Protection against lethal sepsis following immunization with Candida species varies by isolate and inversely correlates with bone marrow tissue damage

ABSTRACT Protection against lethal Candida albicans (Ca)/Staphylococcus aureus (Sa) intra-abdominal infection (IAI)-mediated sepsis can be achieved by a novel form of trained innate immunity (TII) involving Gr-1+ myeloid-derived suppressor cells (MDSCs) that are induced by inoculation (immunization) with low virulence Candida species [i.e., Candida dubliniensis (Cd)] that infiltrate the bone marrow (BM). In contrast, more virulent Candida species (i.e., C. albicans), even at sub-lethal inocula, fail to induce similar levels of protection. The purpose of the present study was to test the hypothesis that the level of TII-mediated protection induced by Ca strains inversely correlates with damage in the BM as a reflection of virulence. Mice were immunized by intraperitoneal inoculation with several parental and mutant strains of C. albicans deficient in virulence factors (hyphal formation and candidalysin production), followed by an intraperitoneal Ca/Sa challenge 14 d later and monitored for sepsis and mortality. Whole femur bones were collected 24 h and 13 d after immunization and assessed for BM tissue/cellular damage via ferroptosis and histology. While immunization with standard but not sub-lethal inocula of most wild-type C. albicans strains resulted in considerable mortality, protection against lethal Ca/Sa IAI challenge varied by strain was usually less than that for C. dubliniensis, with no differences observed between parental and corresponding mutants. Finally, levels of protection afforded by the Ca strains were inversely correlated with BM tissue damage (R 2 = −0.773). TII-mediated protection against lethal Ca/Sa sepsis induced by Candida strain immunization inversely correlates with BM tissue/cellular damage as a reflection of localized virulence.

Our laboratory has been studying Ca/Sa polymicrobial IAI using an experimental mouse model that results in 80-90% mortality by 48 to 72 h post-inoculation (22)(23)(24).Characterization of host responses during Ca/Sa polymicrobial IAI revealed that mortality is associated with robust inflammation, with elevated levels of hallmark sepsis proinflam matory cytokines (IL-6, TNF-α, and IL-1β), both locally and systemically, as early as 4 h and continuing through 24 to 48 h post-inoculation.On the other hand, at similar time points there were equivalent microbial burdens in non-lethal monomicrobial and lethal polymicrobial infections, both locally in the peritoneal cavity, as well as in adjacent (spleen and kidney) and non-adjacent organs (brain), indicating that robust inflamma tion (sepsis) is the key factor in the lethal outcome (24).
Subsequent studies investigated whether all Candida species that cause human disease also lead to synergistic mortality during polymicrobial infection.IAI inoculation with Sa and a variety of non-albicans Candida (NAC) species resulted in varying levels of mortality.Co-infections with Candida glabrata (Cg) or Candida dubliniensis (Cd) and Sa result in no mortality, whereas co-infections with Candida krusei (Ck) or Candida tropicalis (Ct) and Sa lead to 80-90% mortality (25).As with Ca, monomicrobial infections with the majority of these Candida species or Sa alone were not lethal out to 5-10 d post-inoculation (23).
We have also shown that protection against Ca/Sa lethal sepsis can be achieved by prior immunization [intraperitoneal (i.p.) or i.v.] with low virulence Candida species, including Cd, Cg, Candida auris, and Ca efg1Δ/Δ cph1Δ/Δ mutant (26).In each case, immunization followed by a lethal Ca/Sa challenge 14 d later resulted in >80% survival.Protection was long-lived (up to 60 d post-challenge) and mediated primarily by Gr-1+ polymorphonuclear leukocytes of the innate response, rather than adaptive immunity (25,26).This innate-mediated protection was suggestive of a form of trained innate immunity (TII).TII was first described in macrophages "trained" by epigenetic reprog ramming leading to enhanced responsiveness to secondary infection (27).Subsequent studies in our laboratory showed that the Gr-1+ cells mediating protection were long-lived myeloid-derived suppressor cells (MDSCs) (28) which have been reported in other models of sepsis and in patients with candidiasis (29,30).MDSCs are induced and expanded in the bone marrow (BM) (31)(32)(33).We have further shown that the Candida species used for the primary immunization access the BM within 24 h consistent as a requirement for MDSC development (26).
Interestingly, immunization with more virulent Candida species (Ck, Ct, and Ca), even at lower inocula that spared any initial mortality over a 14-d period, resulted in much lower protection against lethal sepsis compared to the low virulent species (40-50% vs 80-100%, respectively), despite similar infiltration into the BM (25,26).We hypothesize that the induction of the protective MDSC response is inversely proportional to the level of tissue damage caused by the Candida species in the BM.The purpose of the present study was to test this hypothesis by comparing protection induced by a variety of Ca isolates with varying levels of reported virulence in other models of infection (24,(34)(35)(36)(37)(38)(39).Accordingly, we employed several Ca strains that were clinical isolates (529L) or mutant strains deficient in one or more virulence factors (TNRG1, efg1Δ/Δ cph1Δ/Δ, and ece1Δ/Δ) together with their isogenic parent strains (TT21, DAY185, and BWPI7 + CIp30), to correlate their ability to protect against lethal sepsis, with immunization-associated tissue damage in the BM.

Mice
Female Swiss Webster mice, 5 to 7 weeks of age, were purchased from Charles River Laboratories.Animals were housed and handled according to institutionally recom mended guidelines.All experiments involving animals were approved by the Tulane University School of Medicine Institutional Animal Care and Use Committee (IACUC).

Strains and growth conditions
C. albicans strain DAY185, a complemented prototroph derived from a triple auxotrophic strain (BWP17; parent, SC5314) was a gift from Dr. Aaron Mitchell (Carnegie Melon University, Pittsburgh, PA).The C. dubliniensis wild-type strain (Wü284) was kindly provided by Dr. Gary Moran (Trinity College, Dublin, Ireland).Yeast-locked strain (TNRG1), the corresponding wild-type strain (TT21), as well as C. albicans mutant efg1Δ/Δ cph1Δ/Δ were provided by Glen Palmer (University of Tennessee Health Science Center, Memphis, TN).The clinical oral isolate, 529L, as well as the ece1Δ/Δ mutant and corresponding wild-type strain (BWPI7 + CIp30) were generously donated by Julian Naglik (King's College London, UK).All frozen stocks were maintained at −80°C and streaked onto yeast peptone dextrose (YPD) agar prior to use.A single colony was transferred to 10 mL of YPD broth and shaken at 30°C for 12-18 h.The methicillin-resistant Sa strain NRS383 used was obtained from the Network on Antimicrobial Resistance in Sa (NARSA) data bank.Frozen stocks were maintained at −80°C and streaked onto Trypticase soy agar (TSA) prior to use.A single colony was transferred to 10 mL of Trypticase soy broth (TSB) and shaken at 37°C overnight.On the following day, the overnight culture was diluted 1:100 in fresh TSB and shaken at 37°C for 3 h until the culture reached the log phase of growth.Prior to inoculation, organisms were washed three times by centrifugation in sterile phosphatebuffered saline (PBS; pH 7.4), counted on a hemocytometer, and diluted in sterile PBS to prepare standardized inocula.To visualize cells, Sa was stained with LIVE/DEAD BacLight Bacterial Viability Kit (Invitrogen, Waltham, MA) 5 min prior to counting.
(ii) Lethal Challenge.Mice were injected i.p. with a lethal challenge of Ca (1.75 × 10 7 /mouse) and Sa (8 × 10 7 /mouse) in a volume of 200 µL of sterile PBS and observed for morbidity (hunched posture, inactivity, and ruffled fur) and mortality up to 10 d after re-challenge.Mice who reached clinical endpoints prior to the study endpoint were humanely euthanized following IACUC-approved euthanasia procedures.Of note, previous studies evaluating sex as a biological variable, using C. dubliniensis as the immunizing strain together with Ca/Sa lethal challenge, showed similar levels of protection against lethal challenge and median day of death in lethal challenge controls in males and females (Noverr, unpublished data).Hence, female mice were used exclusively in the present study.

Histological analysis of whole femur for evidence of tissue damage
(i) Bone architecture by Periodic Acid-Schiff (PAS).One or 13 d after immunization, mice (n = 5/group) were euthanized and whole femurs were excised (40).Bones were cleaned of all soft tissue, placed in tissue cassettes, and immersed in 10% neutral buffered formalin.After 7 d, selected bones (24 h post-immunization) were sent for histological preparation (paraffinembedding, sectioning, deparaffinization, and rehydration) and stained using the standard Periodic AcidSchiff staining technique (GNO Histology Consultants, LLC, Marrero, LA).
(ii) Ferroptosis by 4-HNE staining.Unstained 24 h or 13 d post-immunization bone sections underwent histological preparation as described above.4-HNE tissue staining was performed as described previously (41).Briefly, antigen retrieval was immediately performed using a citrate-based Antigen Unmasking Solution (Vector Laboratories Inc., Burlingame, CA) heated to 95°C.Slides were placed in the pre-heated solution for 2 min and then rinsed with distilled water followed by PBS.Sections were blocked using 1% BSA in PBST (PBS + 0.1% Tween-20) for 30 min at room temperature.Excess liquid was drained, and sections were incubated with goat anti-mouse 4-HNE antibody (ab46545; Abcam, Boston, MA) overnight at 4°C in a humidified chamber.Slides were washed three times in PBS (5 min/wash) followed by incubation with conjugated donkey anti-goat IgG secondary antibody (Alexa Fluor 488; Abcam) for 1 h at room temperature in the dark.Slides were washed three times in PBS (5 min/wash) followed by mounting/coun terstaining with DAPI-Aqueous Mounting Media with Fluoroshield (Abcam).All tissue images were captured with the Lionheart FX Automated Microscope (BioTek Instruments, Winooski, VT) using the same image acquisition settings for each slide.Mean fluores cence intensity of each image was calculated using Gen5 microscopy and imaging software (BioTek).

Statistics
All statistical analyses were performed using Prism software (GraphPad, San Diego, CA).Survival curves were compared using the log-rank (Mantel-Cox) test.Mean fluorescence intensities of 4-HNE were analyzed by Student's t test.The Pearson product-moment correlation coefficient (R 2 , P value) was calculated to determine correlations.Significance was defined as P < 0.05.

Protection against lethal challenge (Ca/Sa) by C. albicans immunization varies by isolate
We next tested the protective potential of the various wild-type C. albicans isolates against lethal IAI challenge.Following the 14-day immunization period with either standard or sub-lethal inocula, surviving mice were injected i.p. with a lethal challenge of Ca (1.75 × 10 7 /mouse) and Sa (8 × 10 7 /mouse) and observed for morbidity (hunched posture, inactivity, and ruffled fur) and mortality over a period of 10 d.Immunization with Cd was used as a positive control for strong protection.Compared to protection elicited by Cd immunization (standard or sub-lethal inocula), immunization with all wildtype Ca isolates resulted in lower levels of protection, with sub-lethal inocula generally resulting in lower levels of protection than that observed in surviving mice given the standard inocula (20-60% vs 40-80%; Fig. 2A and B).Despite these differences, all strains induced significant protection compared to lethal challenge control, with the exception of C. albicans 529L at the sub-lethal inocula immunization.Interestingly, C. albicans 529L exhibited the lowest level of protection at both the standard and sub-lethal inocula immunizations (40% and 20%, respectively).No significant differences in levels of protection were detected between the various isolates, both at the standard or sublethal inocula.

C. albicans mutant strain defects have little effect on protection against lethal challenge (Ca/Sa)
We next compared the relative levels of protection between each wild-type and isogenic mutant strain at both standard and sub-lethal immunizing inocula.For immunization with wild-type Ca TT21 and yeast-lock mutant TNRG1, a range of partial protection (30-70%) was observed at either inocula (Fig. 3A) with no significant differences observed between wild-type and mutant or between standard or sub-lethal inocula of each isolate.In the case of immunization with wild-type parental Ca DAY185 and corresponding efg1Δ/Δ cph1Δ/Δ mutant, the standard inocula conferred 80% and 90% protection, respectively, while immunization with the sub-lethal inocula resulted in lower levels of protection (50% and 20%, respectively) but no significant differences between wild-type and mutant or between the wild-type at standard or sub-lethal inocula.A significantly lower level of protection was observed for the mutant strain at the sub-lethal compared to the standard inocula (P = 0.0003) (Fig. 3B).Similarly, for immunization with either Ca wild-type strain BWP17 + Clp30 or the corresponding mutant ece1Δ/Δ, standard inocula immunization with either isolate resulted in 80% protection, while sub-lethal inocula resulted in lower levels of protection (40% and 30%, respectively) but with no significant differences between wild-type and mutant or between the wild-type at standard or sublethal inocula.A significantly lower level of protection was observed for the mutant strain at the sub-lethal compared to the standard inocula (P = 0.030) (Fig. 3C).Included in these experiments were an additional set of mice (n = 4/group) given standard or sub-lethal inoculations of wild-type/parental strains (C.albicans DAY185; BWP17 + Clp30), including the isogenic mutant strains (C.albicans efg1Δ/Δ cph1Δ/Δ; ece1Δ/Δ), to confirm bone marrow infiltration at 24 h post-immunization.While there were consistently lower levels of infiltration with sub-lethal inocula, there were no statistically significant differences between standard and sub-lethal inocula for each isolate or between isolates, with the exception of standard vs sub-lethal inocula for C. albicans DAY185 (P = 0.015) (Fig. S1).

Ferroptosis in bone marrow post-immunization with Candida strains as a measure of tissue damage
As ferroptosis can serve as a quantitative measure of tissue damage (41,54,55), following both standard and sub-lethal inocula immunization of the various wild-type or mutant isolates, mice were sacrificed at various times post-immunization and femurs excised and processed for BM analysis.Naïve mice, mice immunized with the standard inocula of Cd or Ca efg1Δ/Δ cph1Δ/Δ mutant, or mice given lethal challenge were included as controls.BM tissue sections were stained with anti-4-HNE antibody (green) followed by DAPI cellular staining (blue).Representative images of BM collected 24 h post-immunization with sub-lethal inocula of the various Ca isolates and standard inocula of the control isolates (C.albicans efg1Δ/Δ cph1Δ/Δ and C. dubliniensis wild-type strain Wü284) are shown in Fig. 4A.Gross visual assessment reveals that BM sections from naïve mice and mice immunized with the standard inocula of Cd or Ca efg1Δ/Δ cph1Δ/Δ mutant have weak 4-HNE staining, while BM sections from mice given the lethal challenge have obvious pervasive green 4-HNE staining.Levels of 4-HNE staining in BM sections of mice immunized with the sub-lethal inocula of the cadre of other Ca wild-type or mutant strains were variable but seemingly inversely correlated with the average level of protection following a lethal challenge (shown in parentheses); BM sections from mice immunized with Ca yeast-lock strain TNRG1 and wild-type Ca DAY185 had weak positive 4-HNE staining (60-70% protection) followed by stronger levels for all immunizing strains that did not result in appreciable protection (TT21, 529L, ece1Δ/Δ, BWP17 + Clp30; 30- 40% protection).The levels of 4-HNE staining were then quantified by calculating green fluorescence intensity in a set of 10 images/inoculation group) and plotted against the average percentage of protection from lethal challenge for each immunization strain.Results in Fig. 4B reveal a significant negative correlation between fluorescence intensity (ferroptosis) and average protection against lethal challenge (R 2 = −0.773;P = 0.0008).
Representative images of 4-HNE staining for BM sections of mice comparing standard vs sub-lethal inocula for several of the Ca isolates (including isogenic parental and mutant pairs) at 24 h post-immunization and also at 13 d following sub-lethal inocula immunization (just prior to lethal challenge) are shown in Fig. 4C.Results show relatively similar 4-HNE staining between standard and sub-lethal inocula for the various isolates.By day 13 post-immunization, 4-HNE staining was relatively weak for all isolates.The quantitative analysis of the entire dataset is shown in Fig. 4D and E. BM sections from naïve mice and mice immunized with the standard inocula of Cd and Ca efg1Δ/Δ cph1Δ/Δ had the lowest levels of fluorescence intensity compared to those from the lethal challenge only control (P < 0.0001).Of the remaining Ca strains used for immu nization, with the exception of TNRG1, all sections had significantly higher levels of fluorescence intensity compared to those from mice immunized with Cd (P < 0.0001).Most of the 4-HNE fluorescence intensity in bone sections from mice immunized with the various Ca strains were not significantly different from mice given the lethal challenge alone.The exception was Ca TNRG1 (yeast-lock mutant; P = 0.0037).Comparisons between 24 h and 13 d post-immunization show that fluorescent intensities (ferroptosis) were significantly reduced at day 13 among the various Ca strains, with the exception of Ca TNRG1, which was already low at 24 h (Fig. 4D).Comparisons between the standard and sub-lethal inocula 24 h post-immunization showed no significant differences for any isolate (Fig. 4E), in support of the similar levels of modest protection against lethal challenge between surviving mice given the standard inocula immunization and those given the sub-lethal inocula immunization (Fig. 3).

Bone marrow architecture post-immunization with Candida strains as additional support of tissue damage
As another assessment of tissue damage in the BM following immunization, femurs were excised and from mice 24 h after sub-lethal inocula immunization with the various isolates and processed for Periodic AcidSchiff staining of BM.Naïve mice or mice given lethal challenge were included as controls.Qualitative scoring of negative space (due to reduced cellularity or cell loss) in the BM histological sections was recorded (0 to ++++).Representative images of the histological sections and associated negative space scoring are shown in Fig. 5. Compared to dense bone histology in naïve mice (0) and mice immunized with Cd (+), lethal challenge results in significant negative space (++++).Mice immunized with the cadre of Ca wild-type or mutant strains show varying levels of bone marrow density (++ and +++), with a notable negative correlation to the average level of protection achieved following lethal challenge (noted in parentheses); femur bones that had the larger areas of negative space were from mice immunized with strains inducing the least protection.Efforts made to visualize Candida yeast/hyphae in the histological sections were not informative due to low levels of infiltrating organisms.Interestingly though, in other experiments designed to track infiltrating organisms in the bone marrow, we showed that GFP-labeled efg1Δ/Δ cph1Δ/Δ, albeit minimal numbers/field, were consistently shown as taken up within leukocytes (presumably macrophages) (data not shown).

DISCUSSION
Candida species have an array of virulence traits that manifest during infections at a variety of anatomical sites (14,(56)(57)(58).This holds true as well for IAI leading to fungal/bacterial lethal sepsis.In the animal model of IAI, co-challenge of S. aureus with C. albicans, C. kruseii, and C. tropicalis all result in a lethal sepsis outcome, whereas co-challenge with C. dubliniensis, C. glabrata, and C. auris fail to promote lethal sepsis (23,25,26).Similar variability in virulence outcomes can occur for different isolates of the same species.For example, C. albicans wild-type strains 529L and SC5314/DAY185 are not optimal colonizers of mucosal tissues (36,37,(45)(46)(47)(48). Organ tropism can also vary between species or isolates of one species.Overall, there is a growing awareness that isolates of the same species can act quite differently in model systems, whether they be laboratory or clinical isolates.Hence, it cannot be assumed that "one strain fits all." Rather, various strains and isolates of a particular species need to be defined by how they function/act at different anatomical sites or in different model systems.
This concept was clearly observed in the IAI model early on when we found that the virulence trait of yeast to hyphal transition for C. albicans was not critical to the lethal sepsis outcome when co-challenged with S. aureus (24).Conversely, protection against lethal fungal/bacterial sepsis by prior immunization with live Candida species is heavily dependent on virulence; low virulent ("live attenuated") Candida species (C.dubliniensis, C. glabrata, and C. auris) are better inducers of the protective response than more virulent Candida species that are usually non-lethal as a monomicrobial insult during the normal IAI observation period (5-10 d) (25,26).This is true despite hyphal formation not serving as a key factor in the protective response; C. dubliniensis that forms true hyphae is as good at inducing the protection response as low virulent species that fail to form hyphae (C.glabrata and C. auris) (25,26).This peculiar observation, together with the recognized varied virulence between various C. albicans isolates in other model systems (24,(34)(35)(36)(37)(38)(39), prompted the use of the various Ca strains/isolates, some as wild-type/mutant pairs, to characterize their level of virulence in the IAI model as well as their ability to induce the protective response against lethal sepsis.We hypothesized that the level of host damage caused by the isolates would be a key factor in how well a particular isolate induced the protective response.Since the bone marrow is the site for the induction of the protective response (26,59) and the fact that the intraperitoneal inoculation of fungal species infiltrates the bone marrow within 24 h (26), we focused on tissue damage in the bone marrow for correlates of protection.While immunization with any of the low virulence isolates is normally performed with the standard inocula used for lethal challenge, recognizing the potential for some late-stage mortality with several C. albicans isolates at the standard inocula, we also incorporated a sub-lethal inocula (17.5 × reduced − 1 × 10 6 CFUs), that is known to result in similar levels of protection by C. dubliniensis (25).
The results overall were interesting and enlightening.As expected, sub-lethal inocula had little to no mortality for all the isolates/strains.Conversely, for the standard inocula some mortality was observed for several C. albicans isolates over the 14-d period.In most cases, the mortality began after 6 d with the majority observed after 8 d.Interestingly, while no statistical significance was noted between isolates, the highest mortality was with DAY185 and both the wild-type parental and ece1 mutant strains.The next highest mortality was observed with 529L.Clearly, lethal sepsis from IAI-associated dissemination is not dependent on candidalysin (ece1) and does not discriminate against isolates that are sub-optimal for mucosal colonization.
In the case of the protection studies, immunizing with the cadre of wild-type C. albicans isolates (surviving mice from standard inocula, and all mice given sub-lethal inocula) resulted in a range of protection that was not statistically significant between isolates, but most often lower than that induced by C. dubliniensis.This was true for immunization using both the standard and sub-lethal inocula.Moreover, protection induced by corresponding mutant strains was very similar to the wild-type parental strain.This was not all that surprising recognizing that the defects often included hyphal transformation which is a non-factor in the induction of the TII response (24).Hence, the "low or avirulent" definition for these mutants does not apply to the TII response and remains inferior to C. dubliniensis.The exception was C. albicans efg1Δ/Δ cph1Δ/Δ that lacks two transcription factors crucial for morphogenesis and other pathways required for virulence and is avirulent in all model infections tested to date, including the IAI sepsis model (38,(49)(50)(51)(52). Accordingly, it induced protection similar to C. dubliniensis.In terms of inocula and level of protection, there appeared to be an effect of inocula when visually comparing protection induced by isolates at standard (surviving mice) vs sub-lethal inocula (Fig. 2).When isolates (wild-type or mutant) were compared directly (Fig. 3), differences were identified for two mutant isolates, efg1Δ/Δ cph1Δ/Δ and ece1Δ/Δ.In these cases, the standard immunizing inocula induced better protection than the corresponding sub-lethal inocula despite similar levels of infiltration into the BM (Fig. S1).The differences for efg1Δ/Δ cph1Δ/Δ can be explained by very weak virulence or strong host antifungal activity resulting in clearance prior to inducing the TII response, despite being present at similar levels to the parental strain in the BM at 24 h.However, this explanation cannot be used for ece1Δ/Δ, which is relatively more virulent in the IAI model.Further studies will be required to assess this interesting dichotomy particularly for downstream effects on epigenetic remodeling.
The inverse correlation of protection to BM tissue damage was borne from the observation that the flushed bone marrow in non-protected mice appeared very red compared to that from protected mice, suggestive of blood vessel damage and/or RBC lysis.Initial evaluations of optical density, lactate dehydrogenase (LDH), and hemoglobin concentrations in the BM samples collected from femurs post-immunization, however, were inconclusive.Conversely, BM histology evaluating ferroptosis by 4-HNE staining (41,54,55,60) and negative space indicative of problems with hematopoietic stem cell function (61)(62)(63)(64) were stronger indicators of tissue damage.Indeed, BM from naïve mice and C. dubliniensis-immunized mice with the highest protection (controls) had low 4-HNE staining and dense histology.In contrast, BM from lethal challenge mice and mice given C. albicans-immunizing isolates resulting in low levels of protection had higher 4-HNE staining and reduced cellularity (negative space) at 24 h post-inoculation.Intermediate levels of 4-HNE staining and cellularity were observed for BM from mice given C. albicans isolates that resulted in higher levels of protection.Quantitative analysis of ferroptosis confirmed the qualitative assessment, with significant differences observed between isolates associated with high (80-90%) vs low (40-60%) levels of protection and a highly significant negative correlation (R 2 = −0.773) between ferroptosis and average levels of protection for the various isolates.Some additional interesting data included similar levels of 4-HNE staining at 24 h between standard and sub-lethal inocula of each isolate that was supported by the similar levels of infiltration into the BM at 24 h (Fig. S1), and relative resolution of tissue damage (low 4-HNE staining) by day 13 post-immunization for all the isolates.These results are in support of the early survival data for immuniza tion with each isolate/inocula and suggest that the tissue damage in surviving mice is not permanent and can ultimately resolve for effective induction of the protective TII response in the BM and/or release/activation of the MDSCs for ultimate function.
An important aspect of these results is speculating on what mechanistically is causing the tissue damage in the BM.One might assume the damage is due to the organism based on the overall virulence attributes of C. albicans vs C. dubliniensis and highly avirulent C. albicans mutants.However, we hypothesize that the damage is primarily due to the host response in the BM based on several pieces of information from this study.Firstly, since relative damage is not dependent on the strong virulence attributes of hyphal formation or the presence of candidalysin, organism-mediated damage is unlikely.Secondly, the lack of any correlation of tissue damage to fungal burden in the BM would not favor organism-mediated damage.Conversely, studies that were designed to track C. albicans yeast form in the BM showed the majority as intracellular in leukocytes (possibly macrophages).Hence, it is more likely that the host response to the more virulent C. albicans is mediating the majority of damage.Future studies will test this hypothesis by evaluating cytokines and immune mediators in the BM following inoculation with the various isolates similar to what we recently reported for immuniza tion with C. dubliniensis (28).In addition, a comprehensive analysis of epigenetic changes associated with the TII response in MDSCs along with correlates to Candida isolate-asso ciated tissue damage are also planned.
In conclusion, our results provide strong support for the concept that levels of protection against lethal sepsis by the MDSC TII response are inversely proportional to the local tissue damage in the BM induced by the immunizing live species/isolate.The action of the various Ca isolates in this model revealed definitive caveats that were not always consistent with reported virulence but nonetheless were consistent with the properties of the model, both for lethal sepsis and for protection against.Moreover, while this immunization may be a potentially exciting vaccine strategy, even the use of low virulence (live-attenuated) fungal strains to immunize against polymicrobial sepsis poses significant safety concerns.However, this may be circumvented based on our recent data revealing that protection against lethal sepsis can be equally achieved by intraperitoneal immunization with abiotic cell wall components of C. albicans (i.e., modified β-glucan or depleted zymosan) (65).Current studies are evaluating this abiotic vaccine strategy by monitoring infiltration of such components into the BM and any correlates of tissue damage with the induction and function of the protective MDSC-mediated TII response.

FIG 2
FIG 2 Protection against lethal challenge (Ca/Sa) following immunization with wild-type C. albicans varies by immunization strain.Mice (n = 10/group) were given a standard (1.75 × 10 7 ) or sub-lethal (1 × 10 6 ) intraperitoneal inoculation (immunization) of C. albicans wild-type strains (DAY185, BWP17 + clP30, TT21, and 529L) followed 14 d later by lethal challenge (1.75 × 10 7 Ca + 8 × 10 7 Sa).Mice were monitored for sepsis and mortality for 10 d post-lethal challenge.(A) Percentage survival of mice given the standard inocula immunization (B) Percentage survival of mice given the sub-lethal inocula immunization.C. dubliniensis wild-type strain (Wü284) at the standard/sub-lethal inocula was included as a positive control for protection.Animals receiving no immunization served as the negative (lethal) control.Results shown are cumulative of four independent experiments.Data were analyzed using the log-rank (Mantel-Cox) test.****P < 0.0001; ***P < 0.001; **P < 0.01; and *P < 0.05 (significance as compared to lethal challenge control).Actual P values are listed in tables.WT, wild-type

FIG 4 (
FIG 4 (Continued) sub-lethal inocula.Femur bones were collected and processed for 4-HNE staining.BM tissue-associated 4-HNE is shown in green, and DAPI counterstain is shown in blue.Representative images of two repeats are shown at 40× magnification.Scale bar 100 µm.Images shown for 24 h sub-lethal immunization are the same as shown in frame A. (D) Quantification of bone marrow tissue damage (ferroptosis) 24 h and 13 d post-immunization: Tissue/cellular damage levels were assessed from tissues taken at 24 h and 13 d post-immunization by calculating 4-HNE staining (green fluorescence intensity) in each image (n = 10 images/inoculation group).Results are expressed as mean fluorescence intensity.Actual P values are listed in tables.Percentages shown in parentheses are average percentage of protection/survival following immunization and lethal challenge.(E) Quantification of bone marrow tissue damage (ferroptosis) 24 h post-immunization comparing standard vs sub-lethal inocula.Results are expressed as 4-HNE mean fluorescence intensity in each image (n = 10 images/inoculation group).Actual P values are listed in tables.Percentages shown in parentheses are average percentage of protection/survival following immunization and lethal challenge.WT, wild-type.

FIG 5
FIG 5 Bone marrow architecture 24 h post-immunization with Candida strains as additional support of tissue damage.Mice (n = 5) were injected i.p. with the standard inocula (1.75 × 10 7 ; C. albicans efg1Δ/Δ cph1Δ/Δ, C. dubliniensis wild-type strain Wü284) or the sub-lethal inocula (1 × 10 6 ; DAY185, SC5314, BWP17 + clP30, TT21, 529L, ece1Δ/Δ, and TNRG1).Animals receiving no inoculation (naïve) and lethal challenge only (1.75 × 10 7 Ca + 8 × 10 7 Sa) served as controls.Mice were sacrificed 24 h after inoculation for the collection of femur bones.Bones were fixed and processed for Periodic AcidSchiff staining.Images shown were captured at 10× magnification and are representative of five femur bones/inoculation groups from two repeats.Qualitative scoring of negative space (+ to ++++ ) was conducted for all fields.Scoring is provided for each image as an overall mean score.Percentages shown in parentheses are corresponding percentage of protection/survival following immunization and lethal challenge in prior experiments.WT, wild-type.Scale bar 200 µm.