Searching for the cause of mare reproductive loss syndrome: Kentucky’s $300 million disaster

In late April and throughout the spring of 2001, what appears to have been a series of catastrophic events resulted in abortions, heart and eye problems in the horse industry of Kentucky, West Virginia and southern Ohio. The problem was termed ‘Mare Reproductive Loss Syndrome’ (MRLS) for the number of early and late-term aborted fetuses.

Estimates of losses from this syndrome go as high as 4000 foals with a University of Louisville study requested by the governor’s office estimating that about 31% of the 2002 thoroughbred foal crop and 20% of the standardbred foal crop for this year were lost to this syndrome (although all breeds of horses were affected). The economic impact of MRLS has been speculated to be between $225 and 300 million. The exact cause remains unknown at this time, but a number of theories exist and continue to be pursued.


The calm before the storm: springtime in the Bluegrass

Following a winter in which November and December were extremely cold, January 2001 temperatures remained in the normal range and February temperatures escalated to about 5ºF above normal. March temperatures did a complete reversal, with temperatures more than 5ºF below normal. During this time, precipitation remained low with the exception of December and February. April of 2001 began unusually dry and warm with temperatures reaching 80ºF and above in Lexington from April 7 through April 12. However, the temperature dropped off rather dramatically on April 17 and 18 with night-time temperatures falling below 30ºF. Accompanying this drop in temperature was a snowfall of approximately 2 inches in parts of the region. Weather data provided by the University of Kentucky Department of Agricultural Engineering for the month of April confirmed the presence of frost on the nights of April 17 and 18. Rainfall was low for the month overall. This weather pattern was a mirror image of weather data observed during April 1981 when a similar syndrome to the MRLS of 2001 was reported in the Bluegrass region of Kentucky. Following an 8-day period of mild temperatures the thermometer again topped 80ºF to peak on May 6 at 89ºF.

This unusual weather pattern, together with the similar pattern observed during the problems encountered in 1981, was thought to somehow cause the problems seen in 2001. Because early April temperatures and the rise in temperature after the frost and through the month had been higher than normal, conditions for pasture growth were ideal.

Indeed, many pastures in the Bluegrass region had lush growth and trees were blooming, but an abnormal increase in clover growth throughout the region also was noted. An additional observation was an unusually heavy concentration of Eastern Tent Caterpillars. In some areas, the ground, wild cherry trees and horse water systems were covered with tent caterpillars and many of the wild cherry trees that these creatures feed upon were stripped of observable foliage.

From the period of April 26 through May 1, area veterinarians observed abnormally high numbers of early fetal loss from ultrasound examinations. The beginnings of what was later to be dubbed MRLS were appearing in the Bluegrass region.


And then the storm

The first Saturday in May is the traditional time for all eyes to be focused on the state of Kentucky because of the ‘Run for the Roses’ – the celebrated Kentucky Derby. In 2001, the eyes of the nation were indeed focused on Kentucky’s horse industry, but MRLS was stealing the show from the first leg of the Triple Crown. By this time, the Kentucky horse industry, equine practitioners, private consultants and University of Kentucky equine faculty and staff were spending sleepless nights wondering what was causing the foal losses and how it might be stopped.

On May 10, a meeting was held at the Keeneland Sales Pavilion in Lexington to inform the horse industry and media of the status of diagnostic and analytical testing and hypotheses as to the cause of the syndrome. More than 1000 people attended the meeting, which was also simulcast on the internet. At this meeting, it was revealed that diagnostic testing had failed to find a viral or bacteriological source for the syndrome. Although Streptococcus and Actinobacillus species had been isolated from a high proportion of late fetal loss (LFL) cases examined at the state’s Livestock Disease and Diagnostic Center, they were considered secondary or opportunistic infections that followed the initial insult.

From April 28 to noon on May 8, the state diagnostic lab received 318 fetuses and stillborn foals. By the close of the business day on May 8, an additional 28 fetuses and foals were submitted. Twenty-five more were added on May 9 and another 15 on May 10, bringing the total to 386 by 2 p.m. May 10. In total, nearly 500 late-term pregnancy losses and approximately 2000 early fetal losses were reported.


Defining MRLS

The symptoms observed in MRLS were unique and led primarily to drawing similarities to other known processes causing abortion and a search for common factors among affected horses. Although reports indicated that abortion was by far the most consistent symptom, in the case of MRLS the aborted early fetuses were observed almost exclusively in mares 40 to 100 days into pregnancy. In these mares, abortion was preceded by fetal death followed by expulsion of the dead fetus.

According to veterinarians, abnormal echogenic fluid (cloudy and flocculent), was prevalent with a mucoid discharge from the mare associated with the fetal membrane. In some cases lethargy and loss of appetite were noted with no manure passage for 18 hrs. Late fetal losses occurred in mares that were near term – about three weeks prior to expected parturition date. Gross and histopathological lesions indicated a relatively consistent pattern in characteristics of the fetus and amniotic sac of the placenta. The lungs of the fetus showed evidence of intrauterine toxic insult and inflammation, as did the amniotic sac. Foals born alive were quite weak and frequently required intensive veterinary care; in many cases these foals suffered from respiratory distress. Toxicological results on fetuses were negative for nitrates and nitrites and serum levels of copper, iron, zinc and selenium in the mare were normal. Virological tests for equine herpes virus, equine arteritis virus and adenovirus as well as tests to discover an as-yetunknown virus also proved negative.

Hidden among the many problems recorded in reports from veterinary clinics in the first two weeks of May was an increased incidence of pericarditis (40) cases and unilateral panophthalmylitis (13) cases across a wide age range of horses and breeds.

In some cases as much as 5 gallons of fibrinous fluid had to be removed from the sac surrounding the heart. The eye problems began with cloudy, fibroid material in the anterior portion of the eye. In yearlings and mature horses, this endoopthalmitis appeared in one, but never both eyes. In these cases, the eye was not responsive to treatment and irreversible blindness in the affected eye resulted. In some foals born during this timeframe, the lesions involved both eyes, but was resolved in surviving foals. Studies on the epidemiology of MRLS and pericarditis implicate a point-source insult.

Information on the relationship of the eye problems to the other presentations of MRLS are currently not available.

At first, the entire abortion episode seemed to be quite clearly defined. Mares were aborting and a number of samples were submitted for analysis of the mycotoxin zearalenone. This Fusarium mycotoxin has been found to be associated with pasture grasses in Australia and New Zealand.

Zearalenone has estrogenic properties although it is not chemically an estrogen. For this reason, females are more susceptible than males. In swine, a swollen, red vulva, which may lead to rectal and vaginal prolapse, is normally only seen in prepubertal gilts. In MRLS, maiden and previously barren mares seemed to be more susceptible to the syndrome. A number of studies in other animal species have demonstrated the effects of zearalenone over extended periods of time; however a more probable scenario under field conditions (and in examining MRLS) is exposure to a dosage of toxin over a short period of time. Studies with 16 gilts given pure zearalenone (108 mg on post-mating days 2-6, 7-10 or days 11-15) showed that gilts given zearalenone on post-mating day 7-10 had reduced embryonic survival compared to control or other treatment groups (1/4 pregnant in the 7-10 day group vs 4/4 pregnant in control and other treatment groups). FSH and estradiol-17β concentrations were unaffected by zearalenone consumption. Serum concentrations of prolactin in the 7-10 day and 11-15 day groups were lower than controls or the days 2-6 postmating group (Long and Diekman, 1986). Because of the relationship between effects of zearalenone and stage of pregnancy, this toxin may have played a critical role in the events described in the spring of 2001. This may account for the observation of affected mares being 40-100 days pregnant or near term.

The early results demonstrating the presence of zearalenone were determined by commercial Enzyme-Linked Immunosorbent Assay (ELISA) kits. These kits are not validated for testing forage, caterpillar and caterpillar frass (excrement) samples. ELISA procedures are prone to false positive results on non-validated matrices and should be confirmed using HPLC and/or TLC testing. When confirmation of these positive results was requested, one of the major problems was the availability and quantity of samples from the critical timeframe. However, no Fusarium toxins were detected in any of the samples submitted for HPLC analysis. What was once a simple task of validation of what was already suspected turned to a, “what do we do now?” situation.

The next likely theory was that fescue toxicosis had appeared early due to the drought conditions and the overall strange weather pattern. The characteristics of fescue toxicity as we know them include placental edema (‘red bag’), prolonged gestation (sometimes as long as 13 months), dysmature foals with long limbs, poor muscular development, agalactia, overgrown hooves, retained placenta and excessive hair growth. In the spring of 2001, the late fetal loss cases showed no placental edema and tended to be three weeks early to term with essentially no dysmaturities. Although the mares characteristically did not have milk at the time of fetal loss, they did start lactating after parturition. A number of fescue samples were tested and found to contain relatively low levels of ergot alkaloid, but very few mares demonstrating MRLS were in pastures with fescue and concentrations of ergot alkaloids in other forage sources were not observed.


Enter wild cherry trees and cyanide

Black cherry tree (Prunus serotina) leaves are known to be toxic to cattle and horses, especially if the leaves have wilted from storm damage, trampling or frost. In the investigations of MRLS, when scientists went on-farm to examine the problem and take samples, cherry trees were observed in close proximity to fields where affected mares were grazing. Also noted was the heavy infestation of Eastern Tent Caterpillars in these trees. In many cases, these trees were virtually devoid of foliage and caterpillars were in the grass.

This observation was commonly made by participants in the farm survey conducted throughout the region. In most instances, a good correlation was observed between the presence of cherry trees and Eastern Tent caterpillars.

Black cherry trees contain the cyanide precursor prunasin, which, when hydrolyzed through tissue damage to the leaf, yields hydrocyanic acid (Poulton and Shin, 1983). Cyanide poisoning is classically manifested as a rapidly acting toxin resulting in hypoxia (deficiency of oxygen reaching the tissues).

The first symptoms appear within a few minutes following consumption of plant material. Affected animals typically exhibit excitement, incoordination, convulsions, rapid and labored breathing, bloating, and coma. Death normally occurs in less than an hour due to internal asphyxiation. Normally, the odor of almonds is noticed at the time of necropsy. The cause of asphyxiation is due to cyanide combining with the cytochrome oxidase system and essentially binding to the ferric ion of hemoglobin so the cell can no longer transport oxygen. A working hypothesis evolved for the cause of MRLS in that levels of cyanide or cyanogenic compounds were ingested or inhaled and subsequently interfered with fetal oxygen availability causing death of the fetus and expulsion by the mare of the dead fetus. Tissue samples were submitted for cyanide analysis with a positive tissue sample being detected. The cyanide in tissue is the first ‘smoking gun’ that the group involved in these studies had to work with; and cyanide or cyanogenic compounds move to the top of the list of suspected agents.

On May 24 another public informational meeting was held to reveal these findings. Because of the known potential of black cherry trees to produce cyanide toxicity, and the frost conditions that occurred in April, these trees and the caterpillars that feed on them were implicated as the probable source of the cyanide-containing compounds. Results of the epidemiological survey showed that risk factors associated with MRLS included cherry trees in and around pastures, cherry tree seedlings in the pastures, deciduous trees stripped of their leaves in pastures and moderate to heavy presence of caterpillars. Since Eastern Tent Caterpillars consume the leaves and effectively convert the cyanogenic compounds to amino acids, detectable levels of cyanide in the frass were not seen. Questions arose as to how something that is traditionally thought of as a poison could only exert toxicity on a developing fetus and not kill the mare. Another unanswered question was why were cattle unaffected given that ruminants have been documented to be more susceptible to cyanide poisoning than monogastric animals. None-the-less, cyanide found in fetal tissues remains to date, the best direct evidence of a toxic insult that may have been responsible for MRLS. Wild cherry trees near horse pastures throughout the Bluegrass state were targeted for removal. The role of the caterpillars in this hypothesis are addressed elsewhere, and therefore will not be examined here.


Other suspects

HEMLOCK

While cyanogenic compounds were moved to the front of the suspect list, the investigation into other possible agents continued as intensely as before the cyanogenic compound was found. Field observations included hemlock (Conium maculatum) in pastures; however initial testing at the state diagnostic center could find no traces of hemlock derivatives in any of the samples examined. At the same time, investigators from South Carolina arrived in Lexington and more closely scrutinized the hemlock to find physical signs of horses grazing the material, and correlated these results to MRLS-affected farms and pastures. As of this writing, these investigators show a strong correlation to hemlock presence and MRLS, although a direct link is not suggested at this time. The predominant hemlock alkaloids, coniine and γ-coniceine, are primarily known for exerting teratogenic effects in many domestic animal species due to the ability of these compounds to block nicotinic receptors. Normally, toxicity is observed in a short period of time when respiratory muscles are affected. Acute toxicity that is not lethal resolves spontaneously if the animal is prevented from further exposure. It has been observed, however, that animals tend to return to this plant to feed. Chronic toxicity seems to be reserved for pregnant animals. During fetal organ development, the offspring are born with malformations and multiple congenital contractures (Lopez et al., 1999). Cleft palates were observed in piglets from gilts fed Conium maculatum seed or plant material between days 30 and 45 of gestation (Panter et al., 1985).

Hemlock has been present in Kentucky pastures for uncountable years; and getting horses to eat the material in scientific studies has been the bane of more than one investigator. It is possible that changes in the plant composition from frost damage may make the plant more appealing to the horse in a manner similar to the sugar release seen in cherry tree leaves with frost, but this is speculation. It was also suggested that growth of other pasture forages was limited and that growth of hemlock, which initiates growth earlier in the season, was more pronounced. However it was in fact a banner year for spring forage growth in the Bluegrass and many area farmers had baled their first cutting of grass hay before the Kentucky Derby.


ESTROGENS OR ANTI-ESTROGENS

Because of the widespread abortive effects associated with MRLS, the role of compounds that might alter mare hormonal levels is also being considered. In addition to zearalenone, phytoestrogens found in soy, alfalfa and clover have been shown to exert estrogenic properties in mammals (Barrett, 1996). As mentioned earlier, white clover growth was unusually good in the spring of 2001. The possibility exists that estrogenic compounds or anti-estrogenic compounds may have played a role in MRLS. However, some of the prominent features of estrogenic ‘overload’ would be changes in behavior, increased edema of the uterus and changes in heat cycles. None of these symptoms were reported; and samples of clover tested (from a later time period) contained low levels of known phytoestrogens. These results do not preclude the possibility of higher levels of phytoestrogens being present in the middle of April.

Antiestrogens interfere with the action of estrogen and are also known as agonists. It seems logical that, when considering the possible role of estrogenic compounds in MRLS, compounds that are agonists to estrogen should also be considered.


STREPTOCOCCI AND ACTINOBACILLI

Bacterial-induced abortion is not a new concept. In approximately 70% aborted late-term fetuses in spring of 2001, streptococci and/or actinobacilli were present. As mentioned earlier, these organisms were thought to be the result of secondary infections after the primary insult. The microaerophilic nature of these organisms led some investigators to consider the possibility that the hypoxia associated with exposure to cyanogenic compounds may have provided a hospitable environment for microbial activity. Unknown to researchers at this time is how these organisms, if exogenous, got to the fetal tissue (normally a sterile environment), or whether these organisms are endogenous intestinal or oral microflora in the adult horse.


CATION IMBALANCE

Another suggested explanation for MRLS was that high levels of potassium were present in the forage as a result of the frost experienced on April 17 and 18. The imbalance caused by a high potassium diet was thought to have resulted in intestinal leakage (explained by the recovery of the above bacterial strains), immune suppression and reproductive loss. It was further suggested that the high potassium concentrations favored the growth of streptococci. However, the suggestion that the high potassium diet favored growth of streptococci is quite flawed.

The reference for this bit of misinformation is growth of streptococci in the presence of the ionophore gramicidin D (Kakinuma, 1998). Since the function of ionophores is to cause cation imbalance across the bacterial membrane, it stands to reason that growth in media containing high potassium would favor survival of the organism. It is also worth noting that a number of the soil and forage samples submitted for mineral analysis had levels of potassium at or near normal.


MOLDS AND MYCOTOXINS

In a number of situations where unexplained phenomena occur, mycotoxins are blamed either rightly or wrongly. The possibility that the forage ecosystem was affected by the dramatic change in weather resulting in a more favorable environment for mold growth is also being explored. A theory put forth recently involves a combination of mold and mycotoxin being responsible for the various clinical manifestations observed during the spring of 2001. Mycotoxins are not transmissible from animal to animal, and do not respond to drug and antimicrobial therapy. In addition, most myocotoxins are toxic, but not lethal, at environmental levels (example: zearalenone has an LD50 greater than 16 g/kg BW).

It is well documented that molds can grow at a wider range of pH values and lower levels of available water than bacteria (Jay, 2000). In addition, many molds can grow at temperatures near or below freezing, or be stimulated to produce toxins by these temperatures. Aspergillus is traditionally considered a tropical environmental contaminant of feedstuffs, however, growth and toxin production in northern Europe and Canada is not uncommon. The mycotoxin-producing species of Aspergillus need similar moisture levels to bacteria for growth. Temperature also plays a role in toxin production; and Schindler (1977) found that 25 different strains of A. flavus and A. parasiticus did not produce aflatoxin at 2oC, 7oC, 41oC or 46oC. Northolt and coworkers (1976) found relatively low growth rate of A. parasiticus at 10-13oC, but relatively high concentrations of aflatoxin were produced at these temperatures when available water was high.

Because of the conditions required for toxin production and a lack of symptoms consistent with aflatoxicosis, this mycotoxin was discarded from consideration in association with MLRS very early in the process. However, Aspergillus and Fusarium are two of the most frequently isolated organisms in cases of ulcerative keratomycosis (Andrew et al., 1998). It should be noted, however, that eye problems seemed to have started from the rear of the eye and worked forward with no external ulceration observed. This observation is inconsistent with traditional fungal eye infections. In addition, endocarditis has been demonstrated from aspergillosis (Pace et al., 1994). The type of pericarditis observed in horses in the spring of 2001 is similar to histoplasmosis-induced pericarditis in humans, but histoplasmosis titers were negative in all the cases observed at a local veterinary clinic (Reimer, personal communication). In a histoplasmosis outbreak in Indiana, pericarditis was observed and fungal cultures of tissues and pericardial fluid were negative, but serological studies provided the basis of the diagnosis (Wheat et al., 1983). It is tempting to consider that exposure to another fungal organism could cause similar symptoms in horses. There are also numerous references to mycotic abortions caused by Aspergillus species, but in none of the cases from April 2001 were fungal organisms isolated from tissues of affected horses. Moniliformin (a Fusarium mycotoxin) has been shown to cause fluid accumulation in the pericardium of poultry, but only at relatively high dietary concentrations and a symptomatic match to MRLS is questionable.

Forage samples taken from the critical time period are scarce; however, Fusarium and Aspergillus species are generally considered ubiquitous in soils and have been present in a high percentage of samples taken since May 2001. In our own studies, caterpillar frass has also been shown to be an excellent growth media for many molds and may explain the correlation between caterpillars, molds, mycotoxins and MRLS.

T-2, deoxynivalenol (DON), diacetoxyscirpenol (DAS), Fusarenon-X and fumonisin have all been shown to be both fetotoxic and to cause abortion. Lesions of the mouth and(or) intestinal tract have also been observed with nivalenol, DAS and T-2 (D’Mello et al., 1999; Sklan et al., 2001). These lesions may help to explain the presence of streptococci and actinobacilli in many cases of late term abortion. Fumonisin is reported to cause unilateral blindness, but this observation is suspect in trials conducted with purified fumonisin (Haschek, personal communication). The role of fumonisin in MRLS is also questionable, since by far the largest collection of scientific studies on mycotoxins in horses involve fumonisin (Ross et al., 1992; Schumaker et al., 1995; Wilson et al., 1990). In addition to Fusarium toxins, Penicillium, Alternaria, Acremonium, Claviceps, Stachybotrys and Aspergillus (other than ochratoxin and aflatoxin), produce mycotoxins that are under varying degrees of investigation as possible agents for MRLS.

When considering molds and mycotoxins as possible causes for MRLS, it is important to realize that presence of molds, fungal growth, or small amounts of mycotoxin are not proof of mycotoxicosis. In many cases, the responsible toxin(s) may have been consumed or undetectable by the time that clinical symptoms appear. Detection of toxins in fluids or feces would be considered as strongly implicating in the presence of symptoms.

It is also important to realize that most mycotoxins in nature do not exist alone and synergism exists between many toxic compounds, which further exacerbates the problem of diagnosis.


Conclusion

The exact cause of MRLS remains, at his time, a mystery. The possibility exists that none of the above-mentioned theories caused the devastating tragedy to the Kentucky horse industry. It should be noted that because of the various hypotheses proposed, there are considerably fewer cherry trees on or near horse pastures. In addition, seminars and websites for proper ways to control Eastern Tent Caterpillars (such as insecticides), increased interest in herbicides for controlling hemlock and mycotoxin binders in most, if not all area feeds have all been observed. At the present time, recommendations issued by the scientific community to aid in the prevention of MRLS include the following:

In the midst of a relatively mild late winter, the projections of another banner year for the Eastern Tent Caterpillar, and with no clearly identifiable cause, prevention, or treatment for MRLS, these are the best available suggestions. In the meantime, the investigations continue as the various hypotheses are examined. Skeptics say we may never determine the cause what happened in the spring of 2001. To those individuals, it is best to remember “The one who says it cannot be done should never interrupt the one doing it.” (Author Unknown).


References