Attempt to classify the clinical impact of DNA viruses according to the ability to activate the innate immune system†
We declare that we have no conflict of interest or funding source related to this manuscript.
Abstract
The innate immune system represents the first barrier a microbe has to meet in order to colonize and infect the host. Even in vertebrates, the adaptive immune system is polarized on the basis of what is sensed by the innate immune system. Viruses interact with cells of the innate immune system mainly via Toll-like receptors (TLRs): in particular, unmethylated CpG dinucleotides in the genome of DNA viruses can bind TLR9 [Krug et al., 2001] in addition to KIR3DL2 [Sivori et al., 2010]. TLR-9 binding can have stimulatory or inhibitory effects according to the exact sequence of the CpG motif [Krieg, 2002], and the cumulative effect of stimulatory and inhibitory motifs within a genome has been called the “CpG index.”
Extending previous observations by other investigators [Hoelzer et al., 2008], we calculated the CpG index for many single- (Anelloviridae, Parvoviridae) and double-stranded (Poliomaviridae and Poxviridae) DNA viruses using genomes from the publicly available reference sequences (RefSeq) in Genbank. A spectrum of indexes ranging from positive (+3.8 for bocavirus) to negative (−6.653 for vaccinia virus) resulted (Table I). An attempt was made to group viruses according to their CpG index and we checked for common denominators in their clinical impact.
DNA viruses with a high CpG index (e.g., adenovirus type 12 [Krieg et al., 1998], herpesviruses, TTV, parvovirus B19, and bocavirus) stimulate strongly innate immunity with a potential of leading to autoimmunity [Maggi and Bendinelli, 2009] and are subject to a strict immunosurveillance. When they are not cleared they usually become latent accompanied by methylation of CpG motifs [Winocour et al., 1965; Gunthert et al., 1976; Szyf et al., 1985; Badal et al., 2003], and increase in replication only during mild immunosuppression [Maggi et al., 2008], but are rarely directly pathogenic per se.
DNA viruses with neutral CpG index (e.g., polyomavirus and parvovirus 4) are likely to become symbionts and make life-threatening damage only during heavy immunosuppression, either iatrogenic or disease associated.
In contrast, DNA viruses with negative CpG indexes (e.g., adenovirus type 2 and 5 [Krieg et al., 1998], and poxviruses), by evading innate immunity, are subject to very poor immune surveillance and can replicate without control, resulting in life-threatening illness or chronic infection even in immunocompetent hosts. Of interest, vaccinia virus requires additional skin damage by scarification for inducing tissue inflammation in order to elicit protective immunity.
It is suggested that a classification of DNA viruses according to their ability to stimulate the innate immune system is now feasible and clinically sound. Although this classification cannot make predictions on cell tropism, prevalence in the healthy population, or ability to be cleared effectively by the adaptive immune system, it may be useful for predicting the pathogenicity of emerging DNA viruses. J. Med. Virol. 83:1060–1062, 2011. © 2011 Wiley-Liss, Inc.
| Family | Virus | No. isolates | Genome | Ratio between observed and expected frequencies of indicated motifsa | CpG indexe | |||
|---|---|---|---|---|---|---|---|---|
| Size (kb) | CpG per kbb | Total CpGc | CpG-Sd | CpG-Id | ||||
| Anelloviridae | TTV | 44 | 3.760f (3.570–3.853) | 48.8 (38.7–61.6) | 74.5 (64.2–84.2) | 80.0 (20.1–125.6) | 75.6 (62.7–85.9) | 2.0 (−24.0 to 18.7) |
| TTMV | 22 | 2.872 (2.761–2.950) | 17.9 (14.4–21.9) | 51.1 (43.3–57.9) | 53.8 (16.6–113.9) | 61.7 (47.7–81.1) | −0.8 (−4.71 to 5.9) | |
| TTMDV | 22 | 3.219 (3.175–3.269) | 28.9 (26.2–34.7) | 63.6 (57.2–73.6) | 57.7 (32.1–77.6) | 53.3 (44.0–67.4) | 0.8 (−4.1 to 4.6) | |
| Parvoviridae | B19 | 5 | 5.557f (5.412–5.596) | 22.08 (19.8–24.1) | 46.0 (42.0–50.0) | 80.0 (12.7–102.0) | 45.0 (0.0–61.9) | 3.7 (1.1 to 4.8) |
| Bocavirusg | 29 | 5.197 (4.948–5.301) | 22.0 (19.6–24.6) | 50.0 (44.0–54.7) | 78.7 (35.9–105.1) | 43.8 (32.4–53.3) | 3.8 (−1.6 to 7.6) | |
| PARVh | 2 | 5.268 | 22.2 | 44.0 | 48.0 | 49.5 | −0.1 | |
| Polyomaviridae | BKV | 16 | 5.134f (5.098–5.162) | 2.5 (2.3–2.7) | 6.4 (6.0–7.0) | 0.0 | 5.8 (5.1–6.6) | −0.009 (−0.01 to −0.007) |
| JCV | 11 | 5.118 (5.105–5.130) | 2.8 (2.7–3.1) | 7.0 (6.7–7.6) | 0.0 | 7.0 (6.8–7.6) | −0.014 (−0.02 to −0.01) | |
| LPV | 1 | 5.089 | 5.9 | 14.3 | 22.8 | 18.9 | 0.033 | |
| MCV | 20 | 5.371 (5.201–5.418) | 6.4 (6.1–6.7) | 15.5 (14.8–16.2) | 14.4 (14.3–14.9) | 18.4 (14.4–21.0) | −0.041 (−0.07 to −0.02) | |
| KI | 8 | 5.041 (5.040–5.050) | 6.3 (6.1–6.5) | 16.3 (16.0–17.0) | 17.4 (16.4–24.4) | 16.2 (15.6–17.0) | 0.014 (−0.01 to 0.1) | |
| WU | 27 | 5.229 (5.227–5.229) | 5.1 (5.0–5.3) | 13.4 (13.0–14.0) | 32.0 (24.0–32.3) | 17.1 (15.8–32.0) | 0.1 (0.05 to 0.16) | |
| Poxviridae | Variola | 9 | 186f (184.9–186.7) | 27.37 (27.18–27.67) | 1.024 (1.015–1.033) | 0.936 (0.876–0.973) | 1.03 (1.015–1.05) | −2.647 (−3.927 to −2.025) |
| Vaccinia | 1 | 194.7 | 30.75 | 1.1 | 0.949 | 1.145 | −6.653 | |
- a Ratio between the frequency of CpG motifs observed and that expected on the basis of genomic G + C content, if the bases were used randomly.
- b No. of CpG hexamer motifs (NNCGNN) normalized to 1 kb and that expected on the basis of genomic G + C content, if the bases were used randomly.
- c Ratio between the frequency of total CpG hexamer motifs (NNCGNN).
- d Consensus CpG-S (RRCGYY) and CpG-I (NCCGNN and NNCGRN) deduced by published papers ([Krieg et al., 1998; Lundberg et al., 2003]).
- e Calculated by the formula under Materials and Methods in Rankin et al. [2001]. A positive value is indicative of a putatively stimulatory genome and a negative value of a putatively inhibitory genome.
- f Median. In parenthesis, the range of values.
- g Included a number of 10, 10, 7, and 2 sequences of bocavirus 1, 2, 3, and 4, respectively.
- h The sequences of PARV4 were identical.
