(Professor Dr. Elena Ryabchikova was with us in December 1994 for the first CBMTS, which took place at SPIEZ LABORATORY, Switzerland. She is one of the world's leading electron microscopists, an international author and a leading authority on Ebola and Marburg as well as Smallpox and has worked extensively with the influenza virus)

Features of Pathology in Mice Experimentally Infected with Highly Pathogenic H5N1 Influenza Virus

Oleg S. Taranov, Elena M. Malkova, Oksana B. Gritsyk, Olga K. Demina, Elena I. Ryabchikova State Research Center of Virology and Biotechnology "Vector", Koltsovo, Novosibirsk region, 630559 Russia

ABSTRACT

          Avian influenza became a new threat and has set people thinking about possibility of new influenza pandemic which may be caused by highly pathogenic H5N1 influenza virus. Influenza virus H5N1 exhibited its deadly essence by taking out many millions of birds in nature and aviculture; millions of other chicks and ducks were killed to prevent spread of the epizootic. Public health specialists worry that the virus could acquire the ability to spread rapidly between the humans, thus starting a new pandemic, which could kill millions. H5N1 influenza virus (IV) strains isolated in Russia belong to Qinghai group of H5N1 IV, and were imported by migratory birds. We examined the time-course changes of blood and lungs in mice after intranasal infection with strains A/Chicken/Kurgan/05/2005, A/Duck/Kurgan/08/2005 and A/Chicken/Suzdalka/Nov-11/2005, viruses differing in virulence for this animal species. There was evidence of leucopenia and severe damage of hemopoiesis in the blood of all mice infected with these strains. We also examined the pathological changes in the lungs during the course of the infection and noted features specific to each strain. The main characteristics of lung pathology in all mice were the focal nature of the changes, severe damage of bronchial epithelium and pronounced alteration of lung vasculature. We noted that the A/Chicken/Suzdalka/Nov-11/2005 strain induced massive apoptosis of infected bronchial cells, which may be a part of mechanism responsible for mouse resistance to this strain. The most interesting finding was the absence of evidence that there was serious direct virus damage of the lung, implying that the host humoral mechanisms has an important role in the pathogenesis of H5N1 influenza in mice.

INTRODUCTION

          Influenza A virus belongs to Orthomyxoviridae family and has one-strand RNA genome with negative polarity composed of 8 segments. Sixteen types of hemagglutinin (HA) and 9 types of neuraminidase (NA) presently are known [Fouchier, 2005]. Wild waterfowl serve as a natural reservoir of influenza A virus, and it has been thought for a long time that only H1, H2 and H3 subtypes are able to infect humans [Webster, 1998]. It was proposed that strains infecting humans evolved from bird strains via an intermediate mammalian host, most probably in a swine [Gorman, 1992]. A new period of influenza history started in 1997, when nearly identical strains of H5N1 IV were isolated from sick chicks and a child in Hong Kong [Bender, 1999]. That was the first documented case of direct human infection by IV without an intermediate host. H5N1 IV includes variants having different pathogenicity for birds, with the most harmful called "Highly pathogenic avian influenza (HPAI) viruses." These affect not only the birds, but also other animals and humans [Peiris, 2007]. H5N1 IV was not of prime scientific interest during 1996-2003, although there were papers reporting the isolation of H5N1 strains from domestic birds in south China and Vietnam, as well as registering sporadic human cases.

          The situation dramatically changed in the end of 2003, when H5N1 IV suddenly widened its natural area and caused unprecedented mass epizootic disease in poultry and new severe cases in humans with a high mortality in East and South-Eastern Asia [WHO, 2005]. Epizootic disease in wild waterfowl in Qinghai Lake (Western China) were a disturbing signal that migratory birds were involved in spreading of H5N1 IV [Chen, 2005; Liu, 2005]. During 2005--2006, H5N1 IV was detected in Asia, Europe, North and West Africa, and Middle East. The outbreaks in bird and human cases were registered in 2007-- 2008 in various countries [WHO, 2008]. In Russia the epizootic disease H5N1 IV began on June 10, 2005 in Suzdalka village (Novosibirsk region), then spread to six regions of the Siberian and Ural Federal districts. Both domestic and wild birds were affected and, the most severe disease was observed in chicks. Examinations showed that the strains isolated in Russia belong to Qinghai group of H5N1 IV, and were imported to Russia by migratory birds.

          H5N1 IV is evolving dynamically, with all possible variations of virus-host interactions. The ability of H5N1 IV to infect different species makes it difficult to predict which of many currently circulating H5N1 strains could evolve into a new pandemic IV, and so it is very important to examine the characteristics of different strains. Three strains isolated in Russia were selected (A/Chicken/Kurgan/05/2005; A/Chicken/Suzdalka/Nov-11/2005, and A/Duck/Kurgan/08/2005). All the strains were isolated from sick birds and belong to HPAI group: the intravenous pathogenicity index (IVPI) for chicks was 3.0.

METHODS

          White outbred mice weighing 12-15 g were infected intranasally with 10 LD50 of one of the three strains: highly pathogenic strain A/Chicken/Kurgan/05/2005 (LD50 for intranasal infection was 1.1±0.4 lg EIU50); strain having middle pathogenicity A/Duck/Kurgan/08/2005 (LD50 for intranasal infection was 4.7±0.6 lg EIU50) and apathogenic strain A/Chicken/Suzdalka/Nov-11/2005 (LD50 for intranasal infection more than 6.7±0.7 lg EIU50). Animals in the groups were sacrificed daily over 8 days post infection (PI). Each mouse was killed by cervical dislocation, the chest was opened, heart taken, right ventriculus cut, and one drop of blood was smeared on three microscopic glasses. The smears were dried in air, fixed by ethanol and stained by May-Grunwald and Romanovsky reagents according to Pappengeym-Kryukov's method.

          Lung samples were collected daily during 8 days PI from each sacrificed mouse. Lungs with trachea were taken and immersed into 4% paraformaldehyde for 48 h at +4C. The samples for light microscopy were routinely processed and embedded in paraffin wax. Paraffin sections were stained with hematoxylin-and-eosin and van Gieson's stain, and examined in Imager Z1 microscope; micrographs were obtained by AxiCam HRc digital camera (ZEISS, Germany). The samples for electron microscopy also were fixed by 4% paraformaldehyde for 48 h at +4C, postfixed by 1% osmium tetraoxide, routinely processed and embedded in epon-araldit mixture. Semithin and ultrathin sections were made on Reichert-Young ultratome. Semithin sections were stained by Azur-2 and examined in light microscope, and areas for pyramids were chosen. Ultrathin sections were contrasted by uranylacetate and lead citrate solutions and examined in Jem 1400 electron microscope (Jeol, Japan). Electron microscopic micrographs were obtained by Jeol digital bottom camera and Veleta (SIS, Germany) side-mounted digital camera.

RESULTS

          The morphology and staining properties of uninfected mouse blood cells are similar to those in other animals; however blood counts in mice are characterized by a high content of lymphocytes (85%). Infection of mice with all strains of H5N1 IV caused pronounced changes in the composition and morphology of blood cells. Figure 1 shows the variations in amounts of lymphocytes, monocytes, eosinophils, basophils and neutrophils on each day post infection (PI). Echynocytes (erythrocytes having surface protrusions) and blood blast cells appeared in mice blood smears 24h PI with highly pathogenic strain A/Chicken/Kurgan/05/2005; the smears of mice infected with other two strains did not show any change in comparison with uninfected animals. Echynocytes and blast cells were observed in blood smears of infected mice 2 days PI with all three strains of H5N1 IV and were appreciably higher for A/Chicken/Kurgan/05/2005 strain.

          Day 3 PI with all three H5N1 IV strains blood smears showed a great number of echynocytes, absence of thrombocytes, increase of blast cells number, and presence of promyelocytes and promonocytes. Day 4 PI, blood smears were characterized by prominent poikilocytosis and presence of large number of blast and immature bone marrow cells. Dramatic changes were noted in the lymphocytes: nuclei had breaks in the envelope, dark coarse gobs and shrinkage; some lymphocytes contained two nuclei or mitotic features. By day 5 PI, blood smears showed severe damage of the blood system. All blood smears contained numerous degenerating blast and immature blood cells showing sequential steps of the destruction. Prominent leucopenia was noted in all smears (Fig. 1). The smears of all mice contained thrombocytes and many degenerating lymphocytes and other leukocytes.

          For the mice, the influenza disease took a turn for the worse days 7--8 days PI Poikilocytosis and leucopenia were prominent in the blood smears, as was leukocyte destruction, numerous immature myelocytes and broken azurophilic nuclei. These pathological changes may represent the inability of infected mice to repair damaged hemopoiesis. Most probably the H5N1 IV infection induced an imbalance in the host's cytokine system. Interaction of H5N1 IV proteins with cytokine system, PKR, MAPK _ PI3K kinases and transcription factors NF-kB and IRF has been proposed as a main mechanism in the disease in humans [Lee, 2007].

The respiratory tract is the main target for IV in most sensitive organisms, including humans. Studies of influenza pathogenesis, testing of a new antivirals and vaccines are performed on various experimental animals, including mice which are the most widely used species. Avian H5N1 IV is able to infect mice without preliminary adaptation and causes severe damage of lungs. The pathogenicity of H5N1 IV strains varies in mice and often corresponds to pathogenicity in the virus primary host [Gao, 1999; Perrone, 2008]. However, there are few published observations of lung damage according to different H5N1 IV strains.

          In our study, replication of H5N1 IV strains was detected in ciliated cells of mouse bronchi (Fig. 2). The cells showed virus-specific structures in the nucleus and cytoplasm that were identical to those seen in MDCK and Vero cells infected with the same strains. Virus budding occurred on the apical plasma membrane of ciliated cells, independent of cilia and microvilli.

          All three H5N1 IV strains caused pronounced focal pathological changes in mice trachea and bronchi 24h PI. Spasms of small bronchi, dystrophic changes and vacuolization of bronchial epithelium, and detachment of some epithelial cells were observed in light microscope. The bronchial lumens contained cellular debris and few blood cells. Light congestion of respiratory tissue, stromal edema, stasis of alveolar capillaries, aggregation of erythrocytes in lung vessels were seen in all mice. Goblet cells were empty and showed clear-cut dystrophic changes. Ciliated cells mostly remained unaltered; however some of them looked swollen.

          Pathological changes in the trachea and bronchi differed with H5N1 strains: the two more pathogenic strains, A/Chicken/Kurgan/05/2005 and A/Duck/Kurgan/08/2005, induced epithelial cell degeneration and necroses. The apathogenic strain, A/Chicken/Suzdalka/Nov-11/2005, induced apoptosis in many airways epithelial cells. The ability of IV to induce apoptosis was shown in various experimental systems both in vitro and in vivo [Fesq, 1994; Mori, 1995; Price, 1997]. There are different notions about the apoptosis role in IV infection. One postulates that apoptosis represents the primary link of unspecific host immune defense and so restricts virus reproduction [Kurokawa, 1999]. Another notion considers apoptosis as a main cause of cell death in influenza A and B infection, and so apoptosis plays negative role [Hinshaw, 1994]. We propose that apoptosis played a positive role by eliminating infected cells in mice infected with A/Chicken/Suzdalka/Nov-11/2005 strain.

          Pathological changes in mice lungs increased during infection with all three IV strains. Light microscopy of the lungs on days 2--3 PI revealed bronchospasm, peribronchial edema, accumulation of blood cells, mucus and cell debris in bronchial lumens; some bronchi were clogged with this material. The bronchi of mice infected with A/Chicken/Suzdalka/Nov-11/2005 strain contained many neutrophils; it was possible to say that the mice had purulent bronchitis. The most pathogenic strain,A/Chicken/Kurgan/05/2005, caused pronounced bronchial spasm. All mice showed foci of atelektasis, thickening of interalveolar septa, hemorrhages into alveolar space, erythrocyte rouleaus in vascular lumens, vascular dystonia.

          H5N1 IV profoundly affected the blood system of the mouse lung. Light microscopy revealed aggregation of blood cells, spasm and plethora of blood vessels 24 h PI in all infected animals. Perivascular edema, fragmentation of vascular elastic membranes and alteration of vascular walls were observed on days 2-3 PI. Changes of blood morphology, perhaps reflecting an alteration in plasma proteins, were apparent in large blood vessels: their lumens contained membrane-like structures and flake-like substances. The blood plasma in uninfected mice is electron lucent and does not contain anything except blood cells, so it is possible to conclude that H5N1 IV seriously altered mice blood system.

          Initial stages of thrombus formation were detected in mice blood vessels on days 2--3 PI with all three strains of H5N1 IV. Aggregation of thrombocytes and other blood cells and their binding with vascular endothelium were observed. Sometimes aggregating cells were "glued" by flake-like substance. Many thrombocytes had damaged morphology; some were completely destroyed. The vascular lumens of mice infected with A/Duck/Kurgan/08/2005 strain contained fibrin bundles and fibrin fibers not associated with blood cells. Erythrocytes with decreased electron density and erythrocytes with electron lucent cytoplasm and granular substance, instead of homogeneous hemoglobin, were seen in blood vessels of all infected mice. All these changes indicate that the erythrocyte functions were altered. In all three strains, holes with electron lucent substances inside formed in mouse arterial vessels 2--3 days PI. The holes raised endothelial cells and separated them from the underlying basal membrane, until finally the cells were broken and detached (Fig. 3).

          The most severe pathological changes in mouse lungs were found on days 4--5 PI. In all three strains, there was pronounced inflammation and focal necroses of airways and respiratory zones, inflammatory infiltration of bronchial wall, intra-alveolar (Fig. 4). All infected mice developed altered blood vessels and prominent thrombosis. Spasm, perivascular edema, aggregation of blood cells, and fragmentation of elastic membranes were seen in lung blood vessels. Blood cells clogged alveolar capillaries. Endothelium detachment from the basal membrane continued in arterial vessels. Areas completely devoid of endothelial cells were found in arterial walls and showed direct contact of blood and endothelial basal membrane. These "naked" areas of basal membranes remained clear; no adhesion of blood cells on these areas was seen.

          The period of H5N1 IV infection days 6--8 PI was critical in disease development: at this time the mice infected with highly pathogenic strains died. Light microscopy revealed large necroses, lipid thromboses of the vessels, prominent infiltration of airways and respiratory zone, perivascular and intra-alveolar hemorrhages. Mice infected with A/Chicken/Suzdalka/Nov-11/2005 strain showed purulent-necrotic bronchitis. Pathological changes were less prominent infections with the A/Duck/Kurgan/08/2005 strain. Regeneration of bronchial epithelium was detected in few mice infected with A/Duck/ Kurgan/08/2005 and A/Chicken/Suzdalka/Nov-11/2005 strains. Changes in blood vessels in mouse lungs on days 6--8 PI were very similar to those observed on days 4--5. Thrombi composed of destroyed cells and plasma alteration were seen in large vessel lumens, desquamation of endothelium was increased in arterial vessels. Perivascular infiltration was rising in all mice. Lungs of mice infected with A/Chicken/Kurgan/05/2005 strain showed further development of inflammation and destruction processes.

CONCLUSIONS

          Distinct tropism of H5N1 IV to lung cell has been mentioned in all published studies of this virus. Our study revealed pronounced damage of airways and respiratory zone, and deep alteration of blood cells and lung blood flow in mice infected with the three strains of H5N1 IV. All mice developed leucopenia and poikilocytosis, and numerous blast cells appeared in the bloodstream. These observations are strong evidence for the very important role of blood system alteration in influenza H5N1 pathogenesis. The current knowledge about H5N1 IV biological properties show apparent necessity of accumulation and systematization of data about pathological changes in organs of experimental animals infected with different H5N1 IV strains. The results of the present study are widening existing knowledge about lung damage by H5N1 IV and will be useful for analysis of the relationships of the virus genome structure and pathogenic properties.

REFERENCES

Bender C. et al. (1999) Virology, 254, 115-123.

Chen H. et al. (2005) Nature, 436, 191-192.

Fesq H. et al. (1994) Immunobiology, 190, 175-182.

Fouchier R.A. et al. (2005) J. Virol. 79, 2814-2822.

Gao P. et al. (1999) J. Virol. 73, 3184-3189.

Gorman O. T. et al. (1992) Curr. Top. Microbiol. 172, 75 - 97.

Hinshaw V.S. et al. (1994) J.Virol. 68, 3667-3673.

Kurokawa M. et al. (1999) Int.J.Mol.Med. 3, 527-530.

Lee D.C.W., Lau A. (2007) Signal Transduction, 7, 64-80.

Liu J. et al. (2005) Science, 309, 1206.

Mori I. et al. (1995) J.Gen. Virol. 76, 2869-2873.

Peiris J.S. et al. (2007) Clin.Microbiol. Rev. 20, 243-267.

Perrone L.A. et al. (2008) PLoS Pathog. 4(8): e1000115. doi:10.1371/journal.ppat.1000115.

Price G.E. et al. (1997) J.Gen.Virol. 78, 2821-2829.

Webster R.G. (1998) Emerg. Infect. Dis. 4, 436-441.

WHO (2005) Emerg. Infect. Dis. 11, 1515-1521.

WHO (2008) N.Engl.J.Med. 358, 261-273.

KEY WORDS

          H5N1 influenza virus, mice, intranasal infection, pathogenesis.

FIGURES

          Figure 1. Changes of blood counts in mice during the infection with H5N1 influenza virus. Number of the cells is shown on abscissa axis. Days post infection are pointed on ordinate axis. All values are average of blood counts of three mice. "0" - uninfected mice (control). (l-r) Blue color shows lymphocytes, yellow - monocytes, green - eosinophils, black - basophils, and brown - neutrophils.

          Figure 2. Budding of H5N1 influenza virus, strain A/Chicken/Suzdalka/Nov-11/2005 on apical surface of mouse bronchial ciliated cell. 24 h post infection.

1. bronchial lumen;
2. cytoplasm of ciliated cell;

• black arrows show cellular microvilli resembling virus particles;
  
• grey arrows show cilia; white arrows show cilia basal

          Figure 3. Detachment of endothelial cells from basal membrane in arterial vessel of mouse lung 5 days post infection with H5N1 influenza virus, strain A/Chicken/Kurgan/05/2005.

1. vascular lumen;
   
2. endothelial cell;
   
3. a hole separating endothelial cell from basal membrane;
   
4. erythrocyte;
   
5. elastic membrane;
6. smooth muscle cell.

          Figure 4. Severe pathological changes in mice lungs 7 days post infection with H5N1 influenza virus, strain A/Duck/ Kurgan/08/2005. Large cellular infiltrate covers bronchus, blood vessels and lung parenchyma.

1. bronchial lumen
   
2. blood vessel

• Arrow shows cell debris. Azur II stained semithin section.

Post Script:

          ASA had requested Prof. Ryabchikova give our readers her thoughts on the the flu virus in general and the blossoming H1N1. This request was prior to WHO declaring that the H1N1 virus had reached pandemic levels. Prof. Ryabchikova's reponse follows:

          "As all are aware this illness, familiarly called the 'flu', visits all countries every year. However, the history of influenza pandemics suggests we tend to stand on ceremony with the influenza virus which could kill millions. This flu virus can exist in hundreds of variants and can bring to naught all of the efforts of the full scientific community in the prediction of a specific viral type and the preparation of an effective vaccine for the season.

          Avian influenza virus H5N1 emerged in 1997 in Hong Kong and is circulating worldwide presently. It has exhibited new biological and horrible clinical feature,s such as an ability to directly infect a human from bird, the involvement of waterfowl in epidemiological links, and a severe disease mortality of approximately 60% in humans. An inconceivable number of possible genetic variations of the influenza virus leads to equally an infinite number of biological and pathological properties.

          The influenza virus H1N1 emerged this April in Mexico and demonstrated yet another face of the infection. Scientists and health authorities are actively discussing what can happen if H5N1 and H1N1 influenza viruses meet and exchange genome fragments, but humanity is only an observer of the evolutionary games of the virus.

          We examined experimental infection in mice caused only by three strains of H5N1 influenza virus and found several different targets for the virus and consequently different clinical course. The most interesting are differences in pattern of blood damage caused by two strains isolated at the same place, Kurgan region. Our findings are direct evidence that there are "non-respiratory" targets for influenza H5N1 virus in mice, and correspond to known clinical parameters of H5N1 influenza infection in humans. Damage of blood system in influenza H5N1 infection calls for special attention because it places "familiar flu" into the category of systemic severe diseases. An analysis of our results and published studies clearly shows that the influenza virus has the ability to alter fine homeostatic mechanisms and thereby produce systemic disease with extremely diverse pathogenesis events.

          Our knowledge of the pathogenetic mechanisms of influenza infection is poor. The existing view of influenza as a local respiratory disease, sometimes complicating with following bacterial infection, is unable to explain the clinical course of H5N1 influenza illness. So, there is a need to collect pathogenesis details of infection caused by various subtypes and strains of influenza A virus in order to develop new concept of modern influenza and elaborate new means and approaches or treatment.

Ed. note: Many thanks to Prof. Ryabchikova for her presentation of this paper at the CBMTS Industry VI in Cavtat, Croatia in April 2009 and for her update to the paper for this ASA edition. The accomplishments of Dr. Elena Ryabchikova and her professional associates at VECTOR have been of exceptional importance to global understanding of viruses and world health.

For the Professional in Government and Industry with an interest in Nuclear, Biological and Chemical Defense, Disarmament and Verification; Emergency and Disaster Medical Planning; Industrial Health and Safety; and Environmental Protection