Chinese Work on H5N1 Transmission in Lactating Dairy Cattle
Does "milk stealing" really explain H5N1 transmission? Read on
A hat tip to Tetano on FluTrackers for posting the following link Tuesday afternoon: Nat Sci Rev . H5N1 virus invades the mammary glands of dairy cattle through “mouth-to-teat” transmission - FluTrackers News and Information. I depend on many sources for new information - FluTrackers is a prime one for research leads.
Here is the link to the research pre-print itself, out of the Harbin Veterinary Research Institue in China: H5N1 virus invades the mammary glands of dairy cattle through “mouth-to-teat” transmission | National Science Review | Oxford Academic. This is a sprawling extended study, starting with oral dairy cow inoculations with 2 H5N1 2.3.4.4b avian strains from China (one proving more pathogenic than the other), then moving on to udder inoculations when oral inoculations failed to induce mastitis.
Next researchers reverse engineered the A/dairy cow/Texas/24-008749-001/2024 (DC/24) virus for further studies. Researchers then inoculated calves with the Texas strain (DC/24) and placed them with uninfected lactating cows, from whom the calves “stole milk” for multiple days. Virus was detected in the milk of the contact cows 9-12 days post mixing and suckling, with the researchers concluding that the udder infection ascended from infected calf suckling.
Finally, vaccine studies were conducted with two vaccines (an inactivated vaccine and an HA-based DNA vaccine) derived from a clade 2.3.4.4b virus detected in a swan in China. Both vaccines provided significant protection.
This study is complex with a lot of missing information and leaps in logic in my estimation. One well-documented finding is that in their hands, the reverse engineered Texas (DC/24) H5N1 virus is quite capable of infection, replication, and shedding from the respiratory and oral tract in dairy cattle:
…we generated the index cattle H5N1 virus, A/dairy cow/Texas/24-008749-001/2024 (DC/24), by reverse genetics and assessed it in lactating dairy cattle. Three cattle were each intranasally inoculated with 2 ml of the virus (10-6 EID50 per ml). Animal treatment and sample collection were the same as those for the TS/23 virus-inoculated animals. One cow had a fever on day 4 p.i. (Table S2). Virus was detected in nasal swabs collected at all time points from the cattle that were euthanized on days 3 and 6 p.i., and in nasal swabs collected between days 1 and 9 p.i. of the animal that was euthanized on day 14 p.i. (Fig. 2a). Virus was also detected in oral swabs from three animals on several occasions between days 1 and 6 p.i. (Fig. 2b), but was not detected in rectal swabs or milk (Fig. S5, a and b). In the animal euthanized on day 3 p.i., virus was detected in the nasal turbinate, larynx, tonsil, submandibular lymph nodes, trachea, and two different lung lobes, but was not detected in any other organs or tissues. In the animal euthanized on day 6 p.i., virus was detected in the nasal turbinate, soft palate, root of tongue, tonsil, larynx, submandibular gland, and trachea, but was not detected in any other organs or tissues. In the animal euthanized on day 14, virus was detected in the tonsil but not in any other organs or tissues tested (Fig. 2c, and Fig. S5c).
Replication of DC/24 virus in dairy cattle after intramammary gland inoculation
The above study indicated that the DC/24 virus cannot reach cattle mammary gland following intranasal infection. We therefore performed the intramammary gland-inoculated study as we did for the TS/23 virus. Two cattle had increased body temperature on days 2 and 3 p.i., and the milk from their virus-inoculated udders became yellow and thick (Table S2). Virus was detected in the milk collected from the infected mammary glands (left front and right rear) of all three animals, and virus shedding in milk lasted for nine days in the animal that was euthanized on day 14 p.i., with peak titers about 10-fold higher than that of the TS/23 virus-inoculated cattle (Fig. 2d versus Fig. 1e). Virus was detected in the left front and right rear mammary glands of the animals that were euthanized on days 3 and 6 p.i. (Fig. 2e), but not in any other samples collected from these two animals (Fig. S5, d to f), or any swabs or organs from the animal that was euthanized on day 14 p.i.(Fig. 2e, and Fig. S5g). We further investigated the minimum dose required for the DC/24 virus to infect the mammary glands. Two cattle were respectively inoculated with three different doses of the virus into three different mammary glands: 1 ml of 10-2 EID50 to the right front teat, 1 ml of 10 EID50 to the left rear teat, and 1 ml of one EID50 to the right rear teat. Virus was detected in the milk and tissues of the right front and left rear udders but was not detected in the milk or tissues collected from the left front and right rear udders (Fig. 2, f and g), indicating that 10 EID50 of the DC/24 virus is required for successfully infection of the mammary gland of cattle. These studies demonstrate that replication of H5N1 influenza viruses in cows varies among strains. Virus infected intranasally can only replicate in tissues of the mouth and respiratory tract of the cattle, and virus inoculated into the mammary gland replicates only in the inoculated gland but does not migrate to neighboring mammary glands of the cattle. Therefore, entry through the teat is the only natural way the virus can infect the mammary gland of cattle.
Sialic acid receptor distribution in different cattle tissues
Our studies indicated that the H5N1 virus could replicate in multiple oral tissues, lungs, and mammary glands. Binding to sialic acid receptor on cell surface is the first step for influenza virus to infect cells, and there are mainly two types of sialic acid receptors: α2,6-linked sialic acids (also known as the human-type receptor) and α2,3-linked sialic acids (also known as the avian-type receptor). We therefore investigated the types of sialic acid receptors in different tissues of cattle. We found that cells in the soft palate, tonsils, root of tongue, sublingual glands, and submandibular glands, lungs, teats, and mammary glands of cattle express both avian-type sialic acid receptors (MAL-I or MAL-II stain-positive) and human-type (SNA stain-positive) sialic acid receptors, whereas the nasal turbinate, larynx, trachea, and parotid glands of cattle only express high level of avian-type receptors (Fig. 3a, and Fig. S6). This receptor distribution allows viruses that can bind to either avian-type or human-type receptors to attach to the cells of oral tissues. The DC/24 virus was detectable in oral swabs of dairy cattle for up to 6 days with a titer of approximately 10-2 EID50 (Fig. 2b). Replication of the virus in oral tissues was further confirmed by immunohistochemistry studies on the virus-positive oral tissues of the DC/24 virus-infected cattle, and the amount of antigen in the root of tongue and submandibular gland was notably higher than that in the tonsil and soft palate (Fig. 3b). The high level of sialic acid receptors and the efficient replication of the virus in certain oral tissues of cattle not only supports influenza virus infection through contaminated feed or water but also makes the oral cavity an ideal site for H5N1 virus to evolve and transmit.
“Mouth-to-teat” transmission of H5N1 virus in cattle
Since the DC/24 virus replicated efficiently in certain tissues of cattle mouth, and given that some lactating cattle "steal milk" through self-nursing or mutual-nursing [20-26] (Fig. S7), we speculated that "mouth-to-teat" transmission may be the route by which the H5N1 virus initially infects the mammary glands of dairy cows. To test this hypothesis, we infected a pair of calves with 10-6EID50 of the DC/24 virus intranasally and another pair orally. Six hours later each pair of calves was housed with a lactating cattle in one room (Fig. S8a), and allowed to suck the teats as frequently as needed. Nasal and oral swabs from the calves and the milk from the lactating cattle were collected every day for 12 days for virus titration in eggs. Body temperature and milk color were monitored daily (Table S3). In one of the intranasally inoculated calves, virus was detected in nasal swabs from days 1 to 7 p.i. and in the oral swab on day 1 p.i.; in the other intranasally inoculated calf, virus was detected in nasal swabs from days 2 to 5 p.i. and in oral swabs on days 2 and 3 p.i.(Fig. 3, c and d), but virus was not detected in the milk or mammary gland tissues collected from the lactating cow that was housed with these two calves (Fig. S8, b and c). In one of the orally inoculated calves, virus was only detected in the oral swabs on day 4 p.i. in one calf, which we euthanized on day 8 p.i. to give the other calf more opportunities to suck. In the other orally inoculated calf, virus was detected in the nasal swabs on day 1 p.i. and days 11 and 12 p.i., and in the oral swabs from day 1 to day 12 p.i. (Fig. 3, e and f). In the lactating cow that was housed with the orally infected calves, virus was detected in the milk from one udder on days 9 to 12 after the calves were infected and another udder on days 11 and 12 after the calves were infected (Fig. 3g). Virus was also detected in the tissues collected from these two udders of the cow that was euthanized at the end of the study (Fig. 3h). Of note, the increase in the virus titer in the oral swabs on days 11 and 12 p.i., as well as the recurrence of the virus in the nasal swabs on days 11 and 12 p.i., of calf 2 (Fig. 3, e and f) may be related to the consumption of milk containing the virus. Our results show that H5N1 virus in the oral cavity of cattle can spread to the mammary glands of cattle during nursing.
This Chinese study adds evidence for respiratory transmission, then assumes that failure to demonstrate mastitis in limited oronasal experimental infections rules out that route for lactogenic H5N1 lesions. U.S. field experience indicates that lactating cows do not universally develop mastitis when infected. It’s not justified for the Chinese researchers to declare: “The above study (3 animals) indicated that the DC/24 virus cannot reach cattle mammary gland following intranasal infection.”
I see several weaknesses and overinterpretations in this process. First of course, is lack of replications, almost unavoidable in BSL-3 facilities. Beyond that, researchers failed to monitor the lactating cows for respiratory infection, despite nose-to-nose contact with the infected calves. It’s plausible that the lactating cows become infected by the oral-respiratory route (unmonitored), with the virus passing to the udder. We cannot assume that just because earlier experiments (3 animals) failed to show udder infection following from respiratory infection, that this could not have occurred here. The delayed time to udder infection documented here (9-12 days) is similar to delays in clinical illness in U.S. dairy herds following initial bulk tank positive results (initial infection and viremia). Direct inoculation of virus into the udder results in clinical illness in 2-3 days, not 9-12. While delayed infection from suckling is certainly possible, there is also no direct evidence that virus travelled directly from the oral cavity to the nurse cows’ udder.
This is really a conclusion via exclusion of an event that can’t be proven - that virus does not transmit systemically to the udder in cattle. Quite frankly, that assertion flies in the face of the following evidence:
Lombard et al have documented substantial H5N1 viremia early in the course of infection. Blood flows to all tissue, especially into lactating udder parenchyma with abundant appropriate sialic acid receptors. It’s certainly plausible that virus may pass into lactogenic tissue from the blood in some percentage of infected cows.
In a paper I reviewed last year, Dmitrov et al documented widespread viral RNA distribution in necropsy tissues, strongly implying a systemic viral infection even before the disease was widely characterized in an early-infected herd in spring 2024 in Ohio.
In a final paper just released this week Nydam, Diel et al showed high seroconversion rates in lactating cows and moderate seroconversion in dry cows in an that same Ohio herd infected by asymptomatic herd additions from Texas 10 days prior to clinical illness. While infection rates were higher in the lactating animals, asymptomatic animals and dry cows also seroconverted, indicating non-milk-associated viral transmission within the infected herd.
I’m actually somewhat agnostic to the possibility that the milking process could transmit H5N1 2.3.4.4b from cow to cow within herds with sufficient viral loads, inadequate udder sanitation, etc. I recall cows with compromised teats on my family dairy farm of my youth, i.e. where natural barriers to ascending infections were compromised by injury. However, in general God gives an udder tremendous capacity to exclude deleterious organism invasions - milk flows one way with a series of mechanical and biological barriers to ascending infection. I just reflexively resist the idea that a new organism has come along that suddenly beats millennia of natural protection against ascending infections on an epidemic basis.
The bigger danger in my view is that this udder centric view of H5N1 dairy cow transmission allows continued “plausible deniability” to an increasingly obvious conclusion that we have classic endemic oronasal influenza virus in cattle (yes, most common in dairy cattle at least for now). Under this paradigm the mastitis and milk infection is a prominent clinical sign in a percentage of infected animals, not an exclusive characteristic of infection. As long as we try to keep this disease exclusively “in the parlor”, we miss the responsibility to look for it and prevent it elsewhere.
Strategic serology and targeted PCR testing could give us some estimates of what our true risks are in other cattle groups and other species. Unfortunately, we are biding our time waiting for poultry or a human zoonotic crossover with ongoing transmission to force our hand. It’s a horrible way to handle a One Health challenge, but the economics and politics of the situation really preclude more proactive approaches at this point.
I’d like to close today with a You Tube video of testimony from Congress - Safeguarding U.S. Agriculture: The Role of the National Animal Health Laboratory Network (NAHLN). Most of you likely know Dr. Annette Jones, California State Veterinarian. As David Letterman used to say, “I wouldn’t give my (her) troubles to a monkey on a rock!”
Annette spoke passionately (violating her time limit) regarding the HPAI debacle her poultry and dairy producers faced last fall due to overwhelming viral loads from infected dairy farms making biosecurity essentially impossible. She pleaded passionately for vaccine approval for dairy herds in order to protect poultry and other livestock farms. Please watch the following from about 55 minutes on to catch her comments. Remember that this state veterinarian managed nearly 1000 dairy and poultry quarantines last year, with some dairy quarantines still in place. Additionally, multiple zoonotic infections, feline deaths, and raw milk challenges were identified in California. Annette has lived the epidemiology of H5N1.
John