclonal Nucleocapsid Protein Antibody (Novus NB100–56576)	Not done	0/3	0/3; RNA scopea	<3 hours
Sekulic et al. (102)	0/2	Not done	Not done	0/2	Not done	29–39 hrs
Rocha et al. (73)	Not given	0/8; antibodies to recombinant SARS-CoV-2 nucleocapsid protein/recombinant SARS nucleocapsid protein, Bioss, Woburn, MA, USA	Not done	Not done	0/10 RNA scopea	Not given
Santoriello et al. (87)	Not given	Not done	Not done	Not done	0/10; RNA scopea	Median: 21.8 hrs
Summary	Total number of patients with positive findingsb	IHC	IF	RT-PCR	In situ hybridization
	102/235 patients (43%)	18/94 patients (19%)	10/13 patients (77%)	71/144 patients (49%)	11/84 patients (13%)

SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; IHC, immunohistochemistry; IF, immunofluorescence.

^a By ACD, Newark, CA, USA.

^b The kidney of some patients was examined by more than one method, but for the total number, every patient was counted just once. Therefore, the total number of patients is lower than when adding up the number of patients examined by each individual method. Two reports (references 73,87) do not specify from how many patients they derived their sample and were excluded for the calculation of the total number of patients examined.

Evidence For and Against SARS-CoV-2 in the Kidney from Patients with COVID-19

In contrast to the lungs, it has been difficult to demonstrate the presence of SARS-CoV-2 in kidneys (Tables 2 and 3 and summarized in Figure 2). Particles that resemble the appearance of coronaviruses are clearly not sufficient to unequivocally document direct viral invasion (80). The family of Coronaviridae, moreover, is very large and common, and other members of the family, such as the common cold virus, could be mistaken for SARS-CoV-2. Nevertheless, it is worthwhile to mention studies that reported such particles in the early descriptions of patients with COVID-19. In total, seven of 21 studies that we found reported potential viral-like particles by EM. Of the 128 patients examined by EM, potential viral-like particles were found in the kidney of 16 patients (13%). In total, 11 samples were from autopsied kidneys (9,13,81) and only five from kidney biopsies performed in living patients; the latter were all in patients who were subsequently diagnosed with collapsing glomerulopathy (35–38).

Table 3.

Findings suggestive of severe acute respiratory syndrome coronavirus 2 in patients with coronavirus disease 2019 (extracted from Table 2)

Finding	Viral Protein by Immunohistochemistry	Viral Protein by Immunofluorescence	Viral RNA by In situ Hybridization	Viral RNA by RT-PCR
Frequency of findings from patients examined	18/94 (19%)	10/13 patients (77%)	11/84 patients (13%)	71/144 (49%)
References showing detection	(13,70,81,82)	(9,13,69 and Ichimura et al., unpublished observations)	(13,69,70,97)	(69,70,81,83–85)
References showing no detection	(25,73,77,97–101)	–	(24,25,28,32,33,73, 77,87,100,101)	(27,34,36,77,99,101,102)

Figure 2.

Summary of data against and in favor of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in kidneys from patients with coronavirus disease 2019 (COVID-19). Number and percentage of patients in whom SARS-CoV-2 spike or nucleocapsid protein or RNA was detected (red) or not (blue). Each circle depicts immunohistochemistry, immunofluorescence, RT-PCR, or in situ hybridization. Data extracted from Tables 2 and 3.

Four of 12 IHC-based studies reported detection of viral protein in the kidney. Of the 94 patients with COVID-19 examined by IHC, 18 showed evidence for the presence of viral proteins in the kidney (19%) (Figure 2). Only postmortem samples stained positive for SARS-CoV-2 protein by IHC (13,70,81,82). Viral protein was accessed in four studies by immunofluorescent microscopy, all of them autopsy studies; viral protein was found in the kidney in 10 of 13 patients examined (77%) (9,13,69,and Ichimura et al., unpublished observations) (Figure 2). As detected by in situ hybridization, SARS-CoV-2 RNA was only found in four of 14 studies and 11 of 84 patients (13%) (Figure 2). One of these studies also used a probe specific for the antisense strand of the SARS-CoV-2 S gene and detected its presence in kidney tubules, strongly suggesting there is viral replication in the kidney (13). RT-PCR was used in 13 reports, and 71 of 144 patients (49%) showed presence of SARS-CoV-2 RNA in the kidney (Figure 2). Detection of SARS-CoV-2 RNA by RT-PCR was only observed in postmortem specimens (69,70,81,83–85). One of these studies also tested the kidney of one patient positive for subgenomic viral RNA transcripts by RT-PCR (84). Subgenomic viral RNA is only transcribed by infected cells and not packaged into virions (86). This can be taken as evidence of active replication of the virus in the kidney (84). However, it cannot be determined which cell type in the kidney was infected.

In total, in kidneys from 102 of 235 patients (43%), the presence of SARS-CoV-2 was suggested by at least one of the methods used (Table 2). Each patient was counted once, even if several methods were applied to the same patient. It must be noted that two studies gave the numbers of samples, but not the exact number of patients examined (73,87). These were excluded in the final calculation. The two methods that were more consistent with the presence of SARS-CoV-2 were IF and RT-PCR. IF was positive in 10 of 13 (77%) and RT-PCR in 71 of 144 (49%) samples examined (Figure 2). It is likely these methods are more sensitive and overestimate the virus presence, particularly because in situ hybridization yielded a much lower positivity rate for SARS-CoV-2 (only 11 of 84 samples; 13%). Unfortunately, many studies that performed RT-PCR did not concomitantly assay for in situ hybridization. Clearly, studies that use more than one method, preferably IF and RT-PCR for sensitivity and in situ hybridization to ensure specificity, are more likely to detect SARS-CoV-2 when present in the kidney.

Urine Viral Studies.

In some viral nephropathies, viruria can be easily demonstrated. For instance, high titers of BK virus in the urine are a frequent finding of patients with BK nephropathy (88–90). In a study of six kidney transplant recipients with BK nephropathy, Howell et al. found polyomavirus in the urine of all six patients by EM, in concordance with the presence of virus in biopsies (90). In contrast, the presence of SARS-CoV-2 in the urine has been difficult to document. A meta-analysis of 30 studies comparing SARS-CoV-2 RNA in urine, blood, and stool (91) found the incidence of detecting SARS-CoV-2 RNA in the urine was 8%, a much lower rate compared with presence in blood (21%) and stool (40%) (91). The presence of SARS-CoV-2 RNA in urine was associated with more severe disease in this meta-analysis (91). A more recent report that used an antigen-capture assay detected SARS-CoV-2 S1 protein in the urine of 25% of patients with COVID-19 (92).

SARS-CoV-2 in the urine has been reported to be infectious. Urine from two patients with COVID-19 was sufficient to transfer SARS-CoV-2 infection to ferrets (93,94), fulfilling one of the the main Kochs postulates. In these studies, SARS-CoV-2 RNA levels in nasal washes of the infected ferrets peaked 4 days after infection (94). In a case report, urine from a patient with COVID-19 was used to infect Vero E6 cells in vitro (95). A cytopathic effect was observed 3 days after infection, which was interpreted as presence of infectious/viable SARS-CoV-2 in the urine of this patient (95). Moreover, timing for successful detection of SARS-CoV-2 in urine might be critical. A study that collected urine of 67 patients with COVID-19 tested 13 of 231 (6%) urine samples positive for SARS-CoV-2 RNA (Tan et al., unpublished observations).

The overall relatively few cases with viruria in patients with COVID-19 are consistent with the low number of cases with virus present in kidney samples (see below). This may be because viruria of SARS-CoV-2 is rare, but could be in part due to the lack of an effective and sensitive detection method. Ultracentrifugation to concentrate viral particles and protein mass spectrometry have been proposed to detect the presence of SARS-CoV-2 in the urine (79,96). Ribonucleases in urine can also potentially degrade viral mRNA leading to false-negative results (91,96).

Discussion and Conclusions

It has been difficult to demonstrate the presence of SARS-CoV-2 in the kidneys of patients with COVID-19. Despite multiple negative studies, there are data that demonstrate kidney tropism of SARS-CoV-2 (Figure 2). Even then, there is no evidence that viral presence is directly the cause of AKI frequently seen in patients with COVID-19. When trying to detect viral presence in organs of autopsied patients who had COVID-19, one might not find viral RNA because the search for the virus was done too late after death when kidney autopsy tissue was available. It must be pointed out, therefore, that autopsy studies from patients who died from severe COVID-19 are far from ideal, yet represent the majority of patients with severe and lethal COVID-19 reported. To date, the presence of SARS-CoV-2 in the kidney has been described mainly post mortem. In kidney biopsies usually performed relatively late after the appearance of symptoms, it may be difficult to detect the virus in the kidney sample. Indeed, we could not find studies where kidney biopsies were done very early in the course of COVID-19–associated kidney disease.

Fulminant viremia and viruria are not typical features of COVID-19. Examining the urine early in the course of the disease with advanced techniques such as protein mass spectrometry or more practically with sensitive ELISA assays may provide additional evidence for SARS-CoV-2 kidney infectivity. The timing of the search for the virus may also be crucial. Availability of more information on viruria detected by sensitive methods as noted above could be used sequentially to attempt to assess possible early-stage viral invasion of the kidney in patients with COVID-19. The method of viral detection is an important consideration as well as the site of detection within the kidney. The danger to mistake internal vesicles or other physiologic parts of the cell for viral particles renders EM alone insufficient to search for SARS-CoV-2 presence in the kidneys. Detection of SARS-CoV-2 protein was more successful by IF (77%) than by IHC (19%). It must be noted, however, that the total number of patients tested by IF is very low (Figure 2). The numbers of patients in which SARS-CoV-2 RNA was detected are higher by RT-PCR than by in situ hybridization. However, in most studies, the two methods were not usually performed concurrently in the same patient. There are also important technical aspects. Spatial detection of SARS-CoV-2 by in situ hybridization, IHC, and IF requires a certain degree of tissue preservation, which can be limited due to autolysis. The overall sensitivity of these methods, therefore, can be lower than for RT-PCR. Limited autopsy material and the sampling bias can also lead to false-negative results. Good evidence for kidney tropism of SARS-CoV-2 in patients with COVID-19 should be provided by plaque assay using biopsy material. To our knowledge, there is only one study that successfully isolated SARS-CoV-2 from an autopsied kidney via plaque assay (83).

In summary, although many studies provide support against viral infectivity, there is also reasonable evidence from some comprehensive studies showing that kidney infectivity by SARS-CoV-2 may occur. These positive studies, so far, have been limited to kidney autopsy material. The timing of detection and the method used seem of critical importance for kidney detection. Correlating transcriptional analysis and viral presence in relatively large cohorts would be needed to gain further insight on SARS-CoV-2 infection in the kidney. In particular, studies involving early kidney biopsy tissue or autopsies done very soon after death would clearly improve our understanding of SARS-CoV-2 kidney infectivity. Assays in urine samples using ELISA would be the easiest way to monitor for SARS-CoV-2 in the kidney, and hopefully they will soon become available.

Disclosures

D.C. Batlle reports consultancy agreements with, and receiving honoraria from, AstraZeneca, Relypsa, and Tricida; reports receiving research funding from AstraZeneca, the Feinberg Foundation, and National Institute of Diabetes and Digestive and Kidney Diseases; reports serving as a scientific advisor or member of Relypsa and Tricida; reports being a founder and main owner of Angiotensin Therapeutics Inc., with no royalties or income at this time; reports being coinventor of an issued patent “Active Low Molecular Weight Variants of Angiotensin Converting Enzyme 2,” and provisional patents “Active low molecular weight variants of Angiotensin Converting Enzyme 2 (ACE2) for the treatment of diseases and conditions of the eye” and “Shorter soluble forms of Angiotensin Converting Enzyme 2 (ACE2) for treating and preventing coronavirus infection.” M.A. Sparks reports employment with Duke University and Durham Veterans Affairs Health Care System; reports receiving research funding from Renal Research Institute; and reports receiving honoraria from Elsevier for Nephrology Secrets. M.A. Sparks reports serving as a scientific advisor or member of American Board of Internal Medicine, Nephrology Board, Board of Director, NephJC, and the Editorial Boards of American Journal of Kidney Diseases, ASN Kidney News, Kidney360, and Kidney Medicine; Council for the Kidney in Cardiovascular Disease (KCVD) Membership & Communications Committee American Heart Association (AHA); KCVD Scientific & Clinical Education Lifelong Learning Committee AHA; and National Kidney Foundation (NKF) North Carolina Medical Advisory Board. P. Welling reports receiving research funding from LeDucq Foundation and National Institutes of Health; reports receiving honoraria from the American Physiological Society; reports serving on the Editorial Board of American Journal of Physiology-Renal Physiology, chair of the American Physiological Society Finance Committee, and Chair of Kidney Molecular Biology and Development, National Institutes of Health. All remaining authors have nothing to disclose.

Funding

None.