Original Research Article

Vaccination with recombinant Streptococcus uberis Adhesion Molecule alters immune response to experimental challenge

Dr Gina Pighetti,

Gina M Pighetti1*, Leszek Wojakiewicz1, Susan I Headrick1, Oudessa Kerro Dego1, Sierra A. Lockwood1, Henry G. Kattesh1, Mark J. Lewis2, Charles D. Young3, Lydia J. Siebert1, Barbara E. Gillespie1, Maria E. Prado1, Raul A. Almeida1 and Stephen P. Oliver1
Department of Animal Science, Institute of Agriculture, The University of Tennessee, Knoxville, TN, USA. Tennessee Dairy Herd Information Association, Knoxville, TN, USA. Zoetis, Florham Park, NJ.

Streptococcus uberis Adhesion Molecule (SUAM) enables S. uberis infection of the bovine mammary gland. Our objective was to evaluate responses to experimental intramammary challenge with S. uberis in Holstein cows following vaccination with recombinant(r) SUAM.

*Corresponding author:

Gina M Pighetti
pighetti@utk.edu

Keywords:

Mastitis, Dairy Cattle, Vaccine, Cytokines

Streptococcus uberis Adhesion Molecule (SUAM) enables S. uberis infection of the bovine mammary gland. Our objective was to evaluate responses to experimental intramammary challenge with S. uberis in Holstein cows following vaccination with recombinant(r) SUAM. Cows were immunized with rSUAM (n=20) or saline (n=20) at 84, 56, and 28 d prior to expected calving and challenged with in 3d of calving. Samples were collected to evaluate S. uberis concentration, clinical signs of inflammation, somatic cell count (SCC), antibody titer, SCCcytokine RNA, and cortisol level. Vaccination with rSUAM significantly increased rSUAM specific sera IgG and colostrum IgM, IgG1, IgG2, and IgA. Greater S. uberis concentration in milk, milk inflammation, and blood cortisol was observed 1.5 to 3 d post-challengein the rSUAM versus saline group. Where as, rectal temperature was lower. IL-1 and IL-17 RNA was greater prior to challenge in the rSUAM group. The rSUAM group also tended towards greater IL-1 and IL-10 overall and significantly lower IFN-γ RNA 2.5 d after challenge. SCC, mammary inflammation score, IL-4 and IL-6 RNA, milk production, and days to return to 200,000 SCC/ ml were similar. These results suggest that vaccination with rSUAM caused shifts in immunity but did not influence overall protection.

Mastitis, an inflammation of the mammary gland that affects a high proportion of dairy cows throughout the world, has been described as the most economically imposing disease facing dairy producers. As control of contagious pathogens has become more effective, the frequency of environmental mastitis pathogen such as Streptococcus uberis, Escherichia coli, and Klebsiella species have become more predominant. In a field survey of 9 dairy herds with low somatic cell count (SCC), an indicator of immune cell migration to the mammary gland, Staphylococcus aureus accounted for <2% of all clinical cases (n=646), whereas environmental streptococci and coliforms constituted 25 and 30% of cases, respectively [1]. One of the most preva- lent causes of clinical mastitis is Streptococcus uberis. This organism has been recovered from 6-18% of clinical cases representing herds from across the world [2-6]. As such, environmental streptococci, particularly S. uberis, represent relatively common pathogens - especially in herds that have contagious pathogens under control. Developing strategies to help the host immune system better resist, tolerate, and eliminate environmental organisms are needed because it is not feasible to isolate the cow from her environment to minimize the risk of exposure.

A bacterial protein referred to as Streptococcus uberis Adhesion Molecule (SuAM) represents a novel virulence factor that enables S. uberis establishment, spread, and persistence in the mammary gland. The fibrillar SUAM protein located on the surface of S. uberis has been demon- strated to bind lactoferrin, a protein found in bovine milk and mammary secretions [7,8]. This interaction creates a molecular bridge that enhances S. uberis adherence to and internalization into mammary epithelial cells most likely via caveolae-dependent endocytosis and potentially allows S. uberis to evade host defense mechanisms [7,8]. A subsequent study created a sua-negative mutant of S. uberis that was defective in adherence to and internalization into mam- mary epithelial cells indicating this is one of the key genes involved with S. uberis binding and internalization [9]. The potential for rSUAM as an antigenic candidate for vaccine-based control of S. uberis mastitis was strengthened when immunization of dairy cows with recombinant (r) SUAM led to significant increases in SUAM-specific antibodies that were capable of reducing S. uberis adherence to and internalization into mammary epithelial cells in vitro [10].

The overall objective of the present study was to evaluate the effectiveness of a rSUAM-based vaccine in protecting cows and also to determine the strength and type of immune response elicited following experimental intramammary challenge with S. uberis. We elected to conduct this experiment during a time when cows are most susceptible to this organism, e.g. the last two months of gestation (non-lactating) and early lactation, in addition to having the greatest concentration of antibodies present. Based on the ability of antibodies against rSUAM to re- duce S. uberis adhesion and internalization in vitro, we hypothesized that vaccination with rSUAM would reduce the number of infected cows and severity of infections following experimental challenge.

Pregnant Holstein cows (n=20 per treatment) near the end of their 1st or 2nd lactation were purchased from an outside vendor in order to obtain a sufficient number of cows that were similar in age (35.6±0.9 mo), in their first or second lactation (1.2±0.06), had similar levels of milk production (projected 305-mature equivalent milk yield of 12,638±239 kg), were expected to calve within a limited time frame, were free of Johne’s, brucellosis, tuberculosis, and bovine leukemia virus (BLV), and had most recent SCC ranging from 7,000 to 187,000 /ml and mean of 36,000±6,172 (standard error). The mastitis history was not available prior to purchase. To minimize the poten- tial for prior infections with S. uberis, baseline titers to S. uberis and SUAM were assessed and those cows with the lowest titers were enrolled in the study [10]. At the time of challenge, all cows were free of major mastitis pathogens. Cows were housed in free-stalls with sand bedding at the East Tennessee Research and Education Center (ETREC) Little River Environmental Unit (Walland, TN), milked twice daily, and were managed according to standard on farm protocols. Experimental cows were milked last and milk discarded during the challenge period and following antimicrobial therapy. All procedures were approved by the University of Tennessee Institutional Animal Care and Use Committee (1982-1110). During the course of the study, 5 cows were treated for lameness (3 were in the rSUAM vaccination group and 2 were in the control group). Causes of lameness included heel warts treated with topical tetracycline (n=3), an abcess treated with ceftiofur (n=1), and possible foot rot treated with ceftiofur (n=1).No clinical signs of metritis or other diseases commonly found post-partum were observed.

Cows were blocked on parity, projected 305 mature equivalent milk yield, and expected calving dates within 5-7 days of each other. Within each block, cows were randomly allocated to each vaccine treatment. Cows were enrolled in one of four blocks over a 21 day period start- ing January 2012. All quarters received the same brand non-lactating cow antibiotic therapy at the completion of the last milking approximately 56 d prior to the estimated calving date. At 84, 56, and 28 days prior to expected calv- ing, pregnant Holstein cows (n=20 per treatment) were vaccinated subcutaneously in alternating sides of the neck with rSUAM (200 µg) in Montanide®ISA70VG (Seppic, Fairfield, NJ) or PBS in Montanide®ISA70VG (Control). Test and control vaccines contained a adjuvant:antigen ratio of 70:30 in a total volume of 2 ml. The proposed vaccination schedule was designed to provide protection during the early dry period and periparturient period as bovine mammary glands are highly susceptible to new S. uberis intramammary infections during these time peri- ods [11]. The third vaccination was delivered 22.9±1.00 days prior to estimated calving date across treatments, 21.3±1.41 within control treatment, and 22.5±1.41 within rSUAM treatment (mean±standard error). The rSUAM was generated as outlined previously [10].

Cows were blocked on parity, projected 305 mature equivalent milk yield, and expected calving dates within 5-7 days of each other. Within each block, cows were randomly allocated to each vaccine treatment. Cows were enrolled in one of four blocks over a 21 day period start- ing January 2012. All quarters received the same brand non-lactating cow antibiotic therapy at the completion of the last milking approximately 56 d prior to the estimated calving date. At 84, 56, and 28 days prior to expected calv- ing, pregnant Holstein cows (n=20 per treatment) were vaccinated subcutaneously in alternating sides of the neck with rSUAM (200 µg) in Montanide®ISA70VG (Seppic, Fairfield, NJ) or PBS in Montanide®ISA70VG (Control). Test and control vaccines contained a adjuvant:antigen ratio of 70:30 in a total volume of 2 ml. The proposed vaccination schedule was designed to provide protection during the early dry period and periparturient period as bovine mammary glands are highly susceptible to new S. uberis intramammary infections during these time peri- ods [11]. The third vaccination was delivered 22.9±1.00 days prior to estimated calving date across treatments, 21.3±1.41 within control treatment, and 22.5±1.41 within rSUAM treatment (mean±standard error). The rSUAM was generated as outlined previously [10].

Cows were challenged with a homologous strain of S. uberis, UT888 that was stored at -80°C in 50% Todd Hewitt Broth/glycerol, thawed in a 37°C water bath, and used to inoculate a blood agar plate. This strain was isolated previously from a case of clinical mastitis and used in an experimental challenge model [12]. After incubation for 23-24 h at 37°C, 3 representative colonies were inoculated into Todd Hewitt broth and incubated for 7 h at 37°C. Serial 10-fold dilutions were prepared in sterile PBS and the 10-5 dilution used to prepare the bacterial suspension for infusion. The dilutions also were plated at the time of infusion to determine the actual concentration of S. uberis delivered to the gland which ranged from 1, 185 to 3, 202 CFU/ml and averaged 2, 112±371 (standard deviation) in a total volume of 5 ml.

Cows were challenged within 72 h after calving with a mean±standard error of 1.2±0.13 days post-calving across treatments, 1.3±0.19 within control treatment, and 1.0±0.19 within rSUAM treatment This time frame was selected because it reflects one of the time periods when cows are most susceptible to S. uberis infection and when antibody titers in milk are greatest because of the need to deliver

Milk and sera samples collected immediately prior to challenge were analyzed for rSUAM specific antibody titers as described elsewhere [10,13,14]. Briefly, 96-well plates were coated with 1µg/ml of rSUAM and incubated at 4°C overnight. After incubation, the coating solution was removed and plates were washed 5X with Tris buffered saline with 0.5% (v/v) Tween-20 (TBS-T) and blocked with TBS-T containing 1% (v/v) gelatin (TBS-TG). Se- rum and whey samples were serially diluted in four-fold increments from 1:100 to 1:1, 638, 400 and 1:40 to 1:655, 360 respectively. Alkaline phosphatase-conjugated sheep anti-bovine IgG (H+L), IgG1, IgG2, IgA and IgM secondary antibodies (Bethel laboratories, NJ, USA) were diluted to 1:2000 in TBS-TG and 100 µl was added to each well and incubated for 1h. After washing, freshly prepared BCIP/ NBT phosphatase substrate solution was added into each well and incubated for ~ 20 min at room temperature until color developed. After incubation, the absorbance was read at wave length of 405 nm. Serum and milk titers were calculated by the intersection of least-square regression of A405 versus logarithm of the dilution. Titers are reported as the inverse of the dilution.

Foremilk samples were collected aseptically from all 4 quarters at least 24 h prior to challenge, immediately prior to challenge, twice daily at milking for the 1st week after challenge, and twice weekly though 28 d after challenge. Microbiological evaluation of milk samples (100 µl) was conducted by the Tennessee Quality Milk Laboratory (www.tqml.utk.edu) following procedures recommended by the NMC [15]. The number of somatic cells in milk was determined at the Dairy Herd Improvement Association Laboratory (Knoxville, TN).

Foremilk was scored using the following milk scoring system: 0=Normal, 1=Flakes, 2=Slugs/Clot, 3=Stringy/ watery/bloody. Clinical assessment of each mammary quarter was evaluated using the following scheme: 0=Nor- mal; the udder is pliable and light in weight when totally milked out. Heat, pain, redness, and/or swelling are not detectable. Cow exhibits no signs of discomfort that can be attributed to the milking process. 1=Slight swelling: the udder is less pliable with some firmness or heavier in weight as if not totally milked out. Redness, heat and pain are generally not detectable. 2=Moderate swelling; the udder is definitely firm, heavy, reddened and warm to the touch. The udder does not return to normal size when milked out. The cow generally exhibits signs of dis- comfort (irritable, performs a stepping motion with feet and/or kicks) during preparation for milking. 3=Severe swelling; the udder is very hard, heavy, red and hot. It is noticeably larger than other quarters before milking with little or no change in size following milking. The cow is extremely uncomfortable, very irritable and manifests pain by kicking and stepping. Rectal temperatures were taken using a digital thermometer prior to challenge and daily for 7 d after challenge.

Mammary quarters were considered infected if S. uberis was isolated at least twice with a corresponding increase in milk SCC or when clinical mastitis resulting from S. uberis infection was detected. Mastitis requiring antibiotic therapy was defined as a clinical assessment score of 2 for both milk appearance and udder evaluation or a score of 3 for either milk appearance or udder evaluation observed during 3 consecutive milkings. Animals were treated by intramammary infusion of ceftiofur hydrochloride using label guidelines for extended 8 day therapy (Spectramast LC, Pfizer Animal Health, Kalamazoo, MI).

Milk leukocytes were isolated from 50 ml of whole milk collected immediately prior to challenge (0), ap- proximately every 12 h through 3 d, and 7 d post-challenge. Milk with scores of 2 or greater was filtered through sterile gauze to remove flakes and clots. Milk leukocytes (~1-2x107) then were pelleted via centrifugation, rinsed with PBS, lysed with 1 ml Lysis buffer and stored at -80°C until RNA was extracted using RNeasy according to manufacturer’s instructions (Qiagen, Valencia, CA). As infection progressed, the volume of lysis buffer used was increased based on pellet size and associated increase in cell concentration. Total RNA quality and quantity was evaluated using the Experion System (BioRad, Hercules, CA). Total RNA (0.5 µg) was heat denatured at 70°C for 2 min prior to reverse transcribing with MMLV (Promega, Madison, WI) in the presence of 1x MMLV buffer, for 60 min at 42°C in a final volume of 20 µl. Controls did not receive reverse transcriptase enzyme to assess genomic DNA contamination. The PCR for each gene was run in triplicate in a 384 well plate and included 2 µl cDNA, 100 nM each of specific forward and reverse primers (Primer sequences are located in table 1; [16-22]), and 1x Sybr Green master mix in a final volume of 5 µl. Conditions of the PCR reaction included an initial 2 min at 50°C, then 95°C for 10 min, followed by 50 cycles of 95°C for 15 s and 60°C for 1 min. A melt curve was run to assess specificity of reaction. Cytokine gene expression was relative to the expression of 3 reference genes (YWHAZ, S24, and PPIA) selected from a panel of 8 (ACTB, GAPDH, PPIA, RPS9, S24, SDHA, UXT, YWHAZ) as exhibiting the least amount of variation in a test panel of milk leukocyte RNA samples [23,24]. An inter-run calibrator was created by pooling an equal volume of all samples. Reference genes and an inter-run calibrator were included on all plates. Specific RNA was detected using SYBR Green (LifeTechnologies, Waltham, MA) on an ABI7900HT Fast Real-Time PCR system (Applied Biosystems®, LifeTechnologies). The mean value for each triplicate sample was normalized to the geometric mean of the three selected reference genes as outlined previously [25] using the formula ∆Cq=Cq target– Cq reference [26]. ∆Cq values within a plate also were normalized to the ∆Cq values of the inter-run calibrator to remove technical variability between plates, resulting in ∆∆Cq. By using the formula X=2(-∆∆Cq) were linearized into a value representing expression of the target gene relative to the three reference genes [26].

Plasma total cortisol concentration was analyzed using an RIA procedure (Coat-A-Count, Diagnostic Products, Los Angeles, CA) as previously reported [27]. Samples were analyzed in duplicate and counted for 1 min using a gamma counter (Cobra II Auto-gamma counter, Model D5005, Packard Instrument Co., Meriden, CT). Intra- and inter-assay CV were 5.7 and 14.7% for low (38.6 nM) and 11.8 and 9.0% for high (149.3 nM) cortisol standards, respectively.

Response variables, S. uberis concentrations, SCC, as well as milk and mammary clinical scores were analyzed using a mixed model ANOVA with repeated measures that included vaccine as a fixed effect and cow nested within vaccine as a random effect. Antibiotic therapy was included in all models as a covariate because cows were provided antibiotic therapy as early as 3 d post challenge based on the severity of clinical mastitis signs presented which would significantly alter subsequent measures. Two additional covariates were included: time in days between 3rd vaccination and calving (vac3calv), as well as time between calving and challenge (challenge interval; 0, 1, 2, or 3 d). Including these two covariates allowed us to account for changes that occur during the periparturient period and improved the fitness of the model. However, limited numbers of animals within subgroups prevented drawing specific conclusions and will not be discussed further. Analyses were performed in SAS (Version 9.3, Cary, NC) with significance declared at p<0.05 and a trend declared at p=0.05-0.10.

Clyto Access

Concentration of antibodies in sera and milk prior to challenge
rSUAM vaccination significantly (p<0.005) increased rSUAM specific antibodies of each isotype examined in milk and sera of cows immediately prior to challenge when compared to controls (table 2). Within milk, the greatest fold change was observed in the IgM isotype, with rSUAM vaccinated cows expressing titers 82-fold greater than control cows. The IgG1 and IgG2 isotypes in milk of rSUAM vaccinated cows also were increased approximately 13 and 22 fold, respectively over controls, with IgG1 concentrations nearly twice that of IgG2 in rSUAM vaccinated cows.
S. Uberis presence in the milk
The definition of infection used in this study was the presence of S. uberis in two consecutive samplings with more than 10 CFU/ml of milk [12]. Based on this defini- tion, 100% of cows became infected following challenge. Time had a significant influence (p<0.0001) on bacterial concentrations, which peaked by day 3 (12, 304±8, 862 CFU/ml) followed by a rapid decrease to 8±7 CFU/ml by day 7 post-challenge. This rapid reduction in mean S. uberis concentrations most likely reflects the administra- tion of antibiotics to 65% of all cows by 7 d post-challenge because of the severity of clinical signs. Although the overall effect of vaccine was not significant (p=0.46), a significant interaction of vaccine with time was observed (p=0.007) for S. uberis concentrations in milk (table 3). Closer examination revealed that cows receiving rSUAM initially experienced similar concentrations of S. uberis in milk. However, by 1.5 d post-challenge the number of S. uberis in milk from cows that received rSUAM rapidly and significantly (p=0.03) increased compared to control cows. S. uberis concentrations peaked for both groups at 3 d post-challenge, with the rSUAM vaccine group having six-fold greater concentrations than the control group (p=0.19). By 28 d, both treatment groups had approximately 15% of cows still shedding S. uberis.
Antibiotic therapy
The majority of cows (28/37 or 73%) required antibiotic therapy as a result of clinical scores and/or continued shed- ding of S. uberis in milk from infected mammary glands. The mean time to antibiotic therapy was 6.6 d for those that required it with a range of 3 to 26 d. With respect to rSUAM vaccine treatment, 82% (14/17) of cows required antibiotic therapy versus 65% (13/20) of control cows (p=0.24). Three rSUAM vaccinated cows were excluded because of early removal from study for tissue collection and did not qualify for antibiotic therapy before removal 4-7 d post-challenge.
Clinical evaluation of mammary in- flammation

Clyto Access

Clyto Access

The severity of inflammation in the mammary gland and milk was evaluated prior to and following experi- mental challenge, with scores of 0 representing no signs of inflammation and 3 representing severe signs. Mammary scores increased significantly with time (p<0.0001) to 0.2±0.11 within 1 d (24 h), increased to 0.9-1.0±0.11 d 2 thru 7 post challenge, started to decrease 7 d post challenge, and returned to pre-challenge scores 18 d post-challenge (0.1±0.12) (figure 1). During the plateau period 55-70% of cows experienced scores >0 with little variation in mean score, regardless of vaccine treatment. Changes in mammary score, preceded changes in milk scores which significantly increased with time (p<0.0001) to 0.4±0.12 by the second day post-challenge, peaked at 1±0.12 by the third day, and decreased to pre-challenge levels by 14 d post-challenge. Although a significant effect of vaccine was not observed relative to milk score, a comparison of vaccine responses at individual time points revealed a significantly higher (p<0.05) milk score in the rSUAM vaccinated cows versus control cows at 2 d (0.7±0.17 vs 0.2±0.15, respectively) and 7 d (1.3±0.18 vs 0.6±0.15, respectively) suggesting a more rapid and extended in- flammatory response by cows vaccinated with rSUAM. This also was evidenced by a significantly greater portion of milk scores greater than 0 for cows receiving rSUAM vaccine vs control, (56.2 vs 43.8 respectively, p=0.01). Milk and mammary scores were correlated with an r2 value of 0.66 (p<0.0001).
Clyto Access
Figure 1: Mammary and milk scores following experimental challenge with S. uberis. For mammary scores, the effect of time was p<0.0001, vaccine was p>0.10, the interaction of the two was 0.95, and antibiotic therapy was p<0.0001. For milk scores, the effect of time was p<0.0001, vaccine was p>0.10, the interaction of the two was 0.19, and antibiotic therapy was p=0.0005. *Indicates a significant difference between vaccine treatments for milk score at that time point (p≤0.05).

Clyto Access
Figure 2:Somatic cell count following experimental challenge with S.uberis. The effect of time was p<0.0001, vaccine was p>0.10, the interaction of the two was 0.99, and antibiotic therapy was p<0.0001

Clyto Access
Figure 3:Rectal temperature (0C) following experimental intramammary challenge with S. uberis. The effect of time was p<0.0001, vaccine was p>0.10, the interaction of the two was p=0.3, and antibiotic therapy was p=0.05. *Indicates a significant difference between vaccine treatments for rectal temperature at that time point (p≤0.05).

Similar increases in SCC were observed regardless of vaccine administration (figure 2), with control cows having an average SCC of 394, 750±73, 769 cells/ml over the 28 d sampling period versus 513, 830±119, 925 for cows receiving rSUAM. For both treatment groups, a significant increase in SCC occurred 1 d after challenge, peaked at 3 d post challenge, before returning to baseline levels 21d after challenge. The incfrease at day1 post-challenge coincides with increased mammary scores and precedes the significant changes in milk scores of inflammation on teh second post challenge. . Clearance of infection based on a return to SCC<200,000 cells/ml of milk was similar between rSUAM vaccinated and control cows at 18.8±4.0 d and 20.3±2.5 d, post-challenge respectively.
Systemic signs of infection
(i) reCtal temperature
Time had a significant influence (p<0.0001) on rectal temperature following S. uberis challenge. The initial tem- perature for all cows was 38.6°C±0.1 and increased signifi- cantly within 0.5 d to 39.0°C±0.1, continued to increase through day 3 where peak temperatures of 40.3°C±0.3 were observed. Temperatures were still slightly elevated at day 7 post-challenge 39.0°C±0.1 when recording was terminated. Vaccine did not have a major influence with an average temperature for each treatment group ranging between 39.3 to 39.4°C±0.1 (figure 3). At 0.5 and 2.5 d post-infection, the control cows had significantly greater rectal temperatures (p<0.05) than cows receiving rSUAM.
(ii) milk weight
A significant influence of time (p<0.0001) but not vac- cine (p=0.4) or antibiotic therapy (p=0.25) was observed for milk yield following S. uberis challenge. Increased milk yield with time is expected as cows typically increase pro- duction through at least 6-8 weeks after calving and is con- sistent with this study. Cows started at 15.7±1.1kg the day of challenge which occurred within 3 d of calving to 32.4±1.1 kg by 28 d after challenge (figure 4). A deviation from this pattern for both treatments occurred with significantly lower production 2 d post-challenge (19.0±1.1 kg) versus the first day post-challenge (20.7±1.1) and coincides with increases in clinical signs observed at that time.
Clyto Access
Figure 4: Milk weight (kg) following experimental in- tramammary challenge with S. uberis. The effect of time was p<0.0001, vaccine was p>0.10, the interaction of the two was p=0.97, and antibiotic therapy was p=0.25. The SE was relatively low (1.3-1.7) and is not depicted on the graph.

(iii) CortiSol
Immediately prior to challenge, plasma cortisol was approximately 28±3.1 nM with a significant effect of time (p<0.0001). Values increased significantly 3 d post-challenge to 43±3.1 nM before significantly decreasing to 21±3.2 nM at 7 d post challenge. A significant interaction of vac- cine and time was observed (p<0.01). Only the rSUAM vaccinated group experienced a significant increase in cortisol 3 d post-challenge (54±4.6 nM) when compared to the control group (32±4.3 nM).
Cytokine responses
Unlike milk and mammary scores, SCC, and S. uberis concentrations, antibiotic therapy did not influence milk leukocyte cytokine RNA expression in the mixed model (p>0.40; figure 5).IL-1 increased significantly within 12 h following challenge, followed by an additional 10 fold increase at 24h, with expression peaking at 2 to 2.5 d post challenge. At 7 d post-challenge, IL-1 levels still remained 34-fold greater than prior to challenge. Vaccination with rSUAM tended to have an effect on milk leukocyte synthesis of IL-1 RNA (p=0.10) with approximately two-fold greater RNA expression than control cows (0.190±0.05 and 0.097±0.02, respectively). Closer examination revealed a significant (p=0.03) 10-fold greater expression of IL-1 RNA prior to challenge by rSUAM vaccinated cows versus control cows. After challenge, this difference dropped to approximately 4-fold through 1.5 d post-challenge, after which IL-1 RNA expression for each treatment group paralleled each other.
Clyto Access
Figure 5: Generation of cytokine RNA by milk somatic cells following experimental intramammary challenge with S. uberis (n=10 cows per vaccine treatment). Cytokine RNA expressed relative to 3 house keeping genes (YWHAZ, S24, and PP1A) after normalizing to an internal calibrator to remove technical variability. Unless otherwise indicated the effect of vaccine, time, vaccine interaction with time, and antibiotic therapy were p>0.10. *Indicates a significant difference between vaccine treatments for cytokine at that time point (p≤0.05). †Indicates a significant difference between vaccine treatments for cytokine at that time point (p>0.05 and p≤0.10).

IL-6 significantly increased 75 fold by 24 h following challenge. This peak was followed by a rapid decrease in 36-48 h, with values dropping to approximately 6-fold greater than baseline 7 d post-challenge. Unlike most other cytokines examined, no variation was observed between treatment groups at individual time points.
A significant increase in IFN-γ RNA synthesis was observed within 12 h of challenge, which increased ap- proximately 10-fold by 2.5 d when peak concentrations were observed prior to decreasing by 7 d post-challenge. At 2.5 d post-challenge control cows demonstrated a significant increase (p=0.05) in IFN-γ expression versus rSUAM vaccinated cows, 0.64±0.25 vs 0.17±0.07.
Similar to IFN-γ a significant, 10-fold increase in IL-17 RNA was observed 1 d following challenge. In contrast where IFN-γ began to decline at 2.5 d post-challenge, IL- 17 continued to increase 345-fold 7 d post-challenge, the final sampling time. Although not statistically significant, control cows had an approximate 3 fold greater increase in IL-17 seven d post-challenge than rSUAM vaccinated cows (2.45±1.2 vs 0.78±0.47). This directly contrasts with expression of IL-17 RNA prior to challenge where the rSUAMvaccinated group had ~ 16 fold greater expression of IL-17 at calving than the control group. This reversal in cytokine responses most likely contributes to the observed interaction between vaccine and time (p=0.06).
IL-4 RNA was expressed approximately 5-fold greater than the other cytokine transcripts examined. Unlike the other cytokines, there was not a significant increase of IL-4 with time following challenge (p=0.2). Although not significant, closer examination of the trends of control versus rSUAM vaccinated cows revealed cows receiving rSUAM produced nearly 3-fold more IL-4 RNA at 2.5 to 3 d post-challenge than control cows.
As with IL-1 and IL-6, time significantly influenced IL-10 levels with an approximate 3 fold rise in IL-10 RNA synthesis by milk leukocytes 24 h following challenge, peaking 2.5 d post-challenge, and dropping to levels ap- proximately 4.5 fold greater than baseline. rSUAM vacci- nated cows experienced significantly greater levels of IL-10 more quickly (within 1 d post challenge) when compared to control cows which took longer to reach comparable levels approximately 2.5 d post challenge. This could be linked partly to the approximately 2.5 fold greater expres- sion of IL-10 by the rSUAM vaccinated group observed immediately prior to challenge. These differences are reinforced by the tendency for the overall effect of vac- cine (p=0.07) where the rSUAM vaccinated group tended to have approximately 1.5 fold more IL-10 RNA than the control group (0.49±0.09 vs 0.30±0.05, respectively).

One of the primary goals of this study was to evalu- ate the effectiveness of rSUAM in protecting cows from developing S. uberis infections during a highly susceptible time period and when antibody titers are the greatest, e.g. the first few days after calving [28,29]. The susceptibility of cows at this time period was evident in this study as S. uberis infected all quarters challenged based on isolation of S. uberis at least twice from the infected mammary gland. This differed significantly from our prior research in which cows challenged at least 11 d after calving demonstrated a 75-80% infection rate using this same definition [12]. The greater susceptibility post-partum most likely stems from a complex series of physiological changes that occur to induce calving and produce more milk for health and well-being of the calf [30-32]. These changes include al- tered hormonal expression, lipid metabolism, and immune function. However, as indicated at the beginning of the paragraph. Our prior research in vitro also indicates that sera containing SUAM-specific antibodies diluted 1/1000 decreased the binding and internalization of S. uberis to mammary epithelial cells to approximately 20% of controls (10). More recently, our research group pre-opsonized S. uberis prior to intramammary challenge of Holstein cows 30-60 days in milk. When compared to non-opsonized S. uberis, clinical signs and bacterial concentrations were significantly lower. The discordant results between those studies and the current study may result from the differ- ences in stage of lactation as outlined earlier. In addition, the two described studies pre-opsonized bacteria in a rela- tively small volume of liquid. Where as, S. uberis infused into the mammary gland is a considerably larger space and may limit the interaction of antibodies with S. uberis.

The host signs of mastitis were similar between rSUAM vaccinated and saline injected control cows when consid- ering the overall effect of vaccine on milk and mammary inflammation, SCC, and rectal temperature. Examination of individual time points relative to vaccination offered additional insight to the infection process between the two treatment groups. At 2 d post-challenge cows vac- cinated with rSUAM experienced a significant increase in milk inflammatory scores relative to the control group. Increased milk scores are noted by larger clots and slugs in the milk indicating a massive influx of sera, complement, and cells into milk. This significant variance in milk scores occurred within 12 h after significantly greater concentra- tions of S. uberis in the rSUAM vaccinated group at 1.5 d post-challenge. Of note, a recent study evaluating vac- cination with cell-free extracts of several S. uberis strains also observed naturally-occurring S. uberis infections two weeks post-calving in 2 of 6 cows, but none of the 6 control animals [33]. These observations beg the question of why would vaccination allow greater bacterial concen- trations and/or infection rates? Based on data collected, two possible scenarios exist which may act independently or concurrently.

The first possible scenario suggests that the vaccine partly performed as expected. At calving, a significant increase in rSUAM specific antibodies was observed in milk/colostrum of rSUAM vaccinated versus control cows. This agrees with our prior research [10]. In vitro, rSUAM specific antibodies decreased adherence and internalization of S. uberis to mammary epithelial cells [10]. Assuming the same occurred in vivo, this would limit entry into mam- mary tissue and leave more S. uberis in milk and would support the observation of greater bacterial concentrations and activation of the inflammatory response evident in greater generation of IL-1 RNA and milk inflammatory scores. Future studies which collect mammary biopsies from both treatment groups to examine bacterial entry would be able to address this question more effectively. A second alternate or contributing scenario for greater S. uberis concentrations in milk from rSUAM vaccinated cows may be tied to shifts in the type of immunity generated by vaccination. Vaccination with rSUAM had the greatest influence on generation of IL-10 by milk leukocytes which started approximately 2.5 fold greater prior to challenge and remained greater than saline injected cows through at least 3 d post challenge. Furthermore, IL-10 was five-fold more abundant than either IFN-γ or IL-17 RNA in milk leukocytes from rSUAM vaccinated cows. The observed cytokine profile suggests that IL-10 forms the dominant response by milk leukocytes to vaccination with rSUAM during later stages of gestation. A similar pattern, with greater IL-10 and lower IFN-γ synthesis was observed in a prior study where vaccination with LACK (Leishmania homolog of the receptor for activated C kinase) induced proliferation of CD4+CD25+ T regulatory 1-like cells [34].Of note, the LACK vaccinated group had significantly greater parasite loads post-challenge and agrees with our observations that rSUAM vaccination increased S. uberis loads in milk. Induction of a predominant IL-10 response could have several implications regarding the defense of the mammary gland against infection and will be briefly summarized in the next paragraph.

In general, IL-10 generates a more suppressive immune environment by altering macrophage killing, proinflam- matory cytokine production, and antigen presentation to protect tissues against damage caused by infection [35]. One of the key changes associated with greater IL-10 expression is decreased killing of organisms such as S. aureus by mononuclear phagocytes [36]. Reduced killing by macrophages can severely hinder the host immune response against S. uberis as mammary macrophages more effectively kill S. uberis than neutrophils when in a milk-based environment [37,38]. Increases in IL-10 also have been linked to reduced IFN-γ, a cytokine which helps increase the killing and antigen presentation abilities of macrophages [35]. In our study, the rSUAM vaccinated group had significantly lower expression of IFN-γ RNA than the saline injected group which agrees with this prior research and also indicates that macrophage function could be impaired in the rSUAM vaccinated group. Other studies have demonstrated that blocking of IL-10 increases the expression of antimicrobial peptides β-defensin and LL-37 by keratinocytes [39]. As epithelial cells are the dominant cell type in the mammary gland, lowering the capability of these cells to aid in defense of the gland could alter the severity or final outcome of infection. Although this is not a full listing of the potential effects of IL-10, it strongly indicates that greater IL-10 RNA concentrations observed following rSUAM vaccination during late gestation could promote immune suppression and alter the ability of the mammary gland to eliminate bacteria and fight off infection.

An additional question to be addressed is why vac- cination with rSUAM would promote an IL-10 dominate immune response. Potential factors would include the antigen itself, the adjuvant, the method of vaccination, and the overall physiological status of individual cows at the time of vaccination. The last factor, physiological status, may be particularly relevant in this study. The vaccination strategy used reflects a common one in the dairy industry that attempts to maximize protection during periods of greater susceptibility to infection that occur when cows stop lactating approximately two months prior to parturition and the immediate 2-3 weeks prior to and after parturi- tion [40]. However, hormonal changes that occur with pregnancy generally promote anti-inflammatory reactions to minimize inflammation and preserve the pregnancy. This type response includes increases in IL-4, IL-10, TH2 and T regulatory cells [41-43] and may not be protective against certain organisms. For example, in pregnant mice the load of Leishmania major parasites increased signifi- cantly in the footpad following experimental challenge when compared to non-pregnant mice [44]. The increase in parasite load also was accompanied by reduced antigen specific generation of IFN-γ and increased IL-4 and IL-10 by spleen cells. In humans, ex vivo stimulation of whole blood at 10, 20, and 30 weeks of gestation also revealed a significant increase in IL-10 generation following LPS stimulation coupled with a decrease in IFN-γ over time [41]. The set of changes observed in these two studies mimics our observations with rSUAM vaccination and subsequent challenge with S. uberis. The influence of pregnancy also would have implications, not just for the current study, but potentially other vaccine protocols administered during late gestation. Future studies that investigate responses in non-pregnant animals would provide additional insight as to whether it is the antigen, method of vaccination, and/ or physiological status of the animal that influenced the immune response generated.

Potential factors influencing cytokine production following vaccination includes the antigen itself. Sig- nificantly greater concentrations of IL-1 and IL-17 RNA were observed in milk leukocytes immediately prior to experimental challenge for cows receiving the rSUAM vaccine versus saline. Similar observations were made in a recent trial where dairy cows were vaccinated with cell free extracts from three strains of S. uberis representing different clonal complexes that theoretically would include SUAM [33]. In this related trial, approximately 2.5 weeks after the second vaccination, a trend towards increased IL-17 expression by antigen stimulated peripheral blood mononuclear cells was noted. In our study, this increase in IL-17 was observed in milk leukocytes prior to challenge approximately 2.5-4 weeks after the third vaccination, with the actual time frame depending on how closely the cow calved relative to the expected calving date. These stud- ies taken together, suggest that rSUAM and/or additional components of S. uberis have the potential to promote IL-17 based adaptive responses. An increase in the pro- inflammatory cytokine IL-17 may seem counterintuitive to generation of IL-10, which is immune suppressive in nature. However, the cytokine milieu in response tor SUAM vaccination could contribute to generation of both IL-17 and IL-10. In IL-6 knockout mice, splenocytes produced both IL-10 and IL-17 in the presence of TGF-β and IL- 6[45]. In addition, the percentage of cells producing IL-10 remained the same while splenocytes expressing both IL-10 and IL-17, as well as IL-17 alone decreased as the concentration of IL-6 decreased. Thus, the combination of cytokines present could influence the degree to which IL-10 and IL-17 are expressed.

Vaccination with rSUAM during late gestation increased antibody titers and the generation of IL-10 by milk leuko- cytes prior to and following intramammary challenge. This combination of responses could have increased bacterial concentrations by limiting S. uberis entry into tissue and suppressing mammary gland immunity, respectively. Future research aimed at identifying the influence of the antigen (rSUAM), adjuvant, physiological status (pregnancy), tim- ing, etc. towards promoting IL-10 dominate responses will help develop vaccines that more effectively protect the mammary gland during periods of greater susceptibility.

This project was supported by the Agriculture and Food Research Initiative Competitive Grant no. 2011-67015- 30168 from the USDA National Institute of Food and Agriculture to S.P. Oliver, G.M. Pighetti, R.A. Almeida, O. KerroDego, and M.E. Prado. AgResearch at University of Tennessee Institute of Agriculture also provided partial support to this project. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Special thanks are given to the staff at the East Ten- nessee Research and Education Center, who provided considerable help during this study.

1. Hogan JS, Smith KL, Hoblet KH, Schoenberger PS, Todhunter DA, Hueston WD, Pritchard DE, Bowman GL, Heider LE, Brock- ett BL et al: Field survey of clinical mastitis in low somatic cell count herds. J Dairy Sci 1989, 72(6):1547-1556.
2. Keane OM, Budd KE, Flynn J, McCoy F: Pathogen profile of clinical mastitis in Irish milk-recording herds reveals a com- plex aetiology. Vet Rec 2013, 173(1):17.
3. Oliver SP, Gillespie BE, Headrick SJ, Moorehead H, Lunn P, Dowlen HH, Johnson DL, Lamar KC, Chester ST, Moseley WM: Efficacy of extended ceftiofur intramammary therapy for treatment of subclinical mastitis in lactating dairy cows. J Dairy Sci 2004a, 87(8):2393-2400.
4. Riekerink RGMO, Barkema HW, Kelton DF, Scholl DT: Inci- dence rate of clinical mastitis on Canadian dairy farms. Jour- nal of Dairy Science 2008, 91(4):1366-1377.
5. Todhunter DA, Smith KL, Hogan JS: Environmental streptococ- cal intramammary infections of the bovine mammary gland. J Dairy Sci 1995, 78(11):2366-2374.
6. Verbeke J, Piepers S, Supre K, De Vliegh- er S: Pathogen-specific incidence rate of clinical mastitis in Flemish dairy herds, severity, and association with herd hy- giene. Journal of Dairy Science 2014, 97(11):6926-6934.
7. Almeida RA, Oliver SP: Trafficking of Streptococcus uberis in bovine mam- mary epithelial cells. Microb Pathog 2006, 41(2-3):80-89.
8. Patel D, Almeida RA, Dunlap JR, Oliver SP: Bovine lactoferrin serves as a mo- lecular bridge for internalization of Streptococcus uberis into bovine mam- mary epithelial cells. Vet Microbiol 2009, 137(3-4):297-301.
9. Chen X, Dego OK, Almeida RA, Fuller TE, Luther DA, Oliver SP: Deletion of sua gene reduces the ability of Strep- tococcus uberis to adhere to and inter- nalize into bovine mammary epithelial cells. Vet Microbiol 2011, 147(3-4):426- 434.
10. Prado ME, Almeida RA, Ozen C, Luther DA, Lewis MJ, Headrick SJ, Oliver SP: Vaccination of dairy cows with recom- binant Streptococcus uberis adhesion molecule induces antibodies that reduce adherence to and internalization of S. uberis into bovine mammary epithelial cells. Vet Immunol Immunopathol 2011, 141(3-4):201-208.
11. Oliver SP: Frequency of isolation of environmental mastitis-causing patho- gens and incidence of new intramam- mary infection during the nonlactating period. Am J Vet Res 1988, 49(11):1789- 1793.
12. Rambeaud M, Almeida RA, Pighetti GM, Oliver SP: Dynamics of leukocytes and cytokines during experimentally in- duced Streptococcus uberis mastitis. Vet Immunol Immunopathol 2003, 96:193- 205.
13. Kerro Dego O, Prysliak T, Potter AA, Per- ez-Casal J: DNA-protein immunization against the GapB and GapC proteins of mastitis isolate of Staphylococcus au- reus. Vet Immunol Immunopathol 2006, 113:125-138.
14. Perez-Casal J, Prysliak T, Kerro Dego O, Potter AA: Staphylococcus aureus GapC/B chimera and its potential use as a component of a vaccine for S. aure- us mastitis. Vet Immunol Immunopathol 2006, 109:85-97.
15. Oliver SP, Gonzalez RN, Hogan JS, Jayarao BM, Owens WE: Microbiologi- cal procedures for the diagnosis of bo- vine udder infection and determination of milk quality, 4th edn: The National Mastitis Council, Inc., Verona, WI; 2004.
16. Bevilacqua C, Helbling JC, Miranda G, Martin P: Translational efficiency of ca- sein transcripts in the mammary tissue of lactating ruminants. Reproduction, nutrition, development 2006, 46(5):567- 578.
17. Rosbottom A, Guy CS, Gibney EH, Smith RF, Valarcher JF, Taylor G, Williams DJ: Peripheral immune responses in preg- nant cattle following Neospora cani- num infection. Parasite Immunol 2007, 29(4):219-228.
18. Leutenegger CM, Alluwaimi AM, Smith WL, Perani L, Cullor JS: Quantitation of bovine cytokine mRNA in milk cells of healthy cattle by real-time TaqMan polymerase chain reaction. Vet Immunol Immunopathol 2000, 77(3-4):275-287.
19. Witchell J, Maddipatla SV, Wangoo A, Vordermeier M, Goyal M: Time depen- dent expression of cytokines in Myco- bacterium bovis infected cattle lymph nodes. Vet Immunol Immunopathol 2010, 138(1-2):79-84.
20. Konnai S, Usui T, Ohashi K, Onuma M: The rapid quantitative analysis of bo- vine cytokine genes by real-time RT- PCR. Vet Microbiol 2003, 94:283-294.
21. Blanco FC, Bianco MV, Meikle V, Gar- baccio S, Vagnoni L, Forrellad M, Klepp LI, Cataldi AA, Bigi F: Increased IL-17 expression is associated with pathology in a bovine model of tuberculosis. Tu- berculosis (Edinb) 2011, 91(1):57-63.
22. Goossens K, Van Poucke M, Van Soom A, Vandesompele J, Van Zeveren A, Peel- man LJ: Selection of reference genes for quantitative real-time PCR in bovine preimplantation embryos. BMC devel- opmental biology 2005, 5:27.
23. Bustin SA, Benes V, Garson JA, Helle- mans J, Huggett J, Kubista M, Mueller R, Nolan T, Pfaffl MW, Shipley GL et al: The MIQE guidelines: minimum infor- mation for publication of quantitative real-time PCR experiments. Clin Chem 2009, 55(4):611-622.
24. Pfaffl MW, Tichopad A, Prgomet C, Neuvians TP: Determination of stable housekeeping genes, differentially regu- lated target genes and sample integrity: BestKeeper--Excel-based tool using pair-wise correlations. Biotechnology letters 2004, 26(6):509-515.
25. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, Spe- leman F: Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple inter- nal control genes. Genome Biology 2002, 3(7):research0034.0031-0034.0011.
26. Livak KJ, Schmittgen TD: Analysis of relative gene expression data using real- time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 2001, 25(4):402-408.
27. Doherty TJ, Kattesh HG, Adcock RJ, Wel- born MG, Saxton AM, Morrow JL, Dailey JW: Effects of a concentrated lidocaine solution on the acute phase stress re- sponse to dehorning in dairy calves. J Dairy Sci 2007, 90(9):4232-4239.
28. Jayarao BM, Gillespie BE, Lewis MJ, Dowlen HH, Oliver SP: Epidemiology of Streptococcus uberis intramammary infections in a dairy herd. Zentralbl Vet- erinarmed B 1999, 46(7):433-442.
29. Smith KL, Todhunter DA, Schoenberger PS: Environmental pathogens and in- tramammary infection during the dry period. J Dairy Sci 1985b, 68(2):402-417.
30. Drackley JK: Biology of dairy cows during the transition period: the final frontier?J Dairy Sci 1999, 82:2259-2273.
31. Goff JP, Horst RL: Physiological changes at parturition and their relationship to metabolic disorders. Journal of Dairy Science 1997, 80(7):1260-1268.
32. Sordillo LM, Raphael W: Significance of metabolic stress, lipid mobilization, and inflammation on transition cow disor- ders. Vet Clin North Am Food Anim Pract 2013, 29(2):267-278.
33. Wedlock DN, Buddle BM, Williamson J, Lacy-Hulbert SJ, Turner SA, Subharat S, Heiser A: Dairy cows produce cytokine and cytotoxic T cell responses following vaccination with an antigenic fraction from Streptococcus uberis. Vet Immunol Immunopathol 2014, 160(1-2):51-60.
34. Stober CB, Lange UG, Roberts MT, Alca- mi A, Blackwell JM: IL-10 from regula- tory T cells determines vaccine efficacy in murine Leishmania major infection. J Immunol 2005, 175(4):2517-2524.
35. Sabat R: IL-10 family of cytokines. Cyto- kine Growth Factor Rev 2010, 21(5):315-324.
36. Roilides E, Anastasiou-Katsiardani A, Dimitriadou-Georgiadou A, Kadiltsoglou I, Tsaparidou S, Panteliadis C, Walsh TJ: Suppressive effects of interleukin-10 on human mononuclear phagocyte function against Candida albicans and Staphylococcus aureus. J Infect Dis 1998, 178(6):1734-1742.
37. Grant RG, Finch JM: Phagocytosis of Streptococcus uberis by bovine mam- mary gland macrophages. Res Vet Sci 1997, 62(1):74-78.
38. Leigh JA, Field TR: Streptococcus uber- is resists the bactericidal action of bo- vine neutrophils despite the presence of bound immunoglobulin. Infect Immun1994, 62(5):1854-1859.
39. Howell MD, Novak N, Bieber T, Pastore S, Girolomoni G, Boguniewicz M, Streib J, Wong C, Gallo RL, Leung DY: Inter- leukin-10 downregulates anti-microbial peptide expression in atopic dermatitis. J Invest Dermatol 2005, 125(4):738-745.
40. Sordillo LM: Factors affecting mam- mary gland immunity and mastitis sus- ceptibility. Livestock Production Science 2005, 98(1-2):89-99.
41. Denney JM, Nelson EL, Wadhwa PD, Waters TP, Mathew L, Chung EK, Gold- enberg RL, Culhane JF: Longitudinal modulation of immune system cyto- kine profile during pregnancy. Cytokine 2011, 53(2):170-177.
42. Wegmann TG, Lin H, Guilbert L, Mos-mann TR: Bidirectional cytokine inter- actions in the maternal-fetal relation- ship: is successful pregnancy a TH2 phenomenon?Immunol Today 1993, 14(7):353-356.
43. Somerset DA, Zheng Y, Kilby MD, San- som DM, Drayson MT: Normal hu- man pregnancy is associated with an elevation in the immune suppressive CD25(+) CD4(+) regulatory T-cell sub- set. Immunology 2004, 112(1):38-43.
44. Krishnan L, Guilbert LJ, Russell AS, Wegmann TG, Mosmann TR, Belosevic M: Pregnancy impairs resistance of C57BL/6 mice to Leishmania major infection and causes decreased antigen- specific IFN-gamma responses and increased production of T helper 2 cytokines. Journal of Immunology 1996,156:644-652.
45. McGeachy MJ, Bak-Jensen KS, Chen Y, Tato CM, Blumenschein W, McCla- nahan T, Cua DJ: TGF-beta and IL-6 drive the production of IL-17 and IL- 10 by T cells and restrain T(H)-17 cell- mediated pathology. Nat Immunol 2007, 8(12):1390-1397.

Published: 15 May 2017

Reviewed By : Dr Angel R. Sánchez Quinche.Dr Ion Pérez Baena.Dr Fernando Lenin Aguilar Gálvez.

Copyright:

Copyright: © 2017 Gina M Pighetti. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms