H5N1 clade 2.3.4.4b dynamics in experimentally infected calves and cows

Abstract

In March 2024, highly pathogenic avian influenza virus (HPAIV) clade 2.3.4.4b H5N1 infections were reported in dairy cows in Texas, USA¹. Rapid dissemination to more than 380 farms in 14 states followed². Here we provide results of two independent clade 2.3.4.4b experimental infection studies evaluating the oronasal susceptibility to and transmission of a US H5N1 bovine isolate, genotype B3.13 (H5N1 B3.13), in calves, and the susceptibility of lactating cows following direct mammary gland inoculation of either H5N1 B3.13 or a current EU H5N1 wild bird isolate, genotype euDG (H5N1 euDG). Inoculation of the calves resulted in moderate nasal replication and shedding with no severe clinical signs or transmission to sentinel calves. In dairy cows, infection resulted in no nasal shedding, but severe acute infection of the mammary gland with necrotizing mastitis and high fever was observed for both H5N1 isolates. Milk production was rapidly and markedly reduced and the physical condition of the cows was severely compromised. Virus titres in milk rapidly peaked at 10⁹ 50% tissue culture infectious dose (TCID₅₀) per ml, but systemic infection did not ensue. Notably, the adaptive mutation E627K emerged in the viral polymerase basic protein 2 (PB2) after intramammary replication of H5N1 euDG. Our data suggest that in addition to H5N1 B3.13, other HPAIV H5N1 strains have the potential to replicate in the udder of cows and that milk and milking procedures, rather than respiratory spread, are likely to be the primary routes of H5N1 transmission between cattle.

Fig. 1. Experimental design for infection with HPAIV H5N1 clade 2.3.4.4b isolates.

Timeline of the experimental study. Top, 12 Holstein calves of mixed sex (the asterisk indicates that one calf was hermaphroditic) were allocated to three experimental groups: (1) principal-infected (n = 6; 2 female, 3 male, 1 hermaphrodite); (2) sentinel (n = 3; 2 female, 1 male); (3) negative control (n = 3; 1 female, 2 male). Negative control calves were euthanized prior to experimental infection and tissues were collected for baseline comparison. Principal-infected calves were oronasally inoculated with 1 × 10⁶ TCID₅₀ per calf of H5N1 B3.13. Sentinel calves were introduced 48 h post infection. Rectal temperatures and clinical samples, including whole blood and urine, and nasal, oral and rectal swabs, were collected daily for 7 (whole blood) or 14 dpi and every 3 days thereafter. Serum was collected at 0, 7, 10, 14, 17 and 20 or 21 dpi. Postmortem examinations and extensive tissue collections were performed on days 7 (n = 2, principal-infected), 14 (n = 2, principal-infected) and 20 or 21 (n = 2/3, principal-infected/sentinel) post infection. Bottom, 7 Holstein–Friesian multiparous lactating dairy cattle were used in this experiment. Three cows were inoculated in the mammary gland with 10^5.9 TCID₅₀ per cow of H5N1 B3.13 (A/Cattle/Texas/063224-24-1/2024 (US group), n = 3) and three cows were inoculated in the mammary gland with 10^6.1 TCID₅₀ per animal of H5N1 euDG (A/wild_goose/Germany-NW/00581/2024 (EU group), n = 3). One cow served as a negative control. Swab samples (nasal, conjunctival and rectal) were taken daily until 9 dpi. EDTA blood samples were taken from individual cattle at 1, 3, 7 and 10 dpi. Urine was taken regularly until 14 dpi. Serum samples were obtained at 7 and 14 dpi and on the day of euthanasia. One cow of each group (no. 47 US and no. 72 EU) reached the humane endpoint at 3 dpi, one further cow reached it at 9 dpi (no. 88 EU) and one additional cow reached it at 13 dpi (no. 92 US); two cows (no. 66 EU and no. 87 US) survived until 21 dpi; All of these were analysed by necropsy. Figure created with BioRender.com under agreement number YH275PUF4T.

Fig. 2. Clinical features of calves and lactating cows infected with influenza A/H5N1 clade 2.3.4.4b virus.

a, Rectal temperature of principal-infected (n = 6) and sentinel (n = 3) calves and lactating cows (EU group, n = 3; US group, n = 3; control, n = 1) prior to and following inoculation. Data are shown as mean ± s.d. b, Average feed intake of each group of calves and cows following H5N1 infection. c, Individual milk production of lactating cows prior to and following experimental infection. Milk production of individual cows was tracked daily from −25 dpi until the end of the experiment at 21 dpi. The control animal never produced high amounts of milk. Dotted lines indicate when cows were euthanized. Source Data

Fig. 3. Viral shedding of influenza AH5N1 clade 2.3.4.4b virus isolates in experimentally infected calves and lactating cows.

a, RT–qPCR was used for the detection of the influenza A M gene (left axis; scatter plot) in nasal, oral and rectal swabs collected from H5N1 B3.13 oronasally infected calves post-inoculation and sentinel calves. Viral titres of nasal swabs (right axis; bars) are represented as the average and standard deviation of positive nasal swabs on each day (n = 3 calves at 7 dpi). The Cq value and titre of the inoculum are shown as asterisks on their respective axes. b, RT–qPCR results from nasal, rectal and conjunctival swabs of lactating cows following intramammary infection with H5N1 B3.13 or H5N1 euDG. c, H5N1 viral genome load (left axis) and corresponding H5-specific antibody titre (right axis) in milk samples over time. All cattle were milked daily and individual pooled milk samples were analysed by RT–qPCR to detect H5N1 viral RNA. Detection of H5-specific antibodies in selected milk samples was achieved using an H5-specific enzyme-linked immunosorbent assay (ELISA) and reported as sample OD₄₅₀ as a percentage of negative control OD₄₅₀ (S/N%). Source Data

Fig. 4. Infectious virus yields in milk and udder tissue of H5N1-infected lactating cows and genetic adaptation over time.

a, H5N1 viral titres recovered from milk samples of individual H5N1 B3.13- and H5N1 euDG-infected lactating cows. LOD, limit of detection. b, Viral titre of H5N1 infectious viral particles from individual udder quarters (FL, front left; BL, back left; FR, front right; BR, back right) collected at their euthanasia timepoint. c, Genetic adaptation at position 627 and 631 in PB2 of H5N1 B3.13 and H5N1 euDG. The sequence logo plot displays the relative proportion of amino acids present at positions 627 and 631 of PB2 from H5N1 B3.13 and H5N1 euDG in milk sampled from cows no. 88 EU and no. 87 US at indicated timepoints. Source Data

Fig. 5. Histological changes observed in respiratory tissues of oronasally infected calves.

Histological changes in calves oronasally infected with H5N1 B3.13 at 7 dpi (a–d) and 14 dpi (e,f). a,b, Haematoxylin and eosin (H&E) staining (a) shows a segmental region of suppurative tracheitis at 7 dpi (no. 6760). Degenerate neutrophils filled the tracheal lumen (arrows). No viral antigen was detected by IHC (b). c,d, In calf no. 712 (7 dpi), there were multiple small and discrete foci of interstitial pneumonia (c) with fibrin filling regional alveolar spaces (asterisks) and small numbers of neutrophils, macrophages and lymphocytes expanding alveolar septa. No viral antigen was detected (d). e,f, In calf no. 754 (14 dpi), bronchioles were frequently lined by hyperplastic epithelium (e), filled with degenerate neutrophils and partially occluded by papillary projections composed of a core of fibrous connective tissue with few inflammatory cells and lined by bronchiolar epithelium (bronchiolitis obliterans, asterisk). Bronchioles were also frequently delimited by prominent lymphoid aggregates (BALT hyperplasia). No viral antigen was detected (f). Scale bars, 100 μm.

Fig. 6. Histopathology and IAV NP detection in the mammary gland of multiparous cattle after intramammary infection with H5N1 B3.13 and H5N1 euDG.

a, IAV NP detection at 3 dpi. Inset, enlarged view of the outlined region showing juxtaposition of intact, lactating alveoli lacking antigen (black asterisk) and affected areas (green asterisk) in a lobular pattern. IHC using AEC chromogen and Mayer’s haematoxylin counterstain. Scale bars: 2.5 mm, main image; 100 µm, inset. b, Full necrosis of the alveolar epithelium, with cellular debris filling the lumen (left) and intralesional detection of IAV antigen (green asterisk) on a consecutive section (right). Some adjacent alveoli remain unaffected (black asterisk). Scale bars, 50 µm. c, Alveoli affected by necrosis with mostly intact basal lamina lined by basal and myoepithelial cells (blue arrows). Scale bar, 25 µm. d, Target cells identified on the basis of morphology following IHC include alveolar secretory epithelium. Scale bar, 50 µm. e, Teat with diffuse degeneration and necrosis of the lining epithelium, subepithelial oedema and mainly neutrophilic infiltrates (inset). Scale bar, 100 µm. f, Target cells identified on the basis of morphology following IHC include teat canal epithelium (inset). Scale bar, 100 µm. g, Necrotic alveoli filled with cellular debris admixed with degenerate neutrophils (blue arrow) in acute lesions and many lymphocytes, and fewer macrophages, neutrophils and plasma cells in the interstitium (green arrow). Scale bar, 50 µm. h, Intralesional IAV NP detection in secretory alveoli, mainly within cellular debris (inset). Scale bar, 50 µm. i, Simultaneous occurrence of intact, lactating alveoli (black asterisk), disruption of alveolar epithelium by necrosis (green asterisk) and beginning of regeneration (blue asterisk). Scale bar, 50 µm. j, Late stage IAV NP detection is restricted to scattered cellular debris. Scale bar, 25 µm. k, Left, mammary alveolus with intraluminal sloughed epithelium and cellular debris (green asterisk). Right, intralesional detection of IAV antigen on a consecutive slide (green asterisk). Scale bar, 50 µm. l, Regenerating alveoli (blue asterisk) with lack of IAV antigen labelling (not shown). Interstitial immune cell infiltrates include many lymphocytes and plasma cells (inset).Scale bar, 50 µm.

Extended Data Fig. 1. Exemplary milk consistency and appropriate CMT of H5N1-infected lactating dairy cows during the animal trial.

A CMT-picture of an H5N1 euDG-infected lactating dairy cow at 2 dpi. FL = udder front left, FR = udder front right, BL = udder back left, BR = udder back right B Milk consistency of H5N1 B3.13 and euDG infected dairy cows during the experiment (4 dpi).

Extended Data Fig. 2. California Mastitis Test (CMT).

A Legend for semi-quantification. Milk samples from individual quarters (front left/right and back left/right were gained and collected on appropriate CMT-plates. CMT-reagent was applied ~ 1:1 to the milk samples and was graded by eye with the help of a defined template. B CMT of milk samples from the uninfected control animal (#80) during the course of the experiment until its euthanasia timepoint.

Extended Data Fig. 3. California Mastitis Test (CMT) of lactating cows infected with H5N1 B3.13 (US-group).

A – C CMT of milk samples from cattle infected with H5N1 B3.13 during the course of the experiment until their respective euthanasia timepoint.

Extended Data Fig. 4. California Mastitis Test (CMT) of lactating cows infected with H5N1 euDG (EU-group).

A – C CMT of milk samples from cattle infected with H5N1 euDG during the course of the experiment until their respective euthanasia timepoint.

Extended Data Fig. 5. Megacor-RAT from milk samples of H5N1 experimentally infected dairy cattle.

A H5-specific Megacor-RAT used for milk samples of H5N1-infected cattle at 1 dpi. Positive samples are depicted with a red cross. B H5-specific Megacor-RAT used for milk samples of H5N1-infected cattle at 2 dpi. All H5N1-infected lactating dairy cattle have become positive via the H5-specific RAT from Megacor already at 2 dpi, irrespective of the H5N1-virus isolate used. C H5-specific Megacor-RAT used for milk samples of H5N1-infected cattle at 6 dpi. D H5-specific Megacor-RAT used for milk samples of H5N1-infected cattle at 10 dpi. All cows have become already negative via the RAT at 10 dpi.

Extended Data Fig. 6. Viral genome load in organ samples and survival data of lactating cows.

Orange: lactating cows infected with H5N1 B3.13. Blue: Lactating cows infected with H5N1 euDG. Grey: Uninfected negative control cow. A Viral genome load in udder organ samples of lactating cows euthanized at 3 dpi (#72 EU, #47 US, #80 Ctrl.) B Viral RNA load in udder organ samples of lactating cows euthanized at 9 dpi (#88 EU) or 13 dpi (#92 US). C Viral genome load in organ samples from neuronal tissues. D Viral RNA load in other internal organs of lactating cows. E Viral genome load in organ samples of the respiratory tract. F Survival curve of lactating cows over the course of the experiment. Source Data

Extended Data Fig. 7. Gross lung pathology of calves.

(A) At 7 dpi, multiple well-defined pulmonary lobules were red and slightly depressed on the right cranial lobe (congestion and partial atelectasis) affecting approximately 20% of the cranial and caudal portions of the right cranial lobe extending into the right middle and caudal lobe of one of the two principal-infected calves (#712). There was a focal area of mild subpleural hemorrhage on the ventral surface of the left caudal lung lobe of animal #6760. (B) At 14 dpi, one of the two principal-infected calves (#754) had multifocal to coalescing red and depressed foci of congestion and atelectasis on the left and right cranial lobes. Approximately 60% of the caudal portion of the left cranial lobe, 55–60% of both the cranial and caudal portions of the right cranial lobe and <5% of the accessory lobe were affected. There were also multiple pleural adhesions to the thoracic wall. (C) At 20 dpi, the two principal-infected calves (#6772 and #697) had either few small red and slightly depressed foci of congestion and atelectasis on the left cranial lobes (#6772), or a focal, similar area on the apical portion of the right middle lobe (#697). (D) Postmortem examinations of the three sentinel animals were performed at 21 dpi and revealed scattered red foci of pulmonary congestion/atelectasis. In animal #748, there were multiple, small foci of mild consolidation in the left and right cranial lobes (5% of lung affected) and few pleural adhesions to the thoracic cavity. For animal #6770, congestion and atelectasis were accompanied by mild to moderate edema affecting predominately the right lung lobes. (E) One of the three negative control calves (#6767) had a small isolated focus of consolidation of the pulmonary parenchyma at the apical margin of the right middle lobe. Gross lesions were not appreciated in the remaining negative control animals.

Extended Data Fig. 8. Influenza A virus nucleoprotein detection using immunohistochemistry in the mammary gland of cattle after intramammary infection with H5N1 B3.13 and H5N1 euDG.

The distribution was graded on an ordinal scale with scores 0 = no antigen, 1 = focal, affected cells/tissue <5% or up to 3 foci per tissue; 2 = multifocal, 6%–40% affected; 3 = coalescing, 41%–80% affected; 4 = diffuse, >80% affected. Representative pictures were taken from the most severely affected quarter from each cow. A Score 4, H5N1 B3.13, 3 dpi. B Score 3, H5N1 euDG, 3 dpi. C Score 1, H5N1 B3.13, 13 dpi. D Score 3, H5N1 euDG, 9 dpi. E Score 1, H5N1 B3.13, 21 dpi. F Score 0, H5N1 euDG, 21 dpi. Scale bar 2.5 mm and 50 µm (inlay).

Extended Data Fig. 9. Histopathology and Influenza A virus nucleoprotein detection (antigen) of cattle after intramammary infection with H5N1 B3.13 and H5N1 euDG including tissue controls.

A Nasal concha: Chronic-active rhinitis (A1) lacking IAV antigen (A2). B: Lung: Chronic bronchointerstitial pneumonia in convalescence phase (B1), lacking IAV NP (B2). C Genitofemoral nerve: No findings (C1), no IAV antigen (C2). D Spinal cord: No findings (D1), no IAV antigen (D2). E Brain, cortex: No findings (E1), no IAV antigen (E2). F Positive control slide, HPAIV infected chicken, lung: abundant IAV antigen. G Negative control slide, uninfected cow, mammary gland: no IAV antigen. H1 Mammary gland: cow infected with H5N1 B3.13, 3 dpi, abundant IAV antigen. H2 Consecutive slide of H1: an irrelevant antibody (anti Sars clone 4F3C4) yielded no immunopositive reaction. Hematoxylin and eosin (HE) stain (A1, B1, C1, E1) and immunohistochemistry (antigen) on consecutive (A2, B2, C2, E2, H1, H2) or independent (F, G) slides. Scale bar 25 µm (A1-2), 50 µm (C1-2, inlay D1, E1, F, G), 100 µm (D2, E2, H1-2), 250 µm (B1-2), 2.5 mm (D1, E1).

Extended Data Fig. 10. Tissue controls used for immunohistochemistry.

The anti-NP antibody used strongly labels influenza virus A H3N2 and H5N1-infected epithelial cells lining bronchioles.