SARS-CoV-2 infection and persistence throughout the human body and brain


 COVID-19 is known to cause multi-organ dysfunction1-3 in acute infection, with prolonged symptoms experienced by some patients, termed Post-Acute Sequelae of SARS-CoV-2 (PASC)4-5. However, the burden of infection outside the respiratory tract and time to viral clearance is not well characterized, particularly in the brain3,6-14. We performed complete autopsies on 44 patients with COVID-19 to map and quantify SARS-CoV-2 distribution, replication, and cell-type specificity across the human body, including brain, from acute infection through over seven months following symptom onset. We show that SARS-CoV-2 is widely distributed, even among patients who died with asymptomatic to mild COVID-19, and that virus replication is present in multiple extrapulmonary tissues early in infection. Further, we detected SARS-CoV-2 RNA in multiple anatomic sites, including regions throughout the brain, for up to 230 days following symptom onset. Despite extensive distribution of SARS-CoV-2 in the body, we observed a paucity of inflammation or direct viral cytopathology outside of the lungs. Our data prove that SARS-CoV-2 causes systemic infection and can persist in the body for months.

Long COVID-with cardiovascular, pulmonary, and neurological manifestations with or without 88 functional impairment [4][5] . While autopsy studies of fatal COVID-19 cases support the ability of 89 SARS-CoV-2 to infect multiple organs 3,7-12 , extra-pulmonary organs often lack histopathological evidence of direct virally-mediated injury or inflammation [10][11][12][13][14] . The paradox of extra-pulmonary 91 infection without injury or inflammation raises many pathogen-and host-related questions. To inform these pathogen-focused questions and to evaluate for the presence or absence 99 of associated histopathology in matched tissue specimens, we performed extensive autopsies on 100 a diverse population of 44 individuals who died from or with COVID-19 up to 230 days 101 following initial symptom onset. Our approach focused on timely, systematic, and 102 comprehensive tissue sampling and preservation of adjacent tissue samples for complementary 103 analyses. We performed droplet digital polymerase chain reaction (ddPCR) for sensitive 104 detection and quantification of SARS-CoV-2 gene targets in all tissue samples collected. To 105 elucidate SARS-CoV-2 cell-type specificity and validate ddPCR findings, we performed in situ 106 hybridization (ISH) broadly across sampled tissues. Immunohistochemistry (IHC) was used to 107 further validate cell-type specificity in the brain where controversy remains on the regional 108 distribution and cellular tropism of SARS-CoV-2 infection. In all samples where SARS-CoV-2 109 RNA was detected by ddPCR, we performed qRT-PCR to detect subgenomic (sg)RNA, an assay 110 suggestive of recent virus replication 15 . We confirmed the presence of replication-competent 111 SARS-CoV-2 in extrapulmonary tissues by virus isolation in cell culture. Lastly, in six 112 individuals, we measured the diversity and anatomic distribution of intra-individual SARS-CoV-113 2 variants using high-throughput, single-genome amplification and sequencing (HT-SGS). 114 We categorized autopsy cases of SARS-CoV-2 infection as "early" (n=17), "mid" 115 (n=13), or "late" (n=14) by illness day (D) at the time of death, being ≤D14, D15-D30, or ≥D31, 116 respectively. We defined persistence as presence of SARS-CoV-2 RNA among late cases. Due to 117 the extensive tissue collection, we analyzed and described the results in terms of grouped tissue 118 categories as the following: respiratory tract; cardiovascular; lymphoid; gastrointestinal; renal 119 and endocrine; reproductive; muscle, skin, adipose, & peripheral nerves; and brain. perimortem plasma was unavailable for the other three cases (P3, P4, P15). Extensive sampling 126 of the brain was accomplished in 11 of the 44 cases (Fig. 1). The cohort was 29.5% female with 127 a mean age of 59.2 years and was diverse across race and ethnicity (Extended Data   Table 3).

150
With a few exceptions, the overall burden of SARS-CoV-2 RNA decreased by a log or 151 more across tissue categories among mid cases, and further decreased among late cases.

152
However, several mid and late cases had high levels (≥5 N gene copies/ng RNA input) detected 153 among multiple tissues (Extended Data Fig. 2). Further, persistence of low-level SARS-CoV-2 154 RNA (0.0004 to <0.5 N gene copies/ng RNA input) was frequently detected across multiple 155 tissue categories among all late cases, despite being undetectable in plasma (Extended Data Fig.   156 2, Supplementary Data 1). Notably, SARS-CoV-2 RNA was detected in the brains of all six late cases and across most locations evaluated in the brain in five of these six, including P42 who 158 died at D230 (Fig. 1   We isolated SARS-CoV-2 in cell culture from multiple pulmonary and extrapulmonary 171 tissues, including lung, bronchus, sinus turbinate, heart, mediastinal lymph node, small intestine, 172 and adrenal gland from early cases up to D7 (P19, P27, P32, P37; Supplementary Data 1). 175 We used HT-SGS to analyze SARS-CoV-2 spike gene variant sequences from a total of 176 46 tissues in six individuals. In five individuals from the early group, predominant spike 177 sequences were largely identical across tissues. In P27, P19, and P18, no non-synonymous virus 178 genetic diversity was detected in pulmonary and extrapulmonary sites despite a high depth of 179 single-molecule sampling (Extended Data Fig. 3). Thus, virus populations that were relatively 180 homogeneous had disseminated in these individuals without coding changes in spike. However, 181 we also noted important patterns of intra-individual virus diversity in several patients from the 182 early group. In P27, although all 4,525 inferred spike amino acid sequences were identical, two 183 virus haplotypes, each with a single synonymous substitution, were preferentially detected in 184 extrapulmonary sites including right and left ventricles and mediastinal LN. In P38, we observed 185 clear virus genetic differences between the lung lobes and the brain, with a D80F residue found 186 in 31/31 pulmonary but 0/490 brain sequences and a G1219V residue that was restricted to brain 187 minor variants. A similar distinction was observed between sequences from dura mater and other 188 sites in P36, albeit at very low sampling depth (n = 2 sequences) from dura mater. Overall, these 189 findings suggested no need for alterations in receptor utilization to permit extrapulmonary 190 dissemination of SARS-CoV-2, while also revealing genetic compartmentalization between 191 viruses in the lung lobes and those in extrapulmonary sites, including the brain.

219
Myocytes within skeletal muscle contained spike RNA in both early (P18) and late (P20) 220 cases. In addition to the organ-specific cell type infection of SARS-CoV-2, endothelium, 221 muscularis of atrial vessels, and Schwann cells were identified as infected throughout the body, 222 and were similarly positive across early and late cases.

223
Spike RNA was found in neurons, glia and ependyma, as well as endothelium of vessels 224 across all lobes of the brain of early, mid, and late cases. Within the cerebellum specifically,  Outside the lungs, histological changes were mainly related to complications of therapy 243 or preexisting co-morbidities: mainly obesity, diabetes, and hypertension. Five cases had old 244 ischemic myocardial scars and three had coronary artery bypass grafts in place. Given the 245 prevalence of diabetes and obesity in our cohort, it was not surprising to find diabetic 246 nephropathy (10 cases, 23%) or steatohepatitis (5 cases, 12%). One case was known to have 247 chronic hepatitis C with cirrhosis, but the other cases of advanced hepatic fibrosis were likely 248 related to fatty liver disease, even if diagnostic features of steatohepatitis were not present.

249
Hepatic necrosis (13 cases, 30%) and changes consistent with acute kidney injury (17 cases, 250 39%) were likely related to hypoxic-ischemic injury in these very ill patients.

251
In the examination of the 11 brains, we found few histopathologic changes, despite the 252 evidence of substantial viral burden. Vascular congestion was an unusual finding that had an 253 unclear etiology and could be related to the hemodynamic changes incurred with infection.

254
Global hypoxic/ischemic change was seen in two cases, one of which was a juvenile (P36) with a 255 seizure disorder who was found to be SARS-CoV-2 positive on hospital admission, but who 256 likely died of seizure complications unrelated to viral infection. Here we provide the most comprehensive analysis to date of SARS-CoV-2 cellular 260 tropism, quantification, and persistence across the body and brain, in a diverse autopsy cohort 261 collected throughout the first year of the pandemic in the United States. Our focus on short post-262 mortem intervals, comprehensive approach to tissue collection, and preservation techniques -

263
RNAlater and flash freezing of fresh tissueallowed us to detect and quantify viral levels with 264 high sensitivity by ddPCR and ISH, as well as culture virus, which are notable differences 265 compared to other studies. 266 We show SARS-CoV-2 disseminates across the human body and brain early in infection 267 at high levels, and provide evidence of virus replication at multiple extrapulmonary sites during 268 the first week following symptom onset. We detected sgRNA in at least one tissue in over half of is due to either residual blood within the tissue 8,17 or cross-contamination from the lungs during 272 tissue procurement 8 , our data rule out both theories. Only 12 cases had detectable SARS-CoV-2 273 RNA in a perimortem plasma sample, and of these only two early cases also had SARS-CoV-2 274 sgRNA in the plasma, which occurred at Ct levels higher than nearly all of their tissues with 275 sgRNA. Therefore, residual blood contamination cannot account for RNA levels within tissues. node and ocular tissue from two early cases (P19, P32).

295
Our use of a single-copy sequencing approach for the SARS-CoV-2 spike allowed us to 296 demonstrate homogeneous virus populations in many tissues, while also revealing informative 297 virus variants in others. Low intra-individual diversity of SARS-CoV-2 sequences has been 298 observed frequently in previous studies [18][19][20] , and likely relates to the intrinsic mutation rate of the 299 virus as well as lack of early immune pressure to drive virus evolution in new infections. It is 300 important to note that our HT-SGS approach has both a high accuracy and a high sensitivity for 301 minor variants within each sample, making findings of low virus diversity highly reliable 21 . The 302 virus genetic compartmentalization that we observed between pulmonary and extrapulmonary 303 sites in several individuals supports independent replication of the virus at these sites, rather than 304 spillover from one site to another. Importantly, lack of compartmentalization between these sites 305 in other individuals does not rule out independent virus replication, as independently replicating 306 populations may share identical sequences if overall diversity is very low. It was also interesting 307 to note several cases where brain-derived virus spike sequences showed non-synonymous 308 differences relative to sequences from other tissues. These differences may indicate differential 309 selective pressure on spike by antiviral antibodies in brain versus other sites, though further 310 studies will be needed to confirm this speculation.  Finally, a major contribution of our work is a greater understanding of the duration and 329 locations at which SARS-CoV-2 can persist. While the respiratory tract was the most common 330 location in which SARS-CoV-2 RNA tends to linger, ≥50% of late cases also had persistence in 331 the myocardium, thoracic cavity lymph nodes, tongue, peripheral nerves, ocular tissue, and in all 332 sampled areas of the brain, except the dura mater. Interestingly, despite having much lower 333 levels of SARS-CoV-2 in early cases compared to respiratory tissues, we found similar levels 334 between pulmonary and the extrapulmonary tissue categories in late cases. This less efficient 335 viral clearance in extrapulmonary tissues is perhaps related to a less robust innate and adaptive 336 immune response outside the respiratory tract. 337 We detected sgRNA in tissue of over 60% of the cohort. While less definitive than viral       Autopsies were performed and tissues were collected as previously described 26

537
Upstream tissue processing and subsequent RNA quantification have been previously 538 described 26 . The QX200 AutoDG Droplet Digital PCR System (Bio-Rad) was used to detect and 539 quantify SARS-CoV-2 RNA in technical replicates of 5.5 uL RNA for fluids and up to 550 ng 540 RNA for tissues as previously described 26 . Results were then normalized to copies of N1, N2, 541 and RP per mL of sample input for fluids and per ng of RNA concentration input for tissues. For 542 samples to be considered positive for SARS-CoV-2 N1 or N2 genes, they needed to mean the 543 manufacturer's limit of detection of ≥0.1 copies/µL and ≥2 positive droplets per well. Over 60 544 control autopsy tissues from uninfected patients, representing all organs collected for COVID-19 545 autopsy cases, were used to validate the manufacturer's EUA published LOD for nasopharyngeal 546 swabs for tissues (Extended Data Table 8). ddPCR data for P3 16 as well as a portion of tissues 547 from the oral cavity 26 have been previously reported.