Longitudinal Analysis of Prototheca zopfii-Specific Immune Responses: Correlation with Disease Progression and Carriage in Dairy Cows

Abstract

In order to characterize the humoral and cellular immune responses to bovine mammary protothecosis, serum and whey samples obtained from 72 dairy cows assigned to four different clinical stages of infection were examined for specific antibodies by indirect enzyme-linked immunosorbent assay techniques. Milk samples were analyzed for the total numbers of excreted algal cells and somatic cells. After characterization of the course of immune induction in bovine protothecal mastitis, a long-term sentinel study was performed in an affected herd in order to investigate disease progression. A total of 61 dairy cows with protothecal mastitis were examined for shedding of algae cells and for local immune responses three times in 6-month intervals. During acute and chronic stages of protothecosis, significantly elevated specific antibody activities in sera were detected. A strong correlation of whey immunoglobulin A (IgA) and whey IgG1 antibody activity with the total counts of somatic cells in milk was observed, whereas only a weak correlation of whey IgA and whey IgG1 concentrations to the number of algal cells excreted with the milk was seen. Our results from the sentinel long-term study of infected cows revealed that 70.5% of the persistently infected animals were continuously shedding the pathogen. About 4.9% of the animals showed an intermittent shedding, whereas 18% of the cows were tested culturally negative throughout the study. It can be assumed that Prototheca zopfii mastitis in dairy cows is maintained on the herd level by subclinically infected alga-shedding cows.

Protothecal mastitis in dairy cows is caused by the only known plant infectious pathogen, the alga Prototheca zopfii. Algae of the genus Prototheca are closely related to the green algal genus Chlorella but lack chlorophyll. At present, the genus Prototheca consists of four different, generally accepted species: P. wickerhamii, P. zopfii, P. ulmae, and P. stagnora. Prototheca spp. are ubiquitous and have been isolated from a variety of environmental sources, including plants, mud, sewage, various water sources (including drinking water), and soil (2, 8, 10, 11, 25). P. zopfii can also be found in the feces of various domestic and wild animals such as pigs, wild boars, horses, and deer. P. zopfii can also be isolated from healthy dairy cows in 20 to 70% of the fecal samples excreted (11).

Prototheca spp. can cause infections in cows and dogs. In dogs cutaneous, occular, enteric, and systemic infections have been observed (6, 16, 33). In addition to these animal infections, cutaneous and systemic infections of humans have also been reported. It is mainly immunocompromised patients, e.g., persons infected with human immunodeficiency virus or treated with cortisone, that are at risk. While infections in animals, however, are mainly caused by P. zopfii, human infections are predominantly associated with P. wickerhamii (4, 7, 32).

In 1952, P. zopfii was first identified as a pathogen associated with mastitis in a cow with reduced milk production characterized by thin watery secretion with white flakes (21). Although in the past only sporadic cases of protothecal mastitis have been observed, this form of mastitis is recognized increasingly today to be endemic worldwide (1, 8, 15, 17). The vast majority of cases of protothecal mastitis as a herd problem have been reported in Brazil, Germany, and the United States. Infection rates are specified between 0.2 to 3.6%, and the prevalence of infection at the herd level may be as high as 39% (3, 8, 26).

The infection depends upon predisposing factors, i.e., poor animal care and poor milking hygiene (17, 31). It has been discussed elsewhere that infections of cattle are predominantly caused by a specific P. zopfii type named variant II (3, 5). Preliminary studies showed that P. zopfii can be differentiated into at least three different biotypes biochemically and serologically (28a). In particular, all mastitis isolates were found to be unique and could all be assigned to biotype (variant) II of P. zopfii. Infection with P. zopfii is associated with an acute to chronic granulomatous mastitis causing a reduction of milk production and often subsequent atresia of the infected quarter. Due to their ability to infect and survive in macrophages and to invade the udder tissue, the pathogens can induce a persistent infection with intermittent shedding (29). Since P. zopfii is highly resistant to all known chemotherapeutic agents, infected cows should be rapidly removed from the herd. Therefore, reliable identification of infected cows by plate culturing of milk specimens or enzyme-linked immunosorbent assay (ELISA) is necessary in order to reduce the risk of infection of uninfected cows or contamination of the farm environment (28).

Our knowledge of local and systemic humoral and cellular immune response in Prototheca mastitis is limited. Only few efforts to develop serodiagnostic tools have been made. The presence of specific immunoglobulin A (IgA) antibodies in whey material from lactating cows could be demonstrated by immunodiffusion. These antibodies persist during the dry standing period and could be detected also in the next lactation (9). Both counterimmunoelectrophoresis tests and ELISA-based detection of anti-Prototheca IgG in serum showed poor sensitivity and specificity and thus are not applicable for routine use (18). Recently, a newly developed ELISA based on the detection of IgA and IgG1 in whey was found to be suitable for discriminating infected and noninfected animals with high sensitivity and specificity (28). Other than these reports, there are no further data available about the humoral and cellular immune responses or the rates of intermittent or persistent shedders in bovine Prototheca mastitis.

The aim of the present study was to investigate the systemic and local antibody responses to P. zopfii during several stages of infection and to correlate these findings with both the number of algae excreted with milk and the somatic cell count. We were able to demonstrate that the majority of all affected animals are unable to eliminate the pathogens from their udder and therefore became persistent or intermittent algae shedders.

MATERIALS AND METHODS

Algae strains.

Reference strain SAG 2021 of P. zopfii was obtained from the Culture Collection of Algae at the University of Göttingen, Göttingen, Germany, and was used as coating antigen for the various ELISAs (28). SAG 2021 was originally isolated from a case of a severe acute mastitis in a lactating cow.

Sample collection and preparation.

Milk samples were collected under sterile conditions. For extraction of whey, milk samples were mixed with 0.1 mU of chymotrypsin (Rennin; Fluka, Ltd., Taufkirchen, Germany) per 10 ml of milk. After an incubation of more than 30 min at 37°C, the samples were centrifuged and stored at 4°C overnight to separate the fat and the casein from the whey.

Classification of P. zopfii infected cows.

A total of 72 animals out of a herd with a history of either acute, chronic, or subclinical Prototheca zopfii-infection were assigned to groups by using discriminatory clinical and cultural properties as outlined by Roesler et al. (28) as follows: group 1 included 12 cows with signs of acute mastitis and positive cultural findings (MA, P⁺); group 2 included 15 cows with subacute to chronic clinical symptoms and positive cultural findings (MC, P⁺); group 3 included 15 cows with chronic clinical signs, an earlier positive cultural finding but with negative cultural detection of pathogens at the time of the present study (MC, P⁻); and group 4 (M⁻, P⁻) included 30 cows without clinical and cultural signs of Prototheca mastitis (Table 1).

TABLE 1.

Local and systemic isotype-specific antibody responses to P. zopfii in naturally infected cows

Group	Clinical and cultural characteristics	No. of cows	Antibody activity (EU) ina:
			Serum IgGb		Whey IgG1		Whey IgA
			Mean ± SD (median)	Range	Mean ± SD (median)	Range	Mean ± SD	Range
1	MA, P+ (acute mastitis, culture positive)	12	180.7 ± 78.7 (159.5)	200	68.7 ± 51.1 (56.0)	140	76.3 ± 59.8 (78.0)	173
2	MC, P+ (chronic mastitis, culture positive)	15	91.9 ± 46.7* (81.0)	192	57.2 ± 68.7 (27.0)	213	97.2 ± 74.3 (122.0)	192
3	MC, P− (chronic mastitis, culture negative)	15	23.7 ± 9.7† (21.0)	28	12.7 ± 35.7† (0.0)	139	9.5 ± 17.2† (2.0)	50
4	M−, P− (clinically healthy, culture negative)	30	18.0 ± 6.4‡ (17.0)	26	0 ± 0‡ (0.0)	0	0.1 ± 0.4‡ (0.0)	2

^aAntibody activities are expressed as the means and standard deviations, median, and range for all stages of infection. Significant differences between group 1 and group 3 (*), group 2 and group 3 (†), and group 2 and group 4 (‡) are indicated (P < 0.05).

^bIn serum samples, specific IgA and IgM antibody activities were not detectable.

Course of P. zopfii mastitis.

A total of 61 animals from a second infected herd were included in a sentinel long-term surveillance in order to characterize clinical, serological, and bacteriological parameters typical for bovine protothecal mastitis. All cows were found to be culturally positive for P. zopfii when they were tested the first time. Second and a third examinations of the milk samples were performed 6 and 12 months later, respectively. In addition to cultural examination of milk samples, whey samples prepared from milk derived from all animals were examined for specific IgA and IgG1 antibodies.

Microbial examination of milk samples.

Milk samples were handled and cultivated as described previously (28). Briefly, aliquots of 50 μl from udder quarter milk samples or composite milk samples were streaked onto Sabouraud-dextrose-agar plates (Difco Laboratories, Detroit, Mich.) and then incubated at 37°C under aerobic conditions. After 72 and 120 h, the plates were examined for the growth of Prototheca. Each colony indicative for Prototheca sp. was subcultured once. Smears were made from colonies of interest and stained with lactophenol cotton blue. Specimens were investigated microscopically for characteristic morphological criteria such as the presence of sporangiospores in the sporangium. P. zopfii isolates tested positive for glucose and glycerol and negative for saccharose and trehalose assimilation (28); these tests were used as determinative biochemical characteristics.

Furthermore, the numbers of CFU per milliliter excreted with milk were determined by dilution series.

Cell counting.

All milk samples were subjected to a somatic cell count by using the automated cell-counting system Fossomatic 15600 (A.S.N. Foss Elektric, Hillerod, Denmark) (19, 27).

ELISAs.

All ELISA procedures applied here have been recently described (28). In brief, strain SAG 2021 was used as whole-cell antigen for indirect ELISA. A serum pool of 15 Prototheca-positive tested cows with chronic manifestations of mastitis was used as the positive reference serum. The negative control consisted of a serum pool of 10 clinical healthy and culture-negative cows from a herd without any history of Prototheca mastitis. Anti-bovine IgG, anti-bovine IgG1, and anti-bovine IgA (Bethyl Laboratories, Inc., Montgomery, Tex.) were used as horseradish peroxidase-conjugated polyclonal conjugates.

ELISAs and statistical analyses.

Extinction data were calculated by using a specially developed computer program for ELISA evaluation by using positive reference standard procedures (23). Relative values for antibody activities in the positive standards were set to a level of 100 ELISA units (EU).

Significant differences in antibody activity, in somatic cellcount, and in alga titer within milk between infected cows at various clinical stages and uninfected animals were tested by Student t tests for unpaired observations or by Welch's test. A P value of ≤0.05 was considered significant. The data were calculated and plotted as notch boxes (22). The median, the upper and lower quartiles, the 95% confidence limits, and the extreme values are shown (22).

In order to determine a correlation of specific antibody concentrations with the amount of excreted somatic cells and with the quantity of excreted pathogens, Spearman correlations with one-tailed significance were calculated by using the SPSS computer program.

RESULTS

Antibody responses.

Analyses of whole-cell antigen ELISAs showed the induction of specific IgA and IgG1 antibodies against P. zopfii in serum and whey samples from affected cows. Antibody concentrations of the different antibody isotypes were found to vary between cows depending on the clinical stage of infection (see Table 1). Specific activities of IgG1 in whey in acutely infected cows (68.7 EU) were in the mean slightly higher than the antibody activities measured in chronically infected culture-positive animals (57.2 EU). Specific whey IgA antibody activities against P. zopfii in acutely infected cows were decided lower than specific IgA concentrations in chronically infected but culture-positive animals. Chronically infected but still culture-negative animals showed activities (12.7 EU at whey-IgA) that did not significantly differ from the antibody activities of healthy cows at 0.1 EU. P. zopfii-specific systemic IgG activities during infection were found to be elevated to levels comparable to those of local IgG1 in the udder. Healthy animals showed a basic IgG1 activity of 18 EU.

Cellular infiltrations.

The numbers of somatic cells in milk were greatly increased in cows suffering from acute mastitis (P < 0.05) compared to chronically infected cows. The content of somatic cells in milk from chronically infected but still culture-negative animals, however, was not significantly increased (Fig. 1). The numbers of pathogens excreted with milk from acutely infected cows were higher than from chronically infected cows, but substantial variation in the total number of CFU per milliliter could be measured.

FIG. 1.

Numbers of algae (A) and somatic cells (B) excreted with milk dependent upon the clinical stages of bovine P. zopfii mastitis. Group 1 (MA, P⁺) included cows with clinical signs of acute mastitis and positive cultural P. zopfii findings. Group 2 (MC, P⁺) included cows with subacute to chronic clinical symptoms of mastitis and positive cultural P. zopfii findings. Group 3 (MC, P⁻) included cows with chronic clinical signs and an earlier positive cultural P. zopfii finding but culture negative for P. zopfii at the time of the present study. Group 4 (M⁻, P⁻) included cows without clinical signs of mastitis and culture negative for P. zopfii. A significant discrimination (P < 0.05) of the MA, P⁺ group versus the MC, P⁺ group is indicated by the symbol “#.” The dagger symbol (†) indicates significant differences (P < 0.05) between the MC, P⁺ group and MC, P⁻ group. A significant discrimination (P < 0.05) of chronically infected cultural positive versus noninfected cows is indicated by the “§” symbol. Asterisks indicate that P. zopfii could not be detected by plate culturing.

Correlation of immune responses with excreted algae.

Data analyses of antibody responses and somatic cellcounts showed a strong correlation of the IgA and IgG1 activities (r = 0.637 and 0.618, respectively) with the total number of somatic cells occurring in milk (Fig. 2 and Table 2). The correlation of somatic cells with IgG in serum was significantly weaker at r = 0.355. Whey IgA antibody responses correlated more strongly with the number of excreted algal cells (r = 0.470) than with the other isotypes. The correlation between the somatic cell count and the number of excreted pathogens was calculated to be r = 0.672 (Table 2).

FIG. 2.

Correlation between specific antibody responses against P. zopfii in serum and whey samples of infected dairy cows on the number of CFU (left) and of the somatic cell counts in milk (right). Regression lines are depicted. (A and D) Serum IgG; (B and E) whey IgA; (C and F) whey IgG1.

TABLE 2.

Correlation of specific antibody responses versus the number of excreted algae and the count of somatic cells in milk of P. zopfii-infected dairy cows

Sample type	Excreted algae		Somatic cells
	No.	P	No.	P
Serum IgG	0.423	<0.05	0.355	<0.01
Whey IgA	0.470	<0.01	0.637	<0.001
Whey IgGa	0.408	<0.05	0.618	<0.001
Somatic cells	0.672	<0.001

Determination of carriage and shedding.

The long-term sentinel observations for the shedding of algae and the induction of antibodies indicate that the most of the infected animals were not able to eliminate the algae, since only 18% of the once-infected cows did not shed Prototheca at the third examination. Nevertheless, specific antibodies were measured in these animals at a rate of 45.5%. The algae could be cultured from milk samples obtained from ca. 70.5% of the affected animals at every time point, indicating that these cows were persistently infected animals. A total of 4.9% of the cows were culture positive at the first and third examinations but not in the second; they were therefore classified as intermittent shedders. Interestingly, 14.8% of all infected dairy cows could be identified as Prototheca positive by ELISA 6 months before they first tested positive by plate culturing. About 6.6% of the animals tested as serologically positive but were culture negative at each examination.

DISCUSSION

The results presented here indicate that P. zopfii-induced mastitis in dairy cows is maintained on the herd level by clinically healthy shedders. To our knowledge this is the first report on bovine protothecosis that documents disease progression and the frequency of carrier animals at the herd level by analyses of milk and serum samples.

Antibody responses.

As a predominant result, local and systemic immune responses induced by P. zopfii are (significantly) dependent upon the stage of infection. Acutely diseased animals showed the highest specific IgG and IgG1 activities in serum and whey samples, respectively. However, specific local IgA concentrations in milk of chronically infected animals were higher than in acutely infected cows. This may be due to a slowly increasing IgA response during the course of protothecal infection. This hypothesis is supported by the results obtained from chronically infected cows. These animals had an increased specific antibody activity in serum, as well as in whey, whereas chronically infected but culture-negative cows did not consistently show elevated specific antibody concentrations. Thus, ca. 50% of the latter animals could not be distinguished serologically from healthy animals. It can be assumed that some of these cows succeeded in eliminating the algae. Hence, a lack of specific antibody responses in these cows indicates convalescence. The results for the remaining approximately 50% serologically positive animals (IgA ELISA cutoff of 1 EU) can be explained by the persistence of the infection and the intermittent shedding of algae (28). These disease progression-related findings could be confirmed by the results from our long-term investigation at the herd level. It was already known that animals and humans develop specific IgA antibodies during persistent infections (13, 14, 20). However, further studies in experimentally infected cows should clarify the role of serologically positive but culture-negative cows for the kinetics of the infectious algae on herd level.

Cellular infiltrations.

The investigation of local cellular immune responses in the infected bovine udder revealed significant differences in the total cell influx at various clinical stages of protothecosis. The weak correlations between antibody activities and the number of excreted algal cells can be explained by the nonhomogeneous distribution of the pathogens in the udder tissue. P. zopfii tends to invade the udder tissue and to congest the alveoli and ductuli within the udder (17). Hence, the number of excreted algae in milk does not correlate with the total number of algae present in the udder. It can be assumed that the net influx of leukocytes in cases of bovine protothecosis is antigen induced and, therefore, the count of somatic cells strongly correlates with the total number of excreted algal cells. However, migration of inflammatory cells into the area of infection is inhibited by the algae, as demonstrated for P. wickerhamii in cases of canine cutaneous protothecosis (24). It is known that a significant part of leukocytes in milk are antibody-secreting cells. This might be the reason for a significant correlation between somatic cells in milk and antibody concentrations in serum and whey. Since chronically infected dairy cows have been found with a significantly elevated content of somatic cells in milk, this clinical microbiological criterion can be used to control P. zopfii mastitis in affected herds.

Shedding and carriage.

In the present study, clinical microbiological characteristics of bovine protothecal mastitis have been evaluated and correlated with disease progression. Using these parameters in a long-term sentinel study of an infected diary herd, we showed that only a few cows (18%) were able to terminate the infection. However, since 45.4% of these animals were found to be positive for specific antibodies in whey and serum at the last examination, the number of truly convalescent animals is probably still lower. Therefore, it can be speculated that persistently infected intermittently shedding animals are part of this group. During disease progression a majority of 70.5% of infected cows were identified as permanent shedders of the algal cells. About 21% out of this group of shedders showed a specific antibody response already 6 months before the first positive cultural result. It can be speculated that these animals might also be intermittent shedders. Since 4.9% of dairy cows have been identified as intermittent shedders by culture during a time frame of 12 months, at least 75% of all of the animals investigated could be regarded as persistently infected. Cows with elevated specific antibody concentrations in whey combined with negative cultural results at all examinations (6.6%) could probably also be assigned to the group of persistently infected cows, since these animals might shed the algae intermittently and the algae might persist within the udder tissue (e.g., in granulomas or abscesses) or in macrophages (12, 18, 30). All of these findings may lead to an underestimation of the total number of intermittent shedders in routine diagnostics.

Conclusions.

In summary, our findings underline some typical characteristics of bovine protothecal mastitis as a chronic, persistent infection of the bovine mammary gland. It is obvious that only a small number of lactating cows can effectively terminate an infection with P. zopfii. Under natural conditions during disease progression the algae are intermittently shed into the milk by a significant number of the infected diary cows. Since P. zopfii infection induces a specific local and systemic antibody response, serological examination can be used to identify acutely and chronically infected animals, as well as intermittent shedders. In the future, a combination of antibody testing and somatic cell counts in milk should increase the sensitivity for diagnosis of bovine protothecal mastitis in affected herds. Experimental infection studies are needed to further elucidate the roles of the number of algae in the inoculum and of the route of infection on the severity and progression of bovine protothecal mastitis. This may in turn lead to eradication strategies that enable sustainable elimination of the pathogen from an infected herd.