Graphical abstract

   graphic file with name fx1.jpg
   [47]Open in a new tab

Highlights

     * •
       Miwi2 upregulates select mitochondrial genes in M2MC cells during
       IAV infection
     * •
       Miwi2 reduces intracellular ROS and ADP/ATP ratios in multiciliated
       cells
     * •
       NuMT derived small RNAs are decreased in Miwi2 expressing cells
     * •
       Miwi2 deficient mice experience decreased viral burden from IAV
     __________________________________________________________________

   Biochemistry; Cell biology; Functional aspects of cell biology

Introduction

   Respiratory infections are among the leading causes of morbidity and
   mortality, with influenza alone causing approximately 35–65 million
   infections and 25,000–72,000 deaths in the US during the 2023–2024
   season.[48]^1 Despite the availability of vaccines and antiviral drugs,
   their effectiveness varies, especially among high-risk
   patients[49]^2^,[50]^3; thus underscoring a critical need to further
   our understanding of the host response during an influenza infection to
   develop more effective therapeutic strategies.

   Host defense against respiratory viruses is reliant on an orchestrated
   release of cytokine and chemokine signals originating from infected
   cells of the pulmonary epithelium, which line the respiratory tract
   from the trachea to alveoli.[51]^4^,[52]^5^,[53]^6 These signals
   initiate and modulate host immune responses; recruit neutrophils,
   T cells, and other immune cells to the site of infection; and activate
   macrophages.[54]^7^,[55]^8^,[56]^9^,[57]^10 The influenza A virus (IAV)
   binds to sialic acid glycoprotein receptors found on the surface of
   airway epithelial cells, which include multiple cell types such as type
   I and II pneumocytes, basal cells, secretory club cells, and
   multiciliated cells.[58]^9^,[59]^11^,[60]^12^,[61]^13 Multiciliated
   cells, which are most abundant in the conducting airways, primarily
   function to maintain the mucociliary escalator over the luminal
   surface.[62]^14^,[63]^15^,[64]^16^,[65]^17 Like other epithelial cell
   types, multiciliated cells are heterogeneous, as demarcated by
   differential gene
   expression.[66]^18^,[67]^19^,[68]^20^,[69]^21^,[70]^22^,[71]^23 This
   heterogeneity confers multiciliated cells with a diverse range of
   functions that extend beyond just ciliary beating, enabling them to
   potentially play broader and more versatile roles in immune
   responses.[72]^18^,[73]^19^,[74]^21

   Previously, we identified the MIWI2 multiciliated (M2MC) cell, a rare
   subpopulation of multiciliated cells that exclusively express MIWI2,
   the mouse ortholog of the human P element-induced wimpy testes like 4
   (PIWIL4) argonaute family protein. Similarly, we found that PIWIL4
   distinguishes a rare subpopulation of multiciliated cells in the human
   airway.[75]^23 Most abundantly expressed in male germline stem cells,
   MIWI2 forms a repressive complex with PIWI-interacting RNAs (piRNAs) to
   suppress retrotransposons (RTs) that are abundant transposable elements
   in mammalian
   genomes.[76]^24^,[77]^25^,[78]^26^,[79]^27^,[80]^28^,[81]^29 This
   complex inhibits many major classes of RTs such as the long terminal
   repeat (LTR) and non-LTR RTs.[82]^30^,[83]^31^,[84]^32 Through
   complementary base-pairing with piRNAs, the RT RNA is cleaved by
   another PIWI protein, MILI, which can then be processed into additional
   piRNAs through ping pong
   amplification.[85]^25^,[86]^33^,[87]^34^,[88]^35 Effector piRNAs loaded
   onto MIWI2 translocate to the nucleus, identifies nascent RT RNA, and
   recruits silencing factors for methylation of the promoter
   regions.[89]^32^,[90]^36 This suppressive function of MIWI2 is
   essential in the male germline, ensuring stability of the genome. In
   cases of MIWI2 deficiency or dysfunction, meiotic arrest is induced in
   spermatogonia leading to male infertility.[91]^37

   Outside of the germline, the role of PIWI-piRNA pathways has remained
   largely unexplored although some reports hint at regulatory functions
   in somatic cells. For example, PIWI-piRNA interactions have been
   examined in the pathogenesis of Alzheimer’s disease, multiple
   sclerosis, cardiovascular disease, and
   cancer.[92]^38^,[93]^39^,[94]^40^,[95]^41^,[96]^42 Additionally,
   PIWI-piRNA pathways have been investigated in viral infections such as
   in Respiratory Syncytial Virus (RSV), Human immunodeficiency virus
   (HIV), SARS-CoV-2, and Human Papillomavirus
   (HPV).[97]^43^,[98]^44^,[99]^45^,[100]^46^,[101]^47 In the context of
   respiratory infections, these mechanisms could influence how host cells
   respond to viral invasion, potentially impacting the severity and
   progression of disease. By investigating MIWI2’s role in viral
   infection, we may gain further insights into broader mechanisms of
   host-pathogen interactions and reveal potential therapeutic targets for
   not only influenza but for other respiratory viruses as well.

   In this report, we elucidate the role of Miwi2 in lung airway
   multiciliated cells on influenza pathogenesis and the host response to
   infection. To further investigate how Miwi2 affects the host defense,
   we used a Miwi2-tdTomato knock-in reporter mouse model to isolate M2MC
   and nonMIWI2 multiciliated (nonM2MC) cells for RNA sequencing analysis
   of both long and small RNAs. The transcriptional profiles provided
   evidence of Miwi2-dependent expression of nuclear mitochondrial (NuMT)
   derived small RNAs and select mitochondrial mRNAs. Moreover, we
   observed a Miwi2-dependent modulation of mitochondrial transcripts that
   was associated with increased reactive oxygen species (ROS) levels and
   ADP/ATP ratios in multiciliated cells. Lastly, mice deficient in Miwi2
   exhibited improved outcomes when infected with IAV suggesting Miwi2 as
   a key host susceptibility factor. Our work advances the understanding
   of Miwi2 in somatic cells and highlights its pivotal role to modulate a
   class of NuMT-derived small RNAs and mitochondrial ROS function during
   infection.

Results

Miwi2 heterozygous and deficient experimental models

   A Miwi2-tdTomato knock-in reporter mouse model, previously described,
   was used to generate Miwi2 heterozygous (Miwi2^+/tom) and deficient
   (Miwi2^tom/tom) mice ([102]Figure 1A).[103]^23^,[104]^48 Combining
   established surface markers (CD45^−EpCAM^+CD24^hi) with tdTomato
   expression enabled the separation of M2MC and nonM2MC cells via flow
   cytometry, including cells from Miwi2^tom/tom mice
   ([105]Figure 1B).[106]^23^,[107]^49 This suggests that while M2MC cells
   are identified by the expression of Miwi2, Miwi2 itself is not
   necessary for the development and presence of this subset.

Figure 1.

   [108]Figure 1
   [109]Open in a new tab

   Schematic of Miwi2-tdTomato mice and experimental strategy

   (A) A tdTomato expression cassette is inserted into the first exon of
   Miwi2, inactivating the allele. Single allele knock-in (Miwi2^+/tom)
   allows male mice to still be fertile, consistent with Miwi2
   haplosufficiency. Double allele knock-in (Miwi2^tom/tom) induces
   infertility in males, consistent with lack of Miwi2 activity. p(A):
   poly(A) tail.

   (B) Using multiciliated cell markers (CD45^−EpCAM^+CD24^hi), the M2MC
   subset can be identified via flow cytometry, even in Miwi2 deficient
   mice.

   (C) Miwi2^+/tom and Miwi2^tom/tom were intratracheally instilled with
   saline or PR8 3 days post infection into the left lobes (n = 4, 5–6
   left lobes collected per n). Single cell suspensions were generated
   through an epithelial enrichment protocol. M2MC and nonM2MC cells were
   sorted based on multiciliated cell markers and tdTomato expression.
   Small and long RNAs were sequenced for each sample.

   (D) For small RNA sequencing, length distribution plots were used to
   filter piRNA-like small RNAs that were 24–35 nucleotides in length.

   (E) Flow cytometry quantification of tdTomato^+ cells as a percentage
   of total multiciliated cells (two-way ANOVA followed by Sidak’s test).
   TdTomato mean fluorescence intensity (MFI) was also measured (n = 5).
   Total multiciliated cells (CD24^hi) were quantified as a percentage of
   total epithelial cells (EpCAM^+CD45^−). Error bars represent mean ±
   standard deviation. ∗p < 0.05.

   To examine transcriptomic changes in both small and long RNA, M2MC and
   nonM2MC cells were isolated from Miwi2^+/tom and Miwi2^tom/tom mice
   intratracheally instilled with saline or mouse-adapted influenza
   A/Puerto Rico/8/34 (PR8) 3 days post infection (dpi)
   ([110]Figure 1C).[111]^50 Small and total long RNA libraries were then
   constructed for each sample. Among all samples that passed quality
   control, miRNAs were the predominant class of small RNAs
   ([112]Figures 1D and [113]S1). Read length distributions were filtered
   between 24 and 35 nucleotides to analyze the piRNA-like small RNAs in
   the multi-species small RNA genomics (MSRG) pipeline.[114]^51^,[115]^52
   MicroRNAs (miRNAs) were separately analyzed through the COMPSRA
   pipeline.[116]^53 We observed more M2MC cells in Miwi2^tom/tom mice
   compared to Miwi2^+/tom mice during PR8 infection without an increase
   in tdTomato detection or significant changes in the overall amount of
   multiciliated cells ([117]Figure 1E).

Somatic retrotransposon expression induced by IAV, but is independent of
Miwi2

   During spermatogenesis, MIWI2 regulates the expression of two major
   classes of retrotransposons: long-terminal repeats (LTRs) and non-LTRs.
   Endogenous retroviruses (ERVs) are a commonly expressed family of LTRs
   that share a genetic structure and mechanism similar to exogenous
   retroviruses.[118]^54^,[119]^55 The most common mammalian non-LTR
   family is LINE-1 that contains 2 open reading frame proteins that
   function in the transcription and integration of the RT into the
   genome.[120]^30^,[121]^56^,[122]^57 In MIWI2 deficient testes, these
   RTs become unregulated, induce meiotic arrest, and ultimately
   infertility.[123]^58^,[124]^59 While MIWI2 represses these sites,
   recent studies suggest that viral infection may derepress them and
   stimulate antiviral immunity through proximal cis-acting enhancement of
   nearby antiviral genes or through activation of Pathogen-associated
   molecular pattern (PAMP) receptors inducing interferon (IFN)
   activation.[125]^60^,[126]^61^,[127]^62^,[128]^63^,[129]^64 Therefore,
   expression of RTs was examined in the long and small RNA sequencing
   dataset derived from the sorted multiciliated cells.

   When aligning total long and piRNA-like small RNAs to RTs of the mouse
   genome using the MSRG pipeline,[130]^52 no Miwi2-dependent changes were
   observed in the long or small RNAs of the highest expressed RTs
   ([131]Figure 2A). Additional analysis of detected RTs also showed no
   genotype or cell type dependent difference ([132]Figure S2A). Notably,
   a PR8-induced upregulation in several RT subfamilies were observed
   across all groups, predominantly ERV subfamilies ([133]Figure 2B). In
   particular, more RT subfamilies were increased in both M2MC and nonM2MC
   cells of Miwi2^tom/tom mice, but only during PR8 infection
   ([134]Figure 2C). Protein expression of the open reading frame 1
   protein (ORF1p) derived from LINE-1 was analyzed in whole lung and
   sorted epithelial cells of saline and PR8 treated wildtype (Miwi2^+/+)
   and Miwi2^tom/tom mice by western blot analyses.[135]^65^,[136]^66 No
   changes in ORF1p were observed in the whole lung and it was
   undetectable in sorted epithelial cells ([137]Figures S2B, [138]S3A,
   and S3B). Taken together, these data indicate that IAV induces RT RNA
   expression in all multiciliated cells. However, it does not appear
   Miwi2 alone impacts the steady-state expression of RT long and small
   RNAs in multiciliated cells.

Figure 2.

   [139]Figure 2
   [140]Open in a new tab

   Retrotransposon expression is independent of Miwi2 but is upregulated
   during PR8 infection

   (A) Long and small RNAs of the top 3 expressed RT subfamilies
   (Student’s t test).

   (B) Comparisons of matched cell types and genotypes between saline and
   PR8 treatments showed a set of mouse RTs whose expression was
   repeatedly enhanced, as marked with red labels for the RT type
   (p < 0.05 by Student’s t test).

   (C) Number of significantly upregulated RTs induced by PR8 infection
   for each represented group. Error bars represent mean ± standard
   deviation. ∗p < 0.05, ∗∗p < 0.005.

Influenza A viral-derived small RNAs are expressed independent of Miwi2

   In insects, somatic cells might utilize the PIWI-piRNA pathway as a
   host defense mechanism against viral infections. Insects can generate
   piRNAs from the viral genome, and this processing of the viral RNA
   suppresses virion production and replication.[141]^67^,[142]^68 We
   investigated whether airway multiciliated cells—among the first
   infected cells of the upper respiratory tract during influenza—may
   similarly generate viral-derived piRNAs, which could regulate RNA
   transcription and replication.[143]^69 We leveraged the MSRG pipeline
   to align long and small RNA reads to all 8 segments of the
   mouse-adapted PR8 genome. We observed no significant changes in total
   long and small RNA counts among the PR8 samples ([144]Figures 3A and
   3B). This was also true when comparing each segment separately. In
   addition, the coverage plots of long and small RNA reads along the
   viral genome were unchanged in all samples ([145]Figures S4 and
   [146]S5). While there may be small RNAs derived from the PR8 genome,
   their generation is independent of Miwi2 and does not affect overall
   viral RNA transcription at early stages of infection.

Figure 3.

   [147]Figure 3
   [148]Open in a new tab

   Miwi2 does not alter PR8-derived RNAs during early infection

   Total (A) long and (B) small RNA reads that aligned to the PR8 genome.
   Reads were also separately quantified for each of the eight segments of
   the viral genome. Error bars represent mean ± standard deviation.

Miwi2 modulates ribosomal and mitochondrial gene expression during acute
infection

   A third possible avenue of MIWI2’s function in airway multiciliated
   cells could be the regulation of host genes. Principal-component
   analyses of the total RNA shows M2MC cells cluster separately from
   nonM2MC cells, independent of Miwi2 expression. The separation is more
   apparent during PR8 infection ([149]Figure 4A). As anticipated, Stat1,
   IFN-stimulated, and IAV genes were upregulated when comparing PR8 to
   saline treated samples ([150]Figure S6). Differential expression
   analysis shows that M2MC cells contain Miwi2 and tdTomato transcripts,
   and are transcriptomically distinct from nonM2MC cells ([151]Figure 4B)
   as seen with our previous studies.[152]^23 This also holds true when
   examining differentially expressed miRNAs ([153]Figure S7A).

Figure 4.

   [154]Figure 4
   [155]Open in a new tab

   Miwi2 deficiency is associated with a downregulation of mitochondrial
   and ribosomal genes in M2MC cells during early PR8 infection

   (A) Principal-component analyses of all long RNA samples.

   (B) Volcano plots comparing differentially expressed genes between
   nonM2MC^+/tom (blue) to M2MC^+/tom (red) cells under saline and PR8
   conditions.

   (C and D) Volcano plots showing differentially expressed gene between
   M2MC^+/tom (blue) and M2MC^tom/tom (red) under (C) saline and (D) PR8
   conditions. FDR < 0.05.

   (E) Pathway analysis on downregulated genes associated with Miwi2
   deficiency in PR8 infected M2MC cells. Dotted line: FDR < 0.05.

   Strikingly, there were no Miwi2 dependent gene changes in M2MC cells
   under saline conditions besides the Miwi2 transcript itself
   ([156]Figure 4C). During PR8 infection, however, select mitochondrial
   and ribosomal genes were downregulated, which was associated with Miwi2
   deficiency ([157]Figure 4D). No Miwi2-dependent changes were observed
   in nonM2MC cells ([158]Figure S7B). Pathway enrichment analysis of the
   downregulated genes in association with Miwi2 deficiency indicated
   pathways involved in influenza viral RNA transcription, mitochondrial,
   and ribosomal function ([159]Figure 4E; [160]Table S1). Pathway
   enrichment analysis was also performed on the upregulated genes, but no
   significant findings were found.

   When examining host-derived small RNAs, a significant proportion were
   initially found to align with the mitochondrial genome in all
   multiciliated cell subtypes ([161]Figures 5A and [162]S8A). However,
   after increasing the alignment stringency, which ensures more precise
   matching, we could also match small RNAs from NuMT DNA in addition to
   the mitochondrial genome. NuMTs are genetic insertions of mitochondrial
   DNA in the nuclear genome and while their functional relevance remains
   debated, their expression has been linked to specific diseases such as
   cancer.[163]^70^,[164]^71 Despite their poorly characterized role,
   these small RNAs mapped to specific loci of established NuMT regions on
   mouse chromosome 1 and 4 ([165]Figures 5B, 5C, [166]S8B, and S8C). The
   abundance of distinct NuMT-derived small RNA sequences, rather than
   broad expression across the entire NuMT region, suggests an active
   regulatory mechanism rather than random transcription or degradation.
   Notably, M2MC^+/tom cells exhibited significantly lower counts of these
   small RNAs compared to M2MC^tom/tom cells under saline conditions.
   Taken together, these findings suggest that Miwi2 is associated with
   the downregulation of NuMT-derived small RNA expression in
   multiciliated cells.

Figure 5.

   [167]Figure 5
   [168]Open in a new tab

   NUMT-derived small RNAs decrease in Miwi2 expressing cells during
   homeostasis

   Representative coverage plots of small RNAs that aligned to the
   positive (red) and negative (blue) strands of the (A) mitochondrial
   genome and NUMT regions in (B) chromosome 1 and (C) chromosome 4. The
   most represented strands (positive/negative) were measured (right).
   Data are shown as a boxplot with the median (center line),
   interquartile range (IQR, box), and whiskers extending to min/max (or
   1.5× IQR) (Student’s t test). ∗p < 0.05.

Enhanced mitochondrial reactive oxygen species in multiciliated cells of
Miwi2 deficient mice

   Mitochondria undergo significant changes during viral infection such as
   elongation and increased oxidant generation.[169]^72^,[170]^73 Given
   the previous observation that Miwi2 could regulate mitochondrial genes
   during PR8 infection, we examined changes in mitochondrial features
   such as mass and intracellular ROS. PR8 infected Miwi2^+/+,
   Miwi2^+/tom, and Miwi2^tom/tom lungs were analyzed via flow cytometry.
   Using multiciliated cell surface markers, mitochondrial mass was
   assessed using mean fluorescence intensity (MFI) of MitoTracker Green
   staining.[171]^74 No Miwi2 dependent changes in mitochondrial mass were
   observed during PR8 infection in multiciliated cells ([172]Figure 6A).

Figure 6.

   [173]Figure 6
   [174]Open in a new tab

   Miwi2 deficiency increases intracellular ROS and ADP/ATP ratios in
   multiciliated cells but has no effect on mitochondrial mass during PR8
   infection

   (A) Flow cytometry analysis assessing mean fluorescence intensity (MFI)
   of mitochondrial mass dye (MitoTracker Green) in PR8 infected
   multiciliated cells from Miwi2^+/+ (n = 5), Miwi2^+/tom (n = 4), and
   Miwi2^tom/tom (n = 3) mice. As a negative control, Miwi2^+/+ cells were
   treated with oxidative phosphorylation uncouple, FCCP, that dissipates
   the mitochondrial membrane.

   (B) MFI of ROS dye (DHR-123) was also assessed in infected
   multiciliated cells from Miwi2^+/+ (n = 6), Miwi2^+/tom (n = 5),
   Miwi2^tom/tom (n = 4) mice. Miwi2^+/+ cells (n = 4) were treated with
   mitochondrial complex I inhibitor, rotenone, and Miwi2^tom/tom cells
   (n = 3) were treated with FCCP as a positive and negative ROS control,
   respectively (one-way ANOVA followed by Sidak’s test).

   (C) ADP/ATP ratios of FACS isolated PR8 infected multiciliated cells
   from Miwi2^+/+ (n = 3) and Miwi2^tom/tom (n = 3) mice (Student’s t
   test). Error bars represent mean ± standard deviation. ∗p < 0.05,
   ∗∗p < 0.005, ∗∗∗p < 0.0005, ∗∗∗∗p < 0.0001.

   Cells were separately stained with DHR-123, which targets mitochondrial
   ROS.[175]^75^,[176]^76 Notably, DHR-123 MFI was significantly higher in
   multiciliated cells of Miwi2^tom/tom mice compared to Miwi2^+/+ mice
   ([177]Figure 6B). As a positive control for ROS generation,
   multiciliated cells from Miwi2^+/+ mice were treated with mitochondrial
   complex I inhibitor rotenone that increased intracellular ROS.[178]^77
   Conversely, in multiciliated cells from Miwi2^tom/tom mice, treatment
   with oxidative phosphorylation uncoupler FCCP significantly decreased
   intracellular ROS.[179]^78 Considering the association of mitochondrial
   oxidant generation with ATP production, we assessed ADP/ATP ratios in
   isolated PR8 infected multiciliated cells. Multiciliated cells of
   Miwi2^tom/tom mice exhibited a significant increase in ADP/ATP ratios
   compared to multiciliated cells of Miwi2^+/+ mice ([180]Figure 6C).
   These findings suggest that while Miwi2 does not affect mitochondrial
   mass, it may alter mitochondrial energy dynamics and oxidant generation
   that can drive intracellular ROS during PR8 infection. Given the
   critical role of ROS production in the innate immune response, this
   alteration may influence the host’s ability to combat viral infections.

Miwi2 as a host susceptibility factor for influenza burden and pathogenesis

   We next examined if Miwi2 influences overall outcomes of influenza
   infection. Miwi2^+/+, Miwi2^+/tom, and Miwi2^tom/tom mice were infected
   at 3, 7, and 14 dpi. No significant changes in viral titers or RNA were
   observed at 3 dpi ([181]Figure 7A). However, at 7 dpi, Miwi2^tom/tom
   lungs contained lower viral titers and RNA compared to Miwi2^+/tom
   lungs ([182]Figure 7B). While viral titers are undetected at 14 dpi,
   viral RNA was still significantly lower in Miwi2^tom/tom and
   Miwi2^+/tom lungs compared to Miwi2^+/+ ([183]Figure 7C). When
   assessing weight, Miwi2^tom/tom mice recovered faster in bodyweight
   compared to Miwi2^+/+ mice ([184]Figure 7D). Although Miwi2 deficiency
   mainly causes infertility in male mice, the observed differences in
   viral load were independent of sex.

Figure 7.

   [185]Figure 7
   [186]Open in a new tab

   Miwi2^tom/tom mice experience decreased viral burden and weight loss

   (A) Quantification of plaque titers and relative expression of PR8
   nucleoprotein (NP) RNA in Miwi2^+/+ (n = 3), Miwi2^+/tom (n = 6),
   Miwi2^tom/tom (n = 3) lungs at 3 dpi.

   (B) The same analysis was performed on Miwi2^+/tom (n = 6),
   Miwi2^tom/tom (n = 7) lungs at 7 dpi (Mann-Whitney U test). Dotted line
   represents initial administered PFU dose.

   (C) NP RNA was still detected 14 dpi in Miwi2^+/+ (n = 3), Miwi2^+/tom
   (n = 6), Miwi2^tom/tom (n = 5) lungs (one-way ANOVA followed by
   Dunnett’s test).

   (D) Weight curves of PR8 infected mice measured every day until 14 dpi
   (mixed-effects analysis). Error bars represent mean ± standard
   deviation ∗p < 0.05, ∗∗p < 0.005.

   It is possible that these differences in viral load could be attributed
   to differences in inflammation and immune cell recruitment in response
   to infection. To address this, we employed an immune cell flow
   cytometry panel for assessing immune cell recruitment in Miwi2^+/+,
   Miwi2^+/tom, and Miwi2^tom/tom mice at 3 dpi. No differences in CD8
   T cell subsets or other immune cell types in recruitment were observed
   ([187]Figure S9). Histopathological evaluations of all lung lobes from
   PR8 infected lungs revealed no significant morphological differences
   between Miwi2^tom/tom and Miwi2^+/+ mice either at 3 dpi or at 7 dpi
   ([188]Figure S10). While Miwi2^tom/tom mice experience improved
   outcomes to IAV infection, these findings suggest that it is not due to
   alterations in immune cell recruitment and histopathological patterns.

Discussion

   MIWI2, the mouse ortholog of human PIWIL4, is an argonaute family
   protein highly expressed in the mammalian germline, where it plays a
   critical role in maintaining genomic integrity. Although somatic
   expression of Miwi2 has been observed, its role in mammalian somatic
   cells has remained largely unexplored.[189]^79^,[190]^80^,[191]^81 In
   this study, we uncover a previously unrecognized role for Miwi2 in the
   host defense against IAV infection, specifically in multiciliated cells
   of the respiratory tract. Unlike the canonical role of Piwi proteins
   regulating retrotransposons in the male germline and viral RNA in
   insects, Miwi2 appears to modulate select mitochondrial genes during
   viral infection through regulation of NuMT-derived small RNAs.
   Additionally, Miwi2 deficient multiciliated cells exhibited increased
   intracellular ROS levels and ADP/ATP ratios during IAV infection, and
   Miwi2 deficient mice show reduced morbidity and viral load. These
   findings suggest that Miwi2 plays a more dynamic and context-dependent
   role in somatic cells, particularly during stress conditions such as
   viral infections, where it may influence mitochondrial function and
   host-pathogen interactions.

   Multiciliated cells are highly metabolically active, largely due to the
   energy demands of ciliary beating.[192]^82 These cells often undergo
   mitochondrial adaptations in response to stress, such as viral
   infections, which include mitochondrial elongation and changes in ROS
   production.[193]^72^,[194]^73^,[195]^83 Our study demonstrated
   increased intracellular ROS in multiciliated cells from Miwi2 deficient
   mice during early IAV infection, while mitochondrial mass remained
   unchanged. The precise mechanism through which Miwi2 or the M2MC
   population influences mitochondrial energy dynamics and increased ROS
   in all multiciliated cells remains unclear. However, Miwi2 deficiency
   appears to contribute to more efficient viral clearance, as evidenced
   by reduced viral titers and RNA at later stages of infection.

   We observed an increase in ADP/ATP ratios in multiciliated cells of
   Miwi2 deficient mice. This could be a result of slower ADP-to-ATP
   conversion or rapid ATP consumption. If the former, Miwi2 deficiency
   may impair mitochondrial function, as genes encoding complexes I, III,
   and V were downregulated during IAV infection ([196]Table S1).
   Inhibiting these complexes, such as with rotenone, disrupts electron
   transport, increasing ROS production while reducing
   ATP.[197]^84^,[198]^85^,[199]^86 Rotenone treatment has been shown to
   normalize oxygenation and airway resistance and reduce IAV-induced
   pulmonary edema.[200]^87 Alternatively, if ATP consumption is driving
   the increased ADP/ATP ratios, Miwi2 may promote mitochondrial
   hyperactivity during infection, further elevating ROS.

   ROS generation, which is closely associated with mitochondrial function
   and reactive nitrogen species (RNS) generation, plays a complex role in
   multiciliated cells. While excess oxidant generation can lead to
   oxidative stress and tissue damage, moderate RNS levels, such as nitric
   oxide, are essential for maintaining planar cell polarity and ciliary
   beat frequency, both of which are necessary for effective airway
   clearance.[201]^82^,[202]^88^,[203]^89^,[204]^90^,[205]^91 Influenza
   infection impairs ciliary beating and damages cilia, compromising
   mucociliary clearance and thus facilitating viral entry and
   propagation.[206]^92^,[207]^93 Therefore, it is possible that Miwi2
   deficiency enhances the resilience or effectiveness of ciliary function
   during infection, potentially contributing to the improved antiviral
   host defense.

   Increased intracellular ROS can promote cellular apoptosis, a
   mitochondrial-mediated cell death pathway crucial for regulating viral
   infections. Previous studies have demonstrated that endoplasmic
   reticulum stress induces piRNA production and that PIWIL4 expression
   drives apoptosis in human airway epithelial cells.[208]^94 In our
   study, we observed a decrease in M2MC cells in Miwi2 heterozygous mice
   compared to deficient mice during PR8 infection ([209]Figure 1E),
   suggesting that Miwi2 may trigger apoptosis in infected cells
   potentially contributing to further tissue injury. Moreover,
   mitochondrial ROS can activate the innate immune response to
   infection.[210]^95^,[211]^96 Although we did not observe differences in
   immune cell recruitment during the early stages of infection,
   alterations in ROS levels could influence the functional efficiency of
   immune cells in clearing the virus. Future studies are needed to
   determine the specific mechanisms by which Miwi2 influences viral
   clearance.

   The evolutionary significance of maintaining a subset of multiciliated
   cells expressing Miwi2, may stem from its potential role in regulating
   mitochondrial function, mitigating excessive oxidative damage, and
   regulating inflammatory responses. While Miwi2 deficiency results in
   increased ROS levels, which could enhance antiviral defenses, sustained
   ROS elevation poses risks of prolonged oxidate damage. This could
   exacerbate inflammatory responses and compromise tissue repair,
   especially in chronic infection or inflammatory conditions.[212]^97 By
   tempering ROS levels in some contexts, Miwi2 may serve to balance
   between effective pathogen clearance and the prevention of excessive
   oxidative damage, highlighting its potential role in preserving
   long-term lung homeostasis and function.

   MIWI2 has been extensively studied for its role in silencing
   retrotransposons in the mammalian male germline. Previous
   RNA-sequencing studies have suggested viral infections may re-activate
   transposable elements, and promote immune defenses by regulating
   expression of nearby antiviral genes or can act as immunostimulatory
   self-derived viral mimetics to trigger defenses.[213]^64^,[214]^98 Our
   findings indicated that IAV infection induces expression of several
   retrotransposons subfamilies, specifically ERVs, in airway
   multiciliated cells. ERVs have genetic structures similar to exogenous
   retroviruses and, upon activation, generate double-stranded RNA species
   than can be detected by cytosolic RNA sensors such as RIG-I, MDA5, or
   toll-like receptors.[215]^99 The reactivation of these ERV subfamilies
   could be a method of strengthening the antiviral response to IAV
   infection. Despite these findings, their expression does not appear to
   be directly associated with Miwi2 expression. However, more RT
   subfamilies are expressed in multiciliated cells of Miwi2 deficient
   mice, including nonM2MC cells. Therefore, it is plausible that Miwi2
   alters the cellular state of multiciliated cells that may facilitate
   increased RT activation during influenza infection through an unknown
   mechanism.

   Viral-derived small RNAs, such as miRNAs from IAV, have been documented
   to regulate viral RNA synthesis and act as molecular switches from
   viral transcription to replication.[216]^100 Similarly, in SARS-CoV
   models, small RNAs derived from the nsp3 and N genomic regions
   contribute to viral pathogenesis.[217]^101 Our data are thus consistent
   with previous studies where we can detect significant amounts of
   viral-derived small RNAs in influenza targeted multiciliated cells.
   Although the production of these small RNAs is independent of Miwi2
   expression, their role in viral replication and pathogenesis remains to
   be determined and may represent another layer of host-pathogen
   interaction in infected multiciliated cells of the airways.

   NuMT-derived small RNAs were also identified in multiciliated cells.
   NuMTs are fragments of mitochondrial DNA that have integrated into the
   nuclear genome throughout evolutionary history.[218]^70^,[219]^102 Some
   studies suggest they are ancient evolutionary insertion events with no
   observed function, while others have associated them with genomic
   instability and cancer.[220]^103^,[221]^104^,[222]^105 Although the
   role of NuMTs in gene regulation is unclear, the selective regulation
   of NuMT RNAs rather than random degradation, hints at an organized,
   potentially piRNA-like mechanism in somatic cells that targets either
   these mitochondrial DNA insertions or their transcripts for
   post-transcriptional repression.

   Our observations revealed a decrease in the NuMT-derived small RNAs in
   Miwi2 expressing cells at baseline, indicating that Miwi2 may regulate
   their biogenesis or turnover during homeostasis. This gene-dependent
   regulation of NuMT-derived small RNAs has not been reported before,
   highlighting a novel aspect of Miwi2’s role in somatic cells.
   Interestingly, although we did not observe Miwi2-dependent changes in
   mitochondrial mRNAs in saline treated M2MC cells, it is possible that
   the functional significance of these small RNAs only becomes evident
   during cellular stress such as during viral infection, when the
   regulation of mitochondrial function is more critical.[223]^106 One
   particular observation is that primary piRNA biogenesis occurs near the
   mitochondria, which are localized proximal to the basal body of
   multiciliated cells. This is coincidentally where MIWI2 is also
   localized.[224]^23^,[225]^33^,[226]^35 Given the repressive nature of
   piRNAs, it is conceivable that NuMT transcripts are directed to the
   mitochondria, where they are processed by MIWI2 to generate piRNAs.
   This spatial proximity may facilitate the regulation of
   mitochondrial-nuclear communication, potentially influencing both
   mitochondrial function and the broader stress response pathways in
   multiciliated cells. Moreover, if these small RNAs are indeed involved
   in stress responses, particularly in regulating mitochondrial activity,
   they could play a critical role in the host antiviral defense
   mechanisms. Further investigation is needed to elucidate the functional
   impact of the NuMT-derived small RNAs and to interrogate their novel
   gene regulatory properties influencing both mitochondrial function and
   host defense pathways.

   In conclusion, our findings suggest that Miwi2/Piwil4 acts as a
   potential host susceptibility factor for severe respiratory infections.
   These studies highlight a novel function of somatic Miwi2 in the lung,
   relevant during IAV infection. By regulating mitochondrial gene
   expression and function, Miwi2 may exacerbate influenza disease through
   its impact on intracellular ROS levels and ATP production in
   multiciliated cells. Further elucidation of this unique somatic
   Miwi2-dependent pathway could reveal potential therapeutic targets
   designed to mitigate the severity of viral respiratory infections and
   enhancing host defense.

Limitations of the study

   A limitation of this study is its focus on an early stage of infection,
   which may not capture the full dynamics of Miwi2’s role over the
   complete course of infection. Moreover, the specific molecular pathways
   linking Miwi2 to mitochondrial function, ROS production, and immune
   response modulation remain to be elucidated. Finally, our
   Miwi2-deficient model was a systemic knockout, so while the changes in
   influenza disease burden are likely attributable to the M2MC cells, the
   possibility that Miwi2 expression in other cells and tissues
   contributes to IAV pathogenesis cannot be excluded. Despite these
   limitations, our findings yield novel insights into the role of Miwi2
   and small RNAs in multiciliated cells during influenza infection.

Resource availability

Lead contact

   Further information and requests for resources should be directed to
   and will be fulfilled by the lead contact, Jhonatan Henao Vasquez
   (jhenao@bu.edu).

Materials availability

   This study did not generate new materials.

Data and code availability

   Bulk small and long RNA sequencing files have been deposited in the
   NCBI Gene Expression Omnibus (GEO) under accession number GEO:
   [227]GSE276578
   . Any additional information required to reanalyze the data reported in
   this paper is available from the [228]lead contact upon request. No
   original code or software was used in this study.

Acknowledgments