Evaluation of human polyomavirus BK as a potential cause of villitis of unknown etiology and spontaneous abortion^†‡

INTRODUCTION

Polyomavirus BK (BKV) is a well-recognized pathogen involved in allograft infection after renal transplantation, and is also known to cause interstitial nephritis and hemorrhagic cystitis in patients with other forms of immunodeficiency [Randhawa et al., 1999; Hirsch and Steiger, 2003; Batal et al., 2009]. Latent BKV infection is widely present in healthy individuals, being detected as early as in childhood [Hirsch and Steiger, 2003; Knowles et al., 2003; Zhong et al., 2009]. In addition to its pathogenic role in immunocompromised hosts, there is evidence of symptomatic infection in healthy individuals; it has also been proposed that BKV may trigger autoimmune processes and neoplasms [Fredriksen et al., 1994; Vats, 2008], and cause chromosome aberrations [Trabanelli et al., 1998].

The exact mechanisms of BKV transmission are unclear, but are believed to primarily involve upper respiratory and oral routes [Goudsmit et al., 1982; Sundsfjord et al., 1994]. Transplacental transmission of murine polyomavirus has been demonstrated [McCance and Mims, 1977; Zhang et al., 2005], and two studies in humans detected BKV genomic sequences in placental and fetal tissues from up to 80% of abortuses [Pietropaolo et al., 1998; Boldorini et al., 2010]. BKV viruria suggestive of reactivation infection has been documented in up to 53% of pregnant women through PCR amplification of viral sequences [Bhattacharjee and Chakraborty, 2004; Boldorini et al., 2008; Kalvatchev et al., 2008], and is supported by the occurrence of high IgM titers in up to 7.5% pregnant women [Taguchi et al., 1975; Shah et al., 1980; Gibson et al., 1981]. However, these studies failed to identify BKV-specific IgM titers or DNA from infants' blood, arguing against vertical transmission of pathologically significant viral loads [Taguchi et al., 1975; Shah et al., 1980; Gibson et al., 1981; Boldorini et al., 2008].

Despite the high prevalence of BKV infection in humans and reactivation during pregnancy, a possible role for BKV in the pathogenesis of placental disorders has not been explored. Although many placental lesions have identifiable causes, the etiology of some disorders remains unexplained. While most spontaneous abortions, for instance, are thought to develop in response to karyotypic abnormalities, the etiology of those spontaneous abortions with normal karyotypes is rarely determined. Another example is villitis of unknown etiology (VUE), a common placental pathological process characterized by a histopathological pattern of focal to patchy inflammation of chorionic villi, which is frequently associated with low birth-weight, stillbirth, and premature delivery [Knox and Fox, 1984; Redline and Patterson, 1994; Becroft et al., 2005]. Although the mechanisms of inflammatory response in VUE have been elucidated [Redline and Patterson, 1993; Myerson et al., 2006], an underlying infectious cause has not been identified, and the eliciting antigen remains unknown [Redline, 2007].

MATERIALS AND METHODS

This study was approved by our institutional board review (IRB), under protocol number PRO09050086. Cases were retrospectively collected through review of 673 pathology reports of third trimester placentas diagnosed as chronic villitis from January 2008 through August 2009. One hundred two cases of VUE were identified based on the following criteria [Redline, 2007]: focal or patchy lymphohistiocytic villitis involving less than 10% of villi in a non-uniform pattern, no involvement of membranes, chorion or cord, and no documented infectious etiology; cases with less than 1% of villi involvement were excluded. In addition, 25 cases of severe, widespread villitis (diffuse villitis) with involvement of 15–75% of villi and/or focal plasmacytic infiltrate were also selected. We also selected 22 cases of spontaneous abortion without an identifiable cause of fetal loss in which minimally degenerated villi were available, corresponding to 15 first trimester and 7 second trimester abortions.

All cases were reviewed microscopically, and the percentage of villi involvement recorded for each case. VUE cases were graded as either high-grade (more than 10 villi involved per focus) or low-grade villitis (less than 10 villi involved per focus) according to the extent of involvement [Redline, 2007]. Representative paraffin blocks were selected from each case, and additional sections obtained. Clinical information obtained from medical charts included maternal age, gestational age, placental weight and results of laboratory screening for infectious agents, when available.

In situ hybridization (ISH) with biotinylated DNA probes against BKV (ENZ-40848, Enzo Life Sciences™, Plymouth Meeting, PA) was performed on representative paraffin-embedded sections from all cases following a standard procedure in use at the University of Pittsburgh Medical Centre In-situ hybridization Laboratory. Renal tissue from a kidney transplant patient with BKV nephropathy was used as a positive control for quality control.

RESULTS

For the VUE cases, maternal age ranged from 16 through 40 years (average 29.4 years), and the gestational age ranged from 30 to 42 weeks (average 37.6 weeks). For cases with diffuse villitis, maternal age ranged from 20 to 40 years (average 29.7 years), and the gestational age ranged from 31 to 41 weeks (average 37.2 weeks). The placental weight was below the 25th percentile in 43 out of 102 VUE specimens (42.15%), and in 14 out of 25 diffuse villitis cases (56%).

Histopathological review of the 102 VUE cases revealed 55 high-grade and 57 low-grade VUE; small foci of plasma cells were seen in 9 cases (8.8%). All 25 cases of diffuse villitis showed at least one focus of plasma cell infiltration, and the average villi involvement by the inflammatory infiltrate in these cases was 30.4% (range 15–75%). Viral inclusions were not seen in any of the cases of VUE, diffuse villitis or spontaneous abortion. All the 102 VUE and 25 diffuse villitis cases were negative for the BKV probe by ISH. In addition, the villi were negative in all the 22 cases of spontaneous abortion.

DISCUSSION

In this study we investigated a potential role of BKV infection in fetoplacental pathology, supported by the previously reported findings of viral reactivation during pregnancy and vertical transmission. Our negative results in a large number of cases of VUE and diffuse villitis suggest that BKV is unlikely to play a significant role in their etiology; similar results were found in a smaller sample of first and second trimester spontaneous abortions.

VUE represents a frequent finding in placental pathology, occurring in up to 15% of all third trimester placentas. It is associated with low birth-weight and intrauterine growth restriction [Knox and Fox, 1984; Redline and Patterson, 1994; Becroft et al., 2005]; severe and/or recurrent VUE has been linked to stillbirth, premature delivery, and neurological impairment [Redline, 2007]. The pathogenesis of VUE involves an inflammatory process mediated mainly by maternal CD8 positive T-cells in terminal villi, with participation of fetal histiocytes (Hofbauer cells) and chemokines [Redline and Patterson, 1993; Myerson et al., 2006; Kim et al., 2009], although the target antigens have not been identified. Increased expression of major histocompatibility complex (MHC) class I and II molecules has been documented in VUE placentas, suggesting a possible allogeneic response as the cause of the inflammation [Labarrere and Faulk, 1990; Kim et al., 2009]. This theory is supported by a similar pattern of chemokine and MHC antigen expression in VUE, allograft rejection and graft-versus-host disease [Kim et al., 2009]. However, the possibility of an underlying infectious cause, with increased expression of MHC types I and II molecules induced by microbial antigens, has not been ruled out.

BKV infection is known to trigger an inflammatory response in the human kidney similar to what is observed in placentas with VUE. This response reflects the display of a variety of epitopes in BKV infected cells, which are recognized by T-cells [Randhawa et al., 2006]. Prior observations regarding BKV reactivation during pregnancy, transplacental transmission and possible induction of chromosomal anomalies made BKV an interesting candidate for the etiology of VUE and spontaneous abortion. This prompted us to perform this study, since documentation of BKV infection in perinatal diseases would have significant implications with respect to clinical diagnosis and therapeutic management.

Unfortunately, our ISH studies showed no viral DNA in the tissues examined. More sensitive assays such as PCR are available for the detection of BKV, although we have detected viral loads as low as 59 copies per cell through ISH in cases of BKV nephropathy [Randhawa et al., 2002]. However, detection of low-level viral loads in this manner does not correlate with clinical disease. In fact, PCR can detect latent BKV in blood samples of up to 55% of healthy subjects [Chatterjee et al., 2000]. The limited specificity of PCR in detecting active infection is also supported by the previous reports of BKV DNA in aborted fetal tissue and placenta, including a recently published study [Pietropaolo et al., 1998; Boldorini et al., 2010]; in both studies, a pathogenic effect of BKV in these tissues could not be inferred despite amplification of viral sequences. Overall our negative studies, taken in conjunction with the published literature, suggest that active BKV infection is unlikely to play a role in the pathogenesis of common fetoplacental disorders.