Kidney Biopsy Findings in Patients with COVID-19

RESULTS

The study group of 17 patients included 12 men, with median age of 54 years (range, 22–72 years) (Tables 1 and 2). Racial demographics included 12 blacks, 3 whites, 1 Asian, and 1 Hispanic. Fourteen patients underwent native kidney biopsies, and three had allograft specimens (including two biopsies and one allograft nephrectomy). Sixteen patients had one or more comorbidities, including 11 with hypertension, 8 with obesity, 3 with diabetes, 3 with history of malignancy (2 prostate and 1 cervix), 4 with solid organ transplants (3 kidney and 1 heart), 1 with SLE, and 1 with untreated hepatitis C infection. Two were former smokers, but none had known preexisting lung disease. In addition to four patients with solid organ transplants, one was receiving immunosuppression for SLE.

Eight patients had mild COVID-19 without pneumonia, including one asymptomatic patient and another with predominantly gastrointestinal symptoms. The other nine had imaging-confirmed COVID-19 pneumonia, including two immunosuppressed patients (with SLE and kidney transplant) who required intubation. Baseline serum creatinine was normal in eight patients. Fifteen patients (88%) presented with AKI, including four with AKI superimposed on CKD. Nine patients (53%) had nephrotic-range proteinuria, including six (35%) with new-onset nephrotic syndrome.

At presentation, the study group had median serum creatinine of 5.7 mg/dl (range, 0.8–20 mg/dl), including seven patients who required dialysis. Median urine protein-creatinine ratio was 7.8 g/g (range, 0.2–21 g/g), and median serum albumin was 2.9 g/dl (range, <1.5–4.5 g/dl). Six patients had microhematuria (more than five red blood cells per high-power field), including two with indwelling urinary catheters. Four patients had peripheral leukocytosis (including two with concurrent lymphopenia), and three had leukopenia. Positive serologies included two patients with antinuclear antibodies (including one with double-stranded DNA antibody), one patient with antiglomerular basement membrane (anti-GBM) antibody, and one with hepatitis C antibody. Inflammatory markers were abnormal in all patients tested, including elevated ferritin (n=13), C-reactive protein (n=10), erythrocyte sedimentation rate (n=9), lactate dehydrogenase (n=8), IL-6 (n=5), and IL-2 receptor (n=1).

Pathologic evaluation revealed five patients with collapsing glomerulopathy accompanied by acute tubular injury (ATI), including one with tubuloreticular inclusions (TRIs) (Figure 1, Table 3). The three patients with collapsing glomerulopathy who consented to genetic studies all had high-risk APOL1 genotypes (two with G1/G1 and one with G1/G2). One patient had minimal change disease accompanied by ATI and endothelial TRI; this patient also had high-risk APOL1 genotype (G1/G1). Among these six podocytopathies, three (50%) did not have imaging-confirmed pneumonia. In four patients, the predominant finding was ATI, including one with pigment casts suggesting myoglobinuria. Four patients had immune-mediated glomerular diseases, including two with membranous glomerulopathy (one with positive tissue staining for phospholipase A2 receptor [PLA2R]), one with crescentic lupus nephritis class 4+5, and one with anti-GBM nephritis. The three kidney transplant recipients had grade 2A acute T cell–mediated rejection, cortical infarction, and ATI, respectively.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Kidney biopsy findings in patients with COVID-19. (A) Light microscopy demonstrates a lesion of collapsing glomerulopathy characterized by hyperplasia of glomerular epithelial cells and collapse of the underlying glomerular capillaries. Jones methanamine silver. Magnification, ×600. (B) Diffuse foot process effacement and endothelial TRIs (arrow and inset) in a patient with minimal change disease. Electron micrograph. Magnification, ×8000. (C) Subepithelial electron dense deposits in PLA2R-associated membranous glomerulopathy. Electron micrograph. Magnification, ×15,000. (D) Multiple glomeruli with circumferential cellular crescents (arrows) in a patient with class 4+5 lupus nephritis. Periodic acid–Schiff. Magnification, ×100. (E) A glomerulus compressed by a crescent with global linear GBM staining for IgG in a patient with anti-GBM nephritis. Immunofluorescence for IgG. Magnification, ×400. (F) Tubular simplification and focal shedding of degenerating epithelial cells into the tubular lumina in a patient with isolated ATI. Hematoxylin and eosin. Magnification, ×400. (G) Severe lymphocytic tubulitis in a patient with acute T cell–mediated rejection. Periodic acid–Schiff. Magnification, ×600. (H) ISH for the virus by automated method showing undetectable viral RNA in the kidney (inset shows positive lung control). Automated ISH with hematoxylin counterstain. Magnification, ×400.

Electron microscopy was performed on 13 biopsies, including 10 with glomeruli available. Ultrastructural examination demonstrated glomerular endothelial TRI in 6 of 10 cases (60%) and absence of definitive virions within renal cells in all 13 cases. Immunohistochemical stains for the spike and nucleocapsid proteins and automated ISH for SARS-CoV-2 RNA, performed in 16 cases, showed no definitive staining. ISH for SARS-CoV-2 RNA performed manually by RNAscope revealed rare, possibly positive tubular cell staining in 2 of 16 patients, both with diagnosis of ATI (Supplemental Figure 1).

Short-term follow-up was available in 16 patients (median, 16 days; range, 1–55 days) (Table 2). One patient died of multiorgan failure 6 days after biopsy. Three patients received no treatment. Nine patients received treatment directed to COVID-19 (including tocilizumab in five, hydroxychloroquine in six, and azithromycin in five). Apart from RRT, seven received specific treatments for their kidney disease (including steroids in five; tacrolimus in one; and plasmapheresis, steroids, and cyclophosphamide in one). At last follow-up, the study group had median serum creatinine of 2.4 mg/dl (range, 0.8–14.2 mg/dl), including ten patients with decrease in serum creatinine and five who remained dialysis dependent. Among five with collapsing glomerulopathy, one with minimal change disease, and two with membranous glomerulopathy, repeat urine protein level was available in six, of which five had reductions in proteinuria. Repeat inflammatory markers decreased in two patients with available data. Repeat PCR testing for SARS-CoV-2 by nasopharyngeal swab, available in eight patients, converted to negative in six.

DISCUSSION

Impaired renal function is a common complication affecting 5%–37% of hospitalized patients with COVID-19.^1–⇓3 SARS-CoV-2 is postulated to cause AKI by diverse mechanisms, including interaction with its cellular receptor angiotensin-converting enzyme 2 (ACE2), viral immune responses, cytokine storm, hypoxemia, reduced oral intake, circulatory collapse, prothrombotic effects, and multiorgan dysfunction.^14–⇓16 Specifically, SARS-CoV-2 may directly infect the kidney via ACE2 widely expressed in proximal tubular cells and podocytes.^5,6,17 Imbalance of the renin-angiotensin-aldosterone system via ACE2 also could exert deleterious hemodynamic effects.¹⁵ Viral-induced cytokine storm causes massive release of granulocyte colony–stimulating factor, various interleukins, and IFN,^18–⇓20 which can injure the kidney directly or indirectly via effects on other organs, such as heart and skeletal muscle. To date, pathologic analyses of the kidney are few and limited to single patient reports or autopsy series.^{5–⇓⇓⇓⇓10} We provide a biopsy-based single-center series exploring kidney pathology in SARS-CoV-2–infected patients.

SARS-CoV-2–infected patients developed diverse glomerular and tubular diseases. The most common glomerular disorder was podocytopathy, including five patients with collapsing glomerulopathy and one patient with minimal change disease; all occurred in black patients (including four with documented APOL1 high-risk genotype) and presented with nephrotic syndrome or nephrotic-range proteinuria and AKI with associated ATI.¹⁰ These findings enlarge the literature on collapsing glomerulopathy^8–⇓10 in the setting of COVID-19 and provide the first example of minimal change disease with APOL1 high-risk genotype. Given the closely related association between IFN therapy and both collapsing glomerulopathy and minimal change disease^21,22 as well as the presence of TRI (so-called IFN footprints), the findings support a role for cytokine-mediated podocyte injury in genetically susceptible individuals with SARS-CoV-2 infection.¹⁹ The lack of demonstrable viral particles in the podocytes by all ancillary studies argues against direct glomerular viral infection.

The inflammatory milieu surrounding COVID-19 also may trigger or exacerbate immune-mediated diseases in predisposed patients. Examples include the crescentic transformation of longstanding preexisting class 2 lupus nephritis and development of acute T cell–mediated rejection in a patient with preformed donor-specific antibodies. IFN and granulocyte colony–stimulating factor play an important role in triggering acute rejection²³ or exacerbating immune complex–mediated GN,²⁴ and both cytokines are known to be elevated in patients with COVID-19.^18–⇓20

Other glomerular diseases included new onset of anti-GBM nephritis in one patient and membranous glomerulopathy of uncertain duration in two. Pulmonary injury from influenza or other insults has been postulated to precede onset of anti-GBM nephritis by exposing the cryptic target Goodpasture antigen,²⁵ consisting of distinct epitopes in COL4A3 and COL4A5, in damaged alveolar capillary basement membranes.²⁶ Conceivably, COVID-19 pneumonia could play a similar priming role. The major target antigen in membranous glomerulopathy, PLA2R, is also expressed in the respiratory tract,²⁷ suggesting a potential source for antigen presentation to incite or potentiate anti-PLA2R autoimmune responses. Coincidental associations with COVID-19 cannot be excluded.

The other major disease category was ATI identified in four native and two allograft kidneys, including one with infarction. Four patients had severe COVID-19 pneumonia (including three with hypoxemia), three patients were on immunosuppression as maintenance for solid organ transplants (two kidney and one heart), four patients had exposure to potentially nephrotoxic or nephromodulatory medications (three tacrolimus and one lisinopril), and one patient had rhabdomyolysis with pigment casts. None had evidence of thrombotic microangiopathy. ATI has also been identified as the predominant finding in autopsy series.⁶ Etiology is likely to be multifactorial with complex interplay of sepsis, hypoxia, hemodynamic instability, nephrotoxin exposure, and multiorgan complications, such as rhabdomyolysis.^6,14,15

In an attempt to detect virus in kidney cells, we used five distinct methodologies, namely immunostains for viral spike and nucleocapsid proteins, ISH for viral RNA (by automated platform and manual RNAScope), and ultrastructural examination, all of which failed to reveal definitive viral particles.^28,29 Whether rare, equivocal staining by manual RNAScope represents true positivity, correlating with low viral abundance,¹⁷ or nonspecific staining requires further study. We cannot rule out the possibility that these techniques lack sufficient sensitivity for definitive viral detection, which may require such methodologies as RT-PCR.^17,30 We doubt that such rare and low abundance of virus is sufficient to account for the pathologic changes and favor predominant roles for cytokine-mediated and other systemic effects.³¹

Our series is limited by its descriptive nature. Coincidental associations with COVID-19 cannot be excluded, and detailed pathogenetic mechanisms will require further investigation. Because of lack of consent, APOL1 genetic testing could not be performed on all black patients with podocytopathy. Clinical and laboratory information provided by referring nephrologists could not be independently verified. Because biopsies were performed for indication and the threshold for kidney biopsy in the setting of COVID-19 varies, our findings may not be generalizable to all patients with COVID-19. Given the recent timing of the pandemic, the follow-up period is necessarily short.

In conclusion, this biopsy series reveals diverse kidney pathology in SARS-CoV-2–infected patients. The findings highlight the potential for viral infection to influence innate or adaptive immune responses that in turn trigger new glomerular diseases (such as podocytopathies and anti-GBM nephritis) or exacerbate preexisting autoimmune or alloimmune conditions (such as lupus nephritis, membranous glomerulopathy, and allograft rejection). ATI is common and likely multifactorial. The lack of definitive virus in kidney cells argues against direct viral infection as the major pathomechanism.

Disclosures

J. Barasch reports that Columbia University has licenses to Abbott and Bioporto related to biomarkers such as NGAL, including EPO 1 616 184; USPO 7,977,110; EPO 1 616 184; and USPO 7,977,110. Bioporto provides royalties to Columbia University. All remaining authors have nothing to disclose.

Funding

None.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2020060802/-/DCSupplemental.

Supplemental Figure 1. In situ hybridization for SARS-CoV-2 by manual RNAScope.