Microneedle array delivery of Yersinia pestis recapitulates bubonic plague

Summary

Fleas transmit Yersinia pestis directly within the dermis of mammals to cause bubonic plague. Syringe-mediated inoculation is widely used to recapitulate bubonic plague and study Y. pestis pathogenesis. However, intradermal needle inoculation is tedious, error prone, and poses a significant safety risk for laboratorians. Microneedle arrays (MNAs) are micron-scale polymeric structures that deliver materials to the dermis, while minimizing the risk of needle sticks. We demonstrated that MNA inoculation is a viable strategy to recapitulate bubonic plague and study bacterial virulence by defining the parameters needed to establish a lethal infection in the mouse model and characterizing the course of infection using live-animal optical imaging. Using MNAs, we also demonstrated that Y. pestis must overcome calprotectin-mediated zinc restriction within the dermis and dermal delivery of an attenuated mutant has vaccine potential. Together, these data demonstrate that MNAs are a safe alternative to study Y. pestis pathogenesis in the laboratory.

Graphical abstract

Highlights

• •MNAs reproducibly deliver Y. pestis to the dermis, mimicking flea transmission
• •MNA inoculation reveals Y. pestis encounters calprotectin in the dermis
• •Calprotectin Zn restriction inhibits dermal colonization and lymph node dissemination
• •MNA intradermal delivery of an attenuated Y. pestis suggests potential for vaccine

Drug delivery system; Microbiology

Introduction

The gram-negative bacterium Yersinia pestis causes the disease bubonic plague, which is responsible for the most fatal pandemic in history known as the Black Death.^1,2,3 Currently, bubonic plague cases still occur in areas in which the bacterium is endemic in animal populations, including the southwestern United States.^4,5 Within these environments, Y. pestis is maintained within an enzootic cycle in which the bacterium is transmitted between mammalian hosts by a flea vector. During the feeding process, infected fleas deposit Y. pestis directly into the intradermal skin layer.^6,7 These bacteria then interact with host immune cells and rapidly disseminate from the skin to the draining lymph nodes.^8,9 Migration to the lymph nodes occurs minutes after infection and appears to be independent of intracellular trafficking by phagocytic cells.^8,9 Upon reaching the lymph nodes, Y. pestis evades clearance by resident immune cells and can replicate to high numbers. Eventually the bacteria breach the lymph node to disseminate into the bloodstream,¹⁰ which is an important step in the transmission of Y. pestis to a new flea vector and continuing the enzootic cycle. Spillover into human populations occurs when individuals come into contact with these infected fleas.^11,12 Despite the persistence of Y. pestis in many countries and potential for human spillover infections, an FDA-approved vaccine is not available to prevent plague. A greater understanding of bacterial transmission and dissemination would significantly affect strategies for preventing and controlling the spread of Y. pestis.

In the past, subcutaneous injection of Y. pestis has been routinely used to model bubonic plague. While this approach has revealed many aspects of Y. pestis pathogenesis, growing evidence suggests that subcutaneous injection of Y. pestis does not completely recapitulate the disease kinetics of bubonic plague and lacks the ability to study important pathogen interactions at the initial site of infection in the dermis.^8,13 Therefore, investigators have started to employ intradermal injection models that have proven to be more biologically reminiscent of flea transmission.^10,13,14 Intradermal administration into the dermis of the ear has revealed that Y. pestis dissemination and kinetics of lethal infection are much quicker than rates observed from subcutaneous injection.^8,10 These studies highlight that, upon Y. pestis intradermal inoculation, a few bacteria travel rapidly to the draining lymph node, where they replicate to high numbers attracting an influx of immune cells that form a bubo.^8,10,15 At the site of inoculation within the dermis, Y. pestis encounters neutrophils, but the bacteria overcome the immune system to avoid clearance and eventually replicate at the infection site.^8,9,16 Neutrophils have important roles in nutritional immunity through the release of proteins that restrict bacterial access to metals, including calprotectin that specifically sequesters zinc, manganese, and iron.^{17,18,19,20,21,22,23} Using subcutaneous administration of Y. pestis, we have previously shown that calprotectin is the primary zinc acquisition barrier for Y. pestis during bubonic plague.²⁴ We also showed that the bacterium overcomes calprotectin metal restriction via the expression of two zinc acquisition systems, ZnuABC and yersiniabactin.²⁴ However, the tissue in which Y. pestis encounters calprotectin restriction during bubonic plague remains to be determined. Therefore, application of a model that more closely mimics natural flea transmission, such as intradermal inoculation, is key to understand where Y. pestis encounters nutritional immunity.

While intradermal injection has proven to be a valuable tool to study plague, it presents a difficult challenge within the Biosafety Level 3 (BSL-3) laboratory. The accurate application of a standard syringe needle to precisely inoculate the dermal layer of the skin while wearing appropriate personal protective equipment proves to be tedious and difficult. Moreover, these needles are significant safety hazards for laboratorians. Finally, common administration volumes needed to ensure reproducible inoculations (>20 μL) can cause tissue damage within the intradermal space that may stimulate an immune response unrelated to bacterial infection.¹³ Therefore, a more efficient and safer technique to perform biologically relevant intradermal infections would significantly impact the adoption of the intradermal model for plague researchers. A unique option is presented by microneedle arrays (MNAs), which are micron-scale needles fabricated from biomaterials for drug and biologic delivery to defined intradermal depths.²⁵ Importantly, because of the micron length of the MNAs, they cannot puncture latex or nitrile gloves under normal application during research, significantly reducing the risk of self-inoculation encountered with standard syringe needles.²⁶ MNAs have been used extensively for vaccine and drug delivery and, more recently, have been adapted to specifically administer bacteria to the intradermal space.^25,27,28,29 For drug delivery, MNAs have been utilized to study mechanisms to combat methicillin-resistant Staphylococcus aureus (MRSA) and distribute antibiotic vancomycin directly to the infected skin area.³⁰ MNAs have been also used for a variety of vaccine studies, including immunizing with influenza antigen and with stabilized dengue virus.^28,31,32 Outside of studying infectious diseases, delivery by MNAs has shown an improvement for distribution of cancer vaccines and immunotherapies.²⁵ Finally, MNAs have been adapted to mimic intradermal tick transmission inoculation of Francisella tularensis.²⁹ The successful and reproducible intradermal delivery achieved in these studies suggests that MNAs may provide a useful tool to mimic flea transmission for the study of bubonic plague pathogenesis in the BSL-3 laboratory.

Here we show that MNA administration of Y. pestis reproducibly establishes intradermal instillation of Y. pestis that leads to lethal bubonic plague with disease kinetics previously reported for intradermal needle and flea inoculation. Furthermore, we demonstrate the utility of MNA administration as a tool to study Y. pestis virulence by using this model to establish that host calprotectin effectively limits Y. pestis zinc acquisition within the dermis. Finally, using an attenuated strain of Y. pestis, we show that MNA administration within the dermis has potential as a vaccine delivery platform against pulmonary infection.

Results

MNAs reproducibly deliver Y. pestis to the dermis

Previous studies indicate that an infectious dose >200 colony-forming units (CFUs) of Y. pestis delivered via syringe inoculation to the ear pinna results in a reproducible lethal infection.^13,33 As such, our first step in adapting MNAs for Y. pestis inoculation was to establish conditions required to achieve reproducible inoculums within this range. The MNAs designed for this study were 650 μm from the MNA base to the tip to achieve penetration of the dermal layer.^34,35 Unlike syringes, solid, non-degradable MNAs do not actively inject materials into the dermal layer. Instead, these MNAs are coated with the transported substances or microorganisms and pressed onto the skin for delivery into the puncture sites.²⁹ Therefore, MNAs were incubated for 30 min with increasing bacterial concentrations (from 3 × 10³ to 3 × 10⁸ CFU/mL) to determine if an inoculum of 200–700 CFU could be achieved. After incubation, the MNA was placed on the bolus of the ear (Figure 1A), applied for 5 s, and then removed. 15 min later, ears were harvested, 70% ethanol was applied with a sterile swab to remove external bacteria, and subsets of ears were either fixed in 10% formalin and imaged by two-photon microscopy or homogenized using a tissue tearer and plated on selective agar to enumerate CFU (the limit of detection was 100 CFUs/tissue). Imaging of the ears revealed uniform needle inoculations across the dermal layer (Figure 1B) that appeared to travel approximately 30 μm into the tissue (Figure 1C and Video S1). We were unable to recover any bacteria from ears inoculated with MNAs incubated with inoculums below 3 × 10⁵ CFU/mL (Figure 1D), indicating that less than 100 bacteria were administrated by MNAs incubated at these concentrations. However, bacteria were recovered from tissues inoculated with MNAs incubated with 3 × 10⁶ CFU/mL and the number of CFU recovered increased as the inoculum increased to 3 × 10⁸ CFU/mL (Figure 1D). Furthermore, we could detect rod-shaped bacteria at the inoculation site within the dermis layer (Figure 1E and Video S2). These data suggested that MNAs incubated with 3 × 10⁷ CFU/mL delivered a dose within our desired range of ∼500 CFU. To determine if these conditions yielded consistent reproducible inoculums, 15 mouse ears were inoculated with MNAs incubated with 3 × 10⁷ CFU/mL and inoculums were calculated 15 min later. Bacterial enumeration showed a mean of 624 ± 129 CFUs recovered from the tissue (Figure 1F). Together, these data indicate that MNAs can consistently deliver Y. pestis into the dermis.

Figure 1.

MNAs can reproducibly deliver Y pestis to the dermis

To determine the number of bacteria delivered by MNA infection, MNAs were incubated with increasing numbers of colony forming units (CFUs) of Y. pestis for 30 min.

(A) Mice are infected by applying the MNA to the left ear and applying pressure for 5 s. MNAs are removed with forceps after infection.

(B and C) Image of an ear inoculated with MNA without bacteria. White arrowheads denote MNA puncture sites.

(D) To determine the number of bacteria delivered to the dermis by MNAs, ears were harvested at 15 min post-infection with microarrays incubated with indicated concentrations of bacteria and recovered CFU were enumerated by serial dilution (mean ± S.D.). Dotted line indicates limit of detection.

(E) Image of MNA inoculation site from infected animal. White arrowheads denote bacteria (red).

(F) Bacterial CFU recovered from ears of mice infected with MNAs incubated with 3 × 10⁷ CFU/mL. Each circle represents a single tissue (n = 15) harvested from two independent experiments.

MNA inoculation results in a lethal Y. pestis infection that mimics that reported for flea inoculation

Intradermal inoculation via flea bite or needle injection results in deposition of Y. pestis into the dermis and rapid dissemination of a small population of bacteria to the draining lymph node.^8,9,16 Despite recruitment of resident neutrophils, the bacteria that remain in the dermis are able to survive and undergo limited proliferation.^8,13,36 To determine if MNA inoculation results in a similar disease manifestation, MNAs were used to inoculate mice with a Y. pestis strain expressing a bioluminescent bioreporter.³⁷ A subset of mice (n = 3) was euthanized at 15 min to confirm the inoculum (723 ± 384 CFUs), and optical imaging was used to monitor bacterial proliferation and dissemination as a function of bioluminescence in the remaining animals (Figures 2A–2C). At 0.25 h, bioluminescent bacteria were detected in the ears but not the cervical lymph nodes. By 12 h post-infection the signal in the ears decreased, most likely due to loss of bacteria that were originally deposited on the ear surface, but we observed bioluminescent signal in the cervical lymph nodes indicating lymph node colonization by this time point. Bioluminescence in both tissues increased over the remaining course of the infection, with the signal from both tissues peaking at 48 h post-inoculation (Figures 2A–2C). Conventional bacterial enumeration at 48 h post-inoculation confirmed robust colonization of both tissues by Y. pestis (Figure 2D). Furthermore, by 48 h post-inoculation, bioluminescent signal was detected in other tissues, indicating bacterial dissemination into the bloodstream (Figure 2C). Animals began to succumb to the infection by 60 h post-inoculation, with all animals meeting endpoint criteria by 72 h post-inoculation (Figure 2E). The kinetics of the infection using MNAs for inoculation closely recapitulates what has been reported for intradermal needle inoculation or flea transmission,^8,36 supporting that MNA inoculation is a viable mechanism to model intradermal infections by Y. pestis.

Figure 2.

MNA inoculation results in a lethal Y pestis infection that mimics flea inoculation

To determine if Y. pestis infection follows similar kinetics of dissemination as a needle inoculation, Swiss Webster mice were infected intradermally with a bioluminescent strain of Y. pestis (combined data from two independent experiments; n = 5 per experiment). Bacterial dissemination and replication as a function of bioluminescence by optical imaging of the (A) ear and (B) draining LN.

(C) Representative mouse during optical imaging.

(D) Measurement of bacterial CFU in the ear and LN at 48 hpi.

(E) Survival curve of infected mice. Each point represents an individual mouse.

MNA administration reveals the importance of calprotectin in restricting Y. pestis in the dermis

Using a subcutaneous injection model, we previously showed that calprotectin is the primary zinc sequestration barrier that Y. pestis needs to overcome to cause a lethal infection in mice.²⁴ While these studies showed that calprotectin protects mice from lethal infection by Y. pestis zinc acquisition mutants, the study design did not inform in which tissues Y. pestis encounters calprotectin-mediated zinc restriction during bubonic plague. Because neutrophils are a primary source of calprotectin³⁸ and rapidly recruited to the dermis during flea transmission of Y. pestis,³⁶ we next investigated if calprotectin restricts zinc availability for Y. pestis in the dermis. To test this hypothesis, C57Bl6/J and S100A9^−/− (deficient in calprotectin) mice were inoculated using the MNAs with Y. pestis or a Y. pestis zinc acquisition mutant (Y. pestis znuA ybtX) expressing a bioluminescent bioreporter.³⁷ A subset of mice was euthanized at 15 min post-inoculation to confirm that equivalent inoculums were delivered (750 ± 117 and 526 ± 183, respectively), and bacterial proliferation and dissemination were monitored in the remaining mice by optical imaging. Greater bioluminescence was detected at each time point in the ears of S100A9^−/− mice compared to C57BL6/J mice infected with Y. pestis (Figures 3A and 3B), suggesting that calprotectin restricts bacterial proliferation within the dermis. Calprotectin restriction was further supported by infection with the znuA ybtX mutant, which was unable to replicate in the dermis of C57Bl/6J mice but robustly replicated in the dermis of S100A9^−/− mice (Figures 3A and 3B). Y. pestis also appeared to more rapidly disseminate to the draining lymph node and/or was better able to proliferate within these tissues in S100A9^−/− mice than C57Bl/6J mice (Figures 3C and 3D). Moreover, S100A9^−/− mice succumbed to infection with Y. pestis quicker than C57Bl/6J mice (mean time to death of 72 h vs. 90 h, respectively; p = 0.109) (Figure 3E). As we reported previously using the subcutaneous injection model,²⁴ S100A9^−/− mice were unable to survive challenge by the znuA ybtX zinc acquisition mutant, while C57Bl/6 mice survived (Figure 3E; p < 0.0001). Together, instillation by MNAs allowed us to demonstrate that calprotectin is an obstacle to Y. pestis infection within the dermis and a potent barrier for bacterial zinc acquisition within this tissue.

Figure 3.

MNA administration reveals the importance of calprotectin in restricting Y pestis in the dermis

To determine tissues that Y. pestis encounters calprotectin, C57BL6/J or calprotectin-deficient (S100A9^−/−) mice (n = 10) were infected intradermally with bioluminescent strains of Y. pestis or a znuA ybtX mutant.

(A) Bacterial dissemination and replication as a function of bioluminescence measured by optical imaging in infected ears. Representative ear imaging for (B) C57BL6/J or (C) S100A9^−/− mice.

(D) Bacterial dissemination and replication as a function of bioluminescence measured by optical imaging in the draining LN. Representative LN imaging for (E) C57BL6/J and (F) S100A9^−/− mice.

(G) Survival curve of infected mice. For optical imaging: Two-tailed t test; ∗ = p < 0.05, ∗∗∗ = p < 0.001, ∗∗∗∗ = p < 0.0001. For survival curves: Log rank test; C57BL6/J Y. pestis vs. C57BL6/J znuA ybtX - p < 0.001; C57BL6/J znuA ybtX vs. S100A9^−/−znuA ybtX - p < 0.001.

MNA delivery of attenuated Y. pestis changes disease kinetics in subsequent challenge

MNA delivery of vaccine candidates to the dermis can induce robust immune responses that protect against pulmonary challenge with influenza A virus and F. novicida.^29,31 Because the znuA ybtX mutant is severely attenuated for bubonic plague, we next investigated if the mice that survived MNA administration with the znuA ybtX mutant might be protected from subsequent pulmonary infection with Y. pestis. Therefore, C57Bl/6J mice were inoculated with Y. pestis or Y. pestis znuA ybtX and monitored for the development of moribund disease for 14 days. As expected, all mice challenged with Y. pestis succumbed to infection while those challenged with the znuA ybtX mutant survived the infection (Figure 4A). Optical imaging showed a similar pattern of proliferation and dissemination in the ears and draining lymph nodes as observed in Figure 3 (Figures 4B and 4C), and, importantly, no bioluminescence signal above background was detected in the tissues of znuA ybtX infected mice (Figures 4B and 4C). On day 14, the mice that survived MNA inoculation with znuA ybtX and a naive group of mice were infected with bioluminescent Y. pestis by intranasal administration and monitored for the development of lethal infection. Optical imaging of the thoracic cavity indicated slower proliferation of Y. pestis in the lungs of animals that were previously exposed to the znuA ybtX mutant, but ultimately the mice were not able to inhibit bacterial proliferation (Figures 4D and 4E). However, slower proliferation in the znuA ybtX-exposed mice directly correlated with a significant delay in lethality compared to naive infected mice (Figure 4E; mean time to death of 96 h vs. 60 h, respectively; p = 0.0002). Together, these data indicate that despite the inability of the znuA ybtX mutant to proliferate and disseminate, mice were mounting a systemic immune response to the bacteria, though a longer time or repeated MNA administration may be required to generate a fully protective response.

Figure 4.

MNA delivery of attenuated Y pestis changes disease kinetics in subsequent challenge

To determine the potential for using MNA infection with the znuA ybtX mutant as a vaccine platform, (A) C57Bl/6J mice (n = 10) were infected via MNAs with Y. pestis and znuA ybtX (inoculum = 1 × 10²) and monitored for 14 days.

(B and C) Bacterial dissemination and replication as a function of bioluminescence measured by optical imaging in infected ears and draining LNs.

(D) Bacterial replication as a function of bioluminescence measured by optical imaging in the thoracic cavity after Y. pestis intranasal inoculation of mice that survived znuA ybtX infection.

(E) Representative mice during optical imaging after reinoculation with Y. pestis.

(F) Survival curve of infected mice. For optical imaging: two-tailed t test. For survival curves: log rank test: p = 0.002.

Discussion

While subcutaneous inoculation has been a valuable tool to study bubonic plague pathogenesis, intradermal inoculation is more biologically relevant to natural flea transmission of Y. pestis.³⁹ Previously, MNAs have been used for mouse infections with F. novicida,²⁹ another bacterium transmitted by an arthropod vector. Building on these studies, we developed and characterized a similar model for Y. pestis inoculation in the ear pinna using MNAs, which is a safer and more efficient tool than typical needle inoculation.²⁷ Our first goal was to ensure that infection with MNAs would provide dissemination kinetics similar to those reported in studies of flea or other intradermal infection models with Y. pestis.^8,9,10 These previous studies suggest that, upon infection in the dermal layer, a few bacteria immediately travel via lymphatic vessels to the draining lymph node and begin to replicate to establish infection.^8,9,10 In a direct comparison between intradermal and subcutaneous routes of infection, Gonzalez et al. demonstrated that lymph node colonization occurred more rapidly with the intradermal infection route.¹³ Using bioluminescent imaging, we recapitulated these results and observed lymph node colonization within 12 h of inoculation, with the peak bioluminescence observed at 48 h post-inoculation. Importantly, these data are also similar to infection with Y. pestis by flea inoculation.⁹ Furthermore, our MNA infection model shows a rapid progression to lethality, which correlates with previous needle intradermal infections for plague pathogenesis.¹³ It has been speculated that lethality in the intradermal bubonic model is faster than subcutaneous injection because of greater presence of, or access to, lymphatic vessels in the dermal layer of the skin.¹³ It is also important to note that the volume administered by needle injection can influence the host response, with larger volumes (>2 μL) causing tissue damage and potential administration beyond the intradermal space.^13,40,41 Not only does a smaller inoculation more closely match a flea bite and cause less tissue damage but it also is less favorable for the bacteria.¹³ Importantly because the microneedles do not deposit large volumes into the tissue, the bacteria are delivered directly within the intradermal space without similar concerns of mislocalization of bacteria due to changes in inoculum volumes.²⁷ Overall, these data support that the MNA infection model is biologically applicable to bubonic plague infection that is administered through a flea bite.

Using MNAs, we recapitulated previous findings that calprotectin is the main obstacle for zinc acquisition during Y. pestis infection (Figure 3).²⁴ However, by combining MNA administration with optical imaging of bacterial mutants, we were able to expand on these initial studies and show that Y. pestis encounters calprotectin-mediated zinc sequestration within the dermis during bubonic plague. These data indicate that Y. pestis needs to rapidly adjust to metal limitation upon flea transmission and that zinc acquisition is essential for the bacterium to colonize the dermis. Moreover, these data also demonstrate that we can use Y. pestis MNA administration to better define host-pathogen interactions specifically within the dermis, which has different cell types and immune responses than other tissues. For example, while neutrophils have been implicated as the primary source for calprotectin during immune responses in the lungs and liver,^42,43,44 a study using skin biopsies of Lyme disease patients suggests that other cell types are the primary source for calprotectin-mediated metal limitation in the dermis.⁴⁵ With the establishment of this reproducible dermal model of infection, moving forward we can use Y. pestis as a tool to better understand the mechanisms involved in nutritional immunity within the dermis.

With plague cases still occurring worldwide, there remains a need for a vaccine. Due to differences in lymphatic vascularization between the dermal layer and subcutaneous level in the skin,⁴⁶ intradermal and subcutaneous inoculations show differences in immune responses, with intradermal vaccines generating a more robust mucosal immune response.^13,47,48,49 Recent successes with MNA-delivered vaccines suggest an MNA platform can provide a strong mucosal immune response.^29,31 While vaccination was not the primary goal of these studies, we were able to show that intradermal delivery of an attenuated Y. pestis strain had an impact on subsequent pulmonary challenge—slower replication and a delay in the development of lethal infection. While a change in overall survival was not observed, likely because of the limited time frame that mice were allowed to recover from challenge and generate a robust immune response, these studies provide preliminary data that MNA delivery of a vaccine may result in effective protection against pneumonic plague and justify future studies specifically designed for vaccination (i.e., prime-boost scenarios, measurements of immune responses, different vaccine formulations, etc.).

Limitations of the study

In this study, we established a model to mimic flea inoculation of Y. pestis using MNAs and applied this model to demonstrate that calprotectin-mediated zinc limitation in the dermis is a barrier to Y. pestis infection. We also used the model to provide preliminary evidence that an attenuated strain of Y. pestis introduced into the dermis via MNA inoculation has the potential to generate protective immune responses against subsequent exposure to pneumonic plague. While MNA inoculation is a model for flea transmission, like other laboratory models, inoculation via MNAs will not be able to account for flea-derived factors that may contribute to transmission during natural flea inoculation. Moreover, while Y. pestis encounters calprotectin in the dermis, the cellular source of calprotectin remains unresolved. Finally, while these data indicate that a protective immune response is generated after exposure to the attenuated Y. pestis zinc acquisition mutant, more refined studies, including direct measurement of immune responses and optimized exposure times, are required to fully understand the adaptive immune response generated via this model and to identify correlates of protection.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-Yersinia pestis serum	This paper.	NA
Alexa Fuor546 Donkey anti-Rabbit IgG381 (H + L)	Invitrogen	A10040; RRID: AB_2534016
Bacterial and virus strains
Yersinia pestis KIM6+	Bearden et al.50	NA
Yersinia pestis znuA ybtX	Bobrov et al.51	NA
Yersinia pestis KIM6+ pGEN-luxCDABE	This paper.	NA
Yersinia pestis znuA ybtX pGEN-luxCDABE	This paper.	NA
Yersinia pestis KIM6+ pGEN222-dsRed	This paper.	NA
Experimental models: Organisms/strains
Mouse: Swiss Webster (CFW)	Charles River	Crl: CFW(SW) Strain Code: 024
Mouse: C57Bl/6J	The Jackson Laboratory	Strain #: 000664
Mouse: C57Bl/6J S100A9−/−	Manitz et al.52	NA
Recombinant DNA
pGEN-luxCDABE	Sun et al.37	NA
pGEN222-dsRed	Sun et al.37	NA
Other
Microneedle Arrays	Chandler et al.29; Shah et al.53	NA

Resource availability

Lead contact

Additional information and requests for resources and reagents used in this study should be directed to and will be fulfilled by the lead contact, Matthew Lawrenz, matt.lawrenz@louisville.edu.

Materials availability

• •This study did not generate new unique reagents.
• •There are restrictions on the availability of Y. pestis strains used in these studies due to Select Agent Regulations.

Data and code availability

• •Data reported in this paper will be shared by the lead contact upon request, Matthew Lawrenz, matt.lawrenz@louisville.edu.
• •This study does not report any original code.
• •Any information related to the data presented in this manuscript can be obtained from the lead contact, Matthew Lawrenz, matt.lawrenz@louisville.edu.

Experimental model and study participant details

Animals

C57BL/6J mice were purchased from The Jackson Laboratory and Swiss Webster were purchased from Charles River. S100A9^−/− mice⁵² were bred at the University of Illinois at Urbana-Champaign prior to transfer to the University of Louisville for infection. Both male and female mice (6–12 weeks in age) were randomly assigned to experimental groups. No sex bias was observed during these studies. Animals were housed in accordance with NIH guidelines and all procedures were approved by the University of Louisville IACUC and the University of Illinois IACUC. Three days prior to challenge with Y. pestis, animals were transferred to the University of Louisville’s Center for Predictive Medicine Regional Biocontainment Laboratory ABSL-3 facilities to acclimate to the facility. After infection, mice were observed for up to 14 days for the development of moribund disease and humanely euthanized if they met predefined endpoints.

Bacterial strains

Y. pestis KIM6+⁵⁰ and Y. pestis znuA ybtX⁵¹ were routinely grown in Difco brain heart infusion (BHI) broth (BD Biosciences). Strains transformed with the pGEN-luxCDABE or pGEN222-dsRed plasmids³⁷ were grown in the presence of kanamycin (50 μg/mL). For infection, bacteria were grown overnight at 26°C in BHI broth prior to intradermal challenge. Prior to intranasal instillation, Y. pestis was grown overnight at 26°C, diluted to 0.05 OD₆₀₀ in BHI broth with 2.5 mM CaCl₂, and grown at 37°C with aeration for 16 to 18 h.^54,55 Bacterial concentrations were determined using a spectrophotometer and diluted to desired concentrations in 1x PBS for intranasal instillation. Concentrations of bacterial inoculums for mouse studies were confirmed by serial dilution and enumeration on agar plates.

Method details

Microneedle array preparation

Microneedle arrays (MNAs) were prepared as previously described.^29,53 Briefly, an MNA containing a total of 77 microneedles in staggered rows (base diameter, 250 μm; spacing between rows, 300 μm; spacing within rows, 1200 μm; height, 650 μm) was replicated using soft lithography. Negative templates were formed using poly(dimethylsiloxane) (Sylgard 184, Dow-Corning). Poly(l-lactide) (PLLA) was then melted through a phase transition in the PDMS molds under vacuum (−25 in. Hg, 200°C, 90 min), then cooled to −20°C for 30 min before separating the cast PLLA MNAs from the PDMS mold. MNAs were then washed in distilled water, followed by 70% ethanol.⁵⁶ MNAs were imaged on a using a JEOL 6700F FEG-SEM scanning electron microscope to confirm their morphology (Figure S1A).⁵³ To characterize the mechanical properties (stiffness and fracture force) of the MNAs, a Dynamic Mechanical Analyzer (DMA) TA Q800 was used in the static force mode using a compression clamp. In brief, the MN array was fixed on the lower plate, and the upper plate was moved toward the MN array with a ramp force of 1 Newton (N) per minute until a force of 8N was reached. The force required to penetrate skin was calculated based on the required force for all 77 projections with this geometry to effectively penetrate skin (Figure S1B).⁵⁷ Following preparation, MNAs were sterilized with ethylene oxide before use in subsequent mouse studies.

MNA incubation with Y. pestis

To load MNAs with bacteria, MNAs were incubated with Y. pestis as previously described.²⁹ Briefly, Y. pestis was diluted to desired concentrations (CFU/mL) in BHI. MNAs were fully submerged in 1 mL of culture in a 24 well-plate and incubated for 30 min with rocking at room temperature. Any bubbles from the culture were gently removed from the MNAs prior to incubation using a pipet tip. Prior to application to the ear, excess medium was removed from the MNAs using a sterile cotton swab.

Animal infections with Y. pestis

Mice were challenged with Y. pestis as previously described.^37,54,55 For MNA inoculation, mice were anesthetized with isoflurane and the MNA was applied to the left ear. MNA was pressed on the ear for 5 s and then removed with forceps. For intranasal challenge, mice were anesthetized with ketamine/xylazene and administered 20 μL of bacteria suspended in 1X PBS to the left nare. For optical imaging, mice were anesthetized with isoflurane and imaged using the IVIS Spectrum imaging system (Caliper Life Sciences, Hopkinton, MA). Average radiance (photons/sec/cm2) was calculated for ear and lymph nodes as previously described.³⁷ Infected mice were monitored for the development of disease symptoms twice daily for 14 days. Moribund animals meeting predefined endpoint criteria were humanely euthanized by CO₂ asphyxiation and scored as succumbing to infection 12 h later.

Measurement of bacterial burden in the ear and lymph node

Bacterial numbers in the ear dermal layer and lymph nodes were determined by conventional bacterial enumeration using serial dilutions. Ears were wiped with 70% ethanol prior to remove viable skin-resident bacteria. All tissues were homogenized in sterile PBS using a tissue homogenizer (Omni). 10-fold serial dilutions were made in PBS, and 10 μL were plated in triplicate onto Difco BHI agar (BD Biosciences). Colony forming units (CFU) were counted after incubation at 26°C for 48 h.

Imaging of ear with two-photon microscopy

Mouse ears were collected immediately after MNA inoculation Y. pestis. Ears, with the MNA attached, were placed into tissue cassettes with sponges to maintain the MNAs within the dermis. Tissues were incubated in 10% formalin for 24 h. Following fixation, tissue cassettes with ears were washed once with 1x PBS and stored in 1x PBS until imaged by two-photon microscopy. For imaging, MNAs were removed and bacteria within the ears were stained by incubation with anti-Y. pestis serum (1:1,000) and Alexa Fuor546 Donkey anti-Rabbit IgG (H + L) (1:500; Invitrogen Cat. No. A10040). Tissues were incubated with antibodies overnight in FBS. Mouse ear structures and immune cells were visualized by autofluorescence in the GFP channel.

Quantification and statistical analysis

Animal experiments were repeated twice to confirm reproducibility and the data is represented as the combination of the two independent experiments. p values were calculated using Student’s t test, one-way analysis of variance (ANOVA) with appropriate post hoc testing as indicated in figure legends. Groups size, definition of center, and dispersion and precision measures are also noted in the figure legends as appropriate. All statistics were completed using GraphPad Prism software.

Published: November 30, 2023