Release of fragmented host, cell-free, genomic DNA into the circulation of pigs during infection by virulent African swine fever virus

African swine fever virus (ASFV) causes a severe hemorrhagic disease in domestic pigs resulting in high case fatality rates. The virus replicates in circulating cells of the monocyte-macrophage lineage and within lymphoid tissues, e.g. tonsils, spleen and lymph nodes. The infection results in high fever and a variety of clinical signs from about 3 days post infection. In this study, it was observed that one of many changes resulting from ASFV- infection within pigs was a large (>1000-fold) increase in the level of circulating cell-free DNA (cfDNA), including the beta-actin gene, derived from the nuclei of host cells, in the serum. This change occurred in parallel with the increase in circulating ASFV DNA. In addition, elevated levels (about 30-fold higher) of host mitochondrial DNA (mtDNA) were detected in serum from ASFV-infected pigs, but with a much higher baseline level of mtDNA in sera from uninfected pigs. The host derived cfDNA is derived from dead cells which may, or may not, have been infected. For comparison, the release of the cellular enzyme, lactate dehydrogenase (LDH), a commonly used marker of cellular damage, was also found to be elevated during the infection. The cfDNA is readily detected in serum and is a more sensitive host marker of ASFV infection than the release of mtDNA or LDH. In addition, sera from pigs infected by classical swine fever virus (CSFV), which causes a clinically similar disease as ASFV, were also tested but this infection did not result in the release of cfDNA, mtDNA or LDH. Author summary African swine fever virus causes a severe hemorrhagic disease in domestic pigs and wild boar, which often leads to death within a week. The infection results in a spectrum of different clinical signs and other changes within infected animals. In this study, we have shown, for the first time, that one consequence of infection by a highly virulent strain of this virus is the release into the blood of host genomic DNA, in a highly fragmented form. We found an increase of >1000-fold in the level of this cell-free DNA within the serum of infected animals. Furthermore, we also showed that the level of the small circular DNA from the cell mitochondria is also elevated in serum from infected animals as is the cellular enzyme lactate dehydrogenase but these changes were less marked and occurred later. The increase in the level of the cell-free host DNA is coincident with the increase in level of the viral DNA within blood and may act as a marker for infection by a highly virulent form of the virus. Remarkably, pigs infected by classical swine fever virus, which produces similar clinical signs, did not have elevated levels of these markers in their serum.

stabilized blood (EDTA-blood) and peripheral blood mononuclear cells (PBMCs)) from these 115 pigs have been described previously [23,25] but no analysis of serum samples, which are 116 described here, has been reported previously. The serum samples were obtained from blood 117 samples collected prior to inoculation at 0 dpi and at 3, 5 and 6 dpi. The pigs were euthanized 118 at 6 dpi. 119 In both experiments, water and a commercial diet for weaned pigs were provided ad 120 libitum. EDTA-blood and unstabilized blood samples (for serum preparation) were collected 121 prior to inoculation on day 0 and at indicated days post inoculation (dpi). All samples were 122 stored at -80•C until further analysis. Rectal temperatures were recorded and a total clinical 123 score was calculated on all sampling days, as described previously [23]. The pigs were 124 euthanized at the end of the study period by intravascular injection of Pentobarbital following 125 deep anesthesia. and at 4, 7, 10, 11 or 22 dpi as indicated [26]. These serum samples were stored frozen at -20 139 •C until further analysis.  3.1. protocol, as described previously [14]. The extracted samples were analyzed for the 145 presence of ASFV DNA by qPCR or CSFV RNA by RT-qPCR using the CFX Opus  Time PCR System (Bio-Rad, Hercules, CA, USA), essentially as described [27,28]. For both 147 assays, a positive result was defined as a threshold cycle value (Ct) at which FAM dye 148 emission appeared above background within 42 cycles. 149 Genomic and mitochondrial genome detection 150 For host DNA detection, the level of the Sus scrofa cytoskeletal β-actin gene in the 151 samples was determined using the ACTB-F and ACTB-R primers as described [27], while the 152 level of the Sus scrofa mitochondrial cytochrome b gene was determined using an assay 153 developed by Forth [29]. The qPCRs were performed using the CFX Opus Real-Time PCR  prior to inoculation (day 0) and at 3, 5, 6 and 7 days post inoculation (dpi). These whole 196 blood samples, from each sampling day, were assayed for the presence of ASFV DNA by 197 qPCR; note that results are presented in the graphs as 42-Ct values (see Figure 1A). As 198 expected, no ASFV DNA was present at 0 dpi. Low levels of ASFV DNA were detected at 3 199 dpi in 3 of the 4 pigs (as observed previously using this virus isolate, [23]) and much higher 200 levels were observed at 5, 6 and 7 dpi. Most of the animals were euthanized at 7 dpi but pig 201 13 was euthanized on day 6 due to severe clinical disease and thus could not be sampled on 202 day 7.

203
The body temperature and a clinical score (determined as described in Materials and 204 Methods) for each pig were also recorded on each of the sampling days (see Figure 2). A 205 marked increase in body temperature (Figure 2A) was apparent in each pig at 5 dpi and 206 various clinical signs (e.g. lethargy and anorexia) also began to appear at this time ( Figure   207 2B) and developed further to give the highest clinical scores at 6 and 7 dpi. Prior to 208 euthanasia, reddening of the skin and neurological signs (unsteady walk and sometimes 209 convulsions) were observed. As indicated above, pig 13 had to be euthanized at 6 dpi. As may be expected, the increased body temperatures and elevated clinical scores coincided with 211 the large increase in the presence of ASFV DNA in the blood ( Figure 1A) essentially from 5 212 dpi.

213
To ensure that the DNA extractions and qPCR assays were functional, the ASFV 214 DNA was assayed as part of a duplex assay including primers and probes that also detected 215 the gene encoding the cytoskeletal β-actin (part of the genomic DNA (gDNA)) [30]. As 216 expected, high levels of this β-actin gene were detected in the whole blood samples that 217 included both the nucleated white blood cells (e.g. PBMCs) and the enucleated erythrocytes 218 that were collected from each pig on each sampling day ( Figure 1A). However, it was noticed 219 that markedly elevated levels of β-actin gDNA were detected in the blood from 2 of the 3 pig 220 samples taken at 7 dpi (i.e. from pigs 14 and 20) when very high levels of ASFV DNA were 221 also detected. The higher levels of β-actin gDNA in the blood were also detected in these two 222 pigs at 6 dpi ( Figure 1A). 223 It seemed possible that the elevated signals for β-actin gDNA at 6 and 7 dpi ( Figure   224 1A) resulted from the destruction of ASFV-infected blood cells that could result in the release 225 of cellular genomic DNA into the blood. To test for the release of cellular DNA into the 226 blood, serum samples (lacking all blood cells) that had been prepared, from the same group 227 of animals, using unstabilized blood samples, collected prior to infection and at euthanasia (at 228 6 or 7 dpi) were also assayed for the presence of β-actin gDNA. At 0 dpi, the levels of β-actin 229 gDNA in the serum were low (Ct values 32.1-34.8) although readily detectable, see Figure   230 1B, but at euthanasia, the level of gDNA was much higher (Ct values 21.1-25.2), i.e. a 231 difference of about 10 cycles (ca. 1000-fold increase, as 2 10 =1024). Consistent with the 232 assays using whole blood, no ASFV DNA was detected in the sera at 0 dpi (no Ct value) but  Figure S1) were converted, by reference to standard curves, to 251 genome copy numbers/ml and are shown in Figure 3A, B, C.

260
The level of the β-actin gene, as cfDNA, was very consistent, on 0 dpi, at about 10 5 261 genome copies/ml (mean = 1.33 x 10 5 copies/ml) in the serum of the 12 pigs ( Figure 3A, B, 262 C). It was little changed at 3 dpi (mean = 1.43 x 10 5 copies/ml) but had markedly increased at 263 5 dpi in the ASFV-infected animals at up to about 10 8 genome copies /ml (mean for pigs 1, 2, 264 4, 7-12 = 2.4 x 10 8 copies/ml) and continued at this high level (mean for pigs 1, 2, 4, 7-12 = 265 4.4 x 10 8 genomes copies/ml) at 6 dpi when these infected pigs also had very high levels of 266 ASFV DNA in their blood. Thus, during the time course of ASFV infection, the mean level 267 of gDNA (as measured by the level of the β-actin gene) in the sera had increased by over 268 3000-fold. Furthermore, there is an apparent correspondance between the accumulation of 269 ASFV DNA in serum and the increased presence of cfDNA, containing the β-actin gDNA.

270
The level of mtDNA in serum (see Figure 3A, B, C) for these 12 pigs at 0 dpi was 271 about 10 6 genome copies/ml (mean value = 2.5 x 10 6 copies/ml), and was similar at 3 dpi 272 (mean value = 1.7 x 10 6 copies/ml). At 5 dpi, some of the pigs had markedly elevated levels 273 of mtDNA (i.e. pigs 2, 4, 10 and 12), with a level of well over 10 7 genomes /ml (mean value 274 for these 4 pigs = 7.5 x 10 7 copies/ml). It is noteworthy that these 4 pigs also had very high 275 levels of ASFV DNA and β-actin gDNA in their serum at this time ( Figure 3A, B, C).

276
Finally, at 6 dpi, most of the pigs that were infected with ASFV also had markedly elevated 277 levels of mtDNA in their serum with nearly 10 8 genomes/ml (mean value for pigs 1, 2, 4, 7-278 12 = 7.5 x 10 7 copies/ml). Thus, between 0 and 6 dpi, the level of mtDNA in the serum of 279 ASFV-infected pigs increased by about 30-fold. It is noteworthy that, in total, 9 of the pigs 280 had markedly increased levels of mtDNA at 6 dpi but, interestingly, only four of these pigs 281 (pigs 2, 4, 10 and 12) had markedly elevated mtDNA at 5 dpi ( Figure 3A, B, C) although some other sera, i.e. from pigs 1, 9 and 11, contained high levels of both ASFV DNA and β-283 actin gDNA on each of these days ( Figure 3A, B, C).

284
These results are similar, both qualitatively and quantitatively, to those observed in 285 the ASFV-infection experiment A (Figure 1 and Figure 3D).

286
It appears that the level of mtDNA in serum is increased by ASFV infection but the 287 extent of this change is less marked than the change in β-actin gDNA (cfDNA), due ,in part, Lactate dehydrogenase (LDH) is a cytoplasmic enzyme that can be released by cells 293 when tissue damage occurs (e.g. following a heart attack). To assess whether the release of 294 genomic DNA into blood was accompanied by release of LDH, the serum samples from the 295 12 ASFV-infected pigs from Experiment B, as analyzed in Figure 3A, B, C, were assayed for 296 the presence of LDH. It was found ( Figure 4) that an increase in the level of LDH activity 297 was apparent at 5 or 6 dpi in many (but not all) of the pigs. Thus LDH release appeared to be 298 a less sensitive marker of cell damage due to ASFV infection than the release of cfDNA or 299 mtDNA.

300
Source of cfDNA in ASFV-infected sera 301 To assess the nature of the cfDNA in the serum of the pigs, selected samples were 302 extracted manually and analyzed to determine the size of the DNA fragments. At 0 and 3 dpi, 303 no DNA fragments were detected in this assay. However, it was found that at 5 or 6 dpi a  (Figures 1 and 3). There was some evidence for bands at about 200 bp and 400 bp but the pattern did not appear to represent only apoptosis. It may be that the DNA fragments 381 detected are generated by both apoptosis and necrosis. Furthermore, some degradation of the 382 fragments may occur within the serum which could increase the heterogeneity of the 383 fragment sizes.

384
The release of LDH into the circulation is a convenient marker for cell damage and is 385 widely used for this purpose. The studies presented here do indicate an increase in LDH 386 release into the serum in ASFV-infected animals ( Figure 4) and Karalyan et al.,[35] have 387 also observed an increase in LDH in sera that resulted from a genotype II ASFV infection in 388 pigs. However, in our studies this change was less consistent than the large increase in the 389 level of cfDNA or the smaller, relative increase in mtDNA identified here. Thus, the 390 production of cfDNA, and to a lesser extent mtDNA, seem to be clear markers for the severe 391 infection within animals produced by the highly virulent genotype II ASFV that is currently 392 circulating in many pig producing countries. The mechanism by which this cfDNA is 393 produced in ASFV-infected pigs is not yet known, however, the absence of such changes in 394 CSFV-infected pigs may suggest a specific role for ASFV-encoded products. Furthermore, 395 the linkage between the virulence of ASFV strains and the release of cfDNA remains to be 396 determined. Potentially, the production of cfDNA may be a useful biomarker for the severity 397 of the infection.    Group 3 (panel C)) as described in Figure 3. The samples were assayed by qPCR for the 614 presence of ASFV DNA, the cellular β-actin gene (gDNA) and mtDNA as indicated. Results

615
are presented as 42-Ct values. Absolute genome copy numbers derived from these data, using 616 standard curves, are presented in Figure 3A, B and C. As indicated in Figure