Antibodies (Ab) are critical components of immune defense against viral pathogens and key mediators of immune control during persistent infections. Evasion of antibodies, particularly in settings of limited antibody diversity, is a potent selective pressure for viral mutations. Notably, outgrowth of escape variants is an Achilles’ heel for monoclonal antibody (mAb) antiviral intervention (Bar et al., 2016; Caskey et al., 2015; Caskey et al., 2017; Mehandru et al., 2007; Toma et al., 2011; Trkola et al., 2005). Viral evolution within a host gives rise to variants that become subject to selection for resistance to antiviral antibodies (Inuzuka et al., 2018; Kinchen et al., 2018; Lynch et al., 2015). For viruses that establish persistent infection, viral variants can accumulate over time and increase the likelihood for emergence of viruses carrying mutations that escape recognition by neutralizing Ab. Polyomaviruses (PyV) persist lifelong in their hosts in a ‘smoldering’ infectious state, a lifestyle conducive for viral evolution even by a DNA virus that commandeers the high-fidelity host cell DNA polymerase. Indeed, the human BK and JC polyomaviruses have been shown to exist as quasispecies (Luo et al., 2012; Takahashi et al., 2016; Van Loy et al., 2015). Whether PyV variants are sculpted by antiviral Abs to escape humoral immunity is an open question.

JCPyV causes progressive multifocal leukoencephalopathy (PML), a frequently fatal demyelinating brain disease associated with T cell immunosuppression resulting from HIV/AIDS, organ transplant immunosuppressants, and certain chemotherapeutic and immunomodulatory therapies (Cortese et al., 2021; Pavlovic et al., 2018). Typically identified by MRI following the onset of neurologic symptoms, PML is diagnosed at a point when limited treatment options exist and patients suffer permanent CNS damage. Defining the early stages of disease involved in the transition of JCPyV from a kidney to brain pathogen would facilitate earlier detection and therapeutic intervention before the development of neurologic disease. Prolonged immune suppression is a necessary antecedent for PML. For instance, high risk for natalizumab-associated PML involves infusion therapy for >24 months and a history of immunosuppression (Berger and Fox, 2016; Bloomgren et al., 2012; Fox and Rudick, 2012). This long period of immune suppression before clinical disease manifests raises the possibility that variants of JCPyV, including those with neurovirulent potential, may emerge over time. Iatrogenic T cell and B cell ablation therapies have independently been associated with PML. Lacking is a conceptual connection between T cell deficiency, B cell deficiency, and prolonged immune suppression in JCPyV evolution that could lead to outgrowth of neurovirulent JCPyV variants.

PML is characterized by emergence of mutations in the major capsid protein, VP1, of JCPyV, which are not found in the circulating (i.e., archetype) strains (Gorelik et al., 2011; Zheng et al., 2005b; Zheng et al., 2005a). The nonenveloped icosahedral viral capsid is comprised of the VP1 protein assembled into 72 pentamers. VP1 mediates attachment to cellular receptors via its four solvent-exposed loops; these loops are also the dominant targets of the host’s neutralizing Ab response (Buch et al., 2015; Lindner et al., 2019; Neu et al., 2010). JCPyV-PML VP1 mutations, which are situated in these loops, mediate resistance to neutralizing antibodies, but also alter receptor binding and viral tropism (Geoghegan et al., 2017; Gorelik et al., 2011; Jelcic et al., 2015; Lauver et al., 2020; Maginnis et al., 2013; Ray et al., 2015). Sera from PML patients, as well as some healthy individuals, fail to neutralize particular JCPyV-PML VP1 mutations, despite these individuals having high antibody titers against WT JCPyV (Ray et al., 2015). The conditions that lead to the outgrowth of these variants in PML patients remain poorly understood, in part due to the lack of animal models for studying the early stages of JCPyV pathogenesis.

Because PyVs only replicate in their host species, we developed mouse polyomavirus (MuPyV) as a model for JCPyV pathogenesis and immunity (Ayers et al., 2021). MuPyV shares several features with JCPyV, including asymptomatic disease in immunocompetent hosts, the kidney as a dominant reservoir of virus persistence, and control by virus-specific T cells. Although both viruses bind to cell surface sialylated glycan receptors, the attachment receptor for JCPyV is the oligosaccharide lactoseries tetrasaccharide c, whereas MuPyV binds the gangliosides GD1a and GT1b (Buch et al., 2015; Neu et al., 2010; Tsai et al., 2003).

We previously identified a neutralizing mAb against MuPyV VP1 (Swimm et al., 2010). The epitope for this mAb is the dominant target of the endogenous VP1 Ab response in mice. Importantly, this mAb failed to recognize several VP1 mutations introduced into WT MuPyV that mapped to mutations seen in JCPyV isolates from PML patients (Lauver et al., 2020). In this study, we utilized passive immunization with this MuPyV VP1 mAb in mice lacking endogenous Abs to model the emergence of VP1 variant viruses under conditions of a limited antiviral Ab response. We found that T cell loss during persistent infection in these mice led to outgrowth of MuPyV variants carrying mutations in similar regions of the external VP1 loops to those seen in PML patients. Although an array of VP1 mutations arose in the kidneys in T-cell-depleted mice, the primary driver selecting replication-competent mutant viruses, including rare neuropathogenic variants, was evasion of Ab neutralization. With the MuPyV infection model, we provide evidence that T cell deficiency, coupled with limited Ab coverage of VP1 epitopes, are both required for outgrowth of Ab-escape viral variants, including those with a potential for neurovirulence.

Although depressed anti-JCPyV T cell immunity is the dominant risk factor for PML, JCPyVs in the CNS of PML patients have VP1 mutations that allow evasion of antiviral antibodies (Cortese et al., 2021; Jelcic et al., 2015; Lauver et al., 2020; Pavlovic et al., 2018; Ray et al., 2015). To connect T cell deficiency and Ab-escape JCPyV variants, we investigated T cell insufficiency in mice infected with MuPyV in the setting of a restricted VP1 antibody response. Our findings allowed us to develop a model integrating T cell and humoral deficiencies with emergence of neurovirulent PyV, as illustrated in Figure 7. First, impaired T cell control results in increase PyV replication in the kidney, the central reservoir for persistent PyV infection. Elevated viral replication is accompanied by low-level, stochastic mutagenesis of the viral genome, where particular mutations in VP1 abrogate binding by VP1-specific Ab. By extension, the host must have an inherited or acquired VP1 Ab response that targets few VP1 epitopes. A subset of viruses carrying Ab-escape VP1 mutations acquire the capacity to replicate in the CNS. T cell deficiency further handicaps a second line of defense by antiviral T cells to control Ab-escape VP1 variant viruses. In summary, T cell deficiency acts at two levels to set the stage for emergence of antibody-escape VP1 mutations endowed with neurotropic potential.

Figure 7

Download asset Open asset

The kidney is the major site of JCPyV persistence (Berger et al., 2017). We observed preferential binding of MuPyV to the DCT, mirroring what has been reported for JCPyV, and virus replication in the DCT following T cell loss (Figures 1D and 5G; Haley et al., 2015). This localization of virus binding to the DCT, despite differences in the sialylated glycans bound by each virus, suggests a conserved site of persistence within the kidney for these viruses. Virus binding and replication within the DCT may provide a level of local protection from neutralizing antibodies, necessitating T cell control to limit virus replication within the local kidney environment. Conditions of immune suppression deprive the kidney of this cellular immune control over PyV infection, leading to elevated virus replication there. Consistent with this are reports of depressed JCPyV-specific T cell responses in individuals treated with natalizumab and increased JCPyV shedding in the urine of immunosuppressed individuals (Behzad-Behbahani et al., 2004; Chen et al., 2009; Delbue et al., 2015; Perkins et al., 2012). T cells are the predominant cell type responsible for the production of interferon gamma (IFN-γ), which limits BKPyV replication in kidney cells and is necessary for effective control of MuPyV kidney infection (Abend et al., 2007; Byers et al., 2007; Wilson et al., 2011). T cell loss during PyV infection thus eliminates the major source of IFN-γ and virus control.

PML is associated with conditions of CD4 T cell impairment, including HIV/AIDS and idiopathic CD4 T cell lymphopenia (Berger et al., 1987; Berger et al., 1998; Pavlovic et al., 2018). Decreases in CD4, but not CD8, T cells are seen in the CSF of natalizumab-treated MS patients and CD4 T cell epitope-escape JCPyVs have been found in PML patients (Jelcic et al., 2016; Schneider-Hohendorf et al., 2014; Stüve et al., 2006). We observed higher kidney virus levels at 10 days post depletion and a higher frequency of viremia in CD4 T cell-depleted mice than in CD8 T cell-depleted mice (Figures 1C and 2B). CD4 T cells can exert direct antiviral effector activities as well as provide help to maintain optimal virus-specific CD8 T cell responses in nonlymphoid organs, including the brain and lung (Ren et al., 2020; Son et al., 2021). Tissue-resident CD4 T cells have also been shown to promote development of protective B cell responses in the lungs of influenza virus-infected mice (Son et al., 2021). It will be important to elucidate how CD4 T cells restrict PyV replication in the kidneys and viremia by Ab-escape variant viruses.

Utilizing the host DNA replication machinery, PyVs are regarded as having highly stable genomes with a low incidence of mutations (Sanjuán and Domingo-Calap, 2016). Clinical and experimental data, however, indicate that mutations arise during the course of PyV infection. JCPyV and BKPyV isolates from PML and polyomavirus-associated nephropathy patients, respectively, are characterized by mutations in both the VP1 capsid protein and the non-coding control region (NCCR) (Boldorini et al., 2009; Gorelik et al., 2011; Krautkrämer et al., 2009; Peretti et al., 2018; Reid et al., 2011; Tremolada et al., 2010; Vaz et al., 2000; Zheng et al., 2005b; Zheng et al., 2005a). Likewise, rearrangements in the NCCR of BKPyV emerge after extended passage in vitro, and VP1 mutations arise in BKPyV and MuPyV after serial passage in the presence of neutralizing mAb (Lauver et al., 2020; Lindner et al., 2019; Zhao and Imperiale, 2021). These data suggest that conditions favoring certain mutations (e.g., mutations enhancing replication or enabling antibody evasion) can promote the emergence and outgrowth of viruses with particular genomic alterations.

Several mechanisms provide avenues for VP1 mutations. Missense mutations may represent failed attempts by the host cell to restrict viral replication. The APOBEC3 family of proteins are cytosine deaminases that induce mutagenic damage in the genomes of DNA viruses as a mechanism of antiviral defense. APOBEC3B is upregulated during BKPyV nephritis and has been implicated in the emergence of BKPyV VP1 mutations in kidney transplant recipients and mutations in the Merkel cell PyV genome (Peretti et al., 2018; Que et al., 2021). The presence of deletion mutations could represent the remnants of double-stranded breaks repaired by nonhomologous end joining (NHEJ). DNA double strand breaks occur during PyV infection, and aspects of the DNA damage response pathway are required for efficient PyV replication (Erickson and Garcea, 2019; Heiser et al., 2016; Jiang et al., 2012; Justice et al., 2019; Sowd et al., 2013). The NCCR of BKPyV undergoes significant recombination events resulting from NHEJ during replication in vitro (Zhao and Imperiale, 2021). It is possible that NHEJ repair of breaks in VP1 during replication results in a variety of recombinations and small deletions. Those that are tolerated by the virus and confer a selective advantage, in this case in-frame deletions conferring Ab escape, are thus able to emerge and be detected.

JCPyV VP1 mutations identified in PML patients are associated with impaired receptor binding and infectivity. Only archetype JCPyV is found in the urine of PML patients, despite the presence of VP1 mutant virus in the blood and CSF of these patients (Geoghegan et al., 2017; Gorelik et al., 2011; Maginnis et al., 2013; Reid et al., 2011; Zheng et al., 2005a). Here, we find that the selective pressure of antibody escape drove outgrowth of VP1 mutants with varying kidney infectivity, implicating immune evasion, rather than tropism, as the primary driver of VP1 mutations under these conditions (Figure 3B–D). The initial emergence of the Δ297 mutation, despite imparting severe defects in kidney tropism, indicates that Ab escape can promote the outgrowth of otherwise crippled PyV variants, a feature seen in several other chronic viral infections (Kalinina et al., 2003; Kinchen et al., 2018; Lynch et al., 2015).

The defect in spread and infectivity in vivo by the Δ297 mutation readily selected for a variety of secondary point mutations that restored or even enhanced virulence, generating viruses able to efficiently spread in vitro. The point mutations found together with the Δ297 mutation all involved a positive net change in charge (Supplementary file 1). These mutations are all located near the receptor binding pocket, and most of the mutated side chains face toward the location of receptor binding. With the exception of E91, however, these residues have not been reported to be involved in receptor binding (Buch et al., 2015; Stehle and Harrison, 1997). Mutation of E91 to glycine causes impaired infection by promoting binding to decoy pseudoreceptors (Bauer et al., 1999; Buch et al., 2015; Qian and Tsai, 2010). Similarly, the D295N mutation alone caused pseudoreceptor binding that impaired infectivity in vitro and in vivo (Figure 5D–H). In the context of the Δ297 mutation, however, these normally deleterious mutations synergized to restore proper receptor binding and even enhance virulence. The D295N/Δ297 double mutant displayed increased infectivity and virus production in vitro (Figure 5D–F). In vivo, the double mutant virus showed similar acute kidney infection to WT virus but decreased kidney infection/pathology during chronic infection (Figures 5H and 6A–D). This impairment of kidney pathology during chronic infection suggests a defect in persistence within the kidney by the D295N/Δ297 mutant virus. Within the brain, D295N/Δ297 caused increased acute infection of the ependyma and heightened chronic infection in the periventricular region, as well as increased CNS pathology (Figure 6E–G). Retention of neurotropism and a loss of kidney tropism was previously seen with the V296F MuPyV VP1 mutant and may be a common feature of PML-associated JCPyV mutations (Lauver et al., 2020). This increased CNS virulence could stem from more efficient receptor binding and release, leading to increased viral dissemination throughout the ventricular system and elevated infection of the ependyma during persistence. Increased spread and pathogenesis resulting from altered receptor binding has been seen with a separate MuPyV HI loop mutation, V296A (Bauer et al., 1995; Bauer et al., 1999; Buch et al., 2015). As both the D295 and H297 residues face into the pore at the center of each VP1 pentamer, these mutations may facilitate more efficient interactions with the VP2/3 minor capsid proteins, whose C-termini extend into the pore (Chen et al., 1998).

The mutant viruses we identified all carried at least one mutation in the HI loop, which is also the most common site of JCPyV VP1 mutations in PML patients (Gorelik et al., 2011; Reid et al., 2011). The epitope of the VP1 mAb has significant contribution from the HI loop and competes with a large portion of the endogenous antibody response generated by MuPyV-infected mice (Lauver et al., 2020). The dominant targets of antibodies in JCPyV-infected individuals have not been determined, but given the frequency of JCPyV mutations seen in the HI loop this region may be a common target of neutralizing Ab across species. PML patients typically carry a mutant JCPyV with a single VP1 mutation, rather than the double mutants we saw emerge in several mice (Supplementary file 1; Gorelik et al., 2011; Reid et al., 2011). This difference may stem from the nature of Ab escape mutation, with substitution to a bulkier side chain in JCPyV (e.g., L55F, S266F/Y, and S268F/Y) rather than the deletion of H297 seen in MuPyV. Although both types of mutations alter/impair infectivity, bulky substitutions in JCPyV may do so by occluding a portion of the receptor binding pocket, where there may be reduced opportunity for compensation by a secondary mutation. Already the smallest external loop in VP1, the HI loop of JCPyV (M261 to S268: 8 residues) is one residue shorter than the HI loop of MuPyV (W288 to V296: 9 residues) (Liddington et al., 1991). Thus, the HI loop of JCPyV may be too short to tolerate even a single amino acid deletion and still allow proper positioning of the H and I β strands.

Insufficient humoral control of PyV mutants opens numerous pathways for viral invasion of the CNS. The lack of neutralization could promote infection of hematopoietic cells that then traffick to the brain (Dubois et al., 1997; Wollebo et al., 2015). Alternatively, the absence of neutralizing antibody could enable virus within the vasculature of the brain to infect the cells of the blood-brain barrier (BBB) or blood-CSF barrier (BCSFB). Brain microvascular endothelial cells are susceptible to JCPyV infection, and could provide viral access into the CNS parenchyma (Chapagain et al., 2007). The BCSFB, formed by the choroid plexus in the brain ventricles, is another possible route of entry by the virus into the brain. JCPyV infects choroid plexus epithelial cells (Corbridge et al., 2019; Haley et al., 2015; O’Hara et al., 2018; O’Hara et al., 2020). Infection of choroid plexus epithelium would enable virus spread into the CSF, dissemination throughout the ventricular system, and access to the brain parenchyma via infection of the ependymal cells lining the ventricles. In line with this possibility is the permissivity of ependymal cells for productive MuPyV infection (Lauver et al., 2020; Mockus et al., 2020).

Our findings demonstrate a ‘2-hit’ requirement for development of VP1 mutant PyVs: (1) a narrow VP1 antibody response; and (2) T cell insufficiency, which acts both to limit virus replication in the kidney and eliminate VP1 variants that evade antiviral antibodies. T cell loss leads to elevated virus replication that in the context of a narrow VP1 antibody response results in viremia by Ab-escape mutant viruses. Among these Ab-escape variants will be those carrying the potential for neurovirulence. The stochastic evolution of VP1 mutations, together with depressed anti-PyV T and B cell immunity, may account for the long timeframe between PML and iatrogenic immunosuppression, the array of PML-associated agents with unrelated mechanisms of action, and the rarity of PML in susceptible hosts.

All data files are uploaded as source data files with this manuscript. Images are deposited with Dryad at https://doi.org/10.5061/dryad.prr4xgxqj.

1. 1. Lukacher AE
   2. Lauver M
   3. Jin G
   4. Ayers K
   5. Carey S
   6. Specht C
   7. Abendroth C

   (2022) Dryad Digital Repository

   T cell deficiency precipitates antibody evasion and emergence of neurovirulent polyomavirus.

   https://doi.org/10.5061/dryad.prr4xgxqj

1. 1. Abend JR
   2. Low JA
   3. Imperiale MJ

   (2007) Inhibitory effect of gamma interferon on BK virus gene expression and replication

   Journal of Virology 81:272–279.

   https://doi.org/10.1128/JVI.01571-06

   • PubMed
   • Google Scholar
2. 1. Ayers KN
   2. Carey SN
   3. Lukacher AE

   (2021) Understanding polyomavirus CNS disease-a perspective from mouse models

   The FEBS Journal 289:5744–5761.

   https://doi.org/10.1111/febs.16083

   • PubMed
   • Google Scholar
3. 1. Baer PC
   2. Nockher WA
   3. Haase W
   4. Scherberich JE

   (1997) Isolation of proximal and distal tubule cells from human kidney by immunomagnetic separation: technical note

   Kidney International 52:1321–1331.

   https://doi.org/10.1038/ki.1997.457

   • PubMed
   • Google Scholar
4. 1. Bar KJ
   2. Sneller MC
   3. Harrison LJ
   4. Justement JS
   5. Overton ET
   6. Petrone ME
   7. Salantes DB
   8. Seamon CA
   9. Scheinfeld B
   10. Kwan RW
   11. Learn GH
   12. Proschan MA
   13. Kreider EF
   14. Blazkova J
   15. Bardsley M
   16. Refsland EW
   17. Messer M
   18. Clarridge KE
   19. Tustin NB
   20. Madden PJ
   21. Oden K
   22. O’Dell SJ
   23. Jarocki B
   24. Shiakolas AR
   25. Tressler RL
   26. Doria-Rose NA
   27. Bailer RT
   28. Ledgerwood JE
   29. Capparelli EV
   30. Lynch RM
   31. Graham BS
   32. Moir S
   33. Koup RA
   34. Mascola JR
   35. Hoxie JA
   36. Fauci AS
   37. Tebas P
   38. Chun T-W

   (2016) Effect of HIV antibody VRC01 on viral rebound after treatment interruption

   The New England Journal of Medicine 375:2037–2050.

   https://doi.org/10.1056/NEJMoa1608243

   • PubMed
   • Google Scholar
5. 1. Bauer PH
   2. Bronson RT
   3. Fung SC
   4. Freund R
   5. Stehle T
   6. Harrison SC
   7. Benjamin TL

   (1995) Genetic and structural analysis of a virulence determinant in polyomavirus VP1

   Journal of Virology 69:7925–7931.

   https://doi.org/10.1128/JVI.69.12.7925-7931.1995

   • PubMed
   • Google Scholar
6. 1. Bauer PH
   2. Cui C
   3. Liu WR
   4. Stehle T
   5. Harrison SC
   6. DeCaprio JA
   7. Benjamin TL

   (1999) Discrimination between sialic acid-containing receptors and pseudoreceptors regulates polyomavirus spread in the mouse

   Journal of Virology 73:5826–5832.

   https://doi.org/10.1128/JVI.73.7.5826-5832.1999

   • PubMed
   • Google Scholar
7. 1. Behzad-Behbahani A
   2. Klapper PE
   3. Vallely PJ
   4. Cleator GM
   5. Khoo SH

   (2004) Detection of BK virus and JC virus DNA in urine samples from immunocompromised (HIV-infected) and immunocompetent (HIV-non-infected) patients using polymerase chain reaction and microplate hybridisation

   Journal of Clinical Virology 29:224–229.

   https://doi.org/10.1016/S1386-6532(03)00155-0

   • PubMed
   • Google Scholar
8. 1. Berger JR
   2. Kaszovitz B
   3. Post MJ
   4. Dickinson G

   (1987) Progressive multifocal leukoencephalopathy associated with human immunodeficiency virus infection: A review of the literature with A report of sixteen cases

   Annals of Internal Medicine 107:78–87.

   https://doi.org/10.7326/0003-4819-107-1-78

   • PubMed
   • Google Scholar
9. 1. Berger JR
   2. Pall L
   3. Lanska D
   4. Whiteman M

   (1998) Progressive multifocal leukoencephalopathy in patients with HIV infection

   Journal of Neurovirology 4:59–68.

   https://doi.org/10.3109/13550289809113482

   • PubMed
   • Google Scholar
10. 1. Berger JR
    2. Fox RJ

    (2016) Reassessing the risk of natalizumab-associated PML

    Journal of Neurovirology 22:533–535.

    https://doi.org/10.1007/s13365-016-0427-6

    • PubMed
    • Google Scholar
11. 1. Berger JR
    2. Miller CS
    3. Danaher RJ
    4. Doyle K
    5. Simon KJ
    6. Norton E
    7. Gorelik L
    8. Cahir-McFarland E
    9. Singhal D
    10. Hack N
    11. Owens JR
    12. Nelson PT
    13. Neltner JH

    (2017) Distribution and quantity of sites of John Cunningham virus persistence in immunologically healthy patients

    JAMA Neurology 74:437.

    https://doi.org/10.1001/jamaneurol.2016.5537

    • PubMed
    • Google Scholar
12. 1. Bloomgren G
    2. Richman S
    3. Hotermans C
    4. Subramanyam M
    5. Goelz S
    6. Natarajan A
    7. Lee S
    8. Plavina T
    9. Scanlon JV
    10. Sandrock A
    11. Bozic C

    (2012) Risk of natalizumab-associated progressive multifocal leukoencephalopathy

    The New England Journal of Medicine 366:1870–1880.

    https://doi.org/10.1056/NEJMoa1107829

    • PubMed
    • Google Scholar
13. 1. Boldorini R
    2. Allegrini S
    3. Miglio U
    4. Paganotti A
    5. Veggiani C
    6. Mischitelli M
    7. Monga G
    8. Pietropaolo V

    (2009) Genomic mutations of viral protein 1 and BK virus nephropathy in kidney transplant recipients

    Journal of Medical Virology 81:1385–1393.

    https://doi.org/10.1002/jmv.21520

    • PubMed
    • Google Scholar
14. 1. Buch MHC
    2. Liaci AM
    3. O’Hara SD
    4. Garcea RL
    5. Neu U
    6. Stehle T
    7. Meyers C

    (2015) Structural and functional analysis of murine polyomavirus capsid proteins establish the determinants of ligand recognition and pathogenicity

    PLOS Pathogens 11:e1005104.

    https://doi.org/10.1371/journal.ppat.1005104

    • PubMed
    • Google Scholar
15. 1. Byers AM
    2. Hadley A
    3. Lukacher AE

    (2007) Protection against polyoma virus-induced tumors is perforin-independent

    Virology 358:485–492.

    https://doi.org/10.1016/j.virol.2006.08.044

    • PubMed
    • Google Scholar
16. 1. Caskey M
    2. Klein F
    3. Lorenzi JCC
    4. Seaman MS
    5. West AP
    6. Buckley N
    7. Kremer G
    8. Nogueira L
    9. Braunschweig M
    10. Scheid JF
    11. Horwitz JA
    12. Shimeliovich I
    13. Ben-Avraham S
    14. Witmer-Pack M
    15. Platten M
    16. Lehmann C
    17. Burke LA
    18. Hawthorne T
    19. Gorelick RJ
    20. Walker BD
    21. Keler T
    22. Gulick RM
    23. Fätkenheuer G
    24. Schlesinger SJ
    25. Nussenzweig MC

    (2015) Viraemia suppressed in HIV-1-infected humans by broadly neutralizing antibody 3BNC117

    Nature 522:487–491.

    https://doi.org/10.1038/nature14411

    • PubMed
    • Google Scholar
17. 1. Caskey M
    2. Schoofs T
    3. Gruell H
    4. Settler A
    5. Karagounis T
    6. Kreider EF
    7. Murrell B
    8. Pfeifer N
    9. Nogueira L
    10. Oliveira TY
    11. Learn GH
    12. Cohen YZ
    13. Lehmann C
    14. Gillor D
    15. Shimeliovich I
    16. Unson-O’Brien C
    17. Weiland D
    18. Robles A
    19. Kümmerle T
    20. Wyen C
    21. Levin R
    22. Witmer-Pack M
    23. Eren K
    24. Ignacio C
    25. Kiss S
    26. West AP
    27. Mouquet H
    28. Zingman BS
    29. Gulick RM
    30. Keler T
    31. Bjorkman PJ
    32. Seaman MS
    33. Hahn BH
    34. Fätkenheuer G
    35. Schlesinger SJ
    36. Nussenzweig MC
    37. Klein F

    (2017) Antibody 10-1074 suppresses viremia in HIV-1-infected individuals

    Nature Medicine 23:185–191.

    https://doi.org/10.1038/nm.4268

    • PubMed
    • Google Scholar
18. 1. Chapagain ML
    2. Verma S
    3. Mercier F
    4. Yanagihara R
    5. Nerurkar VR

    (2007) Polyomavirus JC infects human brain microvascular endothelial cells independent of serotonin receptor 2A

    Virology 364:55–63.

    https://doi.org/10.1016/j.virol.2007.02.018

    • PubMed
    • Google Scholar
19. 1. Chen XS
    2. Stehle T
    3. Harrison SC

    (1998) Interaction of polyomavirus internal protein VP2 with the major capsid protein VP1 and implications for participation of VP2 in viral entry

    The EMBO Journal 17:3233–3240.

    https://doi.org/10.1093/emboj/17.12.3233

    • PubMed
    • Google Scholar
20. 1. Chen Y
    2. Bord E
    3. Tompkins T
    4. Miller J
    5. Tan CS
    6. Kinkel RP
    7. Stein MC
    8. Viscidi RP
    9. Ngo LH
    10. Koralnik IJ

    (2009) Asymptomatic reactivation of JC virus in patients treated with natalizumab

    New England Journal of Medicine 361:1067–1074.

    https://doi.org/10.1056/NEJMoa0904267

    • PubMed
    • Google Scholar
21. 1. Corbridge SM
    2. Rice RC
    3. Bean LA
    4. Wüthrich C
    5. Dang X
    6. Nicholson DA
    7. Koralnik IJ

    (2019) Jc virus infection of meningeal and choroid plexus cells in patients with progressive multifocal leukoencephalopathy

    Journal of Neurovirology 25:520–524.

    https://doi.org/10.1007/s13365-019-00753-y

    • PubMed
    • Google Scholar
22. 1. Cortese I
    2. Reich DS
    3. Nath A

    (2021) Progressive multifocal leukoencephalopathy and the spectrum of JC virus-related disease

    Nature Reviews. Neurology 17:37–51.

    https://doi.org/10.1038/s41582-020-00427-y

    • PubMed
    • Google Scholar
23. 1. Delbue S
    2. Elia F
    3. Carloni C
    4. Pecchenini V
    5. Franciotta D
    6. Gastaldi M
    7. Colombo E
    8. Signorini L
    9. Carluccio S
    10. Bellizzi A
    11. Bergamaschi R
    12. Ferrante P

    (2015) Jc virus urinary excretion and seroprevalence in natalizumab-treated multiple sclerosis patients

    Journal of Neurovirology 21:645–652.

    https://doi.org/10.1007/s13365-014-0268-0

    • PubMed
    • Google Scholar
24. 1. Dialynas DP
    2. Quan ZS
    3. Wall KA
    4. Pierres A
    5. Quintáns J
    6. Loken MR
    7. Pierres M
    8. Fitch FW

    (1983)

    Characterization of the murine T cell surface molecule, designated L3T4, identified by monoclonal antibody GK1.5: similarity of L3T4 to the human leu-3/T4 molecule

    Journal of Immunology 131:2445–2451.

    • PubMed
    • Google Scholar
25. 1. Dubois V
    2. Dutronc H
    3. Lafon ME
    4. Poinsot V
    5. Pellegrin JL
    6. Ragnaud JM
    7. Ferrer AM
    8. Fleury HJ

    (1997) Latency and reactivation of JC virus in peripheral blood of human immunodeficiency virus type 1-infected patients

    Journal of Clinical Microbiology 35:2288–2292.

    https://doi.org/10.1128/jcm.35.9.2288-2292.1997

    • PubMed
    • Google Scholar
26. 1. Erickson KD
    2. Garcea RL

    (2019) Viral replication centers and the DNA damage response in JC virus-infected cells

    Virology 528:198–206.

    https://doi.org/10.1016/j.virol.2018.12.014

    • PubMed
    • Google Scholar
27. 1. Fox RJ
    2. Rudick RA

    (2012) Risk stratification and patient counseling for natalizumab in multiple sclerosis

    Neurology 78:436–437.

    https://doi.org/10.1212/WNL.0b013e318245d2d0

    • PubMed
    • Google Scholar
28. 1. Geoghegan EM
    2. Pastrana DV
    3. Schowalter RM
    4. Ray U
    5. Gao W
    6. Ho M
    7. Pauly GT
    8. Sigano DM
    9. Kaynor C
    10. Cahir-McFarland E
    11. Combaluzier B
    12. Grimm J
    13. Buck CB

    (2017) Infectious entry and neutralization of pathogenic JC polyomaviruses

    Cell Reports 21:1169–1179.

    https://doi.org/10.1016/j.celrep.2017.10.027

    • PubMed
    • Google Scholar
29. 1. Golstein P
    2. Goridis C
    3. Schmitt-Verhulst AM
    4. Hayot B
    5. Pierres A
    6. Agthoven AV
    7. Kaufmann Y
    8. Eshhar Z
    9. Pierres M

    (1982) Lymphoid cell surface interaction structures detected using cytolysis-inhibiting monoclonal antibodies

    Immunological Reviews 68:5–42.

    https://doi.org/10.1111/j.1600-065X.1982.tb01058.x

    • PubMed
    • Google Scholar
30. 1. Gorelik L
    2. Reid C
    3. Testa M
    4. Brickelmaier M
    5. Bossolasco S
    6. Pazzi A
    7. Bestetti A
    8. Carmillo P
    9. Wilson E
    10. McAuliffe M
    11. Tonkin C
    12. Carulli JP
    13. Lugovskoy A
    14. Lazzarin A
    15. Sunyaev S
    16. Simon K
    17. Cinque P

    (2011) Progressive multifocal leukoencephalopathy (PML) development is associated with mutations in JC virus capsid protein VP1 that change its receptor specificity

    The Journal of Infectious Diseases 204:103–114.

    https://doi.org/10.1093/infdis/jir198

    • PubMed
    • Google Scholar
31. 1. Haley SA
    2. O’Hara BA
    3. Nelson CDS
    4. Brittingham FLP
    5. Henriksen KJ
    6. Stopa EG
    7. Atwood WJ

    (2015) Human polyomavirus receptor distribution in brain parenchyma contrasts with receptor distribution in kidney and choroid plexus

    The American Journal of Pathology 185:2246–2258.

    https://doi.org/10.1016/j.ajpath.2015.04.003

    • PubMed
    • Google Scholar
32. 1. Heiser K
    2. Nicholas C
    3. Garcea RL

    (2016) Activation of DNA damage repair pathways by murine polyomavirus

    Virology 497:346–356.

    https://doi.org/10.1016/j.virol.2016.07.028

    • PubMed
    • Google Scholar
33. 1. Inuzuka T
    2. Ueda Y
    3. Arasawa S
    4. Takeda H
    5. Matsumoto T
    6. Osaki Y
    7. Uemoto S
    8. Seno H
    9. Marusawa H

    (2018) Expansion of viral variants associated with immune escape and impaired virion secretion in patients with HBV reactivation after resolved infection

    Scientific Reports 8:18070.

    https://doi.org/10.1038/s41598-018-36093-w

    • PubMed
    • Google Scholar
34. 1. Jelcic I
    2. Combaluzier B
    3. Jelcic I
    4. Faigle W
    5. Senn L
    6. Reinhart BJ
    7. Ströh L
    8. Nitsch RM
    9. Stehle T
    10. Sospedra M
    11. Grimm J
    12. Martin R

    (2015) Broadly neutralizing human monoclonal JC polyomavirus VP1-specific antibodies as candidate therapeutics for progressive multifocal leukoencephalopathy

    Science Translational Medicine 7:306ra150.

    https://doi.org/10.1126/scitranslmed.aac8691

    • PubMed
    • Google Scholar
35. 1. Jelcic I
    2. Jelcic I
    3. Kempf C
    4. Largey F
    5. Planas R
    6. Schippling S
    7. Budka H
    8. Sospedra M
    9. Martin R

    (2016) Mechanisms of immune escape in central nervous system infection with neurotropic JC virus variant

    Annals of Neurology 79:404–418.

    https://doi.org/10.1002/ana.24574

    • PubMed
    • Google Scholar
36. 1. Jiang M
    2. Zhao L
    3. Gamez M
    4. Imperiale MJ

    (2012) Roles of ATM and ATR-mediated DNA damage responses during lytic BK polyomavirus infection

    PLOS Pathogens 8:e1002898.

    https://doi.org/10.1371/journal.ppat.1002898

    • PubMed
    • Google Scholar
37. 1. Justice JL
    2. Needham JM
    3. Thompson SR
    4. Banks L

    (2019) BK polyomavirus activates the DNA damage response to prolong S phase

    Journal of Virology 93:e00130-19.

    https://doi.org/10.1128/JVI.00130-19

    • PubMed
    • Google Scholar
38. 1. Kalinina T
    2. Iwanski A
    3. Will H
    4. Sterneck M

    (2003) Deficiency in virion secretion and decreased stability of the hepatitis B virus immune escape mutant G145R

    Hepatology 38:1274–1281.

    https://doi.org/10.1053/jhep.2003.50484

    • PubMed
    • Google Scholar
39. 1. Kinchen VJ
    2. Zahid MN
    3. Flyak AI
    4. Soliman MG
    5. Learn GH
    6. Wang S
    7. Davidson E
    8. Doranz BJ
    9. Ray SC
    10. Cox AL
    11. Crowe JE
    12. Bjorkman PJ
    13. Shaw GM
    14. Bailey JR

    (2018) Broadly neutralizing antibody mediated clearance of human hepatitis C virus infection

    Cell Host & Microbe 24:717–730.

    https://doi.org/10.1016/j.chom.2018.10.012

    • PubMed
    • Google Scholar
40. 1. Krautkrämer E
    2. Klein TM
    3. Sommerer C
    4. Schnitzler P
    5. Zeier M

    (2009) Mutations in the BC-loop of the BKV VP1 region do not influence viral load in renal transplant patients

    Journal of Medical Virology 81:75–81.

    https://doi.org/10.1002/jmv.21359

    • PubMed
    • Google Scholar
41. 1. Lauver MD
    2. Goetschius DJ
    3. Netherby-Winslow CS
    4. Ayers KN
    5. Jin G
    6. Haas DG
    7. Frost EL
    8. Cho SH
    9. Bator CM
    10. Bywaters SM
    11. Christensen ND
    12. Hafenstein SL
    13. Lukacher AE

    (2020) Antibody escape by polyomavirus capsid mutation facilitates neurovirulence

    eLife 9:e61056.

    https://doi.org/10.7554/eLife.61056

    • PubMed
    • Google Scholar
42. 1. Liddington RC
    2. Yan Y
    3. Moulai J
    4. Sahli R
    5. Benjamin TL
    6. Harrison SC

    (1991) Structure of simian virus 40 at 3.8-A resolution

    Nature 354:278–284.

    https://doi.org/10.1038/354278a0

    • PubMed
    • Google Scholar
43. 1. Lindner JM
    2. Cornacchione V
    3. Sathe A
    4. Be C
    5. Srinivas H
    6. Riquet E
    7. Leber XC
    8. Hein A
    9. Wrobel MB
    10. Scharenberg M
    11. Pietzonka T
    12. Wiesmann C
    13. Abend J
    14. Traggiai E

    (2019) Human memory B cells harbor diverse cross-neutralizing antibodies against BK and JC polyomaviruses

    Immunity 50:668–676.

    https://doi.org/10.1016/j.immuni.2019.02.003

    • Google Scholar
44. 1. Lukacher AE
    2. Wilson CS

    (1998)

    Resistance to polyoma virus-induced tumors correlates with CTL recognition of an immunodominant H-2dk-restricted epitope in the middle T protein

    Journal of Immunology 160:1724–1734.

    • PubMed
    • Google Scholar
45. 1. Luo C
    2. Hirsch HH
    3. Kant J
    4. Randhawa P

    (2012) VP-1 quasispecies in human infection with polyomavirus BK

    Journal of Medical Virology 84:152–161.

    https://doi.org/10.1002/jmv.22147

    • PubMed
    • Google Scholar
46. 1. Lynch RM
    2. Boritz E
    3. Coates EE
    4. DeZure A
    5. Madden P
    6. Costner P
    7. Enama ME
    8. Plummer S
    9. Holman L
    10. Hendel CS
    11. Gordon I
    12. Casazza J
    13. Conan-Cibotti M
    14. Migueles SA
    15. Tressler R
    16. Bailer RT
    17. McDermott A
    18. Narpala S
    19. O’Dell S
    20. Wolf G
    21. Lifson JD
    22. Freemire BA
    23. Gorelick RJ
    24. Pandey JP
    25. Mohan S
    26. Chomont N
    27. Fromentin R
    28. Chun TW
    29. Fauci AS
    30. Schwartz RM
    31. Koup RA
    32. Douek DC
    33. Hu Z
    34. Capparelli E
    35. Graham BS
    36. Mascola JR
    37. Ledgerwood JE
    38. VRC 601 Study Team

    (2015) Virologic effects of broadly neutralizing antibody VRC01 administration during chronic HIV-1 infection

    Science Translational Medicine 7:319ra206.

    https://doi.org/10.1126/scitranslmed.aad5752

    • PubMed
    • Google Scholar
47. 1. Maginnis MS
    2. Ströh LJ
    3. Gee GV
    4. O’Hara BA
    5. Derdowski A
    6. Stehle T
    7. Atwood WJ

    (2013) Progressive multifocal leukoencephalopathy-associated mutations in the JC polyomavirus capsid disrupt lactoseries tetrasaccharide C binding

    MBio 4:e00247-13.

    https://doi.org/10.1128/mBio.00247-13

    • PubMed
    • Google Scholar
48. 1. Maru S
    2. Jin G
    3. Desai D
    4. Amin S
    5. Lauver MD
    6. Lukacher AE

    (2017) Inhibition of retrograde transport limits polyomavirus infection in vivo

    MSphere 2:e00494-17.

    https://doi.org/10.1128/mSphereDirect.00494-17

    • PubMed
    • Google Scholar
49. 1. Mehandru S
    2. Vcelar B
    3. Wrin T
    4. Stiegler G
    5. Joos B
    6. Mohri H
    7. Boden D
    8. Galovich J
    9. Tenner-Racz K
    10. Racz P
    11. Carrington M
    12. Petropoulos C
    13. Katinger H
    14. Markowitz M

    (2007) Adjunctive passive immunotherapy in human immunodeficiency virus type 1-infected individuals treated with antiviral therapy during acute and early infection

    Journal of Virology 81:11016–11031.

    https://doi.org/10.1128/JVI.01340-07

    • PubMed
    • Google Scholar
50. 1. Mockus TE
    2. Netherby-Winslow CS
    3. Atkins HM
    4. Lauver MD
    5. Jin G
    6. Ren HM
    7. Lukacher AE
    8. Banks L

    (2020) CD8 T cells and STAT1 signaling are essential codeterminants in protection from polyomavirus encephalopathy

    Journal of Virology 94:e02038-19.

    https://doi.org/10.1128/JVI.02038-19

    • PubMed
    • Google Scholar
51. 1. Neu U
    2. Maginnis MS
    3. Palma AS
    4. Ströh LJ
    5. Nelson CDS
    6. Feizi T
    7. Atwood WJ
    8. Stehle T

    (2010) Structure-function analysis of the human JC polyomavirus establishes the lstc pentasaccharide as a functional receptor motif

    Cell Host & Microbe 8:309–319.

    https://doi.org/10.1016/j.chom.2010.09.004

    • PubMed
    • Google Scholar
52. 1. O’Hara BA
    2. Gee GV
    3. Atwood WJ
    4. Haley SA

    (2018) Susceptibility of primary human choroid plexus epithelial cells and meningeal cells to infection by JC virus

    Journal of Virology 92:e00105-18.

    https://doi.org/10.1128/JVI.00105-18

    • PubMed
    • Google Scholar
53. 1. O’Hara BA
    2. Morris-Love J
    3. Gee GV
    4. Haley SA
    5. Atwood WJ

    (2020) Jc virus infected choroid plexus epithelial cells produce extracellular vesicles that infect glial cells independently of the virus attachment receptor

    PLOS Pathogens 16:e1008371.

    https://doi.org/10.1371/journal.ppat.1008371

    • PubMed
    • Google Scholar
54. 1. Pavlovic D
    2. Patel MA
    3. Patera AC
    4. Peterson I
    5. Consortium PML

    (2018) T cell deficiencies as a common risk factor for drug associated progressive multifocal leukoencephalopathy

    Immunobiology 223:508–517.

    https://doi.org/10.1016/j.imbio.2018.01.002

    • PubMed
    • Google Scholar
55. 1. Peretti A
    2. Geoghegan EM
    3. Pastrana DV
    4. Smola S
    5. Feld P
    6. Sauter M
    7. Lohse S
    8. Ramesh M
    9. Lim ES
    10. Wang D
    11. Borgogna C
    12. FitzGerald PC
    13. Bliskovsky V
    14. Starrett GJ
    15. Law EK
    16. Harris RS
    17. Killian JK
    18. Zhu J
    19. Pineda M
    20. Meltzer PS
    21. Boldorini R
    22. Gariglio M
    23. Buck CB

    (2018) Characterization of BK polyomaviruses from kidney transplant recipients suggests a role for APOBEC3 in driving in-host virus evolution

    Cell Host & Microbe 23:628–635.

    https://doi.org/10.1016/j.chom.2018.04.005

    • PubMed
    • Google Scholar
56. 1. Perkins MR
    2. Ryschkewitsch C
    3. Liebner JC
    4. Monaco MCG
    5. Himelfarb D
    6. Ireland S
    7. Roque A
    8. Edward HL
    9. Jensen PN
    10. Remington G
    11. Abraham T
    12. Abraham J
    13. Greenberg B
    14. Kaufman C
    15. LaGanke C
    16. Monson NL
    17. Xu X
    18. Frohman E
    19. Major EO
    20. Douek DC

    (2012) Changes in JC virus-specific T cell responses during natalizumab treatment and in natalizumab-associated progressive multifocal leukoencephalopathy

    PLOS Pathogens 8:e1003014.

    https://doi.org/10.1371/journal.ppat.1003014

    • PubMed
    • Google Scholar
57. 1. Pullen GR
    2. Fitzgerald MG
    3. Hosking CS

    (1986) Antibody avidity determination by ELISA using thiocyanate elution

    Journal of Immunological Methods 86:83–87.

    https://doi.org/10.1016/0022-1759(86)90268-1

    • PubMed
    • Google Scholar
58. 1. Qian M
    2. Tsai B

    (2010) Lipids and proteins act in opposing manners to regulate polyomavirus infection

    Journal of Virology 84:9840–9852.

    https://doi.org/10.1128/JVI.01093-10

    • PubMed
    • Google Scholar
59. 1. Que L
    2. Li Y
    3. Dainichi T
    4. Kukimoto I
    5. Nishiyama T
    6. Nakano Y
    7. Shima K
    8. Suzuki T
    9. Sato Y
    10. Horike S
    11. Aizaki H
    12. Watashi K
    13. Kato T
    14. Aly HH
    15. Watanabe N
    16. Kabashima K
    17. Wakae K
    18. Muramatsu M

    (2021) IFN-γ‒induced APOBEC3B contributes to merkel cell polyomavirus genome mutagenesis in merkel cell carcinoma

    The Journal of Investigative Dermatology 142:1793–1803.

    https://doi.org/10.1016/j.jid.2021.12.019

    • PubMed
    • Google Scholar
60. 1. Ray U
    2. Cinque P
    3. Gerevini S
    4. Longo V
    5. Lazzarin A
    6. Schippling S
    7. Martin R
    8. Buck CB
    9. Pastrana DV

    (2015) Jc polyomavirus mutants escape antibody-mediated neutralization

    Science Translational Medicine 7:306ra151.

    https://doi.org/10.1126/scitranslmed.aab1720

    • PubMed
    • Google Scholar
61. 1. Reid CE
    2. Li H
    3. Sur G
    4. Carmillo P
    5. Bushnell S
    6. Tizard R
    7. McAuliffe M
    8. Tonkin C
    9. Simon K
    10. Goelz S
    11. Cinque P
    12. Gorelik L
    13. Carulli JP

    (2011) Sequencing and analysis of JC virus DNA from natalizumab-treated PML patients

    The Journal of Infectious Diseases 204:237–244.

    https://doi.org/10.1093/infdis/jir256

    • PubMed
    • Google Scholar
62. 1. Ren HM
    2. Kolawole EM
    3. Ren M
    4. Jin G
    5. Netherby-Winslow CS
    6. Wade Q
    7. Rahman ZSM
    8. Evavold BD
    9. Lukacher AE

    (2020) Il-21 from high-affinity CD4 T cells drives differentiation of brain-resident CD8 T cells during persistent viral infection

    Science Immunology 5:eabb5590.

    https://doi.org/10.1126/sciimmunol.abb5590

    • PubMed
    • Google Scholar
63. 1. Sanjuán R
    2. Domingo-Calap P

    (2016) Mechanisms of viral mutation

    Cellular and Molecular Life Sciences 73:4433–4448.

    https://doi.org/10.1007/s00018-016-2299-6

    • PubMed
    • Google Scholar
64. 1. Schneider-Hohendorf T
    2. Rossaint J
    3. Mohan H
    4. Böning D
    5. Breuer J
    6. Kuhlmann T
    7. Gross CC
    8. Flanagan K
    9. Sorokin L
    10. Vestweber D
    11. Zarbock A
    12. Schwab N
    13. Wiendl H

    (2014) Vla-4 blockade promotes differential routes into human CNS involving PSGL-1 rolling of T cells and MCAM-adhesion of th17 cells

    The Journal of Experimental Medicine 211:1833–1846.

    https://doi.org/10.1084/jem.20140540

    • PubMed
    • Google Scholar
65. 1. Son YM
    2. Cheon IS
    3. Wu Y
    4. Li C
    5. Wang Z
    6. Gao X
    7. Chen Y
    8. Takahashi Y
    9. Fu YX
    10. Dent AL

    (2021) Tissue-resident CD4⁺ T helper cells assist the development of protective respiratory B and CD8⁺ T cell memory responses

    Science Immunology 6:eabb6852.

    https://doi.org/10.1126/sciimmunol.abb6852

    • Google Scholar
66. 1. Sowd GA
    2. Li NY
    3. Fanning E

    (2013) ATM and ATR activities maintain replication fork integrity during SV40 chromatin replication

    PLOS Pathogens 9:e1003283.

    https://doi.org/10.1371/journal.ppat.1003283

    • PubMed
    • Google Scholar
67. 1. Stehle T
    2. Harrison SC

    (1997) High-resolution structure of a polyomavirus VP1-oligosaccharide complex: implications for assembly and receptor binding

    The EMBO Journal 16:5139–5148.

    https://doi.org/10.1093/emboj/16.16.5139

    • PubMed
    • Google Scholar
68. 1. Stüve O
    2. Marra CM
    3. Bar-Or A
    4. Niino M
    5. Cravens PD
    6. Cepok S
    7. Frohman EM
    8. Phillips JT
    9. Arendt G
    10. Jerome KR
    11. Cook L
    12. Grand’Maison F
    13. Hemmer B
    14. Monson NL
    15. Racke MK

    (2006) Altered CD4+/CD8+ T-cell ratios in cerebrospinal fluid of natalizumab-treated patients with multiple sclerosis

    Archives of Neurology 63:1383–1387.

    https://doi.org/10.1001/archneur.63.10.1383

    • PubMed
    • Google Scholar
69. 1. Sunyaev SR
    2. Lugovskoy A
    3. Simon K
    4. Gorelik L

    (2009) Adaptive mutations in the JC virus protein capsid are associated with progressive multifocal leukoencephalopathy (PML)

    PLOS Genetics 5:e1000368.

    https://doi.org/10.1371/journal.pgen.1000368

    • PubMed
    • Google Scholar
70. 1. Swimm AI
    2. Bornmann W
    3. Jiang M
    4. Imperiale MJ
    5. Lukacher AE
    6. Kalman D

    (2010) Abl family tyrosine kinases regulate sialylated ganglioside receptors for polyomavirus

    Journal of Virology 84:4243–4251.

    https://doi.org/10.1128/JVI.00129-10

    • PubMed
    • Google Scholar
71. 1. Takahashi K
    2. Sekizuka T
    3. Fukumoto H
    4. Nakamichi K
    5. Suzuki T
    6. Sato Y
    7. Hasegawa H
    8. Kuroda M
    9. Katano H
    10. Banks L

    (2016) Deep-sequence identification and role in virus replication of a JC virus quasispecies in patients with progressive multifocal leukoencephalopathy

    Journal of Virology 91:e01335-16.

    https://doi.org/10.1128/JVI.01335-16

    • PubMed
    • Google Scholar
72. 1. Tokonami N
    2. Takata T
    3. Beyeler J
    4. Ehrbar I
    5. Yoshifuji A
    6. Christensen EI
    7. Loffing J
    8. Devuyst O
    9. Olinger EG

    (2018) Uromodulin is expressed in the distal convoluted tubule, where it is critical for regulation of the sodium chloride cotransporter NCC

    Kidney International 94:701–715.

    https://doi.org/10.1016/j.kint.2018.04.021

    • PubMed
    • Google Scholar
73. 1. Toma J
    2. Weinheimer SP
    3. Stawiski E
    4. Whitcomb JM
    5. Lewis ST
    6. Petropoulos CJ
    7. Huang W

    (2011) Loss of asparagine-linked glycosylation sites in variable region 5 of human immunodeficiency virus type 1 envelope is associated with resistance to CD4 antibody ibalizumab

    Journal of Virology 85:3872–3880.

    https://doi.org/10.1128/JVI.02237-10

    • PubMed
    • Google Scholar
74. 1. Tremolada S
    2. Delbue S
    3. Castagnoli L
    4. Allegrini S
    5. Miglio U
    6. Boldorini R
    7. Elia F
    8. Gordon J
    9. Ferrante P

    (2010) Mutations in the external loops of BK virus VP1 and urine viral load in renal transplant recipients

    Journal of Cellular Physiology 222:195–199.

    https://doi.org/10.1002/jcp.21937

    • PubMed
    • Google Scholar
75. 1. Trkola A
    2. Kuster H
    3. Rusert P
    4. Joos B
    5. Fischer M
    6. Leemann C
    7. Manrique A
    8. Huber M
    9. Rehr M
    10. Oxenius A
    11. Weber R
    12. Stiegler G
    13. Vcelar B
    14. Katinger H
    15. Aceto L
    16. Günthard HF

    (2005) Delay of HIV-1 rebound after cessation of antiretroviral therapy through passive transfer of human neutralizing antibodies

    Nature Medicine 11:615–622.

    https://doi.org/10.1038/nm1244

    • PubMed
    • Google Scholar
76. 1. Tsai B
    2. Gilbert JM
    3. Stehle T
    4. Lencer W
    5. Benjamin TL
    6. Rapoport TA

    (2003) Gangliosides are receptors for murine polyoma virus and SV40

    The EMBO Journal 22:4346–4355.

    https://doi.org/10.1093/emboj/cdg439

    • PubMed
    • Google Scholar
77. 1. Van Loy T
    2. Thys K
    3. Ryschkewitsch C
    4. Lagatie O
    5. Monaco MC
    6. Major EO
    7. Tritsmans L
    8. Stuyver LJ

    (2015) Jc virus quasispecies analysis reveals a complex viral population underlying progressive multifocal leukoencephalopathy and supports viral dissemination via the hematogenous route

    Journal of Virology 89:1340–1347.

    https://doi.org/10.1128/JVI.02565-14

    • PubMed
    • Google Scholar
78. 1. Vaz B
    2. Cinque P
    3. Pickhardt M
    4. Weber T

    (2000) Analysis of the transcriptional control region in progressive multifocal leukoencephalopathy

    Journal of Neurovirology 6:398–409.

    https://doi.org/10.3109/13550280009018304

    • PubMed
    • Google Scholar
79. 1. Wilson JJ
    2. Lin E
    3. Pack CD
    4. Frost EL
    5. Hadley A
    6. Swimm AI
    7. Wang J
    8. Dong Y
    9. Breeden CP
    10. Kalman D
    11. Newell KA
    12. Lukacher AE

    (2011) Gamma interferon controls mouse polyomavirus infection in vivo

    Journal of Virology 85:10126–10134.

    https://doi.org/10.1128/JVI.00761-11

    • PubMed
    • Google Scholar
80. 1. Wilson JJ
    2. Pack CD
    3. Lin E
    4. Frost EL
    5. Albrecht JA
    6. Hadley A
    7. Hofstetter AR
    8. Tevethia SS
    9. Schell TD
    10. Lukacher AE

    (2012) Cd8 T cells recruited early in mouse polyomavirus infection undergo exhaustion

    Journal of Immunology 188:4340–4348.

    https://doi.org/10.4049/jimmunol.1103727

    • PubMed
    • Google Scholar
81. 1. Wollebo HS
    2. White MK
    3. Gordon J
    4. Berger JR
    5. Khalili K

    (2015) Persistence and pathogenesis of the neurotropic polyomavirus JC

    Annals of Neurology 77:560–570.

    https://doi.org/10.1002/ana.24371

    • PubMed
    • Google Scholar
82. 1. Zhao L
    2. Imperiale MJ

    (2021) A cell culture model of BK polyomavirus persistence, genome recombination, and reactivation

    MBio 12:e0235621.

    https://doi.org/10.1128/mBio.02356-21

    • PubMed
    • Google Scholar
83. 1. Zheng HY
    2. Ikegaya H
    3. Takasaka T
    4. Matsushima-Ohno T
    5. Sakurai M
    6. Kanazawa I
    7. Kishida S
    8. Nagashima K
    9. Kitamura T
    10. Yogo Y

    (2005a) Characterization of the VP1 loop mutations widespread among JC polyomavirus isolates associated with progressive multifocal leukoencephalopathy

    Biochemical and Biophysical Research Communications 333:996–1002.

    https://doi.org/10.1016/j.bbrc.2005.06.012

    • PubMed
    • Google Scholar
84. 1. Zheng HY
    2. Takasaka T
    3. Noda K
    4. Kanazawa A
    5. Mori H
    6. Kabuki T
    7. Joh K
    8. Oh-Ishi T
    9. Ikegaya H
    10. Nagashima K
    11. Hall WW
    12. Kitamura T
    13. Yogo Y

    (2005b) New sequence polymorphisms in the outer loops of the JC polyomavirus major capsid protein (VP1) possibly associated with progressive multifocal leukoencephalopathy

    The Journal of General Virology 86:2035–2045.

    https://doi.org/10.1099/vir.0.80863-0

    • PubMed
    • Google Scholar

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting the paper "T Cell Deficiency Precipitates Antibody Evasion and Emergence of Neurovirulent Polyomavirus" for consideration by eLife. Your article has been reviewed by three peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by a Senior Editor.

The reviewers all agreed that the results are fascinating and important, due to the role of compromised immune systems for polyoma and indeed other viruses. Nonetheless, as you will see below, the reviewers had detailed issues with experimental justification for some of the conclusions, overly complicated prose, and organization. In the cases of immune depletion, additional controls are requested.

The policy of eLife is not to accept papers that require more than a few months of work. Considering the requests for additional experimentation and, just as importantly, revision of the flow and logic of the manuscript, it is not accepted for publication or revision at this point. However, the Editors may be willing to consider a revised version if comments such as requests for genetic reconstructions, confirmations of antibody depletion, and careful editing interpretations can be fully addressed.

Below, the comments of the reviewers are combined, in general, and figure-by-figure. This organization was necessary because the writing made it difficult to distinguish which experiments supported which conclusion.

Of all manuscripts submitted to eLife, only a minority will ultimately be accepted for publication. The policy of eLife is to reject manuscripts that require major revision, to facilitate the authors' submitting the manuscript elsewhere. In the case of this interesting manuscript, we encourage resubmission provided you and your colleagues are willing to address the critiques below. Otherwise, we hope that the reviews and our feedback will help you in publishing your manuscript. Thank you for giving us the opportunity to review this work.

Summary

This is an interesting study of mouse polyomavirus (MuPyV) as a model for the human pathogen JCPyV. The latter can spread to the brain in immunocompromised individuals, where replication leads to progressive multifocal leukoencephalopathy (PML). This replication is often accompanied by mutations in the major viral capsid protein, VP1, that make it resistant to certain neutralizing antibodies. The selection for such resistance can be recapitulated in the mouse model. The authors examine the contributions of the T cell response to the appearance of these resistant viruses and examine the biology of an interesting mutant that they discovered. Thus, this study addresses an important question, especially given a resurgence in PML in recent years due to the increasing use of immunomodulatory monoclonal antibodies to treat various diseases. JC polyomavirus causes Progressive Multifocal Leukoencephalopathy (PML) in T-cell immunosuppressed patients. The host-pathogen interaction behind this condition is not fully understood. Using a complicated murine infection model, the authors address the particular role of T cells in this evolution. The model includes commercially available B-cell depleted mice, first infected with MuPyV, then treated with an anti-VP1mAb, and 16 days later with IgG control or anti-CD4 and/or anti-CD8 antibodies to induce T-cell depletion. Viral replication is monitored in the kidney and blood at different times post-infection in T-cell-depleted versus control mice and the emergence of resistant variants is characterized. Finally, the fitness of a resistant variant presenting two mutations is compared in cells, tissues, and mice. The authors show that CD4 T-cells play a critical role in limiting or delaying the emergence of resistant variants. In addition, they highlight that variants that can emerge in absence of T-cell control show fitness loss in kidney but increased neurovirulence. These observations shed some light on JC virus pathogenesis in T-cell suppressed patients. The conclusions are overall supported by the data and confirm the expected critical role of T-cells in controlling viral replication. However, concerns about the presentation will require additional editing and experimental work.

Comments concerning the presentation

1. An overall concern I have is that the first part of the introduction and the entire abstract sound as though JCPyV itself is under investigation. These possibly misleading sections need to be rewritten. There is no need to be concerned that the relevance of their model is going to be questioned.

2. The relevance of these mutations relative to existing JC mutants is not made clear. In fact, two out of three reviewers missed the point that the mouse model involved a different virus from the human one. Clarified prose is needed.

3. It would be good to give some background on JCPyV receptor usage and also if the receptor(s) is (are) conserved between mouse and human and in the different tissues.

4. The last paragraph of the introduction somewhat exaggerates the findings, particularly the sentence beginning "In this study…." It is not clear that Ab-deficient mice are really the equivalent of PML patients since the patients DO produce antibodies against the virus. The model is not being questioned but it should not be oversold.

Comments concerning the figures

Figure 1: Shows that treatment with anti CD4⁺ T cell antibody, but not anti-CD8⁺ T cell antibody or IgG control antibody, reduces control of the virus in a mouse model that contains B cells.

1. The tissue should be clearly indicated in the legend and ideally on the panel.

2. The word 'resurgence' is not appropriate because the data in Figure 2B do not show early viremia.

3. The original comment from all three reviewers was that fluorescent cytometry is needed to confirm T-cell depletion, especially in the case of CD8⁺ T cells. However, it is stated somewhere that these are commercially obtained preparations. If this is the case and there is a validation of the depletions, that needs to be clearly stated. These controls are important, however they are obtained.

Supplemental Figure 1: Throughout the manuscript, the term 'blind spot' is used. This colloquial language implies a structural mechanism of neutralization that is often the case but not always.

1. Instead, more precise evolutionary language (especially for a general audience such as that of eLife) such as "anti-VP1 treatment allows the selection of VP1-antibody resistant viruses.

2. The data themselves show convincingly that the anti-VP1 antibody specifically neutralizes wild-type virus but not the A2Δ295 virus. At this point in the manuscript, the A2D295 virus has not been mentioned so it is not clear why it is S.Figure 1.

3. In (E), the y-axis of 'neutralizing titer' is not clear. Is it 'extent depletion'? Is it the

required serum dilution?

Figure 2 shows convincingly that the combination of anti-CD4⁺ and anti-CD8⁺ antibodies greatly reduce viral control in kidney, spleen, and brain.

1. Fluorescent cell quantitation is not provided to ensure the depletion of the relevant T cell populations,

2. The data in (B) cannot be clearly deciphered and the x-axis should be expanded.

However, the remaining panels in B, dissect the data in the time course. As in Figure 1, these data show that there is little control of the virus in the mice treated with anti-CD8⁺ antibodies.

Figure 3 is the strongest figure in the paper, showing that the sequences of viruses that grow upon anti-VP1 monoclonal antibody treatment of B-cell-deficient mice show a variety of sequence changes. Interestingly, a Δ295 and other amino acid 295 mutations are selected under all conditions and thus, as the authors conclude, confer resistance to the anti-VP1 antibody under all conditions. However, more mutations seem to be required in the absence of T cell depletion, and different spectra of mutations are found under each condition. Panel B shows the location in the VP1-antibody complex of several of the amino acids involved. Panel C shows a similar lack of neutralization for all selected viruses in cultured cells, and Panel D shows that there is different tropism and fitness for the various viruses in the kidney and spleen of infected B-cell-deficient mice in the presence of the anti-VP1 antibody.

1. It is not clear whether the mutations of interest have been installed in the otherwise wild-type virus free of other mutations that might have been selected; if it is not, this should be clearly stated.

Figure 4 shows explicit challenges with the 295 virus.

1. This virus is, presumably, reconstructed so that only this mutation is present. If it is not, it should be, and several points brought up by one of the reviewers is then relevant. Specifically, "The extensive analysis of the A2.D295N, A2.Δ297, and double mutant viruses is impressive. Given that the double mutant has enhanced replication properties, was the entire genome sequenced to confirm nothing has happened in the non-coding control region or the early region. It is not clear from the methods. Another sequencing-related question relates to the paragraph on lines 214-221: how many clones did they analyze, in order to rule out the presence of the D295N/A mutations alone? Could there have been a low-frequency present that was missed (and that might only be detected by NGS)?"

2. The legend does not describe this but the text (line 152) states that the mice used were B-cell deficient. If that is the case, this experiment with the B-cell escape variant is the same experiment done with wild-type virus previously and thus does not enhance the findings of the paper at this point, unless this reviewer is missing the point. Perhaps it was performed in the B-cell-deficient mouse supplemented with the specific antibody?

3. Why are infections done in the hind footpad in the initial protocol and then by IV for the challenge or to test the fitness of the different mutants? The infection route should be clearly described and justified in the result section.

Figure 5 switches up the viruses under study. It shows a very good experiment in which the viruses have now clearly been reconstructed in isolation. However, several inconsistencies in the data likely make it necessary to repeat several aspects of these experiments.

1. 297, D295N, and double mutant viruses were measured by somewhat indirect means such as fold increase over input and fixed time points at two different MOIs, rather than growth curves at a fixed MOI so that any differences in total yield can be clearly assessed.

2. In Panel I, the frequencies of the D295N/Δ297 double mutation in the original mice in which they were isolated, as well as a different combination of D295A/297 from another mouse, are shown. These data do not contribute mechanistically to Figure 5 and should be moved to Panel 3 where the mutations are first discussed. They do, however, provide a basis for discussion of the potential order of mutational events, which should be in the discussion.

3. The double mutant virus seems fitter than the wt. However, the abundance of this variant is very low in Fig5I, how do the authors explain that?

4. Figure 5 D to F and Figure 5 sup 1 are showing very close assays. In addition, the results are not always consistent depending on the readout: in Figure 5 D to F D295N is consistently less fit than A2 while in the supplementary figure, in panel A it looks fitter. Is it a transfected virus in the sup panel A? Why is A2 del 297 not shown in figure 5 panel A? Again, in panel E of Figure 5, this variant is as fit as wt but not in the other assay.

5. Sensitivity to neuraminidase: based on Figure 5 G and 5 sup D, del295 seems insensitive to sialidase pre-treatment, then why in panels E and F of Figure 5 sup it is sensitive?

6. With the way results are presented, it is not clear that panel H represents kidney viral loads after mice infection.

7. In figure 5H, the double mutant is as fit as the wt in the kidney. This is not reproduced in Figure 6.

Figure 6.

1. Why did the authors perform IC inoculation rather than IV or inoculation in the hind footpad as before? There are likely reasons for this, but they need to be explained.

2. The authors show in several figures that the double mutant has an advantage in cells, and this is shown here in the mouse brain. If such variants present increased neurovirulence, the authors should try to understand by what mechanism. Is binding to the receptor changed in the brain? Binding assay in brain tissues should be done in the same way as with kidney tissues.

Supplement to Figure 6 presents further exploration into the comparative phenotypes of the four viruses presented in Figure 5.

Reviewer #1 (Recommendations for the authors):

– In the introduction, it would be good to give some background on JCPyV receptor usage and also if the receptor(s) is(are) conserved between mouse and human and in the different tissues.

– Figures 1 and 2 should be grouped in a single and simplified figure. The experimental setup is the same and only the site of genome replication and the time of analysis differ. In figure 1, the tissue should be clearly indicated in the legend and ideally on the panel.

– Figure 4: Viral loads are not different than in absence of virus challenge in figure 3, the authors should comment.

– Why are infections done in the hind footpad in the initial protocol and then by IV for the challenge or to test the fitness of the different mutants? The infection route should be clearly described and justified in the result section.

– Figure 5 and the supplementary figure 5 are confusing and several points should be clarified in the result section:

o the double mutant virus seems fitter than the wt. However, the abundance of this variant is very low in Fig5I, how do the authors explain that?

o Figure 5 D to F and Figure 5 sup 1 are showing very close assays. In addition, the results are not always consistent depending on the readout: in Figure 5 D to F D295N is consistently less fit than A2 while in the supplementary figure, in panel A it looks fitter. Is it a transfected virus in the sup panel A? Why is A2 del 297 not shown in figure 5 panel A? Again, in panel E of Figure 5, this variant is as fit as wt but not in the other assay.

o Sensitivity to neuraminidase: based on Figure 5 G and 5 sup D, del295 seems unsensitive to sialidase pre-treatment, then why in panels E and F of Figure 5 sup it is sensitive?

o With the way results are presented, it is not clear that panel H represents kidney viral loads after mice infection.

o In figure 5H, the double mutant is as fit as the wt in the kidney. This is not reproduced in Figure 6

– Figure 6: Why did the authors perform IC inoculation rather than IV or inoculation in the hind footpad as before? Also, the authors show in several figures that the double mutant is an advantage, in cells already and here in mice brain. If such variants present increased neurovirulence, the authors should try to understand by what mechanism. Is binding to the receptor changed in the brain? Binding assay in brain tissues should be done in the same way as with kidney tissues.

Reviewer #2 (Recommendations for the Authors):

1. One overall concern I have is that in the first part of the introduction and the entire abstract, they make it sound as if they are studying JCPyV. These sections need to be rewritten because I think they are misleading. I don't think they need to be concerned that the relevance of their model is going to be questioned, at least not by this reviewer.

2. I found the middle of the last paragraph of the introduction to be another example of "exaggeration," particularly the sentence beginning "In this study…." I'm not sure Ab-deficient mice are really the equivalent of PML patients since the patients do produce antibodies against the virus. Again, I am not questioning their model but they should not oversell it.

3. What does "un-neutralized virus" on line 121 mean?

4. Do the authors have any sense of when during the persistent phase of the infection the mutations are arising? Is it before or after reactivation of all-out replication, for example?

5. The extensive analysis of the A2.D295N, A2.Δ297, and double mutant viruses is impressive. Given that the double mutant has enhanced replication properties, I am wondering whether they sequenced the entire genome to confirm nothing has happened in the non-coding control region or the early region. It is not clear from the methods. Another sequencing-related question relates to the paragraph on lines 214-221: how many clones did they analyze, in order to rule out the presence of the D295N/A mutations alone? Could there have been a low-frequency present that was missed (and that might only be detected by NGS)?

Reviewer #3 (Recommendations for the authors):

The narrative of this manuscript is difficult to follow although it contains much interesting biology. It begins by showing that T-cell control of JC polyomavirus in mice is important in the presence and absence of a B-cell response. Then, the authors show that the relevant T cells for viral control are CD4⁺, although controls for specific cell depletion are not provided. These data are presented in Figures 1, S.Figures 1, Figure 2, and one panel of Figure 3. The antibody-escape mutations in VP1 described in Figure 3 are very interesting and form the heart of this paper.

In Figure 5, four viruses are defined genetically and their properties compared in antibody-binding studies, antibody-neutralization studies, and in B-cell depleted mice. The viruses are A2 wild-type, Δ297, D295N, and double mutant. The authors show that the D295N virus (as does the Δ295 virus they have been studying up until this point) confers resistance to a particular neutralizing antibody. However, the Δ297 mutation is required to increase its fitness. Thus, Figure 5 is where all the pieces are together to ask about the properties of these viruses in the presence and absence of the neutralizing antibody and CD4⁺ T cells.

By stopping where they do in the narrative, the message of the manuscript is that compensatory mutations are often needed to increase the fitness of antibody- or drug-resistant mutations. This has been shown many times, but, as such, remains of interest to a specialized audience. ,

The writing is so confusing that suggestions, below, are to highlight the findings and make suggestions to increase their clarity and persuasiveness.

Figure 1: Shows that treatment with anti- depletion of CD4⁺ T cell antibody, but not anti-CD8⁺ T cell antibody or IgG control antibody, reduces control of the virus in mouse model that contains B cells.

1. The word 'resurgence' is not appropriate because data in Figure 2B do not show early viremia.

2. Fluorescent cytometry is needed to confirm T-cell depletion, especially in the case of CD8⁺ T cells.

Supplemental Figure 1: Throughout the manuscript, the term 'blind spot' is used. This colloquial language implies a structural mechanism of neutralization that is often the case but not always.

1. Instead, more precise evolutionary language (especially for a general audience such as that of eLife) such as "anti-VP1 treatment allows the selection of VP1-antibody resistant viruses.

2. The data themselves show convincingly that the anti-VP1 antibody specifically neutralizes wild-type virus but not the A2Δ295 virus. At this point in the manuscript, the A2D295 virus has not been mentioned so it is not clear why it is S.Figure 1.

3. In (E), the y-axis of 'neutralizing titer' is not clear. Is it 'extent depletion'? Is it the

required serum dilution?

Figure 2: (A) Shows a time course of viremia in B cell-deficient mice after treatment with IgG, anti-CD4⁺ , anti-CD8⁺, and a combination of anti-CD4⁺ and anti-CD8⁺ T cell antibodies. The finding that CD4⁺ , but not CD8⁺ T cells can control viremia in the absence of B cells is interesting but not pursued mechanistically.

1. Fluorescent cell quantitation is provided to ensure the depletion of the relevant T cell populations,

2. The data in (B) cannot be clearly deciphered and the x-axis should be expanded.

However, the remaining panels in B, dissect the data in the time course. As in Figure 1, these data show that there is little control of the virus in the mice treated with anti-CD8⁺ antibodies. The remaining panels show convincingly that the combination of anti-CD4⁺ and anti-CD8⁺ antibodies greatly reduce viral control in kidney, spleen, and brain.

Figure 3 is the strongest figure in the paper, showing that the sequences of viruses that grow upon anti-VP1 monoclonal antibody treatment of B-cell-deficient mice show a variety of sequence changes. Interestingly, a Δ295 and other amino acid 295 mutations are selected under all conditions and thus, as the authors conclude, confer resistance to the anti-VP1 antibody under all conditions. However, more mutations seem to be required in the absence of T cell depletion, and different spectra of mutations are found under each condition. Panel B shows the location in the VP1-antibody complex of several of the amino acids involved. Panel C shows a similar lack of neutralization for all selected viruses in cultured cells, and Panel D shows that there is different tropism and fitness for the various viruses in the kidney and spleen of infected B-cell-deficient mice in the presence of the anti-VP1 antibody.

1. It is not clear whether the mutations of interest have been installed in the otherwise wild-type virus free of other mutations that might have been selected; if it is not, this should be clearly stated.

Figure 4 shows an explicit challenge with the Δ295 virus, presumably specifically reconstructed so that only this mutation is present. Panel (B) shows viral growth following T cell depletion. The legend does not describe this but the text (line 152) that the mice used were B-cell deficient.

1. Therefore this experiment with the B-cell escape variant is the same experiment done with wild-type virus previously and thus does not enhance the findings of the paper at this point.

Figure 5 switches up the viruses under study. It shows a very good experiment in which the viruses have now clearly been reconstructed in isolation. A substitution at residue 295 (D295N), rather than the deletion that has been previously investigated, is reconstructed individually and together with a different single-amino acid deletion, Δ297, which was found in one of the natural isolates presented in Figure 3. They find (A) that only the Δ297 mutation confers neutralizing antibody resistance. This can be explained by the observed failure of either the Δ297 mutant virus or the double mutant virus to bind to the neutralizing antibody. In the other figures, the relative fitness of the wild-type A2, Δ297, D295N, and double mutant viruses were measured by somewhat indirect means such as fold increase over input and fixed time points at two different MOIs, rather than growth curves at a fixed MOI. However, the case is made convincingly that the double mutant virus grows better under all conditions in cultured cells. The authors show that these data can clearly be modified in animals, because, in the mouse kidney (H), a double mutant is present at equivalent abundance to wild-type.

1. In Panel I, the frequencies of the D295N/Δ297 double mutation in the original mice in which they were isolated, as well as a different combination of D295A/Δ297 from another mouse, are shown These data do not contribute mechanistically to Figure 5 and should be moved to Panel 3 where the mutations are first discussed. They do, however, provide a basis for discussion of the potential order of mutational events, which should be in the discussion.

Supplement to Figure 6 presents further exploration into the comparative phenotypes of the four viruses presented in Figure 5.

1. Could the authors discuss why the double-mutant virus generates the most virus but shows considerably less CPE than the D295N mutation?

Figure 6 shows pathogenesis experiments that compare wild-type and the D295N/Δ297 double mutant in B cell-deficient mice in the absence of T cell depletion. The double mutant shows similar viremia but increased growth in brain tissues. It is part of the interesting biology of this paper, that the mutations were isolated for their antibody-escape properties, but show differences in mice with no B cells. Thus, this provides the perfect jumping-off point to compare the properties of the viruses in the presence and absence of neutralizaing antiboy and T cells.

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Comments concerning the presentation

1. An overall concern I have is that the first part of the introduction and the entire abstract sound as though JCPyV itself is under investigation. These possibly misleading sections need to be rewritten. There is no need to be concerned that the relevance of their model is going to be questioned.

The abstract has been modified and the introduction has been reworked to clarify that MuPyV is used for this study.

2. The relevance of these mutations relative to existing JC mutants is not made clear. In fact, two out of three reviewers missed the point that the mouse model involved a different virus from the human one. Clarified prose is needed.

We have extensively revised the abstract and added sections to the Introduction to clarify that this study exclusively uses MuPyV, an experimental setup necessitated by the fact the JCPyV only replicates in humans. Further discussion of the nature of the mutations found in this study in comparison to those found in PML patients has been added in the Discussion (Lines 401-407), which includes the following text “The mutant viruses we identified all carried at least one mutation in the HI loop, which is also the most common site of JCPyV VP1 mutations in PML patients (Gorelik et al., 2011; Reid et al., 2011). The epitope of the VP1 mAb has significant contribution from the HI loop and competes with a large portion of the endogenous antibody response generated by MuPyV-infected mice (Lauver et al., 2020). The dominant targets of antibodies in JCPyV-infected individuals have not been determined, but given the frequency of JCPyV mutations seen in the HI loop this region may be a common target of neutralizing Ab across species.”

3. It would be good to give some background on JCPyV receptor usage and also if the receptor(s) is (are) conserved between mouse and human and in the different tissues.

A paragraph on receptor usage by JCPyV and MuPyV has been added to the Introduction (Lines 78-84). We also included a discussion of the localization of JCPyV and MuPyV binding to the distal tubules in the kidney (Lines 308-312).

4. The last paragraph of the introduction somewhat exaggerates the findings, particularly the sentence beginning "In this study…." It is not clear that Ab-deficient mice are really the equivalent of PML patients since the patients DO produce antibodies against the virus. The model is not being questioned but it should not be oversold.

In retrospect, we can appreciate how this paragraph could lead to this misperception. There is evidence (Ray et al., 2015; discussed in Lines 72-74) that PML patients have select deficiencies in their JCPyV VP1 Ab repertoire for VP1 mutations, a situation implying that a narrow VP1 repertoire sets the stage for selecting outgrowth of such variant viruses. The last paragraph of the introduction has been rewritten to clarify the experimental setup – B cell-deficient mice given an MuPyV-specific monoclonal VP1 antibody – as a model to investigate if a monospecific VP1 response will drive selection of VP1 variants (Line 88-93).

Comments concerning the figures

Figure 1: Shows that treatment with anti CD4⁺ T cell antibody, but not anti-CD8⁺ T cell antibody or IgG control antibody, reduces control of the virus in a mouse model that contains B cells.

1. The tissue should be clearly indicated in the legend and ideally on the panel.

The tissue is now identified as kidney in both the legend and the figure.

2. The word 'resurgence' is not appropriate because the data in Figure 2B do not show early viremia.

We have removed this term throughout the manuscript and replaced it with “increased or elevated replication.”

3. The original comment from all three reviewers was that fluorescent cytometry is needed to confirm T-cell depletion, especially in the case of CD8⁺ T cells. However, it is stated somewhere that these are commercially obtained preparations. If this is the case and there is a validation of the depletions, that needs to be clearly stated. These controls are important, however they are obtained.

We apologize for not including this data in the original submission. Flow cytometry data confirming the depletions of CD4 and CD8 T cells in Figures 1, 2, and 4 have been added as Figure 1—figure supplement 3, Figure 2—figure supplement 1, and Figure 4—figure supplement 1.

Supplemental Figure 1: Throughout the manuscript, the term 'blind spot' is used. This colloquial language implies a structural mechanism of neutralization that is often the case but not always.

1. Instead, more precise evolutionary language (especially for a general audience such as that of eLife) such as "anti-VP1 treatment allows the selection of VP1-antibody resistant viruses.

The term “blind spot” was used repeatedly to describe the inability of sera IgG from PML patients to neutralize JCPyV VP1 mutant viruses (Ray et al., 2015). We can see the reviewer’s concern with this colloquialism, and have replaced with “an inability to neutralize.”

2. The data themselves show convincingly that the anti-VP1 antibody specifically neutralizes wild-type virus but not the A2Δ295 virus. At this point in the manuscript, the A2D295 virus has not been mentioned so it is not clear why it is S.Figure 1.

Further clarification has been added to the text (Lines 110-114) regarding the use of the A2.Δ295 virus, which we previously published (Lauver et al., 2020) as a mutant virus that is resistant to neutralization by the VP1 mAb used in this study.

3. In (E), the y-axis of 'neutralizing titer' is not clear. Is it 'extent depletion'? Is it the

required serum dilution?

The y-axis and legend for supplemental-figure 1E have been updated to indicate the neutralization titer is the Log₁₀ IC50 of the neutralization curves in C and D.

Figure 2 shows convincingly that the combination of anti-CD4⁺ and anti-CD8⁺ antibodies greatly reduce viral control in kidney, spleen, and brain.

1. Fluorescent cell quantitation is not provided to ensure the depletion of the relevant T cell populations,

Flow cytometry data confirming T cell depletion has been added as Figure 2—figure supplement 1 (see response to comment #3 concerning the figures above).

2. The data in (B) cannot be clearly deciphered and the x-axis should be expanded.

However, the remaining panels in B, dissect the data in the time course. As in Figure 1, these data show that there is little control of the virus in the mice treated with anti-CD8⁺ antibodies.

The x-axis has been expanded for Figure 2B as requested to more clearly show the time course for detection of viremia in individual mice for each of the conditions depicted in the key to the far right of this panel. This is the raw data used to generate the middle panel showing virus levels at the peak of viremia in each mouse and the right panel collating the time course of viremia per treatment group. The interpretation of this data, as the reviewer correctly points out, is that CD4 T cells rather than CD8 T cells are primarily responsible for controlling MuPyV viremia.

Figure 3 is the strongest figure in the paper, showing that the sequences of viruses that grow upon anti-VP1 monoclonal antibody treatment of B-cell-deficient mice show a variety of sequence changes. Interestingly, a Δ295 and other amino acid 295 mutations are selected under all conditions and thus, as the authors conclude, confer resistance to the anti-VP1 antibody under all conditions. However, more mutations seem to be required in the absence of T cell depletion, and different spectra of mutations are found under each condition. Panel B shows the location in the VP1-antibody complex of several of the amino acids involved. Panel C shows a similar lack of neutralization for all selected viruses in cultured cells, and Panel D shows that there is different tropism and fitness for the various viruses in the kidney and spleen of infected B-cell-deficient mice in the presence of the anti-VP1 antibody.

1. It is not clear whether the mutations of interest have been installed in the otherwise wild-type virus free of other mutations that might have been selected; if it is not, this should be clearly stated.

Clarification has been added to the Results to indicate that all the mutant viruses used in this study were generated de novo by inserting the mutations identified in viremic mice into the WT MuPyV genome. In particular, the lines “To exclude possible effects of other mutations in the viral genome, we introduced several of these single and dual VP1 mutations into WT MuPyV using site-directed mutagenesis” and “We generated viruses individually carrying the D295N or Δ297 mutations in the WT A2 genome” have been added to the Results (Lines 156-158 and 189-190).

Figure 4 shows explicit challenges with the 295 virus.

1. This virus is, presumably, reconstructed so that only this mutation is present. If it is not, it should be, and several points brought up by one of the reviewers is then relevant. Specifically, "The extensive analysis of the A2.D295N, A2.Δ297, and double mutant viruses is impressive. Given that the double mutant has enhanced replication properties, was the entire genome sequenced to confirm nothing has happened in the non-coding control region or the early region. It is not clear from the methods. Another sequencing-related question relates to the paragraph on lines 214-221: how many clones did they analyze, in order to rule out the presence of the D295N/A mutations alone? Could there have been a low-frequency present that was missed (and that might only be detected by NGS)?"

All VP1 mutant viruses used in the study were constructed by inserting mutations in the VP1 gene found in viremic mice into a WT genome. As noted by this reviewer, this approach was essential to exclude the possibility that changes in the readout assays could be influenced by potential mutations in other parts of the genome, such as the non-coding control region. This has been clarified in the Results as “To exclude possible effects of other mutations in the viral genome, we introduced several of these single and dual VP1 mutations into WT MuPyV using site-directed mutagenesis” (Lines 156-158).

The text and legend for Figure 5I have been updated to include that 50 VP1 clones from each mouse were sequenced (Line 254). We added the sentence “Although more sensitive sequencing methods may detect VP1 sequences with D295N/A mutations without Δ297, the high frequency of the Δ297 mutation strongly suggests this was the initial mutation” (Lines 258-260) to acknowledge the possibility that single mutations at D295 could be detected with a more sensitive sequencing technique. However, we feel that the abundance of single Δ297 mutations in the absence of clones containing an isolated D295N/A mutation strongly supports the likelihood that this was the initial mutation.

2. The legend does not describe this but the text (line 152) states that the mice used were B-cell deficient. If that is the case, this experiment with the B-cell escape variant is the same experiment done with wild-type virus previously and thus does not enhance the findings of the paper at this point, unless this reviewer is missing the point. Perhaps it was performed in the B-cell-deficient mouse supplemented with the specific antibody?

We expanded the text to state that these are B cell-deficient mice given the VP1 mAb. Further justification for the experiment has been added to the Results, in particular the text “Mice were challenged i.v. with a lower titer inoculum of the A2.Δ295 mutant virus to mimic the development of viremia with an Ab escape mutant virus. This experimental setup allowed us to separate the function of T cells in preventing the generation of VP1 mutations from the ability of T cells to control the outgrowth of a VP1 mutant virus” (Lines 175-178).

3. Why are infections done in the hind footpad in the initial protocol and then by IV for the challenge or to test the fitness of the different mutants? The infection route should be clearly described and justified in the result section.

We have added an explanation to the Results why IV infection was used for the subsequent infections comparing the mutant viruses and for the challenge experiments. In the section detailing the comparison of mutant viruses, the text “The viruses were injected i.v. to examine tropism when virus is spreading in the blood, which is the condition under which these mutations were identified, and to avoid the possibility of the viruses having impaired spread from the site of s.c. inoculation” has been added (Lines 160-163). In the section containing the challenge experiment, the sentence “Mice were challenged i.v. with a lower titer inoculum of the A2.Δ295 mutant virus to mimic the development of viremia with an Ab escape mutant virus” has been added (Lines 175-176).

Figure 5 switches up the viruses under study. It shows a very good experiment in which the viruses have now clearly been reconstructed in isolation. However, several inconsistencies in the data likely make it necessary to repeat several aspects of these experiments.

1. 297, D295N, and double mutant viruses were measured by somewhat indirect means such as fold increase over input and fixed time points at two different MOIs, rather than growth curves at a fixed MOI so that any differences in total yield can be clearly assessed.

To address this concern, we expanded the time points in Figure 5F to show virus spread over time following a low MOI infection, where the double mutant virus was first detected two days before either single mutant virus and even the WT virus.

2. In Panel I, the frequencies of the D295N/Δ297 double mutation in the original mice in which they were isolated, as well as a different combination of D295A/297 from another mouse, are shown. These data do not contribute mechanistically to Figure 5 and should be moved to Panel 3 where the mutations are first discussed. They do, however, provide a basis for discussion of the potential order of mutational events, which should be in the discussion.

We appreciate the reviewer’s comment, however we feel that the results in Figure 5I are thematically distinct from the results in Figure 3. Figure 5I shows the order of the emergence of the double mutations, which logically follows once data showing differences in the functions of each of these mutations have been described in the preceding panels of Figure 5.

3. The double mutant virus seems fitter than the wt. However, the abundance of this variant is very low in Fig5I, how do the authors explain that?

We recognize that the low frequency of double mutant sequences in the samples seems to be at odds with the increased spread and replication observed in vivo. We have added further clarification to this section to address this in Lines 252-253 “Each of these mice rapidly became morbid after the detection of viremia necessitating their euthanasia within 30 days of the emergence of viremia” and Lines 251-252 “The rapid morbidity associated with the emergence of the mutation at D295 and subsequent viremia likely limited the accumulation of these double mutant viruses.”

4. Figure 5 D to F and Figure 5 sup 1 are showing very close assays. In addition, the results are not always consistent depending on the readout: in Figure 5 D to F D295N is consistently less fit than A2 while in the supplementary figure, in panel A it looks fitter. Is it a transfected virus in the sup panel A? Why is A2 del 297 not shown in figure 5 panel A? Again, in panel E of Figure 5, this variant is as fit as wt but not in the other assay.

We can now appreciate how the data in these figures would appear to be discrepant. This apparent difference is due to the timing of the assays. The assays in Figure 5D to F are short-term experiments, which demonstrate significant delays in virus production, infection, and spread by the D295N virus compared to WT virus. Figure 5-supplement 1 shows an image of a single plaque at 6 days post infection, at which point the D295N virus has been able to spread sufficiently to form plaques. The new time course data in Figure 5F shows that the D295N virus has significantly delayed ability to spread than WT virus, but does eventually spread.

5. Sensitivity to neuraminidase: based on Figure 5 G and 5 sup D, del295 seems insensitive to sialidase pre-treatment, then why in panels E and F of Figure 5 sup it is sensitive?

We now recognize that this was not fully explained in the text, and appreciate the reviewer pointing this out. Figure 5G and 5 Sup 1D show neuraminidase-insensitive binding by the D295N mutant. However, Figure 5 Sup 1E-F show that infection by the D295N mutant remains dependent on sialic acid, which indicates that the neuraminidase-insensitive binding is non-productive. Further discussion of the difference in neuraminidase sensitivity for binding vs infection by A2.D295N has been added, including “The dependence on sialic acid for infection, but not for binding, by A2.D295N indicates that by itself the D295N mutation mediates binding to a non-sialyated, non-productive receptor” (Lines 226-228).

6. With the way results are presented, it is not clear that panel H represents kidney viral loads after mice infection.

To further clarify that 5H is kidney viral loads, we added the sentence “We then assessed virus levels in the kidneys of mice infected i.v. with the mutant viruses” (Lines 239-240) to this part of the Results.

7. In figure 5H, the double mutant is as fit as the wt in the kidney. This is not reproduced in Figure 6.

Figure 5H shows similar levels of mRNA during acute (4 dpi) infection of the kidney, whereas Figure 6A-D shows pathology induced by chronic kidney infection. The similar levels of infection early but reduced infection/pathology during chronic infection by the double mutant suggest a defect in viral persistence within the kidney. The differences seen between acute and chronic kidney infection with the double mutant is now addressed in the Discussion, with “in vivo, the double mutant virus showed similar acute kidney infection to WT virus but decreased kidney infection/pathology during chronic infection (Figures 5H and 6A-D). This impairment of kidney pathology during chronic infection suggests a defect in persistence within the kidney by the D295N/Δ297 mutant virus” (Lines 385-389).

Figure 6.

1. Why did the authors perform IC inoculation rather than IV or inoculation in the hind footpad as before? There are likely reasons for this, but they need to be explained.

IC inoculation is the conventional injection route for inducing MuPyV brain infection, which is not seen with s.c. or i.v. routes of inoculation. The use of IC inoculation to investigate MuPyV brain infection is now explained in the Results. “we infected WT mice intracranially (i.c.), a route of inoculation which mediates efficient MuPyV infection of the brain (Lauver et al., 2020; Mockus et al., 2020)” (Lines 272274).

2. The authors show in several figures that the double mutant has an advantage in cells, and this is shown here in the mouse brain. If such variants present increased neurovirulence, the authors should try to understand by what mechanism. Is binding to the receptor changed in the brain? Binding assay in brain tissues should be done in the same way as with kidney tissues.

Understanding the mechanism responsible for the neurovirulence of the double mutant virus is of keen interest to us as well. To this end, we found that transfection of mutant viral DNA yielded increased virus production compared to WT DNA (Figure 5—figure supplement 1B). Although this does not rule out superior/altered virus binding by the mutant, it does show that the mutant virus more efficiently completes its replication cycle at a post entry step. We speculated in the Discussion about the mechanism by which these mutations may facilitate viral assembly; however, demonstrating this would require extensive structural analysis best done in a separate study. We have invested considerable effort to carry out virus binding assays on brain sections. However, this has turned out to be a technically difficult. While MuPyV stains kidney sections well, staining of brain tissue sections is weak and nonspecific. We have attempted to improve virus binding in brain section by antigen retrieval methods, but these have proven unsuccessful. We hope the reviewer understands our inability to compare binding by the viruses in the brain due to this technical impasse.

Supplement to Figure 6 presents further exploration into the comparative phenotypes of the four viruses presented in Figure 5.

The reviewer is correct to point out that the plaque size for the double mutant is smaller. We have added a new panel to this figure supplement quantifying plaque area. Given the data indicating the double mutant produces more virus, the difference in plaque area is most likely explained by the double mutant having an increased affinity for binding host cell receptors, limiting its spread. This phenotype has previously been observed with other MuPyV mutants possessing altered receptor binding (Bauer et al. 1999). This is now noted in the Results (Lines 199-201).

Author details

1. Matthew D Lauver

   Department of Microbiology and Immunology, Pennsylvania State University, Hershey, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7001-9730

2. Ge Jin

   Department of Microbiology and Immunology, Pennsylvania State University, Hershey, United States

   Contribution

   Investigation

   Competing interests

   No competing interests declared

3. Katelyn N Ayers

   Department of Microbiology and Immunology, Pennsylvania State University, Hershey, United States

   Contribution

   Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-6156-8685

4. Sarah N Carey

   Department of Microbiology and Immunology, Pennsylvania State University, Hershey, United States

   Contribution

   Investigation

   Competing interests

   No competing interests declared

5. Charles S Specht

   Department of Pathology and Laboratory Medicine, Penn State Milton S. Hershey Medical Center, Hershey, United States

   Contribution

   Investigation

   Competing interests

   No competing interests declared

6. Catherine S Abendroth

   Department of Pathology and Laboratory Medicine, Penn State Milton S. Hershey Medical Center, Hershey, United States

   Contribution

   Investigation

   Competing interests

   No competing interests declared

7. Aron E Lukacher

   Department of Microbiology and Immunology, Pennsylvania State University, Hershey, United States

   Contribution

   Conceptualization, Supervision, Funding acquisition, Visualization, Project administration, Writing - review and editing

   For correspondence

   alukacher@pennstatehealth.psu.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7969-2841

• 250

  Page views

• 46

  Downloads

• 0

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.