Article Text

Download PDFPDF

Paper
The consequences of vaccination with the Johne's disease vaccine, Gudair, on diagnosis of bovine tuberculosis
  1. M. Coad, MSc1,
  2. D. J. Clifford, BVMS, MRCVS2,
  3. H. M. Vordermeier, PhD1 and
  4. A. O. Whelan, PhD1
  1. 1Department of Bovine Tuberculosis, Animal Health and Veterinary Laboratories Agency, New Haw, Addlestone, Surrey KT15 3NB, UK
  2. 2Animal Sciences Unit, AHVLA, New Haw, Addlestone, Surrey KT15 3NB, UK
  1. E-mail for correspondence: michael.coad{at}ahvla.gsi.gov.uk

Abstract

The single intradermal comparative cervical tuberculin skin-test (SICCT) remains the primary surveillance tool to diagnose bovine tuberculosis (BTB) in the UK. Therefore, understanding the potential confounding influences on this test is important. This study investigated the effects of vaccination against Johne's disease (JD) on the immunodiagnosis of BTB using a Mycobacterium bovis BCG vaccination model as a surrogate of M bovis infection. Calves were vaccinated with either BCG (an attenuated live vaccine) or the JD vaccine, Gudair (a heat-inactivated suspension of Mycobacterium avium subspecies paratuberculosis), or a combination of both, and SICCT responses were measured approximately six and 12 weeks postvaccination. Animals vaccinated with Gudair only were negative to the SICCT test, thus supporting the specificity of the SICCT test following Gudair vaccination. However, while animals vaccinated with BCG-only demonstrated a bovine tuberculin-biased response as expected, covaccination with Gudair resulted in a bias towards avian tuberculin in the SICCT test. Therefore, our model demonstrates the potential of the Gudair vaccine to reduce the sensitivity of the SICCT. In addition, while we also demonstrate that Gudair vaccination can compromise the specificity of serological tests to detect JD, the specificity of defined M bovis antigens in serological or interferon gamma-based blood assays was not compromised by the vaccine.

View Full Text

Statistics from Altmetric.com

Introduction

Bovine tuberculosis (BTB) is a disease of economic and zoonotic importance caused by the bacterial pathogen Mycobacterium bovis. In the UK, control of BTB centres on the application of the single intradermal comparative cervical tuberculin (SICCT) test and the subsequent removal of animals found to be positive (reactors). Despite this strategy, the incidence of BTB has been on the rise since 1988 (Krebs 1997) and continues to be a major economic burden to the farming community.

Another important disease caused by a mycobacterial pathogen is Johne's disease (JD), a chronic infection of the intestine of ruminants caused by Mycobacterium avium subspecies paratuberculosis (MAP), which results in significant loss in milk yield and, occasionally, wasting and eventual death. While the true incidence of JD in the UK cattle population is not known, a recent DEFRA-funded survey estimated that 34.7 per cent (95 per cent CI 27.6 per cent to 42.5 per cent) of UK dairy herds were infected with MAP (DEFRA 2009).

An available vaccine against JD which is authorised for use in sheep and goats is the Gudair vaccine, a heat-inactivated suspension of MAP. Due to the absence of a vaccine against JD for use in cattle in the Veterinary Medicines Directorate (VMD) has, since 2005, permitted the importation of the Gudair vaccine for use in cattle under its Special Import Certificate (SIC) scheme (Price 2011). Records maintained by the VMD indicate that during each of the last five years, approximately 15 000 doses of Gudair have been approved for import into the UK for use in cattle (A Fawcus, VMD, personal communication). While it is unclear exactly how the Gudair vaccine is being used in these animals, its use could have implications for the diagnosis of BTB.

Vaccination against JD has long been known to exert a confounding influence on skin test-based diagnosis of BTB in cattle (Hebert and others 1959, Inglis and Weipers 1963). Indeed, the product information provided with the vaccine acknowledges that the vaccination with Gudair can induce skin-test reactions to avian and bovine tuberculin-purified protein derivatives (PPD-A and PPD-B, respectively). This is not surprising given that MAP is a subspecies of the M avium family, and that PPD-A is a relatively crude antigenic product derived from a M avium culture filtrate. Since the use of the SICCT to detect BTB relies on differential responses to bovine and avian PPDs, vaccine-induced elevation of PPD-A responses has the potential to mask a BTB-induced PPD-B response, thereby generating a false negative outcome.

The primary aim of this study was to investigate the effect of Gudair vaccination on the diagnosis of BTB. Since vaccination of cattle with the tuberculosis vaccine strain M bovis BCG can induce PPD-B-biased skin-test responses, (reviewed by Vordermeier and others 2011) this model was used as a diagnostic surrogate of M bovis infection. We investigated skin test and blood-based immune responses in calves that were vaccinated neonatally with combinations of BCG and/or Gudair vaccines, and report on the consequences of Gudair ­vaccination on the immunodiagnosis of BTB.

Materials and methods

Animals

Holstein-Friesian calves (n = 42) were sourced from BTB-free farms at approximately two weeks of age and accommodated at the Animal Health and Veterinary Laboratories Agency (AHVLA) for the duration of the experiment. All cattle experiments were approved by local ethical review, and animal procedures were performed under a UK Home Office project licence within the conditions of the Animals (Scientific Procedures) Act 1986.

Calves were divided into six groups of seven animals and were vaccinated with the following vaccine combinations, the primary vaccination being administered to calves at six weeks of age, and a second vaccination (where indicated) at 12 weeks of age; Group 1: non-vaccinated; Group 2: BCG at week 6; Group 3: Gudair at week 6, BCG at week 12; Group 4: BCG at week 6, Gudair at week 12; Group 5; Gudair at week 6; Group 6: BCG and Gudair covaccination at week 6. One animal was withdrawn from Group 6 due to the need for the administration of postvaccination antibiotics for animal welfare reasons.

Blood and serum samples were collected at three-weekly intervals throughout the study, serum samples were stored at −80°C for later testing.

Vaccinations

BCG vaccinations were performed using BCG SSI (Statens Serum Institute, Copenhagen, Denmark) which was prepared fresh from lyophilised stocks on the day of use according to manufacturer's instructions. Vaccine was administered subcutaneously in a 0.5 ml volume (equivalent to five human doses). Retrospective culture of each batch of BCG on modified 7H11 agar (Gallagher and Horwill 1977) determined the administered dose to be approximately 1.8 × 106 units (colony forming units) per calf.

Gudair (CZ Veterinaria, Spain) is an emulsified suspension of inactivated MAP, and vaccinations were performed subcutaneously in a 2 ml volume.

Tuberculin skin tests

The SICCT test was performed according to Annexe B of the consolidated Council Directive 64/432/EEC, using avian and bovine tuberculins (PPD-A, PPD-B) (Prionics Lelystad BV, Lelystad, Netherlands). The first skin test (ST1) was performed on all animals at 18 weeks of age, and a second skin test (ST2) was performed at a following interval of no less than 42 days (median 42 days, range 42–63 days) in line with OIE recommendations on repeat testing (OIE 2011). All tests were performed by the same veterinary officer at AHVLA.

Under the standard interpretation of the SICCT skin test, an animal was deemed positive if the thickness of the reaction to PPD-B exceeds that of PPD-A by more than 4 mm. Under the severe interpretation of the SICCT, an animal was classed as a positive either if the PPD-B reaction was positive (greater than 2 mm or presence of oedema) and the PPD-A reaction is negative (no swelling or a small increase in skin thickness not exceeding 2 mm), or the PPD-B reaction exceeded the PPD-A reaction by more than 2 mm.

When applying the single intradermal tuberculin test (SIT) interpretation to SICCT data, an animal was considered positive if the thickness of the PPD-B reaction was 4 mm or more (OIE 2011).

Whole blood interferon (IFN-γ) assay

Heparinised whole blood cultures were stimulated with either PPD-A (10  ∝ g/ml), PPD-B (10  ∝ g/ml), Esat-6/CFP-10 peptide cocktail (5  ∝ g/ml), Rv3615c peptide cocktail (5  ∝ g/ml) or staphylococcal enterotoxin B (1  ∝ g/ml, Sigma-Aldrich, St Louis, Missouri, USA) as a positive control. Cultures were set up within eight hours of blood collection and incubated overnight at 37°C/5 per cent CO2. Antigen-stimulated interferon (IFN)-γ was measured in culture supernatants using the commercially available BOVIGAM ELISA (Prionics AG, Zurich, Switzerland). Results are expressed as the background-corrected optical density measured at 450 nm (OD450 nm). A positive comparative PPD response was defined as the response to PPD-B minus the response to PPD-A ≥ 0.1 OD450 nm. A positive response to the defined antigens was defined as the response to the antigen minus the nil-antigen response ≥0.1 OD450 nm.

Serological assays

Idexx M bovis antibody ELISA

The Idexx M bovis Antibody Test Kit (Idexx, Westbrook, Maine, USA) is a commercial serological diagnostic test used to detect antibodies to the serodominant M bovis antigens MPB70 and MPB83 (Waters and others 2011). Assays were performed according to the manufacturer's instructions. Briefly, test serum was diluted 1 : 50 in sample buffer and tested in duplicate. Optical densities were measured at 450 nm, and sample to positive ratios were calculated in accordance with the manufacturer's recommendations (Sample Mean − Negative Control Mean)/(Positive Control Mean − Negative Control Mean). Samples were considered positive if the sample to positive ratios exceeded 0.3.

Parachek Johne's disease ELISA

The Parachek ELISA (Prionics AG) is a commercial serological diagnostic test for JD in cattle (Cox and others 1991). Briefly, serum was diluted 1:20 and first incubated with Mycobacterium phlei to remove cross-reacting antibodies, then tested in duplicate according to the manufacturer's instructions. Colour development was read at 450 nm. A positive result was indicated if the sample absorbance value exceeded the cut-off value defined by the mean negative control value plus 0.1 (OD450 nm).

Results

Tuberculin skin-test

A summary of the SICCT skin-test responses measured at week 18 (ST1) and after week 24 (ST2) for each of the vaccine groups is shown in Fig 1. Only 1/7 of the non-vaccinated calves (Group 1) induced a measurable skin-test response to PPD-A (1 mm), and this was only at ST1. None of these control calves developed any measurable skin-test response to PPD-B at either skin test. In calves that were vaccinated with BCG only (Group 2), all calves developed a PPD-B response at ST1 (range, 3–9 mm). A PPD-B response was still measurable in each of these calves at ST2 (range, 2–10 mm). When comparing skin-test responses induced by PPD-B with those induced by PPD-A in the BCG-only vaccine group, 5/7 and 7/7 calves demonstrated a PPD-B-biased response at ST1 and ST2, respectively. For the two animals that did not demonstrate a PPD-B-biased response at ST1, the PPD-B- and PPD-A-induced responses were equivalent. By contrast, most calves in the groups that received Gudair vaccination (Groups 3–6) developed a strong PPD-A-biased skin-test reaction (Fig 1). In those calves vaccinated with Gudair and BCG, only 3/20 animals demonstrated a PPD-B-biased response at either of the skin tests, two animals in Group 3 and one animal in Group 6.

FIG 1:

Skin-test responses measured as the purified protein derivatives (PPD)-B reaction (in mm) minus the PPD-A reaction. Animals are grouped by vaccinate group, and results are shown for the first skin test (solid bars), and the second skin test (shaded bars)

Responder frequency analysis using UK standard and severe interpretations of the SICCT are presented in Table 1. Four animals in the BCG-only group would have been classed as reactors using standard SICCT interpretation at ST1, one of which was still a reactor at ST2. When applying the severe interpretation of the SICCT, 6/7 calves in the BCG-only group were reactors at ST1 with five of these animals retaining their severe reactor status at ST2. By contrast, even when applying the severe SICCT test interpretation, only 1/20 of the animals which received BCG and Gudair vaccination (in any combination) was SICCT test positive. When using the OIE interpretation of the SIT, with the exception of the non-vaccinated control group, positivity was seen in most animals in all vaccine groups (Table 1).

TABLE 1:

 ​ ​Responder frequencies by vaccination group to the tuberculin skin test

Whole blood IFN-γ responses

Antigen-stimulated IFN-γ responses were measured in whole blood using the BOVIGAM assay, as shown in Fig 2. These responses were only measured at the final sampling time point (approximately week 24) since application of the BOVIGAM assay is not recommended in animals less than six months of age due to increased non-specific IFN-γ responses in young calves (Olsen and others 2005). At the time of sampling, the median age of the calves was 162 days (range, 159–194). Elevated PPD-A- and PPD-B-induced IFN-γ responses were observed even in the non-vaccinated calves. As a consequence of elevated PPD-A responses above equivalent PPD-B responses, only 3/7 of the BCG-only vaccinated calves provided a positive test response when applying the current UK test interpretation. None of the calves that received Gudair vaccination, either alone or in ­combination with BCG, presented with a positive comparative PPD-B-biased response. Responses to the defined M bovis antigens ESAT-6, CFP10 and Rv3615c were also investigated since this combination of antigens have previously been identified as promising candidates to allow differentiation of M bovis infected from BCG vaccinated animals ­(so-called DIVA antigens (Sidders and others 2008)). Compared with responses observed to PPD antigens, responses to these DIVA antigens were minimal. Only 2/42 animals provided a positive response to the ESAT-6/CFP10 peptide cocktail, one animal in each of the non-vaccinated and BCG-only vaccinated groups.

FIG 2:

Interferon-γ (BOVIGAM) responses to purified protein derivatives (PPD)-A, PPD-B, Esat-6/CFP10 peptide cocktail and Rv3615c peptide cocktail measured at the last time point (preskin test 2) shown as optical readings at 450 nm for antigen-stimulated samples minus the corresponding nil antigen reading. Samples are grouped by vaccinate group

Serological assays

Idexx ELISA

Serum antibody responses to the M bovis antigens MPB83 and MPB70 were tested using the Idexx ELISA. Responses were measured in blood samples collected at four time points: Prefirst vaccination (week 6), presecond vaccination (week 12), prefirst skin test (week 18) and presecond skin test (week 24). Responses (OD450 nm) measured at weeks 18 and 24, representing preskin test and postskin test samples, are presented in Fig 3A. Prior to the skin test, there were no responses in any of the groups at week 18 (Fig 3A) or at the earlier time points (data not shown). However, approximately six weeks following the first skin test, detectable responses were observed in some animals in all groups except the non-vaccinated controls (Fig 3A). The strongest responses were observed in Group 3 where calves had received Gudair as the primary vaccine, and BCG as the second vaccine. Calculation of the reactor to positive ratios confirmed all animals to be negative for the first three preskin test time points (data not shown), and that some animals in all vaccinated groups induced positive responses at week 24, Group 3 demonstrated the highest responder frequency.

FIG 3:

Serological test results by vaccinate group for samples taken prior to the first (ST1, solid bars) and second skin test (ST2, shaded bars) for Idexx bovine tuberculosis serology ELISA (A) expressed as sample to positive ratio and Parachek ELISA (B) expressed as optical density readings at 450 nm. Lines indicate positive/negative cut-offs as defined by manufacturer's instructions (0.3 Idexx, 0.17 Parachek)

Parachek ELISA

Serum samples were tested using the Parachek MAP ELISA at weeks 18 and 24. Responses (OD450 nm) at these time points are shown in Fig 3B. Responder frequency analysis demonstrated that all animals in the BCG-only and non-vaccinated groups were negative. By contrast, all four groups of animals which received Gudair contained animals with positive reactions. In these Gudair vaccinated groups, there was a trend towards stronger antibody responses at the later, postskin-test bleed (Fig 3B). More than half the Gudair vaccinated cattle were positive at week 18, with more animals converting to positivity at week 24. All animals that were positive at the first time point were also positive at the second.

Discussion

In view of evidence suggesting that the killed MAP vaccine, Gudair, is being used off-licence in UK cattle, the primary aim of this study was to investigate the effect of MAP vaccination on the diagnosis of BTB. Within the scope of this study we were unable to vaccinate M bovis-infected cattle with Gudair and, therefore, used BCG vaccination to provide a diagnostic surrogate of BTB infection. A previous study demonstrated that neonatal vaccination of calves with BCG induces strong tuberculin skin test sensitisation for at least six months (Whelan and others 2011). In the current study, consistent with the expectations from this earlier study, most calves that were vaccinated with only BCG developed PPD-B-biased skin-test reactions. Furthermore, comparable numbers of calves in both studies provided positive SICCT responses at the standard test interpretation approximately 12 weeks postvaccination (ca. 60 per cent). Therefore, these data supported the use of BCG vaccination to address the objectives of the current study while once again confirming the potential of BCG vaccination to compromise the specificity of the SICCT to diagnose BTB.

Vaccination of calves with the Gudair vaccine resulted in strong PPD-A skin test sensitisation as expected. More importantly, in those BCG vaccinated animals that also received Gudair vaccine, a bias towards stronger PPD-A responses was observed. Despite all BCG-only vaccinated cattle demonstrating a PPD-B-biased skin reaction in at least one skin test, only 3/20 of the covaccinated calves still demonstrated this bias. Even when BCG was administered six weeks after Gudair vaccination, and within six weeks of the first skin test (Group 3), most animals in this group still demonstrated a PPD-A-biased response. It is already acknowledged that MAP vaccination can interfere with skin testing against BTB (OIE 2011). However, previous data on how MAP vaccination specifically affects the SICCT skin test format, which is most relevant to the UK testing situation, is very limited. While a recent study demonstrated that a killed MAP vaccine (Mycopar) did not compromise the specificity of the SICCT (Stabel and others 2011), its effect on test sensitivity was not addressed. Inglis and Weipers (1963) reported that 2/4 naturally BTB infected cattle which had been previously MAP vaccinated would have passed a SICCT three months after infection. Their study demonstrated the potential for MAP vaccination to compromise SICCT test sensitivity, albeit using an M tuberculosis-derived PPD and a vaccine strain which is no longer available. In the absence of more recent experimental studies, our data confirm that Gudair vaccination can alter a PPD-B-biased skin reaction to a PPD-A-biased one. Therefore, it seems likely that Gudair ­vaccination of a BTB infected animal might also result in a change in the comparative magnitude of PPD-B and PPD-A responses giving rise to the potential for false negative SICCT test responses, which would be consistent with the data from Inglis and Weipers (1963). Further studies in infected cattle would be of benefit to support and extend these findings.

The SIT format of the skin test is still used widely in many countries with national BTB surveillance programmes. The negative confounding effect of MAP vaccination on the specificity of the SIT has been reported previously. For example, Lesslie and others (1975) studied a cohort of BTB-free cattle which included 61 MAP vaccinated animals (vaccination strain not specified). They reported that the mean SIT PPD-B response (Rotterdam PPD) exceeded the response for the non-vaccinated group. In our study, 6/7 animals that received only Gudair provided false positive SIT reactions, dropping to 4/7 at the second test. Therefore, this study confirms and extends the body of data demonstrating that MAP vaccination compromises the specificity of the SIT for BTB diagnosis while providing novel data relating specifically to the use of the Gudair vaccine. By contrast, our results also show that the high specificity of the SICCT test for BTB is not reduced in herds vaccinated against MAP. None of the ‘BTB-free’ calves receiving Gudair at six weeks of age reacted to this test, either at the standard or the severe interpretation.

Measurement of IFN-γ in blood is a valuable adjunct to the tuberculin skin test for diagnosing BTB (Wood and others 1991; reviewed by Schiller and others 2010). While strong PPD-A-induced IFN-γ responses were observed in calves vaccinated with Gudair, comparable or PPD-A-biased responses were also observed in the BCG-only and non-vaccinated control cattle. Therefore, unlike the skin test data, we are unable to reliably conclude a confounding diagnostic effect of Gudair vaccination based on the measurement of IFN-γ responses. It has previously been reported that non-specific IFN-γ responses are more common in young calves (Olsen and others 2005). Since the age of calves ranged from five to six months at the time of the IFN-γ test, their young age may have been a contributing factor for the relatively strong PPD-A and PPD-B responses observed in our study. Interestingly, the non-vaccinated calves provided little or no skin-test response to either PPD-A or PPD-B. While it is generally accepted that the IFN-γ test is more sensitive than the skin test for diagnosing BTB (reviewed by de la Rua Domenech and others 2006), our data suggest that the reduced sensitivity of the skin test may have specificity advantages when used in young calves. We also took the opportunity of assessing the specificity of the candidate DIVA antigens ESAT-6, CFP10 and Rv3615c in the MAP vaccinated calves. Although the ESAT-6/CFP10 peptide cocktail induced an IFN-γ response in 1/7 animals in each of the non-vaccinated and BCG-only groups, encouragingly, no responses were observed in any of the Gudair vaccination groups. This was consistent with the data from Stabel and others (2011) who also recently reported low IFN-γ responses to ESAT-6 and CFP10 in calves vaccinated with a different killed MAP vaccine (Mycopar). Together they demonstrate that the specificity of these candidate DIVA antigens is not compromised by sensitisation to MAP antigens.

The MPB83-based Idexx antibody, ELISA, has been proposed as a supplemental serological test to tuberculin skin testing for BTB diagnosis (Waters and others 2011). In the current study, no Idexx-positive responses were observed prior to the skin test, but serum samples tested 42 days after a SICCT test showed that 14/34 cattle ­vaccinated with either BCG or Gudair (or both) gave false positive results. Boosting of serum IgG responses by the tuberculin skin test has been reported previously in BTB infected cattle (Harboe and others 1990, Lightbody and others 1998, Lyashchenko and others 2004). While Waters and others (2011) did not assess the specificity of the Idexx test in BCG vaccinated cattle, their study did include MAP infected cattle. They reported high specificity (>98 per cent) prior to the SICCT test which was consistent with this study but did not report a loss of specificity following the test. One possible explanation for this latter difference could be that vaccination with the killed MAP vaccine Gudair in this study resulted in stronger or more rapid serological responses compared with those observed in the experimental MAP infected ­animals used by Waters and others (2011), at least at the time points studied. Despite Waters and others (2011) reporting that test specificity was not compromised following the SICCT test, a noticeable increase in the magnitude of the post-test serological response in their cattle was observed. Regardless of the reasons for these differences, our data highlight the importance of evaluating the specificity of serological tests after a tuberculin skin test.

The Parachek ELISA has been used in the UK in JD surveillance. All the BCG-only vaccinates were negative indicating that BCG vaccination does not compromise the specificity of the Parachek ELISA. By contrast, the majority of animals which received the Gudair vaccine (23/27) gave false positive results even prior to the skin test. An increased responder frequency was observed at the postskin-test bleed, possibly as a consequence of skin test-induced IgG boosting. The observation that MAP vaccination compromises the specificity of the Parachek test is consistent with the data of Spangler and others (1991) who reported ELISA positivity developing in 13/15 vaccinated cattle over a period of two to six months following vaccination. Therefore, these data highlight and confirm diagnostic implications of MAP vaccination for JD surveillance as well as for BTB surveillance.

This study has investigated diagnostic consequences of vaccination against JD using a killed MAP vaccine, and demonstrated its potential to confound BTB diagnosis, particularly with respect to its potential for reducing SICCT test sensitivity. It is likely that these confounding diagnostic influences could also be extrapolated to cattle which are infected with MAP. This would be of particular concern in coinfected herds. Like MAP vaccinations, MAP infection will give rise to the development of strong PPD-A responses in cattle, demonstrated recently by Barry and others (2011). In MAP infected herds, these data raise the possibility that BTB infected cattle might provide false negative SICCT test responses as a consequence of elevated PPD-A responses induced by sensitisation to MAP antigens. While there is currently a paucity of data to confirm this possibility in cattle, studies in goats have clearly demonstrated that MAP coinfection can reduce the sensitivity of skin test and IFN-γ diagnostic assays to detect TB (review by Bezos and others 2012). In view of the 34.7 per cent estimated prevalence of JD in the UK dairy herds (DEFRA 2009), and the continued prevalence of BTB in the UK, the possibility of coinfection could, therefore, have significant diagnostic consequences for TB surveillance.

In summary, using tuberculin PPDs and a MAP vaccine strain which are currently used in the UK, this study demonstrates the possible confounding BTB diagnostic consequences of vaccinating cattle against JD. In particular, these data demonstrate that sensitisation of cattle by MAP antigens has the potential to compromise the sensitivity of the SICCT test for diagnosing BTB, and therefore, highlight the importance of considering the consequences of MAP vaccination and MAP infections for BTB surveillance operations.

Acknowledgments

The authors would like to thank the staff of the AHVLA Animal Services Unit for their dedication to animal welfare, and the AHVLA Laboratory Services Department, for performing the Parachek ELISA. This work was funded by the UK Department for Environment, Food and Rural Affairs (Defra project SE3227).

References>

View Abstract

Footnotes

  • Provenance: Not commissioned; externally peer reviewed

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.