CD169-mediated restrictive SARS-CoV-2 infection of macrophages induces pro-inflammatory responses

SARS-CoV-2 Spike protein can mediate ACE2-independent entry into macrophages

To examine the role of macrophages in SARS-CoV-2 infection and COVID-19 pathogenesis, we differentiated primary MDMs from multiple donors by culturing CD14⁺ monocytes in the presence of human AB-serum and M-CSF for 6 days (55). Compared to a control HEK293T/ACE2 cell line retrovirally transduced to stably express human ACE2, ACE2 expression in primary human MDMs was under the detection limit (Fig. 1A and B). However, similar to HEK293T/ACE2 cells (Fig. 1C), MDMs from multiple donors were robustly infected with a SARS-CoV-2 S-pseudotyped lentivirus (Fig. 1D), suggesting that macrophages can support ACE2-independent S-pseudotyped virus entry. Furthermore, SARS-CoV-2 S-pseudotyped infections in MDMs were blocked by pre-treatment with a cathepsin inhibitor (E64D) but not a TMPRSS2 inhibitor (Camostat) (Fig. 1D), suggesting that SARS-CoV-2 S facilitates endosomal viral entry into ACE2-deficient MDMs. Preferential engagement of endosomal entry mechanism for SARS-CoV-2 S pseudotyped lentiviruses (LVs) in MDMs correlated with lack of active and cleaved form of serine protease TMPRSS2 expression in THP1/PMA and primary macrophages ((56), Fig. S1A), and robust cathepsin-L expression in THP1/PMA and MDMs (Fig. S1B). Strikingly, pre-treatment with neutralizing antibodies targeting the N-terminal domain (NTD) of SARS-CoV-2 S that do not compete with ACE2 binding by S (57), led to significant reduction in S-pseudotyped lentivirus infection of primary MDMs, suggesting that specific interaction between SARS-CoV-2 S and ACE2-independent entry factors are essential to mediate entry and endosomal fusion in macrophages (Fig. S2).

CD169 is a SARS-CoV-2 attachment and entry factor in macrophages

Expression of CD169, an ISG, is significantly upregulated in monocytes and alveolar macrophages isolated from COVID-19 patients, and its expression enhancement correlates with COVID-19 disease severity (7, 53). While co-expression of CD169 with ACE2 can enhance ACE2-mediated SARS-CoV-2 S entry in HEK293T cells (16), whether CD169 plays a role in SARS-CoV-2 infection of ACE2-deficient macrophages is unclear. To address this question, THP1 cells stably expressing wildtype (wt) CD169, mutant CD169/R116A which displays an attenuated ability to bind sialylated glycoconjugates (39, 58, 59), ACE2, or both wt CD169 and ACE2 (CD169/ACE2) (Fig. S3A) were incubated with recombinant SARS-CoV-2 S protein, and relative S binding was determined by flow cytometry. In contrast to parental THP1 monocytes, THP1 cells expressing wt CD169 (THP1/CD169) displayed robust S binding comparable to levels observed with THP1/ACE2 cells (Fig. 2A). There was a significant reduction in S binding to THP1/CD169-R116A cells, indicating that the sialylated S protein is recognized by CD169 (Fig. 2A). Co-expression of CD169 and ACE2 in THP1 cells further enhanced S binding compared to cells expressing only ACE2 or CD169, suggestive of cooperative binding of CD169 and ACE2 to SARS-CoV-2 S.

We next sought to determine the role of CD169 in mediating SARS-CoV-2 infection of macrophages. PMA-differentiated THP1 macrophages expressing wt or mutant CD169, ACE2, or CD169/ACE2 were infected with SARS-CoV-2 S-pseudotyped lentiviruses. Expression of wt CD169 on THP1/PMA macrophages enhanced S-pseudotyped lentiviral infection bŷ30-fold compared to parental THP1/PMA macrophages, similar to the levels of infection observed with ACE2⁺ THP1/PMA macrophages (Fig. 2B). In contrast expression of CD169/R116A did not enhance S-pseudotyped lentiviral infection of THP1/PMA cells, confirming that recognition of sialylated motifs on SARS-CoV-2 S is essential for CD169-mediated infection of macrophages (Fig. 2B). Co-expression of wt CD169 and ACE2 further enhanced S-pseudotyped lentiviral infection by greater than 11-fold and 3-fold when compared to cells expressing CD169 or ACE2, respectively (Fig. 2B), confirming that CD169 facilitates S-mediated entry into macrophages in the absence of ACE2 and enhances entry in the presence of ACE2. To confirm the role of CD169 in SARS-CoV-2 S-mediated infection in primary human MDMs, and given the relatively low and variable expression of endogenous CD169 in primary MDMs (Fig. S3B), we used lentiviral transduction to overexpress either wt CD169 or ACE2 (Fig. S3C and D). Following infection of transduced primary macrophages with S-pseudotyped lentivirus, we observed that CD169 or ACE2 overexpression in MDMs from multiple donors significantly enhanced SARS-CoV-2 S-pseudotyped lentivirus infection compared to control MDMs transduced with empty vector (Fig. 2C). Crucially, pre-treatment with anti-CD169 blocking mAb (7D2) prior to infection with S-pseudotyped lentivirus significantly attenuated infection of untransduced primary MDMs when compared to non-specific IgG1 pretreatment (Fig. 2D). These findings suggest that CD169 facilitates SARS-CoV-2 S-dependent fusion and entry into both THP1/PMA macrophages and primary MDMs in the absence of ACE2.

SARS-CoV-2 establishes abortive infection in macrophages lacking ACE2

To investigate whether CD169 expression is sufficient to establish productive SARS-CoV-2 infection and replication in ACE2-deficient macrophages, we infected THP1/PMA and primary MDMs overexpressing CD169, ACE2, or parental cells (lacking both CD169 and ACE2) with replication-competent SARS-CoV-2 (Washington isolate, NR-52281), as previously described (60). To evaluate productive infection, we examined the temporal expression of double-stranded RNA (dsRNA), a viral replication intermediate (61), as well as viral nucleocapsid (N) protein. SARS-CoV-2-infected THP1/PMA macrophages were fixed at various time points post infection and subjected to immunofluorescence analysis using antibodies against dsRNA and SARS-CoV-2 N. In contrast to parental THP1/PMA macrophages that showed background staining of dsRNA and no expression of SARS-CoV-2 N at any time point post infection (Fig. 3A), CD169-expressing THP1/PMA cells showed low levels of dsRNA production and small puncta staining of SARS-CoV-2 N that did not significantly increase over the course of infection (Fig. 3B). However, both ACE2⁺ (Fig. 3C) and CD169/ACE2 double-positive (Fig. 3D) THP1/PMA macrophages showed robust dsRNA and SARS-CoV-2 N production starting as early as 2–4 hours post infection (hpi), which significantly increased over the course of infection. There was a clear distinction in the spatial distribution of viral N protein in infected ACE2⁺ and CD169⁺/ACE2⁺ THP1/PMA macrophages over time, as noted by a transition from 6 hpi onward from small cytosolic N protein puncta to homogenous distribution throughout the cytosol (Fig. 3C and D). Furthermore, there was extensive co-localization between perinuclear dsRNA foci and N staining in ACE2⁺ and CD169⁺/ACE2⁺ THP1/PMA macrophages at 6 and 24 hpi. In accordance with our S-pseudotyped lentivirus infection data (Fig. 2B), co-expression of CD169 and ACE2 led to an increase in SARS-CoV-2 infection compared to ACE2-expressing cells, with higher levels and slightly earlier production of dsRNA and SARS-CoV-2 N (compare Fig. 3C and D). In contrast, N protein staining was not dispersed in the cytoplasm of CD169⁺ THP1/PMA macrophages and remained as small cytosolic puncta over the course of infection with minimal co-localization with dsRNA staining (Fig. 3B). These findings suggest that CD169-mediated SARS-CoV-2 entry in THP1/PMA macrophages results in initiation of viral transcription and moderate viral protein production.

Similar to our findings in THP1/PMA macrophages, we observed SARS-CoV-2 N expression in primary MDMs overexpressing ACE2, whereas we were not able to detect N protein in primary untransduced MDMs or those overexpressing CD169 at 24 hpi (Fig. S4). The lack of detection of either N (Fig. S4) or dsRNA (data not shown) in primary MDMs compared to the low level of N and dsRNA expression observed in THP1/CD169 macrophages (Fig. 3B) could be explained by the inefficient lentivector transduction efficiency and low expression of CD169 in primary MDMs (Fig. S3B and C) compared to the robust CD169 expression in retrovirally transduced THP1 cells (Fig. S3A). We further quantified infectious SARS-CoV-2 particle production from both THP1/PMA macrophages (Fig. 3E) and primary MDMs (Fig. 3F) by TCID50 assay and detected no infectious virions in culture supernatants from CD169-expressing cells. However, both THP1/PMA and primary MDMs expressing ACE2 showed robust production of infectious SARS-CoV-2 particles, with a marked increase in infectious virus production in THP1/PMA cells co-expressing CD169 and ACE2 compared to ACE2 alone (Fig. 3E). These results suggest an absence of post-entry restrictions to SARS-CoV-2 replication in macrophages, if virus entry is mediated by ACE2.

CD169-mediated SARS-CoV-2 infection of macrophages results in de novo expression of viral mRNAs

To explore the spatial and temporal distribution of SARS-CoV-2 RNA at single cell level, single-molecule RNA FISH (smFISH) analysis (62) was performed in infected THP1/PMA macrophages. Individual fluorescent spots corresponding to viral gRNAs were detected in CD169⁺, ACE2⁺ and CD169⁺/ACE2⁺ THP1/PMA macrophages, as early as 1 hpi (Fig. S5). By 6 hpi, we observed diffuse cytosolic staining and perinuclearly localized bright gRNA foci (Fig. S5), which further increased in in ACE2⁺ and CD169⁺/ACE2⁺ but not CD169⁺ THP1/PMA macrophages (Fig. S5). In order to distinguish between gRNAs and N gene transcripts, we probed cells at 24 hpi with two probe sets labeled with distinguishable dyes one against SARS-CoV-2 ORF1a and the other against N gene. The former detects only gRNA and the later reports both gRNA and N sub-genomic transcripts (Fig. 4A). The staining, particularly that of N gene, showed diffused cytoplasmic distribution, only in ACE2⁺ and CD169⁺/ACE2⁺ THP1/PMA macrophages, especially in the advanced stages of infection.

Dispersion of N transcripts (gRNA and sgRNA) in the cytosol of ACE2⁺ and CD169⁺/ACE2⁺ THP1/PMA macrophages mirrors the temporal localization phenotype of viral N protein (Fig. 3C and D), and might be indicative of formation of N positive viral replication compartments (63). While a majority of virus-exposed CD169⁺ THP1/MA macrophages expressed fluorescent puncta (gRNA) at 24 hpi (indicated by the white arrowheads showing both ORF1a and N expression, Fig. 4A and B), transition to bright gRNA foci or cytosolic expansion was not observed, suggesting that CD169 expression in THP1/PMA cells increased uptake of virus, but without ACE2, failed to establish viral replication foci. It should be noted that the formation of distinct fluorescent puncta in CD169⁺ THP1/PMA macrophages at 24 hpi is suggestive of de novo RNA synthesis and not virus inoculum, since these puncta were not detected at 1 hpi (Fig. S5). While viral RNAs in the CD169-expressing cells are localized in few distinct granulated puncta, ACE2-expressing cells exhibited inclusion-like structures, suggesting that differential engagement of viral entry receptors (such as CD169 and ACE2) leads to altered fate of the incoming viral genome and subsequent steps in the viral replication cycle.

To further investigate the step at which SARS-CoV-2 replication is restricted in macrophages, THP1/PMA macrophages or those expressing either CD169, ACE2, or both CD169 and ACE2 were infected with SARS-CoV-2 and cells lysed and harvested at 2, 4, 6, and 24 hpi to quantify the level of total SARS-CoV-2 N transcripts, +gRNA and sgRNA, as well as negative sense antigenomic RNA, by RT-qPCR. We detected increasing levels of total SARS-CoV-2 N transcripts at early time points (2–6 hpi) in THP1/PMA macrophages expressing CD169, ACE2, or both CD169 and ACE2, whereas parental THP1/PMA macrophages showed no significant increase in viral N RNA levels compared to mock infection controls (Fig. 4C). Furthermore, there were no significant differences in gRNA levels at early times post infection (up to 6 hpi) between CD169⁺, ACE2⁺ or CD169⁺/ACE2⁺ THP1/PMA macrophages suggesting an absence of cell-intrinsic restrictions to early steps of SARS-CoV-2 replication, such as attachment and fusion, in CD169+ THP1/PMA macrophages. Interestingly, SARS-CoV-2 N transcripts peaked at 6 hpi in CD169+ THP1/PMA cells, followed by a progressive decline at 24 hpi (Fig. 4C). In contrast, there was a temporal increase in SARS-CoV-2 N transcripts in cells expressing ACE2, with markedly higher expression at 24 hpi compared to CD169-expressing cells. In concordance with SARS-CoV-2 N protein expression (Fig. 3D), THP1/PMA cells expressing both CD169 and ACE2 expressed the highest levels of SARS-CoV-2 N transcripts at 24 hpi compared to those expressing ACE2 or CD169 alone.

After virus entry and capsid uncoating, the next steps in the SARS-CoV-2 replication cycle are translation of the viral gRNA followed by the formation of viral replication-transcription complexes, which enables synthesis of the negative sense antigenomic RNA (64). Negative sense antigenomic RNAs are present at significantly lower levels than +gRNA and sgRNA, and are templates for synthesis of additional positive sense gRNA and sgRNAs (64). We employed strand-specific RT-qPCR analysis to detect negative sense viral RNA (ORF1b) and confirmed viral replication (at 2 hpi) in CD169⁺ THP1/PMA cells, compared to complete absence of negative sense viral RNAs in parental THP1/PMA macrophages (Fig. 4D). Interestingly, negative sense RNA expression in CD169⁺ THP1/PMA macrophages plateaued at 6 hpi, suggesting that CD169-mediated SARS-CoV-2 entry promotes initial steps of viral replication. In contrast, there was a progressive increase in negative sense RNA levels in ACE2⁺ THP1/PMA macrophages particularly at 24 hpi, and a further enhancement was observed upon co-expression of CD169 and ACE2 in THP1/PMA macrophages at later times post infection (Fig. 4D).

To confirm that temporal increases in viral transcripts in CD169⁺ or ACE2⁺ THP1/PMA macrophages were generated from ongoing virus transcription, THP1/PMA cells (±CD169 ±ACE2) were pre-treated with remdesivir (RDV), a known inhibitor of SARS-CoV-2 RNA synthesis, prior to infection with SARS-CoV-2 (24, 63). RT-qPCR analysis of both gRNA (N transcripts) (Fig. 4E) and sgRNA (E transcripts) (Fig. 4F) transcripts harvested at 24 hpi revealed that RDV pre-treatment completely blocked the increase in viral RNA expression in THP1/PMA macrophages expressing CD169, ACE2, or both CD169 and ACE2, to levels observed in parental THP1/PMA macrophages. Furthermore, levels of viral gRNA (Fig. 4G) and sgRNA (Fig. 4H) in SARS-CoV-2 infected untransduced MDMs progressively increased over time, and these increases in viral gRNA and sgRNAs were also suppressed upon RDV pre-treatment. Notably, primary MDMs harbored higher levels of viral transcripts compared to parental THP1/PMA macrophages at all times post infection (Fig. 4E–H), suggesting that endogenous CD169 expression was sufficient to facilitate entry and initiation of virus replication in primary MDMs. The marked lower levels of viral transcripts in primary MDMs (Fig. 4G and H) compared to CD169⁺ THP1/PMA macrophages (Fig. 4E and F) might be correlated to the lower CD169 expression in primary MDMs compared to CD169⁺ THP1/PMA macrophages cells which were selected for high CD169 expression (Fig. S3A). Taken together, these results suggest that CD169-mediated viral entry enables initiation of viral RNA replication and transcription in ACE2-deficient CD169⁺ macrophages even in the absence of establishment of robust viral RNA replication foci in macrophages, and that SARS-CoV-2 infection of macrophages is restricted at a post-entry step of virus replication.

Low level expression of SARS-CoV-2 genomic and sub-genomic RNAs is sufficient to induce pro-inflammatory cytokines in non-productively infected macrophages

Since CD169 expression in macrophages was sufficient to permit SARS-CoV-2 entry and initiate restricted viral RNA expression, we investigated whether de novo viral RNA synthesis in the absence of productive virus replication can trigger innate immune responses. SARS-CoV-2 infected THP1/PMA cells (±CD169 ±ACE2) were lysed at 24 hpi, and total RNA was analyzed by RT-qPCR for inflammatory cytokine expression (Fig. 5). We observed significant induction of expression of IL-6, TNFα, IL-1β, and IL-18 mRNA in non-productively infected CD169⁺ THP1/PMA macrophages compared to the parental THP1/PMA cells (Fig. 5A). IL-6 and TNFα mRNA expression in SARS-CoV-2 infected THP1/CD169 cells was similar to the levels observed in productively infected THP1/ACE2, but lower than that observed in THP1/CD169/ACE2 cells. Intriguingly, induction of pro-inflammatory cytokines, IL-1β and IL-18, was only consistently observed in non-productively infected THP1/CD169 macrophages but not in the productively infected THP1/ACE2 or THP1/CD169/ACE2 macrophages (Fig. 5A). A delayed or impaired type I IFN response is thought to be a critical mechanism for COVID-19 pathogenesis, though impaired induction of type I IFN response has mostly been reported in infected lung epithelial cells (65). Interestingly, we observed muted induction of IFNβ, IP-10 and Viperin in SARS-CoV-2 infected CD169⁺ THP1/PMA macrophages. In contrast, IFNβ, IFNλ1, IP-10 and Viperin mRNA expression was dramatically upregulated upon establishment of productive SARS-CoV-2 infection in THP1/ACE2 and THP1/CD169/ACE2 macrophages (Fig. 5A) and correlated with the extent of viral RNA expression and virus replication in these cells (Fig. 3).

Since previous reports have suggested that macrophage exposure to recombinant S protein may trigger inflammatory cytokine production (17, 66), we infected parental, CD169⁺, ACE2⁺, or CD169⁺/ACE2⁺ THP1/PMA cells with S-pseudotyped lentivirus but observed no induction of pro-inflammatory cytokines (IL-6, TNFα, IL-1β, and IFNλ1) (Fig. S6). To confirm that the induction of pro-inflammatory cytokine and ISG expression was due to viral replication and sensing of viral RNA, we infected THP1/PMA macrophages with SARS-CoV-2 in the presence of RDV. Pre-treatment with RDV which abolished virus replication (Fig. 4E and F) completely abrogated IL-6, TNFα, IL-1β and IFNλ1 induction in CD169⁺, ACE2⁺, and CD169⁺/ACE2⁺ THP1/PMA macrophages (Fig. 5B), suggesting that de novo viral RNA transcription and sensing of virus replication intermediates are required for the induction of pro-inflammatory cytokines and ISGs in macrophages. Temporal analysis of inflammatory cytokine induction revealed that while SARS-CoV-2 infection of parental THP1/PMA cells did not result in induction of pro-inflammatory responses (Fig. 5C), infection of CD169⁺ THP1/PMA cells led to rapid induction of IL-6, TNFα, IL-1β, and IFNλ1 mRNA expression (Fig. 5D). In fact, fold-induction in levels of pro-inflammatory cytokines at early times post infection (4–6 hpi) was similar between CD169⁺ and ACE2⁺ macrophages (Fig. 5D and E), indicating that early viral RNA production is the key trigger of innate immune activation. Accelerated kinetics and highest magnitude of induced IFNλ1 expression was observed in SARS-CoV-2 infected CD169⁺/ACE2⁺ THP1/PMA macrophages (Fig. 5F, note difference in scale on y axis), which correlated with greater level of negative sense antigenomic RNA (Fig. 4D) and highest levels of virus replication in CD169⁺/ACE2⁺ THP1/PMA macrophages (Fig. 3E).

To further confirm the findings obtained from CD169⁺ THP1/PMA macrophages, we pre-treated primary MDMs from multiple donors with RDV for 30 minutes prior to infection with SARS-CoV-2. Total RNA was harvested and subsequently analyzed by RT-qPCR to determine fold-induction of IL-6, TNFα, IL-1β, and IFNλ1 mRNA expression at indicated timepoints (Fig. 5G). Similar to the observations in CD169⁺ THP1/PMA macrophages (Fig. 5B), SARS-CoV-2 infection induced IL-6, TNFα, IL-1β, and IFNλ1 mRNA in primary MDMs (Fig. 5G), with proinflammatory cytokine induction observed as early as 2 hpi. Furthermore, induction of pro-inflammatory responses in primary MDMs was significantly suppressed by RDV pre-treatment (Fig. 5G), confirming that establishment of infection and initiation of viral transcription is required for macrophage-intrinsic innate immune activation. Collectively, these findings suggest that CD169-mediated restricted SARS-CoV-2 infection of macrophages induces robust inflammatory responses.

Cytosolic RNA sensing by RIG-I and MDA-5 is required for SARS-CoV-2 induced inflammation in macrophages

Since innate immune activation in THP1/CD169 cells and primary MDMs requires the expression of de novo viral RNAs (Fig. 5B and G), we next sought to delineate the nucleic acid sensing mechanism required for the detection of SARS-CoV-2 gRNA and sgRNAs in CD169⁺ THP1/PMA macrophages. Depending on the specific pathogen-derived cytosolic nucleic acids, numerous host sensors can detect and trigger innate immune activation via the MAVS and/or STING pathways (67). Viral RNAs can be sensed by RIG-I-like receptors (RLRs) or endosomal toll-like receptors (TLRs) to activate MAVS or TRIF, respectively, leading to induction of pro-inflammatory cytokines (68). Innate immune sensing of SARS-CoV-2 RNAs by RLRs such as RIG-I and MDA-5, or endosomal TLRs (TLR7/8), has been previously proposed (69, 70). It has also been hypothesized that SARS-CoV-2-induced mitochondrial damage and release of mitochondrial DNA into the cytosol, a cellular stress response, could trigger cGAS/STING sensing pathway in infected cells (71, 72). To investigate which of the nucleic acid sensing pathways are involved in sensing abortive SARS-CoV-2 infection in macrophages, we stably knocked-down expression of either RIG-I, MDA-5, UNC93B1 (a protein required for TLR3/7/9 trafficking to endosomes, (73)), MAVS or STING in CD169⁺ THP1/PMA cells through shRNA-based lentiviral transduction. Upon successful selection of transduced cells targeted for knockdown of individual host proteins, we observed robust decrease in both mRNA (Fig. 6A) and protein (Fig. 6B) expression. Knock-down of RIG-I, MDA-5, UNC93B1, MAVS or STING in THP1/CD169 macrophages did not impact infection efficiency of SARS-CoV-2, as shown by RT-qPCR analysis of viral gRNA and sgRNAs in each cell line (Fig. 6C and D). While knockdown of UNC93B1 in virus-infected THP1/PMA CD169⁺ cells had negligible impact on induction of pro-inflammatory cytokines (IL-6, TNFα, IL-1β and IFNλ1) compared to scramble control cells, depletion of either RIG-I or MDA-5 led to dramatic reduction in pro-inflammatory cytokine expression to near background levels, suggesting both cytosolic viral RNA sensors, but not endosomal TLRs, are required for innate immune sensing of SARS-CoV-2 transcripts (Fig. 6E–H). Furthermore, knock-down of MAVS but not STING, completely abrogated SARS-CoV-2-induced innate immune activation in THP1/PMA macrophages, further confirming the requirement of cytosolic viral RNA sensing for induction of pro-inflammatory cytokines.

MAVS is essential for SARS-CoV-2 RNA-induced innate immune activation in both non-productively and productively infected THP1/PMA macrophages

We next sought to determine if MAVS was required for induction for pro-inflammatory responses not only in CD169-mediated abortive infection but also in ACE2-mediated productive infection of macrophages with SARS-CoV-2. Parental THP1 cells or those expressing CD169, ACE2, or both CD169 and ACE2 were transduced with lentivectors expressing scramble shRNAs (control) or MAVS-specific shRNAs, which led to robust knock-down of MAVS protein expression in all cell lines (Fig. 7A). While MAVS depletion did not affect subsequent SARS-CoV-2 infection of CD169⁺, ACE2⁺ or CD169⁺/ACE2⁺ THP1/PMA cells, as quantified by total SARS-CoV-2 N gRNA transcripts (Fig. 7B), induction of pro-inflammatory cytokines (IL-6, TNFα, IL-1β and IFNλ1) was significantly attenuated in both productively (ACE2⁺, CD169⁺/ACE2⁺) and abortively (CD169⁺) infected THP1/PMA macrophages upon MAVS knockdown (Fig. 7C). These results suggest that MAVS plays a pivotal role in pro-inflammatory cytokine induction in both ACE2-dependent and independent (CD169-mediated) SARS-CoV-2 infection of THP1/PMA macrophages. Taken together, these findings suggest that CD169-dependent establishment of abortive SARS-CoV-2 infection in macrophages triggers RIG-I/MDA-5 mediated sensing of viral gRNA and sgRNAs and MAVS-dependent inflammatory cytokines induction, which might contribute to the dysregulated hyper-immune phenotype of inflammatory macrophages and severity of COVID-19 disease.

Ethics statement.

This research has been determined to be exempt by the Institutional Review Board of the Boston University Medical Center since it does not meet the definition of human subjects research, since all human samples were collected in an anonymous fashion and no identifiable private information was collected.

Plasmids

The SARS-CoV-2 S/gp41 expression plasmid was a gift from Dr. Nir Hachoen at the Broad Institute, and has previously been described (74). For ACE2 lentiviral transduction in THP1 cells, we used a lentiviral plasmid expressing ACE2 and the puromycin resistant gene (Addgene #145839, hereafter pLenti-ACE2-IRES-puro). Generation of wildtype and mutant (R116A) human CD169 plasmids (LNC-CD169) was previously described and validated (39). For transduction of primary MDMs, a 3’ LTR-restored lentiviral expression vector (Addgene #101337, hereafter LV-3’LTR) expressing a GFP reporter was used to express ACE2 or CD169. For ACE2 cloning, the NotI-XhoI fragment from pLenti-ACE2-IRES-puro was inserted into the LV-3’LTR backbone. For cloning CD169 into LV-3’LTR vector, a BglII-AgeI fragment from LNC-CD169 was inserted into LV-3’LTR vector. HIV-1 packaging plasmid psPAX2 and VSV-G expression constructs have been previously described (39). All lentiviral vectors (pLKO.1) expressing shRNAs used for knockdown of host proteins were purchased from Sigma.

Cells

HEK293T cells (ATCC) were maintained in DMEM (Gibco) containing 10% heat-inactivated fetal bovine serum (FBS) (Gibco) and 1% pen/strep (Gibco) (39, 40, 97). Vero E6 cells (ATCC CRL-1586) were maintained in DMEM supplemented with 10% FBS and 100 μg/mL primocin. THP1 cells (ATCC) were maintained in RPMI/1640 (Gibco) containing 10% FBS and 1% pen/strep (50). THP1 cells stably expressing CD169 have previously been described (39). To generate HEK293T/ACE2⁺, THP1/ACE2⁺ and THP1/CD169⁺/ACE2⁺ cells, HEK293T, THP1 or THP1/CD169 cells were transduced with pLenti-ACE2-IRES-puro lentivector and cultured in puromycin-containing media (2 μg/ml). Cells with robust surface expression of ACE2 and CD169/ACE2 (double-positive) were sorted using a MoFlo cell sorter (Beckman Coulter) and cultured in puromycin-supplemented media. All cell lines are routinely tested for mycoplasma contamination and confirmed negative. For THP1 monocyte to macrophage differentiation, THP1 cells were stimulated with 100 nM PMA (Sigma-Aldrich) for 48 hours. Human monocyte-derived macrophages (MDMs) were derived from CD14⁺ peripheral blood monocytes by culturing cells in RPMI/1640 (Gibco) media containing 10% heat-inactivated human AB serum (Sigma-Aldrich) and recombinant human M-CSF (20 ng per ml; Peprotech) for 5–6 days, (55). To generate MDMs expressing ACE2 or overexpressing CD169, cells were co-infected with ACE2 or CD169 expressing lentiviruses (100 ng based on p24^gag ELISA per 1×10⁶ cells) and SIV3+ (Vpx expressing) VLPs. ACE2 and CD169 surface expression was determined by flow cytometry 3 days post transduction.

Viruses

SARS-CoV-2 stocks (isolate USA_WA1/2020, kindly provided by CDC’s Principal Investigator Natalie Thornburg and the World Reference Center for Emerging Viruses and Arboviruses (WRCEVA)) were grown in Vero E6 cells (ATCC CRL-1586) cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 2% fetal FBS and 100 μg/mL primocin. To remove confounding cytokines and other factors, viral stocks were purified by ultracentrifugation through a 20% sucrose cushion at 80,000×g for 2 hours at 4°C (60). SARS-CoV-2 titer was determined in Vero E6 cells by tissue culture infectious dose 50 (TCID₅₀) assay using the Spearman Kärber algorithm. All work with SARS-CoV-2 was performed in the biosafety level 4 (BSL4) facility of the National Emerging Infectious Diseases Laboratories at Boston University, Boston, MA following approved SOPs. Generation of SARS-CoV-2 S-pseudotyped lentiviruses expressing spike glycoprotein has previously been described (74). Briefly, HEK293T cells were co-transfected with HIV-1 reporter plasmid containing a luciferase reporter gene in place of the nef ORF and SARS-CoV-2 S (74). To generate ACE2 or CD169 expressing recombinant lentiviruses, LV-3’LTR lentivectors expressing either empty vector, ACE2, or CD169, were co-transfected with psPax2 (HIV Gag-pol packaging plasmid) and H-CMV-G (VSV-G envelope-expressing plasmid) in HEK293T cells by calcium phosphate-mediated transfection (98). Virus-containing supernatants were harvested 2 days post-transfection, cleared of cell debris by centrifugation (300×g, 5 min), passed through 0.45 μm filters, aliquoted and stored at −80 °C until use. Lentivirus titers were determined by a p24^gag ELISA (98).

Infection

For RNA analysis, 1×10⁶ cells (THPI/PMA, MDMs, HEK293T) were seeded in 12-well plates. For smFISH analysis, 1×10⁶ cells were seeded in 6-well plates containing 3–4 coverslips per well. The next day, cells were infected with purified SARS-CoV-2 at the indicated multiplicity of infection (MOI). At indicated time points, the cells were either lysed with TRIzol (for total RNA analysis) or fixed in 10% neutral buffered formalin for at least 6 hours at 4°C and removed from the BSL-4 laboratory for staining and imaging analysis in accordance with approved SOPs. For SARS-CoV-2 S-pseudotyped lentiviral infections of THP1/PMA macrophages or MDMs, 1×10⁵ cells were seeded in 96-well plates, and infected via spinoculation the following day with 10–20 ng (p24^Gag) of purified S-pseudotyped lentivirus and SIV_mac Vpx VLPs (1 hr at RT and 1100 × g), as previously described (55). Incubation with virus was continued for 4 additional hours at 37°C, cells were then washed to remove unbound virus particles, and cultured for 2–3 days. For CD169 blocking experiments, primary MDMs from 3 different donors were pre-incubated with 20 μg/ml anti-CD169 antibody (HSn 7D2, Novus Biologicals) or IgG1k (P3.6.2.8.1, eBioscience) for 30 min at 4°C prior to infection. For anti-spike neutralizing experiments, virus-containing media were pre-incubated with antibodies targeting SARS-CoV-2 spike NTD for 30 min at 37°C prior to infection. At indicated timepoints, cells are lysed, cell lysates were analyzed for luciferase activity using the Bright-Glo luciferase assay kit (Promega), as previously described (36). The SARS-CoV-2 neutralizing antibodies was previously characterized (57) and were a kind gift from Dr. Duane Wesemann at Harvard Medical School.

S binding

To evaluate SARS-CoV-2 S binding to various THP1 monocytes expressing different surface receptors, approximately 0.25×10⁶ cells from parental THP1 or those expressing wt CD169, mutant CD169 (R116A), ACE2, or both wt CD169 and ACE2 were incubated for 30 min at 4 °C with 2 μg of spike glycoprotein (stabilized) from Wuhan-Hu-1 SARS-CoV-2 containing a C-terminal Histidine Tag, recombinant from HEK293F cells (BEI resources, #NR-52397). This is followed by secondary staining for 30 min at 4°C with APC-conjugated mouse anti-His antibody (BioLegend, #362605, 1:50) or isotype control. Cells were fixed with 4% PFA (Boston Bioproducts) for 30 min, and analyzed with BD LSRII (BD). Data analysis was performed using FlowJo software (FlowJo).

Immunofluorescence

In brief, the cells were permeabilized with acetone-methanol solution (1:1) for 10 min at − 20°C, incubated in 0.1 M glycine for 10 min at room temperature and subsequently incubated in 5% goat serum (Jackson ImmunoResearch) for 20 minutes at room temperature. After each step, the cells were washed three times in PBS. The cells were incubated overnight at 4°C with a rabbit antibody directed against the SARS-CoV nucleocapsid protein (Rockland; 1:1000 dilution in 5% goat serum), which cross-reacts with the SARS-CoV-2 nucleocapsid protein, as previously described (99). The cells were washed four times in PBS and incubated with goat anti-rabbit antibody conjugated with AlexaFluor594 for 1 hour at room temperature (Invitrogen; 1:200 dilution in blocking reagent). 4’,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich) was used at 200 ng/ml for nuclei staining. For dsRNA staining (61), anti-dsRNA (Pan-Enterovirus Reagent, clone 9D5, Light Diagnostics, Millipore) antibody was used 1:2 overnight and anti-mouse-AF488 (Invitrogen) 1:200 dilution as secondary antibody with DAPI. Images were acquired using a Nikon Eclipse Ti2 microscope with Photometrics Prime BSI camera and NIS Elements AR software.

RNA isolation and RT-qPCR

Total RNA was isolated from infected cells using TRIzol reagent (Invitrogen). Reverse transcription (RT) from purified RNAs was performed using oligo(dT)₂₀ primer (Superscript III, Invitrogen) or strand-specific RT primers as previously described (100). Target mRNAs were quantified using Maxima SYBR Green (Thermo Scientific), using the primer sets shown in Table 1. Primer sequences for GAPDH, IL-6, TNFα, IL-1β, IP-10, IFNλ1, IL-18, Viperin and IFNβ have been described previously (97). The C_T value was normalized to that of GAPDH and represented as a relative value to a ‘mock’ control using the 2^−ΔΔC_T method as described (97, 101).

shRNA mediated knockdown

Stable knockdown of host proteins in THP1 cells was carried out by transduction with lentivectors expressing individual shRNAs (pLKO.1, 400 ng p24^Gag (as measured by ELISA) per 1 × 10⁶ cells) in the presence of polybrene (Millipore). Cells were washed and cultured for 5–7 days in the presence of puromycin (2 μg/ml, InvivoGen). Selected cells were expanded, and knockdown confirmed and quantified by RT-qPCR or western blotting, prior to any downstream experiments. All shRNA target sequences are listed in Table 2.

Flow cytometry

To examine cell surface expression of CD169 or ACE2 in transduced THP1 or primary MDMs, approximately 0.5×10⁶ cells were harvested with CellStripper (Corning), stained with Zombie-NIR (BioLegend, #423105, 1:250) followed by staining for 30 min at 4°C with the following antibodies; Alexa647-conjugated mouse anti-CD169 antibody (BioLegend, #346006, 1:50), Alexa647-conjugated mouse anti-ACE2 antibody (R&D systems, 1:200), or unconjugated goat anti-ACE2 polyclonal antibody (R&D systems, #AF933, 1:200) followed by Alexa488-conjugated chicken anti-goat antibody (Invitrogen, #A-21467, 1:100). Cells were fixed with 4% PFA (Boston Bioproducts) for 30 min, and analyzed with BD LSRII (BD). Data analysis was performed using FlowJo software (FlowJo).

smFISH

Immunoblot Analysis

To assess expression of endogenous or transduced proteins, cell lysates containing 30–40 μg total protein were separated by SDS-PAGE, transferred to nitrocellulose membranes and the membranes were probed with the following antibodies: mouse anti-TMPRSS2 (Santa Cruz, #515727, 1:1000), mouse anti-Cathepsin-L (Santa Cruz, #32320, 1:1000), goat anti-ACE-2 (R&D systems, #AF933, 1;1000), rabbit anti-STING (Cell Signaling, #13647, 1:1000), rabbit anti-MAVS (Thermo Fisher, #PA5–17256, 1:1000), mouse anti-RIG-I (AdipoGen, #20B-0009, 1:1000), rabbit anti-MDA-5 (Proteintech, #21775–1-AP, 1:1000), rabbit anti-UNC93B1 (Invitrogen, #PA5–83437, 1:1000), rabbit anti-IRF1 (Cell Signaling, #8478S, 1:1000). Specific staining was visualized with secondary antibodies, goat anti-mouse-IgG-DyLight 680 (Thermo Scientific, #35518, 1:20000), goat anti-rabbit-IgG-DyLight 800 (Thermo Scientific, #SA5–35571, 1:20000), or a donkey anti-goat-IgG-IR-Dye 800 (Licor, #926–32214, 1:20000). As loading controls, actin or tubulin expression was probed using a rabbit anti-actin (Sigma-Aldrich, A2066, 1:5000), mouse anti-actin (Invitrogen, #AM4302, 1:5000), or rabbit anti-tubulin (Cell Signaling, #3873, 1;5000). Membranes were scanned with an Odessy scanner (Li-Cor).

Statistics

All the statistical analysis was performed using GraphPad Prism 9. P-values were calculated using one-way ANOVA followed by the Tukey-Kramer post-test (symbols for p-values shown with a line) or the Dunnett’s post-test (comparing to mock), symbols for p-values shown with a bracket), or a two-tailed paired t-test (comparing two samples, symbols for two-tailed p-values shown with a line bracket). Symbols represent, *: p < 0.05, **: p < 0.01, ***: p < 0.001, ****: p < 0.0001. No symbol or ns: not significant (p ≥ 0.05).