To examine the susceptibility and pathogenicity of pneumococcus coinfection during SARS-CoV-2 infection, K18-hACE2 mice (male and female, 10 to 13 weeks old) were infected with 250 PFU of SARS-CoV-2 or PBS followed by 10³ CFU of pneumococcal strain D39 (coinfected) or PBS (mock coinfected) at either 3, 5, or 7 d pvi. During mock coinfection, the selected viral dose was lethal in 35% of mice (Figure 1A) and caused weight loss from 5 to 11 d pvi with maximum weight loss (average 7%) at 8 d pvi (Figure 1B) and clinical scores peaking at 6 d pvi (Figure 1C). In the absence of viral infection, the selected bacterial dose was lethal in 1/6 mice (17% lethality) at 4 d post bacterial infection (pbi) (Figure S1A) and caused only mild, transient weight loss (~3%) (Figure S1B) and increased temperatures (Figure S1C) after 1 to 2 d pbi.

When the bacterial coinfection was initiated at 3 d pvi, lethality was not enhanced (P = 0.73) (Figure 1A). Interestingly, weight loss in coinfected animals was reduced at 1 d (P = 0.03) and 2 d (P = 0.04) pbi (Figure 1B) and the cumulative clinical score was lower at 2 d pbi (P = 0.03) (Figure 1C) compared with mock coinfected controls. In addition, the temperature of coinfected animals was higher at 2 d (P = 0.003) and 3 d (P = 0.01) pbi and lower at 5 d (P = 0.02) and 8 d (P = 0.045) pbi (Figure 1D). A coinfection initiated at 5 d pvi was slightly more lethal than the SARS-CoV-2 infection alone, where additional mortality was observed at 5 to 6 d pbi, but this was not statistically significant (P = 0.14) (Figure 1A). The average weight loss was reduced (P = 0.01) and temperature was increased (P = 0.001) at 1 d pbi in the coinfected animals (Figures 1B, D). Coinfected animals lost more weight than animals infected with SARS-CoV-2 alone at 5 d pbi (P = 0.03) (Figure 1B), but no significant difference in their clinical scores was detected (Figure 1C). Comparatively, a coinfection at 7 d pvi was significantly more severe than SARS-CoV-2 infection alone (P = 0.03) and resulted in additional lethality at earlier times than the coinfection at 5 d pvi, with additional animals succumbing to the infection within 1, 3, or 4 d pbi (Figure 1A). Significantly more weight loss at 3 d (P < 0.001) and 4 d (P = 0.002) pbi (Figure 1B) and higher clinical scores at 3 d pbi (P = 0.01) (Figure 1C) occurred without altering temperature (Figure 1D).

To evaluate whether SARS-CoV-2-bacterial coinfection alters pathogen burden, we measured viral loads in the lung and bacterial loads in the lung and blood of infected animals. In mice infected with bacteria alone or with SARS-CoV-2 followed by bacteria at 3 d pvi, no bacteria were recovered from the lungs of 7/8 mice at 24 h pbi (Figures 2A, S1D). However, when the bacteria was introduced at 5 d pvi, bacterial loads in the lung remained at a level similar to the inoculum in 7/8 mice and was cleared in 1/8 mice (Figure 2A). Bacteria were not detected in the blood of mice infected with bacteria alone (data not shown) or SARS-CoV-2-bacteria coinfected at 3 or 5 d pvi (Figure 2B). However, in mice coinfected at 7 d pvi, significant bacterial growth occurred in the lungs of all animals (P = 0.02; Mann-Whitney test) and the blood of some animals (3/7) with titers reaching 4.4 to 7.9 log₁₀ CFU/lung (Figure 2A) and 4.1 to 6.6 log₁₀ CFU/mL (Figure 2B), respectively, within 24 h pbi.

Pulmonary viral loads were unchanged by bacterial coinfection whether coinfection was initiated at 3 d (P = 0.12) or 5 d (P = 0.18) pvi (Figure 2C) and the amount and distribution of viral antigen in the lung tissue were also unchanged (Figures 2D, E). Although some areas of the lung contained colocalized virus and bacteria, both intracellular and extracellular bacterial antigen were detected in areas containing no viral antigen (Figure 2F). The virus had cleared by 8 d pvi in the groups that were mock coinfected or bacterial coinfected at 7 d pvi (Figure 2C). No significant differences were found in viral or bacterial loads between males and females.

To investigate whether bacterial coinfection altered immune response dynamics, several immune cells, cytokines, and chemokines were quantified in the lung 24 h after mock coinfection or bacterial coinfection in SARS-CoV-2 infected mice (Figures 3, 4, S3, S6). In animals infected with SARS-CoV-2 only, natural killer (NK) T cells (Figure S3D) and total CD19⁺ B cells (Figure 3E) were reduced at 4 d pvi compared with naïve (P = 0.007 and P = 0.018, respectively). The absolute numbers of other cells were unchanged at this time point (Figures 3, S3); however, increases in the proportion of activated (CD69⁺) immune cells were evident (Figure S4). SARS-CoV-2 infection also resulted in many cytokines and chemokines above baseline levels (all P < 0.05) throughout the infection, including IFN-γ, IL-1β, IL-4, IL-28, CXCL10, GM-CSF, LIF, CCL2, CCL7, MIP-1α, MIP-1β, RANTES, IFN-α, and IFN-β. IL-5, IL-6, IL-15, IL-18, M-CSF, and TNF-α were elevated at both 4 d and 6 d pvi while CXCL5, CXCL1, G-CSF, IL-3, IL-13, and IL-17A were increased only at 6 d pvi. MIP-2α, IL-2, and IL-22 were elevated at 6 d and 10 d pvi, and increased IL-10 and IL-23 were detected only at 8 d pvi (absolute values of cytokines are in Figure 4, S5; log₂ changes over naïve in Figure S6).

As expected, a significant influx of CD45⁺ immune cells was evident at 6 and 8 d pvi in animals infected with SARS-CoV-2 only (both P < 0.001) (Figure S3A), including neutrophils (Ly6G^hi; both P < 0.01; Figure 3A), the F4/80^midCD11c^midCD11b⁺ monocyte/macrophage subset (both P < 0.001; Figure 3B), inflammatory macrophages (F4/80^hiCD11c^hiCD11b⁺, iMΦ; P = 0.02 and P < 0.001, respectively; Figure 3C), F4/80^midCD11c^- cells (both P < 0.001; Figure S3B), NK cells (both P < 0.001; Figure S3C), CD4⁺ T cells (P = 0.02 and P < 0.001, respectively; Figure 3F), and CD8⁺ T cells (both P < 0.001; Figure 3G). Unlike the pathogen loads, some of the immune cells were different between males and female that were mock coinfected at 5 d pvi, including neutrophils (P = 0.047), resident alveolar macrophages (F4/80^hiCD11c^hiCD11b^-MHC-II^low/-, AMΦ; P = 0.047), CD4⁺ T cells (P = 0.02), NK cells (P = 0.03), and NK T cells (P = 0.02), which were higher in females than males.

In the groups coinfected with bacteria at 3 d pvi, no changes were observed in the absolute number (Figures 3, S3) or activation (Figure S4) of any quantified immune cell subset or the amount of cytokines and cytokines (Figures 4, S5) within 24 h pbi compared with mock coinfection. A bacterial coinfection at 5 d pvi resulted in fewer total CD45+ cells (P = 0.03; Figure S3A), including neutrophils (Figure 3A), CD19⁺ B cells (Figure 3E), CD8⁺ T cells (Figure 3G), and F4/80^midCD11c^- cells (Figure S3B) (all P < 0.05) compared with the mock coinfected groups. In addition, iMΦ (P = 0.01) and AMΦ (P = 0.047) were again higher in females than males following coinfection at 5 d pvi (Figures 3C, D). The extent of activation was not different between the mock coinfection and bacterial coinfection at 5 d pvi (Figure S4), but reduced IL-6, IL-18, LIF (all P = 0.04), and IL-15 (P = 0.02) was observed at 24 h pbi (Figures 4A-D).

Coinfection at 7 d pvi induced a significant increase in neutrophils at 24 h pbi (P < 0.001) (Figure 3A) without altering the number or activation of any other immune cell quantified (Figures 3, S3, S4). AMΦ were reduced in the mock coinfected group compared with naïve animals (P = 0.001) but were not different between the mock coinfection and bacterial coinfection (P = 0.29) (Figure 3D). Absolute cell numbers and activation did not differ between male and female mice following coinfection at 7 d pvi (Figures 3, S3, S4). Perhaps unexpectedly, none of the measured cytokines were significantly different between animals that were mock coinfected and animals that were bacterial coinfected at 7 d pvi (Figure 4 and Figure S5).

To investigate whether bacterial coinfection altered immune cell dynamics and activation in recovered animals, pulmonary immune cells, cytokines, and chemokines were quantified at 17 d pvi following mock coinfection or bacterial coinfection at 3, 5, or 7 d pvi. The number of iMΦ (P = 0.01) (Figure 3C) and CD8⁺ T cells (P = 0.02) (Figure 3G), as well as the activated proportion of iMΦ (P = 0.004), CD8⁺ T cells (P = 0.001), CD4⁺ T cells P 0.001), and CD19+ B cells (P = 0.005) (Figure S4), remained increased above naïve levels in the lungs of animals that recovered from SARS-CoV-2 infection alone. These changes were accompanied by elevated IFN-γ, CXCL10, and RANTES (P = 0.01, P = 0.03, and P = 0.04, respectively) at 17 d pvi compared to naïve (Figures 4, S5, S6). However, many measured cytokines and chemokines were below naive levels at 17 d pvi in the lungs of animals infected with SARS-CoV-2 only, including eotaxin, IL-2, IL-3, IL-17A, IL-22, IL-27, IL-28, M-CSF, and MIP-2α (all P < 0.05) (Figures 4, S5, S6).

A sustained increase in immune cell accumulation and activation was evident in animals that recovered from SARS-CoV-2-pneumococcal coinfection. At 17 d pvi, an increased absolute number and activated proportion of F4/80^midCD11c^midCD11b⁺ monocytes/macrophages (P = 0.01; Figures 3B, 4B), iMΦ (P = 0.01; Figures 3C, S4C), and CD4⁺ and CD8⁺ T cells (P = 0.03 and 0.02, respectively; Figures 3F, G, S4F, G) were present in coinfected mice compared with mock coinfected mice. Comparison between the coinfected groups indicated that more CD8⁺ T cells were present at 17 d pvi in mice that were coinfected at 3 d or 7 d pvi than those coinfected at 5 d pvi (both P = 0.02; Figure 3G). In addition, animals that recovered from a coinfection at 7 d pvi had more activated neutrophils or iMΦ than those who recovered from a coinfection at 3 d pvi (P = 0.04) or 5 d pvi (P = 0.03), respectively (Figures S4A, C). These changes were accompanied by higher levels of CXCL-10 (P < 0.001), MIP-2α (P = 0.04), IL-3 (P = 0.001), IL-22 (P < 0.008), IL-28 (P = 0.01), and RANTES (P < 0.001) in the lungs of mice that had recovered from a bacterial coinfection compared with those recovered from SARS-CoV-2 alone (17 d pvi; Figure 4E–J). In addition, select cytokines and chemokines were reduced in animals that recovered from bacterial coinfection compared with those that were mock coinfected, including CXCL-1 (P = 0.01), IL-1α (P = 0.04), IL-6 (P = 0.03), IL-9 (P = 0.03), IL-10 (P < 0.001), IL-13 (P < 0.001), IL-15 (P = 0.001), IL-18 (P < 0.001), G-CSF (P = 0.03), and TNF-α (P =0.02) (17 d pvi; Figures 4, S5). These cytokines, except for IL-1α (P = 0.19) and IL-18 (P = 0.09), were also below baseline levels (all P < 0.05). In addition, IL-2 (P = 0.02), IL-5 (P = 0.02), IL-17A (P = 0.04), and eotaxin (P = 0.01) were below baseline in both the bacterial coinfected and mock coinfected groups (Figures 4, S5).

To examine whether lung pathology was enhanced during SARS-CoV-2-pneumococcal coinfection, we assessed seven pathological features (endothelial hypertrophy/margination, peribronchiolar/perivascular lymphoid cells, interstitial inflammation/septal thickening, alveolar inflammation, alveolar edema/hemorrhage, the extent of alveolar involvement, and consolidation (Figure 5). There were no significant differences in any of these measurements between mock coinfected animals and those coinfected with bacteria at 3 or 5 d pvi at either 24 h pbi or 17 d pvi.

Adult (10-13 week old) male and female K18-hACE2 transgenic mice (B6.Cg-Tg(K18-ACE2)2Prlmn/J) were obtained from Jackson Laboratories (Bar Harbor, ME). Mice were housed in groups of 4 in solid–bottom polysulfone individually ventilated cages (Allentown BCU) in rooms maintained on a 12:12-hour light:dark cycle at 22 ± 2°C with 30-70% humidity in the Regional Biocontainment Laboratory (animal biosafety level 3 facility) at UTHSC (Memphis, TN). Mice were acclimated for 1 day before being lightly anesthetized with 2% inhaled isoflurane (Baxter, Deerfield, IL) and implanted subcutaneously with an IPTT300 transponder (Bio Medic Data Systems, Seaford, DE) for identification and temperature monitoring, followed by an additional 3 days of acclimation before inclusion in the experiments. Envigo irradiated rodent diet (catalog no. 7912) and autoclaved water were available ad libitum during the acclimation and study periods; gel food and hydrogel were provided at the time of infection. All experimental procedures were performed under protocol 20-0132 approved by the Institutional Animal Care and Use Committee at University of Tennessee Health Science Center (UTHSC) under relevant institutional and American Veterinary Medical Association (AVMA) guidelines and were performed in a animal biosafety level 3 facility that is accredited by the American Association for Laboratory Animal Science (AALAS).

All experiments were done using 2019-nCoV/USA-WA1/2020 (BEI Resources NR-52281) (SARS-CoV-2) and type 2 pneumococcal strain D39. The viral infectious dose [plaque forming units (PFU)] was determined by plaque assay of serial dilutions on Vero E6 cells. Virus seed stocks were sequenced using next-generation sequencing with ARTIC primers on the Illumina MiSeq. Bacterial infectious dose [colony forming units (CFU)] was determined by using serial dilutions on tryptic soy agar plates supplemented with 3% sheep erythrocytes (TSA). Doses of virus and bacteria were selected that elicited mild-moderate disease independently to ensure that changes in disease severity following coinfection would be evident. Frozen stocks were diluted in sterile PBS and administered intranasally to groups of 4 mice, lightly anesthetized with 2.5% inhaled isoflurane (Baxter, Deerfield, IL) in a total volume of 50 µl (25 µl per nostril). Mice were inoculated with either PBS or SARS-CoV-2 at day 0 then with 10³ CFU of D39 or PBS, either 3 or 5 days later. Assessment of symptom severity was performed twice daily after the onset of symptoms by assigning a score (scale 0-3) to clinical features, including weight loss (0, <15%; 1, 15-20%; 2, 21-25%; 3, >25%), temperature change (0, >34°C; 1, 34-31°C; 2, 30-26°C; 3, <26°C), body condition/appearance (0, normal; 1, roughened fur; 2, roughened fur, hunched posture, mild grimace, active; 3, roughened fur, hunched posture, grimace, inactive, conjunctivitis, head-tilt), respiratory effort (0, normal; 2, increased respiratory rate and effort; 3, weak, intermittent breathing), behavior (0, normal; 1, slow, unprovoked movement; 2, slow, provoked movement; 3, minimal response/unresponsive or spinning), and dehydration (0, normal; 1, ≤ 2 second skin tent, mildly sunken eyes; 2, 2-3 second skin tent, sunken eyes; 3, > 3 second skin tent, sunken eyes). Mice were euthanized if they lost >25% of their starting body weight or became moribund based on clinical scores (a score of 3 in any single category or a cumulative score of ≥9 in respiratory effort, dehydration, temperature reduction, behavior, body condition/appearance).

Mice were euthanized by 33% isoflurane inhalation. Lungs were aseptically harvested, washed in PBS, and fixed in 10% neutral buffered formalin for histology or digested with collagenase (1 mg/ml, Sigma C0130) and physical homogenization against a 40 µm cell strainer for immune cell staining. Lung digest supernatants were used to quantify cytokines and chemokines and to determine viral and bacterial titers as above; bacterial titers were also measured in peripheral blood. Following red blood cell lysis, lung cells were washed in staining buffer (PBS, 5mM EDTA, 10mM HEPES, and 0.5% bovine serum albumin), counted with trypan blue exclusion using a Cell Countess System (Invitrogen, Grand Island, NY), and prepared for flow cytometric analysis as described below.

Flow cytometry (BD FACSAria; San Jose, CA) was performed on single cell suspensions after Fc receptor blocking (TruStainFcX, Biolegend) and viability staining (Zombie Violet Fixable Viability, Biolegend), 25 min surface staining, and fixation (BD Cytofix). The followed anti-mouse antibody panels were used for cell subset analysis: CD45 (clone 30-F11, Pe-Cy7, Biolegend), CD3e (clone 145-2C11, FITC, Biolegend), CD4 (clone RM4-5, V500, BD Biosciences), CD8α (clone 53-6.7, PerCP-Cy5.5, Biolegend), CD19 (clone 6D5, PE, Biolegend), CD335 (clone 29A1.4, APC-Fire750, Biolegend), and CD69 (clone H1.2F3, APC, Biolegend) or CD45 (clone 30-F11, Pe-Cy7, Biolegend), Ly6G (clone 1A8, PerCP-Cy5.5, Biolegend), F4/80 (clone BM8, PE, eBioscience), CD11b (clone M1/70, V500, BD Biosciences), CD11c (clone N418, APC-Fire750, Biolegend), MHC-II (clone I-A/I-E, FITC, eBioscience), and CD69 (clone H1.2F3, APC, Biolegend). The data were analyzed using FlowJo 10.7.2 (Tree Star, Ashland, OR). Data were cleaned using the flowAI application (159) followed by gating viable cells from a forward scatter/side scatter plot, singlet inclusion, and viability dye exclusion. CD45⁺ cells were selected for further analyses. Neutrophils (Ly6G^hi), alveolar macrophages (AMΦ) (F4/80^hiCD11c^hiCD11b^-MHC-II^low/-), inflammatory/exudate macrophages (iMΦ) (F4/80^hiCD11c^hiCD11b⁺MHC-II^mid/hi), other monocyte/macrophage populations (F4/80^midCD11c^midCD11b⁺ and F4/80^midCD11c^-CD11b^+/-), NK cells (CD3e^-CD19^-CD335⁺), CD4 T cells (CD3⁺CD8^-CD4⁺CD335^-), CD8 T cells (CD3⁺CD8⁺CD4^-CD335^-), NK T cells (CD3e⁺CD335⁺), B cells (CD3e^-CD19⁺), and recently activated subsets thereof (CD69⁺) were gated as in Figure S2.

Cytokines G-CSF (CSF-3), GM-CSF, IFN-γ, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12p70, IL-13, IL-15/IL-15R, IL-17A (CTLA-8), IL-18, IL-22, IL-23, IL-27, IL-28, IL-31, LIF, MCP-3 (CCL7), M-CSF, TNF-α) and chemokines (ENA-78 (CXCL5), eotaxin (CCL11), GROα (CXCL1), IP-10 (CXCL10), MCP-1 (CCL2), MIP-1α (CCL3), MIP-1β (CCL4), MIP-2α (CXCL2), RANTES (CCL5) were measured in lung supernatant by Luminex and ELISA (IFN-α,β). Before use, cell debris and aggregates were removed by centrifugation at 4°C, 400 x g. ProcartaPlex magnetic bead cytokine/chemokine plates (Invitrogen) were prepared according to the manufacturer’s instructions. Data were acquired using a MagPix (Luminex) with Luminex xPonent software (v4.2) and analyzed with the ProcartaPlex Analysis App (ThermoFisher Connect). ELISAs for IFNα and IFNβ (PBL Assay Science) were prepared according to the manufacturer’s instructions, read at 450 nm, and analyzed using GraphPad Prism 9.2.0. Mean concentrations of duplicate samples were calculated by the construction of standard curves using a weighted 5PL and 4PL regression for the ProcartaPlex and ELISA data, respectively. Absolute quantities of each cytokine/chemokine were calculated based on the mean concentration of replicate samples normalized to the lung supernatant volume collected during tissue processing. Internal plate controls were used to adjust values obtained between plates and fold changes in cytokine and chemokine quantities were calculated for each animal, normalized to the average of naïve controls (pooled males/females).

Following euthanasia and tissue removal as above, lungs were continually fixed in 10% neutral-buffered formalin solution (NBF; ThermoFisher Scientific, Waltham, MA) before being embedded in paraffin, sectioned at 4μm, and mounted on positively charged glass slides (Superfrost Plus; Thermo Fisher Scientific, Waltham, MA). Tissue sections were stained with hematoxylin and eosin (H&E) or subjected to immunohistochemical (IHC) staining to detect SARS-CoV-2 antigen or pneumococcus. Tissue sections were deparaffinized and rehydrated before undergoing antigen retrieval in a citrate-based solution (pH 6.0) at 97°C for SARS-CoV-2 detection or a tris-based solution (pH 9.0) for pneumococcal detection (Vector Laboratories, Burlingame, CA). For IHC, a primary monoclonal antibody against SARS-CoV-2 nucleoprotein (NP) (Sino Biological, Wayne, PA) or a rabbit polyclonal antibody against pneumococcus (Novus Biologicals, Littleton, CO) was used at 1:1000 followed by a biotinylated anti-rabbit antibody (Vector Laboratories, Burlingame, CA) at 1:200, the Vectastain Elite ABC-HRP kit (Vector Laboratories, Burlingame, CA), and 3,3’-Diaminobenzidine (DAB) solution development. Stained sections were counterstained with hematoxylin, dehydrated, and examined by a pathologist blinded to the experimental group assignments. Pathology was scored on a scale from 0-5, where 0 = normal, no tissue affected; 1 = minimal: rare or inconspicuous lesions; 2 = mild: multifocal or small, focal, or widely separated, but conspicuous lesions; 3 = moderate: multifocal, prominent lesions; 4 = marked: extensive to coalescing lesions or areas of inflammation with some loss of structure; 5 = severe: diffuse lesion with effacement of normal structure. Intermediate severity grades were assigned where necessary. To quantify the extent of viral infection in the lungs, digital images of whole lung sections stained for viral antigen were first captured using the Aperio ScanScope XT Slide Scanner (Aperio Technologies, Inc., Vista, CA). The areas of both the entire lung parenchyma (alveoli and bronchioles) and the virus-positive regions were outlined manually with areas determined using ImageScope software (Aperio Technologies, Inc.). Representative images and quantitative analyses of viral spread and lung pathology during infection are shown in Figures 2, 5, respectively.

Significant differences in Kaplan-Meier survival curves were calculated using the log-rank test. Linear values of lung and blood bacterial loads, viral loads, immune cells, and cytokines/chemokines were compared using an unpaired t test with Welch correction except where the Mann-Whitney test was used due to unequal variances (GraphPad Prism 9.2.0 and Rv4.0.3). The confidence interval of significance was set to 95%, and P ≤ 0.05 was considered significant.