Pulmonary Expression of Interleukin-17 Contributes to Neutrophil Infiltration into the Lungs during Pneumonic Plague

ABSTRACT Inhalation of respiratory droplets infected with Yersinia pestis results in a rapidly progressing and lethal necrotic pneumonia called primary pneumonic plague. Disease manifests as biphasic, with an initial preinflammatory phase with rapid bacterial replication in the lungs absent readily detectable host immune responses. This is followed by the onset of a proinflammatory phase that sees the dramatic upregulation of proinflammatory cytokines and extensive neutrophil accumulation in the lungs. The plasminogen activator protease (Pla) is an essential virulence factor that is responsible for survival of Y. pestis in the lungs. Our lab recently showed that Pla functions as an adhesin that promotes binding to alveolar macrophages to facilitate translocation of effector proteins called Yops into the cytosol of target host cells via a type 3 secretion system (T3SS). Loss of Pla-mediated adherence disrupted the preinflammatory phase of disease and resulted in early neutrophil migration to the lungs. While it is established that Yersinia broadly suppresses host innate immune responses, it is not clear precisely which signals need to be inhibited to establish a preinflammatory stage of infection. Here, we show that early Pla-mediated suppression of Interleukin-17 (IL-17) expression in alveolar macrophages and pulmonary neutrophils limits neutrophil migration to the lungs and aids in establishing a preinflammatory phase of disease. In addition, IL-17 ultimately contributes to neutrophil migration to the airways that defines the later proinflammatory phase of infection. These results suggest that the pattern of IL-17 expression contributes to the progression of primary pneumonic plague.

clear precisely which signals are necessary to suppress to establish a preinflammatory disease phase. The finding that deletion of Pla partially alleviates this early immune suppression afforded us the opportunity to identify those specific host molecules that are targeted to limit early neutrophil infiltration in the lungs. We predicted that measuring expression of proinflammatory cytokines and chemokines in the lungs of mice infected with Y. pestis lacking Pla (Dpla) would reveal specific molecules that are otherwise suppressed during the earliest stages of wild-type infection (,24 h) that we could evaluate in greater detail. We analyzed expression of a number of host proinflammatory cytokines known to contribute to neutrophil chemotaxis via reverse transcriptionquantitative PCR (qRT-PCR) during the early stages of infection with wild-type CO92 Y. pestis, Dpla, or bacteria lacking the pCD1 virulence plasmid (27)(28)(29)(30). Infection of mice with pCD1 Y. pestis occurs in the absence of T3S and therefore serves as a positive control for expression of target molecules. While these molecules had various levels of expression, we consistently observed a significant increase in IL-6, MCP-1, and IL-17 expression in the absence of Pla compared to wild-type infection (Fig. 1), suggesting that expression of these molecules is suppressed by Y. pestis to establish an early preinflammatory disease phase. We therefore sought to determine how expression of IL-6, MCP-1, and IL-17 impacted the biphasic progression of pneumonic plague.
IL-17 drives early neutrophil infiltration into the lungs in the absence of Pla. We recently showed that deletion of Pla results in the early migration of neutrophils into the lungs that is detected in the bronchoalveolar lavage fluid of infected mice as early as 12 h post infection (hpi) (13). This coincided with a significant increase in expression of IL-6, MCP-1, and IL-17 ( Fig. 1). IL-6 and IL-17 are known to contribute to neutrophil chemotaxis during multiple infections/diseases (28,(31)(32)(33)(34). IL-6 is promptly synthesized in infected tissue and contributes to neutrophil migration into tissue by modulating expression of chemokines, including MCP-1 and endothelial adhesion molecules (35) or by inducing neutrophil mobilization from the bone marrow (36,37). Similarly, IL-17 has been shown to impact neutrophil recruitment, largely through its impact on chemokine expression and the release of neutrophil-mobilizing factors in local cells (38,39). While it is known that MCP-1 contributes to migration and infiltration of monocytes/macrophages in a variety of diseases, its expression can also contribute to neutrophil chemotaxis during pulmonary infection (28). Previous work showed that pulmonary infection with a Y. pestis mutant strain that saw reduced neutrophil infiltration into the airways also saw reduced expression of four specific proinflammatory molecules (40). These included IL-6, MCP-1 and IL-17, further correlating expression of these molecules with neutrophil chemotaxis to the lung. We therefore hypothesized that the absence of one or more of these molecules would eliminate early aberrant neutrophil chemotaxis to the lung in the absence of Pla. To test this, we FIG 1 Regulation of proinflammatory cytokine expression in the absence of Pla. RNA isolated from whole lungs of C57BL/6J mice (n = 3) infected intranasally (i.n.) with 10 6 CFU of WT CO92 Y. pestis, CO92 Y. pestis lacking pCD1 -, or CO92 Dpla 12 h postinfection (hpi) was used for qRT-PCR analysis of the relative expression of tumor necrosis factor alpha (TNF-a), IL-6, MCP-1, granulocyte-macrophage colony-stimulating factor (GM-CSF) and IL-17A during infection with pCD1or Dpla strains relative to wild-type infection. Error bars represent standard deviations. The graph shows a representative experiment. **, P # 0.01; ***, P # 0.001; ****, P # 0.0001 (2-way analysis of variance).  Fig. 2A and B). In contrast, the absence of IL-17 abrogated neutrophil migration to the lungs, and levels of lung neutrophils were significantly reduced in IL-17 KO mice during infection with Dpla Y. pestis compared to WT mice ( Fig. 2A and B). These results indicate that IL-17 contributes to neutrophil infiltration in the lung in the absence of Pla and suggests that the targeted suppression of IL-17 by Y. pestis aids in establishing an early preinflammatory disease phase. Absence of IL-6, MCP-1, or IL-17 did not impact growth of Dpla Y. pestis in the lung. Deletion of Pla results in decreased bacterial burdens in the lungs during i.n. infection of mice that is detectable as early as 12 hpi (12,13). Depletion of neutrophils in mice prior to infection rescued growth of Dpla Y. pestis at 24 hpi, indicating that the FIG 2 IL-17 is responsible for the influx of neutrophils during the preinflammatory phase seen in the absence of Pla. (A) F4/80neutrophil gate. Flow plots represent neutrophils as a percentage of live F4/80cells in the whole lung of IL-6, MCP-1, and IL-17 knockout mice or C57BL/6 WT mice infected i.n. with 10 4 CFU of CO92 WT or CO92 Dpla 24 hpi. (B) Quantification of the abundance of neutrophils (F4/80 -CD11b high Ly-6G 1 ) in the whole lungs of mice. The percentage of neutrophils out of all live cells is shown. Cy7, cyanine 7; PE, phycoerythrin. Experiments were performed $3 times, and graphs show the means from 2 to 3 pooled experiments; error bars represent the standard deviation (n = 10 to 15 mice). *, P # 0.05; **, P # 0.01 (2-way analysis of variance). The significance of the difference in lung neutrophil levels between IL-17 knockout mice and wild-type C57BL/6 mice infected with Dpla Y. pestis was confirmed by Welch's t test at P # 0.05.

IL-17 Contributes to the Progression of Pneumonic Plague
Infection and Immunity early influx of neutrophils likely controls Y. pestis in the absence of Pla (13). We sought to determine if the absence of either IL-6, MCP-1, or IL-17 alone could rescue growth and/or dissemination of Dpla Y. pestis. We predicted that the absence of IL-17 would rescue Dpla bacterial burdens due to reduced numbers of neutrophils in the lungs. As expected, we observed a significant decrease in lung bacterial burdens in WT B6 mice infected with Dpla compared to WT Y. pestis at 24 hpi (Fig. 3A). The absence of IL-6, MCP-1, or IL-17 did not impact the growth of Dpla Y. pestis relative to WT bacteria, indicating that neither of these cytokines alone drives responses capable of controlling infection (Fig. 3A). While the loss of MCP-1 and IL-17 had no effect on lung bacterial burden for either strain, the loss of IL-6 resulted in a slight but significant increase in the numbers of WT bacteria (Fig. 3A), indicating that IL-6 contributes to control of bacterial survival in the lung early during infection. While there is minimal dissemination of bacteria beyond the lung during the preinflammatory disease phase (12), we also evaluated whether loss of a particular cytokine/ chemokine of interest impacted dissemination of bacteria from the lung to other tissues as measured by the presence of bacterial burdens in the spleen. We observed no difference in dissemination in either CO92 WT or Dpla Y. pestis in the absence of IL-6 or MCP-1 at 24 hpi (Fig. 3B). However, the loss of IL-17 resulted in increased dissemination of CO92 WT Y. pestis to the spleen, where 80% of mice harbored detectable bacterial burdens, suggesting a role for IL-17 in controlling bacterial dissemination from the lung during the preinflammatory phase (Fig. 3B). In summary, while the absence of IL-6, MCP-1, or IL-17 did not rescue growth of Dpla Y. pestis, mice lacking IL-6 showed a slight increase in wild-type bacterial growth in the lung, and mice lacking IL-17 saw increased dissemination of wild-type bacteria to the spleen, indicating that these molecules contribute to bacterial control during pulmonary infection.

IL-17 Contributes to the Progression of Pneumonic Plague
Infection and Immunity Alveolar macrophages and neutrophils contribute to the production of IL-17 during the preinflammatory phase of disease. To better understand the expression of IL-17, we used flow cytometry coupled with intracellular cytokine staining to identify the cell types responsible for IL-17 production during the preinflammatory phase of infection. Previous work utilizing two strains of the Microtus biovar of Y. pestis showed increased expression of IL-17 produced by CD45 1 cells between 12 and 24 hpi and determined that neutrophils are a major source of IL-17 during infection (24). At 12 hpi, IL-17 production was split equally between CD45 1 CD3 1 and CD45 1 CD3populations during WT infection ( Fig. 4A and B). CD45 1 CD3 1 cells include lymphocyte Th17 and natural killer T cell populations that are well-established producers of IL-17 (41). CD45 1 CD3populations include resident lung leukocytes such as alveolar macrophages, interstitial macrophages, and dendritic cells, as well as neutrophils and inflammatory monocytes. Interestingly, at 12 hpi with Dpla Y. pestis there was a shift to a significantly higher percentage of IL-17 expression for CD45 1 CD3cells compared to CD45 1 CD3 1 populations ( Fig. 4A and B). Using a fluorescent antibody panel to identify CD45 1 CD3leukocyte populations in the lung, we found that alveolar macrophages were significant producers of IL-17 at 12 hpi during both WT and Dpla infection (Fig. 4D). At 24 hpi CD45 1 CD3 1 and CD45 1 CD3remained the primary producers of IL-17 (Fig. 4C), and both Dpla and WT infection saw CD45 1 CD3cells producing a significantly higher percentage of IL-17 than that in CD45 1 CD3 1 populations. We found that in addition to alveolar macrophages, neutrophils were also significant producers of IL-17 at 24 hpi (Fig. 4E). This coincides with the migration of neutrophils to the alveolar space that is observed at 12 to 24

IL-17 Contributes to the Progression of Pneumonic Plague
Infection and Immunity hpi (13). Of note, at 24 hpi we began to observe IL-17 production from a population of CD45-CD3cells, indicating that endothelial/epithelial cells may play a role in IL-17 signaling during the mid/later stages of the preinflammatory disease phase. We also observed IL-17 expression from dendritic cell (DC) populations at both time points, highlighting the importance of the initial signaling by lung antigen-presenting cells during the earliest stages of infection. These results suggest that early during infection alveolar macrophages produce IL-17 to recruit an initial wave of neutrophils that also produce IL-17. IL-17 contributes to neutrophil recruitment during the proinflammatory phase of infection. The proinflammatory disease phase of pneumonic plague is marked by a dramatic influx of neutrophils into the lungs accompanied by the onset of a proinflammatory cytokine storm (3,8,42). It remains unclear, though, precisely which host signals drive neutrophil infiltration and progression into the proinflammatory phase of disease. In addition to early events in the lung during Dpla infection, expression of IL-6, IL-17, and MCP-1 has been previously shown to correlate with levels of pulmonary neutrophils during the later proinflammatory stage of infection (40). We therefore sought to examine the role of IL-6, IL-17, and MCP-1 in the transition to a proinflammatory state in the lung during infection. The Dpla strain has attenuated growth in the lung beginning at 12 hpi and does not progress to the proinflammatory stage of disease; therefore, WT infection alone was used to evaluate the later proinflammatory stage of infection. Mice lacking IL-6, IL-17, or MCP-1 showed no difference in lung bacterial burden compared to WT mice infected with Y. pestis, nor was there a difference in dissemination to the spleen, suggesting that IL-17 does not contribute to control of bacterial survival or dissemination during the proinflammatory disease phase (Fig. 5A and B). However, the absence of IL-17 saw significantly decreased levels of neutrophils in the lungs compared to WT mice ( Fig. 6A and B). Despite a roughly 20% decrease in lung neutrophil levels, the absence of IL-17 did not appear to impact the presence and/or

IL-17 Contributes to the Progression of Pneumonic Plague
Infection and Immunity organization of the neutrophil-packed inflammatory lesions that are a hallmark of pneumonic plague (Fig. 7A) (4). While inflammatory lesions appeared to be less densely packed and borders were more poorly demarcated, we did not detect a difference in the number or size of pulmonary lesions between WT mice and IL-17 2/2 mice ( Fig. 7B and C). These results suggest that IL-17 contributes to the recruitment of neutrophils during the proinflammatory disease phase, but IL-17 alone is not responsible for the inflammatory pathology seen during the later stages of disease.

DISCUSSION
The protease Pla functions as an adhesin to facilitate T3S of the Yop effectors into host alveolar macrophages to silence host innate immune responses (11,13,14,18). The absence of Pla results in increased neutrophil influx into the lung early during infection (13). This positioned Pla as an important factor in maintaining the preinflammatory phase of Y. pestis infection. While Y. pestis is efficient at silencing host inflammatory responses overall, the specific molecules that need to be suppressed to limit early neutrophil infiltration into the lungs remain unknown and therefore represent a knowledge gap in the field. In this study, we used Y. pestis lacking Pla as a tool to identify host factors that are typically "silenced" in the lung by Pla-mediated T3S during the earliest stages of infection. Using an i.n. murine infection model of pneumonic plague, we found that partially alleviating early suppression of host responses by deleting Pla resulted in increased expression of host IL-6, MCP-1, and IL-17 during the early preinflammatory phase of disease. We coupled infection using knockout mice and a

IL-17 Contributes to the Progression of Pneumonic Plague
Infection and Immunity bacterial mutant strain to characterize the role of each of these molecules in the host responses that define the biphasic progression of pneumonic plague. While the absence of IL-6 and MCP-1 did not significantly alter the progression of disease, we identified IL-17 as a key cytokine that impacts host inflammatory responses during the progression of primary pneumonic plague. IL-17 is a proinflammatory cytokine that has neutrophil chemotactic properties and has previously been shown to be highly expressed during the proinflammatory phase of pneumonic plague (13,29,30,41). IL-17 has also been shown to be an important mediator of protection in vaccination studies of pneumonic plague, typically in synergy with other molecules such as gamma interferon (IFN-g ) (43,44). Our data show that IL-17 is suppressed by Pla-mediated T3S early during infection, contributing to inhibition of neutrophil migration to the lungs. A previous study by Bi et al. showed that IL-17 KO mice have increased lung bacterial burdens compared to WT mice during pneumonic plague (24). However, we did not find that IL-17 impacted bacterial survival. A likely explanation for this is that Bi et al. utilized strains from a different biovar of Y. pestis with a different lethal dose and different growth kinetics during infection (24). Bi et al. used strains of the biovar Microtus that saw peak bacterial burdens of roughly 10 8 CFU at 4 days postinfection in the lungs of infected mice and nearly 100% survival using similar inoculating doses as our study (24). In this scenario, the investigators found that IL-17 was protective and prevented lethality. In contrast, we used CO92 of the Orientalis biovar, infection with which results in 100% lethality by 3 days postinfection (8,12,45) and .10 12 CFU in the lungs at 48 hpi. In our study IL-17 is not protective and, alternatively, may contribute to disease severity later during infection through its contribution to neutrophil accumulation in the lungs. These data suggest that IL-17 may play different roles depending on the Yersinia biovar or relative lethality of the infecting bacterial

IL-17 Contributes to the Progression of Pneumonic Plague
Infection and Immunity strains. For less lethal biovars such as Microtus, IL-17-driven neutrophil chemotaxis may be protective, while during infection with Orientalis strains neutrophil infiltration likely contributes to the severity of pulmonary disease and therefore lethality. Further work will need to be done to further characterize this phenomenon, but it highlights the fact that specific host responses can be protective or exacerbate disease depending on the lethality/severity of the infecting strain or the susceptibility of the host. Previous data from our lab showed that the increased neutrophil infiltration in the lungs upon deletion of Pla limited replication of the Pla mutant during the preinflammatory stage of infection (13). Antibody-mediated depletion of neutrophils prior to infection rescued growth of Dpla Y. pestis by 24 hpi, and we showed that Pla contributes to bacterial resistance to neutrophil-mediated killing. However, in this study decreased neutrophils in the lung due to the loss of IL-17 did not impact the growth of CO92 Dpla. The significance of this finding is not yet clear. Perhaps neutrophil populations other than those recruited by IL-17 (which are therefore still present in IL-17 KO mice) are responsible for limiting the growth of Dpla Y. pestis. Bi et al. showed that IL-17 signaling contributes to M1 macrophage programming that is protective during pneumonic plague in mice (24). The Y. pestis CO92 strain, though, may require a combination of additional mechanisms/signals to limit bacterial growth. Though significant, the increase in neutrophil infiltration during infection with Dpla is subtle, especially compared to the dramatic influx of innate immune populations later during infection. It may be that this increase and the subsequent decrease in the absence of IL-17 are not sufficient to impact bacterial growth and/or survival. Regardless, the precise mechanism responsible for controlling Dpla Y. pestis remains unclear and is likely multifactorial. The absence of IL-17 did, though, result in enhanced WT bacterial burdens in the spleen, indicating that IL-17 contributes to controlling bacterial dissemination from the lung to other tissues. Also, loss of IL-6 resulted in a slight but significant increase in lung bacterial burden during wild-type Y. pestis infection. Current work is focused on characterizing the impact of IL-17 on downstream innate immune signaling during the early stages of pneumonic plague and the combination of host signals that need to be suppressed by Y. pestis to establish a replicative niche in the lung.
IL-17 has been described as being secreted by various cell types in the lung, including Th17, CD8 1 T cells, innate immune cells, etc.; however, the source of IL-17 during infection varies from pathogen to pathogen (19)(20)(21)(22)(23). Our study examined the sources of IL-17 during infection with Y. pestis and found that the cells secreting IL-17 during the initial stages of infection include a significant contribution from alveolar macrophages. This aligns with previous data from our lab and suggests that reduced targeting of alveolar macrophages for T3S likely results in IL-17 secretion (14). Alveolar macrophages are resident to the lung (46) and one of the first cells to make contact with the bacteria (8), thus initiating the production of IL-17 to recruit the first wave of neutrophils, which in turn, assume a role in secreting IL-17. This aligns with the data presented by Yujing Bi and colleagues, who described neutrophils as secreting IL-17 during Y. pestis infection (24). We also observed IL-17 secretion for DCs, potentially highlighting a yet to be described role for these cell populations in host responses that define the progression of pneumonic plague. While our study only expanded on the CD45 1 CD3leukocyte populations that produce IL-17, it is of note that there are additional populations that contribute to IL-17 expression that have yet to be characterized in detail during pneumonic plague. While WT infection saw equal contributions from CD45 1 CD3 1 and CD45 1 CD3populations in IL-17 production at 12 hpi, the absence of Pla saw a significant shift in that balance, with CD45 1 CD3populations producing increased IL-17 compared to CD45 1 CD3 1 populations. Interestingly, at 24 hpi we observed a contribution from CD45 -CD3cells in IL-17 production, which includes endothelial/epithelial populations. This gives insight into the temporal production of IL-17, where the earliest time point sees resident alveolar macrophages as the primary producer, and a shift to include neutrophils and, potentially, CD45 -CD3endothelial/epithelial cells beginning 12 h later. Future studies will be performed to gain a clear picture of the players and timing involved in the secretion of IL-17 during the critical preinflammatory stage of pneumonic plague.
A hallmark of primary pneumonic plague is the progression into a lethal proinflammatory phase of infection (3)(4)(5)(6)(7). The proinflammatory phase occurs around 36 h in mice and is marked by uncontrolled proinflammatory cytokine production leading to a rapid influx of neutrophils, damage to the alveolar architecture, and diminished pulmonary function and ultimately resulting in death (3)(4)(5)(6)(7). The onset of a proinflammatory cytokine storm during pneumonic plague is well established, and nearly any proinflammatory cytokine or chemokine measured during this stage of infection shows dramatically increased expression (3,4,7,12,40). It is not known, though, precisely which of these molecules/signals is responsible for the dramatic neutrophil chemotaxis to the lungs that defines the proinflammatory stage of infection. IL-17 has been described as one of the cytokines most highly upregulated during the proinflammatory disease phase (12,24), and we discovered that while it does not control bacterial survival in the lung, it does contribute to the massive influx of neutrophils during the proinflammatory disease phase (1,3,8,12). This implicates IL-17 as a player in the transition to a lethal proinflammatory state of infection in the lung and has therapeutic implications, as limiting pulmonary inflammation (including neutrophil infiltration) has been shown to enhance treatment of late-stage infection (47). However, the loss of IL-17 did not impact the formation and organization of inflammatory pulmonary lesions, not surprisingly indicating that progression into the proinflammatory disease phase is multifactorial and involves expression of IL-17 in tandem with other signals.
This work has shown that alveolar macrophage-and neutrophil-derived IL-17 is suppressed by Pla-mediated T3S early during infection to establish an initial preinflammatory stage of infection. Further, the upregulation, or rather, "derepression" of IL-17 expression contributes to neutrophil infiltration in the lungs and progression to a lethal proinflammatory stage of infection. While previous work had defined a protective role for neutrophil-derived IL-17 during infection with Yersinia of the Microtus biovar, we did not find that IL-17 controls bacterial growth and/or survival, and it is likely not protective during infection with the highly lethal CO92 strain of the Orientalis biovar. In summary, this work indicates that IL-17 is a critical cytokine, the expression of which contributes to the biphasic progression of pneumonic plague. Targeting IL-17 in tandem with antibiotic therapy may aid in treating late-stage pneumonic plague or pneumonia caused by other agents in which IL-17-driven neutrophil infiltration in the airways contributes to pathogenesis.

MATERIALS AND METHODS
Bacterial strains. Y. pestis fully virulent strain CO92 (biovar Orientalis), pCD1-CO92, and Dpla CO92 (12) were obtained from William E. Goldman (University of North Carolina at Chapel Hill). All experiments were performed in a biosafety level 3 facility at the University of Arkansas for Medical Sciences. Y. pestis strains were grown on brain heart infusion (BHI) agar (Difco) at 26°C for 2 days. For infection, Y. pestis strains were grown in 10 mL of BHI broth containing 2.5 mM calcium chloride for 16 h at 37°C with constant shaking.
Animal infections. All animal experiments were conducted as approved by the University of Arkansas for Medical Sciences (UAMS) Institutional Animal Care and Use Committee. Female C57BL/6, IL-6 knockout (B6.129S2-Il6 tm1Kopf /J), MCP-1 knockout (B6.129S4-Ccl2 tm1Rol /J), and IL-17 knockout (B6.129 (SJL)-Il17a tm1.1 (icre)Stck /J) mice (4 to 6 weeks old) were obtained from Jackson Laboratories. Mice were given food and water and maintained at 25 to 26°C with 40 to 70% humidity. For infection, mice were anesthetized via the intraperitoneal (i.p.) route using ketamine-xylazine. Anesthetized mice were inoculated via the intranasal (i.n.) route with a lethal dose (10 4 to 10 6 CFU) of Y. pestis delivered in 20 mL phosphate-buffered saline (PBS). Mice were sacrificed using a lethal dose of sodium pentobarbital injected i.p. Bacterial burdens were determined via harvesting and homogenizing the lungs and spleens of mice at 24 h and 48 h postinfection. Organ homogenates were serially diluted and plated on BHI agar to determine CFU.
Quantitative reverse-transcription PCR. Groups of mice were inoculated i.n. with 10 6 CFU of CO92 WT, CO92 pCD1--, or CO92 Dpla Y. pestis, and at 12 hpi lungs were harvested, placed in 1 mL of PBS, and homogenized using a tissue tearer. Homogenized lungs were centrifuged at 500 Â g for 5 min, supernatant was removed, and the pellet was lysed in TRIzol reagent (Thermo Fisher). Total RNA was then isolated from lung homogenates per the manufacturer's instructions and treated with DNase I using the TURBO DNA-free kit (Invitrogen). Then, 1 mg of RNA was reverse transcribed to generate cDNA using the Superscript III polymerase kit (Invitrogen). Real-time quantitative reverse transcription PCR was performed using PowerUp SYBR green master mix (Thermo Fisher) with a QuantStudio 6 system, and fold changes were calculated using the delta delta cycle threshold method normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) from the same sample.
Statistical analysis. All statistical analyses were done using 2-way analysis of variance or Welch's t test. All statistical analyses were performed using GraphPad Prism software v9.4.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 1.5 MB.